text
stringlengths 7
107k
|
|---|
Note that, unlike Cas9, Cas12a cleaves distally
490
from its PAM and seed regions. The preferred PAM recognition sequence of commonly studied
491
Cas12a orthologs is TTTV. However, as shown, JH4, the most frequently used JH gene in all
492
species, contains optimally located GTTC and TTCC PAM sequences, located 3’ of the HCDR3-
493
encoding sequence but oriented Cas12a cleavage within the this sequence. This PAM, sequence
494
of the gRNA, and the Cas12a cut sites are indicated. (B) To identify a Cas12a ortholog efficient
495
at cleaving these non-canonical PAM motifs, the human B-cell line Jeko-1 was transfected with
496
plasmids encoding BsCas12a, TsCas12a, Mb2Cas12a, or Mb3Cas12a. Targeting efficiency was
497
measured by flow cytometry as loss of IgM expression. Among these Cas12a orthologs,
498
Mb2Cas12 most efficiently cleaved the J-chain region initiated with GTTC and TTCC (orange). 499
Error bars indicted range of two independent experiments, and asterisks indicate statistical
500
significance relative to controls. Statistical difference were determined by non-paired Students t-
501
test, (****, p<0.0001). (C) Mb2Cas12 RNP were compared with commercial AsCas12a RNP for
502
their ability cleave four distinct regions in the HCDR3-encoding region of Jeko-1 cells. Loss of
503
IgM expression indicates successful introduction of a double-strand break and inexact NHEJ. (D)
504
Results of three experiments similar to that shown in panel C. Error bars indicate standard error
23
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 505
(SEM). Asterisks indicted significant differences from the canonical TTTG PAM (Mb2Cas12a-
506
RNP or AsCas12a, respectively). Statistical difference were determined by non-paired Students
507
t-test, (****, p<0.0001). 508
Figure 2. Optimization of ssDNA templates for Mb2Cas12a-mediated editing the HCDR3-
509
encoding region of a human B cell line. (A) A diagram representing four HDR templates
510
(HDRT) used in panels B-E. Specifically, sense and anti-sense forms of HDRT-A, were used to
511
replace a 9-nucleotide (nt) region (grey) with 39-nt insert (green), and both forms of HDRT-B
512
were used to replace a 36-nt region with a 69-nt region. 50-nt homology arms of the sense and
513
antisense forms are represented in red and blue, respectively. SpCas9 (cyan) and Mb2Cas12
514
(orange) cleavage sites of the target strand (complementary to grNA) are indicated by arrows. 515
Note that paired Cas9 and Cas12a cleavage sites are separated by at most five nucleotides. (B) A
516
representative example of an experiment used to generate panels C-E in which editing efficiency
517
of MbCas12A or SpCas9 RNP is monitored through recognition of an HA tag introduced into the
518
HCDR3 of the Jeko-1 cell BCR by flow cytometry.
|
Control cells were electroporated with
519
Mb2Cas12a RNP without an HDRT. (C) A comparison of Mb2Cas12a (Mb2) and SpCas9
520
(Cas9) knock-in efficiencies, measured as described in panel B, for all four sites shown in panel
521
A. Differences between Mb2 and Cas9, and among the four sites, are not significant (n.s.). The
522
same data generated for panel C was replotted according to whether the sense or anti-sense
523
HDRT were used (D), or whether the HDRT complemented the gRNA target or non-target
524
strand. (E) Non-target strand is the PAM containing strand, and the target strand is the strand
525
annealed to gRNA. Again, as indicated, most differences were not significant. However, the
526
HDRT complementary to the Mb2Cas12a gRNA target strand were slightly more efficient than
527
those complementary to the non-target strand (p=0.027). Dots in (C)-(E) represent pooled data
24
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 528
from two independent experiments. Statistical significance was calculated by one-way ANOVA
529
with Tukey’s multiple comparison test. 530
Figure 3. The length of the 3’ mismatch tail determines replacement efficiency with short
531
single-stranded HDRT. (A) A model showing where a 3’ mismatch tail occurs. A cut site
532
(yellow) is introduced into a region of the gene targeted for replacement (grey), asymmetrically
533
dividing this region. Efficient 5’ to 3’ resection exposes two 3’ ends. An HDRT can complement
534
a strand with a short (left figures) or long 3’-mismatch tail (right figures), which must be
535
removed before the remaining 3’ end can be extended to complement the HDRT insert region
536
and its distal homology arm. We propose that the removal of this 3’ mismatch tail is a rate-
537
limiting step determining editing efficiency when genomic sequences are replaced. (B) The
538
predicted length of the 3’-mismatch tail in experiments presented in Figure 2 are plotted against
539
the efficiency with which an HA-tag is introduced into the HCDR3 region, as determined by
540
flow cytometry. Error bar indication SD from two independent experiments. (C) A comparison
541
of editing efficiency between those with short (<10 nt) or long (>10 nt) 3’ mismatch tails. 542
Editing by SpCas9 or Mb2Cas12a is significantly more efficient with short 3’ mismatch tails, as
543
determined by one-way ANOVA with Tukey’s multiple comparison test (p<0.0001). Dots
544
represent pool data from two independent experiments. 545
Figure 4. The BCR specificity of Jeko-1 cells can be reprogrammed with a novel HCDR3. 546
(A) The amino-acid sequence of the native Jeko-1 cell HCDR3 region and those of the HIV-1
547
neutralizing antibodies PG9 and PG16 are shown.
|
In addition the amino-acid translations of four
548
HDRT used in the subsequent panels are represented in green, in the context of the remaining
549
Jeko-1 region. (B) Mb2Cas12a RNP targeting the GTTC PAM of Site 4 in Jeko-1 cells shown in
550
Figure 2B were co-electroporated with the indicated HDRT. Editing efficiency was monitored on
25
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 551
the vertical axis by flow cytometry with fluorescently labeled PSG2, an antibody that recognizes
552
sulfotyrosines within the PG9 and PG16 HCDR3 region, a similarly labeled HIV SOSIP or E2p. 553
The horizontal axis indicates IgM expression, and its loss indicates imprecise NHEJ after
554
Mb2Cas12a-mediated cleavage. Note that introduction of a PG16 HCDR3 was efficient, as
555
indicated by PSG2 recognition, but unlike the PG9 HCDR3, it did not bind the Env trimer. Cells
556
edited to express an HA tag did not bind any reagent. SOSIP proteins were derived from the
557
BG505 HIV-1 isolate. (C) A summary of three independent experiments similar to that shown in
558
panel B. flow cytometric studies used to generate panel B. Error bars indicate SD. (D) Jeko-1
559
edited with PG9-CAR HDRT were enriched by FACS with the anti-sulfotyrosine antibody
560
PSG2. (E) Cells enriched in panel D were analyzed two weeks later by flow cytometry for their
561
ability to bind PSG2, a BG505-based nanoparticle (BG505-E2p), SOSIP trimers derived from
562
the indicated HIV-1 isolate, or an V2 apex negative mutant (dBG505-SOSIP). Grey control
563
indicates wild-type Jeko-1 cells. (F) Unedited Jeko-1 cells and those edited with PG9-CAR
564
HDRT without sorting, or sorted with PSG2 or with E2p, were analyzed by next-generation
565
sequencing (NGS) of the VDJ region. Sequences were divided into four categories, depending on
566
whether the edited sequence matched exactly the HDRT (Perfect HDR), whether HDRT
567
sequence was visible but modified (Imperfect HDR), whether the original Jeko-1 HCDR3 region
568
was intact (Original), or whether this region was modified by NHEJ as indicated by the presence
569
of insertions or deletions (Indel). Representative examples of each category are shown below the
570
charts. 571
Figure 5 Editing primary human B-cells with HDRT recognizing consensus sequences of
572
multiple VH families. (A) A panel of PG9-CAR HDRT with homology arms complementary to
573
JH4 and to consensus VH1-,VH3-, and VH4-family sequences were evaluated for their ability to
26
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
|
It is made
available under a
CC-BY-NC-ND 4.0 International license . 574
edited primary human B cells. Cells electroporated with Mb2Cas12a RNP and PG9-CAR HDRT
575
were analyzed by flow cytometry with the anti-sulfotyrosine antibody PSG2 modified with two
576
distinct fluorophores to eliminate non-specific binding from either fluorophore, (B) A summary
577
of results from experiments similar to that shown in panel A, using primary B cells from three
578
independent donors. Note that a mixture of three HDRT edited more cells than any individual
579
HDRT. Null indicates that cells were not electroporated and control indicates cells electroporated
580
with Mb2Cas12a RNP and an HDRT that is not homologous to any sequence in the human
581
genome. Mix indicates cells electroporated with RNP and an equimolar mixture of HDRT with
582
VH1-, VH3- and VH4-specific homology arms. Error bars indicted range of three independent
583
experiments, and asterisks indicate statistical significance calculated by one-way ANOVA with
584
Tukey’s multiple comparison test (*, p<0.5; **, p<0.01; ****, p<0.0001). (C) NGS analysis of
585
primary B cells from two human donors, quantified as described in Figure 4F except that the
586
VH-family of edited cells was also counted. 587
Figure 6. Reprogrammed primary human B cells retain V-gene diversity. Primary cells were
588
electroporated with Mb2Cas12a RNP and HDRT encoding an HA tag (A) or the HCDR3 regions
589
of the HIV-1 neutralizing antibodies CH01 (B) and PG9 (C), with the same mixture of homology
590
arms as those used in Figure 5. Cells were sorted with an anti-HA antibody (HA tag, panel A) or
591
a SOSIP trimer derived from the CRF_AG_250 isolate (panels B and C). Edited cells were
592
analyzed by NGS before and after sorting, and the frequency of each VH1-,VH3-, and VH4-
593
family gene was measured. Flow cytometry histograms displays one of two experiments with
594
similar results, and bar graphs indicate the mean of those two experiments. (D) Antibodies
595
composed the heavy-chains expressed from the indicated VH genes enriched in panel C or that
596
of PG9, the PG9 HCDR3, a transmembrane domain, and the native PG9 light chain were
27
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 597
expressed on the surface of 293T cells and analyzed by flow cytometry. One of two
598
representative experiments is presented. Mature indicates expression of the original PG9
599
antibody. (E) The mean of two experiments shown in panel D is presented. (F) The IC50 values
600
of soluble forms of the antibodies characterized in panel D against indicated HIV-1 isolates is
601
represented. 602
603
Reference:
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632
1.
|
West Jr AP, Scharf L, Scheid JF, Klein F, Bjorkman PJ, Nussenzweig MC. Structural insights on the role of antibodies in HIV-1 vaccine and therapy. Cell. 2014;156(4):633-48. 2. Mascola JR, Haynes BF. HIV‐1 neutralizing antibodies: understanding nature's pathways. Immunological reviews. 2013;254(1):225-44. 3. Burton DR, Hangartner L. Broadly neutralizing antibodies to HIV and their role in vaccine design. Annual review of immunology. 2016;34:635-59. 4. Escolano A, Dosenovic P, Nussenzweig MC. Progress toward active or passive HIV-1 vaccination. Journal of Experimental Medicine. 2017;214(1):3-16. 5. Jardine JG, Kulp DW, Havenar-Daughton C, Sarkar A, Briney B, Sok D, Sesterhenn F, Ereño- Orbea J, Kalyuzhniy O, Deresa I. HIV-1 broadly neutralizing antibody precursor B cells revealed by germline-targeting immunogen. Science. 2016;351(6280):1458-63. 6. Steichen JM, Lin Y-C, Havenar-Daughton C, Pecetta S, Ozorowski G, Willis JR, Toy L, Sok D, Liguori A, Kratochvil S. A generalized HIV vaccine design strategy for priming of broadly neutralizing antibody responses. Science. 2019;366(6470). 7. Klein F, Diskin R, Scheid JF, Gaebler C, Mouquet H, Georgiev IS, Pancera M, Zhou T, Incesu R-B, Fu BZ. Somatic mutations of the immunoglobulin framework are generally required for broad and potent HIV-1 neutralization. Cell. 2013;153(1):126-38. 8. Walker LM, Huber M, Doores KJ, Falkowska E, Pejchal R, Julien J-P, Wang S-K, Ramos A, Chan-Hui P-Y, Moyle M. Broad neutralization coverage of HIV by multiple highly potent antibodies. Nature. 2011;477(7365):466-70. 9. Huang D, Tran JT, Olson A, Vollbrecht T, Tenuta M, Guryleva MV, Fuller RP, Schiffner T, Abadejos JR, Couvrette L. Vaccine elicitation of HIV broadly neutralizing antibodies from engineered B cells. Nature communications. 2020;11(1):1-10. 10. Voss JE, Gonzalez-Martin A, Andrabi R, Fuller RP, Murrell B, McCoy LE, Porter K, Huang D, Li W, Sok D. Reprogramming the antigen specificity of B cells using genome-editing technologies. Elife. 2019;8:e42995. 11. Hartweger H, McGuire AT, Horning M, Taylor JJ, Dosenovic P, Yost D, Gazumyan A, Seaman MS, Stamatatos L, Jankovic M. HIV-specific humoral immune responses by CRISPR/Cas9-edited B cells. Journal of Experimental Medicine. 2019;216(6):1301-10. 28
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676
12. Moffett HF, Harms CK, Fitzpatrick KS, Tooley MR, Boonyaratanakornkit J, Taylor JJ. B cells engineered to express pathogen-specific antibodies protect against infection. Science immunology. 2019;4(35). 13. Nahmad AD, Raviv Y, Horovitz-Fried M, Sofer I, Akriv T, Nataf D, Dotan I, Carmi Y, Burstein D, Wine Y.
|
Engineered B cells expressing an anti-HIV antibody enable memory retention, isotype switching and clonal expansion. Nature communications. 2020;11(1):1-10. 14. Giudicelli V, Duroux P, Ginestoux C, Folch G, Jabado-Michaloud J, Chaume D, Lefranc M- P. IMGT/LIGM-DB, the IMGT® comprehensive database of immunoglobulin and T cell receptor nucleotide sequences. Nucleic acids research. 2006;34(suppl_1):D781-D4. 15. Rolink A, Melchers F. Molecular and cellular origins of B lymphocyte diversity. Cell. 1991;66(6):1081-94. 16. Elhanati Y, Sethna Z, Marcou Q, Callan Jr CG, Mora T, Walczak AM. Inferring processes underlying B-cell repertoire diversity. Philosophical Transactions of the Royal Society B: Biological Sciences. 2015;370(1676):20140243. 17. Schatz DG, Ji Y. Recombination centres and the orchestration of V (D) J recombination. Nature Reviews Immunology. 2011;11(4):251-63. 18. Bassing CH, Swat W, Alt FW. The mechanism and regulation of chromosomal V (D) J recombination. Cell. 2002;109(2):S45-S55. 19. Mesin L, Ersching J, Victora GD. Germinal center B cell dynamics. Immunity. 2016;45(3):471-82. 20. Leslie A, Pfafferott K, Chetty P, Draenert R, Addo M, Feeney M, Tang Y, Holmes E, Allen T, Prado J. HIV evolution: CTL escape mutation and reversion after transmission. Nature medicine. 2004;10(3):282-9. 21. Abbott RK, Lee JH, Menis S, Skog P, Rossi M, Ota T, Kulp DW, Bhullar D, Kalyuzhniy O, Havenar-Daughton C. Precursor frequency and affinity determine B cell competitive fitness in germinal centers, tested with germline-targeting HIV vaccine immunogens. Immunity. 2018;48(1):133-46. e6. 22. Dosenovic P, Kara EE, Pettersson A-K, McGuire AT, Gray M, Hartweger H, Thientosapol ES, Stamatatos L, Nussenzweig MC. Anti–HIV-1 B cell responses are dependent on B cell precursor frequency and antigen-binding affinity. Proceedings of the National Academy of Sciences. 2018;115(18):4743-8. 23. Andrabi R, Voss JE, Liang C-H, Briney B, McCoy LE, Wu C-Y, Wong C-H, Poignard P, Burton DR. Identification of common features in prototype broadly neutralizing antibodies to HIV envelope V2 apex to facilitate vaccine design. Immunity. 2015;43(5):959-73. 24. Lee JH, Andrabi R, Su C-Y, Yasmeen A, Julien J-P, Kong L, Wu NC, McBride R, Sok D, Pauthner M. A broadly neutralizing antibody targets the dynamic HIV envelope trimer apex via a long, rigidified, and anionic β-hairpin structure. Immunity. 2017;46(4):690-702. 25. Zetsche B, Strecker J, Abudayyeh OO, Gootenberg JS, Scott DA, Zhang F. A Survey of Genome Editing Activity for 16 Cpf1 orthologs. bioRxiv. 2017:134015. 26. Doudna JA, Charpentier E. The new frontier of genome engineering with CRISPR-Cas9. Science. 2014;346(6213). 27. Wang Y, Liu KI, Sutrisnoh N-AB, Srinivasan H, Zhang J, Li J, Zhang F, Lalith CRJ, Xing H, Shanmugam R. Systematic evaluation of CRISPR-Cas systems reveals design principles for genome editing in human cells. Genome biology. 2018;19(1):1-16. 29
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021.
|
The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722
28. Liang X, Potter J, Kumar S, Ravinder N, Chesnut JD. Enhanced CRISPR/Cas9-mediated precise genome editing by improved design and delivery of gRNA, Cas9 nuclease, and donor DNA. Journal of biotechnology. 2017;241:136-46. 29. Briney B, Inderbitzin A, Joyce C, Burton DR. Commonality despite exceptional diversity in the baseline human antibody repertoire. Nature. 2019;566(7744):393-7. 30. Volpe JM, Kepler TB. Large-scale analysis of human heavy chain V (D) J recombination patterns. Immunome research. 2008;4(1):1-10. 31. Zetsche B, Gootenberg JS, Abudayyeh OO, Slaymaker IM, Makarova KS, Essletzbichler P, Volz SE, Joung J, Van Der Oost J, Regev A. Cpf1 is a single RNA-guided endonuclease of a class 2 CRISPR-Cas system. Cell. 2015;163(3):759-71. 32. Jinek M, Chylinski K, Fonfara I, Hauer M, Doudna JA, Charpentier E. A programmable dual- RNA–guided DNA endonuclease in adaptive bacterial immunity. science. 2012;337(6096):816- 21. 33. Yamano T, Zetsche B, Ishitani R, Zhang F, Nishimasu H, Nureki O. Structural basis for the canonical and non-canonical PAM recognition by CRISPR-Cpf1. Molecular cell. 2017;67(4):633- 45. e3. 34. Shalem O, Sanjana NE, Hartenian E, Shi X, Scott DA, Mikkelsen TS, Heckl D, Ebert BL, Root DE, Doench JG. Genome-scale CRISPR-Cas9 knockout screening in human cells. Science. 2014;343(6166):84-7. 35. Hur JK, Kim K, Been KW, Baek G, Ye S, Hur JW, Ryu S-M, Lee YS, Kim J-S. Targeted mutagenesis in mice by electroporation of Cpf1 ribonucleoproteins. Nature biotechnology. 2016;34(8):807-8. 36. Paix A, Folkmann A, Goldman DH, Kulaga H, Grzelak MJ, Rasoloson D, Paidemarry S, Green R, Reed RR, Seydoux G. Precision genome editing using synthesis-dependent repair of Cas9- induced DNA breaks. Proceedings of the National Academy of Sciences. 2017;114(50):E10745- E54. 37. Richardson CD, Ray GJ, DeWitt MA, Curie GL, Corn JE. Enhancing homology-directed genome editing by catalytically active and inactive CRISPR-Cas9 using asymmetric donor DNA. Nature biotechnology. 2016;34(3):339-44. 38. Walker LM, Phogat SK, Chan-Hui P-Y, Wagner D, Phung P, Goss JL, Wrin T, Simek MD, Fling S, Mitcham JL. Broad and potent neutralizing antibodies from an African donor reveal a new HIV-1 vaccine target. Science. 2009;326(5950):285-9. 39. Sok D, van Gils MJ, Pauthner M, Julien J-P, Saye-Francisco KL, Hsueh J, Briney B, Lee JH, Le KM, Lee PS. Recombinant HIV envelope trimer selects for quaternary-dependent antibodies targeting the trimer apex. Proceedings of the National Academy of Sciences. 2014;111(49):17624- 9. 40. Sanders RW, Derking R, Cupo A, Julien J-P, Yasmeen A, de Val N, Kim HJ, Blattner C, de la Peña AT, Korzun J.
|
A next-generation cleaved, soluble HIV-1 Env trimer, BG505 SOSIP. 664 gp140, expresses multiple epitopes for broadly neutralizing but not non-neutralizing antibodies. PLoS pathog. 2013;9(9):e1003618. 41. de la Peña AT, Julien J-P, de Taeye SW, Garces F, Guttman M, Ozorowski G, Pritchard LK, Behrens A-J, Go EP, Burger JA. Improving the immunogenicity of native-like HIV-1 envelope trimers by hyperstabilization. Cell reports. 2017;20(8):1805-17. 42. Hoffhines AJ, Damoc E, Bridges KG, Leary JA, Moore KL. Detection and purification of tyrosine-sulfated proteins using a novel anti-sulfotyrosine monoclonal antibody. Journal of Biological Chemistry. 2006;281(49):37877-87. 30
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749
43. He L, De Val N, Morris CD, Vora N, Thinnes TC, Kong L, Azadnia P, Sok D, Zhou B, Burton DR. Presenting native-like trimeric HIV-1 antigens with self-assembling nanoparticles. Nature communications. 2016;7(1):1-15. 44. Briney B, Sok D, Jardine JG, Kulp DW, Skog P, Menis S, Jacak R, Kalyuzhniy O, De Val N, Sesterhenn F. Tailored immunogens direct affinity maturation toward HIV neutralizing antibodies. Cell. 2016;166(6):1459-70. e11. 45. Escolano A, Steichen JM, Dosenovic P, Kulp DW, Golijanin J, Sok D, Freund NT, Gitlin AD, Oliveira T, Araki T. Sequential immunization elicits broadly neutralizing anti-HIV-1 antibodies in Ig knockin mice. Cell. 2016;166(6):1445-58. e12. 46. Tian M, Cheng C, Chen X, Duan H, Cheng H-L, Dao M, Sheng Z, Kimble M, Wang L, Lin S. Induction of HIV neutralizing antibody lineages in mice with diverse precursor repertoires. Cell. 2016;166(6):1471-84. e18. 47. Pâques F, Haber JE. Two pathways for removal of nonhomologous DNA ends during double- strand break in Saccharomyces cerevisiae. Molecular and Cellular Biology. repair 1997;17(11):6765-71. 48. Anand R, Beach A, Li K, Haber J. Rad51-mediated double-strand break repair and mismatch correction of divergent substrates. Nature. 2017;544(7650):377-80. 49. Masella AP, Bartram AK, Truszkowski JM, Brown DG, Neufeld JD. PANDAseq: paired-end assembler for illumina sequences. BMC bioinformatics. 2012;13(1):1-7. 50. Shugay M, Britanova OV, Merzlyak EM, Turchaninova MA, Mamedov IZ, Tuganbaev TR, Bolotin DA, Staroverov DB, Putintseva EV, Plevova K. Towards error-free profiling of immune repertoires. Nature methods. 2014;11(6):653-5. 51. Shugay M, Bagaev DV, Turchaninova MA, Bolotin DA, Britanova OV, Putintseva EV, Pogorelyy MV, Nazarov VI, Zvyagin IV, Kirgizova VI. VDJtools: unifying post-analysis of T cell receptor repertoires. PLoS computational biology. 2015;11(11):e1004503.
|
52. Montefiori DC. Measuring HIV neutralization in a luciferase reporter gene assay. HIV protocols: Springer; 2009. p. 395-405. 750
31
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . Figure 1
A
IgH germline:
HCDR3
B
VHn
Dn
JHn
CH
20
****
BsCas12a
TsCas12a
JH4
JH4
Y F D Y W G Q G T L V T V S S
...ACTACTTTGACTACTGGGGCCAAGGAACCCTGGTCACCGTCTCCTCAG... ...TGATGAAACTGATGACCCCGGTTCCTTGGGACCAGTGGCAGAGGAGTC...
...ACTACTTTGACTACTGGGGCCAAGGAACCCTGGTCACCGTCTCCTCAG... ...TGATGAAACTGATGACCCCGGTTCCTTGGGACCAGTGGCAGAGGAGTC...
t u o k c o n k M g I %
15
10
5
Mb2Cas12a
Mb3Cas12a
Cut site
gRNA
PAM
0
G T T C
C ontrol G T T C T T C C
C ontrol G T T C T T C C
C ontrol G T T C T T C C
C ontrol T T C C
C
D
EMX1-TTTG
JH4-TTTG
JH4-GTTA
100
Mb2Cas12a-RNP
JH4-GTTC
Mb2
JH4-TTCC
Mb2
JH4-TTTG
Mb2
t u o k c o n k M g I %
80
60
40
20
****
AsCas12a-RNP
****
0
JH4-GTTA
Mb2
JH4-GTTC
Mb2
JH4-TTCC
As
T T T G
G T T A
G T T C
C ontrol T T C C
T T T G
G T T A
G T T C
C ontrol T T C C
C S S
As
As
As
IgM
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . Figure 2
A
B
A H
IgM
C
D
E
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . Figure 3
A
B
20
C
Mb2Cas12a
n k c o n k A H %
i
15
10
SpCas9
5
0
0
10
20
30
40
50
3’ mismatch tail (nt)
3’ mismatch tail
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . Figure 4
A
HCDR3 sequence: Jeko-1 CARIRGFGVVNLPDYWGQG PG9 PG16
CVREAGGPDYRNGYNYYDFYDGYYNYHYMDVWGKG CAREAGGPIWHDDVKYYDFNDGYYNYHYMDVWGQG
B
PG9-CVR
PG9-CAR
PG16
HA tag
HDRT: PG9-CVR CVREAGGPDYRNGYNYYDFYDGYYNYHYMDYWGQG PG9-CAR CAREAGGPDYRNGYNYYDFYDGYYNYHYMDYWGQG PG16 CAREAGGPIWHDDVKYYDFNDGYYNYHYMDYWGQG HA tag CARGGAGGYPYDVPDYAGGAGGYFDYWGQG
2 G S P
C
e v i t i s o P %
15
10
5
PG9-CVR PG9-CAR PG16 HA tag
P S O S
I
0
2 G S P
P S O S
I
p 2 E
2 G S P
P S O S
I
p 2 E
2 G S P
P S O S
I
p 2 E
2 G S P
P S O S
I
p 2 E
p 2 E
IgM
D
No editing
Before sort
Post sort
F
No editing
Before sort
PSG2 sort
E2p sort
2 G S P
IgM
E
Representative sequences:
CARIRGFGVVNLPDYW
Original
PSG2
CRF250 (SOSIP)
CARIGVVNLPDHW
Indel
CAREAGGPDYRNGYNYYDFYDGYYNYHYMDYW
Perfect HDR
BG505 (SOSIP)
WITO (SOSIP)
CAREAGGPDYRNGYNYYLPDYW
Imperfect HDR
CAREAGGPDYRNGYSYYDFYDGYYNYHYMDYW
Imperfect HDR
dBG505 (SOSIP)
BG505 (E2p)
CAREAGGPDYRNGYNYYDFYDGYYNWGFGVVNLPDYW
Imperfect HDR
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021.
|
The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . Figure 5
A
P C r e P - 2 G S P
PSG2-APC
B
C
e v i t i s o p 2 G S P %
1.5
1.0
0.5
**
****
n.s. R D H %
1.2
1.0
0.8
0.6
0.4
VH4 VH3 VH1
0.2
0.0
0.0
l l
u n
l
o r t n o c
1 H V
3 H V
4 H V
x m
i
1
2
3741 0422
Human donor
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.04.01.437943
;
this version posted April 3, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
available under a
CC-BY-NC-ND 4.0 International license . Figure 6
A
Control template
)
%
15
HA-HCDR3
Post sort
Before sort
( y c n e u q e r F
10
5
0
Anti-HA
8 1 - 1 V H G
I
2 - 1 V H G
I
4 2 - 1 V H G
I
3 - 1 V H G
I
6 4 - 1 V H G
I
8 5 - 1 V H G
I
9 6 - 1 V H G
I
8 - 1 V H G
I
1 1 - 3 V H G
I
5 1 - 3 V H G
I
0 2 - 3 V H G
I
1 2 - 3 V H G
I
3 2 - 3 V H G
I
0 3 - 3 V H G
I
3 - 0 3 - 3 V H G
3 3 - 3 V H G
I
3 - 8 3 - 3 V H G
D 3 4 - 3 V H G
8 4 - 3 V H G
I
9 4 - 3 V H G
I
3 5 - 3 V H G
I
4 6 - 3 V H G
I
6 6 - 3 V H G
I
7 - 3 V H G
I
4 7 - 3 V H G
I
9 - 3 V H G
I
4 - 0 3 - 4 V H G
1 3 - 4 V H G
I
4 3 - 4 V H G
I
2 - 8 3 - 4 V H G
9 3 - 4 V H G
I
9 5 - 4 V H G
I
1 6 - 4 V H G
I
I
B
I
I
I
I
Post sort
Control template
Before sort
)
%
( y c n e u q e r F
20
15
10
5
0
CH01-HCDR3
CRF250 SOSIP
8 1 - 1 V H G
I
2 - 1 V H G
I
4 2 - 1 V H G
I
3 - 1 V H G
I
6 4 - 1 V H G
I
9 6 - 1 V H G
I
8 - 1 V H G
I
1 1 - 3 V H G
I
3 1 - 3 V H G
I
0 2 - 3 V H G
I
1 2 - 3 V H G
I
3 2 - 3 V H G
I
0 3 - 3 V H G
I
3 - 0 3 - 3 V H G
3 3 - 3 V H G
I
3 - 8 3 - 3 V H G
3 4 - 3 V H G
I
8 4 - 3 V H G
I
3 5 - 3 V H G
I
4 6 - 3 V H G
I
7 - 3 V H G
I
9 - 3 V H G
I
1 3 - 4 V H G
I
4 3 - 4 V H G
I
9 3 - 4 V H G
I
9 5 - 4 V H G
I
1 6 - 4 V H G
I
1 - 4 - 7 V H G
I
I
I
C
Control template
)
20
PG9-HCDR3
Post sort
Before sort
%
( y c n e u q e r F
15
10
5
0
CRF250 SOSIP
8 1 - 1 V H G
I
2 - 1 V H G
I
4 2 - 1 V H G
I
3 - 1 V H G
I
6 4 - 1 V H G
I
8 5 - 1 V H G
I
9 6 - 1 V H G
I
8 - 1 V H G
I
1 1 - 3 V H G
I
5 1 - 3 V H G
I
0 2 - 3 V H G
I
1 2 - 3 V H G
I
3 2 - 3 V H G
I
0 3 - 3 V H G
I
3 - 0 3 - 3 V H G
3 3 - 3 V H G
I
3 - 8 3 - 3 V H G
D 3 4 - 3 V H G
8 4 - 3 V H G
I
9 4 - 3 V H G
I
3 5 - 3 V H G
I
4 6 - 3 V H G
I
6 6 - 3 V H G
I
7 - 3 V H G
I
4 7 - 3 V H G
I
4 - 0 3 - 4 V H G
1 3 - 4 V H G
I
4 3 - 4 V H G
I
2 - 8 3 - 4 V H G
9 3 - 4 V H G
I
9 5 - 4 V H G
I
1 6 - 4 V H G
I
I
I
I
I
I
D
E
P S O S
I
No VH
VH3-11
VH3-33 (original)
VH3-23
VH3-30
VH4-59
y t i s n e t n I
I
F M
d e z
i l
a m r o N
)
G g I /
P S O S
I
(
0.18
0.12
0.06
0.00
IgG
V H 3-33
V H 3-30
V H 3-11
V H 3-23
V H 4-59
N o V H
F
HIV isolate 16055 25710 Bal.26 BG505 CRF_AG_250
Clade C C B A AG
VH3-33 1.81 34.94 >50 8.02 0.57
VH3-30 0.95 4.89 2.30 2.29 0.20
VH3-23 1.67 >50 >50 32.22 0.86
VH3-11 3.99 >50 >50 >50 4.20
VH4-59 3.15 >50 >50 >50 1.49
|
1 2 3 4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. No substantial pre-existing B cell immunity against SARS-CoV-2 in healthy adults
Meryem Seda Ercanoglu1,8, Lutz Gieselmann1,2,8, Sabrina Dähling1, Nareshkumar
Poopalasingam1, Susanne Detmer1, Manuel Koch3,4, Michael Korenkov1, Sandro Halwe6,7,
Michael Klüver6,7, Veronica Di Cristanziano1, Hanna Janicki1, Maike Schlotz1, Johanna
Worczinski1, Birgit Gathof5, Henning Gruell1,2, Matthias Zehner1,2, Stephan Becker6,7, Kanika
Vanshylla1, Christoph Kreer1, Florian Klein1,2,3,9,*
1Laboratory of Experimental Immunology, Institute of Virology, Faculty of Medicine and
University Hospital Cologne, University of Cologne, 50931 Cologne, Germany
2German Center for Infection Research, Partner Site Bonn-Cologne, 50931 Cologne, Germany
3Center for Molecular Medicine Cologne (CMMC), University of Cologne, 50931 Cologne,
Germany
4Institute for Dental Research and Oral Musculoskeletal Biology and Center for Biochemistry,
University of Cologne, 50931 Cologne, Germany
5Institute of Transfusion Medicine, Faculty of Medicine and University Hospital Cologne, 50937
Cologne, Germany
6Institute of Virology, Philipps University Marburg, Hans-Meerwein-Straße 2, 35042 Marburg,
Germany
7German Center for Infection Research, Partner Site Giessen-Marburg-Langen, 35043
Marburg, Germany
8These authors contributed equally
9Lead Contact
Correspondence: [email protected], Tel. : +49 221 478 89693
Summary
Pre-existing
immunity against SARS-CoV-2 may have critical
implications
for our
understanding of COVID-19 susceptibility and severity. Various studies recently provided
evidence of pre-existing T cell immunity against SARS-CoV-2 in unexposed individuals. In
contrast, the presence and clinical relevance of a pre-existing B cell immunity remains to be
fully elucidated. Here, we provide a detailed analysis of the B cell response to SARS-CoV-2 in
unexposed individuals. To this end, we extensively investigated the memory B cell response
to SARS-CoV-2 in 150 adults sampled pre-pandemically. Comprehensive screening of donor
plasma and purified IgG samples for binding and neutralization in various functional assays
revealed no substantial activity against SARS-CoV-2 but broad reactivity to endemic
betacoronaviruses. Moreover, we analyzed antibody sequences of 8,174 putatively SARS-
CoV-2-reactive B cells on a single cell level and generated and tested 158 monoclonal
1
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder.
|
All rights reserved. No reuse allowed without permission. antibodies. None of the isolated antibodies displayed relevant binding or neutralizing activity
against SARS-CoV-2. Taken together, our results show no evidence of relevant pre-existing
antibody and B cell immunity against SARS-CoV-2 in unexposed adults. Introduction
The current pandemic of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2)
represents a global health emergency that challenges health care systems throughout the
world. SARS-CoV-2 infections appear with a broad spectrum of clinical manifestations ranging
from asymptomatic infections to life-threatening acute respiratory distress syndrome (ARDS),
multi organ failure, septic shock, and death. Although influenced by multiple contributing
factors, disease severity is substantially shaped by innate and adaptive immune responses. Pre-existing immunity against SARS-CoV-2 could represent a crucial determinant of disease
severity and clinical outcome. For example, recognition of SARS-CoV-2 by a pre-existing
background immunity could limit disease severity by rapidly mounting specific immune
responses. However, pre-existing immunity may also be detrimental for the clinical course by
mechanisms of antibody-dependent enhancement (ADE) (Khurana et al., 2013; de Alwis et al.,
2014; Katzelnick et al., 2017; Arvin et al., 2020) or original antigenic sin (OAS) (Vatti et al.,
2017), which have been previously described for other viral pathogens, such as dengue (de
Alwis et al., 2014; Katzelnick et al., 2017; Mongkolsapaya et al., 2003; Midgley et al., 2011;
Rothman, 2011) and influenza viruses (Linderman and Hensley, 2016; Zhang et al., 2019;
Arevalo et al., 2020). Pre-existing T cell immune responses against SARS-CoV-2 have been observed in unexposed
individuals (Mateus et al., 2020; Grifoni et al., 2020; Le Bert et al., 2020; Braun et al., 2020;
Bacher et al., 2020; Weiskopf et al., 2020; Echeverría et al., 2021). In these studies, T cell
reactivity against the spike (S) and nucleocapsid (N) proteins as well the non-structural
proteins NSP7 and NSP13 was determined using antigen peptide pools (Grifoni et al., 2020;
Le Bert et al., 2020; Mateus et al., 2020). Importantly, pronounced T cell reactivity was detected
against S protein peptides exhibiting a high degree of homology to endemic ‘common cold’
human coronaviruses (HCoV), including HCoV-OC43, HCoV-HKU-1, HCoV-NL63, and HCoV-
229E. Therefore, pre-existing T cell immunity against SARS-CoV-2 is hypothesized to originate
from prior exposure to endemic HCoVs (Mateus et al., 2020; Grifoni et al., 2020; Le Bert et al.,
2020; Braun et al., 2020; Weiskopf et al., 2020). Pre-existing B cell immunity may be already germline-encoded in the naïve B cell repertoire or
originate from cross-reactive immune responses against related pathogens or variants. Many
potent SARS-CoV-2 neutralizing antibodies (nAbs) exhibit binding by germline-encoded amino
2
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021.
|
The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. acid residues within the complementarity-determining regions 1 and 2 (CDRH1 and CDRH2)
(Barnes et al., 2020a; Hurlburt et al., 2020; Shi et al., 2020; Wu et al., 2020; Yuan et al., 2020),
are restricted to specific heavy chain V genes (Brouwer et al., 2020; Cao et al., 2020; Ju et al.,
2020; Robbiani et al., 2020; Rogers et al., 2020; Seydoux et al., 2020; Wu et al., 2020; Zost et
al., 2020), or exhibit a low degree of somatic mutations (Barnes et al., 2020b; Kreer et al.,
2020a; Robbiani et al., 2020; Seydoux et al., 2020). This suggests that near-germline B cell
receptor (BCR) sequences with close similarity to SARS-CoV-2 reactive antibodies are already
encoded in the naïve B cell repertoire and can be readily selected to mount a potent B cell
response without further affinity maturation. In line with this, we previously identified potential
heavy and/or light-chain precursor sequences of SARS-CoV-2- binding as well as -neutralizing
antibodies by deep sequencing of naïve B cell receptor repertoires sampled before the SARS-
CoV-2 pandemic (Kreer et al., 2020a). Beyond germline encoded immunity, cross-reactive immune responses to endemic HCoVs
may also account for a pre-existing humoral or B cell immunity against SARS-CoV-2. Recent
studies provide controversial results regarding the frequency of cross-reactive antibodies in
the sera of unexposed individuals (Anderson et al., 2021; Ng et al., 2020; Nguyen-Contant et
al., 2020; Poston et al., 2020; Shrock et al., 2020; Song et al., 2021) and their association with
protection from disease severity or hospitalization (Anderson et al., 2021; Gombar et al., 2021;
Sagar et al., 2021). Whereas one study found SARS-CoV-2 reactive antibodies in a
considerable number of unexposed individuals -particularly among children, adolescents and
pregnant women (Ng et al., 2020)– other studies did not find comparable evidence (Anderson
et al., 2021; Nguyen-Contant et al., 2020; Poston et al., 2020; Song et al., 2021). All studies to
date depended on the investigation of SARS-CoV-2 reactivity in serum/plasma or affinity-
enriched or secreted IgG fractions of unexposed individuals (Anderson et al., 2021; Ng et al.,
2020; Nguyen-Contant et al., 2020; Poston et al., 2020; Shrock et al., 2020; Song et al., 2021). These conflicting results call for more comprehensive and detailed investigations which
beyond plasma and IgG fractions also involve BCR sequence analysis of SARS-CoV-2
reactive B cells and characterization of recombinant monoclonal antibodies. In order to investigate the presence of a relevant pre-existing SARS-CoV-2 B cell immunity,
we extensively investigated plasma samples, single B cells, and monoclonal antibodies
isolated from 150 SARS-CoV-2 unexposed individuals. We found no evidence of a pre-existing
B cell immunity that may account for the broad clinical spectrum of SARS-CoV-2 infections.
|
Instead, our results indicate that rare naïve B cell precursors are selected to mount a SARS-
CoV-2 directed antibody response upon infection. Results
3
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Pre-pandemic naïve B cell receptor repertoires encode for heavy and light
chains of SARS-CoV-2 reactive antibodies
By performing NGS analyses on healthy individuals, we previously identified rare heavy and
light chain variable regions of pre-pandemic naive B cells that closely resembled near-germline
SARS-CoV-2-reactive antibodies (Kreer et al., 2020a). This raised the question, whether naïve
B cells can encode for antibodies that do not require further affinity maturation for a potent
SARS-CoV-2 reactivity. To test this hypothesis, we cloned and expressed 31 pre-pandemic
heavy or light chains obtained from naïve B cells together with heavy or light chains from
SARS-CoV-2-reactive mAbs derived from convalescent individuals and tested these chimeras
for S protein binding and SARS-CoV-2 neutralization activity (Figure 1). V gene and CDR3
differences between NGS-derived and original chains ranged from 0 to 5 and 0 to 3 amino
acids, respectively (Table S2). Except for mAbs CnC2t1p1_B4 and MnC2t1p1_C12, pairing of
pre-pandemic light chains with the original heavy chains did retain binding and/or neutralization
activity. For 20 out of 23 chimeric mAbs, pairing of pre-pandemic heavy chains with the original
light chain led to a loss of binding and neutralization activity. However, for one antibody
(MnC2t1p1_C12) pairing of the original light chain with 3 out of 13 pre-pandemic heavy chains
was able to retain binding activity suggesting that these heavy chains can assemble SARS-
CoV-2 spike protein reactive antibodies. We conclude that some naïve B cells can express
heavy or light chains that can be components of SARS-CoV-2-reactive antibodies without the
need for further affinity maturation. Pre-pandemic plasma samples and polyclonal IgG exhibit no significant
reactivity to SARS-CoV-2
To investigate whether SARS-CoV-2-reactive antibodies are present in the plasma of
unexposed individuals, we investigated pre-pandemic blood samples from 150 donors. Samples were collected between August and November 2019 and were studied for binding
and neutralization activity against SARS-CoV-2 (Figure 2A and B; Table S1). Donors were
between 18 and 66 years old (with a mean/median age of 30.6/27 years) (Figure 2A and Table
S1). 49.3 % of donors were male and 50.7 % female (Figure 2A and Table S1). Plasma
samples of all 150 donors were tested for binding to the soluble full trimeric SARS-CoV-2 S
ectodomain (S1/S2) or the S1 subunit (S1) by commercially available (com.
|
IA) and in-house
(ELISA) immunoassays. (Figures 2B, C and Figure S1). In addition, binding activity was
assessed against cell surface expressed full-length SARS-CoV-2 S protein by flow cytometry
(cell assay, CA; Figures 2B and C). Plasma samples showed mostly no or only minimal
binding of IgG, IgM and IgA to soluble or cell surface expressed SARS-CoV-2 S proteins
(Figures 2C and Figure S1). Only in a few samples (Pre033 IgA, Pre051 IgG, Pre074 IgA,
4
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Pre084 IgM and Pre097 IgA) notable binding was detected in one single assay (Figure 2C and
Figure S2). However, binding detected by immunoassays could not be confirmed by flow-
cytometry against cell surface expressed full-length S protein (Figure 2C). In CAs, one donor
(Pre004) exhibited weak plasma IgG activity to full lengths S protein. Furthermore, 1:10
dilutions of 150 plasma samples were screened for neutralization activity against SARS-CoV-
2 wildtype- (WT) and pseudovirus (PSV) (Figures 2C and Figure S2). Samples from two
donors neutralized SARS-CoV-2 pseudovirus by 50 – 60 %. However, neutralization activity
of these samples could not be confirmed by testing of serial dilutions or wildtype neutralization
assays (Figure 2C and Figure S2). Next, polyclonal IgGs (pIgG) were purified from pre-pandemic plasma samples and tested for
binding to soluble full trimeric S proteins of SARS-CoV-2 and endemic human-pathogenic
betacoronaviruses (HKU1 and OC43) (Figures 2C and Figure S2). Although pIgG showed
strong reactivity to HKU1 and OC43 S proteins, no binding to SARS-CoV-2 trimeric S protein
was detected. Furthermore, neutralization activity of pIgG against SARS-CoV-2 pseudovirus
was determined at a concentration of 1 mg/ml. PIgG of six donors exhibited neutralization
activity of SARS-CoV-2 pseudovirus by ≥ 50 %. However, neutralization activity of these
samples could not be confirmed by testing serial dilutions (Figure 2C and Figure S2). We
conclude that pre-pandemic plasma and polyclonal IgG samples obtained from 150 adults
before the onset of the pandemic exhibit only minimal reactivity to SARS-CoV-2 in few samples
and overall no SARS-CoV-2-neutralizing activity. No detection of SARS-CoV-2-reactive B cells in pre-pandemic samples
The lack of binding or neutralization activity against SARS-CoV-2 on plasma level does not
exclude the existence of SARS-CoV-2-reactive B cells in unexposed individuals. To investigate
the presence of SARS-CoV-2 specific B cells, we performed single B cell sorts of 40 donors
sampled before the pandemic (Figure 3 and Figure S3). Using the same analysis gate as for
COVID-19 convalescent donors, frequencies of SARS-CoV-2-reactive IgG+ and IgG- B cells
isolated from pre-pandemic blood samples were significantly lower (p value < 0.0001).
|
For
COVID-19 convalescent donors, frequencies ranged from 0.002 to 0.065 % for IgG+ (median
0.02 %) and 0.007 to 0.39 % for IgG- (median 0.031 %) B cells (Figure 3A). Applying the same
analysis gate of the COVID-19 convalescent samples to our pre-pandemic samples,
frequencies ranged from 0 to 0.0013 % for IgG+ (median 0.0001 %) and from 0 to 0.016 for
IgG- (median 0.003 %) B cells (Figure 3A and Figure S3). Therefore, we conclude that, if
present at all, SARS-CoV-2-reactive B cells have a significantly lower frequency in pre-
pandemic samples. We reasoned that gate settings applied for COVID-19 convalescent
samples may exclude reactive B cells with low spike affinity which may be present in individuals
5
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. unexposed to SARS-CoV-2. To assure that such cells do not remain undetected, we adjusted
the actual sorting gate (Figure 3A) to isolate a total of 8,174 putatively SARS-CoV-2-reactive
single B cells, of which 3,852 were IgG+ and 4,322 IgG- cells. Of those we amplified a total of
5,432 productive heavy chain sequences, of which 2,789 sequences accounted for IgG and
2,643 for IgM heavy chains, respectively (Figure 3B). Sequence analyses showed that in each
individual 0 to 58 % of the sequences were clonally related with a mean clone number of 8.1
clones per individual and a mean clone size of 2.9 members per clone (Figure 3B). Heavy
chain variable (VH) gene segment distribution, heavy chain complementarity-determining
region 3 (CDRH3) length and VH gene germline identities showed no notable divergence to a
reference memory IgG and naïve IgM repertoire data set (Kreer et al., 2020a) (Figure 3C). The lack of this divergence in combination with the FACS data suggest the absence of SARS-
CoV-2-reactive B cells in pre-pandemic samples. Monoclonal antibodies isolated from pre-pandemic samples are not reactive
against SARS-CoV-2
To confirm the absence of SARS-CoV-2-reactive B cells in pre-pandemic samples on a
functional level, we selected 200 antibody candidates among 36 donors from single cell sorted
B cells for production and functional testing (Figure 4A). Selection was conducted based on
sequence similarity (see Methods section) to 920 SARS-CoV-2-reactive antibodies deposited
in the CoVAbDab (n = 18) and on random sequence selection (n = 182) (Figure 4). Criteria for
the similarity selection were identical VH/JH combinations, low level of CDRH3 length
differences, and Levenshtein distances between isolated and deposited BCR sequences. The
random selection included clonal as well as non-clonal sequences and was performed to
ensure an equal selection of BCR sequences among different donors and clonotypes.
|
In total,
we successfully produced 158 monoclonal antibodies (81 IgM-, 77 IgG-derived) as IgG1
isotypes for functional testing. First, we determined binding activity of these antibodies to
SARS-CoV-2 S protein and evaluated cross-reactivity to HKU1 and OC43 S proteins by ELISA
(Figure 4B). Monoclonal antibodies showed no relevant binding or cross-reactivity against
SARS-CoV-2, HKU-1 or OC43 S proteins. Next, we tested all 158 monoclonal antibodies for
neutralization activity against SARS-CoV-2 pseudovirus in single concentrations of 50 µg/ml
(Figure S4) and in serial dilutions (Figure 4B). In line with the lack of binding activity, none of
the produced antibodies showed neutralizing activity up to concentrations of 50 µg/ml (Figure
4B and Figure S4). We conclude that putatively SARS-CoV-2-S+ B cells from pre-pandemic
samples do not encode for SARS-CoV-2-reactive B cell receptors. Discussion
6
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Adaptive immune responses against a pathogen are shaped by the naïve immune repertoire
and imprinted by previous encounters to the same pathogen or a related variant. Investigation
of pre-existing immunity to SARS-CoV-2 can advance our understanding of protective
immunity, susceptibility to infection, disease severity and guide the development of vaccination
strategies (Imkeller and Wardemann, 2018; Niu et al., 2020; Schultheiß et al., 2020). Although
various studies already provide evidence for pre-existing T cell immunity against SARS-CoV-
2, the presence of a pre-existing B cell immunity remains to be elucidated (Grifoni et al., 2020;
Le Bert et al., 2020; Mateus et al., 2020; Sette and Crotty, 2021). Recently published studies that investigated a pre-existing B cell immunity in unexposed
individuals provide partially conflicting results. For instance, one study based on the
examination of serum samples reported detection of SARS-CoV-2 S protein cross-reactive
antibodies and pre-existing humoral immunity in uninfected children, adolescents, or pregnant
women (Ng et al., 2020). Cross-reactive humoral immune responses were mainly observed
against S protein structures within the S2 subunit that are relatively conserved among endemic
HCoVs and SARS-CoV-2. The authors of this study hypothesize that prior immunity against
HCoVs may modify COVID-19 disease severity and susceptibility as well as seasonal and
geographical transmission patterns. In contrast, other studies only rarely detected cross-
reactive serum and B cell responses against SARS-CoV-2 in pre-pandemic samples and argue
against a broad and clinically relevant pre-existing B cell immunity (Anderson et al., 2021;
Nguyen-Contant et al., 2020; Poston et al., 2020; Shrock et al., 2020; Song et al., 2021).
|
Of
note, all studies to date only investigated of plasma/serum and enriched or PBMC secreted
IgG fractions of pre-pandemic samples (Anderson et al., 2021; Ng et al., 2020; Nguyen-
Contant et al., 2020; Poston et al., 2020; Shrock et al., 2020; Song et al., 2021). They did not
involve the analysis of naïve B cell receptor repertoires and putatively SARS-CoV-2-reactive
B cells or the characterization of recombinant monoclonal antibodies derived from these cells. Therefore, such plasma-based studies may miss pre-existing potent naïve B cell precursors
and low frequency memory B cells which do not contribute sufficient amounts of functionally
detectable antibodies to plasma Ig fractions. In our study we cover not only an extensive
examination of plasma and IgG fractions but reach out further to investigate the presence of
SARS-CoV-2-reactive B cell precursors, memory B cells and the characterization of respective
monoclonal antibodies. We recently isolated SARS-CoV-2-reactive antibodies from convalescent individuals and
identified highly similar heavy and/or light chain sequences in naive B cell receptor repertoires
from pre-pandemic samples (Kreer et al., 2020a). However, our previous publication did not
cover any functional data. Here, we show that some of these chains can replace the original
heavy or light chain in SARS-CoV-2-reactive antibodies without altering their functionality. This
7
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. suggests that the readily detected antibody response in some individuals might use existing
antibody heavy or light chains with distinct CDR3 recombination patterns or germline encoded
sequence features. In line with this, SARS-CoV-2 neutralizing antibodies were successfully
isolated from human naïve antibody gene libraries using phage display (Bertoglio et al., 2021). The most promising antibody candidate STE73-2E9 isolated from these libraries specifically
targets the ACE2-RBD interface without cross-reactivity to other coronaviruses and neutralizes
authentic SARS-CoV-2 wildtype virus with an IC50 of 0.43 nM. Since phage display relies on
random sequence recombination, this study does not provide evidence that SARS-CoV-2-
reactive antibodies naturally occur in the naïve B cell repertoire. Using various immunological and functional assays assessing SARS-CoV-2 binding and
neutralization activity as well as cross-reactivity to endemic beta coronaviruses, we found no
evidence of significant plasma or IgG reactivity against SARS-CoV-2 in pre-pandemic
samples. We comprehensively investigated plasma binding activity of IgG, IgM and IgA
immunoglobulin isotypes against diverse beta coronavirus S proteins (SARS-CoV-2 S1 and
S1/S2, HCoV-HKU1 and HCoV-OC43 S) by in house and commercially available ELISAs.
|
Moreover, we validated our ELISA results against cell-surface-expressed S protein using flow
cytometry and by applying SARS-CoV-2 pseudo- as well as wildtype neutralization assays. Our results are consistent with recently published studies that disagreed on a broad pre-
existing B cell background immunity (Anderson et al., 2021; Ng et al., 2020; Nguyen-Contant
et al., 2020; Poston et al., 2020; Shrock et al., 2020; Song et al., 2021). However, the findings
of our plasma screening may not be directly comparable to the one recent study that indeed
reported pre-existing humoral immunity in unexposed children, adolescents, and pregnant
women (Ng et al., 2020) since our cohort does not encompass such particular groups of
individuals. To investigate potential pre-existing immunity on a molecular level, we applied antigen-specific
single cell sorts to 40 donors and isolated a total of 8,174 putatively SARS-CoV-2-reactive B
cells. Consistent with our findings from binding and neutralization screening of plasma samples
and polyclonal IgG, frequencies of putatively SARS-CoV-2-reactive B cells in pre-pandemic
samples were very low especially when compared to COVID-19 convalescent samples. The
choice of an appropriate bait protein for single cell sorting strategies critically determines the
isolation of antigen-specific B cells. Immunogenic structures that are not displayed by the
chosen bait protein are excluded from isolation. To ensure the comprehensive isolation of
antigen-specific B cells, we therefore chose to apply the native, full trimeric SARS-CoV-2 S
protein and to adjust the sorting gate. With regard to the low frequencies of SARS-CoV-2-
reactive B cells in our study, it can be argued that application of alternative bait proteins e.g. the S2 subunit would have been favorable (Ng et al., 2020; Nguyen-Contant et al., 2020;
8
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Shrock et al., 2020; Song et al., 2021). Antibody sequence analyses of isolated putatively
SARS-CoV-2-reactive B cells revealed diverse heavy chain V gene usage, normally distributed
CDRH3 lengths and VH gene germline identities similar to the naïve BCR receptor repertoire
of healthy individuals sampled before the pandemic. In particular, we found no evidence of
enrichment of specific V regions which were preferentially used in antibodies of SARS-CoV-2
convalescent individuals such as IGHV3-30, IGHV3-53 or IGHV3-63 (Barnes et al., 2020b,
2020a; Robbiani et al., 2020). Furthermore, we detected no relevant reactivity of 158
recombinantly produced monoclonal antibodies against SARS-CoV-2 and endemic HCoV S
proteins. A pre-existing B cell immunity against pathogens can be a critical determinant of clinical
outcome of infections and vaccination strategies.
|
Recent studies provide conflicting results
regarding the presence and importance of a pre-existing B cell immunity against SARS-CoV-
2 in unexposed individuals. While a pre-existing B cell immunity against SARS-CoV-2 was
reported in special study cohorts including children, adolescents or pregnant women, our
detailed analysis of the B cell response yielded no comparable evidence in a large population
of healthy adults. Acknowledgments
We thank all study participants for supporting our research by blood donation; members of the
Klein and Becker laboratories for their support and inspiring discussions; Raiees Andrabi,
Victor Corman, Jason McLellan for sharing and providing the SARS-CoV-2 and endemic
HCoVs S ectodomain plasmids; Daniela Weiland and Nadine Henn for lab management and
assistance; Birgit Gathof and Sabine Adam of the Institute of Transfusion Medicine of the
University Hospital of Cologne for organizing and providing blood donations of study
participants; technical assistants of the serology department at the Institute of Virology of the
University Hospital of Cologne for assistance with ELISA testing of plasma samples; as well
as Stephan Becker and Verena Kraehling for sharing VeroE6 cells and Susanne Berghöfer for
excellent technical assistance with neutralization assays. This work was funded by grants from
the German Center of Infection Research (DZIF) to F.K and S.B., the German Research
Foundation (DFG) CRC1279 and CRC1310, European Research Council (ERC) ERC-
stG639961, and COVIM: NaFoUniMed-Covid19 to F. Klein. Author Contributions
Conceptualization – F.K., M.S.E., L.G., C.K. and M.Z. ; Methodology – M.S.E., L.G., S. Dähling,
N.P., S. Dettmer, M. Korenkov, K.V., V.D.C., S.H, Michael Klüver, C.K. ; Formal analysis –
9
315
316
317
318
319
320
321
322 323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. M.S.E., L.G., C.K., V.D.C, S.H., V.K., S. Dähling; Investigation – M.S.E., L.G., C.K., V.D.C.,
S.H., V.K., S. Dähling, S. Dettmer, M. Korenkov, and N.P. ; Resources – B.G. ; Writing original
draft - M.S.E., L.G., C.K. and F.K. ; Writing reviewing and editing – all authors; Supervision –
F.K., C.K. ; Funding acquisition – F.K.. These authors contributed equally to the work: M.S.E. and L.G. Declaration of Interests
The authors have no competing interests to declare. Figure Legends
Figure 1: Pre-pandemic naïve B cells express SARS-CoV-2-reactive heavy or
light chains
(A) Selection of heavy and light chain sequences in a pre-pandemic naïve B cell repertoire that
closely resemble heavy and light chains of SARS-CoV-2 reactive antibodies (Abs). (B) Binding
and neutralization plots against SARS-CoV-2 of 6 previously described SARS-CoV-2-reactive
antibodies (red) and 31 chimeric antibodies composed of heavy or light chain of the original
SARS-CoV-2-reactive antibodies paired with a NGS-derived heavy (darker grey) or light (grey)
chain.
|
Activity of mAbs against SARS-CoV-2 was determined by serial dilution ELISA and
pseudotyped neutralization assays (PsV-NA). Each antibody was produced in at least
duplicates. ELISAs and neutralization assays were performed with biological replicates. Symbols depict means and error bars indicate standard deviation. Figure 2: Screening of pre-pandemic samples from 150 adults reveal no relevant
reactivity against SARS-CoV-2
(A) Timeline of blood collections and demographic characteristics of 150 donors sampled
before the SARS-CoV-2 pandemic. Pre-pandemic blood samples were collected as buffy coats
between August and November of 2019. Gender and age distribution of donors are illustrated
as pie chart and bar plots. (B) Pre-pandemic plasma samples and purified IgGs (pIgG) were
tested for binding and neutralization using different experimental approaches. Plasma IgG, IgM
and IgA as well as pIgG were tested for binding to SARS-CoV-2, HKU1 and OC43 S proteins
using in house (ELISA) and commercially available immunoassays (com. IA). Binding of
plasma IgG and IgM to cell surface expressed SARS-CoV-2 S protein was also determined by
10
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. FACS analysis (CA). Neutralization activity was determined against SARS-CoV-2 pseudo- and
wildtype virus (PSV or WT). (C) Heat map visualizing binding (AUC or AU) and neutralization
activity (% or CPE) of pre-pandemic plasma samples and pIgG against SARS-CoV-2 and
endemic HCoVs (HKU1 and OC43) S proteins or SARS-CoV-2 wildtype (WT) and pseudovirus
(PSV Wu_01), respectively (see also Figure S1 and S2). Immunoassays were performed in
duplicate experiments and the AUC is presented as geometric mean of duplicates. Neutralization activity was first determined for single sample concentrations. Samples that
displayed neutralization activity of ≥ 50% in single concentrations were repeatedly investigated
in serial dilutions (x) (see also Figure S2). Samples were tested in duplicates in single
experiments. The average of neutralization is presented and each row represents one donor. Figure 3: Lack of SARS-CoV-2-reactive B cells and distinct antibody sequence
features in pre-pandemic samples
(A) Representative dot plots of SARS-CoV-2-reactive, CD19+CD20+, IgG+ and IgG- B cells of
COVID-19 samples compared to pre-pandemic samples. Depicted numbers indicate
frequencies of S protein reactive B cells (see also Figure S3). Red colored gate indicates
gating strategy for analysis and doted gate indicates actual sorting gate. Dot plot bar graph
displays the mean ± SD frequency of SARS-CoV-2-reactive, IgG+ and IgG- B cells in 40 pre-
pandemic and 23 COVID-19 samples (*P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001;
unpaired two-tailed t test).
|
(B) Clonal relationship of heavy chain sequences amplified from
single SARS-CoV-2-reactive IgG+ and IgG- B cells isolated from 40 donors sampled before the
pandemic. Individual clones are colored in shades of blue, grey and white. In the center of
each pie chart, numbers of productive heavy chain sequences are illustrated. Presentation of
clone sizes are proportional to the total number of productive heavy chain sequences per
clone. (C) VH gene distribution, VH gene germline identity and CDRH3 length distribution in
amino acids (AA) were separately determined for IgG and IgM. Distributions were calculated
per individual. Bar and line plots show mean ± SD. Figure 4: Monoclonal antibodies isolated from pre-pandemic samples are not
reactive to SARS-CoV-2 and endemic HCoVs
(A) Illustration demonstrating the sequence selection for downstream antibody production. Sequences were selected for antibody production based on similarity to antibody sequences
deposited in the CoVAbDab and on random selection. From 8,174 SARS-CoV-2-reactive
IgG+ and IgG- B cells, 5,432 productive heavy chain sequences were amplified. For antibody
11
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. production, 18 HC sequences were selected based on similarity selection and 182 HC
sequences were selected based on random selection. In total 158 antibodies were produced. (B) Heat map visualizing binding (AUC) and neutralization activity (%) of monoclonal
antibodies isolated from pre-pandemic blood samples against SARS-CoV-2 and endemic
HCoVs (HKU1 and OC43) S proteins or SARS-CoV-2 pseudovirus (PSV Wu_01),
respectively (see also Figure S4). Each row represents one monoclonal antibody. ELISAs
were performed in duplicate experiments and the AUC is presented as geometric mean of
duplicates. Neutralization activity was determined for single sample concentrations. Samples
were tested in duplicates and the average of neutralization is presented. (C) Neutralization
activity against SARS-CoV-2 pseudovirus (PSV Wu_01) was verified for all mAbs in serial
dilutions. STAR Methods
RESOURCE AVAILABILITY
Lead contact
Further information and requests for resources and reagents should be directed to and will be
fulfilled by the Lead Contact, Florian Klein ([email protected]). Materials availability
There are restrictions to the availability of SARS-CoV-2-reactive antibodies due to limited
production capacities and ongoing consumption. Reasonable amounts of antibodies will be
made available by the Lead Contact upon request under a Material Transfer Agreement (MTA)
for non-commercial usage. Nucleotide sequences and expression plasmids will be shared
upon request.
|
Data and code availability
Nucleotide sequences of isolated monoclonal antibodies will be deposited at GenBank after
review or acceptance of the manuscript. Further heavy and light chain sequences as well as
NGS data of healthy individuals will be shared by the Lead Contact upon request. EXPERIMENTAL MODEL AND SUBJECT DETAILS
Study participants and collection of clinical samples
12
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Study participants were recruited at the University Hospital Cologne. Buffy coats from study
participants were collected according to a study protocol approved by the Institutional Review
Board of the University of Cologne (study protocol 16-054). Buffy coats were provided as
residual products in the context of regular blood donations from the Institute for Transfusion
Medicine at the University Hospital of Cologne. All study participants provided informed
consent. The study cohort comprised 150 individuals with 50.7% female and 49.3% male
adults. All study participants were ≥ 18 years old. Blood samples were collected between
August and November 2019. Cell lines
HEK293T cells were maintained in Dulbecco’s Modified Eagle Medium (DMEM, Thermo
Fisher) supplemented with 10 % fetal bovine serum (FBS, Sigma-Aldrich), 1 x antibiotic-
antimycotic (Thermo Fisher), 1 mM sodium pyruvate (Gibco) and 2 mM L-glutamine (Gibco) at
37 °C and 5 % CO2. HEK293-6E cells were maintained in FreeStyle 293 Expression Medium
(Life Technologies) supplemented with 0.2 % penicillin/streptomycin under constant shaking
at 90 – 120 rpm, 37 °C and 6 % CO2. VeroE6 and HEK293T-ACE2 cells were maintained in
Dulbecco’s Modified Eagle Medium (Gibco) supplemented with 10 % fetal bovine serum (FBS,
Sigma-Aldrich), 1 x penicillin-streptomycin (Gibco), 1 mM sodium pyruvate (Gibco) and 2 mM
L-glutamine (Gibco) at 37 °C and 5 % CO2. The sex of HEK293T, HEK293-6E and VeroE6 cell
lines is female. Cell lines were not specifically authenticated. METHOD DETAILS
Isolation of PBMCs, plasma and total IgG
Peripheral blood mononuclear cells (PBMCs) were isolated from buffy coats and whole blood
by density gradient separation using Histopaque separation medium (Sigma-Aldrich) and
Leucosep cell tubes (Greiner Bio-one) according to the manufacturer’s instructions. Isolated
PBMCs were cryopreserved at -150 °C in 90 % FBS supplemented with 10 % DMSO until
further use. Plasma was collected and stored at -80 °C. Plasma samples were heat-inactivated
at 56 °C for 40 min prior to further use. For IgG isolation, 1 ml of heat-inactivated plasma was
incubated with Protein G Sepharose (GE Life Sciences) overnight at 4 °C (GE Life Sciences)
under constant rotation.
|
Protein G beads were transferred to chromatography columns and
washed twice with sterile PBS. IgGs were eluted from columns using 0.1 M glycine (pH = 3.0)
and buffered in 0.1 M Tris (pH = 8.0). Buffer exchange to PBS and concentration of IgGs was
performed by centrifugation using 30 kDa Amicon spin membranes (Millipore). The final
13
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. concentration of purified IgGs was determined by UV/Vis spectroscopy using a Nanodrop
(A280). Subsequently, purified IgGs were stored at 4 °C. Expression and purification of viral surface proteins
Constructs encoding the stabilized spike protein ectodomain of the endemic HCoV-HKU1
(amino acids 1 – 1295, GenBank ID. : ABD75497) and HCoV-OC43 (amino acids 1 – 1300,
GenBank ID. : AAX84792) were kindly provided by Raiees Andrabi (California, USA) and Victor
Corman (Berlin, Germany) and described previously (St-Jean et al., 2004; Woo et al., 2005;
Corman et al., 2012; Kreye et al., 2020; Song et al., 2021). Constructs encoding the prefusion
stabilized SARS-CoV-2 S ectodomain (amino acids 1 – 1207; GenBank ID. : MN908947) were
kindly provided by Jason McLellan (Texas, USA) and Florian Krammer (New York, USA) and
previously described (Stadlbauer et al., 2020; Wrapp et al., 2020). All three constructs contain
the following mutations: two proline substitutions for prefusion state stabilization (SARS-CoV-
2: residues 986 and 987; OC43: residues 1078 and 1079; HKU1: residues 1071 and 1072)
and the furin cleavage sites were mutated (SARS-CoV-2: ‘‘GGGG’’ substitution at residues
682–685; OC43: ‘‘GSAS’’ substitution at residues 762-766; HKU1‘‘GSAS’’ substitution at
residues 756-760). Ebola surface glycoprotein (EBOV Makona, GenBank ID. : KJ660347;
amino acids 1 – 651) was used as a negative control for HCoV S protein ELISAs. The Ebola
surface glycoprotein was stabilized with GCN4 trimerization domain and expressed without the
transmembrane domain as previously described (Ehrhardt et al., 2019). The different
coronavirus ectodomains were amplified from the synthetic gene plasmids by PCR and
subsequently cloned into a modified sleeping beauty transposon expression vector containing
a C-terminal T4 fibritin trimerization motif (foldon) followed by a Twin-Strep-Tag purification
tag. For the recombinant protein productions, stable HEK293 EBNA cell lines were generated
employing the sleeping beauty transposon system (Kowarz et al., 2015). Briefly, expression
constructs were transfected into the HEK293 EBNA cells using FuGENE HD transfection
reagent (Promega). After selection with puromycin, cells were induced with doxycycline. Cell
supernatants were filtered and the recombinant proteins purified via Strep-Tactin XT (IBA
Lifescience) resin.
|
Proteins were then eluted by biotin-containing TBS-buffer (IBA Lifescience),
and dialyzed against TBS-buffer. Isolation of single SARS-CoV-2-reactive B cells
CD19+ B cells were enriched from PBMCs by immunomagnetic cell separation using CD19
microbeads (Miltenyi Biotec) according to the manufacturer’s protocol. Enriched B cells were
labeled for 20 min on ice with 4’,6-Diamidin-2-phenlindol (DAPI; Thermo Fisher Scientific), anti-
human CD20-Alexa Fluor 700 (BD), anti-human IgG-APC (BD), anti-human CD27-PE (BD)
14
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. and DyLight488-labeled SARS-CoV-2 S protein (10 µg/ml). DAPI-, CD20+, SARS-CoV S
protein+, IgG+/IgG- single cells were sorted into 96 well plates using a FACSAria Fusion (Becton
Dickinson). Each well of the 96 well plate was prefilled with 4 µl of sorting buffer consisting of
0.5 x PBS, 0.5 U/µl RNAsin (Promega), 0.5 U/µl RNaseOut (Thermo Fisher Scientific), and 10
mM DTT (Thermo Fisher Scientific). Plates were cryopreserved at -80 °C immediately after
sorting. B cell receptor amplification and sequence analysis
Generation of cDNA and amplification of antibody heavy and light chain genes from sorted
single cells was performed as previously described (Gieselmann et al., 2021; Kreer et al.,
2020a; Schommers et al., 2020). For reverse transcription, sorted cells were incubated with 7
µl of a random-hexamer-primer master mix (RHP mix) consisting of 0.75 µl Random Hexamer
Primer (Thermo Fisher Scientific), 0.5 µl NP-40 (Thermo Fisher Scientific), 0.15 µl RNaseOut
(Thermo Fisher Scientific), and 5.6 µl of nuclease-free H2O at 65 °C for 1 min. Subsequently,
samples were incubated with a reverse-transcription master mix (RT mix) consisting of 3 µl of
5 x Superscript IV RT buffer (Thermofisher Scientific), 0.5 µl dNTPs (Thermo Fisher Scientific),
1 µl DTT (Thermo Fisher Scientific), 0.1 µl of RNasin (Promega), 0.1 µl of RNaseOut (Thermo
Fisher Scientific), 2.05 µl of nuclease-free H2O and 0.25 µl of Superscript IV (Thermofisher
Scientific) per well and incubated at 42 °C for 10 min, 25 °C for 10 min, 50°C for 10 min and
94 °C for 5 min. Heavy and light chains were amplified from cDNA by semi-nested, single cell
PCRs using optimized V gene-specific primer mixes(Kreer et al., 2020b) and Platinum Taq
DNA Polymerase or Platinum Taq Green Hot Start Polymerase (Thermo Fisher Scientific) as
previously described (Kreer et al., 2020a; Schommers et al., 2020; Kreer et al., 2020b;
Gieselmann et al., 2021). PCR products were analyzed by agarose gel electrophoresis for
correct product size and subsequently send for Sanger sequencing. Only chromatograms with
a mean Phred score of 28 and sequences with a minimal length of 240 nucleotides were
selected for downstream sequence analyses.
|
Filtered sequences were annotated with IgBlast
(Ye et al., 2013) according to the IMGT system (Lefranc et al., 2009) and only the variable
region from FWR1 to the end of the J gene was extracted. Base calls within the variable region
with a Phred score below 16 were masked and sequences with more than 15 masked
nucleotides, stop codons, or frameshifts were excluded from further analyses. Sequence
analyses to inform on sequence clonality were performed separately for each study participant. All productive heavy chain sequences were grouped by identical VH/JH gene pairs and the
pairwise Levenshtein distance for their CDRH3s was determined. Starting from a random
sequence, clone groups were assigned to sequences with a minimal CDRH3 amino acid
identity of at least 75 % (with respect to the shortest CDRH3). 100 rounds of input sequence
randomization and clonal assignment were performed and the result with the lowest number
15
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. of remaining unassigned (non-clonal) sequences was selected for downstream analyses. All
clones were cross-validated by the investigators taking shared mutations and light chain
information into account. Sequence selection for antibody production
Antibody selection for cloning was done by two approaches. First, a similarity search was
performed against 868 and 52 SARS-CoV-2-binding antibodies from the CoVAbDab
(Raybould et al., 2020) (retrieved on 21.08.20) and Kreer et al 2020 (Kreer et al., 2020a),
respectively. B cell receptor sequences with identical VH/JH combination, a CDRH3 length
difference ≤ 2 AA, and a CDRH3 Levenshtein distance ≤ 3 AA in comparison to at least one
SARS-CoV-2-binding antibody were selected as similar (18 in total). Second, a random
selection was performed to yield at least 6 random antibody sequences per individual including
at least 3 different clones and at least 3 non-clonal sequences. Random non-clonal sequences
were used to fill up the selection if less than 3 clones were available. Cloning and production of monoclonal antibodies
Heavy and light chain variable regions of selected antibodies were cloned into expression
vectors by sequence and ligation independent cloning (SLIC) (von Boehmer et al., 2016) as
previously described (Tiller et al., 2008; Schommers et al., 2020; Kreer et al., 2020a;
Gieselmann et al., 2021). 1st PCR product was amplified using Q5 Hot Start High Fidelity DNA
Polymerase (New England Biolabs) and specific forward and reverse primers including adaptor
sequences which are homologous to the restriction sites of the antibody expression vector
(IgG1, IgL, IgK (Tiller et al., 2008)). Forward primers were designed according to 2nd PCR
primers (Kreer et al., 2020b) and encode for the complete native leader sequence of all heavy
and light chain V genes, whereas reverse primers bind to the conserved sequence motifs at
the 5’ end of heavy and light chain immunoglobulin constant regions.
|
PCR was run at 98 °C
for 30 s; 35 cycles of 98 °C for 10 s, 72 °C for 45 s; and 72 °C for 2 min. Subsequently, PCR
products were purified with 96-well format silica membranes, cloned by SLIC (von Boehmer et
al., 2016) into linearized expression vectors with T4 DNA polymerase (NEB) and transformed
into chemically competent Escherichia coli (DH5α). Heavy chain variable regions of IgG-
B cells
were cloned into IgG1 expression vectors. Correct insertion of the V region sequence into the
expression vector was examined by colony PCR and Sanger sequencing. Positive colonies
were propagated in midi cultures and plasmids purified. For production of recombinant monoclonal antibodies, HEK293-6E suspension cells were co-
transfected chemically with human heavy chain and corresponding light chain antibody
expression vectors using 25 kDa branched polyethylenimine (PEI; Sigma-Aldrich). Transfected
16
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. cells were propagated in FreeStyle 293 Expression Medium (Thermo Fisher Scientific)
supplemented with 0.2 % penicillin/streptomycin (Thermo Fisher Scientific) at 37 °C and 6 %
CO2 under constant shaking at 90 – 120 rpm for seven days. Cell supernatants were harvested
by centrifugation, filtered through PES filters and incubated overnight at 4 °C under constant
rotation with Protein-G-coupled Sepharose beads (GE Life Sciences). The suspension was
transferred to chromatography columns, washed twice with sterile PBS and antibodies were
eluted with 0.1 M glycine (pH = 3) and buffered using 1 M Tris (pH = 8). Buffer exchange to
PBS and concentration of antibodies was performed by centrifugation using 30kDa Amicon
spin membranes (Millipore). Antibodies were filtered through Ultrafree-MC 0.22 µm
membranes (Millipore) and stored at 4 °C. The final concentration of purified antibodies was
determined by UV/Vis spectroscopy using a Nanodrop (A280). Subsequently, purified
antibodies were stored at 4 °C. Cell-surface-expressed S protein immunoassay
HEK293T cells were transfected with plasmids encoding the full-length SARS-CoV-2 S protein
(GenBank ID. : MN908947) using TurboFectTM transfection reagent (Thermo Fisher Scientific). After incubation at 37 °C and 5 % CO2 for 48 h, adherent cells were detached with PBS
supplemented with 1 mM EDTA (pH = 7.4) and resuspended in FACS buffer (1x PBS
supplemented with 2 % FCS and 2 mM EDTA). 3 x 104 cells were distributed in 50 µl of FACS
buffer to each well of a V-bottom-shaped 96 well plate. Starting with an initial dilution of 1:25,
plasma samples were prepared in a 4-fold serial dilution for a total of 6 dilutions. Cells were
incubated with 50 µl/well of diluted plasma samples and incubated for 30 min on ice.
|
After
incubation, cells were washed once with 100 µl/well of FACS buffer and stained in 50 µl/well
of 1:160 dilution of BV421 IgG and 1:100 dilution of APC IgM on ice for 30 min. After staining,
cells were washed once with 100 µl/well of FACS buffer and analyzed on a FACS BD Aria
Fusion. Evaluations were performed using FlowJo10 software. Geometric Mean values of all
cells/single cells/mCherry positive (transfected cells) in APC-A channel (for IgM) and BV421-
A channel (for IgG) were determined and graphically displayed using GraphPad Prism. SARS-CoV-2 and HCoV S protein immunoassays
ELISA plates (Greiner Bio-One 655092) were coated with 2 µg/ml protein (spike ectodomains
of SARS-CoV-2, HKU1, OC43) for IgG measurement or 5 µg/ml (spike ectodomain of SARS-
CoV-2) for IgM and IgA measurements in PBS at 4 °C overnight. Plates were blocked with
blocking buffer (BB) consisting of PBS supplemented with 5 % nonfat dried milk powder (Carl
Roth T145.2) for 60 min at RT. Monoclonal antibodies were tested with a starting concentration
of 10 µg/ml in PBS, polyclonal IgGs with 500 µg/ml in BB and plasma samples with a starting
17
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. dilution of 1:20 for IgG detection and 1:10 for IgM and IgA detection in BB. Monoclonal
antibodies were serially diluted 1:5. Polyclonal IgGs and plasma were diluted in 1:4 serial
dilutions. After incubation with samples for 90 min at RT, plates were incubated with anti-
human IgG-HRP (Southern Biotech 2040-05) diluted 1:2500 in BB, anti-human IgM-HRP
(Thermo Fisher Scientific A18835) diluted 1:2000 in BB or anti-human IgA-HRP (Thermo
Fisher Scientific A18781) 1:2000 in BB. Plates were developed using ABTS solution (Thermo
Fisher Scientific 002024) and absorbance was measured at 415 nm and 695 nm by a
Microplate Reader (Tecan). Anti-SARS-CoV-2 S1/S2 IgG and IgM antibody titers of plasma samples were also assessed
using the automated DiaSorin’s LIAISON® SARS-CoV-2 S1/S2 protein ELISA kit according to
the manufacturer’s instructions. IgG and IgM result values were interpreted with the following
cut-off values: negative < 12.0 AU/ml, equivocal > 12.0 - < 15.0 AU/ml, and positive > 15.0
AU/ml. Anti-SARS-CoV-2 S1 IgA antibody titers of plasma samples were also measured using
the Euroimmun anti-SARS-CoV-2 ELISA on the automated system Euroimmun Analyzer I and
S/CO values interpreted with following cut-off values: negative S/CO < 0.8, equivocal S/CO >
0.8 - < 1.1, positive S/CO > 1.1. SARS-CoV-2 pseudovirus neutralization assays
SARS-CoV-2 pseudovirus expressing the Wu01 spike (EPI_ISL_40671) was generated by co-
transfection of individual plasmids encoding HIV Tat, HIV Gag/Pol, HIV Rev, luciferase
followed by an IRES and ZsGreen, and the SARS-CoV-2 spike protein into HEK 293T cells
using the FuGENE 6 Transfection Reagent (Promega).
|
Cell culture supernatants containing
pseudovirus particles were harvested and stored at -80°C till use. The pseudovirus was titrated
by infecting HEK293T cells expressing human ACE2 (Crawford et al., 2020). Following a 48-
hour incubation at 37°C and 5% CO2, luciferase activity was determined by addition of
luciferin/lysis buffer (10 mM MgCl2, 0.3 mM ATP, 0.5 mM Coenzyme A, 17 mM IGEPAL (all
Sigma-Aldrich), and 1 mM D-Luciferin (GoldBio) in Tris-HCL) using a microplate reader
(Berthold). For neutralization assays, a virus dilution with a relative luminescence unit (RLU)
of approximately 1000-fold in infected cells versus non-infected cells was selected. For testing neutralization at a single dilution, polyclonal IgG samples at a concentration of 1000
µg/ml, plasma samples at a dilution of 1:10, or mAbs at a concentration of 50 µg/ml, were co-
incubated with pseudovirus supernatants for 1 h at 37°C, following which 293T-ACE-2 cells
were added. After a 48 h incubation at 37 °C and 5 % CO2, luciferase activity was determined
using the luciferin/lysis buffer. After subtracting background RLUs of non-infected cells, % of
neutralization was calculated and the mean value was used for reporting. Each sample was
tested in duplicates. 18
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. To determine IC50 values for mAbs a dilution series of the antibody was performed starting with
50 µg/ml. IC50 values were calculated as the antibody concentration causing a 50 % reduction
in signal compared the virus-only controls using a dose-response curve in GraphPad Prism. Authentic virus neutralization assays
Authentic virus neutralization was tested using a virus previously grown out from an oro-/naso-
pharyngeal swab using VeroE6 cells (Vanshylla et al., 2021). For testing neutralization, plasma
samples at a single dilution of 1:10 were co-incubated with the 200 TCID50 virus for 1 h at 37°C,
following which VeroE6 cells were added. After 4 days, cytopathic effects (CPE) were analysed
under a bright-field microscope and neutralization was determined as the absence of CPE. Cells without any virus served as reference for lack of CPE and cells with virus only served as
reference for positive CPE. QUANTIFICATION AND STATISTICAL ANALYSIS
Flow cytometry analyses and quantification were performed using FlowJo10 software. Statistical tests and analyses were done with GraphPad Prism (v7 and v8), Python (v3.6.8), R
(v4.0.0) and Mircosoft Excel for Mac (v14.7.3 and 16.4.8). CDRH3 lengths, V gene usage and
germline identity distributions for clonal sequences (Figure 3C) were assessed for all input
sequences without further collapsing. A two-tailed unpaired t test (Prism, GraphPad) was
performed to test for the frequency differences of SARS-CoV-2-reactive IgG+ and IgG- B cells
between pre-pandemic and convalescent blood samples (Figure 3A).
|
KEY RESOURCES TABLE
Supplementary Figure Legends
Figure S1: Binding of plasma samples and polyclonal IgG, related to Figure 2
Binding of plasma IgG, IgM and IgA to SARS-CoV-2 S protein determined by serial
dilution (A) or automated ELISAs (B) (DiaSorin’s LIAISON® and Euroimmun Analyzer). (C) Binding of purified, polyclonal IgGs (pIgG) to SARS-CoV-2, HKU1 and OC43 S
proteins determined by serial dilution ELISA. ELISAs were performed in duplicate
experiments. Circles depict means and error bars indicate standard deviation. IgG titers
19
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. were interpreted as negative < 12.0 AU/ml, equivocal > 12.0 - < 15.0 AU/ml, and positive >
15.0 AU/ml. IgM titers were interpreted as negative < 1.1 IgM index and positive > 1.1. IgA
values interpreted with following cut-off values: negative < 0.8 IgA ratio, equivocal > 0.8 - < 1.1
IgA ratio, positive > 1.1 IgA ratio. Figure S2: Neutralization activity of plasma and polyclonal IgGs against SARS-
CoV-2, related to Figure 2
Neutralization activity of plasma samples (A) and polyclonal IgGs (B) against SARS-
CoV-2 wildtype (WT) and/or pseudovirus (PSV). Samples were tested in duplicate
experiments (wildtype) or in duplicates in a single experiment (pseudovirus). The
SARS-CoV-2 neutralizing antibody C6 was used as a positive control. Bars and circles
of graph plots show means and error bars indicate standard deviation. Figure S3: Gating strategy and single cell sorts of SARS-CoV-2-reactive B cell
subsets, related to Figure 3
(A) FACS plots illustrating the gating strategy for single cell sorts of SARS-CoV-2-
reactive, IgG+ and IgG- B cells. (B) Individual FACS plots depicting sorting gates and
frequencies of SARS-CoV-2-reactive, IgG+ and IgG- B cells from 40 donors. Figure S4: Binding of monoclonal antibodies isolated from pre-pandemic blood
samples, related to Figure 4
(A) Binding of monoclonal antibodies to SARS-CoV-2, HKU-1 and OC43 S proteins
determined by serial dilution ELISA. ELISAs were performed in duplicate experiments. Circles depict means and error bars indicate standard deviation. Supplementary Tables
Table S1: Demographical characteristics of blood donors, related to Figure 2
Table S2: Sequence information of pre-pandemic and original heavy and light
chains, related to Figure 1
References
20
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved.
|
No reuse allowed without permission. de Alwis, R., Williams, K.L., Schmid, M.A., Lai, C.-Y., Patel, B., Smith, S.A., Crowe, J.E., Wang,
W.-K., Harris, E., and de Silva, A.M. (2014). Dengue viruses are enhanced by distinct
populations of serotype cross-reactive antibodies in human immune sera. PLoS Pathog. 10,
e1004386. Anderson, E.M., Goodwin, E.C., Verma, A., Arevalo, C.P., Bolton, M.J., Weirick, M.E., Gouma,
S., McAllister, C.M., Christensen, S.R., Weaver, J., et al. (2021). Seasonal human coronavirus
antibodies are boosted upon SARS-CoV-2 infection but not associated with protection. Cell
184, 1858-1864.e10. Arevalo, C.P., Le Sage, V., Bolton, M.J., Eilola, T., Jones, J.E., Kormuth, K.A., Nturibi, E.,
Balmaseda, A., Gordon, A., Lakdawala, S.S., et al. (2020). Original antigenic sin priming of
influenza virus hemagglutinin stalk antibodies. Proc. Natl. Acad. Sci. 117, 17221–17227. Arvin, A.M., Fink, K., Schmid, M.A., Cathcart, A., Spreafico, R., Havenar-Daughton, C.,
Lanzavecchia, A., Corti, D., and Virgin, H.W. (2020). A perspective on potential antibody-
dependent enhancement of SARS-CoV-2. Nature 584, 353–363. Bacher, P., Rosati, E., Esser, D., Martini, G.R., Saggau, C., Schiminsky, E., Dargvainiene, J.,
Schröder, I., Wieters, I., Khodamoradi, Y., et al. (2020). Low-Avidity CD4+ T Cell Responses
to SARS-CoV-2 in Unexposed Individuals and Humans with Severe COVID-19. Immunity 53,
1258-1271.e5. Barnes, C.O., West, A.P., Huey-Tubman, K.E., Hoffmann, M.A.G., Sharaf, N.G., Hoffman,
P.R., Koranda, N., Gristick, H.B., Gaebler, C., Muecksch, F., et al. (2020a). Structures of
Human Antibodies Bound to SARS-CoV-2 Spike Reveal Common Epitopes and Recurrent
Features of Antibodies. Cell 182, 828-842.e16. Barnes, C.O., Jette, C.A., Abernathy, M.E., Dam, K.-M.A., Esswein, S.R., Gristick, H.B.,
Malyutin, A.G., Sharaf, N.G., Huey-Tubman, K.E., Lee, Y.E., et al. (2020b). SARS-CoV-2
neutralizing antibody structures inform therapeutic strategies. Nature 588, 682–687. Bertoglio, F., Meier, D., Langreder, N., Steinke, S., Rand, U., Simonelli, L., Heine, P.A.,
Ballmann, R., Schneider, K.-T., Roth, K.D.R., et al. (2021). SARS-CoV-2 neutralizing human
recombinant antibodies selected from pre-pandemic healthy donors binding at RBD-ACE2
interface. Nat. Commun. 12, 1577.
von Boehmer, L., Liu, C., Ackerman, S., Gitlin, A.D., Wang, Q., Gazumyan, A., and
Nussenzweig, M.C. (2016). Sequencing and cloning of antigen-specific antibodies from mouse
memory B cells. Nat. Protoc. 11, 1908–1923. 21
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Braun, J., Loyal, L., Frentsch, M., Wendisch, D., Georg, P., Kurth, F., Hippenstiel, S.,
Dingeldey, M., Kruse, B., Fauchere, F., et al.
|
(2020). SARS-CoV-2-reactive T cells in healthy
donors and patients with COVID-19. Nature 587, 270–274. Brouwer, P.J.M., Caniels, T.G., van der Straten, K., Snitselaar, J.L., Aldon, Y., Bangaru, S.,
Torres, J.L., Okba, N.M.A., Claireaux, M., Kerster, G., et al. (2020). Potent neutralizing
antibodies from COVID-19 patients define multiple targets of vulnerability. Science 369, 643–
650. Cao, Y., Su, B., Guo, X., Sun, W., Deng, Y., Bao, L., Zhu, Q., Zhang, X., Zheng, Y., Geng, C.,
et al. (2020). Potent Neutralizing Antibodies against SARS-CoV-2 Identified by High-
Throughput Single-Cell Sequencing of Convalescent Patients’ B Cells. Cell 182, 73-84.e16. Corman, V.M., Müller, M.A., Costabel, U., Timm, J., Binger, T., Meyer, B., Kreher, P., Lattwein,
E., Eschbach-Bludau, M., Nitsche, A., et al. (2012). Assays for laboratory confirmation of novel
human coronavirus (hCoV-EMC) infections. Euro Surveill. Bull. Eur. Sur Mal. Transm. Eur. Commun. Dis. Bull. 17. Crawford, K.H.D., Eguia, R., Dingens, A.S., Loes, A.N., Malone, K.D., Wolf, C.R., Chu, H.Y.,
Tortorici, M.A., Veesler, D., Murphy, M., et al. (2020). Protocol and Reagents for Pseudotyping
Lentiviral Particles with SARS-CoV-2 Spike Protein for Neutralization Assays. Viruses 12. Echeverría, G., Guevara, Á., Coloma, J., Ruiz, A.M., Vasquez, M.M., Tejera, E., and de Waard,
J.H. (2021). Pre-existing T-cell immunity to SARS-CoV-2 in unexposed healthy controls in
Ecuador, as detected with a COVID-19 Interferon-Gamma Release Assay. Int. J. Infect. Dis. 105, 21–25. Ehrhardt, S.A., Zehner, M., Krähling, V., Cohen-Dvashi, H., Kreer, C., Elad, N., Gruell, H.,
Ercanoglu, M.S., Schommers, P., Gieselmann, L., et al. (2019). Polyclonal and convergent
antibody response to Ebola virus vaccine rVSV-ZEBOV. Nat. Med. 25, 1589–1600. Gieselmann, L., Kreer, C., Ercanoglu, M.S., Lehnen, N., Zehner, M., Schommers, P., Potthoff,
J., Gruell, H., and Klein, F. (2021). Effective high-throughput isolation of fully human antibodies
targeting infectious pathogens. Nat. Protoc. 16, 3639–3671. Gombar, S., Bergquist, T., Pejaver, V., Hammarlund, N.E., Murugesan, K., Mooney, S., Shah,
N., Pinsky, B.A., and Banaei, N. (2021). SARS-CoV-2 infection and COVID-19 severity in
individuals with prior seasonal coronavirus infection. Diagn. Microbiol. Infect. Dis. 100, 115338. Grifoni, A., Weiskopf, D., Ramirez, S.I., Mateus, J., Dan, J.M., Moderbacher, C.R., Rawlings,
S.A., Sutherland, A., Premkumar, L., Jadi, R.S., et al. (2020). Targets of T Cell Responses to
22
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. SARS-CoV-2 Coronavirus in Humans with COVID-19 Disease and Unexposed Individuals. Cell 181, 1489-1501.e15. Hurlburt, N.K., Seydoux, E., Wan, Y.-H., Edara, V.V., Stuart, A.B., Feng, J., Suthar, M.S.,
McGuire, A.T., Stamatatos, L., and Pancera, M. (2020).
|
Structural basis for potent
neutralization of SARS-CoV-2 and role of antibody affinity maturation. Nat. Commun. 11, 5413. Imkeller, K., and Wardemann, H. (2018). Assessing human B cell repertoire diversity and
convergence. Immunol. Rev. 284, 51–66. Ju, B., Zhang, Q., Ge, J., Wang, R., Sun, J., Ge, X., Yu, J., Shan, S., Zhou, B., Song, S., et al. (2020). Human neutralizing antibodies elicited by SARS-CoV-2 infection. Nature 584, 115–
119. Katzelnick, L.C., Gresh, L., Halloran, M.E., Mercado, J.C., Kuan, G., Gordon, A., Balmaseda,
A., and Harris, E. (2017). Antibody-dependent enhancement of severe dengue disease in
humans. Science 358, 929–932. Khurana, S., Loving, C.L., Manischewitz, J., King, L.R., Gauger, P.C., Henningson, J., Vincent,
A.L., and Golding, H. (2013). Vaccine-induced anti-HA2 antibodies promote virus fusion and
enhance influenza virus respiratory disease. Sci. Transl. Med. 5, 200ra114. Kowarz, E., Löscher, D., and Marschalek, R. (2015). Optimized Sleeping Beauty transposons
rapidly generate stable transgenic cell lines. Biotechnol. J. 10, 647–653. Kreer, C., Zehner, M., Weber, T., Ercanoglu, M.S., Gieselmann, L., Rohde, C., Halwe, S.,
Korenkov, M., Schommers, P., Vanshylla, K., et al. (2020a). Longitudinal Isolation of Potent
Near-Germline SARS-CoV-2-Neutralizing Antibodies from COVID-19 Patients. Cell 182, 843-
854.e12. Kreer, C., Döring, M., Lehnen, N., Ercanoglu, M.S., Gieselmann, L., Luca, D., Jain, K.,
Schommers, P., Pfeifer, N., and Klein, F. (2020b). openPrimeR for multiplex amplification of
highly diverse templates. J. Immunol. Methods 480, 112752. Kreye, J., Reincke, S.M., Kornau, H.-C., Sánchez-Sendin, E., Corman, V.M., Liu, H., Yuan, M.,
Wu, N.C., Zhu, X., Lee, C.-C.D., et al. (2020). A Therapeutic Non-self-reactive SARS-CoV-2
Antibody Protects from Lung Pathology in a COVID-19 Hamster Model. Cell 183, 1058-
1069.e19. 23
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Le Bert, N., Tan, A.T., Kunasegaran, K., Tham, C.Y.L., Hafezi, M., Chia, A., Chng, M.H.Y., Lin,
M., Tan, N., Linster, M., et al. (2020). SARS-CoV-2-specific T cell immunity in cases of COVID-
19 and SARS, and uninfected controls. Nature 584, 457–462. Lefranc, M.-P., Giudicelli, V., Ginestoux, C., Jabado-Michaloud, J., Folch, G., Bellahcene, F.,
Wu, Y., Gemrot, E., Brochet, X., Lane, J., et al. (2009). IMGT, the international
ImMunoGeneTics information system. Nucleic Acids Res. 37, D1006-1012. Linderman, S.L., and Hensley, S.E. (2016). Antibodies with “Original Antigenic Sin” Properties
Are Valuable Components of Secondary Immune Responses to Influenza Viruses. PLoS
Pathog. 12, e1005806. Mateus, J., Grifoni, A., Tarke, A., Sidney, J., Ramirez, S.I., Dan, J.M., Burger, Z.C., Rawlings,
S.A., Smith, D.M., Phillips, E., et al.
|
(2020). Selective and cross-reactive SARS-CoV-2 T cell
epitopes in unexposed humans. Science 370, 89–94. Midgley, C.M., Bajwa-Joseph, M., Vasanawathana, S., Limpitikul, W., Wills, B., Flanagan, A.,
Waiyaiya, E., Tran, H.B., Cowper, A.E., Chotiyarnwon, P., et al. (2011). An In-Depth Analysis
of Original Antigenic Sin in Dengue Virus Infection. J. Virol. 85, 410–421. Mongkolsapaya, J., Dejnirattisai, W., Xu, X., Vasanawathana, S., Tangthawornchaikul, N.,
Chairunsri, A., Sawasdivorn, S., Duangchinda, T., Dong, T., Rowland-Jones, S., et al. (2003). Original antigenic sin and apoptosis in the pathogenesis of dengue hemorrhagic fever. Nat. Med. 9, 921–927. Ng, K.W., Faulkner, N., Cornish, G.H., Rosa, A., Harvey, R., Hussain, S., Ulferts, R., Earl, C.,
Wrobel, A.G., Benton, D.J., et al. (2020). Preexisting and de novo humoral immunity to SARS-
CoV-2 in humans. Science 370, 1339–1343. Nguyen-Contant, P., Embong, A.K., Kanagaiah, P., Chaves, F.A., Yang, H., Branche, A.R.,
Topham, D.J., and Sangster, M.Y. (2020). S Protein-Reactive IgG and Memory B Cell
Production after Human SARS-CoV-2 Infection Includes Broad Reactivity to the S2 Subunit. MBio 11. Niu, X., Li, S., Li, P., Pan, W., Wang, Q., Feng, Y., Mo, X., Yan, Q., Ye, X., Luo, J., et al. (2020). Longitudinal Analysis of T and B Cell Receptor Repertoire Transcripts Reveal Dynamic
Immune Response in COVID-19 Patients. Front. Immunol. 11, 582010. Poston, D., Weisblum, Y., Wise, H., Templeton, K., Jenks, S., Hatziioannou, T., and Bieniasz,
P. (2020). Absence of Severe Acute Respiratory Syndrome Coronavirus 2 Neutralizing Activity
24
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. in Prepandemic Sera From Individuals With Recent Seasonal Coronavirus Infection. Clin. Infect. Dis. ciaa1803. Raybould, M.I.J., Kovaltsuk, A., Marks, C., and Deane, C.M. (2020). CoV-AbDab: the
Coronavirus Antibody Database. Bioinforma. Oxf. Engl. Robbiani, D.F., Gaebler, C., Muecksch, F., Lorenzi, J.C.C., Wang, Z., Cho, A., Agudelo, M.,
Barnes, C.O., Gazumyan, A., Finkin, S., et al. (2020). Convergent antibody responses to
SARS-CoV-2 in convalescent individuals. Nature 584, 437–442. Rogers, T.F., Zhao, F., Huang, D., Beutler, N., Burns, A., He, W., Limbo, O., Smith, C., Song,
G., Woehl, J., et al. (2020). Isolation of potent SARS-CoV-2 neutralizing antibodies and
protection from disease in a small animal model. Science 369, 956–963. Rothman, A.L. (2011). Immunity to dengue virus: a tale of original antigenic sin and tropical
cytokine storms. Nat. Rev. Immunol. 11, 532–543. Sagar, M., Reifler, K., Rossi, M., Miller, N.S., Sinha, P., White, L.F., and Mizgerd, J.P. (2021). Recent endemic coronavirus infection is associated with less-severe COVID-19.
|
J. Clin. Invest. 131, 143380. Schommers, P., Gruell, H., Abernathy, M.E., Tran, M.-K., Dingens, A.S., Gristick, H.B., Barnes,
C.O., Schoofs, T., Schlotz, M., Vanshylla, K., et al. (2020). Restriction of HIV-1 Escape by a
Highly Broad and Potent Neutralizing Antibody. Cell 180, 471-489.e22. Schultheiß, C., Paschold, L., Simnica, D., Mohme, M., Willscher, E., von Wenserski, L., Scholz,
R., Wieters, I., Dahlke, C., Tolosa, E., et al. (2020). Next-Generation Sequencing of T and B
Cell Receptor Repertoires from COVID-19 Patients Showed Signatures Associated with
Severity of Disease. Immunity 53, 442-455.e4. Sette, A., and Crotty, S. (2021). Adaptive immunity to SARS-CoV-2 and COVID-19. Cell 184,
861–880. Seydoux, E., Homad, L.J., MacCamy, A.J., Parks, K.R., Hurlburt, N.K., Jennewein, M.F., Akins,
N.R., Stuart, A.B., Wan, Y.-H., Feng, J., et al. (2020). Analysis of a SARS-CoV-2-Infected
Individual Reveals Development of Potent Neutralizing Antibodies with Limited Somatic
Mutation. Immunity 53, 98-105.e5. Shi, R., Shan, C., Duan, X., Chen, Z., Liu, P., Song, J., Song, T., Bi, X., Han, C., Wu, L., et al. (2020). A human neutralizing antibody targets the receptor-binding site of SARS-CoV-2. Nature 584, 120–124. 25
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Shrock, E., Fujimura, E., Kula, T., Timms, R.T., Lee, I.-H., Leng, Y., Robinson, M.L., Sie, B.M.,
Li, M.Z., Chen, Y., et al. (2020). Viral epitope profiling of COVID-19 patients reveals cross-
reactivity and correlates of severity. Science 370, eabd4250. Song, G., He, W., Callaghan, S., Anzanello, F., Huang, D., Ricketts, J., Torres, J.L., Beutler,
N., Peng, L., Vargas, S., et al. (2021). Cross-reactive serum and memory B-cell responses to
spike protein in SARS-CoV-2 and endemic coronavirus infection. Nat. Commun. 12, 2938. Stadlbauer, D., Amanat, F., Chromikova, V., Jiang, K., Strohmeier, S., Arunkumar, G.A., Tan,
J., Bhavsar, D., Capuano, C., Kirkpatrick, E., et al. (2020). SARS-CoV-2 Seroconversion in
Humans: A Detailed Protocol for a Serological Assay, Antigen Production, and Test Setup. Curr. Protoc. Microbiol. 57. St-Jean, J.R., Jacomy, H., Desforges, M., Vabret, A., Freymuth, F., and Talbot, P.J. (2004). Human Respiratory Coronavirus OC43: Genetic Stability and Neuroinvasion. J. Virol. 78,
8824–8834. Tiller, T., Meffre, E., Yurasov, S., Tsuiji, M., Nussenzweig, M.C., and Wardemann, H. (2008). Efficient generation of monoclonal antibodies from single human B cells by single cell RT-PCR
and expression vector cloning. J. Immunol. Methods 329, 112–124. Vanshylla, K., Di Cristanziano, V., Kleipass, F., Dewald, F., Schommers, P., Gieselmann, L.,
Gruell, H., Schlotz, M., Ercanoglu, M.S., Stumpf, R., et al.
|
(2021). Kinetics and correlates of
the neutralizing antibody response to SARS-CoV-2 infection in humans. Cell Host Microbe 29,
917-929.e4. Vatti, A., Monsalve, D.M., Pacheco, Y., Chang, C., Anaya, J.-M., and Gershwin, M.E. (2017). Original antigenic sin: A comprehensive review. J. Autoimmun. 83, 12–21. Weiskopf, D., Schmitz, K.S., Raadsen, M.P., Grifoni, A., Okba, N.M.A., Endeman, H., van den
Akker, J.P.C., Molenkamp, R., Koopmans, M.P.G., van Gorp, E.C.M., et al. (2020). Phenotype
and kinetics of SARS-CoV-2-specific T cells in COVID-19 patients with acute respiratory
distress syndrome. Sci. Immunol. 5, eabd2071. Woo, P.C.Y., Lau, S.K.P., Chu, C., Chan, K., Tsoi, H., Huang, Y., Wong, B.H.L., Poon, R.W.S.,
Cai, J.J., Luk, W., et al. (2005). Characterization and Complete Genome Sequence of a Novel
Coronavirus, Coronavirus HKU1, from Patients with Pneumonia. J. Virol. 79, 884–895. Wrapp, D., Wang, N., Corbett, K.S., Goldsmith, J.A., Hsieh, C.-L., Abiona, O., Graham, B.S.,
and McLellan, J.S. (2020). Cryo-EM structure of the 2019-nCoV spike in the prefusion
conformation. Science 367, 1260–1263. 26
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Wu, Y., Wang, F., Shen, C., Peng, W., Li, D., Zhao, C., Li, Z., Li, S., Bi, Y., Yang, Y., et al. (2020). A noncompeting pair of human neutralizing antibodies block COVID-19 virus binding
to its receptor ACE2. Science 368, 1274–1278. Ye, J., Ma, N., Madden, T.L., and Ostell, J.M. (2013). IgBLAST: an immunoglobulin variable
domain sequence analysis tool. Nucleic Acids Res. 41, W34-40. Yuan, M., Liu, H., Wu, N.C., Lee, C.-C.D., Zhu, X., Zhao, F., Huang, D., Yu, W., Hua, Y., Tien,
H., et al. (2020). Structural basis of a shared antibody response to SARS-CoV-2. Science 369,
1119–1123. Zhang, A., Stacey, H.D., Mullarkey, C.E., and Miller, M.S. (2019). Original Antigenic Sin: How
First Exposure Shapes Lifelong Anti–Influenza Virus Immune Responses. J. Immunol. 202,
335–340. Zost, S.J., Gilchuk, P., Chen, R.E., Case, J.B., Reidy, J.X., Trivette, A., Nargi, R.S., Sutton,
R.E., Suryadevara, N., Chen, E.C., et al. (2020). Rapid isolation and profiling of a diverse panel
of human monoclonal antibodies targeting the SARS-CoV-2 spike protein. Nat. Med. 26, 1422–
1427. 27
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A
Pre-pandemic healthy
Naive BCR repertoire
Alignment
Similiar sequences
Chimeras
) d e r i a p n u (
HC
LC
Naive BCR S-reactive BCR Alignment
VH
CDRH3
JH
Naive BCR chains
23 pre-pandemic HCs
SARS-CoV-2+
Sequences of S-reactive Abs
HC
Naive BCR S-reactive BCR Alignment
VL
CDRL3
JL
31
LC
aa match
aa miss-match
Sequences of S-reactive Abs
8 pre-pandemic LCs
B
SARS-CoV-2- reactive antibody
Original heavy chain Pre-pandemic light chain
Pre-pandemic heavy chain Original light chain
ELISA
PsV-NA
ELISA
PsV-NA
CnC2t1p1_B4
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
LC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
LC1
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
HC1 HC2
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
HC1 HC2
HbnC3t1p1_G4
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
LC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
LC1
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
HC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
HC1
HbnC3t1p2_B10
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
LC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
LC1
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
HC1 HC2 HC3 HC4 HC5
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
HC1 HC2 HC3 HC4 HC5
HbnC4t1p1_A1
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
LC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
LC1
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
HC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
HC1
HbnC4t1p1_C11
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
LC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
LC1
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
HC1
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
HC1
MnC2t1p1_C12
d e z
i l
a m r o N
m n 5 9 6 -
m n 5 1 4 D O
1.0 0.8 0.6 0.4 0.2 0.0
10-4
10-3
LC1 LC2 LC3
10-2
10-1
0 10
1 10
)
%
( n o i t a z
i l
a r t u e N
100 80 60 40 20 0
10-2
10-1
LC1 LC2 LC3
0 10
1.0 0.8 0.6 0.4 0.2 0.0
m n 5 9 6 -
d e z
i l
a m r o N
m n 5 1 4 D O
1 10 Concentration ( g/ml)
2 10
10-4
10-3
HC1-13 HC6 HC13 HC14
10-2
10-1
0 10
1 10
)
%
( n o
i t
a z
i l
a r t u e N
100 80 60 40 20 0
10-2
10-1
HC1-13
0 10
1 10
2 10
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021.
|
The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Figure 1: Pre-pandemic naïve B cells express SARS-CoV-2 reactive heavy or light chains (A) Selection of heavy and light chain sequences in a pre-pandemic naive B cell repertoire that closely resemble heavy and light chains of SARS-CoV-2 reactive antibodies (Abs). (B) Binding and neutralization plots against SARS-CoV-2 of 6 previously described SARS-CoV-2-reactive antibodies (red) and 31 chimeric antibodies composed of heavy or light chain of the original SARS-CoV-2-reactive antibodies paired with a NGS-derived heavy (darker grey) or light (grey) chain. Activity of mAbs against SARS-CoV-2 was determined by serial dilution ELISA and pseudotyped neutralization assays (PsV-NA). Each antibody was produced in at least duplicates. ELISAs and neutralization assays were performed with biological replicates. Symbols depict means and error bars indicate standard deviation. bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A
n=150
s e p m a s
l
f o . o N
20
15
10
5
0
g u A
t p e S
2019
t c O
v o N
c e D
9 1 - D V O C
I
k a e r b t u o
0 2
n a J
s e p m a s
l
f o
. o N
20
15
10
5
0
0
20
40
Age (years)
60
80
150
Male (49.3%) Female (50.7%)
B
com. IA
in house ELISA
SARS-CoV-2 Cell assay
SARS-CoV-2
g n d n B
i
i
s y a s s a
Anti- IgG IgM IgA
Anti-IgG
Anti-IgM
n o
i t
a z
i l
a r t
u e N
s y a s s a
1.Screening 2.IC50
CPE
S1/S2-CoV-2
S1/S2-CoV-2
HKU1 OC43
Pseudovirus
Wildtype
Plasma
pIgG
C
Immunoassays
Cell assay (CA)
Neutralization assays
Plasma
pIgG
Plasma
Plasma
pIgG
IgG IgM
IgA
IgG IgM
IgA
IgG IgM
IgA
A
A
A
A
A
A
A
A
A
I
A S I L E
. m o c
I
. m o c
A S I L E
I
. m o c
A S I L E
I
. m o c
A S I L E
pos neg
I
. m o c
A S I L E
I
A S I L E
. m o c
I
. m o c
A S L E
I
I
. m o c
A S L E
I
I
. m o c
A S L E
I
2 S / 1 S
1 U K H
3 4 C O
2 1 S U / K 1 H S pos neg
3 4 C O
2 S / 1 S
1 U K H
3 4 C O
G g I
M g I
M G g g I I pos neg
G g I
M g I
V S P
T W
V T S W P pos neg
V S P
T W
V S P
V S P pos neg
V S P
001 - 050 (1) 051 - 100 (2) 101 - 150 (3)
(1)
(2)
(3)
(1)
(2)
(3)
(1)
(2)
(3)
(1)(2)(3)
ELISA AUC
com. IA IgG AU
com. IA IgM AU
com. IA IgA AU
pIgG AUC
Cell assay (CA) AUC
Neutralization PSV Wu_01 (%)
WT
0 - 0.01 0.01 - 0.02 0.02 - 0.03 0.03 - 0.04 0.04 - 0.10
0 - 12 12 - 14 14 - 16 16 - 20 20 - 25
0 - 1.1 1.1 - 1.5 1.5 - 2.0 2.0 - 2.5 2.5 - 3.0
0 - 0.8 0.8 - 1.1 1.1 - 1.5 1.5 - 2.0 2.0 - 2.6
0 - 50 50 - 200 200 - 400 400 - 600 600 - 800
0 - 1,000 1,000 - 2,000 2,000 - 3,000 3,000 - 4,000 4,000 - 7,000
< 50 50 - 60 60 - 80 80 - 100 > 50, but no neutralization in serial dilution assays
no yes
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder.
|
All rights reserved. No reuse allowed without permission. Figure 2: Screening of pre-pandemic samples from 150 adults reveal no relevant reactivity against SARS-CoV-2 (A) Timeline of blood collections and demographic characteristics of 150 donors sampled before the SARS-CoV-2 pandemic. Pre-pandemic blood samples were collected as buffy coats between August and November of 2019. Gender and age distribu- tion of donors are illustrated as pie chart and bar plots. (B) Pre-pandemic plasma samples and purified IgGs (pIgG) were tested for binding and neutralization using different experi- mental approaches. Plasma IgG, IgM and IgA as well as pIgG were tested for binding to SARS-CoV-2, HKU1 and OC43 S proteins using in house (ELISA) and commercially available immunoassays (com. IA). Binding of plasma IgG and IgM to cell surface expressed SARS-CoV-2 S protein was also determined by FACS analysis (CA). Neutralization activity was determined against SARS-CoV-2 pseudo- and wild-type virus (PSV and WT). (C) Heatmap visualizing binding (AUC or AU) and neutralization activity (% or CPE) of pre-pandemic plasma samples and pIgG against SARS-CoV-2 and endemic HCoVs (HKU1 and OC43) S proteins or SARS-CoV-2 wildtype (WT) and pseudovirus (PSV Wu_01), respectively (see also Figure S1 and 2). Immunoassays were performed in duplicate experiments and the AUC is presented as geometric mean of duplicates. Neutralization activity was first determined for single sample concentrations. Samples that displayed neutralization activity of ≥ 50% (x) were repeatedly investigated in serial dilutions (see also Figure S2). Samples were tested in duplicates and in single experiments. The average of neutralization is presented and each row represents one donor. bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A
HbnC1
0.03
0.016
HbnC2
0.03
0.03
HbnC3
0.03
0.07
0.30
****
****
)
9 1 - D V O C
I
c m e d n a p e r P
i
s e p m a s
l
s e p m a s
l
n=23
n=40
n e
i
t
o r p S 2 - V o C S R A S
Pre040
0.005 0.04
IgG
0.0003 0.04
Pre041
0.008 0.058
0.0008 0.053
Pre042
0.007 0.035
0.0001 0.017
%
( s
l l
e c B + S 2 - V o C S R A S
0.20
0.15
0.10
0.05
0.00
IgG +
IgG -
Subset
Pre-pandemic (n=40) COVID-19 (n=23)
Analysis gate
Sort gate
B
024
025
026
027
028
029
030
031
clonal
66
126
107
99
128
206
141
259
IgG
5,432
032
282
033
213
034
136
035
75
036
131
037
127
038
135
039
212
IgM
non- clonal
040
041
042
043
044
045
046
047
30
15
048
135
049
224
050
159
051
128
052
122
053
262
054
124
055
123
t n u o C
20
10
10
5
116
115
101
260
79
245
129
120
0
Number of clones
0 Clone size
056
057
058
059
060
061
E05
E06
45
98
95
6
34
53
116
100
C
)
%
20
10
Memory IgG reference IgG
t n u o C
20 10 0 10 20
( y c n e u q e r F
0
10
20
30
VH
IgM Naive IgM reference
2 - 1
3 - 1
8 - 1
8 1 - 1
4 2 - 1
5 4 - 1
6 4 - 1
8 5 - 1
9 6 - 1
2 - 9 6 - 1
5 - 2
6 2 - 2
0 7 - 2
7 - 3
9 - 3
1 1 - 3
3 1 - 3
5 1 - 3
0 2 - 3
1 2 - 3
3 2 - 3
0 3 - 3
3 - 0 3 - 3
3 3 - 3
3 4 - 3
D 3 4 - 3
8 4 - 3
9 4 - 3
3 5 - 3
4 6 - 3
D 4 6 - 3
6 6 - 3
2 7 - 3
3 7 - 3
4 7 - 3
4 - 4
8 2 - 4
2 - 0 3 - 4
4 - 0 3 - 4
1 3 - 4
4 3 - 4
2 - 8 3 - 4
9 3 - 4
9 5 - 4
1 6 - 4
1 - 0 1 - 5
1 5 - 5
1 - 6
1 - 4 - 7
0 5 10 15 20 25 30 35 CDRH3 length (AA)
t n u o C
20 0 20 40 60
75 80 85 90 95 100 VH gene germline identity (%)
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021.
|
The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Figure 3: Lack of SARS-CoV-2-reactive B cells and distinct antibody sequence features in pre-pandemic samples (A) Representative dot plots of SARS-CoV-2-reactive, CD19+CD20+, IgG+ and IgG- B cells of COVID-19 samples compared to pre-pandemic samples. Depicted numbers indicate frequencies of S protein reactive B cells (see also Figure S3). Red colored gate indicates gating strategy for analysis and doted gate indicates actual sorting gate. Dot plot bar graph displays the mean ± SD frequency of SARS-CoV-2-reactive, IgG+ and IgG- B cells in 40 pre-pandemic and 23 COVID-19 samples (*P < 0.05, **P < 0.01, ***P < 0.001 and ****P < 0.0001; unpaired two-tailed t test). (B) Clonal relationship of heavy chain sequences amplified from single SARS-CoV-2-reactive IgG+ and IgG- B cells isolated from 40 donors sampled before the pandemic. Individual clones are colored in shades of blue, grey and white. In the center of each pie chart, numbers of productive heavy chain sequences are illustrated. Presentation of clone sizes are proportional to the total number of productive heavy chain sequences per clone. (C) VH gene distribution, VH gene germline identity and CDRH3 length distribution in amino acids (AA) were separately determined for IgG and IgM. Distributions were calculated per individual. Bar and line plots show mean ± SD. bioRxiv preprint
doi:
https://doi.org/10.1101/2021.09.08.459398
;
this version posted September 8, 2021. The copyright holder for this
preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A
Selection criteria
Similarity (n = 18)
CoVAb Dab
identical VH/JH genes similar CDRH3 lengths similar CDRH3 sequence
Random (n = 182)
1 - 3 clonals
3 - 6 mAbs
1 - 3 singles
n=36
n=7,829
n=5,223
n=200
n=158
B
C
pos neg
024
025
026
2 S / 1 S
1 U K H
3 4 C O
034
035
036
037
2 S / 1 S
1 U K H
ELISA
3 4 C O
042
043
044
045
2 S / 1 S
1 U K H
3 4 C O
051
052
053
2 S / 1 S
1 U K H
3 4 C O
)
%
(
n o i t a z
i l
a r t u e N
100 80 60 40 20 0 -20 -40
pos. ctrl. Pre024-034
Pre035-042
027
028
029
030
031
032
033
034
AUC
0 - 1
038
039
040
041
042
1 - 2
046
047
048
049
050
051
2 - 4
4 - 8
054
055
056
057
058
E06
8 - 12
)
%
(
n o
i t
a z
i l
a r t u e N
100 80 60 40 20 0 -20 -40
0.1
Pre043-051
1
10
100 mAb concentration (µg/ml)
0.1
Pre052-E06
1
10
100
Figure 4: Monoclonal antibodies isolated from pre-pandemic samples are not reactive to SARS-CoV-2 and endemic HCoVs (A) Illustration demonstrating the sequence selection for downstream antibody production. Sequences were selected for antibody production based on similarity to antibody sequences deposited in the CoVAbDab and on random selection. From 7,829 SARS-CoV-2 reactive IgG+ and IgG- B cells, 5,223 productive heavy chain sequences were amplified. For antibody production, 18 HC sequences were selected based on similarity selection and 182 HC sequences were selected based on random selection.
|
In total 158 antibodies were produced. (B) Heat map visualizing binding (AUC) and neutralization activity (%) of monoclonal antibodies isolated from pre-pandemic blood samples against SARS-CoV-2 and endemic HCoVs (HKU1 and OC43) S proteins or SARS-CoV-2 pseudovirus (PSV Wu_01), respectively (see also Figure S4). Each row represents one monoclonal antibody. ELISAs were performed in duplicate experi- ments and the AUC is presented as geometric mean of duplicates. Neutralization activity was determined for single sample concentrations. Samples were tested in duplicates and the average of neutralization is presented. (C) Neutralization activity against SARS-CoV-2 pseudovirus (PSV Wu_01) was verified for all mAbs in serial dilutions.
|
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. A new platform for high-throughput therapy testing on iPSC-derived, immature airway from Cystic Fibrosis Patients
Jia Xin Jiang1, Leigh Wellhauser1, Onofrio Laselva1,2, Irina Utkina1,3, Zoltan Bozoky1, Tarini Gunawardena1, Zoe Ngan4, Sunny Xia1, Paul D.W. Eckford1, Felix Ratjen5,6, Theo J. Moraes5,6, John Parkinson1,3,8,7, Amy P. Wong4 and *Christine E. Bear1,8,9
1Programme in Molecular Medicine, Hospital for Sick Children, Toronto, Canada
2Department of Medical and Surgical Sciences, University of Foggia, Foggia, Italy
3Department of Molecular Genetics, University of Toronto, Toronto, Canada
4Programme in Developmental & Stem Cell Biology, Hospital for Sick Children, Toronto, Canada
5Programme in Translational Medicine, Hospital for Sick Children, Toronto, Canada
6Department of Paediatrics, University of Toronto, Toronto, Canada
7Department of Computer Science, University of Toronto, Toronto, Canada
8Department of Biochemistry, University of Toronto, Toronto, Canada
9Department of Physiology, University of Toronto, Toronto, Canada
To whom correspondence should be addressed: Christine E. Bear, Molecular Medicine, Research Institute, Hospital for Sick Children Peter Gilgan Centre for Research and Learning, 686 Bay St., Room 20.9420, Toronto, ON, M5G 0A4, Canada; phone: 1-416-816-5981; email:[email protected]
1
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. One Sentence Summary:
We describe a fluorescence-based platform that enables high-throughput Cystic Fibrosis therapy testing using iPSCs differentiated to immature lung. Abstract: Induced pluripotent, stem cell (iPSC)-derived models of airway tissue have successfully modeled the primary defect in regulated chloride conductance caused by the major Cystic Fibrosis causing mutation, F508del. However, it remains unclear if iPSC-derived airway cultures can be used in high-throughput therapy development for F508del and rarer mutations. There is an urgent need for airway tissue models that reflect the variability of patient-specific responses and are scalable for therapy development. In the current work, we describe a robust, high-throughput fluorescence assay of mutant CFTR function in iPSCs differentiated to immature airway epithelium. This assay measures reproducible functional responses to modulators targeting either the major CF mutant F508del or the nonsense mutant: W1282X-CFTR. We show that the ranking of patient-specific responses to interventions in this stem-cell based model recapitulates the ranking observed in primary nasal epithelial cultures obtained from the same individuals.
|
In summary, these proof-of-concept studies show that this novel platform has the potential to support therapy development and precision medicine for Cystic Fibrosis patients. Main text:
INTRODUCTION
Mutations in the CFTR gene can result in the disease called Cystic Fibrosis (CF). The major mutation, called F508del leads to CFTR protein misassembly, misprocessing, mistrafficking and altered function as a phosphorylation regulated chloride channel (1-3). Combinations of small molecule modulators called correctors, that rescue misassembly, together with potentiators that increase the probability of channel opening have been shown to be effective in improving lung function in individuals harboring the major mutation, F508del (4-7). However, not all individuals with this mutation show the same level of clinical improvement, supporting the ongoing need for novel therapy development. In addition, there are multiple, rarer CF-causing mutations for which there are no approved therapies. For example, W1282X causes the loss of CFTR mRNA expression due to the triggering of nonsense mediated decay (NMD), a mechanism thought to vary amongst individuals (8). The process of NMD hampers therapies promoting read-through of premature
2
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. stop codons and in the case of W1282X, protein modulators known to rescue the altered functional expression of the shorter protein (9-11). Therefore, there is an urgent need for novel therapy development for the treatment of Cystic Fibrosis. Tissue models that reflect the variability amongst patient-specific responses and are scalable to high-throughput formats would facilitate such therapy development. Traditionally, in-vitro studies of CFTR modulators, have employed bronchial epithelial tissue cultures from explanted CF lungs obtained at the time of transplant. Such “gold-standard” studies were, and continue to be, very instructive with respect to understanding the mechanism of modulator activity and evaluating their efficacy in relevant tissues. The functional consequences of modulator treatment are examined in direct measurements of ion conductance in the Ussing chamber apparatus (12-14). Although informative, these methods are low-throughput and not readily conducive to comparative studies of multiple interventions simultaneously. In addition, as the cell donor has undergone a lung transplant, the potential for directing individualized therapy is limited. Further, for rare mutations, for which currently no effective therapies are available, reduced availability of explanted tissue harboring the appropriate genotype impedes drug development. In lieu of bronchial epithelial cells, the development of systems for the culturing of the nasal epithelium and rectal organoids, greatly enhanced access to patient specific tissues harboring different disease-causing mutations (15-22).
|
For example, Ussing chamber studies of patient specific nasal epithelial cultures have been used to show the relative efficacy of the new modulator combination TRIKAFTATM in rescuing F508del-CFTR and certain rare mutations (23, 24). In addition, the potential efficacy of novel compounds from academic groups or different pharmaceutical companies have also been revealed in Ussing chamber studies of patient-derived primary nasal epithelial cultures. Interestingly, our group and others observed that patients with the same genotype can exhibit different in-vitro drug responses (11, 25-30). The molecular basis for interpatient variations in in-vitro response size remains unknown but is thought to be a harbinger of the extent of variability in clinical response size (15). Thus, there is a need to develop a higher-throughput testing platform of patient-specific airway tissue that enables therapy development, particularly for those CF-causing mutations like W1282X for which no therapies exist. Unfortunately, nasal epithelial cell cultures have a limited ability to expand in culture and lose CFTR functional expression with progressive population doublings (16). These properties are limiting to combinatorial studies of investigational compounds. Robust protocols have been developed for the differentiation of CF patient derived iPSCs (31-37) thereby providing the potential for a renewable source of patient derived tissue that is scalable to high-throughput testing. In proof of concept studies, the functional rescue of
3
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. F508del-CFTR chloride channel activity by CFTR modulators was shown for lung tissue differentiated at the air-liquid interface (ALI) using either the Ussing chamber or a membrane potential sensitive dye (FLiPR) (35, 38). However, the protocol for differentiating mature airway at ALI is lengthy and not readily scalable. In the current project, we asked if a truncated protocol, generating immature airway tissue from patient-derived iPSCs could be used for drug profiling in a higher throughput format. We show here that this abbreviated differentiation protocol generated relevant CFTR expressing cell types, enabled comparative studies of multiple interventions targeting either F508del or W1282X-CFTR and recapitulated patient specific responses observed in gold-standard Ussing chamber studies of primary nasal epithelial cultures in proof-of-concept studies. We anticipate that these findings will enable the use of patient- specific iPSCs for CF precision medicine approaches. RESULTS
Differentiation of IPS cells to immature lung in high-throughput format. We employed the protocol developed by Wong et al (35), Fig. 1A) to differentiate non-CF and CF iPS cells to immature lung tissue in a 96 well plate format.
|
These cultures remained submerged under differentiation media hence, we refer to them as “submerged cultures”. As expected, these cultures express NKX2.1, a marker of lung progenitors and pan-cytokeratin – an epithelial cell marker, (Fig. 1B). We conducted transcriptomic analyses in order to further characterize the properties of submerged cultures. Bulk RNAseq was performed for: 1) 3 undifferentiated iPS cells, each with 3 replicate cultures; 2) 9 iPS cell lines differentiated under “submerged” conditions to create immature lung cultures (4 non-CF and 5 CF); and non-CF differentiated bronchial airway cultures (n=3) as positive controls (NIH Tissue Core- University of Iowa), (Fig. 1C and D). Principal component analyses showed that the immature lung cultures were intermediate (as determined by PC1) between the iPS cells and differentiated mature bronchial epithelium with respect to gene expression of the 100 genes displaying the greatest variability across samples (Fig. 1C). Further, at least with these samples, CFTR genotype (with or without a CF causing mutation), did not change the distribution of cell sub-populations across the different samples. The heatmap in Fig. 1D, confirms that the immature lung cultures express genes specific for multiple airway cell types, including basal cells, early club cells and goblet cells (39). Interestingly, after differentiation to immature lung cultures, CFTR is expressed, close to levels comparable to those measured in mature non-CF bronchial cultures differentiated at the air-liquid interface. 4
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 1: Differentiation of patient derived iPS cells to immature airway. (A) Schematic of differentiation protocol and timeline: human iPSCs were directed to definitive endoderm and passaged onto 96 well plates during the anterior foregut endoderm differentiation. The submerged cultures were differentiated for 10 more days into immature lung cells. (B) Immunofluorescence images of submerged, immature cultures. Scale bar, 20 µm. Most cells stained positive for TTF1 (NKX2-1) and pan-cytokeratin. (C) Principal component analysis (PCA) comparing iPS cell lines, immature lung cultures differentiated from iPS cell lines and primary bronchial cultures. Both CF and non-CF (including mutation corrected) iPS cell lines studied. (D) Heatmap of gene expression clustered according to cell types using marker genes (39). The columns correspond to different donors and whether lines are CF or non-CF (these include mutation corrected (MC). Supplementary Table 1, provides a legend for colours used with CFTR genotype, for each of the samples. Columns are also clustered as iPS cell lines, immature lung (IM) differentiated from iPS cell lines and primary bronchial cultures. Relative CFTR expression across cultures is shown in the bottom row of the heat map.
|
5
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Immature lung tissue differentiated from iPSCs express functional CFTR. Cyclic AMP activated Wt-CFTR channel function was detected for immature airway cultures differentiated in wells of a 96 well plate, using a fluorescence-based assay of membrane potential changes (FLiPR) (40). The trace in Fig. 2A, shows representative forskolin mediated membrane depolarization in the presence of an outward chloride gradient (mean change -/+SEM for 4 wells). This response was inhibited by the CFTR inhibitor, CFTRInh-172 as expected for CFTR mediated channel activity. Expression of the mature CFTR protein as a 180 kD polypeptide was confirmed for cultures grown in this format (Fig. 2B). The heat map in Figure 2C, shows well scans of FLiPR dye fluorescent change with depolarization caused by forskolin or the DMSO vehicle added to alternating rows in quadruplicate. Red shows the highest responses and dark blue, the lowest responses. Interestingly, for certain wells, there are focal hot spots of activity, reflecting heterogeneity of the culture. The open circles in the bar graph of Fig.2D, show the mean peak CFTR channel activity after DMSO or forskolin measured by FliPR for each well in a 96 well plate. The solid symbol represents the mean of all the technical replicates for either DMSO or forskolin on a plate. This FLiPR assay of CFTR mediated chloride conductance in immature airway cultures exhibits reproducibility in a 96 well plate format with a Z-factor score of 0.33 and an SSMD (Strictly standardized mean difference) score of 4.59. These metrics support the claim that this is a good to excellent assay platform for CFTR channel function (41, 42). 6
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 2: Immature lung cultures express Wt CFTR and channel activity in fluorescence-based assay. (A) Representative FLiPR trace (mean-/+SEM) responses of 4 wells plated with immature airway cultures stimulated by 10 µM forskolin and inhibited by 10 µM CFTR Inh-172. This cell line was derived from donor=CF2MC for which the F508del mutation was corrected to Wt. The naming of lines is consistent with Figure 1. (B) Western blot shows MW markers and mature CFTR expression (180 kD). Calnexin (CNX) was used as loading control. (C) Sample study for establishing assay statistics (Z-factor and SSMD). Multiple wells seeded with iPSC differentiated to immature airway epithelium. Alternating rows (each containing 4 replicates) show responses to agonist (FSK) or vehicle control (DMSO). The response size is colour-coded as shown in the scale bar, red showing the highest response.
|
(D) Bar graph shows the peak responses of
7
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Wt CFTR function (after vehicle (DMSO) or agonist (FSK) in each well (open circle) of a 96 well plate. Bar graphs summarize data from four, 96 well plates generated from 2 differentiations from the iPSC line, CF2MC. Data from each plate grouped with a bracket. The solid symbol represents the mean for each condition in a plate. Immature lung tissue differentiated from CF iPSCs models primary defects caused by F508del. Having shown that the FLiPR assay of immature airway differentiated from iPSCs is suitable for a high-throughput assay of Wt-CFTR, we applied it to study of the mutation, F508del-CFTR, and its modulation by small molecules. In Fig. 3, we show that the primary defects caused by the major mutation, F508del are recapitulated in immature lung cultures generated from two different patient donors who are homozygous for this mutation. The donors identification numbers are consistent those shown in Fig.1. For immature airway cultures from both CF donors: CF2 and CF4, the mean residual forskolin activated CFTR channel activities, (5.67+/-2.16 (SD), n=4, and 7.25+/-0.91 (SD), n=4) were significantly reduced relative to that exhibited by non-CF culture, (37.50+/- 8.02 (SD, n=4) as shown in Figs 3 and 2 respectively. As expected, the abundance of mature band C was also reduced for the major mutation in iPSC-derived airway epithelium in the absence of small molecule modulators (Fig. S1). Treatment of immature lung cultures from donor, CF2, with small molecule modulator combinations led to reproducible responses in the FLiPR assay (Fig. 3A). Relative to the vehicle (DMSO) control, treatment with the corrector, lumacaftor, i.e., VX-809, led to a partial rescue of the functional expression for CF2 immature lung cultures relative to vehicle (DMSO) in the presence of the potentiator (VX-770). As expected, there were robust functional responses to investigational modulator combinations: AC1+AC2-1 or AC1+AC2-2 and the potentiator: AP2 using the FLIPR assay on immature lung cultures (Fig. 3A). The efficacy of these modulator combinations was published previously for studies on primary nasal epithelial cultures (20, 25, 43). We compared modulator responses immature lung cultures generated from donor, CF4, another individual who is homozygous for F508del. Whilst the ranking of modulator efficacies was similar for the two donors; CF2 and CF4, the responses to the modulator combination: AC1+AC2-2 was higher in cultures from donor CF4 than in donor CF2 (p=0.022). Together, these findings suggest that donor-specific differences in in-vitro response to modulators can be measured using this platform. 8
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021.
|
The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 3: Immature lung cultures generated from iPSCs from two different donors homozygous for F508del exhibit robust, responses to modulators (A) Left panel: Representative FLiPR traces of cultures derived from iPSCs with F508del mutation after chronic rescue (48 hours) with DMSO (0.1%) or small molecules as defined in the legend with concentrations indicated in Table 1. After 5 min baseline, the cells were stimulated with DMSO or FSK +/- 1µM VX-770 (or 1.5 µM AP2). Right panel: Bar graphs show peak responses from each modulator combination. Each solid circle represents mean peak responses of 4 wells
9
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. in a 96 well plate for 4 plates of immature airway epithelium generated from a single differentiation of iPSC line: CF2. The naming of lines is consistent with Figure 1. (B) Left: FLIPR traces showing responses to small molecule modulators on immature airway epithelium differentiated from an iPSC line derived from donor: CF4. Right: Bars show reproducibility of FLIPR assay. Each solid dot represents the mean peak response for 4 technical replicates (wells) of a 96 well plate and there were 4 plates generated from a single differentiation of iPSC line, CF4. (C) Correlation between mean donor specific activations measured using FLiPR (mean values, -/+ interventions, extracted from bars above) and mean, donor specific changes measured in the Ussing chamber, delta Ieq (µA/cm2) after forskolin and treatment as reported in (44). In Fig. 3C, we show that the responses to modulators observed for immature lung cultures generated from CF2 and CF4, correlate (r=0.95) with the in-vitro responses previously measured for well differentiated primary nasal cultures from the same donors (44). The bioelectric responses to modulators were previously measured for primary nasal cultures in Ussing chambers, considered the “gold standard” for measuring CFTR channel function. Interestingly, as in the FLiPR assay of immature lung, the fully differentiated nasal epithelial cultures generated from donor CF4 consistently exhibited as higher functional response to AC1-+AC2+2 than the nasal cultures generated from donor CF2 (p=0.022). Hence the patient-specific responses to modulators of the major mutation (F508del) observed in the high-throughput, FLiPR based assay recapitulated those observed in the gold-standard, but, low-through put Ussing chamber assay. Immature lung tissue differentiated from CF iPSCs models primary defects caused by W1282X and report differential responses to investigational rescue compounds. Having shown that the FLiPR assay of immature lung cultures differentiated from iPSC lines faithfully recapitulates therapeutic responses to modulator drugs targeting the major mutation: F508del, we then tested the fidelity with which this novel platform can used to measure patient- specific responses to investigational compounds targeting the rare, nonsense mutation: W1282X.
|
The format for these phenotypic assays is shown in Fig 4A. Here, we show a heat map for the FLiPR assay of submerged lung cultures. Wells (24) seeded with submerged lung containing Wt- CFTR (CF2 mutation corrected (MC), are displayed on the left. Wells (32) seeded with immature lung differentiated from iPSCs generated from a donor who is homozygous for the rare nonsense mutation: W1282X (CF7) are shown on the right. The colour scale on the right of the wells corresponds to the peak FLiPR fluorescence after activation and potentiation of CFTR channel activity. The wells seeded with non-CF (Wt), immature lung tissue exhibit robust activation by forskolin and the peak corresponds to red. On the other hand, the colour of the wells containing immature lung cultures from a donor homozygous for the nonsense mutation varied (from dark blue to red, reflecting low to high responses) according to the treatments provided. A panel of small molecule interventions was tested for efficacy in rescuing the functional expression of W1282X in three donors, all homozygous for this mutation. These interventions
10
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. were chosen based on previous experiments in a bronchial cell lines (8) and our previous bioelectric studies of primary nasal epithelial cultures (43). Based on these previous studies, we expect that only those cultures receiving SMG1i, a small molecule inhibitor of nonsense mediated decay together with protein modulators of the assembly and function (8), would enable functional rescue. As shown in the bar graphs in 4B, the responses observed for each patient-specific culture are reproducible. Interestingly, there were variable responses to the same panel of modulators across the 3 donors. For example, for donor CF5, the magnitude of the response to the modulator combinations is muted relative to the cultures derived from CF6 and CF7. As expected, significant functional rescue was only observed in immature cultures treated with the nonsense mediated decay inhibitor, SMG1i in combination with correctors and potentiator of the truncated W1282- CFTR protein. Interestingly, the premature termination codon (PTC) read-through agent (G418) was not effective augmenting functional rescue of W1282X mediated by SMG1i, as suggested in some but not all studies of primary cultures harboring this nonsense mutation (42). Hence, we show that combinatorial interventions targeting W1282X-CFTR can be tested in patient derived cultures using the FLiPR based high-throughput assay. Further, we show that nonsense mediated decay is limiting modulator activity in these cultures and finally, that different responses are observed for different patient-specific cultures. As for the studies of F508del-CFTR, we observed a strong correlation (r=0.93) between the patient-specific response to interventions measured by FLiPR in immature lung and responses measured in primary nasal epithelial cultures in the Ussing chamber (data from (43) Figure 4C).
|
Unfortunately, due to limitations related to expansion of primary nasal epithelial cultures, the correlation for all conditions, for all three donors, could not be assessed. However, it is clear that the best functional responses in both iPS-derived immature lung and primary nasal epithelial cultures, is achieved after inhibition of nonsense mediated decay using SMG1i. 11
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Fig. 4: Immature lung cultures generated from iPSCs from three different donors with the nonsense mutation, W1282X, exhibit differential phenotypic responses to modulators. (A) Heatmap of peak responses generated from non-CF culture stimulated with 10µM FSK (left) and CF culture (W1282X) treated with combination of small molecules (A-H, concentrations defined in Table 1). (B) Bar graph shown the peak response of W1282X-CFTR treated with DMSO (0.1%) or small molecules (defined on y axis with concentrations indicated in Table 1). After 5 min baseline, the cells were stimulated with DMSO or FSK +/-
12
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 1µM VX-770 (VX) or 1.5µM AP2 (AC). Mean peak responses (of 4 wells in a 96 well plate) after agonist and potentiator shown for each small molecule combination as a solid circle. Three patients homozygous for W1282X were studied (CF5,6 and 7). For CF5, 5 plates generated from 4 differentiations from iPSCs were studied. For CF6, 4 plates from 1 differentiation were studied and for CF7, 5 plates from 2 differentiations were studied. CFTR modulators (VX-809 or AC1+AC2-2), in combination with SMG1i, were effective in increasing the abundance of the truncated W1282X protein in immature lung cultures (Supp Fig 2A). (C) Correlation plot between FLIPR peak response and Ussing chamber studies (data from (43)). (D) Basal W1282X-CFTR transcript (left) expression in primary nasal epithelial cultures (nas) and iPSC-derived immature lung cultures (IL) from three donors. The mutant CFTR transcript abundance (expressed relative to adult lung, shown in Fig. 4D, upper graph), is similar for both the iPSC derived immature lung cultures and the matched nasal cultures. DISCUSSION
In these studies, we showed that fluorescence-based channel activity assays of immature lung cultures generated from iPSCs, have the potential to support therapy development for individuals with CF. Such immature lung cultures possess disease-relevant cell types, such as secretory cells, and can be scaled up to enable robust phenotypic screening of therapeutic interventions. As suggested in our proof-of-concept studies, functional responses seen in this novel assay platform recapitulate patient specific responses measured by low-throughput Ussing chamber studies of primary nasal epithelial cultures, thereby supporting the future development of this novel screening platform to aid in precision medicine development for CF individuals currently lacking therapeutic options.
|
A caveat of our work, relates to the relatively immature airway tissue employed for the phenotypic screen. Although transcriptomic studies show that the submerged, immature lung tissue expresses CFTR and contains CF-relevant cells types including secretory cells such as club and goblet cells (Figure 1), we acknowledge that there will be a deficit of certain other cell types, including ionocytes. Hence, we suggest that the medium to high-throughput iPSC-based screen described here can be used to identify promising compounds that target defects in RNA and protein processing exhibited by CF causing nonsense mutations. Subsequently, compounds identified from these primary screens should be validated in bioelectric assays of fully differentiated primary nasal epithelial tissues as in the current work. 13
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Interestingly, at least for the cultures studied here, there was no difference in the expression of cell type markers between immature airway cultures derived from CF and non-CF iPS cells. These studies support previous publications, suggesting that CFTR mutations do not change cell type distribution (45). However, additional transcriptomic studies of lung tissues modeling early lung development for individuals harbouring nonsense mutations, expected to reduce CFTR mRNA by triggering NMD, are required for a more fulsome analysis of the role of CFTR expression on early lung development. Similarly, the impact of CFTR nonsense mutations on proximal airway maturation after transitioning to the air/liquid interface are required in order to test the impact of CFTR expression on airway differentiation. Although an extremely active area of research, there are no approved treatments targeting the primary defects caused by nonsense mutations in CFTR. Small molecules that facilitate read- through of premature termination codons (PTCs) to enable production of full length CFTR protein are undergoing development. However, only one such compound, called ELOX-2 is currently in (46), clinical trial https://www.cff.org/Trials/Pipeline. An alternative approach is to suppress NMD using small molecule inhibitors which has been shown to augment abundance of the truncated W1282X protein in the current work and previous papers (8, 43, 47); though such inhibitors may have nonspecific, deleterious effects. Further, such NMD inhibitors will need to be combined with protein modulators in order to correct the primary defects in W1282-CFTR function. Hence, combinatorial treatments aimed at promoting PTC readthrough, inhibition of NMD and modulation of the stability and function of the mutant protein will need to be tested to identify the most effective interventions. We anticipate that the iPS cell phenotypic platform that we describe here will facilitate such urgently needed therapy development by facilitating comparative analysis of multiple therapeutic strategies.
|
for subjects carrying the CF-causing nonsense mutation, G542X
Individuals who are homozygous for nonsense mutations are rare world-wide (188 patients, http://cftr2.org) and as a result, relevant airway tissue models for therapy testing are limited. Immature lung differentiated from iPSC lines are infinitely renewable in principle and we show that after differentiation to an immature tissue, together with the FLiPR assay of CFTR channel activity, they can be employed for testing therapeutic strategies. Importantly, immature lung cultures from individuals homozygous for F508del exhibited variable phenotypic responses to modulators targeting its mutation class. Likewise, immature lung cultures from individuals homozygous for W1282X also exhibited variable patient-specific responses to small molecules targeting this nonsense mutation. Hence, we provide a resource with which to identify therapeutic targets other than mutant CFTR that may inform companion therapies. Finally, our findings suggest that patient-specific genetic background plays a role in
14
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. conferring variable drug effect size and highlights the importance of iPS-derived lung cultures for advancing therapy development and precision medicine. MATERIALS AND METHODS:
Cell culture: Differentiation to submerged immature lung culture: Human iPS cells were obtained from the Cystic Fibrosis Individualized Therapy (CFIT) program (48). The submerged immature lung cultures were generated from iPS cells as previously described (35). Human iPS cells were grown on six-well plates (Corning) coated with Matrigel (Corning) and maintained with mTeSR (Stem Cell Technologies). Cultures were expanded weekly with Gentle Cell Dissociation Buffer (GCDR, Stem Cell Technologies) at 70-90% confluency at a 1:10 ratio. For definitive endoderm (DE) induction, single-cell suspensions were generated from five-minute incubation at 37°C followed by scraping and gentle trituration. Cells were plated onto six-well plates in media supplemented with 10 µM Y27632 compound (Stem Cell Technologies) for 24 hours. DE cultures were generated using the StemDiff Definitive Endoderm Kit (Stem Cell Technologies) as per manufacturer’s protocol for 5 days. To differentiate anterior foregut endoderm (AFE) culture, cells were treated with differentiation basal medium (KnockOut DMEM, 10% KnockOut serum replacement, 1% penicillin-streptomycin, 2mM Glutamax, 0.15 mM monothioglycerol, and 1 mM non-essential amino acid) supplemented with FGF2 (500 ng/mL) and SHH (500 ng/mL) for 24 hours. On the second day of AFE differentiation, cells were dissociated into single-cell suspensions and plated onto type IV collagen coated (60 ug/mL, Sigma) 96-well plates at a density of 25,000 cells per well.
|
The media was supplemented with 10 µM Y27632 compound for 24 hours and was changed every 48 hours for an additional three days. For directed differentiation to lung progenitor cells and immature lung cells, cultures were overlayed with differentiation basal medium supplemented with FGF7 (50 ng/mL), FGF10 (50 ng/mL) and BMP4 (5 ng/mL) for 5 days, and then FGF7 (10 ng/mL), FGF10 (10 ng/mL) and FGF18 (10 ng/mL) for 5 days. Membrane potential based functional assays: Apical Chloride Conductance (ACC) Assay for CFTR function The ACC assay was used to assess CFTR mediated changed in membrane depolarization using methods as previously described (40, 49). In summary, iPSC derived- submerged lung cultures were incubated with zero sodium, chloride and bicarbonate buffer (NMDG 150 mM, Gluconic acid lactone 150 mM, Potassium Gluconate 3 mM, Hepes 10 mM, pH 7.42, 300 mOsm) containing 0.5 mg/ml of FLIPR® dye for 30 mins at 37°C. Wt-CFTR function was measured after acute addition of Fsk (10 µM) or 0.01% DMSO control. Cells were chronically rescued with corrector compounds
15
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. for 24 hours. Post drug rescue, F508del-CFTR function was measured after acute addition of Fsk (10 µM) and VX-770 (1 µM) or AP-2 (1.5 µM). CFTR functional recordings were measured using the FLIPR® Tetra High-throughput Cellular Screening System (Molecular Devices), which allowed for simultaneous image acquisition of the entire 96 well plate. Images were first collected to establish baseline readings over 5 mins at 30 second intervals. Modulators were then added to stimulate CFTR mediated anion efflux. Post drug addition, CFTR mediated fluorescence changes were monitored and images were collected at 15 second intervals for 70 frames. CFTR channel activity was terminated with addition of Inh172 (10 µM) and fluorescence changes were monitored at 30 second intervals for 25 frames. Analysis and heatmap generation Experiments were exported as multi frame TIFF images of which every frame recorded the entire plate. Pixels outside of well areas were filtered out using the initial signal intensities and wells containing opened organoids were separated. All traces were normalized to the last point of the baseline intensity. Peak response for each pixel was calculated as the maximum deviation from baseline. During the stimulation segment, fluorescence intensity increased for CFTR function. Heatmap representation was generated from the peak response of each pixel and the mean response trace of wells was generated by averaging the corresponding pixel traces. Real-time Quantitative PCR: As previously described (21), total mRNA was extracted using RNeasy® Plus Micro Kit, following enclosed instructions. After measuring the spectrophotometric quality of extracted RNA through 260/280 ratios of 2.0 and 260/230 ratios of 1.8-2.2, mRNA samples used to reverse transcribe 1 µg of cDNA using iScriptTM cDNA Synthesis Kit.
|
Quantitative real-time PCR was performed with PowerUP SYBR Green Mastermix Master Mix on ViiA7 (Applied Biosystems). Gene expression is normalized to house-keeping gene GAPDH and expressed relative to control human tissue RNA extracts (2^ΔΔCT). A total run of 40 cycles. Cycle threshold (CT) values above 38 were considered “not expressed”. The primers used for amplification are described in the following table. Primer Sequences CFTR CFTR Alpha-ENaC Alpha-ENaC FOXI1 FOXI1 FOXJ1 FOXJ1
Fwd: 5’- CTATGACCCGGATAACAAGGAGG-3’ Rev: 5’- CAAAAATGGCTGGGTGTAGGA-3’ Fwd: 5’- TTGACGTCTCCAACTCACCG-3’ Rev: 5’- GGCAGAGGAGGACAAAGGTC-3’ Fwd: 5’- CGGGCAAAGGGAATTACTGG-3’ Rev: 5’- AGGCTCCATCCAAGATGTCC-3’ Fwd: 5’- GAGCGGCGCTTTCAAGAAG-3’ Rev: 5’- GGCCTCGGTATTCACCGTC-3’
16
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. KRT5 KRT5 MUC5ac MUC5ac SCGB1A1 SCGB1A1 TRP63 TRP63 GADPH GADPH
Fwd: 5’- GGAGTTGGACCAGTCAACATC-3’ Rev: 5’- TGGAGTAGTAGCTTCCACTGC-3’ Fwd: 5’- CCATTGCTATTATGCCCTGTGT-3’ Rev: 5’- TGGTGGACGGACAGTCACT-3’ Fwd: 5’- TTCAGCGTGTCATCGAAACCC-3’ Rev: 5’- ACAGTGAGCTTTGGGCTATTTTT-3’ Fwd: 5’- ACTTCACGGTGTGCCACCCT -3’ Rev: 5’- GAGCTGGGGTTTCTACGAAACGCT -3’ Fwd: 5’- CTGGGCTACACTGAGCACC -3’ Rev: 5’- AAGTGGTCGTTGAGGGCAATG -3’
RNA Sequencing and Analysis: RNA samples were extracted using methods described previously (50, 51). RNA samples with an RNA integrity number (RIN) greater than 8.5 was submitted to The Centre of Applied Genomics (TCAG) at SickKids for bulk RNA sequencing. In brief, RNA libraries were generated using NEB Ultra II Directional mRNA with an average of 69,954,038 reads from each library on the Illumina HiSeq 2500 platform using high-throughput V4 flowcells. Post sequencing quality control was performed using open-source software FastQC (https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Trim galore (52) was used to remove low quality sequences and trim adapters. Paired-end reads were aligned to the human reference genome (hg38) using STAR (version 2.7.1a) (53). The resulting bam files containing aligned sequences were subsequently processed using SAMtools (54), and raw counts generated with featureCounts (55) were used for downstream analysis (transcripts per gene). R package DESeq2 (v.1.24.0) (56) was used to calculate size factors for each sample and perform regularized-logarithm rlog transformation of read counts. The 100 genes with the highest variance in expression across all samples were subjected to principal component analysis (PCA). Immunofluorescence: Samples were fixed in 4% paraformaldehyde and then washed three times with PBS, 5 mins per wash at room temperature. Cell permeabilization was performed using 0.05% TritonX-100 followed by three PBS washes. Samples were blocked using 5% BSA for 1 hour and incubated with primary antibody against TTF1 (NKX2-1) or EPCAM overnight.
|
After removal of primary antibody, samples were washed 3 times with PBS, 5 mins per wash and incubated with secondary antibodies and nuclear marker DAPI for 1 hour. Samples were then washed 3 times with PBS, 5 mins per wash at room temperature. Images were acquired on the SP8/STED confocal microscope (Leica). 17
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Western blotting: Samples were collected in ice cold PBS and pelleted through centrifugation at 4°C (500g for 7 in 200µL of modified mins). Post centrifugation, the cell pellet was re-suspended radioimmunoprecipitation assay butter (50 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, pH 7.4, 0.2% (v/v) SDS and 0.1% (v/v) Triton X-100) containing a protease inhibitor cocktail for 10 min. After centrifugation at 13,000 rpm for 5 min, the soluble fractions were analyzed by SDS-PAGE on 6% Tris-Glycine gel. After electrophoresis, proteins were transferred to nitrocellulose membranes and incubated in 5% milk and CFTR bands were detected using the mAb 596. Calnexin (CNX) was used as a loading control and detected using a Calnexin-specific rAb (1:5000). The blots were developed with using the Li-Cor Odyssey Fc (LI-COR Biosciences, Lincoln, NE, USA) in a linear rage of exposure (1-20 min). Relative levels of CFTR protein were quantitated by densitometry of immunoblots using ImageStudioLite (LI-COR Biosciences, Lincoln, NE, USA) (57, 58). Supplementary Materials:
Supplementary Figure 1: Representative F508del-CFTR protein expression in lung submerge after 48h pre-treatment with DMSO (0.1%), 3µM VX-809, 0.5µM AC1 + 3µM AC2-1 or 0.5µM AC1 + 3µM AC2-2. Supplementary Figure 2: Correction of W1282X mutation by CRISPR-Cas9 editing, confers CFTR channel activity in submerged, immature lung cultures from donor #4. CFTR channel activity was measured using the FLiPR assay and the bars represent mean +/- SD in 4 technical replicates. Table 1: Concentrations employed for small molecules
ACKNOWLEDGMENTS:
iPS cells and primary nasal cell cultures were obtained through the CF Canada-SickKids Program for Individualized CF Therapy (CFIT). The SMG1i was obtained through Cystic Fibrosis Foundation Therapeutics and the small molecules AC1, AC2-1, AC2-2 and AP2 were provided by AbbVie. We thank Ashvani Singh for reviewing this manuscript. Declaration of competing interest. There are no competing interests. Funding: This work was supported by CFIT program with funding provided by CF Canada and the SickKids Foundation, by the Government of Canada through Genome Canada and Ontario
18
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
|
Genomics Institute (OGI-148) and a grant to CEB from Medicine by Design. This work was also supported by the Cystic Fibrosis Foundation (OOC 590131) This study was supported by a grant from the Government of Ontario. TJM was also supported by funding from Emily’s Entourage. Author contributions: Jia Xin Jiang: designed and performed experiments, revised the manuscript Leigh Wellhauser: performed experiments Onofrio Laselva: performed experiments and revised the manuscript Irina Utkina: Data analysis Zoltan Bozoky: Data analysis Tarini Gunawardena: Performed experiments Zoe Ngan: Performed experiments Sunny Xia: Data collection Paul D.W. Eckford: Documentation oversight, reviewed and edited the manuscript Felix Ratjen: Provided scientific insight, reviewed and edited the manuscript Theo Moraes: Provided scientific insight, reviewed and edited the manuscript John Parkinson: Data analysis and interpretation, reviewed the manuscript Amy P. Wong: Data analysis and interpretation, reviewed the manuscript Christine E. Bear: Conceived and designed the work, drafted and revised the manuscript
Data and materials availability: Raw sequence data is available through the NCBI sequence read archive (https://www.ncbi.nlm.nih.gov/sra) with the BioProject ID PRJNA721455. Bibliography:
1. 2. 3. 4. 5. S. H. Cheng et al., Defective intracellular transport and processing of CFTR is the molecular basis of most cystic fibrosis. Cell 63, 827-834 (1990). B. H. Qu, P. J. Thomas, Alteration of the cystic fibrosis transmembrane conductance regulator folding pathway. J Biol Chem 271, 7261-7264 (1996). G. L. Lukacs, A. S. Verkman, CFTR: folding, misfolding and correcting the DeltaF508 conformational defect. Trends Mol Med 18, 81-91 (2012). C. E. Wainwright, J. S. Elborn, B. W. Ramsey, Lumacaftor-Ivacaftor in Patients with Cystic Fibrosis Homozygous for Phe508del CFTR. N Engl J Med 373, 1783-1784 (2015). J. L. Taylor-Cousar et al., Tezacaftor-Ivacaftor in Patients with Cystic Fibrosis Homozygous for Phe508del. N Engl J Med 377, 2013-2023 (2017). 19
bioRxiv preprint
6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26.
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. P. G. Middleton et al., Elexacaftor-Tezacaftor-Ivacaftor for Cystic Fibrosis with a Single Phe508del Allele. N Engl J Med 381, 1809-1819 (2019). S. Cuevas-Ocana, O. Laselva, J. Avolio, R. Nenna, The era of CFTR modulators: improvements made and remaining challenges. Breathe (Sheff) 16, 200016 (2020). H. C. Valley et al., Isogenic cell models of cystic fibrosis-causing variants in natively expressing pulmonary epithelial cells. J Cyst Fibros 18, 476-483 (2019). P. M. Haggie et al., Correctors and Potentiators Rescue Function of the Truncated W1282X- Cystic Fibrosis Transmembrane Regulator (CFTR) Translation Product.
|
J Biol Chem 292, 771-785 (2017). M. A. Aksit et al., Decreased mRNA and protein stability of W1282X limits response to modulator therapy. J Cyst Fibros 18, 606-613 (2019). V. Mutyam et al., Novel Correctors and Potentiators Enhance Functional Rescue of CFTR Nonsense Mutation Translational Readthrough. Am J Respir Cell Mol Biol, (2021). F. Van Goor et al., Correction of the F508del-CFTR protein processing defect in vitro by the investigational drug VX-809. Proc Natl Acad Sci U S A 108, 18843-18848 (2011). G. Veit et al., Structure-guided combination therapy to potently improve the function of mutant CFTRs. Nat Med 24, 1732-1742 (2018). L. Guerra et al., The preclinical discovery and development of the combination of ivacaftor + tezacaftor used to treat cystic fibrosis. Expert Opin Drug Discov 15, 873-891 (2020). I. M. Pranke et al., Correction of CFTR function in nasal epithelial cells from cystic fibrosis patients predicts improvement of respiratory function by CFTR modulators. Sci Rep 7, 7375 (2017). J. J. Brewington et al., Brushed nasal epithelial cells are a surrogate for bronchial epithelial CFTR studies. JCI Insight 3, (2018). G. Berkers et al., Rectal Organoids Enable Personalized Treatment of Cystic Fibrosis. Cell Rep 26, 1701-1708 e1703 (2019). J. F. Dekkers et al., A functional CFTR assay using primary cystic fibrosis intestinal organoids. Nat Med 19, 939-945 (2013). O. Laselva et al., The CFTR Mutation c.3453G > C (D1152H) Confers an Anion Selectivity Defect in Primary Airway Tissue that Can Be Rescued by Ivacaftor. J Pers Med 10, (2020). O. Laselva et al., Preclinical Studies of a Rare CF-Causing Mutation in the Second Nucleotide Binding Domain (c.3700A>G) Show Robust Functional Rescue in Primary Nasal Cultures by Novel CFTR Modulators. J Pers Med 10, (2020). H. Cao et al., A helper-dependent adenoviral vector rescues CFTR to wild-type functional levels in cystic fibrosis epithelial cells harbouring class I mutations. Eur Respir J 56, (2020). A. S. Ramalho et al., Correction of CFTR function in intestinal organoids to guide treatment of cystic fibrosis. Eur Respir J 57, (2021). G. Veit et al., Allosteric folding correction of F508del and rare CFTR mutants by elexacaftor- tezacaftor-ivacaftor (Trikafta) combination. JCI Insight 5, (2020). O. Laselva et al., Rescue of multiple class II CFTR mutations by elexacaftor+ tezacaftor+ivacaftor mediated in part by the dual activities of Elexacaftor as both corrector and potentiator. Eur Respir J, (2020). O. Laselva et al., Emerging preclinical modulators developed for F508del-CFTR have the potential to be effective for ORKAMBI resistant processing mutants. J Cyst Fibros 20, 106-119 (2021). S. V. Molinski et al., Orkambi(R) and amplifier co-therapy improves function from a rare CFTR mutation in gene-edited cells and patient tissue. EMBO Mol Med 9, 1224-1243 (2017). 20
bioRxiv preprint
27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48.
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021.
|
The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Y. S. Wu et al., ORKAMBI-Mediated Rescue of Mucociliary Clearance in Cystic Fibrosis Primary Respiratory Cultures Is Enhanced by Arginine Uptake, Arginase Inhibition, and Promotion of Nitric Oxide Signaling to the Cystic Fibrosis Transmembrane Conductance Regulator Channel. Mol Pharmacol 96, 515-525 (2019). G. de Wilde et al., Identification of GLPG/ABBV-2737, a Novel Class of Corrector, Which Exerts Functional Synergy With Other CFTR Modulators. Front Pharmacol 10, 514 (2019). P. W. Phuan et al., CFTR modulator therapy for cystic fibrosis caused by the rare c.3700A>G mutation. J Cyst Fibros, (2020). P. W. Phuan et al., Combination potentiator ('co-potentiator') therapy for CF caused by CFTR mutants, including N1303K, that are poorly responsive to single potentiators. J Cyst Fibros 17, 595-606 (2018). S. Simsek et al., Modeling Cystic Fibrosis Using Pluripotent Stem Cell-Derived Human Pancreatic Ductal Epithelial Cells. Stem Cells Transl Med 5, 572-579 (2016). A. L. Firth et al., Generation of multiciliated cells in functional airway epithelia from human induced pluripotent stem cells. Proc Natl Acad Sci U S A 111, E1723-1730 (2014). S. X. Huang et al., Efficient generation of lung and airway epithelial cells from human pluripotent stem cells. Nat Biotechnol 32, 84-91 (2014). S. Merkert, C. Bednarski, G. Gohring, T. Cathomen, U. Martin, Generation of a gene-corrected isogenic control iPSC line from cystic fibrosis patient-specific iPSCs homozygous for p.Phe508del mutation mediated by TALENs and ssODN. Stem Cell Res 23, 95-97 (2017). A. P. Wong et al., Efficient generation of functional CFTR-expressing airway epithelial cells from human pluripotent stem cells. Nat Protoc 10, 363-381 (2015). M. Ogawa et al., Directed differentiation of cholangiocytes from human pluripotent stem cells. Nat Biotechnol 33, 853-861 (2015). M. Ogawa et al., Generation of functional ciliated cholangiocytes from human pluripotent stem cells. bioRxiv, 2021.2003.2023.436530 (2021). A. P. Wong et al., Directed differentiation of human pluripotent stem cells into mature airway epithelia expressing functional CFTR protein. Nat Biotechnol 30, 876-882 (2012). M. Deprez et al., A Single-Cell Atlas of the Human Healthy Airways. Am J Respir Crit Care Med 202, 1636-1645 (2020). S. Ahmadi et al., Phenotypic profiling of CFTR modulators in patient-derived respiratory epithelia. NPJ Genom Med 2, 12 (2017). X. D. Zhang, A pair of new statistical parameters for quality control in RNA interference high- throughput screening assays. Genomics 89, 552-561 (2007). H. Gubler, Assay data quality assessment. Methods Mol Biol 552, 79-95 (2009). O. Laselva et al., Functional rescue of c.3846G>A (W1282X) in patient-derived nasal cultures achieved by inhibition of nonsense mediated decay and protein modulators with complementary mechanisms of action.
|
J Cyst Fibros 19, 717-727 (2020). O. Laselva, Bartlett C., Popa A., Ip W., Ouyang H., Moraes J.T., Gonska T., Bear C.E, paper presented at the 15th ECFS Basic Science Conference Loutraki, Greece, 21-24 March, 2018 2018. K. Okuda et al., Secretory Cells Dominate Airway CFTR Expression and Function in Human Airway Superficial Epithelia. Am J Respir Crit Care Med, (2020). D. K. Crawford et al., Targeting G542X CFTR nonsense alleles with ELX-02 restores CFTR function in human-derived intestinal organoids. J Cyst Fibros, (2021). M. M. Keenan et al., Nonsense-mediated RNA Decay Pathway Inhibition Restores Expression and Function of W1282X CFTR. Am J Respir Cell Mol Biol 61, 290-300 (2019). P. D. W. Eckford et al., The CF Canada-Sick Kids Program in individual CF therapy: A resource for the advancement of personalized medicine in CF. J Cyst Fibros 18, 35-43 (2019). 21
bioRxiv preprint
49. 50. 51. 52. 53. 54. 55. 56. 57. 58.
doi:
https://doi.org/10.1101/2021.07.05.451013
;
this version posted July 5, 2021. The copyright holder for this preprint (which
was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. S. Erwood et al., Allele-Specific Prevention of Nonsense-Mediated Decay in Cystic Fibrosis Using Homology-Independent Genome Editing. Mol Ther Methods Clin Dev 17, 1118-1128 (2020). O. Laselva, T. A. Stone, C. E. Bear, C. M. Deber, Anti-Infectives Restore ORKAMBI((R)) Rescue of F508del-CFTR Function in Human Bronchial Epithelial Cells Infected with Clinical Strains of P. aeruginosa. Biomolecules 10, (2020). M. Di Paola et al., SLC6A14 Is a Genetic Modifier of Cystic Fibrosis That Regulates Pseudomonas aeruginosa Attachment to Human Bronchial Epithelial Cells. mBio 8, (2017). K. F., Trim Galore! : a wrapper tool around Cutadapt and FastQC to consistently apply quality and adapter trimming to FastQ files. In. A. Dobin et al., STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21 (2013). H. Li et al., The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078-2079 (2009). Y. Liao, G. K. Smyth, W. Shi, featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923-930 (2014). M. I. Love, W. Huber, S. Anders, Moderated estimation of fold change and dispersion for RNA- seq data with DESeq2. Genome Biol 15, 550 (2014). S. Chin et al., Cholesterol Interaction Directly Enhances Intrinsic Activity of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR). Cells 8, (2019). O. Laselva, S. Erwood, K. Du, Z. Ivakine, C. E. Bear, Activity of lumacaftor is not conserved in zebrafish Cftr bearing the major cystic fibrosis-causing mutation. FASEB Bioadv 1, 661-670 (2019). 22
|
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license . 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Ineffective neutralization of the SARS-CoV-2 Mu variant by convalescent and vaccine sera
Keiya Uriu1#, Izumi Kimura1#, Kotaro Shirakawa2, Akifumi Takaori-Kondo2, Taka-aki Nakada3, Atsushi Kaneda3, The Genotype to Phenotype Japan (G2P-Japan) Consortium, So Nakagawa4, Kei Sato1*
1 Institute of Medical Science, University of Tokyo, Tokyo, Japan 2 Kyoto University, Kyoto, Japan 3 Chiba University, Chiba, Japan 4 Tokai University, Kanagawa, Japan # Keiya Uriu and Izumi Kimura contributed equally to this letter. Correspondence: [email protected] (Kei Sato)
Conflict of interest: The authors declare that no competing interests exist. 1
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license . 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33
Abstract On August 30, 2021, the WHO classified the SARS-CoV-2 Mu variant (B.1.621 lineage) as a new variant of interest. The WHO defines “comparative assessment of virus characteristics and public health risks” as primary action in response to the emergence of new SARS-CoV-2 variants. Here, we demonstrate that the Mu variant is highly resistant to sera from COVID-19 convalescents and BNT162b2-vaccinated individuals. Direct comparison of different SARS-CoV-2 spike proteins revealed that Mu spike is more resistant to serum-mediated neutralization than all other currently recognized variants of interest (VOI) and concern (VOC). This includes the Beta variant (B.1.351) that has been suggested to represent the most resistant variant to convalescent and vaccinated sera to date (e.g., Collier et al, Nature, 2021; Wang et al, Nature, 2021). Since breakthrough infection by newly emerging variants is a major concern during the current COVID-19 pandemic (Bergwerk et al., NEJM, 2021), we believe that our findings are of significant public health interest. Our results will help to better assess the risk posed by the Mu variant for vaccinated, previously infected and naïve populations. 2
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license . 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71
Text During the current pandemic, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the causative agent of coronavirus disease 2019 (COVID-19), has considerably diversified.
|
As of September 2021, the WHO has defined four variants of concern (VOC), Alpha (B.1.1.7), Beta (B.1.351), Gamma (P.1) and Delta (B.1.617.2 and AY lineages), as well as five variants of interest (VOI), Eta (B.1.525), Iota (B.1.526), Kappa (B.1.617.1), Lambda (C.37), and Mu (B.1.621).1
The Mu variant represents the most recently recognized VOI.1 Until August 30, 2021, this VOI was detected in 39 countries (Table S1). The epicenter of the Mu variant is Colombia, where it was first isolated on January 11, 2021 (GISAID ID: EPI_ISL_1220045; Figure 1A and Table S2). This country has experienced a huge COVID-19 surge from March to August 2021 that has peaked at 33,594 cases per day (on June 26, 2021; Figure 1A). Although the Gamma VOC was dominant during the initial phase, the Mu VOI outcompeted all other variants including the Gamma VOC in May 2021 and has driven the epidemic in Colombia since then (Figure 1A). Newly emerging SARS-CoV-2 variants need to be carefully monitored for a potential increase in transmission rate, pathogenicity and/or resistance to immune responses. For example, the resistance of VOC/VOIs to humoral immunity elicited by natural SARS-CoV-2 infection or vaccination may allow significant spread of the virus in populations that were initially thought to be protected.2 Resistance to COVID- 19 convalescent and vaccine recipient sera can be attributed to a variety of mutations in the viral spike protein.2 The majority of Mu variants harbors the following eight mutations in spike: T95I, YY144-145TSN, R346K, E484K, N501Y, D614G, P681H, and D950N (Tables S3 and S4). These include mutations commonly identified in VOCs: E484K (shared with Beta, Gamma), N501Y (shared with Alpha), P681H (shared with Alpha) and D950N (shared with Delta) (Table S5). Of those, the E484K change has been shown to reduce sensitivity towards antibodies induced by natural SARS-CoV-2 infection and vaccination.3,4
To assess the sensitivity of the Mu variant to antibodies induced by SARS- CoV-2 infection and vaccination, we generated pseudoviruses harboring the spike proteins of Mu or the other VOC/VOIs. Virus neutralization assays revealed that the Mu variant is 12.4-fold more resistant to sera of eight COVID-19 convalescents, who were infected during the early pandemic (April–September, 2020), than the parental virus (P=0.0078; Figure 1B). Also, the Mu variant was 7.6-fold more resistant to sera obtained from ten BNT162b2-vaccinated individuals compared to the parental virus (P=0.0020; Figure 1C). Notably, although the Beta VOC was thought to be the most resistant variant to date,3,4 Mu pseudoviruses were significantly more resistant to convalescent serum-mediated neutralization than Beta pseudoviruses (P=0.031;
3
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license .
|
72 73 74 75
Figure 1B). Thus, the Mu variant shows a pronounced resistance to antibodies elicited by natural SARS-CoV-2 infection and the BNT162b2 mRNA vaccine. Since breakthrough infections are a major threat of newly emerging SARS-CoV-2 variants,5 we strongly suggest to further characterize and monitor the Mu variant. 4
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license . 76 77 78 79 80 81 82 83 84 85 86
References 1. https://www.who.int/en/activities/tracking-SARS-CoV-2-variants/. 2021. 2. mutations and immune escape. Nat Rev Microbiol 2021;19:409-24. 3. to mRNA vaccine-elicited antibodies. Nature 2021;593:136-41. 4. B.1.351 and B.1.1.7. Nature 2021;593:130-5. 5. with SARS-CoV-2 variants. N Engl J Med 2021;384:2212-8.
WHO. “Tracking
SARS-CoV-2
variants”. Harvey WT, Carabelli AM, Jackson B, et al. SARS-CoV-2 variants, spike
Collier DA, De Marco A, Ferreira I, et al. Sensitivity of SARS-CoV-2 B.1.1.7
Wang P, Nair MS, Liu L, et al. Antibody resistance of SARS-CoV-2 variants
Hacisuleyman E, Hale C, Saito Y, et al. Vaccine breakthrough infections
5
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license . 87
88 89 90 91 92 93 94 95 96
Figure 1. Characterization of the Mu variant. (A) SARS-CoV-2 epidemic in Colombia. New COVID-19 cases per day (black line, left y-axis) and percentage of different SARS-CoV-2 variants spreading in Colombia (right y-axis) are shown. The daily frequency of Gamma (P.1), Delta (B.1.617.2, AY.4, AY.5, AY.12), Lambda (C.37), Mu (B.1.621), and other variants are shown in the indicated colors. Note that there are a few Delta VOC (the currently most dominant variant in the world) and Lambda VOI (a variant mainly spreading in South American countries) have been isolated in this country so far. The date when the Mu variant
6
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made
doi:
https://doi.org/10.1101/2021.09.06.459005
;
this version posted September 7, 2021.
available under a
CC-BY-NC-ND 4.0 International license . 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112
was first isolated (January 11, 2021) is indicated in the figure. The raw data are summarized in Table S2 in the Supplementary Appendix. (B and C) Virus neutralization assays. A neutralization assay was performed using pseudoviruses harboring the SARS-CoV-2 spike proteins of the Alpha, Beta, Gamma, Delta, Epsilon, Lambda, Mu variants or the D614G-harboring parental virus.
|
Eight COVID-19 convalescent sera (B) and ten sera from BNT162b2-vaccinated individuals (C) were tested. The assay of each serum was performed in triplicate to determine the 50% neutralization titer, and each data point represents the 50% neutralization titer obtained with a serum sample against the indicated pseudovirus. The bar graphs indicate geometric mean titers with 95% confidence. The numbers over the bars indicate geometric mean titers. The numbers over the bars in parentheses (with "X") indicate the average of fold change in neutralization resistance of the indicated spike variants compared to that with the parental spike in each serum. Statistical analysis was performed with the use of the Wilcoxon signed-rank test. Horizontal dashed lines indicate limit of detection. The raw data are summarized in Tables S6 and S7 in the Supplementary Appendix. 7
|
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. The Visual Dictionary of Antimicrobial Stewardship, Infection Control, and Institutional Surveillance
Julia Keizer1*, Christian F. Luz2*, Bhanu Sinha2, Lisette van Gemert-Pijnen1, Casper Albers3, Nienke Beerlage-de Jong4†, Corinna Glasner2†
1 University of Twente, Department of Psychology, Health and Technology, Centre for eHealth and Wellbeing Research, PO box 217, 7500AE Enschede, The Netherlands
2 University of Groningen, University Medical Center Groningen, Department of Medical Microbiology and Infection Control, Hanzeplein 1, 9700RB Groningen, The Netherlands
3 University of Groningen, Heymans Institute for Psychological Research, Grote Kruisstraat 2/1, 9712 TS Groningen, The Netherlands
4 University of Twente, Department of Health Technology and Services Research, Technical Medical Center, Faculty of PO box 217, 7500AE Enschede, The Netherlands
,† Equal contribution
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Abstract
Objectives
Data and data visualization are integral parts of (clinical) decision-making in general and
stewardship (antimicrobial stewardship, infection control, and institutional surveillance) in
particular. However, systematic research on the use of data visualization in stewardship is
lacking. This study aimed at filling this gap by creating a visual dictionary of stewardship
through an assessment of data visualization in stewardship research. Methods
A random sample of 150 data visualizations from published research articles on stewardship
were assessed. The visualization vocabulary (content) and design space (design elements)
were combined to create a visual dictionary. Additionally, visualization errors, chart junk, and
quality were assessed to identify problems in current visualizations and to provide
improvement recommendations. Results
Despite a heterogeneous use of data visualization, distinct combinations of graphical elements
to reflect stewardship data were identified. In general, bar (n=54; 36.0%) and line charts (n=42;
28.1%) were preferred visualization types. Visualization problems comprised colour scheme
mismatches, double y-axis, hidden data points
through overlaps, and chart
Recommendations were derived that can help to clarify visual communication, improve colour
use for grouping/stratifying, improve the display of magnitude, and match visualizations to
scientific standards. Conclusions
Results of this study can be used to guide data visualization creators in designing
visualizations that fit the data and visual habits of the stewardship target audience.
|
Additionally, the results can provide the basis to further expand the visual dictionary of
stewardship towards more effective visualizations that improve data insights, knowledge, and
clinical decision-making. junk. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Introduction
The amount of and reliance on data increases with the increase of scientific publications and
information technologies in healthcare practice [1,2]. The increased complexity of big data
raises various issues to be resolved by innovative big data analytics. This includes integrating,
analysing and visualizing data to translate into meaningful information [3,4]. Translating raw
data into meaningful information and communicating it to specific target groups is a challenge
[1]. Without this translation and communication, researchers and practitioners cannot
optimally use the information, so that the true value of the data remains hidden. Data
visualization, here defined as the graphical representation of quantitative information, can
facilitate the transformation, memorisation, and communication of data to understandable and
actionable information. Data visualization also aids in the interpretation of increasingly large
and complex datasets (big data) and in the understanding of sophisticated statistical models
(machine learning) and their results - two rising trends over the last decades [5,6]. The
importance of data visualization can, once again, be observed in the COVID-19 pandemic with
the ubiquitous presence of charts, figures, and dashboards that aim to inform and support
decision-making for a wide variety of target audiences [7]. Data visualization is a very active (research) field in itself and is generally part of typical
statistical software used in the data analysis process (e.g. R, SPSS, SAS, STATA, Excel). Information and recommendations for the data visualization process are numerous and can
be transferred between research fields or domains [8–11]. However, research on the visual
domain context within a research field is often lacking, i.e. what the target audience is
accustomed to see and expects in terms of content and design, and how this influences the
perception and interpretation of data visualizations from different perspectives [12]. Common
data visualization practices in a specific domain can be identified by studying the visualization
design space [13]. This visual design space can be described as “an orthogonal combination
of two aspects”, namely marks (i.e. graphical elements such as points, lines and areas) and
visual channels to control their appearance (i.e. aesthetic properties such as colour, size and
shape) [13]. To clarify the conceptual definitions for assessing and describing data visualizations a
linguistic analogy can be used: a dictionary describes language in terms of both vocabulary
(i.e.
|
the set of words familiar in a language) and grammar/punctuation (i.e. the set of structural
rules and supporting marks that control the composition and navigability of sentences,
phrases, and words). Similarly, the visual dictionary describes visualizations in terms of both
visual vocabulary (i.e. the domain content in terms of visualized data attributes) and visual
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. design space (i.e. graphical elements and supporting aesthetic properties). The language or
visual domain context is an overarching concept that represents language/visualization in
practice, i.e. expectations and customs of the target audience, and how this affects their
perception and interpretation of data visualizations (see also Figure 1). The visual domain
context is, just as language, subject to changes over time and subject to interpretation
differences based on varying perspectives. Figure 1. Conceptual framework used in this study to clarify the definitions and interrelations between
the visual domain context, the visual dictionary and the visual domain vocabulary and visual design
space. Data and data visualization play important parts in the field of infectious diseases and
antimicrobial resistance (AMR) for the reporting on the growing burden on health and
healthcare systems [14,15]. Comprehensible and actionable information on antimicrobial
consumption, pathogen distribution, or incidence and prevalence of (multi-) drug resistant
microorganisms are vital to design interventions to tackle the AMR challenge [16]. On the
hospital level, antimicrobial and diagnostic stewardship, infection control, and institutional
surveillance (further summarised under ‘stewardship’) are the core components of strategies
that promote the responsible use of antimicrobials and improve the quality and safety of patient
care [17,18]. Data visualization is an integral part of these strategies, as it unveils the local
situation and the drivers of AMR and can have a significant impact on the use of antimicrobials
[19,20]. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. It has been shown how important it is to study data and data visualization experiences and
perceptions in the medical domain and how this can influence the interpretation of data [21,22]. Depending on the type of visualization used, identifying the key messages from the data
visualization can be substantially hindered.
|
The audience’s background and its familiarity with
data visualization (the visual domain context) have to be taken into account in the design
process to avoid these obstacles. Example studies that identified the visual domain context
by studying the design space can be found in the field of genomic epidemiology and genomic
data visualization [23,24]. Although, some recommendations and best practices exist that are
helpful in the data visualization creation process, common data visualizations practices in the
field of stewardship have yet to be revealed [25,26]. The visual domain context and the use of
data visualization in the field are unstudied - a systematic approach to define the design space
is missing. In this study, we aim to fill these gaps by assessing and defining the design space of data
visualization in stewardship and to create a visual dictionary. The results of this study can help
data visualization creators, such as healthcare/AMR/data professionals and scientists, to
anticipate the visual domain context of the target audience and link it with existing
recommendations for the data visualization process. This could benefit both research and
clinical decision-making in the translation and communication of data to understandable and
actionable information needed to tackle the AMR challenge, thereby improving the quality and
safety of health and healthcare. Methods
Study data
This study was based on a previous mapping study that clustered the field of AMR into 88
topics [27]. The map was generated by assessing the entire body of AMR literature available
on PubMed between 1999 and 2018 consisting of 152780 articles. The identification of the 88
topics within the field was performed based on the title and abstract text using a machine
learning algorithm (STM) [28]. The present study used all articles of three of the identified
topics: stewardship (n = 3383 articles), infection control (n = 1687 articles), and institutional
surveillance (n = 2176 articles). Within the corpus of the 88 topics, these three topics reflect
the core components of an integrated, comprehensive stewardship concept in institutional
healthcare as defined above [18]. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. For each topic, a sample of 60 articles that contained at least one data visualization was
randomly drawn. Data visualization was defined as the graphical representation of quantitative
data. Geographical maps and flowcharts were excluded. From the sampled articles, one
visualization per article was randomly sampled resulting in 180 data visualizations. The study
design is shown in Figure 2. To analyse reliability, ten randomly picked data visualizations of each topic were analysed in
duplicate to calculate the inter-rater reliability (joint probability of agreement)
Subsequently, 150 visualizations were included in the final analysis.
|
Figure 2. Study design. IRR = inter-rater reliability
[29]. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Data visualization analysis
The extracted data visualizations were analysed based on the nomenclature and
categorization developed by Munzner and further adapted for this study [13]. This approach
dissected data visualizations into visual characteristics:
Attributes (or variables, parameters, features): the underlying data labelled as
categorical, ordered, or quantitative. Marks: the basic geometric element (points, lines, or areas)
Channel: channels control the visual appearance of marks
Position: horizontal, vertical, both
Colour
Shape
Tilt
Size: length, area, volume
Channel effectiveness
Magnitude: ordered attributes can be expressed in ranks from most effective to
least effective: position on common scale (most effective) > position on
unaligned scale >
length >
tile/angle > area > depth > colour
luminance/saturation > curvature/volume (least effective)
Identity: the effectiveness to express categorical attributes can also be ordered:
colour hue > shape
In addition, data visualizations were labelled with the visualization type used (e.g., bar chart,
line chart, scatter plot, etc.) and the use of faceting (multiple linked visualizations in a design
grid). Each visualization was assessed upon its interpretability without additional text (yes, if
interpretable without additional information; partially, if a description was given in a caption;
not all, if a description was absent or only available in the article text). Visualization quality was captured by rating the first and last impression during the analysis
process on a scale form 1 (poor) to 5 (good). The choice of the visualization type given the
underlying data was rated on a scale from 1 (poor) to 5 (good). In addition, free, written text
was recorded to capture comments and remarks about the data visualization. A structured assessment form (supplementary materials S1) was developed comprising all the
above mentioned elements. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. The form was discussed within a multidisciplinary team of data-visualization and AMR experts. The assessment form was applied to each data visualization in a two reviewer (JK, CFL)
process. First, the assessment form was used for training the analysis process with ten data
visualizations not part of the final study data. Next, each reviewer analysed 50% of the study
data visualizations followed by a re-review through the other researcher.
|
Consensus was
reached through discussion if the first assessment differed. Quantitative analysis
Results from the data visualization analysis step that were analysed with descriptive statistics
comprised visualization type, number of attributes, faceting, rating, and visualization type
choice. Attributes were analysed for pairwise co-occurrence and presented if a combination
occurred more than twice in total. Visual dictionary
The visual dictionary was created based on the visual vocabulary content (stewardship-related
data) and the visual design space (characteristics used to design the visualization). The
content was analysed by identifying the attributes and grouping the attribute names using
inductive coding. This part built the vocabulary of visualized stewardship data. Next, stratified
quantitative analyses of visual characteristics (channel, marks, etc.) per attribute were
performed, thereby adding the visual design space to the vocabulary to create the visual
dictionary. Linking text and visual vocabulary enabled the creation of a visual dictionary to help
identify attributes (e.g., resistance) with associated channels (e.g., points and lines on a
common scale). Qualitative analysis
Comments and remarks about the data visualizations were coded in Microsoft Excel by two
researchers (CL and JK). An open coding round was followed by axial coding to discover
related concepts in the sub-codes. Differences were discussed until consensus was reached,
which increased the internal validity [30]. Next to improvements, CL and JK coded remarks
about chart junk (i.e. the unnecessary and/or redundant use of visualization embellishments)
[11]. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Results
In total, 150 visualizations were analysed (IRR: 87% joint probability of agreement). The
following sections are separated into visualization vocabulary (content) and visual dictionary
with results stratified by identified attributes. These sections are followed by visualization
ratings,
identified visualization problems
(including chart
junk), and suggested
recommendations for visualization creators and users. Visual vocabulary (content)
In total, 48 different coded attributes were identified. The majority (54.7%) of visualizations
used three attributes. Two or four attributes were used in 18.0% and 20.7% of all
visualizations, respectively. Few of the studied visualizations (6.7%) used more than 4
attributes. The ten most used attributes were time (n=69, 46.0%), setting (n=43, 28.7%), antimicrobial
consumption (n=32, 21.3%), resistance (n=31, 20.1%), antimicrobials (n=27, 18.0%),
percentage (n=26, 17.3%), count (n=24, 16.0%), incidence (n=24, 16.0%), numeric value
(n=20, 13.3%), and bacteria (n=12, 8.0%).
|
Attributes can be grouped into objects (e.g. bacteria) and measurements (e.g. percentage). However, the following analysis focuses on
attribute combinations and attributes are thus kept ungrouped. Attributes showed different co-occurrence patterns (Figure 2). The ten most frequent
combinations were time and antimicrobial consumption (n=21), time and incidence (n=18),
antimicrobial consumption and antimicrobials (n=12), antimicrobials and resistance (n=12),
time and resistance (n=12), time and antimicrobials (n=11), antimicrobial consumption and
setting (n=10), resistance and setting (n=9), time and setting (n=9), and percentage and
setting (n=8). bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Figure 3. Attribute combinations in visualizations (combination count ≥ 3), thickness of lines
corresponds to combination count. Orange points and labels represent attributes related to
measurements; blue points and labels represent attributes related to objects. Visual dictionary
Visualization types
Fourteen different visualization types were identified of which bar charts (n=54, 36.0%) and
line charts (n=42, 28.1%) were predominantly used. Bar charts were most frequently
associated with the attributes antimicrobials, bacteria, cohorts, compliance, counts, diagnosis,
errors, percentages, resistance, setting, and survey answers. Line charts were predominantly
associated with antimicrobial consumption, costs, cut-off, incidence, numeric values,
regression, statistics, and time (detailed results available in the supplementary materials S2)
Different visualization types combined in one visualization were used in 10.7% (n = 16) of all
visualizations. In these, visualization types that were combined more than once were bar
charts with line charts (n=5, 31.3%) and stacked bar charts with line charts (n=2, 12.5%). In
41 visualization (27.3%) facets were used, i.e., one visualization split into a matrix of
visualizations using the same axes. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Visual design space
Different patterns of visual characteristics could be identified for different attributes (detailed
counts and percentages in supplementary materials S3). 1. Position: Horizontal axes were mostly used for Antimicrobials, bacteria, confidence
intervals, counts, cut-offs, diagnoses, events, numeric values, settings, similarity, and
time. In contrast, vertical axes were mostly used for antimicrobial consumption, cases,
cohorts, counts, errors, incidence, percentages, regression, resistance, samples,
statistics, and survey answers.
|
2. Marks, colour, shape: Attributes also differed in their use of marks. Some attributes
had clear associations with mark types, e.g. time was visualised with lines in all
instances. Areas as marks were seldomly used, e.g. for antimicrobial consumption,
counts, cut-offs, incidence, numeric values, percentages, and resistance. Colour and
shape channels were frequently used in most attributes. A detailed colour and shape
channel analysis is available in the supplementary materials S3. 3. Size: Size was most often visually reflected through length. Area to reflect size was
used for antimicrobial consumption, count, cut-off, incidence, numeric values,
percentages, and resistance. Volume was rarely used (count, percentages). 4. Ordering: Position on a common scale was mostly used in quantitative and ordered
attributes reflecting the best channel effectiveness for these attribute types. Categorical attributes mostly used colour hue, which is preferred over the less effective
use of shapes. A detailed channel effectiveness analysis is available in the
supplementary materials S4. Ratings, problems, and chart junk
Visualization ratings
Overall, 55.3% (n=83) of all visualizations were interpretable without additional text (in caption
or in the manuscript text). The overall choice of visualization type for the presented data was
rated with a mean of 4.62 (SD: 0.9) on a scale from 1 (poor) to 5 (good). The general
assessment of the visualization quality (scale 1=poor to 5=good) was rated with a mean of 3.6
(SD:1.2). Identified problems (and recommendations) are described below. Identified problems
The coding of the identified problems are presented in the coding scheme in Table 1, including
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. axial codes, open codes and frequencies. The axial and open codes are further elaborated
upon below the table. In supplementary materials S5, additional illustrative quotes per code
are presented. Table 1. Identified problems in data visualization. Code (axial) Missing labels, annotations, legend and/or abbreviation explanations
Code (open) Legend/caption is missing/unclear Labels for lines/points are missing/unclear Labels for axes are missing/unclear Annotation/direct labelling overflow Abbreviations not explained Use of colours not explained
Count 26 23 20 14 12 7
Subtotal 102
Axes not readable
Axis intervals uneven (within visualization and between faceted visualizations) Axes texts not clearly readable Too short/dense axes/intervals Uneven bar placement Axis intervals not logical (within visualization and between faceted visualizations)
17
11 5 1 1
Colour scheme mismatch Groups not distinguishable by colours
Non-intuitive colour schemes used Categorical colours used for ordered attribute Groups not distinguishable from background
Hidden data points by overlaps
Overlap of shapes problematic
Subtotal 35 14 6 5 2 Subtotal 27 7
Using suboptimal channel effectiveness
Groups not distinguishable by shapes Sub-effective channel is chosen
Subtotal 7 12 3 Subtotal 15
Size scale indistinguishable Differences in size not clear
Groups not distinguishable by shape size Contrasts between groups not clear
Missing channel
Line types not used to distinguish between groups Colours not used to compare between visualization/groups
10 3 2 Subtotal 15 2 2 Subtotal 4 21
Visualization type does not (optimally) fit data
Other visualization type preferred
Data points/lines on double axes
Double Y-axes difficult to read
Subtotal 21 11
Subtotal 11
Channel overflow
Double use of shape and colour Unnecessary use of shape sizes Unnecessary use of colour Too many colours
8 1 1 1 Subtotal 11
Attribute overflow
Too many attributes Relating attributes that are not related
2 1 Subtotal 3
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
|
It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Information sparsity
Incoherent ordering
Could be text
Data not ordered coherently
1 Subtotal 1 1 1
Subtotal
Grand total 253
Most problems were related to the clarification of the visualization, because of missing or
unclear labels, annotations, legend, captions and/or abbreviation explanations (n=102;
68.0%). Other problems concerned the axis readability (n=35; 23.3%) for example due to
uneven axis-intervals within a visualization and between
faceted visualizations
comparison. Problems related to distinguishing data points and groups in visualizations were
detected, for example with mismatches in colour scheme (n=27; 18.0%), hidden data points
by overlaps (n=7; 4.7%), and using suboptimal channel effectiveness (n=7, 4.7%). In some
cases, data points and groups were not clearly distinguishable because of the size scale
(n=15; 10.0%) or because of missing channels (e.g. line type or colour, n=4; 2.7%). Furthermore, problems identified were the suboptimal or wrong choice in type of visualization
(n=21; 14.0%) and the confusing use of double y-axis (n=11; 7.3%). Some visualizations were
overcrowded, either in terms of channel overflow (e.g. using both colour and shape, n=11;
7.3%) or attribute overflow (e.g. too many attributes, n=3; 2.0%). On the contrary, the
information in one visualization was sparse enough to be written in text (i.e. no added value
of a visualization). Lastly, one problem related to the incoherent ordering of data. Chart junk
Most chart junk represented text that cluttered the visualization (n=8), for example with
redundant direct labels for each data point. Other chart junk was found in visualizations using
unnecessary 3D (n=8), background colours (n=6), shadow (n=4), and colour/shape filling
(n=4). Examples and recommendations
To illustrate problems in data visualization, we designed a visualization that exhibits several
of the identified problems based on simulated data (Figure 4). Figure 5 proposes an alternative
to Figure 4 where the identified problems were avoided. Of note, data such as the simulated
data in these figures can be visualised in many different ways, depending on the underlying
research questions. for
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Figure 4. Resistance to amoxicillin in Escherichia coli and consumption of cefuroxime (black) and
piperacillin/tazobactam (blue) across hospital departments in 2020. This data visualization (simulated
data) shows several problems identified in this study: Axes not starting at zero, use of double y-axes,
background colours, hidden data points by overlaps, colour scheme mismatch (blue and black difficult
to distinguish), unequal axis steps on x-axis, missing legend, incomplete axis labels (abbreviation not
explained).
|
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Figure 5. Resistance to amoxicillin in Escherichia coli and consumption of cefuroxime and
piperacillin/tazobactam across hospital departments in 2020. These data visualizations use the same
data as in Figure 4 (simulated data), but propose an improved visualization. Figure 6 summarises the results of this study and presents the visual dictionary of stewardship. In addition, it provides a set of recommendations to avoid the most common problems in data
visualizations as identified in this study. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Figure 6. The visual dictionary of stewardship (antimicrobial stewardship, infection control, and institutional surveillance). bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Discussion
This study systematically analysed the visual domain context of stewardship (i.e. antimicrobial
stewardship, infection control, and institutional surveillance). Stewardship healthcare experts
and scientists that create data visualizations can benefit from the revealed visual domain
context, since it allows them to anticipate the visual habits of their target audience. The results
of this study can serve as the basis to inform visualization creators to optimise visual
communication in the field and to guide user-centred design, e.g., in clinical decision support
systems. Implications for data visualization creators
With the systematic analysis of the visual domain context of Stewardship we revealed common
practices and identified problems with current visualizations. In this study we identified 14
different visualization types used in the visual domain context of the field. However, more than
80% of all visualizations used classical (stacked) bar or line charts; quite homogenous design
choices. We argue that the visualization type choice is based on tradition and habits as a
systematic approach to data visualization in the field was missing until now [26]. For
researchers in the field that communicate their findings to various stakeholders (e.g. stewardship professionals, policy makers, epidemiologists) the described visual domain
context in this study provides guidance to match their visualizations with the audience’s visual
expectations and habits.
|
Especially the visual dictionary, the link between often used attributes
(i.e. content) and associated design choices (e.g. visualization type or marks), will help to
compose visualizations that fit with common practice. However, given the wide variety of data
in the field and the increased complexity that big data will add (in terms of volume, velocity,
variety, veracity, validity, volatility and value), more “visual variability” might be expected and
even needed in the future [3,31,32]. Informing and teaching visualization creators and users
about data visualization design alternatives is an important step in this process. A lack of
awareness and knowledge about data visualization design alternatives might lead to
suboptimal data visualizations. Examples from our findings were the use of less effective
visual channels, suboptimal plot types for the presented data, or mismatches in colour choices
for different data types. These are examples of instances where the respective data
visualization creators require more support in visualization design choices. We see a clear
role here for data visualization experts and software developers to cocreate open-
source/access tools that support visualization creators in their visualization choices (e.g. reminders for adding labels and legends, suggestions for optimal colour schemes, warnings
in case of chart junk). Our results and findings from similar studies in other fields [23,24] can
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. support them in doing so by providing an overview of what is already used, including potential
pitfalls. Of note, academic journals play an important part in this process by providing the
platform for data visualizations and should be encouraged to promote high quality data
visualization practices. Improving data visualization practices
In general, the use of data visualizations for communicating data is highly encouraged. It
greatly supports the interpretation, memorisation, and communication of insights and
knowledge gained from data. Based on the identified problems with data visualizations in this
study, several recommendations can be made to improve data visualization practices both in
general and for stewardship specifically. Some recommendations relating to identified
problems were already depicted in Figure 5 and 6 and are elaborated and extended upon
below. Colours
The use of colours in data visualization is highly complex. Colours can make a plot more
appealing. Colours are also more effective than shape to distinguish categorical data [13]. Yet, shapes for categorical data were still widely used in the studied visualizations. This could
reflect the need to provide visualizations that are black-and-white compatible (printable),
although this has become less important with most of the scientific content being accessible
online.
|
Several aspects are key to consider when designing data visualizations with colour:
colour-blindness, distortion through uneven colour gradients, or the perceived order of colours
[33]. While field-specific colour codes might exist (e.g., red colour to represent resistance),
general recommendations for the use of colours in scientific publications are available and are
applicable across fields [33]. Extensive information on the use of colours in data visualization
can also be found in online blogs from designers in the data visualization community (e.g.,
https://blog.datawrapper.de/which-colour-scale-to-use-in-data-vis/). Adding statistics
Common scientific visualization types such as heatmaps or boxplots were rarely used in the
studied data visualizations. In general, statistical aggregate parameters were often lacking. This would often have improved the visualizations under study. Boxplots are a classical
example. However, this visualization type is rightfully criticised to conceal individual data
points and could be misleading [34]. We also identified difficulties in displaying individual data
points as one of the main problems. This problem was caused by overlaps, problematic scale
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. sizes, or missing channels to distinguish data points. We highly advise to stick to the mantra
of “above all show the data”, to avoid overlapping or concealing graphical elements [11], and
to carefully balance the number of visualised attributes per visualization with simplicity. Standardizing the visual dictionary
Although a large variety of data were displayed in the studied visualizations (48 different
attributes such as antimicrobials, bacteria, or time), we observed some prominent patterns in
the content and purpose of data visualization. Changes over time, e.g. time series, were part
of 43.3% of all studied visualization. Twenty percent and 19.3% included antimicrobial
consumption or resistance, respectively. This is not surprising given the importance of these
data in the field. It could be worth considering to standardise data visualizations for these data
types and contents similar to the consensus of international guidelines committees in the field,
e.g. the European Committee of Antimicrobial Susceptibility Testing (EUCAST) or the Clinical
& Laboratory Standards Institute (CLSI). This could help ensure high quality data
visualizations for reliable insights in AMR- and stewardship-related data. Such initiatives to
standardise data visualizations have already been taken by bodies in other fields, e.g. the
Intergovernmental Panel on Climate Change (IPCC) [35]. Another example is the development
and evaluation of a standardised medical data visualization method based on the ISO13606
data model [36].
|
Limitations and strengths
This study has several limitations. Despite sampling from a comprehensive set of articles that
cover the field of stewardship (antimicrobial stewardship, infection control, and institutional
surveillance), only a limited sample of data visualizations were included. Data visualizations
for content (attributes) not covered in this study could have been missed. However, the
homogeneity of the identified data visualization types suggests that saturation was reached
regarding the visual design space in the field. Another limitation is that we included data
visualizations from scientific publications and not from other sources relevant to stewardship
data visualizers (e.g. data systems used in practice [12,37] and AMR policy reports [38,39]). As a result, our findings might be more applicable to stewardship researchers and data
visualization experts than healthcare professionals. Another use of data visualization, namely
the exploration of increasingly complex big data, was outside the scope of the included articles
[40,41]. Subsequent research into the visual domain context of stewardship should include
these additional data visualization sources and applications to ensure a more comprehensive
picture for healthcare professionals. Even though the extracted data visualizations were
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. systematically analysed using a structured assessment form based on existing data
visualization nomenclature and categorization [13], the analyses relied on the subjective
interpretation and rating by the coding researchers. Several measures were taken to validate
our findings, including discussing the assessment form and results within a multidisciplinary
team of data-visualization and AMR experts, analysing the interrater-reliability, and comparing
our findings to other data visualization studies. Our study is one of the first empirical studies
that explores the use of data visualization in stewardship, thereby adding to the few review
studies providing primers for data visualization recommendations and best practices in the
field of antimicrobial stewardship, infection control, and institutional surveillance [25,26]. Future perspectives
Future research can build upon our results by studying and expanding the use and design of
data visualizations beyond the basic visual dictionary provided here. Two important future
research directions are elaborated upon below. Studying the visual domain context is as important as studying data visualizations themselves. This includes studying expectations and customs of the target audience, how this affects their
perception and interpretation of data visualizations, and how this consequently impacts their
decision-making or behaviour.
|
The importance of assessing visual habits and perceptions in
data visualization has been demonstrated before in other medical fields [21,42]. It was shown
that personal preferences and the familiarity of a target audience with certain visualization
types can result in tensions with data visualization recommendations and novel data
visualization approaches. An example is the use of pie charts, which is often discouraged in
the data visualization community. The target audience might still favour this type of
visualization because of its apparent simplicity, despite the fact that pie charts are less
accurately interpreted as angles and wedges are difficult to compare [8,43]. We strongly
believe that incorporating best practices and data visualization recommendations are essential
but advocate that these should be carefully balanced with visual habits and expectations in
the field, and the message to be transported. An exemplary study is published by Aung et al. focusing on data visualization interpretation capacity and preferences in their target audience
by combining interviews on interpretability and card-sorting of preferred visualizations [21]. Additionally, research is needed to better understand how data visualizations impact the
viewers/users in terms of changes in opinions or attitudes that direct decision-making or
behaviour changes [44]. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. In future research special attention should be paid to matching the visual dictionary and the
context in which the visualization will be used in terms of users and their tasks and current
practices (e.g. studying questions like “What kinds of visualizations are currently used?” and
“How do they support to do current tasks?”) [45]. We see a clear parallel with user-centred
eHealth design that emphasises the need of a holistic understanding of the interrelations
between technology, people and their context [46]. Both qualitative (e.g. interviews) and
quantitative (e.g. eye-tracking in current data visualizations) study designs can contribute to
such a holistic understanding, which in turn can inform or improve the design of visualizations
(or eHealth) in terms of required content, functionalities and usability [47]. Therefore,
complementing research on data visualizations, as the current study and many other studies
do, with research that primarily focuses on the interaction between people, their context and
how data visualizations can support them, is needed [45]. Conclusion
In this study, we analysed the visual domain context of stewardship (antimicrobial stewardship,
infection control, and institutional surveillance). We successfully created a visual dictionary
that can support the process of creating and using tailor-made data visualizations in the field.
|
Thereby, our results allow data visualization creators to learn the visual language of the
diverse field of stewardship. As data-driven solutions for stewardship are of increasing
importance, effective processes of transforming this data to insights and knowledge is
essential. Data visualization supports and enables this transformation and our results can
guide the optimal visualization design choices that are grounded on expectations and habits
in the field. In the future, our study can provide the basis to further expand the visual dictionary
of antimicrobial stewardship towards more effective data visualizations that improve data
insights, knowledge, and decision-making. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Acknowledgement
The authors would like to thank Anamaria Crisan, Tamara Munzner, and Nils Gehlenborg for
the inspiration provided for this study. Funding
This research was supported by the INTERREG-VA (202085) funded project EurHealth-
1Health (http://www.eurhealth1health.eu), part of a Dutch-German cross-border network
supported by the European Commission, the Dutch Ministry of Health, Welfare and Sport, the
Ministry of Economy, Innovation, Digitalisation and Energy of the German Federal State of
North Rhine-Westphalia and the Ministry for National and European Affairs and Regional
Development of Lower Saxony. In addition, this study was part of a project funded by the
European Union's Horizon 2020 research and innovation programme under the Marie
Sklodowska-Curie grant agreement 713660 (MSCA-COFUND-2015-DP "Pronkjewail"). References
[1] Murdoch TB, Detsky AS. The inevitable application of big data to health care. JAMA
2013;309:1351–2. https://doi.org/10.1001/jama.2013.393. [2] Wang Y, Kung L, Byrd TA. Big data analytics: Understanding its capabilities and potential benefits for healthcare organizations. Technol Forecast Soc Change 2018;126:3–13. https://doi.org/10.1016/j.techfore.2015.12.019. [3] Khan MA-U-D, Ali-ud-din Khan M, Uddin MF, Gupta N. Seven V’s of big Data
understanding big data to extract value. Proceedings of the 2014 Zone 1 Conference of the American Society for Engineering Education, Bridgeport, USA; 2014. https://doi.org/10.1109/aseezone1.2014.6820689. [4] Ambigavathi M, Sridharan D. Big data analytics in healthcare. 2018 Tenth International
Conference on Advanced Computing (ICoAC), Chennai, India; IEEE; 2018. https://doi.org/10.1109/icoac44903.2018.8939061. [5] Adam Bohr KM. The rise of artificial intelligence in healthcare applications. Artificial
Intelligence in Healthcare 2020:25. https://doi.org/10.1016/B978-0-12-818438-7.00002- 2. [6] Bailly S, Meyfroidt G, Timsit J-F. What’s new in ICU in 2050: big data and machine learning.
|
Intensive Care Med 2018;44:1524–7. https://doi.org/10.1007/s00134-017- 5034-3. [7] Comba JLD. Data visualization for the understanding of COVID-19. Computing in Science Engineering 2020;22:81–6. https://doi.org/10.1109/MCSE.2020.3019834. [8] Evergreen SDH. Effective data visualization: the right chart for the right data. Los
Angeles: SAGE Publications; 2017. [9] Gatto MAC. Making research useful: current challenges and good practices in data
visualisation. Oxford: Reuters Institute for the Study of Journalism; 2015. [10] Kelleher C, Wagener T. Ten guidelines for effective data visualization in scientific
publications. Environmental Modelling & Software 2011;26:822–7. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. https://doi.org/10.1016/j.envsoft.2010.12.006. [11] Tufte ER. The Visual display of quantitative information. Cheshire: Graphics Press;
2001. [12] Sedrakyan G, Mannens E, Verbert K. Guiding the choice of learning dashboard
visualizations: linking dashboard design and data visualization concepts. Journal of Computer Languages 2019;50:19–38. https://doi.org/10.1016/j.jvlc.2018.11.002. [13] Munzner T. Visualization analysis and design. CRC Press; 2014. [14] OECD. Stemming the superbug tide. OECD Health Policy Studies 2018:224.
https://doi.org/10.1787/9789264307599-en. [15] World Health Organization. Global antimicrobial resistance surveillance system
(GLASS) report: early implementation 2020. Geneva: World Health Organization; 2020. [16] World Health Organization. Global action plan on antimicrobial resistance. Geneva:
World Health Organization; 2015. [17] Dyar OJ, Huttner B, Schouten J, Pulcini C, ESGAP (ESCMID Study Group for
Antimicrobial stewardshiP). What is antimicrobial stewardship? Clin Microbiol Infect 2017;23:793–8. https://doi.org/10.1016/j.cmi.2017.08.026. [18] Dik J-WH, Poelman R, Friedrich AW, Panday PN, Lo-Ten-Foe JR, van Assen S, et al. An integrated stewardship model: antimicrobial, infection prevention and diagnostic (AID). Future Microbiol 2016;11:93–102. https://doi.org/10.2217/fmb.15.99. [19] Luz CF, Berends MS, Dik J-WH, Lokate M, Pulcini C, Glasner C, et al. Rapid analysis of diagnostic and antimicrobial patterns in R (RadaR): interactive open-source software app for infection management and antimicrobial stewardship. J Med Internet Res 2019;21:e12843. https://doi.org/10.2196/12843. [20] Graber CJ, Jones MM, Goetz MB, Madaras-Kelly K, Zhang Y, Butler JM, et al. Decreases in antimicrobial use associated with multihospital implementation of electronic antimicrobial stewardship tools. Clin Infect Dis 2020;71:1168–76. https://doi.org/10.1093/cid/ciz941. [21] Aung T, Niyeha D, Shagihilu S, Mpembeni R, Kaganda J, Sheffel A, et al. Optimizing data visualization for reproductive, maternal, newborn, child health, and nutrition (RMNCH&N) policymaking: data visualization preferences and interpretation capacity among decision-makers in Tanzania.
|
Glob Health Res Policy 2019;4:4. https://doi.org/10.1186/s41256-019-0095-1. [22] Aung T, Niyeha D, Heidkamp R. Leveraging data visualization to improve the use of
data for global health decision-making. J Glob Health 2019;9:020319. https://doi.org/10.7189/jogh.09.020319. [23] Crisan A, Gardy JL, Munzner T. A systematic method for surveying data visualizations and a resulting genomic epidemiology visualization typology: GEViT. Bioinformatics 2019;35:1668–76. https://doi.org/10.1093/bioinformatics/bty832. [24] Nusrat S, Harbig T, Gehlenborg N. Tasks, techniques, and tools for genomic data
visualization. Comput Graph Forum 2019;38:781–805. https://doi.org/10.1111/cgf.13727. [25] Carroll LN, Au AP, Detwiler LT, Fu T-C, Painter IS, Abernethy NF. Visualization and
analytics tools for infectious disease epidemiology: a systematic review. J Biomed Inform 2014;51:287–98. https://doi.org/10.1016/j.jbi.2014.04.006. [26] Salinas JL, Kritzman J, Kobayashi T, Edmond MB, Ince D, Diekema DJ. A primer on data visualization in infection prevention and antimicrobial stewardship. Infect Control Hosp Epidemiol 2020;41:948–57. https://doi.org/10.1017/ice.2020.142. [27] Luz CF, van Niekerk JM, Keizer J, Beerlage-de Jong N, Braakman-Jansen LMA, Stein A, et al. Mapping twenty years of antimicrobial resistance research trends. Cold Spring Harbor Laboratory 2021:2021.03.01.433375. https://doi.org/10.1101/2021.03.01.433375. [28] Roberts ME, Stewart BM, Tingley D. stm: An R package for structural topic models. Journal of Statistical Software, Articles 2019;91:1–40. https://doi.org/10.18637/jss.v091.i02. [29] O’Connor C, Joffe H. Intercoder reliability in qualitative research: Debates and practical
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. guidelines. Int J Qual Methods 2020;19:160940691989922. https://doi.org/10.1177/1609406919899220. [30] Patton MQ. Enhancing the quality and credibility of qualitative analysis. Health Serv Res
1999;34:1189–208. [31] Galletta A, Carnevale L, Bramanti A, Fazio M. An innovative methodology for big data
visualization for telemedicine. IEEE Trans Ind Inf 2019;15:490–7. https://doi.org/10.1109/TII.2018.2842234. [32] Gotz D, Borland D. Data-driven healthcare: challenges and opportunities for interactive
visualization. IEEE Comput Graph Appl 2016;36:90–6. https://doi.org/10.1109/MCG.2016.59. [33] Crameri F, Shephard GE, Heron PJ. The misuse of colour in science communication. Nat Commun 2020;11:5444. https://doi.org/10.1038/s41467-020-19160-7. [34] Weissgerber TL, Winham SJ, Heinzen EP, Milin-Lazovic JS, Garcia-Valencia O,
Bukumiric Z, et al. Reveal, don’t conceal: transforming data visualization to improve transparency. Circulation 2019;140:1506–18. https://doi.org/10.1161/CIRCULATIONAHA.118.037777.
|
[35] Gomis MI, Pidcock R. IPCC visual style guide for authors. Stockholm:
Intergovernmental Panel on Climate Change; 2018. [36] Kopanitsa G, Veseli H, Yampolsky V. Development, implementation and evaluation of an information model for archetype based user responsive medical data visualization. Journal of Biomedical Informatics 2015;55:196–205. https://doi.org/10.1016/j.jbi.2015.04.009. [37] Huber TC, Krishnaraj A, Monaghan D, Gaskin CM. Developing an interactive data
visualization tool to assess the impact of decision support on clinical operations. J Digit Imaging 2018;31:640–5. https://doi.org/10.1007/s10278-018-0065-z. [38] Anderson M, Clift C, Schulze K, Sagan A, Nahrgang S, Ait Ouakrim D, et al. Averting the AMR crisis: what are the avenues for policy action for countries in Europe? Copenhagen: European Observatory on Health Systems and Policies; 2019. [39] World Health Organization. Central Asian and European Surveillance of Antimicrobial
Resistance. Geneva: World Health Organization; 2020. [40] Fisher D. Big data exploration requires collaboration between visualization and data infrastructures. Proceedings of the Workshop on Human-In-the-Loop Data Analytics, New York, NY, USA; Association for Computing Machinery; 2016,1–5. https://doi.org/10.1145/2939502.2939518. [41] Kaysi B, Kesler EK. Human Computer Interaction and Visualization Tools in Health Care
Services. In: Arabnia HR, Hashemi RR, G. TF, editors. Proceedings of the International Conference on Information and Knowledge Engineering (IKE), Athens, Greece; CSREA Press 2018;55–61. [42] Backonja U, Haynes SC, Kim KK. Data visualizations to support health practitioners’
provision of personalized care for patients with cancer and multiple chronic conditions: user-centered design study. JMIR Hum Factors 2018;5:e11826. https://doi.org/10.2196/11826. [43] Cleveland WS, McGill R. Graphical perception: Theory, experimentation, and
application to the development of graphical methods. J Am Stat Assoc 1984;79:531. https://doi.org/10.2307/2288400. [44] Pandey AV, Manivannan A, Nov O, Satterthwaite M, Bertini E. The persuasive power of
data visualization. IEEE Trans Vis Comput Graph 2014;20:2211–20. https://doi.org/10.1109/TVCG.2014.2346419. [45] Lam H, Bertini E, Isenberg P, Plaisant C, Carpendale S. Empirical studies in information
visualization: seven scenarios. IEEE Trans Vis Comput Graph 2012;18:1520–36. https://doi.org/10.1109/TVCG.2011.279. [46] van Gemert-Pijnen L, Kelders SM, Kip H, Sanderman R. eHealth research, theory and
development: a multi-disciplinary approach. Abingdon: Routledge; 2018. [47] Keizer J, Beerlage-De Jong N, Al Naiemi N, van Gemert-Pijnen JEWC. Finding the match between healthcare worker and expert for optimal audit and feedback on
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
.
|
antimicrobial resistance prevention measures. Antimicrob Resist Infect Control 2020;9:125. https://doi.org/10.1186/s13756-020-00794-7. Supplementary materials
S1. Data Visualization Assessment Form
Assessor
o Assessor 1
o Assessor 2
Data visualization ID
▼ ID1 … ID180
Year article
________________________________________________________________
First impression
o 1: poor
o 2
o 3
o 4
o 5: good
Viz type:
▢ Violin
▢ Density plot
▢ Boxplot
▢ Histogram
▢ Scatter plot
▢ Connected scatter plot
▢ Bubble plot
▢ Area plot
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. ▢ Stacked area plot
▢ Stream plot
▢ Line chart
▢ Ridge line
▢ Correlogram
▢ Heatmap
▢ Dendrogram
▢ Bar chart
▢ Stacked bar chart
▢ Lollipop chart
▢ Doughnut chart
▢ Treemap
▢ Circular packaging plot
▢ Venn diagram
▢ Sunburst diagram
▢ Spider plot
▢ Sankey diagram
▢ Network plot
▢ Chord diagram
▢ Arc diagram
▢ Hive plot
▢ Hierarchical edge bundling
▢ Other ________________________________________________
▢ Combined
Faceting (multiple groups in a grid)
o Yes
o No
Actions
▢ Analyze
▢ Search
▢ Query - Identify
▢ Query - Compare
▢ Query - Summarize
Attributes
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. o 1
o 2
o 3
o 4
o 5
o 6
Attribute names and types
Attribute name
Attribute type
Name
Categorical
Ordinal
Quantitative
NA
Mismatch
Attr. 1
▢
▢
▢
▢
▢
Attr. 2
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
Attr. 4
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
Marks
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Points
Lines
Areas
NA
Mismatch
Attr. 1
▢
▢
▢
▢
▢
Attr. 2
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
Attr. 4
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
Channels / position
Horizontal
Vertical
Both
NA
Mismatch
Attr. 1
▢
▢
▢
▢
▢
Attr. 2
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Attr. 4
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
Channels / colour - shape - tilt
Color
Shape
Tilt
NA
Mismatch
Attr.
|
1
▢
▢
▢
▢
▢
Attr. 2
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
Attr. 4
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Channels / size
Length
Area
Volume
NA
Mismatch
Attr. 1
▢
▢
▢
▢
▢
Attr. 2
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
Attr. 4
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
Channels effectiveness for ordered attributes :
Positi on on com mon scale
Position on unaligne d scale
Length (1D size)
Tilt/angle Area (2D
size)
Depth (3D position)
Color luminanc e / Color saturatio n
Curvatur e / Volume (3D size)
NA Mismatc
h
Attr. 1
▢
▢
▢
▢
▢
▢
▢
▢
▢
▢
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Attr. 2
▢
▢
▢
▢
▢
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
▢
▢
▢
▢
▢
Attr. 4
▢
▢
▢
▢
▢
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
▢
▢
▢
▢
▢
Channels effectiveness for categorical attributes:
Spatial position
Color hue
Motion
Shape
NA
Mismatch
Attr. 1
▢
▢
▢
▢
▢
▢
Attr. 2
▢
▢
▢
▢
▢
▢
Attr. 3
▢
▢
▢
▢
▢
▢
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Attr. 4
▢
▢
▢
▢
▢
Attr. 5
▢
▢
▢
▢
▢
Attr. 6
▢
▢
▢
▢
▢
Describe what is visualized:
________________________________________________________________
Interpretability without text
o 1 Not at all
o 2 Partially
o 3 Yes
Choice of plot type
o 1: poor
o 2
o 3
o 4
o 5: good
What can be improved? ________________________________________________________________
________________________________________________________________
________________________________________________________________
________________________________________________________________
________________________________________________________________
Remarks (e.g. chart junk)
________________________________________________________________
▢
▢
▢
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. ________________________________________________________________
________________________________________________________________
________________________________________________________________
________________________________________________________________
Last impression
o 1: poor
o 2
o 3
o 4
o 5: good
S2.
|
Visualization types per attribute
Table S2
Attribute
Line chart
Bar chart
Stacked bar chart
Dendrogram Forest
plot
Range plot
Bubble plot
Pie chart
Tornado plot
Scatter plot
Bma image plot
Heatmap Stacked area plot
Antimicrobial
31.03%
34.48%
24.14%
6.90%
3.45%
Antimicrobial consumption
46.15%
23.08%
25.64%
2.56%
2.56%
Bacteria
8.33%
58.33%
8.33%
8.33%
8.33%
8.33%
Case
16.67%
16.67%
66.67%
Confidence Interval
18.18%
27.27%
45.45%
9.09%
Cohort
20.00%
40.00%
20.00%
20.00%
Compliance
11.11%
77.78%
11.11%
Cost
40.00%
40.00%
20.00%
Count
14.81%
33.33%
18.52%
3.70%
7.41%
14.81%
7.41%
Cut-off
30.77%
15.38%
15.38% 15.38%
7.69%
7.69%
7.69%
Diagnosis
44.44%
33.33%
11.11%
11.11 %
Error
80.00%
20.00%
Event
25.00%
25.00%
25.00%
8.33% 16.67%
Incidence
37.04%
22.22%
22.22%
3.70%
7.41%
3.70%
3.70%
Numeric value
47.83%
26.09%
13.04%
4.35%
4.35%
4.35%
Correlogram
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. Percentage
7.41%
51.85%
22.22%
3.70%
11.11 %
3.70%
Regression
62.50%
12.50%
25.00%
Resistance
35.14%
40.54%
10.81%
2.70%
5.41%
5.41%
Sample
8.33%
8.33%
50.00%
16.67%
8.33%
Setting
15.22%
41.30%
19.57%
2.17%
8.70%
4.35%
2.17%
4.35%
2.17%
Similarity
80.00%
10.00%
Statistics
33.33%
33.33%
16.67%
8.33%
8.33%
Survey answer
50.00%
37.50%
12.50%
Time
52.63%
25.00%
10.53%
6.58%
2.63%
1.32%
1.32%
S3. Visual characteristics per attribute
Table S3
Channel - position
Channel - marks
Channel - colour/shape/tilt
Channel - size
Attribute
Horizontal Vertical
Both
Areas
Lines
Points
Color
Shape
Tilt
Area
Length Volume
Antimicrobial
72.73% 27.27%
64.71% 35.29%
Antimicrobial consumption
3.13% 96.88%
3.03% 48.48% 48.48%
33.33% 66.67%
6.25% 93.75%
Bacteria
71.43% 14.29% 14.29%
80.00% 20.00%
Case
20.00% 80.00%
Confidence interval
60.00% 30.00% 10.00%
81.82% 18.18%
100.00%
100.00%
Cohort
40.00% 60.00%
100.00%
75.00% 25.00%
Compliance
100.00%
83.33% 16.67%
100.00%
100.00%
Cost
40.00% 60.00%
50.00% 50.00%
100.00%
Count
23.81% 76.19%
5.26% 63.16% 31.58%
100.00%
18.75% 75.00% 6.25%
Cut-off
66.67% 11.11% 22.22%
12.50% 75.00% 12.50%
100.00%
50.00% 50.00%
Diagnosis
80.00% 20.00%
75.00% 25.00%
Error
40.00% 60.00%
100.00%
100.00%
100.00%
Event
100.00%
33.33% 66.67%
25.00% 75.00%
100.00%
Incidence
8.70% 91.30%
5.00% 45.00% 50.00%
28.57% 71.43%
9.09% 90.91%
Numeric value
55.00% 40.00% 5.00%
6.67% 40.00% 53.33%
33.33% 66.67%
Percentage
34.78% 60.87% 4.35%
12.50% 75.00% 12.50%
66.67% 33.33%
8.70% 78.26% 13.04%
Regression
71.43% 28.57%
71.43% 28.57%
33.33% 66.67%
Resistance
7.41% 92.59%
3.70% 51.85% 44.44%
77.78% 22.22%
7.14% 92.86%
Sample
100.00%
100.00%
Setting
60.87% 39.13%
33.33% 66.67%
59.38% 40.63%
Similarity
100.00%
100.00%
100.00%
Statistics
22.22% 77.78%
50.00% 50.00%
66.67% 33.33%
100.00%
Survey answer
40.00% 60.00%
100.00%
75.00% 25.00%
Time
100.00%
100.00%
55.56% 33.33% 11.11%
100.00%
8.33%
10.00%
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity.
|
It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. S4. Channel effectiveness
for quantitative, ordinal and
categorical attributes
Table S4
Effectiveness
Quantitative
Ordinal
Categorical
Position on common scale
82.8%
83.0%
Position on unaligned scale
1.1%
1.0%
Length (1D size)
3.4%
2.0%
Area (2D size)
6.7%
2.0%
Depth (3D position)
0.4%
Color luminance/ Color saturation
5.6%
12.0%
Spatial position
1.3%
Color hue
68.9%
Shape
29.8%
S5. Problems and illustrative quotes. Table S5. Problems and illustrative quotes
Code (axial) Missing labels, annotations, legend and/or abbreviation explanations
Code (open) Legend/caption is missing/unclear Labels for lines/points are missing/unclear Labels for axes are missing/unclear Annotation/direct labeling overflow Abbreviations not explained Use of colours not explained Not clear what colours mean in
Illustrative quotes Legend is missing. Line is not labeled and it is unclear what it represents. Y-axis label is missing. Difficult to read which sites are statistically significantly different. Abbreviations not explained. Count 26
23
20
14
12 7
visualization. Axes not readable
Subtotal 102 17
Axis intervals uneven (within visualization and between faceted visualizations) Axes texts not clearly readable Too much text in y-axis labels. Too short/dense axes/intervals Lines are cut above 100. Uneven bar placement
Plots use different y-axis, so it is difficult to compare them. 11 5 1
White separation lines not shown for each bar. X-axis intervals are not easy to read/compare. Axis intervals unlogical (within visualization and between faceted visualizations)
1
Colour scheme mismatch Groups not distinguishable by Use of similar colours, so hard to
Subtotal 35 14
bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. colours Non-intuitive colour schemes used
Categorical colours used for ordered attribute Groups not distinguishable from background
distinguish. Counter-intuitive coding (colour luminance levels for min max should be switched). Mismatch of categorical colours for order attribute. Colour for one group should not be the same as background (white). 6
5
2
Hidden data points by overlaps
Overlap of shapes problematic Points could be double, but not
Subtotal 27 7
differentiable because of overlap. Using suboptimal channel effectiveness
Groups not distinguishable by shapes Sub-effective channel is chosen
Subtotal 7 12
Shapes are very difficult to differentiate. Shape/annotation not most effective in distinguishing between colonized/infected (colour higher effectiveness)
3
Size scale indistinguishable
Differences in size not clear
Groups not distinguishable by shape size
Contrasts between groups not clear
Subtotal 15 10
Standard errors difficult to read (too small).
|
Difference in study weight (represented with square size) not readable. Difficult to follow one patient line. 3
2
Missing channel
Line types not used to distinguish between groups Colors not used to compare between visualization/groups
Subtotal 15 2
Different line types could improve distinguishing the different groups. Colours could have been used to make comparison of isolates across groups easier. 2
Visualization type does not (optimally) fit data
Other visualization type preferred
Subtotal 4 21
Too many categories for the plot type (e.g. bar chart) would be better). Data points/lines on double axes
Subtotal 21 11
Double Y-axes difficult to read Double y-axis confuse (not easy to
apprehend which line is on which axis). Channel overflow
Attribute overflow
Subtotal 11 8
Double use of shape and colour Unnecessary use of shape sizes Unnecessary use of colour
Unnecessary use of two channels for one attribute. Unequal size of point marks (without meaning). Unnecessary colour channel for intervention period. Too many colours. 1
1
Too many colours
1 Subtotal 11 Too many attributes 2 Relating attributes that are not Stacked bar chart used for unrelated 1
Plot is overloaded with attributes. bioRxiv preprint The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is
doi:
https://doi.org/10.1101/2021.05.19.444819
;
this version posted May 20, 2021.
made available under a
CC-BY 4.0 International license
. related
attributes (stacking does not make sense)
Information sparsity
Incoherent ordering
Subtotal 3 1 Subtotal 1 1
Could be text
Almost too little information. Data not ordered coherently Colonized/infected not consequently
ordered in bars. Subtotal 1
Grand Total 253
|
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. Structure of benzothiadiazine at zwitterionic phospholipid cell membranes
Zheyao Hu, Jordi Mart´ı(cid:63)∗ and Huixia Lu† (Dated: August 1, 2021)
The use of drugs derived from benzothiadiazine, which is a bicyclic heterocyclic benzene deriva- tive, has become a widespread treatment for diseases such as hypertension (treated with diuretics such as bendroflumethiazide or chlorothiazide), low blood sugar (treated with non-diuretic diazox- ide) or the human immunodeficiency virus, among others. In this work we have investigated the interactions of benzothiadiazine with the basic components of cell membranes and solvents such as phospholipids, cholesterol, ions and water. The analysis of the mutual microscopic interactions is of central importance to elucidate the local structure of benzothiadiazine as well as the mecha- nisms responsible for the access of benzothiadiazine to the interior of the cell. We have performed molecular dynamics simulations of benzothiadiazine embedded in three different model zwitteri- onic bilayer membranes made by dimyristoilphosphatidylcholine, dioleoylphosphatidylcholine, 1,2- dioleoyl-sn-glycero-3-phosphoserine and cholesterol inside aqueous sodium-chloride solution in or- der to systematically examine microscopic interactions of benzothiadiazine with the cell membrane at liquid-crystalline phase conditions. From data obtained through radial distribution functions, hydrogen-bonding lengths and potentials of mean force based on reversible work calculations, we have observed that benzothiadiazine has a strong affinity to stay at the cell membrane interface al- though it can be fully solvated by water in short periods of time. Furthermore, benzothiadiazine is able to bind lipids and cholesterol chains by means of single and double hydrogen-bonds of different characteristic lengths. I. INTRODUCTION
Cell membranes are responsible of the separation and exchange of cell contents from external environments. Membranes only allow the movement of compounds gen- erally with small molecular size (within the nanometric range) in or out the cell. For such a reason, the knowl- edge of the mechanisms responsible for the exchange of small peptides and drugs inside the membrane are of greatest scientific interest. The principal components of human cellular membranes are phospholipids, choles- terol and proteins, always inside electrolyte aqueous so- lution. Phospholipid membranes consist of two leaflets of amphiphilic lipids with a hydrophilic head and one or two hydrophobic tails which self-assemble due to the hy- drophobic effect[1, 2]. Among a wide variety of lipids, dimyristoilphosphatidylcholine (DMPC, C36H72N O8P ) are saturated phospholipids incorporating a choline as a headgroup and a tailgroup formed by two myristoyl chains.
|
They are usually synthesised to be used for re- search purposes (studies of liposomes and bilayer mem- branes). Their properties are very similar to those of dipalmitoylphosphatidylcholine (DPPC, C40H80N O8P ) which has the same structure but slightly longer tails, being a major constituent of pulmonary surfactants of lungs (about 40%). Differently, 1,2-dioleoyl-sn-glycero-3- phosphocholine (DOPC, C44H84N O8P ) and 1,2-dioleoyl- sn-glycero-3- phospho-L-serine (DOPS, C42H77N O10P )
∗ [email protected],
of Physics, Technical University of Catalonia-Barcelona Tech, B5-209 Northern Campus UPC, 08034 Barcelona, Catalonia, Spain (Corresponding author)
[email protected]; Department
† [email protected]; School of Pharmacy, Shanghai Jiaotong
are unsaturated species very common in the tissues forming the most essential human organs. Cholesterol (C27H46O), is a sterol playing a central role in main- taining the structure of the membrane and regulating their functions[3, 4]. It induces the membrane to adopt a liquid-ordered phase with positional disorder and high lateral mobility[2] and also regulates the fluidity of the membrane. There exists a big variety of experimental techniques useful to explore membrane organisation and molecular interactions of small probes within the membrane, such as: NMR, neutron diffraction, X-ray scattering or IR, Raman and fluorescence spectroscopy[5–8]. Among the latest techniques, fluorescence-lifetime microscopy[9] can be combined with spectral information to report basic information of aspects such as metabolic profiles, pho- tophysics or dipolar relaxations[10]. However, the direct information on atomic interactions and local structures can only by accessed by computer simulations at the all- atom level, as it is the aim of the present work. Heterocyclic molecules play an important role in medicine and are closely related to human life[11–13], agriculture[14] and industry[15], as the main fields of ap- plication. Such compounds are common parts of com- mercial drugs having multiple applications based on the control of lipophilicity, polarity and molecular hydro- gen bonding capacity. Among them, 3,4-dihydro-1,2,4- benzothiadiazine-1,1-dioxide (C7H8N2O2S, DBD, ben- zothiadiazine from now on) is a bicyclic heterocyclic benzene derivative containing two nitrogen atoms and one sulphur within the heterocyclic group. Benzoth- iadiazine derivatives have wide pharmacological appli- cations, such as diuretic[16, 17], antiviral[18–21], anti- inflammatory[22], anticancer[23, 24]or the regulation of the central nervous system[25, 26] to mention a few. Be-
University, Shanghai,China
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. cause of its importance as the main common structure of the DBD family, in this work we have selected the 3,4- dihydro-1,2,4-benzothiadiazine-1,1-dioxide species as our target to explore its affinity to the cell membrane and its local structure and hydrogen-bonding characteristics.
|
Up to the best of our knowledge, no microscopical studies of DBD at the interface of phospholipid membranes are yet available. So the main goal of the present study is to report structural and energetic results obtained from all- atom computer simulations of three classes of membranes for the first time. In this paper we provide the details of the simulations in section II and explain the main results of the work in section III, focusing our attention especially on the local structures of DBD (section III A) and also on the free-energy barriers of DBD and specific atomic sites, described in section III B. Finally, some concluding remarks are outlined in section IV. TABLE I. Characteristics of the systems simulated in this work. 10000 water molecules considered in all cases. ions Lipids 0 200 DMPC 0 160 DOPC, 40 DOPS
System 1 2 3
112 DOPC, 28 DOPS, 60 Cholesterol 55 Na+, 27 Cl −
whereas the pressure was controlled by a Nos´e-Hoover Langevin barostat[33] with a damping time of 50 fs. In the isobaric-isothermal ensemble, i.e. at the condition of constant number of particles (N), pressure (P) and temperature (T), equilibration periods for all simulations were of more than 50 ns. In all cases, we recorded sta- tistically meaningful trajectories of more than 200 ns through several production runs. The simulation boxes had different sizes because of the composition of the membrane. For instance, in system 1, the size was of 77.9 ˚A× 77.9 ˚A× 85.1 ˚A. For system 2 the size was of 87.1 ˚A× 87.1 ˚A× 72.9 ˚A, whereas for system 3 it was of 71.3 ˚A× 71.3 ˚A× 101.7 ˚A. II. METHODS
Three models of lipid bilayer membranes in an aqueous solution have been constructed using the CHARMM-GUI web-based tool[27, 28]. The membrane components and the amount of particles of each class are summarised in Table I. The lipids have been distributed in two leaflets embedded inside the electrolyte solution. We have con- sidered three different setups, namely: (1) a bilayer mem- brane formed only by DMPC, that we will label as ’sys- tem 1’; (2) a membrane formed by neutral DOPC and DOPS, labelled as ’system 2’ and finally (3) a membrane made by neutral DOPC and DOPS associated to Na+, to- gether with cholesterol in ionic solution, labelled as ’sys- tem 3’. System 1 is the simplest prototype cell membrane design, whereas system 3 is the most reallistic, complete one of the three, including electrolyte sodium-chloride so- lution at 0.15 M concentration, typical of human body. System 2 is a model with two clases of lipids, a kind of intermediate setup between the other two. Sketches of all species are reported in Fig. 1. Each phospholipid was described with atomic resolution (DMPC had 118 sites, DOPC 138 sites, DOPS 131 sites and cholesterol 74 sites). In all simulations, a single prototype drug, ben- zothiadiazine (20 sites, also represented in Fig. 1) has been chosen for the evaluation of its interactions with the main components of the membrane. Water has been represented by rigid TIP3P[29] molecules.
|
In all cases, we run the simulations at the fixed pressure of 1 atm and at the temperature of 303.15 K, well above the crossover temperatures for pure DMPC, DOPC and DOPS needed to be at the crystal liquid phase (295, 253 and 262 K, respectively)[30]). Molecular dynamics (MD) simulations were performed by means of the NAMD2 simulation package[31]. In all cases, the temperature was controlled by a Langevin thermostat[32] with a damping coefficient of 1 ps−1,
We have considered periodic boundary conditions in the three directions of space. The simulation time step was fixed to 2 fs. The CHARMM36 force field[34, 35] was adopted for lipid-lipid and lipid-protein interactions. We selected the version CHARMM36m[36], able to re- produce the area per lipid for the most relevant phos- pholipid membranes, in excellent agreement with exper- imental data. All bonds involving hydrogens were set to fixed lengths, allowing fluctuations of bond distances and angles for the remaining atoms. Van der Waals in- teractions were cut off at 12 ˚A with a smooth switch- ing function starting at 10 ˚A. Finally, long ranged elec- trostatic forces were computed using the particle mesh Ewald method[37], with a grid space of about 1 ˚A, up- dating such electrostatic interactions every time step of each simulation. III. RESULTS AND DISCUSSION
The three classes of bilayer phospholipid membranes considered in this work were previously simulated and their main characteristics were thoroughly analysed (see for instance Ref. [38] for DMPC and Ref. [39] for DOPC- DOPS), finding suitable values of the area per lipid A (by means of the deuterium order parameter) and thick- ness of the membranes (defined by the average distance between phosphorus of both leaflets of the membrane), in good agreement with available experimental and sim- ulation data. For the calculation of A we considered the total surface along the XY plane (plane parallel to the bilayer surface) divided by the number of lipids (eventu- ally plus cholesterol) in one lamellar layer[40]. 2
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 3
O
O
O
O
O
O
O
H3N
O
O
O
O
O
O
Na
P
O
P
cholesterol
NHNHS
H
H
H
O
N
O
OH
N
P
3,4-dihydro-1,2,4-benzothiadiazine-1,1-dioxide (DBD)1,2-dimyristoyl-sn-glycero-3-phosphocholine (DMPC)1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC)1,2-dioleoyl-sn-glycero-3-phospho-L-serine, sodium salt (DOPS-Na)322213141112322211121314131411122232121124
O
H
O
O
O
O
O
O
O
O
O
O
O
FIG. 1. Sketches of the backbone structures of DBD, DMPC, DOPC, DOPS and cholesterol. Hydrogens bound to carbon not shown. The highlighted sites of each species will be re- ferred in the text by the labels defined here. A. Local structure of benzothiadiazine
where nB(r) is the number of atoms of species B sur- rounding a given atom of species A inside a spherical shell of width ∆r.
|
V stands for the total volume of the system and NB is the total number of particles of species B. gAB(r) physical meaning stands for the probability of finding a particle B at a given distance r of a particle A. With the usual normalisation taken here, RDF will tend to 1 at long distances, when reaching the average density of the system. We have evaluated the local structure of the DBD molecule solvated by lipids and cholesterol. Among the myriad of possible RDF that we can compute, we will restrict ourselves to report a selection of the most rele- vant ones. The results are presented in Figures 2, 4 and 5, corresponding to the simulated systems 1, 2 and 3 as described in Table I. All RDF show some fluctuations in their profiles due to statistical noise. As a general fea- ture, the active sites of DBD capable of forming hydro- gen bonds (HB) are hydrogens ’H2’ and ’H4’ and oxygens ’O11’ and ’O12’ (see Fig. 1). We will see below that the role of such atomic sites is not unique, participating in a wide variety of HB of different length, ranging between 1.7 and 2.1 ˚A. We can observe a clear first coordination shell in all cases associated to the binding of DBD to lipid and cholesterol species (and eventually water as well in the case of system 3) together with lower second shells centred around 4 ˚A. 1.7 Å2.05 Å
024681012g(r)
H2-O11
H2-O13
H4-O13
012345678910r / Å
H4-O11
H2-O22
1. Radial distribution functions
The local structure of a multicomponent system is usu- ally analysed by means of the (normalised) atomic radial distribution functions (RDF) gAB(r). For a species B close to a tagged species A, they are given by
FIG. 2. Radial distribution functions for DBD with DMPC (system 1). gAB(r) =
V (cid:104)nB(r)(cid:105) 4 NB πr2 ∆r
,
(1)
The structure of DBD in system 1 is the simplest, as expected. We can distinguish DBD’s hydrogens H2 and H4 both forming HB with DMPC. H2 is able to bind oxygens ’O13’, ’O22’ and ’O11’ of DMPC, located in dif-
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 4
ferent regions of the lipid structure (see Fig. 1). In many cases, the HB length (location of first maximum of the RDF) is a short distance of 1.7 ˚A, typical of the binding of small-molecules to cell membranes, such as tryptophan to dipalmytoilphosphatidylcholine (DPPC, see for instance the review [41]). It should be pointed out that using flu- orescence spectroscopy, Liu et al. [8] obtained values for the hydrogen-bond lengths of tryptophan-water between 1.6 and 2.1 ˚A, i.e of the same range than those reported here. The largest peak (see Fig. 2) corresponds to H2 forming HB with oxygens of the phosphoryl group O13 and O14, since both oxygen sites are sharing a negative charge and their contributions have been averaged and labelled ’O13’, for the sake of simplicity.
|
The same rule has been applied to oxygens ’O22’ and ’O32’, although labels ’O11’ and ’O12’ have been analysed separately, given their different location in the lipid structure. This way of grouping equivalent lipid sites has been applied to all lipid species (DMPC, DOPC and DOPS). Con- cerning hydrogens H4 of DBD, they can be either con- nected to O13 or O11, but not to O22. Interestingly, in the case of DBD’s H4, HB lengths are in the range of 2-2.1 ˚A, significantly longer than those formed by H2 (range around 1.6 to 1.8 ˚A). We have analysed the fre- quency of such HB and observed that they tend to oc- cur in simultaneous pair structures, as it is sketched in Fig. 3. So, in plots (A), (B) and (C) we can observe that DBD’s H2 hydrogen-bonded to DMPC’s O13-O14 and DBD’s H4 bonded to DMPC’s O11 happen simulta- neously. Further, the angles formed by O-H-N (oxygen from lipids and H-N from DBD) are essentially flat, with values around 180◦ ± 30◦ in a rough estimation. We observed such double bindings for periods of time longer than 5 ns (considering statistical average along the 200 ns long MD trajectory). The group of eight atomic sites involved in such double hydrogen-bond, closed-ring struc- tures reassembles the special structure observed in some oncogenic proteins (KRas-4B) where salt bridges were found as the main responsible of the anchoring mecha- nism of the protein in the cell membrane[39]. This fact is qualitatively different of the single HB formed between small-molecules and lipid bilayer membranes observed in previous works (see Refs. [42, 43] for instance) where the small-molecule (melatonin or tryptophan) was able to bridge several lipid units and keep them together for sev- eral ns, probably due to the relative large size of the small-molecule and to the situation of their binding sites quite far away each other, differently of the very close binding sites of DBD. (A) (B) ti (C) (D)
FIG. 3. Snapshots of typical DBD-lipid and DBD-cholesterol bonds. Hydrogen bonds depicted in red, oxygens in red, ni- trogen in blue, carbon in cyan, phosphorus in brown and sulphur in yellow. (A) DBD-DMPC, (B) DBD-DOPC, (C) DBD-DOPS, (D) DBD-cholesterol. We have not found significant evidence of H4-water oxy- gen hydrogen-bonding. Regarding DBD-lipid bindings, we have observed that DBD’s H2 is more likely to es- tablish HB with DOPC’s O13 and O22, unlikely with O11 or O12. When analysing DOPS, H2 is also able to hydrogen-bond to O13 but not to O22 and, conversely, it can produce stable HB with O11. So the interactions of DBD are slightly different regarding DOPC or DOPS. However, the binding of DBD’s H4 are essentially the same for the two lipid species: it can be bound to O13 and O11 in both cases. The signature HB distances have been determined to be almost the same as in the DMPC case: 1.75 ˚A for H2 bindings and 2.0 ˚A for the H4 ones. All secondary shells have been also found centred around 4 ˚A. In system 2 we have also detected double HB between DBD and the two lipid classes (see Fig.
|
3, snapshots (B) and (C) where DOPC and DOPS, respec- tively are shown): H2 with O13 together with H4 with O12. Similar findings hold for system 2, where DBD is lo- cated at the interface of a DOPC-DOPS bilayer mem- brane. According to the results reported in Figure 4, we have distinguished three RDF for DBD-water, DBD- DOPC and DBD-DOPS. In this setup, DBD is able to be solvated by water for periods of the order of 20 ns. In such a case, we have located the corresponding HB formed by DBD’s H2 and oxygens of water, with the characteristic signature length of the water’s HB of 1.85 ˚A(see [44, 45]). Finally, the most complete and more realistic setup with system 3 has revealed several RDF similar to sys- tem 2, but some new features: (1) DBD’s H2 is able to establish some HB with DOPC’s O11, but with maxima less marked than H2-O13; (2) H2 is only bound to O13 of DOPS and (3) H4 connections are less and weaker than in system 2. We believe that this general weakening of the DBD-DOPC and DBD-DOPS hydrogen-bond struc- tures should be attributed to the presence of cholesterol. bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 5
H2-O13
11.522.533.544.555.56r / Å
H4-O13
H4-O12
H2-O22
024681012g(r)
0510152025g(r)
H2-OWater
1.75 Å1.85 Å2.0 ÅDOPCDOPS
H2-O11
00.511.522.5g (r)
H4-OChol. H4-O22
H2-O13
0510g(r)
123456r / Å
H2-OWater
H2-O13
04812g(r)
H2-O22
H2-O11
01020g(r)
O-HChol. H2-O22
H4-O12
DOPCDOPS
012g (r)
H2-OChol. FIG. 4. Radial distribution functions for DBD with water, DOPC and DOPS (system 2). We have located three possible HB between DBD and cholesterol: for both H2 and H4 of DBD with hydroxyl’s oxygens of cholesterol as well as oxygens of DBD with hy- droxyl’s hydrogen of cholesterol. The distances are still typical between 1.7 and 2.1 ˚A, but as it can be seen in snapshot (D) of Fig.3, none of them is due to double HB but to single ones. These facts can be probably relevant regarding the interactions of drugs from the DBD family with cell membranes. 2. Atomic site-site distances
After evaluating local structure of DBD, we have made a first step into the analysis of HB dynamics estimat- ing the lifetime of some of the HB reported by RDF. No other dynamical properties involving time-correlation functions such as power spectra[46, 47], relaxation times or self-diffusion coefficients[48, 49] have been considered here. In order to estimate the averaged time intervals for DBD-lipid association, we display the time evolution of selected atom-atom distances d(t) in Figures 6 and 7, for systems 1 and 3, respectively. Results for system 2 are in full agreement with those for system 3 and will be numerically reported in Table II. We can see that typ- ical hydrogen-bonding distances between 1.7-2.1 ˚A are reached in all cases. For instance, sites O13 and O14 of DMPC are reported independently in order to provide information on the relative distances between DBD’s H2 site and these two zwitterionic sites of DMPC.
|
Interest- ingly, we can observe that DBD’s ’H4’ is able to bind
FIG. 5. Radial distribution functions for DBD with water, DOPC, DOPS and cholesterol(system 3). DMPC’s ’O11’ but not ’O12’. Typical HB time lengths of about 5 ns are observed. Conversely, the other two selected distances are significantly longer (12 ns for H2 bound to DMPC’s O22) or shorter (2 ns for H2 bound to O32) and correspond to limit boundaries. The val- ues reported in Fig. 6 are representative of the average values collected in Table II over full trajectories. At the cholesterol-free membrane (system 1) we get typical HB lifetimes between 5 and 8 ns, whereas in systems 2 and 3 the range is broader, between 2 and 8 ns for lipids DOPC and DOPS and, interestingly, of only 1-2 ns for DBD-cholesterol hydrogen bonds. These shorter HB are represented in Fig. 7, where the HB dynamics indicates a larger extent of fluctuations, with bonds continuously formed and broken. When analysing the hydrogen-bonding of DBD with cholesterol (system 3), we have observed that some peri- ods of hydrogen-bonding are established between oxygens ’O11’and ’O12’ of DBD and the hydroxyl’s hydrogen of cholesterol (’HChol.’). In order to compute more precise hydrogen-bond lifetimes, correlation functions should be used (see Ref. [50] for instance), but these are out of the scope of this paper. A closer look indicates that in the case of DBD-cholesterol interactions, the longest living are the HB formed by hydrogens of a DBD (H2, H4 acting as donors) and the oxygens of cholesterol, the acceptors. bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 6
H2-O13
024681012141618t / ns
H4-O12
024681012d (t)
H2-O14
H2-O22
H4-O11
0246810d (t)
H2-O32
1.75 Å2.05 Å
H4-O11
H2-OChol. O12-HChol. 0246810d (t)
DOPCDOPSCholesterol2.05 Å1.85 Å
O11-HChol. H2-O14
H2-O12
H2-O14
051015
H4-OChol. H2-O32
H4-O32
05101520
051015t (ns)
FIG. 6. Distance distribution of selected sites in DBD-DMPC bonding (system 1) as a function of simulation time. Top: Distances of DBD’s H2 with DMPC’s sites (O13, black circles; O14, red squares; O22, green diamonds and O23, blue trian- gles). Bottom: Distances of DBD’s H4 with DMPC’s sites (O11, black circles; O12, red squares). Dashed lines indicate typical HB distances obtained from Fig.2 (1.75, 2.05 ˚A ). FIG. 7. Distance distribution of selected sites in DBD-DOPC, DBD-DOPS and DBD-cholesterol (system 3) as a function of simulation time. Left: Distances of DBD-DOPC (H2- O14, black circles; H2-O12, green diamonds and H4-O11, or- ange stars). Middle: Distances of DBD-DOPS (H2-O14, red squares; H2-O32, green diamonds and H4-O32, blue trian- gles). Right: Distances of DBD-cholesterol (H2-Ochol., black circles; H4-Ochol., red squares; O11-Hchol., green diamonds and O12-Hchol., blue triangles). Dashed lines indicate typical HB distances obtained from Fig.5 (1.85 and 2.05 ˚A ).
|
“Reverse” hydrogen-bonding composed by cholesterol’s hydrogen as the donors and oxygens O11-O12 of DBD (acceptors) is also possible but it is weaker than the for- mer, with significantly smaller free-energy barriers as we will see below. B. Potentials of mean force between DBD and lipids
The use of a variety of one-dimensional methods to compute the potentials of mean force (PMF) between two tagged species has been extensively discussed in the liter- ature (see Ref. [51] for a review), where up to twelve meth- ods were applied to the benchmark case of a methane pair in aqueous solution. The authors concluded that the best choice is a constraint-bias simulation combined with force averaging for Cartesian or internal degrees of freedom. The results from unbiased simulations, as those reported in the present work, were considered good at the qualitative level, with the PMF reasonably well repro- duced. However, the use of one-dimensional reaction co- ordinates is simply an approximation to the real ones[52], which may be in general multidimensional, presumably involving a limited number of water molecules and, even- tually coordinates or distances to the other species of the system. Methods which do not assume any preconceived
reaction coordinates such as transition path sampling[53– 55] or, those allowing to consider collective variables, such as metadynamics[39, 56] would be in order to ob- tain much more accurate free-energy landscapes but they require a huge amount of computational time. Since the determination of reaction coordinates for the adsorption of DBD at zwitterionic membranes is out of the scope of this paper, we will consider the radial distances between two species as our order parameters useful to work as reaction coordinates of unbiased simulations. In such case, we can easily obtain a good approxima- tion of the PMF through the so-called reversible work WAB(r) required to move two tagged particles (A,B) from infinite separation to a relative separation r (see for instance Ref. [57], chapter 7):
WAB(r) = −
1 β
ln gAB(r),
where β = 1/(kBT ) is the Boltzmann factor, kB the
Boltzmann constant and T the temperature. We show PMF for system 1 in Fig. 8. A free-energy barrier defined by a neat first minimum and a first max- imum of W (r) is clearly seen in all cases, with size ∆W1
(2)
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 7
TABLE II. Distances between selected sites of DBD and the membrane. Continuous time intervals are obtained from av- eraged computations. DBD site System Membrane Lipid sites Distance (˚A) Lifetime (ns)
H2 H2 H4 H4 H2 H2 H2 H2 H4 H2 H2 H2 H2 H4 H2 H2 H4 H2 H4 O11-O12
1 1 1 1 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3
DMPC DMPC DMPC DMPC DOPC DOPC DOPC DOPC DOPC DOPS DOPS DOPC DOPC DOPC DOPS DOPS DOPS
O13-O14 O22-O32 O11 O12 O13-O14 O22-O32 O11 O12 O11-O12 O13-O14 O22-O32 O13-O14 O22-O32 O11 O13-O14 O22-O32 O22-O32
Cholesterol OChol.
|
Cholesterol OChol. Cholesterol HChol. 1.7 1.7 1.7 4.0 1.85 1.85 4.0 1.75 2.0 1. 7 1.7 1.85 1.85 2.0 1.85 1.85 2.1 1. 85 2.1 1.85
5 8 5 5 3 2.75 6 6 3 2 8 2.5 4 5 2.5 2 3 2 2 1
3-2-1012345W(r) / kBT H2-O13 H2-O22 H2-O11 H4-O13 H4-O11 1.7 Å2.0 ÅDW1DW2
234567r / Å
(H2 hydrogens of DBD) and ∆W2 (H4 hydrogens of DBD). As expected the locations directly match the first maxima of the corresponding RDF. At first sight we can observe that ∆W1 > ∆W2 in all cases, which is a clear indication that H2 bonding is significantly stronger than that of H4. The full set of positions and free-energy barri- ers for the three systems has been reported in Table III. There we can observe overall barriers between 1.3 and 4.6 kBT , what correspond to 0.77-2.74 kcal/mol. We ob- serve stable binding distances very close to the typical hydrogen-bond distances. Furthermore, since the typi- cal energy of water-water hydrogen-bonds estimated from ab-initio calculations is of about 5 kcal/mol[58], we could conclude that our result are probably underestimated. Given the absence of DBD free-energy barriers in the literature (up to our knowledge), and for the sake of com- parison with other similar systems, the PMF of tryp- tophan in a di-oleoyl-phosphatidyl-choline bilayer mem- brane shows a barrier of the order of 4 kcal/mol[59], whereas the barrier for the movement of tryptophan at- tached to a poly-leucine α-helix inside a DPPC mem- brane was reported to be of 3 kcal/mol[60]. Finally, neurotransmitters such as glycine, acetylcholine or glu- tamate were reported to show small barriers of about 0.5-1.2 kcal/mol when located close to the lipid glycerol backbone[61]. These values could further indicate that our estimations match at least the order of magnitude of the free-energy barriers. IV. CONCLUDING REMARKS
A series of molecular dynamics simulations of a 3,4- dihydro-1,2,4-benzothiadiazine-1,1-dioxide molecule em-
FIG. 8. Potentials of mean force for the adsorption of DBD to DMPC (system 1). bedded in phospholipid bilayer membranes in aqueous ionic solution have been performed by molecular dynam- ics using the CHARMM36m force field. We have focused our analysis on the local structure of DBD, when asso- ciated to lipids, water and cholesterol molecules. After the systematic analysis of meaningful data, we noted rel- evant changes in local structure and dynamics of DBD. The location of DBD in the interface of the membrane is permanent when the bilayer is formed only with DMPC lipids. However, when DOPC and DOPS replace the DMPC backbone of the membrane, DBD is able to make excursions to the solvent water. The same feature has been observed in the most realistic case, when a mem- brane formed by 56% of DOPC, 14% of DOPS and 30% of cholesterol in sodium chloride solution has been set. We have computed radial distribution functions de- fined for the most reactive particles, namely hydrogens ’H2’ and ’H4’ and oxygens of DBD, according to labels defined in Figure 1 as well as selected sites of lipids and cholesterol able to form hydrogen bonds with DBD.
|
Our data revealed the existence of a strong first coordi- nation shell and a milder second coordination shell for DBD-lipid structures. Such first shell is due to hydro- gen bonds of variable lengths, between 1.7 and 2.1 ˚A , in good agreement with experimental data obtained from fluorescence measurements[8] for similar small-molecule- membrane systems. A direct analysis based on monitor- ing the relative distances between tagged sites of DBD and lipids has revealed that the lifetime of such hydro- gen bonds ranges (obtained by averaging data from the 200 ns MD trajectories simulated) between 1 ns for DBD-
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. TABLE III. Free-energy barriers ∆W for the binding of DBD to water and lipids. In order to quantify the height of all barriers, 1 kBT = 0.596 kcal/mol. Site ’1’ corresponds to DBD and ’Site 2’ to water or a lipid. and Knowledge (grant PGC2018-099277-B-C21, funds MCIU/AEI/FEDER, UE). ZH is the recipient of a grant from the China Scholarship Council (number 202006230070). System Site 1
1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3 3 3 3 3 3 3 3 3 3 3
H2 H2 H2 H4 H4 H2 H2 H2 H2 H4 H4 H2 H2 H2 H4 H2 H2 H2 H2 H4 H2 H2 H4 H2 H4 O11-O12
Site 2 O13-O14 (DMPC) O22-O23 (DMPC) O11 (DMPC) O13-O14 (DMPC) O11 (DMPC) O (Water) O13-O14 (DOPC) O22-O23 (DOPC) O11 (DOPC) O13-O14 (DOPC) O11 (DMPC) O13-O14 (DOPS) O22-O23 (DOPS) O11 (DOPS) O13-O14 (DOPS) O (Water) O13-O14 (DOPC) O22-O23 (DOPC) O11 (DOPC) O11 (DOPC) O13-O14 (DOPS) O22-O23 (DOPS) O22 (DOPS) OChol. OChol. HChol. rmin. (˚A) rmax. (˚A) ∆W (kB T ) 2.46 2.67 2.46 3.6 3.44 2.55 2.6 2.7 2.5 3.56 2.9 2.55 2. 95 2.8 3.56 2.55 2.47 2.57 2.33 2.76 2.36 2. 45 2.97 2.86 3.95 2.94
1.75 1.75 1.75 2.0 2.0 1.85 1.75 1.8 1.75 1.95 1.95 1.75 1.75 2.0 2.0 1.85 1.75 1.75 1.75 1.97 1.75 1.75 1.97 1.87 2.05 1.87
3.9 3.8 3.4 1.8 2.0 2.5 4.1 3.8 2.9 2.4 2.3 4.6 4.1 1.3 1.7 2.5 3.8 3.9 2.5 1.7 3.6 4.0 3.1 3.5 2.2 1.65
DATA AVAILABILITY
The data that support the findings of this study are available from the corresponding author upon reasonable request. cholesterol and up to 8 ns in the case of the HB formed between H2 of DBD and sites O22-O32 of DMPC and DOPC. Our results indicate that most of stable bonds formed by DBD and lipids usually involve sites H2 and H4 of DBD in double bonding with selected sites, as indicated in Figure 3. Only in the case of DBD-cholesterol bonds these are formed in a single-bond status, either between the hydrogen from the hydroxyl group of cholesterol and DBD’s oxygens or between the oxygen from the hydroxyl group of cholesterol and DBD’s species H2 and H4. Fi- nally, from the analysis of potentials of mean force based on reversible work calculations, we have estimated the free-energy barriers of the HB reported above (see Ta- ble III), where the strongest corresponded to the asso- ciation between DBD’s H2 and oxygens sites O13-O14 of DOPS in the absence of cholesterol.
|
The influence of cholesterol has been especially noted in the weakening of the DBD-lipid hydrogen bond connections and it may be a relevant factor when the interaction of drugs from the DBD family is considered from a pharmaceutical per- spective. ACKNOWLEDGMENTS
The authors gratefully acknowledge financial sup- port provided by the Spanish Ministry of Economy
8
bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. 9
[1] J. F. Nagle and S. Tristram-Nagle, BBA-Rev. Biomem-
branes 1469, 159 (2000). L. Danober, B. Pirotte, et al., ACS Chemical Neuro- science 7, 378 (2016). [2] O. G. Mouritsen, Life-As a Matter of Fat (Springer,
[27] S. Jo, T. Kim, V. G. Iyer, and W. Im, J. Comput. Chem. 2005). 29, 1859 (2008). [3] H. Ohvo-Rekil¨a, B. Ramstedt, P. Leppim¨aki, and J. P.
[28] S. Jo, J. B. Lim, J. B. Klauda, and W. Im, Biophys. J.
Slotte, Prog. Lipid Res. 41, 66 (2002). 97, 50 (2009). [4] T. P. McMullen, R. N. Lewis, and R. N. McElhaney, Cur- rent opinion in colloid & interface science 8, 459 (2004). [5] H. I. Petrache, S. W. Dodd, and M. F. Brown, Biophys. [29] W. L. Jorgensen, J. Chandrasekhar, J. D. Madura, R. W. Impey, and M. L. Klein, J. Chem. Phys. 79, 926 (1983). [30] W. Chen, F. Duˇsa, J. Witos, S.-K. Ruokonen, and S. K.
J. 79, 3172 (2000). Wiedmer, Scientific reports 8, 1 (2018). [6] K. M. Sanchez, G. Kang, B. Wu, and J. E. Kim, Biophys. J. 100, 2121 (2011). [7] P. Leyton, J. Brunet, V. Silva, C. Paipa, M. V. Castillo, and S. A. Brand´an, Spectrochimica Acta Part A: Molec- ular and Biomolecular Spectroscopy 88, 162 (2012). [8] H. Liu, H. Zhang,
and B. Jin, Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 106, 54 (2013). [9] M. A. Digman, V. R. Caiolfa, M. Zamai, and E. Gratton,
Biophys. J. 94, L14 (2008). [10] L. Malacrida, D. M. Jameson, and E. Gratton, Scientific
[31] J. C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, C. Chipot, R. D. Skeel, L. Kal´e, and K. Schulten, J. Comput. Chem. 26, 1781 (2005). [32] H. J. Berendsen, J. v. Postma, W. F. van Gunsteren, A. DiNola, and J. Haak, J. Chem. Phys. 81, 3684 (1984). [33] S. E. Feller, Y. Zhang, R. W. Pastor, and B. R. Brooks,
J. Chem. Phys. 103, 4613 (1995). [34] J. B. Klauda, R. M. Venable, J. A. Freites, J. W. O’Connor, D. J. Tobias, C. Mondragon-Ramirez, I. Vorobyov, A. D. MacKerell, and R. W. Pastor, J. Phys. Chem. B 114, 7830 (2010). Reports 7, 9215 (2017). [35] J. B. Lim, B. Rogaski, and J. B. Klauda, J. Phys. Chem. [11] D. A. Horton, G. T. Bourne, and M. L. Smythe, Chem-
B 116, 203 (2012). ical reviews 103, 893 (2003). [36] J. Huang and A. MacKerell Jr, Journal of computational
[12] V. Sharma, R. Kamal, and V. Kumar, Current Topics
chemistry 34, 2135 (2013). in Medicinal Chemistry 17, 2482 (2017). and A. Matter, Nature reviews Drug discovery 1, 493 (2002).
|
[14] M. Matheron and M. Porchas, Plant Disease 88, 665
[13] R. Capdeville, E. Buchdunger, J. Zimmermann,
[37] U. Essmann, L. Perera, M. L. Berkowitz, T. Darden, H. Lee, and L. G. Pedersen, J. Chem. Phys. 103, 8577 (1995). [38] J. Yang, C. Calero, and J. Mart´ı, J. Chem. Phys. 140,
(2004). 104901 (2014). [15] J.-P. Corbet and G. Mignani, Chemical reviews 106, 2651
[39] H. Lu and J. Mart´ı, The Journal of Physical Chemistry
(2006). Letters 11, 9938 (2020). [16] M. M. Platts, British Medical Journal 1, 1565 (1959). [17] R. M. Taylor and T. H. Maren, Journal of Pharmacology
[40] P. R. Pandey and S. Roy, J. Phys. Chem. B 115, 3155
(2011). and Experimental Therapeutics 140, 249 (1963). [41] J. Mart´ı and H. Lu, International journal of molecular
[18] A. Martinez, A. I. Esteban, A. Castro, C. Gil, S. Conde, G. Andrei, R. Snoeck, J. Balzarini, and E. De Clercq, Journal of medicinal chemistry 42, 1145 (1999). sciences 22, 2842 (2021). [42] H. Lu and J. Mart´ı, The Journal of chemical physics 149,
164906 (2018). [19] A. Martinez, C. Gil, A. Castro, C. Perez, M. Witvrouw, C. Pannecouque, J. Balzarini, and E. De Clercq, Antivi- ral Chemistry and Chemotherapy 12, 347 (2001). [20] D. Das, J. Hong, S.-H. Chen, G. Wang, L. Beigelman, S. D. Seiwert, and B. O. Buckman, Bioorganic & medic- inal chemistry 19, 4690 (2011). [43] H. Lu and J. Mart´ı, PloS one 14, e0224624 (2019). [44] J. Mart´ı, J. Chem. Phys. 110, 6876 (1999). [45] K. Modig, B. G. Pfrommer, and B. Halle, Physical re-
view letters 90, 075502 (2003). [46] J. Mart´ı, J. Padr´o, and E. Gu`ardia, Molecular Simula-
tion 11, 321 (1993). [21] V. M Patil, S. Prakash Gupta, S. Samanta, N. Masand, Medicinal Chemistry 8, 1099 (2012). [22] A. Tait, A. Luppi, A. Hatzelmann, P. Fossa,
and L. Mosti, Bioorganic & medicinal chemistry 13, 1393 (2005). and
[47] J. Padr´o and J. Mart´ı, The Journal of Chemical Physics
120, 1659 (2004). [48] J. Padr´o, J. Mart´ı, and E. Gu`ardia, Journal of Physics:
Condensed Matter 6, 2283 (1994). [49] J. Mart´ı and M. Gordillo, Chemical physics letters 354,
[23] A. Kamal, M. N. A. Khan, Y. Srikanth, K. S. Reddy, A. Juvekar, S. Sen, N. Kurian, and S. Zingde, Bioorganic & medicinal chemistry 16, 7804 (2008). 227 (2002). [50] J. Mart´ı, Phys. Rev. E 61, 449 (2000). [51] D. Trzesniak, A.-P. E. Kunz, and W. F. van Gunsteren,
[24] X. Ma, J. Wei, C. Wang, D. Gu, Y. Hu,
and R. Sheng, European journal of medicinal chemistry 170, 112 (2019). ChemPhysChem 8, 162 (2007). [52] P. L. Geissler, C. Dellago, D. Chandler, J. Hutter, and
M. Parrinello, Science 291, 2121 (2001). [25] A.-B. Nørholm, P. Francotte, L. Olsen, C. Krintel, K. Frydenvang, E. Goffin, S. Challal, L. Danober, I. Botez-Pop, P. Lestage, et al., Journal of Medicinal Chemistry 56, 8736 (2013). [26] A. P. Larsen, P. Francotte, K. Frydenvang, D. Tap- ken, E. Goffin, P. Fraikin, D.-H. Caignard, P. Lestage,
[53] J. Mart´ı, F. S. Csajka, and D. Chandler, Chem. Phys. Lett. 328, 169 (2000). [54] P. G. Bolhuis, D. Chandler, C. Dellago,
and P. L.
Geissler, Annu.
|
Rev. Phys. Chem. 53, 291 (2002). [55] J. Mart´ı and F. S. Csajka, Phys. Rev. E 69, 061918
(2004). bioRxiv preprint
doi:
https://doi.org/10.1101/2021.08.12.456125
;
this version posted August 13, 2021. The copyright holder for this preprint
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. [56] A. Barducci, G. Bussi, and M. Parrinello, Phys. Rev. [60] A. J. de Jesus and T. W. Allen, BBA-Biomembranes
Lett. 100, 020603 (2008). 1828, 864 (2013). [57] D. Chandler, Introduction to Modern Statistical Mechan-
ics (Oxford University Press, 1987). [58] M. W. Feyereisen, D. Feller, and D. A. Dixon, J. Phys. [61] G. H. Peters, M. Werge, M. N. Elf-Lind, J. J. Madsen, G. F. Velardez, and P. Westh, Chemistry and physics of lipids 184, 7 (2014). Chem. 100, 2993 (1996). [59] J. L. MacCallum, W. D. Bennett, and D. P. Tieleman,
Biophysical Journal 94, 3393 (2008). 10
|
bioRxiv preprint
doi: was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. https://doi.org/10.1101/2021.04.19.440495
;
this version posted April 20, 2021. The copyright holder for this preprint (which
Ligands binding to the cellular prion protein induce its protective proteolytic release
with therapeutic potential in neurodegenerative proteinopathies
Luise Linsenmeier1§, Behnam Mohammadi1§, Mohsin Shafiq1, Karl Frontzek2, Julia Bär3,15, Amulya N. Shrivastava4,#, Markus Damme5, Alexander Schwarz6, Stefano Da Vela7, Tania Massignan8, Sebastian Jung9, Angela Correia10, Matthias Schmitz10, Berta Puig11, Simone Hornemann2, Inga Zerr10, Jörg Tatzelt9,16, Emiliano Biasini8, Paul Saftig5, Michaela Schweizer15, Dimitri Svergun7, Ladan Amin12, Federica Mazzola13, Luca Varani13, Simrika Thapa14, Sabine Gilch14, Hermann Schätzl14, David A. Harris12, Antoine Triller4, Marina Mikhaylova3,15, Adriano Aguzzi2, Hermann C. Altmeppen1§*, Markus Glatzel1*
Affiliations:
1) Institute of Neuropathology, University Medical Center Hamburg-Eppendorf (UKE), Hamburg, Germany 2) Institute of Neuropathology, University of Zurich, Zürich, Switzerland 3) Institute of Biology, Humboldt-Universität zu Berlin, Berlin, Germany 4) École Normale Supérieure, Institut de Biologie de l'ENS (IBENS), INSERM, CNRS, PSL Research University, Paris, France 5) Institute of Biochemistry, Christian Albrechts University, Kiel, Germany 6) Institute of Nanostructure and Solid State Physics, Universität Hamburg, Hamburg, Germany 7) European Molecular Biology Laboratory (EMBL), Hamburg, Germany 8) Dulbecco Telethon Laboratory of Prions and Amyloids, CIBIO, University of Trento, Trento, Italy 9) Institute of Biochemistry and Pathobiochemistry, Ruhr University Bochum, Bochum, Germany 10) Department of Neurology, University Medical Center Göttingen, Göttingen, Germany 11) Department of Neurology, Experimental Research in Stroke and Inflammation, UKE, Hamburg, Germany 12) Department of Biochemistry, Boston University School of Medicine, Boston, MA, USA 13) Institute for Research in Biomedicine, Università della Svizzera italiana, Bellinzona, Switzerland 14) Calgary Prion Research Unit, University of Calgary, Calgary, Alberta, Canada 15) Center for Molecular Neurobiology Hamburg (ZMNH), UKE, Hamburg, Germany 16) Cluster of Excellence RESOLV, Bochum, Germany #) current affiliation: UCB Pharma SRL, Braine l’Alleud, Belgium
§Authors contributed equally
Corresponding authors: [email protected] & [email protected]
bioRxiv preprint
doi: was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission. https://doi.org/10.1101/2021.04.19.440495
;
this version posted April 20, 2021. The copyright holder for this preprint (which
Abstract
The cellular prion protein (PrPC) is a central player in neurodegenerative diseases caused by protein misfolding, such as prion diseases or Alzheimer`s disease (AD). Expression levels of this GPI-anchored glycoprotein, especially at the neuronal cell surface, critically correlate with various pathomechanistic aspects underlying these diseases, such as templated misfolding (in prion diseases) and neurotoxicity and, hence, with disease progression and severity.
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.