bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 The copyright holder for this preprint (which was not certified by peer review) is the author/funder.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.
PD-1 checkpoint blockade disrupts CD4 T cell regulated
adaptive B cell tolerance to foreign antigens
Chad R. Dufaud*, Andrew G. Shuparski*, Brett W. Higgins*,
Louise J. McHeyzer-Williams, Michael G. McHeyzer-Williams¶
Department of Immunology and Microbiology,
The Scripps Research Institute, La Jolla, CA, USA
* Equal contribution
¶ Lead contact and Correspondence: [email protected]
Keywords
PD-1 checkpoint blockade; IgG1 B cells; germinal centers; plasma cells; adaptive immune
tolerance; CD4 T cells; TFH cells; TFR cells; single-cell RNAseq
1 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 The copyright holder for this preprint (which was not certified by peer review) is the author/funder.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.
PD-1 checkpoint blockade disrupts CD4 T cell regulated adaptive B cell tolerance to foreign antigens
Graphical Abstract
Self
Cancer Surveillance Tolerance Autoimmunity
Activate PD-1 Inhibit
Infection Immunity Adaptation Allergy
Foreign
Steady Checkpoint State Blockade antibody
PD-1 PD-1
TFR TFR Plasma FDC TFH TFH cell Plasma LZ cell LZ GC B GC B
IgG1 IgG1
DZ DZ Memory Memory B cell B cell Germinal Center Germinal Center
In Brief PD-1 controls an adaptive B cell tolerance checkpoint in steady-state germinal centers to inhibit the maturation and production of IgG1 antibody with pre-existing foreign specificities.
Highlights - Acute PD-1 blockade induces extensive IgG1 PC differentiation at homeostasis - PD-1 blockade releases an IgG1 GC B cell checkpoint that drives expansion and PC formation - No adjuvant effect on foreign antigen but expansion of pre-existing IgG1 specificities to non-self - PD-1 exerts CD4 T cell dependent tolerance in the GC to restrict IgG1 maturation to non-self
2 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 The copyright holder for this preprint (which was not certified by peer review) is the author/funder.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.
SUMMARY
Adaptive B cell immunity to environmental antigens must be regulated by multiple CD4 T cell
dependent tolerance mechanisms. Using integrated single cell strategies, we demonstrate that
acute PD-1 blockade induces extensive and selective local anti-inflammatory IgG1 plasma cell
(PC) differentiation. Expansion of pre-existing IgG1 germinal center (GC) B cell and enhanced
GC programming without memory B cell involvement reveals an isotype-specific GC checkpoint
that blocks steady-state IgG1 antibody maturation. While there was no adjuvant impact on
immunization, acute PD-1 checkpoint blockade exaggerates anti-commensal IgG1 antibody
production, alters microbiome composition and exerts its action in a CD4 T cell dependent
manner. These findings reveal a PD-1 controlled adaptive B cell tolerance checkpoint that
selectively constrains maturation of pre-existing anti-inflammatory antibodies to prevent over-
reaction to steady-state foreign antigens.
3 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
INTRODUCTION
Adaptive B cell immunity involves a range of highly orchestrated cell fate decisions
dynamically controlled by specialized CD4 follicular helper T (TFH) cell compartments (Crotty,
2019; Cyster and Allen, 2019). These TFH cells induce antibody class switch upon antigen-
activation and promote affinity maturation within GC B cells to support long-term immune
protection against a multitude of foreign antigens (McHeyzer-Williams et al., 2012; Mesin et al.,
2016; Roco et al., 2019; Victora et al., 2010; Vinuesa et al., 2016). During development, some
aspects of these CD4-regulated B cell compartments must also become tolerate to commensal
organisms and a broad array of benign environmental antigens (Donaldson et al., 2018; Li et al.,
2020; Morais et al., 2021; Yilmaz et al., 2021). Immune checkpoints regulate the quality and
magnitude of CD4 T cell mediated signals that ultimately dictate B cell fate and impact immune
function (ElTanbouly and Noelle, 2021; Liu et al., 2021; Qi, 2016; Sharpe and Pauken, 2018; Shi
et al., 2018; Tan et al., 2021). It remains important to dissect how these immune checkpoints
regulate adaptive B cell tolerance to environmental antigens.
PD-1 checkpoint blockade is a highly effective immune therapeutic that re-activates
exhausted CD8 T cells during persistent infection (Barber et al., 2006; Hashimoto et al., 2018)
and across a variety of cancers (Sharma et al., 2021; Wei et al., 2017). Despite their non-
exhausted state, TFH cells (Ansel et al., 1999; Fazilleau et al., 2009; Vinuesa et al., 2005) and
their follicular regulatory counterpart (TFR) (Gonzalez-Figueroa et al., 2021; Linterman et al.,
2011; Wu et al., 2020) both express PD-1 with highest levels on GC localized subsets (Haynes
et al., 2007; Shulman et al., 2013). Genetic ablation of PD-1 results in increased TFH and TFR
cell expansion following immunization (Good-Jacobson et al., 2010; Hams et al., 2011; Sage et
al., 2013), in the gut (Kawamoto et al., 2012) and in autoimmune models (Nishimura et al., 1999;
Nishimura et al., 2001). These germline PD-1 deficiencies have variable impact on the magnitude,
maturation and outcome of GC B cell selection (Good-Jacobson et al., 2010; Hams et al., 2011;
4 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Sage et al., 2016; Shi et al., 2018) with TFR-focused deletion influencing early pre-GC B cell
activity and immune tolerance as well as reducing B cell recall responses (Clement et al., 2019;
Lu et al., 2021; Tan et al., 2021).
Acute PD-1 blockade can enhance B cell responses (Dyavar Shetty et al., 2012;
Mylvaganam et al., 2018; Velu et al., 2009) and promote TFH and B cell presence in the tumor
microenvironment (Hollern et al., 2019; Sharonov et al., 2020) with a positive correlation to the
outcome of immune therapy (Cabrita et al., 2020; Helmink et al., 2020; Petitprez et al., 2020).
However, multiple adverse immune reactions to PD-1 blockade can involve autoimmune reactivity
and inflammatory responses targeting barrier surfaces such as the gut (Dougan et al., 2021;
Pauken et al., 2019). Identifying cellular pathways and steady-state molecular targets for PD-1
blockade remains an important fundamental need with high potential utility for understanding the
mechanism of action as well as these adverse reactions of immune therapy.
Here, we use high-fidelity single cell sorting and targeted RNA sequencing to reveal the
cellular and molecular mechanisms of acute PD-1 blockade on adaptive B cell tolerance. This
CD4 T cell-dependent mechanism dysregulates pre-existing anti-inflammatory IgG1 GC B cell
maturation to produce massive local expansion and differentiation of IgG1 PC and antibodies.
Released specificities target pre-existing foreign antigens with no evidence for autoimmunity or
overt adjuvant action in a vaccine context. We propose that this sub-class specific GC B cell
checkpoint attenuates the maturation of pre-existing foreign specificities to avoid over-reacting to
environmental antigens and maintain immune homeostasis.
5 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
RESULTS
Acute PD-1 blockade induces extensive IgG1 PC differentiation
To assess the role of the PD-1 pathway on CD4-regulated B cell immunity, we focus on
IgG class-switched B cell fate and function in the steady-state spleen. Six days after single anti-
PD-1 treatment across three different blocking antibody clones, we observed significant increases
in IgG1 antibody-secreting cells in the spleen by functional analysis (Figure 1A and 1B). Direct
assessment of isotype-specific PC numbers in vivo by high resolution flow cytometry revealed a
rapid and selective increase in splenic IgG1-expressing PC (Figure 1C - 1F). Thus, without
intentional immunization acute PD-1 blockade alone substantially disrupted isotype-specific
antibody homeostasis and significantly enhanced the size of the local IgG1 PC compartment.
To discriminate between recent PC migration and rapid PC differentiation we sorted IgG-
expressing splenic PC before and after checkpoint blockade for gene expression analysis.
Individual cells from these rare compartments were uniquely bar-coded during cDNA synthesis,
and gene targeted amplification of ~500 mRNA species for sequencing utilizing digital
deconvolution for single cell quantification (quantitative and targeted RNA sequencing; qtSEQ)
(Figure S1A - S1D). By avoiding over-abundant immunoglobulin (Ig) mRNA, this single cell
qtSEQ strategy provided clear access to many elements of the class-switched PC transcriptional
program (Figure S1E - S1G). Pseudo-time TSCAN re-construction (Ji and Ji, 2016) orders and
clearly separates naive B cells, steady-state IgG PC and anti-PD-1 treated IgG1 PC based on
their targeted transcriptional profile (Figure 1G). Relative expression levels per cell of central PC-
associated transcription factors (TF) Xbp-1 and Prdm-1 are similar between steady-state and anti-
PD-1 treated PC, while proliferation associated Mki67 and Ig co-chaperone Mzb-1 are markedly
up-regulated after treatment indicating recent cell expansion and IgG1 PC differentiation (Figure
1H).
6 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Closer scrutiny of the induced IgG1 PC program reveals a multitude of transcriptional
changes related to the steady-state IgG PC program that are significantly enhanced by acute
checkpoint blockade. IgG1 PCs exhibit enhanced capacity for inter-cellular communication with
increased expression of adhesion molecules (Cd44, Cd48, Spn, Lgals1, Cd9), SLAM family
members (Slamf6, Slamf9), and TNF ligand and receptor superfamily members (Tnfsf5, Tnfsf9,
Tnfsf18, Tnfrsf4, Tnfrsf13c) (Figure S1H). Acute changes in expression for antigen presentation
(H2.Ab1, Ly75, Cd81), cell fate determination (Notch2, Plxdn1) and a range of known inhibitory
receptors (Havcr1, Timd2) and co-stimulators (Cd69, Cd28, Cd24a) indicate major recent
alterations in IgG1 PC development and function.
Enhanced expression of the intracellular machinery managing endoplasmic reticuluum
(ER; Mzb1, Spsc1, Ssr2, Edem1), protein modification (Slpi, Fut8, Txk), chaperone function
(Igbp1), signaling (Blk, Nfkbia, Pten), antigen processing (H2.Dma, Cd74, Lamp1) and
metabolism (Ctsz, Mpc2) along with changes in secreted factors (Il1a, Il22, Il23a, Ccl21a, Ltbp3)
were evident after treatment (Figure 1I and S1I). Highly significant increases in expression across
multiple TF families (C2H2ZF: Bcl6b, Ikzf2, Ikzf4, Klf4, Klf7; WHforkhead: Irf2, Irf4, Spib) as well
as upregulation of gene indicating recent PC differentiation activity (Xbp1, Prdm1, Pou2af1,
Bhlha15) and PC response programs (Gata3, Sta5b, Smad4, Skil) were also induced by PD-1
checkpoint blockade (Figure 1J and S1J). Most informative are the dominant increases in cell
cycle machinery (Ccdn2, Cdkn1b, Pcna, Ube2c, Mki67) and substantial elevation of DNA
replication and repair elements (Mcm2, Mcm4, Top2a, Polh). Overall, these data indicate that PD-
1 blockade drove recent and selective expansion of a PC precursor and then differentiation into
a new anti-inflammatory IgG1 secreting PC compartment.
IgG1 GC B cell expansion without memory B cell involvement
7 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Class-switched IgG-expressing memory and GC B cells can be found in the steady-state
spleen, but these small fractions of IgG1+ memory (GL7-CD38+) B cells remained unchanged after
PD-1 blockade (Figure 2A to 2C). In contrast, there was a 30-50 fold selective expansion of IgG1-
expressing GC (GL7+CD38-) B cells (Figure 2D to 2F). Significant IgG1 GC B cell or PC increases
were not seen before day 6 after treatment (Figure S2A and S2B), with no evidence for early B
cell activation (GL7+CD38+) or recruitment of IgG1+ memory B cells throughout treatment (data
not shown). Single cell qtSEQ analysis with pseudo-time reconstruction (using TSCAN) clearly
separates the transcriptional impact of checkpoint blockade from the steady-state IgG1 GC B cell
program (Figure S2C to S2E; Figure 2G) with enhancement of major GC molecular program
modifiers (Bcl6, Aicda) and evidence of recent proliferation (Mki67)(Figure 2H). There was a shift
in GC B cell transcriptional activity with significant changes in expression for surface molecules
involved in modifying BCR signals (Cd79b, H2.Aa, Ptprc, Cd81), TFH contact (Cd244, Cd276,
Spn, Havcr1) and TNF receptor/ligand superfamily (Tnfsf9, Tnfrsf25, Tnfrsf8)(Figure 2I and S2F).
There were also increases in expression of intracellular signaling components (Dock8, Blnk,
Aurkb, Blk) and antigen processing machinery (Cd74, H2.Dmb2) that may accompany heightened
selection activity of the GC cycle (Figure S2G).
Similar to the IgG1+ PC compartment, there were substantial increases in expression for
cell cycle machinery (Mki67, Pcna; Ube2c, Trp53, E2f4), DNA replication (Mcm2, Mcm4, Top2a)
and DNA repair components (Aicda, Polh) typically associated with GC B cell expansion, somatic
hypermutation and GC BCR diversification (Figure 2J and S2H). Clear expression increases in
a multitude of TFs (Bcl6, Bach2, Pou2af1, Id3) highlight the induction (Spi1, Irf2, Klf13, Bhlha15)
and orchestration of multiple complex facets (Ikzf4, Tcf3, Ikzf3, Ets1) of the ongoing IgG1 GC
cycle after checkpoint blockade (Figure 2J and S2H). Using STREAM trajectory analysis (Chen
et al., 2019), we display the distribution of the steady-state IgG1 GC program in contrast to the
amplified and more clear segregation of GC cycle activities after treatment (Bcl6, Bach2, Mki67,
8 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
CD24a) (Figure 2K). In the absence of non-GC B cell activation or memory B cell involvement,
these data are consistent with the acute enhancement of a pre-existing IgG1 GC B cell cycle that
skews preferentially towards IgG1 PC differentiation upon PD-1 blockade.
Checkpoint blockade exaggerates IgG1 GC B cell programming
To further understand GC cellular dynamics, we used cell surface markers to interrogate
the impact of PD-1 blockade on the zonal distribution of IgG GC B cell fate and function. The GC
light zone (LZ; CXCR4loCD86hi) B cells are selected on antigen-binding variant BCR and then with
cognate TFH contact, which promotes cell cycle and BCR diversity upon re-entry into the GC dark
zone (DZ; CXCR4hiCD86lo)(Victora et al., 2010). As expected, we observe that class-switched
non-IgG1 GC B cells favor DZ phenotype with no change after anti-PD-1 treatment (Figure 3A).
In contrast, IgG1 GC B cells skewed heavily into the LZ at homeostasis and then all IgG1 GC B
cells expanded with a significant shift into the DZ upon PD-1 blockade (Figure 3B). Using single
cell qtSEQ with STREAM analysis, we visualized IgG1 GC B cells with a more organized and
enhanced distribution of mRNA for CD86 and CXCR4 after anti-PD1 treatment (Figure 3C). Using
this strategy, we identified a range of highly significant divergent genes expressed in the same
DZ-associated state that were heavily skewed towards cell cycle (Pcna, Mki67) and DNA
replication and repair (Mcm2, Mcm4, Top2a, Aicda, Polh) released upon PD1 blockade (Figure
3D). Increases in expression of surface molecules (Tnfsf9, CD24a, H2-Ab1), signaling machinery
(Blnk, Aurkb, Dock8) and TFs (Pou2af1, Bcl6, Bhlha15, Bach2) also accompanied this anti-PD1
driven amplification of DZ activity (Figure 3D).
TSCAN trajectories can be used to dissect further the organization and molecular
components of anti-PD-1 driven GC cycle amplification. At steady-state, two major regions of cells
displaying LZ and DZ associated activities were found with a minor separate 'transitional' subset
(Figure 3E). After PD-1 blockade and IgG1 GC B cell expansion, LZ and DZ regions remained
9 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
high in frequency and level of gene expression; however, a prominent 'transitional' zone (TZ) was
discernable at equivalent frequencies to LZ and DZ subsets (Figure 3F). Examples of significantly
divergent genes across pseudo-time are presented in heatmaps to highlight assortment in LZ-
associated activities of antigen processing and presentation (H2.Dmb2, H2-Aa, H2-Ob), and BCR
signaling (Blk, Igbp1) that increase (H2.Dmb2, H2.Aa) or change (Cd74, Cd79b, Blnk, Tnfrsf13c,
H2.K1) upon PD-1 blockade. The range of functions released in the transitional state include
notable changes in TF expression (FoxJ1, Spi1, Gata1), surface and secreted molecules (Plxnd1,
Spp1, Tgfb2, Xcl1) with evidence for recent exaggerated cell cycle entry (Aurkb, MKi67, Top2a).
These TZ enhancements are further reinforced by increases of DZ-associated gene expression
for cell surface and intracellular molecules (Tnfsf9, Cd24a, Blk), TF expression (Pou2af1), and
DNA replication and repair machinery (Mcm2, Mcm4, Polh, Pcna, Trp53). Overlay STREAM
reconstructions reinforce the induction of an exaggerated IgG1 GC program upon PD-1 blockade
by using Cd83 and Polh mRNA to discriminate LZ/DZ regions respectively (Figure 3G), to
highlight significant changes in transition genes due to PD-1 blockade (Figure 3H).
Taken together, we propose that the steady-state IgG1 GC transcriptional state constitutes
a tolerance program that exerts LZ constraint of the multiple pre-existing IgG1 GC B cell
specificities. Acute PD-1 blockade releases this inhibitory program to exaggerate typical
molecular actions of an active GC cycle that accompanies BCR maturation, ongoing cellular
selection and exit into long-lived post-GC B cell compartments. Here we demonstrate the highly
skewed preference towards post-GC IgG1 PC production.
IgG1 PC transition from anti-PD-1 dysregulated GC B cell cycle
Next, we considered the molecular changes that occur between rapid IgG1 GC B cell
expansion and IgG1 PC production induced by PD-1 blockade. Bray-Curtiss dissimilarity plots
display high dimensional approximations of clusters for GC and PC subsets based on gene
10 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
expression with further separation due to isotype and anti-PD-1 treatment (Figure 4A).
Considering PD-1 treatment alone, TSCAN pseudo-time projections place the induced IgG1 PC
program downstream from the IgG1 GC program with skewing of particular genes into IgG1 GC
(Bcl6, Aicda, Bach2), IgG1 PC (Xbp1, Irf4, Xbp1, Prdm1) and across both (Pcna, Mki67, Batf and
Bhlha15)(Figure 4B and S3A).
Closer scrutiny of these two programs reveals major differences in expressed genes with
abundant state-specific and highly significant alterations that would be required to convert IgG1
GC B cells into the IgG1 PC pathway. At the cell surface, IgG1 GC B cells must lose growth factor
receptors (Il2ra, ifngr1, Flt3, Il31ra), BCR co-receptors (Cd79a, Cd19, Ptprc) and T cell interactors
(Cd72, Havcr1, Cd69), while gaining new expression of inter-cellular communicators (Notch2,
Cd48, Cd9, Lgals1) and new means of cell contact (Cd44, Ly75, Cd28, Cd40lg) in this transition
(Figure 4C and S3C). Shifts in intracellular machinery include GC decrease of BCR signaling
components (Blk, Bank1, Dock8) and antigen processing (H2.Dmb2, H2.Oa, Cd74) in favor of
high-fidelity protein management (Edem1, Spcs1, Fut8), efficient secretory machinery
(Mzb1,Ssr2) with changes in metabolism (Mpc2, Ctsz) and secretion of soluble regulators
(Ccl21a, Il22,, Il1a, Ltbp3, Slpi) emphasized in the IgG1 PC compartment (Figure S3B).
Transitions in usage of major transcriptional regulators for GC (Bcl6, Bach2, Id3, Pax5, Ets1,
Ikzf1, Ikzf3, Spib, Irf5) and PC (Bcl6b, Bhlha15, Gata3, Ikzf4, Klf7, Atf4, Irf4, Runx2) cohorts
(Figure 4D and S3D) highlight the highly significant changes that are needed to drive IgG1-
expressing GC B cell undergoing BCR maturation towards IgG1-secreting PC differentiation.
Using Slingshot trajectory inference (Street et al., 2018), we recreate the IgG1 GC to PC
transition that accompanies PD-1 blockade and define two separable IgG1 PC subsets. We
included the steady-state IgG1 GC for these pseudotime projections and separate them with
overlap from the anti-PD-1 treated IgG1 GC compartments (Figure 4E). These data further
11 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
segregate increased transcriptional activity following treatment that is restricted to the IgG1 GC
(Aicda, Bcl6, Bach2), and extended into the PC compartment (Pou2af1, Cd79b) or expressed
largely in the PC compartment (Xbp1)(Figure 4E). The steady-state PC compartment exists as a
singular pathway while using STREAM projections, the PD-1 treated PC emerge from the GC
across one node that then branches out across two forks based on transcriptional divergence
(Figure 4F). Some genes expressed in the lower fork (Pcna, Ssr2) and the upper fork (Notch2,
Bcl6b) can also be seen in slingshot projections to exemplify multiple differences in anti-PD-1
induced PC subsets (Figure 4G and S3E). Among many transcriptional differences, the strong
indication of recent cell cycle activity with evidence of DNA replication and repair machinery
upregulation is consistent with a recent GC origin, while PC population S4 has an expression
profile more similar to expanded and fully differentiated IgG1 PC compartments induced by PD-1
blockade.
PD-1 blockade provides no adjuvant action in a vaccine response.
While the IgG1 selectivity has not been previously seen, the release of GC B cell control
and exaggerated GC B cell expansion is predicted from previous studies of PD-1 function. Hence,
we reasoned that anti-PD-1 could be used as an isotype-specific adjuvant to enhance IgG1 GC
B cell maturation in the context of vaccination. We have extensive experience in the model antigen
response to NP-KLH using a TLR4 agonist known to induce both IgG1 and IgG2 antibody isotypes
(McHeyzer-Williams et al., 2015; Pelletier et al., 2010; Wang et al., 2012). At day 10 after priming,
anti-PD-1 was introduced and shown to significantly decrease PD-1 expression on follicular T cell
compartments 6 days later (Figure S4A and S4B). Using this strategy, we quantified the impact
of acute PD-1 blockade on antigen-specific B cells from newly-formed GC (day 16 primary),
persistent GC (day 48 primary) and newly-formed memory GC (day 6 recall) together with the
outcomes of antigen-specific memory B cells and PC differentiation (Figure S4C to S4I).
Surprisingly, there was no impact of PD-1 blockade on the vaccine-driven antigen-specific B cell
12 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
response across all compartments and over all timepoints, regardless of antibody isotype. We
thought this may be due to the overwhelming effect of the TLR4 agonist, however the pre-existing
non-specific IgG1 splenic GC and PC response remained equally susceptible to PD1 blockade
(at day16 and day 48 primary) but not measurable above the elevated baseline at recall (Figure
S4J and S4K). Using high resolution flow cytometry, we can reproducibly measure even the very
small antigen-specific response to NP-KLH alone without adjuvant (Figure S4L). Even under
these sensitive circumstances with no added inflammatory drive, we saw no adjuvant impact of
anti-PD1 on the resultant newly-formed antigen-specific B cell response.
Acute checkpoint blockade releases pre-existing foreign IgG1 specificities
While PD-1 blockade did not enhance the demonstrable immune response to foreign
antigen, we were able to detect a small level of pre-existing IgG1 binding to the J43 hamster
antibody before treatment that assorted across all mature B cell compartments (Figure S5A).
This specific J43 binding increased significantly over 6 days of treatment displaying the
characteristic IgG1 selectivity for expansion in the GC and PC compartments (Figure S5A). Using
ELISPOT analysis, we determined that J43 binding accounted for <30% of the total IgG1 PC
expansion driven by acute blockade (Figure S5B). Next, we attached the foreign hapten NP to
J43 under different conjugation ratios and failed to drive an anti-NP response even in the presence
of the extensive anti-J43 reaction (Figure S5C). Taken together, we propose that PD-1 blockade
does not enhance a newly-forming anti-foreign immune response, but selectively targets the
release of pre-existing foreign or cross-reactive self-specificities of IgG1 isotype within the steady-
state splenic GC reaction.
Broad screening for IgG1 self-specificities by immunofluorescence staining showed no
detectable labeling of intracellular components from syngeneic cell lines (data not shown).
However, labeling live cecal bacteria by flow cytometry captured significant increases in IgG1
13 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
binding from C57Bl/6 animals >14 weeks of age after PD-1 blockade (Figure 5A and 5B). To
determine the specificities of IgG1 antibodies released by PD-1 blockade, we used freshly isolated
cecal bacteria with detectable hamster control binding but 100-fold higher intensity binding after
PD-1 blockade and sorted these bacteria for Oxford Nanopore based 16S rRNA sequencing
(Figure 5C and 5D). Notably, there was significant enrichment in IgG1 binding to members of the
order Bacteroidales and genus Helicobacter, with especially high binding to murine pathobiont
species H. typhlonius using a steady-state CD45.1 congenic C57BL/6 microbiome containing this
species (Figure 5E). Higher representation of binding to the genus Lactobacillus and species L.
johnsonii before treatment indicates the shift away from these IgG1 specificities upon release of
anti-inflammatory IgG1 antibodies with blockade. Furthermore, acute anti-PD-1 treatment
significantly altered the presence of multiple gut microbial species (Figure 5F). H. typhlonius was
not present in the CD45.2 steady-state C57Bl/6 microbiomes used in the latter experiments and
did not emerge over the 6 days of treatment. While not significant, Akkermansia muciniphila,
known to induce T-dependent IgG1 (Ansaldo et al., 2019), displayed the highest level of decrease
over the 6 days of treatment (17-fold) in these studies. Hence, PD-1 mediated inhibition can
control a subset of established anti-inflammatory anti-microbial IgG1 specificities and acute
disruption significantly alters circulating antibody levels and impacts steady-state gut microbial
diversity.
CD4 T cell dependent PD-1 tolerance checkpoint.
The ligand PD-L1 is expressed on all memory B cell subsets and many non-IgG1 PC;
however, expression is very low on all splenic GC B cells regardless of IgG subclass as well as
IgG1 PC (Figure 6A). Hence, the selective inhibitory impact of PD-1 on IgG1 GC B cells was
unlikely to involve direct signals through the PD-L1 pathway on GC B cells. Using acute depletion
of CD4, we established the requirement of CD4 T cells as key regulators of the PD-1 GC B cell
checkpoint (Figure 6B). CD4 T cell depletion did not significantly impact existing splenic IgG1 PC
14 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
or memory B cell numbers; however, the enhancement of IgG1 PC by PD-1 blockade was
completely abolished (Figure 6C). Furthermore, >95% of pre-existing IgG1 GC B cells were lost
with anti-CD4 treatment, but this loss was not overcome by PD-1 blockade. These trends further
reinforce the likelihood that pre-existing IgG1 GC B cells are the immediate cellular precursor of
both expanded IgG1 GC B cells and post-GC IgG1 PC cell fate induced with PD-1 blockade.
These data also broadly implicate CD4 T cells as the direct target of anti-PD-1 checkpoint
blockade.
As briefly mentioned, anti-PD-1 treatment downregulates surface expression of PD-1
itself, even in the context of foreign antigen vaccination (Figure S4B). Here, we examined this
more closely using this non-cross blocking anti-PD1 antibody with similar results at 24hr and 6
days after treatment (Figure 6D and 6E). In contrast to germline genetic ablation, acute PD-1
blockade had no impact on the number or broad composition of the pre-existing activated splenic
CD4 T helper cell compartment at steady-state (Figure 6F). Even sub-fractions of conventional
regulatory T cells and the two main follicular fractions of TFH and TFR cells known to express the
highest levels of PD-1 show no significant alteration in numbers due to PD-1 blockade (Figure
6G). Hence, we propose there must be a significant acute alteration in the local transcriptional
program of follicular T cells to selectively target the anti-inflammatory IgG1 pathway and release
local control on IgG1 GC maturation to produce these extensive changes in the IgG1-secreting
PC compartment.
15 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
DISCUSSION
At homeostasis, the adaptive immune system remains dynamically poised between
immunity and tolerance to both self and foreign antigens. Our studies reveal an acquired B cell
tolerance mechanism that uses the PD-1 pathway on CD4 T cells to control B cell fate and
antibody homeostasis. This PD-1 driven mechanism targets the anti-inflammatory IgG1 pathway
at steady-state to constrain GC maturation and PC formation without memory B cell involvement.
Using single cell qtSEQ, we define molecular components of the IgG1 GC B cell and PC tolerance
mechanism and demonstrate how it is released upon acute PD-1 checkpoint blockade. We find
no evidence of this activity during adjuvant-induced immunity but enhancement of pre-existing
IgG1 specificities, including binding to members of the commensal microbiome, was evident upon
checkpoint blockade. We propose this adaptive B cell tolerance mechanism is operative in healthy
individuals to guard against over-reaction to antigens within the environment.
PD-1 blockade provides one mechanism for driving the acute progression of IgG1 GC B
cells towards PC differentiation. We developed a custom qtSEQ strategy to quantify changes in
the transcriptional program of individually sorted and indexed members of these rare lymphocyte
sub-populations. High-fidelity cell purification is directly coupled to sensitive and targeted cDNA
library formation to provide ~10-fold increased detection per single cell of targeted mRNA species
compared to global strategies (Fan et al., 2015; Gao et al., 2020; Milpied et al., 2018; Peng et al.,
2021; Wu et al., 2014). PCs have been difficult to sequence due to over-abundance of Ig mRNA
signals per PC impacting resolution of gene expression studies (Higgins et al., 2019; Nutt et al.,
2015). Nevertheless, RNAseq strategies have defined core plasma cell programs of differentiation
(Low et al., 2019; Roy et al., 2019; Shi et al., 2015), at rest and during infection in humans (King
et al., 2021; Sokal et al., 2021), staged transcriptional states in the context of malignancy (Holmes
et al., 2020; Ledergor et al., 2018) and implicated B cell activity and PC function in tumor
microenvironments (Hollern et al., 2019; Sharonov et al., 2020). Here, we reveal acute induced
16 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
alterations in intracellular machinery, changes in capacity for intercellular contact and control, and
substantial shifts in key nuclear processes and TF expression that accompany IgG1 GC B cell re-
programming towards PC differentiation. In the absence of IgG1 B cell activation or recruitment
of IgG1 memory B cells, together with the concurrent expansion of IgG1 GC B cells, we conclude
that massive local PC expansion is a directed post-GC B cell fate. Unbiased clustering and
pseudo-time reconstruction provide molecular detail to progression of the IgG1 post-GC PC
differentiation program that is released after acute PD-1 blockade.
Complex and progressive molecular events drive the CD4 regulated GC B cell cycle of
affinity maturation (McHeyzer-Williams et al., 2012; Mesin et al., 2016; Roco et al., 2019; Victora
et al., 2010; Vinuesa et al., 2016). Recent single cell molecular approaches use a range of tissue
sources containing active GC to unravel the cyclic progression of GC function (Holmes et al.,
2020; Kennedy et al., 2020; King et al., 2021). In the current study, the cellular distribution of
splenic GC B cells suggested that the steady-state IgG1 GC B cell compartment was selectively
constrained in the LZ and that PD-1 expressed on T cells was needed to restrict ongoing DZ re-
entry with its associated cell cycle and BCR diversification machinery (Kennedy et al., 2020;
Mesin et al., 2016; Victora et al., 2010). While the steady-state IgG1 GC B cell transcriptional
program was significantly altered with treatment, it remains important to ascertain which of the
many pre-existing traits were central to maintaining the adaptive tolerant state. Class-switched
memory B cells and tertiary lymphoid structures (Cabrita et al., 2020; Helmink et al., 2020) have
been linked to immunotherapy efficacy and patient survival (Petitprez et al., 2020) with TFH and
B cell compartments also implicated (Hollern et al., 2019). In the current study, PD-1 blockade
releases a vigorous and progressive IgG1 GC program with exaggerated transition towards
ongoing IgG1 GC B cell DZ maturation and PC terminal differentiation. We predict that release of
IgG1 GC B cell maturation can impact efficacy of checkpoint blockade in the clinic and will vary
depending on the range of pre-existing GC BCR specificities in each individual.
17 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
We did not find evidence for the release of self-reactive IgG1. Here, we reveal the role of
PD-1 in regulating acquired tolerance to non-self antigens (Goodnow, 2021). Unlike recent
tolerance models that address the role of the GC in preventing self-reactivity (Brink and Phan,
2018; Burnett et al., 2018; Degn et al., 2017; Sabouri et al., 2014), we propose a separate
mechanism driving tolerance to foreign antigen specificities encountered during normal
development. This CD4 T cell-controlled tolerance mechanism constrains maturation to pre-
existing GC B cell specificities to promote or permit commensalism for example. PD-1 blockade
enhanced pre-existing IgG1 binding to a foreign antigen, a range of commensals with evidence
for increased reactivity to a known pathobiont. We do see smaller, more variable changes to IgA
GC in the Peyer's Patch with small but significant increases in IgA PC in the bone marrow (but
not spleen or PP) after acute PD-1 blockade, however the impact on the local PP IgA GC program
remains uncharacterized. While IgA is more typically considered the humoral regulator of the
microbiome (Donaldson et al., 2018; Fagarasan et al., 2010; Guzman-Bautista et al., 2020; Huus
et al., 2021), anti-microbial IgG also impacts gut microbial composition (Chen et al., 2020;
Macpherson et al., 2018; Slack et al., 2009; Yilmaz et al., 2021).
The anti-inflammatory role of murine IgG1 is similar to IgG4 in humans (Strait et al., 2015)
hence restraining evolution of these GC B cell specificities offers one measure of homeostatic
regulation. Checkpoint inhibitor efficacy has also been linked to microbiome composition
(Gopalakrishnan et al., 2018; Matson et al., 2018; Routy et al., 2018). More recent studies connect
TFH and B cell activation, together with antibody production as predictors of checkpoint inhibitor
efficacy (Hollern et al., 2019). Therefore, understanding the impact of the IgG1 GC B cell
checkpoint in the context of health, or its release in disease or during clinical intervention remains
an important implication of our current studies.
18 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
PD-1 is highly expressed on sub-populations of CD4 TFH (Ansel et al., 1999; Fazilleau et
al., 2009; Vinuesa et al., 2005) and TFR cells responsible for the proliferation and differentiation
of class-switched B cells (Gonzalez-Figueroa et al., 2021; Linterman et al., 2011; Wu et al., 2020).
Typically, higher levels of PD-1 are expected on the GC-localized fractions of these cells (Haynes
et al., 2007; Shulman et al., 2013) and help to explain the indirect impact on GC B cells by acute
blockade. Unlike genetic modification of these subsets in the germline (Good-Jacobson et al.,
2010; Hams et al., 2011; Sage et al., 2013), there is no evidence of significant expansion within
the steady-state TFH or TFR in the spleen in the current study. Acute ablation of CD4 T cells
blocked the PD-1 enhanced IgG1 GC and PC expansion indicating the requirement of CD4s in
this B cell effector phenomenon. However, we predict that the CD4 T cell program controlling
IgG1 isotype-specific GC B cell and PC differentiation must be altered substantially to account for
the change in adaptive B cell immune homeostasis.
The selective targeting of PD-1 blockade to the IgG1 B cell response provides one clear
example of isotype-specific CD4 T cell immune regulation (Higgins et al., 2019; McHeyzer-
Williams et al., 2012). We and others have proposed that TFH (Gowthaman et al., 2019; Wang et
al., 2012; Zhang et al., 2020) and potentially TFR function (Gonzalez-Figueroa et al., 2021;
Vinuesa et al., 2016) is organized and delivered in an isotype-specific manner. We have
demonstrated the IgG2a-selective memory B cell requirement for T-bet expression, not only for
class-switch, but also for survival into memory and re-expansion at antigen recall (Wang et al.,
2012) and attribute this requirement to differential memory TFH organization. On the T cell
perspective, Eisenbarth and colleagues (Gowthaman et al., 2019) recently discovered a TFH
subtype with a selective role in controlling the IgE allergic. Here, we isolate the IgG1 pathway as
susceptible to PD-1 blockade with further preference for GC function and PC formation without
memory B cell involvement. The murine IgG1 isotype is known to be anti-inflammatory (Strait et
al., 2015), induced by the same CD4 T cell signals that drive IgE class switch (Snapper and Paul,
19 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
1987; Yang et al., 2020) and importantly, IgG1 GC B cells are considered the 'reservoir' B cell
state immediately upstream of IgE-secreting PCs (He et al., 2017; Turqueti-Neves et al., 2015).
Thus, we propose that this PD-1 adaptive B cell tolerance checkpoint operates as an antigen-
specific rheostat for IgG1 GC B cell maturation that prevents hypersensitivity to the commensal
microbiome and other environmental antigens.
Our observations identify PD-1 as a key tolerance checkpoint in the adaptive IgG1 isotype-
specific GC B cell to PC axis of differentiation. We found no consistent evidence of dysregulated
IgE GC or PC production in our studies or extending into Balb/c animals that favor IgE responses.
Hence, the broader connection to allergy must be explored with the notion that a second
checkpoint exists for the separate and subsequent induction of IgE producing antibodies in
allergy. We provide evidence for an IgG1-specific tolerance mechanism and how it is released
upon checkpoint blockade across both states of IgG1 GC B cell and IgG1 PC based immunity.
We propose molecular links between these functionally distinct but temporally related B cell
immune fate and functions. The molecular targets revealed could prove effective for clinical
modification at times of checkpoint blockade or otherwise. It will be important to extend these new
principles and further studies into humans at health, disease and following checkpoint blockade.
20 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
ACKNOWLEDGEMENTS
Graphical Abstract constructed using BioRender.com. This work was supported by the US
National Institutes of Health (AI047231, AI040215 and AI071182) and Bill & Melinda Gates
Foundation (BMGF 0PP1154835) to M.G.M.-W.
AUTHOR CONTRIBUTIONS
Conceptualization: C.R.D., L.M.W., M.M.W.; Design and Methodology: C.R.D., A.G.S., B.W.H.,
L.M.W., M.M.W.; Acquisition, Analysis and Interpretation of Data: C.R.D., A.G.S., B.W.H., L.M.W.,
M.M.W.; Writing and Review of manuscript: C.R.D., A.G.S., B.W.H., L.M.W., M.M.W.;
Supervision: L.M.W., M.M.W.; Funding Acquisition: M.M.W
DECLARATION OF INTERESTS
C.R.D. current affiliation is Flagship Labs 78 Inc., Cambridge, Massachusetts, USA.
21 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
REFERENCES
Ansaldo, E., Slayden, L.C., Ching, K.L., Koch, M.A., Wolf, N.K., Plichta, D.R., Brown, E.M.,
Graham, D.B., Xavier, R.J., Moon, J.J., et al. (2019). Akkermansia muciniphila induces
intestinal adaptive immune responses during homeostasis. Science 364, 1179-1184.
Ansel, K.M., McHeyzer-Williams, L.J., Ngo, V.N., McHeyzer-Williams, M.G., and Cyster, J.G.
(1999). In Vivo–Activated Cd4 T Cells Upregulate Cxc Chemokine Receptor 5 and Reprogram
Their Response to Lymphoid Chemokines. The Journal of Experimental Medicine 190, 1123-
1134.
Babicki, S., Arndt, D., Marcu, A., Liang, Y., Grant, J.R., Maciejewski, A., and Wishart, D.S. (2016).
Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res 44, W147-153.
Baldridge, J.R., and Crane, R.T. (1999). Monophosphoryl Lipid A (MPL) Formulations for the Next
Generation of Vaccines. Methods 19, 103-107.
Barber, D.L., Wherry, E.J., Masopust, D., Zhu, B., Allison, J.P., Sharpe, A.H., Freeman, G.J., and
Ahmed, R. (2006). Restoring function in exhausted CD8 T cells during chronic viral infection.
Nature 439, 682.
Brink, R., and Phan, T.G. (2018). Self-Reactive B Cells in the Germinal Center Reaction. Annu
Rev Immunol 36, 339-357.
Bunker, J.J., Erickson, S.A., Flynn, T.M., Henry, C., Koval, J.C., Meisel, M., Jabri, B.,
Antonopoulos, D.A., Wilson, P.C., and Bendelac, A. (2017). Natural polyreactive IgA antibodies
coat the intestinal microbiota. Science 358, eaan6619.
Burnett, D.L., Langley, D.B., Schofield, P., Hermes, J.R., Chan, T.D., Jackson, J., Bourne, K.,
Reed, J.H., Patterson, K., Porebski, B.T., et al. (2018). Germinal center antibody mutation
trajectories are determined by rapid self/foreign discrimination. Science 360, 223-226.
Cabrita, R., Lauss, M., Sanna, A., Donia, M., Skaarup Larsen, M., Mitra, S., Johansson, I., Phung,
B., Harbst, K., Vallon-Christersson, J., et al. (2020). Tertiary lymphoid structures improve
immunotherapy and survival in melanoma. Nature 577, 561-565.
22 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Chen, H., Albergante, L., Hsu, J.Y., Lareau, C.A., Lo Bosco, G., Guan, J., Zhou, S., Gorban, A.N.,
Bauer, D.E., Aryee, M.J., et al. (2019). Single-cell trajectories reconstruction, exploration and
mapping of omics data with STREAM. Nat Commun 10, 1903.
Chen, K., Magri, G., Grasset, E.K., and Cerutti, A. (2020). Rethinking mucosal antibody
responses: IgM, IgG and IgD join IgA. Nat Rev Immunol 20, 427-441.
Clement, R.L., Daccache, J., Mohammed, M.T., Diallo, A., Blazar, B.R., Kuchroo, V.K., Lovitch,
S.B., Sharpe, A.H., and Sage, P.T. (2019). Follicular regulatory T cells control humoral and
allergic immunity by restraining early B cell responses. Nat Immunol.
Crotty, S. (2019). T Follicular Helper Cell Biology: A Decade of Discovery and Diseases. Immunity
50, 1132-1148.
Cyster, J.G., and Allen, C.D.C. (2019). B Cell Responses: Cell Interaction Dynamics and
Decisions. Cell 177, 524-540.
Degn, S.E., van der Poel, C.E., Firl, D.J., Ayoglu, B., Al Qureshah, F.A., Bajic, G., Mesin, L.,
Reynaud, C.-A., Weill, J.-C., Utz, P.J., et al. (2017). Clonal Evolution of Autoreactive Germinal
Centers. Cell 170, 913-926.e919.
Donaldson, G.P., Ladinsky, M.S., Yu, K.B., Sanders, J.G., Yoo, B.B., Chou, W.-C., Conner, M.E.,
Earl, A.M., Knight, R., Bjorkman, P.J., et al. (2018). Gut microbiota utilize immunoglobulin A
for mucosal colonization. Science 360, 795-800.
Dougan, M., Luoma, A.M., Dougan, S.K., and Wucherpfennig, K.W. (2021). Understanding and
treating the inflammatory adverse events of cancer immunotherapy. Cell 184, 1575-1588.
Dyavar Shetty, R., Velu, V., Titanji, K., Bosinger, S.E., Freeman, G.J., Silvestri, G., and Amara,
R.R. (2012). PD-1 blockade during chronic SIV infection reduces hyperimmune activation and
microbial translocation in rhesus macaques. J Clin Invest 122, 1712-1716.
ElTanbouly, M.A., and Noelle, R.J. (2021). Rethinking peripheral T cell tolerance: checkpoints
across a T cell's journey. Nat Rev Immunol 21, 257-267.
23 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Fagarasan, S., Kawamoto, S., Kanagawa, O., and Suzuki, K. (2010). Adaptive Immune
Regulation in the Gut: T Cell–Dependent and T Cell–Independent IgA Synthesis. Annual
Review of Immunology 28, 243-273.
Fan, H.C., Fu, G.K., and Fodor, S.P. (2015). Expression profiling. Combinatorial labeling of single
cells for gene expression cytometry. Science 347, 1258367.
Fazilleau, N., McHeyzer-Williams, L.J., Rosen, H., and McHeyzer-Williams, M.G. (2009). The
function of follicular helper T cells is regulated by the strength of T cell antigen receptor binding.
Nature Immunology 10, 375-384.
Finak, G., McDavid, A., Yajima, M., Deng, J., Gersuk, V., Shalek, A.K., Slichter, C.K., Miller, H.W.,
McElrath, M.J., Prlic, M., et al. (2015). MAST: a flexible statistical framework for assessing
transcriptional changes and characterizing heterogeneity in single-cell RNA sequencing data.
Genome Biol 16, 278.
Gao, C., Zhang, M., and Chen, L. (2020). The Comparison of Two Single-cell Sequencing
Platforms: BD Rhapsody and 10x Genomics Chromium. Curr Genomics 21, 602-609.
Gonzalez-Figueroa, P., Roco, J.A., Papa, I., Nunez Villacis, L., Stanley, M., Linterman, M.A.,
Dent, A., Canete, P.F., and Vinuesa, C.G. (2021). Follicular regulatory T cells produce neuritin
to regulate B cells. Cell.
Good-Jacobson, K.L., Szumilas, C.G., Chen, L., Sharpe, A.H., Tomayko, M.M., and Shlomchik,
M.J. (2010). PD-1 regulates germinal center B cell survival and the formation and affinity of
long-lived plasma cells. Nature Immunology 11, 535-542.
Goodnow, C.C. (2021). COVID-19, varying genetic resistance to viral disease and immune
tolerance checkpoints. Immunol Cell Biol 99, 177-191.
Gopalakrishnan, V., Spencer, C.N., Nezi, L., Reuben, A., Andrews, M.C., Karpinets, T.V., Prieto,
P.A., Vicente, D., Hoffman, K., Wei, S.C., et al. (2018). Gut microbiome modulates response
to anti–PD-1 immunotherapy in melanoma patients. Science 359, 97-103.
24 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Gowthaman, U., Chen, J.S., Zhang, B., Flynn, W.F., Lu, Y., Song, W., Joseph, J., Gertie, J.A.,
Xu, L., Collet, M.A., et al. (2019). Identification of a T follicular helper cell subset that drives
anaphylactic IgE. Science 365.
Guzman-Bautista, E.R., Suzuki, K., Asami, S., and Fagarasan, S. (2020). Bacteria-immune cells
dialog and the homeostasis of the systems. Curr Opin Immunol 66, 82-89.
Hafemeister, C., and Satija, R. (2019). Normalization and variance stabilization of single-cell
RNA-seq data using regularized negative binomial regression. Genome Biol 20, 296.
Hams, E., McCarron, M.J., Amu, S., Yagita, H., Azuma, M., Chen, L., and Fallon, P.G. (2011).
Blockade of B7-H1 (Programmed Death Ligand 1) Enhances Humoral Immunity by Positively
Regulating the Generation of T Follicular Helper Cells. The Journal of Immunology 186, 5648-
5655.
Hashimoto, M., Kamphorst, A.O., Im, S.J., Kissick, H.T., Pillai, R.N., Ramalingam, S.S., Araki, K.,
and Ahmed, R. (2018). CD8 T Cell Exhaustion in Chronic Infection and Cancer: Opportunities
for Interventions. Annual Review of Medicine 69, 301-318.
Haynes, N.M., Allen, C.D.C., Lesley, R., Ansel, K.M., Killeen, N., and Cyster, J.G. (2007). Role of
CXCR5 and CCR7 in Follicular Th Cell Positioning and Appearance of a Programmed Cell
Death Gene-1High Germinal Center-Associated Subpopulation. The Journal of Immunology
179, 5099-5108.
He, J.-S., Subramaniam, S., Narang, V., Srinivasan, K., Saunders, S.P., Carbajo, D., Wen-Shan,
T., Hamadee, N.H., Lum, J., Lee, A., et al. (2017). IgG1 memory B cells keep the memory of
IgE responses. Nature Communications 8, 641.
Helmink, B.A., Reddy, S.M., Gao, J., Zhang, S., Basar, R., Thakur, R., Yizhak, K., Sade-Feldman,
M., Blando, J., Han, G., et al. (2020). B cells and tertiary lymphoid structures promote
immunotherapy response. Nature 577, 549-555.
Higgins, B.W., McHeyzer-Williams, L.J., and McHeyzer-Williams, M.G. (2019). Programming
Isotype-Specific Plasma Cell Function. Trends in Immunology 40, 345-357.
25 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Hollern, D.P., Xu, N., Thennavan, A., Glodowski, C., Garcia-Recio, S., Mott, K.R., He, X., Garay,
J.P., Carey-Ewend, K., Marron, D., et al. (2019). B Cells and T Follicular Helper Cells Mediate
Response to Checkpoint Inhibitors in High Mutation Burden Mouse Models of Breast Cancer.
Cell 179, 1191-1206 e1121.
Holmes, A.B., Corinaldesi, C., Shen, Q., Kumar, R., Compagno, N., Wang, Z., Nitzan, M.,
Grunstein, E., Pasqualucci, L., Dalla-Favera, R., et al. (2020). Single-cell analysis of germinal-
center B cells informs on lymphoma cell of origin and outcome. J Exp Med 217.
Huus, K.E., Petersen, C., and Finlay, B.B. (2021). Diversity and dynamism of IgA-microbiota
interactions. Nat Rev Immunol.
Ji, Z., and Ji, H. (2016). TSCAN: Pseudo-time reconstruction and evaluation in single-cell RNA-
seq analysis. Nucleic Acids Res 44, e117.
Kawamoto, S., Tran, T.H., Maruya, M., Suzuki, K., Doi, Y., Tsutsui, Y., Kato, L.M., and Fagarasan,
S. (2012). The Inhibitory Receptor PD-1 Regulates IgA Selection and Bacterial Composition in
the Gut. Science 336, 485-489.
Kennedy, D.E., Okoreeh, M.K., Maienschein-Cline, M., Ai, J., Veselits, M., McLean, K.C.,
Dhungana, Y., Wang, H., Peng, J., Chi, H., et al. (2020). Novel specialized cell state and spatial
compartments within the germinal center. Nat Immunol 21, 660-670.
King, H.W., Orban, N., Riches, J.C., Clear, A.J., Warnes, G., Teichmann, S.A., and James, L.K.
(2021). Single-cell analysis of human B cell maturation predicts how antibody class switching
shapes selection dynamics. Sci Immunol 6.
Ledergor, G., Weiner, A., Zada, M., Wang, S.Y., Cohen, Y.C., Gatt, M.E., Snir, N., Magen, H.,
Koren-Michowitz, M., Herzog-Tzarfati, K., et al. (2018). Single cell dissection of plasma cell
heterogeneity in symptomatic and asymptomatic myeloma. Nat Med 24, 1867-1876.
Li, H., Limenitakis, J.P., Greiff, V., Yilmaz, B., Scharen, O., Urbaniak, C., Zund, M., Lawson,
M.A.E., Young, I.D., Rupp, S., et al. (2020). Mucosal or systemic microbiota exposures shape
the B cell repertoire. Nature 584, 274-278.
26 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Linterman, M.A., Pierson, W., Lee, S.K., Kallies, A., Kawamoto, S., Rayner, T.F., Srivastava, M.,
Divekar, D.P., Beaton, L., Hogan, J.J., et al. (2011). Foxp3+ follicular regulatory T cells control
the germinal center response. Nature Medicine 17, 975-982.
Liu, B., Lin, Y., Yan, J., Yao, J., Liu, D., Ma, W., Wang, J., Liu, W., Wang, C., Zhang, L., et al.
(2021). Affinity-coupled CCL22 promotes positive selection in germinal centres. Nature 592,
133-137.
Low, M.S.Y., Brodie, E.J., Fedele, P.L., Liao, Y., Grigoriadis, G., Strasser, A., Kallies, A., Willis,
S.N., Tellier, J., Shi, W., et al. (2019). IRF4 Activity Is Required in Established Plasma Cells to
Regulate Gene Transcription and Mitochondrial Homeostasis. Cell Rep 29, 2634-2645 e2635.
Lu, Y., Jiang, R., Freyn, A.W., Wang, J., Strohmeier, S., Lederer, K., Locci, M., Zhao, H., Angeletti,
D., O'Connor, K.C., et al. (2021). CD4+ follicular regulatory T cells optimize the influenza virus-
specific B cell response. J Exp Med 218.
Macpherson, A.J., Yilmaz, B., Limenitakis, J.P., and Ganal-Vonarburg, S.C. (2018). IgA Function
in Relation to the Intestinal Microbiota. Annual Review of Immunology 36, 359-381.
Matson, V., Fessler, J., Bao, R., Chongsuwat, T., Zha, Y., Alegre, M.-L., Luke, J.J., and Gajewski,
T.F. (2018). The commensal microbiome is associated with anti–PD-1 efficacy in metastatic
melanoma patients. Science 359, 104-108.
McDonald, D., Clemente, J.C., Kuczynski, J., Rideout, J.R., Stombaugh, J., Wendel, D., Wilke,
A., Huse, S., Hufnagle, J., Meyer, F., et al. (2012). The Biological Observation Matrix (BIOM)
format or: how I learned to stop worrying and love the ome-ome. GigaScience 1.
McHeyzer-Williams, L.J., Milpied, P.J., Okitsu, S.L., and McHeyzer-Williams, M.G. (2015). Class-
switched memory B cells remodel BCRs within secondary germinal centers. Nature
Immunology 16, 296-305.
McHeyzer-Williams, M., Okitsu, S., Wang, N., and McHeyzer-Williams, L. (2012). Molecular
programming of B cell memory. Nature Reviews Immunology 12, 24-34.
27 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
McHeyzer-Williams, M.G., Nossal, G.J., and Lalor, P.A. (1991). Molecular characterization of
single memory B cells. Nature 350, 502-505.
McMurdie, P.J., and Holmes, S. (2013). phyloseq: an R package for reproducible interactive
analysis and graphics of microbiome census data. PLoS One 8, e61217.
Mesin, L., Ersching, J., and Victora, Gabriel D. (2016). Germinal center B cell dynamics. Immunity
45, 471-482.
Milpied, P., Cervera-Marzal, I., Mollichella, M.-L., Tesson, B., Brisou, G., Traverse-Glehen, A.,
Salles, G., Spinelli, L., and Nadel, B. (2018). Human germinal center transcriptional programs
are de-synchronized in B cell lymphoma. Nature Immunology 19, 1013.
Morais, L.H., Schreiber, H.L.t., and Mazmanian, S.K. (2021). The gut microbiota-brain axis in
behaviour and brain disorders. Nat Rev Microbiol 19, 241-255.
Mylvaganam, G.H., Chea, L.S., Tharp, G.K., Hicks, S., Velu, V., Iyer, S.S., Deleage, C., Estes,
J.D., Bosinger, S.E., Freeman, G.J., et al. (2018). Combination anti-PD-1 and antiretroviral
therapy provides therapeutic benefit against SIV. JCI Insight 3.
Nishimura, H., Nose, M., Hiai, H., Minato, N., and Honjo, T. (1999). Development of Lupus-like
Autoimmune Diseases by Disruption of the PD-1 Gene Encoding an ITIM Motif-Carrying
Immunoreceptor. Immunity 11, 141-151.
Nishimura, H., Okazaki, T., Tanaka, Y., Nakatani, K., Hara, M., Matsumori, A., Sasayama, S.,
Mizoguchi, A., Hiai, H., Minato, N., et al. (2001). Autoimmune Dilated Cardiomyopathy in PD-
1 Receptor-Deficient Mice. Science 291, 319-322.
Nutt, S.L., Hodgkin, P.D., Tarlinton, D.M., and Corcoran, L.M. (2015). The generation of antibody-
secreting plasma cells. Nat Rev Immunol 15, 160-171.
Pauken, K.E., Dougan, M., Rose, N.R., Lichtman, A.H., and Sharpe, A.H. (2019). Adverse Events
Following Cancer Immunotherapy: Obstacles and Opportunities. Trends Immunol 40, 511-523.
28 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Pelletier, N., McHeyzer-Williams, L.J., Wong, K.A., Urich, E., Fazilleau, N., and McHeyzer-
Williams, M.G. (2010). Plasma cells negatively regulate the follicular helper T cell program.
Nature Immunology 11, 1110-1118.
Peng, M., Wei, G., Zhang, Y., Li, H., Lai, Y., Guo, Y., Chen, Y., Liu, L., Xiao, H., Guan, H., et al.
(2021). Single-cell transcriptomic landscape reveals the differences in cell differentiation and
immune microenvironment of papillary thyroid carcinoma between genders. Cell Biosci 11, 39.
Petitprez, F., de Reynies, A., Keung, E.Z., Chen, T.W., Sun, C.M., Calderaro, J., Jeng, Y.M.,
Hsiao, L.P., Lacroix, L., Bougouin, A., et al. (2020). B cells are associated with survival and
immunotherapy response in sarcoma. Nature 577, 556-560.
Qi, H. (2016). T follicular helper cells in space-time. Nature Reviews Immunology 16, 612-625.
Roco, J.A., Mesin, L., Binder, S.C., Nefzger, C., Gonzalez-Figueroa, P., Canete, P.F., Ellyard, J.,
Shen, Q., Robert, P.A., Cappello, J., et al. (2019). Class-Switch Recombination Occurs
Infrequently in Germinal Centers. Immunity 51, 337-350 e337.
Routy, B., Chatelier, E.L., Derosa, L., Duong, C.P.M., Alou, M.T., Daillère, R., Fluckiger, A.,
Messaoudene, M., Rauber, C., Roberti, M.P., et al. (2018). Gut microbiome influences efficacy
of PD-1–based immunotherapy against epithelial tumors. Science 359, 91-97.
Roy, K., Mitchell, S., Liu, Y., Ohta, S., Lin, Y.S., Metzig, M.O., Nutt, S.L., and Hoffmann, A. (2019).
A Regulatory Circuit Controlling the Dynamics of NFkappaB cRel Transitions B Cells from
Proliferation to Plasma Cell Differentiation. Immunity 50, 616-628 e616.
Sabouri, Z., Schofield, P., Horikawa, K., Spierings, E., Kipling, D., Randall, K.L., Langley, D.,
Roome, B., Vazquez-Lombardi, R., Rouet, R., et al. (2014). Redemption of autoantibodies on
anergic B cells by variable-region glycosylation and mutation away from self-reactivity.
Proceedings of the National Academy of Sciences 111, E2567-E2575.
Sage, P.T., Francisco, L.M., Carman, C.V., and Sharpe, A.H. (2013). The receptor PD-1 controls
follicular regulatory T cells in the lymph nodes and blood. Nature Immunology 14, 152-161.
29 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Sage, P.T., Ron-Harel, N., Juneja, V.R., Sen, D.R., Maleri, S., Sungnak, W., Kuchroo, V.K.,
Haining, W.N., Chevrier, N., Haigis, M., et al. (2016). Suppression by TFR cells leads to durable
and selective inhibition of B cell effector function. Nature Immunology 17, 1436-1446.
Sharma, P., Siddiqui, B.A., Anandhan, S., Yadav, S.S., Subudhi, S.K., Gao, J., Goswami, S., and
Allison, J.P. (2021). The Next Decade of Immune Checkpoint Therapy. Cancer Discov 11, 838-
857.
Sharonov, G.V., Serebrovskaya, E.O., Yuzhakova, D.V., Britanova, O.V., and Chudakov, D.M.
(2020). B cells, plasma cells and antibody repertoires in the tumour microenvironment. Nat Rev
Immunol 20, 294-307.
Sharpe, A.H., and Pauken, K.E. (2018). The diverse functions of the PD1 inhibitory pathway.
Nature Reviews Immunology 18, 153-167.
Shi, J., Hou, S., Fang, Q., Liu, X., Liu, X., and Qi, H. (2018). PD-1 Controls Follicular T Helper
Cell Positioning and Function. Immunity 49, 264-274.e264.
Shi, W., Liao, Y., Willis, S.N., Taubenheim, N., Inouye, M., Tarlinton, D.M., Smyth, G.K., Hodgkin,
P.D., Nutt, S.L., and Corcoran, L.M. (2015). Transcriptional profiling of mouse B cell terminal
differentiation defines a signature for antibody-secreting plasma cells. Nat Immunol 16, 663-
673.
Shulman, Z., Gitlin, A.D., Targ, S., Jankovic, M., Pasqual, G., Nussenzweig, M.C., and Victora,
G.D. (2013). T follicular helper cell dynamics in germinal centers. Science 341, 673-677.
Slack, E., Hapfelmeier, S., Stecher, B., Velykoredko, Y., Stoel, M., Lawson, M.A.E., Geuking,
M.B., Beutler, B., Tedder, T.F., Hardt, W.-D., et al. (2009). Innate and Adaptive Immunity
Cooperate Flexibly to Maintain Host-Microbiota Mutualism. Science 325, 617-620.
Snapper, C.M., and Paul, W.E. (1987). Interferon-γ and B Cell Stimulatory Factor-1 Reciprocally
Regulate Ig Isotype Production. Science 236, 944-947.
30 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Sokal, A., Chappert, P., Barba-Spaeth, G., Roeser, A., Fourati, S., Azzaoui, I., Vandenberghe,
A., Fernandez, I., Meola, A., Bouvier-Alias, M., et al. (2021). Maturation and persistence of the
anti-SARS-CoV-2 memory B cell response. Cell 184, 1201-1213 e1214.
Strait, R.T., Posgai, M.T., Mahler, A., Barasa, N., Jacob, C.O., Köhl, J., Ehlers, M., Stringer, K.,
Shanmukhappa, S.K., Witte, D., et al. (2015). IgG1 protects against renal disease in a mouse
model of cryoglobulinaemia. Nature 517, 501-504.
Street, K., Risso, D., Fletcher, R.B., Das, D., Ngai, J., Yosef, N., Purdom, E., and Dudoit, S.
(2018). Slingshot: cell lineage and pseudotime inference for single-cell transcriptomics. BMC
Genomics 19, 477.
Stuart, T., Butler, A., Hoffman, P., Hafemeister, C., Papalexi, E., Mauck, W.M., 3rd, Hao, Y.,
Stoeckius, M., Smibert, P., and Satija, R. (2019). Comprehensive Integration of Single-Cell
Data. Cell 177, 1888-1902 e1821.
Tan, C.L., Kuchroo, J.R., Sage, P.T., Liang, D., Francisco, L.M., Buck, J., Thaker, Y.R., Zhang,
Q., McArdel, S.L., Juneja, V.R., et al. (2021). PD-1 restraint of regulatory T cell suppressive
activity is critical for immune tolerance. J Exp Med 218.
Turqueti-Neves, A., Otte, M., Schwartz, C., Schmitt, M.E., Lindner, C., Pabst, O., Yu, P., and
Voehringer, D. (2015). The Extracellular Domains of IgG1 and T Cell-Derived IL-4/IL-13 Are
Critical for the Polyclonal Memory IgE Response In Vivo. PLoS Biol 13, e1002290.
Velu, V., Titanji, K., Zhu, B., Husain, S., Pladevega, A., Lai, L., Vanderford, T.H., Chennareddi,
L., Silvestri, G., Freeman, G.J., et al. (2009). Enhancing SIV-specific immunity in vivo by PD-1
blockade. Nature 458, 206-210.
Victora, G.D., Schwickert, T.A., Fooksman, D.R., Kamphorst, A.O., Meyer-Hermann, M., Dustin,
M.L., and Nussenzweig, M.C. (2010). Germinal center dynamics revealed by multiphoton
microscopy with a photoactivatable fluorescent reporter. Cell 143, 592-605.
31 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Vinuesa, C.G., Cook, M.C., Angelucci, C., Athanasopoulos, V., Rui, L., Hill, K.M., Yu, D.,
Domaschenz, H., Whittle, B., Lambe, T., et al. (2005). A RING-type ubiquitin ligase family
member required to repress follicular helper T cells and autoimmunity. Nature 435, 452-458.
Vinuesa, C.G., Linterman, M.A., Yu, D., and MacLennan, I.C.M. (2016). Follicular helper T cells.
Annual Review of Immunology 34, 335-368.
Wang, N.S., McHeyzer-Williams, L.J., Okitsu, S.L., Burris, T.P., Reiner, S.L., and McHeyzer-
Williams, M.G. (2012). Divergent transcriptional programming of class-specific B cell memory
by T-bet and RORα. Nature Immunology 13, 604-611.
Wei, S.C., Levine, J.H., Cogdill, A.P., Zhao, Y., Anang, N.A.S., Andrews, M.C., Sharma, P., Wang,
J., Wargo, J.A., Pe'er, D., et al. (2017). Distinct Cellular Mechanisms Underlie Anti-CTLA-4 and
Anti-PD-1 Checkpoint Blockade. Cell 170, 1120-1133 e1117.
Wu, A.R., Neff, N.F., Kalisky, T., Dalerba, P., Treutlein, B., Rothenberg, M.E., Mburu, F.M.,
Mantalas, G.L., Sim, S., Clarke, M.F., et al. (2014). Quantitative assessment of single-cell
RNA-sequencing methods. Nat Methods 11, 41-46.
Wu, X., Wang, Y., Huang, R., Gai, Q., Liu, H., Shi, M., Zhang, X., Zuo, Y., Chen, L., Zhao, Q., et
al. (2020). SOSTDC1-producing follicular helper T cells promote regulatory follicular T cell
differentiation. Science 369, 984-988.
Yang, Z., Wu, C.M., Targ, S., and Allen, C.D.C. (2020). IL-21 is a broad negative regulator of IgE
class switch recombination in mouse and human B cells. J Exp Med 217.
Yilmaz, B., Mooser, C., Keller, I., Li, H., Zimmermann, J., Bosshard, L., Fuhrer, T., Gomez de
Aguero, M., Trigo, N.F., Tschanz-Lischer, H., et al. (2021). Long-term evolution and short-term
adaptation of microbiota strains and sub-strains in mice. Cell Host Microbe 29, 650-663 e659.
Zhang, B., Liu, E., Gertie, J.A., Joseph, J., Xu, L., Pinker, E.Y., Waizman, D.A., Catanzaro, J.,
Hamza, K.H., Lahl, K., et al. (2020). Divergent T follicular helper cell requirement for IgA and
IgE production to peanut during allergic sensitization. Sci Immunol 5.
32 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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
Ham IgG anti-PD-1 A 105 B 106 C *** * PC PC ** 105 4±1 9±1 104 104 (Elispot) 103 (Elispot) 103 CD138 2 2 Total IgG1 secreting cells Total 10 Total IgG1 secreting cells Total 10 RMP1 29F Rat Steady Ham J43 B220 IgG2a anti-PD-1 State IgG anti PD-1
Plasma Cells D Ham IgG anti-PD-1 E F 105 *** Ham IgG ** 4±1 60±4 60 anti-PD-1
104 40
103 20 9±1 9±1 Plasma Cells (%) IgG1
2 Frequency of switched 0
Total IgG1 Plasma Cells Total 10 Ham anti IgG3 IgG2b IgG2a IgG1 IgG2b IgG PD-1
G H Xbp1 Prdm1 Mzb1 Mki67 State 1 Naive B State 2 Steady State IgG1 PC IgG PC State 3 TSCAN PCA2 anti PD-1 IgG1 PC Expression Expression Expression Expression TSCAN PCA1 marker TSCAN 0 6 0 5 0 8 0 6
Cytosol & Secreted Bank1 Nucleus Bach2 I Eif2ak3 J Ets1 anti PD-1 Bcl2 anti PD-1 Atf4 Naive H2-Oa Naive Ell2 100 PC Fut8 100 PC Pou2af1 Eif2ak3 Mpc2 Bach2 Prdm1 Bcl2 Spcs1 Rel Irf2 H2-Oa Edem1 -10 Xbp1 Gnai2 10 Mcm2 -10 Bank1 Nfkbia Mcm4 10 Pten Mzb1 Smad4 Ikzf4 Bax -20 Klf7 Lsp1 Rexo2 10 Mcm4 Ube2c Casp4 Ube2c Mcm2 Txk Cd74 Pcna -20 Tgfb2 Top2a 10 Eef2 Blnk 10-30 Ccnd2 Bcl6b Ccl21a Lsp1 Ikzf4 Ciita Il22 Eef2 Bcl6b Klf7 Il1a Slpi Atf4 Pparg -30 H2-DMa Igbp1 -40 Pcna Ikzf2 10 Lamp1 H2-DMa 10 Prdm1 Mki67 Fut8 Il23a Skil Top2a Ltbp3 Pten Pparg Klf4 Ctsz Tgfb2 -50 Relb Skil Ssr2 Il1a 10 Pou2af1 Cdkn1b 10-40 Blnk Lamp1 Irf4 Relb Nfkbia Ltbp3 Cdkn1b Gata3 10-50 Edem1 Bax -60 Ikzf2 Bhlha15 Slpi Ssr2 10 Gata3 Runx2 10-60 Aurkb Ctsz Mki67 Spi1 p-value (Mast Negative Binomial) Mpc2 Gnai2 p-value (Mast Negative Binomial) Ell2 Ccnd2 10-70 Spcs1 Txk Bhlha15 Smad4 -80 Xbp1 -80 Mzb1 Il22 10 Irf4 10 Rexo2 Ccl21a 10-100 Polh Polh Steady anti Steady anti -2-4 0 24 6 8 Naive State PD-1 -1 0 1 -2-4 0 24 6 Naive State PD-1 -1 0 1 Fold Change Log2 Plasma Cells Fold Change Log2 Plasma Cells
33 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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. Acute PD-1 blockade induces extensive IgG1 Plasma Cell differentiation
(A) Total IgG1 secreting splenic cells by Elispot from mice treated 6 days prior with isotype control
(Rat IgG2a, clone 2A3, n=4) or anti-PD-1 antibodies (clones, RMP1-14 or 29F.1A12, n=4 each),
mean±SEM. (B) Untreated mice (Steady State, n=4) and mice treated 6 days prior with isotype
control (Hamster IgG, n=4) or anti-PD-1 (clone J43, n=4) total IgG1 secreting cells, mean±SEM.
(C) Flow cytometric analysis of class-switched Plasma Cells (PC: IgD- IgM- CD138+) and (D)
distribution of IgG1 and IgG2b with (E) quantification for IgG1 (n=9) and (F) distribution of IgG
sub-classes, (n=6) mean±SEM.
(G) TSCAN pseudotime analysis of qtSEQ generated transcriptional programming from single
FACS sorted Naïve B cells, IgG PC from Steady State, and IgG1 PC from anti-PD1 treated mice
with (H) marker TSCAN gene expression. See also Figure S1.
(I) Volcano plot of Naïve B cells compared to anti-PD1 IgG1 PC with averaged single cell
heatmap (including Steady State PC) from cytosol and secreted gene products and (J) nucleus.
See Figure S1 for cell membrane gene products and single cell heatmaps.
34 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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
A Ham IgG anti-PD-1 B C 5 10 Ham IgG GC GC 12±2 19±3 60 anti-PD-1
40 104
Total IgG1 Total 20
Mem Mem Memory B cells
GL7 60±2 52±2 Memory B Cells (%) 103 Frequency of switched 0 CD38 Ham anti IgG PD-1 IgG3 IgG2b IgG2a IgG1
D Germinal Center Cells E F Ham IgG anti-PD-1 5 *** 10 Ham IgG 5±1 42±5 60 anti-PD-1 ***
40 104
Total IgG1 Total 20 12±1 8±1 GC B Cells (%) IgG1 Frequency of switched Germinal Center B cells 103 0 IgG2b Ham anti IgG3 IgG2b IgG2a IgG1 IgG PD-1
G H Cd79a Bcl6 Aicda Mki67 State 1 Naive B
State 2 Steady State IgG1 GC State 3 anti PD-1 TSCAN PCA2 IgG1 GC
Expression Expression Expression Expression TSCAN PCA1 marker TSCAN 0 5 0 4 0 4 0 4
Membrane Selplg Nucleus Tcf3 I Steady anti H2-Aa J Steady anti Irf5 Havcr1 State PD-1 Pou2af1 State PD-1 Cd24a Ikzf4 0 Irf2 10 GC GC Ptprc 0 GC GC Cd276 10 Nfatc3 10-1 Tnfsf9 Ube2c Cd79b -2 Pax5 10 E2f4 -2 Lgals1 Icam1 Nr4a1 Bhlha15 10 Selplg Tgfbr1 Smad4 Klf13 Trp53 H2-M1 Tnfrsf25 Fcrla 10-4 Bcl6 -3 Tnfrsf25 Bcl6 10 Tgfbr1 Irf2 Pcna Havcr1 Tnfrsf8 Bach2 Cd79a -6 Polh 10-4 Flt3 10 Ikzf1 Pparg Ptprc Lgals1 Aicda Mcm2 Il2rg E2f4 -5 Spn 10-8 Id3 10 Il1r2 Trp53 Ikzf3 H2-Aa Notch1 Pcna Tnfrsf11a Bach2 -6 Cd69 -10 Nfatc3 Ikzf1 10 Notch1 Flt3 10 Tcf3 Il2ra Smad4 Bhlha15 -7 Cd81 Spi1 10 Fcrla Tfrc 10-12 Ube2c Aicda Cd69 Ly9 Top2a Mcm4 10-8 Cd48 Spn Ets1 Mki67 Cd79a Notch2 10-14 Mcm4 Polh Top2a -9 Tnfsf9 Slc7a5 Klf13 10 Cd244 Mki67 H2-K1 -16 Pparg Ets1 p-value (Mast Negative Binomial) Il1r2 Cd48 p-value (Mast Negative Binomial) 10 Ell2 -10 Il2ra Mcm2 10 Cd79b -20 Pou2af1 Runx1 10-20 Cd24a H2-K1 10-40 Ikzf2 10-40 Tmem176b 10 Jund -1-2 0 1 2 Steady anti -1 0 1 -1-2 0 1 2 Steady anti -1 0 1 Fold Change Log2 Naive State PD-1 Fold Change Log2 Naive State PD-1
IgG1 GC Steady anti Bcl6 Bach2 Mki67 Cd24a K State PD-1 6 4 4 20
Steady State
0 0 0 0 Naive
6 6 32 6 anti PD-1
0 0 0 0 Pseudotime
35 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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. Selective expansion of IgG1 Germinal Center B cells without memory B cell
involvement
(A) Flow cytometric analysis from mice 6 days after control IgG or anti-PD1 treatment displaying
class-switched (IgD- IgM- CD138-) GC (CD38- GL7+) and memory (CD38+ GL7-) B cells, with
(B) memory B cell quantification of IgG1, (C) distribution of memory IgG sub-classes and (D) GC
distribution of IgG1 and IgG2b with (E) quantification for GC IgG1 (n=9) and (F) distribution of GC
IgG sub-classes, (n=6) mean±SEM.
(G) TSCAN pseudotime analysis of qtSEQ from single FACS sorted Naïve B cells, IgG1 GC
from Steady State and anti-PD1 treated mice with (H) marker TSCAN gene expression. See
also Figure S2.
(I) Volcano plot of Steady State compared to anti-PD1 IgG1 GC with averaged single cell
heatmap from cell membrane gene products and (J) nucleus. See Figure S2 for cytosol and
secreted gene products and single cell heatmaps.
(K) Stream developmental trajectories of Naïve B cells with Steady State IgG1 GC compared to
anti-PD-1 IgG1 GC with sample UMI gene expression.
36 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 A Non-IgG1 GC B cells IgG1+ GC B cells Steady State anti-PD-1 B Steady State anti-PD-1 1.5 1.5 *** LZ LZ LZ LZ 44±1 42±3 72±5 41±1 1.0 1.0
0.5 0.5 DZ DZ DZ DZ
52±1 54±3 19±1 53±1 Ratio IgG1 DZ:LZ CD86 CD86 Ratio non-IgG1 DZ:LZ Germinal Center B cells 0 Germinal Center B cells 0 CXCR4 Steady anti CXCR4 Steady anti State PD-1 State PD-1
C IgG1 GC Cd86 Cxcr4 D anti-PD-1 IgG1 GC cells Differential Gene Expression 2.4 10-60 Naive 0.6 S2 to S3:DZ
Steady -40 State 10 Starting 0 0 State S4 5 0.8 10-20 S2 anti
PD-1 p-value (Stream)
S3 p=0.0001 100 0 0 Il10 Tcf3 Blnk Bcl6 Spi1 Polh Eef2 E2f4 Pcna Cd74 Mzb1 Mpc2 Aicda Trp53 Aurkb Log2 Top2a Mki67 Pparg Gnai2 Mcm4 Mcm2 Mcm3 Tnfsf9 H2-Aa Dock8 Bach2 Ube2c Cd24a Cd79b Lamp1 Smad4 H2-Ab1 Bhlha15 Single Cell Trajectory Pou2af1
Steady State IgG1 GC Igbp1 Rc3h1 IgG1 GC H2.DMb2 Nfkbia E Steady State H2.Aa H2.Ab1 State 1 H2.Ob Light Zone Blk Ltb 2 Polh 1 Cd74 Mzb1 Lsp1 0 Relb Spib 3 Cd79a -2 2 State 2 Lamp1 Transitional TSCAN PCA2 Pou2af1 Aurkb Tgfb2 -2 0 2 4 Foxj1 Itch TSCAN PCA1 Pparg Lgals1 State 1 2 3 Spp1 Tnfsf9 MST 321 Mki67 path State 3 Mcm2 Dark Zone Mcm4 Pcna Row Z-score Pseudotime 0 2-2 anti PD-1 IgG1 GC Cd74 CD79a IgG1 GC H2.DMb2 H2.K1 anti PD-1 H2.DMa F Blnk H2.Ab1 4 State 1 H2.Aa Light Zone Cd79b Tnfsf13c 2 Foxj1 2 Spi1 Xcl1 0 Spp1 Tgfb2 3 1 Plxnd1 -2 Gata1
State 2 Top2a TSCAN PCA2 Mki67 Transitional Ube2c -4 -2 0 2 4 Aurkb Tnfsf9 TSCAN PCA1 Mpc2 Blk State 1 2 3 Polh Trp53 MST 321 Cd24a Mcm2 path State 3 Mcm4 Dark Zone Pcna Pou2af1 Row Z-score Pseudotime 0 2-2 Transition Genes G H 10-20 S4 64% Cd83 Polh S3 10-15 33% S5 85% -10 1 10 4.5 S0 6% S6 32% 10-5 p-value (Stream) 100 0 0 S1 46% Itgb2 Pcna Aicda Aurkb Aurkb Top2a Top2a Mki67 Mki67 Mcm4 Bach2 Ube2c Ube2c Nfatc3 Ube2c Cd24a Cd24a Lamp3 Smad4 Pseudotime Pseudotime Pseudotime Smad4 Pou2af1 Naive GC Steady anti State PD-1 S1 S3 S4 S5 S6
37 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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. Exaggerated IgG1 GC B cell programing upon PD-1 blockade
(A) Flow cytometric analysis from Steady State mice or 6 days after anti-PD1 treatment displaying
class-switched non-IgG1 and (B) IgG1 GC B cells distributed across light zone (CD86+CXCR4-)
and dark zone (CD86-CXCR4+) with relative ratios (n=4) mean±SEM.
(C) Stream developmental trajectories of single cell qtSEQ for Naïve B cells with Steady State
IgG1 GC compared to anti-PD-1 IgG1 GC with distribution of Cd86 and Cxcr4 gene expression.
Identification of starting state with diverging branches S2, S3, and S4 defined for the anti-PD-1
IgG1 GC. (D) Differential gene expression of S2 to S3 dark zone displaying the most significant
gene products.
(E) TSCAN pseudotime reconstruction of minimum spanning tree (MST) path and single cell
heatmap displayed across each of the 3 states (State 1 LZ, State 2 Transitional, State 3 DZ) of
IgG1 GC for Steady State compared to (F) anti-PD-1 treated mice. Gene products in blue are
represented in both compartments.
(G) Stream trajectory of combined Naïve, IgG1 GC Steady State and anti-PD1 treated mice for
Cd83 and Polh gene products. (H) Stream map with each state identified for each diverging
branch (S0 to S6) and the percent contribution of anti-PD1 GC B cells, the most significant
transition genes detected for each state is displayed at right.
38 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 4
A B anti-PD-1 Bcl6 Bach2 Mki67 PC 2.5 GC GC 2
0 1
UMAP 2 UMAP -2.5 3 TSCAN PCA2 Naive Naive -5 PC UMAP 1 Steady State Naive -5 -2.5 0 2.5 5 anti-PD-1 TSCAN PCA1 0 4 0 4 0 5 Xbp1 Irf4 Bhlha15 Pou2af1 UMAP
PC GC Steady State anti PD-1 PC non IgG1 non IgG1 non IgG1 PC IgG1 Expression IgG1 IgG1 PC IgG1 GC IgG1 0 5 0 4 0 6 GC non IgG1 marker TSCAN 0 6 GC IgG1
Membrane Cd69 Nucleus Rel Il2ra Id3 C Ifngr1 D Bach2 anti PD-1 Fas anti PD-1 Bcl6 0 0 GC PC Flt3 10 GC PC Ets1 10 H2-Ab1 Rara Pcna Cux2 Ptprc Irf5 Mki67 Aicda Il31r Cux2 Klf7 Ikzf1 Cd79a 10-10 Spib Lef1 Pax5 -5 Cxcr3 Havcr1 Ikzf3 Skil Spib 10 Il2ra Cd274 Ifitm2 Mcm2 Runx2 Mcm3 Ifngr1 Cd80 Cxcr5 Rel Ciita Tcf3 Fas Cd24a -20 Tcf3 Atf4 Ifnar2 Cd19 10 Mcm3 Prdm1 Ikzf3 Cd79b Cd72 Skil -10 Cxcr5 Cd9 Tle1 Ccnd2 10 Flt3 Tnfsf14 Ltb Ikzf1 Ikzf4 Bhlha15 Ifitm2 Tnfsf11 Cd24a Ets1 Bcl6b Cdkn1b Tnfrsf4 Cd79b -30 Pax5 Klf7 Cd19 10 Id3 Relb Havcr1 Ly6c1 Cd48 Prdm1 -15 Cd48 Ly6c1 Bhlha15 Irf4 10 Il31r Cd44 Cd44 Runx2 Cd69 Tnfsf10 -40 Cdkn1b Cd72 Cd28 10 Xbp1 Ifnar2 Bach2 Gata3 Ccnd2 Ptprc Cd28 Irf4 Cd79a Cd81 Cd274 Ell2 10-20 Tmem176b Ly75 Atf4 Ltb Spn Notch2 10-50 Gata3 Lgals1 Lgals1 Bcl6b Selplg Cd9 Bcl6 Ikzf4 Cd40lg Cd81 Ikzf2 Mki67 -50 H2-Ab1 Ly75 Il1r2 Ell2 Pcna p-value (Mast Negative Binomial) p-value (Mast Negative Binomial) 10 Il1r2 Cd40lg Xbp1 Mcm2 Notch2 H2-K1 10-100 Aicda Irf5 H2-K1 Cd80 Polh 10-100 10-150 Rara -1 0 1 -1 0 1 -6 -2-4 0 24 6 Naive GC PC -6 -2-4 0 24 6 Naive GC PC Fold Change Log2 Fold Change Log2
E Pseudotime Cell Type Cd79b Bcl6 Xbp1
Naive anti PD-1 0 3010 GC PCGC 0 2-1 0 2 0 4
F G Pcna Notch2 Steady State
PC
GC
0 20 0 3 0 10 0 3 anti PD-1 PC Ssr2 Bcl6b
GC
Pseudotime Naive GC PC 0 8 0 4 0 5 0 3
39 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 4. Anti-PD-1 induced transition of IgG1 GC B cell to PC program
(A) Single cell qtSEQ UMAP of Naïve B cell, GC and PC from Steady State and 6 days after anti-
PD1 treated mice.
(B) TSCAN pseudotime reconstruction of minimum spanning tree (MST) path for Naïve B cells
and IgG1 GC and PC from anti-PD-1 treated mice with representative marker TSCAN gene
expression. See also Figure S3-A.
(C) Volcano plot comparing IgG1 GC and PC from anti-PD-1 treated mice with averaged single
cell heatmap of cell membrane gene products and (D) nucleus. See Figure S3-B for cytosol and
secreted gene products and Figure S3-C,D single cell heatmaps.
(E) Slingshot cell lineage and pseudotime inference with sample temporal gene expression of
Naïve B cell, Steady State IgG1 GC and anti-PD-1 IgG1 GC and PC.
(F) Stream map of Naïve B cells with Steady State IgG1 GC and PC compared to Naïve B cells
with anti-PD-1 IgG1 GC and PC along pseudotime. (G) Comparative gene expression overlaid on
pseudotime distribution with Stream and Slingshot for distinguishing sample genes across PC
anti-PD1 subsets. See also Figure S3-E for volcano comparison.
40 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 5 Ham IgG anti-PD-1 A 1± 3±1 0.3
Live Bacteria 26±1 IgG1 Viability (dead) Viability
SYTO BC Viability (dead)
6 B Ham IgG * anti-PD-1 4 4 3 2 2 1 serum IgG1 (% of live bacteria) (% of live bacteria) 0 0 Cecum bacteria bound No 10 1413 1615 Cecum bacteria bound No Ham anti serum Age of mice (weeks) serum IgG PD-1
C No serum Ham IgG serum anti-PD-1 serum 0.03 7±1 12±3 34±3
0 SYTO BC SYTO IgG1
FSC-A Viability (dead)
anti-PD-1 treated anti-PD-1 treated D Post sort Ham IgG treated serum IgG1 E No serum IgG1 serum IgG1 Increased with serum Increased with Ham IgG treated anti-PD-1 treated 0 anti-PD-1 0 anti-PD-1 serum IgG1 serum IgG1 Lactobacillus Muribaculum Bacteroides Ruminococcus reuteri intestinale -2 champanellensis Parabacteroides 76±6 92±2 -5 Barnesiella Lactobacillus Ethanoligenens Lactobacillus viscericola harbinense johnsonii Porphyromonas -4 Bacteroides Roseburia gingivalis cellulosilyticus [Eubacterium] Lactobacillus -10 rectale Bacteroidales johnsonii Clostridium Flavonifractor -6 Helicobacter plautii -15 Helicobacter Helicobacter hepaticus Bacteroides cellulosilyticus p adj value (log10) -8 Helicobacter p adj value (log10) Helicobacter IgG1 typhlonius -20 -20 typhlonius Viability (dead) -1-2-4 0 1 2 3 -2-4-9 0 2 4 6 Fold Enrichment (log2) Fold Enrichment (log2)
F G Decreased with Increased with anti-PD-1 anti-PD-1 100 Ham IgG
10-1 Clostridium formicaceticum [Clostridium] Blautia hansenii
anti PD-1 Thermaerobacter Cloacibacillus -2 marianensis porcorum 10 Ruminococcus bicirculans Paenibacillus sp. FSL R7-0273 Bifidobacterium Luteitalea pratensis animalis 10-3 Alistipes shahii Blautia hansenii Prevotella fusca Peptostreptococcaceae Alistipes finegoldii
[Eubacterium] sulci bacterium oral taxon 929 Acholeplasma oculi Luteitalea pratensis Tannerella forsythia Tannerella p-value (Mann-Whitney) Bacillus weihaiensis Paenibacillus larvae Ammonifex degensii Truepera radiovictrix Truepera [Clostridium] bolteae Ruminococcus albus Desulfallas gibsoniae Eubacterium limosum Salimicrobium jeotgali Anaerococcus prevotii Burkholderia sp. BDU8 Butyrivibrio fibrisolvens Clostridium perfringens -4 Cloacibacillus porcorum
Planococcus rifietoensis Candidatus Acholeplasma brassicae Bifidobacterium animalis [Clostridium] sphenoides Streptococcus pyogenes
Clostridium argentinense 10 Treponema sp. OMZ 838 Treponema Laceyella sp. FBKL4.010
Novibacillus thermophilus Ishikawaella capsulata Fimbriimonas ginsengisoli Megamonas hypermegale Ruminococcus bicirculans Butyrivibrio proteoclasticus Planococcus sp. MB-3u-03 Anaerotignum propionicum Clostridium formicaceticum Alkaliphilus metalliredigens -6 Syntrophobotulus glycolicus Lactobacillus paraplantarum Christensenella massiliensis Pseudoxanthomonas spadix Desulfotomaculum reducens
Desulfitobacterium hafniense 10 Geosporobacter ferrireducens Thermaerobacter marianensis Synechococcus sp. PCC 7502 Synechococcus sp. PCC 7336 Paenibacillus sp. FSL R7-0273 Paenibacillus sp. FSL
Z-score Geobacillus thermodenitrificans Burkholderiales bacterium YL45 Ruminiclostridium thermocellum -10 -2-4-6 0 2 4 6 8 Tannerella sp.oral taxon HOT-286 Tannerella [Clostridium] thermosuccinogenes Clostridiaceae bacterium 14S0207 Candidatus Ishikawaella capsulata Ruminococcaceae bacterium CPB6 Fold Change
Peptostreptococcaceae oral taxon 929 -1 1
41 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 5. Pre-existing gut microbiota specificities released by checkpoint blockade
(A) Flow cytometry analysis of IgG1 binding to cecal bacteria isolated from untreated mice and
incubated with serum from Hamster IgG or anti-PD-1 treated mice. (B) Quantification of frequency
of live bacteria coated with serum IgG1 as defined in (A). Serum samples are from 5 independent
experiments (n=3 for each age). Pooled data from mice 14 weeks or older (n=9).
(C) Sorting strategy for IgG1-bound freshly isolated cecal bacteria after incubation with control
serum or serum from 6 days after anti-PD-1 treatment. (D) Post-sort analysis and volcano plot
showing comparison of serum IgG1 specificity from control-treated (n=3) to anti-PD-1 (n=4) mice
or (E) no serum to anti-PD1 treated serum IgG1.
(F) Heatmap displaying relative z-score for changes in species abundance and (G) volcano plot
with fold change and significant difference (n=3).
42 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 6
A Ham IgG anti-PD-1 B Steady State anti-CD4
0.2± 0.2± Memory B 55±1 0.001 0.03 0.02 ±0
87±1 89±1 CD4
CD8 3±1 32±6 Plasma Cells 105 C Steady 104 State anti-CD4 66±1 32±6 IgG1 Total anti-CD4 &
Memory B cells anti-PD-1 103 104 105 6±1 32±3 GC B * * * 104 103 103 4±0.4 4±0.4 IgG1
2 IgG1 GC B cells Total 2
Total IgG1 Plasma Cells Total 10 10 PD-L1 D Ham IgG anti-PD-1 F Ham IgG anti-PD-1 24hr Treg TFR 22±2 19±1 7±2 0.1±0.02 28±2 24±1
TFH Foxp3 44±2 15±1 34±3 14±2
6 days Cxcr5
13±2 4±0.2 7 G 10 Ham IgG anti-PD-1
106 PD-1 Total CD4 cells Total Cxcr5 105 ETH Treg TFR TFH E 16 ** 500 400 *** 12 * *** 300 *** 8 200 CD4 T cells CD4 T
Frequency of 4 Ham IgG 100
Activated cells (%) anti-PD-1 PD-1 high CD4 T cells T PD-1 high CD4
0 PD-1 expression (MFI) 0 24hr 6 days ETH Treg TFR TFH Time after treament
43 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 6. CD4 T cell dependent mechanism without local follicular T cell expansion
(A) Flow cytometry analysis of PD-L1 expression on splenic class-switched memory B cells, PC
and GC B cells after 6 days of isotype control or treatment with anti-PD-1.
(B) Mice treated with CD4-depleting antibody GK1.5 (anti-CD4) and (C) quantification of total
IgG1 memory B cells, PC and GC with or without anti-PD-1 after CD4 depletion (n=3) mean±SEM.
(D) Splenic PD-1 levels on activated CD4+ T cells (TCRb+ CD8- CD4+ CD44hi) following 24 hours
or 6 days after control IgG or anti-PD-1 treatment with (E) quantification of frequency changes.
(F) Intracellular flow cytometry of activated CD4 T cell subsets (eFluor780 viability-, GR1-, B220-
TCRb+ CD8- CD4+ CD44 hi) separated into major compartments after control antibody or anti-
PD-1 treatment 6 days prior and (G) quantification of total cell numbers and PD-1 expression,
(n=3) mean±SEM.
44 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
EXPERIMENTAL MODEL DETAILS
Mice.
Male 8-18 week old C57BL/6 and congenic C57BL/6-Ly5.1 (B6.SJL-Ptprca Pepcb/BoyJ) mice
were bred and housed under specific-pathogen-free (SPF) conditions at The Scripps Research
Institute. Mice were euthanized with isoflurane prior to sample collection and analysis. All
experiments were performed with protocols approved by Institutional Animal Care and Use
Committee at The Scripps Research Institute.
METHOD DETAILS
In vivo Antibody treatments & Immunization.
For in vivo blockade of PD-1, 300µg anti-mouse PD-1 antibody (clone J43, 29F.1A12 or RMP1-
14, BioXcell) was administered intraperitoneal (IP). Control mice were given 300µg polyclonal
Armenian hamster IgG, or rat IgG2a isotype control, anti-trinitrophenol (clone 2A3, BioXcell). For
CD4 depletion, 300µg anti-mouse CD4, GK1.5 monoclonal antibody (BioXcell) was administered.
Mice were analyzed 6 days (or as indicated) post-blockade treatment.
The hapten 4-hydroxy-3-nitrophenylacetyl (NP, Biosearch) was conjugated to keyhole limpet
hemocyanin (KLH, Pierce; NP-KLH, in house preparation, (Baldridge and Crane, 1999;
McHeyzer-Williams et al., 1991), or anti-PD-1 antibody clone J43 (BioXcell; NP-J43).
Immunizations were performed by subcutaneous (sc) or IP injection of 300 or 400µg immunogen
with & without Monophosphoryl lipid A (MPL)-based adjuvant (2% squalene, 0.5 mg MPL, in
house preparation,(Baldridge and Crane, 1999), as indicated.
Flow cytometry.
Single cell suspensions from spleen were prepared (Wang et al., 2012). After lysing with ACK red
cell lysis buffer (Lonza) and passed through a 70 µm cell strainer (Fisher Scientific), cells were
45 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
washed and resuspended at 4 x 108 cells/mL in FACS buffer: 2% FCS, 10 mM HEPES in phenol
red-free DMEM with L-glutamine (Gibco). Draining lymph nodes for base of tail immunization
(inguinal and periaortic lymph nodes, LN) were combined and processed together. All processing
and staining were performed on ice. Prior to staining, Fc receptors were blocked with 10 µg/mL
of purified antibody clone 2.4G2 (BioXcell) for 10 min.
For surface staining, a master mix containing a panel of fluorophore- or biotin-labeled antibodies
was mixed at pre-titrated concentrations in Brilliant Violet Staining Buffer (BD Biosciences) at 2X,
and added to suspended cells at a 1:1 ratio for a final staining concentration of 2 x 108 cells/mL
for 30-45 min. For secondary staining of biotin-labeled antibodies, cells were washed and
resuspended in FACS buffer. Fluorophore-conjugated streptavidin was added for 15-30 min.
Cells were washed and resuspended in FACS buffer before analysis.
For panels including surface staining of IgG, cells were first stained with single isotype-specific
fluorophore- or biotin-conjugated antibodies for 30 min. After washing and resuspending in FACS
buffer, cells were blocked with 1% normal rat serum and 1% normal mouse serum (Jackson
ImmunoResearch) for 10 min. Next, staining using the master mix of additional antibodies and
streptavidin reagents was performed as above. If the isotype-specific antibody was biotin-
conjugated, the fluorophore-conjugated streptavidin was included with the master mix stain.
For intracellular FACS, surface stains were completed as described above. Cells were incubated
with eBioscience Fixable Viability Dye eFluor 780 (Invitrogen) for 10 min, washed with FACS
buffer, then fixed and permeabilized using the eBioscience Foxp3 / Transcription Factor Staining
Buffer Set (Invitrogen). Cells were incubated in 1X Fixation/Permeabilization buffer for 25 min and
washed with 1X Permeabilization buffer. Prior to intracellular staining, nonspecific binding was
blocked by incubation with 1% normal mouse serum and 1% normal rat serum for 10 min. Cells
46 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 then stained with fluorophore-conjugated antibodies specific for intracellular antigens.
Finally, cells were washed once with 1X Permeabilization buffer, and once with FACS buffer prior
to resuspension in FACS buffer for analysis on the flow cytometer.
All flow cytometry sample data was collected using a BD FACSAria III on a workstation running
FACSDiva software version 8.0.1 (BD Biosciences). Cytometry data analysis was performed
using FlowJo version 10 (TreeStar). FACS data is presented with the following gates: plasma
cells (Gr1- CD3e- IgD- IgM- CD138+), germinal center B cells (Gr1-CD3e-IgD-IgM-CD138-
CD19+B220+CD38-GL7+), memory B cells (Gr1-CD3e-IgD-IgM-CD138-CD19+B220+GL7-
CD38+), NP-specific B cells (Gr1-CD3e-CD19+ and/or CD138+IgD-IgM-NP+Lambda1+),
activated CD4 T cells (eFluor780 Viability-Gr1-B220-TCRβ+CD8-CD4+CD44hi).
Elispot.
As previously described (Pelletier et al., 2010), multiScreen-HA sterile 0.45 µm surfactant-free
mixed cellulose ester membrane filter plate (Millipore) wells were coated with 50 µl of 25 µg/mL
unlabeled goat anti-mouse IgG (SouthernBiotech), anti-mouse IgG1 (RMG1-1, BioLegend),
polyclonal Hamster IgG (BioXcell), or J43 (BioXcell), for 4 hours at room temperature or overnight
at 4°C. The plate was washed 3 times with sterile PBS prior to plating samples. Cells were
resuspended in RPMI medium with L-glutamine (Gibco) containing 5% fetal calf serum (FCS),
10 mM HEPES, 100 units/mL penicillin, and 100 µg/mL streptomycin, and plated in two-fold
dilutions. Negative control wells were included, for which medium only was plated. After
incubation overnight at 37°C and 5% CO2, the plate was washed 3 times with 0.1% Tween 20 in
PBS and 3 times with PBS, then incubated with horse radish peroxidase (HRP) conjugated goat
anti-mouse Fc-specific antibodies in FACS buffer with 5% FCS and 10% dry instant nonfat milk.
The plate was incubated for 4 hours at room temperature, washed as in the previous step, and
developed with chromogen substrate: 0.03% 3-amino-9-ethylcarbazole (Sigma-Aldrich) and
47 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
0.015% H2O2 in 1M sodium acetate, pH 4.8-5.0. Spots were enumerated under a dissection
microscope, and plasma cell numbers were determined by extrapolation from total cell counts. All
positive wells containing discernible spots were averaged together to obtain per-sample final
counts.
FACS analysis of microbiome-specific serum Ig.
Bacterial flow cytometry was performed as previously described (Bunker et al., 2017). Cecum
contents were squeezed into a sterile 2 mL tube on ice with 1 mL FACS Buffer and homogenized
by vortexing horizontally for 5 min at 3,000 rpm. After centrifugation for 400 x g for 5 min to
eliminate large debris, the supernatant was centrifuged once again at 8000 x g for 5 min to pellet
cecum bacteria. The supernatant was then aspirated, and the bacterial pellet washed twice with
FACS buffer. The washed bacteria were resuspended in 1 mL FACS buffer and quantified by
taking the optical density at 600 nm using the DU 530 Life Science UV/Vis Spectrophotometer
(Beckman). Bacterial concentration was determined assuming that:
C OD = # !"" 5 × 10$
where C# represents the concentration in bacteria/mL. Bacteria were resuspended in FACS buffer
and all subsequent staining steps were performed at 2 x 108 cells/mL. All processing was
performed on ice.
Bacteria were first stained with eBioscience Fixable Viability Dye eFluor 780 (Invitrogen) and
STYO BC Green Fluorescent Nucleic Acid Stain (Invitrogen) at 1:750 dilution each for 30 min.
Bacteria were washed with FACS buffer and centrifuged at 8,000 x g for 5 min. To assess serum
Ig binding, bacteria were incubated with two-fold serial dilutions of serum, beginning with a
1:12.5 dilution to a 1:X dilution, for 45 min. Bacteria were washed and stained with phycoerythrin
48 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
(PE)-conjugated anti-IgG1 antibody for 20 min. After a final wash step, bacteria were
resuspended in FACS buffer and analyzed using a FACSAria III flow cytometer.
To sort bacteria for downstream 16S rRNA sequencing, bacteria were sorted into 5 mL tubes
containing 500 µl DMEM (phenol red-free, no antibiotics). Sorted samples were centrifuged at
13,000 x g for 10 min. The supernatant was aspirated, and the pellet was frozen at -80°C until
genomic DNA library preparation (at least overnight).
16S rRNA sequencing analysis.
Genomic DNA was purified from sorted bacteria, or raw cecum or feces. Raw cecum or feces was
frozen for at least overnight at -80°C before DNA purification. Bacterial DNA was purified using
the DNeasy PowerSoil kit (Qiagen) as per manufacturer’s instructions, under sterile conditions.
DNA was eluted off the spin column with sterile nuclease-free water (Teknova) and quantified
using the Qubit dsDNA High Sensitivity Assay Kit (Invitrogen). A microbial community standard
(Zymo Research) was included in genomic DNA purification steps and subsequent library
preparation and analysis.
16S rRNA library preparation was performed using the 16S Barcoding Kit from Oxford Nanopore
Technologies (#SQK-RAB204), as per manufacturer’s instructions. Twelve samples including the
microbial community standard and a reagent-only control were processed separately in parallel.
Briefly, 10 ng of genomic DNA was amplified using LongAmp Taq 2X Master Mix (New England
BioLabs) and the provided barcoded 16S primers. PCR was performed using an initial
denaturation for 1 min at 95°C, followed by 25-30 cycles of: 20 sec at 95°C, 30 sec at 55°C, and
2 min at 65°C, and a final extension of 5 min at 65°C. Cleanup of PCR product was performed
using AMPure XP beads (Beckman Coulter), and the concentration of amplified DNA for each
sample was measured using the Qubit dsDNA HS Assay Kit as above. Assuming a PCR product
49 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
length of 1500Kb, all samples were pooled in roughly equal molar proportion. A MinION Flow Cell
version R9.4.1 was prepared for 1D sequencing as per manufacturer’s instructions, and a total
amount 50-100 fmol of DNA was sequenced using the MinION Mk1B and MinKNOW (Oxford
Nanopore Technologies). Fast5 files generated over the course of the 48-hour sequencing run
were transferred to a computing cluster for downstream basecalling and analysis.
Pre-processing of reads collected using MinKNOW version 3.1.19 was performed using Albacore
version 2.3.3 (Oxford Nanopore Technologies) and custom scripts to basecall, de-barcode, and
convert to FASTQ separate directories of 40,000 reads in parallel. These FASTQ files were
concatenated by barcode and trimmed using Porechop version 0.2.3
(https://github.com/rrwick/Porechop) with the default options selected.
Pre-processing of reads collected using MinKNOW version 3.1.19 was performed using the
guppy_basecaller module of Guppy version 2.3.5 (Oxford Nanopore Technologies) using the
following options: qscore_filtering, min_qscore 7, calib_detect, records_per_fast 0. De-barcoding
was performed separately using the guppy_barcoder module of Guppy, specifying the SQK-
RAB204 kit and -q value of 0.
Alignment of basecalled, trimmed, and de-barcoded reads was performed using Kraken version
2.0.7-beta (https://ccb.jhu.edu/software/kraken2). The Kraken 2 reference library used for
sequence assignment was built from Archaea and Bacteria sequences downloaded from the
NCBI database on August 12, 2018, using the default settings. Sequence assignment reports
were converted to BIOM format (McDonald et al., 2012) using kraken-biom version 1.0.1
(https://github.com/smdabdoub/kraken-biom) and then converted to a DESeq object using the
phyloseq R package version 1.29.0 (McMurdie and Holmes, 2013). DESeq2 was used to
calculate differential representation of assigned taxa between groups, using the default multiple-
50 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
inference correction (Benjamini-Hochberg). Prior to analysis, geometric means of counts were
used to estimate size factors for normalization. Expression-based heatmap visualization was
preformed using Heatmapper (Babicki et al., 2016) (http://www.heatmapper.ca) with average
linkage clustering and Pearson distance measurement.
Integrated Single Cell RNA-seq (qtSEQ).
Single cell transcriptional analysis was carried out using a custom RNA sequencing
protocol called quantitative targeted single cell sequencing (qtSEQ). This technique
sequences polyA mRNA using a nested set of 3’ targeted primers to amplify specific gene
sets.
Single naïve B cells (Gr1-FceR1a-IgD+IgMlowCD138-GL7-CD79b+CD19+CD38+), plasma cells
(Gr1-FceR1a-IgD-IgM-CD138+IgG1+), germinal center cells (Gr1-FceR1a-IgD-IgM-CD138-
CD79b+CD19+CD38-GL7+IgG1+) from 8 individual mice were sorted into 384 well plates at 4C.
Doublets were excluded using SSC-A vs SSC-H. Each well contained 1µL of reverse transcription
(RT) master mix consisting of 0.03μL SuperScript II taq, 0.2μL SuperScript II 5x buffer, 0.012μL
of 25μM dNTPs, 0.03μL DTT (0.1M stock), 0.03μL RNaseOUT (Invitrogen), 0.19μL
DNase/RNase Free Water and 0.5μL of extended oligoDT primers (1μM stock, custom made,
IDT). To allow assessment of contamination from sample processing, 8 negative control well with
no cell sorted were interspersed across each plate. Immediately after sorting, reverse
transcription (RT) was performed at 42˚C for 50 mins then at 80˚C for 10 mins.
Following RT, all 384 wells were pooled within each plate and excess oligoDT primer removed
using ExoSAP-IT Express (Affymetrix) at 37C for 20mins. cDNA was purified and volume reduced
using a 0.8x AMpure XP SPRIselect magnetic bead (BeckmanCoulter) solution to 1x library ratio.
51 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 each plate two nested PCR reactions were used to amplify specific gene targets then a final
PCR reaction perfomed to attach Illumina sequencing adapters. Library preparation was carried
out on each plate separately to preserve cell labeling.
PCR 1 was performed in a 60μL reaction: 35μL cDNA library in H2O, 1.5μL RNaseH (5000U/mL,
NEB), 1.2μL Phusion DNA polymerase (2000U/mL, NEB), 12μL 5X Phusion HF Buffer (NEB),
0.85μL of 25μM dNTPS (Invitrogen), 1.2μL of 100μM RA5-overhang primer (custom made,
5’AAGCAGTGGTGAGTTCTACAGTCCGACGATC 3’) and 25nM final concentration for each
specific gene targeting external primer using the following thermocycling conditions: 37ºC for 20
minutes, 98ºC for 3 minutes, followed by 10 cycles of 95ºC for 30 seconds, 60ºC for 3 minutes,
72 ºC for 1 minute and final elongation at 72ºC for 5 minutes. Removal of excess primers and
volume reduction was performed again using 0.9x AMpure XP SPRIselect beads to 1x library
ratio.
PCR2 was performed in a 20μL reaction: 10μL PCR1 reaction elute, 0.4μL Phusion DNA
polymerase (2000U/mL, NEB), 4μL 5X Phusion Buffer (NEB), 0.3μL of 25μM dNTPs (Invitrogen),
2μL of 20μM RA5 (custom made, 5’ GAGTTCTACAGTCCGACGATC 3’), and 25nM final
concentration for each specific gene targeting internal primer using the following thermocycling
conditions: 95ºC for 3 minutes followed by 10 cycles of 95ºC for 30 seconds, 60ºC for 3 minutes,
72 ºC for 1 minute and final elongation at 72ºC for 5 minutes. Removal of excess primers and
volume reduction was performed using AMpure XP SPRIselect beads at a 0.9x bead to 1x library
ratio.
PCR3 was performed in a 20μL reaction: 10μL PCR2 reaction elute, 0.4μL Phusion DNA
polymerase (2000U/mL, NEB), 4μL 5X Phusion Buffer (NEB), 0.3μL of 25μM dNTPs (Invitrogen),
2μL of 10μM RP1 primer (Illumina), 2μL of 10μM RPI library specific primer (Illumina) and 1.3μL
52 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
water using the following thermocycling conditions: 95ºC for 3 minutes followed by 8 cycles of
95ºC for 15 seconds, 60ºC for 30 seconds, 72 ºC for 30 seconds and final elongation at 72ºC for
5 minutes. Removal of excess primers and volume reduction was performed using 0.7x AMpure
XP SPRIselect beads to 1x library ratio and then eluted in 20μL H2O for quantification and library
sequencing preparation.
Final cDNA amplified library concentration and quantity was determined using a Qubit 4
Fluorometer (Invitrogen Technologies) and 2100 BioAnalyzer (Agilent). Separately labeled cDNA
libraries were mutliplexed according to manufacturer’s protocol (Illumina, San Diego CA).
Libraries were sequenced using an Illumina NextSeq 500 (Read 1: 19 cycles; Index Read 1: 6
cycles; Read 2: 67 cycles).
QUANTIFICATION AND STATISTICAL ANALYSIS
Data Processing and Normalization.
Bcl2fastq (Illumina) was used to demultiplex sequencing libraries using plate barcodes and assort
reads by well barcodes prior to sequence alignment using Bowtie2 (v2.2.9) to a custom genome
consisting of the amplicon regions for the specifically targeted genes used in the experiment
based on the murine genome GRCm38.p6, version R97 (Ensembl). HTseq was used for read
and UMI reduced tabulation. Custom scripts for these steps available upon request.
Quality control excluded cells that contained fewer than 24 UMI counts and had fewer than 14
unique genes detected. A secondary quality control filter included naïve B cells that contained 2
of 3 genes (Cd79a, Cd79b, Ptprc) n=158, germinal center cells containing 2 of 5 genes (Cd79a,
Cd79b, Aicda, Bcl6, Fas) n=623, and plasma cells containing 2 of 5 genes (Cd79a, Cd79b,
Prdm1, Irf4, Xbp1) n=266. A total 1047 B cells from 2 experiments were analyzed.
53 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Data from the two separate experiments was pooled into a single count matrix for processing
using the R package Seurat (v3.2.3) (Stuart et al., 2019); scaling and normalization was carried
out separately by cell type using the ScTransform package (v0.3.2.9002, (Hafemeister and Satija,
2019) with the batch_var option set to the metadata variable “Experiment”. This normalized UMI
count matrix was then used for downstream analysis.
scRNA Sequencing Computational Analysis.
Using the gene expression UMI normalized count matrices, Tools for Single Cell Analysis
(TSCAN) (v1.0, (Ji and Ji, 2016) constructed cluster centroids, minimum spanning tree (MST)
pseudo-temporal path, identified underlying marker genes and heatmap expression of top
significant genes (FDR p<0.05) that increased along each pseudotime MST path. For the
heatmaps generated, each column represents a cell that was mapped and ordered by its
pseudotime value with each row denoting significant genes along the MST path.
For differential gene expression analyses, MAST (Model-based Analysis of Single Cell
Transcriptomics) package (v1.16.0 https://github.com/RGLab/MAST/ (Finak et al., 2015)) using a
hurdle model was used to generated fold change and p-values for all genes. Volcano plots were
then created using Prism v9.0 (GraphPad Software). The R package pheatmap (v1.0.12) was
used to plot heatmaps of averaged UMI counts with euclidean distance hierarchical clustering.
Single cell heatmaps were generated using the “Doheatmap” function from the Seurat package.
Normalized UMI count matrices were used for psuedotime trajectory analysis with STREAM
(Single-cell Trajectories Reconstruction, Exploration And Mapping of omics data)(Chen et al.,
2019)(https://stream.pinellolab.partners.org/) and Slingshot (Street et al., 2018). Slingshot
trajectory analysis was performed using the “dyno” analysis package
(https://github.com/dynverse/dyno).
54 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Uniform Manifold Approximation and Projection (UMAP) was performed using Biovinci (v3.0.9
Bioturing) on all genes. Statistical tests were calculated and graphed using Prism v9.0 (GraphPad
Software). A P value *<0.05, **<0.01, ***<0.001 was considered statistically significant.
55 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 S1
A 106 B 106 C D 3000 500
1000 300 104 104
300 Total number Total number Total number Total count per cell of UMI per cell
2 2 Normalized UMI of Reads per cell 10 of Reads per cell 10 100 100
30
4 samples 4 samples Steady anti Steady anti 4 samples 4 samples Steady anti Steady anti 4 mice 4 mice Naive State PD-1 State PD-1 4 mice 4 mice Naive State PD-1 State PD-1 B B Experiments GC PC Experiments GC PC E Membrane 105
104
103
102
101 Reads per Plasma Cel l 100 Cd24a Lgals1 Cd81 CD79b Cd28 Cd9 Ly75 Cd69 Havcr1 Spn Cd48 Slamf9 Itgal Itgb2 Cd44 Icam1 H2-K1 H2-Aa Tnfrsf13c Fcrla Tmem H2-Ab1 Notch1 Selplg Il1r2 Folr1 Tnfrsf1b 176b BCR Co-receptors & Carbohydrate Receptors Cell Adhesion MHC, TNF-family, Growth Factor Receptors F Cytosol & Secreted 105
104
103
102
101 Reads per Plasma Cel l 100 Mzb1 Lamp1 Igbp1 Edem1 Eef2 Fut8 Ssr2 Itch Slpi Blnk Aurkb Txk NFKBia Gnai2 H2-DMb2 H2-DMa Cd74 Mpc2 Lsp1 Spcs1 Rexo2 Ctsz Tgfb2 Protein Regulation Signal Transduction Antigen Processing Metabolism, Cytoskeleton, Secreted Nucleus G 105
104
103
102
101 Reads per Plasma Cel l 100 Mcm4 Polh Ube2c Mki67 Pcna Ccnd2 Irf4 Spi1 Smad4 Skil Ikzf4 Klf11 Pou2af1 Pparg Xbp1 Ell2 Plac8 Bhlha15 Atf4 Jund Relb DNA Replication Cell Cycle Transcription Factors: WHForkhead, Smad, C2H2ZF Transcription Factors: other & Repair
Membrane Membrane Membrane Cytosol & Secreted Nucleus H anti Cd38 Cd38 I Bank1 J Bach2 PD-1 Cxcr5 Cxcr5 Eif2ak3 Ets1 Naive Sell Sell Bcl2 Atf4 100 PC H2-Ab1 H2-Ab1 H2-Oa Ell2 Havcr1 Havcr1 Fut8 Pou2af1 -2 Tnfrsf13c Tnfrsf13c Prdm1 10 Cxcr5 Mpc2 Ly6c1 Plxnd1 Plxnd1 Spcs1 Irf2 -4 Ccr7 Slamf6 Tnfsf18 Tnfsf18 Edem1 Xbp1 10 Slamf6 Mcm4 Cd38 Tnfrsf13c Slamf6 Nfkbia Timd2 Cd44 Cd44 Ikzf4 -6 Sell Mzb1 10 Tnfsf18 Cd69 Cd69 Rexo2 Ube2c Cd44 Tmem176b Tmem176b Casp4 Mcm2 10-8 Cd9 Cd81 Cd81 Cd74 Pcna Slamf9 Itgb2 Itgb2 Blnk Bcl6b 10-10 Fcrla H2-K1 H2-K1 Lsp1 Ciita Tnfrsf4 Cd28 Cd28 Eef2 Klf7 Plxnd1 Slamf9 Slamf9 Slpi Pparg 10-20 Cd28 Cd9 Cd9 Igbp1 Ikzf2 Tmem176b Fcrla Fcrla H2-DMa Mki67 -30 Cd48 Cd24a Cd24a Il23a Top2a 10 Cd81 Itgal Itgal Pten Klf4 Lgals1 Lgals1 Tgfb2 Skil -40 Tnfsf9 10 Spn Il1r2 Il1r2 Il1a Cdkn1b Cd40lg Ly75 Ly75 Lamp1 Relb -50 Lgals1 Notch2 Notch2 Ltbp3 Gata3 10 Selplg Selplg Bax Bhlha15 p-value (Mast Negative Binomial) Ly75 Spn Spn Runx2 10-60 Cd24a Ssr2 H2-K1 Tnfrsf4 Tnfrsf4 Ctsz Spi1 Timd2 Timd2 Ccnd2 -70 Notch2 Gnai2 10 Il1r2 Cd40lg Cd40lg Txk Smad4 Cd48 Cd48 Il22 Irf4 -2-4 0 24 6 Tnfsf9 Tnfsf9 Ccl21a Polh Fold Change Log2 Steady anti -1 0 1 Steady anti 0 2-11 Steady anti 0 2-11 Steady anti 0 2-11 Naive State PD-1 Naive State PD-1 Naive State PD-1 Naive State PD-1 Plasma Cells Plasma Cells Plasma Cells Plasma Cells
Figure S1. Integrated single cell RNA-seq analysis (qtSEQ) of Plasma Cell differentiation after PD-1 blockade, Related to Figure 1 (A) Single-cell RNA-sequencing with qtSEQ violin plots showing total reads per cell from sorted cells that passed quality control divided across experiments and (B) cell type. (C) Total UMI counts per cell across experiments and (D) cell type after normalization using SCTransform. (E) Reads per cell from sorted anti-PD-1 IgG1 PCs for selected assays targeting gene products from cell membrane, (F) cytosol and secreted, and (G) nucleus. (H) Volcano plot of Naïve B cells compared to anti-PD1 IgG1 PC with averaged single cell heatmap (including Steady State PC) and single cell heatmap from UMI counts for cell membrane gene products, (I) cytosol and secreted and (J) nucleus.
56 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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.
Day 0 Day 2 Day 4 Day 6 Figure S2 A 105 * Plasma 3±2 2±1 4±3 56±7 Cells 104
3
Total IgG1 Total 10 Plasma Cells
IgG1 102 IgG2b 0 2 4 6 Days post anti-PD-1 Day 0 Day 2 Day 4 Day 6 B 105 ** GC 5±0.5 8±1 10±4 50±4 Cells
104 Total IgG1 Total IgG1 Germinal Center B cells 103 0 2 4 6 IgG2b C Membrane Days post anti-PD-1 105
104
103
102
101
100 Reads per Germinal Center Cel l Cd79b Cd79a Cd81 Ptprc Cd24a Itgb2 Itgal Cd48 Spn Cd9 Havcr1 Cd72 Icam1 Lgals1 Timd2 H2-K1 H2-Aa H2-Ab1 Tnfrsf13c Ltb Tnfrsf10b Fcrla Flt3 Ifngr1 Cxcr5 BCR Co-receptors Cell Adhesion Carbohydrate Receptors MHC TNF-family Cytokine / Chemokine Receptors
Cytosol & Secreted D 104
103
102
101
100 Reads per Germinal Center Cel l Igbp1 Mzb1 Lamp1 Eef2 Cd74 Itch Pten Eif2a Blk Blnk Aurkb Rgs10 Lck Dock8 Gnai2 H2-DMb2 H2-DMa H2-Ob Lsp1 Mpc2 Sat1 Ctsz Protein Regulation Signal Transduction Antigen Processing Metabolism, Cytoskeleton E Nucleus
104
103
102
101
100 Reads per Germinal Center Cel l Mcm4 Mcm2 Top2a Aicda Mcm3 Mki67Ube2c Pcna Ccnd2 Bach2 Bcl6 Ikzf1 Batf Spi1 Ets1 Irf2 Irf5 Spib Pou2af1 Id3 Pax5 Nfatc3 Relb Stat3 Stat5bSmad4 Trp53 Kdm2bPparg DNA Replication & Repair Cell Cycle Transcription Factors: WHForkhead Homeobox, HLH, Rel, Stat, Smad Tumor suppressor, C2H2ZF & BLZ family Paired-box Chromatin binding, Nuclear receptor
Membrane Cytosol & Secreted Cytosol & Secreted Cytosol & Secreted Nucleus F Selplg G Bank1 Bank1 H Tcf3 H2-Aa Steady anti Eif2ak3 Eif2ak3 Irf5 Havcr1 State PD-1 H2-DMa H2-DMa Pou2af1 Cd24a Igbp1 Igbp1 Ikzf4 Ptprc 100 GC GC Ctsz Ctsz Irf2 Cd276 Gnai2 Gnai2 Nfatc3 Tnfsf9 Lck Lck Ube2c -2 Cd74 Pax5 Cd79b 10 Bcl2 Cd74 E2f4 Icam1 Il27 Eif2ak3 Dock8 Dock8 Smad4 Tgfbr1 Fgl2 Rexo2 Blnk Blnk Trp53 Eef2 Fcrla 10-4 Dusp2 Eef2 Bcl6 Tnfrsf25 Mpc2 Il10 Il10 Pcna Tnfrsf8 Igbp1 Lamp1 Lamp1 Polh Cd79a -6 Bank1 Mpc2 Mpc2 Pparg Lgals1 10 H2-DMa Prmt5 Prmt5 Mcm2 Il2rg Eif2a Edem1 Edem1 Id3 Il1r2 Lsp1 Lsp1 Ikzf3 -8 Spcs1 Notch1 10 Dock8 Stk11 Stk11 Bach2 Flt3 H2-DMb2 Eif2a Eif2a Ikzf1 Cd81 Ero1l Ero1l Bhlha15 Gnai2 Spi1 Tfrc 10-10 Lamp1 Mtor Mtor Aurkb Aicda Ly9 Lsp1 Aurkb Mcm4 Spn Blk Mzb1 Mzb1 Mki67 Notch2 Aurkb Spcs1 Spcs1 Top2a
Slc7a5 p-value (Mast Negative Binomial) Rexo2 -20 Blnk Rexo2 Klf13 Cd244 10 Cd74 Gclc Gclc Ets1 Cd48 Eef2 Ssr2 Ssr2 Ell2 Il2ra Mzb1 Blk Blk Runx1 H2-K1 10-30 Dusp2 Dusp2 Ikzf2 Tmem176b -1-2 0 1 2 H2-DMb2 H2-DMb2 Jund Steady anti Fold Change Log2 Steady anti -1 0 1 Steady anti Steady anti Naive State PD-1 0 2-11 Naive State PD-1 Naive State PD-1 0 2-11 Naive State PD-1 0 2-11 GC Cells GC Cells GC Cells GC Cells
Figure S2. Kinetics of PD-1 blockade and single cell qtSEQ of the Germinal Center, Related to Figure 2 (A) Timecourse analysis of mice either untreated (day 0) or treated with anti-PD-1 for splenic PC and (B) GC B cells at various days with quantification of total IgG1 cells (n=3) mean±SEM. (C) Reads per cell from sorted anti-PD-1 IgG1 GC for selected assays targeting gene products from cell membrane, (D) cytosol and secreted, and (E) nucleus. (F) Single cell heatmap from UMI counts for cell membrane gene products and (G) Volcano plot of Steady State compared to anti-PD1 IgG1 GC with averaged single cell heatmap and single cell heatmap from cytosol and secreted gene products. (H) Single cell heatmap from gene products from nucleus. 57 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 S3
A Aicda Prdm1 Batf Cdkn1b Pcna
Expression marker TSCAN 0 4 0 5 0 3 0 4 0 6
Cytosol & Screted Cytosol & Secreted Cytosol & Secreted B H2-Ob H2-Ob anti PD-1 Blk Blk 0 GC PC Bank1 Bank1 10 Rc3h1 Rc3h1 Tgfb3 Dusp2 Dusp2 H2-Oa H2-Aa H2-Aa -10 H2-Ob Tgfb3 Tgfb3 10 Dusp2 Blnk Il10 Il10 Dock8 Aurkb H2-DMb2 H2-DMb2 Bank1 Txk H2-Oa H2-Oa Sat1 Stk38 Stk38 -20 Cd74 Ctsz 10 Il10 Fut8 Sat1 Sat1 Il22 Cd74 Cd74 Rc3h1 Ccl21a Eef2 Eef2 Blk Dock8 Dock8 10-30 Ssr2 Eif2a Eif2a Ltbp3 Txk Txk Edem1 Nfkbia Nfkbia H2-Aa Il1a Spcs1 Spcs1 10-40 Edem1 Edem1 Nfkbia Fut8 Fut8 Ltbp3 Ltbp3 Slpi Ssr2 Ssr2 10-50 Il22 Il22 Mpc2 Slpi Slpi Rexo2 Rexo2 p-value (Mast Negative Binomial) Spcs1 Ccl21a Ccl21a Il1a Il1a -100 Rexo2 10 Mzb1 Mzb1 Mzb1 10-150 Ctsz Ctsz 0 2-2-4 4 Mpc2 Mpc2 GC PC GC PC 0 2-11 Fold Change Log2 Naive -1 0 1 Naive anti PD-1 anti PD-1
Membrane Nucleus anti PD-1 Plasma Cells C Cd69 D Rel Il2ra Id3 10-1 S5: Long Arm S4: Short Arm Ifngr1 Bach2 E Fas Bcl6 Ebi3 Flt3 Ets1 H2-Ab1 Cux2 Socs2 Ptprc Aicda 10-2 Il31r Ikzf1 H2-DMb2 Cd79a Pax5 Rgs10 Havcr1 Spib anti PD-1 S4 Cd44 Ifitm2 Mcm3 Tnfrsf13b Sdc1 Cxcr5 Tcf3 PC 10-3 Selpg S1pr4 Cd19 Ikzf3 S100a9 Cd72 Skil S5 Smad4 Ikzf2 Icam1 Ltb Bhlha15 Cd24a Cdkn1b Tlr5 Spp1 Cd79b Klf7 -4 Mcm3 Kmt2d Runx2 Cd48 Prdm1 GC 10 Ccl21a Folr1 Ly6c1 Irf4 H2-Ab1 Xcl1 Klf4 Cd44 Runx2 p-value (Stream) Ikzf4 Cd28 Xbp1 Ifnar2 Ccnd2 Pseudotime -5 Tnfsf11 Cd274 Ell2 10 Gnai2 Il22 Ly75 Atf4 Naive GC PC Ssr2 Cd22 Notch2 Gata3 Cd48 Il1a Lgals1 Bcl6b 10-6 Mki67 Cd9 Ikzf4 Ube2c Bcl6b Notch2 Cd81 Mki67 10-10 Mcm4 Il1r2 Pcna Mcm2 Il1r2 Cd40lg Mcm2 Slpi H2-K1 Irf5 10-14 Ltbp3 Cd80 Rara Top2a Pcna Cd40lg GC PC 0 2-11 GC PC 0 2-11 3 10-18 Naive Naive anti PD-1 anti PD-1 -2-4-6 0 2 4 6 Fold Change Log2
Figure S3. Single cell qtSEQ comparison of GC to PC program after PD-1 blockade, Related to Figure 4 (A) TSCAN pseudotime reconstruction for Naïve B cells and IgG1 GC and PC from anti-PD-1 treated mice with representative marker TSCAN gene expression. (B) Volcano plot of IgG1 GC and PC from anti-PD-1 treated mice with averaged single cell heatmap from cytosol and secreted gene products and the single cell heatmap. (C) Single cell heatmap of gene products from cell membrane and (D) nucleus. (E) Stream map with states identified for diverging branches S4 and S5 for the IgG1 anti PD-1 PC subsets. Volcano plot displayed for differential gene expression (p<0.05) across S4 and S5.
58 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 S4 A B Day 16 Ham IgG anti-PD-1 Treat anti-PD-1 Newly-formed 6 Ag-specific GC 7±2 0.3±0.1 10 day 10 16 * 5 Treat anti-PD-1 10 Persistent
Ag-specific GC Total Immunize day 42 48 104 NP-KLH PD-1 TFH cells Treat anti-PD-1 3
PD-1 10
Memory Recall Node draining Lymph Ham anti Ag-specific GC IgG PD-1 day 42 Recall Boost NP-KLH day 6 Cxcr5
Day 16 C Ham IgG anti-PD-1 F I 6±1 6±1 105 106
104
103 105
2
10 Spleen Total Ham IgG Ham IgG
anti-PD-1 NP-specific cells
NP anti-PD-1 IgG1 Memory Cells 1 ND 4 Total dLN NP-specific Total 10 10 Naive 16 48 6 16 6 Lambda1 Primary Recall Primary Recall NP-specific Day 16 D Ham IgG anti-PD-1 G J 105 106 PC PC ** 17±4 22±4 104 105
103 104
2 3
10 Spleen Total 10 Ham IgG Ham IgG anti-PD-1 anti-PD-1 IgG1 Plasma Cells IgG1 Plasma Cells 1 ND 2
Total dLN NP-specific Total 10 10 CD138 Naive 16 48 6 16 6 Primary Recall Primary Recall B220 NP-specific Day 16 Ham IgG anti-PD-1 E H 5 K 6 GC GC 10 10 40±4 41±5 ** 104 105
103 104
102 103 Ham IgG Spleen Total Ham IgG IgG1 GC B cells anti-PD-1 IgG1 GC B cells anti-PD-1 1 ND 2 Mem Mem dLN NP-specific Total 10 10 GL7 44±4 42±3 Naive 16 48 6 16 6 Primary Recall Primary Recall CD38
Day 6 Day 6 J43 alone NP-KLH & Ham IgG NP-KLH & J43 NP-KLH & MPL L 106 ** 0.02± 1±0.1 1±0.1 2±0.4 ** 0.01 105 **
104 Total
103 NP NP-specific B cells 102 J43 Ham Lambda1 alone IgG J43 MPL with NP-KLH
Figure S4. Vaccine-induced IgG1 GC B cells are not susceptible to PD-1 blockade (A) Schematic of experimental procedure. Mice were immunized with NP-KLH in MPL adjuvant base of tail and treated with anti-PD-1 or isotype control antibody at the indicated timepoints. (B) Flow cytometry of activated CD4+ T cells (TCRb+ CD8- CD4+ CD44hi) in the draining lymph nodes following immunization and subsequent control IgG or anti-PD-1 treatment at day 10 with analysis at day 16. (C) Antigen-specific B cells (NP+ Lambda1+) from draining lymph nodes of (B), (D) antigen-spe- cific PC, and (E) GC and memory B cells. (F) Quantification for the primary and recall response of total antigen-specific IgG1 memory B cell, (G) IgG1 PC, and (H) IgG1 GC B cells in the draining lymph nodes. (I) Total antigen-specific cells in the spleen. (J) All splenic IgG1 PC and (K) GC after control IgG or anti-PD-1 treatment. (L) Mice were immunized IP with anti-PD-1 (J43) alone, NP-KLH with isotype control antibody, NP-KLH with anti-PD1 (J43), or the NP-KLH with vaccine adjuvant TLR4 agonist (MPL). Antigen-specific B cells in the spleen with quantification (n=3) mean±SEM.
59 bioRxiv preprint doi: https://doi.org/10.1101/2021.06.10.447979; this version posted June 10, 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 S5
A Switched B cells J43-specific 56±7 16±7 B 0.06± Untreated Elispot 0.02 105 *
104 * 81±8 25±9
70±7 Day 6 103 74±4 Plasma Cells 8±2 J43
Total IgG1+ secreting Total Untreated J43 102 N.D. Total Ham J43 IgG1 IgG J43
CD138 24±4 GL7 9±2 Specificity IgG1 B220 CD38
NP-KLH with MPL NP-J43 (8:1) NP-J43 (4:1) J43 C 2.4 0.12 0.02 0.1 NP
Lambda1
Figure S5. Pre-existing J43-specific IgG1 binding is enhanced with PD-1 blockade (A) Steady state mice untreated or treated with anti-PD1 antibody (J43) and analyzed 6 days later for J43-binding. (B) Elispot quantification of specific binding to total IgG1, Ham IgG and J43 by secreting splenic B cells. (C) Mice treated with NP-KLH with MPL, NP-J43 (8:1 and 4:1 conjugation ratio) and J43 alone and assayed for antigen-specific B cells in the spleen.
60