Autoreactive T cells preferentially drive differentiation of non-responsive memory B cells at the expense of germinal center maintenance

B cell fate decisions within a germinal center (GC) are critical to determining the outcome of the immune response to a given antigen. Here, we characterize GC kinetics and B cell fate choices in a response to the autoantigen myelin oligodendrocyte glycoprotein (MOG), and compare them the response to a standard model foreign antigen (NP-haptenated ovalbumin, NPOVA). Both antigens generated productive primary responses, as evidenced by GC development, circulating antigen-specific antibodies, and differentiation of memory B cells. However, in the MOG response the status of the cognate T cell partner drove preferential B cell differentiation to a memory phenotype at the expense of GC maintenance, resulting in a truncated GC. Reduced plasma cell differentiation was largely independent of T cell influence. Interestingly, memory B cells formed in the MOG GC were unresponsive to secondary challenge and this could not be overcome with T cell help.

Digital Droplet PCR (ddPCR): Tfh and naïve T cells were sorted by flow and RNA was #+*! extracted from cells using a RNeasy Plus Micro Kit (QIAGEN, Hilden, Germany) and #"+! immediately converted into cDNA using a Superscript VILO cDNA Synthesis Kit (Invitrogen).

Immunization with MOG autoantigen results in an atypical, unsustained GC response:
#&"! In order to identify and track responding B and T cells throughout an immune response to two #&#! different antigens, GFP + B cells (either NP-specific B1-8 + J! -/or MOG-specific IgH MOG ) and #&$! RFP + T cells (either OVA-specific OTII or MOG-specific 2D2) were isolated from mutant mice #&%! and transferred into wild type C57BL/6, non-fluorescent recipients ( Figure 1A). Two days post

Levels of T cell activation do not explain the differential B cell response between the
%+*! different model systems:

%"+!
In an attempt to understand the underlying mechanism behind the differential outcome of the GC %""! response in the different model antigen systems, antigen-specific Tfh cells (CXCR5 + PD-1 hi %"#! RFP + ) were FACS sorted from lymph nodes of mice 10d post immunization with NPOVA, %"$! mMOGtag, or haMOGtag ( Figure 5A, B). mRNA was isolated for quantitative digital droplet PCR  We consistently observed that the absolute number of Tfh cells was greater in the NPOVA vs %#*! MOG systems ( Figure 5B, and also reflected in Figure 1C) and that haMOGtag immunization %$+! produced intermediate numbers of Tfh cells ( Figure 5B, and also reflected in Figure 4C). This %$"! ! #"! resulted in the GC B cell:Tfh cell ratio remaining the same across model antigen systems (one %$#! example presented in Figure 5E). To determine if the size of the GC response was simply linked %$$! to the size of the T cell response to a given antigen, different numbers of 2D2 T cells were %$%! transferred along with equal numbers of MOG-specific B cells into SMARTA recipient mice.

%$&!
While immunization with mMOGtag resulted in a significantly larger antigen-specific T cell %$'! response in mice that received more cells, there was no similar increase in the number of Tfh %$(! cells, nor was there an alteration in the GC response ( Figure 5F). Here, we use manipulatable antigen model systems as a novel approach to investigate how the %**! immune system controls B cell fate choice and differentiation to produce different GC outcomes &++! tailored to the specific antigen. The response to NPOVA and other NP haptenated proteins is &+"! well characterized (Shlomchik and Weisel, 2012;Weisel et al., 2016), and in many ways is &+#! considered to represent the default response to a foreign antigen. We and others have shown that &+$! the anti-NP GC consistently forms 4-5d after exposure to antigen, peaks ~2 wks post exposure,