Queen’s Gambit: B Cell to Follicle

The first time one stains a spleen with B and T cell–specific Abs, the clear-cut compartmentalization of these cells within the white pulp becomes startingly obvious as the colorimetric readout develops in real time. It is astounding that lymphocytes, which tirelessly wander the body, nonetheless know exactly where to locate within these organs. During homeostasis, B cell aggregates, commonly known as follicles, generally exclude most T cells. This theme of cellular compartmentalization is recapitulated during an immune response, for example in the segregation of light and dark zones in the germinal center. In this Pillars of Immunology article, we mark the achievements of Jason Cyster and colleagues in ascertaining the signals that keep naive B cells in their rightful place in the follicle through the chemokine CXCL13 (1). From there, we consider how B cells integrate input from multiple chemokine gradients at homeostasis and during an immune response.

How naive B cells make their move into the follicle was a very hot topic in the late 1990s, and investigation of the rules governing lymphocyte positioning continue to occupy scientists today. By the turn of the century, efforts from the groups of Reinhold Förster, Martin Lipp, Marco Baggiolini, and Jason Cyster coalesced to elucidate the role of the CXCR5-CXCL13 chemokine system in B cell positioning. Going back in time to the late 1990s, the importance of homeostatic chemokines was emerging, as nicely articulated in another Pillars of Immunology article by Gwen Randolph (2). These were heady times, with very rapid advances arising from the intersection of widespread chemokine gene identification and genetic knockouts. For B cells, numerous clues were on the table, notably, the identification by Lipp and colleagues of the receptor CXCR5 (BLR1 at that time), and its relatively selective lymphocyte expression (3). By 1998, the Baggiolini and Cyster groups (the latter in collaboration with Michael Gunn and Lewis Williams at University of California, San Francisco) characterized CXCL13 (also known as BCA-1 or BLC) as the chemokine that engaged with CXCR5 (4, 5). B cells in CXCL13^−/− mice did not reside in distinct follicles, but rather formed a ring abutting the marginal zone in the spleen (1). This phenotype was very similar to that in CXCR5^−/− mice, thus laying the foundation for the CXCR5/CXCL13 axis in B cell positioning.

There was additional gold in the Cyster group study, showing that B cell regions in CXCL13^−/− mice also lacked a follicular dendritic cell (FDC) stromal network. Prior work by the Chaplin, Flavell, and Pfeffer groups revealed that developmental deficiency in lymphotoxin-β receptor (LTβR) signaling resulted in the absence of FDC (6–8). Moreover, Fabienne Mackay and colleagues showed that LTβR signaling was obligatory for FDC maintenance in the adult organism (9). Impaired lymph node development in both CXCR5 and CXCL13 knockout mice (similar to mice deficient in LTβR signaling) was a second major clue. Cyster’s group put these observations into a model in which CXCL13 and LT ligand expression were connected. Indeed, they observed that loss of CXCL13 led to loss of surface LTα₁β₂ expression on B cells, and CXCL13 chemokine could induce LTα₁β₂ expression in vitro on B cells. A feedback loop was thus defined, in which a B cell is lured into the follicle by CXCL13, then gains surface LTα₁β₂ expression, which, in turn, maintains FDC in a fully differentiated state. As prior work by Cyster had shown that FDC are a major source of CXCL13 in the follicle (5, 10), the loop was closed. Indeed, several years later, the groups of Shin-Ichi Nishikawa and Daniela Finke showed that CXCL13 attracts lymphoid tissue–inducer cells to developing lymphoid tissue anlage (11, 12).

Since this seminal study (1), a finer appreciation for the localization of chemokines has emerged. Many chemokines bind tightly to glycosaminoglycans, whereas others can remain more soluble (e.g., the CCL19 and CCL21 pair) (13). It is possible that these binding interactions are regulated either by proteolytic cleavage of the chemokine itself or modification of matrix elements to create microniches (14, 15). Nature has clearly expended considerable effort to ensure precise and constitutive expression of homeostatic chemokines, and those that regulate B cell positioning within the follicle are no exception. Teleologically, this could be a strategy to minimize random, noncognate encounters or “bystander” events that result in unwanted T or B cell activation.

This brings us to the next chapter: how do multiple chemokine networks influence B cell positioning in the follicles at homeostasis or during an immune response? Returning to the study of Ansel et al. (1), it was baffling that disruption of the CXCL13-lymphotoxin feedback loop did not result in more obviously aberrant B cell positioning within the spleen; B cells still managed to congregate at edges of the white pulp. This riddle went unexplained for a decade until additional chemokines gradients were defined, notably the EBI-2 receptor and its oxysterol ligands (16–19). Analysis of three chemokine receptor systems (CXCR5, CCR7, and EBI-2) led to a more complete picture of mechanisms guiding B cell positioning within the follicles, as well as in germinal centers (GCs) (17). Through the following decades, Cyster and his team have burrowed deep into the lipid gradients that affect lymphocyte positioning. The groundwork was laid for sphingosine-phosphate’s role in lymphocyte egress from the lymph node (20–22) and, more recently, a new lipid mediator S-geranylgeranyl-glutathione (GGG) was defined by his group (23). Now, it is recognized that lymphocytes integrate signaling through multiple G protein–coupled receptors to make retention or exit decisions.

What about immune responses? We now know that BCR triggering on the surface of follicle-resident B cells results in a transient upregulation of CCR7, which draws them to the T/B border. As CCR7 is subsequently downregulated, some B cells are drawn into an extrafollicular response via EBI2, an essential way to generate a “quick and dirty” humoral response to a threat. In contrast, other B cells return to the follicle via CXCR5 activation to seed a GC (17). Using CXCR4^−/−, CXCR5^−/−, and CXCL13^−/− mice, Christopher Allen, then in the Cyster laboratory, showed that chemokine gradients further segregate dark and light zone GC B cells (24). Using scRNA sequencing, Pikor and Ludewig subsequently showed that topological remodeling of light and dark zone FDC promoted such micropolarities within the GC (25).

A story that started as B cells making their opening chess move to the follicle via the LT/CXCL13 loop has evolved considerably. Nature has devised a geometric solution to the “encounter problem”: how are minute amounts of Ag “found” by lymphocytes with the appropriately rearranged T or B cell receptors, and how do these B and T cells subsequently “find” each other? A beautiful, highly orchestrated series of concentric chemokine loops compartmentalize these chess pieces into their appropriate immune compartments, thereby enforcing proper encounters. Approximately 4% of the protein encoding human genome is encompassed by G protein–coupled receptors. Thus, it is possible, if not likely, that many more systems are awaiting definition. These receptors are ultimately druggable and, therefore, the future is bright for specific control of their functions. The field of immunology is deeply indebted to Jason Cyster’s contributions to this biological paradigm, highlighted by this Pillars of Immunology article.

• Copyright © 2021 by The American Association of Immunologists, Inc.