In the most general terms, the work of the Rothstein laboratory is directed toward elucidation of the function and role of B lymphocytes in health and disease.  Over the years our efforts have focused on a number of specific, interrelated areas.  These areas currently include studies of BCR signaling, B-1 cell function, B cell/T cell interactions, and Apoptosis, and include a Program Project on Molecular Determinants of B-1 and B-2 Cell Responses.

B cell receptor (BCR) signaling

The activity and fate of B cells is determined by antigen binding to the B cell receptor.  We have studied early events in BCR signaling over many years time, including examination of transcription factor activation and pharmacologic mimicry of signal propagation. During this time a general consensus has evolved regarding the absolute requirement for certain signaling mediators grouped together as the signalosome (eg, Btk, PI-3K, BLNK, PLC2, PKC) to mediate BCR-triggered downstream events. Recently we found that prior engagement of certain non-BCR receptors results in the generation of a new signalosome-independent alternate pathway for subsequent BCR signaling as a result of receptor crosstalk. Thus, when B cells are treated first with CD40L, or IL-4, or LPS, or CpG, then washed, and then stimulated with anti-Ig, BCR triggered ERK phosphorylation is resistant to signalosome inhibitors such as LY294002 (PI-3K), U73122 (PLC) and Go6976 (PKC), in contrast to the sensitivity to these agents of anti-Ig-stimulated naïve B cells (1-6).  We have studied the IL-4-induced alternate pathway for BCR signaling most extensively, and there we found that the signalosome-dependent (classical) BCR signaling pathway, which is the only pathway present in naïve B cells, operates in parallel with the signalosome-independent (alternate) pathway in IL-4-treated B cells (5).  That is, after IL-4 treatment, two pathways co-exist, so that blockade of both is required to interrupt BCR signaling for ERK phosphorylation. Thus, in naïve B cells BCR-triggered pERK is inhibited by LY294002, but in IL-4-treated B cells, LY294002 does not block BCR-triggered pERK, nor does rottlerin (an inhibitor of the alternate pathway), but the combination of LY294002 plus rotterlin does completely block BCR-triggered pERK. Along the same lines, we found that anti-Ig fails to induce ERK phosphorylation in naïve B cells deficient in PKC (a signalosome mediator), but does so in the same PKC-deficient B cells when those B cells have been previously treated with IL-4, again demonstrating the ability of the IL-4-induced alternate pathway to bypass the need for signalosome elements. The IL-4-induced alternate pathway further differs from the classical pathway in requiring Lyn (which the classical pathway does not) and in failing to encompass NF-B activation (which the classical pathway does) (5).  Of note, however, there are at least two alternate pathways in B cells, inasmuch as the LPS-induced alternate pathway is not inhibited by rottlerin but does entail NF-B activation.  Although the precise mediators comprising the alternate pathway have not been identified beyond Lyn, these results suggest that: 1) what is known about BCR signaling may only be correct for naïve B cells and not for B cells exposed to T cell- or bacterially-derived products in the midst of an ongoing immune response; and, 2) alternate pathway signaling may represent a B cell adaptation designed to strengthen BCR-triggered signaling when products produced by activated T cells or bacteria are sensed in the environment.  Our future work will focus on identifying the precise components of the alternate pathways we have identified and determining the role of alternate pathway signaling in B cell responses.  

B-1 cell function

B-1 cells represent a phenotypically distinct (CD5-expressing) subset of B cells, in both mouse and human, that manifests a number of unusual functional features that are distinct from the characteristics of the more numerous conventional (B-2) cell population.  Among these features is the spontaneous and constitutive secretion of IgM that is known as “natural” immunoglobulin and serves as an initial serological line of defense against pathogens.  Many other distinctive features have been described, including repertoire skewing of VH11/VH12, enhanced allogeneic T cell stimulation, and constitutive activation of STAT3 and ERK, which are only activated in B-2 cells after stimulation. We have spent much time identifying differences between B-1 and B-2 cells in a number of ways, examining transcriptomic differences by DNA microarray and proteomic differences using mass spectrometry (7-9).  In the course of this work we found that murine splenic and peritoneal B-1 cells differ from each other, as well as from B-2 cells, by numerous criteria (7, 10, 11).

Some of our recent work on B-1 cells has focused on immunoglobulin secretion. Here we found that immunoglobulin secretion by unstimulated B-1 cells occurs in the context of undetectable (ie, very, very low) levels of Bcl-6, Blimp-1, Pax-5, and XBP-1 (12).  This is very much unlike B-2 cells, which in the resting state express high levels of Bcl-6 and Pax-5, but after LPS stimulation to induce differentiation to immunoglobulin secretion lose Bcl-6 and Pax-5 and express high levels of Blimp-1 and XBP-1. Our demonstration that B-1 cells secrete immunoglobulin without the levels of Blimp-1 required for B-2 cell Ig secretion is important to understanding both B-1 cell physiology and the mechanism of immunoglobulin secretion; in contrast, others have claimed that B-1 cells either fail to secrete immunoglobulin at all, or fail to do so in the absence of Blimp-1 (13, 14). However, Nutt and colleagues recently reported on a Blimp-1-independent pathway for B-2 cell immunoglobulin secretion (15), thereby confirming that Blimp-1 is dispensable for immunoglobulin secretion in certain situations and lending support for our earlier finding regarding B-1 cells.

In our continuing evaluation of B-1 cells we recently found that a subset of B-1 cells expresses the B7 family member PD-L2 (16), display of which up until the present time has been said to be limited to activated macrophages and dendritic cells.  PD-L2 expression divides B-1 cells into PD-L2+ and PD-L2- subsets.  Remarkably, PD-L2+ B-1 cells manifest, to a much greater extent than PD-L2- B-1 cells, several well-accepted features attributed to B-1 cells, including enhanced allogeneic stimulation and repertoire skewing.  This raises the possibility that PD-L2+ and PD-L2- B-1 cells may differ in origin or function.  Our future studies of B-1 cells will focus on the regulation of immunoglobulin secretion, the origin of constitutively phosphorylated STAT3 and ERK, and the regulation and role of PD-L2. 

B cell/T cell interactions

It is not always remembered that B cells function as antigen presenting cells.  Anti-inflammatory regulatory Th cells (Treg cells) and IL-17-secreting pro-inflammatory Th cells (Th17 cells) have recently been shown to play key roles in aberrant (auto-) immune function.  Both Treg and Th17 cells can be induced from naïve T cells that are TCR-stimulated and given the right cytokines, that include TGF and TGF/IL-6/IL-23, respectively.  Because differences have previously been shown in the efficiency with which B-1 and B-2 cells allogeneically stimulate T cells, we evaluated the ability of antigen-presenting B cells to induce Treg conversion and/or Th17 cell differentiation.  We found that B-2 cells strongly induced Treg conversion, as effectively as dendritic cells, despite relatively poor stimulation of CD4+ T cell activation and proliferation, whereas the inverse was true for B-1 cells that poorly induced Treg cells but strongly stimulated T cell proliferation.  Conversely, we found that B-1 cells strongly induced Th17 cell differentiation, whereas the opposite was true for B-2 cells that poorly induced Th17 cells (17).  Thus, B cell lineage matters in determining the nature of Th cell differentiation that follows antigen presentation, and it may be that B cells play a key role in specifying the anti- and pro-inflammatory nature of Th cell responses.  This work suggests a larger role for B cells in regulating T cell activity than has heretofore been accepted, and raises the possibility of a new avenue by which B-1 cells may influence autoimmunity.  Our future work will entail studies of the differences between B-1 and B-2 cells that might account for lineage-specific differences in inducible Th cell differentiation.

Apoptosis

B cells, like other cell types, are susceptible to apoptosis by engagement of the Fas (CD95) death receptor.  We demonstrated some time ago that engagement of various other receptors regulates the susceptibility of B cells to Fas-mediated apoptosis.  In that work, we found that treatment of B cells with CD40L upregulated Fas expression and produced marked susceptibility to FasL killing; in contrast, B cell treatment with anti-Ig produced resistance to FasL killing, when administered coincidentally with CD40L or even 24 hours after CD40L (at which time CD40-mediated upregulation of Fas has already occurred), implying that BCR triggering is capable of reversing already-established Fas-sensitivity (18, 19).  This BCR-induced Fas-resistance may come into play to protect antigen-specific B cells from activated T cells during B:T interactions that accompany immune responses.  In subsequent work we found that IL-4 also induces Fas-resistance (20) and, incidentally, we found that IL-4 transgenic mice express serological autoreactivity (20, 21), suggesting that physiological Fas-resistance, like Fas mutation in lpr mice, results in loss of tolerance and autoantibody production.

To identify molecules that might be responsible for inducible Fas-resistance, we undertook differential display, comparing B cells stimulated with CD40L/anti-Ig with those stimulated by CD40L alone.  This led to the cloning and characterization (and patenting) of a novel anti-apoptotic gene termed Fas Apoptosis Inhibitory Molecule (FAIM) (22).  FAIM is highly conserved from fly to human but contains no known effector motifs.  Subsequent work showed that two alternatively spliced forms of FAIM exist, a “short” form and a “long” form, the latter being expressed only in the brain (and being longer by an additional 22 N-terminal amino acids) (23).  FAIM-S acts to enhance NF-B activation produced by NGF stimulation of PC12 cells, as shown by Comella and colleagues, who recently showed that FAIM-L is normally expressed in neurons (as opposed to other brain cells), and produces neuronal resistance to apoptosis (24, 25). We are currently in the process of analyzing a FAIM KO mouse we constructed (lacking exons 1-3, encompassing both start sites).  Our future work will entail producing FAIM-L and FAIM-S transgenic mice, including FAIM-L transgenic mice in which overexpression is limited to the brain.  This will provide a model with which to examine the role of FAIM in neurodegenerative diseases.

Program Project: Molecular Determinants of B-1 and B-2 Cell Responses

B-1 cells differ dramatically from B-2 cells in terms of mitogenic responsiveness, in that B-1 cells enter S phase in response to phorbol ester (eg, PMA) alone (which is an incomplete mitogen for B-2 cells and only stimulates B-2 cells in conjunction with a calcium ionophore), whereas B-1 cells fail to proliferate in response to anti-Ig (a complete mitogen for B-2 cells) (26, 27).  Following our description of these anomalies, we collaborated with Dr. Thomas C. Chiles (Department of Biology, Boston College) to determine the molecular origin of PMA responsiveness.  Together we found that PMA produces early induction of cyclin D2 in B-1, but not B-2 cells, and complete, Rb-phosphorylating cyclin D3-cdk4 complexes in B-1 cells, whereas similar complexes produced in B-2 cells lack activity (28, 29).  To facilitate our collaboration, we initiated a program project that focuses on this and other issues in B-1 cell biology and cell cycle progression, including the role of the peritoneal environment in specifying B-1 cell characteristics and the physiology of cdk4 activation and function. In examining the role of cyclin D3 in B-1 cells, we evaluated cell cycle progression in D3-deficient B cells. Although cyclin D3 is necessary for PMA-induced cell cycle progression in normal B-1 cells, we found that PMA still stimulated proliferation in cyclin D3-deficient B-1 cells.  We found that this resulted from a compensatory increase and prolongation of cyclin D2 activation (30).  These results point out that lymphocytes in which particular molecules have been knocked out may contain additional molecular alterations that, without further investigation, could well complicate the interpretation of data so generated.  Our future studies will focus on cdk4 and the peritoneal environment. 
 
As part of these signaling studies we re-examined Lck, which was reported to be elevated in B-1 cells and to be responsible for the failure of BCR signaling (31).  In extensive study of Lck, we found that levels of Lck were less than, rather than greater than, Lck levels in B-2 cells (32).  Thus, there is no evidence that Lck is responsible for aberrant BCR signaling in B-1 cells.