Systemic lupus erythematosus is characterized by production of autoantibodies and glomerulonephritis. The murine chronic graft-vs-host (cGVH) model of systemic lupus erythematosus is induced by allorecognition of foreign MHC class II determinants. Previous studies have shown that cGVH could not be induced in CD4 knockout (CD4KO) mice. We have further explored the role of host CD4 T cells in this model. Our studies now show that B cells in CD4KO mice have intrinsic defects that prevent them from responding to allohelp. In addition, B cells in CD4KO mice showed phenotypic differences compared with congeneic C57BL/6 B cells, indicating some degree of in vivo activation and increased numbers of cells bearing a marginal zone B cell phenotype. The transfer of syngeneic CD4 T cells at the time of initiation of cGVH did not correct these B cell abnormalities; however, if CD4 T cells were transferred during the development and maturation of B cells, then the B cells from CD4KO mice acquire the ability to respond in cGVH. These studies clearly indicate that B cells need to coexist with CD4 T cells early in their development to develop full susceptibility to alloactivation signals.
Systemic lupus erythematosus (SLE)3 is an autoimmune disease characterized by autoantibodies against multiple specificities, including chromatin, Sm, DNA, and other nuclear Ags, as well as deposition of Ig in various organs, such as the kidneys and the skin. Anti-dsDNA Abs in particular are the hallmark of the disease, because they are produced in many patients but rarely in normal individuals. Moreover, some of the pathology of SLE, especially nephritis, may be mediated by anti-dsDNA Abs (1).
The understanding of SLE has been facilitated by the definition of several excellent murine models of this disease. Our laboratory has been particularly interested in the chronic graft-vs-host (cGVH) model. Transfer of I-A-incompatible spleen cells from non-autoimmune B6.C-H2bm12/KhEg (bm12) mice into co-isogenic C57BL/6 (B6) recipients results in cGVH reactions that closely resemble SLE in the spectrum of autoantibodies and immunopathology (2, 3, 4, 5, 6, 7). The bm12 strain has a mutant form of I-A that differs by 3 aa from the I-Ab β-chain of B6 mice (8). This difference is sufficient to induce a full alloreactivity in vitro. After in vivo transfer, the cognate recognition of recipient B cells by alloreactive donor CD4 T cells generates a cGVH reaction. Although this cGVH model does not postulate any essential role for the recipient’s endogenous T cells, several lines of evidence have suggested that recipient T cells may be involved in the autoimmune syndrome. Rolink et al. (6) reported that adult thymectomized, irradiated, bone marrow-reconstituted recipients of alloreactive T cells had a more severe cGVH syndrome than intact recipients. An in vitro system studied by Merino and coworkers (9) demonstrated that the ability of B cells from F1 hybrid mice to respond to allogeneic help from parental T cells depended on the presence of CD4+ T cells in the F1 progeny. More recent studies from our laboratory indicated that endogenous (host) CD4+ T cells play an essential role in the cGVH autoimmune syndrome, because cGVH could not be induced in CD4 knockout (CD4KO) recipients (10). Collectively, these studies suggest that the cellular interactions that induce the production of autoantibodies in the cGVH may require recipient CD4 T cells.
In the present studies, we have begun to explore the mechanism of the effect of endogenous CD4 T cells on the response of B cells to allohelp. Using adoptive cell transfer experiments, we have found that absence of CD4 T cells leads to functional alterations and intrinsic changes in B cells, rendering them resistant to allostimulation. These studies show an unsuspected critical role of CD4 T cells during B cell ontogeny, in determining their future susceptibility to alloreactive stimulus.
Materials and Methods
C57BL/6-Cd4tm1Mak (CD4KO), C57BL/6J (B6), C57BL/6J-Igha (B6.C20), and B6.C-H2bm12/KhEg (bm12) mice were originally obtained from The Jackson Laboratory. All mice were subsequently bred and maintained in our mouse colony at the University of Pennsylvania Medical Center. Recipient and donor mice were sex and age matched within each independent experiment. All of the experimental procedures performed on these animals were conducted according to the guidelines of the Institutional Animal Care and Use Committee.
Experimental cGVH disease protocol
cGVH disease was induced as previously described (2, 10). Briefly, recipient mice between 2 and 5 mo of age were injected (i.p.) with single-cell suspensions of 1 × 108 donor splenocytes, prepared by pressing donor spleens through a wire mesh screen in HBSS. Blood samples were obtained from experimental mice before the induction of cGVH disease and at 2- to 4-wk intervals thereafter. Sera were stored at −20°C for later analysis.
Detection of autoantibodies
Autoantibodies were assessed by ELISA, as previously described (3, 10, 11). Briefly, plates were coated with optimal concentration of autoantigens: 1) chromatin, purified from chicken erythrocyte nuclei, was used at 5 μg/ml; 2) dsDNA from calf thymus DNA (Sigma-Aldrich) was extracted with chloroform, precipitated by addition of ethanol, treated with S1 nuclease for 45 min at 37°C to remove single-strand regions, and used at 3 μg/ml. Ags were diluted in borate-buffered saline (BBS), added to polyvinyl microtiter plates (Dynatech Laboratories), and incubated overnight at 4°C. For the anti-dsDNA ELISA, plates were first coated with poly-l-lysine (1 μg/ml) (Sigma-Aldrich), before incubating with the autoantigen. The plates were washed with BBS and blocked with BBT (BBS, 0.4% Tween 80, 0.5% BSA, and 0.1% NaN3) for 1 h at room temperature. Serum samples, diluted 1/250 in BBT, were added in duplicate and incubated overnight at 4°C. Biotinylated goat anti-mouse IgG (pFc′ specific; Jackson ImmunoResearch Laboratories) was added as secondary Ab. For reference, standard serum from a diseased MRL/lpr mouse with high-titer autoantibodies was tested at serial 2-fold dilutions from 1/250 to 1/128,000. The plates were washed and incubated for 1 h at room temperature with avidin-alkaline phosphatase (Zymed Laboratories). The plates were washed again, and para-nitrophenyl phosphate substrate (Sigma-Aldrich), 1 mg/ml in 0.01 M diethanolamine, pH 9.8, was added. The plates were read at various time points with an automated ELISA reader (Molecular Devices).
The allotypes of IgG2a anti-dsDNA and anti-chromatin Abs were tested by assays similar to those for anti-dsDNA and anti-chromatin as described above, except that the assays were developed with rabbit anti-mouse preabsorbed allotype reagents (anti-IgG2aa or anti-IgG2ab; Accurate Chemical and Scientific) and detected with alkaline phosphatase-conjugated anti-rabbit IgG Ab (Jackson ImmunoResearch Laboratories). For “a” allotype standard, serum from an old MRL/lpr was used, and for “b” allotype standard, serum from an old B6/lpr (Ighb) was used. Both the reference sera were used at serial 2-fold dilutions starting from 1/250 to 1/128,000.
Total IgM was measured by ELISA, as described above. Instead of autoantigens, plates were coated with goat anti-mouse IgM (Jackson ImmunoResearch Laboratories) at 4 μg/ml. Biotinylated Bet-2 F(ab′)2 was used as the second Ab, and mouse IgM (clone CBPC 112) was used as standard in these assays.
The following conjugated Abs were purchased from BD Pharmingen: allophycocyanin anti-CD19 (1D3), FITC anti-CD21 (7G6), FITC anti-B7.1 (16-10A1), FITC anti-class II (AF6-120.1), PE anti-B7-2 (GL1), PE anti-CD23 (B3B4), PE anti-CD24 (M1/69), PE anti-Fas (Jo2), PE anti-CD44 (IM7), biotin anti-CD9 (KMC8), and streptavidin-CyChrome. Anti-FcγR (2.4G2), used for blocking, was grown in our laboratory. Cell surface staining was routinely performed with age- and sex-matched controls, as previously described (11, 12). A total of 1.5 × 106 cells were blocked with 50 μl of 2.4G2 culture supernatant. The cells were then incubated with directly labeled Abs for 30 min and washed. An additional 20-min incubation with streptavidin-CyChrome was performed to detect biotinylated Abs. Cells were fixed in PBS containing 1% paraformaldehyde and analyzed on a BD Biosciences FACScan. Relative fluorescence intensity was plotted on a logarithmic scale using FlowJo software.
Cell separation using magnetic beads
The magnetic beads, anti-CD4, anti-CD43, anti-B220, and CD4+ T cell isolation kits were purchased from Miltenyi Biotec, and the AutoMACS was used for cell separation. In most cases, the depletion cycle (negative selection) was used in the AutoMACS to prevent any cell activation due to separation. Briefly, splenic cell suspensions were incubated with magnetic beads at 6–12°C for 15–20 min at a concentration of 10 μl of beads/107 cells in 90 μl of MACS buffer (PBS plus 0.5% BSA plus 2 mM EDTA). The cells were washed after labeling and resuspended in the MACS buffer before proceeding for magnetic separation. The purity of cell separation was checked by flow cytometry.
Adoptive cell transfers
CD4KO recipients were irradiated at 3 Gy (using 137Cs) before performing adoptive cell transfer experiments. In most cases, 30 × 106 cells were transferred i.v. For the irradiation-autoreconstitution experiments, CD4KO recipients were given a sublethal dose of 5 Gy, according to Allman et al. (13, 14).
Statistical analysis was performed using Student’s t test. A value of p < 0.05 was considered to be significant. For analyzing samples in multiple experiments, as in analysis of cell surface markers at 2 mo of age, ANOVA was used.
cGVH cannot be induced in CD4KO recipients
As we have previously published (10), CD4KO mice do not develop cGVH after transfer of 1 × 108 age/sex-matched bm12 spleen cells. Fig. 1 shows that no anti-dsDNA Ab could be detected at any time in CD4KO recipients after challenge with bm12 spleen cells, whereas the positive control group (bm12⇒B6) produced autoantibodies and showed signs of autoimmune disease, such as hair loss and skin lesions (data not shown). This confirmed that host endogenous CD4 T cells play a critical role in the development of cGVH.
We have also previously found that the B cells in cGVH mice show up-regulation of cell surface markers, such as MHC class II, Fas, and B7.2 (11, 15). This suggested a general activation of B cells by allostimulation. We asked whether the B cells in CD4KO mice acquired these phenotypic changes after challenge with bm12 spleen cells. As positive controls, age- and sex-matched C57BL/6 mice also received bm12 cells. After 4 wk, individual mice from each group were sacrificed, and splenocytes were analyzed by flow cytometry for developmental and activation markers on B cells. As seen in Fig. 2,A, activation markers MHC II, B7.2, and Fas were elevated on gated B cells upon induction of cGVH in our positive control group (bm12⇒B6), in accordance with our prior reports (11, 15), whereas the expression of B7.1 remained unchanged. In contrast, with the exception of Fas, there was little change in expression of activation markers on B cells from CD4KO mice following induction of cGVH. The cell surface analysis also provided insights into the developmental status of the B cell populations. Fig. 2,B shows that splenic B cells in the bm12⇒CD4KO group did not down-regulate CD24, CD23, or CD21, compared with the positive control group (bm12⇒B6). In fact, splenocytes from bm12⇒CD4KO mice had slightly higher levels of CD21. Taken together, these data suggest that, in the absence of endogenous CD4 T cells, the B cells in CD4KO mice are not activated upon receiving an allogeneic stimulus, which is consistent with the serological data in Fig. 1.
B cells in naive CD4KO mice have an activated phenotype and are enriched for B cells bearing markers of the marginal zone (MZ) compartment
In the course of these studies, it became apparent that B cells in naive CD4KO mice differed from B cells from C57BL/6 mice in their expression of certain cell surface markers. Presumably, the absence of CD4 T cells was reflected in their altered phenotype and was related to their deficient response to allohelp. We thus proceeded to study by flow cytometry splenic B cells in CD4KO mice at different ages, along with age/sex-matched C57BL/6 mice. Figs. 3 and 4 show representative analyses of a panel of markers. No differences were seen in the size of B cells between C57BL/6 and CD4KO mice, as represented by forward- vs side-scatter plots (data not shown). At 1 mo of age, B7.1 and B7.2 were marginally elevated on B cells from CD4KO mice (Fig. 3), but otherwise the levels of MHC II, CD24, CD44, and Fas remained almost identical in both groups of mice. At 2 mo of age, B cells in CD4KO mice began to show altered expression of certain markers, compared with normal C57BL/6. CD24 was down-regulated, whereas Fas and B7.2 were up-regulated. At 3 mo, the differences in level of B7.2 expression on B cells between CD4KO and age-matched B6 mice were more pronounced. The statistical analyses of these data are shown in the table. These results suggest that B cells in naive CD4KO mice had undergone some degree of in vivo activation.
Because up-regulation of B7.2 is characteristic of the phenotype of MZ B cells (16), we compared the MZ B cell population in CD4KO and normal C57BL/6 mice. Splenocytes were first gated on B cells by CD19 expression and then analyzed for CD21 and CD23. The CD21highCD23low/− population was gated as MZ B cells (17, 18). As depicted in Fig. 4,A, CD4KO mice had a greater percentage of MZ B cells compared with B6 mice, and the numbers increased with the age of mice, in parallel with augmented B7.2 expression. To confirm our findings further, we analyzed the gated MZ B cells for expression of CD9, a cell surface glycoprotein that serves as an exclusive marker for MZ B cells (19). Fig. 4 B shows that MZ B cells in CD4KO mice had significant increase in levels of CD9 compared with normal B6, and this became more prominent with increasing age. Taken together, these data suggest that, in the absence of CD4 T cells, the B cells in CD4KO mice show an aberrant phenotype, including the expansion of MZ B cells.
Adoptive transfer of normal C57BL/6 splenocytes can confer autoreactivity on CD4KO recipients following cGVH
We wanted to know whether the added presence of normal spleen cells might render CD4KO mice susceptible to cGVH. Therefore, we adoptively transferred 30 million splenic cells from normal C57BL/6 mice into lightly irradiated (3 Gy) CD4KO recipients and challenged the recipients with 1 × 108 bm12 spleen cells. One control group received only bm12 cells, whereas another group received only B6 spleen cells. Anti-chromatin (Fig. 5 A) and anti-dsDNA (B) Ab titers, measured by ELISA, were indicative of a cGVH response. Reconstitution of CD4KO mice with C57BL/6 splenocytes resulted in autoimmune reactions to an allogeneic stimulus. This permitted us to further dissect the cellular mechanisms involved in this alloreaction.
Purified syngeneic B cells, but not CD4 T cells, are able to reconstitute the autoimmune responses of CD4KO recipients to an allogeneic stimulus
We next tested which component of the transferred B6 splenocytes was responsible for permitting a cGVH response in CD4KO recipients. B cells were purified from normal C57BL/6 mice by negative selection with anti-CD43 magnetic beads. CD43 is expressed on almost all cells except on naive resting B cells (20). The purity of the B cell population was >90%, as determined by flow cytometry (not shown). Thirty million purified B cells (CD43-ve fraction) were injected i.v. into irradiated (3 Gy) CD4KO recipients. The non-B cell (CD43+ve) fraction was also injected into another group of recipients. The positive control group received unfractionated C57BL/6 spleen cells. All of the groups were challenged with 1 × 108 bm12 spleen cells on the following day. One control group received only B6 spleen cells but was not challenged with bm12 cells. Another control group did not receive B6 cell transfer, but did receive bm12 cells. Our data showed that irradiated CD4KO recipients that received adoptively transferred B6 B cells (CD43-ve fraction) had significant disease characterized by high titers of anti-dsDNA Abs (Fig. 6 A) and anti-chromatin Abs (B). In contrast, recipients that received the CD43+ve fraction (the non-B cell component) had no evidence of disease. Of the different control groups, only the positive control (transfer of unfractionated spleen cells followed by induction of cGVH) produced autoantibodies, whereas none of the negative control recipients made any autoantibodies.
Additional cell transfer experiments were performed using other separation strategies. B cells from normal C57BL/6 or from CD4KO mice were positively selected using anti-B220 magnetic beads. CD4 T cells were similarly purified from normal C57BL/6 mice by using anti-CD4 beads. Both the positive and negative fractions were injected into irradiated CD4KO recipients. cGVH was induced in all of the groups of mice by transferring 1 × 108 bm12 spleen cells. Consistent with our previous data, the adoptive transfer of B6 B cells (either B220+ve fraction or CD4-ve fraction) (Fig. 6,C) permitted the development of autoimmune responses to alloreactive T cells in CD4KO recipients. However, the transfer of B cells from CD4KO mice into other CD4KO recipients did not allow alloreactive-autoimmune reactions, clearly indicating that the B cells in CD4KO mice have certain deficiencies that prevent them from responding to a cGVH stimulus. Surprisingly, transfer of B6 CD4 T cells (CD4+ve fraction) (Fig. 6 C) failed to allow graft-vs-host (GVH)-driven autoantibody production and disease in CD4KO recipients. Based on these data, it is possible to conclude that exogenous syngeneic CD4 T cell help could not immediately correct the B cell nonresponsiveness to an allostimulus in CD4KO mice. Also, it is interesting to note that lack of host CD4 T cells did not impair the autoreactive potential of the adoptively transferred B6 B cells. This suggested that host CD4 T cells are most likely required during the development of B cells, so that the B cells are able to respond to alloreactive donor cells, but the presence of host CD4 T cells is redundant at the time of the actual autoimmune cGVH response.
Transferred donor B cells are exclusively responsible for secreting autoantibodies following induction of cGVH in CD4KO recipients
An alternative, but unlikely, possibility, was that the transferred B6 B cells somehow allowed the endogenous B cells to respond to allohelp. Therefore, we wanted to confirm that the donor B cells were the exclusive source of autoantibodies. We performed cell transfer experiments using purified B cells from the B6 Igh allotype congenic strain, B6.C20 (allotype a). CD4KO recipients that received B cells purified from B6.C20 donors, produced “a” allotype autoantibodies during cGVH, whereas recipients with transferred B6 B cells produced autoantibodies of only “b” allotype (Fig. 7). These results confirmed our hypothesis that the transferred donor B cells were exclusively responsible for secreting autoantibodies and ruled out the possibility of autoantibody production by the host (CD4KO) B cells under the influence of transferred B cells and cGVH. They also ruled out the possibility that the transferred bm12 B cells might play a role in production of autoantibodies, because these B cells also bear the “b” allotype.
B cells in CD4KO mice can become autoreactive under the influence of a polyclonal activator
It was possible that the inability of CD4KO B cells to produce autoantibodies in the cGVH was due to a failure to develop an autoreactive Ig repertoire. We therefore used another method to induce autoreactivity in this population. Studies have shown that the in vivo challenge with LPS in non-autoimmune B6 mice results in lupus-like features with production of anti-dsDNA Abs of both IgM and IgG isotypes (21, 22). We therefore injected a single dose of 100 μg of LPS (Salmonella minnesota Re) into CD4KO mice. One group of CD4KO mice received only saline and served as a negative control, whereas a group of B6 mice received a similar dose of LPS, and served as a positive control. Our data (Fig. 8) show that B cells from CD4KO mice initially responded to LPS just like B6 B cells and produced anti-dsDNA Abs. However, the Ab levels decreased more rapidly in CD4KO recipients than in B6 mice. These data clearly established that the B cells in CD4KO mice have autoreactive potential and could generate autoantibodies if provided with certain stimuli.
Timing of the effect of CD4 T cells
Our cell transfer experiments indicated that exogenous CD4 T cell help could not immediately correct the inability of B cells to respond to an allogeneic stimulus. Therefore, we hypothesized that CD4 T cell help was most likely necessary during the development and maturation of peripheral B cells in order for them to be responsive to alloreactive T cells. To test this hypothesis, we used the irradiation-autoreconstitution model established by Allman et al. (13, 14). CD4KO mice were subjected to a sublethal dose of irradiation (5 Gy). Fifty million CD4 T cells purified from B6 mice were then transferred to one group of irradiated recipients (experimental group), whereas another irradiated group of recipients received no cell transfer (control group). cGVH was induced in both the groups at days 1, 7, or 21 postirradiation, by challenging with 1 × 108 bm12 cells. Our data (Fig. 9) showed that the experimental group in which GVH was induced at day 21 postirradiation manifested an autoimmune response by producing autoantibodies, whereas experimental recipients challenged at day 1 or day 7 showed very little response. The control groups that received no CD4 T cells did not respond to cGVH stimulus at any time points. Another control group that received only CD4 T cells but were not challenged with bm12 cells also produced no autoantibodies (data not shown). In analogous replicate experiments, irradiated recipients that received B6 CD4 T cells and were challenged with bm12 cells on day 14 postirradiation showed an intermediate autoantibody response (data not shown). Therefore, our data indicate that the transferred CD4 T cells could influence the development of B cells during the process of autoreconstitution, so that now the B cells in CD4KO recipients could produce autoantibodies in response to cGVH. These results also throw light upon the timing of the effect of the CD4 T cells, suggesting that the influence occurred late in B cell maturation, but this remains to be defined further.
The autoimmune manifestations of cGVH result from cognate interactions of host B cells with alloreactive donor T cells (3, 4). Although such a model does not postulate an essential role for the recipient’s endogenous T cells, our previous work has indicated that host CD4 T cells were essential (10). In this study, we further investigated how the absence of endogenous CD4 T cells modulates the cGVH response. Our results show that B cells from CD4KO mice have certain underlying aberrations that prevent them from reacting to allostimulus. However, if the B cells are allowed to develop and mature in the company of syngeneic CD4 T cells, then, they can respond to alloreactive stimulus.
CD4KO mice have been used to address different aspects of T cell development (23) and T cell responses to viral (24) and parasitic infections (25). In some cases of viral infections, CD4 Th cells are required to prime CD8 T cells for cytolytic (CTL) responses. Recent reports using CD4KO mice suggest that CD4 T cells are needed during the development (26), secondary expansion (27), and maintenance of functional CD8 memory (28).
Less is known about the influence of CD4 T cells on B cells. CD4KO mice have impaired humoral responses and produce significantly reduced levels of Abs to T-dependent Ags. In a recent study, it was shown that CD4-deficient mice formed 30–40% fewer germinal centers in the spleen, compared with wild-type controls (29). The role of CD4 T cells in autoimmunity has been explored using the MRL/lpr mice. These mice spontaneously develop severe autoimmune disease characterized by autoantibody production and glomerulonephritis. However, when these mice were bred in combination with the CD4KO locus so that they lacked CD4 T cells, the severity of the spontaneous disease was greatly reduced (30). In contrast to MRL/lpr CD4KO mice, studies of other autoimmune disease models, i.e., collagen-induced arthritis (31) and experimental allergic encephalomyelitis (32), showed that lack of CD4 T cells only moderately affected the progression and severity of the disease.
In the present work, we confirmed that B cells from CD4KO mice on a B6 background failed to respond to a cGVH stimulus, unlike the C57BL/6 control mice, and did not produce autoantibodies (10). We further characterized the phenotypic changes associated with induction of cGVH in CD4KO mice. Notably, although decreased levels of CD21 and CD23 have been seen in cGVH (11), there was little change in CD23 and CD21 in the splenic B cell population in CD4KO mice following induction of GVH. In addition, the changes in MHC class II, B7.2, and CD24 that characterized the cGVH response in the C57BL/6 control mice, failed to occur in the CD4KO recipients. These results indicate that the polyclonal phase of the cGVH reactions, which affects all B cells, is omitted in CD4KO mice, as is the autoantigen-specific phase.
The absence of GVH in CD4KO mice, together with the failure of their B cells to undergo phenotypic changes in response to allogeneic stimulation, was surprising in view of current thinking about cGVH. In this model, the recipients’ MHC II molecules on B cells should be recognized by alloreactive donor T cells. This allogeneic effect then delivers T cell help to the B cells and drives them to produce autoantibodies (2, 3). Our data imply that some interaction between recipient’s B cells and its CD4 T cells is required before the B cells can respond to alloreactive donor T cells.
It is possible that cognate interactions between recipient CD4 T cells and B cells are required to prime B cells to respond to alloantigens. Steele et al. (33) showed that if recipient mice were thymectomized and treated with anti-CD4 mAb before grafting MHC-mismatched skin, the recipients produced very low or undetectable levels of cytotoxic alloantibody. They postulated that recipient CD4 T cells stimulated the B cells to produce alloantibody through both cognate and noncognate interactions. Based on this report, we had hypothesized that exogenous syngeneic CD4 T cells transferred along with the alloreactive bm12 cells, would permit the B cells in CD4KO mice to respond in a cGVH. Surprisingly, the addition of such “direct” syngeneic CD4 T cell help was not sufficient to stimulate B cells in CD4KO mice to respond to an allogeneic stimulus. In contrast, normal B6 B cells adoptively transferred to CD4KO mice responded robustly to a cGVH stimulus and produced autoantibodies, whereas B cells from CD4KO mice did not (Fig. 6). Our experiments with Ig allotypes confirmed that the donor B cells were solely responsible for autoantibody production after induction of cGVH. Thus, the B cells from the CD4KO mouse were themselves defective in their response to allogeneic T cells.
We therefore favor the possibility that B cells in CD4KO mice acquire this intrinsic defect, because they develop and mature in the absence of CD4 T cells. In vitro studies by Merino’s group have also indicated that B cells would respond to an allogeneic stimulus only if CD4 T cells were present during B cell development. They showed that cultured B cells from CB6F1 nu/nu mice or euthymic CB6F1 mice depleted from birth of CD4+ cells by Ab administration failed to proliferate or produce Ig after allogeneic stimulation with BALB/c CD4 T cells. This B cell defect was completely restored by neonatal syngeneic thymic engraftment that partially reconstituted the mature T cell populations (9). Similarly, we have shown that adoptive transfer of CD4 T cells to CD4KO mice during the autoreconstitution of B cells after sublethal irradiation restored the ability of B cells to respond to an allogeneic stimulus. It is important to note that the presence of CD4 T cells was required for nearly 3 wk before the maturing B cells could react to alloreactive donor T cells. This is in contrast to the findings from Merino and coworkers (9) suggesting that CD4 T cell help is critical only during the early stages of B cell development. We are pursuing additional experiments in our model to determine more precisely at what stage of B cell development the influence of CD4 T cells is critical.
Our hypothesis that lack of CD4 T cells leads to abnormalities in the B cell compartment is further reinforced by our flow cytometry data. B cells in naive CD4KO mice showed elevated expression of costimulatory molecules, mainly B7.2. Also, the MZ B cell population was significantly expanded compared with normal C57BL/6 mice. Up-regulation of B7.2 is normally seen after cognate T-B interactions (34); therefore, it was paradoxical that, in the absence of CD4 T cells, the level of B7.2 expression was enhanced.
Elevated expression of the costimulatory molecules, B7.1 and B7.2, is characteristic of certain autoimmune-prone strains of mice like the NZB and NZB/W mice, which develop lupus-like diseases (35). In a recent study, Wither et al. (16) found that NZB and NZB/W mice have an increased proportion of splenic B cells expressing B7.1 and B7.2 and displayed phenotypic characteristics typical of MZ B cells. Studies from our laboratory have shown up-regulation of both B7.1 and B7.2 on peripheral lymphocytes in autoimmune B6/gld mice (36). Taken together, all of these studies indicate that augmented levels of costimulatory molecules on B cells reflect the ongoing B cell activation characteristic of chronic systemic autoimmunity. Thus, it is surprising that CD4KO mice, which are not spontaneously autoimmune-prone and cannot respond with autoimmunity in a cGVH reaction, should still show evidence of B cell activation.
We hypothesize that this could be attributed to the presence of a subpopulation of class II MHC-restricted T cells that are CD4−CD8−TCRαβ+. This double-negative (DN) T cell population can produce IFN-γ upon parasite challenge (25), can mediate Ab class switching (25, 37), can support somatic hypermutation and affinity maturation of germinal center B cells (29), and is known to produce IL-2 upon ex vivo stimulation (38). Based on these observations, we speculate that this DN TCRαβ+ cell population could be responsible for the polyclonal B cell activation seen in CD4KO mice. It is possible that these DN T cells provide tonic stimulation to the B cells that is sufficient to up-regulate B7.2, but that in the absence of signals mediated by CD4 coreceptor and MHC II interaction, B cells are rendered nonresponsive, especially to an allostimulus. This lack of response to allostimulus by B cells from CD4KO mice is not similar to the phenomenon of B cell anergy as illustrated by Ig/HEL models, because recent work from our laboratory showed that induction of cGVH could abrogate B cell anergy in Ig/HEL system (15). An alternative mechanism may involve the lack of CD4+CD25+ T regulatory cells in CD4KO mice.
How could CD4 T cells influence B cell responses to alloantigens? We postulate that both cognate and noncognate interactions between CD4 T cells and B cells are required to permit an alloresponse. Cognate interactions would be mediated by CD40-CD40L. CD40, a member of TNFR family, is expressed by a variety of cells including B cells, macrophages, dendritic cells, and endothelial cells. The ligand for CD40, CD154 (CD40L), is transiently expressed on activated T cells, mainly CD4 subsets. The interaction of CD40 on B cells with its ligand (CD40L) on T cells provides a B cell costimulatory signal that induces B cell proliferation, Ig production, class switching, germinal center formation, and B cell memory (39). The importance of CD40-CD154 ligation in the development of autoimmune disease has been illustrated in several murine models of autoimmunity using blocking Abs and knockout mice (40, 41, 42). Recent data from SLE patients and murine lupus models have demonstrated that lupus T cells have prolonged expression of CD40L, and this probably leads to excessive B cell activation (43, 44, 45). Based on these, we hypothesize that direct interactions between CD4 T cells and B cells through CD40/CD40L could activate B cells to react with alloreactive donor cells. Whether this implies a specificity of recognition of the CD4 T cells for an Ag expressed on the B cell remains to be determined.
Noncognate interactions would also occur mainly through cytokines like IL-2, IL-4, IL-7, and IL-15, which are known to play an integral role in B cell development. Of these, IL-4 plays a crucial role in promoting Ab responses. IL-4 is known to enhance the B cell immune response by releasing B cells from CD22 and FcγRIIb-mediated inhibition, and the IL-4 signal is mediated through activation of Stat6 (46, 47). Merino and coworkers (9) have also reported that the ability of CD4 T cell-deprived B cells to respond to alloreactive help could be restored by preincubation with modest concentrations of rIL-4 over 18 h in vitro. There are reports suggesting that in vivo IL-4 treatment promotes migration of circulating B cells to the spleen and enhances survival and maturation of autoreactive B cells (48). All of these suggest that secretion of IL-4 by endogenous CD4 T cells may be crucial for proper maturation of B cells and their ability to react to an allostimulus. Currently, our laboratory is conducting experiments to explore these possibilities.
Our data emphasize that host CD4 T cells are required for pathogenesis of autoimmune disease induced by cGVH and the absence of CD4 T cells causes inherent changes in B cells such that they overexpress B7.2, and are skewed toward the MZ compartment. This intrinsic B cell defect could be corrected only if syngeneic CD4 T cell help was provided during the development and maturation of B cells, so that they could now react to allostimulus. These findings may also provide insights into role of CD4 T cells in other models of autoimmunity and even in HIV. HIV infection induces a wide array of B cell dysfunctions (49). B cells from HIV-infected patients show poor proliferative responses to CD4 T cell help, due to low expression of CD25 (50). Based on our findings it is tempting to speculate that low CD4 T cell counts, associated with high levels of viremia, may be one of the major factors that impair the B cell responses to further CD4 T cell help.
We thank Magda M. Cuevas and Shuqin Wang for technical assistance, Dr. Fangqi Chen for useful discussions and technical assistance, and Dr. Mary Putt for helping with statistical analysis.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health (R01-AR34156, R01-AR26574, U19-AI-46358), Lupus Research Institute, Alliance for Lupus Research, U.S. Department of Veteran Affairs, and Lupus Foundation of South New Jersey. A.C. was supported by a postdoctoral fellowship from Arthritis Foundation.
Abbreviations used in this paper: SLE, systemic lupus erythematosus; GVH, graft-vs-host; cGVH, chronic GVH; CD4KO, CD4 knockout; MZ, marginal zone.