The autoimmune cGVH syndrome is due to cognate interactions of host B cells with allo-reactive donor T cells [2
]. Although such a model does not postulate an essential role for the recipient’s endogenous T cells, our previous work has shown that host CD4 T cells were essential [10
]. B cells from CD4KO mice failed to generate a cGVH response, even in the presence of syngeneic CD4 T cells. However, if syngeneic T cells are present during the ontogeny of B cells, they become capable of responding to the cGVH reactions [11
]. In this study we further investigated how endogenous CD4 T cells "nurture" developing B cells and confer upon them the ability to react to cGVH stimulus. Our results show that this "nurturing" process does not require antigen-specific cognate interactions between CD4 T cells and B cells, but does require IL-4. The CD4 T cell nurturing may also be accomplished by large doses of recombinant IL-4 and or agonistic anti-CD40 mAb.
Our previous studies showed that cGVH could be induced only when B cells matured in the presence of CD4 T cells over 21 days [11
]. However, the studies did not define the precise timing of the CD4 help. Our present studies further elucidate this issue. We found that CD4 T cells are needed very early during B cell maturation. When syngenic CD4 T cells were transferred to a heterogeneous pool of immature or mature B cells seven days after sub-lethal irradiation of CD4KO mice, the B cell response to cGVH was greatly impaired. Since we induce cGVH in all the recipients at the same time, it is possible to argue that the disparity in B cell response in different groups of recipients is mainly due to difference in duration of T-B interactions, and does not reflect the need for CD4 T cells in early B-cell development. To address this, we transferred Thy-1 congenic donor B6 T cells on day 1 and then, depleted the CD4 T cells at different time points. This approach provided various windows during which CD4 T cells would be present during peripheral B cell maturation. Our data demonstrated clearly that eradicating CD4 T cells at seven days severely attenuates the ability of B cells to react to allogeneic stimuli and become autoreactive, while removing them at 14 days had a much smaller effect. Therefore, the presence of CD4 T cells through much of the maturation of peripheral B cells is imperative for B cells to become autoreactive under cGVH stimulus.
Activation of B cells usually requires cognate interactions with CD4 T cells through multiple molecular interactions, the most important being the specific interactions of MHC II with the TCR, with the additional interaction of the CD4 co-receptor [25
]. Studies have also shown that CD4-mediated help available for B cells responding to allo-antigens involves both a non-cognate interaction that activates B cells and a cognate interaction that is required for differentiation and IgG alloantibody production [26
]. To explore these possibilities in our model, we used B6 donor CD4 T cells from AND x RAG −/− mice, in which all T cells bore the same TCR and endogenous TCR rearrangements was prevented. Impressively, these tg CD4 T cells provided very robust nurturing to B cells in their response to allo-stimulation, and the mice produced high titers of autoantibodies after induction of cGVH. Thus, the nurturing of B cells during their development does not require cognate recognition of the B cells by a polyclonal repertoire of CD4 T cells. On the other hand, the fact that the AND TCR is positively selected in the presence of I-Ab
suggests that it must recognize this molecule, probably in the presence of some autologous peptides, and probably with low affinity, since it is not subsequently negatively selected. It is possible then that this level of recognition may play a role in B cell nurturing.
Our conception of the potential mechanisms for the B-cell nurturing must take into account our results with substituting soluble molecules for CD4 T cells. It is known that T-cell derived cytokines profoundly affect the pathogenesis of autoimmune diseases, such as lupus. Although IFN-γ, IL-6 and IL-10 are known to play pathogeneic role in development of lupus, our studies showed that these cytokines did not have any "nurturing" effect on autoreactive B cells and their absence also did not affect B cell response to cGVH stimulus. On the other hand, IL-4 seems to play a pivotal role in mediating B cell development and their response to allo-signal.
IL-4 is a multifunctional cytokine that is mainly secreted by CD4+ T cells. IL-4 has multiple effects on the antibody responses of B cells and most of these are mediated through the Stat6 pathway. IL-4 up-regulates the expression of several cell surface markers that are involved in T cell mediated B cell stimulation, such as MHC II, CD40, CD23, mIgM and the IL-4R. It also regulates isotype switching to IgG1 and IgE in B cells, promotes migration of circulating B cells to the spleen, inhibits spontaneous apoptosis of cultured B cells and decreases B-cell susceptibility to Fas-mediated killing during cognate interactions with CD4 T cells. Therefore, IL-4 plays a key role in providing T-cell derived help to B cells. IL-4 is also known to influence B-cell functions indirectly through its effects on T cells, mast cells, macrophages and dendritic cells [27
]. Because IL-4 is known to stimulate B cells, it is not surprising that it has a role in mediating lupus like conditions. MRL/lpr mice treated with IL-4 antagonists had lower anti-dsDNA titers and associated renal disease. When MRL/lpr mice were back-crossed to IL-4−/− mice, they developed significantly reduced lymphadenopathy and end-organ disease. Treatment with anti-IL-4 mAb before the onset of lupus abrogated IgG anti-dsDNA antibody production in lupus-prone NZB/W F1 mice [28
]. Even in an inducible lupus model, Deocharan et al.
, showed that absence of IL-4 resulted in decreased autoantibody response [30
]. All these reports highlight the importance of IL-4 in B cell activation. It is therefore not surprising that our data support the possibility that IL-4 is one of the key T-cell mediated cytokines that is required to nurture B cells and make them receptive to allo-signal. CD4KO recipients that received donor CD4 T cells from IL-4−/− mice, produced no autoantibodies upon induction of cGVH. In addition, treatment with large amounts rIL-4 could substitute for CD4 T cell help and nurture the B cells and predispose them to become autoreactive. It is possible that this effect of IL-4 is mediated by reducing expression of inhibitory receptors on B cells [31
] and allowing the immature B cells to mature to a developmental stage that is more resistance to cell death and increasing the survival of autoreactive B cells in the spleen [20
]. Interestingly, our data in fact showed that rIL-4 treatment generated more robust anti-dsDNA antibody responses than anti-chromatin responses.
It is known that pathogenic autoantibodies in lupus-prone mice generally belong to the IgG2a and IgG3 subclasses and these are generally suppressed by type 2 cytokines such as IL-4 [32
]. However, in one study it was shown that over-expression of IL-4 by B cells of (NZW x C57BL/6.Yaa
)F1 mice resulted in lower titers of IgG2a and in an increase in IgG1 sub-classes of anti-dsDNA antibodies [33
]. Also, in some inducible lupus models (mercuric-chloride induced, pristane induced) the autoantibodies produced are mostly of IgG1 isotypes and are dependent on IL-4 [30
]. However, our data showed that rIL-4 treatment did not necessarily promote IgG1 subclasses and deviate the antibody response to type 2; rather all IgG subclasses, including IgG2c, IgG2b and IgG3 were significantly elevated. It is possible that upon activation by allo-T cells, the B cell autoreactivity is so intense that it drives the antibody response predominantly to Th1-type. It is interesting to note that the autoantibody levels in rIL-4 treated recipients abated significantly by 4 weeks after induction of cGVH, unlike the positive control recipients that were treated with syngeneic CD4 T cells. One possible explanation is that the syngeneic CD4 T cells persist in the recipients (data not shown) and can subsequently continue to nurture newly emerging immature B cells. On the other hand, the rIL-4 treatment was stopped before induction of cGVH, and it would not remain in circulation for more than 3 days [14
]. Therefore, autoantibody responses in recipients with rIL-4 treatment gradually diminished.
Various studies have shown that IL-4 also affects lupus nephritis. NZM.2410 mice that over-express IL-4 in vivo
develop severe glomerulosclerosis and renal disease [35
]. Our studies also showed that IL-4 treated recipients had remarkably high urine protein concentrations. It is possible to argue that the autoantibody production and proteinuria scores in our IL-4 treated recipients have been in part augmented by effects of the IL-4 beyond the nurturing effect on B cells. It is also interesting to note that even though autoantibody titers subsided in rIL-4 treated recipients, the proteinuria level continued to remain high. Recent studies from Shu Man Fu's group using the (SWR x NZB) F1 mouse model, have demonstrated that lupus nephritis can persist even in the absence of anti-dsDNA antibodies [36
]. In another study, it was shown that STAT-4 deficient NZM2410 mice developed renal disease despite decreases in anti-dsDNA antibody titers [35
]. We do not know whether the progressive renal disease in our model is in part driven by autoantibody specificities other than anti-DNA or anti-chromatin or whether ongoing antibody-independent inflammatory processes are initiated in the damaged glomeruli.
One intriguing finding of our study was that several weeks after induction of cGVH in rIL-4 treated CD4KO recipients, many of the mice developed patches of skin de-pigmentation and vitiligo like conditions. This was not observed in controls that received only rIL-4 or recipients that received syngeneic CD4 T cells, followed by allo-T cells. Our previous studies in cGVH models have documented severe skin lesions with progression of the disease [10
]. We did observe skin lesions in our recipients with donor syngeneic T cells, but no skin discoloration. Vitiligo has been frequently reported in association with autoimmune diseases such as lupus, thyroid disease, diabetes mellitus and alopecia areata [37
]. Many pro-inflammatory cytokines such as IL-1, IL-2, IL-4 and IL-13 are considered key players mediating the disease [38
]. This is the first time we have observed vitiligo like conditions in our cGVH mice after rIL-4 treatment. It is also interesting to note that when we tried to establish cGVH responses immediately after stopping IL-4 treatment, most of our recipients perished within a week. This was not due to a direct toxic effect of IL-4, since the control recipients that received IL-4 but no allo-T cells, survived well. It is known through in vivo
studies that IL-4 treatment strongly stimulate CD8+
T cell proliferation. Also, IL-4 can act rapidly through Stat6-dependent pathway to induce NK and NKT cells to produce IFN-γ [39
]. Work from Via's group have shown that IFN-g production and CD8+
T cell activation lead to up-regulation of Fas [40
] and to an acute GVH response in the parent-into-F1 model of GVH [41
]. Although there was no obvious source of MHC class I directed allo-reactivity in our model, we speculate that our recipients might have suffered from acute GVH instead of a chronic one when a strong allo-stimulus was given immediately after rIL-4 treatment.
Signaling through CD40-CD40L interaction is considered an important aspect of cognate contact-dependent help for antibody production. CD40 is a 48kDa transmembrane glycoprotein cell surface receptor that shares sequence homology with TNF receptor family and 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. Engagement of CD40 on B cells with its ligand (CD40L) on T cells provides a B cell co-stimulatory signal that induces B cell proliferation, Ig production, class switching, germinal center formation, affinity maturation and generation of long lived plasma cells [23
]. 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 [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 [44
]. CD40 agonists have long been used to boost immune responses [46
]. Our data also demonstrate that engagement of CD40 by using an agonistic CD40 mab could effectively nurture B cells in CD4KO mice and prime them for allo-reactivity, in the absence of syngeneic CD4 T cells. Interestingly, upon induction of cGVH in recipients that received agonist CD40 mab, the disease severity was exacerbated. The mice had severe skin lesions (data not shown), elevated protein concentration in their urine and high rate of mortality. We postulate that heightened engagement of CD40 provides very robust nurturing effect to the B cells, and this is reflected in the magnified cGVH response in the recipients. This situation is perhaps analogous to the response we saw to pharmacologic doses of IL-4.
Although it is well documented that T-B cell interactions are needed to activate mature B cells in a TD response, our data further document an essential role for T-B interactions in order to nurture B cells and make them receptive to allo-stimuli. Our data elucidate that this T cell effect is required not only during the initial stages of B cell development, but also needs to be sustained over 2–3 weeks, ie., throughout the maturation of peripheral B cells. This process of nurturing is achieved in a polyclonal fashion and is not dependent on specific receptor interactions. It can also be effectively achieved through engagement of CD40-CD40L. Our data also highlight the importance of cytokines such as IL-4 in this nurturing process. We propose that a low affinity T-B interaction, analogous to positive selection of T cells as discussed above, would allow local interactions through CD40 and a local high concentration of IL-4, to signal B cells to become receptive to an eventual allo-stimulus. Thus, in the absence of T cells, massive amounts of IL-4 or a CD40 agonistic mab could provide the same signals.
The requirement for CD4 T cells during the early ontogeny of B cells in order to permit the B cells to respond to allo-stimulation reflects fundamental aspects of B-cell development. More recently, we have found that B cells from CD4KO mice are also incapable of responding to an exogenous T dependent antigen (Choudhury et al.
manuscript in preparation). Since B cell development is an ongoing, life long process, this hitherto unrecognized T-B interaction may have implications beyond lupus and other autoimmune diseases. For example, these findings may also provide insights into role of CD4 T cells in HIV. HIV infection induces a wide array of B cell dysfunctions [47
]. Although the absence of immediate CD4 T cell help certainly must contribute to some of the immunodeficiency in these cases, our results suggest that the B cells themselves may become intrinsically defective, based on their ontogeny in a CD4 T cell deficient environment. HIV is also characterized by dysregulation of APC functions due to impaired CD40L function [49
]. It is therefore tempting to speculate that agonistic CD40 mab may abrogate some of the immune dysfunctions in this disease.