In this study, we questioned how ATG therapy affected naïve and memory T cells including Tregs with naïve or memory phenotypes. Until now, it was unclear to what extent ATG treatment affected naive and memory T cells pools. Resolving this issue is of great significance in managing ATG therapy in autoimmune diseases as well as in allogeneic transplantation. In line with the previous reports , we found that ATG therapy markedly depleted CD4+ and CD8+ T cells from the peripheral blood, and largely spared B cells and granulocytes. By day 3 post-treatment, T cell numbers reached their nadir. By day 22, CD4+ T cells recovered to within normal ranges, but CD8+ T cells remained lower than baseline. It has been suggested that ATG therapy may have differential effects in depleting T cells in peripheral blood and lymphoid organs dependent on dosing . In this study, we tested our treatment protocol with optimal efficacy in preventing or reversing type 1 diabetes (500 μg/mouse × 2 doses 3 days apart) as described in our other reports [11, 12]. To determine the efficiency of our ATG treatment protocol in depleting T cells from the lymphoid organs, we examined splenic CD4+ and CD8+ T cells in both groups (ATG versus isotype IgG) at day 3 post-ATG therapy. We found that the depletion of both CD4+ and CD8+ T cells was less efficient in spleen than from peripheral blood. The similar results were obtained in lymph nodes such as inguinal or pancreatic lymph nodes (data not shown). Of interest, it appears that ATG therapy preferentially depletes CD62L+ naive T cells from the blood because the proportion of CD62L+CD4+ naive T cells was markedly reduced while CD44+CD62L- CD4+ memory T cells as a fraction of total CD4+ T cells were increased. We observed a similar significant trend in the change of naive and memory T cells in spleen as well. Whether this differential depleting effect of ATG exhibits in local lymph nodes, especially in pancreatic lymph nodes is of interest to be further addressed. It is unlikely that the increase of memory T cells at day 3 post-ATG therapy is due to the conversion from naïve T cells [20, 21] because ATG is still depleting T cells during this short period of time post-ATG therapy, and homeostatic proliferation unlikely leads to much in the way of T cell conversion.
The relative resistance of memory T cells to ATG-induced T cell depletion would allow for survival of memory T cells which potentially could lead to the recurrence of allograft rejection or autoimmunity after reconstitution of immune system post-ATG therapy. Consistent with this, we demonstrated that the proliferation of spleen cells from mice receiving ATG and de novo KLH immunization was as high as that of spleen cells from isotype IgG treated animals in KLH recall responses in vitro. We also found that β cell antigen-primed T cells during ATG therapy could survive ATG depletion as well. However, despite unaffected T cell proliferation in response to antigen stimulation post-ATG therapy, the T cell cytokine-producing profile in ATG treated animals indicated that ATG therapy skewed Th2 and possibly IL-10-producing Tr1, and reduced IFN-γ-producing Th1 responses. We had previously shown long-term reversal of diabetes in NOD mice using ATG or ATG in combination with G-CSF [11, 12] and our current findings in this report provide a mechanistic basis for this in the skewing toward Th2 and/or IL-10-producing Tr1 responses under the regimen of ATG. In this study, although we focused our studies on CD4+ T cells, it is also important to study phenotypic and functional alterations of CD8+ T cells by the ATG therapy, given the pathogenic role of CD8+ T cells in type 1 diabetes, which will be addressed in the following future studies. The memory T cell phenotypic characteristics, as well as the functional alterations post-ATG therapy, may allow the modulated antigen-primed T cells to efficiently exert their regulatory functions in the periphery through affecting their migrating and homing capabilities, thereby preventing the recurrence of autoimmunity in autoimmune diseases and allogeneic rejection in allogeneic transplantation. These findings also implicate that ATG therapy plus antigen vaccination could lead to synergistic effect on induction of antigen-specific immune tolerance. Such information would be of great significance for developing antigen-based immunotherapeutic strategy for autoimmune diseases such as type 1 diabetes.
Prior to this effort, several mechanisms underlying ATG immune modulation have been proposed. A common belief is that ATG therapy works by T cell depletion through complement-mediated cell lysis and activation-induced cell death. However, another view regarding ATG therapy is that this agent exerts immunosuppressive function beyond that of simple T cell depletion [7, 9]. ATG therapy may modulate immune response in vivo through inhibiting chemokine-driven T cell chemotaxis . It may also influence the interaction between T cells and endothelial cells through modulating expression of adhesion molecules . Our recent study showed that ATG therapy eliminated certain subset of dendritric cells and induced tolerogenic dendritic cells . In addition, ATG therapy may facilitate tolerance induction through ATG-mediated apoptosis of T cells; because T cell apoptosis induced by anti-CD3 therapy was recently demonstrated to be associated with CD3 antibody therapy-induced immune tolerance . The skewing of antigen-specific Th2 and IL-10-producing regulatory T cells (i.e., Tr1) by ATG therapy demonstrated in the current study suggests that the non-depleted antigen-responding T cells, instead of causing immune attack, may lead to antigen-specific restoration of immune tolerance, which implies that ATG works as immune modulator rather than immune suppressant.
As suggested previously, Foxp3+ Tregs may play a major role in preventing autoimmune diabetes during ATG therapy [11, 12]. However, it is incompletely understood whether ATG therapy depletes Tregs differently than conventional T cells and how ATG affects the distribution of Tregs in different lymphoid tissues. It is also unclear whether ATG therapy affects naïve and memory Tregs differently. In the present study, we demonstrated that ATG therapy was less efficient in depleting CD4+Foxp3+Tregs and as a result, the proportion of CD4+Foxp3+ Tregs in CD4+ T cells was significantly increased in ATG treated animals compared to controls. This increase is even more dramatic in lymph nodes with greater than a doubling in the frequency of Tregs within total CD4+ T cells in ATG treated as compared to isotype IgG treated animals. In some animals, the percentage of Foxp3+ Tregs reaches 30% of total lymph node CD4+ T cells. Unlike equivalent absolute numbers of splenic Tregs in both groups, the absolute number of Tregs in lymph nodes was significantly higher in ATG than in Isotype IgG treated group at 3 days post treatment, suggesting that more Tregs were recruiting to the lymph nodes besides resistance to ATG depletion. The increase of Tregs in lymph nodes may be of great immunological significance for ATG to control local antigen-specific immune responses in the settings of autoimmunity such as type 1 diabetes, as well as in allogeneic transplantation. This increase of Tregs 3 days post-ATG therapy is unlikely due to the preferential proliferation of Tregs in the ATG-therapy induced lymphopenic animals  because the proliferation is limited in this short period of time especially still under active T cell depletion. Of interest, by day 22 post-ATG treatment when CD4+ T cells return to the normal levels, Tregs remained proportionately higher in the ATG group than in control group, which may be attributable to a faster proliferation of Tregs than conventional T cells  because Tregs possess superior capability to utilize IL-2 to conventional T cells . This may also explain why a short-term ATG therapy offers a long-term protection in type 1 diabetes [11, 12] and in allogeneic transplantation [2, 26]. Intriguingly, Tregs with memory T cell phenotype were preferentially preserved in ATG therapy, which suggests that the preserved memory Tregs specific to certain antigens would be more potent in suppressing effector T cells reactive to the same antigens. As suggested recently , the memory Tregs may home to areas with active immunological reaction to quickly exert their regulatory function preferentially to naïve Tregs. This also explains the findings in our recent report that the post-therapy Tregs gain heightened immunosuppressive capacity . There is evidence that the progression of autoimmunity in NOD mice leads to memory-like CD8+ Tregs which can be expanded in vivo by stimulation of nanoparticles coated with MHC-carried autoantigenic peptides. Of note, injection of these nanoparticles not only prevented T1D but also reversed overt diabetes in NOD mice . Thus, the quantitative and qualitative changes of Tregs post ATG therapy may play an important role in suppressing antigen-specific effector T cells. Although Tregs are generally thought to suppress T cell responses in an antigen non-specific manner, emerging evidence shows that antigen-specific Tregs are more potent in suppressing antigen-specific T cell responses [29–31]. Lu, et al. reported recently that ATG therapy indeed induced self-antigen-specific Tregs in vivo that could provide long-term T1D protection in NOD mice . Whether the increased memory Tregs post ATG therapy plus antigen challenge in our experimental settings contain more antigen-specific Tregs needs to be further explored.