Logo of immunologyLink to Publisher's site
Immunology. 2011 Sep; 134(1): 50–59.
PMCID: PMC3173694

Induction of self-antigen-specific Foxp3+ regulatory T cells in the periphery by lymphodepletion treatment with anti-mouse thymocyte globulin in mice


Lymphodepletion therapies are increasingly tested for controlling immune damage. One appealing premise for such a therapy is that it may ‘reboot’ the immune system and restore immune tolerance. However, the tolerogenic potential of lymphodepletion therapies remains controversial. The debate is exemplified by conflicting evidence from the studies of anti-thymocyte globulin (ATG), a prototype of immunodepleting drugs, in particular on whether it induces CD4+ CD25+ Foxp3+ regulatory T (Treg) cells. To understand the impact of ATG on T cells at a clonal level in vivo, we studied the effect of anti-mouse thymocyte globulin (mATG) in a reductionist model in which the T-lymphocyte repertoire consists of a single clone of pathogenic T effector (Teff) cells specific to a physiological self-antigen. The mATG treatment led to peripheral induction of antigen-specific Treg cells from an otherwise monoclonal Teff repertoire, independent of thymic involvement. The de novo induction of Treg cells occurred consistently in local draining lymph nodes, and persistence of induced Treg cells in blood correlated with long-term protection from autoimmune destruction. This study provides in vivo evidence for clonal conversion from a pathogenic self-antigen-specific Teff cell to a Treg cell in the setting of immunodepletion therapies.

Keywords: autoimmunity, CD4 T cells, immunotherapeutics, regulatory T cells, tolerance


More than a decade of intensive studies on CD4+ CD25+ Foxp3+ regulatory T (Treg) cells have demonstrated their critical role in immunological tolerance and their potential in immune therapies.1 However, adoptive Treg-cell therapy remains a clinical challenge and its efficacy has yet to be shown in clinical trials. On the other hand, enhancing endogenous Treg-cell-based immune regulation with clinically ready agents may emerge as a potential alternative to realize the tolerogenic potentials of Treg cells. A number of studies have demonstrated that the ratio of Treg cells to effector T (Teff) cells in vivo can be readily altered by immunomodulation, but it remains debatable whether de novo induction of Treg cells specific to physiological antigens occurs in vivo under therapeutic settings. Definitive evidence in this regard will have important clinical implications because de novo induction of Treg cells, or conversion of Teff cells to Treg cells, creates new specificities, and therefore new protective potentials, for the Treg-cell population.

Lymphodepletion approaches have been tested to eliminate a variety of pathogenic immune cells implicated in immune damage and allow the generation of a new repertoire of lymphocytes (i.e. ‘rebooting’ the immune system). A prototype of lymphodepleting agents is anti-thymocyte globulin (ATG).2,3In vitro, ATG treatment of human peripheral blood mononuclear cells (PBMCs) from healthy individuals increased the percentage of CD4+ CD25hi Foxp3+ cells, and CD4+ CD25+ cells from ATG-treated PBMC culture suppressed T-cell proliferation in vitro.4,5 However, the notion of Treg-cell induction by ATG was contradicted by a recent study showing that while ATG treatment of in vitro culture of human PBMCs enhanced Foxp3 expression, the effect of ATG on human T cells was associated with activation of the T cells, but not with inducing suppressive activity of the cells.6

To facilitate the mechanistic studies of ATG therapeutic effect, anti-mouse thymocyte globulin (mATG) was produced using mouse thymocytes as an immunogen in a process analogous to that used for producing the clinically used ATG. The mATG treatment of in vitro cultured splenocytes indeed caused profound depletion of T cells, but residual T cells are induced to proliferate in a later phase of culture.7 The splenocytes treated with mATG suppressed T-cell proliferation in vitro and graft-versus-host response in vivo in mice. However, the suppressive elements induced in vitro by mATG in the study did not involve any Foxp3 induction.7

In vivo, patients that received Thymoglobulin™ (Genzyme Corporation, Framingham, MA) treatment had a transient increase in the percentage of CD4+ CD25+ Foxp3+ cells in peripheral blood.8 This increase may be the result of the resistance of Foxp3+ natural Treg (nTreg) cells to immunodepletion, and is possibly contributed by induced Treg (iTreg) cells with new specificities. Increased proportions of the CD4+ Foxp3+ or CD4+ CD25+ cells were also observed in mouse models receiving mATG treatment,911 but those studies employed animal models that harboured a polyclonal repertoire of T cells and a functional subset of pre-existing Treg cells. Given the documented evidence that the depletion effect of mATG preferentially spared Treg cells,9 it was difficult to discern whether de novo induction of Treg cells occurred in those studies, and if it did, whether an antigen-specific Teff cell could be converted to an antigen-specific Treg cell.

T-cell antigen receptor (TCR) transgenic mouse models remain definitive and instrumental tools to study immune tolerance induction at T-cell clonal levels.12 To examine the tolerogenic potential of mATG depletion therapy, we resorted to the BDC2.5 TCR transgenic mouse model.13 The BDC2.5 line expresses a transgenic TCR originated from a CD4+ T helper type 1 clone specific to a natural self-antigen expressed by pancreatic beta cells. As the result of ‘leaky’ allelic exclusion by the transgenic TCR, the transgenic mouse still harbours a polyclonal T-cell repertoire, although the majority of the T cells express the BDC2.5 TCR.14 In this model, CD4+ CD25+ Fopx3+ Treg cells are critical for controlling immune destruction,15 and even a moderate reduction of Treg cells can dramatically unleash immune pathology progression.16 When the BDC2.5 line is crossed to a Rag-deficient background, the resulting model, the BDC2.5/NOD.Rago/o mouse, like other lines of Rag-deficient TCR transgenic mice such as DO11.10, is devoid of Treg cells.15,17,18 The genetic deficiency of the Rag recombinase in this mouse precludes the rearrangement of the endogenous TCR locus. The TCR transgene allows the development of a monoclonal repertoire of Teff cells specific to a self-antigen. Hence, mATG depletion treatment in the BDC2.5/NOD.Rago/o mouse model allows a definitive assessment of its potential in de novo induction of Treg cells specific to a natural antigen. In particular, this model will allow us to examine the effect of mATG on naturally arising T cells at a clonal level in its physiological niche, rather than its effect on adoptively transferred cells in a new host. In vivo experimental evidence, or lack thereof, for Treg-cell induction in lymphodepletion treatment may not only contribute to the conceptual possibility of therapeutic conversion from a Teff cell to a Treg-cell clone, but also yield critical insight for assessing the immunological impact in clinical applications of ATG-based depletion therapy.

Materials and methods


BDC2.5/NOD.Foxp3sf and BDC2.5/NOD.Rago/o mice were described previously.15,17 NOD/Lt and C57BL/6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME). All animals were maintained in a specific pathogen-free barrier facility and the studies are approved by the institutional review committee at the University of Miami.

mATG, control rabbit IgG, and treatment regimen

Rabbit mATG7 was prepared by immunizing rabbits with mouse thymocytes. The IgG from the serum of the immunized animals or normal controls was purified according to a process similar to that used for production of the clinically used Thymoglobulin™. Both mATG and control rabbit IgG were provided by Genzyme Corporation. BDC2.5/NOD.Rago/o and BDC2.5/NOD.Foxp3sf mice were treated with mATG beginning at about 10 days of age, before the activation of specific T cells and onset of inflammatory pathology in pancreatic islets. The treatment was administered by intraperitoneal injection every 3–4 days for a total of four doses. The standard dose was 20 mg/kg according to the manufacturer's suggestion, but doses at 5, 10 or 40 mg/kg were also tested.

Flow cytometry analyses of PBMCs and lymphoid organs

Single cell suspensions of immune cells from blood or lymphoid organ samples were incubated with anti-CD16/32 (clone 2.4G2) and normal mouse sera (Jackson ImmunoResearch, West Grove, PA) to block non-specific binding of antibodies to Fc receptors. The samples were then stained with antibody conjugates using a standard flow cytometry procedure. The antibody conjugates include: Alexa 405-conjugated anti-CD4; FITC-conjugated anti-CD103, anti-GITR, TCR-β, anti-Foxp3, and rat IgG2a isotype (eBioscience, San Diego, CA); phycoerythrin (PE) -conjugated anti-CD69 (eBiosciences) and anti-CTLA-4 (BD Biosciences, San Jose, CA); PE-Alexa Fluor®610-conjugated anti-CD4 (Invitrogen, Carlsbad, CA); PE-Cy5-conjugated anti-CD8α; peridinin chlorophyll protein (PerCP) -Cy5.5-conjugated anti-CD25 and anti-CD127; allophycocyanin (APC) -conjugated anti-CD62L and anti-Foxp3 (eBioscience). Intracellular staining with the anti-Foxp3 and anti-CTLA-4 antibodies was performed according to manufacturer recommendations with rat IgG2a and hamster IgG as isotype controls, respectively. Stained samples were analysed with an EPICS® XL™ (Beckman Coulter, Miami, FL) or LSRII (BD Biosciences) flow cytometer.

Diabetes monitoring and pathology examination

Urine glucose levels were tested every 2–3 days with Uristix® (Bayer Diagnostics, Elkhart, IN). Diabetes was diagnosed if positive urine glucose readings were confirmed in two consecutive measurements and verified by blood glucose reading (> 250 mg/dl) with an Ascensia Elite™ XL reader (Bayer Diagnostics). Pancreata of the experimental mice were harvested and fixed in formalin solution (Sigma, St. Louis, MO). Paraffin-embedded sections were stained with haematoxylin and eosin and examined by microscopy for inflammatory cell infiltration into the pancreatic islets.

Adoptive transfer for in vivo assessment of suppressive activity

Spleen and lymph node cells from BDC2.5/NOD and mATG-treated BDC2.5/NOD.Rago/o were blocked against Fc-binding as stated above, stained with Alexa 405-conjugated anti-CD4, APC-conjugated anti-CD25, PE-conjugated anti-CD69, FITC-conjugated anti-CD8 and CD11b and sorted using a FACS Aria cell sorter (BD Biosciences) for CD8 CD11b CD69 CD4+ cells that were either CD25+ or CD25 populations (purity > 98%). Sorted cells were resuspended in PBS and injected intraperitoneally into recipient BDC2.5/NOD.Rago/o or NOD.Rago/o mice 8–13 days of age. The recipient animals were then monitored for diabetes development by urine and blood glucose measurement as described above.


Suppression of autoimmune diabetes by mATG and its gender effect in immune tolerance induction

To assess mATG-mediated inhibition of pathogenic Teff cells at a clonal level, we used a line of TCR transgenic mice on a Rag-deficient background, the BDC2.5/NOD.Rago/o mice.15,17 Some animals were treated with mATG or non-specific rabbit IgG control at a standard dose of 20 mg/kg. Doses of 5, 10 and 40 mg/kg were also tested. The treatment was a short course, administered between 10 and 22 days of age in four doses. As shown in Fig. 1(a), BDC2.5/NOD.Rago/o mice exhibited aggressive autoimmune damage to pancreatic beta cells, such that all animals suffered from complete beta cell destruction and became diabetic between 20 and 30 days of age. Treatment with mATG substantially inhibited the development of autoimmunity, with 20 mg/kg apparently being an optimal dose. This effect is not the result of systemic introduction of polyclonal IgG, as control rabbit IgG had no effect on the course of autoimmune damage (Fig. 1a). This inhibition of overt diabetes was consistent with pathological analyses revealing a delay of inflammatory pathology in the pancreatic islets (data not shown). In the initial dosage experiments, we observed that one of 10 animals in the 20 mg/kg mATG treatment group was protected from diabetes for more than 70 days. The protected animal was a male, an observation that led us to further examine whether there was a gender-biased effect in mATG treatment. Interestingly, mATG exhibited a more protective efficacy in male animals than in their female counterparts (Fig. 1b). Of 17 female mice treated with mATG, none remained free of autoimmune diabetes beyond 55 days of age. On the other hand, about one-quarter of the mATG-treated male animals (n = 13), had achieved ‘operational’ immune tolerance, i.e. remaining diabetes-free long term (> 160 days of age) without the need for further immunomodulation. The frequency of long-term protection by mATG in male mice was significantly increased over that in female animals (P < 0·05). It is noteworthy that as opposed to the standard NOD mice, which exhibit a gender-bias in autoimmune diabetes frequency, the BDC2.5/NOD.Rago/o mice did not exhibit gender-bias in autoimmune diabetes development unless they were treated with mATG (Fig. 1b). Gender-bias has been observed in natural development of a number of autoimmune diseases; several mechanisms have been suggested, including regulation of T helper subset differentiation by hormonal factors that are genetically determined by chromosomal differences.19 It remains to be addressed how mATG treatment elicits a gender-biased effect in BDC2.5/NOD.Rago/o mice. We did not detect a difference in the extent of T-cell depletion between male and female animals. Whether ATG has a gender effect in humans remains to be examined in clinical trials.

Figure 1
Lymphodepletion by anti-mouse thymocyte globulin (mATG) inhibited autoimmunity with a gender effect. The mATG was administered to female or male BDC2.5/NOD.Rago/o mice, which harbour a monoclonal T-cell repertoire consisting of a single pathogenic effector ...

mATG depleted peripheral lymphocytes without apparent effect on thymocytes

The mATG was generated in a similar fashion to the clinically used ATG, by immunizing rabbits with thymocytes.7 Interestingly, mATG treatment in the BDC2.5/NOD.Rago/o mice did not reduce thymocyte cell counts per organ (179 ± 92 in controls versus 215 ± 41 in mATG-treated animals, n = 4–6), nor did it substantially alter CD4 and CD8 expression profiles (Fig. 2a). In the young BDC2.5/NOD.Rago/o group, individual animals may vary slightly in percentages of thymocyte subsets defined by CD4 and CD8 markers. The mATG treatment did not cause a consistent change of subset profiles (Fig. 2a,b and data not shown). The lack of depletion effect on thymocytes was not unique to the BDC2.5/NOD.Rago/o reductionist model, because the agent did not substantially reduce the number of thymocytes in standard NOD or C57BL/6 mice either (data not shown), consistent with a previous report showing that mATG treatment did not deplete thymocytes in C57BL/6 mice, probably as the result of reduced accessibility to the thymus for mATG.11 The mATG treatment did profoundly deplete peripheral T cells in BDC2.5/NOD.Rago/o mice in peripheral blood and all secondary lymphoid organs tested. The T-cell population recovered in 2–3 weeks after discontinuation of depletion treatment (Fig. 2a,b).

Figure 2
Anti-mouse thymocyte globulin (mATG) depleted peripheral T lymphocytes but not thymocytes. The mATG-treated BDC2.5/NOD.Rago/o mice were analysed at the time of active mATG-mediated depletion (within 2–3 days after the last mATG dose) (middle panels, ...

De novo induction of CD4+ Foxp3+ Treg cells by mATG treatment in vivo in the periphery

In standard C57BL/6 and NOD mice, mATG treatment increased the percentage of CTLA4+ Foxp3+ Treg cells by about twofold in lymph nodes, spleen and peripheral blood of the treated animals, although the total numbers of Treg cells were substantially reduced (data not shown), probably by the depletion effect of mATG, which might preferentially, but not completely, spare Treg cells.11 The increased Treg-cell percentage could still be contributed by de novo Treg-cell induction, but this possibility is very difficult to address without employing a reductionist, antigen-specific system.

The BDC2.5/NOD.Rago/o mouse can be instrumental for studying Treg-cell induction. It harbours a single clone of pathogenic Teff cells as its entire T-cell repertoire and is devoid of Treg cells.15,17 As shown in Fig. 3, the animals that received control rabbit IgG or no treatment did not harbour detectable Foxp3+ T cells when analysed at pre-diabetic stages around 20–30 days of age; the very low percentages, 0·1% to 0·3%, of cells that were stained by anti-Foxp3 antibodies (Fig. 3a) were similar to that stained with isotype control antibodies irrelevant to the Foxp3 protein (data not shown). However, when the T-cell population rebounded after mATG treatments, the pancreatic draining lymph nodes of all treated animals contained a substantial frequency of Foxp3+ Treg cells that can be readily detected by flow cytometry (Fig. 3a,b). Most of the induced Foxp3+ cells also expressed CTLA4 (Fig. 3a), a molecule that is essential for Treg-cell function.20 As the Rag deficiency in these animals precludes the production of any T-cell specificity other than that encoded by the BDC2.5 pathogenic Teff-cell clone, the newly emerged Foxp3+ T cells must carry the same TCR as the pathogenic Teff-cell clone, possibly induced by de novo generation in the thymus or conversion of Teff cells to Treg cells in the periphery. The iTreg cells did not seem to have originated from the thymus, as Foxp3+ cells were not detected in the thymus (Fig. 3a,b), in agreement with the lack of a depletion effect on this primary lymphoid organ by mATG (Fig. 2). For the majority of the mATG-treated animals, the induction of Treg cells was most evident in the pancreatic draining lymph nodes, the local draining lymph nodes of the pancreatic islet beta cells, the specific target cells of the BDC2.5 T-cell clone (Fig. 3a,b). However, from the animals that achieved long-term protection from autoimmunity by the short course of mATG treatment, iTreg cells were detected in other peripheral lymphoid organs and in the blood circulation (Fig. 4 and data not shown).

Figure 3
Local induction of CD4+ Foxp3+ CTLA4+ regulatory T cells in the periphery from a monoclonal Teff-cell repertoire. anti-mouse thymocyte globulin (mATG) -treated BDC2.5/NOD.Rago/o mice were analysed 3 weeks after the depletion treatment (bottom panels), ...
Figure 4
Appearance of induced Foxp3+ regulatory T (Treg) cells in peripheral blood correlated with induction of immune tolerance in a model of aggressive autoimmune damage. (a) A representative profile of Foxp3 expression by induced Treg cells (iTreg) in peripheral ...

The detection of iTreg cells in the blood prompted us to further analyse the kinetics and phenotypes of these cells in peripheral blood, in anticipation that such analyses might also aid in the assessment of Treg-cell induction in clinical settings, because peripheral blood is the most accessible portal for clinical analyses. Indeed, induced Treg cells persisted in the peripheral blood of mATG-tolerized animals throughout the monitoring period, with variable ratios from about 10% to 50% of the total CD4+ T cells (Fig. 4). On average, about 10 iTreg cells per μl of peripheral blood could be detected.

We then analysed iTreg cells in peripheral blood of mATG-treated, long-lived animals for the expression of cell surface and intracellular Treg-cell marker proteins, in comparison to that of nTreg cells in the blood of BDC2.5/NOD TCR transgenic mice and standard NOD mice. As shown in Fig. 5, the percentages of CD103+ cells among iTreg cells in the BDC2.5/NOD.Rago/o mice was doubled compared with that of nTreg cells in BDC2.5/NOD mice (60 ± 8 versus 28 ± 5; n = 4–7; P < 0·0001; Student's t-test). [We did not find substantial differences between the iTreg cells and nTreg cells in the levels of CD25, CTLA4, GITR and CD127 (not shown)]. Intracellular staining of Ki-67, a nuclear antigen marker for cellular proliferation, revealed that Treg cells induced by mATG treatment and natural Treg cells had similar frequencies of Ki-67+ cells (not shown), indicating that the high percentage of iTreg cells in mATG-treated, long-lived animals was not the result of an excessive proliferation rate.

Figure 5
The induced Treg cells in the peripheral blood of anti-mouse thymocyte globulin (mATG) -treated, long-lived BDC2.5/NOD.Rago/o male mice expressed higher levels of CD103. Blood samples from the mATG-protected animals (right) were analysed by flow cytometry, ...

The tolerogenic effect of mATG depended on the induction of Foxp3+ Treg cells

Whereas the induced Treg cells in the long-lived BDC2.5/NOD.Rago/o mice may indeed be critical for maintaining immune tolerance status by suppressing the pathogenic Teff cells, mATG could also be directly anergizing the Teff cells in the animals. To dissect these possibilities, CD4+ CD25+ cells and CD4+ CD25 cells were purified by flow cytometry sorting from the spleen and lymph nodes of mATG-tolerized BDC2.5/NOD.Rago/o mice (> 160 days diabetes-free), and adoptively transferred into new recipients, at a dose of 7 × 104 per animal, to assess their pathogenic and regulatory capacity. Groups of BDC2.5/NOD.Rago/o recipients (n = 5 each) were injected with CD4+ CD25+ cells from mATG-tolerized BDC2.5/NOD.Rago/o mice (iTreg cells), or with CD4+ CD25+ cells from BDC2.5/NOD mice (nTreg cells). The protective efficacies of iTreg cells and nTreg cells are very similar, conferring at least short-term protection from autoimmune damage for all recipients. Two of the five animals receiving iTreg cells had long-term (> 150 days) protection. The same frequency of long-term protection also occurred in the group that received nTreg cells (Fig. 6a). On the other hand, CD4+ CD25 cells from the tolerized animals caused fulminate autoimmune damage in Rag-deficient recipients, with all recipients becoming diabetic within 2 weeks, in a course similar to that caused by CD4+ CD25 Teff cells isolated from non-manipulated BDC2.5/NOD mice (Fig. 6b). Furthermore, CD4+ CD25 cells from mATG-tolerized animals did not show any protective effect when transferred into new BDC2.5/NOD.Rago/o hosts (Fig. 6a).

Figure 6
Essential role of induced Foxp3+ regulatory T (Treg) cells for the long-term protective effect of anti-mouse thymocyte globulin (mATG). (a) Adoptive transfer experiments to test the efficacy of mATG-induced Treg (iTreg) cells from long-lived, male BDC2.5/NOD. ...

The in vivo suppressive activity of iTreg cells from mATG-treated animals, together with the phenotypes shown in Figs 14, strongly suggest that the long-lasting effect of mATG treatment may depend on the induction of Foxp3+ Treg cells. However, lymphodepletion has long been thought to be the major effect of ATG. We examined the contribution of lymphodepletion in the absence of Treg-cell induction, by treating BDC2.5/NOD.Foxp3sf mice, which carry a null mutation of the Foxp3 gene.15 This genetic deficiency of Foxp3 made it impossible to induce Foxp3+ Treg cells in these animals. The mATG treatment of BDC2.5/NOD.Foxp3sf mice (n = 5) did delay autoimmune damage, but the delay only lasted for about 5 days compared with the control treatment (Fig. 6c). Hence, mATG treatment could only achieve short-term protection from autoimmune damage in a model wherein the induction of Foxp3+ Treg cells was genetically blocked. Although it remains possible that additional factors, other than the Treg-cell deficit, might contribute to the relative resistance of BDC2.5/NOD.Foxp3sf mice to mATG treatment, the results from this experiment, taken together with results from the adoptive transfer experiments described above (Fig. 6a,b), strongly suggest that the tolerogenic effect of mATG may depend on induction of Foxp3+ Treg cells.


Anti-thymocyte globulin is a prototype of lymphodepletion therapies. It was developed primarily for immunosuppressive conditioning in transplantation, but is now actively tested for autoimmune disorders.2,3,21 Using a reductionist mouse model, we found that mATG treatment in mice, besides its effect in T-cell depletion, indeed induced Treg cells de novo (Fig. 2). Prevalence of induced Treg cells in peripheral blood correlated with induction of immune tolerance (Fig. 3). Despite the fact that mATG is generated against thymocytes as an immunogen, in a similar fashion as for the clinically used ATG, thymocytes were not depleted in mATG-treated animals, in contrast to the efficient depletion in peripheral lymphoid organs and blood. The apparent lack of depletion effect in the thymus was consistent with the absence of Treg-cell induction in the thymus. Instead, Treg-cell induction occurred consistently in the periphery, in all treated animals in the local draining lymph nodes of the organ that was subject to immune damage (Fig. 2).

The development of nTreg cells is thought to be shaped by a repertoire of TCRs with specificities towards self-antigens.22 Peripheral induction of Treg cells remains a topic of debate.23 Strong evidence of Treg-cell induction has been produced in studies with artificially engineered antigens.24 However, a study using BDC2.5/NOD models found no evidence of peripheral induction of Treg cells specific to natural self-antigens under physiological conditions.25 Conversion of adoptively transferred CD4+ CD25 or CD4+ Foxp3 cells to CD4+ CD25+ Foxp3+ cells have been well-documented in many studies, but the clonal diversity of T cells in those studies preclude a definitive conclusion on the possibility of converting a pathogenic Teff-cell clone into a specific Treg cell. Our studies used a monoclonal derivative of the BDC2.5/NOD model that harbours one known Teff-cell clone. The model was subjected to a pharmacological intervention that is clinically relevant to Thymoglobulin depletion therapy, and the potential conversion of a Teff-cell clone was studied in the absence of in vitro manipulation and adoptive cell transfer. Immunodepletion therapy mediated by mATG indeed ‘rebooted’ T cells at a clonal level and so reset an immunoregulatory balance. This study indicates that it is possible to convert a bona fide pathogenic Teff-cell clone into a protective Treg-cell clone with the same specificity in the periphery by lymphodepletion therapy, a possibility also clearly illustrated by a recent study of non-depleting anti-CD3 monoclonal antibody therapies in BDC2.5/NOD.Rago/o models.26 An earlier study has also demonstrated that monospecific CD4 T cells against the male antigen Dby could be converted to Foxp3-expressing Treg cells by treatment with non-depleting anti-CD4 monoclonal antibodies in a transplantation setting.27 The non-depleting nature of the monoclonal anti-CD3 or anti-CD4 antibodies used in those studies contrasts the profound depleting effect by mATG. However, given the polyclonal nature of the mATG product, it is possible that a subset of the mATG mixture contains non-depleting anti-CD3 and/or anti-CD4 antibodies that might produce an effect(s) similar to what has been reported in previous studies with monoclonal anti-CD3 or anti-CD4 antibodies,26,27 with a shared mechanism(s). Nevertheless, treatment by mATG and anti-CD3 monoclonal antibodies could induce Treg cells through different mechanisms, because anti-CD3 antibody therapies, as opposed to mATG treatment (Fig. 3), could induce Foxp3+ cells in the thymus as well as in peripheral lymphoid organs.26

It should be noted, however, that a recent study26 demonstrated that BDC2.5/NOD.Rago/o.Foxp3GFP mice had a frequency of 0·15% Foxp3GFP+ cells among the CD4+ subset. The GFP reporter enabled examination of anti-CD3 therapies for conversion of the adoptively transferred BDC2.5 Foxp3GFP− Teff-cell clone into Foxp3GFP+ cells in pancreatic lymph nodes.26 Our studies of the mATG effect relied on flow cytometry analyses of Foxp3 protein expressed by natural Foxp3 allele. We could not detect a Foxp3+ population in untreated BDC2.5/NOD.Rago/o mice above the background of intracellular flow cytometry staining (approximately 0·2%). In a previous study, it was shown that Foxp3-deficient and Foxp3-sufficient BDC2.5/NOD.Rago/o mice had a virtually identical course of autoimmune diabetes development, with all mice becoming diabetic by about 20 days of age.15 That genetic study suggests that there is no functional Foxp3+ Treg-cell subset in the BDC2.5/NOD.Rago/o model, at least before the onset of autoimmune diabetes.15 In the mATG-tolerized mice, the induced Treg cells were present at unusually high frequencies (as high as approximately 50%) with a distinct phenotype (Figs 4 and and5).5). Although it remains possible that a very low frequency of pre-existing Foxp3+ cells may expand and contribute to the pool of the induced Foxp3+ Treg cells, the expanded cells should only constitute a minor portion among the high percentage of Foxp3+ cells detected, as there was no substantial evidence that the Treg cells detected in the mATG-treated animals proliferated faster than their natural counterpart. Hence, without employing a Foxp3GFP reporter, the evidence gathered so far suggests that the Treg-cell populations observed in the mATG-treated BDC2.5/NOD.Rago/o mice are most likely derived from conversion of an otherwise pathogenic Teff-cell clone.

It remains to be studied how mATG-mediated immunodepletion led to the conversion of a self-antigen-specific pathogenic Teff cell into a Treg cell. The mATG has been found to affect multiple types of immune cells. It has been reported that mATG modulated the frequency of dendritic cells, altered their expression of co-stimulatory molecules that are important for T-cell activation, and changed the profiles of cytokines secreted by dendritic cells.10,28 The altered dentritic cells have been shown to bias the differentiation of BDC2.5 T cells from Th1 to Th2 subsets in vitro and attenuate the in vivo pathogenic capacity of T cells when adoptively transferred.28 In our studies, the BDC2.5 Teff-cell population isolated from mATG-tolerized animals caused immune destruction at a similar tempo as unmanipulated BDC2.5 Teff cells, and evidence strongly suggests that iTreg cells are essential for immune tolerance induction by mATG (Fig. 6). It remains to be addressed whether alteration of dendritic cells might contribute to the de novo induction of Treg cells by inducing Foxp3 expression,5,10 or induce suppressive activity of T cells independent of Foxp3 expression.7 In addition, lymphodepletion creates a lymphopenic environment. Such an environment is known to alter the differentiation of naive T cells through a perturbed cytokine milieu.29

The lack of nTreg cells in the BDC2.5/NOD.Rago/o model afforded us an opportunity to characterize the iTreg cells versus their natural counterparts. CD103 expression was found to be increased in iTreg cells over the nTreg cells. CD103 has been associated with the activation or memory differentiation status of Treg cells, as well as with homing to non-lymphoid tissues.30,31 It remains to be studied if increased CD103 expression in mATG-induced iTreg cells has functional impact. In adoptive transfer experiments, the iTreg cells did not exhibit superiority in suppressing autoimmune diabetes (Fig. 6). In tumour-bearing mice, adoptively transferred CD4+ CD25 cells converted into CD4+ CD25+ cells, which also exhibited higher levels of CD103 expression.32 This study, with an experimental analogue to the clinically used ATG, demonstrates that it is indeed possible for lymphodepletion to induce immune tolerance, even in a setting of aggressive autoimmune damage, but it requires substantial induction of antigen-specific Treg cells. On the other hand, lymphodepletion is certainly far from inducing immune tolerance by default, but rather it might occur only in a minority of circumstances, as demonstrated in this study. As a matter of fact, it is well recognized that homeostatic proliferation ensuing T-cell depletion enhances the formation of memory T cells,33 which are resistant to tolerogenic manipulation. This study, albeit with an artificial reductionist model, suggests that circulation of iTreg cells in the blood might portend ongoing tolerance induction, and it might be possible to distinguish induced and natural Treg cells in the blood circulation. It remains to be understood what factors conduct lymphodepletion to immune tolerance induction, and what parameters responding to the depletion treatment might predict successful tolerance induction.


This work was in part supported by start-up funds from the University of Miami Miller School of Medicine and the Diabetes Research Institute Foundation, and the National Institute of Diabetes and Digestive and Kidney Diseases (award number DP3DK085696). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Diabetes and Digestive and Kidney Diseases or the National Institutes of Health. The authors thank Andrew Ferguson and Melanie Ruzek (Genzyme Corporation) for their critical review of the manuscript.


The authors have no conflict of interest to disclose.


1. Sakaguchi S, Yamaguchi T, Nomura T, Ono M. Regulatory T cells and immune tolerance. Cell. 2008;133:775–87. [PubMed]
2. Beiras-Fernandez A, Thein E, Hammer C. Induction of immunosuppression with polyclonal antithymocyte globulins: an overview. Exp Clin Transplant. 2003;1:79–84. [PubMed]
3. Mohty M, Gaugler B. Mechanisms of action of antithymocyte globulin: old dogs with new tricks! Leuk Lymphoma. 2008;49:1664–7. [PubMed]
4. Lopez M, Clarkson MR, Albin M, Sayegh MH, Najafian N. A novel mechanism of action for anti-thymocyte globulin: induction of CD4+ CD25+ Foxp3+ regulatory T cells. J Am Soc Nephrol. 2006;17:2844–53. [PubMed]
5. Feng X, Kajigaya S, Solomou EE, et al. Rabbit ATG but not horse ATG promotes expansion of functional CD4+ CD25high FOXP3+ regulatory T cells in vitro. Blood. 2008;111:3675–83. [PMC free article] [PubMed]
6. Broady R, Yu J, Levings MK. ATG-induced expression of FOXP3 in human CD4+ T cells in vitro is associated with T-cell activation and not the induction of FOXP3+ T regulatory cells. Blood. 2009;114:5003–6. [PubMed]
7. Ruzek MC, Waire JS, Hopkins D, et al. Characterization of in vitro antimurine thymocyte globulin-induced regulatory T cells that inhibit graft-versus-host disease in vivo. Blood. 2008;111:1726–34. [PubMed]
8. Serban G, Whittaker V, Fan J, et al. Significance of immune cell function monitoring in renal transplantation after Thymoglobulin induction therapy. Hum Immunol. 2009;70:882–90. [PubMed]
9. Chung DT, Korn T, Richard J, Ruzek M, Kohm AP, Miller S, Nahill S, Oukka M. Anti-thymocyte globulin (ATG) prevents autoimmune encephalomyelitis by expanding myelin antigen-specific Foxp3+ regulatory T cells. Int Immunol. 2007;19:1003–10. [PubMed]
10. Simon G, Parker M, Ramiya V, et al. Murine antithymocyte globulin therapy alters disease progression in NOD mice by a time-dependent induction of immunoregulation. Diabetes. 2008;57:405–14. [PubMed]
11. Ruzek MC, Neff KS, Luong M, et al. In vivo characterization of rabbit anti-mouse thymocyte globulin: a surrogate for rabbit anti-human thymocyte globulin. Transplantation. 2009;88:170–9. [PubMed]
12. Allen PM. Defining yourself: tolerance development in the immune system. J Immunol. 2006;177:1369–72. [PubMed]
13. Katz JD, Wang B, Haskins K, Benoist C, Mathis D. Following a diabetogenic T cell from genesis through pathogenesis. Cell. 1993;74:1089–100. [PubMed]
14. Kanagawa O, Militech A, Vaupel BA. Regulation of diabetes development by regulatory T cells in pancreatic islet antigen-specific TCR transgenic nonobese diabetic mice. J Immunol. 2002;168:6159–64. [PubMed]
15. Chen Z, Herman AE, Matos M, Mathis D, Benoist C. Where CD4+ CD25+ T reg cells impinge on autoimmune diabetes. J Exp Med. 2005;202:1387–97. [PMC free article] [PubMed]
16. Brode S, Raine T, Zaccone P, Cooke A. Cyclophosphamide-induced type-1 diabetes in the NOD mouse is associated with a reduction of CD4+ CD25+ Foxp3+ regulatory T cells. J Immunol. 2006;177:6603–12. [PubMed]
17. Gonzalez A, Andre-Schmutz I, Carnaud C, Mathis D, Benoist C. Damage control, rather than unresponsiveness, effected by protective DX5+ T cells in autoimmune diabetes. Nat Immunol. 2001;2:1117–25. [PubMed]
18. Itoh M, Takahashi T, Sakaguchi N, Kuniyasu Y, Shimizu J, Otsuka F, Sakaguchi S. Thymus and autoimmunity: production of CD25+ CD4+ naturally anergic and suppressive T cells as a key function of the thymus in maintaining immunologic self-tolerance. J Immunol. 1999;162:5317–26. [PubMed]
19. Rubtsov AV, Rubtsova K, Kappler JW, Marrack P. Genetic and hormonal factors in female-biased autoimmunity. Autoimmun Rev. 2010;9:494–8. [PMC free article] [PubMed]
20. Wing K, Onishi Y, Prieto-Martin P, Yamaguchi T, Miyara M, Fehervari Z, Nomura T, Sakaguchi S. CTLA-4 control over Foxp3+ regulatory T cell function. Science. 2008;322:271–5. [PubMed]
21. Suzuki J, Ricordi C, Chen Z. Immune tolerance induction by integrating innate and adaptive immune regulators. Cell Transplant. 2010;19:253–68. [PMC free article] [PubMed]
22. Hsieh CS, Liang Y, Tyznik AJ, Self SG, Liggitt D, Rudensky AY. Recognition of the peripheral self by naturally arising CD25+ CD4+ T cell receptors. Immunity. 2004;21:267–77. [PubMed]
23. Germain RN. Special regulatory T-cell review: a rose by any other name: from suppressor T cells to Tregs, approbation to unbridled enthusiasm. Immunology. 2008;123:20–7. [PMC free article] [PubMed]
24. Apostolou I, Sarukhan A, Klein L, von Boehmer H. Origin of regulatory T cells with known specificity for antigen. Nat Immunol. 2002;3:756–63. [PubMed]
25. Wong J, Mathis D, Benoist C. TCR-based lineage tracing: no evidence for conversion of conventional into regulatory T cells in response to a natural self-antigen in pancreatic islets. J Exp Med. 2007;204:2039–45. [PMC free article] [PubMed]
26. Nishio J, Feuerer M, Wong J, Mathis D, Benoist C. Anti-CD3 therapy permits regulatory T cells to surmount T cell receptor-specified peripheral niche constraints. J Exp Med. 2010;207:1879–89. [PMC free article] [PubMed]
27. Cobbold SP, Castejon R, Adams E, Zelenika D, Graca L, Humm S, Waldmann H. Induction of foxP3+ regulatory T cells in the periphery of T cell receptor transgenic mice tolerized to transplants. J Immunol. 2004;172:6003–10. [PubMed]
28. Huang Y, Parker M, Xia C, et al. Rabbit polyclonal mouse antithymocyte globulin administration alters dendritic cell profile and function in NOD mice to suppress diabetogenic responses. J Immunol. 2009;182:4608–15. [PMC free article] [PubMed]
29. Ramsey C, Rubinstein MP, Kim DM, Cho JH, Sprent J, Surh CD. The lymphopenic environment of CD132 (common gamma-chain)-deficient hosts elicits rapid homeostatic proliferation of naive T cells via IL-15. J Immunol. 2008;180:5320–6. [PubMed]
30. Huehn J, Siegmund K, Lehmann JC, et al. Developmental stage, phenotype, and migration distinguish naive- and effector/memory-like CD4+ regulatory T cells. J Exp Med. 2004;199:303–13. [PMC free article] [PubMed]
31. Izcue A, Coombes JL, Powrie F. Regulatory T cells suppress systemic and mucosal immune activation to control intestinal inflammation. Immunol Rev. 2006;212:256–71. [PubMed]
32. Valzasina B, Piconese S, Guiducci C, Colombo MP. Tumor-induced expansion of regulatory T cells by conversion of CD4+ CD25 lymphocytes is thymus and proliferation independent. Cancer Res. 2006;66:4488–95. [PubMed]
33. Goldrath AW, Bogatzki LY, Bevan MJ. Naive T cells transiently acquire a memory-like phenotype during homeostasis-driven proliferation. J Exp Med. 2000;192:557–64. [PMC free article] [PubMed]

Articles from Immunology are provided here courtesy of British Society for Immunology
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...