We have recently described two independent mouse models in which the administration of diphtheria toxin (DT) leads to specific depletion of regulatory T cells (Tregs) due to expression of DT receptor-enhanced GFP under the control of the Foxp3 promoter. Both mouse models develop severe autoimmune disorders when Foxp3+ Tregs are depleted. Those findings were challenged in a recent study published in this journal suggesting the expression of Foxp3 in epithelial cells as the cause for disease development. By using genetic, cellular, and immunohistochemical approaches, we do not find evidence for Foxp3-expression in nonhematopoietic cells. DT injection does not lead to a loss of epithelial integrity in our Foxp3-DTR models. Instead, Foxp3 expression is Treg-specific and ablation of Foxp3+ Tregs leads to the induction of fatal autoimmune disorders. Autoimmunity can be reversed by the adoptive transfer of Tregs into depleted hosts, and the transfer of Foxp3-deficient bone marrow into T cell-deficient irradiated recipients leads to full-blown disease development.
The existence of a dedicated population of suppressor T cells has long been surmised. However, the immunology community en large treated this notion with contempt because the markers and other molecular features for the presumed suppressor T cell population were not known. The identification of CD25+CD4+ T cells followed by the discovery of the transcription factor Foxp3 as a faithful marker for regulatory T cells (Tregs)6 allowed for a large body of published reports demonstrating the suppressive activity of Foxp3-expressing T cells (1, 2, 3, 4, 5). A fatal early onset autoimmune syndrome in mice or humans harboring a nonfunctional Foxp3 allele is manifest to the vital significance of Treg-mediated suppression. Collectively, the general consensus in the field is that the high level of Foxp3 expressed by Tregs is essential for their suppressive activity. In accordance with this idea, the loss of Foxp3 expression in hematopoietic cells, more specifically in T cells (1), or the loss of Tregs induces an autoimmune syndrome (6, 7) similar to that in mice with germline mutations in Foxp3 (8). However, this model was recently challenged by Liu and colleagues (9, 10). According to their studies, Foxp3 expression is not restricted to Treg cells, let alone hematopoietic cells, but is widespread in variety of epithelial tissues including thymic, mammary gland, lung, and prostate epithelial cells (9, 10, 11, 12). These investigators argued that the loss of epithelial-specific Foxp3 expression is the cause of life-threatening autoimmunity in immunodysregulation, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX) patients, Foxp3-deficient mice, or mice engineered to eliminate Foxp3-expressing cells, whereas Foxp3 expression in hematopoietic cells, including Treg cells, is not essential (10). Therefore, we revisited the issue of expression of Foxp3 in nonlymphoid tissues using several strains of mice with targeted mutations of the Foxp3 locus and monoclonal and polyclonal Foxp3 Abs.
Materials and Methods
DEREG (depletion of regulatory T cells; Ref. 7), C57BL/6, and BALB/c mice were bred at the animal facility of the Institute for Medical Microbiology, Immunology, and Hygiene (Technical University, Munich, Germany). Foxp3GFP (13) reporter mice and Foxp3DTR (where DTR is diphtheria toxin receptor; Ref. 6) mice were bred at the animal facility of the Department of Immunology, University of Washington (Seattle, WA). All animal experiments were performed under specific pathogen-free conditions and in accordance with institutional, state, and federal guidelines.
175) (dilution 1/200; Cell Signaling), androgen receptor (dilution 1/100; catalog no. ab47563, Abcam), and estrogen receptor (dilution 1/20; catalog no. ab21232, Abcam) biotinylated rabbit anti-rat (DakoCytomation) or donkey anti-rabbit (Dianova) secondary Abs were used followed by the streptavidin alkaline phosphatase kit (catalog no. K5005, DakoCytomation). Alkaline phosphatase was revealed by Fast Red as chromogen and peroxidase was developed with a highly sensitive diaminobenzidine chromogenic substrate for 10 min.
DT treatment and Treg cell transfer
For immunohistochemistry, Tregs were depleted from DEREG (7) and control mice by i.p. injection of DT (50 mg/kg body weight; Merck) on two consecutive days, day −2 and day −1 before tissue collection. For Treg transfer, Treg cells were eliminated in 5- to 6-wk-old Foxp3DTR recipient mice (6). In brief, 50 mg/kg DT (Sigma-Aldrich or Calbiochem) was injected i.p. into Foxp3DTR mice on two consecutive days, day 1 and day 2. CD4+ Foxp3GFP+ cells (5 × 105) isolated from Foxp3GFP reporter mice were injected i.v. on day 3 of the experiment (13). DT was injected on days 4, 6, and 8, and mice were euthanized on day 9 of the experiment.
Results and Discussion
Central to their model of Foxp3 function in epithelial cells, Liu and colleagues reported Foxp3 protein expression in epithelial cells by immunohistochemical staining (10). According to their arguments, eliminating Foxp3-expressing cells in Foxp3DTR mice is expected to dramatically destroy the epithelial tissue architecture (10). To investigate whether tissue destruction was a consequence of DT-mediated depletion of Foxp3+ cells, we injected Foxp3DTR (DEREG) mice with DT twice before the analysis of several organs by H&E staining. No tissue destruction could be observed in any of the analyzed organs (Fig. 1⇓). In sharp contrast to the aforementioned observations, high levels of Foxp3 protein were detected in cells with lymphoid but not epithelial morphology (Fig. 1⇓⇓). The latter was in full agreement with a previous report (14), and staining was also not detectable even when Abs were used at a 10-fold higher concentration (data not shown). Similar results were obtained using a polyclonal, affinity-purified, rabbit anti-mouse Foxp3 Ab (supplemental Fig. 1A).7 In addition to immunohistochemical staining of tissue sections, we also prepared whole tissue extracts from colon, lung, prostate, and spleen of wild-type, Rag1−/−, and Foxp3−/− mice for Western blot analysis of Foxp3 expression. We failed to detect Foxp3 protein in lung and prostate tissues isolated from Rag-deficient mice, further corroborating our findings through an independent method of Foxp3 detection (Fig. 2⇓). Furthermore, eliminating Foxp3-expressing cells in Foxp3DTR (DEREG) (7) mice resulted in the loss of Foxp3 signal among lymphoid cells, whereas very weak background levels of nonspecific cytoplasmic Foxp3 staining in epithelial cells remained unchanged (Fig. 1⇓). Importantly, the overall tissue architecture supported by epithelial cells was unperturbed in the DT-treated Foxp3DTR (DEREG) mice. DT-treated animals were followed closely in various experimental settings, and we never observed tissue destruction. Furthermore, DEREG mice have been used by different groups (see for example Refs. 15, 16, 17, 18, 19) with no reports of epithelial tissue damage.
Because all studies by Liu et al. were performed in BALB/c mice, we crossed our C57BL/6 Foxp3DTR mice onto a BALB/c background for 10 generations to exclude strain-specific differences (9, 10, 11, 12). Identical to our results on the C57BL/6 background, nonhematopoietic tissues were not affected by DT administration in BALB/c Foxp3DTR (DEREG) mice (Fig. 1⇑). This was further confirmed by cleaved caspase-3 labeling of prostate tissue showing, according to Ki-67 staining, very little ongoing apoptosis and a low rate of regular epithelial proliferation, both of which remained unchanged in the different mouse genotypes with or without DT injection (Fig. 1⇑).
Additionally, in contrast to another study that reported broad Foxp3 expression by breast epithelial cells and its role as a cancer suppressor (12), we did not detect Foxp3 expression in normal breast tissue using the monoclonal (Fig. 3⇓) or polyclonal Ab (supplemental Fig. 1B). Again, tissue morphology was not altered following DT treatment in this organ.
Altogether, we have been unable to independently confirm the assertion that Foxp3 is highly expressed by epithelial cells. Therefore, we conclude that autoimmunity induced by elimination of Foxp3-expressing cells does not represent the loss of epithelial cells or its integrity (or architecture). To account for these discordant results, we speculate that the polyclonal rabbit anti-Foxp3 Ab used in previous studies of Foxp3 expression in epithelial cells nonspecifically cross-reacts with an unidentified protein Ag in addition to or instead of Foxp3.
The strongest argument for the contention that Foxp3 expression in epithelial cells is biologically important is derived from earlier observation that the transfer of Foxp3-deficient bone marrow into Foxp3-sufficient, Rag-deficient recipient mice does not induce autoimmunity (11). This result suggested that Foxp3 deficiency in bone marrow-derived cells was not primary to autoimmune disease instigation, challenging the current dogma. In a stark contrast to these findings, we and other groups have reported that bone marrow transfer from Foxp3-deficient mice into Rag−/− recipients leads to autoimmune lymphoproliferative disease (2, 20, 21). Additionally, wild-type bone marrow transferred into Foxp3sf (where sf is scurfy) Rag−/− mice does not cause the disease (22). The disease was not caused by mature pathogenic T cells possibly contaminated within the Foxp3sf donor bone marrow cells because the transfer of fetal liver cells from Foxp3sf mice or bone marrow cells from Foxp3sf nude mice, both of which do not contain any mature pathogenic T cells, also caused identical disease in irradiated Rag−/− recipients (14, 22). The usage of recipient mice with a genetic deficiency in T cell generation is essential for assessing the contribution of Foxp3 expression in hematopoietic cells (22). As we demonstrated, the use of T lymphocyte-deficient recipients is necessary because radiation-resistant Treg cells in wild-type recipients rapidly reconstitute the Treg cell compartment after irradiation (22). Thus, expanded host-derived Treg cells spare irradiated recipient mice from lethal autoimmunity, irrespective of the donor bone marrow genotype. Notably, an early bone marrow transfer study cited by Liu and colleagues to independently support their model for Foxp3 action in nonhematopoietic tissues used irradiated wild-type mice as recipients of Foxp3-deficient bone marrow (23). Therefore, the result obtained by Liu and colleagues has not, to date, been independently reproduced. It should also be pointed out that the bone marrow transfers in a study that championed a role for Foxp3 expression in nonhematopoietic cells lacked a positive control, i.e., bone marrow transfer from Foxp3− mice into Foxp3−/−Rag−/− mice, obfuscating the interpretation of the key finding (11). In contrast, our studies tested all combinations of donor and host Foxp3 genotypes and demonstrated no contribution of Foxp3 deficiency in nonhematopoietic tissues to disease development (22). Finally, bone marrow transplantation serves as an effective treatment for IPEX patients in agreement with our bone marrow chimera studies in mice (24).
In addition to bone marrow transfer studies, a large body of genetic evidence has established that Foxp3 expression in T cells and more specifically in Treg cells is required to prevent fatal autoimmunity. Mice with a Foxp3 deficiency restricted to the T cell lineage through CD4-Cre-mediated deletion of a conditional Foxp3 allele are indistinguishable from mice with the germline ablation of the Foxp3 gene (1, 14). In contrast, near complete deletion of Foxp3 in thymic epithelial cells was inconsequential, i.e., no changes in thymocyte development and no signs of autoimmunity were observed (14). To accommodate these findings, Liu and colleagues suggested that CD4-Cre is expressed and mediates Foxp3 deletion in thymic epithelial cells and that the latter, not Foxp3 deficiency in T cells, causes the autoimmune syndrome in these mice (11). However, extensive genetic analyses of Cre-mediated recombination in CD4-Cre transgenic mice using a highly sensitive reporter allele failed to detect Cre expression in thymic epithelial cells, and genetically controlled immunhistochemical analysis failed to detect Foxp3 protein expression in thymic epithelial cells (14). Lastly, we have demonstrated that Treg cell-specific Foxp3 deletion via retroviral Cre delivery to purified Treg cells isolated from conditional Foxp3-knockout mice abrogates Treg cell suppressive activity, causing fulminating autoimmunity when these cells were transferred into lymphopenic recipients either alone or together with T cells from Foxp3−/− mice (25). Hence, fatal autoimmunity in mice and humans with Foxp3 mutations or in mice engineered to inducibly eliminate Foxp3-expressing cells is ascribed to the essential role of Foxp3 in Treg cells.
If autoimmunity in Foxp3-deficient mice and humans is indeed caused by the lack of Treg cells, restoration of the Treg cell compartment is predicted to cure disease. In agreement with this notion, we and five other groups have demonstrated that injecting purified Treg cells into Foxp3-deficient mice is sufficient to prevent life-threatening autoimmunity (1, 21, 22, 26, 27, 28). Additionally, the transfer of Treg cells into Foxp3DTR mice treated with DT to eliminate Foxp3-expressing cells inhibits tissue pathology in the liver, lung, and skin (Fig. 4⇓). In contrast to these results, Liu and colleagues recently demonstrated that injecting sorted Treg cells into scurfy mice did not alleviate morbidity (9). Thus, our results and the published data from a number of other groups are at odds with those generated in the Liu laboratory.
In summary, a significant body of genetic, immunohistochemical, and genetically controlled functional studies by us and others is irreconcilable with the view that Foxp3 function in epithelial cells, including the thymic epithelium, is a key factor in the prevention of severe autoimmune lesions associated with Foxp3 mutations.
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.
↵1 This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB/TR22 and SFB 587).
↵2 J.K. and K.L. contributed equally to this work.
↵3 Current address: Genentech, 1 DNA Way, South San Francisco, CA 94080.
↵4 Current address: Immunology Program, Memorial Sloan Kettering Cancer Center, Box 212, 1275 York Avenue, New York, NY 10065.
↵5 Address correspondence and reprint requests to Dr. Tim Sparwasser, Institute of Infection Immunology, TWINCORE, Center for Experimental and Clinical Infection Research, Feodor-Lynen-Straße 7, 30625 Hannover, Germany. E-mail address:
↵6 Abbreviations used in this paper: Treg, regulatory T cell; DEREG, depletion of regulatory T cells; DT, diphtheria toxin; DTR, DT receptor; IPEX, immunodysregulation, polyendocrinopathy, enteropathy, X-linked syndrome; sf, scurfy.
↵7 The online version of this article contains supplemental material.
- Received December 23, 2008.
- Accepted October 13, 2009.
- Copyright © 2009 by The American Association of Immunologists, Inc.