Addition of rapamycin to cultures of expanding natural CD4+CD25+Foxp3+ T regulatory cells (Tregs) helps maintain their suppressive activity, but the underlying mechanism is unclear. Pim 2 is a serine/threonine kinase that can confer rapamycin resistance. Unexpectedly, pim 2 was found to be constitutively expressed in freshly isolated, resting Tregs, but not in CD4+CD25 T effector cells. Introduction of Foxp3, but not Foxp3Δ2, into effector T cells induced pim 2 expression and conferred preferential expansion in the presence of rapamycin, indicating that Foxp3 can regulate pim 2 expression. Finally, we determined there is a positive correlation between Treg expansion and Foxp3 expression in the presence of rapamycin. Together, these results indicate that Tregs are programmed to be resistant to rapamycin, providing further rationale for why this immunosuppressive drug should be used in conjunction with expanded Tregs.

The ability of natural CD4+CD25+Foxp3+ T regulatory cells (Tregs)3 to suppress immune responses has generated interest in harnessing their therapeutic power to treat autoimmune disease and enable transplants (1, 2). Successful therapeutic application of Tregs will likely require significant ex vivo expansion. The inability to isolate pure Treg populations coupled with their relative ex vivo proliferative disadvantage has made expansion of functional Tregs isolated from peripheral blood problematic (1). A considerable breakthrough occurred when Roncarolo and colleagues observed that addition of rapamycin to expanding murine Treg cultures significantly and consistently increased the yield of Foxp3-expressing cells with suppressive activity (3). Similar findings have been described in human Treg culture systems. Furthermore, in vivo administration of rapamycin preferentially preserves Treg function (4, 5, 6).

The mechanism by which rapamycin maintains the suppressive activity of expanding Tregs is unclear. One study demonstrated that Tregs were resistant to rapamycin-induced apoptosis and thus selectively expanded in the presence of rapamycin (7), while another study suggested that rapamycin induced a transient Treg phenotype in T effector cells (8). Resolving this controversy has important clinical implications. If rapamycin only temporally endows T effector cells with regulatory activity, as the latter study suggests, then its clinical utility is questionable. In contrast, if rapamycin does preferentially select for Tregs at the expense of effector cells, then it would be an important component of Treg culture systems for adoptive T cell therapy.

Pim 2 is a transcriptionally regulated serine/threonine kinase discovered as a proviral integration site for the Moloney murine leukemia virus (9, 10). Functionally, pim 2 has considerable overlap with Akt and, by extension, mTOR. Akt and pim 2 share many common downstream targets including Bad and 4E binding protein-1 (11). In effector lymphocytes, pim 2 expression is tightly regulated by cytokine-induced JAK/STAT pathways and its expression rapidly disappears upon cytokine removal (12). In murine lymphocytes and cell lines, pim 2 can mediate resistance to rapamycin (13). In this study, we show that pim 2 is regulated in a fundamentally different way in Tregs. Foxp3, the master regulator of Tregs, induces pim 2 expression in Tregs. This permits constitutive pim 2 expression in resting Tregs, conferring a replicative advantage in cultures containing rapamycin. These results argue that natural, Foxp3-expressing T cells are indeed selected for in the presence of rapamycin and thus the use of rapamycin in expanding Treg cultures is a promising way to enable adoptive Treg cell therapy.

Primary human CD4+ T cells from healthy donors were purified by negative selection as previously described (14). Tregs were purified by CD25+ or CD127CD25+ selection using magnetic beads as per the manufacturer’s suggestions (Miltenyi Biotech). CD4 T cells were expanded with anti-CD3/anti-CD28 Ab-coated beads (15) or by coculture with irradiated, anti-CD3 Ab-loaded, lentiviral vector-transduced K562 artificial APCs (aAPCs) expressing CD64 and CD86 as described previously (16). T cells were cultured in the presence of rhIL-2 (300 U/ml; Chiron) and, where indicated, rapamycin (100 ng/ml, Calbiochem). To determine relative cell expansion, the number of cells at the end of culture was divided by the number of cells that initiated the culture. Fifty nanograms of PMA (Sigma-Aldrich) and 500 ng of ionomycin (Calbiochem) were added to the cells before performing intracellular cytokine staining.

Surface staining for CD4 and CD25 (BD Pharmingen) was performed according to the manufacturer’s recommendations. Intracellular staining was performed using the FOXP3 Fix/Perm kit (Biolegend) for Foxp3 and the Caltag Fix & Perm kit (Invitrogen) for IL-2 (BD Pharmingen) as per each manufacturer’s recommendation. All flow cytometry was analyzed using FACSCalibur (BD Biosciences) and FlowJo software (Tree Star).

RNA was purified, reverse transcribed, amplified, and analyzed as previously described (14) using the ABI Prism 7900HT sequence detection system (Applied Biosystems). Primers and probes to detect 28 S ribosomal RNA and pim 2 were designed using Primer Express software (Applied Biosystems) and are available upon request.

Cell lysis, electrophoresis, and immunoblotting were performed as described previously (15). Anti-human pim 2 (C-20) and anti-human actin (I-19) were purchased from Santa Cruz Biotechnology.

GFP, yellow fluorescent protein (YFP) 2A Foxp3, and YFP 2A Foxp3Δ2 were cloned upstream of the EF-1a promoter in a previously described lentiviral vector (15) so that all transduced cell populations could be detected by FL1. The 2A sequence used to allow coexpression of YFP and Foxp3 was GSGEGRGSLLTCGDVEENPGP. High titer vector was used to transduce T cells as previously described (15).

Pim 2 expression confers rapamycin resistance to murine T cells (14). We first confirmed that human pim 2 also confers rapamycin resistance. To do this, we transduced primary human CD4 T cells with a lentiviral vector expressing pim 2 and noted that pim 2 promoted the expansion of CD4 T cells in the presence and absence of rapamycin (data not shown), making it an attractive target to study in Tregs. Studies have shown that the addition of rapamycin at the initiation of a Treg culture preserves the suppressive function of the expanded cells (3, 5, 8). This suggests that if pim 2 confers rapamycin resistance in Tregs, it should be expressed in resting Tregs. To investigate this, we purified Tregs and conventional T cells from freshly isolated leukapheresis products and examined the populations for pim 2 mRNA and protein expression. Measurement of Foxp3 expression by flow cytometry confirmed that we successfully enriched for Tregs (Fig. 1,A). Surprisingly, pim 2 mRNA (Fig. 1,B) and protein (Fig. 1 C) were readily found in the CD4+CD25+ T cells and virtually undetectable in the CD4+CD25 population. This was unexpected, because both pim 2 mRNA and protein are highly labile (12), making it well suited for its role as an environmental sensor. Indeed, the first insights into pim 2 function were revealed in a DNA microarray screen searching for transcripts acutely regulated by cytokine withdrawal (12). Further studies demonstrated that multiple cytokines and growth factors can induce pim 2 expression via JAK/STAT signaling pathways (11, 17). In fact, pim 2 expression has not been described in primary cells in the absence of growth factors and cytokines. Given the number of microarrays used to find differences between conventional T cells and Tregs, it is surprising that pim 2 expression in these cells is not more widely appreciated. One study did observe that resting Tregs express ∼2-fold more pim 2 than resting CD4+CD25 T cells (18), whereas other studies that compared murine Tregs with effector CD4+ T cells did not observe any differences (19, 20). Our data suggest that there is a cytokine-independent pathway to induce pim 2 expression in Tregs. Alternatively, these results could reflect differences in how Tregs and T effectors respond to minute levels of cytokine or Ag stimulation.

Foxp3 expression is the hallmark of Tregs and its ectopic expression in CD4+ T effector cells results in repression of IL-2 production and up-regulation of Treg cell surface markers, including CTLA-4, GITR, and CD25(21). One striking difference between human and murine Tregs is the approximately equal expression of the full-length and a truncated, exon 2-deleted form of Foxp3 (Foxp3Δ2) in human cells. To date, the functional significance of Foxp3Δ2 has not been elucidated. One study suggested that coexpression of both isoforms slightly increases suppressive activity (21), but the mechanism underlying this observation remains unclear. Thus, we investigated whether both Foxp3 isoforms induced pim 2 in human CD4 T cells. CD4+ CD25 T cells were activated with CD3/CD28-coated beads and transduced with GFP, YFP-2A-Foxp3, or YFP-2A-Foxp3Δ2 expression vectors (Fig. 2, A and B). Because suppression of IL-2 production is indicative of Foxp3 activity (22), we measured IL-2 production by Foxp3- or Foxp3Δ2-expressing cells (Fig. 2 C). Upon stimulation with PMA and ionomycin, 60% of untransduced cells produced IL-2. Thus, the ratio of IL-2 producing to IL-2 nonproducing cells was 1.5. This ratio was inverted in both Foxp3- and Foxp3Δ2-expressing cells (0.43 and 0.47, respectively). Additionally, both Foxp3 and Foxp3Δ2-expressing cells inhibited IL-2 production from nontransduced cells in the same culture (0.48 and 0.28, respectively), as previously reported (21). These data suggest that our Foxp3 and Foxp3Δ2 expression vectors produce functional Foxp3 isoforms.

Next, we examined the ability of both Foxp3 isoforms to induce pim 2 in human CD4+ T cells. We observed that Foxp3, but not Foxp3Δ2, can induce pim 2 in CD4+CD25 T cells (Fig. 2,D). To demonstrate that Foxp3 expression leads to preferential T cell expansion in the presence of rapamycin, we transduced primary human CD4+CD25 T cells with Foxp3, Foxp3Δ2, or control GFP expression vectors as described above. After 3 days of expansion, we split the cultures so that the cells were expanded either in the presence or absence of rapamycin for an additional 14 days. In the absence of rapamycin, both Foxp3- and Foxp3Δ2-expressing cells were at a replicative disadvantage relative to their untransduced counterparts and were diluted out (Fig. 2 E). Foxp3-transduced cells expanded in the presence of rapamycin were enriched in these cultures. In contrast, Foxp3Δ2-expressing cells were not enriched in the presence of rapamycin and were equally diluted out as they were in the absence of rapamycin. We did not observe the induction of Foxp3 expression in GFP-transduced control cells (data not shown), suggesting that rapamycin selects for Foxp3-expressing cells rather than inducing Foxp3 expression. These data demonstrate that both Foxp3 isoforms can suppress proliferation of primary human CD4+CD25 T cells. However, only the full-length Foxp3 isoform induces pim 2 expression, allowing for preferential expansion in the presence of rapamycin.

Foxp3 interacts with multiple partners, including both NFAT (23) and Runx1(24), suggesting that the molecular mechanism by which it regulates pim 2 expression is likely to be complex. However, recent data addressing the differential effects of mTOR inhibition on Tregs and effector cells may provide some clues (6). These studies illustrate that Treg activation leads to prolonged STAT5 phosphorylation rather than PI3K/AKT/mTOR activation as compared with CD4+CD25 T cells. Because STAT5 phosphorylation can also induce pim 2 expression (25), this might be one mechanism by which Tregs up-regulate pim 2. While providing some insight, this propensity of Tregs to activate STAT5 does not fully explain constitutive pim 2 expression. Nor does it fully delineate the relationship between Foxp3, STAT5, and pim 2.

Our data up to this point confirm previous work showing that forced Foxp3 overexpression hinders the expansion of T cells (21). More importantly, our results indicate that in the presence of rapamycin, Foxp3 expression aids in the expansion of T cells by up-regulating pim 2. Thus, we predicted that the degree of expansion of freshly isolated natural Tregs would positively correlate with Foxp3 expression if cultured with rapamycin. Similarly, in the absence of rapamycin we would expect the opposite to hold true. To test this, we expanded enriched Treg cells (40–80% Foxp3 positive) isolated from healthy donors using a previously described K562 cell-based aAPC. These KT64 86 aAPCs express CD64 to load an anti-CD3 agonist Ab and CD86 to engage CD28 (16). Enriched Tregs from nine donors were expanded by anti-CD3-loaded KT64 86 cells at a 2:1 ratio (Treg:aAPC) in the absence of rapamycin. Tregs from the same nine donors plus an additional six (15 total) were similarly expanded in the presence of rapamycin. Both groups were cultured for 14–20 days, after which regression analysis was performed to determine the correlation between Foxp3 expression and relative expansion. Representative data for one donor are shown in Fig. 3, A and B. We observed enrichment for Foxp3-positive cells in the presence of rapamycin and dilution of Foxp3-positive cells in the absence of rapamycin. It is important to emphasize that Tregs are not immune to all of the effects of rapamycin nor do they expand as well as effector cells ex vivo (Fig. 3,B). Rather, Tregs, because of their Foxp3-mediated pim 2 expression, are less sensitive to the immunosuppressive effects of rapamycin and preferentially expand in the presence of rapamycin. Additionally, as others have shown (3, 4, 5, 7), Tregs that expanded for extended periods in the presence of rapamycin retain suppressive ability whereas Tregs grown in its absence lose suppressive function (data not shown). In rapamycin-containing Treg cultures, we observed a positive correlation (R = 0.438) between Foxp3 expression and the degree of expansion (Fig. 3,C). Similarly, Tregs cultured without rapamycin exhibited a negative correlation (R = 0.349) between Foxp3 expression and the magnitude of expansion (Fig. 3 D). Taken together, these studies illustrate that pim 2 expression by Tregs grants a growth advantage in the presence of rapamycin. These data also strengthen the argument for the use of combination therapies employing both rapamycin and Tregs to suppress unwanted immune responses.

In summary, pim 2 is constitutively expressed in Tregs in a Foxp3-dependent manner and this expression allows for a selective growth advantage in the presence of rapamycin. These data further demonstrate that only full-length Foxp3 protein can induce pim 2, while the equally expressed Foxp3Δ2 is unable to do so. Thus, it is reasonable to conclude that factors that interact with exon 2 in Foxp3 are necessary to induce pim 2 expression. Our data demonstrate that functional Tregs are indeed selected by rapamycin. Furthermore, employing rapamycin for the selective ex vivo expansion of Tregs for adoptive T cell immunotherapy is an attractive strategy.

We are grateful to members of the Juvenile Diabetes Research Foundation Collaborative Center for Cell Therapy for helpful suggestions and discussions, Drs. Richard Carroll, and Gwen Binder for proofreading the manuscript, Dr. Steve Ziegler for providing the Foxp3 cDNAs, Abraham Chacko for help in creating the Foxp3 lentiviral vectors, the Penn Center for AIDS Research Immunology Core for providing primary human T cells.

Drs. June and Riley received a research grant from Becton Dickinson.

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 Juvenile Diabetes Research Foundation (JDRF) Collaborative Center for Cell Therapy, the JDRF Autoimmunity Center on Cord Blood Therapy for Diabetes, and National Institutes of Health Grants R01 CA105216 and R01 AI057838. S.B. received support from National Institutes of Health Grant T32 CA101968.

3

Abbreviations used in this paper: Treg, T regulatory cell; aAPC, artificial APC; mTOR, mammalian target of rapamycin; YFP, yellow fluorescent protein.

1
Bluestone, J. A..
2005
. Regulatory T-cell therapy: is it ready for the clinic?.
Nat. Rev. Immunol.
5
:
343
-349.
2
Roncarolo, M. G., M. Battaglia.
2007
. Regulatory T-cell immunotherapy for tolerance to self antigens and alloantigens in humans.
Nat. Rev. Immunol.
7
:
585
-598.
3
Battaglia, M., A. Stabilini, M. G. Roncarolo.
2005
. Rapamycin selectively expands CD4+CD25+FoxP3+ regulatory T cells.
Blood
105
:
4743
-4748.
4
Battaglia, M., A. Stabilini, B. Migliavacca, J. Horejs-Hoeck, T. Kaupper, M. G. Roncarolo.
2006
. Rapamycin promotes expansion of functional CD4+CD25+FOXP3+ regulatory T cells of both healthy subjects and type 1 diabetic patients.
J. Immunol.
177
:
8338
-8347.
5
Coenen, J. J., H. J. Koenen, E. van Rijssen, L. B. Hilbrands, I. Joosten.
2006
. Rapamycin, and not cyclosporin A, preserves the highly suppressive CD27+ subset of human CD4+CD25+ regulatory T cells.
Blood
107
:
1018
-1023.
6
Zeiser, R., D. B. Leveson-Gower, E. A. Zambricki, N. Kambham, A. Beilhack, J. Loh, J. Z. Hou, R. S. Negrin.
2007
. Differential impact of mammalian target of rapamycin inhibition on CD4+CD25+Foxp3+ regulatory T cells as compared with conventional CD4+ T cells.
Blood
111
:
453
-462.
7
Strauss, L., T. L. Whiteside, A. Knights, C. Bergmann, A. Knuth, A. Zippelius.
2007
. Selective survival of naturally occurring human CD4+CD25+Foxp3+ regulatory T cells cultured with rapamycin.
J. Immunol.
178
:
320
-329.
8
Valmori, D., V. Tosello, N. E. Souleimanian, E. Godefroy, L. Scotto, Y. Wang, M. Ayyoub.
2006
. Rapamycin-mediated enrichment of T cells with regulatory activity in stimulated CD4+ T cell cultures is not due to the selective expansion of naturally occurring regulatory T cells but to the induction of regulatory functions in conventional CD4+ T cells.
J. Immunol.
177
:
944
-949.
9
Allen, J. D., E. Verhoeven, J. Domen, M. van der Valk, A. Berns.
1997
. Pim 2 transgene induces lymphoid tumors, exhibiting potent synergy with c-myc.
Oncogene
15
:
1133
-1141.
10
van der Lugt, N. M., J. Domen, E. Verhoeven, K. Linders, H. van der Leuven, J. Allen, A. Berns.
1995
. Proviral tagging in Eu-myc transgenic mice lacking the Pim-1 proto-oncogene leads to compensatory activation of Pim-2.
EMBO J.
14
:
2536
-2544.
11
White, E..
2003
. The pims and outs of survival signaling: role for the Pim-2 protein kinase in the suppression of apoptosis by cytokines.
Genes Dev.
17
:
1813
-1816.
12
Fox, C. J., P. S. Hammerman, R. M. Cinalli, S. R. Master, L. A. Chodosh, C. B. Thompson.
2003
. The serine/threonine kinase Pim-2 is a transcriptionally regulated apoptotic inhibitor.
Genes Dev.
17
:
1841
-1854.
13
Fox, C. J., P. S. Hammerman, C. B. Thompson.
2005
. The Pim kinases control rapamycin-resistant T cell survival and activation.
J. Exp. Med.
201
:
259
-266.
14
Riley, J. L., P. J. Blair, J. T. Musser, R. Abe, K. Tezuka, T. Tsuji, C. H. June.
2001
. ICOS costimulation requires IL-2 and can be prevented by CTLA-4 engagement.
J. Immunol.
166
:
4943
-4948.
15
Chemnitz, J. M., A. R. Lanfranco, I. Braunstein, J. L. Riley.
2006
. B and T lymphocyte attenuator-mediated signal transduction provides a potent inhibitory signal to primary human CD4 T Cells that can be initiated by multiple phosphotyrosine motifs.
J. Immunol.
176
:
6603
-6614.
16
Suhoski, M. M., T. N. Golovina, N. A. Aqui, V. C. Tai, A. Varela-Rohena, M. C. Milone, R. G. Carroll, J. L. Riley, C. H. June.
2007
. Engineering artificial antigen-presenting cells to express a diverse array of co-stimulatory molecules.
Mol. Ther.
15
:
981
-988.
17
Aho, T. L. T., R. J. Lund, E. K. Ylikoski, S. Matikainen, R. Lahesmaa, P. J. Koskinen.
2005
. Expression of human pim family genes is selectively up-regulated by cytokines promoting T helper type 1, but not T helper type 2, cell differentiation.
Immunology
116
:
82
-88.
18
Learn, C. A., P. E. Fecci, R. J. Schmittling, W. Xie, I. Karikari, D. A. Mitchell, G. E. Archer, Z. Wei, H. Dressman, J. H. Sampson.
2006
. Profiling of CD4+, CD8+, and CD4+CD25+CD45RO+FoxP3+ T cells in patients with malignant glioma reveals differential expression of the immunologic transcriptome compared with T cells from healthy volunteers.
Clin. Cancer Res.
12
:
7306
-7315.
19
Sugimoto, N., T. Oida, K. Hirota, K. Nakamura, T. Nomura, T. Uchiyama, S. Sakaguchi.
2006
. Foxp3-dependent and -independent molecules specific for CD25+CD4+ natural regulatory T cells revealed by DNA microarray analysis.
Int. Immunol.
18
:
1197
-1209.
20
Zheng, Y., S. Z. Josefowicz, A. Kas, T. T. Chu, M. A. Gavin, A. Y. Rudensky.
2007
. Genome-wide analysis of Foxp3 target genes in developing and mature regulatory T cells.
Nature
445
:
936
-940.
21
Allan, S. E., L. Passerini, R. Bacchetta, N. Crellin, M. Dai, P. C. Orban, S. F. Ziegler, M. G. Roncarolo, M. K. Levings.
2005
. The role of 2 FOXP3 isoforms in the generation of human CD4+ Tregs.
J. Clin. Invest.
115
:
3276
-3284.
22
Ziegler, S. F..
2006
. FOXP3: of mice and men.
Annu. Rev. Immunol.
24
:
209
-226.
23
Wu, Y., M. Borde, V. Heissmeyer, M. Feuerer, A. D. Lapan, J. C. Stroud, D. L. Bates, L. Guo, A. Han, S. F. Ziegler, et al
2006
. FOXP3 controls regulatory T cell function through cooperation with NFAT.
Cell
126
:
375
-387.
24
Ono, M., H. Yaguchi, N. Ohkura, I. Kitabayashi, Y. Nagamura, T. Nomura, Y. Miyachi, T. Tsukada, S. Sakaguchi.
2007
. Foxp3 controls regulatory T-cell function by interacting with AML1/Runx1.
Nature
446
:
685
-689.
25
Mizuki, M., J. Schwable, C. Steur, C. Choudhary, S. Agrawal, B. Sargin, B. Steffen, I. Matsumura, Y. Kanakura, F. D. Bohmer, et al
2003
. Suppression of myeloid transcription factors and induction of STAT response genes by AML-specific Flt3 mutations.
Blood
101
:
3164
-3173.