Transcriptional regulator early growth response gene 2 (Egr2) is required for T cell anergy in vitro and in vivo

In the search for transcription factors that regulate DGK-α expression during anergy induction, we identified an Egr binding site located between −909 and −901 bp of mouse DGK-α transcriptional start site, and a corresponding binding site was found in the homologous region of human sequence (Fig. 1 A). This was of interest because our previous gene expression profiling analyses on anergic T cells identified Egr2 as a highly up-regulated transcriptional factor in the anergic state (Zha et al., 2006). Kinetic analysis upon anergy induction revealed that Egr2 mRNA peaked as early as 1–3 h after anti-CD3 mAb stimulation (Fig. 1 B). In comparison, DGK-α transcription occurred later, coinciding with the reported onset of hyporesponsiveness (Gajewski et al., 1995). The expression of both Egr2 and DGK-α was substantially lower in T cells fully activated with anti-CD3 + anti-CD28 mAbs (unpublished data). These observations suggest that Egr2 could be involved in the transcription of DGK-α in the anergic state.

To determine whether Egr2 could induce the activity of DGK-α promoter, we cloned mouse DGK-α promoter region from −1054 to 192 bp, which contains the putative Egr binding site (−909 to −901 bp), into a luciferase reporter vector and cotransfected this with an Egr2-expressing vector into Jurkat cells. A control reporter (−54 to 190 bp), lacking the putative Egr binding site, and an empty expression vector (empty vector [EV]) were included as controls. As seen in Fig. 1 C, the luciferase activity of the DGK-α reporter increased in the presence of Egr2, suggesting that Egr2 is capable of driving DGK-α transcription. Next, we performed a ChIP assay, and the particular DGK-α promoter region was coprecipitated with an anti-Egr2 Ab in anergic cells (Fig. 1 D). Thus, upon anergy induction Egr2 associates with the DGK-α promoter and can directly regulate its expression.

To determine the necessity of Egr2 in the expression of DGK-α upon anergy induction, we developed a system to be able to delete Egr2 conditionally from Th1 clones, to be able to use a monoclonal population of T cells, and to use the Th1 cellular context, which is the best characterized model for the anergy process (Schwartz, 2003). To achieve targeted gene deletion, we used a mouse that expresses Coxsackie/adenovirus receptor as a transgene (CAR Tg) in T cell lineage, along with a Cre-expressing adenovirus (Wan et al., 2000). Egr2^flox/flox mice were interbred with CAR Tg mice to generate CAR Tg × Egr2^flox/flox mice, from which OVA-specific Th1 cell clones were generated (Fitch et al., 2006). To delete Egr2, the Th1 T cell clones were transduced with the Cre adenovirus. As assessed by qRT-PCR and immunoblotting, the Egr2 expression normally induced during anergy induction by anti-CD3 mAb was reduced to minimal levels (Fig. 2, A and B).

Cre adenovirus-transduced T cell clones were anergized by immobilized anti-CD3 mAb. As shown in Fig. 2 (C and D), deletion of Egr2 abrogated detectable binding to DGK-α promoter, which, as expected, was associated with markedly reduced levels of DGK-α transcripts. Therefore, Egr2 is necessary for DGK-α up-regulation under anergizing conditions.

To determine whether Egr2 was essential for the functional characteristics of T cell anergy, CAR Tg × Egr2^flox/flox T cell clones were transduced with the EV or Cre adenovirus and left untreated or anergized with immobilized anti-CD3 mAb. Upon rechallenge with anti-CD3 + anti-CD28 mAbs, the EV-transduced cells showed diminished IL-2 production and proliferation consistent with the hyporesponsive state of anergy (Fig. 3, A and B). However, with Egr2 deletion, the T cells remained active with nearly normal IL-2 secretion and proliferation. Of note, although the nonanergic control Cre-transduced cells produced a modestly increased amount of IL-2 upon anti-CD3 + anti-CD28 mAb stimulation compared with the control EV-transduced cells, the difference between the anergic Cre- and EV-transduced cells was more than sevenfold. Thus the restored IL-2 production found in the anergic Cre-transduced cells was not simply a result of hyperactivity; rather, Egr2 deletion rendered the cells less susceptive to anergy induction. Similar functional results were observed with additional independently derived CAR Tg × Egr2^flox/flox Th1 clones (unpublished data). In agreement with these functional results, the blunted ERK phosphorylation normally observed in anergic cells was restored with Egr2 deletion (Fig. 3 C). Quantitative densitometry averaged over three independent experiments confirmed the statistical significance of this difference (P < 0.05).

To further confirm the functional relationship between Egr2 and DGK-α, we reintroduced DGK-α into the Egr2-deleted T cells by transduction with a DGK-α–expressing adenovirus (Fig. 3 D). In fact, directed expression of DGK-α dominated over Egr2 deletion, resulting in reinhibition of IL-2 production and proliferation (Fig. 3, E and F). Thus, DGK-α is operational downstream of Egr2 and is one of the critical anergy-associated genes that contribute to T cell dysfunction.

We chose a well established model of superantigen-induced anergy to examine whether the Egr2 pathway was important for anergy induction in vivo (Kawabe and Ochi, 1990; Rellahan et al., 1990). Two methods were used. In the first, CD4-Cre × Egr2^flox/flox mice were generated by interbreeding CD4-Cre Tg mice and Egr2^flox/flox mice, and the lack of Egr2 expression in T cells from these mice was confirmed by immunoblot (Fig. 4 A). CD4-Cre × Egr2^flox/flox mice and control mice were intraperitoneally injected with staphylococcal enterotoxin B (SEB) or PBS. 7 d later, splenic T cells were isolated, comparable numbers of Vβ8⁺ T cells were restimulated with SEB ex vivo, and IL-2 production was measured. Consistent with the superantigen-induced anergy model, T cells from SEB-treated control mice produced significantly less IL-2 than those obtained from PBS-treated counterparts (Fig. 4 B). In contrast, there was no reduction of IL-2 by T cells from conditionally Egr2-deficient mice treated with SEB, indicating that Egr2-deficient T cells were resistant to superantigen-induced anergy in vivo. This effect was not simply an artifact of Cre expression in T cells because CD4-Cre Tg mice remained susceptible to SEB-induced anergy induction (unpublished data).

Because of a concern that Egr2 deletion, even late in thymic development, using Cre driven by the CD4 promoter might yield an abnormal peripheral T cell compartment, a second method was used. T cells from CAR Tg × Egr2^flox/flox mice were purified, and Egr2 was deleted in vitro using the Cre-expressing adenovirus. After verification of Egr2 deletion by intracellular flow cytometry (Fig. 4 C), T cells were adoptively transferred into OT-1 TCR Tg × Rag2^−/− mice. The rationale for using these recipients is to block homeostatic proliferation of the transferred T cells, which itself could interrupt or reverse anergy but without inducing immune-mediated regression of CAR-expressing T cells (Brown et al., 2006; Kline et al., 2008). In addition, the β chain of OT-1 TCR is Vβ5, which does not respond to SEB superantigen (Vβ8). 1 d after adoptive transfer, recipient mice were injected intraperitoneally with SEB, and 7 d later T cells were isolated and analyzed as described in the previous section. As with the CD4-Cre model, Cre-transduced T cells did not show reduced IL-2 production after SEB treatment in vivo (Fig. 4 D), confirming the importance of Egr2 for anergy induction in this model.

To further examine the functional role for Egr2 in vivo, CD4-Cre × Egr2^flox/flox mice and control mice matched for age and sex were injected subcutaneously with B16.SIY (SIY-expressing B16 melanoma cells). This is a tumor model system in which anergy appears to be one mechanism of immune escape in vivo (Kline et al., 2008). As shown in Fig. 4 E, tumor growth was significantly slowed in Egr2-deleted mice. Although the frequency of SIY-specific T cells in CD4-Cre × Egr2^flox/flox and control mice were similar (Fig. 4 F), Egr2 deletion led to markedly augmented SIY-specific IFN-γ production ex vivo assessed by ELISPOT (Fig. 4 G). These results argue for increased functional capacity of the endogenously activated anti-tumor T cells and indicate that T cell–intrinsic Egr2 expression can contribute to tumor escape from immune destruction in vivo. We did not observe any obvious evidence of spontaneous autoimmunity in CD4-Cre × Egr2^flox/flox mice up to 12 mo of age (unpublished data).

In the present study, we demonstrate that Egr2 directly regulates the expression of DGK-α in anergic T cells. Furthermore, conditional deletion of Egr2 in T cells abolished anergy-induced DGK-α expression indicating the necessity of Egr2 in this process. In contrast, transduction of DGK-α back into the Egr2-deleted T cells resulted in decreased IL-2 production and T cell proliferation. These data suggest that DGK-α acts downstream of Egr2 and is one of the critical anergy-associated genes responsible for maintaining the anergic state. Besides DGK-α, multiple anergy-associated genes have been identified previously including DGK-ζ, Cbl-b, Itch, GRAIL, Tob1, and Dtx1, which also contribute to T cell–intrinsic hyporesponsiveness and the regulation of peripheral tolerance. In fact, using ChIP assay, we found that Egr2 bound the regulatory regions of these additional anergy-associated genes (unpublished data), and conditional deletion of Egr2 prevented their up-regulation as determined by qRT-PCR (the fold increase upon anergy induction of these genes was compared in EV and Cre adenovirus-infected CAR Tg × Egr2^flox/flox T cells, and the results are DGK-ζ: 5.64 ± 0.08 vs. 2.19 ± 0.01; Cbl-b: 5.62 ± 0.05 vs. 1.65 ± 0.01; Itch: 6.54 ± 0.47 vs. 3.41 ± 0.18; Tob1: 2.56 ± 0.11 vs. 1.07 ± 0.01; Dtx1: 2.80 ± 1.16 vs. 0.73 ± 0.65; all P values <0.05). We have also performed gene expression profiling of control versus Egr2-deleted T cells and a genome-wide ChIP-SEQ analysis, which have identified 46 genes that represent the identifiable Egr2 transcriptome in T cell anergy (unpublished data). The functional importance of several of these new candidates is currently being characterized. Thus, Egr2 is a major transcriptional regulator of the full anergy program.

T cell anergy can be viewed as a differentiation state of activated T cells, and previous studies had suggested that anergic T cells are not inert but can have additional functional roles, such as the ability to function as suppressor cells (Chai et al., 1999). The identification of transcriptional regulators for other T cell differentiation states (e.g., Tbet for Th1, GATA-3 for Th2, RORγT for Th17, etc.) has provided tremendous information about the biology of those subsets in immune models in vivo. The identification of Egr2 as a central factor for anergic T cells should similarly place studies of the anergic state on firmer footing for in vivo studies, and pursuit of additional gene targets of Egr2 might identify molecules that could be used as markers for anergic T cells for ex vivo identification.

Egr2 is the first molecule in our hands which, when deleted from or blocked in peripheral T cells, results in anergy resistance in anti-CD3 induced in vitro anergy model as well as in the superantigen SEB-induced in vivo anergy models. Increased activities of protein tyrosine kinase Fyn, and Rap1 GTPase, which, driven by guanine nucleotide exchange factor C3G bound to adapter protein CrkL, have been described in anergic T cells (Gajewski et al., 1994; Boussiotis et al., 1997). However, Fyn- and CrkL-deficient T cells remained subject to anergy induction, arguing that those pathways are dispensable (Fields et al., 1996a; Zha et al., 2006). Primary T cells from Cbl-b knockout and PTEN conditional knockout mice had been reported to be hyperresponsive and relatively anergy resistant (Bachmaier et al., 2000; Suzuki et al., 2001). However, introduction of a dominant-negative form of Cbl-b, or conditional deletion of PTEN directly in the peripheral T cells, led to hyperactivation but intact anergy susceptibility (Zha and Gajewski, 2007; unpublished data). Thus, not every genetic manipulation of peripheral T cells that can lead to increased activation results in anergy resistance, and these observations together argue more strongly for the particular importance of the Egr2 pathway.

In the anergy models using immobilized anti-CD3 or SEB, Egr2 deletion prevented anergy induction. However, Ramón et al. (2010) recently reported that CD4-Cre × Egr2^flox/flox mice did not demonstrate increased immune responses against minor histocompatibility antigens, Toxoplasma gondii infection, or lymphocytic choriomeningitis virus. These data suggest that anergy is not a dominant mechanism of immune dysfunction in those models, and support the notion that Egr2 is not simply a global negative regulator of T cell activation. It will be of interest to examined the role of Egr2 and its target genes in other in vivo models in which anergy is thought to contribute to peripheral tolerance, such as costimulation blockade-induced immune suppression for solid organ transplantation (Alegre and Najafian, 2006).

Egr2^flox/flox and CD4-Cre Tg mice were gifts from H. Singh and F. Gounari (University of Chicago, Chicago, IL), respectively. CAR Tg mice expressing the extracellular domain of CAR under control of a Lck promoter/CD2 enhancer were generated as previously described (Wan et al., 2000). All three mouse strains have been backcrossed with C57BL/6 mice for more than eight generations. C57BL/6 mice were purchased from Taconic. All mice were housed in pathogen-free conditions at the University of Chicago, and all animal protocols were approved by the Institutional Animal Care and Use Committee. To generate CAR Tg × Egr2^flox/flox Th1 clones, CAR Tg × Egr2^flox/flox mice were immunized in the hind footpads with chicken OVA (Sigma-Aldrich) emulsified in complete Freund’s adjuvant (Sigma-Aldrich). 7 d later, the draining lymph nodes were harvested, and CD4⁺ Th1 cell clones were derived and maintained as previously described (Gajewski and Fitch, 1990; Fitch et al., 2006).

The mouse DGK-α promoter region from −1054 to 192 bp containing a putative Egr2 binding site (−909 to −901 bp) was amplified from genomic DNA by PCR using the following primers: forward, 5′-ATCACTCGAGTTCCAGCTGTCAACGCTTCCTTCT-3′; and reverse, 5′-CTATGGTACCTCTCTCTCTGACCCTCACTGGACC-3′. As control, the region from −54 to 190 bp was amplified using the following primers: forward, 5′-ATCGACTCGAGCCAGCTGTCAACGCT-3′; and reverse, 5′-TCGTAGGTACCCAGTGTGTCGTCAGGA-3′. The fragments were then cloned into a pGL4luc2 vector (Promega) via KpnI and XhoI sites (underlined) to make DGK-α and control luciferase reporter vectors. A plasmid encoding Egr2 was a gift from J. Powell (Johns Hopkins School of Medicine, Baltimore, MD). Jurkat cells (12 × 10⁶ cells in 300 µl of RPMI1640 with 10% FBS) were cotransfected with 10 µg EV or Egr2-expressing vector, along with 10 µg control or DGK-α luciferase reporter vector by electroporation (300 V and 800 µF) using a Gene Pulser II electroporator (Bio-Rad Laboratories). 16 h later, the cells were lysed in Glo Lysis Buffer (Promega), and luciferase activity was measured by Bright-Glo Luciferase Assay System (Promega).

Cre- and DGK-α–expressing adenoviruses were generated as previously described (Zha et al., 2006, 2008). For T cell transduction, cells were suspended at the high density of 10 × 10⁶/ml in DMEM with 2% FBS, incubated with an EV or the Cre adenovirus at 37°C for 50 min, transferred to DMEM with 10% FBS, and cultured for another 16 h at low density of 10⁶/ml.

In vitro anergy was induced by treating cells overnight with 1 µg/ml of immobilized anti-CD3 mAb (145-2C11; BioXCell). The cells were then harvested, washed, and rested for 1–3 d before analysis or rechallenge. T cell rechallenge was performed using 1 µg/ml anti-CD3 + 1 µg/ml anti-CD28 (PV-1; BioXCell) mAbs, either immobilized to plastic plates or bound to beads (Dynabeads; Invitrogen). IL-2 production was assessed at 24 h by ELISA, and proliferation was measured at 48 h by [³H] thymidine incorporation.

In vivo anergy was induced by superantigen SEB using two different models. In the first model, CD4-Cre × Egr2^flox/flox or control mice were injected intraperitoneally with SEB (20 µg/mouse; Toxin Technology) or PBS. 7 d later, total T cells were purified from spleens using a Pan T cell Isolation kit (Miltenyi Biotec). The number of Vβ8⁺ cells was calculated by multiplying total number of T cells by percentage of Vβ8⁺ cells determined by flow cytometry. For rechallenge, 25 × 10³ Vβ8⁺ cells were stimulated with 10 µg/ml SEB and 250 × 10⁴ irradiated T cell–depleted splenocytes. IL-2 production was assessed at 48 h by ELISA. In the second model, total T cells from CAR Tg × Egr2^flox/flox mice were purified, transduced with the EV or Cre-expressing adenovirus, and cultured for 3 d with 1 µg/ml IL-7 (R&D Systems). 5 × 10⁶ cells were then adoptively transferred into OT-1 TCR Tg × Rag2^−/− mice. The next day, the mice were treated by SEB or PBS injection, and 7 d later the T cell collection and rechallenge were performed as described.

ChIP assays were conducted according to the manufacturer’s protocol (Millipore). In brief, 2.5 × 10⁶ cells were lysed in 500 µl SDS lysis buffer, and cellular DNA was sheared with six sets of 15-s pulses plus a 60-s rest using a Misonix Sonicator 3000 (Qsonica). For immunoprecipitation, 200 µl of cell lysate supernatant was diluted fivefold in ChIP dilution buffer, and an anti-Egr2 Ab was added at 1:100 (Covance). SYBR green qPCR was conducted using primers specific for DGK-α promoter (forward, 5′-CCCGCCCCAAACCCACGACTAACT-3′; reverse, 5′-TGCTCTCCCACTCCTTTCTATTCC-3′). Primers specific for GJA5 were used as controls (forward, 5′-ACCATGGAGGTGGCCTTCA-3′; reverse, 5′-CATGCAGGGTATCCAGGAAGA-3′).

Total RNA was purified using an RNeasy Mini kit (QIAGEN). RNA was reverse transcribed by M-MLV RT (Invitrogen). The primers and probes were purchased from IDT and Roche, and Applied Biosystems. qRT-PCR used primers and probes specific for Egr2 (forward, 5′-CTACCCGGTGGAAGACCTC-3′; reverse, 5′-AATGTTGATCATGCCATCTCC-3′) and DGK-α (forward, 5′-CTGGGCACTGGAAATGATCT-3′; reverse, 5′-AATCTTTCTCAAATTCTCACCTTCATA-3′). Relative RNA abundance was determined based on control 18S RNA (Hs99999901_s1; Applied Biosystems).

Equal numbers of T cells were resuspended in ice-cold lysis buffer containing 50 mM Tris, pH 7.6, 5 mM EDTA, 150 mM NaCl, 0.5% Triton X-100, 1 mM PMSF, 10 mM NaF, 1 mM Na₃VO₄, and 1× protease inhibitor mixture (Roche). After a 30-min incubation on ice, the cells were spun for 10 min at top speed at 4°C, and supernatant was collected. The cellular lysate was loaded into 10% Tris-HCL gels (Bio-Rad Laboratories), separated by SDS-PAGE, and transferred to PVDF membranes (Millipore). Proteins were detected using primary antibodies against Egr2 (1:1,000), phospho-ERK (1:1,000; Cell Signaling Technology), total ERK (1:1,000; Invitrogen), and Myc-tag (1:1,000; Cell Signaling Technology). Secondary antibodies were HRP-linked anti–mouse IgG or anti–rabbit IgG (1:3,000; GE Healthcare). Detection was performed using an ECL Detection kit (GE Healthcare).

Antibodies specific for CD3, CD4, CD8, and Vβ8 were purchased from BD, eBioscience, and BioLegend. For Egr2 intracellular staining, cells were permeated for 1 h (Foxp3 Buffer Set; eBioscience), stained with the anti-Egr2 rabbit polyclonal Ab (1:100) for 1 h at 25°C, followed by a secondary antibody Alexa Fluor 647 goat anti–rabbit (1:100; Invitrogen) for 1 h at 4°C.

B16. SIY tumor cells were washed three times with DPBS, and 2 × 10⁶ cells were injected subcutaneously in 100 µl DPBS at the flank. Tumor size was measured twice per week, and tumor area was calculated by multiplying the longest diameter and the shortest diameter of the tumor. 14 d later, the mice were sacrificed, and the percentage of SIY-specific CD8⁺ T cells was determined by SIY-K^b pentamer staining according to the manufacturer’s protocol (Proimmune). An irrelevant pentamer OVA-K^b was used as a negative control. To determine functional tumor antigen-specific immune responses by ELISPOT, splenocytes were plated at 1 × 10⁶ cells/well and stimulated overnight with medium alone or 80 nM SIY peptide. IFN-γ–producing cells were assessed using the manufacturer’s protocol (mouse IFN-γ ELISPOT kit; BD).

Data from independent groups were analyzed using a Student’s t test. Western blot data were analyzed using ANOVA after digital densitometry.

Table S1 shows ChIP Assay primers used to confirm Egr2 association. Table S2 shows qRT-PCR primers and probes used to confirm dependency on Egr2 for expression in anergy.

We thank Dr. Harinder Singh and Dr. Fotini Gounari (University of Chicago, Chicago, IL) for providing Egr2^flox/flox and CD4-Cre Tg mice; Dr. Jonathan Powell (Johns Hopkins School of Medicine, Baltimore, MD) for providing an Egr2 plasmid; Fenge Gao, Mihir Vohra, and Long Zhang for technical assistance; and Drs. Justin Kline, Maria-Luisa Alegre, Mercedes B. Fuertes, and Robbert M. Spaapen for critical review of the manuscript.

This work was supported by R01 AI080745 and R01 CA118153 from the National Institutes of Health.

The authors have no competing financial interests.