Hypoxia controls CD4⁺CD25⁺ regulatory T-cell homeostasis via hypoxia-inducible factor-1α^†

The effect of hypoxia and HIF-1α overexpression on Foxp3 expression in Jurkat T cells

We first determined the effect of hypoxic conditions (1% O₂) and HIF-1α stabilization on the expression of Foxp3 in human Jurkat T cells. Incubation of Jurkat cells under hypoxia for 16 h resulted in a marked increase in HIF-1α activity (eight fold; p<0.005) as determined by ELISA (Fig. 1A). Hypoxia-induced HIF-1α increased activity was abrogated by retroviral infection with short-interfering RNAs (siRNA) directed against HIF-1α but not with control siRNA. Flow cytometry analysis revealed a significant increase in Foxp3 expression in hypoxic cultures, compared with normoxic controls (Fig. 1B). This effect was markedly reduced in the presence of siHIF-1α, clearly pointing at a HIF-1α-dependent effect on Foxp3 expression. We then assessed Foxp3 expression in Jurkat cells following retrovirus-mediated overexpression of HIF-1α. To circumvent von Hippel-Lindau-mediated ubiquitination of HIF-1α, we employed a mutant form of HIF-1α, HIF-1αP564A, in which mutation of proline 564 to alanine leads to enhanced HIF-1α stability 13. Jurkat cells were infected with different dilutions (1:0, 1:1, 1:2, 1:4) of pMX-HIF-1αP564A virus (∼6×10⁶ IU/mL). A dose-dependent increase in mRNA level of HIF-1α and Foxp3 was evident by semi-quantitative RT-PCR (Fig. 1C). The increase in Foxp3 mRNA following HIF-1α overexpression was further validated by quantitative real-time PCR (Fig. 1D). At the protein level, enhanced HIF-1α activity (7.94-fold; p<0.005) was validated by ELISA (Fig. 1E). Finally, FACS analysis confirmed a direct dose-dependent association between HIF-1α and Foxp3 expression (Fig. 1F). Altogether, these findings pointed at clear causal relationship between HIF-1α and Foxp3 expression.

The effect of hypoxia on the number and function of primary mouse and human Treg

To confirm the validity of the above findings in primary cells we tested the effects of hypoxia in human peripheral blood mononuclear cells (PBMC) and mouse splenocytes. Human PBMC as well as mouse splenocytes were subjected to different periods of hypoxia and analyzed by ELISA for HIF-1α DNA binding activity (Fig. 2A). Annexin V-PE/7-AAD staining showed no significant differences in apoptosis between normoxic and hypoxic cell groups (data not shown). A time-dependent increase in HIF-1α activity was demonstrated, showing a peak elevation after 16 h (6.41- and 5.32-fold in PBMC and splenocytes, respectively, p<0.005). Thus, human PBMC and mouse splenocytes were exposed for 16 h hypoxia, allowed to recover for 24 h, and analyzed by flow cytometry for CD4⁺CD25⁺ expression (Fig. 2B). The Foxp3⁺ population among CD4⁺CD25⁺ suppressor cells was found to increase from 30.56±6.3 to 48.90±4.9% in human cells (p<0.05) (Fig. 2C; upper panels). Likewise, in mouse splenocytes hypoxia treatment increased Foxp3⁺CD4⁺CD25⁺ T cells from 57.95±6.2 to 70.05±5.1% (p<0.05) (Fig. 2C; lower panels).

On the functional aspect, hypoxia-treated Treg more potently suppressed the proliferation of CD4⁺CD25⁻ responders as compared with normoxic CD4⁺CD25⁺ cells, in a dose-dependent manner (Fig. 2D).

The effect of in vivo overexpression of HIF-1α in mouse splenocytes on the levels of Foxp3⁺CD4⁺CD25⁺ Treg

We next explored the effect of HIF-1α on Treg in vivo. For this purpose, we employed the hydrodynamic naked DNA injection in order to induce HIF-1α overexpression in mouse splenocytes 14. A noticeable expression of human HIF-1α mRNA (Fig. 3A) and enhanced HIF-1α activity (Fig. 3B) were evident in the splenocytes 24 h after injection. Foxp3 mRNA levels were elevated in the HIF-1α-injected group 48 h after DNA administration (Fig. 3A). FACS analysis of spleen-derived Treg 72 h after injections showed significantly elevated levels of Foxp3⁺CD4⁺CD25⁺ cells in HIF-1α-injected mice (80.16±2.26%) compared with controls (68.98±2.25%) (p<0.005) (Fig. 3C). These overall findings support the assumption that HIF-1α activation contributes to an increase in the expression of Foxp3 as well as to an increase in the number of functionally active Foxp3-Treg.

Cell culture

Experiments using animals were done with the permission of the local Ethical Committee. Local Ethical Committee approval was received also for human studies and informed consent of all participating subjects was obtained. Male C57BL/6J mice were obtained from Harlan Laboratories. Human peripheral blood mononuclear cells PBMC were isolated from healthy subjects by Ficoll density-gradient centrifugation (Sigma-Aldrich) of 20 mL blood samples. Mouse spleens were explanted, mechanically minced and red cells were lysed using ammonium chloride solution. Jurkat T cells, human PBMC and mouse RBC-depleted splenocytes were cultured in complete RPMI medium supplemented with 10% FBS, 5% CO₂ at 37°C. For hypoxic preconditioning, cells were placed in a hypoxic chamber (Billups-Rothenberg) and a gas mixture of 94% N₂, 1% O₂ and 5% CO₂ was infused for 4 min at a rate of 2 psi. The chamber was placed in an incubator maintained for the indicated time courses at 37°C.

Retroviral transduction

pMX-eGFP constructs were kindly provided by Dr. I. Spiegel. pCDNA3-hHIF-1αP564A was a gift from Dr. M. Safran. The HIF-1α mutant form was inserted into the BamH1/NotI sites of pMX-eGFP, excluding eGFP, to yield pMX-hHIF-1αP564A. pMX-eGFP served as control vector. siRNA that correspond to the HIF-1α and GFP genes, cloned into pSuperRetro vector (OligoEngine, WA, USA), were kindly provided by Prof. R. Ratan 15. All plasmids were confirmed by DNA sequencing (Hylabs).

Retroviral plasmids were transfected into Phoenix packaging cells (ATCC) using standard Ca₂PO₄ protocol. Medium was changed after 24 h. Cells were grown for additional 24 h and virus collected and centrifuged 5 min/1500 rpm/4°C. Viral titer was determined as >6×10⁶ IU/mL using NIH 3T3 cells. Jurkat cells were incubated with viral supernatant for 6 h in the presence of 4 μg/mL polybrene (Sigma-Aldrich). Infection was repeated after 24 h.

Active HIF-α ELISA

Human/Mouse Active HIF-1α DuoSet® Intracellular ELISA (R&D) was used according to manufacturer's instructions. Briefly, a biotinylated oligonucleotide containing the consensus hypoxia response element HIF-1α binding site was incubated with nuclear extracts. HIF-1α-ds oligonucleotide complexes were captured by immobilized anti-HIF-1α antibody. Unlabeled oligonucleotide was used to test assay's specificity. Detection was performed utilizing Streptavidin-HRP and ELISA reader.

Reverse transcription-polymerase chain reaction

Total RNA was extracted from cultured 3−4×10⁶ cells by phenol/chloroform/isoamyl alcohol (Biological Industries). RNA was treated with DNase I (Ambion). RNA (2 μg/reaction) was transcribed using AMV-RT (Promega). Amplification reactions were carried out using REDTaq ReadyMix (Sigma-Aldrich). Reaction conditions were calibrated for each gene and primers. The following primers were used: hHIF-1α fw′-CCACCTATGACCTGCTTGGT, rev′-GGCCAGCAAAGTTAAAGCAT; mHIF-1α fw′-TGCTCATCAGTTGCCACTTC, rev′-CTTCCACGTTGCTGACTTGA; hFoxp3 fw′-CTCTTCTTCCTTGAACCCCA, rev′-CACTTGCAGACACCATTTGC; mFoxp3 fw′-GGCCCTTCTCCAGGACAGA, rev′-GCTGATCATGGCTGGGTTGT; h/mGA3PDH fw′-TCCACCACCCTGTTGCTGTA, rev′-ACCACAGTCCATGCCATCAC.

Real-time polymerase chain reaction

RNA was transcribed as described above. Foxp3 mRNA expression was quantified using ABsolute SYBER Green ROX mix (Thermo) according to manufacturer's instruction and using default cycling parameters on an ABI Prism 7200 Sequence Detector. The gene-expression results were expressed as arbitrary units relative to the expression of the gene encoding GA3PDH, as the mean±SD. Sequences for Foxp3 primers were the same used for semi-quantitative RT-PCR.

FACS analysis

After Ficoll (Sigma-Aldrich) separation, human PBMC were stained with combinations of the following monoclonal antibodies and corresponding isotype controls (eBioscience): APC- anti-human CD4, FITC-anti-human CD25. Intra-cytoplasmic Foxp3 staining was performed using the PE-anti-human Foxp3 staining kit according to the manufacturer's instructions. Mouse splenocytes were stained using the following antibodies (eBioscience): anti-mouse FITC-CD4, anti-mouse PE CD25, anti-mouse APC-Foxp3 staining kit and corresponding isotype controls.

In both cases, 8×10⁴ cells were acquired by flow cytometry (FACSCalibur, Becton Dickinson). To calculate Foxp3 percentage in CD4⁺CD25⁺ cells, the analysis was performed as follows (Fig. 2B): CD4⁺ PBMC or splenocytes were gated on CD25 marker. Then, CD4⁺CD25⁺ gated cells were presented as Side Scatter versus Foxp3 (Fig. 2C human, Fig. 2D mouse). Gating on Foxp3 was done with respect to isotype control. Jurkat T cells were stained with anti-human PE-Foxp3 staining kit and 1×10⁴ cells were acquired for flow cytometry.

Apoptosis in the different cell groups was assessed using SouthernBiotech ApoScreen Annexin V Apoptosis detection kit (Annexin V-PE, 7-AAD solution and Annexin V binding buffer).

Cell sorting

The CD4⁺CD25 (responders) and CD4⁺CD25⁺ (suppressors) populations were isolated from 10⁸ RBC-depleted splenocytes by FACS (FACSAria, Becton Dickenson) after FITC-anti-CD4 and PE-anti-CD25 staining for 30 min.

Functional suppression assay

Ninety-six well plates (CoStar, Corning) were coated with 1 μg/mL anti-CD3 monoclonal antibody (17A2, eBioscience) overnight at 4°C and washed. CD4⁺CD25⁻ and CD4⁺CD25⁺ T cells (10⁴ cells/well) were cultured in RPMI 10% FBS medium at different responder–suppressor ratios (1:1, 1:, 1: and 1:). All cells were cultured in a final volume of 200 μL and 2.5 μg/mL anti-CD28 monoclonal antibody (eBioscience). After 72 h, [³H]thymidine (1 μCi/well) was added for 16 h before proliferation was assayed by scintillation counting (beta counter). Percent inhibition of proliferation was determined from the following formula: 1−(median [³H]thymidine uptake of 1:1 CD4⁺CD25⁺:CD4⁺CD25⁻ coculture/median [³H]thymidine uptake of CD4⁺CD25⁻ cells).

Hydrodynamic naked DNA systemic injections and gene expression analysis

Male C57BL/6J mice (18–20gr) were injected via tail vein up to 5 s with 12 μg plasmid of pCDNA3 empty vector or pCDNA3-HIF-1αP564A in a volume of 1.6 mL of normal saline, as previously described 14. Animals were sacrificed at different time points (24, 48, 72 h; n=4 per group at each point) and spleens were harvested using standard surgical protocol. Splenocytes were then isolated and gene expression was assessed by RT-PCR. HIF-1α activity in splenocytes was evaluated by ELISA, as well as CD4, CD25 and Foxp3 expression was assessed by flow cytometry as described in section “FACS analysis”.

Statistical analysis

All comparisons were performed using a one-way ANOVA test. p<0.05 was accepted as statistically significant.