Hypoxia-Inducible Factor Regulates Survival of Antigen Receptor-Driven T Cells

Abs and reagents

mAb against human CD3 (UCHT1) was purchased from BD PharMingen (San Diego, CA). Anti-human Fas Ab (CH11) was from Medical & Biological Laboratories (Nagoya, Japan). Anti-human HIF-1α Ab (Ab463) was from Abcam (Cambridge, U.K.). Recombinant human IL-2 was obtained from PeproTech (London, U.K.) and PHA-M was from Sigma-Aldrich (St. Louis, MO). Synthetic peptides of human AM and human calcitonin gene-related peptide (CGRP) 8–37 were provided by Peptide Institute (Minoh, Japan). Other chemicals were purchased from Sigma-Aldrich unless specified.

Cell culture and activation

Fresh PBMC were prepared from heparinized blood of healthy volunteers by Ficoll-Paque Plus (Amersham Biosciences, Uppsala, Sweden) density gradient centrifugation and suspended in RPMI 1640 medium (Sigma-Aldrich) supplemented with 10% FCS and antibiotics. Adherent cells were removed by incubation on plastic dishes for 1 h at 37°C, and the rest of the cells were separated on nylon wool columns to obtain T cell-rich fraction. For activation, T cell blasts were generated by stimulation of the T cell fraction (1 × 10⁶ cells/ml) with 5 μg/ml PHA-M for 48 h. The mitogen was then washed out, and the cells were maintained in the medium with 10 ng/ml recombinant human IL-2. TCR/CD3 engagement of the cells (1 × 10⁶ cells/ml) was performed in the presence of 10 ng/ml IL-2 on 96-well plates coated with CD3 Ab (5 μg/ml). Jurkat T cells were obtained from American Type Culture Collection (Manassas, VA) and maintained in fully supplemented RPMI 1640 medium. A Jurkat T cell derivative stably integrated with cDNA encoding HIF-1α/1-396/VP16 AD chimeric protein, Jurkat-HIF-1α, was established by means of G418 selection. Exposure of the cells to various oxygen concentrations was carried out as described previously (14).

Determination of cell death and viability

The cells were harvested after culture under the indicated conditions and incubated with Annexin V^FITC (0.25 μg/ml) and propidium iodide (PI) (50 μg/ml) in binding buffer (10 mM HEPES, 140 mM sodium chloride, 2.5 mM calcium chloride) 30 min before flow cytometric analysis for detection of apoptotic/necrotic cells. Numbers and the ratio of the viable cells were determined by means of trypan blue dye exclusion and counting on a hemocytometer.

Immunoblot assay

Immunodetection of HIF-1α protein was performed as described previously (16). Briefly, whole cell extracts of T cells were prepared in lysis buffer (25 mM HEPES, 100 mM NaCl, 5 mM EDTA, 100 μM orthovanadate, and 0.5% Triton X-100, pH 7.9), followed by centrifugation for 30 min at 14,000 rpm. One hundred micrograms of whole cell extract were separated by SDS-PAGE and blotted onto polyvinylidine difluoride filters. The filters were incubated with anti HIF-1α Ab (Abcam) in TBS containing 1% nonfat milk at 4°C overnight, followed by anti-mouse Ig-HRP conjugate (Amersham Biosciences) in the same buffer. Immunocomplexes were visualized by ECL (Amersham Biosciences).

Immunohistochemistry

Samples of the synovial tissue were fixed overnight in 4% parafolmaldehyde at 4°C, dehydrated in alcohol, and embedded in paraffin. Four-micrometer sections were then deparaffinized and incubated with monoclonal anti-HIF-1α Ab and alkaline phosphatase-conjugated anti mouse Ig Ab, or polyclonal anti-human CD3 Ab (DAKO, Glostrup, Denmark) and histofine simple stain MAX PO (Nichirei, Tokyo, Japan). Nonimmunized mouse Ig was used as a negative control for the primary Ab.

RT-PCR analysis

Total RNA was isolated from T cells by guanidine isothiocyanate lysis/phenol chloroform extraction, followed by removal of contaminating DNA. First-strand cDNA was synthesized using 2 μg of DNase-treated total RNA as a template in 20 μl of reaction mixture containing 50 mM Tris-HCl (pH 8.3), 3 mM MgCl₂, 10 mM DTT, 75 mM KCl, 1 mM dNTPs, 0.1 mM oligo dT (25) primer, and 50 U Superscript II (Invitrogen, Carlsbad, CA) at 42°C for 50 min. PCR was carried out in a total volume of 30 μl in a mixture composed of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 200 μM dNTPs, 0.25 μM each of the sense and antisense primers, 1 U ExTaq DNA polymerase (TaKaRa, Ohtsu, Japan). The amount of cDNA, as judged by the intensity of the amplified β-actin signal, was comparable among the preparations. Amplification by 27 cycles of 94°C for 30 s, 50°C for 30 s, and 72°C for 1 min was performed after 3 min of denaturing of the samples at 94°C, and shown to be within linear range or nonsaturated conditions for β-actin amplification. Identities of the PCR products were confirmed by sequencing. PCR primer pairs for amplification of each gene are as following; Fas, sense: 5′-CCAAGTGACTGACATCAACTC-3′, antisense: 5′-ACTCTTTGCACTTGGTGTTGC-3′, Fas-L, sense: 5′-GGAATGGGAAGACACCTATG-3′, antisense: 5′-GCACTGGTAAGATTGAACAC-3′; bcl-2, sense: 5′-TGAACTGGGGGAGGATTGTG-3′, antisense: 5′-GCCAGGAGAAATCAAACAGA-3′; bcl-x_L sense: 5′-CCCAGAAAGGATACAGCTGG-3′, antisense; 5′-GCGATCCGACTCACCAATAC-3′; vascular endothelial growth factor (VEGF), sense: 5′-TGCCTTGCTGCTCTACCTCC-3′, antisense: 5′-TCACCGCCTCGGCTTGTCAC-3′; glucose transporter (GLUT)-1, sense: 5′-CTTTCTCCAGCCAGCAATGA-3′, antisense: 5′-TGGATCCTGAGTCGAAGTCT-3′, GLUT-3, sense: 5′-GATGCTGGAGAGGTTAAGGT-3′, antisense: 5′-ACTTCCACCCAGAGCAAAGT-3′; AM, sense: 5′-AAGAAGTGGAATAAGTGGGCT-3′, antisense: 5′-TGGCTTAGAAGACACCAGAGT-3′; calcitonin receptor-like receptor (CRLR), sense: 5′-GATGCTCTGTGAAGGCATTT-3′, antisense: 5′-CAGAATTGCTTGAACCTCTC-3′, receptor-activity-modifying protein (RAMP)1, sense: 5′-GAGACGCTGTGGTGTGACTG-3′, antisense: 5′-TCGGCTACTCTGGACTCCTG-3′; RAMP2, sense: 5′-GGACGGTGAAGAACTATGAG-3′, antisense: 5′-ATCATGGCCAGGAGTACATC-3′; β-actin, sense: 5′-CCTCGCCTTTGCCGATCC-3′, antisense: 5′-GGATCTTCATGAGGTAGTCAGTC-3′.

DNA microarray analysis

DNA microarrays (human genome U95A v2 array; Affymetrix, Santa Clara, CA) were hybridized with labeled cRNA derived from T cells with various treatments as recommended by the manufacturer. Differential expression, as analyzed by GeneChip 3.1 software (Affymetrix), was assessed by pairwise comparisons of samples from each treatment with untreated subject. The criterion for differential expression required an over 2-fold increase or decrease in mRNA expression levels.

Expression of HIF-1α in T cells accumulating in the interstitium of inflammatory tissue

We first examined expression of HIF-1α in T cells accumulating in a hypoxic inflammatory tissue. Immunohistochemical analysis demonstrated expression of HIF-1α in infiltrating CD3-positive T cells in the synovial tissue of a patient with rheumatoid arthritis (Fig. 1⇓). The largest portion of HIF-1α staining was found in the cytoplasmic compartment of T cells, and a nuclear signal was only occasionally detected (Fig. 1⇓B). The affected skin in a patient with dermatomyositis also revealed accumulation of CD3-positive T cells expressing HIF-1α (data not shown). Given these results, we hypothesized that HIF-1α might have a role in regulation of T cell-mediated immune reactions in the inflammatory tissues.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Expression of HIF-1α in CD3-positive T cells accumulating in inflammatory tissues. Sections from the synovium from a patient with rheumatoid arthritis were stained for expression of HIF-1α and CD3. A, Low magnification; B, high magnification. Double staining with the Abs against HIF-1α (blue staining by alkaline phosphatase/fast blue) and CD3 (brown staining by peroxidase/3,3′-diaminobenzidine tetrahydrochloride). Arrows indicate CD3-positive cells with cytoplasmic HIF-1α expression, and the arrowhead shows CD3-positive cells with nuclear signals of HIF-1α. Scale bars in A, and B represent 100 and 20 μm, respectively. C, Staining with anti-CD3 Ab and mouse IgG as negative control for a primary Ab (scale bar, 20 μm).

Hypoxia represses AICD of human peripheral T cells with concomitant expression of HIF-1α

To explore the potential role of HIF-1α in regulation of T cell function in the microenvironment of the tissue, we investigated the effect of hypoxia on AICD of peripheral T cells. PHA-activated human peripheral blood T cells were maintained with IL-2 supplementation, followed by incubation with an immobilized mAb against CD3 either under normoxic (21% O₂) or hypoxic (1% O₂) conditions. Staining with annexin V and PI revealed massive apoptotic cell death of T cells induced by TCR/CD3-mediated activation under normoxic conditions, indicating the induction of AICD. To our surprise, this TCR/CD3-triggered AICD of T cells was dramatically attenuated by exposure of the cells to hypoxia (Fig. 2⇓A). When the cells were incubated at various oxygen concentrations (ranging from 21 to 1% O₂), the survival rates of T cells after TCR/CD3 ligation were gradually increased as oxygen tension decreased, whereas survival of T cells without TCR/CD3 ligation showed proportional reduction (Fig. 2⇓B). This influence of the oxygen concentration on T cell viability became evident when the oxygen concentration was decreased to 7 or 5%, corresponding to that often encountered in the tissue interstitium. These results indicate that not only activation through the Ag receptor but also the oxygen concentration play critical roles in determination of peripheral T cell life.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Oxygen concentration regulates survival of TCR/CD3-stimulated T cell. A, Hypoxia and TCR/CD3 stimulation repress AICD in peripheral T cells. PHA-preactivated human peripheral T cells maintained in the medium supplemented with IL-2 (PHA-activated T cells) were cultured in the absence (−) or presence of immobilized CD3 Abs (5 μg/ml) either at 21% O₂ (CD3) or at 1% O₂ (CD3/hypoxia). Apoptotic cell death was detected by staining the cells with Annexin V^FITC and PI followed by flow cytometric analysis, and percentages of Annexin V^FITC and/or PI-positive cells are indicated. Three independent experiments were performed and representative results are shown. B, Role of oxygen concentration in T cell survival. PHA-activated T cells were incubated in the presence (▪) or absence (□) of CD3 Abs at 21, 10, 7, 5, 3, and 1% O₂, respectively, for 72 h, and cell viability was determined by the trypan blue-dye exclusion method. Cell counting was performed in triplicate, and means ± SD of percentages of the viable cells are shown. ∗, p < 0.01, compared to the cells without CD3 stimulation at 21% O₂; ∗∗, p < 0.01, compared to the cells with CD3 stimulation at 21% O₂. C, Hypoxia- and TCR/CD3 stimulation-dependent expression of HIF-1α protein in T cells. PHA-activated T cells were incubated in the absence (−) or presence of immobilized CD3 (CD3) or soluble Fas (Fas) Abs (100 ng/ml) at indicated oxygen concentrations, and expression of protein (top panel, filled arrowhead) and mRNA (middle panel, open arrow head) levels of HIF-1α and β-actin mRNA (bottom panel, ∗) were determined by immunoblot and RT-PCR, respectively. Numbers (kDa) at left side of the top panel depict the molecular masses of the reference proteins run in parallel. D, Role of hypoxia in HIF-1α protein expression in PMA/ionomycin-activated T cells. PHA-activated T cells were incubated in the presence (P/I) or absence (−) of PMA (10 ng/ml)/ionomycin (1 μM) at the indicated oxygen concentrations for 24 h and cell lysates were subjected to immunoblot analysis for HIF-1α expression (filled arrowhead). Open arrowheads indicate nonspecific bands. Numbers (kDa) depict the molecular masses of the reference proteins run in parallel. Representative results of triplicated experiments are shown. E, Oxygen concentration-dependent inhibition of PMA/ionomycin-induced AICD of T cells. PHA-activated T cells were incubated in the presence of PMA/ionomycin at 21, 10, 7, 5, 3, and 1% O₂ for 24 h, cell viability was determined by trypan blue-dye exclusion, and results of three independent assays were shown as described above. ∗, p < 0.01, compared to the cells at 21% O₂.

HIF-1α mRNA was constitutively expressed in human peripheral T cells (Fig. 2⇑C). In contrast, it was not possible to detect HIF-1α protein by immunoblot in the extracts from unstimulated T cells even if the cells had been exposed to hypoxia (1% O₂ (Fig. 2⇑C). However, Ab-mediated TCR/CD3 engagement resulted in dramatic accumulation of HIF-1α protein upon exposure of the cells to hypoxic conditions corresponding to tissue oxygen levels (below 7% O₂), indicating that expression of HIF-1α protein in peripheral T cells requires not only hypoxia but also additional signals following TCR/CD3-mediated stimulation (Fig. 2⇑C and data not shown). Because Ab-cross-link of Fas molecules did not induce HIF-1α protein accumulation either under normoxic or hypoxic conditions (Fig. 2⇑C), it may not be the cell death executing components but activation pathways after TCR/CD3 ligation that are involved in HIF-1α induction. Activation of T cells by a combination of PMA and a calcium ionophore ionomycin mimicked TCR/CD3 ligation with regard to cooperation with hypoxia in HIF-1α protein accumulation (Fig. 2⇑D). PMA/ionomycin-mediated AICD was gradually decreased along with the reduction of oxygen levels and expression of HIF-1α (Fig. 2⇑E). Taken together, we may conclude that HIF-1α plays an essential role in regulation of AICD of T cells in the hypoxic tissue microenvironment.

AM suppresses AICD of peripheral T cells via autocrine regulatory mechanism

To examine the mechanism of HIF-1α-mediated attenuation of AICD, we performed gene expression profiling experiments using DNA microarray analysis and RT-PCR assays of peripheral T cells cultured in the absence or presence of TCR/CD3 ligation and exposure to hypoxia. As shown in Table I⇓, a variety of changes in expression profile of T cell-related genes were induced by those treatments. Among them, expression of Fas mRNA was detected irrespective of cellular treatment. Fas-L mRNA expression was induced by TCR/CD3 ligation, whereas hypoxic culture had no effect on this induction response, suggesting that hypoxia does not likely interfere with TCR/CD3-mediated activation of T cells leading to Fas-L expression. mRNA levels of Bcl-2 remained constant, whereas Bcl-x_L mRNA expression was induced by TCR/CD3 activation but not by hypoxia. These classical apoptosis-regulating factors, thus, do not show any correlation with HIF-1α protein levels. Expression of mRNA for GLUT-1, -3, and VEGF were induced by hypoxia but not altered after TCR/CD3 stimulation. In contrast, hypoxia had only a modest effect on AM mRNA induction, and the combination of hypoxia and TCR/CD3 stimulation resulted in a robust enhancement of AM gene expression in T cells (Fig. 3⇓A and Table I⇓). These results may indicate a close correlation between AM expression and HIF-1α protein accumulation in T cells. It is therefore likely that AM may represent a HIF-1α target gene involved in determination of cell survival of activated T cells. In fact, RT-PCR analysis demonstrated mRNA expression of the components of the receptor complex for AM, CRLR and RAMP 1 and 2 (26) in T cells (Fig. 3⇓B). Blocking of the AM receptor by a fragment of the calcitonin gene-related peptide CGRP 8-37 abolished the effect of hypoxia on T cell survival (Fig. 3⇓C), strongly suggesting that AM receptor-mediated signal transduction is essential for hypoxia-dependent rescue of T cells from AICD. In excellent agreement with this model, incubation of T cells with synthetic AM peptides bypassed requirement of hypoxia for survival, and inhibited AICD under normoxic conditions (Fig. 3⇓D). Taken together, these results suggest that HIF-1α target gene product AM plays an important role in T cell survival counteracting AICD under hypoxic conditions possibly via an autocrine loop mechanism.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

AM mediates the effect of hypoxia to protect T cells from AICD through an autocrine regulatory loop. A, AM represents an HIF-1α target gene in TCR/CD3-activated T cells. PHA-activated T cells were incubated in the presence (CD3) or absence (−) of immobilized CD3 Abs (5 μg/ml) at 21 or 1% O₂ for 24 h, and mRNA expression of the indicated genes was examined by RT-PCR. Three independent experiments have shown almost identical results and representative results are shown. B, TCR/CD3-activated T cells express components of the AM receptor complex. T cells incubated as described above were examined for mRNA expression of the indicated genes by RT-PCR. C, Blockade of the AM receptor abrogates hypoxia-mediated life support of T cells. PHA-activated T cells were incubated with immobilized CD3 Abs and 10 μM CGRP8-37 at 21% O₂ (□) or 1% O₂ (▪) as indicated for 72 h, and cell viability was determined. The results of triplicated experiments are presented as described in Fig. 2⇑B. ∗, p < 0.05; ∗∗, p < 0.01. D, Synthetic AM peptides mimic the effect of hypoxia in protection of T cells against AICD. PHA-activated T cells with TCR/CD3 stimulation were cultured in the absence (□) or presence (▪) of the indicated concentrations of AM peptides for 72 h, and cell viability was determined. Experiments were repeated for three times. ∗, p < 0.01; ∗∗, p < 0.005.

Constitutively active HIF-1α induces AM expression under normoxic conditions to protect T cells from AICD

To directly monitor the involvement of HIF-1α mediated induction of AM expression in the regulation of AICD in T cells, we generated Jurkat T cells which demonstrate functional activity of HIF-1α even under normoxic conditions. To this end, we constructed a chimeric protein, HIF-1α/1-396/VP16AD, consisting of the N-terminal bHLH/PAS domain of HIF-1α and the constitutively active TAD of the viral transcription factor VP16 (Fig. 4⇓A). Transient expression of HIF-1α/1-396/VP16AD in Jurkat cells resulted in constitutive activation of HIF-1α regulated reporter gene expression (Fig. 4⇓B). Jurkat cells stably expressing HIF-1α/1-396/VP16AD, designated Jurkat-HIF-1α cells, also showed enhanced expression of HIF-1-regulated reporter gene activity at normoxia (Fig. 4⇓C). In a similar fashion, Jurkat-HIF-1α cells expressed not only AM but also CRLR and RAMP2 mRNA even under normoxic conditions (Fig. 4⇓D). In clear contrast to the parental Jurkat cells, Jurkat-HIF-1α cells showed resistance to TCR/CD3-triggered cell death at normoxia. Moreover, blocking of the AM receptor by CGRP 8–37 impaired the survival of Jurkat-HIF-1α cells (Fig. 4⇓E). In conclusion, sequential expression of HIF-1α and AM under conditions of hypoxia critically controls the viability of Ag receptor-activated T cells, participating in maintenance of T cell-mediated immune reactions in the hypoxic microenvironment.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Regulation of AM by HIF-1α mediates protection of T cells against AICD. A, Schematic representation of the primary structure of the wild-type HIF-1α and chimeric HIF-1α/1-396/VP16AD proteins. N-TAD, N-terminal TAD; VP16 AD, VP16 TAD. B, HIF-1α/1-396/VP16 AD is a constitutively active regulator of the hypoxia response element (HRE)-driven gene expression. HeLa cells were transfected with the expression plasmid for HIF-1α/1-396/VP16 AD and the HRE-luciferase reporter gene and incubated at either 21% (□) or 1% (▪) O₂ for 24 h, whereafter cellular luciferase activity was determined. Experiments were performed in triplicate and means ± SD of fold induction relative to the cells transfected with only the reporter gene at 21% O₂ are shown. C, Establishment of Jurkat T cell-derived cells expressing a constitutively active form of HIF-1α. Parental Jurkat T cells (Jurkat), mock-transfected Jurkat cells (Jurkat-mock), and Jurkat-HIF-1α cells stably expressing HIF-1α/1-396/VP16 AD were transfected with the HRE-luciferase reporter gene and cultured for 24 h at 21% O₂. Cellular luciferase activity was monitored and means ± SD of fold induction relative to parental cells are shown. D, Expression of AM and AM receptor in Jurkat-HIF-1α cells under normoxic conditions. Total RNA was prepared from the parental Jurkat T cells or Jurkat-HIF-1α cells, and RT-PCR analysis was performed to monitor mRNA expression of the indicated gene. A representative result is shown. E, Blocking of AM signals abrogates the survival of Jurkat T cells overexpressing the constitutively active HIF-1α. Parental Jurkat T cells and Jurkat-HIF-1α cells with (▪) or without (□) CD3 stimuli (5 μg/ml) were cultured in the presence or absence of 10 μM CGRP 8–37 at 21% O₂ for 48 h whereafter cell viability was examined. Results from three independent experiments are presented as described in Fig. 2⇑B. ∗, p < 0.05; ∗∗, p < 0.01.