CD69 is a direct HIF-1α target gene in hypoxia as a mechanism enhancing expression on tumor-infiltrating T lymphocytes

ABSTRACT

CD69 is an early activation marker on the surface of T lymphocytes undergoing activation by cognate antigen. We observed intense expression of CD69 on tumor-infiltrating T-lymphocytes that reside in the hypoxic tumor microenvironment and hypothesized that CD69 could be, at least partially, under the control of the transcriptional hypoxia response. In line with this, human and mouse CD3-stimulated lymphocytes cultured under hypoxia (1% O₂) showed increased expression of CD69 at the protein and mRNA level. Consistent with these findings, mouse T lymphocytes that had recently undergone hypoxia in vivo, as denoted by pimonidazole staining, were more frequently CD69⁺ in the tumor and bone marrow hypoxic tissue compartments. We found evidence for HIF-1α involvement both when using T-lymphocytes from inducible HIF-1α^−/− mice and when observing tumor-infiltrating T-lymphocytes in mice whose T cells are HIF-1α^−/−. Direct pro-transcriptional activity of HIF-1α on a newly identified hypoxia response element (HRE) found in the human CD69 locus was demonstrated by ChIP experiments. These results uncover a connection between the HIF-1α oxygen-sensing pathway and CD69 immunobiology.

Introduction

The C-type lectin CD69 is the first membrane-attached glycoprotein induced upon T and NK cell activation.^1,2 For this reason, its expression is frequently used as a read out for recent T cell activation.³ The transcriptional control of CD69 has been studied in depth and gene expression is mainly driven by NF-κB, EGR, ATF/CREB and AP-1 binding to the promoter of both human and mouse CD69.^4-6

In spite of its rapid induction, the functions of CD69 are difficult to interpret. To begin with, the ligand or the ligands for CD69 have remained elusive for a long time and only recently CD69 has been reported to bind galectin-1⁷ and to S100A8/S100A9 both in a glycosylation-dependent manner.⁸ The outcome of such ligations is even less clear. When CD69 is perturbed by monoclonal antibodies, ligation can either enhance⁹ or inhibit T lymphocyte activation and proliferation⁹ but the downstream events have been only poorly clarified. Availability of CD69 knockout mice¹⁰ showed that the phenotype is quite unaltered baseline, although a regulatory role in the differentiation was unraveled by controlling Th17 differentiation in pathogenically relevant disease models.^11,12 Recently, CD69 has been found associated to the amino acid transporter LAT1-CD98, regulating Tryptophan uptake, AhR activation and IL-22 secretion in skin T lymphocytes, contributing to the development of psoriasis.¹³ In addition, a repressive role on NK function has been reported and exploited for antitumoral effects in RMA-S lymphoma-engrafted mice.¹⁴

A role in T cell trafficking was suggested by experiments showing that CD69 downregulates the chemotactic receptor S1P1 from the cell surface acting in cis.¹⁵ According to such a model, CD69 upregulation would prevent T cells from exiting lymph nodes in response to sphingosine-1-phospate (S1P) gradients to transiently ensure permanence in the lymphoid tissue during antigen-driven activation.¹⁶ Interestingly, tissue-resident memory T cells express CD69 and this S1P1-CD69 functional crosstalk can be involved in preventing recirculation of such a memory non-migratory T cell subset.^17,18 Other authors have reported that CD69 is involved in the generation and persistence of long-lasting memory T cells in the bone marrow microenvironment.¹⁹

Hypoxia is known to profoundly affect the physiology of cells of the immune system through the heterodimeric HIF(α/β) transcription factors.^20,21 The HIF-1α and HIF-2α system mainly senses hypoxia by means of control of their post-translational degradation. Prolyl hydroxylases (PDH 1–3), functionally sensitive to availability of O₂ at physiologic levels, hydroxylate HIF-1α and HIF-2α and as a result these proteins are targeted for proteasomal degradation following a K48 poly-ubiquitination reaction catalyzed by von Hippel–Lindau tumor suppressor protein (VHL), an E3 ubiquitin ligase.^22-24 Therefore, in hypoxic conditions, HIFα cannot be hydroxylated resulting in the stabilization of the HIFα subunits. Furthermore, pharmacological inhibitors of PHDs, such as dimethyloxaloylglycine (DMOG), as well as Vhl or PHDs gene inactivation result in constitutive stabilization of the HIF transcription factors in normoxic conditions.^23,25-27 Moreover, HIF-1α activation can be also triggered by NF-κB-dependent activation of HIF-1α proximal promoter, which has been particularly relevant in both myeloid^28,29 and lymphoid cells.^30-32 An additional layer of modulation of HIF-1α levels in immune cells is controlled by metabolic changes in the microenvironment, including adenosine concentration ³³ and oxygen-free radicals.³⁴ Upregulation by HIF-1α has also been reported for OX40 and CD137 as activation-promoting cell surface markers expressed by T cells.^35,36 Such an effect of hypoxia explains more intense patterns of expression on the surface of tumor-infiltrating T-lymphocytes since tumors are hypoxic as a result of insufficient blood supply in a high-demanding metabolic tissue environment.³⁶

In this study, we report that CD69 is a direct transcriptional target of HIF-1α under physiologically relevant conditions. Hypoxia upregulates CD69 expression explaining its presence on tumor-infiltrating T lymphocytes, suggesting that the functions of CD69 could be involved in the adaptation of activated T and NK lymphocytes to low O₂ availability.

Results

Tumor-infiltrating T lymphocytes express CD69

The presence and function of tumor-infiltrating T lymphocytes has been found to be critical for the outcome of human neoplasia as originally reported in ovarian and colorectal cancer.^37,38 At least a fraction of tumor-infiltrating T lymphocytes is known to be recognizing tumor antigens³⁹ and is expected to undergo TCR-CD3-mediated activation.

Flow cytometry analyses of CD4⁺ and CD8⁺ T lymphocytes-infiltrating mouse tumors showed CD69 on their plasma membrane at higher levels than those obtained from spleen lymphocytes (Fig. 1). CD69 expression was maximal in MC38 colon carcinoma, although with variable intensity, in every case, the presence of CD69 was detectable (Fig. 1). We have previously reported that because of hypoxia, CD137 is upregulated on T lymphocytes.³⁶ The fact that CD137 and CD69 are frequently co-stained on tumor-infiltrating T lymphocytes (Fig. S1) prompted us to explore whether CD69 expression was also somehow connected to hypoxia.

Figure 1.

Tumor-infiltrating T lymphocytes in hypoxic tumors express surface CD69. (A, B) Tumors from mice engrafted with the indicated syngeneic tumors or developing spontaneous breast cancer in Her2/NeuT transgenic mice, were surgically excised when reaching a diameter above 10 mm, conditions that we previously reported to result in hypoxia. CD69 expression was assessed by flow cytometry on single-cell suspensions derived from the tumors upon gating of CD4⁺ and CD8⁺ T lymphocytes. MFIs (Mean of fluorescence intensity) are shown for individual mice (B) and representative FACS histograms are shown (A). n = 5 tumor bearing mice per experiment (Mean ± SEM).

Hypoxia upregulates CD69 in human and mouse T lymphocytes

To explore if hypoxia regulates CD69 expression, human peripheral blood mononuclear cells (PBMCs) and mouse splenocytes were cultured for 24 h and 48 h under normoxia conditions and in a 1% O₂ atmosphere. No significant upregulation of surface CD69 was observed (Fig. S2), a finding similar to that we also reported for CD137.³⁶

Next, we assessed if hypoxia can cooperate with a CD3-driven T cell activation to promote CD69 expression. As shown in Fig. 2, CD69 was upregulated at the surface protein and at the mRNA levels both in human T cells (Figs. 2 A and B) and mouse splenic T-cells when exposed to hypoxia during 48 h (Figs. 2C and D). Likewise, activated human CD4⁺FOXP3⁺ also increased CD69 protein expression when PBMCs were cultured under hypoxic conditions (Figs. S3A and B). Interestingly, hypoxia upregulated CD69 surface expression as a single stimulus in human purified NK cells (Fig. S2C).

Figure 2.

Hypoxia upregulates CD69 expression on T lymphocytes undergoing stimulation via TCR-CD3. Human PBMCs (A) or mouse splenocytes (B) were cultured for 48 h under normoxia or hypoxia while being activated with a plate-bound anti-CD3ε mAb. Histograms show a representative experiment out of six performed indicating the mean of intensity of CD69 immunofluorescence on gated CD4⁺ and CD8⁺ T cells. (C) and (D) represent a quantitative RT-PCR analysis of CD69 mRNA expressed in the corresponding human and mouse samples. Bar diagrams represent six cases each (Mean ± SEM). *p <0.05 (Mann–Whitney test).

Taken together, these results highlighted a link between hypoxia and CD69 expression. Hypoxia can potentially exert such a function through a variety of mechanisms. To gain further insight into the underlying mechanism, we substituted hypoxia by DMOG chemical inhibition of the prolyl hydroxylases (PHD 1–3),⁴⁰ that are critical in the control of HIF stability and function,⁴¹ thereby mimicking the effect of hypoxia. We observed that DMOG treatment replicated the findings observed under low O₂ conditions in human and mouse CD4⁺ and CD8⁺ T lymphocytes (Fig. 3) pointing to the involvement of the HIF transcriptional machinery in CD69 upregulation. It is of note that we were not able to substantiate functional consequences for the different levels of surface CD69 expression achieved in normoxia and hypoxia including in chemotaxis to sphingosine-1-phosphate and in galectin-1 binding assays (data not shown).

Figure 3.

DMOG upregulates CD69 mimicking hypoxia by inhibition of prolyl hydroxylases. Experiments were performed as in Fig. 2 but culturing the indicated human PBMCs (A) and mouse splenocytes (B) in the presence of DMOG to inhibit PHDs 1–3. Histograms show an experiment representative of three performed. (C, D) Quantitative RT-PCR assessment of the CD69 mRNA levels in the three cases (Mean ± SEM).

HIF-1α controls CD69 expression on T lymphocytes undergoing TCR-CD3 activation

To ascertain the involvement of HIF-1α in the regulation of CD69 expression,³⁶ we compared HIF-1α wild type versus HIF-1α-deleted splenocytes T cells from HIF-1α inducible knockout mice subjected to 1% O₂. Hif1a deletion prevented the surface upregulation of CD69 (Fig. 4A) as well as CD69 mRNA levels (Fig. 4B). The classical HIF-1α target gene PGK1 also lost its induction under hypoxia as a control (Fig. 4C).

Figure 4.

CD69 induction by hypoxia is curtailed in T lymphocytes from inducible HIF-1α^−/− T cells. Splenocytes from HIF-1α^floxed-UBC-Cre-ER or HIF-1α^WT-UBC-Cre-ER mice were cultured in the presence of 4-hydrotamoxifen for 48 h, subsequently washed and activated for 48 h with plate-bound CD3 mAb and soluble CD28 mAb under normoxia or hypoxia as indicated. (A) CD69 staining on gated CD4⁺ and CD8⁺ T lymphocytes in the resulting cultures is shown as FACS histograms representative of 7 independent experiments. Mean intensity of fluorescence is shown in each condition. (B, C) Analyses of CD69 mRNA expression by quantitative RT-PCR (B) and assessment of the classical HIF-1α target gene PGK1 in the same experiments as in (B) as a positive control (C). Mean ± SEM *p <0.05 (n = 7, Wilcoxon signed rank test).

One of the possible explanations of CD69 upregulation by HIF-1α would be direct binding of HIF-1α/HIF-1β to regulatory elements in the CD69 locus. Sequence analysis revealed that the human CD69 proximal promoter contains a potential HIF binding site at position −593 bp upstream of the point of transcription initiation (Fig. 5A). ChIP assays were performed by quantitative PCR in sequences flanking the candidate HRE in the nuclei of human T lymphocytes activated by anti-CD3ε mAb while cultured under normoxic or hypoxic conditions. Chip analyses to assess HIF-1α binding to the well-recognized HIF-1α responsive gene PDK1 was used as a positive control (Fig. 5B). Our results indicate that HIF-1α binds to the CD69 proximal promoter when T lymphocytes undergo hypoxia (Fig. 5B).

Figure 5.

HIF-1α binding to the human CD69 promoter is contingent on hypoxia. (A) Schematic representation of the human CD69 locus with a newly identified HRE candidate indicating the primers used for DNA amplification in the ChIP assays, its sequence and position at the proximal promoter. Amplicon generated by PCR is shown and the arrows indicate where the primers are localized. (B) ChIP assays using anti-HIF-1α mAb on genomic DNA from PBMCs activated with plate-bound CD3 for 12 h under hypoxia or normoxia. Results represent percentage with respect to total unprecipitated DNA input using primers flanking an HRE in the PDK1 promoter as a positive control and a region without HRE in the same gene as a negative control. The graph shows data from four independent experiments (Mean ± SEM). *p <0.05 (Mann–Whitney test).

Hypoxia correlates with CD69 expression in vivo

Tissues undergo hypoxia as a result of pathological and physiologic conditions. We have previously shown by pimonidazole-based F-MISO positron emission tomography (PET) that isograft murine tumor models are hypoxic.³⁶ To track hypoxic lymphocytes isolated from different tissue compartments in mice, we used pimonidazole staining followed by flow cytometry. B16-OVA melanoma-bearing mice were adoptively transferred with TCR-transgenic OT1 CD45.1⁺ T cells that recognize OVA as a surrogate tumor antigen, and pimonidazole staining was used to detect lymphocytes with a recent history of hypoxia.

Fig. 6A shows that transferred T lymphocytes that become CD69⁺ in vivo when reaching either the bone marrow or tumor tissue were more intensely stained by pimonidazole. Endogenous CD69⁺ CD8⁺ T lymphocytes showed brighter pimonidazole signals in the bone marrow but not always in the tumor microenvironment.

Figure 6.

Tissular hypoxia determines CD69 expression in vivo. (A) Mice-bearing B16-OVA subcutaneous tumors were adoptively transferred with total splenocytes from TCR-transgenic OT1 mice congenic for the CD45.1 allele. Forty-eight hours following adoptive transfer, mice were killed and cell suspension from the tumor and bone marrow were rapidly stained with anti-CD69 and pimonidazole. CD45.1⁺ and CD45.2⁺ CD8⁺-gated T cells were analyzed for CD69 surface expression. Dots represent individual mice (Mean ± SEM). *p <0.05 (Mann–Whitney test). (B) Five HIF-1α^floxed-dLck-Cre and five HIF-1α^WT-dLck-Cre mice were subcutaneously engrafted with Lewis lung carcinoma syngeneic tumors and killed when these reached over 10 mm in diameter. Cell suspensions from the tumors were retrieved and CD8⁺ T lymphocytes were assessed for CD69-specific immunofluorescence. B shows representative SSC/CD69 contour plots indicating percentage of positive CD69⁺ events and mean ± SEM of the five analyzed individual cases per condition. (C) CD69 assessment as in B representing CD69⁺ CD8⁺ T lymphocytes referred to total CD45⁺ leukocytes infiltrating the tumor. Results come from two pooled experiments and are represented as mean ± SEM, **p <0.01 (Unpaired test).

To further study the correlation of CD69 expression and pimonizole positive hypoxic areas in the B16-OVA tumor microenvironment, we performed multiplex immunofluorescence on frozen tissue sections that confirmed that the areas with pimonidazole staining contained CD45.1 OT-1 T lymphocytes recognizing OVA with greater levels of CD69 expression (Fig. S4). In these studies, we excluded intravascular lymphocytes.

To study the involvement of HIF-1α in the control of CD69 expression in the tumor microenvironment, we performed experiments using Lewis lung carcinoma-bearing mice whose loxP-flanked Hif1a locus is conditionally deleted by a cre recombinase expressed under the control of the distal lck promoter (dLck-Cre) thereby targeting all T cell lineages. When retrieved from tumors, HIF-1α^−/− T cells showed weaker and less frequent CD69 staining than WT counterparts (Figs. 6B and C). These results show that T lymphocytes which have experienced and sensed hypoxia in their recent past upregulate CD69 in vivo in an HIF-1α-dependent fashion.

Discussion

The functions of T lymphocytes are executed in tissue locations with variable and usually low availability of O₂ and the immune response is known to be strongly influenced by hypoxia, not only acting on T lymphocytes but also on myeloid leukocytes ⁴² and B cells.⁴³ In this complex scenario, our study highlights that CD69, an early C-lectin receptor, considered a hallmark of T cell activation upon antigen recognition, is co-regulated by HIF-1α sensing hypoxia as a direct transcriptional target.

Our observations in mouse and human T lymphocytes argue in favor of hypoxia-mediated upregulation of CD69 being a conserved function likely to have an important role in physiology and pathology. Indeed, tissues such as the bone marrow are hypoxic under physiologic conditions.⁴⁴ In infectious diseases, autoimmunity and cancer the inflammatory infiltrate is very often under severe O₂ deprivation.⁴⁵ Our in vivo results clearly demonstrate that lymphocytes residing in hypoxic tissue or with a recent history of hypoxia, as denoted by pimonidazole staining, expressed higher molecular surface densities of CD69.

Our finding of CD69 upregulation by hypoxia was prompted by observations in tumor-infiltrating T lymphocytes which are very often CD69⁺ and become hypoxic to some extent when entering into the hypoxic tumor microenvironment. However, hypoxia by itself does not upregulate CD69, since in our hands concomitant activation via CD3-TCR is absolutely required. By contrast, NK cells apparently upregulate surface CD69 to some extent under hypoxia without requiring further exogenous. The need for CD3-TCR stimulation probably reflects the interplay of the various transcription factors jointly involved in the transcriptional regulation of the CD69 promoter.^4-6

The function/s of CD69 besides denoting recent T-cell activation is/are difficult to address. Reportedly, CD69 expression inhibits lymphocyte migration through a crosstalk with the S1P₁ receptor. According to this model, CD69⁺ lymphocytes fail to chemotactically respond to sphingosine-1-phosphate (S1P) gradients and thereby do not depart from lymphoid or inflamed tissues.^16,46,47 These findings are consistent with the expression of CD69 on CD103⁺ resident memory CD8⁺ T cells that remain in tissue without recirculation via afferent lymphatic vessels or the blood stream.⁴⁸ In our hands, recently activated T lymphocytes through CD3 stimulation are very poorly attracted by S1P in any circumstance, thus precluding comparative chemotaxis experiments of lymphocytes cultured under hypoxia or normoxia.

Another limitation to understanding the function of CD69 is that ligands for CD69 have not been widely studied. Recent evidence suggests specific glycosylation-dependent binding of CD69 to galectin-1 ⁷ but the functional outcome of such interaction has only been established for Th17 cells. In our hands, the enhancement of CD69 expression under hypoxia conditions does not result in stronger binding of soluble galectin-1 to the cell surface (data not shown).

Previous evidence in CD69KO mice suggests an involvement of CD69 in the differentiation of Th17 cells¹² with implications in animal models of autoimmunity.¹¹ Very recently, a role for CD69 in skin gamma/delta T cells and the pathogenesis of psoriasis in an IL-23-induced mouse model has been described.¹³ How hypoxia shapes these disease models will be an area to explore in the future. Interestingly, many tissues under self-inflicted damage by autoimmunity become hypoxic.⁴⁹ In the context of cancer and autoimmunity, the overall picture is that CD69 behaves as a checkpoint inhibitor that contributes to decreasing the intensity of local inflammation and tissue damage. Hence, our results suggest that tissue O₂ could play a role in such functions as a result of a mechanism of CD69 upregulation conserved between humans and rodents. This mechanism conceivably would mitigate tissue damage in areas of ischemia and hypoxia to favor T cell residence in stressed hypoxic tissues that are likely to be challenged by pathogens.

The control of T cell gene expression³⁵ and functions⁵⁰ by hypoxia is an entangled but very active area of research. It is becoming very clear that energy metabolism is a key feature regulated in T cells that dictates, among others, effector functions, apoptosis susceptibility and T cell memory differentiation.⁵⁰ The hypoxic HIF-mediated regulation of CD69 brings another piece to the puzzle of the adaptive response of T lymphocytes to hypoxia as observed in tumor-infiltrating T lymphocytes. Implications and consequences of CD69 control by the HIF system are therefore potentially far reaching for T cell immunology and immunotherapy.

Material and methods

Mice and cell lines

Female C57BL/6 and BALB/c wild-type mice (6–7 weeks old) were purchased from Harlan Laboratories. C57BL/6 Tg(TcraTcrb)1100Mjb(OT-1) x B6SJL-Ptprc^aPep3^b/boyJ(CD45.1) mice were bred in our laboratory. BALB/c MMTV-NeuT transgenic mice were purchased from Jackson Laboratories and bred in our facilities. Hif-1α^floxed-UBC-Cre-ER^T2 mice and their counterparts Hif-1α^WT-UBC-Cre-ER^T2 were generated by crossing B6.129-Hif-1α^tm3Rsjo/J mice (Jackson Laboratories, stock no. 007561) with Tg(UBC-Cre/ER^T2)1Ejb/J mice (Jackson Laboratories, stock no. 008085). Mice carrying Hif-α loxP-flanked alleles were crossed with dLck-Cre mice to obtain T cell-specific gene deletion. Mice were backcrossed over 10 generations to the C57Bl6 background. All animal procedures were approved and conducted under institutional ethics committee guidelines (study number 003/16).

Mouse CT26 colon carcinoma and RENCA renal carcinoma cell lines were obtained from American Type Culture Collection (purchased from ATCC year 2007 to generate a master bank in CIMA). MC38 cells were provided by Dr Karl E. Hellström (University of Washington, Seattle, WA) in September 1998. B16-OVA cells were a kind gift from Dr Lieping Chen (Yale University, New Haven, CT) in November 2001, these cell lines were authenticated by Idexx Radil (case 6592–2012). Cell lines were cultured in RPMI 1640 medium (Gibco) supplemented with 10% FBS (Sigma-Aldrich), 100 IU/mL penicillin and 100 ug/mL streptomycin (Gibco) and 5×10⁻⁵ M 2-mercaptoethanol (Gibco), with the exception of LLC that was cultured with high glucose DMEM (Gibco) supplemented with 10% FBS. B16OVA cultures were additionally supplemented with Geneticin (400 μg/mL, Gibco). The LLC cell line was purchased from ATCC in year 2015.

In vitro T-lymphocytes activation studies

Splenocytes obtained from C57Bl6 mice were activated with plate-bound anti-CD3ε (0.5 μg/mL, clone 145–2C11, Biolegend) and, when indicated, with soluble anti-CD28 (1 μg/mL, clone 37.51) at 2.5×10⁶ cells/mL. To silence HIF-1α gene, splenocytes from Hif-1α^floxed-UBC-Cre-ER^T2 mice and their corresponding controls were cultured with 5 μM 4-hydroxytamoxifen (Merck) for 48 h following activation under either 21% or 1% O₂ atmospheres (altitude 449 m over sea level).

PBMCs were obtained from healthy donors and isolated from total blood by Ficoll gradients. Afterwards, cells were stimulated in 12-well plates precoated with anti-CD3ε (1 μg/mL, clone OKT3) at 2.5×10⁶ cells/mL. Human NKs cells were isolated from peripheral blood by Ficoll gradients, following purification with NK Cell Isolation kit by negative selection in an automacs Device (Miltenyi Biotec).

In both mice and human experiments, cells were cultured for indicated times at 37°C in a 1% O₂ atmosphere in the H35 Hypoxystation (Don Whitley, West Yorkshire, UK) incubator to study hypoxia conditions. In hypoxic mimicking experiments, dimethyloxaloylglycine (DMOG, Enzo Life Sciences, NY, USA) was added to lymphocyte cultures at a concentration of 0.2 mM for 48 h.

In vivo studies

BALB/c, C57Bl6 or conditional HIF-1α^−/− mice were inoculated subcutaneously in the flank with 5×10⁵ tumor cells of the indicated origin in 50 μL of PBS. When tumor areas reached 100 mm², including the spontaneous breast carcinomas from Her2/NeuT, mice were killed and the spleen and tumor were excised to obtain splenocytes and tumor-infiltrating T lymphocytes. Isolated tumors were incubated with Collagenase-D and DNase-I (Roche) for 15 min at 37°C, followed by mechanical disaggregation and filtration in a 70-μm cell strainer (BD Falcon, BD Bioscience). Tumor-infiltrating lymphocytes were isolated from stromal cells in a 35% Percoll gradient.

For pimonidazole experiments, B16OVA-bearing mice for 12 d were transferred intravenously with splenocytes obtained from OT1CD45.1. Two days following lymphocyte transfer, mice were injected i.p. with the hypoxic marker pimonidazole hydrochloride (60 mg/kg, Hydroxyprobe-1 Plus Kit, Hydroxyprobe Inc.) and 4 h later, mice were killed to prepare single-cell suspensions from bone marrow and tumor. To detect lymphocytes that had undergone hypoxia, cells were fixed with Cytofix/Cytoperm (BD Bioscience), washed with PermWash (BD Bioscience) and incubated with a FITC-MAb1⁵¹ after a surface staining of CD69.

Flow cytometry and antibodies and reagents

After treating with FcR-Block (anti-CD16/32 clone 2.4G2; BD Biosciences PharMingen), mouse T cell suspensions were extracellularly stained with the following antibodies purchased from Biolegend: CD3-PEC7 or FITC (17A2), CD8-BV510 (53–6.7), CD4-BV421 (RMA-5), CD45.1-PerCPC5.5 (A20), CD45.2-APC (104) CD69 PE (H1.2F3), CD137-biotin (17B5), Syrian Hamster IgG Biotin (SHG-1) and Armenian Hamster IgG PE (HTK888) as isotype-matched negative control.

Cultured PBMCs were pretreated with Beriglobin and surface stained with the following antibodies: CD4-BV421 (RPA-T4), CD8-BV510 (5K1) purchased from BD Bioscience, CD3-FITC (UCHT-1), CD4-PerCPC5.5 (OKT4), CD69-PE (FN50), NKp46-APC (9E2), CD16-PB (3G8), FOXP3-AF647 (150D), CD25-APC (BC96) and mouse IgG1-PE and AF647 (MOPC-21) purchased from Biolegend.

Either Zombi NIRFixable viability kit (Biolegend) or 7AAD (Biolegend) were used as a live/dead marker. True-Nuclear™ Transcription Factor Buffer Set (Biolegend) was used for FOXP3 staining experiments. Cell acquisition was performed with FACSCanto II and Fortessa (BD Biosciences) and FlowJo (Treestar) software was used for data analysis and presentation.

RNA purification, reverse transcription and qRT-PCR assays

Total RNA was extracted from splenocytes or PBMCs using a Maxwell 16LEV simply RNA tissue kit (Promega). Reverse transcriptions were performed with M-MLV reverse transcriptase (Invitrogen). Quantitative RT-PCR (qRT-PCR) was performed with iQ SYBR green supermix in a CFX real-time PCR detection system (Biorad). PCRs included primers for mouse CD69 cDNA (fw: 5´- AGGCTTGTACGAGAAGTTGGA-3´, rev: 5´-AGTTCACCAGAATATCGCTTCAG-3´), mouse PGK1 cDNA (fw: 5´- GTTCCTATGAAGAACAACCAG-3´, rev: 5´- CATCTTTTCCCTTCCCTTCTTCC-3´) and human CD69 cDNA (fw: 5´-AAATCTGTGTCAGTGGATGC-3´, rev: 5´-TCATTCTTCTCATTCTTGGG -3´). Expression data were normalized by comparison with levels of RPLP0 (mouse and human: fw: 5´-AACATCTCCCCCTTCTCCT-3´, rev: 5´- GAAGGCCTTGACCTTTTCAG-3´). The expression of each transcript was represented according to this formula 2 ^{ΔCt (Ct RPLP0 - Ct CD69 or PGK1)}, where C_t corresponds to cycle number.

Chromatin Immunoprecipitation Assay (ChIP)

For ChIP assays, human T lymphocytes were activated by anti-CD3ε mAb and cultured under normoxic or hypoxic conditions for 12 h. Subsequently, cells were fixed with formaldehyde that was stopped by the addition of glycine. Cell pellets were resuspended in a membrane lysis buffer. Nuclei pellets from these lysates were resuspended in a SDS sonication buffer and were sonicated to shear the DNA under conditions established. Next, samples were diluted in a Triton dilution buffer and pre-cleared with protein G sepharose. An “input sample” was removed and stored from each sample, while the rest was immunoprecipitated with a rabbit polyclonal anti-HIF-1α antibody (Abcam, ab2185) or a rabbit normal IgG control antibody (Cell Signaling Technology, 2729). Immunocomplexes were recovered by addition of protein G sepharose to the samples that were then sequentially washed with several buffers and eluted with an elution buffer. DNA-protein crosslinking was reversed in the input and eluted samples and DNA was purified and resuspended in water. Immunoprecipitated DNA was quantified by PCR using the following primers: CD69 proximal promoter: forward 5´-CAAGCTTTCTGTTTCCTGCATTC-3´; reverse 5´-TCGCTTCTTCCCTGGTGACT-3´; PDK1 positive control: forward 5´-CGCGTTTGGATTCCGTG-3´; reverse 5´-CCAGTTATAATCTGCCTTCCCTATTATC-3´. PDK1 negative control: forward 5´-GTGGGATGGTATCGTGATGGT-3´; reverse 5´-TTTGGCCAACCTCCTTCCT-3´.

Immunofluorescence of frozen tissue

Immediately after euthanizing tumor-bearing mice preinjected with pimonidazole 3 h earlier, tumors were embedded in OCT (Sakura) and frozen on dry ice. 10-µm tissue sections were fixed in 2% PFA (Sigma) for 10 min, and washed three times with PBS. Primary antibodies were incubated in BSA (Sigma) 0.5% Triton X-100 (Thermo) 0.3% in PBS at 4°C overnight. After three washes with PBS, samples were stained one hour at RT with secondary antibodies. Tissue slices were washed three additional times, labeled during 5 min with Hoetsch 33258 (LifeTechnologies) and mounted with Diamond prolong anti fade mounting media (LifeTechnologies). Goat anti-mCD69 (R&D) and FITC anti Pimonizadole were used at a 1:100 dilution and anti-CD45.1 PE (clon A20, Biolegend) at 1:50. Donkey anti goat Alexa fluor 647 (LifeTechnologies) at 1:200 was used as secondary antibody. Images were acquired in a LSM 800 confocal microscope (Zeiss) equipped with a 40X Apocromat (N.A 1.4). For quantification, Z stacks tile scans of the whole tumor slice were taken. Individual stacks were then analyzed with Image J for fluorescence intensity in the different channels. Pimonidazole positive areas were manually gated based in fluorescence signal. Gamma and brightness correction was performed in the images with Imaris (Bitplane) software for better visualization.

Statistics

Prism software (GraphPad Software) was used to analyze statistical differences of CD69 mRNA and protein expression by applying the Mann–Whitney U test or the Wilcoxon paired test. p values <0.05 were considered significant.

Disclosure of potential conflicts of interest

IM is a consultant for Bristol Myers Squibb, Roche-Genentech, Lilly, Incyte, Merck Serono, Boehringer Ingelheim, Alligator, Bioncotech.

Acknowledgments

We are grateful to Eneko Elizalde for excellent animal facility maintenance and mouse husbandry. Critical reading and scientific discussion by Drs Roncal, Rouzaut, Pérez-Gracia and Rodriguez-Ruiz are much appreciated.

Authors will deeply miss the late Dr Manuel Ortiz de Landázuri as a dedicated scientist, as a friend and as a mentor. His enthusiasm and his insightful advices will remain with us forever.

Funding

This work was supported by the MINECO (SAF2011–22831, SAF2014–52361-R), Gobierno de Navarra (Salud), Fundación BBVA, Asociación Española contra el cáncer; Spanish Ministry of Economy and Competitiveness (SAF2013–46058-R), Comunidad de Madrid/Fondo Social Europeo (S2010/BMD-2542 “Consepoc-CM”) and Red de Cardiovascular (RD12/0042/0065); Spanish Ministry of Economy and Competitiveness (SAF 2014–55579-R), grant INDISNET-S2011/BMD-2332 from the Comunidad de Madrid; Red Cardiovascular RD 12–0042–0056 from Instituto Salud Carlos III (ISCIII), and ERC-2011-AdG294340-GENTRIS. Asis Palazon is a recipient of a Marie Curie IEF fellowship FP7.