Both naturally occurring and immunotherapy-induced tumor elimination relies on the coordinated function of different immune cells and a series of events referred to as the cancer-immunity cycle [1
]. Briefly, the antitumor immune response begins at the tumor microenvironment (TME), where dendritic cells (DCs) acquire tumor-associated antigens (TAA) and are activated by the recognition of activator signals arising from stressed or dying cancer cells [2
]. After this, DCs migrate to tumor-draining lymph nodes where they induce the differentiation and activation of naïve TAA-specific CD8+
T cells, via the productive engagement of T cell receptors (TCR) by major histocompatibility class I (pMHC-I) or class II (pMHC-II)-antigen peptide complexes, respectively, in conjunction with the recognition of appropriate signals sensed by T cells during the establishment of an immunological synapse [3
]. Finally, effector CD8+
cytotoxic T lymphocytes (CTL) traffic to and infiltrate the TME, where they kill target cancer cells through mechanisms elicited by the formation of cytotoxic immunological synapses and polarized release of perforin and granzymes-containing cytotoxic granules [4
]. Therefore, regardless of their nature, cancer immunotherapy success mainly relies on the elimination of tumor cells by the activities of DCs and CTLs, which are also strongly associated with good prognosis in human cancer patients [5
Unfortunately, cancer cells may develop numerous immuno-evasive mechanisms that allow them to resist natural or therapy-induced immune attacks. Through these mechanisms, tumor cells are capable of modulating themselves and their surroundings in order to promote their survival, growth and invasion, even under a persistent immune pressure. Indeed, stressors present in the TME, such as chronic hypoxia, play determinant roles in promoting tumor cell plasticity and heterogeneity, which finally leads into the acquisition of immune tolerance and tumor progression [7
]. For example, it has been shown that hypoxia increases the expression of the immune checkpoint programmed death-ligand 1 (PD-L1) on melanoma cells and tumor infiltrating immune cells, via the transcriptional activity of hypoxia-inducible factor-1α (HIF-1α), leading to lower levels of CTL activation and tumor elimination [8
]. Accordingly, tumor hypoxia has been associated with decreased antitumor immunity and poor clinical responses to programmed death-1 (PD-1) blockade therapy in melanoma patients [9
MicroRNAs (miRNAs) are 19–22 nucleotide long non-coding RNAs that regulate gene expression at a post-transcriptional level. The profile and levels of miRNAs are frequently altered in cancer cells, affecting tumorigenesis and progression, via the induction of sustained proliferation, resistance to cell death, acquisition of invasive phenotypes, and regulating anti-tumor immune responses [10
]. Alterations in miRNA profiles in cancer cells can be a consequence of deregulation of specific signaling pathways and/or alterations of their expression in response to TME factors, such as hypoxia [11
]. Some of these hypoxia-responsive miRNAs (HRM) have been implicated in cancer progression through numerous cancer cell autonomous mechanisms, including the acquisition of immune-resistance phenotypes. For example, miR-210, a signature of hypoxia in many cell types and a robust target of HIF-1α, decreases the susceptibility of lung cancer and melanoma cells to CTL-mediated lysis, through the down-regulation of different cancer cell genes [12
]. Importantly, cancer cell-expressed HRMs also can modulate immune cells of the TME via non-cell-autonomous mechanisms, including the direct mobilization of HRM in extracellular vesicles (EV) (mainly exosomes) from cancer to immune cells, and the subsequent gene targeting in the recipient cells [13
]. For instance, it has been shown that hypoxic lung cancer-derived miRNA-containing EV negatively impact the antitumor activities of natural killer (NK) cells and macrophages, by directly targeting CD107a and phosphatase and tensin homolog (PTEN) expression, respectively [15
]. Therefore, there are numerous evidences indicating that hypoxic tumor cells directly reprogram surrounding immune cells in the TME, via exosome-mediated transfer of HRMs, in order to support cancer progression.
Interestingly, the mobility of miRNAs between cells is not limited only to an exosome-mediated mechanism. Indeed, recent data suggests that miRNAs could be transferred between contacting cells via gap junctions (GJs), and it has been proposed as a novel pathway for intercellular regulation of gene expression [17
]. GJs are connexin (Cx)-formed channels that through the cell-to-cell shuttling of small molecules regulate a plethora of physiological functions, including immunity [19
]. Connexin-43 (Cx43), the main GJ protein expressed in immune cells [19
], has been shown by us and other groups, to accumulate at immunological synapses, allowing GJ-mediated intercellular communications (GJIC) between lymphocytes and DCs [20
], between cytotoxic cells (CTLs and NK cells) and tumor cells [22
], and between DCs and tumor cells [26
]. Interestingly, we previously reported that Cx43 expression is transcriptionally induced by HIF-1α in hypoxic melanoma cells [24
]. However, whether hypoxic melanoma cells can transfer HRMs via Cx43-GJs to immune cells remains to be elucidated. In this report, using human and murine melanoma cell lines, we aimed to evaluate the involvement of Cx43 GJIC in the mobility of miRNAs induced by hypoxia, from melanoma cells to DCs and melanoma-specific CTLs. We identified that miR-192-5p was strongly induced in hypoxic melanoma cells, and it was acquired by CTLs and DCs after being co-cultured with hypoxic melanoma, via a cell contact and Cx43-dependent mechanism. Interestingly, the levels of miR-192-5p in CTLs co-cultured with hypoxic melanoma cells inversely correlated with their cytotoxic activity and ZEB2 mRNA expression, an immune-related and validated target of miR-192-5p. Altogether, our data suggest that hypoxic melanoma cells may suppress CTL cytotoxic activity by transferring hypoxia-induced mi-192-5p through Cx43-GJs. As far as we know, we reported, for first time, a Cx43-mediated miRNA transfer between hypoxic tumors and immune cells, and provided evidence that support a Cx43-dependent mechanism of immune-evasion used by hypoxic melanoma cells.
Recent advances in immunotherapeutic approaches have raised unprecedented clinical achievements in the treatment of cancer, particularly in melanoma [47
]. However, a considerable fraction of treated patients remain as non-responders and/or acquire resistance to immunotherapy [48
]. Moreover, the induction of tumor-specific immune responses by immunotherapies is not always followed by tumor rejection (objective response) and long-term patient survival, mainly because cancer cells acquire the capacity to avoid normalizing cues from their microenvironment, including evasive mechanisms that render them resistant to immune cytotoxic attack. In this line, since more than five decades ago, it is known that cancer cells isolate themselves from their cellular environment down-regulating GJIC [49
]. Indeed, aberrant Cx43 expression and its mislocalization, which leads to reduced GJIC, are frequent in different tumors, including melanoma, and predict cancer patient outcome [19
]. Numerous evidences have shown that Cx43 suppresses tumors via different heterotypic (cancer cell-immune cell or cancer cell-stromal cell) GJICs, as we previously reviewed [19
]. Indeed, Cx43-CJICs have been implicated in different antitumor roles that affect the cancer-immunity cycle (Figure S4
). For example, it has been observed that Cx43-GJICs increase tumor immunogenicity via STING-mediated production of type I interferon (IFN) by colorectal tumor associated-DCs that acquire cancer cell-generated cyclic guanosine monophosphate-adenosine monophosphate (cGAMP) by Cx43-GJs [51
Nevertheless, recent evidence also indicates that GJICs and Cxs could also promote cancer progression and tumor aggressiveness, depending on cancer type, disease stage, and Cx isotype [52
]. Indeed, depending on these variables, expression of Cxs in tumor biopsies could be associated with good or bad prognosis in cancer patients [19
]. Therefore, Cxs are currently considered as conditional tumor suppressor genes. In fact, heterotypic Cx43-GJICs among cancer cells and immune cells can also promote tumor progression [21
]. For example, Cx43-GJs allow brain metastatic cancer cells transferring cGAMP to astrocytes, leading to the activation of the STING pathway and the subsequent production of IFN-α and tumor necrosis factor by the cGAMP-receiving astrocytes [53
]. These cytokines then lead to paracrine activation of STAT1 and NF-κB pathways in brain metastatic cells, supporting tumor growth and chemoresistance [53
]. This pro-tumor role of Cx43-GJ-mediated cGAMP transfer contrasts with its anti-tumor activities observed in colorectal cancer [51
], highlighting the aforementioned context-dependent role of Cx43 in cancer immunity and tumor progression.
Interestingly, hypoxia, a microenvironmental condition that promotes tumor progression and immune evasion, also modulates Cx43 expression and function in cancer and other biological contexts. Interestingly, recent data shows a significantly decrease in Cx43 expression in the left ventricular of HIF-1α conditional knockout mice, which may contribute to impaired heart contractility [54
]. This HIF-1α-dependent Cx43 expression in cardiac cells supports our previous evidence showing that HIF-1α binds GJA1 promoter, inducing Cx43 expression in hypoxic melanoma cells [24
]. In the present study, we confirmed that hypoxia promotes Cx43 expression in additional human and mouse melanoma cell lines. Interestingly, our data also suggest that this phenomenon could occur in patients, given the positive correlation of HIF1A and GJA1 expressions observed in metastatic melanoma TCGA dataset (Figure 1
). Additional reports also indicate that hypoxia modulates Cx43 and GJIC in other cancer types. For example, it has been shown that hypoxia induces Cx43 expression and GJIC in pancreatic ductal adenocarcinoma cells, promoting tumor progression by different metabolic cell coupling-mediated mechanisms [55
]. We have previously identified that the pre-incubation of melanoma cells to hypoxic conditions, decreases their susceptibility to NK cell cytotoxic attack by interfering with Cx43-GJICs via autophagy-mediated Cx43 degradation at the cytotoxic immunological synapse [24
]. In the present report, we observed that hypoxic melanoma cells can establish more Cx43-mediated GJICs with DCs and CTLs than normoxic melanoma cells (Figure 2
), allowing the transfer of specific hypoxia-induced miRNAs (HRMs) from tumor to immune cells (Figure 4
and Figure 5
). Moreover, we generated evidence that suggests that this kind of tumor cell Cx43-dependent modulation of miRNA content in CTLs could also occur in vivo (Figure 6
The first studies suggesting a role of Cx43-GJs in the intercellular transfer of functional RNA molecules were performed using exogenously introduced small interfering (siRNA) or shRNAs [57
]. Similarly, Katakowski and collaborators showed that the delivery of functional and exogenously introduced miRNAs between rat gliosarcoma cells could be inhibited by CBX [59
]. Interestingly, Cx43 channels have higher permeability to different miRNAs than channels constituted by other Cxs, like Cx26, Cx30, and Cx31 [60
]. Given the cell-specific expression of Cxs, this differential permeability of different Cxs may have important implications in the non-cell autonomous miRNA-mediated gene regulation in different cells. Additional reports have also suggested a physiological role for the GJ-mediated miRNA shuttling, including in the context of cancer. For example, bone marrow stromal cells can transfer miR-127, miR-197, miR-222, and miR-223 to breast cancer cells via GJ channels, targeting CXCL12 expression, inhibiting cancer cell proliferation and impacting metastatic dormancy [61
]. Likewise, human macrophages can deliver miR-142 and miR-223 to hepatocarcinoma cells by a contact-dependent mechanism that is inhibited by different GJ chemical blockers [62
]. These miRNAs targeted the expression of stathmin-1 and insulin-like growth factor-1 receptor in the tumor cells, functionally inhibiting their proliferation level [62
]. More recently, it was reported that the inhibition of glioma cell proliferation by miR34a can be increased by allowing its transfer via Cx43-GJs, which suggests that antitumor miRNAs can be combined with Cx43 enhancers in order to improve miRNA therapeutic effects [31
The deregulation of miRNA expression has been implicated in tumor initiation, progression and resistance to therapy or immune attack in most cancers, including in melanoma [63
]. Many of these miRNA deregulations are a result of adaptation pathways to hypoxia and/or are regulated by HIF-1α [64
]. Some of these hypoxia-induced miRNAs have been implicated in cancer progression, promoting immune-resistance phenotypes in cancer cells. For example, in a very elegant study, Wu and collaborators showed that two hypoxia-induced miRNAs, miR-25 and miR-93, repressed cyclic GMP-AMP synthase (cGAS) expression during hypoxia, via targeting the epigenetic factor NCOA3 in breast cancer cells, which lead to hypoxic tumor cells to escape immunological responses induced by damage-associated molecular patterns [65
]. As mentioned before, HRMs can also modulate immune cells of the TME via non-cell-autonomous mechanisms, including the direct mobilization of the miRNAs in EV. Very interestingly, it has been reported that EV, mainly exosomes, contain functional Cx43-channels that allow the exchange of material between exosomes and recipient cells [44
]. In addition, a putative role for Cx43 channels in the recruitment of miRNA into exosomes has also been proposed, given that Cx43 contains several RNA-binding motifs [67
]. Another Cx, Cx46, which is strongly induced by hypoxia in breast cancer cells [68
], was recently observed to be contained as functional channels in breast cancer derived EV [69
]. These antecedents suggest that Cx43 channels may also participate in the transfer of hypoxia-induced miRNAs, contained in tumor derived EV, to immune cells. Although our results indicate a cell-contact dependence in the Cx43-mediated transfer of miR-192-5p from melanoma to CTLs, we cannot discard the participation of melanoma Cx43+
exosomes in the transfer of this (and/or others) miRNA in different experimental conditions.
Our results also suggest that the identity of the transferred miRNAs depends on the nature of the interacting cells more than the levels of miRNAs in those cells. In our experimental conditions, miR-192-5p was transferred from hypoxic melanoma cells to DCs, as well as to CTLs. The hypothetical functional relevance of these observations was suggested by the analysis of the expression of a strongly validated miR-192-5p target, ZEB2, a transcription factor that regulates the expression of many important genes (like GZMA) necessary for the formation of the terminal effector a state in CD8+
T cells [36
]. Accordingly, we observed that after co-culture with hypoxic B16F10 cells, CTL have elevated levels of miR-192-5p, decreased levels of ZEB2 and low cytotoxic activity. Interestingly, we also observed that cancer patients with this expression pattern have poor prognosis, indicating that the miR-192-5p/ZEB2 axis could be relevant in the immune evasion of tumors during cancer progression. Although we do not provide mechanistic data of the effects of miR-192-5p in the recipient immune cells, recent evidence has shown that miR-192 contained in circulating EV regulates cytokine expression in EV-receiving macrophages [70
]. In addition, other report has shown that miR-192 suppresses CD4+
T follicular helper cell differentiation by targeting CXCR5 [71
]. Whether miR-192-5p directly targets ZEB2 (or other mRNA genes) on DCs and CTLs remains to be elucidated in further investigations. Finally, as far as we know, we reported, for the first time, a Cx43-mediated miRNA transfer between hypoxic tumors and immune cells, and provided evidence that support a Cx43-dependent mechanism of immune-evasion used by hypoxic melanoma cells (Figure S5
). Altogether, this evidence can be useful for the designing of target therapies aimed to improve immunotherapy outcomes for hypoxic cancers.
4. Materials and Methods
Wild-type C57BL6 and transgenic pMEL-1 (C57BL6 background) mice [72
] were bred at Bioterio de alta seguridad (BAS) at Universidad de Chile. For all experiments, mice between 6 and 12 weeks of age were bred in specific pathogen-free conditions. All animal experiments were performed in accordance with institutional guidelines for animal care and were approved by the Ethical Review Committees at the Faculty of Medicine of the Universidad de Chile, ethical number: CBA FMUCH 0906 (approval date: 13 December 2016).
4.2. Cell Lines
CRL-6475™) and Mel1, Mel2 and Mel3 [73
] melanoma cell lines were maintained in RPMI-1640 medium supplemented with 10% fetal bovine serum (FBS) and 1% streptomycin/penicillin (all from Corning, Manassas, VA, USA) at 37 °C under 5% CO2
and 95% relative humidity.
4.3. Hypoxia Treatments
For the induction of hypoxia, melanoma cell cultures were incubated for 24 (B16F10) to 72 h (Mel3) in a Modulator Incubator Chamber (Billups-Rothenberg, Inc., San Diego, CA, USA) in a humidified atmosphere containing 5% CO2, 1% O2, and 94% N2 (special gas cylinder supplied by Linde Chile) at 37 °C. In some experiments, CoCl2 (Sigma-Aldrich, St. Louis, MO, USA), a chemical inductor of hypoxia, was used at 150 μM final concentration. Hypoxia induction was routinely tested by flow cytometry using Hypoxia Green Reagent (HGR, Invitrogen, Paisley, UK) according to the manufacturer instructions. HGR is a membrane-permeant probe that releases rhodamine as O2 levels decrease, resulting in a fluorogenic response.
4.4. pMEL-1 CTL Differentiation
pMEL-1 CTLs were differentiated as described before [25
]. Briefly, total splenocytes of pMEL-1 mice were obtained by perfusion of spleens and then cultured for six days in RPMI GlutaMAX™ culture medium supplemented with 10% FBS (HyClone™, South Logan, UT, USA), 1% streptomycin/penicillin (Corning), and 55 μM-mercaptoethanol (Gibco, Carlsbad, CA, USA) and stimulated with gp100 peptide (1 μM KVPRNQDWL; Genetel Laboratories LLC, Madison, WI, USA) and recombinant human (rh) IL-2 (30 IU/mL; Miltenyi Biotec, Bisley, UK). Fresh culture medium with rhIL-2 was added on days 2 and 4. At day 6, the phenotype of the cells was analyzed by flow cytometry as described before [25
4.5. moDC and BM-DC Differentiation
Murine BM-DCs were differentiated from femurs and tibias of C57BL/6 mice by culturing bone marrow cells in the presence of 20 ng/mL mouse recombinant GM-CSF (Biolegend) for 8 days. Fresh culture medium with cytokine was added on day 3, and, on day 6, the medium was refreshed. The phenotype of the cells was analyzed by flow cytometry as described before [74
]. Human peripheral blood moDCs were differentiated from peripheral blood mononuclear cells (PBMCs), as described before [73
4.6. Total RNA Transfer Assay
Mel3 or B16F10 melanoma cells were cultured in normoxic (21% O2
) or hypoxic (1% O2
) conditions for 72 (Mel3) or 24 h (B16F10). Intracellular total RNA molecules were stained by incubating the melanoma cells with 5 μM Syto RNASelect (Invitrogen, Paisley, UK) [29
] for 20 min and washed according to manufacturer’s instructions. mo-DC and pMEL-1 mouse-derived CTLs were loaded with the CellTracker Violet BMQC for 10 min according to manufacturer’s instructions (10 μM for DCs and 2.5 μM for CTLs; Invitrogen), the reaction was stopped by adding FBS for 1 min, and then the cells were washed. SytoRNA donor Mel3 or B16F10 cells were co-cultured for 4 h with the Violet BMQC preloaded mo-DC or pMEL-1 CTLs, respectively, at a 1:3 ratio. GJs were inhibited including in the co-cultures 50 μM of CBX (Sigma-Aldrich, St. Louis, MO, USA), a Cx43 inhibitory mimic peptide named 1848 (300 μM), or its respective controls (Ctrl), vehicle or gap20 peptide, as described before [21
]. After co-culture, the acquisition of RNA in Violet BMQC+
recipient cells was determined by flow cytometry.
4.7. miRNA Transfer Assays
Melanoma cells were cultured in normoxic and hypoxic conditions for 24 (B16F10, shCx43-B16F10, shCtrl-B16F10) or 72 h (Mel3). Then, 1 × 105
melanoma cells were co-culture with BM-DC, pMEL-1 CTLs or moDC at 1:2 cell ratios for 2 h. After co-culture, CD11c+
cells (BM-DC, moDC) or CD8+
cells (CTL) were isolated from CD11c-
cells (melanoma) using specific magnetic cell sorting according to manufacturer’s instructions (MACS, Miltenyi Biotec, Bisley, UK). Total RNA samples, including miRNA, were isolated from pre-co-culture (named culture) and post-co-culture (named co-culture) using miRNeasy Mini kit according to manufacturer’s instructions (Qiagen, Hilden, Germany). Quality and concentration of RNA samples were determined by Agilent 2200 TapeStation analysis using RNA ScreenTapes according to manufacturer instructions (Agilent Technologies, Santa Clara, CA, USA). Only RNA samples with RNA integrity number (RIN) values >8 were used. cDNAs were generated by reverse transcription using miScript II RT Kit and the miScript HiFlex buffer according to manufacturer instructions (Qiagen). Expressions of different mature miRNAs were determined using miScript SYBR Green PCR Kit and specific mouse or human miScript primer assays according to the manufacturer’s protocols (Qiagen). Levels of each miRNA were expressed relative to the level of small nuclear RNA RNU6-2 (miScript primer assays) and analyzed according to the Δ/ΔCt method. PCRs were performed in a LightCycler 480 Instrument and LightCycler 480 Software (Roche, Basel, Switzerland). As controls of mature miRNA transfer or cell isolation, precursor miRNA expression levels or CD8 and gp100 mRNA expressions were determined in culture and co-cultured cells, using specific miScript primer assays (for precursor miRNAs) or designed primers for mRNA (see Section 4.12
) (Figure S6
For in vivo determination of miR-192-5p expression, 6–12 weeks old C57BL6 mice were subcutaneously challenged with 1.5 × 105
B16F10 cells expressing wild type (Cx43WT
) or a dominant negative Cx43 mutant (Cx43DN
). Fourteen days after tumor challenge, spleens and tumors were collected and treated as described before [74
] and non-immune (CD45−
) and CD45+
T cells were isolated using FACS sorting. RNA samples were obtained and miR-192-5p expression analyzed as described before.
4.8. Pantoxilux Cytotoxic Assay
B16F10 cells expressing wild type (Cx43WT) or a dominant negative Cx43 mutant (Cx43DN) were cultured for 24 h upon normoxic or hypoxic conditions. Then, pre-conditioned melanoma cells were co-cultured for 2 h with pMEL-1 CTLs at a 1:2 cell ratio, allowing or not (Transwell plates, Corning) cell-to-cell contact. After co-culture, the cytotoxic activity of the CTLs against B16F10 cells was determined by flow cytometry using the Pantoxilux® (OncoImmunin, Inc., Gaithersburg, MD, USA) granzyme and upstream caspase activity kit, following manufacturer’s instructions.
4.9. Flow Cytometry
Flow cytometry experiments were performed as previously described [30
]. Intracellular hypoxia was determined by hypoxia green reagent (HGR) staining, according to manufacturer’s instructions (Thermofisher, Waltham, MA, USA). To detect Cx43, a rabbit polyclonal anti-human/mouse Cx43 antibody, directed to the C-terminal domain (C6219; Sigma-Aldrich), and a secondary donkey anti-rabbit fluorescein isothiocyanate (FITC)-conjugated antibody (Poly4064; BioLegend, San Diego, CA, USA) were used. The following monoclonal antibodies were used for CTL staining: CD8-V500 (dilution 1:500; clone 53-6.7), CD25-APC (dilution 1:1500; clone PC61), CD44- phycoerythrin (PE)-Cy7 (dilution 1:500; clone IM7), Vβ13-V450 (dilution 1:400; clone MR 12.3), all from BioLegend. 7-AAD (1:1500; BioLegend) staining was used for dead cell exclusion. For mo-DC phenotype evaluation, the following antibodies were used: IA/IE-APC-Cy7 (dilution 1:1500; clone M5/114.15.2; Biolegend), CD11c-APC (dilution 1:800; clone N418; eBioscience™). LIVE/DEAD™ Fixable Aqua (dilution 1:800; Thermofisher) was used for dead cell exclusion. Samples were acquired on a FACSVerse. (BD Biosciences, Hershey, PA, USA) and analyzed using the FlowJo software (version 10.0.7; Tree Star Inc., Ashland, OR, USA). For CD8+
FACS sorting, the following monoclonal antibodies were used: CD3-AlexaFluor488 (dilution 1:400; clone 17A2), CD8-PE (dilution 1:200; clone 53-6.7), CD45-PerCP-Cy5.5 (dilution 1:400; clone 30-F11), all from BioLegend. 7-AAD (1:1500; BioLegend) staining was used for dead cell exclusion.
Confocal microscopy of fixed B16F10 cells was performed as previously described [24
]. Briefly, B16F10 cells were cultured overnight at 37 °C and 5% CO2
on poly-l-lysine-coated slides, using 5 × 104
cells per slide in 100 μL of complete culture medium. Cells were then cultured in absence (N) or presence (H) of 150 μM CoCl2
(Sigma-Aldrich) for additional 48 h and gently washed twice with phosphate-buffered saline (PBS) and fixed with 4% paraformaldehyde for 15 min at room temperature in the dark. After gentle washing with PBS, cells were incubated in ammonium chloride (50 mM) for 10 min. Then, cells were permeabilized for 10 min (0.5% Triton X-100 and 0.5% FBS) and blocked with 0.5% bovine serum albumin (BSA) for 15 min. Cells were stained with 1:500 dilutions of FITC-coupled monoclonal antibody anti-Cx43 (D-7; Santa Cruz Biotechnology, Dallas, TX, USA) or a PE-coupled monoclonal antibody anti-HIF-1α (28b, Santa Cruz Biotechnology) for 1.25 h at room temperature, and then stained with 1:1000 dilution of Hoechst 33,342 (Invitrogen) for 15 min. Cells were analyzed with a C2+ confocal microscope (1000, Nikon Instruments Inc., Melville, NY, USA) and a Spinning Disk Olympus BX61WI microscope (400, Center Valley, PA, USA). Cx43 and HIF-1α expression and nucleus areas were quantified using ImageJ software (version number 1.48v, National Institutes of Health, Bethesda, MD, USA).
For knocking down Cx43 expression, B16F10 and Mel3 melanoma cells were stably transfected with lentiviral vector plasmids encoding a pool of 4 Gja1 (sc-35091-SH) or GJA1 (sc-29276-SH)-specific shRNAs (plus hairpin), using plasmid transfection reagent (sc-108061) and shRNA plasmid transfection medium (sc-108062) according to the manufacturer’s instructions (Santa Cruz Biotechnology). A control shRNA plasmid (sc-108060, Santa Cruz Biotechnology) was used for generating stable transfection in the control cells. Puromycin (5 μg/mL; sc-108071, Santa Cruz Biotechnology) was added to the culture medium for selection of transfected cells.
B16F10 cells were stably transfected with pcDNA3.1 + NeGFP plasmids containing wild type or mutant dominant negative Cx43 (T154A) (synthesized by GenScript, NJ, USA) using Lipofectamine (Invitrogen) and G418 neomycin antibiotic as selection pressure (Sigma-Aldrich).
RNA samples were isolated and treated as described in Section 4.7
. Human and mouse Cx43, Cx26 and actin expression levels were determined by qRT-PCR using specific QuantiTect Primer Assays and miScript SYBR Green PCR Kit according to the manufacturer’s protocols (Qiagen) and analyzed according to the Δ/ΔCt method. PCRs were performed in a LightCycler 480 Instrument and LightCycler 480 Software (Roche). Primers for ZEB2 (forward: 5′-CAT GAA CCC ATT TAG TGC CA-3′, reverse: 5′-AGC AAG TCT CCC TGA AAT CC-3′); CD8 (forward: 5′-GCT CAG TCA TCA GCA ACT CG-3′, reverse: 5′-ATC ACA GGC GAA GTC CAA TC-3′); and gp100 (forward: 5′-AGC ACC TGG AAC CAC ATC TA-3′, reverse: 5′-CCA GAG GGC GTT TGT GTA GT-3′) were synthetized by IDT Integrated DNA Technologies (Fermelo, Providencia, Chile). Human Hypoxia Signaling miScript miRNA PCR Arrays (MIHS-121ZF-2; Qiagen) were used for determination of hypoxia-induced miRNAs in pooled RNA samples of Mel1, Mel2, and Ml3 melanoma cells following the manufacturer protocols.
4.13. Western Blot
Cells were lysed in radioimmunoprecipitation assay (RIPA) lysis buffer supplemented with 5 mM EDTA and protease/phosphatase inhibitors cocktail (all from Thermo Fisher Scientific, Waltham, MA, USA) for 30 min on ice. Western blotting was conducted as described previously [25
], loading 30 μg of total proteins per well. For protein detection HRP-conjugated anti-Cx43 (dilution 1:500; clone F-7; Santa Cruz Biotechnology), and HRP-conjugated anti-β-actin (dilution 1:10,000; clone C4; Santa Cruz Biotechnology) antibodies were used. The membranes were revealed using the SuperSignal®
West Pico Chemiluminescent Substrate kit (Thermo Fisher Scientific) according to manufacturer’s instructions and analyzed in an ImageQuant LAS 500 (GE Healthcare, Uppsala, Sweden). Images were analyzed using ImageJ software (National Institutes of Health). Unedited image of the western blot is shown in Figure S7
Data were analyzed using GraphPad Prism 7 (GraphPad Software Inc., San Diego, CA, USA). Statistical analyses were performed using one-way ANOVA, Tukey’s multiple comparison tests, or a two-tailed Student’s t-test where appropriate. Differences were considered statistically significant at p < 0.05.