Hepatocellular carcinoma (HCC) is the sixth most common cancer type and the fourth leading cause of cancer-related death. This cancer appears with higher incidence in men and during obesity; however, the specific mechanisms underlying this correlation are unknown. Adipose tissue, a key organ in metabolic syndrome, shows evident gender disparities in the production of adipokines. Levels of the important adipokine adiponectin decrease in men during puberty, as well as in the obese state. Here, we show that this decrease in adiponectin levels is responsible for the increased liver cancer risk in males. We found that testosterone activates the protein JNK in mouse and human adipocytes. JNK-mediated inhibition of adiponectin secretion increases liver cancer cell proliferation, since adiponectin protects against liver cancer development through the activation of AMP-activated protein kinase (AMPK) and p38α. This study provides insight into adipose tissue to liver crosstalk and its gender relation during cancer development, having the potential to guide strategies for new cancer therapeutics.

Hepatocellular carcinoma (HCC) is the fourth leading cause of cancer-related death (Bray et al., 2018), and its incidence is rising worldwide due to the increased prevalence of obesity (Parkin et al., 2005). Therapeutic options for HCC are limited, and survival after diagnosis is poor. Better preventive, diagnostic, and therapeutic tools are therefore urgently needed, particularly in view of the important contribution of obesity to HCC incidence worldwide (Bakiri and Wagner, 2013). In addition, epidemiological studies have shown a higher incidence of HCC in men than in women (Bosch et al., 2004), a dimorphism also observed in mouse models (Ghebranious and Sell, 1998; Naugler et al., 2007). Adipose tissue is one of the most important players in the adaptation to obesity, through the regulation of fuel metabolism storage, the release of nutrients and, indirectly, the production of circulating adipokines (Havel, 2002). However, this tissue responds differently in males and females, due to the different genetic make-up and differences in the inputs that act on it (Fuente-Martín et al., 2013). The adipokine adiponectin improves hepatic insulin sensitivity and fuel oxidation (Whitehead et al., 2006) and is considered to play a protective role in cancer (Dalamaga et al., 2012). However, its role in HCC is controversial and requires further investigation (Cheung and Cheng, 2016). Circulating adiponectin levels have been reported to show a gender disparity, being higher in females than in males (Yannakoulia et al., 2003). It has been suggested that testosterone might play a role in adiponectin levels, whereas estrogens do not seem to mediate this effect (Nishizawa et al., 2002). Here, we demonstrate that lower adiponectin levels in males account for their higher prevalence of liver cancer. We show that p38α and AMP-activated protein kinase (AMPK) activation in hepatocytes by adiponectin results in a protective effect against tumor growth. Furthermore, we found that testosterone activates JNK in human and mouse adipocytes, and genetic deletion of JNK1 in mouse adipose tissue results in higher adiponectin levels and protection against HCC. Our results unravel a clear crosstalk between sex hormones and adipocytes signaling and physiology, clarifying the mechanism underlying gender disparity in liver cancer development.

Increased levels of adiponectin in females protect against liver cancer

Using a cohort of healthy subjects, we found that adiponectin levels are higher in women compared with men (Fig. 1 a). Quantification of circulating adiponectin in 11–12-wk-old C57BL/6J mice confirmed previous findings, detecting more than twice the level in females than in males (Fig. S1 a). This correlated with the stronger growth of subcutaneously implanted mouse HCC-derived tumor cells in males than in females (Fig. 1 b). In addition, gender differences were HCC-specific, given that growth of subcutaneously injected colon adenocarcinoma-derived tumor cells (Fig. S1 b) or melanoma-derived tumor cells (Fig. S1 c) was not different between males and females. To further study how gender could affect adiponectin levels and, as a consequence, tumor growth, we assessed the effect of castration on blood adiponectin levels. Castrated males presented similar adiponectin levels to those found in females (Fig. 1 c). In line with the important role of adiponectin in gender disparity, castrated males developed smaller tumors in allograft experiments (Fig. 1 d). To determine whether the gender differences in adiponectin production influences tumor growth, we performed allograft experiments in male and female adiponectin KO (Adipoq−/−) and WT mice. In WT mice, we observed larger tumors in males than in females, whereas Adipoq−/− mice showed no gender differences (Fig. 1 e). Adiponectin cellular function is mainly mediated by two transmembrane receptors, known as adiponectin receptors 1 and 2 (AdipoR1/2; Wang and Scherer, 2016). Increased expression of both receptors were found in HCC-derived cells in comparison with colon adenocarcinoma and melanoma cells (Fig. S2 a). In the liver, hepatocytes are the main source of AdipoR2, while AdipoR1 is expressed in hepatocytes, Kupffer cells (KCs), and endothelial cells (ECs; Fig. S2 b). In addition, hepatocytes increase the expression of both receptors when they become tumoral (Fig. S2 c).

To evaluate the importance of these receptors in the protection observed in females, we used shRNAs against them in Hep53.4 cells and performed an allograft experiment in females (Fig. S2 d). Only depletion of AdipoR2 was sufficient to block the protection found in females, as no differences in tumor growth were found between males and females injected with Hep53.4 AdipoR2-shRNA–treated cells (Fig. S2 e).

The contribution of adiponectin to HCC development was evaluated by the diethylnitrosamine (DEN)-induced hepatocarcinogenesis model (Heindryckx et al., 2009). At postnatal day (P) 1, male mice were injected with a serotype 8 adeno-associated virus overexpressing adiponectin under the control of the adipocyte aP2 promoter (AAV-aP2-Adipoq) or carrying an empty control vector (AAV-aP2-CTRL). After 14 d, these mice received i.p. DEN injections to induce HCC. Adiponectin overexpression was confirmed by measurement of adiponectin plasma levels 3 wk after virus injection (Fig. S3 a), as well as 8.5 mo after virus injection (Fig. S3 b). Quantification of tumors 8 mo after DEN treatment revealed a significant reduction in the number of tumors in adiponectin-overexpressing males (Fig. 1, f–i), indicating that adiponectin reduces tumor progression and may be a potential treatment for HCC. To further evaluate the protective effects of adiponectin in females, we used the DEN-induced HCC model in WT and Adipoq−/− mice. Results showed that lack of adiponectin slightly increases the number of tumors 8 mo after DEN treatment (Fig. S3, c and d), corroborating that high levels of adiponectin in females contribute to the protection against HCC.

Lack of JNK1 in the adipose tissue protects against HCC progression

It is known that adipose tissue JNK controls the production of important adipokines that regulate liver metabolism (Sabio et al., 2008; Manieri and Sabio, 2015). We therefore tested whether sex-dependent differences in JNK activation in adipose tissue could account for the higher adiponectin levels in females. Western blot analysis revealed a higher activation of adipose tissue JNK in males than in females, both in mouse and rat models, which correlated with lower adiponectin levels (Fig. 2 a; Fig. S1 a; and Fig. S4, a and b), suggesting a general effect conserved among species. In line with these results, testosterone activated JNK in both mouse and human adipocytes (Fig. 2 b and Fig. S4 c). To check whether testosterone was also able to regulate JNK phosphorylation in vivo, we castrated WT male mice and performed testosterone replacement. Adipose tissue JNK activation was significantly reduced in castrated males, whereas testosterone injection was enough to increase its activation to noncastrated mice levels within 2 wk (Fig. 2 c). Importantly, the observed changes in JNK phosphorylation were in concordance with adiponectin levels, which decreased to the levels of noncastrated mice when testosterone replacement was performed (Fig. 2 d). We next examined serum concentrations of adiponectin in males with adipose tissue-specific deletion of JNK1 (FKO) or control mice (FWT; Fig. S5 a). Blood adiponectin levels were significantly higher in FKO males than in FWT males (Fig. S5 b), whereas no between-genotype differences were found in the levels of TNFα, IL-1β, or IL-6 (Fig. S5 c). We next performed allograft tumor formation assays in FWT and FKO males to investigate the role of adipose tissue JNK1 in tumor growth. Consistent with the higher adiponectin levels found in FKO males (Fig. S5 b), tumor growth in these mice was slower than in FWT controls, and postmortem inspection revealed smaller tumors (Fig. 3 a). To further evaluate these results, FKO and FWT males were injected with DEN, and tumors were examined 8 mo later. We found that HCC was strongly suppressed in FKO mice (Fig. 3 b). Moreover, FKO mice had significantly fewer tumors than FWT mice (Fig. 3 c), and tumor size and maximum tumor size were significantly lower in FKO mice (Fig. 3, d and e).

Higher levels of adiponectin in FKO mice are essential for their protection against HCC

To evaluate the potential role of adiponectin in the protection against HCC in FKO males, we crossed WT and JNK1-deficient mice with mice deficient in adiponectin (Adipoq−/−), and repeated the allograft assays. Tumor growth was identical in FWTAdipoq−/− and FKOAdipoq−/− males (Fig. 4 a), consistent with a central role for adiponectin in the FKO phenotype, and suggesting that the elevated adiponectin levels in FKO males were responsible for their reduced tumor growth. To corroborate this hypothesis, we analyzed DEN-induced liver cancer in FWTAdipoq−/− and FKOAdipoq−/− males. Tumor number, size, and maximum size after 8 mo did not differ significantly between the two groups of mice, suggesting an important role for adiponectin in the protection of FKO males against DEN-induced HCC (Fig. 4, b–e).

AMPKα and p38α activation by adiponectin reduces tumor progression

Adiponectin transduces its signal in the liver through the activation of two pathways: p38 MAPK (Mao et al., 2006a,b) and AMPK (Yamauchi et al., 2002; Polyzos et al., 2010). p38α suppresses liver cancer development by reducing cell proliferation, and consequently, mice with hepatic deletion of p38α show enhanced hepatocyte proliferation and tumor development (Hui et al., 2007). AMPK also suppresses cancer, and treatment with the AMPK activator metformin reduces cancer incidence (Chen et al., 2013) and may improve survival among liver cancer patients (Ma et al., 2016). Consequently, AMPK is emerging as an important metabolic tumor suppressor and a promising target for cancer prevention and therapy (Luo et al., 2010). We evaluated the activation of AMPK and p38α in allografts from male and female mice. Immunoblot analysis detected higher levels of AMPK and p38α activation (phosphorylation) in females (Fig. 5 a), correlating with the higher adiponectin levels in female mice (Fig. S1 a). This gender difference in p38α or AMPK activation was not observed in Adipoq−/− mice (Fig. 5 b), suggesting that higher activation of these kinases in females is the result of the higher levels of circulating adiponectin. To evaluate whether p38α and AMPK activation is implicated in the protection against liver cancer in females, we implanted Hep53.4 cells subcutaneously into male mice and activated each pathway by two different strategies. In the first strategy, animals were treated with metformin to activate AMPK. In the second one, allografts were injected with a retrovirus containing active p38α or a control construct on days 9 and 22 after transplantation. As expected, metformin treatment significantly reduced tumor volume (Fig. 5 c), indicating that AMPK activation confers protection against HCC. Similarly, tumors infected with active p38α were significantly smaller (Fig. 5 d), suggesting that the adiponectin-mediated protection against HCC is also mediated by p38α activation.

We next investigated the extent of p38α and AMPK phosphorylation in livers of FKO mice. Consistent with the high levels of adiponectin in FKO males, liver activation of p38α and AMPK was significantly higher in DEN-treated FKO male mice than in their FWT counterparts (Fig. S5, d and e). To evaluate whether AMPK activation is necessary for adiponectin-mediated protection in FKO mice, we treated Hep53.4 cells with shRNA against AMPK or a scrambled control sequence. As expected, AMPK knockdown in Hep53.4 cells resulted in larger tumors after implantation in FWT animals. More importantly, FKO males were no longer protected against tumor growth, indicating that AMPK activation in hepatocytes is necessary for the protection observed in FKO mice (Fig. 5 e).

Liver cancer is the fourth most common cause of death from cancer worldwide (Marquardt et al., 2015). The incidence and mortality rates of liver cancer have increased rapidly, coinciding with the rising prevalence of obesity (Larsson and Wolk, 2007). Accumulating epidemiological evidence indicates that excess of body weight may be a risk factor for liver cancer; however, the link between both phenomena remains elusive.

Adipose tissue is recognized as an active endocrine organ that mediates a variety of physiological functions via secretion of adipokines. Here, we found that gender differences in adipocytes are important players in HCC progression and can contribute to the increased incidence of HCC observed in males. Our study demonstrates that adiponectin plays an important role in protecting female mice against HCC. In men, adiponectin levels decrease with puberty, whereas its levels are maintained in women (Böttner et al., 2004). A similar gender disparity occurs in mice (Combs et al., 2003). We found that the gender bias in tumor growth is strongly dependent on adiponectin, and that adiponectin KO mice showed no gender differences in tumor progression. We also demonstrate that activation of JNK in adipose tissue correlates with reduced levels of adiponectin in males, and that males with JNK1 deficiency in adipose tissue have increased adiponectin secretion, protecting against HCC. Moreover, castration reduces JNK phosphorylation in the adipose tissue while testosterone replacement increases it, together with a reduction in adiponectin to the levels before castration. These results indicate that activation of JNK in adipose tissue inhibits adiponectin secretion, promoting tumor growth in males. Epidemiological studies suggest the association between low adiponectin levels and liver tumorigenesis (Kotani et al., 2009). In line with this, adiponectin has an anti-oncogenic potential in HCC cell lines (Sharma et al., 2010). However, opposite results were reported for HCC associated with hepatitis C (Arano et al., 2011). These discrepancies may be attributable to the different HCC etiologies examined in those studies (Dalamaga et al., 2012). Here, using mouse models, we demonstrated that adiponectin protects against HCC by activating p38α and AMPK. Our results suggest that adiponectin and metformin could be of use in the treatment of HCC.

Study population and sample collection

For the analysis of human levels of blood adiponectin, individuals were recruited among patients who underwent laparoscopic cholecystectomy for gallstone disease.

Blood was extracted using straight needles (21G butterfly) and Vacuette Z Serum Sep Clot Activator tubes. After 30 min, these tubes were centrifuged at 1,500 rpm for 10 min at room temperature to separate serum, which was divided into aliquots and stored at −80°C until further analysis. The study was approved by the Ethics Committee of the University Hospital of Salamanca, and all subjects provided written informed consent to participate. Fasting blood samples were collected for adiponectin, glucose, and lipid analysis. Patients were excluded if they had a history of alcohol use disorders or excessive alcohol consumption (>30 g/d in men and >20 g/d in women), chronic hepatitis C or B, or body mass index ≥ 35 (Table S1).

Animals

Adiponectin KO mice (B6;129-Adipoqtm1Chan/J) and FabP4 cre mice (B6.Cg-Tg(Fabp4-cre)1Rev/J) were purchased from the Jackson Laboratory and backcrossed to the C57BL/6 background for 10 generations. Mice with specific deletion of JNK1 in adipocytes (FKO) were as described (Sabio et al., 2008), and FWT littermates were used as controls. Mice were housed randomly in a pathogen-free animal facility and maintained on a 12-h light/dark cycle at constant temperature and humidity. Genotypes were identified by PCR analysis of genomic DNA isolated from mouse tails. For tumor studies, mice at P14 received one i.p. injection of DEN (Sigma-Aldrich) at 50 mg/kg body weight dissolved in saline. After 8 mo, tumors were extracted and measured with a caliper. In all cases, mice were euthanized after overnight starvation. For testosterone deprivation studies, castration or sham operations were performed in mice at 4–5 wk of age. 23 d after castration, testosterone propionate (Desma) diluted in corn oil (5 µg/g body weight) or vehicle alone was subcutaneously injected in castrated or sham-operated mice, every other day for 3 wk. Blood samples were collected 2 wk after the first injection, and testosterone propionate (50 µg/g body weight) was injected into some castrated mice 30 min before sacrifice, at the end of the experiment. For healthy and tumor hepatocytes analysis, C57BL/6J mice were injected with a single dose of DEN at day 14 postnatally, followed by administration of thioacetamide (TAA; 300 µg/liter) in the drinking water during 26 wk as described in Salguero Palacios et al. (2008). Control mice received a single injection of PBS at day 14 and normal drinking water. The parenchymal and the nonparenchymal compartment was isolated from vehicle (healthy) or DEN/TAA (tumor) mice. All animal experiments conformed to European Union Directive 2010/63EU and Recommendation 2007/526/EC, enforced in Spanish law under Real Decreto 1386/2018. All experiments were approved for the Centro Nacional de Investigaciones Cardiovasculares Carlos III ethics committee and Comunidad de Madrid.

Adeno-associated virus

Male mice were injected intravenously at P1 with 0.5 × 1011 viral particles of a serotype 8 adeno-associated virus overexpressing adiponectin under the control of the adipocyte-specific aP2 mini promoter (lentiviral vector tailor-designed and purchased from Vector Builder) or carrying an empty control vector.

Serum analysis

Serum cytokine concentrations were measured by multiplexed ELISA in a Luminex 200 analyzer (Millipore).

Adiponectin was detected in plasma samples by Western blot. Samples were diluted in PBS (1:10) and buffered with 2× Native Tris-Glycine Sample Buffer (Invitrogen). Diluted samples were loaded on precast native gels to preserve adiponectin multimers (NativePAGE Bis-Tris gel system; Thermo Fisher Scientific) and run with Novex Tris-Glycine Native Running Buffer (Invitrogen). In other cases, plasma samples were prepared as described but diluted in PBS, Milli-Q water, and Laemmli Sample Buffer and separated by SDS-PAGE. Membranes were blocked with 10% milk prepared in PBS without detergents (in the first case) or with PBS-Tween 0.1% (in the second case), and probed with a primary antibody to adiponectin (1:500 dilution; PA1-054; Thermo Fisher Scientific). After washing, membranes were incubated with a fluorescent secondary antibody (goat anti-rabbit 680 nm 926–32221; Odyssey) or a horseradish peroxidase–conjugated secondary antibody (GE Healthcare), and proteins were visualized and quantified using the Odyssey LI-COR imaging system (LI-COR) or an enhanced chemiluminescent substrate (Clarity Western ECL substrate; Bio-Rad), respectively. We assured the specificity of the antibody using serum from Adiponectin KO mice. Results were normalized to the control group (male mice or FWT or control mice, as indicated in the figure legends).

Biochemical analysis

Total proteins from different organs or from adipocytes were extracted in lysis buffer (50 mM Tris-HCl, pH 7.5, 1 mM EGTA, 1 mM EDTA, pH 8.0, 50 mM NaF, 1 mM sodium glycerophosphate, 5 mM pyrophosphate, 0.27 M sucrose, 1% Triton X-100, 0.1 mM PMSF, 0.1% 2-mercaptoethanol, 1 mM sodium orthovanadate, 1 µg/ml leupeptin, and 1 µg/ml aprotinin) and centrifuged at 8,600 × g for 20 min at 4°C. Extracts were separated by SDS-PAGE and transferred to 0.2-µm-pore-size nitrocellulose membranes (Bio-Rad). Blots were probed with primary antibodies to phospho-JNK (Thr183/Tyr185; 4668), total JNK (9252), phospho-AMPKα (Thr172; 2531), total AMPKα (2603), and phospho-p38 (Thr180/Tyr182; 9211; all from Cell Signaling Technology), p38 from Santa Cruz Biotechnology (sc-535), and GAPDH (sc-25778) and vinculin (V4505; Sigma-Aldrich). All antibodies were used at 1:1,000. After washes, membranes were incubated with an appropriate horseradish peroxidase–conjugated secondary antibody (GE Healthcare), and signals were detected using an enhanced chemiluminescent substrate (Clarity Western ECL substrate; Bio-Rad). Protein levels were analyzed by optical density measured with ImageJ (National Institutes of Health). Results were normalized to the not phosphorylated form.

Allografts

The Hep53.4 cell line was purchased from CLS-Cell Lines Service and tested against Mycoplasma (MycoAlert Detection Kit; Lonza). Cells were cultured in DMEM supplemented with 10% FBS, L-glutamine, and antibiotics. The MC-38 and B16-F10 cell lines originally were tested against Mycoplasma. MC-38 cells were cultured in DMEM supplemented with 10% FBS, non-essential amino acids, sodium pyruvate, Hepes, L-glutamine, and antibiotics, whereas B16-10’s culture medium only contained DMEM supplemented with 10% FBS, sodium pyruvate, and antibiotics. For allograft assays, Hep53.4 cells (5 × 104 cells for usual experiments, or 1 × 106 for short hairpin RNA experiments) were mixed 1:1 with Matrigel Matrix (Corning) and subcutaneously injected into each flank of anesthetized 8–12-wk-old mice. For MC-38 or B16-F10 allograft assays, 5 × 105 cells without Matrigel were subcutaneously injected into each flank. In both cases, tumor growth was monitored by measuring length and width every 3–4 d for a period of 2, 3, or 5 wk. At the end of the experiment, mice were sacrificed and tumors extracted and measured. To study AMPKα activation by metformin, mice were treated with metformin (300 mg/d · kg body weight), starting 1 wk before allograft injection. In allografts treated with active p38α, pBABE p38α D176A/F327S retrovirus was injected intratumorally on days 9 and 22.

Adipocytes cell culture

Immortalized white preadipocytes from C57BL/6 male mice were maintained in DMEM-F12 medium with 8% FBS, 200 mM L-glutamine, and 10,000 U/ml penicillin/streptomycin. These cells were differentiated to white adipocytes for 10 d with the previous medium supplemented with 5 µg/ml insulin, 25 µg/ml 3-isobutyl-1-methylxanthine, 1 µg/ml dexamethasone, and 1 µM troglitazone. The cells were starved overnight and treated with 300 nM testosterone (Sigma-Aldrich) for 2 h.

Human visceral white preadipocytes were purchased from Innoprot, and culture was set up following the manufacturer’s instructions. Preadipocyte differentiation medium from Innoprot was used for the differentiation to white adipocytes during 8 d, followed by an overnight starvation and treatment with 1,200 nM testosterone or vehicle for 30 min.

Liver cell populations isolation

To isolate the different hepatic cell types, a perfusion was performed in a C57BL/6 mouse. The liver was perfused with each solution (A, B, and C; see Table S3) separately for 5 min or 25 ml. After the final solution C, the perfusion was stopped and the liver was dissected and transferred into solution D and incubated for 20 min at 37°C. To ensure proper digestion of the liver, solution D was inverted several times during the incubation time. Next, solution D containing the digested liver was filtered through a 70-µm cell strainer, and the filtered solution was centrifuged at 50 × g for 1 min at 4°C. The pellet containing hepatocytes was kept for further analysis, and the supernatant was transferred to a new conical tube and subsequently centrifuged at 720 × g for 8 min at 4°C. The supernatant was removed and the pellet was resuspended in 10 ml Gey’s Balanced Salt Solution buffer (GBSS-B) solution containing 150 µl DNase I stock solution, and the conical tube was filled up to 50 ml with GBSS-B solution. In the next step, the solution was centrifuged at 720 × g for 8 min at 4°C, and the supernatant was removed. Next, the pellet was resuspended in 10 ml GBSS-B containing 150 µl DNase I stock solution. In addition, 24 ml GBSS-B and 14 ml of the Nycodenz 1 solution were added, and the solution was carefully mixed. Further, 4 ml of the Nycodenz 2 solution was transferred into six 15-ml falcon tubes. Next, 8.3 ml of the mixed cell solution containing GBSS-B, DNase, and Nycodenz 1 was carefully laid onto the Nycodenz 2 solution, and the GBSS-B solution was used to gently overlay the cell suspension to obtain a final volume of 15 ml. Consequently, the gradient solution was centrifuged without brakes at 3,000 × g for 20 min at 4°C. After the centrifugation step, hepatic stellate cells (HSCs) are found in the upper gradient phase as a white ring, and KCs are found in the lower gradient phase. Each cell type was carefully transferred to a new tube and washed with GBSS-B solution and subsequently centrifuged at 720 × g for 8 min at 4°C to pellet the cells. The supernatant was removed, and the cells were checked for their purity under a microscope and subsequently frozen in liquid nitrogen and stored at −80°C for further RNA isolation. See Tables S3 and S4 to check the solutions used for the isolation of different hepatic cell types.

Histological analysis

Histopathological examination of H&E-stained livers was performed in an Eclipse CiL light microscope (Nikon), using 10–20× magnification with a numerical aperture of 0.25–0.40 of the optic lenses. Microphotographs were taken with a Nikon Camera DS-Fi3 and acquired using the NIS-Elements Microscope Imaging Software (Nikon). Histopathological grading of chronic liver disease was assessed based on the score system previously published (Goodman, 2007), but modified for nonalcoholic fatty liver disease and HCC. No insult (−) or different grades of injury (mild [+], moderate [++], or marked [+++]) were noted in at least seven view fields per slide.

RNA analysis

Total RNA was isolated from healthy or tumorigenic hepatocytes, other cell populations in the liver (KCs, ECs, and HSCs), as well as from different tumor cell lines (Hep53.4, MC-38, and B16-F10) using the RNeasy Mini Kit (Qiagen). Complementary DNA was synthesized with the High-Capacity Complementary DNA Reverse Transcription Kit (Applied Biosystems). Sequences of primers used for quantitative real-time PCR (qRT-PCR) are provided in Table S2. Expression levels were normalized to Gapdh mRNA. qRT-PCR was performed using Fast SYBR Green.

Statistical analysis

Differences between groups were examined for statistical significance using the two-tailed Student’s t test, Student’s t test coupled with Welch’s correction in case of not equal variances, or one- or two-way ANOVA coupled to Bonferroni’s or Dunnett’s post-test.

Online supplemental material

Fig. S1 shows the mouse gender differences of adiponectin levels and the lack of differences in the growth of colon adenocarcinoma and melanoma cells between males and females. Fig. S2 shows the analysis of adiponectin receptors expression in different tumor cells and liver cell populations, along with the growth of HCC-derived tumor cells with reduced AdipoR1 or AdipoR2 levels. Fig. S3 shows the increased adiponectin levels observed in males infected with the adeno-associated virus overexpressing adiponectin and the maintenance of these high levels until the end of the experiment, as well as the effect of lack of adiponectin in females on tumor number. Fig. S4 shows the rat gender differences of adiponectin levels, the mouse gender differences in fat JNK activation, and its activation in human adipocytes upon testosterone stimuli. Fig. S5 shows the specific deletion of JNK1 in the adipose tissue, adiponectin and cytokine levels in FWT and FKO mice, and higher AMPKα and p38α activation in FKO compared with FWT mice. Table S1 shows the characteristics of women and men population used in the study. Table S2 shows the sequences of the primers used for the qRT-PCRs. Tables S3 and S4 show the solutions used for the isolation of different hepatic cell types, for both hepatic perfusion and hepatic cell isolation.

We thank K. McCreath and S. Bartlett for English editing. We are grateful to Dr. R.J. Davis (University of Massachusetts Medical School, Worcester, MA) for the FKO animals and critical reading of the manuscript. We thank Dr. A. Nebreda (Institute for Research in Biomedicine, Barcelona, Spain) for the active pBABE p38α D176A/F327S plasmid. We also thank Dr. David Sancho (Centro Nacional de Investigaciones Cardiovasculares [CNIC], Madrid, Spain) for the MC-38 and B16-F10 cell lines, and Centro Nacional de Investigaciones Cardiovasculares Carlos III Animal Facility for the castration of mice.

G. Sabio is an investigator on the Ramón y Cajal Program. E. Manieri is a La Caixa Foundation fellow. L. Herrera-Melle is a fellow of the Ministerio de Educación, Cultura y Deporte (FPU15-05802). This study was funded by the following grants: G. Sabio was funded by the European Research Council (ERC 260464), European Foundation for the Study of Diabetes–Lilly, Ministerio de Ciencia, Innovación y Universidades (MICINN/SAF2016-79126-R), Comunidad de Madrid (B2017/BMD-3733), and BBVA Becas Leonardo a Investigadores y Creadores Culturales (Investigadores-BBVA-2017; IN[17]_BBM_BAS_0066); M. Marcos was funded by Instituto de Salud Carlos III and Federación Española de Enfermedades Raras (PI16/01548); and J.L. Torres was funded by Junta de Castilla y León GRS (1587/A/17). F.J. Cubero is a Ramón y Cajal Researcher (RYC-2014-15242) and a Gilead Liver Research Scholar 2018, and his work is supported by the Ministerio de Economia y Competitividad Retos (SAF2016-78711), Comunidad de Madrid (S2017/BMD-3727), The Alan Morement Memorial Fund Cholangiocarcinoma Charity (2018/117), the European Cooperation in Science and Technology Action (CA17112), and the European Foundation for Alcohol Research (EA14/18). L. Morán is a Comunidad de Madrid fellow (S2017/BMD-3727). The CNIC is supported by the Ministerio de Ciencia, Innovación y Universidades and the Pro CNIC Foundation, and is a Severo Ochoa Center of Excellence (SEV-2015-0505).

The authors declare no competing financial interests.

Author contributions: G. Sabio conceived, designed, and supervised this project. E. Manieri, L. Herrera-Melle, and A. Mora designed the project, developed the hypothesis, performed the experiments, analyzed the data, and prepared figures. A. Tomás-Loba, L. Leiva-Vega, D.I. Fernández, M.E. Rodríguez, L. Morán, L. Hernández-Cosido, J.L. Torres, L.M. Seoane, F.J. Cubero, and M. Marcos participated in the experiments. E. Manieri, L. Herrera-Melle, and G. Sabio wrote the manuscript with input from all authors.

Arano
,
T.
,
H.
Nakagawa
,
R.
Tateishi
,
H.
Ikeda
,
K.
Uchino
,
K.
Enooku
,
E.
Goto
,
R.
Masuzaki
,
Y.
Asaoka
,
Y.
Kondo
, et al
2011
.
Serum level of adiponectin and the risk of liver cancer development in chronic hepatitis C patients
.
Int. J. Cancer.
129
:
2226
2235
.
Bakiri
,
L.
, and
E.F.
Wagner
.
2013
.
Mouse models for liver cancer
.
Mol. Oncol.
7
:
206
223
.
Bosch
,
F.X.
,
J.
Ribes
,
M.
Díaz
, and
R.
Cléries
.
2004
.
Primary liver cancer: worldwide incidence and trends
.
Gastroenterology.
127
(
5
,
Suppl 1
):
S5
S16
.
Böttner
,
A.
,
J.
Kratzsch
,
G.
Müller
,
T.M.
Kapellen
,
S.
Blüher
,
E.
Keller
,
M.
Blüher
, and
W.
Kiess
.
2004
.
Gender differences of adiponectin levels develop during the progression of puberty and are related to serum androgen levels
.
J. Clin. Endocrinol. Metab.
89
:
4053
4061
.
Bray
,
F.
,
J.
Ferlay
,
I.
Soerjomataram
,
R.L.
Siegel
,
L.A.
Torre
, and
A.
Jemal
.
2018
.
Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries
.
CA Cancer J. Clin.
68
:
394
424
.
Chen
,
H.P.
,
J.J.
Shieh
,
C.C.
Chang
,
T.T.
Chen
,
J.T.
Lin
,
M.S.
Wu
,
J.H.
Lin
, and
C.Y.
Wu
.
2013
.
Metformin decreases hepatocellular carcinoma risk in a dose-dependent manner: population-based and in vitro studies
.
Gut.
62
:
606
615
.
Cheung
,
O.K.
, and
A.S.
Cheng
.
2016
.
Gender Differences in Adipocyte Metabolism and Liver Cancer Progression
.
Front. Genet.
7
:
168
.
Combs
,
T.P.
,
A.H.
Berg
,
M.W.
Rajala
,
S.
Klebanov
,
P.
Iyengar
,
J.C.
Jimenez-Chillaron
,
M.E.
Patti
,
S.L.
Klein
,
R.S.
Weinstein
, and
P.E.
Scherer
.
2003
.
Sexual differentiation, pregnancy, calorie restriction, and aging affect the adipocyte-specific secretory protein adiponectin
.
Diabetes.
52
:
268
276
.
Dalamaga
,
M.
,
K.N.
Diakopoulos
, and
C.S.
Mantzoros
.
2012
.
The role of adiponectin in cancer: a review of current evidence
.
Endocr. Rev.
33
:
547
594
.
Fuente-Martín
,
E.
,
P.
Argente-Arizón
,
P.
Ros
,
J.
Argente
, and
J.A.
Chowen
.
2013
.
Sex differences in adipose tissue: It is not only a question of quantity and distribution
.
Adipocyte.
2
:
128
134
.
Ghebranious
,
N.
, and
S.
Sell
.
1998
.
Hepatitis B injury, male gender, aflatoxin, and p53 expression each contribute to hepatocarcinogenesis in transgenic mice
.
Hepatology.
27
:
383
391
.
Goodman
,
Z.D.
2007
.
Grading and staging systems for inflammation and fibrosis in chronic liver diseases
.
J. Hepatol.
47
:
598
607
.
Havel
,
P.J.
2002
.
Control of energy homeostasis and insulin action by adipocyte hormones: leptin, acylation stimulating protein, and adiponectin
.
Curr. Opin. Lipidol.
13
:
51
59
.
Heindryckx
,
F.
,
I.
Colle
, and
H.
Van Vlierberghe
.
2009
.
Experimental mouse models for hepatocellular carcinoma research
.
Int. J. Exp. Pathol.
90
:
367
386
.
Hui
,
L.
,
L.
Bakiri
,
A.
Mairhorfer
,
N.
Schweifer
,
C.
Haslinger
,
L.
Kenner
,
V.
Komnenovic
,
H.
Scheuch
,
H.
Beug
, and
E.F.
Wagner
.
2007
.
p38alpha suppresses normal and cancer cell proliferation by antagonizing the JNK-c-Jun pathway
.
Nat. Genet.
39
:
741
749
.
Kotani
,
K.
,
K.
Wakai
,
A.
Shibata
,
Y.
Fujita
,
I.
Ogimoto
,
M.
Naito
,
Y.
Kurozawa
,
H.
Suzuki
,
T.
Yoshimura
,
A.
Tamakoshi
, and
J.S.
Group
.
JACC Study Group
.
2009
.
Serum adiponectin multimer complexes and liver cancer risk in a large cohort study in Japan
.
Asian Pac. J. Cancer Prev.
10
(
suppl
):
87
90
.
Larsson
,
S.C.
, and
A.
Wolk
.
2007
.
Overweight, obesity and risk of liver cancer: a meta-analysis of cohort studies
.
Br. J. Cancer.
97
:
1005
1008
.
Luo
,
Z.
,
M.
Zang
, and
W.
Guo
.
2010
.
AMPK as a metabolic tumor suppressor: control of metabolism and cell growth
.
Future Oncol.
6
:
457
470
.
Ma
,
S.J.
,
Y.X.
Zheng
,
P.C.
Zhou
,
Y.N.
Xiao
, and
H.Z.
Tan
.
2016
.
Metformin use improves survival of diabetic liver cancer patients: systematic review and meta-analysis
.
Oncotarget.
7
:
66202
66211
.
Manieri
,
E.
, and
G.
Sabio
.
2015
.
Stress kinases in the modulation of metabolism and energy balance
.
J. Mol. Endocrinol.
55
:
R11
R22
.
Mao
,
X.
,
J.Y.
Hong
, and
L.Q.
Dong
.
2006
a
.
The adiponectin signaling pathway as a novel pharmacological target
.
Mini Rev. Med. Chem.
6
:
1331
1340
.
Mao
,
X.
,
C.K.
Kikani
,
R.A.
Riojas
,
P.
Langlais
,
L.
Wang
,
F.J.
Ramos
,
Q.
Fang
,
C.Y.
Christ-Roberts
,
J.Y.
Hong
,
R.Y.
Kim
, et al
2006
b
.
APPL1 binds to adiponectin receptors and mediates adiponectin signalling and function
.
Nat. Cell Biol.
8
:
516
523
.
Marquardt
,
J.U.
,
J.B.
Andersen
, and
S.S.
Thorgeirsson
.
2015
.
Functional and genetic deconstruction of the cellular origin in liver cancer
.
Nat. Rev. Cancer.
15
:
653
667
.
Naugler
,
W.E.
,
T.
Sakurai
,
S.
Kim
,
S.
Maeda
,
K.
Kim
,
A.M.
Elsharkawy
, and
M.
Karin
.
2007
.
Gender disparity in liver cancer due to sex differences in MyD88-dependent IL-6 production
.
Science.
317
:
121
124
.
Nishizawa
,
H.
,
I.
Shimomura
,
K.
Kishida
,
N.
Maeda
,
H.
Kuriyama
,
H.
Nagaretani
,
M.
Matsuda
,
H.
Kondo
,
N.
Furuyama
,
S.
Kihara
, et al
2002
.
Androgens decrease plasma adiponectin, an insulin-sensitizing adipocyte-derived protein.
Diabetes
.
51
:
2734
2741
.
Parkin
,
D.M.
,
F.
Bray
,
J.
Ferlay
, and
P.
Pisani
.
2005
.
Global cancer statistics, 2002
.
CA Cancer J. Clin.
55
:
74
108
.
Polyzos
,
S.A.
,
J.
Kountouras
,
C.
Zavos
, and
E.
Tsiaousi
.
2010
.
The role of adiponectin in the pathogenesis and treatment of non-alcoholic fatty liver disease
.
Diabetes Obes. Metab.
12
:
365
383
.
Sabio
,
G.
,
M.
Das
,
A.
Mora
,
Z.
Zhang
,
J.Y.
Jun
,
H.J.
Ko
,
T.
Barrett
,
J.K.
Kim
, and
R.J.
Davis
.
2008
.
A stress signaling pathway in adipose tissue regulates hepatic insulin resistance
.
Science.
322
:
1539
1543
.
Salguero Palacios
,
R.
,
M.
Roderfeld
,
S.
Hemmann
,
T.
Rath
,
S.
Atanasova
,
A.
Tschuschner
,
O.A.
Gressner
,
R.
Weiskirchen
,
J.
Graf
, and
E.
Roeb
.
2008
.
Activation of hepatic stellate cells is associated with cytokine expression in thioacetamide-induced hepatic fibrosis in mice
.
Lab. Invest.
88
:
1192
1203
.
Sharma
,
D.
,
J.
Wang
,
P.P.
Fu
,
S.
Sharma
,
A.
Nagalingam
,
J.
Mells
,
J.
Handy
,
A.J.
Page
,
C.
Cohen
,
F.A.
Anania
, and
N.K.
Saxena
.
2010
.
Adiponectin antagonizes the oncogenic actions of leptin in hepatocellular carcinogenesis
.
Hepatology.
52
:
1713
1722
.
Wang
,
Z.V.
, and
P.E.
Scherer
.
2016
.
Adiponectin, the past two decades
.
J. Mol. Cell Biol.
8
:
93
100
.
Whitehead
,
J.P.
,
A.A.
Richards
,
I.J.
Hickman
,
G.A.
Macdonald
, and
J.B.
Prins
.
2006
.
Adiponectin–a key adipokine in the metabolic syndrome
.
Diabetes Obes. Metab.
8
:
264
280
.
Yamauchi
,
T.
,
J.
Kamon
,
Y.
Minokoshi
,
Y.
Ito
,
H.
Waki
,
S.
Uchida
,
S.
Yamashita
,
M.
Noda
,
S.
Kita
,
K.
Ueki
, et al
2002
.
Adiponectin stimulates glucose utilization and fatty-acid oxidation by activating AMP-activated protein kinase
.
Nat. Med.
8
:
1288
1295
.
Yannakoulia
,
M.
,
N.
Yiannakouris
,
S.
Blüher
,
A.L.
Matalas
,
D.
Klimis-Zacas
, and
C.S.
Mantzoros
.
2003
.
Body fat mass and macronutrient intake in relation to circulating soluble leptin receptor, free leptin index, adiponectin, and resistin concentrations in healthy humans
.
J. Clin. Endocrinol. Metab.
88
:
1730
1736
.

Author notes

*

E. Manieri and L. Herrera-Melle contributed equally to this paper.

E. Manieri’s present address is Dana-Farber Cancer Institute, Boston, MA.

This article is distributed under the terms of an Attribution–Noncommercial–Share Alike–No Mirror Sites license for the first six months after the publication date (see http://www.rupress.org/terms/). After six months it is available under a Creative Commons License (Attribution–Noncommercial–Share Alike 4.0 International license, as described at https://creativecommons.org/licenses/by-nc-sa/4.0/).

Supplementary data