Metabolic reprogramming is critical for the polarization and function of tumor-associated macrophages (TAM) and hepatocarcinogenesis, but how this reprogramming occurs is unknown. Here, we showed that receptor-interacting protein kinase 3 (RIPK3), a central factor in necroptosis, is downregulated in hepatocellular carcinoma (HCC)–associated macrophages, which correlated with tumorigenesis and enhanced the accumulation and polarization of M2 TAMs. Mechanistically, RIPK3 deficiency in TAMs reduced reactive oxygen species and significantly inhibited caspase1-mediated cleavage of PPAR. These effects enabled PPAR activation and facilitated fatty acid metabolism, including fatty acid oxidation (FAO), and induced M2 polarization in the tumor microenvironment. RIPK3 upregulation or FAO blockade reversed the immunosuppressive activity of TAMs and dampened HCC tumorigenesis. Our findings provide molecular basis for the regulation of RIPK3-mediated, lipid metabolic reprogramming of TAMs, thus highlighting a potential strategy for targeting the immunometabolism of HCC.
Hepatocellular carcinoma (HCC) is one of the most lethal human malignancies worldwide. It commonly develops in patients with underlying chronic liver inflammation, is a typical inflammation-associated tumor (1). Among the infiltrating immune cells in the HCC microenvironment, tumor-associated macrophages (TAM) are one of the most prominent components, playing a crucial role in the hepatocarcinogenesis (2, 3). The majority of TAMs polarizes toward an M2 state and promotes HCC progression by secreting tumor-promoting and proangiogenic factors and by suppressing the activation of tumor-infiltrating T cells (4). In contrast, M1 TAMs possess antitumor capabilities by releasing reactive oxygen species (ROS), IL1β, and other proinflammatory cytokines (5, 6). Because of their opposing functions in tumor microenvironment (TME), selective deletion of M2 TAMs or switching to a M1 phenotype is considered a therapeutic approach to inhibit HCC progression (6).
Macrophage polarization is associated with metabolic reprogramming (7). For example, lipopolysaccharides (LPS)/IFNγ-activated M1 macrophages display enhanced glycolysis and fatty acid synthesis (FAS). Conversely, IL4/IL10-activated M2 macrophages have increased fatty acid oxidation (FAO; refs. 8–10). Inhibition of FAO or lysosomal lipolysis impairs IL4-induced M2 activation (8). In the TME, TAMs undergo metabolic reprogramming to survive and acquire protumoral or immunosuppressive properties in the nutrient- and oxygen-limited environment (11, 12). The FAS and PPAR pathways are activated in TAMs and contribute to tumor growth (7, 13, 14). However, the molecular mechanisms reprogramming lipid metabolism of TAMs remain unclear.
Receptor-interacting protein kinase 3 (RIPK3) is a serine/threonine kinase that plays pivotal roles in necroptosis and inflammation (15–17). RIPK3 can be initiated by diverse innate immune signaling pathways including toll-like receptors, tumor necrosis factor receptors, and IFN receptors, suggesting that RIPK3 may regulate immune responses (18). RIPK3 is downregulated in colorectal cancer infiltrating myeloid-derived suppressor cells (MDSC), which contributes to enhanced tumor-induced immune suppression (19). However, the expression and function of RIPK3 in TAMs have not yet been studied.
Here, we showed that RIPK3 was downregulated in the TAMs from mouse and human HCC samples. RIPK3 deficiency promoted the M2 polarization of TAMs by enhancing FAO in TAMs through ROS–Caspase1–PPAR pathway. Upregulation of RIPK3 or inhibition of FAO significantly reversed the immunosuppressive activity of TAMs and suppressed HCC tumorigenesis. These results indicated that RIPK3 regulates fatty acid metabolism and thus the plasticity of TAMs, suggesting a potential therapeutic target for HCC.
Materials and Methods
Human samples and databases
Five human HCC and adjacent tissues were collected fresh during surgery from the Department of Hepatobiliary Surgery of Xinqiao Hospital, Third Military Medical University (Chongqing, China). The stages of HCC were assessed by an experienced pathologist (Supplementary Table S1). The samples were frozen with liquid nitrogen or were fixed with formalin for indicated experiments. The inclusion criteria were: patients with advanced HCC and ages 45 to 65 years old. Healthy controls excluded diabetes, metabolic syndrome, hypertension, inflammatory bowel disease, liver disease, and cancer. The exclusion criteria were: (i) patients with chronic viral hepatitis (except hepatitis B virus–related HCC), autoimmune hepatitis, cholestatic diseases such as primary bile cholangitis or primary sclerosing cholangitis, and hereditary liver diseases that cause cirrhosis, such as hemochromatosis, Wilson disease, and alpha-1 antitrypsin deficiency; (ii) subjects who have taken antibiotics, probiotics, prebiotics, proton pump inhibitors, and laxatives, any other hepatotoxic drugs, or have received antiviral or immune therapy in the past 6 months; (iii) affected by diseases that may affect the composition of the intestinal microbiota (e.g., diabetes, celiac disease, inflammatory bowel disease, and diverticulosis); and (iv) patients with a history of cancer.
All experiments involving human subjects were conducted in accordance with local, national, and international regulations and were approved by the Ethics Committee of the Third Military Medical University (Chongqing, China). All patients provided written informed consent in accordance to the Declaration of Helsinki before enrolling in the study. The expression of RIPK3 in normal liver, cirrhosis, and HCC tissues were determined through analysis of Wurmbach Liver cancer datasets, which are available at Oncomine (http://www.oncomine.org/). All available The Cancer Genome Atlas (TCGA) data on HCC were obtained from the TCGA data portal (TCGA group). The mRNA RNA-seq data of patients with liver cancer (n = 440) were retrieved from http://www.cbioportal.org on February 15, 2017.
C57BL/6 wild-type (WT) mice were purchased from the Chinese Academy of Medical Sciences. RIPK3−/−-knockout (KO) mice were described previously (19). Mice were maintained at the specific pathogen–free animal facilities of the Clinical Medicine Research Center of Xinqiao Hospital, Third Military Medical University, under a 12-hour light cycle, and were given a regular chow diet (Chow). All animal studies were conducted in accordance with the national and international Guidelines for the Care and Use of Laboratory, and with the approval of the Animal Care and Use Committee of Third Military Medical University (Chongqing, China, AMUWEC2017294) and complied with the Declaration of Helsinki. HCC model was established by injection of diethylnitrosamine (DEN, catalog no. N0756, Sigma-Aldrich, diluted in PBS at 1:495 v/v i.p.) into 2-week-old male pups at a dose of 20 μg/g and followed by feeding with choline-deficient amino acid defined (CDAA) diet (catalog no. 518753, Dyets) from 4-week-old for 8 months. Then, the mice were sacrificed 8 months post CDAA administration and the livers were collected (20).
For in vivo treatments of etomoxir (20 mg/kg, catalog no. HY-50202, Med Chem Express) and GW9662 (1 mg/kg, catalog no. HY-16578, Med Chem Express), compounds were diluted in PBS and given intraperitoneally every 3 days until the mice were sacrificed.
Cell culture and treatments
Eight- to 10-week-old mice were euthanized and their femur and tibia were crushed and flushed with RPMI medium (Gibco) to collect bone marrow cells using a syringe. Bone marrow (BM) cells were cultured with RPMI1640 (Gibco) containing 10% (v/v) FBS (Gibco) and 20 ng/mL of mouse M-CSF (catalog no. 315-02, PeproTech) for 6 days to obtain bone marrow–derived macrophages (BMDM). Primary mouse peritoneal macrophages (PM) were obtained from the peritoneal exudates of 8-week-old mice. The peritoneal exudate cells were washed twice with PBS solution and adjusted to 1 × 106 cells/mL in RPMI1640 and cultured at 37°C in a 5% CO2 for 3 to 4 hours. The nonadherent cells were removed by washing twice with warm PBS. The purity of macrophages was analyzed by flow cytometry (FCM) as described below, which was higher than 95% (Beckman). To generate M1 and M2 macrophages, BMDMs or PMs were cultured in respective media supplemented with 10% FBS, and stimulated with LPS (20 ng/mL; catalog no. L4524, Sigma-Aldrich) + IFNγ (20 ng/mL; catalog no.315-05; PeproTech), IL4 (20 ng/mL, catalog no. 214-14; PeproTech), IL10 (20 ng/mL, catalog no. 210-10; PeproTech), or IL13 (20 ng/mL, catalog no.210-13; PeproTech) for 24 hours, respectively.
Murine H22 cells were purchased from Type Culture Collection of the Chinese Academy of Sciences in 2015 and cultured in RPMI1640 containing 10% FBS and antibiotics. H22 cells were routinely verified Mycoplasma-free using MycAwayTM-Color One-Step Mycoplasma Detection Kit (Yeasen Biotechnology) and the most recent date of testing was April 6, 2019. These cells were authenticated and certified by Chengdu Nuohe Biotech Co., Ltd. In the coculture assay, purified PMs or BMDMs (106 cells/well) were seeded in a 6-well plate and 2 × 105 H22 cells seeded onto the membranes of the top chambers of a 0.4-μm pore Boyden Chamber Insert (Corning). After coculture for 48 hours, macrophages were collected and treated with etomoxir (200 μmol/L), Wy14643 (200 μmol/L, catalog no. HY-16995, Med Chem Express), GW9662 (10 nmol/L), 5-(Tetradecyloxy)-2-furoic acid (TOFA; catalog no. HY-101068, Med Chem Express), GSK872 (3 μmol/L, catalog no. 530389, Merck Millipore), orlistat (100 μmol/L, catalog no. HY-B0218, Med Chem Express), N-acetyl-L-cysteine (NAC; 10 mmol/L, catalog no. S0077, Beyotime), VX765 (1 μmol/L, catalog no. HY-13205, Med Chem Express), and decitabine (0.02 μg/mL, catalog no. HY-10586, Med Chem Express) for indicated time intervals.
Cells were lysed by RIPA lysate buffer, and the cell lysates were incubated on ice for 30 minutes and centrifuged at 13,000 × g, 4°C for 15 minutes before the supernatant was collected. Western blot analysis was performed as described previously (19). The primary antibodies included RIPK3 (1:1,000; catalog no. 2283, ProSci), CPT1A (1:1,000; catalog no. 15148-1-AP, Proteintech), FASN (1:1,000; catalog no. 3180, CST), ACC (1:1,000; catalog no. 3676, Cell Signaling Technology), SCD1 (1:1,000; catalog no. 2794, Cell Signaling Technology), PPARα (1:500; catalog no. ab24509, Abcam), PPARγ (1:500; catalog no. 2435, Cell Signaling Technology), p-Stat3 (1:1,000; catalog no. 9145, Cell Signaling Technology), Stat3 (1:1,000; catalog no. 4904, Cell Signaling Technology), p-Stat6 (1:1,000; catalog no. 56554, Cell Signaling Technology), anti-Stat6 (1:1,000; catalog no. 5397, Cell Signaling Technology), ASC (1:200; catalog no. 67824, Cell Signaling Technology), NLRP3 (1:500; catalog no. 15101, Cell Signaling Technology), and Caspase1 (1:500; catalog no. AF1681, Beyotime) were used.
The fresh HCC tissues were mechanic dispersed and digested using Type IV collagenase (1 mg/mL; catalog no. C5138, Sigma-Aldrich) for 30 minutes at 37°C. After dissociation, the tumor suspensions were filtered, washed with cold PBS, and the single-cell suspension was collected for further FCM. For extracellular staining, 1 × 106 cells were preincubated in a mixture of PBS, 1% FBS, and FcgIII/IIR-specific antibody (1:100; catalog no. 130-092-575, Miltenyi Biotec) to block nonspecific binding. Then, cells were labeled with the indicated antibodies (1:100) for 30 minutes at 4°C. FCM was performed on BD FACS Canto II platforms and results were analyzed using FlowJo Software Version 10.0.7 (TreeStar). The panel of antibodies used in these experiments included CD11b (catalog no. 101208), F4/80 (catalog no. 123116), CD68 (catalog no. 333806), CD86 (catalog no. 105006), Gr1 (catalog no. 108426), CD11c (catalog no. 117308), MHC II (catalog no. 107622), CD103 (catalog no. 121414), CD206 (catalog no. 141704), CD3 (catalog no. 100206), CD4 (catalog no. 100412), and CD8α (catalog no. 100706) and CD19 (catalog no. 152406), all from BioLegend. Dead cells were excluded by using a Fixable Viability Dye EFluor 780 (catalog no. 65-0865-14, eBioscience) following the manufacturer's instructions. For staining of RIPK3, ARG1 (catalog no. 42284, GeneTex), iNOS (catalog no. MA5-17139, Thermo Fisher Scientific), IFNγ (catalog no. 505810, BioLegend), granzyme B (GzmB; catalog no. 515403, BioLegend), CPT1A, PPARα, CASPASE1, FASN, ACC, SCD1 and PPARγ, cells were stained surface markers, then fixed and permeabilized with Foxp3/Transcription factor staining buffer (catalog no. 00-5523-00, eBioscience), followed by primary antibodies Anti-rabbit IgG (H+L) Alexa Fluor 488 Conjugate (catalog no. 4412, CST) or Anti-rabbit IgG (H+L) Alexa Fluor 647 (catalog no. 4414, CST) staining according to the manufacturer's protocols. The proliferation and functional assay of CD8+ T cells were performed as described previously (19). CFSE probe was obtained from Dojindo (catalog no. C309). Bodipy558/568 (catalog no. D3835) probe was obtained from Thermo Fisher Scientific. Mito-Tracker Green (Mitogreen; catalog no. C1048) and DCFH-DA (catalog no. S0033) probes were from Beyotime.
Metabolic flux analysis
PMs from WT or RIPK3−/− mice were seeded into 6-well plate at the density of 2 × 106 cells/well in RPMI1640 medium supplemented with 10% FBS overnight. FAO metabolites were determined using [U-13C]-palmitate as the substrate (Cambridge Isotope Laboratories, CLM-3943; refs. 21, 22). Extraction of acetyl-CoAs and acetyl-CoAs from cells was carried out as previously described with some modifications (21). Briefly, 300 μL of extraction buffer containing isopropanol, 50 mmol/L KH2PO4, and 50 mg/mL BSA (25:25:1 v/v/v) acidified with glacial acetic acid was added to cells. Next, 19:0-CoA was added as an internal standard and lipids were extracted by incubation at 4°C for 1 hour at 1,500 rpm. Following this, 300 μL of petroleum ether was added and the sample was centrifuged at 12,000 rpm for 2 minutes at 4°C. The upper phase was removed. The samples were extracted two more times with petroleum ether as described above. To the lower phase remaining, 5 μL of saturated ammonium sulfate was added followed by 600 μL of chloroform:methanol (1:2 v/v). The sample was then incubated on a thermomixer at 450 rpm for 20 minutes at 25°C, followed by centrifugation at 12,000 rpm for 5 minutes at 4°C. Clean supernatant was transferred to fresh tube and subsequently dried in the SpeedVac under OH mode (Genevac). Dry extracts were resuspended in appropriate volume of methanol:water (9:1 v/v) prior to LC/MS analyses on a Thermo-Fisher U3000 DGLC coupled to Sciex QTRAP 6500 Plus.
Bone marrow reconstitution assay
Five- to six-week-old C56BL/6 or RIPK3-KO mice were irradiated with 9.5 Gy (MultiRad 225, Faxitron). Subsequently, 1 × 107 BM cells (diluted in100 μL PBS) from WT or RIPK3-KO mice were injected intravenously via the tail vein. HCC experiments were initiated 5 weeks after BM reconstitution. For 2 weeks after BM transfer, mice were given antibiotic water (containing 10 mg/mL trimethoprim and sulfamethoxazole). After 7 weeks, the peripheral blood was sampled, red blood cells were lysed in Ammonium-Chloride-Potassium Lysing Buffer (catalog no. A1049201, Thermo Fisher Scientific) for 10 minutes, and RIPK3 expression in macrophages were analyzed by FCM.
Transduction of macrophages
For production of knockdown lentivectors for PPARα, PPARγ, and Caspase1, three pLKD-CMV-EGFP-2A lentivectors containing short hairpin RNAs (shRNA) targeting different regions of PPARα (GCTTCTTTCGGCGAACTAT, GGAGCTGCAAGATTCAGAA, and GCTAAAGTACGGTGTGTAT), PPARγ (GGATGTCTCACAATGCCATCA, GGAAAGACAACGGACAAATCA, and GGAAGCCCTTTGGTGACTTTA), and Caspase1 (GGAAGAACAGAACAAAGAA, GCCCAAGCTTGAAAGACAA, and GCATTAAGAAGGCCCATAT) or a lentivector containing scrambled shRNA were obtained from GenePharma. Mouse BMDMs were plated at 1 × 106 cells per mL in 12-well plates and transduced with lentiviral particles (at multiplicity of infection of 100) with 5 μg/mL Polybrene (GenePharma). Macrophages were harvested and used for further experiments 3 days after transduction.
Mice tissues were 4% formaldehyde fixed for 15 minutes and tissue sections were then incubated in 10% normal goat serum for 1 hour. The cells were then incubated with the primary antibodies F4/80 (1:100, catalog no. ab6640, Abcam) and RIPK3 overnight at 4°C. The secondary antibodies Anti-rat IgG (H+L) Alexa Fluor 488 Conjugate (catalog no. 4416, Cell Signaling Technology) and Anti-rabbit IgG (H+L) Alexa Fluor 647 (catalog no. 4414, Cell Signaling Technology) were used at a 1:200 dilution for 1 hour. DAPI was used to stain the nucleus at a concentration of 100 ng/mL. Then, the sections were imaged on a Leica TCS SP5 Confocal Microscope (Leica Microsystems). The colocalization was assessed using Leica LASX (Microsystems Software).
Total RNA was extracted from cells using the TRizol Reagent Kit (catalog no. 12183555, Invitrogen) and the RNA concentration in the samples was measured using NanoDrop 2000 (Thermo Fisher Scientific). One microgram total RNA was converted to cDNA using the PrimeScript RT-PCR Kit (RR014A, Takara) according to the manufacturer's instructions. qPCR was performed using TB Green Fast qPCR Mix Kit (RR430A, Takara) on a CFX384 system (Bio-Rad), and the relative quantification (2−ΔΔCt) method was used to analyze gene expression. β-Actin mRNA was used as a reference for mRNA quantification. All qPCR experiments were repeated at least three times. Primer sequences were as follows: Arg1-F CTCCAAGCCAAAGTCCTTAGAG, Arg1-R AGGAGCTGTCATTAGGGACATC; Nos2-F GTTCTCAGCCCAACAATCCAAGA, Nos2-R GTGGACGGGTCGATGTCAC; Mrc1-F CTCTGTTCAGCTATTGGACGC, Mrc1-R CGGAATTTCTGGGATTCAGCTTC; IL10-F CCCATTCCTCGTCACGATCTC, IL10-R TCAGACTGGTTTGGGATAGGTTT; Ptgs2-F CTCCAAGCCAAAGTCCTTAGAG, Ptgs2-R AGGAGCTGTCATTAGGGACATC; Il1b-F TACGGACCCCAAAAGATGA, Il1b-R TGCTGCTGCGAGATTTGAAG; Cpt1a-F TGGCATCATCACTGGTGTGTT, Cpt1a-R GTCTAGGGTCCGATTGATCTTTG; Cpt1b-F GCACACCAGGCAGTAGCTTT, Cpt1b-R CAGGAGTTGATTCCAGACAGGTA; Acadvl-F CTACTGTGCTTCAGGGACAAC, Acadvl-R CAAAGGACTTCGATTCTGCCC; Cd36-F GAGGCATTCTCATGCCAGT, Cd36-R ACGTCATCTGGGTTTTGCAC; actin-F TCCATCATGAAGTGTGACGT, actin-R TACTCCTGCTTGCTGATCCAC; Hadh-F TCAAGCATGTGACCGTCATCG, Hadh-R TGGATTTTGCCAGGATGTCTTC; Fabp4-F AAGGTGAAGAGCATCATAACCCT, Fabp4-R TCACGCCTTTCATAACACATTCC; Ppara-F TACGGACCCCAAAAGATGA, Ppara-R TGCTGCTGCGAGATTTGAAG; Pparg-F TATGGAGTGACATAGAGTGTGCT, Pparg-R CCACTTCAATCCACCCAGAAAG; Fasn-F CTCTGTTCAGCTATTGGACGC, Fasn-R CGGAATTTCTGGGATTCAGCTTC; Acaca-F CTCCCGATTCATAATTGGGTCTG, Acaca-R TCGACCTTGTTTTACTAGGTGC; Scd1-F TCCATCATGAAGTGTGACGT, Scd1-R TACTCCTGCTTGCTGATCCAC; Acsl1-F TGCCAGAGCTGATTGACATTC, Acsl1-R GGCATACCAGAAGGTGGTGAG; Dgat1-F TCCGTCCAGGGTGGTAGTG, Dgat1-R TGAACAAAGAATCTTGCAGACGA; Lpl-F GGGAGTTTGGCTCCAGAGTTT, Lpl-R TGTGTCTTCAGGGGTCCTTAG; Lipe-F TTCAACACACTCTATCACTGGC, Lipe-R AGAAGCGTTTGCGGTACTCAT; and Abhd5-F TGGTGTCCCACATCTACATCA, Abhd5-R CAGCGTCCATATTCTGTTTCCA.
Macrophages were seeded at 2 × 105 cells/well in an ultra-low attachment 12-well plate (Corning) for 3 to 4 hours to allow adherence to the plate. Then, opsonized (incubation in 14% normal human serum for 30 minutes at 37°C) pHrodo GFP+ Pseudomonas aeruginosa (P. aeruginosa) were added to macrophages at a ratio of 10:1 and incubated at 37°C for 15 minutes. Phagocytosis was terminated by the addition of 1 mL ice-cold sterile PBS. Cells were harvested and washed in ice-cold PBS three times and subsequently analyzed by FCM as described above.
Measurements of cellular ATP
Cellular ATP was measured by an ATP Assay Kit (catalog no. S0026, Beyotime) according to the manufacturer's protocol.
Seahorse XFp metabolic assays
Macrophages were seeded at 2 × 104 cells/well in 96-well plates for 3 to 4 hours to allow adherence to the plate. Then, they were treated with or without IL4 (20 ng/mL, diluted in RPMI1640), etomoxir (200 μmol/L, diluted in RPMI1640), TOFA (20 μmol/L, diluted in RPMI1640), and orlistat (100 μmol/L, diluted in RPMI1640) overnight, the cells were changed to unbuffered assay media, and incubated in a non-CO2 incubator at 37°C for 1 hour. Oxygen consumption rates (OCR) were measured using a XF96 extracellular flux analyzer (Seahorse Bioscience) after the sequential addition of oligomycin (1 μmol/L, diluted in Seahorse XFp Base media, 495455, Sigma), FCCP (1.5 μmol/L, diluted in Seahorse XFp Base media, C2920, Sigma), and antimycin/rotenone (2 μmol/L, diluted in Seahorse XFp Base media, 35410/R8875, Sigma).
The concentration of IL1β in the supernatants of the WT and KO TAMs was determined with ELISA (catalog no. PI301, Beyotime) according to the manufacturer's instructions.
Caspase-1 activity assay
Caspase-1 activity was measured using Caspase 1 Assay Kit (catalog no. ab39412, Abcam) according to the manufacturer's protocol.
Statistical parameters and n values are indicated in figure legends and below. All results were confirmed in at least three independent experiments and were expressed as means ± SEM. Student t test and one- or two-way ANOVA were used for calculation statistical significance with GraphPad Prism software (version 7.0). For correlation analysis, Pearson correlation coefficient was used. P < 0.05 was considered statistically significant.
RIPK3 was downregulated in HCC-associated TAMs
To evaluate RIPK3 expression in the HCC tissues, we examined the expression of RIPK3 in human tissues with TCGA and the Oncomine database, and found that RIPK3 expression was dramatically increased in cirrhosis and decreased in HCC (Fig. 1A; Supplementary Fig. S1A). We also collected HCC patient tissues and observed increased accumulation of TAMs, but decreased expression of RIPK3 in these TAMs (Fig. 1B and C; Supplementary Table S1). In the HCC mouse model established with DEN plus CDAA diet (Fig. 1D; Supplementary Fig. S1B; ref. 20), the expression of RIPK3 and the phosphorylation of its downstream target Mixed Lineage Kinase Domain Like Pseudokinase (p-MLKL) in liver tissues was increased during cirrhosis, whereas they were significantly suppressed in the HCC tissues (Supplementary Fig. S1C). In addition, macrophages significantly accumulated in HCC tissues compared with adjacent tissues (Fig. 1E). These TAMs were predominantly a M2 phenotype (CD206+ Arg1+; Supplementary Fig. S1D and S1E). RIPK3 was lower in TAMs compared with Kupffer cells (KC; Fig. 1F and G). Of note, RIPK3 was upregulated in M1 macrophages, but did not show significant change after IL4, IL10, and IL13 stimulation (Supplementary Fig. S1F and S1G). Coculture with H22 cells and TGFβ treatment significantly reduced the expression of RIPK3 in BMDMs (Fig. 1H; Supplementary Fig. S1F and S1H). These results demonstrated that RIPK3 expression was downregulated in HCC tissues and TAMs.
RIPK3 deficiency exacerbated DEN plus CDAA–induced HCC
To investigate the role of RIPK3 in hepatocarcinogenesis, RIPK3−/−(KO) mice were used. After CDAA diet treatment, more macroscopic tumors, intensified hepatocyte ballooning, and inflammatory cell infiltration were observed in KO mice livers as compared with WT, whereas no tumors developed in the Chow diet WT or KO mice (Fig. 2A and B). Consistently, KO mice exhibited increased serum concentrations of alanine aminotransferase (ALT) and aspartate aminotransferase (AST; Fig. 2C), as well as increased liver/body weight ratio (Fig. 2D), suggesting that RIPK3 was protective against hepatocarcinogenesis.
Given the importance of immunoediting in tumorigenesis, we next investigated the accumulation of immune cells within HCC tissues. Compared with WT HCC tissues, the KO group had increased TAM infiltration (Fig. 2E–G). However, the fractions of T cells, B cells, dendritic cells, and MDSC did not show significant change in KO mice (Fig. 2F; Supplementary Fig. S2A–S2E). We also observed a higher fraction of KCs in adjacent tissues from KO mice with CDAA diet treatment, whereas other types of inflammatory cells remain unchanged (Supplementary Fig. S2F–S2L). We next asked whether the exacerbated hepatocarcinogenesis in KO mice was due to loss of RIPK3 in hematopoietic cells because RIPK3 was mostly expressed in TAMs or KCs (Fig. 1F). We generated chimeras by injection of either WT or KO BM cells into irradiated WT or KO recipients (Supplementary Fig. S2M and S2N). Subsequently, HCC tumors were induced in these animals via DEN plus CDAA. The recipients exhibited more macroscopic tumors, more severe liver injury, and increased liver/body weight ratios if they were engrafted with myeloid cells from RIPK3-KO donors, compared with that from WT donors (Fig. 2H–J). Moreover, the percentage of the infiltrated TAMs was higher in the mice that received KO BM cells, compared with that of mice that received WT BM (Fig. 2K). Collectively, these data indicated that deficiency of RIPK3 in TAMs contributed to the accumulation of TAMs and the tumorigenesis of HCC.
RIPK3 deficiency promoted the immunosuppressive activity of TAMs
We next hypothesized that RIPK3 could affect the phenotype of TAMs. As expected, TAMs and PMs from KO mice expressed higher M2 markers such as Arg1 and CD206, but lower M1 markers including inducible nitric oxide synthase (iNOS or NOS2), MHC II, CD11c, and CD86 (Fig. 3A; Supplementary Fig. S3A). RIPK3 inhibitor GSK872 also significantly enhanced CD206 and Arg1 expression but decreased CD11c expression in PMs (Supplementary Fig. S3B and S3C). In addition, the coculture of BMDMs with H22 cells elevated the mRNA expression of Arg1, Mrc1, and Ptgs2 in macrophages (Supplementary Fig. S3D–S3H). We also assessed Stat3 and Stat6 signaling pathways, involved in the polarization of M2 phenotype (23), which remained unaffected in KO TAMs (Supplementary Fig. S3I). These data indicated that RIPK3-deficient macrophages were prone to M2 polarization.
PD-L1 contributes to the immunosuppressive activity, while PD-1 expression correlates with impaired phagocytosis of TAMs (24). We found that RIPK3-KO TAMs expressed significantly higher PD-L1 and PD-1 than WT (Fig. 3B). Similar results were also observed in TAMs from the H22 coculture system (Supplementary Fig. S3J). The Ripk3-KO TAMs had decreased phagocytic activity compared with WT (Fig. 3C). Coculture of RIPK3-KO TAMs significantly dampened the proliferation and activation of CD8+ T cells (Fig. 3D–F). These results demonstrated that TAMs obtained more potent immunosuppressive activity and showed impaired phagocytosis activity in RIPK3-KO mice.
RIPK3 deficiency–induced M2 activation depended on fatty acid metabolism
Given that lipid metabolism determines the polarization of macrophages (25), we wondered whether RIPK3 deficiency–induced M2 polarization was regulated by lipid metabolism. Compared with WT TAMs, more lipid droplets deposition was observed in RIPK3-KO TAMs (Fig. 4A). Coculture of TAM with H22 cells induced increased accumulation of lipid droplets in RIPK3-KO TAMs (Fig. 4B). Notably, FCM revealed that TAMs had more mitochondria than KCs, and RIPK3-KO TAMs contained significantly more mitochondria than WT TAMs (Fig. 4C). Consistent with these results, TAMs possessed more lipid droplets and mitochondria than KCs in clinical HCC samples (Supplementary Fig. S4A and S4B). In metabolic assay experiments, RIPK3-KO TAMs had enhanced mitochondrial OCR and markedly increased spare respiratory capacity (SRC), which was similar to IL4-treated WT macrophages (Fig. 4D and E). ATP production was higher in RIPK3-KO TAMs, suggesting increased oxidative phosphorylation (Fig. 4F).
Using metabolic flux assays, we observed that the incorporation of 13C from [U-13C]-palmitate into fatty acids intermediates was higher in RIPK3-KO than in WT TAMs, indicating enhanced FAO in RIPK3-KO TAMs (Fig. 4G). CPT1 is a well-known rate-limiting enzyme in FAO (26). We found that the expression of Cpt1A was higher in KO TAMs compared with WT TAMs (Fig. 4H–J). Genes involved in FAO, including Cpt1a, Cpt1b, Acadvl, and Hadh, were upregulated in KO TAMs (Fig. 4J). Of interest, we found that the expression of key genes in FAS (Fasn, Acsl1, Acaca, Scd1, and Dgat1) and lipolysis (Lpl, Lipe, and Abhd5) were also enhanced in KO TAMs (Supplementary Fig. S4C–S4F). In addition, the CPT1 inhibitor etomoxir reduced OCR and SRC, and decreased ATP production in KO TAMs (Fig. 4K and L). Both the FAS inhibitor TOFA and the lipolysis inhibitor orlistat abolished ATP production in KO TAMs (Supplementary Fig. S4G–S4J). Etomoxir and orlistat inhibited the expression of Arg1, and enhanced the expression of iNOS in KO TAMs (Fig. 4M; Supplementary Fig. S4K), delineating the link between fatty acid metabolism and M2 activation. Together, these results indicated that fatty acid metabolism was essential for RIPK3 deficiency–induced M2 activation.
RIPK3 deficiency activated PPARs to regulate fatty acid metabolism
PPAR pathway is a well-known signaling pathway involved in fatty acid metabolism (27). KO TAMs showed significantly elevated expression of PPARα and PPARγ compared with WT TAMs (Fig. 5A–C), accompanied with their downstream molecules such as Ppargc1a, Ucp2, Cd36, and Fabp4 (Fig. 5C; Supplementary Fig. S5A). Consistently, GSK872 moderately increased Ppara and Pparg in macrophages (Supplementary Fig. S5B). Inhibition of PPARs by GW9662 markedly decreased OCRs and SRC (Fig. 5D). Knockdown (KD) of Ppara and Pparg or GW9662 treatment also decreased the expression of genes involved in FAO pathway including Cpt1a, Cpt1b, Hadh, Acadvl, and PPAR target genes Cd36 and Fabp4 (Fig. 5E; Supplementary Fig. S5C and S5D). Conversely, the activation of PPARs by Wy-14643 induced CD206 and Arg1 expression, while KD or inhibition of PPARs suppressed CD206 and Arg1 expression in KO macrophages (Fig. 5F; Supplementary Fig. S5E and S5F). Administration of etomoxir, TOFA, and orlistat also inhibited the expression of Ppara and Pparg (Fig. 5G). These findings demonstrated that RIPK3 reduction in TAMs promoted fatty acid metabolism via activating PPAR pathway.
RIPK3 suppressed PPARs through ROS–Caspase1 signaling
RIPK3 signaling causes mitochondrial ROS generation, which is essential for necroptosis (28, 29). We found that TAMs produced lower ROS compared with adjacent KCs, and that KO TAMs produced even lower ROS than WT (Fig. 6A and B). KO PMs also showed a significantly decreased ROS (Supplementary Fig. S6A), suggesting that RIPK3 was essential for macrophage ROS production. The antioxidant NAC markedly induced the expression of PPARα, PPARγ, and Cpt1A in macrophages (Fig. 6C and D). However, the mRNA and protein expression of RIPK3 was not altered after NAC administration (Supplementary Fig. S6B and S6C), demonstrating that ROS was the downstream of RIPK3. NAC treatment significantly increased Arg1 and CD206 expression with a concomitant decrease in iNOS and CD11c in macrophages (Fig. 6E; Supplementary Fig. S6D). Moreover, coculture with NAC-pretreated TAMs markedly inhibited the proliferation and activation of CD8+ T cells (Fig. 6F–H; Supplementary Fig. S6E–S6G), indicating that lower ROS facilitated M2 polarization.
RIPK3 is required for caspase-1–dependent IL1β activation in BMDMs (30). Consistent with previous reports, the expression of caspase-1, NLR family pyrin domain containing 3 protein (NLRP3), apoptosis-associated speck-like protein containing caspase-recruitment domain (CARD) (ASC), and the ASC substrate IL1β were lower in KO TAMs compared with WT (Fig. 6I and J; Supplementary Fig. S6H–S6J). We found that NAC markedly reduced macrophage caspase-1 activity induced by H22 coculture (Fig. 6J), therefore, RIPK3-dependent ROS production was required for caspase-1–dependent inflammasome activation and IL1β processing in TAMs.
Caspase-1 mediates PPARγ cleavage and impairs the transcriptional activity of PPARγ (31). Here, the cleavage of PPARα and PPARγ was reduced in KO TAMs, compared with that in WT TAMs (Fig. 6K; Supplementary Fig. S6K). The NAC and caspase-1 inhibitor VX765 or caspase-1 KD blunted the cleavage of PPARα and PPARγ in TAMs, and suppressed the proliferation and activation of cocultured CD8+ T cells (Fig. 6K–O; Supplementary Fig. S6L–S6N). These results indicated that ROS-mediated caspase-1 cleavage was essential for RIPK3-regulated PPAR pathway and TAM polarization.
Upregulation of RIPK3 reversed TAM polarization and attenuates HCC
RIPK3 is silenced in a methylation-dependent manner in cancer cells (32). We wondered whether RIPK3 hypomethylation would affect lipid metabolism and polarization of macrophages. Decitabine is a hypomethylating drug for patients with liver metastasis or colorectal cancer in a phase II clinical trial (NCT number: NCT02264873). We found that decitabine treatment in macrophages increased the protein and mRNA expression of RIPK3 and reduced the expression of PPARα and PPARγ, but increased their cleaved forms. In addition, Cpt1a and Cpt1b were suppressed by decitabine treatment, indicating that RIPK3 inhibited FAO in macrophages (Fig. 7A–C). Decitabine also increased iNOS and decreased Arg1 in TAMs (Fig. 7B), indicating that upregulation of RIPK3 reversed TAM polarization.
We next speculated that FAO blockade in TAMs in vivo would alleviate HCC tumorigenesis. As expected, etomoxir and GW9662 treatment significantly suppressed tumorigenesis and protected against DEN plus CDAA diet–induced liver injury (Fig. 7D–F). Etomoxir and GW9662 effectively reversed the accumulation of TAMs in HCC tissues (Fig. 7G and H) and dramatically decreased Arg1 and increased iNOS in KO TAMs (Fig. 7I and J), suggesting that ablation of FAO switched KO TAM polarization from M2 to M1. Finally, TCGA analysis showed that RIPK3 expression negatively correlated with CD68, PPARA, and PPARG, while positively correlated with Caspase1 (Supplementary Fig. S7). These results demonstrated that RIPK3 regulated the polarization of TAMs and hepatocarcinogenesis via Caspase1-PPAR pathway (Fig. 7K).
TAMs play an essential role in tumor growth, invasion, angiogenesis, and metastasis (6). TAMs undergo metabolic reprogramming to adapt the oxygen- and nutrient-limited microenvironment (25). In our study, we found that RIPK3 was downregulated in HCC and TAMs. RIPK3 deficiency promoted the infiltration and M2 differentiation of TAMs by metabolic reprogramming of fatty acids in TME, which contributed to hepatocarcinogenesis.
RIPK3 is a well-known regulator of necroptosis and inflammation. Its expression increases in patients with chronic hepatitis and multiple liver injuries, and in inflammation models (33–35). RIPK3 plays a protective role in TAK1-induced HCC model (36), as well as in high-fat diet induced liver injury and steatosis (35). Our current data suggested that RIPK3 expression was elevated in cirrhosis but decreased in human HCC. The exacerbated hepatocarcinogenesis and increased ALT/AST in RIPK3-KO mice after been fed with CDAA diet indicated opposing expression modes of RIPK3 during HCC development.
RIPK3 activates pyruvate dehydrogenase complex E3 subunit thus increased anaerobic glycolysis (37), suggesting that RIPK3 played a role in glucose metabolism. Because macrophages M1 activation largely relies on aerobic glycolysis, RIPK3 may favor M1 polarization and enhance inflammation by promoting glycolysis in macrophages. However, TAMs are mostly M2 as they experience hypoxia and insufficient nutrients (especially glucose) in TME, thus are forced to adapt their metabolism to fulfill their energy requirements by enhancing FAO and glutaminolysis (25, 38). ROS production plays an important role in RIPK3-induced necroptosis (15). Our data showed that RIPK3-dependent ROS production was required for the activation of caspase-1 and cleavage of PPARα and PPARγ in TAMs. We found that inhibition of ROS generation was sufficient to induce M2 phenotype and PPARα and PPARγ expression in WT cells. In addition, FAO, FAS, and lipolysis were enhanced in RIPK3-KO TAMs. The fatty acids generated from lipolysis of triacylglycerols and FAS were the source for FAO. They could also function as ligands for PPARs, which regulated the expression of genes encoding molecules used for FAO and were critical for M2 activation (13). Therefore, these findings indicated that RIPK3 deficiency in TAMs facilitated fatty acid metabolism and M2 activation through ROS–caspase1–PPAR pathway in macrophages.
RIPK3-derived necroptosis and pyroptosis facilitate inflammation (39). Its effect on inflammation in the TME was carefully considered in our study. First, we monitored immune cell infiltration during the hepatocarcinogenesis and found that RIPK3 deficiency correlated with the accumulation and M2 polarization of TAMs. Second, we observed a significantly suppressed MLKL phosphorylation in the tumor tissues. Third, we proved that RIPK3-dependent ROS production was required for caspase-1–dependent inflammasome activation and IL1β processing in TAMs, that is, RIPK3 deficiency blunted the ROS-mediated inflammasome activation in TAMs. Fourth, RIPK3-deficient TAMs showed enhanced PPAR activation, fatty acid uptake, FAS, and FAO, but reduced glycolysis. Therefore, RIPK3-mediated necroptosis and pyroptosis contributed the initiation of hepatitis, while during hepatocarcinogenesis TME-derived factors such as TGFβ led to RIPK3 deficiency which promoted FAO in TAMs.
In conclusion, the loss of RIPK3 in TAMs reprograms the fatty acids metabolism by ROS–caspase1–PPAR pathway. This signaling axis played an essential role in the accumulation and M2 polarization of TAMs in TME thereby accelerated HCC growth. Importantly, targeting FAO decreased TAMs infiltration and differentiation toward M2 phenotype and inhibited HCC tumorigenesis in RIPK3-KO mice. Our data also demonstrated that hypomethylation of RIPK3 by decitabine inhibits FAO and reverses TAMs polarization. Therefore, decitabine may be a HCC treatment by modulating fatty acid metabolism and enhancing the antitumor immunity of macrophages. These findings provide molecular basis for the RIPK3-mediated lipid metabolic reprogramming and polarization of TAMs, which suggest a potential strategy for tumor immunotherapy and fatty acid metabolism–targeted treatment.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Conception and design: L. Wu, Y. Li
Development of methodology: L. Wu, X. Zhang, L. Zheng, H. Zhao, G. Yan, Q. Zhang, J. Zhang, J. Wang, R. Xin, L. Jiang, Q. Chen, G. Shui
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): L. Zheng, Y. Zhou, J. Lei, J. Zhang, L. Jiang, S.M. Lam
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): L. Wu, H. Zhao, G. Yan, Q. Zhang, Y. Zhou, J. Lei, J. Zhang, S.M. Lam, G. Shui, Y. Li
Writing, review, and/or revision of the manuscript: L. Wu, Y. Li
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): L. Wu, X. Zhang, L. Zheng, J. Zhang, J. Wang, R. Xin, L. Jiang, J. Peng, Q. Chen, H. Miao
Study supervision: Y. Li
The authors thank Hua Yu and Yi Zhou for their assistance in Western blot analysis and flow cytometry procedure. This work was supported by the National Natural Science Foundation of China (81472435, 81671573, and 81920108027 to Y. Li and 81901624 to L. Wu), Frontiers in Military Medicine (2018YQYLY008 to Y. Li and 2019XQN10 to L. Wu), and Science Foundation for Post Doctorate Research of China (2017M613344 to L. Wu). The authors sincerely acknowledge the anonymous reviewers for their insights and comments to further improve the quality of this article.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.