Hypoxia inducible factor-1 (HIF-1)/ENTPD2 maintained hepatocellular carcinoma an immune-suppressive microenvironment

Abstract

Introduction: Mortality rate of hepatocellular carcinoma (HCC) is extremely high due to late symptom presentation and lack of curative therapy. Understanding the biology of HCC will accelerate the development of more effective therapies. Hypoxia (Oxygen O2 deprivation) is frequently found in regions of HCC due to abnormal tumor vasculature. Methods: To study the effects of hypoxia on HCC cells, transcriptome sequencing of 6 human HCC cell lines exposed to 20% O2 and 1% O2 conditions and 16 pairs HCC and non-tumourous liver tissues were performed. An ectoenzyme, ectonucleoside triphosphate diphosphohydrolase 2 (ENTPD2/ CD39L1), was found to be consistently upregulated by hypoxia and over-expressed in HCC. ChIP assay confirmed the binding of HIF-1 to ENTPD2. ENTPD2 expression was evaluated in HIF-knockdown or knockout HCC cells established by shRNA or CRISPR-Cas9 approach. ENTPD2 knockdown and knockout HCC cells were established for functional studies. Liquid Chromatography-Mass Spectrometry (LC-MS) was employed to quantitate the level of extracellular metabolites (ATP, AMP) in the ENTPD2 knockdown or knockout HCC subclones. To study the functions of extracellular AMP in myeloid-derived suppressor cells (MDSCs), we purified MDSCs from HCC-bearing mice by magnetic bead sorting followed by in vitro culturing in the presence of absence of AMP. Syngeneic and hepatocarcinogen-induced mouse HCC models were employed to evaluate the effects of the HIF-1/ENTPD2 pathway in HCC. ENTPD2 inhibitors were used alone or in combination with immune therapies (anti-CTLA4 and anti-PD1) in HCC-bearing mice. Results: We first showed that hypoxia, through stabilizing hypoxia-inducible factor 1 (HIF-1), transcriptionally activated an ectoenzyme, ectonucleoside triphosphate diphosphohydrolase 2 (ENTPD2/ CD39L1) in HCC cells. We confirmed that ENTPD2 on hypoxic HCC cells hydrolyzes extracellular ATP to AMP, thereby preventing the differentiation of a population of immunesuppressive cells, MDSCs, and maintaining them in the tumor niche. Co-culturing of HCC-associated MDSCs with CFSE-labelled T cells revealed that MDSCs directly inhibited T cell proliferation. In vivo, we showed that knockdown or knockout of ENTPD2 in HCC cells reduced MDSC accumulation and increased T cell infiltration within the tumors. ENTPD2 inhibitors, POM1 and ARL67156, significantly reduced the number of tumor-associated MDSCs and repressed HCC progression in vivo. More excitingly, ENTPD2 inhibitor worked in synergy with anti-CTLA4 and anti-PD1 monoclonal antibodies to suppress HCC progression in mice. Clinically, we further found that ENTPD2 was frequently over-expressed in human HCC. IHC staining showed that MDSCs were preferentially found in hypoxic regions. Upregulation of ENTPD2 was found to be associated with poor aggressive HCC clinicopathological features such as presences of direct liver invasion, tumor microsatellite formation, and venous invasion, as well as absence of tumor encapsulation. Upregulation of ENTPD2 was also associated with shorter survivals in HCC patients. Conclusion: Our study indicates that ENTPD2 might be a good prognostic marker and therapeutic targets especially for HCC patients who are considering immune checkpoint inhibitors. Our study also reveals a novel molecular mechanism by which hypoxia maintains MDSCs at the tumor niche through HIF-1/ENTPD2 by AMP generation, thereby allowing cancer cells to escape immunosurveillance. Our study also highlights how the external microenvironment (hypoxia) affects metabolism and immunity of HCC.