Mobilization of myeloid-derived suppressor cells through the CXCL1-CXCR2 signaling facilitates tumor recurrence after liver surgery

Abstract

Background & aims Liver cancer is the sixth most commonly diagnosed cancer and the fourth leading cause of cancer death worldwide. Curative surgery is a first-line treatment for liver cancer patients, although tumor recurrence remains a concern for long-term survival. Surgical stress is a significant risk promoting tumor recurrence after surgery, proven in mouse models and clinical trials. Immunotherapy is one of the therapeutic strategies for HCC, as the immune system is a major defense against cancer by attacking dysplastic cells. Emerging evidence suggests that myeloid-derived suppressor cells (MDSCs) accumulate in the tumor microenvironment (TME) by suppressing immune resistance. However, its role in the hepatic microenvironment post-surgery and surgical stress-induced hepatocellular carcinoma (HCC) recurrence remains poorly defined. Materials and Methodology The phenotypes of MDSCs and their clinical association were analyzed by flow cytometry in Study Cohort 1 of 49 HCC patients after liver resection. The expression and regulation of several chemokines were detected to explore the chemokines related to HCC recurrence. The distribution characteristics and immunosuppressive function of MDSCs subsets in the peripheral blood and paired non-tumor/tumor tissues were investigated in Study Cohort 2 of 176 HCC patients who underwent liver surgery. The underlying mechanisms of surgical-induced MDSCs recruitment were explored in the tumor-bearing mouse model, and the mouse model underwent hepatic ischemia/reperfusion injury (IRI) together with major hepatectomy. The role of MDSCs accumulation on the efficacy of anti-PD-1 immunotherapy was investigated in Study Cohort 3 of 22 patients after Nivolumab treatment. Results The significant enrichment of circulating MDSCs was observed in HCC patients with tumor recurrence before surgery and at the early time points during follow-up and was associated with poor prognosis. The upregulated CXCL1 in the tumor was associated with HCC recurrence. In non-tumor tissues, the CXCL1 expression was positively correlated with MDSCs. Moreover, overexpression of CXCR2, the receptor of CXCL1, was detected on PMN-MDSCs with a high level of IL-10. The enrichment of PMN-MDSCs was especially observed in non-tumor tissues in HCC patients with tumor recurrence. In the hepatic ischemia/reperfusion injury plus hepatectomy mouse model, upregulated CXCL1 was also found in the liver at 6hrs post-surgery. The increase of circulating M-MDSCs in blood and increases in hepatic PMN-MDSCs were detected 24hrs post-surgery, indicating the CXCL1-CXCR2 signaling might induce the mobilization of immunosuppressive CXCR2+ PMN-MDSCs in the liver post-surgery. Furthermore, the CXCR2+ PMN-MDSCs showed a stronger immunosuppressive function with the overexpression of PD-L1 in patients with progressive disease after the anti-PD-1 therapy. Conclusions The accumulation of circulating MDSCs before resection is associated with poor prognosis in HCC, indicating its predictive value in HCC recurrence. The proportion of CXCR2+ PMN-MDSCs with stronger immunosuppressive function is accumulated in non-tumor tissues to inhibit T cell infiltration and diminish the efficacy of anti-PD-1 therapy. Surgical stress-induced upregulated CXCL1 in non-tumor tissues after liver surgery results in the CXCR2+ PMN-MDSCs mobilization via the CXCL1/CXCR2 signaling. Targeting MDSCs mobilization via the CXCL1-CXCR2 signaling may provide a novel strategy to enhance the immunotherapeutic efficacy of HCC.