Metabolic signatures of resistance to Sorafenib and Immune Checkpoint Inhibitors (ICIs) in Hepatocellular Carcinoma

Hepatocellular carcinoma (HCC) is the most common primary liver cancer and is associated with high mortality worldwide. Treatment options depend on disease stage and include surgical resection, liver transplantation, or local ablative approaches for early-stage disease, whereas systemic therapies are used for advanced or unresectable HCC. These include multikinase inhibitors, such as sorafenib, and immune checkpoint inhibitors (ICIs), including atezolizumab plus bevacizumab, durvalumab, and tremelimumab. However, both therapeutic strategies can be limited by the development of drug resistance, ultimately compromising treatment efficacy. In this project, we investigated two distinct therapeutic settings in HCC. In the first study, we used targeted metabolomics and untargeted lipidomics to dissect the metabolic adaptations underlying sorafenib response and resistance, and to identify circulating biomarkers of drug activity. In the second, ongoing study, we explored metabolic alterations associated with response to ICIs in patient plasma samples. In the sorafenib study, we demonstrated that sorafenib impairs mitochondrial oxidative metabolism in HCC cells, promoting metabolic reprogramming toward glycolysis and lipid remodeling. In sensitive cells, increased glycolytic flux leads to the accumulation of toxic methylglyoxal, which is detoxified through the glyoxalase pathway to produce D-lactate. D-lactate accumulation is associated with sorafenib activity and contributes to cell death through ferroptosis-related mechanisms. In contrast, resistant cells escape this vulnerability through metabolic rewiring involving the glycerol-3-phosphate pathway, which supports NAD+ regeneration, glycerolipid metabolism, and membrane remodeling, thereby protecting cells from ferroptotic damage. These mechanistic findings led to the identification and validation of circulating biomarkers in plasma from HCC patients treated with sorafenib: increased D-lactate was associated with treatment response, whereas elevated glycerol was linked to resistance and disease progression. In a separate, ongoing study focused on ICI-treated HCC patients, preliminary targeted metabolomics analyses of plasma samples indicate that the kynurenine/tryptophan pathway may be involved in treatment response. In responding patients, kynurenine levels increased from baseline, possibly reflecting IFNγ-driven activation of the indoleamine 2,3-dioxygenase pathway following T-cell reactivation. This suggests that kynurenine may represent a potential marker of an active antitumor immune response. In contrast, non-responders showed a lower kynurenine/tryptophan ratio and reduced tryptophan levels compared with baseline. In conclusion, our sorafenib study identifies D-lactate and glycerol as promising circulating biomarkers for monitoring sorafenib activity in HCC, with D-lactate associated with response and glycerol with resistance. In parallel, our ongoing investigation of ICI-treated patients highlights the kynurenine/tryptophan pathway as a potentially relevant metabolic axis associated with immunotherapy response. Further studies are required to validate these preliminary findings and to assess their clinical utility as predictive or pharmacodynamic biomarkers in HCC.