ZEB1 stratifies the response to Sorafenib and Mdivi-1 combination therapy in hepatocellular carcinoma

Abstract

Primary liver cancer is one of the most frequently diagnosed and deadliest cancers. Zinc finger E-box binding homeobox 1 (ZEB1) is negative prognostic factor in liver cancer by promoting therapy resistance and tumorigenesis. Interfering in pathways of cellular metabolism emerges as a potent strategy to overcome tumor cells resistance to therapy. Our study aims to investigate the apoptotic potential of pharmacological inhibition of mitochondrial metabolism in primary liver cancer and whether this strategy can reduce ZEB1-associated resistance to standard of care chemotherapy Sorafenib. ZEB1 mRNA levels in patient samples and cancer cell lines were computationally screened using public datasets. Transcript and protein abundance of ZEB1 and regulators of mitochondrial fission and fusion were quantified in patient-matched tumor and non-tumor tissues of hepatocellular carcinoma (HCC) and cholangiocellular adenocarcinoma (CCA) from our clinic and common liver cancer cell lines. Functional assays on cell models with varying ZEB1 expression exposed to mitochondrial division inhibitor Mdivi-1 included mitochondrial mass quantification, mitochondrial membrane potential examination, apoptosis, extracellular flux, and cell growth analyses. Mdivi-1 treatment in human-physiological achievable concentration effectively induced apoptosis in all tested cell models. ZEB1 expression was heightened in younger patients, and dynamin-related protein 1 (Drp1) protein abundance was elevated in a subgroup of tumor tissue compared to healthy tissue. Cancer cell lines with high ZEB1 expression showed increased mitochondrial fission marker Drp1 and larger total mitochondrial mass, preserved membrane potential with reduced mitochondrial ATP production and respiration, resulting in an overall reduced mitochondrial fitness. Pharmacological inhibition of mitochondrial fission using Mdivi-1 reduced HCC resistance to Sorafenib in ZEB1-driven liver cancer. A subset of HCC cell lines with elevated ZEB1 levels exhibit increased mitochondrial mass and reduced metabolic activity. Targeting mitochondrial division in HCC by treatment with Mdivi-1 in combination with Sorafenib demonstrates a synergistic therapeutic effect in hepatocellular carcinoma (HCC) cell lines characterized by high ZEB1 expression. Further in vivo validation is needed to confirm these findings and evaluate this potential combined treatment option.

Keywords: EMT; Hepatocellular carcinoma; Mdivi-1; Mitochondrial metabolism; Sorafenib; ZEB1.

Fig. 1

Bioinformatic analysis of ZEB1 expression correlated with genes of citric acid cycle and fatty acid metabolism, ZEB1 and Drp1 expression and protein abundance in liver cancer cells. (A,B) KEGG metabolic pathways database was used for pathway annotation and transcriptional data were extracted from The Cancer Genome Atlas, top 5 positively correlated genes in HCC and CCA. Statistical analysis was conducted using Spearman’s method. Statistically significant was defined as p ≤ 0.05 and effect size ‘r’ defining a medium effect from 0.3–0.49 and a large effect ≥ 0.5. (C) ZEB1 expression in our liver cancer cell lines on transcriptomic and protein levels. Differentiation of high and low ZEB1 cell lines was classified by at least 3-fold higher transcriptional expression of ZEB1 compared to low ZEB1 cell lines. (D) DRP1 expression in our liver cancer cell lines on transcriptomic and protein levels. The full western blot images can be obtained in the supplementary dataset.

Fig. 2

ZEB1 and Drp1 protein and RNA abundance in healthy and tumor liver tissue samples of patients. (A) ZEB1 and Drp1 expression on RNA levels. Normal liver tissue classified as ‘Healthy’ comprises of 5 biological replicates of HCC and CCA patients, ‘HCC’ transcriptional expression of 4 biological replicates, ‘CCA’ transcriptional expression of 3 biological replicates (B) Drp1 expression in patient matched healthy and tumor tissue samples on protein levels. Quantitative analysis of the corresponding Western blot of DRP1 in (C). Biological replicates: ‘Healthy’ n = 7, ‘HCC’= 5, CCA = 3 (C) Corresponding Western Blot results of ZEB1 and Drp1 protein abundance in healthy and tumor liver tissue. β-actin was used as housekeeping protein. ZEB1 protein abundance was observed in CCA-3 and HCC-5. The full western blot images can be obtained in the supplementary dataset.

Fig. 3

Evaluation of mitochondrial fitness and glucose metabolism in ZEB1^high and ZEB1^low liver cancer cell lines. (A–C) metabolic profile and metabolic balance for HCC cell lines, ZEB1^high cell lines exhibit a decreased metabolic balance compared to ZEB1^low cell lines. (D) Parameters of mitochondrial fitness, such as basal respiration, maximal respiration, resp. ATP and spare resp. cap. (E) Extracellular acidification rate as an indicator for glucose metabolism in all cell lines (F) Various parameters of glucose metabolism, such as glycolysis, glycolytic capacity and glycolytic reserve in HCC cell lines. (G) MitoTracker results showing mitochondrial content with significant reduction in ZEB1^Low cancer cell lines. Statistical analysis was performed using non-parametric Wilcoxon test, * = p < 0.05 (H) Mitochondrial membrane potential assessed in FACS analysis using JC-1.

Fig. 4

Apoptosis and mitochondrial membrane potential in MDIVI-1 treated cells and Combination Index plots for cells treated with a combination treatment of MDIVI-1 and Sorafenib. (A) Cell viability and apoptosis levels in cell lines after 48 h of treatment with MDIVI-1, assessed in FACS analysis. JC-1 aggregate levels indicating mitochondrial membrane potential after MDIVI-1 treatment, assessed in FACS analysis (B) Sorafenib titration results to determine the IC50 used in our combination treatment. Cell viability after 72 h measured by fluorescence is displayed. (C) Combination index < 1 indicates synergism, > 1 indicates antagonism, CI = 1 indicates additive effects. Our ZEB1^high cell lines exhibit a stronger synergism, also in lower drug concentrations compared to ZEB1^low cell lines with only additive or at lower levels antagonistic effects. Each data point represents the effect of a specific concentration of Sorafenib in combination with the indicated concentration of MDIVI-1 and is based on two technical replicates.