HLD patients exhibit a lower incidence of liver cancer.
Liver fibrosis is an important nexus of liver lesions caused by a variety of factors. Our previous clinical observations have shown that fibrosis induced by HLD is less likely to lead to liver cancer than fibrosis induced by the hepatitis B virus (HBV). Based on this, we have presented cartoon diagrams and pathological slides of different stages in Fig. 1A. There was no difference in blood fibrosis indices (data not shown) and imaging between HBV and HLD in the liver fibrotic stage; however, HBV-induced liver fibrosis is prone to develop into hepatocarcinoma (Fig. 1B). Based on the above observations, data from more than 500 patients were collected from the First Affiliated Hospital of Anhui University of Chinese Medicine in 2019–2023, including HLD (191 cases) and HBV (117 cases), a total of 308 patients exhibited no significant difference in physiological and pathological indicators during liver fibrosis. Subsequent follow-up studies showed a significantly lower incidence of cancer in HLD patients compared to HBV patients (HLD: 0.52%; HBV: 11.97%), especially regarding hepatocellular carcinoma (HLD: 0.52%; HBV: 9.4%) (Fig. 1C). Statistically, HBV patients accounted for 50–80% of HCC cases worldwide[46], which is consistent with our clinical observations. Liver tissues from patients with hepatic hemangioma, HLD and HBV were collected during surgery (Fig. S1A-I), Masson's staining showed normal liver tissues in patients with hepatic hemangioma (Fig. S1B-C), while fibrous tissues in HLD and HBV patients were extensively proliferated, hepatic lobe space was enlarged, hepatocyte edema with a large number of inflammatory cell infiltration (Fig. S1E-F, H-I). The above results indicate that despite the presence of fibrotic changes in liver tissue among HLD and HBV patients, the progression of liver fibrosis in these groups differs. Moreover, the incidence of liver cancer in HLD patients was significantly lower than that in HBV patients. In conclusion, according to the clinical statistics, CT reports, pathological specimens and multiple studies we collected, it was confirmed that HBV-induced liver fibrosis tends to develop into hepatocarcinoma, while HLD-induced liver fibrosis inhibits the occurrence of hepatocarcinoma and tends to liver failure, which is significantly different.
Table 1
The demographics and clinical characteristics of individuals
Baseline characteristics | HLD (n = 191) | HBV (n = 117) | P value |
Female/Male | 79/112 | 47/70 | 0.403 |
Mean age | 54.49 ± 9.96 | 56.25 ± 11.36 | 0.905 |
BMI | 17.5 ± 6.54 | 16.4 ± 5.34 | 0.341 |
ALT | 35.81 ± 12.91 | 36.41 ± 14.00 | 0.324 |
AST | 45.61 ± 16.63 | 49.11 ± 13.57 | 0.463 |
TBIL | 37.63 ± 17.65 | 34.66 ± 19.88 | 0.653 |
TBA | 43.22 ± 18.82 | 39.62 ± 11.39 | 0.295 |
Table 2
Cancer types | HLD (n = 191) | HBV (n = 117) | P value |
Lung | 0 | 1 | 0.201 |
Stomach | 0 | 2 | 0.070 |
Liver | 1 | 11 | < 0.001 |
Pancreas | 0 | 0 | -- |
Total | 1 | 14 | < 0.001 |
Proteomic differences between HLD and NOR patients.
To further investigate the differential changes in liver fibrosis between patients with HLD or NOR, we conducted proteomic and transcriptomic analyses on fresh tissue samples. Based on the proteomic analysis of 3 cases of HLD-induced cirrhosis and 3 cases of normal liver, we identified 305 significantly differentially expressed proteins between the two groups. (top 15 in Heat map is shown in Fig. 2A and top 15 in Volcano plot is shown in Fig. 2B). Specifically, there were 219 proteins upregulated and 86 proteins downregulated in HLD patients compared to NOR patients. (Fig. 2C). Subsequent Gene Ontology (GO) analysis highlighted significant differences in biological processes, molecular function and cellular component between the two groups, with a particular focus on abnormal metabolism (Fig. 2D and Fig. S2A). KEGG pathway analysis revealed that the enriched pathways associated with the differentially expressed proteins were predominantly related to metabolism. These findings suggest the presence of distinct metabolic differences in the liver between the two groups of patients, which could potentially contribute to the observed divergent clinical outcomes observed (Fig. 2E, Fig. S2B). The results from protein-protein interaction network (PPI) analysis indicated a significant coordinated interaction between metabolic pathways in the two groups (Fig. 2F). The consistent results of protein quality control across different patient groups confirm the reliability of our analysis and hence the robustness of the results (Fig. S2C-S2H).
Differentially expressed miRNAs between HLD and NOR patients.
To compare the differences in miRNAs between patients with HLD-induced cirrhosis and normal tissues, RNA from liver tissue samples from both groups underwent sequencing. The RNA quality control of the isolated tissues from each patient group met the standard and is suitable for subsequent analysis (Fig. S2A and S2B). The expression profiles of miRNAs between the two groups revealed that 21 miRNAs were differentially expressed (Heat map was shown in Fig. 3A and Volcano plot was shown in Fig. 3B). Among them, 18 miRNAs were up-regulated and 3 miRNAs were down-regulated in the liver tissue of HLD patients compared to NOR patients (Fig. 3C and Fig. S3C). GO functional analysis of the differentially expressed miRNAs showed that the differences between the two groups were observed in the Biological Process category, indicating their involvement in the regulation of biological functions (Fig. 3D, Fig. S3D and S3E). KEGG pathway enrichment analysis revealed that the differences between the two groups were mainly associated with cancer development. This finding may provide insight into why patients with HLD-induced cirrhosis and HBV-induced cirrhosis exhibit different outcomes. Moreover, significant differences were observed in the Hippo and AMPK pathways between the two groups (Fig. 3E). Network of miRNAs analysis revealed that the AMPK and Hippo signaling pathways interact with numerous cancer-related pathways (Fig. 3F). The expression results of key miRNAs in tissues detected by qRT-PCR technology were consistent with the results of omics analysis (Fig. 3G). Through these experiments, we have identified differences in key metabolic pathways, specifically the AMPK and Hippo signaling pathways, between the two patient groups, which interact with pathways involved in cancer development.
Transcriptomic analysis of mRNA and its integration with miRNA transcriptomics and proteomic analysis.
Omics analysis of mRNAs in two groups of patients with HLD-induced cirrhosis or normal liver tissues revealed a large number of differentially expressed mRNAs between the two groups (Heat map is shown in Fig. 4A and Volcano plot is shown in Fig. 4B). GO functional analysis revealed that the differentially expressed mRNAs between the two groups were mainly enriched in stress response, which may be caused by liver cell damage in liver cirrhosis (Fig. 4C). KEGG enrichment analysis identified enrichment in the AMPK signaling pathway and other pathways related to carbohydrate and glycolipid metabolism (Fig. 4D), which is consistent with our findings in miRNA omics analysis. Previous studies have demonstrated the close association between glycolipid metabolism and the Hippo signaling pathway[47–49]. The combined analysis of miRNA and mRNA omics suggests the presence of 12 upregulated miRNAs and 3 downregulated miRNAs that target the differentially expressed mRNAs. These changes in miRNAs resulted in the upregulation of 29 mRNAs and the downregulation of 22 mRNAs (Fig. 4E), which is in line with our findings in the miRNA omics analysis. The intersection analysis of mRNA and protein omics results revealed an overlap of 131 mRNAs and proteins that were upregulated and 12 that were downregulated in both datasets, confirming the consistency between mRNA and protein omics results (Fig. S4A -S4D). Combining these results with the differential expression of upstream miRNAs, we identified two consistently altered upstream miRNAs were found to target 5 downstream mRNAs and proteins (Fig. 4F). Western blot (WB) results validated the expression of key proteins in the differentially regulated pathways identified by the omics analysis (Fig. 4G). The results of the mRNA omics analysis further supported the consistent regulation of metabolic signaling pathways observed in the miRNA omics and protein omics results, suggesting that these changes may be related to the stress response function.
The Hippo and AMPK pathways are activated in liver samples from HLD patients.
The above omics analysis revealed that the protein and miRNA differences between HLD and HBV were mainly enriched in glycolipid metabolism, and the Hippo and AMPK signaling pathways, and the Hippo signaling pathway is closely associated with tumor progression[50–52]. It was found that the AMPK and Hippo signaling pathways were activated in the liver tissues of HLD patients compared to the control group (Fig. 5A), which was consistent with the omics results. Periodic Acid Schiff stain (PAS) staining showed a significant increase in glycogen granule deposition in liver tissue from HLD patients compared to HBV patients (Fig. 5B). Since HLD is a copper metabolism disorder caused by ATP7B mutation or deletion, we identified R778L as the major mutation site based on ATP7B sequencing of 20 HLD patients, and subsequently constructed ATP7B R778L mutant lentivirus (ATP7B mut) accordingly (Fig. 5C). To further investigate how ATP7B mutation or deletion affects tumor metabolism, we overexpressed ATP7B mut in two hepatocellular carcinoma cell lines, MHCC-97H and Huh-7, and treated the cells with the AMPK pathway activator (metformin) or AMPK pathway inhibitor (compound C) under energy-sufficient or energy-deficient conditions, respectively. As shown in Fig. 5, overexpression of the ATP7B mut activated the Hippo signaling pathway (lane 1 vs lane 2 and 4 in Fig. 5D-E), resulting in increased level of Yes-associated protein (YAP)/Phosphor-YAP (P-YAP) and transcriptional coactivator with PDZ-binding motif (TAZ)/Phosphor-TAZ (P-TAZ). Although the ATP7B mut led to activation of AMPK phosphorylation in hepatocytes (data not shown), it had a negligible effect on AMPK phosphorylation in tumor cells (Fig. S5). Furthermore, due to metabolic reprogramming in tumor cells, AMPK phosphorylation varied dramatically between different tumor cells (Fig. S5A vs S5B). Although the effect of ATP 7B mut on AMPK phosphorylation was not significant in tumor cells, the AMPK inhibitor Compound C still inhibited the activation of the Hippo signaling in MHCC-97H and Huh-7 cells (Fig. 5D-E, lane 4 vs 5). Similarly, the Hippo signaling was further activated by metformin (lane 3 vs lane 4 in Fig. 5D-E). These results suggest that ATP7B mutants, which mimic HLD, induce a specific activation of the Hippo signaling pathway in HCC cells, and that this activation is dependent on the phosphorylation status of the metabolic stress sensor AMPK.
Mutation or deletion of ATP7B mimicking HLD inhibits tumor cell progression.
Mutation or deletion of the ATP7B gene, mimicking HLD, activates the Hippo signaling pathway, which is often involved in tumor growth and proliferation. However, is unclear how the Hippo signaling activated by mutation or deletion of ATP7B affects the progression of HCC. Two types of hepatocellular carcinoma cells, MHCC-97H and Huh-7, were cultured under energy-sufficient or energy-deficient conditions. It was found that the proliferative capacity of these cells was significantly higher than that of the ATP7B knockdown cells, suggesting that loss of ATP7B inhibits tumor cell growth (Fig. 6A). MHCC-97H and Huh-7 cells with or without ATP7B knockdown were cultured under low (2% FBS) serum conditions and treated with the Hippo signaling inhibitor Verteporfin or the AMPK signaling inhibitor Compound C (Compd C), respectively. In both MHCC-97H and Huh-7 cells, the knockdown of ATP7B eliminated the low serum viability of these tumor cells, while the use of Verteporfin could restore this low serum viability to some extent. Whereas the use of Compound C after ATP7B knockdown did not restore the low serum viability of MHCC-97H and Huh-7 cells (Fig. 6B). Further focus formation assays showed that knockdown of ATP7B significantly restored cell-to-cell contact inhibition in MHCC-97H or Huh-7 cells, thereby inhibiting foci formation (Fig. 6C, column 1 vs column 2). Verteporfin inhibition of the Hippo pathway caused MHCC-97H and Huh-7 cells to lose cell-cell contact inhibition to some extent (Fig. 6C and Fig. S6A, column 2 vs column 3), partially reserving the malignant transformation of the tumor cells. On the other hand, Compound C also significantly inhibited the formation of foci (Fig. 6C, column 1 vs column 4), indicating that AMPK phosphorylation is required for the malignant transformation of MHCC-97H and Huh-7 cells. In anchorage-independent growth assays, knockdown of ATP7B significantly inhibited colony formation of MHCC-97H or Huh-7 cells in soft agar (Fig. 6D Fig. S6B, column 1 vs column 2). Verteporfin, the Hippo signaling inhibitor, caused MHCC-97H and Huh-7 cells to regain the ability to form colonies (Fig. 6D, column 2 vs column 3). Compound C also significantly inhibited colony formation in soft agar (Fig. 6D, column 1 vs column 4). In conclusion, mutations or deletions in the ATP7B gene that mimic HLD activate the Hippo signaling pathway and inhibit the proliferation and transformation of HCC cells. This may partly explain why HLD patients exhibit a lower incidence of hepatocellular carcinoma. Figure 6E presents a model demonstrating how ATP7B mutation or deletion in HLD affects the process of liver fibrosis.