HCC is a malignant tumor with high incidence and mortality rates worldwide [27]. The occurrence and progression of HCC are complex process, which is modulated through various numbers of oncogenes and anti-tumor genes. ADHs are huge family covering 7 members that are mainly involved in the conversion between alcohol and acetaldehyde, and also correlated to several hepatic diseases [14, 28]. However, the prognostic value of ADHs in the patients with HCC is still unclear. In the present study, we performed comprehensive analysis to investigate ADH genes association with the progression and prognosis of the patients with HCC, and to explore a series of diagnostic biomarkers of HCC.
The ADH family members are widely expressed in human liver without ADH7 [12, 13]. Therefore, ADH1A-ADH6 were selected to evaluate the prognostic value of ADHs in the patients with HCC. Primarily, our results revealed that the expression levels of ADH1A-ADH4, and ADH6 were significantly decreased in HCC tissues compared to normal liver tissues (Fig 1), which were similar to their expression in non-small cell lung cancer (NSCLC) and pancreatic adenocarcinoma [19, 29]. Therefore, we speculate that ADH1A-ADH4, and ADH6 may serve as tumor suppressors in HCC. Whereas, the expression of ADH5 were significantly upregulated in HCC tissues compared to normal liver tissues (Fig 1E), consistent with what is already known in NSCLC, gastric cancer, and pancreatic adenocarcinoma [19, 20, 29]. In addition, previous studies have reported the positive correlation of ADHs with each other in numerous cancers [19, 20]; here, our results also confirmed that in HCC (Figure 2).
Smoking is an important risk factor for lung cancer, a recent study reported that the expression levels of ADHs without ADH1A were significantly associated with smoking status of the NSCLC patients [19]. Interestingly, alcohol consumption status is considered as a primary cause for HCC; here, our results revealed that the expression levels of ADH1A-ADH6 were obviously increased in alcohol consumption HCC patients (Fig 3). Moreover, accumulating evidences have demonstrated that ADH family members were correlated with clinical stages, pathological grades, and TNM classifications in several cancers [19, 29, 30]. In our present study, it showed that the expression levels of ADH1A-ADH1C, and ADH6 were remarkably downregulated according to the pathologic T stage progression of HCC (Fig 4). In addition, the expression of ADH1A and ADH4 was significantly associated with pathologic T stage (Table 1). Altogether, ADH family members play an important role in the development process of HCC. It is meaningful to evaluate the prognostic value for HCC patients.
In recent years, large numbers of investigations have reported that gene polymorphism of ADHs was correlated with cancer risk [15, 17, 18, 31, 32] . Moreover, the activity of ADH isoenzymes was significantly higher in liver cancer tissues than in healthy tissues [33], suggested the diagnostic value of ADH for the patients with liver cancer. However, rare studies have been performed to evaluate the prognostic value of ADHs mRNA expression in HCC. In the present study, our univariate and multivariate Cox regression analyses results indicated that high ADH1A, ADH1C, ADH4, and ADH6 levels independently predicted improved OS and RFS in HCC patients; whereas, high ADH1B levels independently predicted improved RFS in HCC patients (Table 2, Fig 6 and 7). Furthermore, our Kaplan-Meier analysis data also revealed that high ADH1A, ADH1C, ADH4, and ADH6 levels predicted good OS and RFS in HCC patients; while, high ADH1B only predicted good RFS in HCC patients (Fig 8 and 9). Recently, the expression of ADH1A was measured by using MS/MS and TMA in CHCC-HBV patients, which indicated the robust prognostic value of ADH1A for potential clinical application [34]. Moreover, Chen Q, et al also reported that high expression of ADH1C was associated with a good prognosis for HCC patients by using the TCGA internal and three GEO (GSE76427, GSE15654, and GSE14520) external validation cohorts [35]. In addition, the prognostic value of ADH4 was also confirmed by immunohistochemical analysis with 91 paraffin-embedded HCC specimens [36]. Although, a recent study reported that decreased ADH5 expression in HBV-related HCC tumor tissue predicted earlier recurrence [30]. Surprisingly, ADH5 expression could not play a significant role in prediction of OS and RFS in HCC patients depending on our present data. Therefore, our data supposed that ADH family members without ADH5 might serve as the potential biomarkers for the patients with HCC.
ADHs play a pivotal role in the metabolic process of ethanol [26]. Our GO enrichment analysis showed that ADHs were contributed to ethanol metabolism, such as ethanol oxidation, ethanol metabolic process, and primary alcohol metabolic process. Moreover, KEGG enrichment analysis indicated that ADHs were involved in fatty acid degradation, retinol metabolism, and so on (Fig 9), constitute with what is investigated using GSEA (Fig 10A and 10B). In addition, our GSEA results also showed that high expression group of ADHs was significantly and negatively associated with pathways in cancer without ADH5 (Fig 10C), which suggested that high expression of ADHs could inhibit cancer related pathways and ADHs presented the tumor suppressor role. Increasing evidences indicated that various numbers of signaling pathways participated in the progression of HCC [37]. In order to determine the signaling pathways related to ADHs in HCC patients, GSEA was performed based on PID dataset [38]. Our results showed that low expression group of ADHs without ADH5 was positively related to various pathways (Fig 10, Table S3 and S4). Interestingly, most of the signaling pathways contributed to promote tumorigenesis, such as ATR pathway [39], FOXM1 and FOXO pathways [40, 41], MTOR pathway [42], NOTCH pathway [43], and P53 downstream pathway [44]. Accordingly, high expression of ADHs could inhibit the malignant process of HCC, which promote the OS and RFS probability of HCC patients. Although low expression group of ADH5 was positively related to oncogenic signaling pathways, such as ATR, FOXM1, MTOR, NOTCH, and P53 downstream pathways, the statistic difference was not significant. Interestingly, low expression group of ADH5 was not positively related to FOXO signaling pathway, which may be associated with the high expression of ADH5 in HCC. By summarizing all the points, ADHs without ADH5 might act as the tumor suppressor via inhibiting oncogenic signaling pathway in HCC.
There were some limitations in our present study, which should be known. Primarily, the clinicopathological information of HCC patients from TCGA website, such as tumor size, hepatitis virus infection, α-fetoprotein, non-alcoholic fatty liver and cirrhosis, was not comprehensive or lost. Secondly, our present study investigates the prognostic value of ADHs for HCC patients at mRNA levels, which is not consummate. Therefore, their prognostic value for HCC patients at protein levels should be further evaluated in our future studies. Thirdly, ADH1A, ADH1B, ADH1C, ADH4, and ADH6 were all associated with the prognosis of HCC patients. In the future, the in vitro and in vivo experiments should be performed to investigate their roles in the progress of HCC, then the best one of them could be selected as diagnostic and prognostic biomarkers for HCC patients. In addition, further effort is also needed to elucidate the mechanisms, which ADHs contribute to the progress of HCC in the future.