The most significant finding of this study is that it reveals some novel mechanisms by which HBV promotes LC, in particular, based on the HB-LC genes, we constructed the prognostic models, analyzed the mechanism of immune infiltration, and proposed the potential therapeutic agents.

As reviews shown, high-risk populations for LC are usually HBV positive ones [15]. Regarding the impacts of HBV on LC, there are dynamic molecular mechanisms to control gene expression, such as the integration of the viral genome with host genome, triggering glycosylation in malignant hepatocytes, impacting metabolomics, changing exosomal characterize, and so on [1]. Previously, scholars have paid attention to the predictive significance of early factors for HB related LC patients, like preoperative serum SOD, GSH, MDA, Urea, AFU, CA153, CA125, and AFP and UA levels [16, 17]. Besides, non-coding RNAs can participate in HBV-driven HCC development [7]. Also, HBV may impact the angiogenesis signals during HCC progression [18]. However, the most critical oncogene driven by HBV infection is still greatly under-recognized. We found eight key HB-LB genes (GPSM2, UBE2S, EZH2, G6PD, CDK1, CDCA5, AURKA, and MKI67) with a clear carcinogenic role and prognostic values.

GPSM2 can serve as an independent prognostic biomarker for LC survival [19]. But very few studies have proposed its molecular mechanisms in LC development [20], and there is a lack of evidence from empirical molecular experiments. In 2018, it has been reported that UBE2S expression is elevated in hepatocellular carcinoma (HCC) and it predicts a poor prognosis [21]. It may exert oncogenic activities through enhancing the ubiquitination of p53 [22]. Recently, a study showed that UBE2S promotes malignant properties via VHL/HIF-1α and VHL/JAK2/STAT3 signaling pathways and decreases sensitivity to sorafenib in HCC [23]. EZH2 is a histone methyltransferase and one of the principal factors that mediates oncogenesis by acting as a driver of epigenetic alternation, and it is a well-known risk factor in HCC progression [24]. And there have been many related cancer-promoting mechanisms, such as the VASH1 Pathway [25], epigenetic silencing of let-7c/miR-99a [26], the miR-361/GPNMB axis [27], epigenetic regulation of TFR2 [28], the STAT3 signaling pathway [29], etc. And, EZH2 inhibition is an effective strategy in HCC treatment [30,31,32,33,34,35]. G6PD (glucose-6-phosphate dehydrogenase) is the rate-limiting enzyme of the pentose phosphate pathway, which is upregulated in HCC [36]. G6PD is a prognostic and diagnostic indicator of HCC [37]. It may promote LC through contribution to migration and invasion by inducing epithelial-mesenchymal transition (EMT) [38]. CDK1 is a well-known cyclin-dependent kinase and a strong oncogene in LC [39,40,41,42,43]. Similar to our conclusion, many scholars have proven its prognostic role and the correlation with immune infiltration [40,41,42,43,44,45]. CDCA5 is another validated prognostic biomarker for HCC [46,47,48,49]. CDCA5 is mediated by CDK1, and it regulates cell cycle and HCC proliferation [50, 51]. Indeed, it partially reflects pathological progression from cirrhosis to HCC [52]. AURKA can promote cell migration, EMT and HCC metastasis [53,54,55]. Finally, MKI67 is also called Ki67, which is the most famous cell proliferation marker. Previous studies have revealed the correlation of MKI67 with prognosis, immune infiltration, and T cell exhaustion in HCC [56]. In summary, all of the above eight key HB-LC genes have been sufficiently studied in the field of HCC. However, whether they are key participants in HBV-driven LC has not been clearly studied; our results are innovative in that we provide a dependable candidate molecular mechanism for the pro-LC effect of HBV.

Given the role of HBV on LC development, it has been recognized that antiviral treatment significantly decreased LC incidence and progression [57]. However, in addition to antivirals, we here propose seven potential anti-LC agents based on the relationship between HB-LC gene expression and drug sensitivity, focusing on HBV-positive LC patients. PHA-793887 is a novel and potent CDK inhibitor, and its application in HCC is uncommon currently, but there have been bioinformatics studies presenting the same view as ours [58, 59]. AZD8055 can induce cell death and activation of AMPK in HCC [60]. TSC2-deficient HCC cell lines were sensitive to AZD8055 [61], and covalent conjugation of AZD8055 with unsaturated fatty acids as the mTOR blocker can increase the therapeutic efficacy of HCC treatment [62]. Our study, on the other hand, suggests the unique sensitizing value of AZD8055 for the HBV + population. AZD7762 is a Chk1 inhibitor, the only one study showing its application on HCC reported that AZD7762 can enhanced the effect of doxorubicin on inhibiting growth of HCC cells in vitro and in vivo [63]. SNX-2112 is an Hsp90 inhibitor showing broad antitumor activity, it can induce apoptosis of human HCC cells; and a possible mechanism is that it may influence the level of endoplasmic reticulum stress [64]. However, the specific therapeutic effects of SNX-2112 on HCC remain to be explored. The application of TPCA-1 (an IKK-2 inhibitor) in the LC field has not been reported in any clear way. It is reasonable to select TPCA-1 as a preferred drug for HBV + LC treatment in the future. Similarly, PIK-93 is a synthetic inhibitor of PI4K and PI3Kα. There is no existing data on the use of PIK-93 in the treatment of HCC. Also, there are no known conclusions about the role of PAC-1 in HCC. In summary, the seven drugs we propose have considerable promise and great scope for exploration in the treatment of HBV + LC.

Still, there are some limitations in this study. Firstly, there are still multiple possibilities for the cancer-promoting mechanisms of these 8 key HB-LC genes, without validation of which mechanisms are the exact core driving mechanisms. We will follow up with cellular and animal experiments to clarify. Secondly, we found that all 8 genes are oncogenes, and the core driving mechanism seems to be related to the cell cycle, which is a very common oncogenic mechanism; however, this type of mechanism usually leads to rapid tumor growth and relatively short survival. Theoretically, prediction models based on this mechanism should be better in the short term than in the medium to long term. However, as a matter of fact, the 8-gene prognostic models we constructed have a better predictive efficacy at 5 years than at 1–2 years. This suggests that the short-term cancer-promoting mechanism of the 8-genes network is more complex than expected, and their interactions need to be elucidated through further regulatory experiments. Thirdly, given that the enhanced expression of eight driver genes are mainly due to HB infection, it was natural to consider their impact on immune cell infiltration. We noticed that the key HB-LC genes are negatively correlated with Th17, MAIT, monocytes, and CD4 Naive cells, while positively correlated with B cells, nTreg cells, and Tr1 cells. This suggests that the HB-LC genes promote progression and are detrimental to survival for reasons associated with stronger immune activation. This is not consistent with the fundamentals of immunotherapy in general. The possible reason for this, naturally, is that overgrowth of the tumor invariably leads to higher levels of immune activation. But it also suggests that modulating immune activation at the right time point may be beneficial to patient survival. The answer to this issue still needs to be based on the accumulation of more clinical data. Finally, we have unearthed seven novel drugs based on the oncogenic mechanism of eight HB-LC genes that may contribute to therapeutic sensitization of HB-LC patients. However, these drugs have yet to be validated for their effects and potential toxicity by animal experiments. This work deserves to be prioritized in subsequent studies.

In conclusion, we identified eight key HB-LC genes that play important cancer-promoting roles in LC, which may be the molecular mechanism by which HBV drives the development of LC. They can affect the immune infiltration level in HCC and can be used to construct LC prognostic models. Seven new drugs may have potentially therapeutic value against HBV-positive LC: PHA-793887, TPCA-1, AZD8055, PIK-93, AZD7762, PAC-1, and SNX-2112.