Background

Hepatocellular carcinoma (HCC) has the third highest rate of mortality worldwide [1]. It manifests frequently in patients with cirrhosis and liver dysfunction, limiting interventions with systemic cytotoxic drugs [2]. To date, four drugs targeting tyrosine kinases have been approved for the treatment of HCC. Lenvatinib has recently been used as a first-line drug [3], and regorafenib and cabozantinib have been approved as second-line drugs after the failure of sorafenib [4], but their use is not widely accepted. Sorafenib is the first approved systemic drug and continues to play the most important role in the management of advanced HCC [5]. However, the known resistance to sorafenib greatly minimizes its therapeutic benefits. Therefore, exploring the mechanisms for sorafenib resistance and seeking novel molecular targets are urgently required.

We have previously demonstrated that the Akt pathway is highly activated in sorafenib-resistant HCC (SR-HCC) cells and the inhibition of Akt could reverse this resistance by switching autophagy from a cytoprotective role to a death-promoting mechanism [6]. MicroRNAs (miRNAs) are a group of small noncoding RNAs, which regulate multiple cellular functions and have emerged as potential targets in the anti-cancer campaign [7]. In exploring the miRNA-related mechanisms that regulate the activation of Akt involved in sorafenib resistance, we found that miR-21 was highly expressed in SC-HCC cells and was able to activate the Akt pathway by dysregulating phosphatase and tensin homolog (PTEN) [8].

Long non-coding RNAs (lncRNAs) play crucial roles in controlling gene expression involving numerous biological processes in many diseases [9]. One of the mechanisms of their action involves the formation of regulatory networks with other RNA species, such as miRNAs and mRNAs [10, 11]. Several lncRNAs have been shown to participate in anti-cancer drug resistance [12]. For instance, lncARSR secreted by exosomes regulated sunitinib resistance by acting as competing endogenous RNAs for miR-34 and miR-449 in renal cancer [13], LncRNA MIR-100HG was involved in cetuximab resistance by regulating the wnt/β-catenin pathway via miR-100 and miR-125b in colorectal carcinoma [14], and LncRNA HULC reduced the sensitivity to cytotoxic drugs by stabilizing Sirt1-induced autophagy in HCC cells [37]. The best characterized mechanism for the interaction of lncRNAs and miRNAs pertains to the role of competitive endogenous lncRNAs, which serve as “sponges” for miRNAs [38]. On the other hand, miRNAs can regulate lncRNAs via a RNA-induced silencing complex, which often takes places in the cytoplasm [39]. In the canonical way, miRNAs are egressed from the nucleus and eventually mature in the cytoplasm, where they exert the post-transcriptional gene regulatory roles [40]. However, recent evidence suggests that certain mature cytoplasmic miRNAs can shuttle back to the nucleus, where they regulate gene expression [41,42,43]. The presence of mature miRNAs in the nucleus has recently been confirmed as a general phenomenon in mammalian cells [39, 41]. MiR-21 is the first miRNA with both a cytoplasmic and nuclear location in HeLa cells, with approximately 20% of mature miR-21 in the nucleus [44]. In agreement, the present study shows that 34.7% of miR-21 resides in nuclei of untreated Huh7 cells, which is supported by our microscopic evidence from in situ hybridization analysis. More interestingly, sorafenib induced the nuclear translocation of miR-21, which contributes to sorafenib resistance by upregulating SNHG1 in HCC cells.

MiR-21 is one of the few widely studied miRNAs expressed in many types of malignances including HCC, and plays a central role in cancer progression [27, 45]. The proposed “onco-miR addition” concept has further strengthened its role in carcinogenesis [45]. MiR-21 has also been identified as a key miRNA associated with drug insensitivity including the resistance to sorafenib [46]. In the conventional way, miR-21 exhibits its function by regulating downstream mRNAs [47]. The present study provides strong evidence that miR-21 could re-enter the nucleus, where it binds to the enhancer/promoter region of lncRNA SNHG1, leading to enhanced expression of SNHG1. This novel regulatory function of miRNAs on lncRNAs has been demonstrated in a previously reported study, where mature miR-140 trafficking in the nucleus can interact with lncRNA NEAT1, resulting in an increase in NEAT1 expression [48].

One weak point of the present study is that the results obtained in cell and animal experiments have not been validated in tumor tissues collected from HCC patients who have developed sorafenib resistance. The guideline of American Association for the Study of Liver Diseases (AASLD) indicates that sorafenib therapy is only approved for treating advanced HCC patients, judged clinically unsuitable for curative therapies including surgery and liver transplantation [49]. Due to the anatomic feature, clinical HCC tissues can be obtained by laparotomy, laparoscopy or fine-needle biopsy. However, the chance for harvesting SR-HCC tissues from late-staged patients is rare because such patients are unable to benefit from these invasive procedures in the view of stringent human ethic requirements, the informed consent and potentially risky complications. Postmortem may be another possible procedure for collecting SR-HCC tissues, but the quality of such tissues is hard to control. However, further validation in clinical SR-HCC tissues should be investigated in the future when such tissues are available since it will increase the clinical relevance and translation of the present results.

Conclusions

The present study has demonstrated that lncRNA SNHG1 contributes to sorafenib resistance by activating the Akt pathway through upregulating SLC3A2 and its nuclear overexpression is promoted by miR-21 in HCC cells. Depletion of SNHG1 inhibited the activation of the Akt pathway, enhancing the efficacy of sorafenib in suppressing SR-HCC cells by promoting apoptosis and autophagy in cultured cells and in animal experiments. Despite the fact that two putative miR-21 and SNHG1 binding sites exist, SNHG1 was shown to be incapable of regulating miRNA expression, which is different from the well characterized role of lncRNAs as sponges for miRNAs. On the other hand, miR-21 did not exhibit a negative regulatory effect on SNHG1, which usually takes place in the cytoplasm via a miRNA-induced silencing complex. However, miR-21 was able to promote the expression of SNHG1 in the nucleus, to which the shuttling of miR-21 was induced by sorafenib in HCC cells. The present study has uncovered a novel regulatory pathway by which miR-21 promotes the expression of SNHG1, leading to the activation of Akt pathway. Our results also suggest that SNHG1 may represent a potentially valuable target for overcoming sorafenib resistance in HCC.