LncRNA SNHG1 contributes to sorafenib resistance by activating the Akt pathway and is positively regulated by miR-21 in hepatocellular carcinoma cells

Human HCC HepG2 and Huh7 cells, and SR-HCC cells (HepG2-SR and Huh7-SR cells established from parental HepG2 and Huh7 cells, respectively) have previously been described [6, 23, 24]. All cell lines were confirmed as negative for mycoplasma infection by using a PCR-based Universal Mycoplasma Detection kit (American Type Culture Collection, Manassas, VA, USA). Cells were routinely cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco BRL, Grand Island, NY, USA) supplemented with 10% fetal bovine serum in a humidified atmosphere of 5% CO₂. The SR-HCC cells were kept by culturing them in the presence of sorafenib. Information for antibodies, reagents and kits is described in details under Additional file 1.

Male BALB/c-nu/nu mice (aging 6–8 weeks) obtained from SLAC laboratory Animal Co., Ltd. (Shanghai, China) were maintained at the Animal Research Center of the First Affiliated Hospital of Harbin Medical University. Animal experiments were performed as described previously [6, 23, 24], according to a permit (No. SYXK20020009, Harbin Medical University) in compliance with the Experimental Animal Regulations by the National Science and Technology Commission, China. Briefly, Huh7-SR cells (5 × 10⁶) were subcutaneously injected into mice receiving daily administration of sorafenib at a low dose of 10 mg/kg, which could help Huh7-SR cell maintain their sorafenib-resistant ability. Mice were monitored and the appearance of tumors recorded. 25 days later, mice bearing subcutaneous tumors (~ 100 mm³ in volume) were selected and randomly assigned to four treatment groups: control, sorafenib, anti-SNHG1 and sorafenib + anti-SNHG1. Sorafenib was suspended in an oral vehicle containing Cremophor (Sigma-Aldrich, Shanghai, China), 95% ethanol and water in a ratio of 1:1:6, and administered to mice in the sorafenib and sorafenib + anti-SNHG1 groups by gavage feeding at a dose of 30 mg/kg daily. Anti-SNHG1 was intratumorally delivered by means of lncRNA Smart Silencer mixed with Lipofectamine2000 (5 pmol/μl of oligonucleotides solution) once every 3 days for a total of five times in the anti-SNHG1 and sorafenib + anti-SNHG1 groups. Mice in the control group received oral vehicle and intratumoral injection of negative control (NC) oligonucleotides. 2 days following intratumoral injections, two mice from the control and anti-SNHG1 groups were sacrificed and tumors harvested for analysis. The remaining mice were further monitored for recording the size of tumors every 5 days and euthanized 21 days after treatments commenced.

These were described in detail under Additional file 1 and have also been described previously [6, 23, 26]

The sorafenib-resistant properties of SR-HCC cells were confirmed by comparing cell viability and apoptosis rates with parental HCC cells following incubation with various concentrations of sorafenib (Additional file 2: Figure S1).

We have previously reported that overexpressed miR-21 contributes to sorafenib resistance by activating the Akt pathway via downregulating PTEN, and anti-miR-21 could reverse this resistance in HCC cells [8]. In accord, the expression levels of miR-21 were significantly increased by sorafenib exposure in both parental and SR-HCC cells, and SR-HCC cells expressed higher levels of miR-21 than their parental sublines, in either the presence or absence of sorafenib (Additional file 2: Figure S2A). PTEN, a known miR-21 targeting gene in HCC cells [27], was downregulated in SR-HCC cells compared to parental sublines, and sorafenib incubation resulted in a significant reduction of PTEN expression in both parental and SR-HCC cells (Additional file 2: Figure S2B and C). In addition, transfection of miR-21 mimics downregulated PTEN expression in Huh7 cells that expressed lower levels of miR-21, while anti-miR-21 upregulated the expression of PTEN in Huh7-SR cells that expressed higher levels of miR-21 (Additional file 2: Figure S2 D and E).

Given that the interaction between lncRNAs and miRNAs plays important roles in gene regulation [28], we sought to identify potential lncRNAs that interact with miR-21 by studying multiple bioinformatics analyses and databases (http://​starbase.​sysu.​edu.​cn, www.​lncrnadb.​org, https://​omictools.​com, https://​bioinfo.​life.​hust.​edu.​cn/​lncRNASNP). Ten potential lncRNAs were screened based on the criteria of free energy <− 10 kcal/mol and score > 140 (Additional file 2: Table S1), and their expression levels were further detected by qRT-PCR with a panel of specific primers (Additional file 2: Table S2). Among the ten candidates, SNHG1 was shown to have the largest folds of alteration between SR-HCC and parental cells and between vehicle- and sorafenib-treated cells, and this change remained consistently elevated in both HepG2 and Huh7 cells (Additional file 2: Figure S3).

We next examined intracellular location of SNHG1 and miR-21 by detecting their expression in nuclear and cytoplasmic fractions isolated from lysates of Huh7 cells that had been incubated in the absence or presence of sorafenib (2.5 μM) for 48 h. Nuclear and cytoplasmic fractions possessed 47.5 and 52.5% of total SNHG1, respectively, in untreated Huh7 cells, while sorafenib incubation led to a significant increase of SNHG1 in nuclear fraction (84.7%) (Fig. 1a), indicating that sorafenib may induce a nuclear accumulation of SNHG1 in HCC cells, in support of a previous study where doxorubicin induced SNHG1 nuclear retention [16]. Intriguingly, sorafenib also significantly increased the proportion of miR-21 in nuclear fractions to 68.6%, compared with 34.7% in untreated Huh7 cells (Fig. 1b). The change of nuclear retention of SNHG1 and miR-21 induced by sorafenib was visualized by in situ hybridization with specific probes in Huh7 cells (Fig. 1c). The nuclear proportion of SNHG1 was also significantly higher in Huh7-SR cells (79.6%) than parental sublines (Fig. 1d). Transfection of an siRNA targeting SNHG1 (SNHG-si) significantly reduced the expression levels of SNHG1 in cytoplasmic fractions, while transfection of SNHG1-Smart Silencer significantly reduced SNHG1 expression in both nuclear and cytoplasmic fractions, compared with mock-treated or NC-transfected cells (Fig. 1d). Therefore, SNHG1-Smart Silencer was used as an anti-SNHG1 tool in the following experiments.

It has been reported that SNHG1 activates the Akt signaling pathway by promoting the transcription of SLC3A2 [20] and Akt is highly activated in SR-HCC cells [6]. We first examined the expression of SLC3A2 in HCC cells, which were incubated in the absence or presence of sorafenib (2.5 μM) for 48 h. SR-HCC cells expressed higher levels of SLC3A2 than their parental sublines, and sorafenib exposure upregulated its expression in both parental and SR-HCC cells (Fig. 2a and b). Depletion of SLC3A2 by an siRNA enhanced the activity of sorafenib in inhibiting the viability (Additional file 2: Figure S4A) and inducing apoptosis (Additional file 2: Figure S4B and C), and abolished the activation of Akt induced by sorafenib and enhanced the ability of sorafenib in inducing caspase-3 cleavage (Additional file 2: Figure S4D).

Anti-SNHG1 reduced the expression of SLC3A2 in both Huh7 and Huh7-SR cells (Fig. 2c and d). The Akt pathway was highly activated in Huh7-SR cells evidenced by the increased expression of p-Akt (ser473) and the sequential alterations of the downstream factors including mTOR (mTOR, mammalian target of rapamycin), p-GSK3β (phosphorylated glycogen synthase kinase 3β), p-S6K (phosphorylated ribosomal protein S6 kinase) and p-4EBP1 (phosphorylated eukaryotic translation initiation factor 4E-binding protein 1) (Fig. 2d), in consistent with our previous reports [6, 23]. Anti-SNHG1 treatment resulted in a significant downregulation of SLC3A2 and p-Akt, leading to the decreased expression of mTOR, p-GSK3β and p-S6K and increased expression of p-4EBP1 in both Huh7 and Huh7-SR cells (Fig. 2d). Inhibition of the Akt pathway by anti-SNHG1 also downregulated cyclin D1 and promoted the cleavage of caspase-9 and sequential caspase-3 (Fig. 2d). PTEN, a well-known upstream regulator that directly inactivates the Akt pathway [29], was downregulated in SR-HCC cells (Fig. 2a and b), in consistent with our previous study [8]. However, Anti-SNHG1 had little effects on the expression of PTEN mRNA or protein (Fig. 2c and d), indicating that SNHG1 may regulate the activation of Akt pathway independent of PTEN.

Mock- and anti-SNHG1-treated SR-HCC cells were incubated with various concentrations of sorafenib for 48 h. Anti-SNHG1 inhibited the viability and promoted the apoptosis of SR-HCC cells in either the presence or absence of sorafenib (Additional file 2: Figure S5). Specifically, both sorafenib and anti-SNHG1 induced cell apoptosis, but anti-SNHG1 exhibited a stronger pro-apoptotic activity than sorafenib, and anti-SNHG1 enhanced the pro-apoptotic ability of sorafenib (Fig. 3a-c). We have previously demonstrated that autophagy plays a death-promoting role in SR-HCC cells [6, 8]. Here we showed that anti-SNHG1 treatment resulted in significantly more acridine orange-stained acidic vesicular organelles (AVOs), while sorafenib only induced a slight increase of AVOs (Fig. 3b). Quantitative analysis confirmed that anti-SNHG1-transfected cells had significantly higher fluorescence intensity (FL3) and the combination of sorafenib and anti-SNHG1 led to even higher FL3 (Fig. 3d). In terms of mechanisms, anti-SNHG1 corrected the enhanced activation of Akt by sorafenib, and led to the increased conversion of LC3 (microtubule-associated protein 1 light chain 3)-I to LC3-II, upregulation of Beclin-1, and an increased cleavage of caspase-3 (Fig. 3e).

Sorafenib-resistant tumors were established as shown in Materials and Methods and also as described previously [6, 23, 24]. 25 days after cell inoculation, 27 mice (27/35) bearing palpable tumors of ~ 100 mm³ in volume were selected and randomly assigned to four groups, which received respective treatments (Fig. 4a). Four mice from the control (n = 2) and anti-SNHG1 (n = 2) groups were sacrificed 2 days after the commencement of treatments, and depletion of SNHG1 by anti-SNHG1 was confirmed by in situ hybridization (Additional file 2: Figure S6A). SR-HCC tumors were shown to be refractory to sorafenib since sorafenib-treated tumors (1157.4 ± 60.6 mm³ in volume, 1097.3 ± 86.8 mg in weight) were only slightly smaller than control tumors (1783.8 ± 55.6 mm³ in volume, 1509.1 ± 60.2 mg in weight) 21 days after the commencement of treatments (Fig. 4b-d). However, anti-SNHG1 treatment significantly reduced the size of tumors (811.3 ± 70.5 mm³ in volume, 812.3 ± 48.7 mg in weight) by 54.5%, and the combinational therapy resulted in a further reduction of tumor size (420.0 ± 32.2 mm³ in volume, 494.3 ± 70.6 mg in weight) by 76.4%, compared with control tumors at the end of experiments (Fig. 4b-d).

The expression of the key proteins in the Akt pathway was analyzed by Western blot analysis (Fig. 4e) and immunohistochemistry (Additional file 2: Figure S6B). In consistent with the in vitro results (Fig. 3e), anti-SNHG1 was shown to correct the upregulation of SLC3A2 and the activation of Akt pathway induced by sorafenib, and combined with sorafenib to upregulate the expression of LC3-II, Belcin-1 and cleaved caspase-9/- 3 (Fig. 4e). Tumors were examined for in situ cell proliferation by immunohistochemistry with an anti-Ki67 Ab, apoptosis by TUNEL staining, and autophagy by electronic microscopy (Fig. 4f). Again, sorafenib had a weak inhibitory effect against cell proliferation and weak activities to induce apoptosis and autophagy in SR-HCC tumors. Anti-SNHG1 significantly reduced cell proliferation, and induced apoptosis and autophagy in situ. Quantitatively, anti-SNHG1 enhanced the effects of sorafenib in inhibiting cell proliferation and promoting apoptosis, since tumors treated with the combinational therapy had a further reduction in proliferation index and a further increase in apoptosis rate, compared with either sorafenib or anti-SNHG1 monotherapy (Fig. 4g and h).

The above results showed that both SNHG1 and miR-21 could activate the Akt pathway in HCC cells, we thus investigated their interactive effects. Two putative miR-21 binding sites were identified in the enhancer/promoter regions of SNHG1 gene at 62622071-62622094 nt and 62,622,142-62,622,162 nt by using the RNAhybrid algorithm (Fig. 5a). A pmiR-RB-REPORT™ dual luciferase reporter system was employed to determine whether the predicted binding sites were involved in SNHG1 regulation. An expression vector containing the enhancer region of wild-type SNHG1 (WT) and three vectors containing mutants at 3′-UTR (Mut1, “TAAGCT” [62622089–62,622,094] mutated to “ATTCGA”; Mut2, “TAAGCT” [62622157–62,622,162] mutated to “ATTCGA”; Mut3, the combination of Mut1 and Mut 2) were constructed (Fig. 5a). MiR-21 mimics and the above vectors were co-transfected into Huh7 cells, and the relative luciferase activities were measured. Transfection of miR-21 mimics resulted in significantly higher luciferase activities than NC oligonucleotides in Huh7 cells, which were co-transfected with WT vector (Fig. 5b). However, this difference was abolished by co-transfection with Mut1 or Mut3 vector, but not by co-transfection with Mut2 vector (Fig. 5b), indicating that miR-21 may promote the expression of SNHG1 through the binding of the sequence “ATTCGA” in miR-21 to the corresponding “TAAGCT” at 3′-UTR of SNHG1 (Fig. 5a). In support, miR-21 mimics significantly elevated, while anti-miR-21 significantly reduced, the expression level of SNHG1 in both parental and SR-HCC cells (Fig. 5c). However, downregulation of SNHG1 by anti-SNHG1 had little effects on the expression levels of miR-21 in either parental or SR-HCC cells (Fig. 5d). The above results indicate that the regulatory effects between miR-21 and SNHG1 may be in a one-way manner. In order to further validate this point, we transfected Huh7-SR cells, which expressed higher levels of miR-21 and SNHG1 (Additional file 1: Figures S2 and S3), with miR-21 and anti-SNHG1. As shown in Fig. 5e and f, anti-miR-21 significantly upregulated the expression of PTEN, leading to the downregulation of p-Akt, while anti-SNHG1 had little effect on PTEN expression, but could downregulate the expression of SLC3A2, resulting in the downregulation of p-Akt. In addition, Huh7 cells, which expressed lower levels of miR-21 and SNHG1 (Additional file 2: Figures S2 and S3), were transfected with miR-21 mimics and anti-SNHG1. MiR-21 mimics downregulated the expression of PTEN, leading to the upregulation of p-Akt, and co-transfection of anti-SNHG1 could not abolish this effect of miR-21 mimics (Fig. 5g and h), indicating that miR-21 displays its regulatory effects on PTEN expression and Akt activation independent of SNHG1. However, miR-21 mimics upregulated the expression of SLC3A2, leading to the upregulation of p-Akt, and anti-SNHG1 could abolish this effect of miR-21 on SLC3A2 and p-Akt (Fig. 5g and h), indicating that miR-21 exhibits its effects on the SLC3A2/Akt pathway through regulating SNHG1.