IFI44L is a novel tumor suppressor in human hepatocellular carcinoma affecting cancer stemness, metastasis, and drug resistance via regulating met/Src signaling pathway

217 HCC tissue microarray slides were obtained from HCC patients receiving surgeries in Changhua Christian Hospital from July 2011 to November 2013 [24]. Paraffin-embedded HCC samples were obtained from Changhua Christian Hospital under the approved Institutional Review Board (IRB) protocol. Clinical patterns and overall survival data were analyzed by SPSS software and chart review. The age of all patients was between twenty-nine and eighty-one years. The clinical characteristics of these 217 patients are shown in Table 1.

The human liver cancer cell lines Hep3B (ATCC number: HB-8064), HepG2 (ATCC number: HB-8065) and PLC (ATCC number: HB-8024) were obtained from the American Type Culture Collection (ATCC, Manassas, VA). All cells were cultured at 37 °C under 5% CO₂ in Dulbecco’s modified Eagle medium (DMEM; Invitrogen) supplemented with 10% fetal bovine serum (FBS; Biological Industries) and 100 units/ml of penicilium and streptomycin (Life Technologies, Carlsbad, CA, USA).

For IFI44L-expressing vector, IFI44L coding sequence was amplified and cloned in pMSCV plasmid. Antibodies for western blotting and immunohistochemistry (IHC) are anti-IFI44L (Abcam), p-Met (Cell signaling, Tyr1234/1235), Met (Cell signaling), Src (Cell signaling) and p-Src (Cell signaling, Tyr416). IFI44L-specific siRNAs were purchased from MDBio, Inc. Detailed sequences for IFI44L siRNA oligonucleotides were shown in Additional file 1: Table S1. For cell sensitivity assays, HCC cells were pretreated with doxorubicin (Sigma-Aldrich) for 18 h (overnight) in serum-free culture medium.

Quantitative RT-PCR (qRT-PCR) was used for gene detection. Detailed procedure of reverse transcription reaction was described elsewhere [25]. qRT-PCR was performed on a CFX96 qPCR detection system (Bio-Rad) with a 1:10 dilution of cDNA by using KAPA SYBR FAST qPCR Kits (KAPA Biosystems). The mRNA levels were normalized to actin mRNA. The primers used for mRNA expression are listed in Additional file 1: Table S1.

Migration and invasion abilities of HCC cells were carried out using the Falcon Cell Culture Inserts with or without Matrigel (BD Biosciences) coating as described previously [27]. Detailed procedures were described elsewhere [25].

Hep3B Cells (1 × 10⁶) with indicated treatments were suspended in phosphate-buffered saline (PBS) and were injected individually into the tail vein of 6- to 8-week-old C.B-17 severe-combined immunodeficient (CB17-SCID) mice. All mice were monitored meticulously and were sacrificed after 40 days of implantation. Tumor growth was observed by live animal BLI (Caliper IVIS system, PerkinElmer).

IHC was performed to detect IFI44L expression from paraffin-embedded HCC specimens. The slides were stained with anti-IFI44L antibody (Bethyl Labs, Montgomery, TX, USA) [28]. The IFI44L antibody was purchased from ThermoFisher (Rock, USA). In liver cancer specimens, the detailed scores for IHC were defined as described previously [24, 29].

In order to enrich for CSCs, parental Hep3B, HepG2, and PLC cells from monolayer were cultured in a stem cell selective condition described in ‘Methods’ to form spheres. Most of the suspended cells underwent apoptosis during the first 2 days of culturing, and the rest of survived cells gradually formed floating spheres. The spheres grew larger and often reached to 50–100 μM in diameter after 4–8 days (Fig. 1a). Overexpression of mRNA of HCC cancer stem cell markers was found in.

Hep3B sphere cells compared with their parental cells. These cancer stem cell markers, including CD24, CD44, CD117, CD133, ALDH, ABCG2, OCT4, and Nanog, were significantly higher in Hep3B sphere cells shown by qRT–PCR analysis (Fig. 1b) [7, 30‐37].

Next, we examined the chemosensitivity of these sphere cells. Parental Hep3B and HepG2 cells and their sphere cells were treated with doxorubicin for 48 h and cell viability was measured by MTT assay. Hep3B and HepG2 sphere cells are found to be more chemoresistant to continuous exposure to various concentrations of doxorubicin (Fig. 1c). Thus, we have successfully enriched HCC cancer stem-like cells from Hep3B, HepG2, and PLC lines displaying cancer stem cell characteristics including sphere-forming, expression of HCC cancer stem cell markers, and more chemoresistant in accordance with established parameters of cancer stem-like cells [38‐40].

Since IFI44L was implied to be correlated with cancer [20], we then investigated the impact of IFI44L on drug resistance. Cells transfected with IFI44L expression plasmid or control plasmid were tested their protein expression of IFI44L to confirm the transfection efficiency. Western blotting showed upregulation of IFI44L protein level in Hep3B and HepG2 cells after transfection with the expression plasmid of IFI44L (IFI44L vector) (Additional file 2: Fig. S1). Our data indicated that Hep3b and HepG2 cells became more chemosensitive to continuous exposure to different doses of doxorubicin after transfection with IFI44L vector (Fig. 2a), whereas IFI44L knockdown restored their chemoresistance (Additional file 3: Figure S2). These data suggested that overexpression of IFI44L significantly decreased chemoresistance of HCC lines towards doxorubicin. To assess whether IFI44L level correlated with cancer stemness in HCC, we examined the protein expression level of IFI44L in HCC lines. Decrease of IFI44L protein level in Hep3b and HepG2 sphere cells was found compared with their parental cells by Western blotting analysis (Fig. 2b). We then investigated if suppression of IFI44L by its small interfering RNAs (siRNA) could inhibit cancer stemness characteristics in HCC lines. Three specific IFI44L-siRNAs were tested for their inhibitory efficacy by analyzing the IFI44L protein levels in Hep3B, HepG2 and PLC cells, IFI44L-siRNA-2 showed the highest knockdown effect in inhibiting IFI44L protein and it was used in the subsequent experiments (Fig. 2c, Additional file 4: Figure S3). Next we tested whether sphere-forming ability of Hep3B, HepG2, and PLC lines could be promoted by knockdown of IFI44L. After 8 days culturing of Hep3b, HepG2, and PLC cells in the stem cell selective condition, sphere number was calculated by visual counting under microscope. Knockdown of IFI44L caused significant increase of sphere number (Fig. 2d). Thus, our data suggested that IFI44L may play as a tumor suppressor role in restoring chemosensitivity and affecting cancer stemness.

Furthermore, we evaluate the tumor suppressor role of IFI44L in regulating cancer metastasis. In Boyden chamber assay, we found that depletion of IFI44L expression significantly promotes Hep3B, HepG2 and PLC cell migration and invasion abilities (Fig. 3a and Additional file 5: Figure S4). To investigate whether IFI44L regulated cancer cell metastasis in vivo, we employed an experimental metastasis model via tail vein injection in SCID mice. In this model, knockdown of IFI44L significantly promoted lung metastasis of Hep3B cells compared with the control group (Fig. 3b). Since ISGs are implied to be correlated with Met pathway [11, 22, 23], we then explored the role of IFI44L in Met signaling pathway. By Western blotting analysis, we found that suppression of IFI44L enhances the phosphorylation of Met and Src in Hep3B and HepG2 cells (Fig. 3c). To further assess the role of IFI44/Met/Src axis in regulating cancer metastasis, we performed additional Western blotting analysis as well as migration and invasion assay. We found that overexpression of IFI44L decreased phosphorylation of Met as well as migration and invasion abilities in Hep3B cell line, whereas ectopic expression of Met reversed IFI44L-mediated inhibition of migration and invasion abilities approximately 50% (Additional file 6: Figure S5). Taken together, these findings reinforced that the functional role of IFI44L as a tumor suppressor and it could implicate in Met/Src signaling pathway in HCC.

To evaluate the correlation of IFI44L with clinical samples, the expression of IFI44L in 217 pairs of normal liver and HCC tumor tissues were analyzed by IHC and Western blotting analysis. The IHC score of IFI44L was significantly higher in normal liver tissues compared with tumor tissues (Fig. 4a). Western blotting analysis also revealed that all of ten pairs of matched HCC tumor tissues expressed lower level of IFI44L in comparison with the matched normal tissues (Fig. 4b). Downregulation of IFI44L expression found in HCC tumor tissues is compatible with the tumor suppressor role in HCC we discovered above.

Furthermore, the correlation between clinicopathological characteristics and IFI44L of these 217 patients were analyzed in Table 1. Among these parameters, age, gender, tumor differentiation, HBV surface antigen, anti-HCV antibody, and the tumor number were not significantly different in patients with low versus high expression levels of IFI44L (Table 1). However, low expression of IFI44L was observed in only 47% (84/180) of the early stages (stage I/II) HCC patents whereas 68% (25/37) of the late stages (stage III/IV) HCC patients expressed low levels of IFI44L (P = 0.029) (Table 1). IHC staining also confirmed that IFI44L protein level decreased markedly in advanced stages in HCC samples (Fig. 5a). Moreover, higher percentage of HCC patients with low expression level of IFI44L had larger tumor size then patients with high expression level of IFI44L (64% vs 36%, P = 0.002) (Table 1). Since CSCs are indicated to be associated with cancer recurrence [2, 38], our previous experiments also indicated that IFI44L affects cancer stemness in HCC cells. In clinic data, we also found that patients with low expression level of IFI44L had significantly higher relapse rate (81% vs 19%, P = 0.002) and shorter relapse-free survival (RFS) (p = 0.0012) than patients with high expression level of IFI44L (Table 1, Fig. 5b).

In survival analysis, the influence of clinicopathological characteristics including IFI44L on patients’ overall survival (OS) was statistically examined by univariate analysis shown in Table 2. Four parameters including advanced stages, larger tumor size, disease relapse, and low expression of IFI44L are significant correlated with shorter median OS (P < 0.001) (Table 2). Kaplan–Meier survival analysis of these 217 patients also revealed that low expression level of IFI44L correlated with poor OS (P < 0.001) (Fig. 5c). These results suggested that downregulation of IFI44L expression levels significantly correlated with larger tumor size, disease relapse, advanced stages, and poor clinical survival in HCC patients and could serve as an important prognostic marker.