SOX30 is a key regulator of desmosomal gene suppressing tumor growth and metastasis in lung adenocarcinoma

The lung cancer cell lines (A549, LTEP-a-2, H520 and H226) were obtained from the Cell Bank of the Chinese Academy of Science (Shanghai, China), cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum (Gibco, CA). All the cells were maintained at 37 °C with 5% CO₂.

Construction of SOX30 expression vector was performed as previously described [19, 22]. For knockdown, two pairs of oligomeric singlestranded oligonucleotides and a pair of negative oligomeric singlestranded oligonucleotides were synthesized, and then inserted into miRNA expression vector pcDNA6.2-GW/EmGFP-miR. Cells were transfected using Lipofectamine2000 Reagent (Invitrogen Preservation, Carlsbad, CA, USA) according to the manufacturer’s instructions. The stably transfected cells were screened under G418 (Calbiochem, La Jolla, CA, USA) or Blasticidin (Sigma). Cell clones were obtained by the limited diluted method.

The RNA of cells was isolated using the Trizol reagent (Invitrogen, Life Technologies), and conversion of total RNA to cDNA was performed with the Reverse Transcription System (Promega, Madison, WI, USA). All qRT-PCR reactions was performed using the C1000 Real-Time Cycler (Bio-Rad Laboratories, Hercules, CA, USA) and qRT-PCR Master mixes (Promega, Madison, WI, USA). Primers for amplification of the DSC1, DSG1, DSC2, JUP, DSP, PKP1, DSC3, DSG3, PKP3 and ACTIN genes are listed in Additional file 1: Table S1. All experiments were carried out in triplicate, and the 2^-ΔΔt method was used to determine expression of the genes of interest.

For knockdown, A549 and LTEP-a-2 cells with SOX30 or empty vector stably expression were plated at 4 × 10³ cells per well on 96-well plates, and transfected with desmosomal gene miRNA or negative control. Cell proliferation was assessed using MTS Reagent (Promega, Madison, WI, USA) on days 1, 2, 3, 4 and 5 after transfection. The assays were performed in triplicate.

Transwell assays were performed by using transwell plates with 8 μm pore (Corning; 3422). For the migration assay, 1 × 10⁴ Cells were resuspended in serum-free RPMI-1640 medium and plated onto the 24-well upper chamber with the lower chamber filled with complete medium. For the invasion assay, the 24 transwell plate with matrigel (BD biosciences) polymerized at the 24-well upper chamber for 2 h at 37 °C as the intervening invasive barrier. After 24 h incubation at 37 °C, the cells on the upper chamber were removed and the number of cells that migrated or invaded to the lower side were fixed in 4% paraformaldehyde and stained with 1% crystal violet and counted at × 200 magnification in 10 different fields of microscope (Leica; D-35578). The results were determined from three repeated experiments.

For the xenograft tumor growth assay, a total of 1 × 10⁶ stable transfected A549 cells suspended in 150 μl PBS were injected subcutaneously into the right flanks of 5-week-old male Balb/c nude mice (n = 4 mice per group), respectively. Tumors size were measured every 3–5 days with calipers after injection, and the tumor volume was calculated based on formula: 0.5 × (length × width²). After 42 days housing, the mice implanted tumors were sacrificed, and liver tissues were dissected and subjected to histological examination. Metastatic nodules were detected by H&E staining and quantified by counting metastatic lesions in each section. The images of the positive areas were taken. All experimental animal procedures were approved by the Institutional Animal Care and Use Committee of Third Military Medical University, China.

WB was performed as previously described [20]. ACTIN was used as a loading control. The following primary antibodies were used: SOX30 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-20,104), DSP rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-33,586), JUP rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-7900), DSC3 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-48,750), CCND1 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-717), MMP7 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-30,071), ERK1/2 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-93), p-ERK1/2 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-101,761) and ACTIN monoclonal antibody (1:2000; Sigma; A5441). ImageJ software was used to quantify the protein expression.

Lung cancer cell lines with SOX30 or vector control stable expression grown on sterile glass coverslips, washed with PBS, fixed in 4% paraformaldehyde for 15 min and permeabilized by 0.5% Triton for 15 min at room temperature. Following blocking with 1% bovine serum albumin for 30 min, rabbit polyclonal antibody SOX30 (1:100; Santa Cruz Biotechnology; sc-20,104), mouse monoclonal antibody DSP (1:100; Santa Cruz Biotechnology; sc-365,981), mouse monoclonal antibody JUP (1:100; Santa Cruz Biotechnology; sc-33,634) and mouse monoclonal antibody DSC3 (1:100; Santa Cruz Biotechnology; sc-81,806) were incubated overnight at 4 °C. After washing three times, the cells were incubated with an appropriate fluorochrome-conjugated secondary antibody (1:300; Invitrogen; A-11010 and A-11003) for 1 h at 37 °C in the dark. After blue nuclear counterstaining with 4, 6-diamidino-2-phenylindole (Beyotime, Shanghai, China; C1006) for 10 min at room temperature, coverslips were mounted and observed with a fluorescence microscope (Zeiss, Oberkochen, Germany; LSM800). ZEN software was used to quantify the protein expression.

Tissue microarrays contained a total of 30 lung cancer patient tissue samples including 15 ADCs and 15 SCCs were obtained from the collaboration (Shanghai Biochip Co Ltd., Shanghai, People’s Republic of China). The rabbit polyclonal antibodies used were SOX30 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-20,104), DSP rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-33,586), JUP rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-7900), DSC3 rabbit polyclonal antibody (1:100; Santa Cruz Biotechnology; sc-48,750). IHC staining was performed as described previously [19].

ChIP analysis was performed using a ChIP Assay Kit (Pierce, Rockford, IL, USA). The immunoprecipitated and input DNA was used as a template for RT–PCR analysis using the primers listed in Additional file 1: Table S1.

C57BL/6 J (SOX30 conditional knockout) mice were obtained from the model animal research center of Nanjing University (Nanjing, China) and maintained in a controlled environment (12-h light/dark cycle, ad libitum access to food and water) at the laboratory animal facility of Third Military Medical University (Chongqing, China). The W/W mice (n = 5) and K/K mice (n = 5) that were used were matched for age (14 weeks) and weight (16–19 g). These mice received a intraperitoneal injection of urethane (1 g/kg in 100 μl saline) once a week for 10 consecutive weeks. Mice were sacrificed after 4 months. Lung neoplastic lesions were observed by H&E staining. All experiments on mice were approved by the Institutional Animal Care and Use Committee of Third Military Medical University, China.

GSE43580 dataset were used to analyze differential gene expression between ADC and SCC. The protein expression of desmosomal genes in human lung tumor tissues were determined from the human protein atlas (www.​proteinatlas.​org). TCGA lung adenocarcinoma RNAseq data (n = 571) and TCGA lung squamous carcinoma RNAseq data (n = 553) were used to compare the expression of SOX30 with desmosomal genes.

To analyze the differential gene expression between ADC and SCC, we first performed pathway analysis of one large-sample lung cancer datasets-GSE43580 and found that the expression of 17 cell junction genes has significant differences between ADC and SCC (Fig. 1a). Gene ontology (GO) analysis was conducted based on the 17 genes identified in the pathway analysis revealing 8 desmosomal genes: DSC1, DSG1, DSC2, JUP, DSP, PKP1, DSC3 and DSG3 (Fig. 1b). To further confirm the GO results, we analyzed the protein expression of these 8 desmosomal genes in clinical specimens from the human protein atlas (www.​proteinatlas.​org). We found that all desmosomal genes were overexpressed in SCC, and underexpressed in ADC (Fig. 1c and Table 1).

Our previous research suggested that SOX30 is a tumor suppressor in ADC but not in SCC [19, 22]. Thus, we suspect that the expression of desmosomal genes are associated with SOX30 expression. Therefore, we analyzed the correlation of SOX30 expression with the 8 desmosomal genes identified in the previous experiments between ADC and SCC by using TCGA data. The results revealed that the expressions of desmosomal genes were significantly positively correlated with SOX30 expression in ADC but not in SCC (Fig. 2a and b). Previous studies revealed that desmosome family members play a crucial role in tumor suppression [21]. Therefore, we hypothesized that SOX30 may inhibit tumor progression through upregulation of the desmosomal gene expression in ADC but not in SCC. To examine this hypothesis, qRT-PCR analyses were performed to measure the expression of desmosomal genes in SOX30-transfected A549 and LTEP-a-2 cells (two human lung adenocarcinoma cell lines), and H520 and H226 cells (two human lung squamous carcinoma cell lines). Consistently, the overexpression of SOX30 upregulated the expressions of the 8 desmosomal genes in the A549 and LTEP-a-2 cells but not in the H520 and H226 cells (Fig. 2c and d). Among these 8 desmosomal genes, the most obvious change in expression is three tumor suppressor genes for lung cancer, DSP, JUP and DSC3. To identify the functional target gene of SOX30, we determined the protein expression of these desmosomal genes by WB analyses and immunofluorescence cell staining. The results showed that these genes were upregulated in the SOX30 high expression group of A549 and LTEP-a-2 cells but not in the H520 and H226 cells (Fig. 2e and f, Additional file 2: Figure S1). To further determine the correction between SOX30 and these genes, we tested for expression of SOX30, DSP, JUP and DSC3 in human lung cancer tissues and adjacent tissues. The expression of DSP, JUP and DSC3 were associated with SOX30 expression in human ADC tissues but not in human SCC tissues (Fig. 2g and h).

Emiko et al. reported that SOX30 was able to specifically recognize the promoter of genes that have an ACAAT motif [23]. Furthermore, SOX30 can transcriptionally activate the target gene by directly binding to the promoter containing the ACAAT motif ref. Subsequently, analyzing the promoter region of DSP, JUP and DSC3, and found their promoter regions all contain an ACAAT motif (Fig. 3a–c). Then, we determined whether SOX30 could activate DSP, JUP and DSC3 transcription by binding to the ACAAT motif on their promoters. Chromatin immunoprecipitation (ChIP) assay demonstrated that SOX30 could directly bound to their promoters in A549 and LTEP-a-2 cells but not in H520 and H226 cells (Fig. 3d).

To define whether DSP, JUP and DSC3 mediates the effect of SOX30 on cell proliferation, migration and invasion in ADC cells, we first constructed two SOX30-overexpressing ADC cell lines, (A549-SOX30(+)) and (LTEP-a-2-SOX30(+)), and their negative controls (A549-Vector) and (LTEP-a-2-Vector). We then blocked either DSP, JUP or DSC3 expression in A549-SOX30(+) and (LTEP-a-2-SOX30(+)) cells by using miRNA; thus establishing ADC cell lines that are SOX30-overexpressing with either DSP, JUP or DSC3 knockdown. The expression of SOX30, DSP, JUP and DSC3 were verified by WB analysis (Fig. 4a). The MTS, colony formation assay revealed that the knockdown of DSP, JUP or DSC3 by miRNA significantly abrogated the effects on cell proliferation induced by SOX30 overexpression in A549 and LTEP-a-2 cells (Fig. 4b–c). The transwell assay was performed to determine whether desmosomal genes are involved in SOX30-mediated migration and invasion of ADC cells. The results showed that interference with the expression of DSP, JUP or DSC3 reversed the inhibitory effect of SOX30 on cell migration and invasion (Fig. 4d).

To further investigate whether upregulation of DSP, JUP and DSC3 are necessary for SOX30-medicated tumorigenesis and metastasis in vivo, we utilized a xenograft tumor model using these stable cell lines. As showed in Fig. 5a–c, xenografted (A549-SOX30(+)/DSP(−)), (A549-SOX30(+)/JUP(−)) and (A549-SOX30(+)/DSC3(−)) cells rapidly proliferated in mice compared with (A549-SOX30(+)) cells, and we found that tumor volume of (A549-SOX30(+)/DSP(−)), (A549-SOX30(+)/JUP(−)) and (A549-SOX30(+)/DSC3(−)) were significantly larger than those with (A549-SOX30(+)) cells. Moreover, H&E staining revealed that knockdown of DSP, JUP or DSC3 dramatically increased the number of metastatic foci in the liver of nude mice (Fig. 5d), and these results were further confirmed by qRT-PCR using a human-specific β2-MG (beta-2-microglobulin) with a mouse-specific β2-MG as an internal control (Fig. 5e) [24]. These results indicate that DSP, JUP and DSC3 play critical roles in SOX30-mediated growth and metastasis of ADC.

Previous research has shown that a member of SOX family mediates the Wnt signaling pathway [25‐27]. According to our observation presented above and considering the facts that DSP and JUP can suppress lung cancer progression by inhibition of the Wnt/β-catenin signaling pathway, we hypothesized that SOX30 can suppress the Wnt/β-catenin signaling pathway through upregulation of DSP and JUP. To test this hypothesis, the protein expression levels of SOX30, DSP, JUP, MMP7 and CCND1 in A549 stable cell lines and xenograft tumors were first examined by WB. Consistent with our theory, the protein expression of MMP7 and CCND1 were downregulated in a SOX30-overexpressing cell line and xenograft tumors; MMP7 and CCND1 expression were increased significantly in SOX30-overexpressing cells, or tumor tissue after DSP or JUP expression was blocked (Fig. 6a). We then conducted a TOP/FOP FLASH reporter assay in SOX30-overexpressing A549 cells. As shown in Fig. 7a, overexpression of SOX30 significantly inhibited the transcription activity of β-catenin/T-cell factor (TCF) compared with cells transfected with the vector control (Fig. 6b). Previous studies have reported that DSC3 acts as a tumor suppressor through inhibiting ERK signaling in human lung cancer [28]. WB assay were performed to test whether these changes affect ERK1/2-related signaling activity in SOX30-transfected A549 cells. The data showed that overexpression of SOX30 significantly reduced the phosphorylation of ERK1/2 whereas the knockdown of DSC3 significantly abrogated the SOX30 overexpression-induced dephosphorylation of ERK1/2, suggesting that DSC3 is involved in the SOX30-regulated dephosphorylation of ERK1/2 in vitro and in vivo (Fig. 6c). Therefore, these results indicated that SOX30 overexpression suppresses Wnt/β-catenin signaling and ERK signaling by increasing the expression levels of desmosomal genes, and this may contribute to inhibition of cell growth, migration and invasion in ADC.

To further define the important role of the “SOX30-desmosomal gene axis” axis in lung carcinogenesis, we first analyzed the correlation of SOX30 expression with DSP, JUP and DSC3 in lung tissues of SOX30 knockout mice by qRT-PCR, WB and IHC analyses. Consistently, the expression levels of DSP, JUP and DSC3 were significantly decreased in lung tissues of SOX30-knockout mice. (Fig. 7a–b and Additional file 3: Figure S2).

Next, we compared the lung carcinogenesis in W/W (wild type of SOX30) and K/K (homozygous deletion of SOX30) mice treated with urethane, an environmental carcinogen that induces ADC in C57BL/6 J mice (data not shown) [29]. H&E staining showed that the number and size of lung tumors were enhanced in K/K mice compared with W/W mice (Fig. 7c). The results suggested that SOX30 plays a vital role in chemical carcinogenesis. To further evaluate whether the promotion of lung tumorigenesis in K/K mice was related to the downregulation of the desmosomal gene expression with activation of the Wnt and ERK pathway, the protein level of SOX30, DSP, JUP, DSC3, CCND1, MMP7, ERK1/2 and p-ERK1/2 were determined by WB in the lung tumor tissues. The protein expression of desmosomal genes were significantly decreased, and CCND1, MMP7, ERK1/2 and p-ERK1/2 were increased in lung tumor tissues of K/K mice compared to W/W mice (Fig. 7d).