Cx58 is associated with the metastasis of non-small cell lung cancer via MEF2B/Cx58 axis

Abstract

Connexins (Cxs), also known as gap junction proteins, are structurally related transmembrane proteins and have been implicated in carcinogenesis. Although some evidence suggests that these proteins are tumor suppressors due to their reduced expression in cancers, recent research indicates their complicated roles in tumor progression during different stages, including metastasis. Here, we show that Cx58, which is upregulated in non-small cell lung cancer (NSCLC), is modulated by myocyte-enhancer binding factor 2B (MEF2B). Either Cx58 or MEF2B knockdown attenuates the migration and invasion of NSCLC cells by inducing cytoskeleton rearrangement. Additionally, the prometastatic role of Cx58 in NSCLC is demonstrated in vivo. In conclusion, our findings suggest that Cx58 is transcriptionally activated by MEF2B and is involved in the metastasis of NSCLC by regulating cytoskeleton organization. Targeting the MEF2B/Cx58 axis may be exploited as a modality for improving NSCLC therapy.

Introduction

Lung cancer remains the number one cause of cancer-related death, causing 1.8 million deaths worldwide each year [1]. Approximately 85% of lung cancers are diagnosed as non-small cell lung cancer (NSCLC) [2]. Great advances have been achieved in lung cancer treatment during the past two decades, especially with the recent use of tyrosine kinase inhibitors and immunotherapy in selected patients [3]. Nevertheless, only a minority of patients have obtained unprecedented survival benefits from the abovementioned new therapies. At present, the overall survival rate for NSCLC patients is still not optimistic, particularly for patients with metastatic NSCLC [4]. Metastasis accounts for the majority of non-small cell lung cancer-related deaths, but its role in cancer biology is poorly understood [ 2, 4, 5]. Thus, novel biomarkers and therapeutic targets are still urgently needed.

Connexins (Cxs), the primary structure of gap junctions, mediate direct cellular communication under various physiological and pathological conditions [6]. Gap junctional intercellular communication (GJIC) is essential for coordinating cell growth, cell differentiation, and tissue homoeostasis [7]. Over the past 50 years, a tremendous amount of data has been generated supporting the hypothesis that a lack of GJIC or reduced expression of Cxs is the structural hallmark of cancer cells [ 8– 12]. Moreover, accumulating evidence has shown that the roles of Cxs in tumors depend on both Cx isoform expression and the tumor type or subtype. The complexity of cancer biology in multiple stages, including metastasis, makes the issue more complicated [ 7, 13– 16]. In addition, more recent data suggest that Cxs could have GJIC-independent functions, playing oncogenic and prometastatic roles in cancer development and metastasis [ 17– 19]. There are 21 members in the human Cx family [20]. Extensive research has explored the link between lung cancer and members of the Cx family, including Cx43, Cx31.1, Cx26, and Cx45 [ 7, 13 , 11]. However, few studies have focused on connexin58 (Cx58). Human Cx58, also known as gap junction alpha-9 protein (GJA9) or gap junction alpha-10 (GJA10) protein, was first identified by genome sequencing in 2003, and this protein has no mouse ortholog in the mouse genome [ 21, 22]. Human Protein Atlas RNA-seq analysis of normal tissues revealed biased expression in the testis, skin and 6 other tissues [23]. Data from The Cancer Genome Atlas (TCGA) revealed significant upregulation of Cx58 mRNA expression in lung adenocarcinoma (3.01-fold) and lung squamous cell carcinoma (1.82-fold), compared with healthy tissue [24]. To date, in addition to mRNA analysis in lung cancer, few studies have investigated the relationship between Cx58 and cancer. Here, we found that the Cx58 protein level was much higher in lung cancer tissues than in adjacent normal lung tissues and that the expression level was positively related to metastatic potential, indicating that Cx58 might be associated with NSCLC metastasis. Moreover, the reduction in Cx58 in NSCLC cells impaired their migration and invasion ability in vitro and suppressed their lung metastasis potential in vivo.

The MEF2B gene encodes a transcriptional activator and is mutated in various cancers [ 25, 26]. MEF2B mutation drives lymphomagenesis in diffuse large B-cell lymphoma by regulating the expressions of genes related to cell migration ability and epithelial‒mesenchymal transition (EMT) [ 27, 28]. Genetic alterations in MEF2B have also been reported in lung carcinoma, indicating that the transcriptional activity of MEF2B may be involved in lung cancer [25]. To date, studies on the role of MEF2B in lung cancer are lacking.

In this study, our data indicated that by regulating Cx58, knockdown of MEF2B could suppress the metastasis of NSCLC, which is involved in cytoskeleton rearrangement. Collectively, our findings suggest that targeting the MEF2B/Cx58 axis has the potential to inhibit the metastasis of NSCLC.

Materials and Methods

Cell lines and cultures

Human NSCLC 95C and 95D cell lines were obtained from the Institute of Biochemistry and Cell Biology (Shanghai, China). HEK293T, A549 and Beas-2b cell lines were purchased from the American Type Culture Collection (ATCC, Manassas, USA). RPMI-1640 medium (Thermo Fisher Scientific, Waltham, USA) was used to culture all the above human lung cell lines. The HEK293T cell line was maintained in DMEM (Thermo Fisher Scientific). All the cells were cultured at 37°C in 5% CO ₂ with the above media supplemented with 10% fetal bovine serum (Corning Co., Corning, USA), 100 U/mL penicillin and 100 U/mL streptomycin. Total RNA and proteins were extracted when the cells reached 80% confluence in culture dishes.

NSCLC patient samples

Totally, 41 patients were confirmed to have NSCLC on the basis of histopathology at the Second Xiangya Hospital of Central South University between 2012 and 2015. The complete clinicopathological characteristics of the patients are listed in Table 1. The human studies were approved by the hospital’s Ethical Review Committee [approval number 2022(387)]. All patients provided written consent.

Table 1 Correlations between Cx58 expression and clinicopatholo-gical features in NSCLC patients

Parameter	Case	Cx58 expression	P
		Low ( n, %)	High ( n, %)
Total	41	12	29
Age				0.8775
< 55	13	4 (30.8)	9 (69.2)
≥ 55	28	8 (28.6)	20 (71.4)
Gender				0.4522
Male	30	9 (30.0)	21 (70.0)
Female	11	4 (36.4)	7 (63.6)
Smoking				0.5392
No smoking	12	4 (33.3)	8 (66.7)
Smoking	29	8 (27.6)	21 (72.4)
Histopathology type				0.7516
Adenocarcinoma	27	7 (25.9)	20 (74.1)
Nonadenocarcinoma	14	4 (28.6)	10 (71.4)
Lymph node metastasis				0.4360
Negative	22	7 (31.8)	15 (68.2)
Positive	19	5 (26.3)	14 (73.7)
TNM stage				0.0035*
Stage I	17	7 (41.2)	10 (58.8)
Stage II–IV	24	5 (20.8)	19 (79.2)
Differentiation				0.1308
Low grade	8	3 (37.5)	5 (62.5)
Middle-high grade	33	9 (27.3)	24 (72.7)

Low 0–4, High 5–12. Two-sided Fisher’s exact test was applied. *significant difference.

Quantitative real-time PCR

Total RNA was extracted from the 95D and A549 cell lines via the TRIzol reagent (Invitrogen). Then, the RNA samples were converted into cDNA via a PrimeScriptTM RT Reagent Kit (TaKaRa, Osaka, Japan). The CFX-96 Real-Time PCR System (Bio-Rad, Richmond, USA) was used for quantitative real-time PCR (qRT-PCR) analysis. The sequences of primer for qRT-PCR are listed in Table 2. The qRT-PCR conditions are as follows: pre-denaturation at 95°C for 30 s; denaturation at 95°C for 5 s; annealing and extension at 60°C for 30 s, repeated for 35 cycles (steps 2 and 3); dissolution curve analysis from 65°C to 95°C for 20 min. The relative levels of the indicated genes were calculated by the 2 ^–ΔΔCt method.

Table 2 The sequences of primers used for PCR in this study

Target	Forward primer (5′→3′)	Reverse primer (5′→3′)
Cx58	TGTACCGACTGAGAGTTCTTGA	ACAGAGCGAGTGAAAATGTGTAT
MEF2B	GATGCGGGAGATCTGGATT	TGATCCTGGAAGAGACTCGG
ACTIN	CTGGAACGGTGAAGGTGACA	AAGGGACTTCCTGTAACAATGCA
Cx58-A ChIP	ATAGAGAATATATACACTCC	AGTGTGAATCTGTTACCAAT
Cx58-B ChIP	GAGTAGCTGGGACTACAAG	TGGATCATTTGAGGTCAGGA
Cx58-C ChIP	TCACCTGAGGTCAGGAGTT	TGAGTAGCTGTGATTACA
Cx58-D ChIP	ATCACTGACTTTTCAAGCCA	AAGGTTTCTAAAACTAGTAAC
GAPDH ChIP	TACTAGCGGTTTTACGGGCG	TCGAACAGGAGGAGCAGAGAGCGA

Western blot analysis

RIPA buffer (Thermo Fisher Scientific) supplemented with protease and phosphatase inhibitors was used to extract proteins. Proteins were separated via 10% SDS-PAGE and then transferred to membranes (Thermo Fisher Scientific). The membranes were incubated with the following primary antibodies: anti-Cx58 (SAB2100922, 1:1000; Sigma-Aldrich, St Louis, USA), anti-MEF2B (AV36916, 1:1000; Sigma-Aldrich), EMT antibody sampler kit (9782, 1:1000; Cell Signaling Technology, Danvers, USA), anti-β-Actin antibody (SC-47778, 1:1000; Santa Cruz Biotechnology, Santa Cruz, USA) and an anti-GAPDH antibody (SC47724, 1:1000; Santa Cruz Biotechnology). The membranes were incubated with the corresponding HRP-conjugated secondary antibodies (7074; Cell Signaling Technology, Danvers, USA) at a dilution of 1:1000 for 60 min at room temperature. Visualization of the results was achieved using enhanced chemiluminescence reagent (Thermo Fisher Scientific) and subsequent analysis was conducted using ImageJ software.

Transient transfection

Transient transfection was performed by transfecting A549 and 95D cells with Cx58 or MEF2B small interfering RNA (siRNA). All siRNAs were provided by RiboBio (Guangzhou, China), and the sequences are as follows: Cx58 siRNA sequence 1: 5′-CCAAATCCTGACAATCATA-3′, Cx58 siRNA sequence 2: 5′-CTAGAGGTCACCGTTCTAT-3′; MEF2B siRNA sequence 1: 5′-CCAGCCCAGATGTGGTATA-3′, MEF2B siRNA sequence 2: 5′-GCATCCTGGACCAAAGGAA-3′. The negative control siRNA (siN0000001-1-5) was also provided by RiboBio. Both 95D cells and A549 cells were transfected with siRNA (50 nM final concentration) using the transfection reagents DharmaFECT 1 (Thermo Fisher Scientific) according to the manufacturer’s protocol. The cells were harvested for RNA extraction, western blot analysis and immunofluorescence staining 48 h later.

Wound healing assay

When the cells reached approximately 80% confluence, a scratch was made via a 200-μL pipette tip, the cell layer was washed with PBS, images of the initial gap was captured using an IX73 microscope (Olympus, Tokyo, Japan), and then the cells were incubated in medium containing 2.5% fetal bovine serum. Gap size was recorded at regular time intervals for 48 h.

Cell invasion assay

Cell invasion activity was measured using transwell chambers (3422; 8 μm; Corning Co.) with filters coated with Matrigel (BD Biosciences, Bedford, USA). A total of 3 × 10 ⁴ cells were seeded in the upper chamber in 200 μL of serum-free medium. The lower chamber was filled with 700 μL of complete medium. After incubation for 48 h, the migrated cells were fixed, stained with 0.1% crystal violet for 15 min, and counted via the IX73 microscope.

Apoptosis detection and flow cytometry assay

Apoptotic analysis was conducted via a fluorescein isothiocyanate (FITC)-Annexin V apoptosis detection kit with propidium iodide (PI) (Biolegend, San Diego, USA) following the manufacturer’s protocol. Cells were stained with PI and Annexin V and subsequently analyzed via flow cytometry. The flow cytometry assay was executed via an LSR-II flow cytometer (BD Biosciences), and the data were analyzed via FlowJo software.

Immunofluorescence staining

The cells seeded on coverslips were fixed with 4% paraformaldehyde for 15 min. After treatment with 2% BSA (Thermo Fisher Scientific) for 20 min at room temperature, the cells were incubated with FITC-conjugated phalloidin (Sigma-Aldrich) at room temperature for 1 h. The nuclei were stained with DAPI (D9542; Sigma-Aldrich). All coverslips were visualized via a DMI8 confocal microscope (Leica, Wetzlar, Germany).

Lentivirus production and infection

Small hairpin RNAs (shRNAs) were cloned and inserted into the pLVX vector (Clontech, Mountain View, USA). Empty vector was used as a negative control. The shRNA-targeting sequences for Cx58 and MEF2B were designed according to the corresponding siRNA sequences. The shRNA sequences targeting Cx58 are as follows: sense: gatccGCCAAATCCTGACAATCATATTCAAGAGATATGATTGTCAGGATTTGGTTTTTTg, antisense: aattcAAAAAAGCCAAATCCTGACAATCATATCTCTTGAATATGATTGTC AGGATTTGGg. The shRNA sequences targeting MEF2B are as follows: sense #1: gatccGCCAGCCCAGATGTGGTATATTCAAGAGATATACCACATCTGGGCTGGTTTTTTg, antisense #1: aattcAAAAAAGCCAGCCCAGATGTGGTATATCTCTTGAATATACCACATCTGGG CTGGg; sense #2: gatccGCATCCTGGACCAAAGGAATTCAAGAGATTCCTTTGGTCCAGGATGCTTTTTTg, antisense #2: aattcAAAAAAGCATCCTGGACCAAAGGAATCTCTTGAATTCCTTTGGTCCAGGA TGCg. HEK 293T cells were co-transfected with the pLVX-shRNA vector together with the psPAX2 packaging plasmid and pMD2.G envelope plasmid (Addgene, Cambridge, USA). After 48 h, the supernatant was collected and then added to the target cells. Puromycin (Gibco, New York, USA) was used to select infected cells. The full-length Cx58 cDNA was subsequently cloned and inserted into the lentiviral vector GV358 (Genechem, Shanghai, China) to construct the recombinant plasmid. A blank GV358 lentiviral vector was used as the negative control. The constructs were subsequently co-transfected into 293T cells. At 48 h post transfection, the lentivirus was collected and used to infect A549 cells to overexpress Cx58.

Chromatin immunoprecipitation (ChIP)-PCR assay

ChIP assays were performed with a kit (EZ-ChIP; Millipore, Billerica, USA). A total of 2 × 10 ⁶ A549 cells in a 10-cm dish were prepared. Formaldehyde (1%) was used for crosslinking. After the formaldehyde was quenched with glycine, the cells were scraped and centrifuged. The crosslinked chromatin was then sonicated to shear the DNA into fragments of 200–1000 bp. One microgram of anti-MEF2B, anti-RNA polymerase II (used as a positive control), or anti-IgG antibody (used as a negative control) was added to three equal aliquots of chromatin supernatant and then rotated overnight at 4°C. Quantitative PCR was performed after elution and reversal of the cross-links. The sequences of the primers used for the Cx58 promoter domain are listed in Table 2.

Luciferase activity assay

Four potential binding sites of MEF2B in the Cx58 promoter were predicted to be distributed at positions –533, –730, –1112, and –1276 bp from the transcription start site (TSS), which were named the A, B, C, and D binding sites, respectively. On the basis of the ChIP data, Cx58 promoter regions with lengths of 1392 bp (including A and B putative binding sites), 615 bp (including a putative binding site) and 528 bp (including a nonbinding site between –1891 bp and –1363 bp from the TSS) were subcloned and inserted into the pGL3 Basic vector (Promega, Madison, USA). The cells were co-transfected with a constructed pGL3-Cx58 promoter vector and Renilla expression vector (used as an internal control). The reporter assay was conducted following the instructions of the Dual-Glo ^® Luciferase Assay System (Promega).

Hematoxylin and Eosin (H&E) staining

The paraffin-embedded tissues were sectioned into slices with a thickness of 5 μm. These sections were subsequently placed in an oven at 70°C for 2 h to facilitate the melting of the paraffin. Following this, the sections underwent a dewaxing process using xylene and ethanol. After staining with hematoxylin and eosin, the sections were washed appropriately. Finally, the sections were mounted with coverslips, sealed with a suitable sealing agent, observed and collected images with the IX73 microscope.

Immunohistochemistry (IHC)

After deparaffinization and rehydration of the paraffin-embedded sections, antigen retrieval was performed in sodium citrate solution using a pressure cooker, followed by blocking endogenous peroxidase with 3% H ₂O ₂ solution and blocking nonspecific binding with 5% goat serum. Then, the primary antibody anti-Cx58 (HPA067850, 1:500; Sigma-Aldrich) was added, and the mixture was incubated at 4°C overnight. The next day, after incubation with the secondary antibody, the sections were stained with 3,3,0-diaminobenzidine. Finally, the sections were counterstained with hematoxylin (ZSGB-Bio, Beijing, China). Quantitative analysis was performed via visual scoring by pathologists (low, 0–4; high, 5–12) [29].

Mice and xenograft tumor models

The metastatic potential of A549-Cx58shRNA or A549-shNC cells in vivo was evaluated via a xenograft tumor model. Female 8-week-old BALB/c nude mice (Slack Jingda Experimental Animal Co., Ltd., Changsha, China) were injected with 1 × 10 ⁶ A549-Cx58 shRNA or A549-shNC cells through the lateral tail vein. Eight weeks later, the mice were euthanized, and the metastatic nodules in organs, including the lungs, liver, and brain, were checked. After the samples were fixed with Bouin’s solution, the number of metastatic nodules on the lung surface was counted. Following formaldehyde fixation, the tissues were prepared for hematoxylin and eosin staining (H&E) and IHC staining. After H&E staining, the tumor slides were observed with the IX73 microscope at a magnification of 400×, and 5 sections were randomly selected (4 slides per mouse). The tumor area was calculated via ImageJ software. All procedures in the animal experiment were approved by the Animal Ethics Committee of Second Xiangya Hospital, Central South University (approval number 2022342).

Statistical analysis

The data in Table 1 were analyzed via two-sided Fisher’s exact tests. Other data were analyzed using a two-tailed Student’s t test for two groups and are presented as the mean ± SD. The statistical software used was GraphPad Prism (version 5; GraphPad software, Inc., San Diego, USA). A P value of less than 0.05 was regarded as statistically significant.

Results

Cx58 is upregulated in human NSCLC tissues and is correlated with poor prognosis.

Given that Cx58 mRNA expression is increased in NSCLC tissue compared with healthy tissue [24], we obtained tumor tissues and adjacent noncancerous tissues from NSCLC patients and measured Cx58 protein levels via immunohistochemical (IHC) staining. Figure 1A shows representative IHC images of adjacent noncancerous tissues and tumor tissues at different stages. Statistical analysis revealed that the expression of Cx58 was significantly upregulated in tumor tissues compared with adjacent noncancerous tissues in the 41 NSCLC patients ( P < 0.0001, Figure 1B). Data analysis via Kaplan-Meier plotter revealed that NSCLC patients with high expression of Cx58 had shorter overall survival than those with low expression of Cx58 did (HR = 1.41, log-rank P = 1.9 × 10 ^–6, Figure 1C) [30]. Furthermore, our data revealed that the association between Cx58 protein expression levels and TNM stage was statistically significant ( P = 0.0035, Figure 1D and Table 1). However, there was no correlation between the Cx58 expression level and age, gender, smoking status, differentiation grade, histopathology type or lymph node metastasis ( Table 1). Collectively, these results showed that Cx58 is upregulated in NSCLC patients and is closely correlated with poor prognosis, suggesting that Cx58 might act as an oncogene and prometastatic gene in NSCLC.

Figure 1 .

Cx58 is upregulated in human NSCLC tissues and is correlated with poor prognosis

(A) Representative immunohistochemistry images of Cx58 expression in adjacent normal lung tissues and NSCLC patient tumor tissues at different stages (stage I, stage II, and III). The scale bars represent 100 μm (upper panel) and 10 μm (lower panel). (B) Statistical analysis of Cx58 expression in tumor tissues (T) and matched adjacent normal tissues (N) from NSCLC patients (n = 41), P < 0.0001. (C) Kaplan-Meier survival analysis of Cx58 expression in lung cancer patients from TCGA array data via kmplotter online. Red line, high Cx58 expression (n=1299); black line, low Cx58 expression (n = 627). (D) Statistical analysis of Cx58 expression between TNM stage I (n = 17) and stage II-IV (n=24) patients. Data are presented as the mean ± SD, **P < 0.01.

Cx58 suppression decreases NSCLC cell migration and invasion by disrupting cytoskeleton rearrangement

We first detected the Cx58 expression level in normal lung epithelial cells (Beas-2b) and lung cancer cells (A549 and 95D), and the western blotting results revealed that, compared with that in normal lung epithelial cells, Cx58 expression was obviously elevated in lung cancer cells ( Supplementary Figure S1A). To address the effect of Cx58 on NSCLC cell metastatic traits, we used the highly metastatic NSCLC cell lines 95D and A549 for subsequent experiments. We then explored the effects of Cx58 on the migration and invasion of NSCLC cells. We used Cx58-specific siRNA to knock down Cx58 in 95D and A549 cells. Then, cell migration (wound healing) and invasion assays were conducted with 95D and A549 cells. There was a significant reduction in the migration ability of 95D cells and A549 cells after the depletion of Cx58 relative to that of control cells ( Figure 2A–D). Similarly, we found a significant decrease in the invasion capacities of 95D and A549 cells after silencing of Cx58 compared with those of control cells ( Figure 2E–H). These results demonstrated that Cx58 silencing attenuated the motility and invasiveness of NSCLC cells. Therefore, Cx58 may promote metastasis in NSCLC cells. To eliminate the potential influence of Cx58 on the apoptosis of lung cancer cells, we used flow cytometry to assess cell apoptosis in the A549 and 95D cell lines. The results revealed no significant difference in the apoptosis rates of 95D and A549 cells following Cx58 silencing compared with those of the control cells ( Supplementary Figure S1 B).

Figure 2 .

Cx58 suppression decreases NSCLC cell migration and invasion by disrupting cytoskeleton rearrangement

(A–D) Migration ability was examined via a wound healing assay in 95D (A,B) and A549 (C,D) control cells and Cx58-silenced cells. The scale bar is 100 μm. Data are presented as the mean ± SD from 3 independent experiments. ***P < 0.001. (E–H) Invasion ability was examined by transwell assay in 95D (E,F) and A549 (G,H) control cells and Cx58-silenced cells. Data are presented as the mean ± SD from 3 independent experiments. **P < 0.01, ***P < 0.001. The scale bar is 100 μm. (I) The cytoskeleton was detected by immunofluorescence assay of F-actin in Cx58-silenced A549 cells and control cells; the scale bar represents 25 μm. (J) Cx58, E-cadherin, N-cadherin, vimentin, and ZO1 protein levels were detected by western blot analysis in A549 and 95D cells with Cx58 silencing and control cells; GAPDH was used as a loading control.

Cancer cell migration and invasion require reorganization of the cytoskeleton or EMT [31]. Cx58, a gap junction protein, may be associated with the progression of EMT and cytoskeleton rearrangement in NSCLC cell lines. To address this possibility, we performed immunofluorescence assays for F-actin staining. As shown in Figure 2I, Cx58 silencing in A549 cells induced obvious changes in the cytoskeleton. F-actin was disrupted after Cx58 depletion. Moreover, we performed western blotting to detect EMT-related protein expression. As shown in Figure 2J, Cx58 silencing induced contradictory changes in EMT-related proteins. When Cx58 was knocked down in 95D cells and A549 cells, both E-cadherin (epithelial marker) and N-cadherin (mesenchymal marker) increased slightly. However, the expression of vimentin (a mesenchymal marker) was decreased to some extent. Notably, Zonula occludens-1 (ZO1), a cytoskeletal protein that is regarded as the most common protein that interacts with Cxs [ 7, 19, 32], was obviously reduced after Cx58 silencing. Collectively, these data suggest that Cx58 knockdown could attenuate the capacity of NSCLC cells to migrate and invade by inducing cytoskeleton rearrangement.

Cx58 inhibition decreases the potential for NSCLC metastasis in vivo

To extend our observations concerning the role of Cx58 in promoting NSCLC metastasis in vitro, we evaluated whether Cx58 could regulate metastasis in vivo. A549-Cx58 shRNA stable cell lines and A549-shNC cells were constructed. We then conducted animal experiments in which 1 × 10 ⁶ cells were injected into BALB/c nude mice through the tail vein to evaluate the role of Cx58 in homing capacity. We observed that silencing of Cx58 decreased the metastatic potential ( P < 0.05), as evidenced by the reduced number of tumor nodules in the lungs harvested on day 60 ( Figure 3A,B). Representative H&E-stained images of tumors from mice are shown in Figure 3C. The tumor area was decreased in the mice injected with A549-Cx58 shRNA compared with the control mice ( P < 0.05, Figure 3D). Moreover, no metastatic nodules were found in other organs, including the liver or brain. IHC staining of Cx58 in the mice is shown in Figure 3E. Cx58 expression in the tumor tissues of the mice injected with A549-Cx58 shRNA was lower than that in the tumor tissues of the mice injected with A549-shNC cells. Figure 3F confirms the Cx58 protein level in A549-Cx58 shRNA cells and A549-shNC cells. Taken together, the above data showed that the knockdown of Cx58 suppressed the metastatic potential of NSCLC in vivo.

Figure 3 .

Cx58 inhibition decreases the potential for NSCLC metastasis in vivo

(A) Representative images of lung tumor nodules in a mouse tail vein injected with A549-Cx58shRNA cells or A549-shNC cells (scale bar: 2 μm). (B) Statistical analysis of lung tumor nodules in mice (n = 7) injected with A549-Cx58shRNA cells and in mice (n = 7) injected with A549-shNC cells (*P < 0.05). (C) Representative H&E images of the lungs of mice injected with A549-Cx58shRNA cells and mice injected with A549-shNC cells (scale bar: 50 μm). (D) Statistical analysis of the lung tumor area/five sections in mice (n = 7) injected with A549-Cx58shRNA cells and in mice (n = 7) injected with A549-shNC cells (*P < 0.05). (E) IHC staining of Cx58 in lung tumor nodules from mice injected with A549-Cx58shRNA cells and mice injected with A549-shNC cells (scale bar of the left panel: 100 μm; scale bar of the right panel: 10 μm). (F) Cx58 protein levels in A549-Cx58shRNA cells and A549-shNC cells were evaluated by western blotting. GAPDH was used as a loading control.

Cx58 expression is regulated by MEF2B

Transcription factors that regulate connexin expression are among the most important points [ 33– 35]. To our knowledge, there are no reports on the transcriptional regulation of Cx58 expression. JASPAR database analysis revealed that Cx58 is a potential target gene of MEF2B [36]. As shown in Figure 4A, there are four potential binding sites of MEF2B in the Cx58 promoter, distributed at –533 bp, –730 bp, –1112 bp and –1276 bp from the transcription start site (TSS), which are referred to as the A, B, C, and D binding sites, respectively. When MEF2B was silenced in A549 cells via specific siRNAs, the mRNA level of Cx58 was significantly reduced ( Figure 4B). Consistent with the changes in Cx58 mRNA levels, the protein level of Cx58 was also decreased when MEF2B was knocked down in A549 cells ( Figure 4C). To determine whether MEF2B regulates the expression of Cx58 by binding directly to putative binding sites, we performed genome-wide chromatin immunoprecipitation (ChIP)-PCR experiments in A549 cells using an anti-MEF2B antibody. The sheared DNA obtained by sonication is shown in Figure 4D. The cross-linked chromatin before sonication was run in the first lane (#1). The sheared DNA shown in the second lane (#2) was between 200 bp and 1000 bp. We then designed four pairs of primers for the potential MEF2B binding sites in the Cx58 promoter. Quantitative PCR analysis of promoter fragments isolated via MEF2B immunoprecipitation revealed that endogenous MEF2B proteins directly bound to sites located at –533 bp (a binding site, fold enrichment: 4.82) and –730 bp (B binding site, fold enrichment: 11.16) in the Cx58 promoter ( Figure 4E). Then, a luciferase activity assay was performed. We constructed three luciferase reporter plasmids: a 1392 bp vector (containing A- and B-binding sites), a 615 bp vector (containing an A-binding site) and a 528 bp vector (no binding sites) ( Figure 4F). In A549-shNC cells, both the 1392 bp luciferase reporter plasmid and the 615 bp luciferase reporter plasmid led to increased reporter activity compared with that in cells transfected with the vector containing no binding sites. The reporter activity decreased in A549 shMEF2B cells relative to that in A549-shNC cells ( Figure 4G). Therefore, these findings suggest that Cx58 is a direct transcriptional target of MEF2B in A549 cells. MEF2B regulates Cx58 expression through a conserved MEF2B binding site in the Cx58 promoter.

Figure 4 .

Cx58 expression is regulated by MEF2B

(A) The JASPAR database shows the potential binding sites of MEF2B in the Cx58 promoter, named A, B, C, and D, in that order. TSS, transcription start site. (B) MEF2B and Cx58 mRNA levels in A549 cells transfected with MEF2B-siRNAs (siRNA1 and siRNA2) or negative controls were detected by qPCR. (C) Cx58 protein levels in A549 cells transfected with MEF2B-siRNA or negative control siRNA were detected by western blotting, and GAPDH was used as a loading control. (D) The DNA for ChIP-qPCR after sonication (#2) was sheared between 200 bp and 1000 bp. The first lane (#1) shows chromatin before sonication. (E) ChIP-qPCR assay was performed to measure the occupancy of the Cx58 promoter by MEF2B in A549 cells. Data are presented as the mean ± SD (n=3). (F) Schematic of pGL3 reporter plasmids carrying potential MEF2B-binding sites. Three pGL3 reporter plasmids carrying putative MEF2B-binding sites were constructed: 1392 bp (containing A- and B-binding sites), 615 bp (containing A-binding sites), and 528 bp (nonbinding sites). (G) Luciferase assay: three pGL3 reporter plasmids were transfected into A549 cells together with renilla as indicated. Luciferase activity was measured in A549-MEF2B-shRNA (shRNA1 and shRNA2) cells and control cells after transfection with the three reporter plasmids. The results are shown as relative luciferase units (RLUs) compared with Renilla luciferase. Data are presented as the mean ± SD (n=3), ***P < 0.001.

MEF2B knockdown inhibits cell invasion and metastasis, which can be reversed by Cx58 overexpression

Increasing evidence has shown that MEF2B is closely related to EMT [ 26, 27, 37]. MEF2B directly regulates a diverse set of genes involved in cell migration, survival and EMT [27]. MEF2B is increased in several types of carcinoma, including NSCLC [ 25, 27]. As shown above, Cx58 is a prometastatic gene that is regulated by MEF2B in NSCLC. Thus, we speculated that MEF2B is involved in NSCLC metastasis. We conducted wound healing and transwell invasion assays after MEF2B silencing. Compared with control treatment, MEF2B suppression impaired the migration and invasion ability of A549 cells ( Figure 5A–D). Moreover, MEF2B suppression also induced F-actin cytoskeleton changes, as shown in Figure 5E. Compared with that in control cells, F-actin in A549 cells was disrupted after the decrease in MEF2B expression. When EMT-associated proteins were detected, markers of EMT presented associated changes to some extent. Notably, ZO1 expression clearly decreased ( Figure 5F). Using the Kaplan-Meier plotter online tool [30], we found that the expression of MEF2B was negatively correlated with the overall survival of patients with NSCLC (HR = 1.4, log-rank P = 2.2 × 10 ^‒5; Figure 5G). Moreover, the coexpression levels of Cx58 and MEF2B in NSCLC patients were negatively correlated with overall survival (HR = 1.63, log-rank P = 2.7 × 10 ^‒9; Figure 5H). In vitro, as shown in Supplementary Figure S1C,D, we conducted wound healing and transwell invasion assays after Cx58 was overexpressed in A549 shMEF2B cells. Compared with the control, MEF2B deletion impaired both the migration ability and invasion ability of A549 cells, and overexpression of Cx58 reversed the promoting effect of MEF2B knockdown on A549 cell metastasis. Therefore, the MEF2B/Cx58 axis may be involved in the metastasis of NSCLC through the induction of cytoskeleton rearrangement. The inhibition of MEF2B/Cx58 is expected to ameliorate the metastasis of NSCLC.

Figure 5 .

MEF2B knockdown inhibits cell invasion and metastasis, which can be reversed by Cx58 overexpression

(A–D) Wound healing assays (A,B) and transwell assays (C,D) were performed in A549 control cells and MEF2B-silenced cells to examine their migration and invasion capabilities (scale bar: 100 μm). Data are presented as the mean ± SD (n = 3), **P < 0.01, ***P < 0.001. (E) The cytoskeleton was detected by an immunofluorescence assay of F-actin in A549 control cells and MEF2B-silenced cells (scale bar: 25 μm). (F) EMT-related proteins, including E-cadherin, N-cadherin, vimentin and ZO-1, were detected by western blot analysis in A549 cells lacking MEF2B and control cells. GAPDH was used as a loading control. (G) Kaplan-Meier survival analysis of the TCGA array data of MEF2B in NSCLC patients via kmplotter online. Red line, high MEF2B expression (n = 1445); black line, low MEF2B expression (n = 481). (H) Kaplan-Meier survival analysis of TCGA array data related to the coexpression of Cx58 and MEF2B in NSCLC patients. Red line, high coexpression (n = 1443); black line, low coexpression (n = 483).

Discussion

The paradigm that Cxs and enhanced GJIC are suppressors of cancers has been challenged by much evidence, suggesting a conflicting role for Cxs in cancers [ 7, 13, 11, 38, 39]. In our study, we identified increased expression of the Cx58 protein in primary NSCLC tissues for the first time. Higher expression of Cx58 was associated with worse prognosis in NSCLC patients. In vitro, when Cx58 was knocked down in NSCLC cells, migration and invasion were inhibited, and the cytoskeleton was reorganized. In addition, inhibition of Cx58 decreases the homing capacity of NSCLC cells in vivo, strengthening that Cx58 is a prometastatic gene in NSCLC. During metastasis, EMT enhances the migratory and invasive potential of cancer cells [ 32, 40]. However, EMT is not a binary biological process [41]. During cancer progression, cancer cells present with different EMT spectra [42]. Recently, increasing evidence suggests that partial EMT, an intermediate status between the epithelial and mesenchymal phenotypes, is associated with a more plastic status during metastasis [43]. Our findings revealed that deletion of Cx58 could induce the coexpression of both epithelial and mesenchymal markers in NSCLC cells into an intermediate state. In the future, more evidence concerning the role of Cx58 in partial EMT in vivo is needed.

Channel activity, including GJIC and hemichannel communication, is the primary function of Cxs [21]. In addition to their function as channels, their interaction with signaling molecules (C-termini), also called the gap junction proteome, is independent of their channel activity [ 7, 19, 44]. A growing amount of evidence has shown that a gap junction channel-independent pathway plays a vital role in tumor development and progression [ 7, 13, 18]. Although the mechanism by which Cx58 regulates migration and invasion is unclear, the current evidence suggests that it might be involved in its channel-independent function [ 19, 32]. We found that the Cx58 protein is located mostly in the cytoplasm rather than on the cell membrane ( Figure 1A) in tumor tissues. These findings indicate that a function independent of the channel of Cx58 might play a role in NSCLC progression, which needs to be investigated in future experiments. ZO-1 is considered one of the most common proteins that interact with Cxs via the PDZ domains [ 7, 11, 32]. It is predicted that Cx58 could also interact with ZO-1 [45]. In our study, ZO-1 expression was significantly reduced at the protein level after silencing of Cx58, suggesting that Cx58 may play a crucial role in the stability of ZO-1. Hence, ZO-1, a key scaffolding and cytoskeletal molecule, may contribute to the evaluation of the role of Cx58 in both lung tissue homeostasis and lung cancer biology in future studies.

The connexin family is under extensive regulation at many different levels, from DNA to RNA to protein [ 35, 46, 45, 47]. Here, we first reported that Cx58 is under direct regulation by MEF2B, a transcriptional activator that may act as an oncogene in lung cancer [26]. Studies on MEF2B have focused mainly on lymphoma. Microarray data have shown that MEF2B target genes are enriched in the regulation of cell migration and the EMT process [27]. We first revealed that MEF2B acts as a prometastatic gene in NSCLC and directly regulates Cx58 transcription. Knockout of either Cx58 or MEF2B attenuated the NSCLC metastatic traits involved in the regulation of cytoskeleton reorganization. The MEF2B/Cx58 axis may help us to better understand the mechanism of NSCLC metastasis. Nevertheless, there are limitations in the evidence concerning the direct regulation between MEF2B and Cx58. The use of mutant constructs and confocal staining of Cx58 and MEF2B in NSCLC tissues performed in future studies will increase the understanding of the role of the MEF2B/Cx58 axis in NSCLC.

In summary, increasing experimental evidence links Cxs or gap junctions to clinical use. Our study improves our understanding of the role of Cxs in NSCLC. We demonstrated that Cx58 promotes NSCLC metastasis by inducing cytoskeleton rearrangement. MEF2B, as a transcriptional activator of Cx58, may play a vital role in this pro-metastatic process, which highlights its potential as a candidate therapeutic target to improve the survival of NSCLC patients with metastatic disease. This study demonstrated the prometastatic role of Cx58 in NSCLC and identified a novel therapeutic target for NSCLC patients.

Supplementary Data

Supplementary data is available at Acta Biochimica et Biophysica Sinica online.

COMPETING INTERESTS

The authors declare that they have no conflict of interest.

Funding Statement

This work was supported by the grants from the National Natural Science Foundation of China (No. 82373089) and the Scientific Research Launch Project for new employees of the Second Xiangya Hospital of Central South University.