Cancerous lung cell lines (CL1-0, CL1-5, A549, and H1299) (gifts from Dr Pan-Chyr Yang) and normal cells (BEAS-2B) (a gift from Dr Pan-Chyr Yang) were cultured in RPMI-1640 medium (Gibco; Thermo Fisher Scientific) with 1% streptomycin/puromycin (Biological Industries) and 10% fetal bovine serum (FBS; Biological Industries). Brain metastatic lung adenocarcinoma Bm7 cells (generated in the authors′ laboratory) were cultured in DME/F12 plus 10% FBS (43). The cultured plates were maintained at 37°C in a humidified atmosphere with 5% CO₂. Another normal lung cell line (MRC-5) and human bronchial epidermal cells (16HBE) (a gift from Dr. Kuo-Ting Chang) were grown in Eagle's Minimum Essential Medium (Gibco; Thermo Fisher Scientific) under the same conditions.

Cell experiments were performed on cells that were passaged <20 times, and were routinely tested for mycoplasma using the PCR Mycoplasma Detection kit (ABM Inc., Vancouver, Canada). The identity of the cell lines was authenticated by short-tandem repeat (STR) analysis (Mission Biotech Inc., Taipei, Taiwan) in February, 2018.

To quantify the transcriptional expression of SEMA5A in different cellular models, total RNA was isolated using TRIzol reagent (Ambion) and purified by precipitation with isopropanol (Sigma-Aldrich). A NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific) was used to assess the purity and quantity of the RNA. A high Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific) was used to synthesize the cDNA from 1 µg of total RNA of each cell line. The final cDNA products were used as the templates for expression analysis using quantitative PCR (qPCR).

The quality and quantity of the RNA were measured using a NanoDrop™ 2000 spectrophotometer (Thermo Fisher Scientific). A total of 1 µg of total RNA from each cell line was reverse transcribed using the High Capacity cDNA Reverse Transcription kit (Thermo Fisher Scientific). The reaction was carried using thermal cycling conditions, including annealing at 25°C for 5 min, an extension temperature at 42°C for 1 h and inactivation temperature at 70°C for 15 min. The final cDNA products were used as the templates for subsequent qPCR with the following thermal cycling conditions. Denaturation temperature at 95°C for 15 sec, anneal/extend temperature: 60°C for 1 min for 40 cycles. RT-qPCR was performed using SYBR-Green (Roche) on an ABI 7900 system (Life Technologies; Thermo Fisher Scientific) according to standard protocols. All individual experiments were carried out in triplicate. Relative quantification, ΔΔCq (cycle threshold) (44), was applied using glyceraldehyde 3-phosphate dehydrogenase (GAPDH) as an internal control and empty group as a reference control.

As a number of tumor suppressor genes are inactivated via hypermethylation within the promoter region (45,46), this study examined the role of methylation in regulating the expression of SEMA5A. The A549, H1299 and BEAS-2B cells were seeded on a 6-well plate, and after 24 h, the cells were treated with 5 or 10 µM of 5-aza-2′-deoxycytidine (5-aza; Sigma-Aldrich). The quantification of the expression levels was carried after 3 days of treatment.

To quantify the methylation levels of multiple CpG sites in the 5′ untranslated region of SEMA5A, bisulfite treatment was first performed with an EpiTect Fast bisulfite conversion kit (Qiagen) according to the manufacturer's instructions, and a pre-designed PyroMark CpG assay (Hs_SEMA5A_03_PM PyroMark CpG assay, PM00021203, Qiagen) was used. Specific CpG sites, including Chr5: 9545395, 9545400, 9545407, 9545409 and 9545417 were examined.

Full-length SEMA5A cDNA (3,225 bp) tagged with a Flag epitope was inserted into a pZeoSV2⁺ viral expression vector with the NotI and AsiSI restriction enzymes (Addgene). The pZeoSV2⁺-SEMA5A-Flag plasmid was transiently transfected into the A549 and H1299 cell lines using TransIT-2020 transfection reagent (MirusBio) according to the manufacturer's instructions. The empty pZeoSV2⁺ viral expression vector was used as a control. All sequences were verified by Sanger sequencing (First Core Laboratory, College of Medicine, National Taiwan University). mRNA levels were quantified by RT-qPCR using SEMA5A-specific primers (forward, 5′-GTCTATACTTACTGCCAGCG-3′ and reverse, 5′-GTTAAATGCCTTGATGGCCTC-3′) and GAPDH-specific primers (forward, 5′-TGCACCACCAACTGCTTAG-3′ and reverse, 5′-GATGCAGGGATGATGTTC-3′). Protein levels were examined by western blot analysis.

To examine the proteolytic effects of ADAM17 on SEMA5A, the H1299 cells were transfected with SEMA5A overexpression plasmid and treated with active recombinant ADAM17 (BioVision). The amount of administering ADAM17 was 0.66 µg for 6-cm dish. Cell lysates and media were collected at 2 days following transfection. Bovine serum albumin (BSA) was added externally to the culture media as the spike-in control, and all proteins in the medium were concentrated using a 10K Acrodisc syringe filter (Pall Life Sciences).

Total cell lysates of A549 and H1299 cells transfected with pZeoSV2⁺-SEMA5A-Flag plasmid or empty vector were prepared. Proteins (30 µg) were separated by 10% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and then electrotransferred onto polyvinylidene difluoride (PVDF) membranes (Bio-Rad Laboratories). After blocking with 5% milk, the membranes were incubated with anti-SEMA5A (1:1,000, Cat. no. PA5-26066, Thermo Fisher Scientific), anti-GAPDH antibody (1:20,000, Cat. no. GTX100118, GeneTex), or anti-BSA antibody (1:5,000, Cat. no. GTX79812, GeneTex). Following incubation overnight at 4°C, the membranes were then incubated with horseradish peroxidase-conjugated anti-IgG (1:5,000, Cat. no. GTX213110, GeneTex) at room temperature for 1 h, and the blots were developed with the chemiluminescent western blotting reagent (Millipore). Western blot images were further analyzed using Gel-Pro Analyzer v6.3 software (Meyer Instruments) to obtain the optical density values of SEMA5A and GAPDH antibodies.

The mRNA detection and analysis were performed as previously described (47). Briefly, the cDNA was synthesized from total RNA primed with T7 Oligo(dT) and amplified using an Illumina TotalPre RNA Amplification kit (Ambion). The second strand of cDNA was synthesized by employing DNA polymerase and RNAase H to simultaneously degrade the RNA and synthesize the second strand of cDNA. Following cleanup, in vitro transcription was conducted to synthesize biotinylated complementary RNA (cRNA). Following amplification, the cRNA was hybridized to Illumina Human HT-12 v4 BeadChips (Illumina) for 16 h. Following hybridization, the BeadChip was washed and stained with streptavidin-Cy3 dye. The intensity of the beads′ fluorescence was detected by HiScan SQ (Illumina), and the results were analyzed using BeadStudio v2011.1 software.

After scanning, the intensity data of Illumina BeadChips were analyzed using Partek v7.0 software (Partek). Background-adjusted signals were normalized by a quantile normalization algorithm. Student′s t-tests and Bonferroni P-value adjustment were utilized to identify differentially expressed genes. Principle component analysis (PCA) was utilized to evaluate the similarity of the gene expression profiles. Hierarchical clustering analysis and the Genesis program were used to generate visual representation of expression profiles. All data have been deposited at the Gene Expression Omnibus (GEO, GSE114578). Furthermore, ingenuity pathway analysis (Ingenuity Systems Inc.) was applied to comprehend the biological functions and signaling pathways of differentially expressed genes.

A total of 3,000 lung adenocarcinoma cells (A549 & H1299) were seeded in 96-well plates in triplicate and incubated for 12 h at 37°C in a CO₂ incubator. The day after seeding, the cells were transfected with pZeoSV2⁺-SEMA5A-Flag plasmid or empty vectors. Following transfection, the cells proliferative activity was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) (EMD Biosciences) assay at 24, 48 and 72 h, respectively. The absorbance of the A549 and H1299 cells was then measured using a microtiter plate reader (BioTek) at 570 nm.

The A549 (3,000 cells) or H1299 (3,000 cells) cells were seeded on silane coated micro slides (Muto pure chemicals Co.) with the same cell density and transfection concentration as MTT assay. The cells were fixed with 4% paraformaldehyde (Sigma-Aldrich) at 48 h following transfection. The cell membrane was damaged by 0.1% Triton X-100 (Mallinckrodt Specialty Chemicals Co) and blocked with 2% bovine serum albumin (Sigma-Aldrich). Ki-67 primary antibody (1:400, #9129, Cell Signaling Technology) and anti-rabbit IgG secondary antibody (1:800, #8889, Cell Signaling Technology) were used for Ki-67 detection at a wavelength of 550/580 nm. The incubation conditions for the primary antibody were 4°C overnight, and for the secondary antibody they were 2 h at room temperature. Cell nuclei were stained with Hoechst 33342 (#B2261, 1 µg/ml, Sigma-Aldrich) and detected at wavelength of 360/460 nm. The incubation conditions for Hoechst 33342 were 30 min at room temperature. Images were acquired using the Zeiss AxioImager A1 system (Carl Zeiss).

The A549 (300 cells), H1299 (300 cells) and Bm7 (300 cells) cells were first seeded in 6-well plates for 24 h, and transfected with pZeoSV2⁺-SEMA5A-Flag plasmid or empty vectors. After 2 weeks, the cells were fixed using 3:1 methanol-acetic acid and stained using 0.1% crystal violet (Sigma-Aldrich) for 10 min at room temperature. Finally, the stained plates were dried and used for image acquisition using microscopic (×100 magnification) evaluation. Colonies containing >50 cells under a stereomicroscope were counted.

Following 24 h of transfection, 2×10⁴ cells of A549 and H1299 were seeded into 24-well plates with a sterile culture insert and grown to ~90% confluence in RPMI-1640 medium containing 10% FBS and changed to 2% FBS when removing the cassettes. The culture inserts were removed and an image of the gap at 0 h was captured. The cells were further incubated at 37°C in a CO₂ incubator for 36 h, and images were captured at 24 and 36 h to measure the progress of gap closure.

Migration assays were carried out using a 24-well Transwell unit (Corning, Inc.). The upper chamber of the Transwell unit was loaded with 4×10⁴ cells/well in 0.2 ml serum-free RPMI-1640 medium, and the lower chambers were loaded with 0.6 ml RPMI-1640 containing 10% FBS as a chemoattractant. The A549 and H1299 cells were then incubated for 24 h at 37°C. A methanol-acetic acid (3:1) mixture was then added to the lower chamber to fix the cells for 20 min at room temperature followed by staining with 0.1% crystal violet for a further 20 min at room temperature. Cells on the upper side of the membrane surface were removed by scraping with a cotton swab, and the cells that passed through the filter were de-stained using 10% acetic acid. The absorbance was measured at 570 nm with an enzyme-linked immunosorbent assay (ELISA) reader (BioTek Instruments). Cells transfected with empty vector were used as the controls.

The experiment was performed as previously described (43). Briefly, cells were cultured on dishes coated with collagen (10 µg/ml) overnight and then cultured in serum-free conditioned medium. Cell movement was detected under A-Plan objectives (X5; 0.55 NA) using an inverted microscope (Axio Observer Z1, Zeiss) in 37°C chambers. Images were collected from CCD video cameras (AxioCam MRm, Zeiss) at 20-min intervals for a total of 16 h using MetaMorph software (Molecular Devices Corp.). The accumulated distance was measured by tracking each cell nuclei for 30 individual cancer cells in each group using the Track Point function of NIH ImageJ v1.43 software.

The SEMA5A-Flag plasmid was transiently transfected into Bm7 lung cancer cells with stable luciferase expression. The following day, the Bm7 cells (2×10⁶ cells in 50 µl of PBS) were mixed with 50 µl of Matrigel and then implanted into 6–8 week-old male SCID mice (weighing 20–25 g; BioLASCO) by the subcutaneous route at two separate sites (right and left side) of the back (each group had 4 mice; 2 groups). Mice were housed in specific pathogen-free rooms with one group in one cage at room temperature (20–23°C), 40–60% relative humidity and under a 12-h light-dark cycle, with free access to food and water. No cachexia and ascites in mice were observed in this study. As Bm7 lung tumors had luciferase, the tumor volume was detected using an IVIS Spectrum Imaging system (Xenogen). Prior to image acquisition, the mice were subsequently administered D-luciferin via intraperitoneal injection and photons emitted from the mice were detected using an imaging system. Tumor size and distribution in vivo were quantified as photons/second. The maximum length and width exhibited by a single subcutaneous tumor in this study did not exceed 1.5 cm. The maximum tumor volume calculated by signal was 500 mm³ using the formula, (length × width²)/2. Anesthesia was induced and maintained with 2.5% isoflurane (Panion & BF Biotech Inc.) in 100% oxygen in an anesthetic chamber when mice were monitored by an IVIS Spectrum Imaging system. The mice were sacrificed by using carbon dioxide (CO₂) at a displacement rate from 10 to 30% of the chamber volume per minute for mouse euthanasia and death was verified by physical methods following euthanasia on day 21, including no heartbeat, no pupillary response to light and no respiratory pattern.

This animal model followed protocols approved by the Institutional Animal Care and Use Committee of China Medical University and Hospital (animal protocol no. 2016-102). Suitable humane endpoints were included in the approved animal experiments. In this study, the mouse experiment was terminated by euthanasia (using carbon dioxide) on day 21 or animals that reached humane endpoints when the tumor size was near 1,000 mm³ or unexpected circumstances, such as illnesses (infection, difficulty breathing, etc.) and injuries (necrosis or bleeding in tumors). The mice were sacrificed by using carbon dioxide (CO₂) at a displacement rate from 10 to 30% of the chamber volume per minute for mouse euthanasia and death was verified by methods, including no heartbeat, no pupillary response to light, and no respiratory pattern for at least 5 min.

Kaplan-Meier survival analyses were conducted using Kaplan-Meier Plotter for lung cancer (48,49), an online platform (http://www.kmplot.com/lung), using SEMA5A probe (Affy ID: 229427_at) and adenocarcinoma in histology restriction. In total, Kaplan-Meier Plotter analyzed 1,715 patients for survival analysis, which included 1,120 patients in 7 GEO datasets (GSE4573, GSE14814, GSE8894, GSE19188, GSE3141, GSE31210, GSE29013 and GSE37745), 462 patients in the caArray, and 133 patients in The Cancer Genome Atlas (TCGA). The log rank test was used to determine differences in the survival rate between the high and low expression groups.

Data are expressed as the means ± standard deviation (SD) of at least 3 independent experiments. Student's t-tests and Bonferroni P-value adjustment were utilized to identify differentially expressed genes. P-values <0.05 were considered to indicate statistically significant differences.

In previous genomic studies, it was found that the downregulation of SEMA5A in non-smoking women with lung adenocarcinoma was associated with a poor overall survival (29,50,51). To validate this phenomenon in a Caucasian population, another publicly available dataset was examined (52). As shown in Fig. 1A, SEMA5A was also significantly (P<0.01) downregulated in cancer tissues as compared to adjacent normal control tissues. Furthermore, the survival of lung adenocarcinoma patients in the new dataset for >15 years was examined using Kaplan-Meier plotter (www.kmplot.com/lung) (49). In total, 1,715 patients were used for survival analysis, which included 1,120 patients in 7 GEO datasets (GSE4573, GSE14814, GSE8894, GSE19188, GSE3141, GSE31210, GSE29013 and GSE37745), 462 patients in the caArray, and 133 patients in The Cancer Genome Atlas (TCGA). Patients were divided into the high and low expression groups based on the median expression value of SEMA5A. The results of Kaplan-Meier survival analysis revealed that the lung adenocarcinoma patients with higher expression levels of SEMA5A had a lower risk of death (hazard ratio, 0.65), i.e., a higher overall survival probability (Fig. 1B), indicating the potential utility of SEMA5A as a prognostic marker for patients with lung adenocarcinoma.

Since SEMA5A was downregulated in non-smoking female lung adenocarcinoma patients (29), the endogenous expression levels of SEMA5A were then examined in 4 lung adenocarcinoma cell lines (CL1-5, CL1-0, A549 and H1299), 2 normal lung cell lines (MRC-5 and BEAS-2B) and human bronchial epithelial cells (16HBE). The results of RT-qPCR revealed that SEMA5A was significantly (P<0.01) downregulated in all lung adenocarcinoma cell lines (Fig. 2A). The results of western blot showed that the endogenous SEMA5A proteins in lung adenocarcinoma cell lines were lower than that of BEAS-2B (Fig. 2B); however, it was not downregulated the 2 other normal control cell lines (MRC-5 and 16HBE). Therefore, the A549, H1299 and BEAS-2B cells were arbitrarily selected for use in the following experiments.

Subsequently, to investigate the mechanisms leading to a lower SEMA5A expression in lung adenocarcinoma cells, we first examined the role of methylation in regulating endogenous expression by treating the A549, H1299 and BEAS-2B cell lines with 5-aza. The expression levels of SEMA5A were significantly (P<0.01) increased in a dose-dependent manner in the A549 and H1299 lung adenocarcinoma cell lines, whereas they were not altered in the BEAS-2B normal cells (Fig. 2C). Moreover, pyrosequencing analysis was performed to identify the specific CpG sites of methylation in SEMA5A. In total, 5 CpG sites in the 5′untranslated region of SEMA5A were examined (Fig. 2D top panel). The 4 CpG sites in the A549 cells and 3 CpG sites in the H1299 cells revealed a significantly higher methylation percentage as compared to the normal BEAS-2B cells (Fig. 2D bottom panel), indicating that methylation plays a role in regulating the expression of SEMA5A in lung adenocarcinoma cells.

Subsequently, SEMA5A was transiently overexpressed in the A549 and H1299 cells. As shown in Fig. 3A, the mRNA levels of SEMA5A in the A549 and H1299 cells following transfection were significantly (P<0.01) increased. Western blot analysis also validated that the SEMA5A protein levels increased in both the A549 and H1299 cells upon transfection with SEMA5A expression plasmid (Fig. 3B). Since a previous study demonstrated that the mature form of SEMA5B was proteolytically processed by ADAM17 (14), this study also examined whether SEMA5A can be cleaved by ADAM17. H1299 cells overexpressing SEMA5A were treated with active recombinant ADAM17. As shown in Fig. 3C and D, the amount of SEMA5A in the cell lysate from the A549 (Fig. 3C) and H1299 (Fig. 3D) cells was lower in the presence of ADAM17. By contrast, the amount of SEMA5A in the medium increased (Fig. 3C and D), suggesting that membrane-bound SEMA5A was released into the medium in the presence of ADAM17.

In order to investigate the functional role of SEMA5A, the functions of SEMA5A-regulated genes were first investigated. Total RNA was extracted 24 h following the transfection of SEMA5A plasmid into the A549 cells. Differentially expressed genes regulated by SEMA5A were identified by Illumina Human HT-12 v4 Bead Chips. The criteria for the selection of differentially expressed genes consisted of a fold change ≥3 and significant differences (P<0.05). In total, 350 differentially expressed probes were identified. The volcano plot illustrated in Fig. 4A represents 350 differentially expressed probes (a.k.a. genes), with 296 probes upregulated (red points) and 54 ones downregulated (green points) (Fig. 4A and Table SI). Genes that did not meet the criteria of differentially expressed genes were shown in gray (Fig. 4A). Principal component analysis, a visualization tool to used illustrate that the expressional profiling of the 350 differentially expressed probes between the SEMA5A group and control group, indicated that the distribution between samples overexpressing SEMA5A (cyan spots) and the controls (black spots) was separated, indicating different expressional profiling of the selected genes between the SEMA5A group and the control group (Fig. 4B).

A heatmap with hierarchical clustering, a common method of visualizing the relative intensity of gene expression by arranging genes together based on the similarity of their expressional levels, is presented in Fig. 4C. The relative intensities of the 296 upregulated probes are shown in red, and those of the 54 downregulated ones are shown in green.

Furthermore, network analyses were performed by ingenuity pathway analysis. The dashed lines indicate indirect interactions between molecules as supported by information in the Ingenuity knowledge base. The results of pathway analysis revealed that these genes were involved in cell signaling, gene expression, cell death and survival, cellular growth and proliferation (Fig. 4D and Table SII). Network analysis also revealed the interactions among some of the differentially expressed genes. In one of the representative networks, although SEMA5A was not enriched in this network, contained genes which were involved in cellular movement, cellular growth and proliferation, as shown in Fig. 4E, which was later validated by further functional analysis.

Based on the function of SEMA5A-regulated genes, the effects of SEMA5A on tumor growth by MTT assays were hence examined. The results revealed a significant decrement in the proliferation of both the A549 and H1299 cells overexpressing SEMA5A (Fig. 5A and B). In agreement with these results, immunohistochemistry assay of Ki-67, a cellular marker for proliferation, also revealed a markedly decreased amount of Ki-67 and cell numbers following transfection with the SEMA5A expression plasmid (Fig. 5C and D). Furthermore, SEMA5A overexpression reduced colony formation in both the A549 (Fig. 5E) and H1299 (Fig. 5F) cells. These results demonstrated the suppressive effects of SEMA5A on the proliferation and colony formation of lung adenocarcinoma cells.

Subsequently, the effects of SEMA5A on the mobility of lung adenocarcinoma cells were investigated by gap closure assay and Transwell migration assays. Both assays revealed that the overexpression of SEMA5A significantly suppressed the migratory abilities of both the A549 (Fig. 5G and I) and H1299 (Fig. 5H and J) cells. Previously, SEMA5A has been shown to impede the motility of human gliomas via the indirect inactivation of RAC1 and FSCN1 (41,42). However, the transcriptional levels of RAC1 and FSCN1 were not altered significantly in this study (Fig. S1).

Lastly, the role of SEMA5A in severe combined immunodeficiency (SCID) mice was examined. Since the growth of tumors in SCID mice using either A549 or H1299 lung cancer cells was not successful in pilot studies by the authors (data not shown), brain metastatic Bm7 lung cancer cells (43) were used to demonstrate the role of SEMA5A in in vivo experiments. Similar results were observed in the in vitro experiments (Fig. 6). SEMA5A overexpression reduced colony formation (Fig. 6A) and cell migration examined by time-lapse video microscopy (Fig. 6B). In the tumor xenograft assays, Bm7 cells overexpressing SEMA5A were injected subcutaneously into the backs of SCID mice to develop subcutaneous tumors. After 13 days, the tumor size was found to be significantly smaller in the SEMA5A group (Fig. 6C). No lymph nodes or distant organ metastases were detected during the experimental period.