CD36 promotes the epithelial–mesenchymal transition and metastasis in cervical cancer by interacting with TGF-β

Streptavidin-perosidase (SP) and diaminobenzidine (DAB) kits were purchased from Maixin Biotechnology Company (Fuzhou, China). The following primary antibodies were used for Western blot analysis: anti-CD36 (Santa Cruz, CA, USA), anti-E-cadherin, anti-TGF-β, anti-vimentin, anti-snail, anti-slug, anti-twist (Proteintech, IL, USA), and anti-GAPDH (Goodhere Biotechnology, Hangzhou, China).

We acquired 133 cases of cervical cancer (CC) and 47 cases of normal cervical tissues between January 2011 to December 2016 from Department of Pathology of the Eighth Affiliated Hospital, Sun Yat-sen University and the Affiliated Hospital of Guangdong Medical University. The diagnoses were conducted by three professional pathologists, and the study was approved by the Institutional Research Ethics Board of the Eighth Affiliated Hospital, Sun Yat-sen University. We purchased the human cervical cancer cell lines C33a, Hce1, HeLa, and SiHa from the China Center for Type Collection (CCTCC) (Wuhan, China). Cell lines were cultured in DMEM (Gibco, CA, USA) medium containing 10% fetal bovine serum (FBS, Sera Gld, Amarica), 100 U/mL of penicillin, and 100 U/mL of streptomycin. All of the cells were incubated in a humidified incubator in 5% CO₂ in compressed air at 37 °C.

Paraffin blocks were cut into 4-μm sections and treated routinely following the reagent instructions. After microwaving in citrate buffer for 5 min, the slides were incubated with anti-CD36 at room temperature. The sections were then incubated with a secondary antibody (MaximBio Company, Fuzhou, China), labeling was monitored using diaminobenzidine (Maxim-Bio Company, Fuzhou, China), and hematoxylin was used to stain the sections. We scored expression in accordance with the intensity (0, no staining; 1, weak staining; 2, moderate staining; 3, strong staining), and the percentage of cervical cancer cells that were stained (0, none stained; 1, < 10% stained; 2, 10–50% stained; 3, > 50% stained; 4, > 75% of all of the cervical cancer cells stained). If the product of multiplying staining intensity by the percentage of positively stained cervical cancer cells was ≥ 2, it was regarded as positive (+).

Homo sapiens CD36 siRNA was obtained from Guangzhou RiboBio Biological Technology (Guangzhou, China), and it targeted the sequence 5′-ACGTATAAGGACCTCTTTG-3′. HeLa and SiHa cells were seeded at 2 × 10⁵ cells/well in six-well plates. We transfected HeLa and SiHa cells with CD36 siRNA or control siRNA (sense strand, 5′- UUCUCCGAACGUGUCACGU TT-3′; antisense, 5′-ACGUGACACG UUCGGAGAATT-3′) with Lipo3000 at a final concentration of 100 nM, and incubated the cells at room temperature for 15 min. The complex was then added to the culture medium for subsequent experiments.

pIRES2-ZsGreen1-CD36 was constructed and amplified by Hanbio Biotechnology (Shanghai China). We selected the cell line with G418 (600 µg/mL) for 3 weeks and expanded it, and C33a cells that overexpressed CD36 were labeled “C33a–CD36” for our study.

The C33a/CD36, C33a/vector (control), SiHa/siRNA, SiHa/nc-siRNA (control), HeLa/siRNA, and HeLa/nc-siRNA (control) cells were seeded into 12-well plates at a density of 1 × 10⁵ cells/well. The cells were then scraped with a 200-μL sterile pipette tip when they formed monolayers. After washing the cells three times with PBS, we used serum-free medium for culture, and photographed the cells at 0 and 48 h.

We performed the invasion assay using transwell plates (Costar, USA). The C33a/CD36, C33a/vector, SiHa/siRNA, SiHa/nc-siRNA, HeLa/siRNA, and HeLa/nc-siRNA cells (each at a density of 1 × 10⁵ cells/well) were added to the upper chamber with 0.2 mL of serum-free RPMI-1640; we added 0.5 mL of 10% FBS medium to the lower chamber. The cells were allowed to invade for 48 h at 37 °C. After removing the cells on the upper surface of the membrane, we stained cells on the lower aspect with trypan blue.

We adjusted the concentrations of C33a/CD36, C33a/vector, SiHa/siRNA, SiHa/nc-siRNA, HeLa/siRNA, and HeLa/nc-siRNA cells to appropriate densities, and then inoculated each culture dish with 200 cells at 37 °C, changing the medium every 4 days. After 2 weeks, cells were stained with trypan blue, and numbers of cell colonies were counted using a light microscope.

Analysis of cellular apoptosis was conducted strictly following the instructions of the apoptosis kit (KeyGEN BioTECH, Nanjing, China). C33a/CD36, C33a/vector, SiHa/siRNA, SiHa/nc-siRNA, HeLa/siRNA, and HeLa/nc- siRNA cells were incubated with the DNA-binding dye propidium iodide (50 ug/mL) and RNase (1.0 mg/mL) for 20 min at 37 °C in the dark. We then washed the cells and analyzed the emitted red fluorescence with a flow cytometer (BD, Heidelberg, Germany).

C33a/CD36, C33a/vector, SiHa/siRNA, SiHa/nc-siRNA, HeLa/siRNA, and HeLa/nc-siRNA cells were cultured in 24-well plates (at 1 × 10⁵/mL) overnight. The cells were washed twice with ice-cold PBS, fixed with 4% paraformaldehyde for 20 min, and then permeabilized with 0.5% Triton X-100 for 10 min at room temperature. The samples were then incubated with primary antibodies, including CD36 (1:50 dilution) and vimentin (1:100 dilution) at 4 °C overnight. After washing 3 times with PBS, we incubated the samples with secondary antibodies (Alexa Fluor 488, Alexa Fluor 594, 1:500 dilution) mixed with DAPI (1:1000 dilution) for 1 h in the dark. After washing three times and covering the samples with anti-fluorescence quencher, we recorded the images using a fluorescence microscope.

We lysed the cells in a lysis buffer after washing twice with ice-cold PBS, and quantified total protein concentrations with a BCA kit (Beyotime Biotechnology, Guangzhou, China). Twenty micrograms of total protein was boiled for 5 min before being loaded onto 10% polyacrylamide gels and transferred to a polyvinylidene fluoride (PVDF) membrane. The membranes were incubated with primary antibody, including anti-CD36 (1:200 dilution), anti-E-cadherin (1:500 dilution), anti-TGF-β (1:500 dilution), anti-vimentin (1:500 dilution), anti-snail (1:500 dilution), anti-slug (1:500 dilution), anti-twist (1:500 dilution), anti-GAPDH (1:1000 dilution), and anti-β-actin (1:1000 dilution) at 37 °C overnight. Next, the membranes were incubated with a secondary antibody for 1 h, and the specific protein bands on the membranes were detected using an enhanced chemiluminescence kit (Beyotime, China).

A nude mouse xenograft model was established using 4-week-old female BALB/C nude mice obtained from Hunan SJA Laboratory Animal Center. C33a/CD36 and C33a/vector cells (2 × 10⁶/mL) were subcutaneously injected into the lower abdomen or tail vein of the nude mice, and the tumor diameters for each mouse were measured weekly. After 5 weeks, we euthanized the mice using anesthesia, and the tumors were removed and measured. All of the animal protocols were conducted in accordance with the Institutional Animal Ethics Care Committee.

All statistical analyses were conducted using SPSS 17.0 (SPSS, Inc., Chicago, IL, USA). We performed experiments in triplicate, and data are presented as mean ± SEM. The χ2 test was used to analyze the relationship between CD36 levels and clinicopathologic characteristics. Data from two groups were analyzed by unpaired t tests; and, if more than two groups, by one-way ANOVA. A P value of < 0.05 was considered statistically significant.

After investigating the expression levels of CD36 in 133 cases of cervical cancer tissues and 47 cases of normal cervical tissues by immunohistochemistry, we found CD36 expression to be primarily located on the cell membrane or in the cytoplasm (Fig. 1a–d). As shown in Table 1, CD36 immunoreactivity was detected in 73.68% (98/133) of cervical cancer cases, whereas CD36 was only expressed in 19.15% (9/47) of normal cervical tissues. Furthermore, when we evaluated the correlation between CD36 expression and various clinicopathologic parameters of patients, we observed that CD36 expression was correlated with tumor differentiation and lymph node metastasis (P < 0.05, Table 2), but not with other clinicopathologic features, such as age and clinical stage. In addition, using Kaplan–Meier survival analysis and TCGA datasets, we showed that patients with cervical cancer and high CD36 expression levels had an unfavorable prognosis relative to those with low or medium CD36 expression (Fig. 1e) (P = 0.02). These results indicated that CD36 might be involved in the progression of cervical cancer.

To further investigate the effects of CD36 on a series of biologic processes in cervical cancer cell lines, we used Western blotting analysis to detect the expression of CD36 in C33a, Hce1, HeLa, and SiHa cells. Our data suggested that the expression of CD36 in Hce1 (96.60 ± 5.66), HeLa (134.24 ± 4.67), and SiHa (95.49 ± 2.49) cells was markedly higher than in C33a (44.77 ± 1.35) cells (Fig. 2a, b). Hence, we selected C33a, HeLa, and SiHa cells to study whether CD36 exerted tumor-promoting effects by assaying for cell migration, invasion, proliferation, and apoptosis. Cells were stably transfected with plasmids overexpressing CD36 or knocking down CD36 expression: C33a–CD36, C33a/vector (control); SiHa/siRNA, SiHa/nc-siRNA (control); and HeLa/siRNA, HeLa/nc- siRNA (control).

As shown in Fig. 2c, d, si-RNA-2 was the most effective si-RNA sequence for interference CD36 expression. After transfected with CD36 siRNA-2 in SiHa and HeLa cells, expression levels of CD36 were significantly decreased (Fig. 2e). However, expression levels of CD36 were significantly enhanced after transfected with plasmid overexpressing CD36 in C33a cells (Fig. 2f). SiHa- and HeLa cells-transfected CD36 siRNA-2 were used to ascertain the effects of CD36 on migration and invasion of cervical cancer cells by trans-well and wound-healing assays. Compared with cells transfected with control siRNA (nc-siRNA), interference of CD36 expression by siRNA inhibited migration both in SiHa and HeLa cells (Fig. 2g–i). Conversely, enhancement of CD36 expression by transfecting a CD36 expression vector promoted migration of C33a cells (Fig. 2g, j). Similarly, overexpression of CD36 promoted invasive ability of C33a cells, while interference of CD36 expression markedly inhibited the invasion-promoting effect in SiHa and HeLa cells (Fig. 3a, b, d, e). Therefore, our evidence verified a vital role for CD36 in the promotion of cervical cancer cell migration and invasion in vitro.

Colony formation assays and flow cytometry were used to examine the proliferative and apoptotic effects of CD36 in cervical cancer cell lines. Our results revealed that high CD36 expression facilitated colony formation in C33a cells (Fig. 3f, j), while interference of CD36 expression attenuated colony formation in SiHa and HeLa cells (Fig. 3f, i). Knocking-down the expression of CD36 increased the rate of apoptosis of SiHa and HeLa cells (Fig. 3c, g, h). In contrast, enhancing CD36 expression inhibited apoptosis in C33a cells (Fig. 3c, k).

In summary, the above results suggest that CD36 plays critical roles in the biologic behavior of cervical cancer cells, promoting proliferation, migration, and invasion, but inhibiting apoptosis.

We found that the morphology of SiHa and HeLa cells (normally manifesting a spindle-invasive phenotype) was replaced by a polygonal phenotype after knockdown of CD36 expression. However, the morphology of C33a cells changed from a polygonal-shaped phenotype to mesenchymal-like morphology under CD36 expression (Fig. 4a).

Having uncovered a close association between CD36 and invasion of cervical cancer cells, we sought to elucidate the effect of CD36 on the EMT in these cells. To evaluate this issue, we compared expression levels of some critical EMT markers with CD36 in 60 samples. As shown in Fig. 4b, the expression of E-cadherin was attenuated with elevated CD36 expression (P < 0.05), while vimentin expression levels were positively correlated with CD36 (P < 0.01).

We used immunofluorescence to detect the expression levels of vimentin in SiHa and HeLa cells after treatment, and demonstrated that knockdown of CD36 expression by transfection with siRNA led to diminished expression of vimentin (Fig. 4c, d). Collectively, our data suggest that CD36 participates in the EMT in cervical cancer.

To study the effects of CD36 on tumor metastasis in vivo, C33a–CD36 and C33a cells were injected into nude mice via the abdomen or tail vein. After 5 weeks (consistent with in vitro data), we observed that overexpression of CD36 significantly promoted tumor growth in female BALB/c-nu mice (Fig. 5a). In addition, although all of the mice that received tail vein injections of C33a–CD36 and C33a cells developed micro-metastases in the liver, more numerous and enlarged metastatic liver foci were observed after overexpression of CD36 (Fig. 5b–g). These results suggest that CD36 promotes the growth and metastasis of cervical cancer cells in vivo.

To further validate CD36 promotion of cervical cancer metastasis by regulating EMT, we attempted to detect the effect of CD36 on some critical EMT markers. As shown in Fig. 6a–d, compared with their respective control cells, knockdown of CD36 led to a dramatic increase in E-cadherin protein expression that resulted in a statistically significant decrease in vimentin, slug, snail, and twist protein expression, both in SiHa and HeLa cells. Conversely, the protein expression levels for vimentin, slug, snail, and twist were upregulated when transfected with exogenous CD36 in C33a cells, although E-cadherin was down-regulated (Fig. 6a–d).

Investigators have recently demonstrated that TGF-β is a vital inducer of EMT. Hence, we further evaluated the association among CD36, TGF-β, and EMT. In the present study we observed that after TGF-β treatment, the expression of CD36 was elevated not only in si-SiHa and si-HeLa cells, but also in C33a–CD36 cells, indicating that TGF-β promoted CD36 expression (Fig. 6a–d). More importantly, we demonstrated that TGF-β treatment attenuated E-cadherin expression and enhanced the expression levels of vimentin, slug, snail, and twist in si-SiHa, si-HeLa, and C33a–CD36 cells (Fig. 6a–d), suggesting that TGF-β synergized with CD36 on EMT via active CD36 expression. We also observed that the expression levels of TGF-β in si-SiHa cells and si-HeLa cells were down-regulated, whereas the expression levels of TGF-β were up-regulated in C33a–CD36 cells (Fig. 6e). These results imply that CD36 and TGF-β interact with each other to promote EMT in cervical cancer.