Stearoyl-CoA desaturase-1 promotes colorectal cancer metastasis in response to glucose by suppressing PTEN

Pairs of cancer and adjacent noncancerous tissues enrolled in this study were collected from patients diagnosed with colorectal cancer, which obtained from Shanghai General Hospital of Shanghai Jiao Tong University. The study protocol was approved by and in accordance with the guidance of the Research Ethics Committee of Shanghai Jiao Tong University of Medicine, Shanghai, China. The Cancer Browser was linked from the UCSC Cancer Genomics Browser (https://​genome-cancer.​ucsc.​edu/​), colorectal cancer was chosen and then downloaded related datasets. SCD1 items were searched and summed up and then we combined SCD1 expression with clinical information, to build overall survival curve and outcomes comparison.

Human colonic carcinoma cell lines HCT116, Caco2, HT29, SW116 and SW620 were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Giboco, USA), which contained 10% FBS, 2 mM L-glutamine and 1% penicillin/streptomycin (Gibco, USA). Cells were incubated in 37 °C, 5% CO2 humidified atmosphere.

Total RNA was isolated by using TRIzol (Invitrogen Life Technologies, USA). cDNA was made using the PrimeScript™ RT reagent kit (Takara Bio Inc., Japan). Real Time-PCR was performed in triplicate utilizing StepOnePlus™ Real-Time PCR System (Applied Biosystems, USA). The primers were as follows: SCD1 forward, 5′- GTCCTTATGACAAGAACATTAGCC -3′ and reverse, 5′- AATCAATGAAGAATGTGGTGAAG -3′; ChREBP forward, 5′- GTGTCTCCCAAGTGGAAGAATTT -3′ and reverse, 5′- GCTCTTCCTCCGCTTCACAT -3′; 18 s rRNA forward, 5′- GTAACCCGTTGAACCCCATT-3′′ and reverse, 5′- CCATCCAATCGGTAGTAGCG-3′; The relative gene expression levels were calculated by the 2-ΔΔCt method. 18 s rRNA were used as an internal control.

The protein lysates of cells or tissue specimens were prepared with RIPA buffer containing protease inhibitors (Thermo Fisher Scientific, USA) and supernatant was collected. Lysates were separated by 8.5% SDS-PAGE, and transferred to PVDF membranes (Millipore). Membranes were probed with the antibodies: anti-SCD1 (Proteintech Group, USA), anti-ChREBP (Santa Cruze, USA), anti-PTEN (Proteintech Group, USA), anti-Ecadherin (Cell Signaling Technology, USA), anti-Vimentin (Cell Signaling Technology, USA), anti-Akt (Cell Signaling Technology, USA); anti-phospho-Akt (Cell Signaling Technology, USA),anti-β-Actin (Cell Signaling Technology, USA), and anti-α-Tubulin(Santa Cruz Bio-technology, USA), followed by incubation of secondary peroxidase-labeled antibody (Cell Signaling Technology, USA). Protein signals were revealed with ECL detection system.

Harvested tumors were fixed in 4% paraformaldehyde, dehydrated, and embedded in paraffin. After deparaffinization and rehydration, the slides were retrieved for 30 min at 95 °C. 10% normal goat serum were used as blocking solution for 1 h. Then sections were subjected to antibody incubation overnight at 4 °C, followed by incubation with secondary antibody incubation. Then chromogenic reaction was performed with 3, 3-diaminobenzidine (Sigma-Aldrich, USA). After counterstaining with hematoxylin, slides were evaluated under a light microscope.

Short hairpin RNA (shRNA) against SCD1 (shSCD1) and nonspecific control shRNA were synthesized (Genepharma, China). The sequences of the two designed SCD1 shRNAs were as follows: shSCD1–1, CAGGACGATATCTCTAGCTCC; shSCD1–2, CCTACCTGCAAGTTCTACACC. The cDNA encoding human SCD1 was obtained from Dr. Jiahuai Han (Xiamen, China) and subcloned into vector pLVX-IRES-ZsGreen1 (Clontech). Cells were transfected with SCD1 shRNA and cDNA using polybrene as previous study [36].

Control and PTEN siRNA were synthesized by GenePharma Co., Ltd. (Shanghai, China). The siRNA sequences were as follows, si1 forward, 5′- GGCUAAGUGAAGAUGACAATT -3′ and reverse, 5′- UUGUCAUCUUCACUUAGCCTT -3′ and si2 forward, 5′- GAAGGCGUAUACAGGAACATT -3′ and reverse, 5′- UGUUCCUGUAUACGCCUUCTT -3′. The siRNA transfection was carried out using Lipofectamine RNAiMAX (ThermoFisher, USA), according to manufacturer’s protocol. The overexpression human PTEN was constructed by pEX-3(pGCMV/MCS/Neo) vector.

24-Well transwell chambers (8 μm pore size; Becton Dickison, Falcon, USA) were used to measure the cell migration and invasion capacity. Approximately 5 × 10⁵ CRC cells in serum-free DMEM were added into the upper chambers with (invasion assays) or without (migration assays) matrigel coating, while complete DMEM was placed in the bottom. After 24 h, cells migrating or invading the bottom layers were fixed with 4% paraformaldehyde, following by staining with 0.1% crystal violet. The nonmigratory cells on the upper side of the membrane were scraped off with a cotton tip. Migrated and invasive cells were photographed under an inverted microscope and counted using the ImageJ software.

BALB/c nude mice and C57BL/6 mice were originally provided by Shanghai Slack in Laboratory Animal Ltd. The animals were maintained under a specific pathogen free (SPF) condition with 12/12 h light/dark cycles. All animal experiments were approved by the Shanghai Jiao Tong University School of Medicine Institutional Animal Care and Use Committee (IACUC).

Male BALB/c nude mice (6 weeks old) were used for tumor metastasis study. Briefly, 2 × 10⁶ HCT116 cells stably transfected with shNC or shSCD1 were suspended in 100 μl PBS and injected through tail vein. Mice were sacrificed 6 weeks later, and lungs were harvested for histological examination of metastasis.

Male mice at 8 weeks of age were used for insulin resistance/T2DM model. Mice were maintained in 12/12 h light/dark cycles and fed either chow diet or high-fat diet (D12492, Research Diets). After 3 months, IPGTT and ITT were performed and mice were sacrificed.

After summarizing the linked fatty acid compositions in lipidomic data, we calculated the p value and log2 (fold change) and made the volcano plot by R-Studio, taking log2 (fold change) as X axis and –log10 (P value) as Y axis.

All experiments were performed in triplicate. All data were present as mean ± standard deviation. All graphing and statistical analyses were performed using GraphPad Prism 6 software (GraphPad Software, La Jolla, CA, USA) and SPSS 19 (IBM SPSS, IBM, Armonk, NY, USA). Correlations between the level of SCD1 in CRC tissues and clinic-pathological parameters were analyzed by Fisher’s exact tests. Comparison of survival between groups was performed using the log-rank test and Kaplan-Meier curves were plotted. The other data statistics were performed with student’s t-test in two groups. P value < 0.05(*), P value < 0.01(**) and P value < 0.001(***) were set as statistical significance.

To determine whether SCD1 might play a role in CRC progression, we examined expression of SCD1 in cancer and adjacent normal samples of pre-treatment patients. The relative expression of SCD1 in CRC tissues was higher when compared with adjacent non-tumor tissues (Fig. 1a-c). SCD1 mRNA expression was significantly higher in 86.36% of CRC tissue samples (19/22), compared with adjacent non-tumor tissue (P = 0.002) (Fig. 1a). Additionally, western blotting analysis and immunohistochemical staining showed significantly increased SCD1 expression in CRC tissues, accompanied with higher expression of Ki67, a proliferation marker (Fig. 1b-c). We found that E-cadherin and vimentin, two proteins frequently altered during epithelial–mesenchymal transition (EMT) [37], was decreased and increased respectively in human colorectal cancer tissues, compared with the adjacent normal tissues (Fig. 1c).

Furthermore, the correlation between SCD1 expression and clinicopathological features of CRC patients was investigated. There was a positive correlation between SCD1 levels and lymph metastasis (P = 0.056) or tumor-node-metastasis (TNM) stages (P = 0.023, Table 1). To assess whether SCD1 might be involved in prognosis of CRC patients, relevance analysis of SCD1 expression with clinical pathological parameters from Cancer Browser was performed (https://​genome-cancer.​ucsc.​edu/). Compared with patients with low SCD1 expression, CRC patients with high SCD1 levels have a poor overall survival (OS) (P = 0.043) (Fig. 1d). Furthermore, similar to the above clinicopathological features, high levels of SCD1 mRNA were associated with advanced TNM stages of CRC patients (P = 0.008) (Fig. 1e). These data indicated that SCD1 might play an important role in CRC progression.

Since SCD1 had a positive correlation with CRC progression, we hypothesized that increasing SCD1 expression might accelerate CRC migration and invasion. To choose suitable CRC cell lines for further experiments, we examined the expression profile of SCD1 in five different human CRC cell lines including SW620, HCT116, Caco2, SW116 and HT29 by western blotting. The expression levels of SCD1 were higher in HCT116 and minimal in CaCo2. Moreover, the protein showed intermediate expression levels in SW620, SW116 and HT29 (Fig. 2a). In order to investigate the effect of suppressing SCD1 on metastatic ability of CRC cells, we designed two different SCD1 shRNAs (shSCD1) to stably knockdown SCD1 expression in HCT116 and SW116 cells (Fig. 2b, Additional file 1: Figure S1A). We found that SCD1 knockdown impaired the migration and invasion ability of CRC cells (Fig. 2c, Additional file 1: Figure S1B). Quantification results revealed a significant decrease in the number of migrated and invaded cells after SCD1 knockdown (Fig. 2d-e, Additional file 1: Figure. S1C-D). We also ectopically expressed SCD1 in Caco2 cells (SCD1) to study how increased SCD1 expression affected metastasis (Fig. 2b). Interestingly, SCD1 overexpression significantly increased migration and invasion rates of Caco2 cells (Fig. 2f-g). All of the results suggest that SCD1 promote migration and invasion of colorectal cancer cells.

To investigate the effect of SCD1 suppression on the metastatic potential of CRC cells in vivo, we employed a mouse tail vein lung metastasis model as described [38]. We found that the incidence of lung metastasis decreased in mice injected with HCT116-shSCD1 cells compared to those injected with HCT116-shNC cells. Histological analysis confirmed that suppression of SCD1 decreased the size and number of lung metastatic tumors (Fig. 2h).

Since SCD1 and EMT markers were simultaneously highly expressed in CRC specimens, we next investigated whether SCD1 could promote metastasis of CRC cells by modulating EMT. Inhibition of SCD1 resulted in significantly elevated expression of E-cadherin and decreased expression of vimentin in HCT116 cells (Fig. 2i). On the other hand, overexpression of SCD1 led to significantly decreased expression of E-cadherin and increased expression of vimentin (Fig. 2j). These data indicate that SCD1 can promote metastasis of CRC by regulating EMT.

As reported, SCD1 plays an important role in fatty acid desaturation, which influences membrane phospholipids unsaturation and signal transduction, contributing to the progression of many cancers [12, 14] . To assess the effect of SCD1 on lipid remodeling and fatty acid ratio, quantitative lipidomic analysis was conducted using extracts of shSCD1 (sh1) and shNC HCT116 cells. The effect of SCD1 knockdown were broad, changing many lipid species measured (Fig. 3a). Levels of some lipids, especially those consist of one or more MUFAs, were significantly reduced following SCD1 knockdown (Table 2). Significant increase in SFA levels and reduction in MUFA levels in phosphatidylethanolamine (PE), monoglyceride (MG) and triglyceride (TG), were observed in shSCD1 HCT116 cells (Fig. 3b-d). Overall, these results are consistent with a role of SCD1 in influencing the composition of lipids by changing MUFA/SFA ratio in CRC cells.

Given the reduction of MUFA in shSCD1 CRC cells, we hypothesized that SCD1 could promote CRC progression through the effect of MUFA. Next, we examined the migration and invasion ability of CRC cells following treatment of oleic acid (OA, C18:1 n-9), the most common MUFA. To choose the best OA response condition, we treated Caco2 cells with different concentrations of OA (0, 0.05, 0.1 mM). After OA treatment, the migration and invasion rates of Caco2 cells were elevated, especially in response to 0.1 mM OA, which was similar to the effect of SCD1 overexpression (Fig. 4a-c). Interestingly, the migration and invasion ability reduced by SCD1 knockdown was reversed after OA treatment in HCT116 and SW116 cells (Fig. 4d-g). These results, together with the lipidomic profiling data, indicate that SCD1 accelerate colorectal cancer cell migration and invasion by increasing MUFA.

SCD1 gene expression is altered by nutrients such as glucose, insulin and lipids [39, 40]. To confirm whether T2DM has a positive connection with colorectal SCD1 expression, we conducted studies in C57BL/6 mice fed chow or high-fat-diet (HFD). Intraperitoneal glucose tolerance test and insulin tolerance test showed HFD group mice had hyperglycemia and insulin resistance (Fig. 5a-b). The protein level of SCD1 in colon of HFD-fed mice is higher than that of the chow-fed mice (Fig. 5c). Importantly, the transfactor, carbohydrate response-element binding protein (ChREBP), which can regulate SCD1 expression, showed increased expression level. The results suggest that T2DM promote expression of SCD1 in colon.

Glucose increases SCD1 transcription partly by activating ChREBP, a glucose-responsive transcription factor [11]. To test whether the original source of diabetic-enhanced SCD1 expression is hyperglycemia, we treated HCT116 cells with 0 mM (G0), 5.5 mM (G5.5), 11 mM (G11) and 25 mM glucose (G25). RT-PCR and western blotting results showed glucose treatment increased the expression of ChREBP and SCD1 in a concentration-dependent manner (Fig. 5d-e).

To determine whether hyperglycemia regulates metastasis of CRC, colorectal cancer cells were grown in the culture medium containing different amount of glucose. Compared with G0 and G5.5 groups, the migration and invasion ability of HCT116 and SW116 cells markedly increased in G11 and G25 groups (Fig. 5f-h, Additional file 2: Figure S2). The results suggest that high glucose accelerates colorectal cancer cells’ progression.

We next investigated whether SCD1 played an important role in glucose-induced CRC migration and invasion. We found that SCD1 knockdown impaired the migration and invasion ability of HCT116 and SW116 cells cultured under different glucose concentrations (0–25 mM) for 24 h (Fig. 6a-c, Additional file 3: Figure S3). Interestingly, compared to the shNC group, colorectal cancer cells transfected with shSCD1 showed more modest increase in migration and invasion in response to glucose (Fig. 6a-c, Additional file 3: Figure S3). Moreover, the enhanced migration and invasion rates of Caco2 cells ectopically expressing SCD1 showed just mild increase after glucose treatment (Fig. 6d-f). All of the results suggest that SCD1 contributes to glucose-induced CRC migration and invasion.

Several signaling pathways such as PTEN, Wnt/β-Catenin, STAT3, mTOR/S6K and JNK have been reported to play important roles in CRC metastasis [41‐47]. We found that the activity of β-Catenin, STAT3, S6K and JNK was unchanged after SCD1 knockdown or overexpression in CRC cells (Additional file 4: Figure S4A). Notably, SCD1 knockdown increased the expression of PTEN and OA treatment could reverse the change in HCT116 cells (Fig. 7a). Conversely, SCD1 overexpression or treatment of OA, the common SCD1 product, significantly decreased the level of PTEN in Caco2 cells (Fig. 7b). Our findings agreed with previous findings of unsaturated fatty acids downregulating PTEN in liver cancer [41, 42]. These data suggest that SCD1-MUFA might act as a negative regulator for PTEN and induce progression of CRC.

We next examined whether upregulation of PTEN was responsible for changes in CRC cell migration and invasion after SCD1 knockdown (Fig. 7c, Additional file 4: Figure S4B). Compared with the G5.5 group, the G25 group showed increased migration and invasion rate (Fig. 7d-f, Additional file 4: Figure S4C-E). Importantly, PTEN knockdown (siPTEN) reduced shSCD1-induced downregulation in migration and invasion of CRC cells, especially in the high glucose condition (Fig. 7d-f, Additional file 4: Figure S4C-E). Conversely, ectopic PTEN expression reduced SCD1-induced and glucose-induced migration and invasion of CRC cells (Fig. 7c, g-i). In HCT116 cells transfected with shSCD1, PTEN was upregulated and phosphate-Akt was reduced. On the other hand, the change of E-cadherin and vimentin expression level induced by high glucose was weakened after SCD1 knockdown (Additional file 4: Figure S4F).

Taken together, our results indicate that SCD1, via MUFA, suppress the expression of PTEN and activity of Akt, promoting the migration and invasion of colorectal cancer cells.