Cottonseed extracts regulate gene expression in human colon cancer cells

Cotton plant provides economically important fiber and cottonseed, but cottonseed contributes 20% of the crop value. Cottonseed value could be increased by providing high value bioactive compounds and polyphenolic extracts aimed at improving nutrition and preventing diseases because plant polyphenol extracts have been used as medicinal remedy for various diseases. The objective of this study was to investigate the effects of cottonseed extracts on cell viability and gene expression in human colon cancer cells. COLO 225 cells were treated with ethanol extracts from glanded and glandless cottonseed followed by MTT and qPCR assays. Cottonseed extracts showed minor effects on cell viability. qPCR assay analyzed 55 mRNAs involved in several pathways including DGAT, GLUT, TTP, IL, gossypol-regulated and TTP-mediated pathways. Using BCL2 mRNA as the internal reference, qPCR analysis showed minor effects of ethanol extracts from glanded seed coat and kernel and glandless seed coat on mRNA levels in the cells. However, glandless seed kernel extract significantly reduced mRNA levels of many genes involved in glucose transport, lipid biosynthesis and inflammation. The inhibitory effects of glandless kernel extract on gene expression may provide a useful opportunity for improving nutrition and healthcare associated with colon cancer. This in turn may provide the potential of increasing cottonseed value by using ethanol extract as a nutrition/health intervention agent.


Results
Effect of cottonseed ethanol extracts on colon cancer cell viability. Before cottonseed extracts on gene expression were analyzed, we evaluated the effect of the ethanol extracts on colon cancer cell growth. Human colon cancer cells (COLO 225) were treated with 10-100 µg/mL of cottonseed extracts for 2 and 24 h. MTT assay was used to estimate the effect of cottonseed extracts on cell viability. MTT assay did not show significant changes in the viability of colon cancer cells under treatments with various concentrations for 2 or 24 h Cottonseed coat or kernel was ground into fine powder and homogenized. The kernel fraction was defatted with chloroform and hexane. The coat fraction was treated with acetic acid followed by autoclave and centrifugation. The defatted materials were extracted with ethanol followed by evaporation to remove acetic acid and ethanol. Ethanol extracts were reconstitution in 100% DMSO (100 mg/mL) and analyzed by HPLC-MS 24 . Table 1. Human mRNA targets analyzed by qPCR; whose levels are regulated by cinnamon extract, gossypol or TTP as indicated in the "Reference" column. www.nature.com/scientificreports/ ( Fig. 2). Similar analysis did not show major effect of these cottonseed extracts on the viability of human lung cancer cells (A549 CCL185) (data not shown). Colon cancer cells were selected for gene expression analysis as described below.

Basal gene expression level in human colon cancer cells
One important factor for relative gene expression evaluation is to get a basic idea about the basal level expression of the genes selected for investigation. The relative mRNA levels of 55 genes (Table 1) were measured in the control cells using the specific qPCR primer pairs as described 53 . SYBR Green qPCR assay showed that BCL2 mRNA C T (cycle of threshold) was one of the least varied mRNAs (Table 2). BCL2 mRNA C T value was 30 ± 1 (mean ± standard deviation, n = 12) ( Table 2). GAPDH and RPL32 mRNA levels were 33 and 51 fold of BCL2 mRNA, respectively. INOS mRNA was undetectable. AHRR1, COX1, CYCLIND1, GLUT4, HUA, ICAM1, IL10, IL12, RAB24, VEGF and ZFP36L2 mRNAs were detected with less than 10% of BCL2 mRNA in the colon cancer cells ( Table 2). The mRNA level of a gene at least twofold or less than 50% of BCL2 mRNA could be interpreted as its expression more or less abundant than that of BCL2 mRNA, respectively. By this standard, 14 genes were expressed more abundantly than BCL2 gene (BCL2L1, BNIP3, CSNK2A1, CTSB, GAPDH, GLUT1, GLUT3, HIF1A,  HMGR, IL6, MAP1LC3B, RPL32, TNFSF10, and ZFAND5) (Table 2). Similarly, 20 genes were expressed less abundantly than BCL2 gene (AHRR1, COX1, CXCL1, CYCLIND1, DGAT2A, DGAT2B, FAS, GLUT4, HUA,  ICAM1, IL2, IL10, IL12, LEPTIN, NFKB, P53, RAB24, TNF, VEGF, and ZFP36L2) ( Table 2). TaqMan qPCR assay showed similar trend of SYBR Green qPCR (data not shown). SYBR Green qPCR assay was chosen to conduct gene expression analysis in the following experiments.

Overall effect of cottonseed ethanol extracts on gene expression in human colon cancer cells.
After we analyzed the basal levels of gene expression and identified the reference gene for qPCR analysis as described previously, we evaluated how these genes might be affected by ethanol extracts by using the pooled qPCR data from 24 samples using BCL2 mRNA as the internal reference and DMSO treatment as the sample control. As shown in Table 3, expression of a number of genes was affected by cottonseed ethanol extracts. There were 3 genes with mRNA levels at least twofold of the DMSO control in the cells treated with glanded coat extract (CYCLIND1, CYP19A1, and LEPTIN) ( Table 3). There were 2 genes with mRNA levels at least twofold of the DMSO control in the cells treated with glanded kernel extract (CYCLIND1 and CYP19A1) ( Table 3). There were 2 genes with mRNA levels at least twofold of the DMSO control in the cells treated with glandless coat extract (CYCLIND1 and CYP19A1) ( Table 3). There were 4 genes with mRNA levels at least twofold of the DMSO control in the cells treated with glandless kernel extract (COX2, CYCLIND1, CYP19A1, and LEPTIN) ( Table 3).
The above results suggest that cottonseed ethanol extracts affected the expression of many genes in the human colon cancer cells. Therefore, we analyzed the mRNA levels of 55 genes in the human colon cancer cells treated with various concentrations of the four cottonseed extracts as described below.
Effect of glanded coat extract on gene expression. Firstly, we analyzed the effect of glanded coat extract on gene expression. Human colon cancer cells were treated with glanded cottonseed coat extract (0, 5, 10, 20, 30, 40, 50 and 100 µg/ml). SYBR Green qPCR analyzed the expression of all 55 genes with BCL2 mRNA as the internal reference and 1% DMSO treatment as the sample control. The expression of some genes was significantly affected by glanded coat extract (Fig. 3). It appeared that the expression of COX2, GLUT1, LEPTIN, TNF, and TNFSF10 was increased by the glanded coat extract (Fig. 3) www.nature.com/scientificreports/ www.nature.com/scientificreports/ ZFP36 (Fig. 3). The expression of the rest of the 55 genes not mentioned above at mRNA levels was not affected by various concentrations of the glanded kernel extract (data not shown).
Effect of glanded kernel extract on gene expression. Secondly, we analyzed the effect of glanded kernel extract on gene expression. Similarly, human colon cancer cells were treated with glanded cottonseed kernel extract. Gene expression was analyzed by qPCR with BCL2 mRNA as the internal reference and 1% DMSO treatment as the sample control. The expression of ELK1, FAS, and GAPDH genes was increased by the glanded kernel extract (Fig. 4). The expression of other genes at mRNA levels was not affected by various concentrations of the ethanol extract.
Effect of glandless coat extract on gene expression. Thirdly, we analyzed the effect of glandless coat extract on gene expression. Human colon cancer cells were also treated with various concentrations of glandless cottonseed coat extract and analyzed gene expression at the mRNA levels by qPCR using BCL2 mRNA as the internal reference and 1% DMSO treatment as the sample control. The expression of FAS, GAPDH, GLUT1, and ZFP36 was increased by the glandless coat extract (Fig. 5), but only CXC1 expression was reduced by the coat extract (Fig. 5). The expression of the rest of the 55 genes not mentioned above at mRNA levels was not affected by various concentrations of the ethanol extract (data not shown).

Effect of glandless kernel extract on gene expression. Finally, we analyzed the effect of glandless
kernel extract on gene expression. Similarly, glandless cottonseed kernel extract treated human colon cancer cells and SYBR Green qPCR analyzed mRNA levels of 55 genes with BCL2 mRNA as the internal reference and 1% DMSO treatment as the sample control. qPCR data indicated that expression of much more genes was affected by the glandless kernel extract. The effect of the glandless kernel extract on gene expression was analyzed in detail according to gene families as described below (Figs. 6, 7, 8).
Glandless kernel extract on reference gene expression. The expression of GAPDH and RPL32 genes, the two well-known reference genes in the literature, was analyzed in the colon cancer cells after treatment with various concentration of glandless kernel extract. The qPCR data showed that glandless kernel extract treatment resulted in a large reduction of both GAPDH and RPL32 mRNA levels in the cells (Fig. 6A).
Glandless kernel extract on GLUT gene expression. Glucose transporter (GLUT) family proteins are responsible for glucose uptake in mammalian cells. Four forms of GLUTs are present in mammalian cells 23 . The glandless kernel extract treatment only decreased GLUT1 mRNA level without much effect on the other GLUT isoforms (Fig. 6D). GLUT4 mRNA level was very low so that it was difficult to be measured with sufficient confidence ( Table 2).
Glandless kernel extract on TTP gene expression. Tristetraprolin (TTP/ZFP36) family proteins regulate mRNA stability 59 . TTP family genes have anti-inflammatory properties with therapeutic potential for inflammationrelated diseases 60,61 . TTP family proteins consist of three members in mammals (ZFP36 or TTP, ZFP36L1 and ZFP36L2) and the fourth member in mouse and rat but not in humans (ZFP36L3) 59,62 . SYBR Green qPCR showed that ZFP36 and ZFP36L1 mRNAs were reduced by the glandless kernel extract (Fig. 7A). ZFP36L2 mRNA levels were too low to be assessed reliably (  68 . SYBR Green qPCR showed that glandless kernel extract increased IL12 mRNA level but decreased IL16 mRNA level (Fig. 7B). IL8 and IL10 mRNA levels were difficult to compare due to their low levels in the colon cancer cells (Table 2).
Glandless kernel extract on other gene expression. A few other gene targets were selected for the analysis of gene expression. The qPCR assays showed that glandless kernel extract decreased the expression of HMGR, INSR, MAPL1C3A, MAPL1C3B, and NFKB mRNA levels (Fig. 8). The effect of glandless kernel extract on ULK2 mRNA level was not much and the effect on CYCLIND1 mRNA level was difficult to assess due to large variation of the results (Fig. 8).

Discussion
Cottonseed accounts for approximately 20% of the crop value. One way to increase cottonseed value is to isolate bioactive materials aimed at improving nutrition and preventing diseases. In this study, we observed that the expression of the majority of genes was significantly reduced by glandless cottonseed kernel extract, although their expression was less affected by three other cottonseed ethanol extracts (glanded cottonseed coat and kernel as well as glandless cottonseed coat extracts). Cottonseed extracts exhibited only minor effect on the viability of human colon cancer cells under the experimental conditions. Our previous study showed that gossypol strongly inhibited human cancer cell viability 24 . The current data confirm our HPLC-MS analyses that the cottonseed extracts are essentially free of the toxic compound gossypol 24 .
Before we examined the effect of cottonseed extracts on gene expression in human colon cancer cells, we evaluated the relative expression levels of 55 genes and selected the internal reference for qPCR analysis since it is important for normalization of gene expression levels [77][78][79][80] . Our study confirmed that BCL2 mRNA was the most stable among the 55 mRNAs analyzed in human colon cancer cells treated with DMSO vehicle or various concentrations of ethanol extracts (Table 2) 53 . We also confirmed that GAPDH and RPL32 mRNAs were not good qPCR assay references for the colon cancer cells since they were most abundant mRNAs with large variations under the cell culture conditions 53 .
The most important observation of this study was that glandless kernel extract decreased the mRNA levels of the great majority of the 55 genes tested, including GAPDH involved in the sixth step of breakdown of glucose in glycolysis 80 and RPL32, a component of the large 60S subunit of ribosomes involved in protein synthesis 77 (Fig. 6A), the genes known to be involved in cancer development, such as BNIP3 involved in the permeability of outer mitochondrial membrane 35 , CYP19A1 localized to the endoplasmic reticulum and catalyzed the last steps of estrogen biosynthesis 30 , FAS, a member of TNF-receptor superfamily playing a key role in programmed cell death 29 , P53 involved in preventing genome mutation 28 , PPARR, a nuclear receptor involved in gene expression regulation 31 and TNFSF10, a TNF super family member functioning as a ligand that induces apoptosis 34 (Fig. 6B), the DGAT family members DGAT1, DGAT2a and 2b responsible for the last and rate-limiting step of triacylglycerol biosynthesis 54,58 (Fig. 6C), and GLUT1 responsible for glucose transport across the plasma membranes 23 (Fig. 6D). In addition, glandless kernel extract reduced ZFP36 mRNA levels in the TTP family which bind to the AU-rich elements of some mRNAs and cause destabilization 60,69 (Fig. 7A). It increased IL12, a T-cell stimulating fsctor 67 but decreased IL16 functions as a chemoattractant, a modulator of T cell activation, and an inhibitor of HIV replication 42 mRNAs levels in the IL family members (Fig. 7B), decreased LEPTIN involved in energy balance 81 and TNF, a cytokine promoting inflammation 64 mRNA levels (Fig. 7C), and appeared to decrease all of the TTP-targeted mRNAs including AHRR1 39 (Fig. 7D). Finally, glandless kernel extract appeared to decrease the expression of HMGR 88 , INSR 21 , MAPL1C3A 89 , MAPL1C3B 89 , and NFKB 90 mRNA levels (Fig. 8).
This study provides valuable information about the effects of cottonseed ethanol extracts on gene expression at the mRNA levels in the human colon cancer cells. Much more investigations need to be conducted in the future. First, it could be a greater addition by confirming the mRNA results with results at the protein levels. Second, the consequence of gene regulation on cellular metabolic levels could be valuable for understanding the molecular mechanism. Finally, additional studies with other cell lines and animals could be required for the potential utilization of cottonseed extracts as viable sources for improving nutrition and preventing diseases.

Conclusions
This study showed that most of the gene expression in human colon cancer cells was not affected by ethanol extracts isolated from glanded cottonseed coat and kernel as well as glandless cottonseed coat, but the expression of the majority of genes was significantly reduced by glandless cottonseed kernel extracts. The inhibitory effects of glandless kernel extract on gene expression in the colon cancer cells may provide a useful opportunity for improving the healthcare associated with colon cancer since it is safe without toxic gossypol contamination and effective in decreasing the expression of so many genes related to cancer development. This in turn may provide the potential of increasing the value of cottonseed by using cottonseed-derived ethanol extracts as a health intervention agent.

Materials and methods
Cottonseed. The  Cottonseed extracts. Seed kernel extracts were isolated by fractionation, defatting, and ethanol extraction, and seed coat extracts were isolated by fractionation, defatting, acetic acid extraction, and ethanol extraction 24 (Fig. 1). Briefly, cottonseed coat or kernel was ground into fine powder and homogenized. The kernel fraction was defatted with chloroform and hexane. The coat fraction was treated with acetic acid followed by autoclave and centrifugation. The defatted materials were extracted with ethanol followed by evaporation to remove acetic acid and ethanol. Ethanol extracts were reconstitution in 100% DMSO (100 mg/mL) and analyzed by HPLC-MS. The ethanol extracts contained trace amount of gossypol (0.82 ng gossypol/mg extract in glanded seed coat, 0.03 ng gossypol/mg extract in glanded seed kernel, 0.37 ng gossypol/mg extract in glandless seed coat and 0 ng gossypol/mg extract in glandless seed kernel) 24 .
Cell culture and chemical treatment. Cell  Cell viability assay. MTT based-In Vitro Toxicology Assay Kit was used to determine cell cytotoxicity 24 .
Cancer cells in 96-well plates (12 wells/treatment) were treated with ethanol extracts and incubated at 37 °C, 5% CO 2 for 2 and 24 h. The cell media were added with 50 µL of MTT assay reagent (thiazolyl blue tetrazolium bromide) and incubated at 37 °C, 5% CO 2 for 2 h before adding 500 µL MTT solubilization solution to each well, shaken at room temperature overnight. The color density in the wells was recorded by Epoch microplate spectrophotometer at A570.

Real-time qPCR primers and probes.
Fifty-five genes were selected for qPCR analysis of their expression in the colon cancer cells as described previously 53 . These genes were shown to be regulated by cottonseedderived gossypol in cancer cells and macrophages or regulated by ZFP36/TTP in tumor cells and macrophages ( RNA isolation and cDNA synthesis. RNA isolation and cDNA synthesis were essentially as described 25 . Human colon cancer cells were treated with various concentrations of cottonseed ethanol extracts for 8 h (triplicate). The cells were lysed directly in the washed dishes with 1 mL of TRI ZOL reagent. RNA was isolated according to the manufacturer's instructions without DNase treatment. RNA concentrations were quantified with an Implen NanoPhotometer (Munchen, Germany). The cDNAs were synthesized from total RNA using Super-Script II reverse transcriptase. The cDNA synthesis mixture contained 5 μg total RNA, 2.4 μg oligo(dT) 12 www.nature.com/scientificreports/ primer, 0.1 μg random primers, 500 μM dNTPs, 10 mM DTT, 40 u RNaseOUT and 200 u SuperScript II reverse transcriptase in 1X first-strand synthesis buffer (20 μL). The cDNA synthesis reaction was performed at 42 °C for 50 min. The cDNA was stored in − 80 °C freezer and diluted with water to 1 ng/µL before qPCR analyses.
Quantitative real-time PCR analysis. The qPCR assays were described 56,78,79,91 and performed according to the MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments 92 . The qPCR assay mixture contained 5 ng of RNA-derived cDNA, 200 nM of forward and reverse primers, and 1 × iQ SYBR Green Supermix. Thermal cycle conditions were 3 min at 95 °C, 40 cycles at 95 °C for 10 s, 65 °C for 30 s and 72 °C for 30 s. BCL2 mRNA was used as the internal reference because it had the minimal variation of gene expression among the 55 genes tested (see "Results" for details). Ribosome protein 32 (RPL32) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNAs were not suitable for qPCR analysis for this cell type due to variations (see "Results" for details) although they were widely used as the reference mRNAs in qPCR analyses 19 . TaqMan qPCR assay confirmed some of the SBYR Green qPCR assays using the same conditions as described 78 .
Data analysis and statistics. The relative expression in fold was determined with 2 −ΔCT or 2 −ΔΔCT equations 93 . The first step was to normalize the threshold cycle (C T ) values of the target mRNAs to the C T values of the internal control BCL2 mRNA (ΔC T = C TTarget − C TBcl2 ). The second step was to normalize treatment ΔC T values with DMSO control ΔC T values (ΔΔC T = ΔC TCottonseed − ΔC TDMSO ). Finally, the fold change in expression was calculated. The data in the figures and tables represent the mean and standard deviation of three and 24 independent samples, respectively. These data were subjected to statistical analysis using ANOVA with SigmaStat 3.1 software (Systat Software). Student-Newman-Keuls method and Tukey test were used to perform multiple comparisons among the treatments with different concentrations of cottonseed extracts 19 . www.nature.com/scientificreports/