* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download STK31 maintains the undifferentiated state of colon cancer cells
Survey
Document related concepts
Transcript
Carcinogenesis vol.33 no.11 pp.2044–2053, 2012 doi:10.1093/carcin/bgs246 Advance Access publication July 25, 2012 STK31 maintains the undifferentiated state of colon cancer cells Kin Lam Fok1,†, Chin Man Chung1,†, Shao Qiong Yi2, Xiaohua Jiang1, Xiao Sun1, Hao Chen1,3, Yang Chao Chen4, Hsiang-Fu Kung5, Qian Tao6, Ruiying Diao1,3, Henry Chan7, Xiao Hu Zhang1, Yiu Wa Chung1, Zhiming Cai3 and Hsiao Chang Chan1,* 1 Epithelial Cell Biology Research Centre, The Chinese University of Hong Kong, Hong Kong SAR, 2Institute of Epidemiology, Academy of Military Medical Sciences, Beijing, P. R. China, 3Shenzhen Second People’s Hospital, The First Affiliated Hospital of Shenzhen University, Shenzhen, P. R. China, 4 School of Biomedical Sciences, 5Stanley Ho Centre for Emerging Infectious Diseases, School of Public Health and Primary Care, 6Department of Clinical Oncology, The Chinese University of Hong Kong, Hong Kong SAR and 7Division of Biological Sciences, University of Chicago, IL, USA *To whom correspondence should be addressed. Tel: +852 26096001; Fax: +852 26037155; Email:[email protected] The expression of serine/threonine kinase (STK) family is frequently altered in human cancers. However, the functions of these kinases in cancer development remain elusive. Here, we report that STK31 is robustly and heterogeneously expressed in colon cancer tissues and plays a critical role in determining the differentiation state of colon cancer cells. Knockdown or overexpression of STK31 induced or inhibited differentiation of colon cancer cells, respectively. Deletion of the STK domain abolished the inhibiting effect of STK31. Associated with differentiation, knockdown of STK31 resulted in significant suppression of tumorigenicity both in vitro and in vivo. Genome microarray analysis showed that knockdown of STK31 altered the expression profile of genes that are known to be involved in germ cell and cancer differentiation. Taken together, these results suggest that STK31 is able to control the differentiation state of colon cancer cells, which critically depends on its STK domain. The present findings may shed light on the new therapeutic approach against cancer by targeting STK31 and cancer differentiation. Introduction Cells within a tumor often exhibit variations in biological properties, including the differentiation status, tumorigenecity and therapeutic resistance. These variables in tumors are referred to as tumor heterogeneity (1–3). Intratumor heterogeneity has long been observed in colon cancers (4,5). Recently, using a high throughput single cell PCR transcription analysis platform, the heterogeneity in colon cancer has been shown to be a corollary of multi-lineage differentiation (6), indicating the presence of cellular hierarchy in colon cancers. Differentiation of colon cancer cells may give rise to enterocytes in the adsorptive lineage or enteroendocrine cells and goblet cells in the secretory lineage (7). TGF-β and BMP pathways are two well-known cascades that regulate colon differentiation. While TGF-β shows a strong cytostatic response in intestinal culture in vitro (8), BMP inhibits intestinal stem cells renewal in vivo (9). These counteracting cascades share the common SMADs protein family as the intracellular messengers. Although the regulatory machinery in colon cancer differentiation has been postulated to recapitulate the signaling mechanisms in normal stem cell counterpart (10), the involvement of reactivated oncogenes in regulating colon-cancer differentiation through intrinsic mechanism remains elusive. Abbreviations: STK, serine/threonine kinase; RT-PCR, Reverse transcription polymerase chain reaction. † These authors contributed equally to this work. Tyrosine kinases and serine/threonine kinases (STKs) are two major classes of kinases in the human kinome. Although extensive studies have been focused on the functions and therapeutic potential of tyrosine kinases in cancers (11,12), a recent study revealed that the expression of STKs is frequently altered in human cancers, suggesting that STKs may play an important role during cancer development. STK31 is a highly conserved member of the STK family and the Tudor Royal family, containing a STK catalytic domain at the C terminus and a tudor domain at the N terminus. The expression of STK31 is testis specific and is restricted to spermatogonia, the stem cell population in the testis (13). Interestingly, in addition to its specific expression in the testis, a recent study reported that STK31 was reactivated in gastrointestinal cancer by promoter hypomethylation, suggesting its potential role in cancer development (14). Despite all these interesting findings, the biological function of STK31 in gastrointestinal cancers has not been investigated. We undertook the present study to investigate the functional role of STK31 and have demonstrated that STK31 plays a critical role in maintaining the undifferentiated state of colon cancer cells, which critically depends on its STK domain. Materials and methods Constructs Full-length STK31 was cloned into pEGFP-C2 vectors (Clontech, Mountain View, CA) and verified by sequencing. For domain deletion constructs, 101– 160 a.a. and 809–872 a.a. were deleted in Δtudor and Δkinase, respectively, by overlapping extension PCR method (15). For RNAi experiment, miR RNA is were designed using web-based software (Invitrogen, Carlsbad, CA). Design targeting bacterial LacZ was used as control. Annealed oligos were cloned into pcDNA6.2 GW/EmGFP-miR vector (Invitrogen, Carlsbad, CA) and verified by sequencing. Primer and oligo sequences are listed in Supplementary Table 1, available at Carcinogenesis Online. Gene expressions RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA). Gene expressions were studied by RT-PCR or real-time PCR using specific primers (Supplementary Table 1, available at Carcinogenesis Online). SYBR green master mix and 7500 Fast Real-Time PCR Systems were used for realtime-PCR (Applied Biosystems, Carlsbad, CA). Glyceraldehyde 3-phosphate dehydrogenase was used as internal control. For microarray analysis, RNA from Caco2 STK31 knockdown and control miR was labeled with Cy3 and Cy5, with dye swapped in duplicate experiment. Labeled cDNA were hybridized to Agilent human whole genome microarray (Agilent, Santa Clara, CA). Data were analyzed by GeneSpring 10.0 software (Agilent, Santa Clara, CA). Expressions of candidate genes in germ cells were obtained from GermSAGE database (16). Clinical samples In tissue array experiment, colon adenocarcinoma tissue array was purchased from US Biomax (US Biomax Inc., Rockville, MD). The tissue array contains 110 colon cancer patients and 10 normal tissues. The 110 patients included 22 well-to-moderately differentiated tumors (Grade 1 or 1–2); 59 moderately differentiated tumors (Grade 2); 22 poorly differentiated tumors (Grade 3) and 7 undefined tumors. The signal intensities were classified into three categories: (− or 0)––no expression; (+ or 1)––weak expression and (++ or 2)––strong expression. Since normal tissue sections have signal intensities ≤1, cancer sections with signal intensities of ≤1 were assigned to low-expression group, whereas those >1 were assigned to high-expression group. Heterogeneous expression was defined by the presence of at least 5% of cells with similar morphology showing differences in signal intensities. Frozen colon cancer samples were collected from Shenzhen Second People’s Hospital, the first affiliated hospital of Shenzhen University. Eight patients were recruited with signed informed consent. Immunostaining and cell imaging For immunofluorescent staining in cell lines, Caco2 and SW1116 cells were grown on coverslips for 48 h and fixed with 4% ice-cooled paraformaldehyde. In both immunofluorescent and immunhistochemical staining experiments, tissue array or cells were stained with antibody against STK31 (1:100, Abnova, © The Author 2012. Published by Oxford University Press. All rights reserved. For Permissions, please email: [email protected] 2044 STK31 regulates colon cancer differentiation Fig. 1. Expression of STK31 in colon cancers. (A) Immunohistochemical staining revealed that STK31 protein was highly expressed in colon cancer patient samples (representative figures are shown), whereas weak signals were observed in normal tissues. Perinuclear granular staining (dashed circle) was observed. 2045 K.L.Fok et al. Taipei, Taiwan). Slides stained omitting primary antibodies were used as negative controls. Cell morphologies were analyzed by MetaMorph software. Three random fields were taken from each stable transfectant. At least six cells per field were randomly picked and the cell membranes were manually traced. The number of pixels within the traced area was used for cell-size estimation. For tumor sections in xenograft experiments, tumor sections were stained with hematoxylin and eosin. All images were taken by Nikon ELIPSE Imaging System (Nikon, Melville, NY). Western blot Cells were lysed in RIPA buffer (50 mM Tris-Cl pH8.0; 150 mM NaCl; 1% NP-40; 0.5% deoxycholic acid and 0.1% SDS) with protease inhibitor cocktail. Proteins were resolved and blotted onto nitrocellulose membrane and probed with indicated antibodies. Antibodies used in this study were anti-STK31 (1:1000, Abnova, Taipei, Taiwan), anti-cleaved caspase-3, anti-cleaved PARP (1:500, both from Cell Signaling, Danvers, MA) and anti-β-tubulin, (1:2000, Santa Cruz Biotech, Santa Cruz, CA). Cultures and transfection Caco2 cells were maintained in DMEM supplemented with 10% FBS. SW1116 cells were maintained in RPMI 1640 supplemented with 10% FBS. All cells were transfected with Lipofectamine 2000 (Invitrogen, Carlsbad, CA) following manufacturer protocol. In miR RNAi experiment, stably transfected clones were selected by Blasticidin (Invitrogen, Carlsbad, CA). In cell-count experiment, 1 × 104 cells per well were seeded into a 24-well plate. Cells were trypsinized and counted by trypan blue exclusion at indicated time point. Flow cytometry Stable transfectants were seeded into 35-mm dishes at three densities: 1 × 105 per dish (sub-confluent), 6 × 105 (confluent) and 1 × 106 cells (post-confluent). Cells were grown for 2 days and collected for flow cytometry. Cells were fixed in chilled 70% ethanol for 30 min on ice. Fixed cell were stained 5 μg/ml propidium iodide (Sigma, St. Louis, MO) and 10 ug/ml RNase A for 2 h at room temperature. All samples were analyzed by Facscalibur flow cytometer (BD Biosciences, New Jersey, NJ). Data were analyzed by FCS express V3 software (Denovo software, Los Angeles, CA). Differentiation assay In spontaneous differentiation, Caco2 cells were cultured for 14 days with medium changed every 3 days. In sodium butyrate–induced differentiation, Caco2 and SW1116 cells were treated with 2 mM sodium butyrate (Sigma, St. Louis, MO). Cells were collected for molecular studies at 24–48 h post-treatment. For alkaline phosphatase activity assay, cells were lysed in lysis buffer (1x PBS; 5 mM EDTA; 1% Triton X-100 with PMSF and PImix). Protein lysate was loaded into pNPP substrate solution, incubated for 30 min and terminated by stopping solution. Optical densities were measured at 405 nm. Soft agar colony formation assay A total of 2–4 × 105 cells were mixed with top agar (0.3%) and overlaid onto bottom agar (0.7%). Colonies were allowed to form for at least 2 weeks followed by staining with 0.005% crystal violet. Number of colonies per plate was counted. Statistical data were obtained from three biological replicates. Xenograft assay Animals were provided by Laboratory Animal Service Centre, CUHK. The protocols were approved by animal ethics committee, LASEC, CUHK (09/005/MIS). For Caco2 cells, 6 × 106 cells were injected subcutaneously into 5–6-week-old female NOD/SCID mice. For SW1116 cells, 1 × 106 cells were injected subcutaneously into 7–8-week-old female athymic nude mice. The mice were then killed after 4–6 weeks. Tumor weight, tumor size and the number of mice that developed solid tumor were determined. Statistical analysis Statistical analyses were performed by Prizm 3.0. Results are presented as mean ± standard error mean (SEM) or standard deviation (SD). Data were compared by student’s t test, Chi Square (χ2) or one-way ANOVA analysis. A P value <0.05 was considered statistically significant. ← Results Reactivation of STK31 in multiple cancers Previous study has demonstrated the promoter hypomethylation and reactivation of STK31 in gastrointestinal cancer by PCR and immunohistochemistry (14). To examine the reactivation of STK31 in cancer development, we studied the expression profile of STK31 in a cohort of cancers. Our RT-PCR results showed that STK31 mRNA was reactivated in nasopharyngeal and liver cancer cell lines (Supplementary Figure S1, available at Carcinogenesis Online). The reactivation of STK31 in a board spectrum of cancers suggests that it may play an important role in cancer development. To consolidate the reactivation of full-length STK31 in colon cancer, we audited the protein expression of STK31 by both immunohistochemistry and western blot. We purchased an antibody raised against human full-length STK31 protein (Abnova) and validated the specificity of the antibody by western blot in STK31-transfected HEK293 cells. The antibody specifically recognizes full-length STK31 protein (116 kDa; Supplementary Figure 2, available at Carcinogenesis Online). Then, we examined the reactivation of STK31 in a cohort of colon cancer samples by immunohistochemistry using the validated antibody. Colon tumor tissue array analysis showed enhanced STK31 expression in the cytoplasm of 57% (63 of 110) human colon adenocarcinoma specimens (Figure 1A; Supplementary Table 2A, available at Carcinogenesis Online). Weak signals were detected in cancer-adjacent and normal colonic tissues. Reactivation of STK31 was not always homogeneous within a tumor, with 22 out of 63 STK31-reactivated patients’ samples showing varied signal intensities in at least 5% of the cancer cells in the samples. The expression of STK31 appeared to be less heterogeneous in moderately to poorly differentiated cancers (31% heterogeneous expression) when compared with well-differentiated cancers (50% heterogeneous expression; P = 0.28); however, the difference was not statistically significant (Supplementary Table 2B, available at Carcinogenesis Online). While the reactivation of STK31 was not correlated to gender and histological grade (P = 0.25 and 0.23, respectively), patients with STK31 reactivation were slightly younger (P = 0.16; Supplementary Table 2B, available at Carcinogenesis Online), which is consistent with the previous report (14). Interestingly, perinuclear granular signals were observed in some of the tumor samples (Figure 1A red circle), which is reminiscent of the localization of its mouse homolog—Stk31—in the RNA processing body of the germ cells (17). To further validate the reactivation of STK31 in a quantitative way, we examined the expression of STK31 by real time-PCR. The results showed that the expression of STK31 transcripts were significantly increased in 5 out of 8 patients (P < 0.05) compared with normal tissue (n = 8; Figure 1B), confirming the reactivation of STK31 in colon cancer. We next examined the expression and localization of endogenous STK31 in two colon cancer cell lines—Caco2 and SW1116—by immunofluorescent staining. Consistent with the IHC results, STK31 was localized in the cytoplasm (Figure 1C). Of note, expression of STK31 was heterogeneous in these two cancer cell lines. Particularly, in Caco2 colonies, high STK31 expression levels were found at the edge of colonies, which became decreased or undetectable in the central part of the colonies, where cells were compacted and were in contact with each other. Given that Caco2 cells are prone to undergo spontaneous post-confluence differentiation (18), the marked decrease of STK31 expression in the confluent cell population suggests that the expression of STK31 is downregulated in a differentiated state. To further demonstrate the correlation between the expression of STK31 and the differentiation of colon cancer cells, Caco2 cells were treated Magnification: ×200 (left panel) and corresponding enlarged images were shown (right panel). Scale bar: 50 µm. (B) Quantitative PCR analysis of STK31 in colon cancer patients. (C) Immunofluorescent staining of STK31 in Caco2 and SW1116 colon cancer cells. STK31 was heterogeneously expressed within a cell line. In Caco2 cells, expression of STK31 was mainly detected on the edge of a colony and decreased in the central population (arrow). In SW1116 cells, a small population of flattened cells (asterisk) demonstrated weaker STK31 expression. Scale bar: 50 µm. (D, E) Alkaline phosphatase activities (D) and protein expression of STK31 (E) in Caco2 cells treated with sodium butyrate. Data are presented as mean ± SEM; t test, *P < 0.05, **P < 0.01. 2046 STK31 regulates colon cancer differentiation Fig. 2. Knockdown of STK31 alters cell morphology and suppresses cell proliferation and cell cycle progression in both Caco2 and SW1116 cells. STK31 RNAi construct was stably transfected into Caco2 and SW1116 cells. (A) Western blot of STK31 showed that the RNAi design could knock down STK31 in both cell lines. 2047 K.L.Fok et al. with sodium butyrate, a metabolite that is known to induce differentiation of colon cancer cells. After sodium butyrate treatment, the activity of the brush border enzyme—alkaline phosphatase—was significantly increased (Figure 1D) while the protein expression of STK31 was markedly downregulated (Figure 1E). This result suggested a negative regulation of STK31 on differentiation of colon cancer cells. Knockdown of STK31 alters cell morphology and cell cycle progression Differentiation of colon adenocarcinoma cells is characterized by cell flattening (19) and growth retardation (20,21). To test whether STK31 is involved in regulating cell differentiation, we designed and constructed an RNAi for stable knockdown of STK31 in Caco2 and SW1116 cells (Figure 2A). Growth and morphology of the gene-manipulated cells were monitored over time. While knockdown of STK31 did not alter morphological or growth properties in subconfluent cultures (Supplementary Figure 3, available at Carcinogenesis Online), cells with knocked-down STK31 became significantly more flattened and spread out on culture dish upon cell confluence as indicated by the average area taken up by a cell (P < 0.01; Figure 2B). The flattened morphology could not be attributed to the increase in cell size as indicated by flow cytometric analysis (Supplementary Figure 4, available at Carcinogenesis Online). More importantly, knockdown of STK31 resulted in a significant decrease in cell proliferation in both cell lines upon confluence (P < 0.05; Figure 2C). Further analysis on cell cycle progression in Caco2 cells revealed that knockdown of STK31 resulted in G1 arrest in both confluence and post-confluence (P < 0.05; Figure 2D). These results indicate that knockdown of STK31 has the potential to enhance differentiation in both colon adenocarcinoma cell lines. Knockdown of STK31 enhances differentiation of colon cancer cells When grown in appropriate culture conditions or treated with inducers, colon cancer cells from primary tumors can differentiate into two lineage-restricted cell types after in vitro expansion—the enterocytes in the absorptive lineage and the goblet cells in the secretive lineage (7,22). Caco2 cells are capable to undergo enterocytic differentiation that is marked by the formation of dome and the expression of villin (an enterocyte marker), whereas SW1116 cells are capable to undergo goblet-like differentiation that is identified by expression of mucin (a goblet cell marker; 7,23,24). Since transition from undifferentiated to differentiated states in these two cell lines seems to recapitulate the growth arrest and commitment of lineage-specific differentiation of colonic cancer cells, we further investigated the potential roles of STK31 in regulating differentiation of the colon carcinoma cell phenotype in these two cell lines. In spontaneous differentiation of Caco2 cells (25), knockdown of STK31 increased both dome (P < 0.05) and pleomorphic glandular formations (P < 0.001; Figure 3A), which are indicative of mature phenotypes of colon cancer cells (22). In keeping with the morphological changes, expression of villin was ~2–3-fold higher in STK31 knockdown cells compared with that in control cells in both spontaneous and sodium butyrate–induced differentiations (P < 0.05; Figure 3B). Besides, in sodium butyrate–induced differentiation of SW1116 cells, knockdown of STK31 resulted in significant increase in the number of cells with vacuolated cytoplasm (P < 0.001; Figure 3C), a morphological characteristic of goblet-like differentiation (7,26). Although differentiation marked by vacuolated cytoplasm had been associated with senescence, it should not have been the case here because ← no massive cell death or senescence was observed (Supplementary Figure 5, available at Carcinogenesis Online). In addition, expression of mucin was significantly higher (~2–3-fold) in STK31 knockdown cells compared with that in control cells (P < 0.01; Figure 3D). These results suggest that STK31 may be required for maintaining the undifferentiated state of the two colon cancer cell lines. STK domain of STK31 is required for maintaining the undifferentiated state To confirm the role of STK31 in maintaining the undifferentiated state of colon cancer cells, we examined whether the overexpression of STK31 would inhibit differentiation. To further identify the functional domains on STK31 responsible for regulating differentiation, we constructed window-deletion mutants in two conserved domain, the N-terminal tudor domain (ΔTudor) and the C-eterminal kinase domain (ΔKinase), of STK31 (Figure 4A). Since the transfection efficiency of Caco2 cells is low, we used SW1116 cells for transient transfection experiments. To induce differentiation, sodium butyrate was added to the transfectants 24 h post-transfection. Overexpression of STK31 and the two mutants was confirmed by western blot (Figure 4B). Differentiation status of transient transfectants was determined by the expression of goblet cell differentiation marker—mucin. As expected, the overexpression of STK31 significantly decreased the expression of mucin in both untreated and butyrate-treated conditions (Figure 4C, P < 0.001 and P < 0.01, respectively), indicating that the STK31-overexpressed groups are less differentiated in both conditions. This result confirmed the role of STK31 in maintaining the undifferentiation state of colon cancer cells. In domain deletion experiment, the expression of mucin in Δtudor mutant-transfected cells was similar to that of full-length STK31 transfectant, suggesting that tudor domain was dispensable in regulating the differentiation of cancer cells (Figure 4C). Interestingly, deleting the STK domain partially reversed the inhibition of differentiation in untreated condition while completely abolishing the effect of STK31 overexpression on butyrate-induced differentiation (Figure 4C). These results suggest that the STK domain of STK31 is required for regulating the differentiation of colon cancer. Knockdown of STK31 suppresses tumorigenicity in vitro and in vivo The observed inhibition in cell proliferation and enhanced differentiation in STK31 knockdown colon cancer cells prompted us to assess the tumorigenicity of STK31 knockdown colon cancer cells in vitro and in vivo by soft agar colony formation assay and mice xenograft assay, respectively. As shown in Figure 5A, the knockdown of STK31 significantly decreased the number of colonies formed in soft agar in both CaCo2 and SW1116 cells (P < 0.05). In SW1116 xenografts, although all mice developed solid tumors, the size of tumors in STK31 knockdown group was significantly smaller than those in the control group (P < 0.05; Figure 5B). Moreover, histochemical staining showed that STK31 knockdown tumors displayed a more differentiated phenotype characterized by the formation of a glandular structure (Figure 5C). Similar results were obtained in Caco2 xenograft experiment where knockdown of STK31 decreased the number of mice developing solid tumors (Supplementary Figure 6A, available at Carcinogenesis Online). Furthermore, while sections of tumor formed by control miR Caco2 cells demonstrated typical welldifferentiated adenocarcinoma morphology, tumor formed by STK31 knockdown Caco2 cells showed remarkable karyorhexis and karyolysis, indicating the occurrence of necrosis (Supplementary Figure 6B, available at Carcinogenesis Online). Together, these results indicate (B) Live cell images of Caco2 and SW1116 stable transfectants showed that the knockdown of STK31 (STK31 kd) resulted in cell-flattened morphology in both cell lines (inset) at post-confluence stage, compared with control (Ctl) RNAi. Magnification: ×100. Quantitative analysis of occupied area by one cell in 2D cultures. The area occupied by STK31 knockdown transfectant in 2D cultures was significantly larger than that in control RNAi (P < 0.001). (C) Cell counting of both control (ctl RNAi) and STK31 knockdown cells (STK31 kd) from Caco2 and SW1116 cultures showed significant decrease in cell proliferation in STK31 knockdown cells after the culture reached confluence (P < 0.05). Data are presented as mean ± SEM; Student t test. (D) Knockdown of STK31 in Caco2 led to a G1/S arrest in confluent and post-confluent culture. Data are presented as mean ± SD; t test, *P < 0.05, **P < 0.01, ***P < 0.001. 2048 STK31 regulates colon cancer differentiation Fig. 3. Knockdown of STK31 enhances differentiation of Caco2 and SW1116 cells. 6 × 104 stable transfected Caco2 and 4 × 104 stable transfected SW1116 cells/well of a 24-well plate cells/well of a 24-well plate were cultured in a serum-containing medium for 14 days for spontaneous differentiation or treated 2049 K.L.Fok et al. Fig. 4. Effect of domain deletion and overexpression of STK31 on SW1116 differentiation. (A) Schematic diagram on the window deletion of STK31. (B) Western blot showing the overexpression of full-length and domain-deleted STK31 in SW1116 cells. The size of Δtudor and Δkinase are slightly smaller than that of the full-length protein. (C) Real time-PCR results showing the expression of goblet cell marker, mucin, was downregulated in STK31-overexpressed SW1116 cells under either untreated or sodium butyrate–induced differentiation condition. Δtudor-expressed cells showed similar mucin expression compared with that of full-length STK31 overexpressed cells. Deleting the STK domain completely abolished the down-regulation of mucin in sodium butyrate–induced condition. Data are presented as mean ± SEM; One-way analysis of variance, *P < 0.05, **P < 0.01, ***P < 0.001. that the knockdown of STK31 decreases tumorigenicity both in vitro and in vivo by promoting differentiation of colon cancer cells. Differential gene expression in STK31 knockdown-mediated colon cancer differentiation We next asked how STK31 could regulate the differentiation status of colon cancer. To answer this question, we compared the transcriptome of STK31 knockdown and control Caco2 cells using whole genome microarrays, and identified 95 genes that were significantly altered (≥2-fold). Interestingly, among the 95 genes regulated by STK31, 10 genes (KIT, NRG1, HDGFRP3, SMAD1, AR, GJA1, CCND2, TREFR1, FABP5 and SCIN) have been implicated both in tumorigenesis and in ← spermatogenesis (Supplementary Table 3, available at Carcinogenesis Online). To validate that genes identified by microarray study are bona fide targets of STK31, we performed real time-PCR to assess the expression levels of genes that (i) their function in tumorigenesis has been reported and with a 10-fold difference in microarray study or (ii) are known to implicate in both spermatogenesis and tumorigenesis. Our results showed that AR, FGF12, GJA1, HDGFRP3, LUM, NRG1, and SMAD1 were significantly downregulated in STK31 knockdown cells from both Caco2 and SW1116 cell lines (Figure 6A and 6B). Downregulation of KIT was only observed in STK31 knockdown Caco2 cells, but not in SW1116 cells probably because SW1116 do not express this receptor. Differential expression of the other two with 2 mM sodium butyrate (NaBT) for 48 h. (A) Knockdown of STK31 enhanced enterocytic differentiation in Caco2 cells as reflected by significant increase in number of dome (dash circles) and glandular crypt formation (arrows). Magnification: ×100. Quantification of data is shown on the right panel. (B) Real time-PCR showed that the expression of an enterocyte marker, villin, was up-regulated in STK31 knockdown Caco2 cells that underwent spontaneous (left) and sodium butyrate–induced differentiation (right) compared with control cells. Data are presented as mean ± SEM; t test, **P < 0.01, ***P < 0.001. (C) Knockdown of STK31 enhanced goblet-like differentiation in SW1116 cells as indicated by significant increase in the number of cells showing vacuole formation in cytoplasm (arrows). Magnification: ×100 (left panel) and ×200 (right panel). Quantification of data is shown on the right panel. (D) Real time-PCR showed that expression of goblet cell marker, mucin, was up-regulated in STK31 knockdown SW1116 cells that underwent sodium butyrate–induced differentiation. Data are presented as mean ± SEM; t test, **P < 0.01, ***P < 0.001. 2050 STK31 regulates colon cancer differentiation Fig. 5. Knockdown of STK31 decreases tumorigenecity with enhanced differenitation. (A) Knockdown of STK31 in both Caco2 and SW1116 cell lines decreases the number of colonies formed in soft agar colony formation assay. Quantification of data is shown below. (B) 1 × 106 STK31 knockdown SW1116 cells were subcutaneously grafted into athymic nude mice. Mice were killed 2 weeks later and the tumors were collected and dissected for histochemical staining. Knockdown of STK31 in SW1116 cells decreased the size of tumor formed (n = 8). (C) Tumor xenograft formed by STK31 knockdown SW1116 cells showed a higher differentiation status compared with that formed by control cells as indicated by the presence of glandular cavities (arrow and inset; scale bar: 100 µm). Data are presented as mean ± SEM; t test, *P < 0.05, ***P < 0.001. genes—ARL11 and PTPN13—was also observed in Caco2 cells only (Figure 6A and 6B). Discussion The present study has demonstrated that STK31 is highly expressed in multiple cancers, including colon cancer. Suppression of STK31 induces differentiation and inhibits tumorigenecity of colon cancer cells, both in vitro and in vivo. In general, tumors are heterogeneous in terms of their surface antigen expression, cell morphology (differentiation stages), proliferation, metastatic capacity, tumorigenecity and therapeutic resistance (1,2), indicating a cellular hierarchy consisting of cancer cells at various stages of abnormal differentiation. This cellular hierarchy is thought to be tightly regulated by homeostasis between proliferation and differentiation (27); however, the detailed regulatory mechanism remains unclear. Here, we have demonstrated that the reactivation of STK31 is crucial in maintaining the undifferentiated state of colon cancer cells. The heterogeneous expression pattern of STK31 in both cancer cell lines and primary tumors, together with the enhanced differentiation upon STK31 knockdown, strongly suggests its possible role in maintaining the primitive state of its expressing cells and thus the cellular hierarchy within the colon cancer cell lines or primary tumors. Moreover, the suppressed cancer growth associated with differentiation in vivo further lends compelling support for the anti-differentiation role of STK31 in colon cancer. STK31 contains two highly conserved domains: a tudor domain and a STK domain. Of interest, we found that the STK domain is required for regulating the differentiation of colon cancer cells while the tudor domain is dispensable for the process. It has been reported that the expression of STKs is frequently altered in human cancers (28). Importantly, the majority of the STKs are reactivated or up-regulated in cancers, suggesting that gaining the activities of STK may favor tumor development. In the present study, we have demonstrated for the first time that the STK domain of STK31 is required for maintaining the undifferentiation state of colon cancer, providing strong support to the importance of STK in regulating tumor development. Since a variety of inhibitors of STKs is available (29), they may be used for cancer therapy in targeting reactivated or up-regulated STKs, including STK31. The restricted expression in normal testis and aberrant expression in many tumors indicates that STK31 can be classified as cancer/testis gene. It has been suggested that spermatogenesis and tumorigenesis share significant similarities in biological and biochemical processes (30), which may be attributable to the reactivation of cancer/testis genes. Previous reports and our preliminary data revealed that STK31 expression was confined to spermatogonia in testis (13), indicating its pivotal role in germ cell differentiation. Therefore, we speculate that the regulation in colon cancer differentiation may resemble that in germ cell differentiation. In support of this notion, our gene expression profiling studies revealed that ~40% (39/95) of the STK31 target genes identified in colon cancer cells were expressed in germ cells. It is particularly noteworthy that several important autocrine/paracrine systems known to be involved in spermatogenesis (including KIT/ SCF, BMPs/SMADs, NRG1/ErbB and possibly hepatoma-derived growth factor HDGF/HDGFRP3) are regulated by STK31 in colon cancer cells. Moreover, most of the STK31 targets (KIT, SMAD, NRG1) in colon cancer cells are specifically expressed in spermatogonia in testis, in which STK31 is normally expressed, and has been implicated in the germ cell differentiation. During spermatogenesis, the receptor tyrosine kinase, KIT, and its ligand, stem cell factor (SCF), are crucial for spermatogonial proliferation and differentiation (31). Knockout of KIT has been shown to abolish spermatogonial 2051 K.L.Fok et al. Fig. 6. Candidate genes regulated by STK31 in Caco2 and SW1116 cells. (A, B) Differentially expressed genes in STK31 knockdown Caco2 cells (A) were identified by whole genome microarray and confirmed by real time-PCR. (B) Differential expressions of candidate genes were confirmed by real time-PCR in SW1116 cells. Data are presented as mean ± SEM; t test, *P < 0.05, **P < 0.01, ***P < 0.001. stem-cell differentiation, which results in sterility (32). Signaling by KIT also plays an important role in cellular transformation and differentiation through multiple signaling cascades, including STAT3, PI3K, PLC and MAPK (33–36). Of particular relevance to this study, human colorectal tumors express c-kit and activation of c-kit protected human colon adenocarcinoma cells against apoptosis enhanced their invasive potential (37). In our study, the downregulation of KIT by STK31 knockdown suggests that repression of KIT may be functionally important in STK31-mediated regulation of differentiation and tumorigenicity in colon cancer. Apart from KIT, BMP signaling plays critical roles both in colon cancer progression (38) and in spermatogonial differentiation (39). Particularly, SMAD1-mediated signaling cascades have been implicated in anchorage-independent growth (40) and tumorigenicity (41). Consistent with these findings, we have found that SMAD1 was downregulated by STK31 knockdown in the two colon cancer cell lines (Figure 6), indicating that SMAD1 is another possible target of STK31 contributing to its suppressive effects on tumorigenicity in vitro and in vivo. In conclusion, our study suggests that reactivation of STK31 in colon cancer maintains the undifferentiated state of colon cancer cells. Acquisition of this anti-differentiation ability could be one of the driving forces of tumorigenesis in colon cancer development. The current findings are of important clinical implications. Traditional cancer therapy has limitations with chemoresistance and cancer relapse, which are largely responsible for the treatment failure (10). With better understating on the molecular mechanisms underlying cellular heterogeneity and hierarchy in tumors, new therapeutic approach targeting undifferentiated tumorigenic tumors cells has evolved. The so-called differentiation therapy 2052 attenuates the tumorigenicity of cells with high differentiation potential through enforced commitment into defined lineages (42,43). Given its demonstrated role in regulating differentiation, together with its expression in a broad spectrum of cancers and specific expression in normal testis, STK31 represents a potent and promising target for the development of differentiation-directed therapy in colon cancer treatment. Supplementary material Supplementary Tables 1–3 and Figures 1–6 can be found at http:// carcin.oxfordjournals.org/ Funding This work was supported by National 973 program (2012CB944903), the Li Ka Shing Institute of Health Sciences and the Focused Investment Scheme of the Chinese University of Hong Kong. Acknowledgements The authors thank Prof XY Guan (Department of Clinical Oncology, The University of Hong Kong) and Prof KW Lo (Department of Anatomical and Cellular Pathology, The Chinese University of Hong Kong) for providing the cell lines cDNA for expression profiling. Conflict of Interest Statement: None declared. STK31 regulates colon cancer differentiation References 1.Ichim,C.V. et al. (2006) First among equals: the cancer cell hierarchy. Leuk. Lymphoma, 47, 2017–2027. 2.Heppner,G.H. (1984) Tumor heterogeneity. Cancer Res., 44, 2259–2265. 3.Marusyk,A. et al. (2010) Tumor heterogeneity: causes and consequences. Biochim. Biophys. Acta, 1805, 105–117. 4.Brattain,M.G., et al. (1984) Heterogeneity of human colon carcinoma. Cancer Metastasis Rev., 3, 177–191. 5.Dexter,D.L., et al. (1981) Heterogeneity of cancer cells from a single human colon carcinoma. Am. J. Med., 71, 949–956. 6.Dalerba,P. et al. (2011) Single-cell dissection of transcriptional heterogeneity in human colon tumors. Nat. Biotechnol., 29, 1120–1127. 7.Vermeulen,L. et al. (2008) Single-cell cloning of colon cancer stem cells reveals a multi-lineage differentiation capacity. Proc. Natl. Acad. Sci. U.S.A., 105, 13427–13432. 8.Sancho,E. et al. (2004) Signaling pathways in intestinal development and cancer. Annu. Rev. Cell Dev. Biol., 20, 695–723. 9.He,X.C. et al. (2004) BMP signaling inhibits intestinal stem cell self-renewal through suppression of Wnt-beta-catenin signaling. Nat. Genet., 36, 1117–1121. 10. Ricci-Vitiani,L. et al. (2008) Colon cancer stem cells. Gut, 57, 538–548. 11.Baselga, J. (2006) Targeting tyrosine kinases in cancer: the second wave. Science, 312, 1175–1178. 12. Krause,D.S. et al. (2005) Tyrosine kinases as targets for cancer therapy. N. Engl. J. Med., 353, 172–187. 13. Wang,P.J. et al. (2001) An abundance of X-linked genes expressed in spermatogonia. Nat. Genet., 27, 422–426. 14. Yokoe,T. et al. (2008) Efficient identification of a novel cancer/testis antigen for immunotherapy using three-step microarray analysis. Cancer Res., 68, 1074–1082. 15. Ho,S.N. et al. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene, 77, 51–59. 16. Lee,T.L. et al. (2009) GermSAGE: a comprehensive SAGE database for transcript discovery on male germ cell development. Nucleic Acids Res., 37, D891–D897. 17. Bao,J. et al. (2011) STK31(TDRD8) is dynamically regulated throughout mouse spermatogenesis and interacts with MIWI protein. Histochem. Cell Biol., 137, 377–389. 18. Ding,Q.M. et al. (1998) Caco-2 intestinal cell differentiation is associated with G1 arrest and suppression of CDK2 and CDK4. Am. J. Physiol, 275, C1193–C1200. 19. Cohen,E. et al. (1999) Induced differentiation in HT29, a human colon adenocarcinoma cell line. J. Cell Sci., 112, 2657–2666. 20. Chung,Y.S. et al. (1985) Effect of growth and sodium butyrate on brush border membrane-associated hydrolases in human colorectal cancer cell lines. Cancer Res., 45, 2976–2982. 21. Tsao,D. et al. (1982) Differential effects of sodium butyrate, dimethyl sulfoxide, and retinoic acid on membrane-associated antigen, enzymes, and glycoproteins of human rectal adenocarcinoma cells. Cancer Res., 42, 1052–1058. 22. Odoux,C. et al. (2008) A stochastic model for cancer stem cell origin in metastatic colon cancer. Cancer Res., 68, 6932–6941. 23. Yedgar,S. et al. (1992) Cyclic AMP-independent secretion of mucin by SW1116 human colon carcinoma cells. Differential control by Ca2+ ionophore A23187 and arachidonic acid. Biochem. J., 283, 421–426. 24. Chantret,I. et al. (1988) Epithelial polarity, villin expression, and enterocytic differentiation of cultured human colon carcinoma cells: a survey of twenty cell lines. Cancer Res., 48, 1936–1942. 25. Mariadason,J.M. et al. (2000) Divergent phenotypic patterns and commitment to apoptosis of Caco-2 cells during spontaneous and butyrate-induced differentiation. J. Cell Physiol, 183, 347–354. 26. Akare,S. et al. (2006) Ursodeoxycholic acid modulates histone acetylation and induces differentiation and senescence. Int. J. Cancer, 119, 2958–2969. 27. Jordan,C.T. et al. (2006) Cancer stem cells. N. Engl. J. Med., 355, 1253–1261. 28. Capra,M. et al. (2006) Frequent alterations in the expression of serine/ threonine kinases in human cancers. Cancer Res., 66, 8147–8154. 29.Scapin,G. (2002) Structural biology in drug design: selective protein kinase inhibitors. Drug Discov. Today, 7, 601–611. 30. Simpson,A.J. et al. (2005) Cancer/testis antigens, gametogenesis and cancer. Nat. Rev. Cancer, 5, 615–625. 31. Mithraprabhu,S. et al. (2009) Control of KIT signalling in male germ cells: what can we learn from other systems? Reproduction, 138, 743–757. 32. Ohta,H. et al. (2003) Proliferation and differentiation of spermatogonial stem cells in the w/wv mutant mouse testis. Biol. Reprod., 69, 1815–1821. 33.Hemesath,T.J. et al. (1998) MAP kinase links the transcription factor Microphthalmia to c-Kit signalling in melanocytes. Nature, 391, 298–301. 34. Chian,R. et al. (2001) Phosphatidylinositol 3 kinase contributes to the transformation of hematopoietic cells by the D816V c-Kit mutant. Blood, 98, 1365–1373. 35. Plo,I. et al. (2001) Kit signaling and negative regulation of daunorubicin-induced apoptosis: role of phospholipase Cgamma. Oncogene, 20, 6752–6763. 36. Ning,Z.Q. et al. (2001) STAT3 activation is required for Asp(816) mutant c-Kit induced tumorigenicity. Oncogene, 20, 4528–4536. 37. Bellone,G. et al. (2001) Aberrant activation of c-kit protects colon carcinoma cells against apoptosis and enhances their invasive potential. Cancer Res., 61, 2200–2206. 38. Hardwick,J.C. et al. (2008) Bone morphogenetic protein signalling in colorectal cancer. Nat.Rev.Cancer, 8, 806–812. 39. Itman,C. et al. (2008) SMAD expression in the testis: an insight into BMP regulation of spermatogenesis. Dev. Dyn., 237, 97–111. 40. Daly,A.C. et al. (2008) Transforming growth factor beta-induced Smad1/5 phosphorylation in epithelial cells is mediated by novel receptor complexes and is essential for anchorage-independent growth. Mol. Cell Biol., 28, 6889–6902. 41. Freudenberg,J.A. et al. (2007) Induction of Smad1 by MT1-MMP contributes to tumor growth. Int. J. Cancer, 121, 966–977. 42. Wicha,M.S. et al. (2006) Cancer stem cells: an old idea—a paradigm shift. Cancer Res., 66, 1883–1890. 43. Visvader,J.E. et al. (2008) Cancer stem cells in solid tumours: accumulating evidence and unresolved questions. Nat. Rev. Cancer, 8, 755–768. Received April 11, 2012; revised June 28, 2012; accepted July 17, 2012 2053