Functional Characterization of the Tumor-Suppressor MARCKS in Colorectal Cancer and Its Association with Survival

ORIGINAL ARTICLE Functional characterization of the tumor-suppressor MARCKS in colorectal cancer and its association with survival

M Bickeböller1,2, KE Tagscherer1,2, M Kloor3, L Jansen4, J Chang-Claude5, H Brenner4, M Hoffmeister4, C Toth2,7, P Schirmacher2, W Roth1,2,8 and H Bläker6,8

The myristoylated alanine-rich C-kinase substrate (MARCKS) acts as a tumor suppressor in a variety of human neoplasms. In colorectal cancers (CRCs), MARCKS has been shown to be a preferential target of mutational inactivation in tumors following the microsatellite instability (MSI-H) pathway but little is known about its impact on intestinal carcinogenesis. To investigate the relevance of MARCKS inactivation in more detail, we analyzed 926 MSI-typed CRCs for MARCKS expression by immunohistochemistry and studied the functional consequences of MARCKS depletion in colorectal cancer cell lines. We found that loss of MARCKS expression was not restricted to MSI-H cancers but also occurred in microsatellite stable (MSS) tumors, where it was associated with an adverse outcome regarding overall survival, cancer-speciﬁc and disease-free survival (P = 0.002, P = 0.0018, P = 0.0001, respectively; univariate analysis). In MARCKS-positive MSS colon cancer cell lines (SW480 and SW707) small interfering RNA (siRNA)-mediated knockdown of MARCKS conferred resistance to tumor necrosis factor-related apoptosis-inducing ligand (TRAIL)-induced apoptosis. This was accompanied by the downregulation of the TRAIL receptors DR4 and DR5 at the cell surface and activation of AKT signaling. Inhibition of AKT signaling and transient overexpression of wild-type MARCKS, but not of MARCKS lacking the effector domain (ED), abolished the anti-apoptotic effect. In conclusion, our data show that inactivation of MARCKS is common in CRCs and is associated with adverse outcome in MSS cancers. The ﬁnding that MARCKS acts as a mediator of apoptosis in MSS CRC cells adds a novel tumor-suppressing function to the so far established roles of MARCKS in cell motility and proliferation and can explain the prognostic effect of MARCKS depletion in MSS CRC.

Oncogene (2015) 34, 1150–1159; doi:10.1038/onc.2014.40; published online 24 March 2014

INTRODUCTION are associated with increased proliferation in different types of The myristoylated alanine-rich C-kinase substrate (MARCKS) transformed cell lines, for example, melanoma cells, transformed protein was identified almost 25 years ago as major specific fibroblasts, neuroblastoma and glioma cells.8–12 In large and small substrate for protein kinase C in bovine forebrain supernatant.1 bowel adenocarcinomas following the microsatellite instability MARCKS is an ubiquitously expressed, rod-shaped protein of 335 (MSI-H) pathway, the MARCKS gene has been identified as a major amino acids,1,2 which contains three domains: an N-terminal target of inactivation through coding MSI.13–15 In line with the myristoylated domain, which mediates binding to membranes, a inactivating nature of coding MSI, loss of MARCKS expression was highly conserved MH2 domain of unknown function and a basic found by immunohistochemical analysis in a series of small bowel effector domain (ED) containing the protein kinase C phosphoryla- adenocarcinomas.13 Tumor-specific expression loss, however, was tion sites as well as binding sites for calmodulin and actin.3–5 not restricted to cancers following the MSI-H pathway but also Dependent on its phosphorylation status, MARCKS cycles between occurred in microsatellite stable (MSS) adenocarcinomas, indicat- the plasma membrane and the cytosol. Upon phosphorylation by ing a role of MARCKS inactivation also in MSS carcinogenesis. protein kinase C, MARCKS translocates to the cytosol where In this study, we have investigated the putative role of MARCKS dephosphorylation results in membrane reassociation.6 At the as a tumor suppressor and regulator of apoptosis in MSS membrane, MARCKS cross-links actin filaments into bundles, colorectal cancer (CRC). We demonstrate that MARCKS inactivation resulting in significant alterations in the actin cytoskeleton,7 in the MSS colon cancer cell lines SW480 and SW707 confers anti- which affect among others cell spreading and cell motility.8,9 apoptotic effects and provide evidence that these effects result MARCKS has been described as a tumor suppressor in various from a reduction of the tumor necrosis factor-related apoptosis- human malignancies. Downregulated MARCKS expression levels inducing ligand (TRAIL) receptors DR4 and DR5 at the cell surface

1Division of Molecular Tumor Pathology, German Cancer Research Center (DKFZ), Heidelberg, Germany; 2Department of General Pathology, Institute of Pathology, University of Heidelberg, Heidelberg, Germany; 3Department of Applied Tumor Biology, Institute of Pathology, University of Heidelberg, Heidelberg, Germany; 4Division of Clinical Epidemiology and Aging Research, German Cancer Research Center, Heidelberg, Germany; 5Division of Cancer Epidemiology, Unit of Genetic Epidemiology, German Cancer Research Center, Heidelberg, Germany and 6Department of General Pathology, Institute of Pathology, University Hospital Charité, Campus Mitte, Berlin, Germany. Correspondence: Professor H Bläker, Department of General Pathology, Institute of Pathology, University Hospital Charité, Campus Mitte, Charitéplatz 1, 10117 Berlin, Germany. E-mail: [email protected] 7Current address: Institute of Pathology, Heinrich Heine University, Düsseldorf, Germany. 8These authors shared senior authorship. Received 29 July 2013; revised 13 December 2013; accepted 12 January 2014; published online 24 March 2014 MARCKS in colorectal cancer M Bickeböller et al 1151 and subsequently decreased activation of caspase 3 as well as MARCKS after adjustment for relevant confounders. Supporting caspase 8. We also demonstrate that MARCKS inactivation this assumption, a trend for an association of MARCKS loss in stimulates pro-survival pathways by increasing phosphorylation MSS CRCs with higher T-stage and nodal involvement (pN+), of AKT. Moreover, we show that loss of MARCKS expression characteristics adjusted for in multivariate analysis, was found in occurs in MSS colon cancers where it is associated with adverse MSS tumors as a potential explanation for the lack of signiﬁcance prognosis. in multivariate analysis (Supplementary Table 2).

MARCKS regulates sensitivity toward apoptosis RESULTS Previous studies pointed to a putative tumor-suppressor role Loss of MARCKS expression in CRC for the MARCKS protein in intestinal cancers.13 To functionally As a first step, we examined expression of MARCKS in samples of characterize MARCKS in CRC, we initially determined MARCKS human CRC and normal mucosa by immunohistochemistry. In protein expression in several CRC cell lines. The three MSS CRC cell the normal mucosa, MARCKS-positive epithelial cells displayed lines SW480, SW707 and HT29 revealed a high expression of cytoplasmic and membranous expression (Figures 1a and b, MARCKS, whereas the MSI CRC cell lines HCT116 and RKO showed Supplementary Figures 1A–D). In addition, MARCKS expression a moderate expression (Figure 2a).15,17,18 As the localization and was also recognized in follicular dendritic cells and ganglionic cells thereby the function of MARCKS depends on its phosphorylation of the submucosal plexus (Figure 1b, arrows, Supplementary status, we also analyzed the pMARCKS expression levels. In the Figures 1 E-H). Most colon cancers displayed strong expression of two MSI-H cell lines, pMARCKS could not be detected (HCT116, MARCKS (Figures 1c and d), whereas loss of expression was found RKO). The MSS-positive SW480 and SW707 cells showed low levels, in 14% of cancers (Figures 1e and f). Tumor heterogeneity with whereas HT29 cells exhibited high levels of pMARCKS (Figure 2a). areas of strong and weak MARCKS expression was observed in Considering the MSS status and the high levels of MARCKS with some cases (Figure 1g) and was scored as MARCKS positive. concurrent low levels of pMARCKS in SW480 and SW707 cells, we Among the cancerous and corresponding normal tissue cores chose these two cell lines for further experiments. In order to mounted on the tissue micro arrays, 977 cases were informative investigate whether MARCKS is able to mediate sensitivity concerning the evaluation of MARCKS expression (patient toward apoptosis induced by different cytotoxic agents, we description in Supplementary Table 1). Out of these, the examined the functional consequences of a decreased expression microsatellite status was known in 926 cases.16 As expected by of MARCKS in SW480 and SW707 cells achieved by specific the frequent targeting of the MARCKS gene through coding MSI, MARCKS-targeting small interfering RNAs (siRNAs; Figure 2b). The loss of MARCKS expression was more frequent in MSI-H cancers siRNA-mediated knockdown of MARCKS resulted in a significant (47 of 96 cases (49%)) than in MSS cancers (79 of 830 cases (9.5%)). decrease of apoptotic cell death induced by oxaliplatin and In univariate analysis of all tumors, loss of MARCKS expression was 5-fluorouracil (5-FU; Figure 2c). As an acquired resistance against associated with reduced overall survival (Table 1). When the therapeutics like 5-FU or oxaliplatin represents a major problem cohort was split according to the microsatellite status, an regarding the treatment of CRC, we further investigated whether association of MARCKS loss with reduced overall, cancer-specific 5-FU or oxaliplatin influences the expression levels of MARCKS and and disease-free survival was found for MSS but not for MSI-H pMARCKS, respectively (Supplementary Figure 2). Treatment with cancers in univariate analysis (P = 0.002, P = 0.0018, P = 0.0001, both oxaliplatin and 5-FU did not have any effects on the respectively; Table 1) indicating that the biologic relevance of expression of MARCKS. However, treatment with 5-FU resulted in MARCKS inactivation may differ between the molecular subtypes decreased pMARCKS levels, pointing to an involvement of 5-FU in of MSI and MSS cancers. In multivariate analysis, however, the the cycling of MARCKS. Besides affecting sensitivity to stimuli of influence of MARCKS loss on survival in MSS cancers lacked the intrinsic apoptotic pathway, downregulation of MARCKS statistical significance. This may be explained by an insufficient further led to a decreased susceptibility to the death ligand sample size to detect a potentially lower independent effect of TRAIL (Figure 2d). Consistently, TRAIL treatment resulted in a

Figure 1. MARCKS expression analysis in normal colon mucosa and CRCs. MARCKS expression in normal colon mucosa (a, b) and CRCs (c–g): normal intestinal epithelium displayed moderate expression of MARCKS comparable to the expression in ganglionic cells of the submucosal plexus (b, arrows). Strong expression of MARCKS in CRCs (c, d), cancer-speciﬁc loss of expression (e, f; arrows), focal loss of MARCKS expression in cancer (g, inset). Original magniﬁcation (a–g: x20; g inset: x100).

Table 1. Survival analysis according to MARCKS expression

All tumors MARCKS N Events Univariate analysis Multivariate analysis

HR 95% CI P-value HRa 95% CI P-value

Overall survival Positive 843 283 (34%) 1.00 Reference 1.00 Reference Negative 134 59 (44%) 1.43 1.08–1.90 0.0122 1.28 0.96–1.70 0.0992 CRC-speciﬁc survival Positive 838 212 (25%) 1.00 Reference 1.00 Reference Negative 134 43 (32%) 1.38 1.00–1.92 0.0523 1.26 0.90–1.77 0.1772 Recurrence-free survival Positive 841 274 (33%) 1.00 Reference 1.00 Reference Negative 134 51 (38%) 1.34 0.98–1.83 0.0657 1.18 0.86–1.63 0.3062

MSS tumors Overall survival Positive 751 257 (34%) 1.00 Reference 1.00 Reference Negative 79 39 (49%) 1.70 1.21–2.38 0.0020 1.33 0.95–1.88 0.1009 CRC-speciﬁc survival Positive 747 196 (26%) 1.00 Reference 1.00 Reference Negative 79 32 (41%) 1.82 1.25–2.64 0.0018 1.39 0.95–2.03 0.0943 Recurrence-free survival Positive 750 252 (34%) 1.00 Reference 1.00 Reference Negative 79 39 (49%) 1.98 1.39–2.81 0.0001 1.30 0.90–1.87 0.1649

MSI-high tumors Overall survival Positive 49 11 (22%) 1.00 Reference 1.00 Reference Negative 47 15 (32%) 1.41 0.65–3.07 0.3891 1.51 0.63–3.60 0.3537 CRC-speciﬁc survival Positive 49 4 (8%) 1.00 Reference 1.00 Reference Negative 47 7 (15%) 1.83 0.54–6.26 0.3351 3.19 0.64–15.85 0.1556 Recurrence-free survival Positive 48 6 (13%) 1.00 Reference 1.00 Reference Negative 47 8 (17%) 1.33 0.46–3.83 0.5992 2.01 0.57–7.04 0.2765 Abbreviations: CI, conﬁdence interval; CRC, colorectal cancer; HR, hazard ratio; MARCKS, myristoylated alanine-rich C-kinase substrate; MSI, microsatellite instability; MSS, microsatellite stable. aHR-adjusted age at diagnosis, sex, UICC stage (International Union Against Cancer), location of CRC (colon and rectum), chemotherapy, neoadjuvant treatment, and the time-dependent effect chemotherapy*log(time).

reduced activation of caspase 8 and caspase 3 after down- and DR5 at the membrane (Figure 4b). However, neither the total regulation of MARCKS. This was accompanied by diminished mRNA level nor the total protein level of the two TRAIL receptors cleavage of the caspase substrate PARP (poly (ADP-ribose) was affected by MARCKS knockdown (Figures 4c and d). Therefore, polymerase) (Figure 2e). we further investigated whether MARCKS has an effect on To investigate the role of MARCKS in apoptosis regulation in receptor localization in general. However, no alterations in the more detail, we transiently overexpressed wild-type MARCKS as surface expression levels of transferrin receptor, epidermal growth well as a MARCKS-ΔED mutant, which is exclusively localized in the factor receptor or integrin alphaVbeta5 receptor after siRNA- cytoplasm (Figures 3a and b) because of the lack of its polybasic mediated knockdown of MARCKS was observed (data not shown). ED. MARCKS-ED contains the protein kinase C phosphorylation In addition, we proved that although siRNA-mediated knockdown sites and therefore determines the localization of the protein.19 of MARCKS led to reduced expression levels of DR4 and DR5 at the Previous studies have shown that the ED is crucial for fulfilling cell surface, receptor internalization upon binding its ligand TRAIL diverse functions of MARCKS.20 To investigate the importance of was not impaired in MARCKS-depleted cells (Supplementary the ED with regard to the function of MARCKS in regulating Figure 4). apoptosis, we decided to use the MARCKS-ΔED mutant as a control in our study. We found that overexpression of MARCKS MARCKS modulates AKT signaling significantly increased TRAIL-induced apoptotic cell death in MSS-positive SW480 and SW707 cells, whereas the mutated As we observed a strong decrease in sensitivity toward oxaliplatin version had negligible effects (Figure 3c). In the MSI-H cell line as well as 5-FU, both triggers of the intrinsic apoptosis pathway in HCT116, however, overexpression of MARCKS had no effect on MSS CRC cells, we further investigated whether this was a result of apoptosis (Supplementary Figure 3). On the basis of these results, increased activation of pro-survival pathways. As MARCKS is able we conclude that MARCKS acts as a regulator of apoptosis in MSS to sequester PIP2 at the plasma membrane, we assumed that the colon adenocarcinoma cells. However, an intact ED and thus knockdown of MARCKS was accompanied by an intensified localization at the plasma membrane is necessary for this effect. phosphoinositide 3-kinase (PI3K)/AKT signaling because of an increased availability of PIP2. Indeed, a distinct activation of AKT was observed after siRNA-mediated knockdown of MARCKS MARCKS regulates cell surface expression of TRAIL receptors (Figure 5a, Supplementary Figure 5, with serum starvation). To Our experiments showed a significant decrease in sensitivity prove whether signaling through the PI3K/AKT pathway was toward the death ligand TRAIL in MSS cancer cells. In order to involved in resistance to apoptosis mediated by downregulation investigate whether this was due to impaired death receptor of MARCKS, we used the selective AKT inhibitors perifosine signaling, we analyzed the cell surface expression levels of DR4 and AKT inhibitor IV, respectively, and LY294002, a reversible and DR5 in SW480 and SW707 cells by flow cytometry. Compared inhibitor of PI3K. Treatment of SW480 and SW707 cells with the with cells transfected with a control siRNA, we observed an explicit different inhibitors diminished pAKT levels effectively, whereas decrease in TRAIL death receptors after knockdown of MARCKS MARCKS and pMARCKS levels remained unchanged (Figure 5b, (Figure 4a). Furthermore, SW480 cells were characterized by Supplementary Figure 6, without serum starvation). As expected, a stronger extent of receptor relocation than SW707 cells. the simultaneous inhibition of AKT abrogated the anti-apoptotic Consistently, cell fractionation of cells transfected with a effect of siRNA-mediated knockdown of MARCKS toward TRAIL MARCKS-specific siRNA also revealed a distinct decrease of DR4 (Figure 5c).

Oncogene (2015) 1150 – 1159 © 2015 Macmillan Publishers Limited MARCKS in colorectal cancer M Bickeböller et al 1153 DISCUSSION recent studies on CRC following the MSI-H pathway revealed a MARCKS has been shown to have an important functional role in high frequency of frameshift mutations in MARCKS.15,26 In analysis the pathogenesis of a variety of malignant tumors.21–25 Moreover, of small bowel cancers, we were able to show that loss of MARCKS

Figure 2. Knockdown of MARCKS decreases sensitivity of CRC cells to apoptosis. (a) Immunoblot analysis of the expression of MARCKS and pMARCKS in various CRC cell lines. (b) SW480 and SW707 cells were transfected with either a nonspeciﬁc control siRNA or two different MARCKS siRNAs (20 nM) and a MARCKS siRNA pool (20 nM), respectively. The successful downregulation of MARCKS was monitored 48 h after transfection by immunoblot analysis. (c) SW480 and SW707 cells were transfected as described in (b). After 48 h of treatment with oxaliplatin (SW480: 150 μM; SW707: 200 μM) and 5-FU (400 μM), respectively, cells were subjected to apoptosis measurement using propidium iodine (PI) staining and ﬂuorescence-activated cell sorting (FACS) analysis. Data represent the mean of four independent experiments (n = 4 ± s.d., Student’s t-test, *Po0.05). (d) SW480 and SW707 cells were transfected as described in (b). After 24 h of treatment with TRAIL, cells were subjected to apoptosis measurement using PI staining and FACS analysis. Data represent the mean of four independent experiments (n = 4 ± s.d., Student’s t-test, *Po0.05). (e) SW480 and SW707 cells were transfected as described in (b). After 6 h of TRAIL treatment (80 ng/ml), whole cellular lysates were generated and subjected to immunoblot analysis for full-length (p32) and cleaved caspase 3 (p19/p17), full-length (p55/p53) and cleaved caspase 8 (p43/41; p30; p18) as well as cleavage of PARP. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as a loading control.

Figure 3. Overexpression of MARCKS increases sensitivity of CRC cells to apoptotic cell death. (a) SW480 and SW707 cells were transiently transfected with MARCKS wt-Flag, MARCKSΔED-Flag or an empty pcDNA3 vector as a control. After 24 h expression of MARCKS, wt-Flag and MARCKSΔED-Flag was determined by immunoblot analysis. The membrane was probed for MARCKS, pMARCKS and FLAG, respectively. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as a loading control. (b) SW480 cells were transiently transfected as described in (a) and subjected to a membrane fractionation. Total cellular membrane proteins (M) as well as cytosolic fraction proteins (c) were used for immunoblot analysis. GAPDH served as a loading control for the cytosolic fraction. EpCAM served as a loading control for the membrane fraction. (c) SW480 and SW707 cells were transiently transfected as described in (a) and treated for 24 h with TRAIL (50 ng/ml). Cells were subjected to apoptosis measurement using ﬂuorescence-activated cell sorting (FACS) analysis as described in the Materials and Methods section. Data represent the percentage of apoptotic cell death in three independent experiments (n = 3 ± s.d., Student’s t-test, *Po0.05).

protein expression was detectable in both MSI-H and MSS cancers. a regulator of apoptosis in MSS CRC. We found that modulation of Although the loss of MARCKS expression correlated with biallelic MARCKS expression significantly changed the sensitivity to the mutational inactivation of MARCKS in MSH-I cancers, the death ligand TRAIL, which was caused by a MARCKS-dependent mechanisms of inactivation in MSS cancers are unknown. In order alteration of the membranous localization of the death receptors to elucidate whether the depletion of MARCKS also has a role in DR4 and DR5. The extent of death receptor relocation was cell line CRC, we evaluated its expression in 926 colorectal tumors dependent and may contribute to the slightly diverging sensitivity collected in the course of a population-based study and previously of SW480 and SW707 cells to TRAIL-induced apoptosis upon typed for their microsatellite status.16 We observed that MARCKS MARCKS knockdown. Our findings may further explain the was expressed in non-neoplastic colorectal mucosa and in the prognostic relevance of MARCKS in MSS CRC. It is already known majority of cancers, whereas MARCKS was not expressed in 47 out that for immune surveillance of tumors and metastases, natural of 96 (49%) MSI-H and in 79 out of 830 (9.5%) MSS tumors. killer cells use the death ligand TRAIL to eliminate tumor cells.32,33 Furthermore, we found the lack of MARCKS expression to be Accordingly, a previous study on CRC has shown that the cell associated with an adverse outcome. However, only in MSS surface expression of DR4 and DR5 is of greater importance for cancers the effect was significant indicating that the biologic prognosis than the total expression level.34 In view of the relevance of MARCKS inactivation differs between the molecular development of TRAIL agonists for therapeutic applications, which subtypes of MSI and MSS cancers as suggested by the diverging selectively induce apoptosis in cancer cells,35 analyzing MARCKS effect of MARCKS overexpression on apoptosis sensitivity in MSS expression in CRC may become a future issue in predictive testing. or MSI-H cell lines. Although the reduction of DR4 and DR5 receptors at the cell Although CRC is one of the most prevalent cancers worldwide, surface after siRNA-mediated knockdown of MARCKS was evident, the functional relevance of MARCKS inactivation during tumor we did not observe altered protein or mRNA expression levels of progression is mostly unknown. Rombouts et al.27 showed that DR4 or DR5 upon MARCKS downregulation. These data suggest downregulation of MARCKS resulted in reduced motility and that MARCKS is involved in regulating TRAIL receptor trafficking invasion in CRC. In other cancer types, MARCKS has been shown to from and to the cell surface, respectively. Although the proper influence cellular processes such as invasion, metastasis, motility, cellular localization of the TRAIL receptors DR4 and DR5 are a adhesion or cell spreading.24,27–31 Recently, a study in gliomas prerequisite for the activation of TRAIL-induced apoptosis, only proposed reduced MARCKS expression as a biomarker for an little is known about TRAIL receptor trafficking and recycling so adverse prognosis.22 far. One possibility to regulate the amount of TRAIL receptors at In view of the prognostic impact of MARCKS expression in MSS the cell surface comprises their endocytotic internalization. cancers and the need of additional functional analyses, we sought As MARCKS is able to sequester PIP2, an important mediator of to elucidate the relevance of MARCKS regarding apoptosis endocytosis,36 it is conceivable that downregulation of MARCKS signaling. Our study provides first evidence that MARCKS acts as leads to an elevation of free PIP2 and consequently to increased

Figure 4. Downregulation of MARCKS leads to a decrease in the expression of TRAIL receptors DR4 and DR5 at the cell surface. (a) Cell surface expression of TRAIL receptors in SW480 and SW707 cells after siRNA-mediated knockdown of MARCKS. The expression of DR4 and DR5 was determined by flow cytometry as described in the Materials and Methods section. Data represent the median fluorescence intensity of fluorescein isothiocyanate (FITC) in three independent experiments (n = 3 ± s.d., Student’s t-test, *Po0.05). (b) SW480 and SW707 cells were transfected with either a nonspecific control siRNA or a MARCKS siRNA pool (20 nM). After 48 h, cells were subjected to membrane fractionation. Total cellular membrane proteins as well as cytosolic fraction proteins were used for immunoblot analysis. (c) Isolation of RNA was carried out 48 h after siRNA-mediated knockdown of MARCKS. Quantitative reverse transcriptase–PCR (RT–PCR) expression levels of MARCKS, DR4 and DR5 mRNA were normalized to 18 S rRNA expression. Data are depicted as the relative changes compared with cells transfected with nonspecific control siRNA and represent the mean of three independent experiments (n = 3 ± s.d., Student’s t-test, *Po0.05). (d) SW480 and SW707 cells were transfected with either a nonspecific control siRNA or two different MARCKS siRNAs (20 nM) and a MARCKS siRNA pool (20 nM), respectively. At 48 h after transfection, total cellular lysates were subjected to immunoblot analysis for endogenous DR4 and DR5 levels. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as a loading control. endocytosis of TRAIL receptors at the plasma membrane. surface and in the endoplasmatic reticulum/trans-Golgi network Consistently, Su et al.37 reported that MARCKS localized at the system.39 It is known that the availability of Fas/CD95 death plasma membrane may inhibit endocytosis by sequestering receptors is regulated at the trans-Golgi network-plasma mem- PIP2. However, we could exclude that the effect of MARCKS on DR brane trafficking step mediated by cargo transport vesicles localization is due to a changed endocytosis rate (data not shown). composed of various coating and adapter proteins.40–42 Recently, Apart from endocytosis, MARCKS may be able to regulate ARAP1, an ArfGAP and RhoGAP adapter protein was discovered 43 further PIP2-dependent mechanisms, for example, exocytotic as a DR4 binding partner. Simova et al. showed that down- trafficking.38 DR4 and DR5 are predominantly localized at the cell regulation of ARAP1 induced a loss of membrane expression of

Figure 5. MARCKS modulates AKT signaling. (a) SW480 and SW707 cells were transfected with either a nonspecific control siRNA or two different MARCKS siRNAs (20 nM) and a MARCKS siRNA pool (20 nM), respectively. At 24 h after transfection, cells were serum starved for additional 24 h before harvesting. Total cellular lysates were subjected to immunoblot analysis. Total AKT and pAKT levels were detected in whole-cell lysates. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) served as a loading control. pAKT expression levels were densitometrically quantified using Image J Software (NIH, Bethesda, MD, USA) (see Supplementary Figure 5). (b) SW480 and SW707 cells were treated with the selective AKT inhibitors perifosine and AKT inhibitor IV, respectively or the PI3K inhibitor LY294002 for 24 h before whole-cell lysates were used to monitor the decrease in phosphorylation of AKT. GAPDH served as a loading control. (c) SW480 and SW707 cells were transfected as described in (a). At 24 h after transfection, cells were treated with AKT inhibitor IV (SW480: 1 μM, SW707: 1.5 μM), perifosine (1.5 μM) and LY294002 (30 μM), respectively. At 24 h after pre-incubation with the various inhibitors, cells were treated with TRAIL (50 ng/ml) for additional 24 h, before apoptosis measurement using propidium iodine (PI) staining and fluorescence-activated cell sorting (FACS) analysis. Data represent the percentage of apoptotic cell death in three independent experiments (n = 3 ± s.d., Student’s t-test, *Po0.05).

DR4 and impaired TRAIL-induced cell death. They further claimed recognizes phospholipids especially in highly curved vesicles,45 it that the protein could be an important mediator of TRAIL receptor is conceivable that the protein serves as an adapter protein trafficking from the intracellular stores, the trans-Golgi network, to having a role in establishing or stabilizing transport vesicles, the membrane. Other studies revealed that ARAP1 specifically thereby influencing the complex process of receptor trafficking binds to phospholipids via its two PH domains.44 As MARCKS also similarly to ARAP1.

Oncogene (2015) 1150 – 1159 © 2015 Macmillan Publishers Limited MARCKS in colorectal cancer M Bickeböller et al 1157 Interestingly, diminished levels of MARCKS increase prolifera- for 24 h. Cell lines were authenticated by single-nucleotide polymorphism tion and radiation resistance by enhancing the activity of the PI3K/ profiling50 and tested regularly for contaminations by multiplex PCR AKT signaling pathway in glioma cell lines.22 A similar effect of performed in the core facility of the German Cancer Research Center 51 decreased MARCKS expression was also found in the CRC cell lines (DKFZ), Heidelberg, Germany. used in our study. In view of the role of unphosphorylated MARCKS in sequestering PIP2 at the cell membrane, we assume Flow cytometry analysis that PIP2 is blocked for its interaction with the PI3K. According to For analysis of cell death, 6 × 105 CRC cells were treated as indicated, this hypothesis, the knockdown of MARCKS would release PIP2. detached and stained with propidium iodine (50 μg/ml in Nicoletti buffer). Subsequently, by phosphorylating PIP2, PI3K generates PIP3, the Cells were subjected to flow cytometry analysis of DNA content using a phospholipid, which recruits the effector kinase AKT to the Becton Dickinson FACScalibur cytometer and Cell Quest Software (BD membrane, where it is activated by other kinases. Indeed, Biosciences, Heidelberg, Germany). immunoblot analysis revealed increased levels of pAKT upon For measurement of cell surface expression of DR4/DR5 and for downregulation of MARCKS. The resistance toward TRAIL was intracellular staining of cleaved caspase 3 and transiently overexpressed Δ 5 abolished by concomitant treatment with different inhibitors of MARCKS wt-Flag and MARCKS ED-Flag, 6 × 10 CRC cells were washed, the PI3K/AKT signaling pathway. We thus speculated whether the incubated with cell dissociation buffer (Gibco cell culture, Darmstadt, Germany) for 30 min at 37 °C, harvested by centrifugation. For intracellular altered localization of TRAIL receptors may be caused by the staining, cells were fixed with 4% paraformaldehyde for 10 min and elevated AKT activity after MARCKS knockdown. However, analysis permeabilized with 90% methanol on ice for 30 min. For cell surface and of the cell surface expression of the TRAIL receptors after for intracellular staining, cells were stained with fluorescently labeled simultaneous downregulation of MARCKS and inhibition of AKT antibodies or mouse IgG1-isotype control antibody for 60 min at room revealed that the reduction of the receptors was not mediated via temperature. Cells were washed, resuspended in phosphate-buffered the AKT signaling pathway in CRC. In accordance with our saline containing 0.5% fetal bovine serum and propidium iodine (1 μg/ml) findings, a study on prostate cancer showed that AKT inhibition and subjected to flow cytometry analysis for cell surface or intracellular staining. For apoptosis measurement after transient overexpression of sensitized prostate cancer cells to TRAIL-induced apoptosis Δ without influencing the TRAIL receptor expression at the cell MARCKS wt-Flag and MARCKS ED-Flag, respectively, cells were stained 46 with an Alexa Fluor-488-labeled anti-flag antibody and analyzed for surface. The fact that MARCKS has an impact on pro-survival Flag expression by fluorescence-activated cell sorting analysis. Flag- AKT signaling pathway may provide a second mechanism of positive cells were gated on scatter plots and analyzed for apoptotic apoptosis regulation in addition to MARCKS effect on DR4 cell death using a cleaved caspase 3 phycoerythrin-conjugated and DR5 localization and the accompanying regulation of the antibody. Fluorescence-activated cell sorting analysis was carried out TRAIL-induced apoptosis. using a Becton Dickinson FACScalibur cytometer and Cell Quest Software Similar to previous studies showing that the ED of MARCKS is (BD Biosciences). pivotal for its function,47–49 we found that the ED was indispensable for MARCKS to execute its regulatory effects on Transfections apoptosis in MSS CRC cells. Overexpression of wild-type MARCKS, but not MARCKS-ΔED, a mutant lacking the ED, significantly CRC cells were transfected with siRNA using Lipofectamine 2000 (Invitrogen). MARCKS siRNA #1, MARCKS siRNA #2 and MARCKS siRNA enhanced apoptosis sensitivity of CRC cells toward TRAIL. As an pool were obtained from Santa Cruz (San Diego, CA, USA) (sc-35857). A unphosphorylated ED is necessary for binding with acidic lipids, nonspecific siRNA served as a control (Dharmacon, Lafayette, CO, USA; primarily phosphoinositides like PIP2, deletion of the ED may result #D-001810-10-20). 5 in a decreased sequestration of PIP2 leading to an activation For transient overexpression of MARCKS, CRC cells (5 × 10 ) were of the PI3K/AKT pathway. Regarding a possible role of MARCKS transfected with an empty control pcDNA3 vector (Stratagene, La Jolla, in receptor exocytosis, a functional ED may further be indis- CA, USA), pcDNA3-MARCKS wt-Flag or pcDNA3-MARCKSΔED-Flag using pensable for proper receptor trafficking. Our results therefore Lipofectamine 2000. indicate that MARCKS-dependent sensitization to TRAIL relies on its ED. Immunoblot analysis In conclusion, the herein presented data substantiate the Immunoblot analysis was carried out as previously described.52 tumor-suppressing functions of MARCKS in MSS CRC. We demonstrate that MARCKS inactivation in MSS colon cancer cell lines confers anti-apoptotic effects and provide evidence that Plasma membrane protein extraction these effects result from the interference of MARCKS with pro- and Cells were processed with the plasma membrane protein extraction kit anti-apoptotic pathways including the AKT signaling and the (BioVision, Zürich, Switzerland) following the manufacturer’s instructions. TRAIL apoptosis pathway. Our data further show that loss of MARCKS expression is a prognostic marker in MSS CRCs. Quantitative PCR analysis Quantitative real-time PCR was performed as described previously.52 For MATERIALS AND METHODS primer sequences, see Supplemental Experimental Procedures. Material The following material was used in this study: soluble human recombinant Association of MARCKS expression in tumor tissue and survival TRAIL (Biomol, Hamburg, Germany), oxaliplatin (Sigma, Deisenhofen, Study design. Patients with CRC from the south-west German population- Germany), propidium iodine (Sigma), 5-FU (AppliChem GmbH, based DACHS study (Darmkrebs, Chancen der Verhütung durch Screening) Darmstadt, Germany), AKT inhibitor IV (Calbiochem, Merck Millipore, were investigated regarding survival. Patients were informed about the Darmstadt, Germany), LY294002 (Biozol, Eching, Germany) and perifosine study by the physicians in charge of their treatment. All patients had a (Cayman Chemicals, Ann Arbor, MI, USA). histologically confirmed first diagnosis of primary CRC and were at least 30 years old, physically and mentally able to participate in an interview, sufficiently proficient in German language, resident in the study region in Cell culture south-western Germany (Rhine-Neckar region) and were treated in 1 of the Human CRC cell lines were cultured in RPMI-1640 medium (GIBCO/ 22 hospitals of the study region. More details about in the DACHS study Invitrogen, Darmstadt, Germany) supplemented with 10% fetal bovine have been reported previously.53,54 serum (PAA Laboratories, Pasching, Austria), 1% penicillin/streptomycin The study was approved by the local ethics committee of the University (GIBCO/Invitrogen) at 37 °C and 5% CO2. Serum starvation was carried out of Heidelberg and the medical boards of Baden-Wuerttemberg and using RPMI-1640 medium supplemented with 1% penicillin/streptomycin Rhineland-Palatinate.

© 2015 Macmillan Publishers Limited Oncogene (2015) 1150 – 1159 MARCKS in colorectal cancer M Bickeböller et al 1158 Data collection and follow-up. The patients provided information during a All analyses were performed with SAS, software version 9.2 (SAS face-to-face interview, which was conducted by trained interviewers. In Institute, Cary, NC, USA). Tests for statistical significance were two-sided addition, discharge letters and pathology reports were collected. On and defined by Po0.05. Hazard ratios and their 95% confidence intervals average 3 years after diagnosis, a questionnaire was sent to the treating were calculated using the SAS procedure PHREG. In both crude and physicians of the patients to collect information on CRC therapy, multivariate models, late entry into the study with potential interim death intermittent diagnoses of concomitant diseases and potential CRC was accounted for. recurrence. For in vitro data, significant differences were identified using the About 5 years after diagnosis, additional information was collected from unpaired two-sided Student's t-test. Throughout, P-values o0.05 were the patients alive, again including questions on newly diagnosed diseases considered significant and are indicated by an asterisk. and recurrences. New diagnoses and cancer recurrences were verified through medical records of the attending physicians. For those alive at 3-year but not at 5-year, follow-up information about recurrence of disease for this period was requested directly from the physicians. Data on vital CONFLICT OF INTEREST status and date of death were obtained from the population registries. The authors declare no conflict of interest. Causes of death were verified by death certificates obtained from the health authorities in the Rhein-Neckar-Odenwald region and coded according to WHO standards. Follow-up time was calculated as the time between the date of ACKNOWLEDGEMENTS diagnosis and the date of event or censoring. Follow-up time of patients We thank the study participants and the interviewers who collected the data. We also without any event of interest (death and recurrence) was censored at the greatly appreciate the help of the hospitals, pathology departments and cooperating fi date of the last follow-up or on 31 December 2011, whichever was rst. institutions in recruiting patients for this study and providing tumor samples. We thank Jutta Richter, Ute Handte-Daub, Bettina Walter, Barbara Schreiber and Marina Collection and processing of tumor samples. Formalin-fixed paraffin- Gernold for their excellent technical assistance. H Bläker has been supported by a embedded samples of the colorectal tumors were collected from the grant from the German Research Council (Deutsche Forschungsgemeinschaft; grant pathology departments of the cooperating clinics and transferred to the number: BL554/3-2). The DACHS study was supported by grants from the German tissue bank of the National Center for Tumor Diseases (NCT) in Research Council (Deutsche Forschungsgemeinschaft, grant numbers BR 1704/6-1, Heidelberg, Germany. Processing of formalin-fixed paraffin-embedded BR 1704/6-3, BR 1704/6-4 and CH 390 117/1-1), the German Federal Ministry of blocks into tissue micro arrays has previously been described for the Education and Research (grant numbers 01KH0404 and 01ER0814), and the 55 DACHS study. Interdisciplinary Research Program of the National Center for Tumor Diseases Immunohistochemistry was performed on 5 μm thick paraffin sections (NCT), Heidelberg, Germany. from the tissue micro array blocks. Slides were incubated with a rabbit monoclonal anti-MARCKS antibody (EP1446Y, Abcam, Cambridge, UK) in a dilution of 1:200 and pre-treated by water bath boiling in DAKO target retrieval buffer, pH 6 (DAKO, Hamburg, Germany). The antibody was REFERENCES incubated for 30 min at room temperature. For visualization purposes of 1 Albert KA, Nairn AC, Greengard P. The 87-kDa protein, a major specific substrate MARCKS, the labeled immunoperoxidase method with AEC as a chromo- for protein kinase C: purification from bovine brain and characterization. Proc Natl gen was applied (DAKO). Acad Sci USA 1987; 84: 7046–7050. MARCKS expression was scored by two pathologists (H Bläker 2 Stumpo DJ, Graff JM, Albert KA, Greengard P, Blackshear PJ. Molecular cloning, and C Toth). Only cases with MARCKS-positive dendritic cells and nerve characterization, and expression of a cDNA encoding the `80- to 87-kDa' fi bers, which were used as an internal control, were scored. Loss of myristoylated alanine-rich C kinase substrate: a major cellular substrate for MARCKS expression was scored when the tumor cells were negative for protein kinase C. Proc Natl Acad Sci USA 1989; 86: 4012–4016. staining and when the intensity of staining in the tumor cells was in strong 3 Seykora JT, Ravetch JV, Aderem A. Cloning and molecular characterization of the contrast to the staining intensity of the internal control cells and could only murine macrophage `68-kDa' protein kinase C substrate and its regulation by fi be detected in × 100 magni cation. Retention of MARCKS expression was bacterial lipopolysaccharide. Proc Natl Acad Sci USA 1991; 88: 2505–2509. fi scored in all cases in which MARCKS expression was identi ed in lower 4 Harlan DM, Graff JM, Stumpo DJ, Eddy RL Jr, Shows TB, Boyle JM et al. The human fi magni cation (25 to 50-fold). Cases were considered noninformative when myristoylated alanine-rich C kinase substrate (MARCKS) gene (MACS). Analysis of the tissue lacked staining of the internal control cells. its gene product, promoter, and chromosomal localization. J Biol Chem 1991; 266: 14399–14405. Microsatellite analysis . Microsatellite status was previously determined in 5 Graff JM, Young TN, Johnson JD, Blackshear PJ. Phosphorylation-regulated the CRCs of the DACHS study. Details about MSI analysis have been calmodulin binding to a prominent cellular substrate for protein kinase C. J Biol 16 reported. In brief, a mononucleotide marker panel (BAT25, BAT26 Chem 1989; 264: 21818–21823. and CAT25) was used to determine MSI-H in DNA extracted from tumor 6 Arbuzova A, Schmitz AA, Vergeres G. Cross-talk unfolded: MARCKS proteins. tissue obtained from tumor block sections. The marker panel used Biochem J 2002; 362:1–12. distinguishes between MSI-H and non-MSI-H tumors with a sensitivity of 7 Laux T, Fukami K, Thelen M, Golub T, Frey D, Caroni P. GAP43, MARCKS, 98.2% and a specificity of 100%, and with 100% concordance of MSI-H and CAP23 modulate PI(4,5)P(2) at plasmalemmal rafts, and regulate cell tumors compared with the National Cancer Institute/International cortex actin dynamics through a common mechanism. J Cell Biol 2000; 149: Collaborative Group on HNPCC (NCI/ICG-HNPCC) marker panel (BAT25, 1455–1472. 56 BAT26, D17S250, D2S123 and D5S346) for the evaluation of MSI in CRC. 8 Brooks G, Brooks SF, Goss MW. MARCKS functions as a novel growth suppressor in cells of melanocyte origin. Carcinogenesis 1996; 17:683–689. Statistical analysis. Overall, MARCKS expression was analyzed in 1260 9 Rose SD, Cook HW, Palmer FB, Ridgway ND, Byers DM. Differential expression of patients with available tissue micro array blocks and follow-up information. MARCKS and other calmodulin-binding protein kinase C substrates in cultured In 283 tumors, staining was not informative, leaving 977 patients with a neuroblastoma and glioma cells. J Neurochem 1994; 63:2314–2323. negative or positive result on MARCKS for the survival analyses (patient 10 Joseph CK, Qureshi SA, Wallace DJ, Foster DA. MARCKS protein is transcriptionally description in Supplementary Table 1). Information on MSI was available down-regulated in v-Src-transformed BALB/c 3T3 cells. J Biol Chem 1992; 267: for 926 of these patients (96 MSI-H and 830 MSS cancers). Associations of 1327–1330. MARCKS expression with T, N and M status were tested in all tumors, MSS 11 Wolfman A, Wingrove TG, Blackshear PJ, Macara IG. Down-regulation of protein tumors and MSI-high tumors, respectively (Supplementary Table 2). kinase C and of an endogenous 80-kDa substrate in transformed fibroblasts. J Biol Using Cox proportional hazards regression models, we estimated crude Chem 1987; 262: 16546–16552. and adjusted hazard ratios and their 95% confidence intervals of overall, 12 Simek SL, Kligman D, Patel J, Colburn NH. Differential expression of an 80-kDa CRC-specific and recurrence-free survival after a median follow-up time protein kinase C substrate in preneoplastic and neoplastic mouse JB6 cells. of 5.0 years, comparing negative MARCKS expression with positive Proc Natl Acad Sci USA 1989; 86: 7410–7414. expression. In the adjusted analyses, we included age at diagnosis, sex, 13 Michel S, Kloor M, Singh S, Gdynia G, Roth W, von Knebel Doeberitz M et al. CRC stage, location of the tumor (colon and rectum), chemotherapy, Coding microsatellite instability analysis in microsatellite unstable small intestinal neoadjuvant treatment and a time-dependent effect of chemotherapy adenocarcinomas identifies MARCKS as a common target of inactivation. (chemotherapy*log(time)). Mol Carcinog 2010; 49:175–182.

Oncogene (2015) 1150 – 1159 © 2015 Macmillan Publishers Limited MARCKS in colorectal cancer M Bickeböller et al 1159 14 Woerner SM, Kloor M, Mueller A, Rueschoff J, Friedrichs N, Buettner R et al. 35 Johnstone RW, Frew AJ, Smyth MJ. The TRAIL apoptotic pathway in cancer onset, Microsatellite instability of selective target genes in HNPCC-associated colon progression and therapy. Nat Rev Cancer 2008; 8: 782–798. adenomas. Oncogene 2005; 24:2525–2535. 36 Honing S, Ricotta D, Krauss M, Spate K, Spolaore B, Motley A et al. 15 Kim NG, Rhee H, Li LS, Kim H, Lee JS, Kim JH et al. Identification of MARCKS, Phosphatidylinositol-(4,5)-bisphosphate regulates sorting signal recognition by FLJ11383 and TAF1B as putative novel target genes in colorectal carcinomas with the clathrin-associated adaptor complex AP2. Mol Cell 2005; 18: 519–531. microsatellite instability. Oncogene 2002; 21: 5081–5087. 37 Su R, Han ZY, Fan JP, Zhang YL. A possible role of myristoylated alanine-rich C 16 Hoffmeister M, Blaker H, Kloor M, Roth W, Toth C, Herpel E et al. Body mass index kinase substrate in endocytic pathway of Alzheimer's disease. Neurosci Bull 2010; and microsatellite instability in colorectal cancer: a population-based study. 26:338–344. Cancer Epidemiol Biomarkers Prev 2013; 22: 2303–2311. 38 Holz RW, Hlubek MD, Sorensen SD, Fisher SK, Balla T, Ozaki S et al. A 17 Koh KH, Kang HJ, Li LS, Kim NG, You KT, Yang E et al. Impaired nonhomologous pleckstrin homology domain specific for phosphatidylinositol 4, 5-bisphosphate end-joining in mismatch repair-deficient colon carcinomas. Lab Invest 2005; 85: (PtdIns-4,5-P2) and fused to green fluorescent protein identifies plasma 1130–1138. membrane PtdIns-4,5-P2 as being important in exocytosis. J Biol Chem 2000; 275: 18 Shin N, You KT, Lee H, Kim WK, Song M, Choi HJ et al. Identification of frequently 17878–17885. mutated genes with relevance to nonsense mediated mRNA decay in the high 39 Zhang XD, Franco AV, Nguyen T, Gray CP, Hersey P. Differential localization and microsatellite instability cancers. Int J Cancer 2011; 128: 2872–2880. regulation of death and decoy receptors for TNF-related apoptosis-inducing 19 Graff JM, Stumpo DJ, Blackshear PJ. Characterization of the phosphorylation sites ligand (TRAIL) in human melanoma cells. J Immunol 2000; 164: 3961–3970. in the chicken and bovine myristoylated alanine-rich C kinase substrate protein, 40 Bennett M, Macdonald K, Chan SW, Luzio JP, Simari R, Weissberg P. Cell surface a prominent cellular substrate for protein kinase C. J Biol Chem 1989; 264: trafficking of Fas: a rapid mechanism of p53-mediated apoptosis. Science 1998; 11912–11919. 282:290–293. 20 Sundaram M, Cook HW, Byers DM. The MARCKS family of phospholipid binding 41 Jones SJ, Ledgerwood EC, Prins JB, Galbraith J, Johnson DR, Pober JS et al. TNF proteins: regulation of phospholipase D and other cellular components. Biochem recruits TRADD to the plasma membrane but not the trans-Golgi network, the Cell Biol 2004; 82: 191–200. principal subcellular location of TNF-R1. J Immunol 1999; 162: 1042–1048. 21 Estrada-Bernal A, Gatlin JC, Sunpaweravong S, Pfenninger KH. Dynamic adhesions 42 Nie Z, Hirsch DS, Luo R, Jian X, Stauffer S, Cremesti A et al. A BAR domain in the N and MARCKS in melanoma cells. J Cell Sci 2009; 122: 2300–2310. terminus of the Arf GAP ASAP1 affects membrane structure and trafficking of 22 Jarboe JS, Anderson JC, Duarte CW, Mehta T, Nowsheen S, Hicks PH et al. MARCKS epidermal growth factor receptor. Curr Biol 2006; 16: 130–139. regulates growth and radiation sensitivity and is a novel prognostic factor 43 Simova S, Klima M, Cermak L, Sourkova V, Andera L. Arf and Rho GAP adapter for glioma. Clin Cancer Res 2012; 18: 3030–3041. protein ARAP1 participates in the mobilization of TRAIL-R1/DR4 to the plasma 23 Jin Cho S, La M, Ahn JK, Meadows GG, Joe CO. Tob-mediated cross-talk between membrane. Apoptosis 2008; 13:423–436. MARCKS phosphorylation and ErbB-2 activation. Biochem Biophys Res Commun 44 Miura K, Jacques KM, Stauffer S, Kubosaki A, Zhu K, Hirsch DS et al. ARAP1: a point 2001; 283: 273–277. of convergence for Arf and Rho signaling. Mol Cell 2002; 9: 109–119. 24 Techasen A, Loilome W, Namwat N, Takahashi E, Sugihara E, Puapairoj A et al. 45 Morton LA, Yang H, Saludes JP, Fiorini Z, Beninson L, Chapman ER et al. Myristoylated alanine-rich C kinase substrate phosphorylation promotes MARCKS-ED peptide as a curvature and lipid sensor. ACS Chem Biol 2013; 8: cholangiocarcinoma cell migration and metastasis via the protein kinase 218–225. C-dependent pathway. Cancer Sci 2010; 101:658–665. 46 Peuhu E, Rivero-Muller A, Stykki H, Torvaldson E, Holmbom T, Eklund P et al. 25 Zhu Y, Dong Q, Tan BJ, Lim WG, Zhou S, Duan W. The PKCalpha-D294G mutant Inhibition of Akt signaling by the lignan matairesinol sensitizes prostate cancer found in pituitary and thyroid tumors fails to transduce extracellular signals. cells to TRAIL-induced apoptosis. Oncogene 2010; 29:898–908. Cancer Res 2005; 65: 4520–4524. 47 Graff JM, Rajan RR, Randall RR, Nairn AC, Blackshear PJ. Protein kinase C substrate 26 Yamaguchi T, Iijima T, Mori T, Takahashi K, Matsumoto H, Miyamoto H et al. and inhibitor characteristics of peptides derived from the myristoylated alanine- Accumulation profile of frameshift mutations during development and rich C kinase substrate (MARCKS) protein phosphorylation site domain. J Biol progression of colorectal cancer from patients with hereditary nonpolyposis Chem 1991; 266: 14390–14398. colorectal cancer. Dis Colon Rectum 2006; 49:399–406. 48 Li Y, Martin LD, Spizz G, Adler KB. MARCKS protein is a key molecule regulating 27 Rombouts K, Carloni V, Mello T, Omenetti S, Galastri S, Madiai S et al. Myristoylated mucin secretion by human airway epithelial cells in vitro. J Biol Chem 2001; 276: alanine-rich protein kinase C substrate (MARCKS) expression modulates the 40982–40990. metastatic phenotype in human and murine colon carcinoma in vitro and in vivo. 49 Morash SC, Douglas D, McMaster CR, Cook HW, Byers DM. Expression of MARCKS Cancer Lett 2013; 333:244–252. effector domain mutants alters phospholipase D activity and cytoskeletal mor- 28 Chen X, Rotenberg SA. PhosphoMARCKS drives motility of mouse melanoma cells. phology of SK-N-MC neuroblastoma cells. Neurochem Res 2005; 30: 1353–1364. Cell Signal 2010; 22: 1097–1103. 50 Castro F, Dirks WG, Fahnrich S, Hotz-Wagenblatt A, Pawlita M, Schmitt M. 29 Myat MM, Anderson S, Allen LA, Aderem A. MARCKS regulates membrane ruffling High-throughput SNP-based authentication of human cell lines. Int J Cancer 2013; and cell spreading. Curr Biol 1997; 7:611–614. 132:308–314. 30 Wang J, Gambhir A, Hangyas-Mihalyne G, Murray D, Golebiewska U, McLaughlin S. 51 Schmitt M, Pawlita M. High-throughput detection and multiplex identification of Lateral sequestration of phosphatidylinositol 4,5-bisphosphate by the basic cell contaminations. Nucleic Acids Res 2009; 37: e119. effector domain of myristoylated alanine-rich C kinase substrate is due to non- 52 Fassl A, Tagscherer KE, Richter J, Berriel Diaz M, Alcantara Llaguno SR, Campos B specific electrostatic interactions. J Biol Chem 2002; 277: 34401–34412. et al. Notch1 signaling promotes survival of glioblastoma cells via EGFR-mediated 31 Glaser M, Wanaski S, Buser CA, Boguslavsky V, Rashidzada W, Morris A induction of anti-apoptotic Mcl-1. Oncogene 2012; 31: 4698–4708. et al. Myristoylated alanine-rich C kinase substrate (MARCKS) produces 53 Brenner H, Chang-Claude J, Seiler CM, Rickert A, Hoffmeister M. Protection from reversible inhibition of phospholipase C by sequestering phosphatidylinositol colorectal cancer after colonoscopy: a population-based, case-control study. 4,5-bisphosphate in lateral domains. J Biol Chem 1996; 271: 26187–26193. Ann Intern Med 2011; 154:22–30. 32 Smyth MJ, Cretney E, Takeda K, Wiltrout RH, Sedger LM, Kayagaki N et al. Tumor 54 Lilla C, Verla-Tebit E, Risch A, Jager B, Hoffmeister M, Brenner H et al. Effect of necrosis factor-related apoptosis-inducing ligand (TRAIL) contributes to interferon NAT1 and NAT2 genetic polymorphisms on colorectal cancer risk associated with gamma-dependent natural killer cell protection from tumor metastasis. J Exp Med exposure to tobacco smoke and meat consumption. Cancer Epidemiol Biomarkers 2001; 193: 661–670. Prev 2006; 15:99–107. 33 Takeda K, Smyth MJ, Cretney E, Hayakawa Y, Kayagaki N, Yagita H et al. Critical 55 Rudolph A, Toth C, Hoffmeister M, Roth W, Herpel E, Jansen L et al. Expression of role for tumor necrosis factor-related apoptosis-inducing ligand in immune oestrogen receptor beta and prognosis of colorectal cancer. Br J Cancer 2012; 107: surveillance against tumor development. J Exp Med 2002; 195: 161–169. 831–839. 34 Kriegl L, Jung A, Horst D, Rizzani A, Jackstadt R, Hermeking H et al. Microsatellite 56 Findeisen P, Kloor M, Merx S, Sutter C, Woerner SM, Dostmann N et al. T25 repeat instability, KRAS mutations and cellular distribution of TRAIL-receptors in early in the 3' untranslated region of the CASP2 gene: a sensitive and specific marker stage colorectal cancer. PLoS One 2012; 7: e51654. for microsatellite instability in colorectal cancer. Cancer Res 2005; 65: 8072–8078.

Supplementary Information accompanies this paper on the Oncogene website (http://www.nature.com/onc)