Containing Truncated Mutant Proteins in the Tumors with High Microsatellite Instability

Published OnlineFirst May 14, 2013; DOI: 10.1158/1078-0432.CCR-13-0684

Clinical Cancer Human Cancer Biology Research

Identiﬁcation and Selective Degradation of Neopeptide- Containing Truncated Mutant Proteins in the Tumors with High Microsatellite Instability

Won Kyu Kim1, Misun Park1, Minhee Park1, Yun Ji Kim1, Nara Shin1, Hyun Ki Kim1, Kwon Tae You2, and Hoguen Kim1

Abstract Purpose: Frameshift mutations in coding mononucleotide repeats (cMNR) are common in tumors with high microsatellite instability (MSI-H). These mutations generate mRNAs containing abnormal coding sequences and premature termination codons (PTC). Normally, mRNAs containing PTCs are degraded by nonsense-mediated mRNA decay (NMD). However, mRNAs containing PTCs located in the last exon are not subject to degradation by NMD (NMD-irrelevant). This study aimed to discover whether genes with frameshift mutations in the last exon generate truncated mutant proteins. Experimental Design: We identified 66 genes containing cMNRs in the last exon by bioinformatic analysis. We found frequent insertion/deletion mutations in the cMNRs of 29 genes in 10 MSI-H cancer cell lines and in the cMNRs of 3 genes in 19 MSI-H cancer tissues. We selected 7 genes (TTK, TCF7L2, MARCKS, ASTE1, INO80E, CYHR1, and EBPL) for mutant mRNA expression analysis and 3 genes (TTK, TCF7L2, and MARCKS) for mutant protein expression analysis. Results: The PTC-containing NMD-irrelevant mRNAs from mutated genes were not degraded. However, only faint amounts of endogenous mutant TTK and TCF7L2 were detected, and we failed to detect endogenous mutant MARCKS. By polysome analysis, we showed that mRNAs from genomic mutant MARCKS constructs are normally translated. After inhibiting 3 protein degradation pathways, we found that only inhibition of the proteasomal pathway facilitated the rescue of endogenous mutant TTK, TCF7L2, and MARCKS. Conclusions: Our findings indicate that cancer cells scavenge potentially harmful neopeptide-containing mutant proteins derived from NMD-irrelevant abnormal mRNAs via the ubiquitin–proteasome system, and these mutant proteins may be important substrates for tumor-specific antigens. Clin Cancer Res; 19(13); 3369–82. 2013 AACR.

Introduction tions inactivate genes by introducing premature termina- Activating mutations in oncogenes and inactivating tion codons (PTC) in mRNA. If mRNAs bearing PTCs are mutations in tumor suppressor genes are hallmarks of not degraded and allowed to be translated normally, then human cancers (1). The common inactivation pathways of genes with nonsense mutations are expected to generate tumor suppressor genes are deletion of one allele and truncated proteins lacking neopeptides, whereas genes with inactivating mutations of the other allele (2). Of these frameshift mutations are expected to generate truncated pathways, nonsense and frameshift mutations are the most proteins containing neopeptides. Generally, most abnor- critical inactivating mutations (3). These 2 types of muta- mal PTC-containing mRNAs are actively degraded by nonsense-mediated mRNA decay (NMD), thus avoiding the potentially deleterious effects associated with the Authors' Afﬁliations: 1Department of Pathology and Brain Korea 21 production of truncated proteins (4, 5). NMD is mediated Projects for Medical Science, Yonsei University College of Medicine; and through the recognition of PTC-containing mRNAs, 2School of Biological Science and Center for National Creative Research, Seoul National University, Seoul, Republic of Korea which are recognized by their position relative to the last exon–exon junction. Mammalian transcripts that contain Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/). PTCs more than 50 to 55 nucleotides (nt) upstream of the last exon–exon junction are degraded by NMD, which Corresponding Author: Hoguen Kim, Yonsei University College of Med- icine, 134 Shinchon-Dong Seodaemun-Gu, Seoul 120-752, Republic of ensures the degradation of most PTC-containing mRNAs. Korea. Phone: 82-2-2228-1761; Fax: 82-2-363-5263; E-mail: However, PTCs located within 50 to 55 nt or downstream [email protected] of the last exon–exon junction are not recognized by doi: 10.1158/1078-0432.CCR-13-0684 NMD and can potentially lead to the generation of 2013 American Association for Cancer Research. mutant proteins (6, 7).

www.aacrjournals.org 3369

Kim et al.

National Research Resource Bank Program of the Korea Translational Relevance Science and Engineering Foundation of the Ministry of Abnormal mRNAs containing premature termina- Science and Technology. Authorization for the use of the tion codon (PTC) derived from frameshift mutation tissues for research was obtained from the Institutional in the last exon are not subject to degradation by Review Board. Conventional pathologic parameters were the nonsense-mediated mRNA decay (NMD) system. examined without prior knowledge of the molecular data These NMD-irrelevant mRNAs are expected to generate (Table 1). neopeptide-containing truncated mutant proteins. We identified mutant genes containing PTC derived from Identification of MSI and mutation analysis the last exon in colon cancers with high microsatellite Genomic DNA and cDNA preparation, analysis of MSI, instability. We showed that the NMD-irrelevant mutant and identification of target gene mutations were conducted mRNAs from the identified genes are normally trans- using a PCR-based assay as described previously (16). lated, but the neopeptide-containing truncated mutant proteins are selectively degraded by the ubiquitin– Semiquantitative RT-PCR and qRT-PCR proteasome system. Our results indicate that mutant The primers for semiquantitative reverse transcription proteins generated from NMD-irrelevant mRNAs can (RT)-PCR and quantitative reverse transcription PCR contribute to the formation of tumor-specific antigens (qRT-PCR) were designed using Primer 3 database and these antigens can be used as specific targets for (http://frodo.wi.mit.edu/primer3/). All RNAs were isolated immunotherapy. from cells using TRIzol (Invitrogen). Reverse transcription was conducted using M-MLV reverse transcriptase (Invitro- gen). For RT-PCR, the reaction was conducted using Ampli- Taq Gold 360 DNA Polymerase (Applied Biosystems). For A subset of colorectal tumors exhibits length alterations qRT-PCR, the reaction was conducted using the ABI PRISM in several coding and noncoding microsatellites, a molec- 7500 Sequence Detector (Applied Biosystems) and SYBR ular phenotype termed high microsatellite instability (MSI- Premix Ex Taq II (TaKaRa). The amount of target mRNA was H; refs. 8, 9). The length alterations in microsatellites of the normalized to that of GAPDH or EGFP mRNA [derived from coding region [coding mononucleotide repeats (cMNRs)] the enhanced GFP (EGFP)-expressing control vector]. The result in frameshift mutations in the affected genes, and sequences of the primers used are listed in Supplementary these mutations are believed to contribute to tumor devel- Table S1. opment and progression (10). Although many reports indicated that numerous genes are frequently mutated in Construction of TTK, TCF7L2, and MARCKS expression their cMNRs in MSI-H cancers, few of these genes have been vectors, RNAi, and transfection reported to express their mutant gene products in MSI-H cDNA expression vectors for TTK [cTTK(WT)], TCF7L2 cancers (11–13). We searched for genes containing cMNRs [cTCF4L2(WT)], and MARCKS [cMARCKS(WT)] were con- in the last exon and analyzed the frequency of mutations in structed by cloning the respective wild-type (WT) genes into these genes. We addressed whether genes with frameshift pcDNA3.1 vectors containing a FLAG tag via amplification mutations generate truncated mutant proteins. We showed of their coding regions using the cDNAs derived from HeLa that mutant proteins are actively translated from genes and WiDr cells. For the genomic DNA form of the MARCKS containing mutations in cMNRs in the last exon but are expression vector [gMARCKS(WT)], all exons and introns rarely detected because these endogenous truncated pro- between the exons of MARCKS were cloned into pcDNA3.1 teins containing neopeptides are extensively degraded by vectors containing a FLAG tag. To generate mutant protein the ubiquitin–proteasome system. expression vectors for TTK [cTTK(2)], TCF7L2 [cTCF7L2 (1)], and MARCKS [cMARCKS(2), gMARCKS(2)], Materials and Methods deletion mutagenesis was conducted. All transfection Tissue samples and cell lines experiments were carried out using Lipofectamine 2000 For the mutation analysis, 12 cell lines were used. DLD1, (Invitrogen), and a CMV10-EGFP vector was used to con- HCT116, HCT-8, LOVO, LS174T, RKO, SNUC2A, SNUC2B, firm the transfection efficiency. The primers used for cloning SNUC4, and SNU407 cells are MSI-H colorectal carcinoma are listed in Supplementary Table S1. The siRNAs against cell lines, whereas WiDr and HeLa cells are microsatellite- TTK, TCF7L2, and MARCKS used in this study (Bioneer) stable (MSS) cell lines, as determined by previous studies were also transfected into cells using Lipofectamine. (14, 15). Cells were grown in RPMI, minimum essential medium, and Dulbecco’s modified Eagle medium supple- Western blotting and mutant protein-specific antibody mented with 10% FBS (Life Technologies), 1% penicillin, generation and streptomycin at 37 Cin5%CO2. About tissue samples, Whole lysates from cells were prepared using passive 19 specimens confirmed as MSI-H colorectal carcinomas lysis buffer (Promega). Membranes were incubated with using BAT25, BAT26, D5S346, D17S25, and D2S123 were primary antibodies against glyceraldehyde-3-phosphate included in this study. Some of the fresh specimens were dehydrogenase (GAPDH; Trevigen), FLAG (Sigma-Aldrich), obtained from the Liver Cancer Specimen Bank of the TTK (Abnova), TCF7L2 (Cell Signaling Technology), and

3370 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Degradation of Truncated Mutants

Table 1. Clinicopathologic features of 19 MSI-H colon cancer tissues and mutation proﬁles of 3 genes

Peritumoral Mutation status Case Age at Anatomic Tumor lymphoid number Sex diagnosis site T N M Stage differentiation reaction TTK TCF7L2 MARCKS 1 F 83 Ascending 3 0 0 II MD 1a www 2 M 75 Sigmoid 3 1 1 IV PD 1 w w w 3 M 71 Ascending 3 0 0 II MD 2 1/w w 1/w 4 F 38 Ascending 3 2 0 III MD 2 w w 1/w 5 M 73 Transverse 3 0 0 II PD 1 w w w 6 F 41 Ascending 3 0 0 II WD 1 w w w 7 F 70 Ascending 3 0 0 II PD 3 w w 1/w 8 M 71 Ascending 3 0 0 II MD 1 w w w 9 M 60 Ascending 3 1 0 III MD 2 1/w w 1/w 10 M 47 Ascending 4 2 1 IV MD 2 1/w 1/w w 11 F 72 Ascending 3 0 0 II MD 2 1/w 1/w 1/w 12 M 52 Ascending 3 0 0 II PD 2 1/w 1/w 1 13 M 47 Ascending 2 0 0 I WD 2 1/w 1/w w 14 M 32 Rectum 2 0 0 I MD 2 1/w w 1/w 15 F 71 Ascending 3 0 0 II PD 2 w 1/w w 16 F 58 Sigmoid 3 0 0 II PD 2 w w 1/w 17 F 38 Ascending 1 0 0 I PD 3 1/w 1/w 1 18 M 55 Ascending 3 0 0 II PD 3 1/w w 1/w 19 F 62 Descending 3 0 0 II MD 3 1/w 1/w þ1/w

Abbreviations: MD, moderate differentiation; PD, poor differentiation; WD, well differentiation. a1, Absent; 2, mild; 3, intense.

MARCKS (Santa Cruz Biotechnology) for 1 hour at room serum albumin for 1 hour. After blocking, the medium was temperature. Antibodies against the neopeptide sequences replaced with the respective primary antibodies, and cells of mutant MARCKS(2) were generated in rabbits. All were incubated overnight. Cells were then washed and antibody generation procedures were conducted according incubated for 1 hour with the appropriate fluorescently to the manufacturer’s manuals (Young In Frontier). More labeled secondary antibodies. For double-labeling experi- detailed information on the antibody generation is includ- ments, cells were simultaneously incubated with the pri- ed on the website (http://www.molpathol.org/). mary and secondary antibodies. Anti-FLAG M2-FITC (Sig- ma; F4049) and anti-g-tubulin (Sigma; T5192) were used in Polysome assay this experiment. All images were obtained using an LSM700 HeLa cells were washed 3 times with ice-cold PBS contain- confocal microscope (Carl Zeiss). ing 100 mg/mL cycloheximide, collected, and then lysed in lysis buffer [15 mmol/L Tris–Cl (pH 7.4), 3 mmol/L MgCl2, Immunoprecipitation and ubiquitination assay 10 mmol/L NaCl, 0.5% Triton X-100, 100 mg/mL cyclohex- Immune complexes of wild-type and mutant proteins imide, 1 mg/mL heparin, and 200 U RNasin (Promega)]. were collected by gently rocking 1 mg of total proteins on an Nuclei and debris were removed by centrifugation at 12,000 orbital shaker with prewashed anti-FLAG M2-agarose affin- g for 2 minutes. One milliliter of each cytoplasmic lysate ity gel (Sigma-Aldrich) at 4C. The immune complexes was layered onto an 10% to 50% sucrose gradient and bound to the affinity gel were washed and then boiled with centrifuged at 4C (39,000 rpm) for 2 hours. Sixteen frac- a 100 mmol/L Tris–HCl–1% SDS solution to elute the tions were collected with concomitant measurements of complexes. Western blotting was conducted using FLAG absorbance at 254 nm by using a fraction collection system. and HA (Santa Cruz Biotechnology) antibodies. The relative RNA was extracted with TRIzol and analyzed by RT-PCR. density of each lane was quantified by ImageJ (NIH, Bethesda, MD) software. Immunofluorescence microscopic examination Subcellular localization of mutant MARCKS was ana- Results lyzed by immunofluorescence staining. The cells attached Systematic search for the genes containing PTC in the to glass coverslips were rinsed with PBS followed by fixation last exon and permeabilization with ice-cold methanol for 10 min- We searched for human genes containing cMNRs longer utes at 20C. Upon the removal of methanol, cells were than 9 nt in SelTarbase (http://www.seltarbase.org), a data- again rinsed. Nonspecific sites were blocked with 2% bovine base providing comprehensive information about human

www.aacrjournals.org Clin Cancer Res; 19(13) July 1, 2013 3371

Kim et al.

Figure 1. Mutation frequencies of the 35 genes containing cMNRs in MSI-H colon cancer cell lines and tumor tissues. A, a pipeline for selecting candidate genes that potentially generate mutant proteins. B, an example of the mutation analysis of TTK, TCF7L2,andMARCKS using a PCR-based assay and sequencing. Gel mobility shifts in the cells with 1-bp insertion (!!), 1-bp deletion (!), and 2-bp deletions (~) are evident (left). The DNA products displaying mobility shifts were conﬁrmed as 1-bp insertions, 1-bp deletion, or 2-bp deletions in cMNR regions by sequence analysis. C, the mutation frequencies of the 35 genes were analyzed in each cell line. The mean number of mutated genes in each cell line was 14.8 5.43, and the mutation frequencies of 29 of these genes ranged from 10% to 60%. D, frequencies of mutated genes in 10 MSI-H cancer cell lines. E, the mutation frequencies in the cMNRs of 3 genes (TTK, TCF7L2, and MARCKS) in 19 MSI-H colon cancer tissues.

mononucleotide microsatellite mutations and genes con- reported to have mutations in their cMNRs in MSI-H cancers taining cMNRs. In total, 447 genes satisfying these criteria (Fig. 1A; http://www.sanger.ac.uk/genetics/CGP; refs. 11, were obtained from SelTarbase. To confirm that these 447 12). Most of the abnormal stop codons were PTCs; however, genes include cMNRs longer than 9 nt, we analyzed the abnormal stop codons after normal stop codons (read- entire genes using bioinformatic tools such as Vector NTI through) were also found in some genes with cMNR muta- (Invitrogen), the Human BLAST database (http://genome. tions. Many of the 66 genes were related to biologically ucsc.edu/), and the National Center for Biotechnology critical reactions, such as apoptosis, cell-cycle regulation, Information gene database. With this approach, 302 of cell proliferation, angiogenesis, and intracellular signaling 447 genes were confirmed to have cMNRs longer than 9 (Supplementary Table S2). nt. On the basis of the NMD-irrelevant condition (mRNAs containing PTCs within 50–55 nt of the last exon junction), Mutation profile of the genes containing cMNRs in the 302 genes were manually analyzed using Vector NTI soft- last exon in MSI-H cancer cell lines and tissues ware. When 1- or 2-bp deletions/insertions were detected in We randomly selected 35 of 66 genes to analyze cMNR the cMNR region, genes that acquired abnormal stop mutations by conducting an isotope PCR-based assay and codons distal to a site 50 to 55 nt from the last exon junction sequencing. We used 10 MSI-H colon cancer cell lines for complex were selected. The number of finally selected genes the mutation search. A MSS colon cancer cell line (WiDr) was 66. Among them, 15 genes had previously been and the HeLa cell line were used as controls. We found

3372 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Degradation of Truncated Mutants

Table 2. Mutation proﬁles of 35 genes containing cMNRs in the last exon

MSS cancer MSI-H cancer cell line cell line

Gene DLD1 HCT116 HCT8 LOVO LS174T RKO SNUC2A SNUC2B SNUC4 SNU407 WiDr HeLa Aim2 1/wa 1/w w 1/1 1/w þ1/1 1/2 1/2w www ASTE1 þ1/w 1/w þ1/w w w 2/2 2/2 2/2 1/1 2/2ww ASH1L wwwwww1/1 1/1 1/w w w w ANKRD49 ww1/w w w w w w w w w w BEND5 wwww1/w w w w w w w w CCDC43 www1/w 1/w w 1/w 1/w w þ1/w w w CCT8L1 w 1/2w1/2 1/2 1/2 2/2 2/2 þ1/3 2/w w w CIR1 wwwwww w w w www CYHR1 wwww1/1 1/w 1/w 1/w þ1/þ2 1/w w w EBPL w þ2/w w 1/þ1w1/w w 1/w w w w w ERCC5 wwwwww w w w www FAM111B 1/w 1/w w 1/1 1/w 1/1 2/w þ1/2 1/1 1/2ww FBXL3 wwwwww w w1/w w w w FLT3LG wwwþ1/w w þ1/w þ1/w w þ1/w w w w GAFA1 w 2/2w2/2 2/2 2/2 2/3 2/3 1/1 1/1ww GBP3 1/w 1/1 1/w w 1/w 1/w w w 1/w w w w GINS1 wwwww1/w w w 1/w w w w HOXA11 wwwwww w w w www INO80E 1/w w 1/w 1/1 1/1 1/1 1/w 1/w 1/1 1/w w w KCTD16 w 1/w w w 1/w w 1/w 1/w w 1/w w w KIAA2018 wwwwww w w w1/w w w LOC643677 wwwwww w w w www LOC100127950 www1/w 1/w w þ1/w 1/w 1/w 1/w w w LOC100128175 wwwwww w w w www LOC100131089 w 1/1w wþ1/þ1 1/1 1/1 1/w 1/w w w w MED8 www1/w 2/w w 1/w 1/w 1/w w w w MARCKS w 1/w w 1/1 1/1 1/w 2/w 2/w 1/1 1/1ww RGS22 www1/w w 1/w w w w w w w RXFP2 wwwwww w w w www SYCP1 1/w w 1/w þ1/þ1w w1/1 1/1 1/1 1/w w w SLAMF1 w 1/w w 1/w w w w w 1/w 1/w w w SFRS12IP1 wwww1/2wþ1/þ1w wþ1/þ1ww TRIM59 www1/w 1/w 1/1 1/w 1/w w 2/2ww TCF7L2 www1/w 1/w w 1/w 1/w 1/1 1/1ww TTK w 1/w w 1/w w 2/w 2/w 2/w þ1/2w ww

aMutation status of each gene: 1, a 1-bp deletion in the cMNR; w, no mutation in the cMNR; þ1, 1-bp insertion in the cMNR. frequent frameshift mutations in the cMNRs of 29 genes displayed frequent and varying mutation incidences rang- in the 10 MSI-H cell lines. The mutation proﬁles and ing from approximately 35% to 60% (Fig. 1E). Informa- status (homozygous vs. heterozygous) are summarized tion about the tissue samples used in this study is pro- in Table 2. To validate the PCR-based mutation analysis, vided in Table 1. we conducted sequence analysis of several genes in the cell lines with homozygous mutations (Fig. 1B). All 10 Expression of mutant mRNAs from genes containing MSI-H cancer cell lines had mutations in more than 4 of PTCs in the last exon 35 genes examined, and 19 mutated genes were found in For the mutant protein expression analysis, we ﬁrstly SNUC2A and SNUC2B cells (Fig. 1C). The mutation selected 7 genes (TTK, TCF7L2, MARCKS, ASTE1, INO80E, frequency of the 29 genes varied from 10% to 90% in CYHR1, and EBPL) displaying frameshift mutations in MSI- the 10 MSI-H cell lines (Fig. 1D). We selected TTK, H cancer cell lines according to the availability of antibo- TCF7L2,andMARCKS and conducted PCR-based muta- dies, presence of homozygous mutations, and cancer rele- tion analysis in 19 MSI-H colon cancer tissues. All 3 genes vance. Before Western blot analysis, the mRNA expression

www.aacrjournals.org Clin Cancer Res; 19(13) July 1, 2013 3373

Kim et al.

of each gene was measured by qRT-PCR and semiquanti- TTK, TCF7L2, and MARCKS showed excellent sensitivity tative RT-PCR analysis. From these experiments, we sought and specificity for the experiments to follow. Then, we to confirm that the mRNAs from mutated genes are not analyzed the expression of endogenous mutant TTK, decayed by NMD or downregulated by other genetic events, TCF7L2, and MARCKS in 9 cell lines (7 MSI-H colon cancer such as deletion or methylation. The mRNA expression of cell lines, a MSS colon cancer cell line, and the HeLa cell the selected genes was similarly quantified in the cell lines line) by Western blotting. There was a minimal size differ- with homozygous mutations and compared with their ence between the mutant and normal proteins generated levels in cells without mutations or those with heterozygous from TTK, whereas considerable size differences were mutations in one allele (Supplementary Fig. S1). SNUC4 observed in the normal and mutant proteins generated cells, a MSI-H cell line with homozygous mutations in TTK from TCF7L2 and MARCKS (Fig. 2D). We expected that if (þ1/2), TCF7L2 (1/1), and MARCKS (1/1), exhib- mutant proteins are expressed, then heterozygous mutation ited relatively higher TTK, TCF7L2, and MARCKS mRNA will cause quantitative and qualitative differences in the levels than the other cell lines (Fig. 2A–C). These results proteins and homozygous mutations will cause qualitative indicate that mRNAs containing PTCs in the last exon are differences in the proteins because no wild-type proteins intact in cells irrespective of the mutation status of genes. can be generated. To confirm our hypothesis, we quantified the normal TTK, TCF7L2, and MARCKS levels in each cell Analysis of endogenous truncated mutant proteins line based on their mutation status. Because of the hetero- from mRNAs containing PTCs in the last exon zygous mRNA expression level in each cell line, we calcu- We tested all the antibodies against the selected 7 genes lated the amount of normal protein by normalizing the by Western blotting, and found that only antibodies against protein level to the mRNA level, and then mean values were

Figure 2. Measurement of mRNA and protein expressions of endogenous mutant TTK, TCF7L2, and MARCKS.A–C, to precisely compare the mRNA expression levels of TTK, TCF7L2, and MARCKS, qRT-PCR was conducted using 10 MSI-H cell lines and 2 MSS cell lines. D, schematic diagram of TTK, TCF7L2, and MARCKS. The cDNA structure of each gene is presented using relative nucleotide numbers. The number and type of cMNR are represented as (A9) and (A11). E, Western blotting was conducted using 7 MSI-H cancer cell lines and 2 MSS cancer cell lines with antibodies against TTK, TCF7L2, and MARCKS. The expected sizes of wild type and mutant protein are marked by black and gray arrows, respectively. Trace amounts of TTK mutant were detected in SNUC4 cells with homozygous TTK mutations. Some faint bands consistent with the expected size of the TCF7L2 mutant were detected only in the cell lines with heterozygous (LS174T) and homozygous 1-bp deletions (SNUC4 and SNU407). MARCKS-mutant proteins were not detected at the expected size. , the mRNA expression level in the cell lines with homozygous mutation. 1, a 1-bp deletion in the cMNR; w, no mutation in the cMNR; þ1, a 1-bp insertion in the cMNR; and 2, the 2-bp deletions in the cMNR.

3374 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Degradation of Truncated Mutants

obtained depending on the mutation status. The result Endogenous mutant TTK, TCF7L2, and MARCKS are revealed that the normal protein level decreased accord- generated, but mostly degraded by the proteasome ing to the mutation status (wild-type > heterozygous system mutation > homozygous mutations; Supplementary Fig. According to polysome analysis, we suspected that S2). For each mutant protein, cell lines with no mutations mutant proteins are generated but mostly degraded through (4 cell lines) or heterozygous mutations (4 cell lines) in proteolytic pathways. To determine which proteolytic path- TTK exhibited a positive band at 90 kDa. The size differ- ways are responsible for degrading mutant proteins, we ence between wild-type and mutant TTK was 0.3 kDa, and designed a rescue assay using several proteolytic inhibitors thus, normal and mutant TTK cannot be distinguished by of 3 major protein degradation pathways: proteasome-, Western blotting in these cell lines. Therefore, SNUC4 autophagy-, and lysosome-mediated degradation (17, cells represented the proper model for confirming wheth- 18). Bafilomycin A1 was used to block lysosomal degrada- er mutant TTK is expressed because these cells have tion, and 3-methyladenine was used to inhibit autophagy- homozygous mutations in TTK. Although SNUC4 cells mediated degradation. Lactacystin and MG132 were used to exhibited relatively higher TTK mRNAexpressionthanthe block the proteasomal pathway. For this experiment, we other cell lines used, only faint mutant TTK expression also generated mutant cDNA vectors of TTK [cTTK(2)], was detected (Fig. 2E). The size difference between wild- TCF7L2 [cTCF7L2(1)], and MARCKS [cMARCKS(2)] to type and mutant TCF7L2 was approximately 24 kDa. We generate control proteins (Fig. 4A). The rescue assays for detected faint bands only in cell lines with 1/1homo- mutant TTK and TCF7L2 were conducted using SNUC4 zygous or 1/w heterozygous deletions (LS174T, SNUC4, cells, and SNUC2B cells were used for mutant MARCKS. The and SNU407) compared with the patterns in cell lines result showed that the levels of endogenous mutant TTK with no mutations (Fig. 2E). The size difference between were significantly increased (approximately 2.5-fold) after wild-type and mutant MARCKS was approximately 35 proteasome inhibition, but no increment was observed kDa. A complete loss of both normal and mutant when autophagy- or lysosome-mediated degradation was MARCKS was observed in the 4 cell lines (LoVo, LS174T, blocked. Expression of the TCF7L2 mutant was also greatly SNUC4, and SNU407) with homozygous mutations (Fig. increased (4-fold) only after proteasome inhibition in 2E). On the basis of these results, we suspected that either SNUC4 cells. For the detection of mutant MARCKS, we translation of the mutant mRNAs is repressed or mutant generated a more sensitive and specific antibody against the proteins are normally generated but extensively degraded neopeptide sequences in the C-terminal region of mutant through protein degradation pathways. MARCKS (Supplementary Fig. S3A). This antibody more specifically and sensitively detected mutant MARCKS than mRNAs containing PTCs from mutant MARCKS are the antibody recognizing the N-terminal region of MARCKS associated with polysomes (Supplementary Fig. S3B). We conducted a mutant To evaluate the efficiency of the translation of mRNAs MARCKS rescue assay using the generated antibody. Inter- containing PTCs in the last exon, we generated the estingly, the expression of mutant MARCKS was increased gMARCKS(WT) vector, a genomic DNA vector construct by more than 100-fold when proteasome degradation was composed of 2 exons and 1 intron of MARCKS.Wealso blocked in SNUC2B cells (Fig. 4B). We clarified these facts generated the gMARCKS(2) vector, a genomic DNA by using RNA interference (RNAi) against TTK, TCF7L2, mutant MARCKS vector missing 2 adenine residues in and MARCKS. In SNUC4 cells treated with siRNA against the cMNR region (A11) of the last exon. We examined the TTK, the proteasome inhibition-induced increase in translation of mutant mRNAs from the gMARCKS(2) mutant TTK expression was reduced to approximately vector by analyzing the distribution of polysomes, using 45% of that in MG132-treated SNUC4 cells, confirming the gMARCKS(WT) vector as a control. Puromycin treat- that these bands represent TTK. To confirm whether ment was used to mimic a condition in which translation MG132 treatment specifically rescues mutant TTK, we is repressed. The results revealed that wild-type MARCKS conducted the same experiment using HeLa cells, which mRNA from HeLa cells transfected with the gMARCKS only express wild-type TTK. The results revealed that (WT) vector was mostly present in the polysome fractions MG132 treatment specifically affects mutant TTK protein (right shifted), and the GAPDH mRNA distribution was (Supplementary Fig. S4A). As observed for mutant TTK, similar, which indicated active translation (Fig. 3A). we showed that the proteasome inhibition–induced Mutant MARCKS mRNA-bearing PTCs from HeLa cells increases in mutant MARCKS and TCF7L2 expression transfected with the gMARCKS(2) vector exhibited a were reduced by approximately 50% upon siRNA treat- similar pattern as the wild-type MARCKS and endogenous ment. However, no changes were observed in HeLa cells GAPDH mRNAs (Fig. 3B). The normal translation of both after proteasome inhibition or siRNA treatment (Supple- wild-type and mutant MARCKS mRNAs was confirmed by mentary Fig. S4B and S4C). We examined the stability of the left shift in the banding pattern after puromycin endogenous mutant TTK in a time course experiment treatment, indicating translational repression (Fig. 3C using SNUC4 (þ1/2) and HeLa cells (wild-type). The and D). On the basis of this polysome analysis, we expression of mutant TTK was dramatically decreased in concluded that both wild-type and mutant MARCKS the absence of MG132, which confirms that endogenous mRNAs are actively translated. mutant TTK is rapidly degraded by the proteasome system

www.aacrjournals.org Clin Cancer Res; 19(13) July 1, 2013 3375

Kim et al.

Figure 3. Mutant mRNAs from the gMARCKS(2) vector construct are associated with the heavy fractions of polysomes. To confirm the translation efficiency of the mRNAs containing PTCs, vector constructs for the genomic DNA form of wild-type [gMARCKS(WT)] and mutant MARCKS [gMARCKS(2)] were generated. A, the mRNAs from gMARCKS(WT) were mainly distributed in fractions 5 to 13, in which polysome peaks were observed. Most GAPDH mRNAs were also found in fractions 5 to 14, indicating that mRNAs from the gMARCKS(WT) construct are normally translated. B, the mRNAs from gMARCKS (2) were distributed in fractions 6 to 14, and GAPDH mRNAs were distributed in a similar pattern. C, cells were treated with puromycin to repress translation. The distributions of mRNAs from gMARCKS(WT) and GAPDH were clearly left-shifted, and the typical fluctuating polysome peaks were not detected, confirming the efficacy of puromycin. D, after puromycin treatment, the distributions of mRNAs from gMARCKS(2) and GAPDH were evidently left-shifted. The gentle and flat polysome peaks were also observed around heavy fractions. The intensities of all of the RT-PCR bands were measured, and they are shown as bars under the bands.

(Fig. 4C). The expression of wild-type TTK was stable when overexpressed in vitro. As expected, the mutant TTK, during the experiment (Fig. 4D). TCF7L2, and MARCKS levels were approximately 19%, 11%, and 13%, respectively, of their wild-type levels. Inhi- Overexpression of mutant MARCKS, TCF7L2, and TTK bition of proteasome degradation by MG132 increased the leads to heavy ubiquitination and localization around expression of the mutants by approximately 3-fold and centrosomes mRNA expression level from the vector constructs was In addition to the rescue assays for endogenous mutant similar (Supplementary Fig. S5A–S5C). To conﬁrm the proteins, we further showed that mutant TTK, TCF7L2, and involvement of the proteasome system, we conducted an MARCKS are more unstable than their wild-type counter- ubiquitination assay using the mutant and wild-type cDNA parts and rapidly degraded by the proteasome system constructs of TTK, TCF7L2, and MARCKS. Consequently,

3376 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Degradation of Truncated Mutants

Figure 4. Endogenous mutant protein rescue assay using inhibitors of protein degradation pathways. A, schematic diagram of the wild-type and mutant TTK, TCF7L2, and MARCKS protein expression vectors used as positive controls. Figures for constructs are presented according to the nucleotide numbers consisting of the coding regions of the genes. , expression vector constructs. B, a mutant protein rescue assay using inhibitors of protein degradation pathways. Baﬁlomycin A1 was used to block lysosomal degradation, and 3-methyladenine was used to inhibit autophagy-mediated degradation. Lactacystin and MG132 were used to block the proteasomal pathway. The positive controls for each mutant were derived from HeLa cells transfected with the respective mutant cDNA constructs (left). Gray arrows indicate the mutant protein sizes, and the black bar graphs show the relative intensities of the bands on the blots (right). Endogenous mutant TTK, TCF7L2, and MARCKS were increased after proteasome inhibition. C, to measure the stability of mutant TTK, SNUC4 cells were incubated with cycloheximide (CHX) in the absence or presence of MG132. The expression of the TTK mutant was dramatically decreased by approximately 4-fold in the absence of MG132, but its expression was stable in the presence of MG132. D, to compare the stability between mutant and wild-type TTK, the same experiment was carried out in HeLa cells expressing wild-type TTK. Wild-type TTK did not display any changes in protein expression upon cycloheximide treatment irrespective of the presence of MG132. 1, 1-bp deletion in the cMNR; 2, 2-bp deletion in the cMNR.

the 3 mutant proteins were more heavily ubiquitinated Enhanced degradation of the truncated mutant (more than 2-fold) than the wild-type proteins in the MARCKS proteins containing neopeptides presence of MG132 (Fig. 5A–C). On the basis of the ubi- All genes containing frameshift mutations in cMNRs in quitination assay, we further examined the localization of the last exon are expected to generate truncated proteins mutant proteins using immunofluorescence microscopy. with variable lengths of neopeptides. After we showed that According to previous studies, we hypothesized that if the endogenous neopeptide-containing truncated mutant mutant proteins are specifically localized around centro- proteins were mostly degraded in the proteasome, we somes, which recruit the proteasomal machinery, then this hypothesized that neopeptides and/or protein truncation further confirms that mutant proteins are actively degraded might enhance mutant protein degradation. To validate our (19, 20). We conducted immunofluorescence staining hypothesis, we constructed another protein expression vec- using antibodies against FLAG and g-tubulin, the latter of tor by using the MARCKS genomic DNA construct. The which was used as a centrosome marker. We chose mutant mutant protein produced from the gMARCKS(2) vector MARCKS constructs for the experiment. Under confocal was composed of 183 amino acids, 27 of which comprised microscopic examination, proteasome inhibition resulted neopeptides in the C-terminal region. We generated an in centrosomal expansion in cells transfected with both additional vector construct by introducing an abnormal wild-type and mutant MARCKS constructs. However, colo- stop codon at amino acid position 183 [gMARCKS(p183)], calization with g-tubulin was detected only in cells trans- which leads to generation of truncated mutant proteins fected with the mutant construct, indicating that mutant lacking neopeptides (Fig. 6A). We evaluated MARCKS proteins are actively recruited to centrosomes after protea- mRNA and protein expression levels before and after some inhibition (Fig. 5D). Taken together, our results MG132 treatment. The expression of mRNAs from the cells suggest that both endogenous and synthetic mutant pro- transfected with the gMARCKS(2) and gMARCKS(p183) teins are more rapidly degraded than wild-type proteins, vectors was similar to that in the cells transfected with the and this process is mediated via the ubiquitin–proteasome wild-type MARCKS vector [gMARCKS(WT); Fig. 6B, right]. pathway. Comparing the protein expression level, we found that the

www.aacrjournals.org Clin Cancer Res; 19(13) July 1, 2013 3377

Kim et al.

Figure 5. Heavy ubiquitination of synthetic mutant TTK, TCF7L2, and MARCKS and their colocalization with centrosomes after proteasome inhibition. A–C, ubiquitination assays of synthetic wild-type and mutant TTK, TCF7L2, and MARCKS generated from cDNA expression vectors in the presence or absence of MG132. Mutant TTK, TCF7L2, and MARCKS were more heavily ubiquitinated than their wild-type counterparts. D, subcellular localization of mutant MARCKS proteins generated from the cDNA expression vectors. Under confocal microscopic examination, proteasome inhibition resulted in centrosomal expansion in cells transfected with both wild-type and mutant MARCKS constructs. Colocalization with g-tubulin was detected only in cells transfected with the mutant constructs, indicating that mutant proteins are actively recruited to the centrosomal region after proteasome inhibition. White arrows denote centrosomes. FITC, ﬂuorescein isothiocyanate.

expression of synthetic wild-type MARCKS protein was of insoluble bodies in cells. To show whether the formation higher than that of synthetic mutant MARCKS proteins of insoluble bodies reduces mutant protein expression, we irrespective of the presence of neopeptides. Interestingly, fractionated the cell lysates into Triton X-100–soluble and cells transfected with gMARCKS(p183) exhibited nearly 2- Triton X-100–insoluble fractions. GAPDH and g-tubulin fold higher mutant protein levels than cells transfected with were used as Triton X-100–soluble markers. Surprisingly, gMARCKS(2) in the absence of MG132. After proteasome signiﬁcant levels of both mutant MARCKS proteins were inhibition, the levels of neopeptide-containing mutant detected in the Triton X-100–insoluble fraction, indicating MARCKS were almost doubled, whereas those of the that the low expression of mutant MARCKS after protea- mutant lacking neopeptides were only increased slightly some inhibition was due to the increased insolubility of the (Fig. 6B, left). The ubiquitination assay indicated that the mutant proteins and subsequent formation of insoluble mutant MARCKS containing neopeptides is more heavily bodies (Fig. 6D). These ﬁndings indicate that truncated ubiquitinated than the mutant MARCKS lacking neopep- mutant proteins containing neopeptides are rarely detected tides, indicating that neopeptides are primarily responsible because of extensive degradation and increased insolubility for the degradation of mutant MARCKS (Fig. 6C). As the (Fig. 6E). expression of mutant MARCKS remained lower than that of wild-type MARCKS even after proteasome inhibition, we Discussion suspected that other factors might be involved in the low Mutations are hallmarks of diseases, especially cancers expression of mutant MARCKS. Thus, we hypothesized that (21, 22). Mutations contribute to cancers via gain-of-func- the changes in mutant protein levels could be related to tion mutations of oncogenes and loss-of-function muta- increases in insolubility, which contributes to the formation tions in tumor suppressor genes (23). The generation and

3378 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Degradation of Truncated Mutants

Figure 6. Expression, degradation, and insolubility of wild-type and truncated mutant MARCKS containing or lacking neopeptides. A, schematic diagram of the genomic DNA vector constructs of wild-type and mutant MARCKS. , expression vector constructs. B, protein expression from each vector construct was analyzed by Western blotting. The expression level of wild-type MARCKS was higher than that of both mutant MARCKS proteins, and the protein expression level of mutant MARCKS lacking neopeptides was approximately 2-fold higher than that of mutant MARCKS containing neopeptides (left). The mRNA expression level was also measured, and all of the constructs were expressed at similar levels (right). C, an ubiquitination assay was conducted. Mutant MARCKS containing neopeptides was more heavily ubiquitinated, in contrast to the slight levels of ubiquitination of mutant MARCKS lacking neopeptides (top). The relative intensities were measured and presented as a bar graph (bottom). D, lysates from cells transfected with MARCKS(WT), MARCKS(2), and MARCKS(183) were prepared from Triton-soluble supernatant and Triton-insoluble pellet fractions and analyzed by Western blotting. Wild-type MARCKS was mostly present in the soluble faction. Conversely, both mutant MARCKS proteins were mostly present in the insoluble faction. E, schematic model for the fate of mutant proteins derived from NMD-irrelevant PTC-containing mRNAs. z, Endogenous wild-type MARCKS. 1, 1-bp deletion in the cMNR; 2, 2-bp deletion in the cMNR. fate of mutant proteins produced from mutated genes vary point mutation in the other allele is very frequent in cancers, according to the gene and type of mutation. Mutations in and mutant p53 overexpression due to the dysregulation of the oncogenes such as KIT and b-catenin result in the over- ubiquitination has been reported (31–33). To this point, expression of activated KIT and b-catenin in many human the protein expression of many other cancer-related genes cancers (24, 25). Overexpression of the wild-type oncopro- with inactivating mutations remains unknown. teins HER2 and cMYC due to gene ampliﬁcation is also well Nonsense and frameshift mutations are important inac- established (26, 27). About tumor suppressor genes, down- tivating mutations in the development of genetic diseases regulation of wild-type proteins is common due to inacti- and human cancers. In particular, in human MSI-H cancers, vation of both alleles (28). The inactivation mechanisms are frequent frameshift mutations have been reported in many deletion, methylation inactivation, and inactivating muta- genes (8, 10, 12). When frameshift mutations induce ran- tion (29, 30). No functional proteins can be expressed from domized nucleotide arrangement after insertion or deletion the tumor suppressor genes with homozygous deletions or sites, the probability of stop codon generation is approxi- those with deletion of one allele and methylation inactiva- mately 3/64 (3 stop codons for every 64 codons), and tion of the other allele. In cases of inactivating mutations, therefore, the average neopeptide is expected to be approx- however, expression of the functional protein might be imately 20 amino acids in length. In 29 genes, in which we variable according to the genes and type of mutations. For showed frameshift mutations, the mean length of neopep- example, p53 ablation via the deletion of one allele and a tides induced by the 1-bp deletion and 2-bp deletion

www.aacrjournals.org Clin Cancer Res; 19(13) July 1, 2013 3379

Kim et al.

mutations was 22.7 17, and 28.5 52.8, respectively data explain why the mutant proteins are barely detected, (Supplementary Fig. S6A–S6C). Because we confirmed that and reveal that tumor cells are protected from potentially mRNAs derived from mutated genes were intact in the harmful mutant proteins by both NMD- and the ubiquitin- cytoplasm, the NMD-irrelevant mRNA model is essential mediated protein degradation mechanism. In addition to to determine if truncated mutant proteins with neopeptides these results, the functional relevance of scant amount of are expressed in tumors. Previously, several studies have detected neopeptide-containing mutant proteins in the attempted to detect truncated mutant proteins derived from tumor progression should be studied in the future. PTC-carrying mRNAs and clarify the roles of these mutant Our findings suggest that the degraded mutant proteins proteins in diseases such as cancers and genetic diseases contribute to the formation of tumor-specific antigens (13, 34–39). Thus far, most of the studies have shown the and these antigens are useful targets for immunotherapy. existence of mutant proteins at the DNA or mRNA level Intense peritumoral and intratumoral lymphocytic infil- (34, 37). Some studies suggested the existence of mutant tration, and its association with favorable prognosis have proteins by in vitro overexpression experiments, but the been reported as the characteristics of MSI-H colon can- constructs used in the previous studies had PTCs in specific cers (8, 45, 46). It is also very clear that the amount of positions of the C-terminal region, which represents non- intracellular mutant proteins, a substrate for tumor anti- sense mutations (38–40). In most cancers, insertion/dele- gen, is closely related to the effective tumor antigen tion mutations usually lead to frameshift mutations that are formation (47). Our results, protein translation from the much more frequent and deleterious than nonsense muta- NMD-irrelevant mutant mRNAs and the generalized deg- tions. Therefore, the identification of endogenous truncated radation of neopeptide-containing mutant truncated pro- proteins containing neopeptides from genes with frameshift teins, provide novel insights that (i) the intracellular mutations is more important, but this has not been studied amount of mutant proteins are scant, but (ii) the degra- at the protein level. We herein validated for the first time the dation of these neopeptide-containing mutant proteins expression of the truncated mutant proteins derived from by proteasome system is directly related to the antigen- the genes with frameshift mutations in protein level. processing and presentation by MHC class I, therefore For the validation of endogenous truncated mutant pro- expected to be effective tumor antigen formation. When tein expression, we chose TTK, TCF7L2, and MARCKS. The we analyzed 19 MSI-H colon cancer tissues, we found a physiologic roles of TTK, TCF7L2, and MARCKS have been significant relationship between the intensity of peritu- studied in several cancers, and these proteins have signif- moral reaction and the mutation status of 3 genes [TTK icant relevance to cancer progression (41–44). We original- (P ¼ 0.01), TCF7L2 (P ¼ 0.46), and MARCKS (P ¼ ly expected that significant amounts of mutant TTK might be 0.002); Table 1]. These findings suggest that the degraded expressed because this mutant protein has minor changes mutant proteins might be related to the regional immune and a readthrough stop codon. To our surprise, endogenous responses of the tumor. A large-scale correlation study mutant TTK was barely detected by Western blotting despite and a study on the immunostimulatory function of the the relatively high mRNA expression level of TTK in SNUC4 neopeptides will be necessary to determine the roles of cells, which have homozygous mutations (þ1/2) in TTK. mutant proteins in tumor antigen formation, and muta- Mutant TCF7L2 was also barely expressed, and mutant tions in other genes containing cMNRs in the last exon MARCKS was not expressed in the tumor cells with homo- also should be evaluated. zygous mutations. After proteasome inhibition by MG132 treatment, large amounts of mutant TTK, TCF7L2, and Disclosure of Potential Conflicts of Interest MARCKS were detected, and we observed these dramatic No potential conflicts of interest were disclosed. increases only in the cells expressing the mutant proteins. Authors' Contributions These findings clearly indicated that the rare expression of Conception and design: W.K. Kim, K.T. You, H. Kim mutant proteins of these 3 genes is mostly due to enhanced Development of methodology: W.K. Kim, K.T. You, H. Kim degradation. In addition to showing the dramatic increment Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): W.K. Kim, M. Park, H.K. Kim, H. Kim of mutant protein expression after proteasome inhibition, Analysis and interpretation of data (e.g., statistical analysis, biosta- we also showed the selective and heavy ubiquitination of the tistics, computational analysis): W.K. Kim, M. Park, N. Shin, H. Kim mutant proteins. Moreover, the colocalization of mutant Writing, review, and/or revision of the manuscript: W.K. Kim, H. Kim Administrative, technical, or material support (i.e., reporting or orga- MARCKS with centrosomes in the presence of MG132 sug- nizing data, constructing databases): W.K. Kim, Y.J. Kim, N. Shin, H. Kim gests that mutant MARCKS is actively degraded via the Study supervision: W.K. Kim, H. Kim proteasomal machinery, which is assembled around centrosomes. By generating 2 different truncated mutant MARCKS Grant Support This work was supported by the National Research Foundation of Korea expression vectors that contained or lacked neopeptides, we grant funded by the Korean government [Ministry of Education, Science, and showed that neopeptide-containing mutant MARCKS pro- Technology (MEST); No. 2012R1A2A2A01005196], a grant of the Korean teins are more extensively degraded. Furthermore, we found Health 21R&D Project, Ministry of Health and Welfare, Republic of Korea (A111218-CP01), Technology and by the Converging Research Center that the neopeptide-lacking mutant proteins were relatively Program through the Ministry of Education, Science and Technology stable compared with the neopeptide-containing mutant (2010K001115), and a grant from the Korea Healthcare Technology R&D Project, Ministry for Health & Welfare Affairs, Republic of Korea (A085136). proteins, and the truncated mutant displayed increased The costs of publication of this article were defrayed in part by the insolubility irrespective of the presence of neopeptides. Our payment of page charges. This article must therefore be hereby marked

3380 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Degradation of Truncated Mutants

advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate Received March 11, 2013; revised April 16, 2013; accepted April 20, 2013; this fact. published OnlineFirst May 14, 2013.

References 1. Stratton MR, Campbell PJ, Futreal PA. The cancer genome. Nature 21. Jones S, Zhang XS, Parsons DW, Lin JCH, Leary RJ, Angenendt P, 2009;458:719–24. et al. Core signaling pathways in human pancreatic cancers revealed 2. Santarius T, Shipley J, Brewer D, Stratton MR, Cooper CS. A census of by global genomic analyses. Science 2008;321:1801–6. amplified and overexpressed human cancer genes. Nat Rev Cancer 22. Hanahan D, Weinberg RA. Hallmarks of cancer: the next generation. 2010;10:59–64. Cell 2011;144:646–74. 3. Stephens PJ, Tarpey PS, Davies H, Van Loo P, Greenman C, Wedge 23. Lengauer C, Kinzler KW, Vogelstein B. Genetic instabilities in human DC, et al. The landscape of cancer genes and mutational processes in cancers. Nature 1998;396:643–9. breast cancer. Nature 2012;486:400–4. 24. Kim WK, Park M, Kim YK, Tae YK, Yang HK, Lee JM, et al. MicroRNA- 4. Brogna S, Wen J. Nonsense-mediated mRNA decay (NMD) mechan- 494 downregulates KIT and inhibits gastrointestinal stromal tumor cell isms. Nat Struct Mol Biol 2009;16:107–13. proliferation. Clin Cancer Res;17:7584–94. 5. Perrin-Vidoz L, Sinilnikova OM, Stoppa-Lyonnet D, Lenoir GM, 25. Clevers H, Nusse R. Wnt/beta-catenin signaling and disease. Cell Mazoyer S. The nonsense-mediated mRNA decay pathway triggers 2012;149:1192–205. degradation of most BRCA1 mRNAs bearing premature termination 26. Meng S, Tripathy D, Shete S, Ashfaq R, Haley B, Perkins S, et al. HER-2 codons. Hum Mol Genet 2002;11:2805–14. gene amplification can be acquired as breast cancer progresses. Proc 6. Zhang J, Sun X, Qian Y, LaDuca JP, Maquat LE. At least one intron is Natl Acad Sci U S A 2004;101:9393–8. required for the nonsense-mediated decay of triosephosphate isom- 27. Persons DL, Borelli KA, Hsu PH. Quantitation of HER-2/neu and c-myc erase mRNA: a possible link between nuclear splicing and cytoplasmic gene amplification in breast carcinoma using fluorescence in situ translation. Mol Cell Biol 1998;18:5272–83. hybridization. Mod Pathol 1997;10:720–7. 7. Zhang J, Sun X, Qian Y, Maquat LE. Intron function in the nonsense- 28. Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100: mediated decay of beta-globin mRNA: indications that pre-mRNA 57–70. splicing in the nucleus can influence mRNA translation in the cyto- 29. Nowak MA, Michor F, Komarova NL, Iwasa Y. Evolutionary dynamics plasm. RNA 1998;4:801–15. of tumor suppressor gene inactivation. Proc Natl Acad Sci U S A 8. Woerner SM, Kloor M, von Knebel Doeberitz M, Gebert JF. Microsat- 2004;101:10635–8. ellite instability in the development of DNA mismatch repair deficient 30. Chim CS, Liang R, Tam CY, Kwong YL. Methylation of p15 and p16 tumors. Cancer Biomarkers 2006;2:69–86. genes in acute promyelocytic leukemia: potential diagnostic and 9. Ionov Y, Peinado MA, Malkhosyan S, Shibata D, Perucho M. prognostic significance. J Clin Oncol 2001;19:2033–40. Ubiquitous somatic mutations in simple repeated sequences reveal 31. Freed-Pastor WA, Prives C. Mutant p53: one name, many proteins. a new mechanism for colonic carcinogenesis. Nature 1993;363: Genes Dev 2012;26:1268–86. 558–61. 32. Muller P, Hrstka R, Coomber D, Lane DP, Vojtesek B. Chaperone- 10. Markowitz S, Wang J, Myeroff L, Parsons R, Sun L, Lutterbaugh J, et al. dependent stabilization and degradation of p53 mutants. Oncogene Inactivation of the type II TGF-beta receptor in colon cancer cells with 2008;27:3371–83. microsatellite instability. Science 1995;268:1336–8. 33. Kim H, Lee JM, Lee G, Bhin J, Oh SK, Kim K, et al. DNA damage- 11. Potocnik U, Glavac D, Ravnik-Glavac M. Identification of novel genes induced RORalpha is crucial for p53 stabilization and increased apo- with somatic frameshift mutations within coding mononucleotide ptosis. Mol Cell 2011;44:797–810. repeats in colorectal tumors with high microsatellite instability. Genes 34. Hall GW, Thein S. Nonsense codon mutations in the terminal exon of Chromosomes Cancer 2003;36:48–56. the beta-globin gene are not associated with a reduction in beta- 12. Woerner SM, Yuan YP, Benner A, Korff S, von Knebel Doeberitz M, mRNA accumulation: a mechanism for the phenotype of dominant Bork P. SelTarbase, a database of human mononucleotide-microsat- beta-thalassemia. Blood 1994;83:2031–7. ellite mutations and their potential impact to tumorigenesis and immu- 35. Jouanguy E, Altare F, Lamhamedi S, Revy P, Emile JF, Newport M, nology. Nucleic Acids Res 2010;38:D682–9. et al. Interferon-gamma-receptor deficiency in an infant with fatal 13. Holbrook JA, Neu-Yilik G, Hentze MW, Kulozik AE. Nonsense- bacille Calmette–Guerin infection. N Engl J Med 1996;335:1956–61. mediated decay approaches the clinic. Nat Genet 2004;36: 36. Jouanguy E, Lamhamedi-Cherradi S, Lammas D, Dorman SE, Fonda- 801–8. neche MC, Dupuis S, et al. A human IFNGR1 small deletion hotspot 14. Ku JL, Yoon KA, Kim DY, Park JG. Mutations in hMSH6 alone are not associated with dominant susceptibility to mycobacterial infection. sufficient to cause the microsatellite instability in colorectal cancer cell Nat Genet 1999;21:370–8. lines. Eur J Cancer 1999;35:1724–9. 37. Millar DS, Kemball-Cook G, McVey JH, Tuddenham EG, Mumford AD, 15. Lengauer C, Kinzler KW, Vogelstein B. DNA methylation and genetic Attock GB, et al. Molecular analysis of the genotype-phenotype instability in colorectal cancer cells. Proc Natl Acad Sci U S A relationship in factor VII deficiency. Hum Genet 2000;107:327–42. 1997;94:2545–50. 38. Fan S, Yuan R, Ma YX, Meng Q, Goldberg ID, Rosen EM. Mutant 16. Kim NG, Rhee H, Li LS, Kim H, Lee JS, Kim JH, et al. Identification of BRCA1 genes antagonize phenotype of wild-type BRCA1. Oncogene MARCKS, FLJ11383 and TAF1B as putative novel target genes in 2001;20:8215–35. colorectal carcinomas with microsatellite instability. Oncogene 39. Sylvain V, Lafarge S, Bignon YJ. Dominant-negative activity of a Brca1 2002;21:5081–7. truncation mutant: effects on proliferation, tumorigenicity in vivo,and 17. Wong E, Cuervo AM. Autophagy gone awry in neurodegenerative chemosensitivity in a mouse ovarian cancer cell line. Int J Oncol diseases. Nat Neurosci 2010;13:805–11. 2002;20:845–53. 18. Shaid S, Brandts CH, Serve H, Dikic I. Ubiquitination and selective 40. Inoue K, Khajavi M, Ohyama T, Hirabayashi S, Wilson J, Reggin JD, autophagy. Cell Death Differ 2013;20:21–30. et al. Molecular mechanism for distinct neurological phenotypes con- 19. Wigley WC, Fabunmi RP, Lee MG, Marino CR, Muallem S, DeMartino veyed by allelic truncating mutations. Nat Genet 2004;36:361–9. GN, et al. Dynamic association of proteasomal machinery with the 41. Chinnadurai G. CtBP, an unconventional transcriptional corepressor in centrosome. J Cell Biol 1999;145:481–90. development and oncogenesis. Mol Cell 2002;9:213–24. 20. Bennett EJ, Bence NF, Jayakumar R, Kopito RR. Global impairment of 42. Duval A, Rolland S, Tubacher E, Bui H, Thomas G, Hamelin R. The the ubiquitin-proteasome system by nuclear or cytoplasmic protein human T-cell transcription factor-4 gene: structure, extensive char- aggregates precedes inclusion body formation. Mol Cell 2005;17: acterization of alternative splicings, and mutational analysis in colo- 351–65. rectal cancer cell lines. Cancer Res 2000;60:3872–9.

www.aacrjournals.org Clin Cancer Res; 19(13) July 1, 2013 3381

Kim et al.

43. Kalwa H, Michel T. The MARCKS protein plays a critical role in cation errors in microsatellite sequences. Am J Pathol 1994;145: phosphatidylinositol 4,5-bisphosphate metabolism and directed cell 148–56. movement in vascular endothelial cells. J Biol Chem 2011;286: 46. Kloor M, Michel S, von Knebel Doeberitz M. Immune evasion of 2320–30. microsatellite unstable colorectal cancers. Int J Cancer 2010;127: 44. Yeh YH, Huang YF, Lin TY, Shieh SY. The cell cycle checkpoint kinase 1001–10. CHK2 mediates DNA damage-induced stabilization of TTK/hMps1. 47. Schwitalle Y, Kloor M, Eiermann S, Linnebacher M, Kienle P, Knaebel Oncogene 2009;28:1366–78. HP, et al. Immune response against frameshift-induced neopeptides in 45. Kim H, Jen J, Vogelstein B, Hamilton SR. Clinical and pathological HNPCC patients and healthy HNPCC mutation carriers. Gastroenter- characteristics of sporadic colorectal carcinomas with DNA repli- ology 2008;134:988–97.

3382 Clin Cancer Res; 19(13) July 1, 2013 Clinical Cancer Research

Identification and Selective Degradation of Neopeptide-Containing Truncated Mutant Proteins in the Tumors with High Microsatellite Instability

Won Kyu Kim, Misun Park, Minhee Park, et al.

Clin Cancer Res 2013;19:3369-3382. Published OnlineFirst May 14, 2013.

Updated version Access the most recent version of this article at: doi:10.1158/1078-0432.CCR-13-0684

Supplementary Access the most recent supplemental material at: Material http://clincancerres.aacrjournals.org/content/suppl/2013/05/15/1078-0432.CCR-13-0684.DC1

Cited articles This article cites 46 articles, 13 of which you can access for free at: http://clincancerres.aacrjournals.org/content/19/13/3369.full#ref-list-1

Citing articles This article has been cited by 4 HighWire-hosted articles. Access the articles at: http://clincancerres.aacrjournals.org/content/19/13/3369.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://clincancerres.aacrjournals.org/content/19/13/3369. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.