Ju Hwa Lee

Frequent frameshift mutations in 2 mononucleotide repeats of RNF43 gene and its regional heterogeneity in gastric and colorectal cancers

RNF43, an E3 ligase, inhibits Wnt signaling by removing Wnt receptors and behaves as a candidate tumor suppressor. Recent studies identified that RNF43 gene was frequently mutated in gastric (GC), colorectal (CRC), and endometrial cancers with high microsatellite instability (MSI-H). The aim of this study is to explore whether RNF43 gene is mutated in GC and CRC in Korean patients and whether the mutations show regional intratumoral heterogeneity (ITH). We analyzed 2 exonic repeats (C6 and G7) of RNF43 in 78 GCs and 130 CRCs by single-strand conformation polymorphism and DNA sequencing analyses. Also, we analyzed regional ITH of RNF43 mutation in 16 CRCs. We found RNF43 frameshift mutation in MSI-H (50/118), the incidence of which was significantly higher than that in microsatellite stable/low microsatellite instability (1/90). GCs showed a significantly higher incidence of the mutation than CRCs (66.7% of GC and 32.9% of CRC with MSI-H). Also, we found that all of the 7 CRCs with the mutations harbored mutational ITH. By immunohistochemistry, we observed that loss of RNF43 expression was significantly more common in those with RNF43 frameshift mutation than those with wild-type RNF43. Our data indicate that RNF43 gene harbored not only exceedingly high mutations but also mutational ITH, which together might play a role in tumorigenesis of GC and CRC. We suggest that regional analysis is required for a more comprehensive evaluation of the mutation status in these tumors.

Frameshift mutation of WISP3 gene and its regional heterogeneity in gastric and colorectal cancers

WISP3 is involved in many cancer-related processes including epithelial-mesenchymal transition, cell death, invasion, and metastasis and is considered a tumor suppressor. The aim of our study was to find whether WISP3 gene was mutated and expressionally altered in gastric (GC) and colorectal cancers (CRCs). WISP3 gene possesses a mononucleotide repeat in the coding sequence that could be mutated in cancers with high microsatellite instability (MSI-H). We analyzed 79 GCs and 156 CRCs, and found that GCs (8.8%) and CRCs (10.5%) with MSI-H, but not those with microsatellite stable/low MSI, harbored a frameshift mutation. We also analyzed intratumoral heterogeneity (ITH) of the frameshift mutation in 16 CRCs and found that the WISP3 mutation exhibited regional ITH in 25% of the CRCs. In immunohistochemistry, loss of WISP3 expression was identified in 24% of GCs and 21% of CRCs. The loss of expression was more common in those with WISP3 mutation than with wild-type WISP3 and those with MSI-H than with microsatellite stable/low MSI. Our data indicate that WISP3 harbored not only frameshift mutation but also mutational ITH and loss of expression, which together might play a role in tumorigenesis of GC and CRC with MSI-H by inhibiting tumor suppressor functions of WISP3. Our data also suggest that mutation analysis in multiregions is needed for a proper evaluation of mutation status in GC and CRC with MSI-H.

Frameshift mutations of a tumor suppressor gene ZNF292 in gastric and colorectal cancers with high microsatellite instability.

A transcription factor‐encoding gene ZNF292 is considered a candidate tumor suppressor gene (TSG). Its mutations have been identified in cancers from liver, colon, and bone marrow. However, ZNF292 inactivating mutations that might suppress the TSG functions have not been reported in gastric (GC) and colorectal cancers (CRC) with microsatellite instability (MSI). In a public database, we found that ZNF292 gene had mononucleotide repeats in the coding sequences that might be mutation targets in the cancers with MSI. In this study, we analyzed 79 GCs and 124 CRCs including high MSI (MSI‐H) and microsatellite stable/low MSI (MSS/MSI‐L) cases for the detection of somatic mutations in the repeats. Overall, we identified frameshift mutations of ZNF292 in 3 (8.8%) GCs and 11 (13.9%) CRCs with MSI‐H (14/113), but not in MSS/MSI‐L cancers (0/90) (p < 0.001). Also, we studied intratumoral heterogeneity (ITH) of the ZNF292 frameshift mutations in 16 CRCs and found that two (12.5%) had regional ITH of the mutations. Our data show that ZNF292 gene harbors not only frameshift mutations but also mutational ITH, which together may be features of GC and CRC with MSI‐H. Based on this, the ZNF292 frameshift mutations may possibly contribute to tumorigenesis by altering its TSG functions in GC and CRC.

Intratumoral Heterogeneity of Frameshift Mutations of GLI1 Encoding a Hedgehog Signaling Protein in Colorectal Cancers

GLI1 is a transcription factor for hedgehog signaling that plays a crucial role in signaling pathways for controlling cell proliferation, alterations of which are known to contribute to tumorigenesis. Aim of this study was to explore whether GLI1 gene is mutated in gastric (GC) and colorectal cancers (CRC). In a public database, we found that GLI1 had a G7 mononucleotide repeat in the coding sequences that could be a mutation target in the cancers with microsatellite instability (MSI). In this study, we analyzed frameshift mutation of GLI1 in 79 GCs and 129 CRCs (high MSI (MSI-H) or microsatellite stable (MSS)) by single-strand conformation polymorphism analysis and DNA sequencing. We found 10 frameshift mutations in the repeat, nine for CRCs and one for GC. All of the mutations were detected in cancers with MSI-H and there was a statistical difference in the frameshift mutation frequencies between the cancers with MSI-H (10/113) and MSS (0/90). We also analyzed intratumoral heterogeneity (ITH) of the frameshift mutation in 16 CRCs and found that the mutations exhibited regional ITH in three of the CRCs (18.8%). Our data indicate GLI1 harbored not only frameshift mutation but also its mutational ITH, which together could be a feature of GC and CRC with MSI-H.

BPTF, a chromatin remodeling-related gene, exhibits frameshift mutations in gastric and colorectal cancers

To the Editor Chromatin remodeling is a dynamic modification of chromatin architecture to allow access of the regulatory transcription proteins to condensed DNA, and thereby control gene expression. In addition to gene expression regulation, chromatin remodeling plays an epigenetic regulatory role in many processes, including DNA repair, cell death, cell cycle, and pluripotency, many of which are related to cancer development (1). Alterations in chromatin-remodeling genes, including ARID1A, MLL2, and ATRX, are found in many cancers (2). Chromatin remodeling is carried out by histone modifications and ATP-dependent chromatin remodeling complexes (1). The chromatin remodelers are subdivided into four families: SWI/SNF, INO80/SWR1, ISWI, and CHD families (1). BPTF (also known as NURF301) is an ISWI-containing ATP-dependent chromatin remodeling factor that mediates a direct preferential association with H3K4me3 tails (3). The essential role of BPTF in embryo development is well known (3), but its role in cancer development remains elusive. Many cancers, including bladder cancers and melanomas harbor somatic mutations of BPTF (4, 5) that are comprised of not only missense mutations but also truncating mutations (5). In a public genome database (http://genome. cse.ucsc.edu/), we found that BPTF gene have mononucleotide repeats in their coding sequences that could be targets for frameshift mutation in cancers with microsatellite instability (MSI). Frameshift mutations of genes with mononucleotide repeats are features of gastric (GC) and colorectal cancers (CRC) with MSI (6). Frameshift mutations in BPTF might cause alterations of their functions and contribute to cancer pathogenesis, but it remains unknown whether it is mutated in GC and CRC with MSI. To see whether BPTF gene harbored frameshift mutations of the repeats in GC and CRC, we analyzed the A9 repeat in exon 8, A8 repeat in exon 10, and A7 repeat in exon 23 in 34 GC with high MSI (MSI-H), 45 microsatellitestable/low MSI (MSS/MSI-L), 79 CRCs with MSIH and 45 CRCs with MSS/MSI-L by polymerase chain reaction (PCR) and single-strand conformation polymorphism (SSCP) assay as described previously (7). The MSI evaluation system used five mononucleotide repeats (BAT25, BAT26, NR-21, NR-24, and MONO-27), tumoral MSI status of which was characterized as: MSI-H, if two or more of these markers show instability, MSI-L, if one of the markers shows instability and MSS, if none of the markers shows instability (8). In cancer tissues, malignant cells and normal cells were selectively procured from hematoxylin and eosin-stained slides by microdissection (7). Radioisotope ([P]dCTP) was incorporated into the PCR products for detection by autoradiogram. The PCR products were subsequently displayed in SSCP gels. After SSCP, direct DNA sequencing reactions were performed in the cancers with mobility shifts in the SSCP as described previously (7). On the SSCP, we found aberrantly migrating bands in eight cases of the A9, two cases of the A8, and one case of the A7 (Fig. 1 and Table 1). DNA from the patients’ normal tissues showed no shifts in SSCP, indicating the aberrant bands had risen somatically. DNA sequencing analysis confirmed that the aberrant bands represented BPTF somatic mutations, which consisted of frameshift mutations by deletion of one base (c.2976delA, c.3259delA, and c.7725delA) or deletion of two bases (c.2975_2976delAA) or duplication of one base (c.2976dupA) in the repeats (Table 1). All of the mutations detected were interpreted as heterozygous according to the SSCP and direct sequencing analyses (Fig. 1A,B). The mutations were detected in cancers with MSI-H, but not in those with MSS/MSI-L. There was a statistical difference in the frameshift mutation frequencies between the cancers with MSI-H (11/113) and MSS/MSI-L (0/90) (Fisher’s exact test, p = 0.001). In terms of tissue origins, there was no statistical difference in mutation frequencies between GCs (2.94%, 1/34) and CRCs (12.7%, 10/79) with MSI-H (Fisher’s exact test, p = 0.084). There was no significant association of the mutations with the clinicopathologic data of the patient (age, sex, histologic grade, and stage). It is now well known that alteration of chromatin remodeling is actively involved in cancer development (1, 4, 5). Based on this, we attempted to disclose whether somatic frameshift mutations of a chromatin-remodeling gene BPTF were present in GC and CRC. We found that 2.9% of GCs and 12.7% of CRCs with MSI-H harbored BPTF frameshift mutations, indicating that BPTF gene is altered in GC and CRC with MSI-H by somatic frameshift mutation that might alter the function of BPTF protein. The mutations we found (Table 1) would delete amino acids after the frameshift mutations and hence would resemble a typical loss-of-function mutation. An earlier study identified that BPTF knockdown led

Intratumoral Heterogeneity of RPL22 Frameshift Mutation in Colorectal Cancers

Dear Editor, RPL22 gene encodes a cytoplasmic ribosomal protein that is a component of the large 60S subunit of ribosome. An earlier study identified that RPL22 protein functions as a haploinsufficient tumor suppressor [1]. RPL22 inactivation promotes transformation by inducing expression of Lin28B [1]. A recent study discovered that RPL22 was frequently mutated in colorectal cancer (CRC) and endometrial cancers by frameshift mutations in A8 repeat, especially those in microsatellite instability high (MSI-H) cancers [2]. Also, another study demonstrated a higher percentage of RPL22 frameshift mutations in MSI-H gastric cancer (GC) [3]. These data suggest that RPL22 is a tumor suppressor that is commonly inactivated in MSI-H cancers by mutations. Intratumoral heterogeneity (ITH) plays an important role in cancer development and progression and impedes proper diagnosis and treatment of cancers [4]. Currently, we are aware of the frequent mutations of RPL22 in MSI-H cancers, but mutational ITH of RPL22 remains elusive. Genes are often observed to harbor frameshift mutations at monocleotide repeats in MSI-H cancers [2, 3]. The present study aimed to find whether RPL22 gene harbored not only frameshift mutations within the A8 repeat but also ITH of the frameshift mutations. We analyzed the A8 repeat in 34 GCs withMSI-H, 45 GCs withMSS, 79 CRCs with MSI-H and 45 CRCs with MSS by polymerase chain reaction (PCR) and single-strand conformation polymorphism (SSCP) assay. After SSCP, Sanger DNA sequencing reactions were performed in the cancers with mobility shifts in the SSCP [5]. We found RPL22 somatic frameshift mutations in 16 CRCs (16/79, 20.3%) and 9 GCs (9/34, 26.5%) with MSI-H, but not in CRCs (0/45) and GCs (0/45) with MSS (Fisher’s exact test, p < 0.001). These mutations were not detected in their normal tissues. The mutations consisted of ‘A’ deletion (c.44delA (p .Lys16Se r f sx4 ) ) , ‘A’ dup l i c a t i on ( c . 44dupA (p.Lys16Glufsx9)) and ‘AA’ deletion (c.43_44delAA (p.Lys15Glufsx9)) in the coding region (Table 1). For ITH of the mutation, we studied 16 cases of CRCs with 4 to 7 regional fragments per CRC. Four of the 16 CRCs (25.0%) showed either the ‘A’ deletion (2 cases) or ‘A’ duplication (one case) or ‘AA’ deletion mutation in different tissue regions. One (case #34) of the 4 CRCs exhibited the ‘A’ duplication in 6 regions as well as the wild type (A8) in the other one region, indicating ITH of the RPL22mutation existed in CRC (Fig. 1). Clinical and histopathological parameters, however, could distinguish neither RPL22 frameshift mutation (+) and (−) cancers, nor the ITH (+) and (−) cancers. Our data here confirm the previous studies on the frequent involvement of RPL22 frameshift mutations in GC and CRC. Furthermore, we report for the first time ITH of the RPL22 frameshift mutation in CRC. The frameshift mutations of RPL22 identified in this study would result in truncation of RPL22 protein, suggesting that RPL22 may be inactivated in MSI-H GCs and CRCs by the frameshift mutations. Based on the tumor suppressor functions of RPL22, the RPL22 frameshift mutations appear to reduce the anti-tumor activities and contribute to tumor pathogenesis. However, ITH of the frameshift mutation in CRC might suggest a possibility that there could be a mixed or ameliorated effect of RPL22 inactivation in MSI-H cancers. However, we were not able to find any distinguished clinicopathologic features of RPL22-mutated or ITH-positive cancers. It was probably due to small number of the mutated cases. Thus, further studies are needed to define the clinical implication of RPL22 mutations and ITH in MSI-H cancers. Ju Hwa Lee and Chang Hyeok An contributed equally to this work.

Intratumoral Heterogeneity of Somatic Mutations for NRIP1, DOK1, ULK1, ULK2, DLGAP3, PARD3 and PRKCI in Colon Cancers

Both NRIP1 and DOK1 genes are considered candidate tumor suppressor genes (TSGs). Also, cell polarity-related genes PARD3, PRKCI and DLGAP3, and autophagy-related genes ULK1 and ULK2 genes are considered to play crucial roles in tumorigenesis. The aim of our study was to find whether these genes were mutated in colorectal cancer (CRC). In a genome database, we observed that each of these genes harbored mononucleotide repeats in the coding sequences, which could be mutated in cancers with high microsatellite instability (MSI-H). For this, we studied 124 CRCs for the frameshift mutations of these genes and their intratumoral heterogeneity (ITH). NRIP1, DOK1, PARD3, PRKCI, DLGAP3, ULK1 and ULK2 harbored 18 (22.8%), 2 (2.5%), 2 (2.5%), 2 (2.5%), 5 (6.3%), 2 (2.5%) and 2 (2.5%) of 79 CRCs with MSI-H, respectively. However, we found no such mutations in microsatellite stable (MSS) cancers in the nucleotide repeats. We also studied ITH for the frameshift mutations in 16 cases of CRCs and detected that the frameshift mutations of NRIP1, DOK1, PARD3, PRKCI, DLGAP3, ULK1 and ULK2 showed regional ITH in 5 (31.3%), 2 (12.5%), 0 (0%), 0 (0%), 1 (6.3%), 1 (6.3%) and 3 (18.8%) cases, respectively. Our data exhibit that candidate cancer-related genes NRIP1, DOK1, PARD3, PRKCI, DLGAP3, ULK1 and ULK2 harbor mutational ITH as well as the frameshift mutations in CRC with MSI-H. Also, the results suggest that frameshift mutations of these genes might play a role in tumorigenesis through their inactivation in CRC.

Rare frameshift mutations of putative tumor suppressor genes CSMD1 and SLX4 in colorectal cancers

CUB and Sushi multiple domains 1 (CSMD1) is mapped to chromosome 8p23 that is frequently deleted in head and neck cancers [1]. Decreased expression of CSMD1 is also common in lung, breast, head/neck and skin cancers [2]. Structure-specific endonuclease subunit (SLX4) encodes a Fanconi anemia-related protein that is required for repair of specific types of DNA lesions and critical for cellular responses to replication fork failure [3]. SLX4 interaction with SUMO plays an important role in maintenance of genome [4], instability of which is a key feature of cancer. Together, these data suggest that both CSMD1 and SLX4 might have tumor suppressor gene (TSG) activities in cells [1,2.4]. However, it remains unknown whether inactivating frameshift mutations of their inactivating mutations are common in colorectal cancer (CRC). About 10–20% third of CRC are classified as high microsatellite instability (MSI-H) cancers [5]. Lots of TSGs harbor frameshift mutations at monocleotide repeats in MSI-H cancers [5]. In the human genome database, we observed that both CSMD1 and SLX4 genes possess nucleotide repeats in coding sequences that might be mutated in MSI-H cancers. In this study, we analyzed a T7 repeat in CSMD1 and a C7 repeat in SLX4 by polymerase chain reaction (PCR)-based single strand conformation polymorphism (SSCP) assay. In this study, we used 124 CRCs. The CRCs were 79 CRCs with MSI-H and 45 CRCs with microsatellite stable (MSS). In cancer tissues, malignant cells and normal cells were selectively procured by microdissection [6]. Radioisotope ([P]dCTP) was incorporated into the PCR products, which were subsequently displayed in SSCP gels and analyzed with direct DNA sequencing [6]. On SSCP, we observed aberrant bands of CSMD1 in one case of CRCs and SRPK1 in two cases of CRCs. Sanger sequencing of them revealed that they were frameshift mutations of CSMD1 and SLX4 (Fig. 1). DNA from the patients’ normal tissues showed no evidence of mutation in both SSCP and Sanger sequencing, indicating the mutations had risen somatically. These mutations were all deletions of one base in the T7 repeat of CSMD1 (c.9722delT) and the C7 repeat of SLX4 (c.140delC) that would result in frameshift mutations (p.Phe3241SerfsX21 of CSMD1 and p.Pro469HisfsX4 of SLX4) (Fig. 1). They were detected in two CRCs (3/79: 3.8%) with MSI-H, but not in those with MSS (0/45). The frameshift mutations of CSMD1 and SLX4 detected in the present study would result in premature stops of amino acid synthesis in CSMD1 and SLX4 proteins, respectively, and hence resemble a typical inactivating mutation. Based on the earlier data that showed TSG activities of PLA2R1

WRN, the Werner Syndrome Gene, Exhibits Frameshift Mutations in Gastric and Colorectal Cancers

To the Editor Werner syndrome, an autosomal recessive disorder causing premature aging, is caused by truncating mutations in Werner syndrome gene (WRN) that encodes a DNA helicase with exonuclease activity [1]. Patients with Werner syndrome have an increased cancer incidence as well, suggesting that the lack normal WRN function affects tumorigenesis [2]. Both helicase and exonuclease activities ofWRN protein contribute to DNA repair in cells [3]. Also, cells with defective WRN show genomic instability [4]. These features are frequently observed in cancers, suggesting a possibility of WRN gene alterations in cancers. However, it remains unknown whether inactivating mutation of WRN is common in gastric cancer (GC) and colorectal cancer (CRC). About one third of GC and CRC are classified as high microsatellite instability (MSI-H) cancers [5]. Many tumor suppressor genes such as BAX and TGFBR2 harbor frameshift mutations at monocleotide repeats in MSI-H cancers [5]. In the human genome database, we observed that WRN gene possesses nucleotide repeats in coding sequences that might be mutated in MSI-H cancers. In this study, we analyzed an A8 repeat in exon 2 and an A7 repeat in exon 28 of WRN by polymerase chain reaction (PCR)-based single strand conformation polymorphism (SSCP) assay. In this study, we used 79 GCs and 124 CRCs. The GCs were 34 GCs with MSI-H, 45 GCs with microsatellite stable/low MSI (MSS/ MSI-L), 79 CRCs with MSI-H and 45 CRCs with MSS/ MSI-L. In cancer tissues, malignant cells and normal cells were selectively procured by microdissection. Radioisotope ([P]dCTP) was incorporated into the PCR products, which were subsequently displayed in SSCP gels and analyzed with direct DNA sequencing [6]. Additionally, to see whether the WRN mutations possess intra-tumor heterogeneity (ITH) that contributes to tumor aggressiveness [7], we studied 16 CRCs with four to seven regional biopsies per CRC. In the SSCP, we found aberrantly migrating bands in three GCs and three CRCs, but not in their matched normal samples. DNA sequencing analysis confirmed that the aberrant bands represented WRN somatic mutations, which consisted of frameshift mutations by a deletion (c.15delA (p. Lys5AsnfsX15)) in exon 2 and another deletion (c.3382delA (p. Ser1128ValfsX34)) in exon 28 within the repeat (Table 1). The mutations were found in GCs (3/34, 8.8%) and CRCs (3/79, 3.8%) with MSI-H (3/113, 5.3%), but not in GCs (0/45) and CRCs (0/45) with MSS/MSI-L (Fisher’s exact test, p = 0.028). The frameshift mutation in exon 2 showed ITH in one of 16 CRCs (6.3%). A CRC (# 41) showed the c.15delA mutation in three regional biopsies (#41–1, 41–3 and 41–4), but there was no such mutation in the two regional biopsies (#41–6 and 41–7) (Fig. 1). We could not find any significant histological difference among the ITH regions in this case. Cancer-related functions (DNA repair and maintenance of genomic stability) and increased cancer incidence in Werner syndrome [1] led to us to analyze inactivating mutations of WRN gene in GC and CRC. In the present study, we found that six cases (5.3%) of GCs and CRCs with MSI-H harbored WRN frameshift mutations, indicating that WRN is mutated Ju Hwa Lee and Sung Soo Kim contributed equally to this work.

Inactivating frameshift mutation of putative tumor suppressor genes PLA2R1 and SRPK1 in gastric and colorectal cancers

To the Editor: Phospholipase A2 receptor 1 (PLA2R1) encodes a transmembrane receptor that plays a role in clearance of phospholipase A2, thereby inhibiting the action of phospholipaseA2 (1). Activation of PLAR1 leads to cellular senescence, apoptosis and inhibition of cell transformation, suggesting its tumor suppressor gene (TSG) activity in cells (1). Decreased PLA2R1 expression is found in many cancers, including colorectal cancer (CRC) (1). Hypermethylation of PLA2R1 promoter that may result in the decreased expression is identified in kidney cancers (1). SRSF protein kinase 1 (SRPK1) encodes a serine/arginine protein kinase specific for the SR (serine/arginine-rich domain) family of splicing factors (2). SRPK1 binds to AKT and either activates or inactivates AKT functions, thus performing dual opposing functions (oncogenic or TSG) depending on cellular context. For example, ablation of SRPK1 induces cellular transformation, confirming its TSG function (3). However, it remains unknown whether inactivating frameshift mutations of PLA2R1 and SRPK1 are common in gastric cancer (GC) and CRC. About one third of GC and CRC are classified as high microsatellite instability (MSI-H) cancers (4). Many TSGs harbor frameshift mutations at monocleotide repeats in MSI-H cancers (4). In the human genome database, we observed that both PLA2R1 and SRPK1 genes possess nucleotide repeats in coding sequences that might be mutated in MSI-H cancers. In this study, we analyzed an A8 repeat in PLA2R1 exon 28 and an A7 repeat in SRPK1 exon 7 by polymerase chain reaction (PCR)-based single strand conformation polymorphism (SSCP) assay. We did not analyze other possible inactivating variants such as nonsense or splicing variants in this study. Previously, the A8 of PLA2R1 and the A7 of SRPK1 harbored deletion mutations in 3 (one GC, one CRC and one melanoma) and one (one CRC) cases, respectively (COSMIC database). In this study, we used 79 GCs and 124 CRCs. The GCs were 34 GCs with MSI-H, 45 GCs with microsatellite stable/low MSI (MSS/MSI-L), 79 CRCs with MSI-H and 45 CRCs with MSS/MSI-L. In cancer tissues, malignant cells and normal cells were selectively procured by microdissection (5). Radioisotope ([P]dCTP) was incorporated into the PCR products, which were subsequently displayed in SSCP gels and analyzed with direct DNA sequencing (5). On SSCP, we observed aberrant bands of PLA2R1 in two cases of CRC and SRPK1 in one case of GC. Sanger sequencing of them revealed that they were frameshift mutations of PLA2R1 and SRPK1. DNA from the patients’ normal tissues showed no evidence of mutation in both SSCP and Sanger sequencing, indicating the mutations had risen somatically. These mutations were all deletions of one base in theA8 repeat of PLA2R1 (c.4113delA) and the A7 repeat of SRPK1 (c.574delA) that would result in frameshift mutations (p.Gly1372AlafsX32 of PLA2R1 and p.Ile192LeufsX32 of SRPK1). They were detected in two CRCs (2/79: 2.5%) and one GC (1/34: 2.9%) with MSI-H, but not in those with MSS (0/90). The frameshift mutation detected in the current study would result in premature stops of amino acid synthesis in PLA2R1 and SRPK1 proteins and hence resembles a typical inactivating mutation. Both of these would be expected to lead to premature stop codons in the mRNAs but not necessarily the production of a truncated protein as these transcripts are more likely to be degraded by nonsense mediated decay. Based on the previous reports that showed TSG activities of PLA2R1 and SRPK1 in cells (1,3), these frameshift mutations could contribute to cancer development by inhibiting their TSG activities. However, the incidence of the mutations is very low and identified only in MSI-H cancers. Our data may suggest that frameshift mutations of PLA2R1 and SRPK1 may not play a principal role in inhibiting their TSG functions in GC and CRC.