Deletion in HSP110 T17: correlation with wild-type HSP110 expression and prognostic significance in microsatellite-unstable advanced gastric cancers

    Deletion in HSP110 T 17 : Correlation with wild-type HSP110 expression and prognostic significance in microsatellite-unstable advanced gastric cancers Kyung-Ju Kim, Tae Hun Lee, Jung Ho Kim, Nam-Yun Cho, Woo Ho Kim, Gyeong Hoon Kang PII: DOI: Reference:

S0046-8177(17)30268-X doi: 10.1016/j.humpath.2017.08.001 YHUPA 4291

To appear in:

Human Pathology

Received date: Revised date: Accepted date:

31 March 2017 12 July 2017 4 August 2017

Please cite this article as: Kim Kyung-Ju, Lee Tae Hun, Kim Jung Ho, Cho Nam-Yun, Kim Woo Ho, Kang Gyeong Hoon, Deletion in HSP110 T17 : Correlation with wildtype HSP110 expression and prognostic significance in microsatellite-unstable advanced gastric cancers, Human Pathology (2017), doi: 10.1016/j.humpath.2017.08.001

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Deletion in HSP110 T17: Correlation with wild-type HSP110 expression and

RI P

T

prognostic significance in microsatellite-unstable advanced gastric cancers

SC

Running title: Deletion of HSP110 T17 in MSI-high advanced gastric cancers

Kyung-Ju Kim1,2, Tae Hun Lee1, Jung Ho Kim1,3, Nam-Yun Cho1, Woo Ho Kim3 and Gyeong

1

MA NU

Hoon Kang*,1,3

Laboratory of Epigenetics, Cancer Research Institute, Seoul National University College of

ED

Medicine, 103 Daehak-ro, Seoul 03080, Republic of Korea; 2Department of Pathology,

PT

Yeungnam University College of Medicine, 170 Hyeonchung-ro, Daegu 42415, Republic of Korea, and 3Department of Pathology, and Seoul National University College of Medicine,

AC

CE

103 Daehak-ro, Seoul 03080, Republic of Korea

*Correspondence:

Gyeong Hoon Kang

Department of Pathology, Seoul National University College of Medicine 103 Daehak-ro, Jongno-gu, Seoul, 110-744, South Korea 82-2-740-8263 [email protected]

1

ACCEPTED MANUSCRIPT Abstract Deletion of the HSP110 T17 mononucleotide repeat has recently been identified as a

T

prognostic marker that is correlated with wild-type HSP110 (HSP110wt) expression in

RI P

microsatellite instability-high (MSI-H) colorectal cancers. The aim of this study was to assess the correlation between deletion of the HSP110 T17 repeat and expression of HSP110wt,

SC

using DNA testing and immunohistochemistry and to determine the prognostic implications

MA NU

of HSP110 T17 deletion in MSI-H advanced gastric cancers (GCs). The status of HSP110wt expression was evaluated by immunohistochemistry using an HSP110wt-specific antibody in 142 MSI-H advanced GCs. The size of the HSP110 T17 repeat deletion was analyzed in 96 MSI-H advanced GCs; deletions were divided into small (0–2 bp) and large deletions (3–5

ED

bp). Low and high expressions of HSP110wt were detected in 38 (26.8%) and 104 (73.2%) of the 142 cases, respectively. The HSP110 T17 deletion was observed in 45 (46.9%) of the 96

PT

MSI-H GC samples. Tumors with high expression of HSP110wt showed a tendency to have

CE

small or no deletion of HSP110 T17. In Kaplan-Meier survival analysis, tumors with a large HSP110 T17 deletion was associated with favorable overall survival and disease free survival

AC

compared with those with small/no deletion of HSP110 T17. However, HSP110 T17 deletion size was not an independent prognostic factor in multivariate analysis. In summary, deletion of the HSP110 T17 repeat was frequently observed in MSI-H GCs, and HSP110 T17 deletion size was inversely correlated with HSP110wt expression status. Large HSP110 T17 was not a prognostic indicator in MSI-H GCs.

Keywords: gastric cancer, microsatellite instability, HSP110, HSP110 T17 deletion, prognosis

2

ACCEPTED MANUSCRIPT Introduction Gastric cancer (GC) is the fifth most common malignancy and is the third most common

T

cause of death in the world [1]. GC is a heterogeneous disease in terms of the molecular

RI P

mechanism of carcinogenesis. Among the potential mechanisms, microsatellite instability (MSI) is characterized by alteration of microsatellite length and is induced by defects in the

SC

DNA mismatch repair (MMR) system, which leads to frameshift mutations of many cancer-

MA NU

related genes that harbor microsatellites in their coding exons [2]. MSI-high (MSI-H) GCs account for approximately 10% of all GCs, and are known to have unique clinicopathological features such as better prognosis, preponderance of antral location, intestinal type according to Lauren classification, differentiated type, mucinous and medullary histologic pattern and

ED

high density of tumor infiltrating lymphocytes (TILs), in comparison with MSI-low or microsatellite-stable (MSS) GCs [3,4]. In addition, increasing evidence indicates that 5-

PT

fluorouracil (5-FU)–based adjuvant therapy shows no or little efficacy in the MSI subset of

CE

tumors compared with MSS tumors, although the issue remains controversial [5-8]. Expression of heat shock proteins (HSPs) is induced by many stress conditions, such as fever,

AC

irradiation, oxidative stress, and cancer [9-11]. HSPs can be classified into several protein families according to their molecular weight, including HSP25/HSP27, HSP40, HSP60, HSP70, and HSP110 (also known as HSP105) [12]. HSPs were reported to play a cytoprotective role under stress conditions via several mechanisms, including as molecular chaperones that facilitate the proper folding and anti-aggregation of proteins, and as antiapoptotic proteins that promote the survival of target cells [11]. Increased expression of HSPs in various cancer cells has been documented in many previous studies, and high expression of HSPs has been shown to be correlated with poor prognosis and tumor metastasis in human 3

ACCEPTED MANUSCRIPT cancers of various tissue types, including lung, breast, ovary, and colorectal cancers [10,12]. Recently, Dorard et al. reported that HSP110 is a putative target gene for mutation in MSI-H

T

colorectal cancers (CRCs). They showed that, in primary CRCs and colon cancer cell lines, a

RI P

mutant isoform of HSP (HSP110ΔE9) was produced by exon 9 skipping as a result of a deletion in the T17 mononucleotide repeat, located in intron 8 of HSP110 [13]. The

SC

HSP110ΔE9 mutant protein, which lacks the substrate-binding domain that is present in the

MA NU

wild-type HP110 (HSP110wt) protein, was reported to perform a beneficial role for MSI-H CRC patients by dose-dependently inhibiting the anti-apoptotic properties and the chaperone activity of HSP110wt in tumor cells [14]. Our previous study revealed that there was an inverse correlation between the length of the deletion in the HSP110 T17 repeat and the

ED

expression status of HSP110wt, and that decreased expression of HSP110wt could be a protective prognostic marker in MSI-H CRCs [15]. Furthermore, overexpression of the

PT

HSP110ΔE9 protein was shown to be able to sensitize cancer cells to anti-cancer agents, such

CE

as 5-fluorouracil (5-FU) and oxaliplatin, which are routinely used in adjuvant treatment of advanced CRCs and GCs [13,14].

AC

Thus, although almost all studies regarding the prognostic and predictive role of decreased HSP110wt expression have been confined to MSI-H CRCs, we hypothesized that similar results might be observed in MSI-H GCs because MSI-H CRCs and GCs share the same molecular mechanism of carcinogenesis. In this study, we aimed to identify whether HSP110 mutations are detected in MSI-H GCs as well, and to assess the correlation between deletion of the HSP110 T17 repeat and expression of HSP110wt, using DNA fragment analysis and immunohistochemistry. We also evaluated the prognostic implications of the HSP110 T17 deletion and its association with various clinicopathological parameters in MSI-H GCs. 4

ACCEPTED MANUSCRIPT Materials and methods Patients and specimens

T

A total of 142 formalin-fixed paraffin-embedded tissue samples of GC were collected from

RI P

the pathology archives of Seoul National University Hospital, Seoul, Korea. All tumor samples originated from patients who underwent surgical resection at Seoul National

SC

University Hospital between 2004 and 2009. During that time, 1,706 GC tissue samples were

MA NU

subjected to MSI analysis in the molecular pathology laboratory and 153 (8.7%) were diagnosed as MSI-H GCs. Among them, 142 cases were available for tissue microarray (TMA)-based immunohistochemical analysis. Of these 142 cases, 96 samples were subjected to further molecular study to determine the size of the deletion in the T17 mononucleotide

ED

repeat, located in HSP110 intron 8. MSI status was evaluated at the loci recommended by the National Cancer Institute: BAT25, BAT26, D2S123, D5S346, and D17S250. We defined

PT

tumors as MSI-H when they had two or more unstable markers, MSI-L when they had one

CE

unstable marker, and MSS when no markers were unstable.

AC

Clinicopathologic data

Clinicopathological information, including sex, age, tumor location, tumor differentiation, TNM stage, Ming classification, Lauren classification, mucinous histology, signet ring cell histology, medullary histology, papillary histology, serrated histology, Crohn-like lymphoid reaction, peritumoral lymphoid reaction, lymphovascular invasion, perineural invasion, date of surgery, date of death, tumor recurrence, and the last follow-up, were collected by reviewing the electronic medical records or by histologic evaluation of hematoxylin and eosin-stained slides. The presence or absence of mucinous histology, signet ring cell histology, 5

ACCEPTED MANUSCRIPT medullary histology, papillary histology and serrated histology were evaluated, regardless of their extent. The evaluation of Crohn-like lymphoid reaction was based on Graham–

T

Appelman criteria [16] in which three categories were defined as follows; absent (no or one

RI P

single lymphoid aggregate (LA) in all sections); mild (occasional LAs with rare or absent germinal centers); marked (numerous LAs with frequent germinal centers). Peritumoral

SC

lymphoid reaction was assessed according to three-tiered grading system; absent, mild and

MA NU

marked. Overall survival (OS) was defined as the duration from the date of surgery to death or the last follow-up before December 31, 2015, and disease-free survival (DFS) was defined as the duration from the date of surgery to tumor recurrence, death, or the last follow-up before December 31, 2015. The mean OS was 83 months (ranging from 1 to 139 months) and

ED

the average DFS time was 30 months (ranging from 1 to 132 months). In this cohort, death and recurrence were observed in 43 (30.3%) and 31 (21.8%) patients out of all MSI-H

PT

advanced GC patients, respectively. Tumor staging and histological grading were assessed

CE

according to the AJCC Staging Manual, seventh edition [17]. This study was approved by the

AC

institutional review board of Seoul National University Hospital.

HSP110 T17 mononucleotide repeat deletion analysis Genomic DNA was isolated from formalin-fixed, paraffin-embedded (FFPE) tissues and fluorescence capillary electrophoresis-based DNA fragment analysis of the mononucleotide repeat located in HSP110 T17 intron 8 was performed as previously described [15]. DNA extracted from normal gastric mucosal tissues (n = 2) and a primary MSI-H CRC cell line (LoVo) was also examined to determine HSP110 T17 repeat size. The primer sequence used for polymerase chain reaction (PCR) amplification was as follows: forward, 5ʹ6

ACCEPTED MANUSCRIPT TGAAAACCCTGTCCATCCAT-3ʹ; reverse, FAM-5ʹ-CTTTAAATGCCGGGGAAAGT-3ʹ. PCR was performed using MightyAmp DNA polymerase Ver. 2 (Takara, Shiga, Japan) under

T

the following conditions: initial pre-denaturation at 98°C for 2 min followed by 40 cycles of

RI P

98°C for 10 s, 57°C for 15 s and 68°C for 30 s, and a final extension of 68°C for 10 min. The size of the deletion in the HSP110 T17 repeat was determined using Peak Scanner Software

SC

(Applied Biosystems, Foster City, CA, USA), by subtracting the size of the DNA fragment

MA NU

with the altered peak from the normal peak size. Representative peak scanner images are presented in Figure 1.

GC tissue microarray and immunohistochemistry

ED

We constructed TMAs following a previously described method [2]. Two representative tumor regions, tumor center and invasive front, were selected and 2-mm tissue cores were

PT

extracted from the donor tissue blocks and transferred to new recipient TMA blocks using a

CE

trephine apparatus. Next, 4-μm-thick TMA sections of the tissues were immunostained using antibodies against MLH1 (DAKO, Glostrup, Denmark), MSH2 (Invitrogen, Camarillo, CA,

AC

USA), PMS2 (Ventana Medical Systems, Tucson, AZ, USA), MSH6 (Ventana Medical Systems, Tucson, AZ, USA), and HSP110 (Leica Biosystems, Newcastle-upon-Tyne, UK). The HSP110 antibody used in this study was specific for HSP110wt but cannot detect HSP110ΔE9, which does not contain the C-terminus of HSP110wt.

Evaluation of HSP110wt and DNA MMR protein expression HSP110wt expression was assessed in a semi-quantitative manner [15]. The intensity of nuclear-to-cytoplasmic staining in tumor cells was initially measured on a scale of 0 to 3 as 7

ACCEPTED MANUSCRIPT follows: negative (score 0), weak (score 1), moderate (score 2), and strong (score 3). The expression of HSP110wt was defined as positive when at least 5% of the tumor cells were

T

immunostained, and staining of less than 5% of the tumor cells was defined as negative.

RI P

When the same tumor exhibited two different HSP110wt expression scores, the higher score was selected. Figure 2 displays representative images of GC cases that were scored from 0 to

SC

3. Tumor cells without nuclear staining for DNA MMR enzymes (MLH1, MSH2, PMS2, or

MA NU

MSH6) in the presence of positive staining in adjacent tumor cell nuclei (i.e., fibroblasts or lymphocytes) were considered deficient for the specific MMR protein [18]. Tumors with expression of the specific MMR protein in less than 5% of tumor cells with nuclear staining in the presence of clear staining of the internal control were also considered deficient for the

Statistical analysis

PT

ED

specific MMR protein.

CE

Pearson’s chi-square test or Fisher’s exact test (for cases with an n-value <10) was performed

AC

for comparison of categorical variables. Differences in mean value were analyzed using 2tailed Student’s t-test. OS and DFS were analyzed using the Kaplan–Meier survival method with log-rank test. Of the total of 142 cases, patients diagnosed with stage IV cancer (n = 12) were excluded from the DFS survival analysis. Statistical analysis was performed using the Statistical Package for the Social Sciences (SPSS) software program (version 20.0; Chicago, IL, USA). All p-values were two-sided, and p < 0.05 was considered statistically significant.

8

ACCEPTED MANUSCRIPT Results

T

Evaluation of HSP110wt expression of MSI-H GCs

RI P

Immunohistochemical staining for HSP110wt was performed in 142 MSI-H GC samples and the expression status of HSP110wt was evaluated and scored from 0 to 3. A frequency

SC

histogram of the scoring is shown in Figure 3A. Next, we divided the expression scores into

MA NU

low (score 0 and score 1) and high expression (score 2 and score 3) groups to evaluate the correlation between HSP110wt expression and HSP110 T17 mononucleotide repeat deletion status. Low and high expression of HSP110wt was detected in 55 (38.7%) and 87 (61.3%) of

ED

142 cases, respectively.

PT

The association between HSP110 T17 deletion and HSP110wt expression

CE

For allelic profiling of this repeat sequence using fluorescence capillary electrophoresis, normal gastric mucosal tissues (n = 2) and a MSI-H CRC cell line (LoVo cell line) were

AC

included as negative and positive controls for HSP110 T17 deletion, respectively. As expected, none of the normal mucosae showed HSP110 T17 deletion, and a 4-base pair (bp)-long deletion of HSP110 T17 was observed in the LoVo cell line. The HSP110 T17 deletion was observed in 45 (46.9%) cases of the 96 MSI-H GC samples, and the deletion size ranged from 1 bp to 5 bp (Figure 3B). More than half of MSI-H GCs (n = 51, 53.1%) showed no deletion of the HSP110 T17 repeat. When we investigated the association between HSP110wt expression score and the deletion size of the HSP110 T17 repeat, tumors with a high HSP110wt expression score showed a tendency toward having deletion of small (1 or 2bp) or no deletion of HSP110 T17 (Figure 3C). Then we divided the HSP110 T17 deletion size into 9

ACCEPTED MANUSCRIPT two subgroups, large deletions (HSP110-dL) (deletion size ≥3 bp; n = 36) or small/no deletions (HSP110-dS) (0≤ deletion size <3 bp; n = 60), for further analysis. This cut-off

T

value was adopted because it provided the maximal survival difference between the two

SC

RI P

groups.

MA NU

Differential clinicopathological and prognostic features according to HSP110 T17 deletion size

No significant difference in clinicopathological parameters was observed between the HSP110-dS and HSP110-dL groups, except in lymphatic invasion (p = 0.033) and HSP110wt

ED

expression (p = 0.015) (Table 1). There was no significant difference in histological features, between the two groups, including mucinous, signet ring cell, medullary, papillary and

PT

serrated histology, as well as Crohn-like lymphoid reaction and peritumoral lymphoid

CE

reaction. However, as predicted, tumors with a HSP110-dL showed a relatively higher

AC

frequency of low HSP110wt expression than those with a HSP110-dS. In Kaplan–Meier survival analysis, the HSP110-dL group was significantly correlated with favorable survival for OS and DFS compared to the HSP110-dS group (p = 0.037 for OS; p = 0.039 for DFS) (Figure 4A & 4B). However, HSP110 T17 deletion was not an independent prognostic factor for MSI-H GCs in multivariate analysis (hazard ratio = 1.863; 95% confidence interval, 0.727 to 4.769; p = 0.195 for OS; hazard ratio = 2.891; 95% confidence interval, 0.829 to 10.177; p = 0.096 for DFS) (Table 2). In addition, no difference in OS or DFS was observed between the low and high HSP110wt expression groups (p = 0.336 for OS; p = 0.551 for DFS) (Figure 4C & 4D). To evaluate the impact of shortening of this 10

ACCEPTED MANUSCRIPT mononucleotide repeat in response to the chemotherapeutic agent, we performed Kaplan– Meier survival analysis on groups who had been administered 5-FU-based adjuvant

T

chemotherapy (n = 37 out of 96 patients). However, no differences in survival dependent on

RI P

HSP110 T17 deletion size were observed in this group (p = 0.630 for OS; p = 0.873 for DFS,

MA NU

SC

data not shown).

Discussion

HSPs are molecular chaperones that are involved in the carcinogenesis of various cancers and that have recently received attention as possible primary targets for anti-cancer therapies

ED

[10,19]. Many HSPs belong to the large chaperon family and are classified by their molecular weight (i.e., HSP27, HSP70, HSP90, and HSP110). Among them, HSP110 has been

PT

recognized to be important in colorectal carcinogenesis, according to some recent studies

CE

[20,21]. Genes that harbor microsatellites in their coding exons are liable to frameshift mutations of the microsatellites due to MMR defects, whereas transcription of genes that

AC

contain microsatellites in intronic regions can undergo alteration of splicing patterns and subsequent exonal skipping provoked by MMR defects [22]. The HSP110 T17 mononucleotide repeat sequence located in intron 8 was reported to be frequently mutated in MSI-H tumors, and its functional activity has been actively studied in MSI-H CRCs. Deletion of the HSP110 T17 repeat sequence results in exon 9 skipping, which leads to production of a truncated protein (known as HSP110ΔE9) that does not contain the substrate-binding domain of the HSP110wt protein. Dorard et al. showed that large deletions (≥5 bp) of the HSP110 T17 repeat contribute to the detriment of the HSP110wt transcript, and that the ratio of 11

ACCEPTED MANUSCRIPT HSP110ΔE9/HSP110wt mRNA is increased in MSI-H CRCs [13]. Collura et al. speculated that as the size of the deletion in this repeat increases, the major gene product switches from

T

HSP110wt to HSP110ΔE9 [14]. Our previous immunohistochemical study, using an antibody

RI P

to HSP110wt, supported this explanation, because MSI-H CRCs with low expression of HSP110wt tended to show larger deletions of the repeat sequence than MSI-H CRCs with

SC

high expression of HSP110wt [15]. In the current study, we found that the correlation

MA NU

between the mutational status of HSP110 T17 and the expression status of HSP110wt protein is also present in MSI-H GCs. Through direct physical interaction between HSP110ΔE9 and HSP110wt at a ratio of 1:1, HSP110ΔE9 is considered to abrogate the normal anti-apoptotic and anti-aggregation activities of HSP110 in cancer cells and to block translocation of

ED

HSP110 from the cytosol to the nucleus in a dominant negative manner [14].

PT

Deletion of HSP110 T17 was observed in 45 of 96 cases (46.9%). This mutation rate is higher than that of other microsatellite targets in MSI-H GCs reported to date (i.e., TGFß-IIR, BAX)

CE

[23]. However, in contrast to previous studies of CRCs, in which all MSI-H CRCs displayed

AC

allelic deletion beyond the polymorphic zone (>3 bp) and a large proportion of CRC samples (88%) showed deletion of this noncoding repeat [13], more than half of MSI-H GCs (n = 51, 53.1%) showed no deletion of this repeat. In addition, the deletion size range was narrower (1 bp to 5 bp) than that of MSI-H CRC samples (1 bp to 7 bp) [14]. This result indicates that some altered microsatellites are not shared between MSI-H CRCs and GCs, although both MSI-H CRCs and GCs are driven by the MSI pathway. In addition, based on the finding that not all tumors with a large deletion in the HSP110 T17 repeat showed low HSP110wt expression, and not all tumors with low expression of HSP110wt had a large HSP110 T17

12

ACCEPTED MANUSCRIPT deletion, we infer that shortening of this repeat does not necessarily produce consistent

T

predictions of low levels of HSP110wt expression, and vice versa.

RI P

Previous studies showed that HSP110 induces tumor promotion in various ways and can be used as a marker of poor prognosis. Kimura et al. demonstrated that high nuclear HSP110

SC

expression is related to reduced overall survival and cancer progression in GC patients [24]. Berthenet et al. revealed that HSP110 promotes STAT3-dependent colon cancer growth

MA NU

through increased STAT3 phosphorylation and transcription factor activity, in an in vitro analysis using the MSI-H colon cancer cell line [25]. It was also shown that secretion of extracellular HSP110 from colorectal cancer cells was diminished by expression of the endogenous HSP110 inhibitor, HSP110ΔE9, and high HSP110wt expression skew the

ED

macrophage polarization towards a pro-tumoral (anti-inflammatory) phenotype that facilitates

PT

tumor progression in the tumor microenvironment [25]. Our previous immunohistochemical study also demonstrated that high HSP110wt expression in MSI-H CRCs is an independent

CE

adverse prognostic marker [14]. In the current study, Kapan-Meier survival analysis revealed

AC

that tumors with a large HSP110 T17 deletion was significantly associated with favorable OS (p = 0.037) and DFS (p = 0.039) compared with tumors with small/no deletion of HSP110 T17, although HSP110 T17 deletion size was not an independent prognostic factor for OS and DFS in multivariate analysis (p = 0.195 for OS; p = 0.096 for DFS). However, there was no survival difference according to the expression of HSP110wt in tumor cells. These inconsistent results between HSP110 T17 deletion status and HSP110wt expression status, regarding prognosis, might be attributed to the fact that evaluation of HSP110wt expression, not HSP110ΔE9, is an indirect and a relatively inaccurate method for estimation of the

13

ACCEPTED MANUSCRIPT genuine biological function of HSP110 T17 repeat deletion (CRCs, ≥5 bp vs. <5 bp; GCs, ≥3

T

bp vs. <3 bp).

RI P

An increased sensitivity to 5-FU-based chemotherapy in association with decreased HSP110wt transcription has been demonstrated by several researchers. The Duval group

SC

showed that expression of mutant HSP110 (HSP110ΔE9) and large deletions of HSP110 T17 can sensitize MSI-H CRC cells to 5-FU-based chemotherapy [13,14]. In addition, Kimura et

MA NU

al. showed that GCs with reduced nuclear HSP110 expression showed higher overall survival than GCs with increased HSP110 expression, among GC patients who received adjuvant chemotherapy [24]. These results indicate that mutant HSP110, lacking the normal chaperone function of wild-type HSP110, can enhance the response of tumor cells to 5-FU-based

ED

chemotherapy, regardless of the underlying genetic alteration. However, in the current study,

PT

there was no association between survival outcome and HSP110 T17 deletion status among patients with MSI-H GC who were treated with adjuvant chemotherapy. This finding has

CE

limited reliability because of the small number of patients studied (n = 37). Thus, the effect of

AC

HSP110 T17 repeat deletion on the response of cancer cells to chemotherapy should be clarified in a larger cohort. Analysis of IHC using TMA technique is a cost-effective method for rapid analysis of a large number of tumor tissues on the same slide under identical assay condition [26]. However, an obvious drawback of TMA technique is derived from intratumoral heterogeneity in expression of certain molecular markers. Intratumoral heterogeneity has been recognized to be an inherent characteristic of many types of cancers and a major concern in accurate diagnosis and molecularly targeted therapy. Many previous studies revealed that GC is also characterized by heterogenous expression of some molecular markers such as Human 14

ACCEPTED MANUSCRIPT epidermal growth factor receptor 2 (HER2) [27] and Fibroblast growth factor receptor 2 (FGFR2) [28]. In the present study, we evaluated HSP110wt expression in two representative

T

tissue cores which are relatively large size (2mm in diameter). Overall concordance was 90.1%

RI P

(n = 128) between two cores. In tumors with high expression of HSP110wt (n = 104), 14 cases (13.5%) showed discordance of HSP110wt expression between two cores. Although the

SC

proportion is minor, such a discordant expression of HSP110wt between two cores indicates

MA NU

that further study using more tissue cores or full tumor sections is warranted for elucidation of HSP110wt expression status as a prognostic marker in MSI-H GCs. In conclusion, we found that deletion of the HSP110 T17 repeat is frequently observed in MSI-H GCs, but frequency and range quite different from those of MSI-H CRCs. And we

ED

identified that the tumors with large HSP110 T17 deletions tended to show low expression of

PT

wild-type HSP110. In multivariate survival analysis, large deletion of HSP110 T17 was not an

CE

independent prognostic indicator of MSI-H GCs.

AC

CONFLICT OF INTEREST The authors declare to no conflict of interest.

FUNDING SOURCES This work was supported by grants from the National Research Foundation (NRF) funded by the Korean Ministry of Science, ICT and Future Planning (2011-0030049 and 2016M3A9B6026921) and the Korean Ministry of Education (2016R1D1A1A02937130), a grant from the Priority Research Centers Program through the NRF (2009-0093820), and a

15

ACCEPTED MANUSCRIPT grant from the Korea Health Technology R&D Project through the Korea Health Industry

AC

CE

PT

ED

MA NU

SC

RI P

T

Development Institute funded by the Korean Ministry of Health and Welfare (HI14C1277).

16

ACCEPTED MANUSCRIPT References [1] Ferlay J, Shin HR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden of

T

cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010; 127, 2893-2917.

RI P

[2] Kim KJ, Lee KS, Cho HJ, et al. Prognostic implications of tumor-infiltrating FoxP3+ regulatory T cells and CD8+ cytotoxic T cells in microsatellite-unstable gastric cancers. Hum Pathol 2014; 45, 285-293.

SC

[3] Kim H, An JY, Noh SH, Shin SK, Lee YC, Kim H. High microsatellite instability predicts good prognosis in intestinal-type gastric cancers. J Gastroenterol Hepatol 2011; 26, 585-592.

MA NU

[4] Sugimoto R, Sugai T, Habano W, et al. Clinicopathological and molecular alterations in early gastric cancers with the microsatellite instability-high phenotype. Inter J Cancer. 2016; 138, 16891697.

[5] Velho S, Fernandes MS, Leite M, Figueiredo C, Seruca R. Causes and consequences of microsatellite instability in gastric carcinogenesis. World J Gastroenterol 2014; 20, 16433-16442. [6] Oki E, Kakeji Y, Zhao Y, et al. Chemosensitivity and survival in gastric cancer patients with

ED

microsatellite instability. Ann Surg Oncol 2009; 16, 2510-2515. [7] An JY, Kim H, Cheong JH, Hyung WJ, Kim H, Noh SH. Microsatellite instability in sporadic

PT

gastric cancer: its prognostic role and guidance for 5-FU based chemotherapy after R0 resection. Int J Cancer 2012; 131, 505-511.

[8] Carethers JM, Chauhan DP, Fink D, et al. Mismatch repair proficiency and in vitro response to 5-

CE

fluorouracil. Gastroenterology 1999; 117, 123-131. [9] Goldstein MG, Li Z. Heat-shock proteins in infection-mediated inflammation-induced

AC

tumorigenesis. J Hematol Oncol 2009; 2, 5. [10] Seigneuric R, Mjahed H, Gobbo J, et al. Heat shock proteins as danger signals for cancer detection. Front Oncol 2011; 1, 37. [11] Rappa F, Farina F, Zummo G, et al. HSP-molecular chaperones in cancer biogenesis and tumor therapy: an overview. Anticancer Res 2012; 32, 5139-5150. [12] Oda T, Morii E, Inoue M, Ikeda J, Aozasa K, Okumura M. Prognostic significance of heat shock protein 105 in lung adenocarcinoma. Mol Med Rep 2009; 2, 603-607. [13] Dorard C, de Thonel A, Collura A, et al. Expression of a mutant HSP110 sensitizes colorectal cancer cells to chemotherapy and improves disease prognosis. Nat Med 2011; 17, 1283-1289. [14] Collura A, Lagrange A, Svrcek M, et al. Patients with colorectal tumors with microsatellite instability and large deletions in HSP110 T17 have improved response to 5-fluorouracil-based chemotherapy. Gastroenterology 2014; 146, 401-411.e401. 17

ACCEPTED MANUSCRIPT [15] Kim JH, Kang GH. Molecular and prognostic heterogeneity of microsatellite-unstable colorectal cancer. World J Gastroenterol 2014; 20, 4230-4243. [16] Kim JH, Kim KJ, Bae JM, et al. Comparative validation of assessment criteria for Crohn-like

T

lymphoid reaction in colorectal carcinoma. J Clin Pathol. 2015; 68, 22-28.

RI P

[17] Edge SB, Byrd DR, Compton CC, et al. AJCC Cancer Staging Manual. 7th ed. New York: Springer-Verlag; 2009. p. 117-26.

[18] Kim KJ, Lee TH, Cho NY, Yang HK, Kim WH, Kang GH. Differential clinicopathologic features

SC

in microsatellite-unstable gastric cancers with and without MLH1 methylation. Hum Pathol 2013; 44, 1055-1064.

MA NU

[19] Jego G, Hazoume A, Seigneuric R, Garrido C. Targeting heat shock proteins in cancer. Cancer Lett 2013; 332, 275-285.

[20] Hwang TS, Han HS, Choi HK, et al. Differential, stage-dependent expression of Hsp70, Hsp110 and Bcl-2 in colorectal cancer. J Gastroenterol Hepatol 2003; 18, 690-700. [22] Slaby O, Sobkova K, Svoboda M, et al. Significant overexpression of Hsp110 gene during colorectal cancer progression. Oncol Rep 2009; 21, 1235-1241.

ED

[22] Yamamoto H, Imai K. Microsatellite instability: an update. Arch Toxicol 2015; 89, 899-921. [23] Shokal U, Sharma PC. Implication of microsatellite instability in human gastric cancers. Indian J

PT

Med Res 2012; 135, 599-613.

[24] Kimura A, Ogata K, Altan B, et al. Nuclear heat shock protein 110 expression is associated with

CE

poor prognosis and chemotherapy resistance in gastric cancer. Oncotarget 2016; 7, 18415-18423. [25] Berthenet K, Boudesco C, Collura A, et al. Extracellular HSP110 skews macrophage polarization in colorectal cancer. Oncoimmunology 2016; 5, e1170264.

AC

[26] Berlth F, Monig SP, Schlosser HA, et al. Validation of 2-mm tissue microarray technology in gastric cancer. Agreement of 2-mm TMAs and full sections for Glut-1 and Hif-1 alpha. Anticancer Res. 2014; 34, 3313-3320. [27] Abrahao-Machado LF, Jacome AA, Wohnrath DR, et al. HER2 in gastric cancer: comparative analysis of three different antibodies using whole-tissue sections and tissue microarrays. World J Gastroenterol. 2013; 19, 6438-6446. [28] Han N, Kim MA, Lee HS, Kim WH. Evaluation of Fibroblast Growth Factor Receptor 2 Expression, Heterogeneity and Clinical Significance in Gastric Cancer. Pathobiology. 2015; 82, 269279.

18

ACCEPTED MANUSCRIPT Figure legends

T

Figure 1. Representative peak scanner images of the HSP110 T17 repeat deletion analysis. (A)

RI P

A case with no HSP110 T17 deletion. (B) A case with small HSP110 T17 deletion (<3 bp). (C)

SC

A case with large HSP110 T17 deletion (≥3 bp).

Abbreviation: bp, base pair

MA NU

Figure 2. Immunohistochemical expression of HSP110wt. Representative cases scored as (A) 0, (B) 1, (C) 2, and (D) 3 (all figures, ×200).

Abbreviation: HSP110wt, wild-type HSP110

ED

Figure 3. Frequencies of HSP110wt expression and HSP110 T17 deletion. (A) Frequencies of

PT

HSP110wt immunohistochemical scores in 142 MSI-H GCs. (B) Frequencies of HSP110 T17 deletion size in 96 MSI-H GCs. (C) Correlation between HSP110wt expression and HSP110

CE

T17 in 96 MSI-H GCs.

AC

Abbreviations: HSP110wt, wild-type HSP110; MSI-H, microsatellite instability-high; GC, gastric cancer

Figure 4. Kaplan–Meier survival analysis with log-rank test of HSP110wt expression and HSP110 T17. (A) Survival curves for the duration of OS according to HSP110 T17 deletion size (No. of patients; HSP110-dS, 60; HSP110-dL, 36). (B) Survival curves for the duration of DFS according to HSP110 T17 deletion size (No. of patients; HSP110-dS, 53; HSP110-dL, 32). Patients who were diagnosed with stage IV TNM were excluded from the survival analysis for DFS. (C) Survival curves for the duration of OS according to HSP110wt 19

ACCEPTED MANUSCRIPT expression (No. of patients; low expression, 38; high expression, 104). (D) Survival curves for the duration of DFS according to HSP110 T17 deletion size (No. of patients; low

T

expression, 35; high expression, 95). Patients who were diagnosed with stage IV TNM were

RI P

excluded from the survival analysis for DFS.

SC

Abbreviations: HSP110wt, wild-type HSP110; OS, overall survival; HSP110-dS, small deletion of HSP110 T17; HSP110-dL, large deletion of HSP110 T17; DFS, disease free

AC

CE

PT

ED

MA NU

survival.

20

AC

Figure 1

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

21

AC

Figure 2

CE

PT

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

22

ED

MA NU

SC

RI P

T

ACCEPTED MANUSCRIPT

AC

CE

PT

Figure 3

23

Figure 4

AC

CE

PT

ED

MA NU

SC R

IP T

ACCEPTED MANUSCRIPT

24

ACCEPTED MANUSCRIPT Table 1. Associations between HSP110 T17 deletion and clinicopathological characteristics. HSP110 T17 deletion size Deletion size ≥3 bp

Case no.

Deletion size <3 bp

96

63.75 ± 11.83

Male

45

25 (41.7)

Female

51

35 (58.3)

16 (44.4)

Polypoid

4

2 (3.3)

2 (5.6)

Ulcerofungating

43

23 (38.3)

20 (55.6)

Ulceroinfiltrative

39

28 (46.7)

11 (30.6)

Diffuse/Unclassifiable

10

7 (11.7)

3 (8.3)

3 (5.0)

0 (0)

0 (0.0)

1 (2.8)

29

21 (35.0)

8 (22.2)

63

36 (60.0)

27 (75)

41

26 (43.3)

15 (41.7)

55

34 (56.7)

21 (58.3)

73

49 (81.7)

24 (66.7)

23

11 (18.3)

12 (33.3)

48

29 (48.3)

19 (52.8)

27

17 (28.3)

10 (27.8)

21

14 (23.3)

7 (19.4)

Absent

66

39 (65.0)

27 (75.0)

Present

30

21 (35.0)

9 (25.0)

Absent

90

54 (90.0)

36 (100.0)

Present

6

6 (10.0)

0 (0.0)

93

59 (98.3)

34 (94.4)

Mean ± SD

SC

Sex

3

Fundus

1

ED

Cardia Body

WD/MD Ming Infiltrative

Intestinal Diffuse

AC

Expanding

CE

PD/Other

PT

Antrum Tumor differentiation

Mixed

MA NU

Gross type

Location

RI P

Age (years)

Lauren

p-value

T

Parameter

67.18 ± 11.07

0.155

20 (55.6)

0.187

0.329

0.141

0.873

0.096

0.884

Mucinous histology 0.367

Signet ring cell histology 0.081

Medullary histology Absent

25

0.554

3

1 (1.7)

2 (5.6)

Absent

91

58 (96.7)

33 (91.7)

Present

5

2 (3.3)

3 (8.3)

Absent

93

58 (96.7)

Present

3

2 (3.3)

Absent

52

36 (60.0)

Mild to marked

44

24 (40.0)

20 (55.6)

Absent

71

45 (75.0)

26 (72.2)

Mild to marked

25

MA NU

ACCEPTED MANUSCRIPT

Present

15 (25.0)

10 (27.8)

19 (31.7)

15 (41.7)

41 (68.3)

21 (58.3)

79

51 (85.0)

28 (77.8)

17

9 (15.0)

8 (22.2)

61

35 (58.3)

26 (72.2)

35

25 (41.7)

10 (27.8)

73

45 (75.0)

28 (77.8)

23

15 (25.0)

8 (22.2)

56

30 (50.0)

26 (72.2)

40

30 (50.0)

10 (27.8)

I/II

54

30 (50.0)

24 (66.7)

III/IV

42

30 (50.0)

12 (33.3)

MLH1 loss/PMS2 loss

89

54 (90.0)

35 (97.2)

MSH2 loss/MSH6 loss

7

6 (10.0)

1 (2.8)

PMS2 loss only

0

0 (0.0)

0 (0.0)

MSH6 loss only

0

0 (0.0)

0 (0.0)

24

10 (16.7)

14 (38.9)

Papillary histology

T SC

Crohn-like lymphoid reaction

Peritumoral lymphoid reaction

Lymphatic invasion Absent

34

Present

62

Vascular invasion

ED

Absent Present

PT

Perineural invasion Absent

T2/3

AC

T4

CE

Present

RI P

Serrated histology

Tumor depth

0.360

35 (97.2)

>0.999

1 (2.8)

16 (44.4)

0.139

0.764

0.321

0.369

0.171

0.810

LN metastasis Absent Present

0.033

AJCC TNM stage 0.111

MMR status 0.250

HSP110wt expression Low (score 0/1+)

26

0.015

ACCEPTED MANUSCRIPT

High (score 2+/3+) 72 Values are presented as the number (%).

50 (83.3)

22 (61.1)

Abbreviations: bp, base pair; WD, well differentiated; MD, moderately differentiated; PD, poorly

T

differentiated; LN, lymph node; AJCC, American Joint Committee on Cancer; MMR, mismatch repair;

AC

CE

PT

ED

MA NU

SC

RI P

HSP110wt, wild-type HSP110

27

ACCEPTED MANUSCRIPT Table 2. Univariate and multivariate survival analysis of patients with MSI-H GCs.

Univariate analysis HR (95% CI)

Parameter

DFS p valu e

Multivariate analysis HR (95% CI)

Univariate analysis HR (95% CI)

Ming 1

Infiltrative

5.716 (1.77818.665)

0.39 5

1 1.710 (0.4975.881)

8.840 (4.08919.112)

III/IV

1

4.852 (1.73613.559)

Lymphatic invasion 0.00 2

1

Present

3.504 (1.5587.879)

Vascular invasion 1

Present

2.416 (1.1894.912)

CE

Perineural invasion

1

0.01 5

HSP110 T17 deletion size deletion size ≥3 bp deletion size <3 bp

0.00 1

2.676 (1.4644.890)

AC

Absent Present

4.572 (1.03820.145)

1

2.527 (1.0226.243)

1.602 0.6424.001)

0.978 (0.4452.147)

0.04 5

1.863 (0.7274.769)

0.02 5

1

1

1 3.389 (0.98711.631)

p valu e

HR (95% CI)

1

0.36 1

2.039 (0.4229.409)

1

0.00 8

1

0.14 3

2.617 (0.7229.483)

0.02 0

1

0.39 3

1.704 (0.5025.783)

0.05 5

1.993 (0.9854.032)

0.19 5

1

1

Multivariate analysis

3.586 (1.48114.202)

2.720 (1.1706.325)

0.95 6

1

<0. 001

2.627 (1.1316.098)

0.31 3

1

PT

Absent

1

5.534 (2.54412.039)

0.04 5

1

ED

Absent

0.00 3

MA NU

<0. 001

1

0.01 7

4.263 (1.29514.034)

AJCC TNM stage I/II

1

SC

0.00 4

Expanding

DFS p valu e

T

OS p valu e

RI P

OS

1

0.30 4

0.589 (0.2141.618)

0.05 2

1

0.09 6

2.891 (0.82910.177)

Abbreviations: OS, overall survival; DFS, disease free survival; MSI-H, microsatellite instability-high; GC, gastric cancer; HR, hazard ratio; AJCC, American Joint Committee on Cancer; bp, base pair

28

ACCEPTED MANUSCRIPT

Highlights 1. Deletion of the HSP110 T17 repeat was frequently observed in microsatellite unstable

RI P

T

gastric cancers.

2. HSP110 T17 deletion size was inversely correlated with wild type-HSP110 expression

SC

status in microsatellite unstable gastric cancers.

3. In multivariate survival analysis, large deletion of HSP110 T17 was not an independent

AC

CE

PT

ED

MA NU

prognostic indicator of MSI-H GCs.

29