Article Search
검색
검색 팝업 닫기

Metrics

Help

  • 1. Aims and Scope

    Gut and Liver is an international journal of gastroenterology, focusing on the gastrointestinal tract, liver, biliary tree, pancreas, motility, and neurogastroenterology. Gut atnd Liver delivers up-to-date, authoritative papers on both clinical and research-based topics in gastroenterology. The Journal publishes original articles, case reports, brief communications, letters to the editor and invited review articles in the field of gastroenterology. The Journal is operated by internationally renowned editorial boards and designed to provide a global opportunity to promote academic developments in the field of gastroenterology and hepatology. +MORE

  • 2. Editorial Board

    Editor-in-Chief + MORE

    Editor-in-Chief
    Yong Chan Lee Professor of Medicine
    Director, Gastrointestinal Research Laboratory
    Veterans Affairs Medical Center, Univ. California San Francisco
    San Francisco, USA

    Deputy Editor

    Deputy Editor
    Jong Pil Im Seoul National University College of Medicine, Seoul, Korea
    Robert S. Bresalier University of Texas M. D. Anderson Cancer Center, Houston, USA
    Steven H. Itzkowitz Mount Sinai Medical Center, NY, USA
  • 3. Editorial Office
  • 4. Articles
  • 5. Instructions for Authors
  • 6. File Download (PDF version)
  • 7. Ethical Standards
  • 8. Peer Review

    All papers submitted to Gut and Liver are reviewed by the editorial team before being sent out for an external peer review to rule out papers that have low priority, insufficient originality, scientific flaws, or the absence of a message of importance to the readers of the Journal. A decision about these papers will usually be made within two or three weeks.
    The remaining articles are usually sent to two reviewers. It would be very helpful if you could suggest a selection of reviewers and include their contact details. We may not always use the reviewers you recommend, but suggesting reviewers will make our reviewer database much richer; in the end, everyone will benefit. We reserve the right to return manuscripts in which no reviewers are suggested.

    The final responsibility for the decision to accept or reject lies with the editors. In many cases, papers may be rejected despite favorable reviews because of editorial policy or a lack of space. The editor retains the right to determine publication priorities, the style of the paper, and to request, if necessary, that the material submitted be shortened for publication.

Search

Search

Year

to

Article Type

Original Article

Split Viewer

Prognostic Significance of ARID1A Expression Patterns Varies with Molecular Subtype in Advanced Gastric Cancer

Jun Yong Kim1 , Cheol Keun Park1 , Songmi Noh2 , Jae-Ho Cheong3 , Sung Hoon Noh3 , Hyunki Kim1

1Department of Pathology, Yonsei University College of Medicine, 2Department of Pathology, CHA Gangnam Medical Center, CHA University, and 3Department of Surgery, Yonsei University College of Medicine, Seoul, Korea

Correspondence to: Hyunki Kim
ORCID https://orcid.org/0000-0003-2292-5584
E-mail hyunkikim@yuhs.ac

Jun Yong Kim and Cheol Keun Park contributed equally to this work as first authors.

Received: August 4, 2022; Revised: November 23, 2022; Accepted: November 23, 2022

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Gut Liver 2023;17(5):753-765. https://doi.org/10.5009/gnl220342

Published online February 15, 2023, Published date September 15, 2023

Copyright © Gut and Liver.

Background/Aims: AT-rich interactive domain 1A (ARID1A) is frequently mutated in gastric cancer (GC), especially Epstein-Barr virus (EBV)-associated and microsatellite instability high GC. The loss of ARID1A expression has been reported as a poor prognostic marker in GC. However, the relationships between ARID1A alteration and EBV-associated and microsatellite instability high GC, which are known to have a favorable prognosis, has hampered proper evaluation of the prognostic significance of ARID1A expression in GC. We aimed to analyze the true prognostic significance of ARID1A expression by correcting confounding variables.
Methods: We evaluated the ARID1A expression in a large series (n=1,032) of advanced GC and analyzed the relationships between expression pattern and variable parameters, including clinicopathologic factors, key molecular features such as EBV-positivity, mismatch repair protein deficiency, and expression of p53 and several receptor tyrosine kinases including human epidermal growth factor receptor 2, epidermal growth factor receptor, and mesenchymal-epithelial transition factor. Survival analysis of the molecular subtypes was done according to the ARID1A expression patterns.
Results: Loss of ARID1A expression was found in 52.5% (53/101) of mutL homolog 1 (MLH1)-deficient and 35.8% (24/67) of EBV-positive GCs, compared with only 9.6% (82/864) of the MLH1-proficient and EBV-negative group (p<0.001). The loss of ARID1A expression was associated only with MLH1 deficiency and EBV positivity. On survival analysis, the loss of ARID1A expression was associated with worse prognosis only in MLH1-proficient and EBV-negative GC. Multivariate analysis revealed that both loss of ARID1A and decreased ARID1A expression were independent worse prognostic factors in patients with advanced GC.
Conclusions: Only in MLH1-proficient and EBV-negative GC, the loss of ARID1A expression is related to poorer prognosis.

Keywords: Stomach neoplasms, ARID1A, Immunohistochemistry, Prognosis

AT-rich interactive domain 1A (ARID1A), a member of the ARID family, is a subunit of SWItch/sucrose non-fermentable complex that remodels histone-DNA interactions in reconstituted nucleosomes in an ATP-dependent manner.1 The SWItch/sucrose non-fermentable complex plays essential roles during lineage specification and in the maintenance of stem cell pluripotency, and malfunctioning ARID1A could potentially trigger persistent proliferative progenitor state.2-6 ARID1A-containing SWItch/sucrose non-fermentable complex is responsible for cell cycle and ARID1A-depletion leads to defective cell cycle checkpoint activation in response to DNA damage.7 Also, regulation of telomere reverse transcriptase by ARID1A through chromatin structure repression was reported.8 Recent studies revealed that the ARID1A gene is frequently mutated in a variety of cancers, such as ovarian clear cell carcinoma, endometrioid adenocarcinoma, and gastric carcinomas.9-12 The mutations of ARID1A found in these types of cancer are mostly insertion/deletion mutations that generate premature stop codon and thus lead to truncation of the ARID1A protein.9-12 Previous studies reported that the knockdown of ARID1A in gastric cancer (GC) cell lines promotes cell proliferation and the forced expression of ARID1A inhibits colony formation and cell growth.12,13 The loss of expression of ARID1A evaluated by immunohistochemistry (IHC) is correlated with the mutation status of ARID1A,11,14 although epigenetic silencing can also induce the loss of ARID1A expression.15 Intriguingly, in GC, mutation of ARID1A and loss of expression are closely associated with Epstein-Barr virus (EBV) positivity and microsatellite instability high (MSI-H) type.11,12,16-18 Several studies have examined the prognostic significance of alterations of ARID1A in GC.11,13,16-22 Among them, Wang et al.11 and Ibarrola-Villava et al.20 reported a favorable outcome for patients with ARID1A-mutated GC. In contrast, other studies reported that the loss of ARID1A expression was associated with poor survival outcome13,16,18,23 or had no prognostic role in GC.19,24 Therefore, the prognostic significance of ARID1A alteration still remains controversial. The conflicting reports of the significance of ARID1A expression on the prognosis of GC patients might result from possible confounding effects of EBV-associated and MSI-H GCs and/or limitations of sample size.

GC is one of the leading causes of cancer death worldwide; even though its incidence has been decreasing,25 the prognosis of advanced GC (AGC) is still dismal. Therefore, a more comprehensive investigation into the underlying molecular mechanisms is mandatory to develop more effective treatment modalities and to better predict the prognosis of patients. The Cancer Genome Atlas project, based on comprehensive genomic, epigenomic, and transcriptomic analyses, proposed four molecular subtypes of GC: (1) GC with EBV positivity, which is characterized by genome-wide CpG island DNA methylation, frequent PIK3CA mutation, and PD-L1 amplification; (2) GC with MSI-H, in which mainly epigenetic silencing of mismatch repair (MMR) genes, such as MLH1 causes a hypermutator phenotype; (3) GC with chromosomal instability, which shows marked aneuploidy, frequent amplification of several receptor tyrosine kinases, including HER2, EGFR, MET, and FGFR, and mutation of TP53; and (4) GC with genomic stability, which is enriched for the diffuse type and mutations or translocation of RHO family genes.26 Recently, we reported the protein expression profile of selected key molecules that might be expected to reflect these four molecular subtypes in a large series of AGC.27 In the present study, we evaluated ARID1A expression in a large series (n=1,032) of AGC to clarify its prognostic significance. We then compared the results with those of our previous study in which the key molecules were four MMR proteins (MLH1, MSH2, PMS2, and MSH6) for MSI-H GC, EBV encoded RNA (EBER) in situ hybridization (ISH) for EBV-associated GC, and several receptor tyrosine kinases including including human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), mesenchymal-epithelial transition factor (MET), and p53 for the chromosomal instability group. Finally, we evaluated the prognostic significance of altered ARID1A expression according to the molecular subtype of GC.

1. Patient selection and data collection

A total of 1,032 patients with AGC who underwent curative radical gastrectomy at Severance Hospital from 2000 to 2003 were consecutively enrolled in this study. Patients who received neoadjuvant chemotherapy or who had undergone surgery for recurrent cancer were excluded. Clinicopathologic information, including age, sex, tumor size, location, and clinical follow-up data were collected from pathologic reports and medical records. Tumor size was divided into two groups, >5 cm and ≤5 cm, as the median tumor size of total cases was 5 cm. Pathologic parameters, including the World Health Organization classification, Lauren classification, lymphovascular invasion (LVI), lymph node metastasis, and TNM stage according to the 7th American Joint Committee on Cancer system, were obtained from pathologic reports and slide review. This study was approved and waived informed consents by the Institutional Review Board of Severance Hospital, Seoul, Republic of Korea (IRB number: 4-2016-0419).

2. Tissue microarray construction

To construct the tissue microarray block, two cores were extracted from a representative tumor area of each case (3-mm diameter), as previously described.28,29 One core of adjacent nonneoplastic gastric mucosa was arrayed in each tissue microarray block as landmark and internal control. Four-micrometer sections from tissue microarray blocks were subjected to IHC and EBER-ISH.

3. Immunohistochemistry

IHC was performed using the Ventana Discovery XT automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA) with anti-ARID1A antibody (polyclonal, 1:200 dilution; Sigma-Aldrich, St. Louis, MO, USA), as previously described.28,30 Details of antibodies for MLH1, PMS2, MSH2, MSH6, HER2, EGFR, MET, and p53 were described previously.27 Immunostained slides were evaluated by two individual pathologists (J.Y.K. and S.N.), and the expression pattern of ARID1A were interpreted as preserved, decreased, or loss of expression. “Preserved” staining was defined as a similar intensity of nuclear staining to that in nonneoplastic cells. “Decreased” expression was defined as a markedly decreased intensity of staining compared to that of stromal cells or normal epithelial cells. “Loss” was defined as no staining in tumor cells, regardless of proportion or intensity. In this study, to compare the ARID1A expression patterns and other molecular profiles of AGCs, we used the previously reported data from our prior study.27 The adopted expression data were those of MLH1, PMS2, MSH2, MSH6, HER2, EGFR, MET and p53. In the data, the expression of EGFR, HER2, and MET was scored according to Hofmann’s criteria, as 0, 1+, 2+, or 3+.31 Cases with complete loss or strong and diffuse (more than 50% of tumor cells) nuclear p53 staining were classified as p53 mutant pattern, and cases with weak and patchy (less than 50% of tumor cells) staining of p53 as wild-type pattern, as described previously.27 Cases with complete negativity for MLH1/PMS2 or MSH2/MSH6 in tumor cells were regarded as MMR deficient, and all others as MMR proficient.

4. EBER-ISH

EBER-ISH was performed with a Ventana BenchMark ISH system (ISH iView kit; Ventana, Tucson, AZ, USA). Sections were deparaffinized with EZ Prep buffer (Ventana), and then digested with protease I for 4 minutes. Probes were hybridized and labeled with fluorescein according to the manufacturer’s protocol, as described previously.29 The slides were counterstained with Nuclear Fast Red for 10 minutes. EBER-ISH was performed in all cases. Also, as previously described, cases showing diffuse strong positivity in the nuclei of all tumor cells were defined as EBER-ISH-positive. Representative image of EBER-ISH-positive cases and EBER-ISH-negative cases are presented in Supplementary Fig. 1.

5. Statistical analysis

Statistical calculation was performed with SPSS version 24.0 (IBM Corp., Armonk, NY, USA). The chi-square or Pearson chi-square tests were used to analyze the relationships between ARID1A expression and variable clinicopathologic parameters. Recurrence-free survival was calculated from the date of operation to the date of first recurrence or death without any type of relapse. Overall survival was calculated from the date of gastrectomy to the date of the last follow-up or death. Survival curves were analyzed using the Kaplan-Meier product limit method with the log-rank test for evaluation of significant differences. For multivariate survival analysis, variables found to be significant on univariate analysis were used with the Cox proportional hazard regression model. Significance statements refer to p-value <0.05.

1. ARID1A expression patterns and related clinicopathologic features in AGC

In nonneoplastic mucosa, diffuse and homogenous nuclear expression of ARID1A was observed in either epithelial cells or stromal cells, such as inflammatory cells and fibroblasts (Fig. 1A). Of 1,032 AGCs, 429 cases (41.6%) showed preserved nuclear expression of ARID1A (Fig. 1B); however, the expression of ARID1A was decreased in 442 cases (42.8%) (Fig. 1C) and absent in 161 AGCs (15.6%) (Fig. 1D).

Figure 1.Representative photographs of immunohistochemical AT-rich interactive domain 1A (ARID1A) expression patterns. (A) Ubiquitous nuclear expression in intestinal metaplastic gastric epithelial cells and stromal cells (lymphocytes, endothelial cells, and fibrocytes). (B) Gastric cancer showing an intensity of ARID1A staining similar to that of nonneoplastic cells, classified as “preserved.” (C) Adenocarcinoma with significantly decreased intensity of ARID1A compared to that of stromal cells, classified as “decreased.” (D) Case demonstrating complete loss of nuclear ARID1A staining (A: ×100, B-D: ×200).

Clinicopathologic and molecular features according to the expression patterns of ARID1A are summarized in Table 1. Decreased expression and loss of ARID1A were more frequently observed in tumors located in the proximal (upper and mid third) stomach (p=0.032 and p=0.019, respectively). The loss of ARID1A was associated with larger tumor size (p=0.003) and intestinal type of Lauren classification (p<0.001). However, except for proximal location, no other parameters were associated with decreased expression of ARID1A (p-value for preserved vs decreased) (Table 1). Consistent with previous reports,16,18,19,23 loss of expression was frequently observed in MLH1-deficient GC (MLH1-loss [53/101, 52.5%] vs MLH1-intact [108/931, 11.6%], p<0.001 for preserved group vs loss group) and EBV-positive GCs (EBV-positive [24/67, 35.8%] vs EBV-negative [137/965, 14.2%], p<0.001 for preserved group vs loss group). However, decreased expression of ARID1A was not associated MLH1 deficiency (MLH1-loss [28/101, 27.7%] vs MLH1-intact [414/931, 44.5%], p=0.279 for preserved group vs decreased group) nor EBV positivity (EBV-positive [22/67, 32.8%] vs EBV-negative [420/965, 43.5%], p=0.955 for preserved group vs decreased group). The loss of ARID1A expression was related to wild-type pattern of p53 (p=0.002 for preserved vs loss and p=0.033 for decreased vs loss) and lower frequency of 2 or 3+ expression of HER2 and MET among total cases.

Table 1. Clinicopathologic and Molecular Characteristics According to ARID1A Expression Patterns in Advanced Gastric Cancer

CategoryVariableNo. of
cases (%)
ARID1A expression, No. (%)p-valuep-value
(preserved vs decreased)
p-value
(preserved vs loss)
p-value
(decreased vs loss)
Preserved (n=429)Decreased (n=442)Loss (n=161)
SexMale677 (65.6)289 (67.4)281 (63.6)107 (66.5)0.485
Female355 (34.4)140 (32.6)161 (36.4)54 (33.5)
Age≤60 yr549 (53.2)232 (54.1)245 (55.4)72 (44.7)0.059
>60 yr483 (46.8)197 (45.9)197 (44.6)89 (55.3)
LocationLower third574 (55.6)259 (60.4)235 (53.2)80 (49.7)0.0260.0320.0190.449
Upper and mid458 (44.4)170 (39.6)207 (46.8)81 (50.3)
Size≤5 cm520 (50.4)219 (51.0)239 (54.1)62 (38.5)0.0030.3720.0070.001
>5 cm512 (49.6)210 (49.0)203 (45.9)99 (61.5)
HistologyWD/MD295 (28.6)135 (31.5)116 (26.2)44 (27.3)0.217
PD/others737 (71.4)294 (68.5)326 (73.8)117 (72.7)
LaurenIntestinal504 (48.8)210 (49.0)194 (43.9)100 (62.1)<0.0010.1340.004<0.001
Diffuse528 (51.2)219 (51.0)248 (56.1)61 (37.9)
LVIAbsent735 (71.2)297 (69.2)316 (71.5)122 (75.8)0.290
Present297 (28.8)132 (30.8)126 (28.5)39 (24.2)
LNMAbsent289 (28.0)122 (28.4)123 (27.8)44 (27.3)0.959
Present743 (72.0)307 (71.6)319 (72.2)117 (72.7)
T stageT2176 (17.1)81 (18.9)74 (16.7)21 (13.0)0.0230.3020.0430.009
T3369 (35.8)152 (35.4)143 (32.4)74 (46.0)
T4487 (47.2)196 (45.7)225 (50.9)66 (41.0)
Overall stageII107 (10.4)49 (11.4)45 (10.2)13 (8.1)0.535
III314 (30.4)133 (31.0)129 (29.2)52 (32.3)
IV611 (59.2)247 (57.6)268 (60.6)96 (59.6)
MLH1Loss101 (9.8)20 (4.7)28 (6.3)53 (32.9)<0.0010.279<0.001<0.001
Intact931 (90.2)409 (95.3)414 (93.7)108 (67.1)
MSH2Loss13 (1.3)4 (0.9)7 (1.6)2 (1.2)0.690
Intact1,019 (98.7)425 (99.1)435 (98.4)159 (98.8)
MMR proteinsDeficient114 (11.0)24 (5.6)35 (7.9)55 (34.2)<0.0010.172<0.001<0.001
Proficient918 (89.0)405 (94.4)407 (92.1)106 (65.8)
EBVPositive67 (6.5)21 (4.9)22 (5.0)24 (14.9)<0.0010.955<0.001<0.001
Negative965 (93.5)408 (95.1)420 (95.0)137 (85.1)
MLH1 & EBVMLH1-loss or EBV+168 (16.3)41 (9.6)50 (11.3)77 (47.8)<0.0010.397<0.001<0.001
MLH1 intact and EBV–864 (83.7)388 (90.4)392 (88.7)84 (52.2)
MMRs & EBVMMR-d or EBV+181 (17.5)45 (10.5)57 (12.9)79 (49.1)<0.0010.269<0.001<0.001
MMR-p and EBV–851 (82.5)384 (89.5)385 (87.1)82 (50.9)
p53 IHC*Wild-type pattern624 (62.0)243 (57.7)268 (62.5)113 (72.0)0.0070.1570.0020.033
Mutant pattern383 (38.0)178 (42.3)161 (37.5)44 (28.0)
HER20 or 1+972 (94.2)393 (91.6)418 (94.6)161 (100.0)<0.0010.084<0.0010.003
2 or 3+60 (5.8)36 (8.4)24 (5.4)0
MET*0 or 1+916 (90.5)381 (90.9)399 (92.1)136 (85.0)0.0290.5230.0390.009
2 or 3+96 (9.5)38 (9.1)34 (7.9)24 (15.0)
EGFR*0 or 1+865 (84.5)346 (81.0)393 (90.1)126 (78.3)<0.001<0.0010.452<0.001
2 or 3+159 (15.5)81 (19.0)43 (9.9)35 (21.7)

ARID1A, AT-rich interactive domain 1A; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; MLH1, mutL homolog 1; MSH2, mutS homolog 2; MMR, mismatch-repair protein; EBV, Epstein-Barr virus; MMR-d, MMR deficient; MMR-p, MMR proficient; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor.

*The cases with inadequate immunohistochemical staining results were excluded.



2. ARID1A expression patterns and related clinicopathologic features in MLH1-proficient and EBV-negative AGC

MLH1-deficient GC is known for its unique clinicopathologic features, including intestinal histology, distal tumor location, Crohn-like peritumoral lymphoid reaction, and genome-wide CpG island hypermethylation.26,32 EBV-positive GC also shows characteristic features, including proximal tumor location, frequent occurrence in remnant stomach, and intratumoral intense lymphocytic infiltration, which contributes to a unique histologic type termed carcinoma with lymphoid stroma.33 Since the loss of ARID1A expression has shown close relationships with MLH1 deficiency and EBV positivity, we analyzed the relationships between ARID1A patterns and clinicopathologic and molecular features in a MLH1-proficient and EBV-negative group to remove any possible confounding effects from those two subtypes (Table 2). In this group, 384 (45.1%), 385 (45.2%), and 82 (9.6%) cases showed preserved, decreased, and loss of ARID1A expression, respectively. An association between the decrease or loss of ARID1A and the proximal location was observed (p=0.01). However, larger tumor size and intestinal-type histology were not associated with the altered pattern. Interestingly, in this analysis, the association between the wild-type pattern of p53 and the loss of ARID1A expression, shown in Table 1, was not observed in the MLH1-proficient and EBV-negative subgroups. The associations between the receptor tyrosine kinases expression and ARID1A loss were conserved, with the exception of MET.

Table 2. Clinicopathologic and Molecular Characteristics According to ARID1A Expression Patterns in MLH1-Proficient and EBV-Negative Advanced Gastric Cancer

CategoryVariableNo. of cases (%)ARID1A expression, No. (%)p-value
Preserved (n=388)Decreased (n=392)Loss (n=84)
SexMale561 (64.9)260 (67.0)244 (62.2)57 (67.9)0.317
Female303 (35.1)128 (33.0)148 (37.8)27 (32.1)
Age≤60 yr476 (55.1)210 (54.1)222 (56.6)44 (52.4)0.680
>60 yr388 (44.9)178 (45.9)170 (43.4)40 (47.6)
LocationLower third487 (56.4)239 (61.6)209 (53.3)39 (46.4)0.010
Upper and mid377 (43.6)149 (38.4)183 (46.7)45 (53.6)
Size≤5 cm456 (52.8)204 (52.6)221 (56.4)31 (36.9)0.005
>5 cm408 (47.2)184 (47.4)171 (43.6)53 (63.1)
HistologyWD/MD242 (28.0)121 (31.2)99 (25.3)22 (26.2)0.169
PD/others622 (72.0)267 (68.8)293 (74.7)62 (73.8)
LaurenIntestinal379 (43.9)180 (46.4)160 (40.8)39 (46.4)0.258
Diffuse485 (56.1)208 (53.6)232 (59.2)45 (53.6)
LVIAbsent616 (71.3)271 (69.8)281 (71.7)64 (76.2)0.494
Present248 (28.7)117 (30.2)111 (28.3)20 (23.8)
LNMAbsent239 (27.7)111 (28.6)109 (27.8)19 (22.6)0.537
Present625 (72.3)277 (71.4)283 (72.2)65 (77.4)
T stageT2149 (17.2)74 (19.1)68 (17.3)7 (8.3)0.104
T3287 (33.2)132 (34.0)121 (30.9)34 (40.5)
T4428 (49.5)182 (46.9)203 (51.8)43 (51.2)
Overall stageII88 (10.2)43 (11.1)41 (10.5)4 (4.8)0.326
III259 (30.0)123 (31.7)112 (28.6)24 (28.6)
IV517 (59.8)222 (57.2)239 (61.0)56 (66.7)
p53 IHC*Wild-type pattern493 (58.6)214 (56.0)227 (59.9)52 (64.2)0.308
Mutant pattern349 (41.4)168 (44.0)152 (40.1)29 (35.8)
HER20 or 1+809 (93.6)355 (91.5)370 (94.4)84 (100.0)0.011
2 or 3+55 (6.4)33 (8.5)22 (5.6)0
MET*0 or 1+781 (92.2)349 (91.8)358 (93.2)74 (89.2)0.427
2 or 3+66 (7.8)31 (8.2)26 (6.8)9 (10.8)
EGFR*0 or 1+741 (86.6)317 (82.1)354 (91.7)70 (83.3)<0.001
2 or 3+115 (13.4)69 (17.9)32 (8.3)14 (16.7)

ARID1A, AT-rich interactive domain 1A; MLH1, mutL homolog 1; EBV, Epstein-Barr virus; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor.

*The cases with inadequate immunohistochemical staining results were excluded.



3. ARID1A expression patterns and clinicopathologic features in MLH1-deficient and EBV-positive AGCs

In this study, MLH1-deficient AGCs and EBV-positive AGCs were mutually exclusive. In MLH1-deficient AGCs (n=101), preserved, decreased, and loss of ARID1A was noted in 20 (19.9%), 28 (27.7%), and 53 (52.5%) cases, respectively (p<0.001 compared to MLH1-proficient GCs) (Table 1). No parameter was associated with altered ARID1A expression (Supplementary Table 1). Among EBV-positive AGCs (n=67), 21 (31.3%), 22 (32.8%), and 24 (35.8%) showed preserved, decreased, and loss of ARID1A expression, respectively. A low frequency of LVI was the only factor associated with decrease or loss of ARID1A expression (p=0.044) (Supplementary Table 2).

4. Prognostic significance of decreased or loss of ARID1a expression in AGC

Kaplan-Meier survival curves according to ARID1A expression showed worse prognosis of the patient group with decreased ARID1A compared to patients with preserved ARID1A (Fig. 2A). However, the ARID1A-loss group exhibited a similar prognosis to that of the ARID1A-preserved group. Since GCs with ARID1A loss were enriched among MLH1-deficient and EBV-positive GCs (Table 1) and the prognosis of GC patients with MLH1-deficiency or EBV-positivity was more favorable than that of patients with EBV-negative and MLH1-proficient AGC (Fig. 2B), we analyzed the prognostic effects of altered ARID1A expression in MLH1-deficient, EBV-positive, and EBV-negative and MLH1-proficient groups, respectively. There were no prognostic differences among the ARID1A preserved, decreased, and loss groups in either MLH1-deficient (Fig. 3A) or EBV-positive GCs (Fig. 3B). However, in the patient group with MLH1-proficient and EBV-negative AGC, the prognosis was significantly different according to the ARID1A expression patterns: patients with ARID1A-loss AGC showed the worst overall survival; and the ARID1A-decreased group had second-worst prognosis (p<0.001) (Fig. 2C).

Figure 2.Kaplan-Meier survival curves for overall survival with AGC according to three ARID1A expression patterns and molecular subtypes. (A) Overall survival curves of the total group according to ARID1A expression pattern show a worse prognosis of the ARID1A-decreased group than the ARID1A-preserved or ARID1A-loss group. (B) Overall survival curves comparing the EBV-positive group, the MLH1-loss group and the MLH-proficient and EBV-negative group reveal a favorable prognosis of AGC with the molecular subtype of the EBV-positive and MLH1-loss group. (C) Among MLH1-proficient and EBV-negative AGCs, survival curves according to ARID1A expression pattern demonstrate a worse prognosis of patients with ARID1A decreased or ARID1A loss AGCs. (D) Among either MLH1-deficient or EBV-positive AGCs, survival curves according to ARID1A expression pattern show no difference in prognosis.
AGC, advanced gastric cancer; ARID1A, AT-rich interactive domain 1A; EBV, Epstein-Barr virus; MLH1, mutL homolog 1.

Figure 3.Kaplan-Meier survival curves for overall survival of MLH1-deficient and EBV-positive AGC. (A) Overall survival curves of MLH1-deficient AGCs according to three ARID1A expression patterns. (B) Overall survival curves of EBV-positive AGCs according to three ARID1A expression patterns.
MLH1, mutL homolog 1; EBV, Epstein-Barr virus; AGC, advanced gastric cancer; ARID1A, AT-rich interactive domain 1A.

In multivariate survival analysis, Cox regression analyses for overall survival and recurrence-free survival were performed using variables that were significant on univariate analysis, including age, tumor location, tumor size, histology, Lauren classification, LVI, lymph node metastasis, overall stage, MLH1 statue, EBV positivity, and ARID1A expression patterns, using a forward conditional method. For overall survival, MLH1 proficiency and EBV negativity were independent worse prognostic factors (hazard ratio [HR], 1.4; p<0.001 and HR, 1.98; p<0.001, respectively). Decrease or loss of ARID1A expression was also revealed as a negative prognostic factor of AGC (HR, 1.47 for decreased and HR, 1.48 for loss pattern, p<0.001). In addition to MLH1, EBV, and ARID1A expression, older age (HR, 1.58; p<0.001), larger tumor size (>5 cm) (HR, 1.24; p=0.008), diffuse type by Lauren (HR, 1.26; p=0.009), presence of LVI (HR, 1.54; p<0.001), and overall stage (HR, 1.69 for stage III and HR, 4.38 for stage IV, p<0.001) were found to be prognostic factors of overall survival. For recurrence-free survival, ARID1A expression pattern (p=0.041) was a prognostic factor, in addition to MLH1 proficiency (p<0.001), EBV negativity (p=0.002), and the features that were shown to be prognostic factors of overall survival, except for older age and proximal tumor location (Table 3).

Table 3. Univariate and Multivariate Survival Analyses

CategoryVariableUnivariate-OSMultivariate-OSUnivariate-RFSMultivariate-RFS
HR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-value
SexMale11
Female1.08 (0.92–1.27)0.3421.18 (0.98–1.42)0.083
Age≤60 yr111
>60 yr1.41 (1.20–1.65)<0.0011.58 (1.34–1.88)<0.0010.96 (0.80–1.15)0.657
LocationLower third111
Upper and mid1.36 (1.16–1.60)<0.0011.25 (1.06–1.48)0.0081.27 (1.05–1.53)0.013
Size≤5 cm1111
>5 cm1.75 (1.49–2.07)<0.0011.24 (1.04–1.48)0.0181.82 (1.51–2.20)<0.0011.32 (1.09–1.61)0.005
HistologyWD/MD11
PD/others1.36 (1.14–1.63)0.0011.60 (1.29–1.99)<0.001
LaurenIntestinal1111
Diffuse1.38 (1.17–1.61)<0.0011.26 (1.06–1.49)0.0091.68 (1.40–2.03)<0.0011.31 (1.08–1.60)0.007
LVIAbsent1111
Present2.21 (1.87–2.60)<0.0011.54 (1.29–1.84)<0.0012.31 (1.92–2.79)<0.0011.49 (1.22–1.82)<0.001
LNMAbsent11
Present3.24 (2.62–4.00)<0.0014.13 (3.15–5.41)<0.001
T stageT211<0.001
T32.44 (1.82–3.28)<0.0013.41 (2.27–5.13)<0.001
T44.96 (3.75–6.56)<0.0017.96 (5.40–11.74)<0.001
Overall stageII11<0.0011<0.0011<0.001
III1.96 (1.32–2.92)0.0011.69 (1.08–2.65)0.0233.01 (1.64–5.51)<0.0012.34 (1.21–4.55)0.012
IV6.03 (4.16–8.73)<0.0014.38 (2.83–6.76)<0.00111.97 (6.73–21.29)<0.0017.78 (4.10–14.76)<0.001
MLH1Loss1111
Intact1.47 (1.08–2.00)0.0141.85 (1.31–2.61)<0.0011.94 (1.31–2.89)0.0011.99 (1.29–3.07)0.002
MSH2Loss11
Intact0.57 (0.31–1.07)0.0800.59 (0.28–1.25)0.166
MMR proteinsDeficient11
Proficient1.30 (0.98–1.72)0.0661.65 (1.16–2.36)0.005
EBVPositive1111
Negative1.98 (1.31–2.97)0.0012.36 (1.54–3.62)<0.0011.98 (1.23–3.17)0.0052.12 (1.31–3.44)0.002
MLH1 & EBVMLH1-loss or EBV+11
MLH1-intact and EBV+1.73 (1.34–2.22)<0.0012.07 (1.51–2.82)<0.001
MMR-d & EBVMMR-d or EBV+11
MMR-p and EBV–1.57 (1.24–2.00)<0.0011.86 (1.39–2.49)<0.001
ARID1APreserved10.0031<0.00110.04110.041
Decreased1.36 (1.14–1.62)0.0011.47 (1.23–1.76)<0.0011.22 (1.00–1.49)0.0541.29 (1.06–1.58)0.013
Loss1.12 (0.87–1.44)0.3881.48 (1.12–1.95)0.0050.88 (0.65–1.19)0.4031.25 (0.90–1.73)0.189
p53Wild-type pattern11
Mutant pattern1.09 (0.92–1.29)0.3061.09 (0.90–1.33)0.355

Human epidermal growth factor receptor 2, epidermal growth factor receptor, and mesenchymal-epithelial transition factor: p>0.05 on univariate analysis (data not shown in table).

OS, overall survival; RFS, recurrence-free survival; HR, hazard ratio; CI, confidence interval; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; MLH1, mutL homolog 1; MSH2, mutS homolog 2; MMR, mismatch-repair protein; EBV, Epstein-Barr virus; MMR-d, MMR deficient; MMR-p, MMR proficient; ARID1A, AT-rich interactive domain 1A.


The loss of ARID1A expression correlates well with mutation status, although mutation does not account for all cases of loss of expression. Mutation or loss of ARID1A in GC are closely related to MSI-H type and EBV positivity.11,16-18,23,34 In this study, 52.5% (53/101) of MLH1-deficient and 35.8% (24/67) of EBV-positive GCs exhibited loss of ARID1A expression, compared with only 9.6% (82/851) of the MLH1-proficient and EBV-negative group (p<0.001) (Table 1). Therefore, we confirmed the strong associations between loss of ARID1A expression and MLH1 deficiency and EBV positivity using the largest sample set to date. Interestingly, in contrast to GCs with ARID1A loss, cases with decreased ARID1A were not associated with MLH1-deficiency or EBV-positivity. With the exception of one study by Kim et al.,18 previous studies that evaluated ARID1A expression in GC categorized the expression as only positive or negative and did not evaluate the association between a decreased pattern of ARID1A expression and EBV-positive or MMR-deficient GC.11,16-19,23,34 Kim et al.18 classified the patterns of ARID1A expression as retained, reduced, complete loss, and partial loss. In their study, reduced ARID1A expression was found in 17.7% of GCs and was not associated with EBV positivity, but was frequently found in the MLH1-deficient group.18 This discrepancy with our results, which revealed no association between the decreased pattern and MLH1 deficiency, might result from the different incidence of MLH1-deficient GC (18.9% vs 9.8% in our study) and of ARID1A decreased cases (17.7% vs 40.9% in our study).

The expression of MSH2, another key MMR protein, was not associated with ARID1A loss or decreased expression, although only limited cases were included in this category. MSH2 deficiency indicates a possible germline mutation of the MSH2 gene (so-called Lynch syndrome), whereas the vast majority of cases with MLH1 deficiency result from epigenetic silencing of MLH1 in the context of a CpG island methylator phenotype35-38 that shows genome-wide hypermethylation.26 Genome-wide hypermethylation is also a well-known characteristic of EBV-positive GC.26,39 Considering this shared epigenetic characteristic between EBV-positive and MLH1-deficient GC and the lack of a relationship between ARID1A alteration and MSH2-deficiency, it might be postulated that the frequent loss of ARID1A expression in MSI-H GC is associated with hypermethylation rather than instability of microsatellites or a mutator phenotype.

A negative association between mutations in TP53 and ARID1A has been reported in GC and endometrial cancer.11 In addition, a relationship between wild-type p53 staining pattern (weak and patch nuclear staining) and loss of ARID1A expression in GC and endometrial cancers was demonstrated.17,40 We also found an association between wild-type pattern of p53 expression and loss of ARID1A (p=0.007) (Table 1). However, in subgroup analysis, this relationship was lost in the MLH1-proficient and EBV-negative groups (Table 2), and was also not observed in MLH1-deficient GCs (Supplementary Table 1) or EBV-positive GCs (Supplementary Table 2). It is well known that EBV-positive and MLH1-deficient GCs are associated with wild-type TP53.26,33 We also found a high frequency of wild-type pattern of p53 staining in MLH1-deficient (81%, 81/100) and EBV-positive GCs (76.9%, 50/65), compared to 58.6% (493/864) of the MLH1-proficient and EBV-negative group. Therefore, it can be assumed that the previously reported association between the wild-type pattern of p53 staining and the loss of ARID1A may reflect the strong relationship between loss of ARID1A and EBV-positive and MLH1-deficient GC. To rule out the possible confounding effects of EBV positivity and MLH1 deficiency on this association, we performed a binary logistic regression analysis using the loss of ARID1 expression as a dependent variable and other parameters that showed a correlation with the loss of ARID1A in Table 1, including EBV positivity, MLH1 deficiency, tumor size, T stage, p53 staining pattern, and HER2 and MET expression, as independent variables. The logistic regression analysis revealed that only EBV positivity (odds ratio, 5.119; p<0.001) and MLH1 deficiency (odds ratio, 8.376; p<0.001) were correlated with the loss of ARID1A (Table 4).

Table 4. Logistic Regression Analysis of Clinicopathologic and Molecular Features Associated with Loss of ARID1A Expression

FeatureOdds ratio (95% CI)p-value
Location (upper and mid)1.35 (0.90–2.01)0.143
Size (>5 cm)1.23 (0.82–1.86)0.315
Lauren type (intestinal)1.38 (0.92–2.07)0.125
T stage0.242
T31.07 (0.58–2.00)0.820
T40.73 (0.48–1.11)0.140
MLH1 (deficient)8.12 (4.82–13.66)<0.001
EBV (positive)4.13 (2.28–7.48)<0.001
p53 (wild-type pattern)1.17 (0.77–1.77)0.459
HER2 (2 or 3+)NA0.997
MET (2 or 3+)1.54 (0.82–2.89)0.175
EGFR (2 or 3+)1.08 (0.62–1.89)0.781

ARID1A, AT-rich interactive domain 1A; CI, confidence interval; MLH1, mutL homolog 1; EBV, Epstein-Barr virus; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor; NA, not available.



Previous studies have shown that the loss of ARID1A expression is correlated with worse prognosis in patients with GC13,16,18,23 or only in patients with EBV-negative and MLH1-preserved GC.16 However, in some studies, the relationship between loss of ARID1A expression and prognosis was not conclusive.11,17,20,24 In our study, the loss of ARID1A showed different prognostic effects in GC patients according to the molecular subtype. In total cases, the prognosis of patients with ARID1A loss was not significantly different from that of patients with preserved ARID1A expression (Fig. 2A). However, in EBV-negative and MLH1-proficient GC patients, the loss of ARID1A expression was associated with the worst prognosis (Fig. 2C). The favorable prognosis of EBV-positive and MLH1-deficient GC patients (Fig. 2B) and the strong correlation between loss of ARID1A and EBV-positive and MLH1-deficient GCs (Table 1) may account for the different prognostic effects of loss of ARID1A among the molecular subtypes. This finding was in good agreement with a previous study by Abe et al.;16 however, in their study, only eight cases of EBV-positive and 36 cases of MLH1-negative GC were enrolled in the survival analysis;16 therefore, there was a limitation in evaluating the prognostic effect of ARID1A loss in patients with EBV-positive and MLH1-deficient GC. In this study, we confirmed that loss of ARID1A expression was an independent negative prognostic factor in the MLH1-proficient and EBV-negative AGC (Table 3, Fig. 2C), but not in EBV-positive or MLH1-deficient AGC (Figs 2D and 3).

It has been reported that IHC for ARID1A can be used as surrogate marker for ARID1A mutation status.11,14,16 However, ARID1A alterations can occur in other ways as well as mutation. It has been reported that ARID1A in cancer can be regulated by way of promotor methylation or post-transcriptional modification.15,41 In addition, there is also a report that ARID1A can be regulated by EBV-encoded miRNA in EBV-positive GCs.42 Due to the large number of the sample of this study, there are limitations in analyzing mutation and epigenetic profiles for every tumors; however, we expect that decreased ARID1A expression is associated with non-mutational alteration of ARID1A and results of this study support the supposition. In this study, patients with decreased expression of ARID1A accounted for 42.8% of total cases and 44.6% of the EBV-negative and MLH1-proficient groups. The incidence of ARID1A mutation in MLH1-proficient and EBV-negative GC has been reported to be 8% to 10%.11,12 However, the proportion of patients with altered expression of ARID1A in this study was 54.8% (45.2% for ARID1A decrease and 9.6% for loss) of total cases, which implied that there might be another mechanism suppressing the expression of ARID1A other than the genetic mutation. In ovarian clear cell carcinoma, biallelic mutations were found in only 30% of cases with ARID1A mutation, whereas 73% of ARID1A-heterozygous cases showed loss of protein expression by IHC.10 In breast cancer, promoter hypermethylation of ARID1A was reported to be strongly correlated with a low level of messenger RNA expression.15 Further investigation into the suppression mechanism of ARID1A might help elucidate the underlying mechanism regulating ARID1A expression in GC. In addition, patients with decreased ARID1A showed worse prognosis among total cases and in the EBV-negative and MLH1-proficient group (Fig. 2). On multivariate analysis, we found that decreased ARID1A was an independent poor prognostic factor in AGC patients, in addition to the loss of ARID1A (Table 3). Therefore, understanding the mechanisms involved in reduced expression of ARID1A is important not only biologically, but also clinically.

Recently, Kim et al.30 reported that synthetic lethality of EZH2, a histone methyl transferase subunit of polycomb repressor complex, is related to ARID1A mutation. Tumor cell lines with ARID1A mutation undergo cell death and are inhibited in their ability of tumor formation in vivo when treated with the EZH2 inhibitor, GSK126.30 Therefore, GC with altered ARID1 expression might be a potential candidate for EZH2-targeted treatment in the future.

In conclusion, the loss of ARID1A expression was associated with larger tumor size, intestinal histology, MLH1 deficiency, EBV positivity, wild-type pattern of p53 staining, and negative or 1+ HER2 and MET expression. However, on logistic regression analysis, only MLH1-deficiency and EBV positivity showed a correlation with the loss of ARID1A. In MLH1-proficient and EBV-negative GC, wild-type pattern of p53 staining was not associated with ARID1A loss. In addition to the loss of ARID1A expression, decreased ARID1A was also revealed as an independent negative prognostic factor in AGC patients. Interestingly, no prognostic significance of altered ARID1A expression was found in MLH1-deficient or EBV-positive GC. Regarding the emerging concept of synthetic lethality associated with ARID1A mutation, GC with reduced or loss of ARID1A expression might be a good candidate for new targeted treatments.

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science and ICT) (2019R1A2C1011355, H.K.).

No potential conflict of interest relevant to this article was reported.

Study concept and design: H.K. Data acquisition: C.K.P., S.N. Data analysis and interpretation: J.Y.K., C.K.P., S.N. Drafting of the manuscript: J.Y.K., S.N. Critical revision of the manuscript for important intellectual content: J.H.C., S.H.N. Statistical analysis: J.Y.K., C.K.P., S.N. Obtained funding: H.K. Administrative, technical, or material support; study supervision: H.K. Approval of final manuscript: all authors.

  1. Dallas PB, Pacchione S, Wilsker D, Bowrin V, Kobayashi R, Moran E. The human SWI-SNF complex protein p270 is an ARID family member with non-sequence-specific DNA binding activity. Mol Cell Biol 2000;20:3137-3146.
    Pubmed KoreaMed CrossRef
  2. Wilson BG, Roberts CW. SWI/SNF nucleosome remodellers and cancer. Nat Rev Cancer 2011;11:481-492.
    Pubmed CrossRef
  3. Nagl NG Jr, Patsialou A, Haines DS, Dallas PB, Beck GR Jr, Moran E. The p270 (ARID1A/SMARCF1) subunit of mammalian SWI/SNF-related complexes is essential for normal cell cycle arrest. Cancer Res 2005;65:9236-9244.
    Pubmed CrossRef
  4. Alver BH, Kim KH, Lu P, et al. The SWI/SNF chromatin remodelling complex is required for maintenance of lineage specific enhancers. Nat Commun 2017;8:14648.
    Pubmed KoreaMed CrossRef
  5. Mathur R, Alver BH, San Roman AK, et al. ARID1A loss impairs enhancer-mediated gene regulation and drives colon cancer in mice. Nat Genet 2017;49:296-302.
    Pubmed KoreaMed CrossRef
  6. Wang Y, Chen SY, Karnezis AN, et al. The histone methyltransferase EZH2 is a therapeutic target in small cell carcinoma of the ovary, hypercalcaemic type. J Pathol 2017;242:371-383.
    Pubmed KoreaMed CrossRef
  7. Shen J, Peng Y, Wei L, et al. ARID1A deficiency impairs the DNA damage checkpoint and sensitizes cells to PARP inhibitors. Cancer Discov 2015;5:752-767.
    Pubmed KoreaMed CrossRef
  8. Suryo Rahmanto Y, Jung JG, Wu RC, et al. Inactivating ARID1A tumor suppressor enhances TERT transcription and maintains telomere length in cancer cells. J Biol Chem 2016;291:9690-9699.
    Pubmed KoreaMed CrossRef
  9. Jones S, Wang TL, Shih IeM, et al. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science 2010;330:228-231.
    Pubmed KoreaMed CrossRef
  10. Wiegand KC, Shah SP, Al-Agha OM, et al. ARID1A mutations in endometriosis-associated ovarian carcinomas. N Engl J Med 2010;363:1532-1543.
    Pubmed KoreaMed CrossRef
  11. Wang K, Kan J, Yuen ST, et al. Exome sequencing identifies frequent mutation of ARID1A in molecular subtypes of gastric cancer. Nat Genet 2011;43:1219-1223.
    Pubmed CrossRef
  12. Zang ZJ, Cutcutache I, Poon SL, et al. Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes. Nat Genet 2012;44:570-574.
    Pubmed CrossRef
  13. Wang DD, Chen YB, Pan K, et al. Decreased expression of the ARID1A gene is associated with poor prognosis in primary gastric cancer. PLoS One 2012;7:e40364.
    Pubmed KoreaMed CrossRef
  14. Maeda D, Mao TL, Fukayama M, et al. Clinicopathological significance of loss of ARID1A immunoreactivity in ovarian clear cell carcinoma. Int J Mol Sci 2010;11:5120-5128.
    Pubmed KoreaMed CrossRef
  15. Zhang X, Sun Q, Shan M, et al. Promoter hypermethylation of ARID1A gene is responsible for its low mRNA expression in many invasive breast cancers. PLoS One 2013;8:e53931.
    Pubmed KoreaMed CrossRef
  16. Abe H, Maeda D, Hino R, et al. ARID1A expression loss in gastric cancer: pathway-dependent roles with and without Epstein-Barr virus infection and microsatellite instability. Virchows Arch 2012;461:367-377.
    Pubmed CrossRef
  17. Wiegand KC, Sy K, Kalloger SE, et al. ARID1A/BAF250a as a prognostic marker for gastric carcinoma: a study of 2 cohorts. Hum Pathol 2014;45:1258-1268.
    Pubmed CrossRef
  18. Kim YB, Ham IH, Hur H, Lee D. Various ARID1A expression patterns and their clinical significance in gastric cancers. Hum Pathol 2016;49:61-70.
    Pubmed CrossRef
  19. Kim KJ, Jung HY, Oh MH, et al. Loss of ARID1A expression in gastric cancer: correlation with mismatch repair deficiency and clinicopathologic features. J Gastric Cancer 2015;15:201-208.
    Pubmed KoreaMed CrossRef
  20. Ibarrola-Villava M, Llorca-Cardeñosa MJ, Tarazona N, et al. Deregulation of ARID1A, CDH1, cMET and PIK3CA and target-related microRNA expression in gastric cancer. Oncotarget 2015;6:26935-26945.
    Pubmed KoreaMed CrossRef
  21. Glückstein MI, Dintner S, Arndt TT, et al. Comprehensive immunohistochemical study of the SWI/SNF complex expression status in gastric cancer reveals an adverse prognosis of SWI/SNF deficiency in genomically stable gastric carcinomas. Cancers (Basel) 2021;13:3894.
    Pubmed KoreaMed CrossRef
  22. Zhu YP, Sheng LL, Wu J, et al. Loss of ARID1A expression is associated with poor prognosis in patients with gastric cancer. Hum Pathol 2018;78:28-35.
    Pubmed CrossRef
  23. Inada R, Sekine S, Taniguchi H, et al. ARID1A expression in gastric adenocarcinoma: clinicopathological significance and correlation with DNA mismatch repair status. World J Gastroenterol 2015;21:2159-2168.
    Pubmed KoreaMed CrossRef
  24. Lee SY, Kim DW, Lee HS, et al. Loss of AT-rich interactive domain 1A expression in gastrointestinal malignancies. Oncology 2015;88:234-240.
    Pubmed CrossRef
  25. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69-90.
    Pubmed CrossRef
  26. Cancer Genome Atlas Research Network. Comprehensive molecular characterization of gastric adenocarcinoma. Nature 2014;513:202-209.
    Pubmed KoreaMed CrossRef
  27. Park CK, Park JS, Kim HS, et al. Receptor tyrosine kinase amplified gastric cancer: clinicopathologic characteristics and proposed screening algorithm. Oncotarget 2016;7:72099-72112.
    Pubmed KoreaMed CrossRef
  28. Bae YS, Kim H, Noh SH, Kim H. Usefulness of immunohistochemistry for microsatellite instability screening in gastric cancer. Gut Liver 2015;9:629-635.
    Pubmed KoreaMed CrossRef
  29. Park JH, Kim EK, Kim YH, et al. Epstein-Barr virus positivity, not mismatch repair-deficiency, is a favorable risk factor for lymph node metastasis in submucosa-invasive early gastric cancer. Gastric Cancer 2016;19:1041-1051.
    Pubmed CrossRef
  30. Kim KH, Kim W, Howard TP, et al. SWI/SNF-mutant cancers depend on catalytic and non-catalytic activity of EZH2. Nat Med 2015;21:1491-1496.
    Pubmed KoreaMed CrossRef
  31. Hofmann M, Stoss O, Shi D, et al. Assessment of a HER2 scoring system for gastric cancer: results from a validation study. Histopathology 2008;52:797-805.
    Pubmed CrossRef
  32. 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.
    Pubmed CrossRef
  33. Shinozaki-Ushiku A, Kunita A, Fukayama M. Update on Epstein-Barr virus and gastric cancer (review). Int J Oncol 2015;46:1421-1434.
    Pubmed CrossRef
  34. Han N, Kim MA, Lee HS, Kim WH. Loss of ARID1A expression is related to gastric cancer progression, Epstein-Barr virus infection, and mismatch repair deficiency. Appl Immunohistochem Mol Morphol 2016;24:320-325.
    Pubmed CrossRef
  35. Hampel H, Frankel WL, Martin E, et al. Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N Engl J Med 2005;352:1851-1860.
    Pubmed CrossRef
  36. Kane MF, Loda M, Gaida GM, et al. Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res 1997;57:808-811.
    Pubmed
  37. Toyota M, Ahuja N, Suzuki H, et al. Aberrant methylation in gastric cancer associated with the CpG island methylator phenotype. Cancer Res 1999;59:5438-5442.
    Pubmed
  38. Kim H, Kim YH, Kim SE, Kim NG, Noh SH, Kim H. Concerted promoter hypermethylation of hMLH1, p16INK4A, and E-cadherin in gastric carcinomas with microsatellite instability. J Pathol 2003;200:23-31.
    Pubmed CrossRef
  39. Zhao J, Liang Q, Cheung KF, et al. Genome-wide identification of Epstein-Barr virus-driven promoter methylation profiles of human genes in gastric cancer cells. Cancer 2013;119:304-312.
    Pubmed CrossRef
  40. Allo G, Bernardini MQ, Wu RC, et al. ARID1A loss correlates with mismatch repair deficiency and intact p53 expression in high-grade endometrial carcinomas. Mod Pathol 2014;27:255-261.
    Pubmed KoreaMed CrossRef
  41. Odnokoz O, Wavelet-Vermuse C, Hophan SL, Bulun S, Wan Y. ARID1 proteins: from transcriptional and post-translational regulation to carcinogenesis and potential therapeutics. Epigenomics 2021;13:809-823.
    Pubmed KoreaMed CrossRef
  42. Kase K, Saito M, Nakajima S, et al. ARID1A deficiency in EBV-positive gastric cancer is partially regulated by EBV-encoded miRNAs, but not by DNA promotor hypermethylation. Carcinogenesis 2021;42:21-30.
    Pubmed CrossRef

Article

Original Article

Gut and Liver 2023; 17(5): 753-765

Published online September 15, 2023 https://doi.org/10.5009/gnl220342

Copyright © Gut and Liver.

Prognostic Significance of ARID1A Expression Patterns Varies with Molecular Subtype in Advanced Gastric Cancer

Jun Yong Kim1 , Cheol Keun Park1 , Songmi Noh2 , Jae-Ho Cheong3 , Sung Hoon Noh3 , Hyunki Kim1

1Department of Pathology, Yonsei University College of Medicine, 2Department of Pathology, CHA Gangnam Medical Center, CHA University, and 3Department of Surgery, Yonsei University College of Medicine, Seoul, Korea

Correspondence to:Hyunki Kim
ORCID https://orcid.org/0000-0003-2292-5584
E-mail hyunkikim@yuhs.ac

Jun Yong Kim and Cheol Keun Park contributed equally to this work as first authors.

Received: August 4, 2022; Revised: November 23, 2022; Accepted: November 23, 2022

This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Background/Aims: AT-rich interactive domain 1A (ARID1A) is frequently mutated in gastric cancer (GC), especially Epstein-Barr virus (EBV)-associated and microsatellite instability high GC. The loss of ARID1A expression has been reported as a poor prognostic marker in GC. However, the relationships between ARID1A alteration and EBV-associated and microsatellite instability high GC, which are known to have a favorable prognosis, has hampered proper evaluation of the prognostic significance of ARID1A expression in GC. We aimed to analyze the true prognostic significance of ARID1A expression by correcting confounding variables.
Methods: We evaluated the ARID1A expression in a large series (n=1,032) of advanced GC and analyzed the relationships between expression pattern and variable parameters, including clinicopathologic factors, key molecular features such as EBV-positivity, mismatch repair protein deficiency, and expression of p53 and several receptor tyrosine kinases including human epidermal growth factor receptor 2, epidermal growth factor receptor, and mesenchymal-epithelial transition factor. Survival analysis of the molecular subtypes was done according to the ARID1A expression patterns.
Results: Loss of ARID1A expression was found in 52.5% (53/101) of mutL homolog 1 (MLH1)-deficient and 35.8% (24/67) of EBV-positive GCs, compared with only 9.6% (82/864) of the MLH1-proficient and EBV-negative group (p<0.001). The loss of ARID1A expression was associated only with MLH1 deficiency and EBV positivity. On survival analysis, the loss of ARID1A expression was associated with worse prognosis only in MLH1-proficient and EBV-negative GC. Multivariate analysis revealed that both loss of ARID1A and decreased ARID1A expression were independent worse prognostic factors in patients with advanced GC.
Conclusions: Only in MLH1-proficient and EBV-negative GC, the loss of ARID1A expression is related to poorer prognosis.

Keywords: Stomach neoplasms, ARID1A, Immunohistochemistry, Prognosis

INTRODUCTION

AT-rich interactive domain 1A (ARID1A), a member of the ARID family, is a subunit of SWItch/sucrose non-fermentable complex that remodels histone-DNA interactions in reconstituted nucleosomes in an ATP-dependent manner.1 The SWItch/sucrose non-fermentable complex plays essential roles during lineage specification and in the maintenance of stem cell pluripotency, and malfunctioning ARID1A could potentially trigger persistent proliferative progenitor state.2-6 ARID1A-containing SWItch/sucrose non-fermentable complex is responsible for cell cycle and ARID1A-depletion leads to defective cell cycle checkpoint activation in response to DNA damage.7 Also, regulation of telomere reverse transcriptase by ARID1A through chromatin structure repression was reported.8 Recent studies revealed that the ARID1A gene is frequently mutated in a variety of cancers, such as ovarian clear cell carcinoma, endometrioid adenocarcinoma, and gastric carcinomas.9-12 The mutations of ARID1A found in these types of cancer are mostly insertion/deletion mutations that generate premature stop codon and thus lead to truncation of the ARID1A protein.9-12 Previous studies reported that the knockdown of ARID1A in gastric cancer (GC) cell lines promotes cell proliferation and the forced expression of ARID1A inhibits colony formation and cell growth.12,13 The loss of expression of ARID1A evaluated by immunohistochemistry (IHC) is correlated with the mutation status of ARID1A,11,14 although epigenetic silencing can also induce the loss of ARID1A expression.15 Intriguingly, in GC, mutation of ARID1A and loss of expression are closely associated with Epstein-Barr virus (EBV) positivity and microsatellite instability high (MSI-H) type.11,12,16-18 Several studies have examined the prognostic significance of alterations of ARID1A in GC.11,13,16-22 Among them, Wang et al.11 and Ibarrola-Villava et al.20 reported a favorable outcome for patients with ARID1A-mutated GC. In contrast, other studies reported that the loss of ARID1A expression was associated with poor survival outcome13,16,18,23 or had no prognostic role in GC.19,24 Therefore, the prognostic significance of ARID1A alteration still remains controversial. The conflicting reports of the significance of ARID1A expression on the prognosis of GC patients might result from possible confounding effects of EBV-associated and MSI-H GCs and/or limitations of sample size.

GC is one of the leading causes of cancer death worldwide; even though its incidence has been decreasing,25 the prognosis of advanced GC (AGC) is still dismal. Therefore, a more comprehensive investigation into the underlying molecular mechanisms is mandatory to develop more effective treatment modalities and to better predict the prognosis of patients. The Cancer Genome Atlas project, based on comprehensive genomic, epigenomic, and transcriptomic analyses, proposed four molecular subtypes of GC: (1) GC with EBV positivity, which is characterized by genome-wide CpG island DNA methylation, frequent PIK3CA mutation, and PD-L1 amplification; (2) GC with MSI-H, in which mainly epigenetic silencing of mismatch repair (MMR) genes, such as MLH1 causes a hypermutator phenotype; (3) GC with chromosomal instability, which shows marked aneuploidy, frequent amplification of several receptor tyrosine kinases, including HER2, EGFR, MET, and FGFR, and mutation of TP53; and (4) GC with genomic stability, which is enriched for the diffuse type and mutations or translocation of RHO family genes.26 Recently, we reported the protein expression profile of selected key molecules that might be expected to reflect these four molecular subtypes in a large series of AGC.27 In the present study, we evaluated ARID1A expression in a large series (n=1,032) of AGC to clarify its prognostic significance. We then compared the results with those of our previous study in which the key molecules were four MMR proteins (MLH1, MSH2, PMS2, and MSH6) for MSI-H GC, EBV encoded RNA (EBER) in situ hybridization (ISH) for EBV-associated GC, and several receptor tyrosine kinases including including human epidermal growth factor receptor 2 (HER2), epidermal growth factor receptor (EGFR), mesenchymal-epithelial transition factor (MET), and p53 for the chromosomal instability group. Finally, we evaluated the prognostic significance of altered ARID1A expression according to the molecular subtype of GC.

MATERIALS AND METHODS

1. Patient selection and data collection

A total of 1,032 patients with AGC who underwent curative radical gastrectomy at Severance Hospital from 2000 to 2003 were consecutively enrolled in this study. Patients who received neoadjuvant chemotherapy or who had undergone surgery for recurrent cancer were excluded. Clinicopathologic information, including age, sex, tumor size, location, and clinical follow-up data were collected from pathologic reports and medical records. Tumor size was divided into two groups, >5 cm and ≤5 cm, as the median tumor size of total cases was 5 cm. Pathologic parameters, including the World Health Organization classification, Lauren classification, lymphovascular invasion (LVI), lymph node metastasis, and TNM stage according to the 7th American Joint Committee on Cancer system, were obtained from pathologic reports and slide review. This study was approved and waived informed consents by the Institutional Review Board of Severance Hospital, Seoul, Republic of Korea (IRB number: 4-2016-0419).

2. Tissue microarray construction

To construct the tissue microarray block, two cores were extracted from a representative tumor area of each case (3-mm diameter), as previously described.28,29 One core of adjacent nonneoplastic gastric mucosa was arrayed in each tissue microarray block as landmark and internal control. Four-micrometer sections from tissue microarray blocks were subjected to IHC and EBER-ISH.

3. Immunohistochemistry

IHC was performed using the Ventana Discovery XT automated staining system (Ventana Medical Systems, Inc., Tucson, AZ, USA) with anti-ARID1A antibody (polyclonal, 1:200 dilution; Sigma-Aldrich, St. Louis, MO, USA), as previously described.28,30 Details of antibodies for MLH1, PMS2, MSH2, MSH6, HER2, EGFR, MET, and p53 were described previously.27 Immunostained slides were evaluated by two individual pathologists (J.Y.K. and S.N.), and the expression pattern of ARID1A were interpreted as preserved, decreased, or loss of expression. “Preserved” staining was defined as a similar intensity of nuclear staining to that in nonneoplastic cells. “Decreased” expression was defined as a markedly decreased intensity of staining compared to that of stromal cells or normal epithelial cells. “Loss” was defined as no staining in tumor cells, regardless of proportion or intensity. In this study, to compare the ARID1A expression patterns and other molecular profiles of AGCs, we used the previously reported data from our prior study.27 The adopted expression data were those of MLH1, PMS2, MSH2, MSH6, HER2, EGFR, MET and p53. In the data, the expression of EGFR, HER2, and MET was scored according to Hofmann’s criteria, as 0, 1+, 2+, or 3+.31 Cases with complete loss or strong and diffuse (more than 50% of tumor cells) nuclear p53 staining were classified as p53 mutant pattern, and cases with weak and patchy (less than 50% of tumor cells) staining of p53 as wild-type pattern, as described previously.27 Cases with complete negativity for MLH1/PMS2 or MSH2/MSH6 in tumor cells were regarded as MMR deficient, and all others as MMR proficient.

4. EBER-ISH

EBER-ISH was performed with a Ventana BenchMark ISH system (ISH iView kit; Ventana, Tucson, AZ, USA). Sections were deparaffinized with EZ Prep buffer (Ventana), and then digested with protease I for 4 minutes. Probes were hybridized and labeled with fluorescein according to the manufacturer’s protocol, as described previously.29 The slides were counterstained with Nuclear Fast Red for 10 minutes. EBER-ISH was performed in all cases. Also, as previously described, cases showing diffuse strong positivity in the nuclei of all tumor cells were defined as EBER-ISH-positive. Representative image of EBER-ISH-positive cases and EBER-ISH-negative cases are presented in Supplementary Fig. 1.

5. Statistical analysis

Statistical calculation was performed with SPSS version 24.0 (IBM Corp., Armonk, NY, USA). The chi-square or Pearson chi-square tests were used to analyze the relationships between ARID1A expression and variable clinicopathologic parameters. Recurrence-free survival was calculated from the date of operation to the date of first recurrence or death without any type of relapse. Overall survival was calculated from the date of gastrectomy to the date of the last follow-up or death. Survival curves were analyzed using the Kaplan-Meier product limit method with the log-rank test for evaluation of significant differences. For multivariate survival analysis, variables found to be significant on univariate analysis were used with the Cox proportional hazard regression model. Significance statements refer to p-value <0.05.

RESULTS

1. ARID1A expression patterns and related clinicopathologic features in AGC

In nonneoplastic mucosa, diffuse and homogenous nuclear expression of ARID1A was observed in either epithelial cells or stromal cells, such as inflammatory cells and fibroblasts (Fig. 1A). Of 1,032 AGCs, 429 cases (41.6%) showed preserved nuclear expression of ARID1A (Fig. 1B); however, the expression of ARID1A was decreased in 442 cases (42.8%) (Fig. 1C) and absent in 161 AGCs (15.6%) (Fig. 1D).

Figure 1. Representative photographs of immunohistochemical AT-rich interactive domain 1A (ARID1A) expression patterns. (A) Ubiquitous nuclear expression in intestinal metaplastic gastric epithelial cells and stromal cells (lymphocytes, endothelial cells, and fibrocytes). (B) Gastric cancer showing an intensity of ARID1A staining similar to that of nonneoplastic cells, classified as “preserved.” (C) Adenocarcinoma with significantly decreased intensity of ARID1A compared to that of stromal cells, classified as “decreased.” (D) Case demonstrating complete loss of nuclear ARID1A staining (A: ×100, B-D: ×200).

Clinicopathologic and molecular features according to the expression patterns of ARID1A are summarized in Table 1. Decreased expression and loss of ARID1A were more frequently observed in tumors located in the proximal (upper and mid third) stomach (p=0.032 and p=0.019, respectively). The loss of ARID1A was associated with larger tumor size (p=0.003) and intestinal type of Lauren classification (p<0.001). However, except for proximal location, no other parameters were associated with decreased expression of ARID1A (p-value for preserved vs decreased) (Table 1). Consistent with previous reports,16,18,19,23 loss of expression was frequently observed in MLH1-deficient GC (MLH1-loss [53/101, 52.5%] vs MLH1-intact [108/931, 11.6%], p<0.001 for preserved group vs loss group) and EBV-positive GCs (EBV-positive [24/67, 35.8%] vs EBV-negative [137/965, 14.2%], p<0.001 for preserved group vs loss group). However, decreased expression of ARID1A was not associated MLH1 deficiency (MLH1-loss [28/101, 27.7%] vs MLH1-intact [414/931, 44.5%], p=0.279 for preserved group vs decreased group) nor EBV positivity (EBV-positive [22/67, 32.8%] vs EBV-negative [420/965, 43.5%], p=0.955 for preserved group vs decreased group). The loss of ARID1A expression was related to wild-type pattern of p53 (p=0.002 for preserved vs loss and p=0.033 for decreased vs loss) and lower frequency of 2 or 3+ expression of HER2 and MET among total cases.

Table 1 . Clinicopathologic and Molecular Characteristics According to ARID1A Expression Patterns in Advanced Gastric Cancer.

CategoryVariableNo. of
cases (%)
ARID1A expression, No. (%)p-valuep-value
(preserved vs decreased)
p-value
(preserved vs loss)
p-value
(decreased vs loss)
Preserved (n=429)Decreased (n=442)Loss (n=161)
SexMale677 (65.6)289 (67.4)281 (63.6)107 (66.5)0.485
Female355 (34.4)140 (32.6)161 (36.4)54 (33.5)
Age≤60 yr549 (53.2)232 (54.1)245 (55.4)72 (44.7)0.059
>60 yr483 (46.8)197 (45.9)197 (44.6)89 (55.3)
LocationLower third574 (55.6)259 (60.4)235 (53.2)80 (49.7)0.0260.0320.0190.449
Upper and mid458 (44.4)170 (39.6)207 (46.8)81 (50.3)
Size≤5 cm520 (50.4)219 (51.0)239 (54.1)62 (38.5)0.0030.3720.0070.001
>5 cm512 (49.6)210 (49.0)203 (45.9)99 (61.5)
HistologyWD/MD295 (28.6)135 (31.5)116 (26.2)44 (27.3)0.217
PD/others737 (71.4)294 (68.5)326 (73.8)117 (72.7)
LaurenIntestinal504 (48.8)210 (49.0)194 (43.9)100 (62.1)<0.0010.1340.004<0.001
Diffuse528 (51.2)219 (51.0)248 (56.1)61 (37.9)
LVIAbsent735 (71.2)297 (69.2)316 (71.5)122 (75.8)0.290
Present297 (28.8)132 (30.8)126 (28.5)39 (24.2)
LNMAbsent289 (28.0)122 (28.4)123 (27.8)44 (27.3)0.959
Present743 (72.0)307 (71.6)319 (72.2)117 (72.7)
T stageT2176 (17.1)81 (18.9)74 (16.7)21 (13.0)0.0230.3020.0430.009
T3369 (35.8)152 (35.4)143 (32.4)74 (46.0)
T4487 (47.2)196 (45.7)225 (50.9)66 (41.0)
Overall stageII107 (10.4)49 (11.4)45 (10.2)13 (8.1)0.535
III314 (30.4)133 (31.0)129 (29.2)52 (32.3)
IV611 (59.2)247 (57.6)268 (60.6)96 (59.6)
MLH1Loss101 (9.8)20 (4.7)28 (6.3)53 (32.9)<0.0010.279<0.001<0.001
Intact931 (90.2)409 (95.3)414 (93.7)108 (67.1)
MSH2Loss13 (1.3)4 (0.9)7 (1.6)2 (1.2)0.690
Intact1,019 (98.7)425 (99.1)435 (98.4)159 (98.8)
MMR proteinsDeficient114 (11.0)24 (5.6)35 (7.9)55 (34.2)<0.0010.172<0.001<0.001
Proficient918 (89.0)405 (94.4)407 (92.1)106 (65.8)
EBVPositive67 (6.5)21 (4.9)22 (5.0)24 (14.9)<0.0010.955<0.001<0.001
Negative965 (93.5)408 (95.1)420 (95.0)137 (85.1)
MLH1 & EBVMLH1-loss or EBV+168 (16.3)41 (9.6)50 (11.3)77 (47.8)<0.0010.397<0.001<0.001
MLH1 intact and EBV–864 (83.7)388 (90.4)392 (88.7)84 (52.2)
MMRs & EBVMMR-d or EBV+181 (17.5)45 (10.5)57 (12.9)79 (49.1)<0.0010.269<0.001<0.001
MMR-p and EBV–851 (82.5)384 (89.5)385 (87.1)82 (50.9)
p53 IHC*Wild-type pattern624 (62.0)243 (57.7)268 (62.5)113 (72.0)0.0070.1570.0020.033
Mutant pattern383 (38.0)178 (42.3)161 (37.5)44 (28.0)
HER20 or 1+972 (94.2)393 (91.6)418 (94.6)161 (100.0)<0.0010.084<0.0010.003
2 or 3+60 (5.8)36 (8.4)24 (5.4)0
MET*0 or 1+916 (90.5)381 (90.9)399 (92.1)136 (85.0)0.0290.5230.0390.009
2 or 3+96 (9.5)38 (9.1)34 (7.9)24 (15.0)
EGFR*0 or 1+865 (84.5)346 (81.0)393 (90.1)126 (78.3)<0.001<0.0010.452<0.001
2 or 3+159 (15.5)81 (19.0)43 (9.9)35 (21.7)

ARID1A, AT-rich interactive domain 1A; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; MLH1, mutL homolog 1; MSH2, mutS homolog 2; MMR, mismatch-repair protein; EBV, Epstein-Barr virus; MMR-d, MMR deficient; MMR-p, MMR proficient; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor..

*The cases with inadequate immunohistochemical staining results were excluded..



2. ARID1A expression patterns and related clinicopathologic features in MLH1-proficient and EBV-negative AGC

MLH1-deficient GC is known for its unique clinicopathologic features, including intestinal histology, distal tumor location, Crohn-like peritumoral lymphoid reaction, and genome-wide CpG island hypermethylation.26,32 EBV-positive GC also shows characteristic features, including proximal tumor location, frequent occurrence in remnant stomach, and intratumoral intense lymphocytic infiltration, which contributes to a unique histologic type termed carcinoma with lymphoid stroma.33 Since the loss of ARID1A expression has shown close relationships with MLH1 deficiency and EBV positivity, we analyzed the relationships between ARID1A patterns and clinicopathologic and molecular features in a MLH1-proficient and EBV-negative group to remove any possible confounding effects from those two subtypes (Table 2). In this group, 384 (45.1%), 385 (45.2%), and 82 (9.6%) cases showed preserved, decreased, and loss of ARID1A expression, respectively. An association between the decrease or loss of ARID1A and the proximal location was observed (p=0.01). However, larger tumor size and intestinal-type histology were not associated with the altered pattern. Interestingly, in this analysis, the association between the wild-type pattern of p53 and the loss of ARID1A expression, shown in Table 1, was not observed in the MLH1-proficient and EBV-negative subgroups. The associations between the receptor tyrosine kinases expression and ARID1A loss were conserved, with the exception of MET.

Table 2 . Clinicopathologic and Molecular Characteristics According to ARID1A Expression Patterns in MLH1-Proficient and EBV-Negative Advanced Gastric Cancer.

CategoryVariableNo. of cases (%)ARID1A expression, No. (%)p-value
Preserved (n=388)Decreased (n=392)Loss (n=84)
SexMale561 (64.9)260 (67.0)244 (62.2)57 (67.9)0.317
Female303 (35.1)128 (33.0)148 (37.8)27 (32.1)
Age≤60 yr476 (55.1)210 (54.1)222 (56.6)44 (52.4)0.680
>60 yr388 (44.9)178 (45.9)170 (43.4)40 (47.6)
LocationLower third487 (56.4)239 (61.6)209 (53.3)39 (46.4)0.010
Upper and mid377 (43.6)149 (38.4)183 (46.7)45 (53.6)
Size≤5 cm456 (52.8)204 (52.6)221 (56.4)31 (36.9)0.005
>5 cm408 (47.2)184 (47.4)171 (43.6)53 (63.1)
HistologyWD/MD242 (28.0)121 (31.2)99 (25.3)22 (26.2)0.169
PD/others622 (72.0)267 (68.8)293 (74.7)62 (73.8)
LaurenIntestinal379 (43.9)180 (46.4)160 (40.8)39 (46.4)0.258
Diffuse485 (56.1)208 (53.6)232 (59.2)45 (53.6)
LVIAbsent616 (71.3)271 (69.8)281 (71.7)64 (76.2)0.494
Present248 (28.7)117 (30.2)111 (28.3)20 (23.8)
LNMAbsent239 (27.7)111 (28.6)109 (27.8)19 (22.6)0.537
Present625 (72.3)277 (71.4)283 (72.2)65 (77.4)
T stageT2149 (17.2)74 (19.1)68 (17.3)7 (8.3)0.104
T3287 (33.2)132 (34.0)121 (30.9)34 (40.5)
T4428 (49.5)182 (46.9)203 (51.8)43 (51.2)
Overall stageII88 (10.2)43 (11.1)41 (10.5)4 (4.8)0.326
III259 (30.0)123 (31.7)112 (28.6)24 (28.6)
IV517 (59.8)222 (57.2)239 (61.0)56 (66.7)
p53 IHC*Wild-type pattern493 (58.6)214 (56.0)227 (59.9)52 (64.2)0.308
Mutant pattern349 (41.4)168 (44.0)152 (40.1)29 (35.8)
HER20 or 1+809 (93.6)355 (91.5)370 (94.4)84 (100.0)0.011
2 or 3+55 (6.4)33 (8.5)22 (5.6)0
MET*0 or 1+781 (92.2)349 (91.8)358 (93.2)74 (89.2)0.427
2 or 3+66 (7.8)31 (8.2)26 (6.8)9 (10.8)
EGFR*0 or 1+741 (86.6)317 (82.1)354 (91.7)70 (83.3)<0.001
2 or 3+115 (13.4)69 (17.9)32 (8.3)14 (16.7)

ARID1A, AT-rich interactive domain 1A; MLH1, mutL homolog 1; EBV, Epstein-Barr virus; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor..

*The cases with inadequate immunohistochemical staining results were excluded..



3. ARID1A expression patterns and clinicopathologic features in MLH1-deficient and EBV-positive AGCs

In this study, MLH1-deficient AGCs and EBV-positive AGCs were mutually exclusive. In MLH1-deficient AGCs (n=101), preserved, decreased, and loss of ARID1A was noted in 20 (19.9%), 28 (27.7%), and 53 (52.5%) cases, respectively (p<0.001 compared to MLH1-proficient GCs) (Table 1). No parameter was associated with altered ARID1A expression (Supplementary Table 1). Among EBV-positive AGCs (n=67), 21 (31.3%), 22 (32.8%), and 24 (35.8%) showed preserved, decreased, and loss of ARID1A expression, respectively. A low frequency of LVI was the only factor associated with decrease or loss of ARID1A expression (p=0.044) (Supplementary Table 2).

4. Prognostic significance of decreased or loss of ARID1a expression in AGC

Kaplan-Meier survival curves according to ARID1A expression showed worse prognosis of the patient group with decreased ARID1A compared to patients with preserved ARID1A (Fig. 2A). However, the ARID1A-loss group exhibited a similar prognosis to that of the ARID1A-preserved group. Since GCs with ARID1A loss were enriched among MLH1-deficient and EBV-positive GCs (Table 1) and the prognosis of GC patients with MLH1-deficiency or EBV-positivity was more favorable than that of patients with EBV-negative and MLH1-proficient AGC (Fig. 2B), we analyzed the prognostic effects of altered ARID1A expression in MLH1-deficient, EBV-positive, and EBV-negative and MLH1-proficient groups, respectively. There were no prognostic differences among the ARID1A preserved, decreased, and loss groups in either MLH1-deficient (Fig. 3A) or EBV-positive GCs (Fig. 3B). However, in the patient group with MLH1-proficient and EBV-negative AGC, the prognosis was significantly different according to the ARID1A expression patterns: patients with ARID1A-loss AGC showed the worst overall survival; and the ARID1A-decreased group had second-worst prognosis (p<0.001) (Fig. 2C).

Figure 2. Kaplan-Meier survival curves for overall survival with AGC according to three ARID1A expression patterns and molecular subtypes. (A) Overall survival curves of the total group according to ARID1A expression pattern show a worse prognosis of the ARID1A-decreased group than the ARID1A-preserved or ARID1A-loss group. (B) Overall survival curves comparing the EBV-positive group, the MLH1-loss group and the MLH-proficient and EBV-negative group reveal a favorable prognosis of AGC with the molecular subtype of the EBV-positive and MLH1-loss group. (C) Among MLH1-proficient and EBV-negative AGCs, survival curves according to ARID1A expression pattern demonstrate a worse prognosis of patients with ARID1A decreased or ARID1A loss AGCs. (D) Among either MLH1-deficient or EBV-positive AGCs, survival curves according to ARID1A expression pattern show no difference in prognosis.
AGC, advanced gastric cancer; ARID1A, AT-rich interactive domain 1A; EBV, Epstein-Barr virus; MLH1, mutL homolog 1.

Figure 3. Kaplan-Meier survival curves for overall survival of MLH1-deficient and EBV-positive AGC. (A) Overall survival curves of MLH1-deficient AGCs according to three ARID1A expression patterns. (B) Overall survival curves of EBV-positive AGCs according to three ARID1A expression patterns.
MLH1, mutL homolog 1; EBV, Epstein-Barr virus; AGC, advanced gastric cancer; ARID1A, AT-rich interactive domain 1A.

In multivariate survival analysis, Cox regression analyses for overall survival and recurrence-free survival were performed using variables that were significant on univariate analysis, including age, tumor location, tumor size, histology, Lauren classification, LVI, lymph node metastasis, overall stage, MLH1 statue, EBV positivity, and ARID1A expression patterns, using a forward conditional method. For overall survival, MLH1 proficiency and EBV negativity were independent worse prognostic factors (hazard ratio [HR], 1.4; p<0.001 and HR, 1.98; p<0.001, respectively). Decrease or loss of ARID1A expression was also revealed as a negative prognostic factor of AGC (HR, 1.47 for decreased and HR, 1.48 for loss pattern, p<0.001). In addition to MLH1, EBV, and ARID1A expression, older age (HR, 1.58; p<0.001), larger tumor size (>5 cm) (HR, 1.24; p=0.008), diffuse type by Lauren (HR, 1.26; p=0.009), presence of LVI (HR, 1.54; p<0.001), and overall stage (HR, 1.69 for stage III and HR, 4.38 for stage IV, p<0.001) were found to be prognostic factors of overall survival. For recurrence-free survival, ARID1A expression pattern (p=0.041) was a prognostic factor, in addition to MLH1 proficiency (p<0.001), EBV negativity (p=0.002), and the features that were shown to be prognostic factors of overall survival, except for older age and proximal tumor location (Table 3).

Table 3 . Univariate and Multivariate Survival Analyses.

CategoryVariableUnivariate-OSMultivariate-OSUnivariate-RFSMultivariate-RFS
HR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-value
SexMale11
Female1.08 (0.92–1.27)0.3421.18 (0.98–1.42)0.083
Age≤60 yr111
>60 yr1.41 (1.20–1.65)<0.0011.58 (1.34–1.88)<0.0010.96 (0.80–1.15)0.657
LocationLower third111
Upper and mid1.36 (1.16–1.60)<0.0011.25 (1.06–1.48)0.0081.27 (1.05–1.53)0.013
Size≤5 cm1111
>5 cm1.75 (1.49–2.07)<0.0011.24 (1.04–1.48)0.0181.82 (1.51–2.20)<0.0011.32 (1.09–1.61)0.005
HistologyWD/MD11
PD/others1.36 (1.14–1.63)0.0011.60 (1.29–1.99)<0.001
LaurenIntestinal1111
Diffuse1.38 (1.17–1.61)<0.0011.26 (1.06–1.49)0.0091.68 (1.40–2.03)<0.0011.31 (1.08–1.60)0.007
LVIAbsent1111
Present2.21 (1.87–2.60)<0.0011.54 (1.29–1.84)<0.0012.31 (1.92–2.79)<0.0011.49 (1.22–1.82)<0.001
LNMAbsent11
Present3.24 (2.62–4.00)<0.0014.13 (3.15–5.41)<0.001
T stageT211<0.001
T32.44 (1.82–3.28)<0.0013.41 (2.27–5.13)<0.001
T44.96 (3.75–6.56)<0.0017.96 (5.40–11.74)<0.001
Overall stageII11<0.0011<0.0011<0.001
III1.96 (1.32–2.92)0.0011.69 (1.08–2.65)0.0233.01 (1.64–5.51)<0.0012.34 (1.21–4.55)0.012
IV6.03 (4.16–8.73)<0.0014.38 (2.83–6.76)<0.00111.97 (6.73–21.29)<0.0017.78 (4.10–14.76)<0.001
MLH1Loss1111
Intact1.47 (1.08–2.00)0.0141.85 (1.31–2.61)<0.0011.94 (1.31–2.89)0.0011.99 (1.29–3.07)0.002
MSH2Loss11
Intact0.57 (0.31–1.07)0.0800.59 (0.28–1.25)0.166
MMR proteinsDeficient11
Proficient1.30 (0.98–1.72)0.0661.65 (1.16–2.36)0.005
EBVPositive1111
Negative1.98 (1.31–2.97)0.0012.36 (1.54–3.62)<0.0011.98 (1.23–3.17)0.0052.12 (1.31–3.44)0.002
MLH1 & EBVMLH1-loss or EBV+11
MLH1-intact and EBV+1.73 (1.34–2.22)<0.0012.07 (1.51–2.82)<0.001
MMR-d & EBVMMR-d or EBV+11
MMR-p and EBV–1.57 (1.24–2.00)<0.0011.86 (1.39–2.49)<0.001
ARID1APreserved10.0031<0.00110.04110.041
Decreased1.36 (1.14–1.62)0.0011.47 (1.23–1.76)<0.0011.22 (1.00–1.49)0.0541.29 (1.06–1.58)0.013
Loss1.12 (0.87–1.44)0.3881.48 (1.12–1.95)0.0050.88 (0.65–1.19)0.4031.25 (0.90–1.73)0.189
p53Wild-type pattern11
Mutant pattern1.09 (0.92–1.29)0.3061.09 (0.90–1.33)0.355

Human epidermal growth factor receptor 2, epidermal growth factor receptor, and mesenchymal-epithelial transition factor: p>0.05 on univariate analysis (data not shown in table)..

OS, overall survival; RFS, recurrence-free survival; HR, hazard ratio; CI, confidence interval; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; MLH1, mutL homolog 1; MSH2, mutS homolog 2; MMR, mismatch-repair protein; EBV, Epstein-Barr virus; MMR-d, MMR deficient; MMR-p, MMR proficient; ARID1A, AT-rich interactive domain 1A..


DISCUSSION

The loss of ARID1A expression correlates well with mutation status, although mutation does not account for all cases of loss of expression. Mutation or loss of ARID1A in GC are closely related to MSI-H type and EBV positivity.11,16-18,23,34 In this study, 52.5% (53/101) of MLH1-deficient and 35.8% (24/67) of EBV-positive GCs exhibited loss of ARID1A expression, compared with only 9.6% (82/851) of the MLH1-proficient and EBV-negative group (p<0.001) (Table 1). Therefore, we confirmed the strong associations between loss of ARID1A expression and MLH1 deficiency and EBV positivity using the largest sample set to date. Interestingly, in contrast to GCs with ARID1A loss, cases with decreased ARID1A were not associated with MLH1-deficiency or EBV-positivity. With the exception of one study by Kim et al.,18 previous studies that evaluated ARID1A expression in GC categorized the expression as only positive or negative and did not evaluate the association between a decreased pattern of ARID1A expression and EBV-positive or MMR-deficient GC.11,16-19,23,34 Kim et al.18 classified the patterns of ARID1A expression as retained, reduced, complete loss, and partial loss. In their study, reduced ARID1A expression was found in 17.7% of GCs and was not associated with EBV positivity, but was frequently found in the MLH1-deficient group.18 This discrepancy with our results, which revealed no association between the decreased pattern and MLH1 deficiency, might result from the different incidence of MLH1-deficient GC (18.9% vs 9.8% in our study) and of ARID1A decreased cases (17.7% vs 40.9% in our study).

The expression of MSH2, another key MMR protein, was not associated with ARID1A loss or decreased expression, although only limited cases were included in this category. MSH2 deficiency indicates a possible germline mutation of the MSH2 gene (so-called Lynch syndrome), whereas the vast majority of cases with MLH1 deficiency result from epigenetic silencing of MLH1 in the context of a CpG island methylator phenotype35-38 that shows genome-wide hypermethylation.26 Genome-wide hypermethylation is also a well-known characteristic of EBV-positive GC.26,39 Considering this shared epigenetic characteristic between EBV-positive and MLH1-deficient GC and the lack of a relationship between ARID1A alteration and MSH2-deficiency, it might be postulated that the frequent loss of ARID1A expression in MSI-H GC is associated with hypermethylation rather than instability of microsatellites or a mutator phenotype.

A negative association between mutations in TP53 and ARID1A has been reported in GC and endometrial cancer.11 In addition, a relationship between wild-type p53 staining pattern (weak and patch nuclear staining) and loss of ARID1A expression in GC and endometrial cancers was demonstrated.17,40 We also found an association between wild-type pattern of p53 expression and loss of ARID1A (p=0.007) (Table 1). However, in subgroup analysis, this relationship was lost in the MLH1-proficient and EBV-negative groups (Table 2), and was also not observed in MLH1-deficient GCs (Supplementary Table 1) or EBV-positive GCs (Supplementary Table 2). It is well known that EBV-positive and MLH1-deficient GCs are associated with wild-type TP53.26,33 We also found a high frequency of wild-type pattern of p53 staining in MLH1-deficient (81%, 81/100) and EBV-positive GCs (76.9%, 50/65), compared to 58.6% (493/864) of the MLH1-proficient and EBV-negative group. Therefore, it can be assumed that the previously reported association between the wild-type pattern of p53 staining and the loss of ARID1A may reflect the strong relationship between loss of ARID1A and EBV-positive and MLH1-deficient GC. To rule out the possible confounding effects of EBV positivity and MLH1 deficiency on this association, we performed a binary logistic regression analysis using the loss of ARID1 expression as a dependent variable and other parameters that showed a correlation with the loss of ARID1A in Table 1, including EBV positivity, MLH1 deficiency, tumor size, T stage, p53 staining pattern, and HER2 and MET expression, as independent variables. The logistic regression analysis revealed that only EBV positivity (odds ratio, 5.119; p<0.001) and MLH1 deficiency (odds ratio, 8.376; p<0.001) were correlated with the loss of ARID1A (Table 4).

Table 4 . Logistic Regression Analysis of Clinicopathologic and Molecular Features Associated with Loss of ARID1A Expression.

FeatureOdds ratio (95% CI)p-value
Location (upper and mid)1.35 (0.90–2.01)0.143
Size (>5 cm)1.23 (0.82–1.86)0.315
Lauren type (intestinal)1.38 (0.92–2.07)0.125
T stage0.242
T31.07 (0.58–2.00)0.820
T40.73 (0.48–1.11)0.140
MLH1 (deficient)8.12 (4.82–13.66)<0.001
EBV (positive)4.13 (2.28–7.48)<0.001
p53 (wild-type pattern)1.17 (0.77–1.77)0.459
HER2 (2 or 3+)NA0.997
MET (2 or 3+)1.54 (0.82–2.89)0.175
EGFR (2 or 3+)1.08 (0.62–1.89)0.781

ARID1A, AT-rich interactive domain 1A; CI, confidence interval; MLH1, mutL homolog 1; EBV, Epstein-Barr virus; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor; NA, not available..



Previous studies have shown that the loss of ARID1A expression is correlated with worse prognosis in patients with GC13,16,18,23 or only in patients with EBV-negative and MLH1-preserved GC.16 However, in some studies, the relationship between loss of ARID1A expression and prognosis was not conclusive.11,17,20,24 In our study, the loss of ARID1A showed different prognostic effects in GC patients according to the molecular subtype. In total cases, the prognosis of patients with ARID1A loss was not significantly different from that of patients with preserved ARID1A expression (Fig. 2A). However, in EBV-negative and MLH1-proficient GC patients, the loss of ARID1A expression was associated with the worst prognosis (Fig. 2C). The favorable prognosis of EBV-positive and MLH1-deficient GC patients (Fig. 2B) and the strong correlation between loss of ARID1A and EBV-positive and MLH1-deficient GCs (Table 1) may account for the different prognostic effects of loss of ARID1A among the molecular subtypes. This finding was in good agreement with a previous study by Abe et al.;16 however, in their study, only eight cases of EBV-positive and 36 cases of MLH1-negative GC were enrolled in the survival analysis;16 therefore, there was a limitation in evaluating the prognostic effect of ARID1A loss in patients with EBV-positive and MLH1-deficient GC. In this study, we confirmed that loss of ARID1A expression was an independent negative prognostic factor in the MLH1-proficient and EBV-negative AGC (Table 3, Fig. 2C), but not in EBV-positive or MLH1-deficient AGC (Figs 2D and 3).

It has been reported that IHC for ARID1A can be used as surrogate marker for ARID1A mutation status.11,14,16 However, ARID1A alterations can occur in other ways as well as mutation. It has been reported that ARID1A in cancer can be regulated by way of promotor methylation or post-transcriptional modification.15,41 In addition, there is also a report that ARID1A can be regulated by EBV-encoded miRNA in EBV-positive GCs.42 Due to the large number of the sample of this study, there are limitations in analyzing mutation and epigenetic profiles for every tumors; however, we expect that decreased ARID1A expression is associated with non-mutational alteration of ARID1A and results of this study support the supposition. In this study, patients with decreased expression of ARID1A accounted for 42.8% of total cases and 44.6% of the EBV-negative and MLH1-proficient groups. The incidence of ARID1A mutation in MLH1-proficient and EBV-negative GC has been reported to be 8% to 10%.11,12 However, the proportion of patients with altered expression of ARID1A in this study was 54.8% (45.2% for ARID1A decrease and 9.6% for loss) of total cases, which implied that there might be another mechanism suppressing the expression of ARID1A other than the genetic mutation. In ovarian clear cell carcinoma, biallelic mutations were found in only 30% of cases with ARID1A mutation, whereas 73% of ARID1A-heterozygous cases showed loss of protein expression by IHC.10 In breast cancer, promoter hypermethylation of ARID1A was reported to be strongly correlated with a low level of messenger RNA expression.15 Further investigation into the suppression mechanism of ARID1A might help elucidate the underlying mechanism regulating ARID1A expression in GC. In addition, patients with decreased ARID1A showed worse prognosis among total cases and in the EBV-negative and MLH1-proficient group (Fig. 2). On multivariate analysis, we found that decreased ARID1A was an independent poor prognostic factor in AGC patients, in addition to the loss of ARID1A (Table 3). Therefore, understanding the mechanisms involved in reduced expression of ARID1A is important not only biologically, but also clinically.

Recently, Kim et al.30 reported that synthetic lethality of EZH2, a histone methyl transferase subunit of polycomb repressor complex, is related to ARID1A mutation. Tumor cell lines with ARID1A mutation undergo cell death and are inhibited in their ability of tumor formation in vivo when treated with the EZH2 inhibitor, GSK126.30 Therefore, GC with altered ARID1 expression might be a potential candidate for EZH2-targeted treatment in the future.

In conclusion, the loss of ARID1A expression was associated with larger tumor size, intestinal histology, MLH1 deficiency, EBV positivity, wild-type pattern of p53 staining, and negative or 1+ HER2 and MET expression. However, on logistic regression analysis, only MLH1-deficiency and EBV positivity showed a correlation with the loss of ARID1A. In MLH1-proficient and EBV-negative GC, wild-type pattern of p53 staining was not associated with ARID1A loss. In addition to the loss of ARID1A expression, decreased ARID1A was also revealed as an independent negative prognostic factor in AGC patients. Interestingly, no prognostic significance of altered ARID1A expression was found in MLH1-deficient or EBV-positive GC. Regarding the emerging concept of synthetic lethality associated with ARID1A mutation, GC with reduced or loss of ARID1A expression might be a good candidate for new targeted treatments.

SUPPLEMENTARY MATERIALS

Supplementary materials can be accessed at https://doi.org/10.5009/gnl220342.

ACKNOWLEDGEMENTS

This work was supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (Ministry of Science and ICT) (2019R1A2C1011355, H.K.).

CONFLICTS OF INTEREST

No potential conflict of interest relevant to this article was reported.

AUTHOR CONTRIBUTIONS

No potential conflict of interest relevant to this article was reported.

Study concept and design: H.K. Data acquisition: C.K.P., S.N. Data analysis and interpretation: J.Y.K., C.K.P., S.N. Drafting of the manuscript: J.Y.K., S.N. Critical revision of the manuscript for important intellectual content: J.H.C., S.H.N. Statistical analysis: J.Y.K., C.K.P., S.N. Obtained funding: H.K. Administrative, technical, or material support; study supervision: H.K. Approval of final manuscript: all authors.

Fig 1.

Figure 1.Representative photographs of immunohistochemical AT-rich interactive domain 1A (ARID1A) expression patterns. (A) Ubiquitous nuclear expression in intestinal metaplastic gastric epithelial cells and stromal cells (lymphocytes, endothelial cells, and fibrocytes). (B) Gastric cancer showing an intensity of ARID1A staining similar to that of nonneoplastic cells, classified as “preserved.” (C) Adenocarcinoma with significantly decreased intensity of ARID1A compared to that of stromal cells, classified as “decreased.” (D) Case demonstrating complete loss of nuclear ARID1A staining (A: ×100, B-D: ×200).
Gut and Liver 2023; 17: 753-765https://doi.org/10.5009/gnl220342

Fig 2.

Figure 2.Kaplan-Meier survival curves for overall survival with AGC according to three ARID1A expression patterns and molecular subtypes. (A) Overall survival curves of the total group according to ARID1A expression pattern show a worse prognosis of the ARID1A-decreased group than the ARID1A-preserved or ARID1A-loss group. (B) Overall survival curves comparing the EBV-positive group, the MLH1-loss group and the MLH-proficient and EBV-negative group reveal a favorable prognosis of AGC with the molecular subtype of the EBV-positive and MLH1-loss group. (C) Among MLH1-proficient and EBV-negative AGCs, survival curves according to ARID1A expression pattern demonstrate a worse prognosis of patients with ARID1A decreased or ARID1A loss AGCs. (D) Among either MLH1-deficient or EBV-positive AGCs, survival curves according to ARID1A expression pattern show no difference in prognosis.
AGC, advanced gastric cancer; ARID1A, AT-rich interactive domain 1A; EBV, Epstein-Barr virus; MLH1, mutL homolog 1.
Gut and Liver 2023; 17: 753-765https://doi.org/10.5009/gnl220342

Fig 3.

Figure 3.Kaplan-Meier survival curves for overall survival of MLH1-deficient and EBV-positive AGC. (A) Overall survival curves of MLH1-deficient AGCs according to three ARID1A expression patterns. (B) Overall survival curves of EBV-positive AGCs according to three ARID1A expression patterns.
MLH1, mutL homolog 1; EBV, Epstein-Barr virus; AGC, advanced gastric cancer; ARID1A, AT-rich interactive domain 1A.
Gut and Liver 2023; 17: 753-765https://doi.org/10.5009/gnl220342

Table 1 Clinicopathologic and Molecular Characteristics According to ARID1A Expression Patterns in Advanced Gastric Cancer

CategoryVariableNo. of
cases (%)
ARID1A expression, No. (%)p-valuep-value
(preserved vs decreased)
p-value
(preserved vs loss)
p-value
(decreased vs loss)
Preserved (n=429)Decreased (n=442)Loss (n=161)
SexMale677 (65.6)289 (67.4)281 (63.6)107 (66.5)0.485
Female355 (34.4)140 (32.6)161 (36.4)54 (33.5)
Age≤60 yr549 (53.2)232 (54.1)245 (55.4)72 (44.7)0.059
>60 yr483 (46.8)197 (45.9)197 (44.6)89 (55.3)
LocationLower third574 (55.6)259 (60.4)235 (53.2)80 (49.7)0.0260.0320.0190.449
Upper and mid458 (44.4)170 (39.6)207 (46.8)81 (50.3)
Size≤5 cm520 (50.4)219 (51.0)239 (54.1)62 (38.5)0.0030.3720.0070.001
>5 cm512 (49.6)210 (49.0)203 (45.9)99 (61.5)
HistologyWD/MD295 (28.6)135 (31.5)116 (26.2)44 (27.3)0.217
PD/others737 (71.4)294 (68.5)326 (73.8)117 (72.7)
LaurenIntestinal504 (48.8)210 (49.0)194 (43.9)100 (62.1)<0.0010.1340.004<0.001
Diffuse528 (51.2)219 (51.0)248 (56.1)61 (37.9)
LVIAbsent735 (71.2)297 (69.2)316 (71.5)122 (75.8)0.290
Present297 (28.8)132 (30.8)126 (28.5)39 (24.2)
LNMAbsent289 (28.0)122 (28.4)123 (27.8)44 (27.3)0.959
Present743 (72.0)307 (71.6)319 (72.2)117 (72.7)
T stageT2176 (17.1)81 (18.9)74 (16.7)21 (13.0)0.0230.3020.0430.009
T3369 (35.8)152 (35.4)143 (32.4)74 (46.0)
T4487 (47.2)196 (45.7)225 (50.9)66 (41.0)
Overall stageII107 (10.4)49 (11.4)45 (10.2)13 (8.1)0.535
III314 (30.4)133 (31.0)129 (29.2)52 (32.3)
IV611 (59.2)247 (57.6)268 (60.6)96 (59.6)
MLH1Loss101 (9.8)20 (4.7)28 (6.3)53 (32.9)<0.0010.279<0.001<0.001
Intact931 (90.2)409 (95.3)414 (93.7)108 (67.1)
MSH2Loss13 (1.3)4 (0.9)7 (1.6)2 (1.2)0.690
Intact1,019 (98.7)425 (99.1)435 (98.4)159 (98.8)
MMR proteinsDeficient114 (11.0)24 (5.6)35 (7.9)55 (34.2)<0.0010.172<0.001<0.001
Proficient918 (89.0)405 (94.4)407 (92.1)106 (65.8)
EBVPositive67 (6.5)21 (4.9)22 (5.0)24 (14.9)<0.0010.955<0.001<0.001
Negative965 (93.5)408 (95.1)420 (95.0)137 (85.1)
MLH1 & EBVMLH1-loss or EBV+168 (16.3)41 (9.6)50 (11.3)77 (47.8)<0.0010.397<0.001<0.001
MLH1 intact and EBV–864 (83.7)388 (90.4)392 (88.7)84 (52.2)
MMRs & EBVMMR-d or EBV+181 (17.5)45 (10.5)57 (12.9)79 (49.1)<0.0010.269<0.001<0.001
MMR-p and EBV–851 (82.5)384 (89.5)385 (87.1)82 (50.9)
p53 IHC*Wild-type pattern624 (62.0)243 (57.7)268 (62.5)113 (72.0)0.0070.1570.0020.033
Mutant pattern383 (38.0)178 (42.3)161 (37.5)44 (28.0)
HER20 or 1+972 (94.2)393 (91.6)418 (94.6)161 (100.0)<0.0010.084<0.0010.003
2 or 3+60 (5.8)36 (8.4)24 (5.4)0
MET*0 or 1+916 (90.5)381 (90.9)399 (92.1)136 (85.0)0.0290.5230.0390.009
2 or 3+96 (9.5)38 (9.1)34 (7.9)24 (15.0)
EGFR*0 or 1+865 (84.5)346 (81.0)393 (90.1)126 (78.3)<0.001<0.0010.452<0.001
2 or 3+159 (15.5)81 (19.0)43 (9.9)35 (21.7)

ARID1A, AT-rich interactive domain 1A; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; MLH1, mutL homolog 1; MSH2, mutS homolog 2; MMR, mismatch-repair protein; EBV, Epstein-Barr virus; MMR-d, MMR deficient; MMR-p, MMR proficient; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor.

*The cases with inadequate immunohistochemical staining results were excluded.


Table 2 Clinicopathologic and Molecular Characteristics According to ARID1A Expression Patterns in MLH1-Proficient and EBV-Negative Advanced Gastric Cancer

CategoryVariableNo. of cases (%)ARID1A expression, No. (%)p-value
Preserved (n=388)Decreased (n=392)Loss (n=84)
SexMale561 (64.9)260 (67.0)244 (62.2)57 (67.9)0.317
Female303 (35.1)128 (33.0)148 (37.8)27 (32.1)
Age≤60 yr476 (55.1)210 (54.1)222 (56.6)44 (52.4)0.680
>60 yr388 (44.9)178 (45.9)170 (43.4)40 (47.6)
LocationLower third487 (56.4)239 (61.6)209 (53.3)39 (46.4)0.010
Upper and mid377 (43.6)149 (38.4)183 (46.7)45 (53.6)
Size≤5 cm456 (52.8)204 (52.6)221 (56.4)31 (36.9)0.005
>5 cm408 (47.2)184 (47.4)171 (43.6)53 (63.1)
HistologyWD/MD242 (28.0)121 (31.2)99 (25.3)22 (26.2)0.169
PD/others622 (72.0)267 (68.8)293 (74.7)62 (73.8)
LaurenIntestinal379 (43.9)180 (46.4)160 (40.8)39 (46.4)0.258
Diffuse485 (56.1)208 (53.6)232 (59.2)45 (53.6)
LVIAbsent616 (71.3)271 (69.8)281 (71.7)64 (76.2)0.494
Present248 (28.7)117 (30.2)111 (28.3)20 (23.8)
LNMAbsent239 (27.7)111 (28.6)109 (27.8)19 (22.6)0.537
Present625 (72.3)277 (71.4)283 (72.2)65 (77.4)
T stageT2149 (17.2)74 (19.1)68 (17.3)7 (8.3)0.104
T3287 (33.2)132 (34.0)121 (30.9)34 (40.5)
T4428 (49.5)182 (46.9)203 (51.8)43 (51.2)
Overall stageII88 (10.2)43 (11.1)41 (10.5)4 (4.8)0.326
III259 (30.0)123 (31.7)112 (28.6)24 (28.6)
IV517 (59.8)222 (57.2)239 (61.0)56 (66.7)
p53 IHC*Wild-type pattern493 (58.6)214 (56.0)227 (59.9)52 (64.2)0.308
Mutant pattern349 (41.4)168 (44.0)152 (40.1)29 (35.8)
HER20 or 1+809 (93.6)355 (91.5)370 (94.4)84 (100.0)0.011
2 or 3+55 (6.4)33 (8.5)22 (5.6)0
MET*0 or 1+781 (92.2)349 (91.8)358 (93.2)74 (89.2)0.427
2 or 3+66 (7.8)31 (8.2)26 (6.8)9 (10.8)
EGFR*0 or 1+741 (86.6)317 (82.1)354 (91.7)70 (83.3)<0.001
2 or 3+115 (13.4)69 (17.9)32 (8.3)14 (16.7)

ARID1A, AT-rich interactive domain 1A; MLH1, mutL homolog 1; EBV, Epstein-Barr virus; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; IHC, immunohistochemistry; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor.

*The cases with inadequate immunohistochemical staining results were excluded.


Table 3 Univariate and Multivariate Survival Analyses

CategoryVariableUnivariate-OSMultivariate-OSUnivariate-RFSMultivariate-RFS
HR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-valueHR (95% CI)p-value
SexMale11
Female1.08 (0.92–1.27)0.3421.18 (0.98–1.42)0.083
Age≤60 yr111
>60 yr1.41 (1.20–1.65)<0.0011.58 (1.34–1.88)<0.0010.96 (0.80–1.15)0.657
LocationLower third111
Upper and mid1.36 (1.16–1.60)<0.0011.25 (1.06–1.48)0.0081.27 (1.05–1.53)0.013
Size≤5 cm1111
>5 cm1.75 (1.49–2.07)<0.0011.24 (1.04–1.48)0.0181.82 (1.51–2.20)<0.0011.32 (1.09–1.61)0.005
HistologyWD/MD11
PD/others1.36 (1.14–1.63)0.0011.60 (1.29–1.99)<0.001
LaurenIntestinal1111
Diffuse1.38 (1.17–1.61)<0.0011.26 (1.06–1.49)0.0091.68 (1.40–2.03)<0.0011.31 (1.08–1.60)0.007
LVIAbsent1111
Present2.21 (1.87–2.60)<0.0011.54 (1.29–1.84)<0.0012.31 (1.92–2.79)<0.0011.49 (1.22–1.82)<0.001
LNMAbsent11
Present3.24 (2.62–4.00)<0.0014.13 (3.15–5.41)<0.001
T stageT211<0.001
T32.44 (1.82–3.28)<0.0013.41 (2.27–5.13)<0.001
T44.96 (3.75–6.56)<0.0017.96 (5.40–11.74)<0.001
Overall stageII11<0.0011<0.0011<0.001
III1.96 (1.32–2.92)0.0011.69 (1.08–2.65)0.0233.01 (1.64–5.51)<0.0012.34 (1.21–4.55)0.012
IV6.03 (4.16–8.73)<0.0014.38 (2.83–6.76)<0.00111.97 (6.73–21.29)<0.0017.78 (4.10–14.76)<0.001
MLH1Loss1111
Intact1.47 (1.08–2.00)0.0141.85 (1.31–2.61)<0.0011.94 (1.31–2.89)0.0011.99 (1.29–3.07)0.002
MSH2Loss11
Intact0.57 (0.31–1.07)0.0800.59 (0.28–1.25)0.166
MMR proteinsDeficient11
Proficient1.30 (0.98–1.72)0.0661.65 (1.16–2.36)0.005
EBVPositive1111
Negative1.98 (1.31–2.97)0.0012.36 (1.54–3.62)<0.0011.98 (1.23–3.17)0.0052.12 (1.31–3.44)0.002
MLH1 & EBVMLH1-loss or EBV+11
MLH1-intact and EBV+1.73 (1.34–2.22)<0.0012.07 (1.51–2.82)<0.001
MMR-d & EBVMMR-d or EBV+11
MMR-p and EBV–1.57 (1.24–2.00)<0.0011.86 (1.39–2.49)<0.001
ARID1APreserved10.0031<0.00110.04110.041
Decreased1.36 (1.14–1.62)0.0011.47 (1.23–1.76)<0.0011.22 (1.00–1.49)0.0541.29 (1.06–1.58)0.013
Loss1.12 (0.87–1.44)0.3881.48 (1.12–1.95)0.0050.88 (0.65–1.19)0.4031.25 (0.90–1.73)0.189
p53Wild-type pattern11
Mutant pattern1.09 (0.92–1.29)0.3061.09 (0.90–1.33)0.355

Human epidermal growth factor receptor 2, epidermal growth factor receptor, and mesenchymal-epithelial transition factor: p>0.05 on univariate analysis (data not shown in table).

OS, overall survival; RFS, recurrence-free survival; HR, hazard ratio; CI, confidence interval; WD, well differentiated; MD, moderately differentiated; PD, poorly differentiated; LVI, lymphovascular invasion; LNM, lymph node metastasis; MLH1, mutL homolog 1; MSH2, mutS homolog 2; MMR, mismatch-repair protein; EBV, Epstein-Barr virus; MMR-d, MMR deficient; MMR-p, MMR proficient; ARID1A, AT-rich interactive domain 1A.


Table 4 Logistic Regression Analysis of Clinicopathologic and Molecular Features Associated with Loss of ARID1A Expression

FeatureOdds ratio (95% CI)p-value
Location (upper and mid)1.35 (0.90–2.01)0.143
Size (>5 cm)1.23 (0.82–1.86)0.315
Lauren type (intestinal)1.38 (0.92–2.07)0.125
T stage0.242
T31.07 (0.58–2.00)0.820
T40.73 (0.48–1.11)0.140
MLH1 (deficient)8.12 (4.82–13.66)<0.001
EBV (positive)4.13 (2.28–7.48)<0.001
p53 (wild-type pattern)1.17 (0.77–1.77)0.459
HER2 (2 or 3+)NA0.997
MET (2 or 3+)1.54 (0.82–2.89)0.175
EGFR (2 or 3+)1.08 (0.62–1.89)0.781

ARID1A, AT-rich interactive domain 1A; CI, confidence interval; MLH1, mutL homolog 1; EBV, Epstein-Barr virus; HER2, human epidermal growth factor receptor 2; MET, mesenchymal-epithelial transition factor; EGFR, epidermal growth factor receptor; NA, not available.


References

  1. Dallas PB, Pacchione S, Wilsker D, Bowrin V, Kobayashi R, Moran E. The human SWI-SNF complex protein p270 is an ARID family member with non-sequence-specific DNA binding activity. Mol Cell Biol 2000;20:3137-3146.
    Pubmed KoreaMed CrossRef
  2. Wilson BG, Roberts CW. SWI/SNF nucleosome remodellers and cancer. Nat Rev Cancer 2011;11:481-492.
    Pubmed CrossRef
  3. Nagl NG Jr, Patsialou A, Haines DS, Dallas PB, Beck GR Jr, Moran E. The p270 (ARID1A/SMARCF1) subunit of mammalian SWI/SNF-related complexes is essential for normal cell cycle arrest. Cancer Res 2005;65:9236-9244.
    Pubmed CrossRef
  4. Alver BH, Kim KH, Lu P, et al. The SWI/SNF chromatin remodelling complex is required for maintenance of lineage specific enhancers. Nat Commun 2017;8:14648.
    Pubmed KoreaMed CrossRef
  5. Mathur R, Alver BH, San Roman AK, et al. ARID1A loss impairs enhancer-mediated gene regulation and drives colon cancer in mice. Nat Genet 2017;49:296-302.
    Pubmed KoreaMed CrossRef
  6. Wang Y, Chen SY, Karnezis AN, et al. The histone methyltransferase EZH2 is a therapeutic target in small cell carcinoma of the ovary, hypercalcaemic type. J Pathol 2017;242:371-383.
    Pubmed KoreaMed CrossRef
  7. Shen J, Peng Y, Wei L, et al. ARID1A deficiency impairs the DNA damage checkpoint and sensitizes cells to PARP inhibitors. Cancer Discov 2015;5:752-767.
    Pubmed KoreaMed CrossRef
  8. Suryo Rahmanto Y, Jung JG, Wu RC, et al. Inactivating ARID1A tumor suppressor enhances TERT transcription and maintains telomere length in cancer cells. J Biol Chem 2016;291:9690-9699.
    Pubmed KoreaMed CrossRef
  9. Jones S, Wang TL, Shih IeM, et al. Frequent mutations of chromatin remodeling gene ARID1A in ovarian clear cell carcinoma. Science 2010;330:228-231.
    Pubmed KoreaMed CrossRef
  10. Wiegand KC, Shah SP, Al-Agha OM, et al. ARID1A mutations in endometriosis-associated ovarian carcinomas. N Engl J Med 2010;363:1532-1543.
    Pubmed KoreaMed CrossRef
  11. Wang K, Kan J, Yuen ST, et al. Exome sequencing identifies frequent mutation of ARID1A in molecular subtypes of gastric cancer. Nat Genet 2011;43:1219-1223.
    Pubmed CrossRef
  12. Zang ZJ, Cutcutache I, Poon SL, et al. Exome sequencing of gastric adenocarcinoma identifies recurrent somatic mutations in cell adhesion and chromatin remodeling genes. Nat Genet 2012;44:570-574.
    Pubmed CrossRef
  13. Wang DD, Chen YB, Pan K, et al. Decreased expression of the ARID1A gene is associated with poor prognosis in primary gastric cancer. PLoS One 2012;7:e40364.
    Pubmed KoreaMed CrossRef
  14. Maeda D, Mao TL, Fukayama M, et al. Clinicopathological significance of loss of ARID1A immunoreactivity in ovarian clear cell carcinoma. Int J Mol Sci 2010;11:5120-5128.
    Pubmed KoreaMed CrossRef
  15. Zhang X, Sun Q, Shan M, et al. Promoter hypermethylation of ARID1A gene is responsible for its low mRNA expression in many invasive breast cancers. PLoS One 2013;8:e53931.
    Pubmed KoreaMed CrossRef
  16. Abe H, Maeda D, Hino R, et al. ARID1A expression loss in gastric cancer: pathway-dependent roles with and without Epstein-Barr virus infection and microsatellite instability. Virchows Arch 2012;461:367-377.
    Pubmed CrossRef
  17. Wiegand KC, Sy K, Kalloger SE, et al. ARID1A/BAF250a as a prognostic marker for gastric carcinoma: a study of 2 cohorts. Hum Pathol 2014;45:1258-1268.
    Pubmed CrossRef
  18. Kim YB, Ham IH, Hur H, Lee D. Various ARID1A expression patterns and their clinical significance in gastric cancers. Hum Pathol 2016;49:61-70.
    Pubmed CrossRef
  19. Kim KJ, Jung HY, Oh MH, et al. Loss of ARID1A expression in gastric cancer: correlation with mismatch repair deficiency and clinicopathologic features. J Gastric Cancer 2015;15:201-208.
    Pubmed KoreaMed CrossRef
  20. Ibarrola-Villava M, Llorca-Cardeñosa MJ, Tarazona N, et al. Deregulation of ARID1A, CDH1, cMET and PIK3CA and target-related microRNA expression in gastric cancer. Oncotarget 2015;6:26935-26945.
    Pubmed KoreaMed CrossRef
  21. Glückstein MI, Dintner S, Arndt TT, et al. Comprehensive immunohistochemical study of the SWI/SNF complex expression status in gastric cancer reveals an adverse prognosis of SWI/SNF deficiency in genomically stable gastric carcinomas. Cancers (Basel) 2021;13:3894.
    Pubmed KoreaMed CrossRef
  22. Zhu YP, Sheng LL, Wu J, et al. Loss of ARID1A expression is associated with poor prognosis in patients with gastric cancer. Hum Pathol 2018;78:28-35.
    Pubmed CrossRef
  23. Inada R, Sekine S, Taniguchi H, et al. ARID1A expression in gastric adenocarcinoma: clinicopathological significance and correlation with DNA mismatch repair status. World J Gastroenterol 2015;21:2159-2168.
    Pubmed KoreaMed CrossRef
  24. Lee SY, Kim DW, Lee HS, et al. Loss of AT-rich interactive domain 1A expression in gastrointestinal malignancies. Oncology 2015;88:234-240.
    Pubmed CrossRef
  25. Jemal A, Bray F, Center MM, Ferlay J, Ward E, Forman D. Global cancer statistics. CA Cancer J Clin 2011;61:69-90.
    Pubmed CrossRef
  26. Cancer Genome Atlas Research Network. Comprehensive molecular characterization of gastric adenocarcinoma. Nature 2014;513:202-209.
    Pubmed KoreaMed CrossRef
  27. Park CK, Park JS, Kim HS, et al. Receptor tyrosine kinase amplified gastric cancer: clinicopathologic characteristics and proposed screening algorithm. Oncotarget 2016;7:72099-72112.
    Pubmed KoreaMed CrossRef
  28. Bae YS, Kim H, Noh SH, Kim H. Usefulness of immunohistochemistry for microsatellite instability screening in gastric cancer. Gut Liver 2015;9:629-635.
    Pubmed KoreaMed CrossRef
  29. Park JH, Kim EK, Kim YH, et al. Epstein-Barr virus positivity, not mismatch repair-deficiency, is a favorable risk factor for lymph node metastasis in submucosa-invasive early gastric cancer. Gastric Cancer 2016;19:1041-1051.
    Pubmed CrossRef
  30. Kim KH, Kim W, Howard TP, et al. SWI/SNF-mutant cancers depend on catalytic and non-catalytic activity of EZH2. Nat Med 2015;21:1491-1496.
    Pubmed KoreaMed CrossRef
  31. Hofmann M, Stoss O, Shi D, et al. Assessment of a HER2 scoring system for gastric cancer: results from a validation study. Histopathology 2008;52:797-805.
    Pubmed CrossRef
  32. 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.
    Pubmed CrossRef
  33. Shinozaki-Ushiku A, Kunita A, Fukayama M. Update on Epstein-Barr virus and gastric cancer (review). Int J Oncol 2015;46:1421-1434.
    Pubmed CrossRef
  34. Han N, Kim MA, Lee HS, Kim WH. Loss of ARID1A expression is related to gastric cancer progression, Epstein-Barr virus infection, and mismatch repair deficiency. Appl Immunohistochem Mol Morphol 2016;24:320-325.
    Pubmed CrossRef
  35. Hampel H, Frankel WL, Martin E, et al. Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N Engl J Med 2005;352:1851-1860.
    Pubmed CrossRef
  36. Kane MF, Loda M, Gaida GM, et al. Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res 1997;57:808-811.
    Pubmed
  37. Toyota M, Ahuja N, Suzuki H, et al. Aberrant methylation in gastric cancer associated with the CpG island methylator phenotype. Cancer Res 1999;59:5438-5442.
    Pubmed
  38. Kim H, Kim YH, Kim SE, Kim NG, Noh SH, Kim H. Concerted promoter hypermethylation of hMLH1, p16INK4A, and E-cadherin in gastric carcinomas with microsatellite instability. J Pathol 2003;200:23-31.
    Pubmed CrossRef
  39. Zhao J, Liang Q, Cheung KF, et al. Genome-wide identification of Epstein-Barr virus-driven promoter methylation profiles of human genes in gastric cancer cells. Cancer 2013;119:304-312.
    Pubmed CrossRef
  40. Allo G, Bernardini MQ, Wu RC, et al. ARID1A loss correlates with mismatch repair deficiency and intact p53 expression in high-grade endometrial carcinomas. Mod Pathol 2014;27:255-261.
    Pubmed KoreaMed CrossRef
  41. Odnokoz O, Wavelet-Vermuse C, Hophan SL, Bulun S, Wan Y. ARID1 proteins: from transcriptional and post-translational regulation to carcinogenesis and potential therapeutics. Epigenomics 2021;13:809-823.
    Pubmed KoreaMed CrossRef
  42. Kase K, Saito M, Nakajima S, et al. ARID1A deficiency in EBV-positive gastric cancer is partially regulated by EBV-encoded miRNAs, but not by DNA promotor hypermethylation. Carcinogenesis 2021;42:21-30.
    Pubmed CrossRef
Gut and Liver

Vol.18 No.5
September, 2024

pISSN 1976-2283
eISSN 2005-1212

qrcode
qrcode

Supplementary

Share this article on :

  • line

Popular Keywords

Gut and LiverQR code Download
qr-code

Editorial Office