Value of HMGB1 expression for assessing gastric cancer severity: a systematic meta-analysis

Abstract

Objective

To evaluate the clinical value of high mobility group box-1 (HMGB1) expression levels in patients with gastric cancer.

Methods

Articles published from January 2000 to August 2022 were searched using PubMed, Google Scholar and Science Direct, Springer, Wiley and NIH to evaluate the clinicopathological significance of HMGB1 expression in gastric cancer.

Results

A total of 156 publications were selected, of which six studies, comprising 846 patients, met the criteria for inclusion in this study. Forest plots of clinicopathological characteristics indicated that HMGB1 expression was not associated with age (odds ratio (OR) = 1.07, 95% confidence interval (CI): 0.89–1.28), sex (OR = 0.90, 95% CI: 0.81–1.00), TNM (OR = 1.39, 95% CI: 0.82–2.37), N stage (OR = 1.42, 95% CI: 0.97–2.07), or tumor differentiation (OR = 0.96, 95% CI: 0.71–1.29), but was highly correlated with pT stage (OR = 1.56, 95% CI: 1.17–2.07). Funnel plots showed no significant publication bias in the included studies in terms of age, sex, TNM, pT stage, N stage, or tumor differentiation.

Conclusion

HMGB1 expression was significantly correlated with tumor pT stage, but not with age, sex, TNM stage, tumor N stage, tumor differentiation, or lymphatic metastasis in patients with GC.

Introduction

Gastric cancer (GC) is one of the most common malignant cancers, accounting for 8% of newly diagnosed cancers and 10% of cancer deaths each year worldwide.¹ Its early diagnosis and treatment are generally beneficial in terms of patient prognosis; however, most cases of GC are diagnosed at an advanced stage,² highlighting the need for new diagnostic and prognostic biomarkers.

High mobility group box-1 (HMGB1) protein belongs to the HMG family of proteins, with two 80-amino acid DNA-binding domains (A-box and B-box) and an acidic carboxyl tail.³ HMGB1 is an inflammation-related protein that plays an important role in many diseases.⁴ Cytosolic HMGB1 was shown to attenuate tissue injury in inflammatory bowel disease and other complex inflammatory disorders through regulating cell autophagy and apoptosis.⁵ In addition, extracellular HMGB1 was involved in NLRP3 inflammasome activity and regulated interleukin-1β-associated sterile inflammation induced by multi-walled carbon nanotubes.⁶ Moreover, HMGB1 is a proinflammatory cytokine that may contribute to many inflammatory diseases, and which has also been reported to play paradoxical roles in promoting both cell survival and cell death by regulating multiple signaling pathways during tumor development.⁷ Recent studies have also identified a role for HMGB1 in cancers, including GC. Zhao et al.⁸ found that co-expression of the receptor for advanced glycation end products and HMGB1 was correlated with disease progression and a poor prognosis in patients with prostate cancer, while another study⁹ found that HMGB1 inhibited the anti-cancer activity of sunitinib by regulating TP53 autophagic degradation. Moreover, HMGB1-knockdown in bladder cancer was associated with a better response to radiotherapy and decreased autophagy.¹⁰ Song et al. also showed that HMGB1-knockdown suppressed cell proliferation and invasion and sensitized cells to apoptosis induced by oxaliplatin in MGC-803 gastric cancer cells.¹¹ However, despite these studies, the clinical significance of HMGB1 in GC is still unclear and controversial.

In the present study, we conducted a meta-analysis to investigate the correlation between HMGB1 expression and progression of GC. We examined the correlations between HMGB1 expression and clinicopathological characteristics including tumor pT stage, age, sex, TNM stage, lymph node metastasis, and degree of tumor differentiation, to provide deeper insights into the clinical value of assessing HMGB1 expression in patients with GC.

Materials and Methods

Publication search strategy

We searched the PubMed, Google Scholar, Science Direct, Springer, Wiley, and NIH databases on 7 August 2020 for related studies published from January 2000 to August 2020 on. The following keywords were used: (high mobility group box 1 or HMGB1) and (gastric cancer or gastric carcinoma). The search was restricted to English language studies in humans. We also searched the reference lists of related publications as well as review articles to identify potentially relevant research. We first checked the abstracts and titles of the publications, followed by the full texts of the remaining articles, to confirm if they met the selection criteria. This study followed the Preferred Reporting Items for Systematic Review and Meta-Analysis (PRISMA) guidelines. The study did not register with PROSPERO, but this will be completed in future studies.

Inclusion and exclusion criteria

The inclusion criteria for the studies were as follows: (1) patients in the studies diagnosed with GC by pathological examination; (2) adequate data to calculate odds ratios (ORs); (3) correlation between HMGB1 expression level and prognosis, clinicopathological features, or patient characteristics analyzed; (4) displaying outcomes in the form of hazard ratios (HRs) with 95% confidence intervals (CI); and (5) full-text of the study available. The exclusion criteria were: (1) reviews, letters, and comments; (2) animal experiments, cancer cell studies, and other laboratory research; (3) articles in language other than English; (4) duplicated articles or data; and (5) insufficient data or information to obtain HRs.

Data extraction

The included studies were first evaluated by two independent investigators using the Newcastle–Ottawa Scale. Studies with a score ≥7 were considered high quality. The following data were extracted from the full text: name of first author, publication year, study period, country of patients, detection method, total number of cases, number of patients with high HMGB1 and low HMGB1, age, sex, TNM stage, tumor differentiation, pT stage, and nodal status.

Statistical analysis

HRs and corresponding 95% CIs were obtained by statistical analysis using Stata 15.0 software (Stata Corporation, College Station, TX, USA). Correlations between HMGB1 expression and clinicopathological features were analyzed. Significant heterogeneity was defined as a Q statistic P value <0.10 or I² value >50%. A fixed-effects model was used if there was no heterogeneity among the included studies; otherwise, a random-effects model was used. Publication bias was evaluated by funnel plots and Egger’s test. A P value <0.05 was considered statistically significant.

Results

Search results and study characteristics

A total of 156 publications were initially identified. Of these, 35 were excluded due to duplication, and 34 were excluded for not being in English or because the full text was not available. Of the remaining 87 articles, 81 full-text articles were excluded because of a lack of sufficient data, not reporting GC, and missing sensitivity, specificity, accuracy, or correlation values. Six articles including 846 patients were therefore finally included for data extraction.^12–17 All the included studies analyzed the clinicopathological value of HMGB1 in GC. The detailed search strategy is shown in Figure 1.

Figure 1.

Flow chart of selection of publications for inclusion in the study.

The six included references were published from 2010 to 2020 (Table 1). The total number of patients with GC was 846 (range 40–414). Five studies were performed in China and one in Egypt. HMGB1 was detected by immunohistochemical staining in four studies and by enzyme-linked immunosorbent assay in one study. Five studies reported the age distribution, sex, TNM, pT stage, and lymph node metastasis, and one study reported the tumor grade.

Table 1.

Clinical and histopathologic data of the patients.

Study	Publication year	Country	Age	Sex	TNM	pT stage	N stage	Histologic grade	Detection method	Cases (n)	NOS score
			≤60 vs. >60 years	male/female	I + II/III + IV	T1 + T2/T3 + T4	negative/positive	G3 + G4/G1 + G2
Bao et al.12	2010	China	H 15/18	H 26/6	H 12/20	H 3/29	H 11/21	H 16/16	IHC	78	6
			L 29/16	L 29/17	L 23/23	L 9/37	L 23/23	L 24/22
Fang et al.13	2019	China	H 98/217	H 207/111	H 126/167	H 78/236	H 102/207	H 188/123	N/A	414	6
			L 35/59	L 61/35	L 46/50	L 32/64	L 29/67	L 56/37
Zhang et al.14	2014	China	H 11/21	H 24/8	H 8/24	H 4/28	H 20	N/A	IHC	50	7
			L 5/13	L 11/7	L 6/12	L 3/15	L 12/6
Zhang et al.15	2014	China	H 10/27	H 29/8	H 13/24	H 4/33	H 7/30	N/A	IHC	88	6
			L 16/35	L 36/15	L 11/40	L 7/44	L 22/29
He et al.16	2013	China	H 45/36	H 59/22	H 17/64	H 9/72	H 19/62	H 56/25	IHC	166	6
			L 39/46	L 54/31	L 41/44	L 25/69	L 44/41	L 57/28
Ghweil et al.17	2020	Egypt	N/A	N/A	N/A	N/A	N/A	N/A	ELISA kits	50	6

H: HMGB1 high expression; L: HMGB1 low expression; NOS, Newcastle–Ottawa Scale; N/A: not available; IHC, immunohistochemistry; ELISA, enzyme-linked immunosorbent assay.

Correlations between HMGB1 expression and clinicopathological characteristics

We analyzed the correlations between HMGB1 expression and clinicopathological characteristics in GC patients. HMGB1 expression was not significantly associated with age (OR = 1.07, 95% CI: 0.89–1.28, I² = 25.8%; Figure 2), sex (OR = 0.90, 95% CI: 0.81–1.00, I² = 0.00%; Figure 3), TNM (OR = 1.39, 95% CI: 0.82–2.37, P < 0.001, I² = 79.6%; Figure 4), N stage (OR = 1.42, 95% CI: 0.97–2.07, P = 0.012, I² = 68.9%; Figure 5), or histologic grade (OR = 0.96, 95% CI: 0.71–1.29, P = 0.022, I² = 68.8%; Figure 6). However, expression level of HMGB1 was highly correlated with pT stage (OR = 1.56, 95% CI: 1.17–2.07, I² = 0.0%; Figure 7).

Figure 2.

Forest plot evaluating association of HMGB1 expression with age.

RR, relative risk; CI, confidence interval.

Figure 3.

Forest plot evaluating association of HMGB1 expression with sex.

RR, relative risk; CI, confidence interval.

Figure 4.

Forest plot evaluating association of HMGB1 expression with TNM.

RR, relative risk; CI, confidence interval.

Figure 5.

Forest plot evaluating association of HMGB1 expression with N stage.

RR, relative risk; CI, confidence interval.

Figure 6.

Forest plot evaluating association of HMGB1 expression with histologic grade.

RR, relative risk; CI, confidence interval.

Figure 7.

Forest plot evaluating association of HMGB1 expression with pT stage.

RR, relative risk; CI, confidence interval.

Publication bias

We subsequently analyzed publication bias of the included articles for all the clinicopathological characteristics, using funnel plots and Egger’s test. Funnel plots suggested that there was no significant publication bias in the meta-analysis in terms of age, sex, TNM, pT stage, N stage, or tumor differentiation (Figures 8–13).

Figure 8.

Funnel plot of studies used in the analysis of age.

se, standard error; RR, relative risk; OR, odds ratio.

Figure 9.

Funnel plot of studies used in the analysis of sex.

se, standard error; RR, relative risk; OR, odds ratio.

Figure 10.

Funnel plot of studies used in the analysis of TNM.

se, standard error; RR, relative risk; OR, odds ratio.

Figure 11.

Funnel plot of studies used in the analysis of N stage.

se, standard error; RR, relative risk; OR, odds ratio.

Figure 12.

Funnel plot of studies used in the analysis of histologic grade.

se, standard error; RR, relative risk; OR, odds ratio.

Figure 13.

Funnel plot of studies used in the analysis of pT stage.

se, standard error; RR, relative risk; OR, odds ratio.

Discussion

Despite surgical and medical developments, the prognosis of GC, especially metastatic GC, remains poor.¹⁸ A better understanding of the molecular mechanisms underlying GC is therefore necessary. The newly identified inflammation-related factor HMGB1 has recently been shown to play important roles in many diseases, including cancer development¹⁹; however, its role in GC is still unclear. In the present study, we conducted a meta-analysis to investigate the correlations between HMGB1 expression and clinical outcomes and prognosis in GC patients, based on six recent references. We demonstrated that HMGB1 was associated with depth of invasion in GC.

The role of HMGB1 in GC has been reported in several studies, both in vitro and in vivo. The release of HMGB1 mediated by autophagy promoted survival of GC cells in vitro.²⁰ Overexpression of extracellular HMGB1 promoted epithelial–mesenchymal transition and increased cell migration/invasion, while HMGB1 knockdown significantly inhibited tumor growth.²¹ Another in vitro and in vivo study also revealed that HMGB1 knockdown suppressed cell growth and invasion in GC via the nuclear factor-κB pathway.¹⁵ In addition to the studies included in the current meta-analysis, Ghweil et al. demonstrated that higher serum levels of both serum amyloid A and HMGB1 reflected advanced tumor stage and higher tumor grade.¹⁷ He et al.¹⁶ also showed that HMGB1 expression was closely associated with TNM stage, pT stage, nodal status, tumor size, metastasis status, and poor prognosis in patients with GC. Moreover, rs1045411 in HMGB1 was associated with clinical outcomes in GC patients in China.²² However, several studies have found apparently conflicting results. Zhang et al.¹⁴ found no association between HMGB1 expression and clinicopathologic features, including TNM stage, metastatic lymph nodes, and overall survival in patients with GC, while another study¹² found that overexpression of HMGB1 was positively associated with cancer-free survival in GC, suggesting that HMGB1 overexpression might be a marker of good prognosis. A meta-analysis was therefore required to clarify the role of HMGB1 in GC. The current meta-analysis demonstrated that HMGB1 expression was correlated with pT stage, but not with age, sex, TNM stage, N stage, tumor differentiation, or lymphatic metastasis.

In addition to GC, HMGB1 has also been associated with other cancers. Tumor-derived HMGB1 promoted the suppressive function of regulatory T cells in patients with head and neck cancer.²³ HMGB1 was also considered to act as a novel tumor suppressor in pancreatic cancer.²⁴ Lee et al.²⁵ found that serum levels of HMGB1 were increased in a subset of colorectal carcinomas, suggesting that it might act as a supportive diagnostic marker for colorectal carcinomas. A previous meta-analysis²⁶ showed that overexpression of HMGB1 was significantly associated with poorer overall and progression-free survival in patients with various types of cancer, including pancreatic cancer, hepatocellular carcinoma, nasopharyngeal carcinoma, colorectal cancer, head and neck cancer, and cervical carcinoma. Another meta-analysis²⁷ found that HMGB1 expression was associated with TNM stages III–IV, and upregulated HMGB1 was also correlated with disease progression and the risk of disease progression in patients with non-small-cell lung cancer. However, there has been no previous meta-analysis of the role of HMGB1 in GC. The current meta-analysis thus confirmed an important role for HMGB1 in GC patients, and indicated that its expression might be correlated with tumor pT stage.

The present study also had some limitations. Notably, the number of references included in the meta-analysis was limited. Further clinical research is therefore needed to confirm the results.

In conclusion, we conducted a meta-analysis to investigate the role of HMGB1 in GC. HMGB1 expression was significantly correlated with tumor pT stage but not with age, sex, TNM stage, lymph node metastasis, or degree of tumor differentiation. This study might further our understanding of the association between HMGB1 expression and GC progression.