The Role of High Mobility Group Box B-1 in the Prognosis of Colorectal Cancer Based on the Changes in the Intestinal Mucosal Barrier

Abstract

Background: To investigate the expression of high mobility group box B-1 (HMGB-1) in patients with colorectal cancer (CRC) and its association with clinicopathological features and prognosis in colorectal carcinoma by combining bioinformatics and clinical data analysis, and to clarify the role of HMGB-1. To examine whether HMGB-1 expression is related to the damage of the intestinal mucosal barrier, and then explore the potential HMGB-1-dependent mechanisms affecting the progression of CRC. Methods: CRC datasets of GSE12945, GSE17536, and GSE17537 from the public gene chip database were screened and downloaded. Clinical information and CRC tissue samples from patients with stage I-III CRC from the hospital were collected. Serum samples of patients were applied by enzyme-linked immunosorbent assay on HMGB-1, and were divided into high and low HMGB-1 expression, which was examined by 16S rDNA sequencing. Immunohistochemistry was performed to examine the relationship between the expression of HMGB-1 and tight junction protein, occludin, tumor necrosis factor-α, and interferon-γ. Results: Based on the Cutoff value of 10.24 ng/mL, the CRC patients were divided into high and low expression groups. In the HMGB-1H patient group, the TNM staging, overall survival, disease-free survival, recurrence, and metastasis were inferior to the HMGB-1L group. The results of 16S rDNA sequencing demonstrated that the Providencia genus was found to be enriched in the HMGB-1L group. Immunohistochemical results showed that HMGB-1 expression was negatively correlated with the expression of ZO-1 and occludin (R = 0.035, R = 0.003, P < .05), but was positively correlated with the expression of TNF-α and IFN-γ (R = 0.016, R = 0.001, P < .05). Conclusion: The survival of CRC patients with positive HMGB-1 expression was significantly shortened, which may be related to the decrease of Rovitensis content, the decreased expression of ZO-1 and occludin, and the increased levels of TNF-α and IFN-γ, which in turn damage the intestinal mucosal barrier, leading to the development of CRC.

Keywords

high mobility group box B-1 (HMGB-1)intestinal flora intestinal mucosal barrier immune barrier

Introduction

Colorectal cancer (CRC) is one of the most common malignant tumors with an incidence rate ranking third and a mortality rate ranking second in the world.¹ Studies demonstrated that even in different patients at the same disease stage, the prognosis varies greatly, which may be related to changes in the tumor microenvironment.² High mobility group protein B-1 (HMGB-1) is a multifunctional non-histone protein located in the nucleus, involved in DNA repair, transcription, and genome stability.³ HMGB-1 is highly expressed in CRC, being involved in lymph node metastasis and distant metastasis in patients with CRC, and can be used as a predictor of poor prognosis.⁴ However, the mechanism by which HMGB-1 affects the progression of CRC remains unclear.

It has been reported that HMGB-1 not only participates in the alteration of the tumor microenvironment but also regulates tumor immune response.⁵ The intestinal mucosal barrier is a part of the intestinal tumor microenvironment, and its destruction can accelerate the invasion and metastasis of CRC cells.⁶ Previous studies confirmed that the destruction of the intestinal mucosal barrier involves intestinal flora and intestinal immune response. For example, the reduction of probiotic flora and the increase of pathogenic flora can lead to the destruction of the intestinal mucosal barrier, thereby promoting the occurrence and development of CRC.^7,8 Tight junction protein are surface indicators of the intestinal mucosal barrier, and changes in their levels are related to the occurrence and development of CRC.⁹ In addition, the increase of the expression of intestinal inflammatory mediators such as tumor necrosis factor-α (TNF-α) and interferon-γ (IFN-γ) can further affect the immune barrier of the intestinal mucosa, thereby promoting the initiation and progression of CRC.¹⁰ Therefore, by examining the relationship between HMGB-1, intestinal flora, and the expression levels of Zonula occludens −1 (ZO-1), occludin, TNF-α and IFN-γ, this study explores the potential mechanism of HMGB-1 affecting the occurrence and development of CRC.

Materials and Methods

Clinical Information

We screened a total of 294 patients from the dataset GSE12945 (including 62 CRC patients), GSE17536 (including 177 CRC patients), and GSE17537 (including 55 CRC patients). Excluding incomplete data, 232 patients were finally included, 120 males and 112 females with the age range 35-82 years (median age 65.3 years), within which 118 patients were >65 and 114 patients were ≤65 years. According to the TNM staging (AJCC No. Version¹¹), 40 cases were diagnosed with stage I, 85 cases with stage II, 85 cases with stage III, and 22 cases with stage IV CRC. A retrospective analysis of 120 patients with CRC admitted to hospital from January 2016 to December 2016 was performed. The study inclusion criteria were as follows: (1) Concurrent chemoradiotherapy with radical surgery; (2) Pathology consistent with CRC; (3) Written informed consent has been obtained; (4) No previous malignant tumors or other accompanying malignant diseases; (5) Complete medical records, imaging data, and follow-up data. The study exclusion criteria were as follows: (1) CRC combined with other malignant tumors; (2) Surgical contraindications; (3) Postoperative survival ≤3 months; (4) Concurrent hematological diseases, acute injury, chronic inflammation, or intestinal perforation; (5) Incomplete patient information. According to the inclusion and exclusion criteria, a total of 110 CRC patients were included in this study, 71 males and 39 females with a median age of 59.61 years, among which 53 patients were ≤60 years, and 57 patients were >60 years. According to the TNM staging (AJCC 8th edition), 10 cases were diagnosed with stage I, 51 cases with stage II, and 49 cases with stage III CRC. Based on the degree of differentiation there were identified: 35 cases of poor differentiation, 62 cases of moderate differentiation, and 13 cases of high differentiation. There were 49 cases of lymphatic metastasis and 61 cases of no lymphatic metastasis. The median tumor diameter was 5.11 cm, 60 cases with a diameter of ≤5 cm and 50 cases with >5 cm. There were 17 cases positive for vascular tumor thrombus and 93 cases negative for vascular tumor thrombus. There were 23 cases with nerve invasion and 87 cases without nerve invasion.

Immunohistochemical Staining

All specimens were fixed in 10% formalin, dehydrated, transparent, soaked in wax, embedded in paraffin, dehydrated, and cut into 3 μm sections on a microtome. Next, antigen retrieval and inactivation of endogenous peroxidase were performed. Then, the samples were incubated in a blocking buffer (goat serum SAP-9100, Chian, Bei Jing, Beijing Zhong Shan-Golden Bridge Biological Technology Co., Ltd), followed by overnight incubation at 4 °C with primary antibody against HMGB-1 (EPR3507, Chian, Shanghai, Abcam; 1:350), ZO-1 (EPR19945-296, Chian, Shang Hai, Abcam; 1:250), occludin (EPR20992, Chian, Shang Hai, Abcam; 1:200), IFN-γ (EPR21704, Chian, Shang Hai, Abcam; 1:250), and TNF-α (P/T2, Chian, Shang Hai, Abcam; 1:100) added dropwise onto the tissue. The next day after the pre-warming of the slides, a secondary antibody (SE134, Chian, Shanghai, Abcam; 1:200) was added. The developing solution was applied to the sections on the slides. In the end, the sections were counterstained, mounted, and coverslips were sealed. The specimens were examined under a microscope by 2 experienced pathologists in a double-blind method. Three high-power fields of view were randomly selected for image collection. We adopted a semi-quantitative scoring method to investigate the expression of HMGB-1, ZO-1, occludin, IFN-γ, and TNF-α proteins using the following point scale: 0—no positive cells or <1% of positive cells, 1—1-10% of positive cells, 2—10-50% of positive cells, 3—50-75% of positive cells, 4— > 75% of positive cells. Color rendering evaluation was as follows: 0—no coloring, 1—light yellow, 2—brownish yellow, and 3—brown. When 2 points were multiplied together, 0-4 points were negative, and 5-12 points were positive.

Serum Specimen Enzyme-Linked Immunosorbent Assay (ELISA)

ELISA kit (E-EL-H1554c, Chian, Wu Han, Elabscience) was used to detect serum HMGB-1 levels in CRC patients. 100 mL/well of standards and samples were loaded into a 96-well plate, and after incubation with biotinylated antibodies, streptavidin-conjugated horseradish peroxidase (HRP) was added to each well and mixed with the HRP substrate solution. The optical density (OD) values were read by a microplate reader at 450 nm.

16S rDNA Sequencing

To detect the intestinal flora of the 2 groups of patients, CTAB or SDS method was used to extract the genomic DNA from patient stool samples. 16S V3-V4 region was amplified and PCR products were verified by agarose gel electrophoresis and purified using magnetic beads. Further, they were quantified by enzyme labeling, mixed in equal amounts according to the concentration of the PCR product, verified on 2% agarose gel, and recovered. TruSeq® DNA PCR-Free Sample Preparation Kit was used to prepare DNA library that was quantified with Qubit and qPCR. After quality control, whole-genome sequencing on NovaSeq 6000 system was performed. For the bioinformatic analysis, the original data was spliced and filtered. UPARSE algorithm was used to cluster the sequences at 97% similarity, and after chimera filtering and Operational Taxonomic Units (OTU) clustering the sequences were analyzed for species diversity (alpha and beta diversity) and abundance. Linear discriminant analysis effect size (LEfSe) analysis was performed using LEfSe software with a screening value of 4. The functional information of the whole genome of prokaryotic organisms in the KEGG database annotated by UProC and PAUDA methods was mapped to the SILVA database to realize the functional annotation with Tax4Fun.

Statistical Analysis

The data statistics were analyzed by SPSS26.0 software. The count data were described by percentages, and the χ² test of 4 table data was used for comparison. The cutoff value of HMGB-1 was calculated by the ROC curve. The prognostic significance of HMGB1 expression on survival in CRC was carried out by Kaplan-Meier survival analysis. The log-rank test was used to conduct univariate analysis on the above indicators, and Cox regression was used to conduct multivariate analysis on the indicators with meaningful univariate analysis. The Pearson method was used to analyze the correlation. P < .05 was considered statistically significant.

Results

Elevated HMGB-1 Levels are Closely Associated With Poor Prognosis in CRC

To understand the impact of HMGB-1 level changes in CRC patients, ROC curve analysis was conducted to compare the predictive ability of HMGB-1 levels in 232 patients from the database for survival. The results showed that the area under the curve of HMGB-1 was 0.587, with a cutoff value of 10.24 ng/mL (Figure 1A). The CRC patients were divided into high (HMGB-1H, > 10.24 ng/mL) and low expression groups (HMGB-1L, ≤ 10.24 ng/mL) with 187 (80.6%) and 45 cases (19.4%), respectively. Compared to the HMGB-1L group, the TNM stage of the CRC patients in the HMGB-1H group was of a higher grade, and they were more likely to relapse, with more deaths and shorter overall survival (OS) and disease-free survival (DFS) (P = .010, P = .006, P < .001, P = .005, and P = .035). There was no significant difference between the 2 groups in gender and age (P > .05) (Table 1). Moreover, the results of Cox regression analysis showed that the high expression of HMGB-1 and recurrence were independent risk factors for the death of CRC patients (P = .013, P < .001), with the death risk in the HMGB-1H group being 6.204 times that of the HMGB-1L group (Table 2). Since December 2021, 170 CRC patients are still alive, with a 5-year OS rate of 73.3%, among which 67.9% and 95.6% for the high and low HMGB-1 expression group, respectively (P = .002) (Figure 1B).

Figure 1.

(A) The ROC curve of HMGB-1 levels. (B) Analysis of the overall survival in colorectal cancer and HMGB-1 levels. (C) HMGB-1 positive group. (D) HMGB-1 negative group. (E) The relationship between HMGB-1 expression and the overall survival rate of patients with colorectal cancer.

Table 1.

The Association of HMGB-1 With Clinical Features and Prognosis of Colorectal Cancer.

Data		Total(n)	HMGB-1		χ²	P
Data		Total(n)	High (187)	Low (45)	χ²	P
Age (years)	>65	118	96 (51.3%)	22 (48.9%)	0.087	.768
Age (years)	≤65	114	91 (48.7%)	23 (51.1%)	0.087	.768
Gender	Male	120	95 (50.8%)	25 (55.6%)	0.328	.567
Gender	Female	112	92 (49.2%)	20 (44.4%)	0.328	.567
Death	YES	62	60 (32.1%)	2 (4.4%)	12.775	<.001
Death	NO	170	127 (67.9%)	43 (95.6%)	12.775	<.001
Relapse	YES	54	51 (27.3%)	3 (6.7%)	7.509	.006
Relapse	NO	178	136 (72.7%)	42 (93.3%)	7.509	.006
OS	>60 m	62	58 (31.0%)	4 (8.9%)	7.974	.005
OS	≤60 m	170	129 (69.0%)	41 (91.1%)	7.974	.005
DFS	>60 m	47	43 (23.0%)	4 (8.9%)	4.468	.035
DFS	≤60 m	185	144 (77.0%)	41 (91.1%)	4.468	.035
TNM	I-II	125	93 (49.7%)	32 (71.1%)	6.671	.010
TNM	III-IV	107	94 (50.3%)	13 (28.9%)	6.671	.010

Abbreviations: HMGB-1, high mobility group box B-1; OS, overall survival; DFS, disease-free survival.

P < .05 was considered statistically significant.

Table 2.

Multivariate Cox Regression Analysis of Death From Colorectal Cancer.

Data	B	SE	Wald	Sig.	Exp(B)	95.0% CI
Data	B	SE	Wald	Sig.	Exp(B)	Lower	Upper
HMGB-1	1.825	0.733	6.197	0.013	6.204	1.474	26.110
Relapse	1.874	0.294	40.679	0.000	6.514	3.662	11.587
TNM	0.216	0.293	0.545	0.460	1.241	0.699	2.203
OS	-12.707	79.394	0.026	0.873	0.000	0.000	1.152E + 62
DFS	-0.993	1.142	0.756	0.385	0.370	0.040	3.473

Abbreviations: HMGB-1, high mobility group box B-1; OS, overall survival; DFS, disease-free survival.

P < .05 was considered statistically significant.

To further examine the influence of HMGB-1 expression on CRC pathophysiology, 110 tissue samples from CRC patients from our hospital were selected for immunohistochemical detection. The results demonstrated that HMGB-1 was mainly expressed in the nuclei of CRC cells, with a brown or brownish-yellow color. The positive expression rate was 66.4% (73/110) (Figure 1C), and the negative expression rate was 33.6% (37/110) (Figure 1D). Furthermore, the HMGB-1 expression was related to the degree of differentiation, lymph node metastasis, vascular tumor thrombus, and TNM stage in CRC patients (P < .05), but not to gender, age, tumor diameter, and nerve invasion (P > .05) (Table 3). The results of Cox regression analysis revealed that HMGB-1 expression and vascular tumor thrombus were independent risk factors for poor prognosis in CRC patients (P < .05) with the death risk in the HMGB- 1H group being 0.437 times that of the HMGB-1L group (P < .036). The death risk of CRC patients with vascular tumor thrombus was 0.461 times that of patients without vascular tumor thrombus (P < .026) (Table 4). Based on the follow-up until December 2021, 62 of the 110 patients survived, and the 5-year OS was 56.4% with 45.2% and 78.4% for the HMGB-1-positive and negative patients, respectively (P = .003) (Figure 1E). Therefore, our findings demonstrate that the increase in HMGB-1 level is closely related to the poor prognosis of CRC.

Table 3.

Comparison of HMGB-1 Expression With Clinical Data and Pathological Characteristics.

		Total(n)	HMGB-1		χ²	P
		Total(n)	Positive	Negative	χ²	P
Age (years)	>60	57	35 (47.9%)	22 (59.5%)	1.304	.254
Age (years)	≤60	53	38 (52.1%)	15 (40.5%)	1.304	.254
Gender	Male	71	49 (67.1%)	22 (59.5%)	0.630	.427
Gender	Female	39	24 (32.9%)	15 (40.5%)	0.630	.427
Size	>5 cm	50	35 (47.9%)	15 (40.5%)	0.543	.461
Size	≤5 cm	60	38 (52.1%)	22 (59.5%)	0.543	.461
Differentiation degree	Poor	35	25 (34.2%)	10 (27.0%)	8.369	.015
	Moderate	62	44 (60.3%)	18 (48.7%)
	Well	13	4 (5.5%)	9 (24.3%)
Lymph metastasis	NO	61	35 (47.9%)	26 (70.3%)	4.954	.026
Lymph metastasis	YES	49	38 (52.1%)	11 (29.7%)	4.954	.026
Nerve invasion	NO	87	56 (76.7%)	31 (83.8%)	0.742	.389
Nerve invasion	YES	23	17 (23.3%)	6 (16.2%)	0.742	.389
Vascular tumor thrombus	NO	91	58 (79.5%)	35 (94.6%)	4.309	.038
Vascular tumor thrombus	YES	17	15 (20.5%)	2 (5.4%)	4.309	.038
TNM	I	10	3 (4.1%)	7 (18.9%)	8.970	.011
	II	51	32 (43.8%)	19 (51.4%)
	III	49	38 (52.1%)	11 (29.7%)

Abbreviations: HMGB-1, high mobility group box B-1.

P < .05 was considered statistically significant.

Table 4.

Multivariate Cox Regression Analysis of Prognostic Factors for Colorectal Cancer.

Clinical data	B	SE	Wald	ddf	Sig.	Exp(B)	95.0% CI
Clinical data	B	SE	Wald	ddf	Sig.	Exp(B)	Lower	Upper
HMGB-1	−0.827	0.394	4.404	1	0.036	0.437	0.202	0.947
Vascular tumor thrombus	−0.775	0.348	4.953	1	0.026	0.461	0.233	0.912
Lymph metastasis	0.330	1.192	0.077	1	0.781	1.392	0.135	14.381
TNM	0.638	1.111	0.330	1	0.566	1.893	0.214	16.700
Differentiation degree DegreeDegree	0.440	0.298	2.183	1	0.140	1.553	0.866	2.783

Abbreviations: HMGB-1, high mobility group box B-1.

P < .05 was considered statistically significant.

Increase in HMGB-1 Expression Affects the Redistribution and Function of Intestinal Flora in Patients With CRC

To explore whether the change in HMGB-1 level is associated with the alteration of intestinal flora, we performed ELISA on serum samples from randomly selected 88 of 110 CRC patients. We found that the average value of HMGB-1 expression was 10.33 ng/mL, which was similar to the database cutoff value (10.24 ng/mL). The 88 samples were divided into high expression group (HMGB-1H) and low expression group (HMGB-1L) with 13.09 ± 2.456 ng/mL and 7.381 ± 1.634 ng/mL, respectively (P < .001) (Figure 2A).

Figure 2.

(A) Comparison of HMGB-1 levels between 2 groups. (B) Venn diagram of intestinal flora among 2 groups. (C) Histogram of relative abundance of genus level in 2 groups. (D) Alpha diversity analysis (Chao 1, Observed species, Shannon, and Simpson) of differential bacteria between groups. (E) Unweighted UniFrac distance analysis in 2 groups. (F) Weighted UniFrac distance analysis in 2 groups. (G) Linear discriminant analysis effect size (LEfSe) for differentially abundant taxa from the 2 groups. (H) Taxonomic cladogram from the LEfSe showing differences in fecal taxa. (I) Tax4fun prediction of the differential microbiome.

Based on the ELISA results, we selected 10 samples from the HMGB-1H group and HMGB-1L group for 16S rDNA analysis. Statistical analysis of OTUs at a 97% similarity level revealed that apart from 468 common species there were 259 and 166 unique species of bacterial flora in the HMGB-1H and HMGB-1L groups, respectively, indicating that the 2 groups have their own unique flora abundance (Figure 2B). The species diversity and richness suggested by alpha and beta diversity analysis were higher in the HMGB-1H group than those of the HMGB-1L group (P < .05) (Figure 2D-F). The species composition was different between the 2 groups. The HMGB-1H group was enriched with Staphylococcus, Faecalibacterium, and Enterococcus, while the HMGB-1L group was enriched with Proteus, Ekmansella, Escherichia coli, and Shigella bacilli (Figure 2C). The results of LEfSe analysis demonstrated that there was no significant difference between the HMGB-1H group and the HMGB-1L group; only Providencia content was found to be significantly increased in the HMGB-1L group. Elevation of HMGB-1 level had little effect on gut microbiota (Figure 2G-H). To explore how the HMGB-1 level affects the gut microbiota, we analyzed the functional predictions of the HMGB-1H and HMGB-1L group differential flora. The results showed that the functional enrichment of the differential flora between the groups was mostly concentrated in DNA damage repair, metabolic changes such as glycolysis and amino acid metabolism, and cell motility (P < .05) (Figure 2I). The above results show that the increase of HMGB-1 levels affects the redistribution and function of intestinal flora.

Increased HMGB-1 Levels Affect the Expression of Proteins Related to the Intestinal Barrier in CRC Patients

To examine the impact of HMGB-1 level changes on the intestinal barrier, we selected ZO-1, occludin, TNF-α, and IFN-γ for immunohistochemical detection. Our results showed that ZO-1, occludin, TNF-α, and IFN-γ were diffusely distributed in the membrane and cytoplasm of CRC cells (brown or brownish yellow), and a small amount was expressed in the nucleus (Figure 3). Among HMGB-1-positive patients, 14 cases (19.2%) were positive for ZO-1 protein, and 59 cases (80.8%) were negative (P = .034). The positive and negative expression rates of occludin in HMGB-1-positive patients were 27.4% and 72.6%, respectively (P = .003). There were 56 cases of TNF-α positive patients, accounting for 76.7% and 17 cases of TNF-α negative patients, accounting for 23.3% (P = .015). In HMGB-1 positive patients, 48 cases were IFN-γ positive, accounting for 65.8%, and 25 patients were IFN-γ negative, accounting for 34.2% (P = .001). Our findings demonstrated that the expressions of ZO-1 and occludin in HMGB-1 positive patients were lower than those in HMGB-1 negative patients (R = 0.035, R = 0.003, P < .05), while the expressions of TNF-α and IFN-γ were higher than those in negative patients (R = 0.016, R = 0.001, P < .05) (Table 5), indicating that the increase in HMGB-1 levels may affect the intestinal mucosal barrier and immune barrier in CRC patients.

Figure 3.

The expression of ZO-1, occludin, TNF-α, and IFN-γ in colorectal cancer tissues (HE, 400 ×). (A) ZO-1 positive group. (B) ZO-1 negative group. (C) Occludin positive group. (D) Occludin negative group. (E) TNF-α positive group. (F) TNF-α negative group. (G) IFN-γ positive group. (H) IFN-γ negative group.

Table 5.

The Association of HMGB-1 With ZO-1, Occludin, TNF-α, and IFN-γ in Colorectal Cancer.

Protein factor		HMGB-1		χ2	R	P
Protein factor		Positive (n = 73)	Negative (n = 32)	χ2	R	P
ZO-1	Negative	59 (80.8%)	23 (62.2%)	4.506	0.035	.034
ZO-1	Positive	14 (19.2%)	14 (37.8%)	4.506	0.035	.034
Occludin	Negative	53 (72.6%)	16 (43.2%)	9.053	0.003	.003
Occludin	Positive	20 (27.4%)	21 (56.8%)	9.053	0.003	.003
TNF-α	Negative	17 (23.3%)	17 (45.9%)	5.903	0.016	.015
TNF-α	Positive	56 (76.7%)	20 (54.1%)	5.903	0.016	.015
IFN-γ	Negative	25 (34.2%)	25 (67.6%)	10.996	0.001	.001
IFN-γ	positive	48 (65.8%)	12 (32.4%)	10.996	0.001	.001

Abbreviations: HMGB-1, high mobility group box B-1; ZO-1, Zonula occludens-1; TNF-α, tumor necrosis factor-α; IFN-γ, interferon-γ.

P < .05 was considered statistically significant.

Discussion

HMGB-1 plays a vital role in the occurrence and development of CRC. Previous studies confirmed that the high expression of HMGB-1 is a poor prognostic factor in CRC. In this study, we reported that the investigated patients with high expression of HMGB-1 had a higher TNM stage. Moreover, their OS and DFS shortened, and they were more likely to relapse and metastasize. Based on the HMGB-1 cutoff value, they were divided into 2 groups, HMGB-1H and HMGB-1L. 16SrDNA sequencing demonstrated that the intestinal flora between these 2 groups had different diversity and richness, among which Providencia was substantially increased in the HMGB-1L group. In addition, this study also found that the increase in HMGB-1 expression may be related to the damage of the intestinal mucosal physical barrier and the change of the local immune barrier.

The analysis of the patient clinical phenotype revealed that the expression of HMGB-1 was related to the death, recurrence, OS, DFS, and TNM stage of CRC patients, and HMGB-1 expression is an independent risk factor for the death and recurrence of CRC patients. The 5-year OS of patients with low HMGB-1 expression was higher than that of CRC patients with high HMGB-1 expression. We further verified that the HMGB-1 expression was related to the degree of differentiation, lymph node metastasis, vascular tumor embolism, and TNM stage of CRC patients. These results correspond to those in the database. The results of this study are also in line with many previous studies, and they confirmed that HMGB-1 can be used as a predictive factor for the CRC prognosis.^12–15 Our research confirms that HMGB-1 is closely related to CRC stage, lymph node metastasis and differentiation, and may be an important marker of its occurrence and development, which will help to improve the accuracy of CRC diagnosis.

Furthermore, we performed 16S rDNA sequencing on the feces of CRC patients according to the HMGB-1 expression. The Venn diagram (Venn) showed that the intestinal flora is a complex micro-ecological environment, and there were many similar species between low and high HMGB-1 expression groups. However, the abundance of specific bacterial flora in the HMGB-1H group was still higher than that in the HMGB-1L group. It has been shown previously that the reduction or depletion of intestinal flora could significantly reduce CRC liver metastasis and primary liver cancer, suggesting that the reduction of the abundance of intestinal flora may be beneficial to the development of CRC.¹⁶ In addition, we also know from the alpha and beta diversity that the species diversity and richness of the HMGB-1H group were greater than those of the HMGB-1L group. It suggests that the reduction of flora species was beneficial to the development of CRC. We used the LEfSe histogram and evolutionary map to perform a high-dimensional classification analysis of the intestinal flora in the 2 groups. At the genus level, only Providencia was enriched in the HMGB-1L group, which suggested that the increase of Providencia content may inhibit the development of CRC. It has been previously reported that the increase of Providencia could lead to massive secretion of lethal swelling toxin, causing cell swelling, and blocking eukaryotic cell proliferation in the G2/M phase, resulting in cell death.¹⁷ In vivo studies have also found that Providencia could translocate from the intestinal lumen to extraintestinal organs, confirming its potential role in disseminating infection.¹⁸ Although previous studies on Providencia considered it pathogenic, the research did not determine a causal link between Providencia species and cancer. Thus, it needs further investigation whether Providencia has a tumor suppressor effect, its content is reduced due to the increase in HMGB-1 level, or produces specific metabolites limiting the CRC progression.

An intact mucosal barrier and mucosal immune system are critical for maintaining intestinal homeostasis.¹⁹ ZO-1 and occludin play an important role in the intestinal inflammatory response and immune response by maintaining the physical barrier function of the intestinal mucosa.²⁰ This study found that the ZO-1 and occludin proteins were reduced in HMGB-1-positive patients, which is consistent with previous research results.^21–24 The possible mechanism may be that HMGB1 stimulates Toll-like receptor-4 (TLR4) leading to a decreased expression of ZO-1 and occludin, thereby damaging endothelial cells and changing their permeability.²⁵ Alternatively, HMGB-1 may stimulate the production of trimethylamine nitrogen oxides, which decreases the expression of ZO-2 and occludin and destroys the tight junction between cells, increasing the physical barrier damage of intestinal mucosa and finally leading to CRC.²⁶ It has been demonstrated that HMGB-1 destroys intercellular connections in a paracrine or autocrine manner, and when the receptor for advanced glycation end products (RAGE) is silenced, the expressions of ZO-1 and occludin increased significantly, thereby weakening the disruption of the intestinal mucosa.²⁷

Imbalances in the structure of the gut barrier may unleash an uncontrolled immune response in the gut microenvironment. This study showed that HMGB-1 could increase TNF-α and INF-γ expression in CRC. Similar to the previous studies, it was found that TNF-α level in the serum of CRC patients was substantially increased, and thus anti-TNF-α agents could be applied in CRC treatment,^28–31 potentially due to the activation of EGFR/ERK1/2 and Wnt/β-catenin signaling.³² HMGB-1 can increase IFN-γ production by promoting the activation of CD8[ + ] T cells induced by dendritic cells, and IFN-γ is positively correlated with HMGB1 at the mRNA level.³³ Therefore, HMGB-1 can participate in inflammatory and immune responses, affecting CRC microenvironment and promoting its initiation and progression.

However, this study also has some limitations. It did not explore the relationship between HMGB-1 and CRC molecular typing, and its influence on the local recurrence rate and distant metastasis rate. Moreover, the association of HMGB-1 with inflammatory markers (TNF-α, and IFN-γ) and the specific mechanism of ZO-1 and occludin should be studied more in-depth.

Conclusion

In summary, we demonstrated that the increase in HMGB-1 levels is closely related to the poor prognosis of CRC and may lead to the decrease of Providencia content, ZO-1, and occludin expression, and the increase of TNF-α, and IFN-γ expression. This can cause damage to the intestinal mucosal barrier ultimately leading to the occurrence and development of CRC.

Footnotes

Abbreviation

Data Availability

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by Training Project of Excellent Clinical Medical Talents in Hebei Provincial Government Funding (No.4).

Statement of Human Rights

This study was approved by the Ethics Committee of the First Affiliated Hospital of Hebei Northern University (K2021125), and all patients signed an informed consent form.

ORCID iD

Weiwei Zhao

References

Sieggel

Miller

Jemal

, et al. Cancer statistics, 2021. CA Cancer J Clin. 2021;71(1):7‐33. doi:10.3322/caac.21654. Epub 2021 Jan 12

Chandra

Karalis

Liu

, et al. The colorectal cancer tumor microenvironment and its impact on liver and lung metastasis. Cancers (Basel). 2021;13(24). doi:10.3390/cancers 13246206

Cheng

Mohammed

Elsa Haniffah

, et al. HMGB-1 an overview of its versatile roles in the pathogenesis of colorectal cancer. Cell Oncol (Dordr). 2020;43(2):177‐193. doi:10.1007 /s13402 −019-00477-5

Zhang

Xia

, et al. Increased HMGB1 expression correlates with higher expression of c-IAP2 and pERK in colorectal cancer. Medicine 2019;98(3):e14069. doi:10.1097/MD.00 00000000014069

Ferrara

Chill

GET

Ruggiere

, et al. Oxidation of HMGB1 is a dynamically regulated process in physiological and pathological conditions. Front Immunol. 2020;11:1122. doi:10.3389/fimm u.2020.01122. eCollection 2020

Wang

Ding

, et al. Cldn-7 deficiency promotes experimental colitis and associated carcinogenesis by regulating intestinal epithelial integrity. Oncoimmunology. 2021;10(1):1923910. doi:10.1080/2162402X.2021.1923910

Dai

Coker

Nakatsu

, et al. Multi-cohort analysis of colorectal cancer metagenome identified altered bacteria across populations and universal bacterial markers. Microbiome. 2018;6(1):70. doi:10.1186/s40168-018-0451-2

Yamamoto

Kinugasa

Hirai

, et al. Heterogeneous distribution of Fusobacterium nucleatum in the progression of colorectal cancer. J Gastroenterol Hepatol. 2021;36(7):1869‐1876. doi:10.1111/jgh.15361. Epub 2020 Dec 14.

Wang

Zhang

Wang

, et al. Lactobacillus coryniformis MXJ32 administration ameliorates azoxymethane/dextran sulfate sodium- induced colitis-associated colorectal cancer via reshaping intestinal microenvironment and alleviating in flammatory response. Eur J Nutr. 2021. doi:10.1007/s00394-021-02627-8. Epub 2021 Jun 29.

10.

Kobelt

Zhang

Clayton-lucey

, et al. Pro-inflammatory TNF-α and IFN-γ promote tumor growth and metastasis viainduction of MA-CC1. Front Immunol. 2020;11:980. doi:10.3389/fimmu.2020.00980. eCollection 2020.

11.

Yao

Liu

. Improvement of prognostic and predictive network of colorectal cancer based upon the 8th edition of AJCC colorectal cancer staging system. Zhonghua Wei Chang Wai Ke Za Zhi. 2017-01-25;20(1):24‐27. doi:10.3760/cma.j.issn.1671-0274.2017.01.006

12.

Cao

Xie

, et al. KRAS Mutation-Responsive miR-139-5p inhibits colorectal cancer progression and is repressed by Wnt signaling. Theranostics. 2020;10(16):7335‐7350. doi:10.7150/thno.45971. eCollection 2020.

13.

Qian

Xiao

Gai

, et al. HMGB1-RAGE Signaling facilitates Rasdependent Yap1 expression to drive colorectal cancer stemnessand development. Mol Carcinog. 2019;58(4):500‐510. doi: 10.1002/mc.22944. Epub 2018 Dec 18.

14.

Chandrasekaran

Sathyanarayanan

Karunagaran

, et al. Downregulation of HMGB-1 by miR-34a is sufficient to suppress proliferation, migration and invasion of human cervical and colorectal cancer cells. Tumour Biol. 2016;37(10):13155‐13166. doi:10.1007/s13277-016-5261-. Epub 2016 Jul 25.

15.

Chen

Liu

, et al. MiR-216b functions as a tumor suppressor by targeting HMGB1-mediated JAK2/STAT3 signaling way in colorectal cancer. Am J Cancer Res. 2017;7(10):2051‐2069. eCollection 2017.

16.

Bullman

Pedamallu

Sicinska

, et al. Analysis of Fusobacterium persistence and antibiotic response in colorectal cancer. Science. 2017;358(6369):1443‐1448. doi:10.1126/science.aal524 0. Epub 2017 Nov 23.

17.

Shima

Hinendya

Askura

, et al. Molecular characterizations of cytolethal distending toxin produced by Providencia alcalifaciens strains isolated from patients with diarrhea. Infect Immun. 2012;80(4):1323‐1332. doi:10.1128/IAI.05831-11. Epub 2012 Jan 17.

18.

Yuan

Wei

Zhang

, et al. Comparative genomic analysis reveals genetic mechanisms of the variety of pathogenicity, antibiotic resistance, and environmental adaptation of Providencia genus. Front Microbiol. 2020;11:572642. doi:10.3389/fmicb.2020.572642. eCollection 2020.

19.

Liang

Chen

, et al. Dendrobium officinale polysaccharides alleviate colon tumorigenesis via restoring intestinal barrier functionand enhancing anti-tumor immune response. Pharmacol Res. 2019;148:104417. doi:10.1016/j.phrs.2019.104417. Epub 2019 Aug 29.

20.

Chuanbin

. Levels of biochemical indicators such as ZO-1 in patients with colorectal cancer and their clinical significance. J Pract Cancer. 2018;33(7):1076‐1079. doi:10. 3969/j.issn.1001-5930.2018.07.009

21.

Zhao

Chen

Zheng

, et al. Citric acid mitigates soybean meal induced inflammatory response and tight junction disruption by altering TLR signal transduction in the intestine of turbot, Scophthalmus maximus L. Fish Shellfish Immunol. 2019;92:181‐187. doi:10.1016/j.fsi.2019.06.004. Epub 2019 Jun 5.

22.

Glavan

Gaulke

Santos Rocha

, et al. Gut immune dysfunction through impaired innate pattern recognition receptor expression and gut microbiota dysbiosis in chronic SIV infection. Mucosal Immunol. 2016;9(3):677‐688. doi:10.1038/mi.2015.92. Epub 2015 Sep 16.

23.

Zhang

Yang

Liang

, et al. Jiawei Baitouweng Decoction affects intestinal mucosal tight junction proteins in rats with ulcerative colitis through p38 MAPK-MLCK signaling pathway. Zhongguo Zhong Yao Za Zhi. 2021;46(21):5719‐5726. doi:10.19540/j.cnki.cjcmm.20210604.401

24.

Kaihara

Kawamata

Imura

, et al. Redifferentiation and ZO-1 reexpression in liver-metastasized colorectal cancer: Possible association with epidermal growth factor receptor-induced tyrosine phosphorylation of ZO-1. Cancer Sci. 2003;94(2):166‐172. doi:10.1111/j.1349-7006.2003.tb01414.x

25.

Gao

Yang

, et al. Sesamin attenuates intestinal injury in sepsis via the HMGB1/TLR4/IL-33 signalling pathway. Pharm Biol. 2020;58(1):898‐904. doi:10.1080/1388020 9.2020.1787469

26.

Singh

Zhang

Boini

, et al. High mobility group box-1 mediates TMAO-induced endothelial dysfunction. Mol Sci. 2019;20(14). doi:10.3390/ijms20143570

27.

Chen

Pitzer

, et al. Instigation of endothelial Nlrp3 inflammasome by adipokine visfatin promotes inter-endothelial junction disruption: Role of HMGB1. Cell Mol Med. 2015;19(12):2715‐2727. doi:10.1111/jcmm.12657. Epub 2015 Aug 20.

28.

Fan

Tseng

, et al. Correlation between the Glasgow prognostic score and the serum cytokine profile intaiwanese patients with colorectal cancer. Int J Biol Markers. 2021;36(2):40‐49. doi:10.1177/17246008211022769. Epub 2021 Jun 6.

29.

Gutierrez-Hurtado

Puebla-Pérez

DelgadoS

, et al. Association between TNF-α-308G > A and-238G > A gene polymorphisms and TNF-α serum levels in Mexican colorectal cancer patients. Genet Mol Res. 2016;15(2). doi:10.4238/gmr.15028199

30.

Wang

Song

, et al. Deficiency of interferon-gamma or its receptor promotes colorectal cancer development. InterferonCytokine Res. 2015;35(4):273‐280. doi:10.1089/jir.2014.0132

31.

Jorgovanovic

Song

Wang

, et al. Roles of IFN-γ in tumor progression and regression: A review. Biomark Res. 2020;8:49. doi:10.1186/s40364-020-00228-x

32.

Kula

Dawidowicz

Mielcarska

, et al. Periostin in angiogenesis and inflammation in CRC-A preliminary observational study. Medicina (Kaunas). 2022;58(1). doi:10.3390/medicina58010 096

33.

Gao

Wang

, et al. A cycle involving HMGB1, IFN-γ and dendritic cells plays a putative role in anti-tumor immunity. Cell Immunol. 2019;343:103850. doi:10.1016/j.cellimm.2018.08.011. Epub 2018 Aug 20.