Objective To investigate the influence of high mobility group protein AT-Hook-2 (HMGA2) expression in pancreatic cancer tissues on prognosis of patients and mechanism of tumor progression promotion.

Methods The retrospective cohort and experiment study was conducted. The clinicopathological data of 161 patients with pancreatic cancer who were admitted to Zhongshan Hospital of Fudan University from January 2018 to February 2025 were collected. There were 102 males and 59 females, aged 64 (range, 56-70)years. Among 161 patients, data were collected and divided into 3 cohorts based on different time periods. Cohort 1 included 29 patients admitted from September 2018 to October 2019, of whom tumor and paracancerous tissue samples were collected to prepare tissue microarrays (TMAs). Cohort 2 included 116 patients admitted from November 2020 to December 2021, and their tumor samples were collected for TMAs preparation. Cohort 3 included 16 patients admitted from January to February 2025, and their tumor samples were collected to make conventional paraffin sections. The immunohistochemical staining method was used to evaluate the expression of HMGA2 in pancreatic cancer and analyze its relationship with the prognosis of patients. The biological function of HMGA2 in pancreatic cancer was analyzed by bioinformatics and verified by quantitative real-time polymerase chain reaction (qRT-PCR), Western blot, cell counting kit-8 (CCK8) assay, flow cytometry. The immunohistochemical staining and flow cytometry were used to identify the correlation between HMGA2 expression and microenvironment of pancreatic cancer. The relationship between HMGA2 and tumor-associated neutrophils (TANs) was analyzed using single cell sequencing data of pancreatic cancer. Observation indicators: (1) expression of HMGA2 in pancreatic cancer and prognosis of patients; (2) functional analysis of HMGA2 in pancreatic cancer; (3) correlation between HMGA2 expression and immune microenvironment in pancreatic cancer; (4) analysis of TANs participation in HMGA2 mediated immune escape. Comparison of measurement data with normal distribution between groups was conducted using the independent sample t test, and one-way analysis of variance was used for multiple group comparisons. Comparison of measurement data with skewed distribution between groups was conducted using the Mann-Whitney U test. The Pearson correlation analysis was used for inter group correlation analysis. The Kaplan-Meier method was used to calculate survival rate and plot survival curves, and Log-rank test was used for survival analysis.

Results (1) Expression of HMGA2 in pancreatic cancer and prognosis of patients. For patients of cohort 1 and 2, immunohistochemical staining analysis showed that HMGA2 mainly localizing in tumor cells, with expression levels of 1 (range, 0-1) in cancer tissues and 2 (range, 1-2) in adjacent tissues, respectively, showing a significant difference between them (Z=-3.77, P<0.05). In cohort 1, the median survival time of pancreatic cancer patients with high and low HMGA2 expression was 17 months and 36 months, respectively, showing a significant difference between them (hazard ratio=3.11, 95% confidence interval as 1.05-9.23, P<0.05). In cohort 2, the median survival time of pancreatic cancer patients with high and low HMGA2 expression was 22 months and not reached, respectively, showing a significant difference between them (hazard ratio=3.47, 95% confidence interval as 1.997-6.016, P<0.05). (2) Functional analysis of HMGA2 in pancreatic cancer. The results of gene set enrichment analysis of patients with high and low HMGA2 expression in pancreatic cancer patients in The Cancer Genome Atlas database showed that the differential genes in GO entry and KEGG pathway were enriched in cell proliferation related processes such as chromatin separation, DNA replication, cell cycle, etc. Patients with high HMGA2 expression had a high KRAS mutation frequency than patients with low HMGA2 expression. In classic-type and basal-type pancreatic cancer, the expression levels of HMGA2 were 0.95 (range, 0.44-1.52) and 1.91 (range, 0.96-2.87), respectively, showing a significant difference between the two groups (Z=-2.41, P<0.05). HMGA2 was positively correlated with the expression of KRT5 in basal-type pancreatic cancer (r=0.38, P<0.05), while it was negatively correlated with the expression of GATA6 in classic-type pancreatic cancer (r=-0.18, P<0.05). Functional experiments were performed in the pancreatic cancer cell line PANC-1, in which the small interfering RNA (siRNA) was used to inhibit HMGA2 expression. Results of qRT-PCR analysis showed that the HMGA2 expression levels in the control group, siRNA#1 group, siRNA#2 group, and siRNA#3 group were 0.99±0.03, 0.43±0.05, 0.76±0.06, and 0.31±0.04, respectively, showing a significant difference among the four groups (F=181.60, P<0.05). Western Blotting revealed that the grayscale values of HMGA2 bands in the siRNA#1 group and siRNA#3 group were lower than those in the control group. Results of the CCK-8 assay showed that at 96 hours after cell seeding, the absorbance values in the control group, siRNA#1 group, and siRNA#3 group were 2.23±0.12, 1.83±0.07, and 1.73±0.06, respectively, showing a significant difference among the three groups (F=46.09, P<0.05). Flow cytometry indicated that taking the control group as a reference, the early and late apoptotic cells rates were 19.99%±1.99% in the siRNA#1 group and 22.94%±1.51% in the siRNA#3 group. (3) Correlation between HMGA2 expression and immune microenvironment in pancreatic cancer. Results of flow cytometry showed that in cohort 1, the infiltration levels of CD8⁺ T cells and tumor-associated neutrophils (TANs) were 76±58 and 118±92 in patients with high HMGA2 expression, versus 119±42 and 58±49 in patients with low HMGA2 expression, respectively,showing significant differences between the two groups (t=-2.30, 2.13, P<0.05). In cohort 3, the expression levels of PD-1 and TIM-3 on the surface of CD8⁺ T cells were 1 398±1 214 and 3 930±2 267 in patients with high HMGA2 expression, versus 263±254 and 2 030±979 in patients with low HMGA2 expression, respectively, showing significant differences between the two groups (t=2.59, 2.18, P<0.05). (4) Analysis of TANs participation in HMGA2 mediated immune escape. Sequencing results of pancreatic cancer cells from The Cancer Genome Atlas, International Cancer Genome Consortium, Gene Expression Omnibus, and European Bioinformatics Institute database showed that HMGA2 expression in pancreatic cancer was positively correlated with TANs infiltration and negatively correlated with CD8⁺ T cell infiltration. Results from single-cell sequencing samples of GSE155698 demonstrated that compared with patients with low HMGA2 expression, those with high HMGA2 expression had a higher TANs infiltration level, the infiltration levels of TANs subsets (TANs_cluster 0, TANs_cluster 1, TANs_cluster 2, and TANs_cluster 3) also increased. Among these subsets, TANs_cluster 2 highly expressed vascular endothelial growth factor.

Conclusions HMGA2 is highly expressed in pancreatic cancer, and the prognosis of patients with high expression is worse. Inhibition of HMGA2 expression can suppress cell proliferation and promote apoptosis in pancreatic cancer. Compared with low HMGA2 expression, pancreatic cancer with high HMGA2 expression exhibits lower infiltration level of CD8⁺ T cells, as well as higher infiltration levels of TANs and their subsets.