High cyclic GMP-AMP synthase and stimulator of interferon genes in cholangiocarcinoma suggest their potential as targets for treatment

Human CCA tissues

Ninety-five CCA tissues were collected from patients diagnosed with CCA and treated at Srinagarind hospital, Khon Kaen University. The paraffin-embedded tissues were obtained from the specimen bank of Cholangiocarcinoma Research Institute, Faculty of Medicine, Khon Kaen University. Written informed consent was obtained from all patients. The selection and exclusion criteria were as previously reported (Saranaruk et al., 2023). The study procedure was approved by the Ethics Committee for Human Research of Khon Kaen University based on the Declaration of Helsinki (HE641574 and HE651124).

CCA cell lines

Three CCA cell lines, KKU-055, KKU-213A (Sripa et al., 2020), KKK-D068 (Vaeteewoottacharn et al., 2019) cells were obtained from the Japanese Collection of Research Bioresources Cell Bank (JCRB) (Osaka, Japan). Cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) and 1% antibiotic-antimycotic in a humidified incubator at 37 °C with 5% CO₂. All cell culture-related reagents were purchased from Gibco (Life Technologies, Grand Island, NY, USA).

Antibodies and reagents

The antibodies were obtained from the following sources: rabbit anti-cGAS (#26416-1-AP), anti-STING (#19851-1-AP), and mouse anti-NF-κB p65 (#66535-1-Ig) were from Proteintech (IL, USA), mouse anti-NF-κB p65 (F6) from Santa Cruz Biotech (CA, USA), mouse anti-GAPDH from EMD Millipore (Darmstadt, Germany), anti-mouse IgG-conjugated with horseradish peroxidase (HRP) (62-6520) from Invitrogen (CA, USA), anti-lamin A/C (4777) and anti-rabbit IgG HRP (7074S) from Cell signaling Technology (MA, USA), anti-mouse (K4001) and anti-rabbit (K4003) EnVision HRP-conjugated antibodies were from DAKO (Glostrup, Denmark).

3,3′-Diaminobenzidine Tetrahydrochloride hydrate (DAB) (D5637) were bought from Sigma-Aldrich; Immunobilion^® Forte Western HRP substrate (WBLUF0500) from EMD Millipore; Quick Start™ Bradford 1x Dye Reagent (5000205) from Bio-Rad; Hoechst33342 (H3570) was from Invitrogen.

mRNA expression analysis

To investigate the expression of the cGAS-STING-related molecules, including CGAS, STING, TANK-binding kinase 1 (TBK1), interferon regulatory factor 3 (IRF3), NF-κB p65 (RELA), NF-κB p50 (NFKB1), interleukin-6 (IL6), tumor necrosis factor (TNF), interferon alpha 1 (IFNA1), and IFN beta 1 (IFNB1), in gastrointestinal cancers compared with comparable normal tissues, we assessed using the Gene Expression Profiling Interactive Analysis (GEPIA, http://gepia.cancer-pku.cn/index.html), an online server for cancer and normal gene expression profiling analysis based on The Cancer Genome Atlas (TCGA) database (Tang et al., 2017).

The expression levels of CGAS, STING, and RELA in CCA tissues were retrieved from Gene Expression Omnibus (GEO: https://www.ncbi.nlm.nih.gov/gds/), including GSE26566 (Andersen et al., 2012), and GSE76297 (Chaisaingmongkol et al., 2017). The GSE26566 dataset included 104 CCA and 71 normal samples. The cohort 91 CCA and 92 normal adjacent tissues were obtained from GSE76297. Log₂ mRNA expressions in cancer and normal tissues were compared.

IHC staining

According to the standard guidance, the expression levels of targeted proteins in the formalin-fixed paraffin-embedded CCA tissues were determined using IHC staining using antibodies against cGAS (1:100), STING (1:200), and NF-κB p65 (1:100). The signal amplification was performed using the EnVision system-HRP system. The immunoreactivity was visualized by DAB.

The cGAS, STING and NF-κB p65 expression levels were assessed using H-scores (Fitzgibbons et al., 2014; Saranaruk et al., 2023). The patients were categorized into low- and high-expressed groups using median H-score value. Two independent evaluators and a pathologist independently evaluated the H-scores while blinded to the clinical information.

Hoechst 33342 staining

KKU-055, KKU-213A, and KKK-D068 cells were seeded at a density of 2 × 10⁴ cells/well into a 24-well plate and incubated overnight. After washing with PBS, cells were fixed with 4% paraformaldehyde for 30 min, then stained with 1 µg/ml Hoechst33342 for 10 min. The micronuclei were observed using a confocal laser scanning fluorescence microscope (Carl Zeiss confocal microscope LSM 800, Jena, Germany) with objective 40X magnification. More than 500 cells from randomly 15–30 fields were counted and calculated for the percentage of MN-containing cells as follows: [numbers of cells with MNs/total cells]*100.

Protein extraction and Western blot analysis

Total protein was extracted using RIPA lysis buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1% NP-40, 1% Na-Deoxycholate, 0.1% SDS) with protease and phosphatase inhibitor cocktail. Nuclear proteins were prepared as described previously (Seubwai et al., 2010). Briefly, cells were washed and lyzed in a hypotonic solution. Nuclei were collected by centrifugation and lyzed by nuclear lysis buffer (50 mM HEPES-KOH, pH 7.9, 1.5 mM MgCl₂, 10 mM KCl, 0.1% NP-40, 0.5 mM DDT) containing protease and phosphatase inhibitors.

Protein concentrations were quantified by Bradford reagent. Proteins were subjected to SDS-PAGE under denaturation conditions and transferred to PVDF membrane. The detection of targeted proteins was performed using specific antibodies: anti-cGAS (1:2000) anti-STING (1:2000), anti-NF-κB p65 (1:1000), and anti-GAPDH (1:5000). Immunoreactivity was detected by Immunobilion^®Forte western HRP substrate and visualized by an ImageQuant LAS 600 (GE Healthcare; Buckinghamshire, UK). Protein expression intensities were analyzed using ImageJ software version 1.53k (Schneider, Rasband & Eliceiri, 2012).

Estimation of immune cell infiltration

The immune cell infiltration (ICI) in CCA tissue was analyzed by a web server, Tumor Immune Estimation Resource (TIMER2.0, http://timer.cistrome.org/), based on computational algorithms and TCGA database (Li et al., 2020). The correlations between CGAS, STING, and RELA mRNA expression levels and the infiltration levels of different immune cells were evaluated by Spearman’s rank correlation coefficient (Rho) and plotted as scatter plots. The statistical significance of correlations is presented as a p-value.

Statistical analysis

Statistical analyses were conducted using SPSS version 28.0.1.0 (SPSS Inc., IL, USA) and GraphPad Prism^®8.0.2 Software (GraphPad Software Inc., La Jolla, CA, USA). Survival analyses between different groups were performed by Kaplan-Meier analysis with a log-rank test. The differences in the clinical features were compared using the Chi-square test. The multivariable Cox regression model was utilized to identify independent clinicopathological characteristics linked to patient survival, followed by testing backward selection to avoid the influence of confounding factors (Grant, Hickey & Head, 2019). The Pearson correlation coefficient (r) was applied to measure a linear correlation between two variables.

The quantitative data were presented as mean ± SD from three independent experiments. Multiple comparisons of means were analyzed by ANOVA, followed by a student’s t-test to identify differences between two groups. p-value less than 0.05 was considered statistically significant.

Micronuclei in CCA cell lines

CIN is one characteristic of CCA (Dalmasso et al., 2015; Jusakul et al., 2017); thus, the presence of MN was expected. MNs were detected in 3 CCA cell lines: KKU-055, KKU-213A, and KKK-D068. The results demonstrated that MNs were commonly visualized in CCA cell lines (Fig. 1A). Among them, KKK-D068 acquired the highest percentage of MN-containing cells, which was statistically significant when compared to the KKU-055 (Fig. 1B). Moreover, mis-segregated chromosome-derived MNs were generally observed in KKK-D068 cells (Figs. 1C, 1D).

The expression levels of cGAS, STING, and NF-κB p65 proteins were assessed by Western blot analysis. The results demonstrated that cGAS was highest in KKU-213A, and STING was elevated in KKK-D068 (Fig. 1E). NF-κB p65 was abundantly expressed in all three cell lines in both whole-cell lysates and nuclear fractions (Figs. 1E, 1F). Notably, only STING levels were correlated with MN formation in cells.

cGAS-STING pathway-related molecules were elevated in CCA and pancreatic adenocarcinoma (PAAD)

GEPIA web-based tool was used to evaluate the mRNA expression of cGAS-STING pathway-related molecules in gastrointestinal cancers, including CCA, PAAD, liver hepatocellular carcinoma (LIHC), colon adenocarcinoma (COAD), esophageal carcinoma (ESCA), rectum adenocarcinoma (READ), and stomach adenocarcinoma (STAD). CGAS, STING, TBK1, IRF3, NF-κB p65 (RELA), and NF-κB p50 (NFKB1), interleukin-6 (IL6), tumor necrosis factor (TNF), interferon alpha 1 (IFNA1), IFN beta 1 (IFNB1) were selected as the candidate molecules. A schematic representation of the cGAS–STING signaling pathway and its related molecules is shown in Fig. 2A. The results revealed that CGAS was highly detected in CCA, PAAD, ESCA, and STAD when compared to the normal counterparts, while STING, TBK1, IRF3, RELA, and NFKB1 were significantly upregulated in CCA and PAAD (p < 0.05). A significant increase in IL6 was observed in PAAD and ESCA (Fig. 2B). These results indicated that cGAS-STING pathway-related molecules were mostly induced in CCA and PAAD. The increased cGAS-STING pathway-related molecules observed in the public database prompt us to determine protein expression in the clinical CCA samples.

Demographic characteristics of CCA

A total of 95 CCA samples were selected to evaluate cGAS–STING pathway-related molecule expression. The patient ages ranged from 37 to 76 years, with a median age of 56. The cohort comprised 62 males and 33 females (a male-to-female ratio is 3.8:2). Anatomically, 33 cases (34.7%) were perihilar CCA and 62 cases (65.3%) were intrahepatic CCA. Fifty-two samples (54.7%) were non-papillary histological subtypes. CCA staging was performed according to the 7^th edition of the American Joint Committee on Cancer (AJCC) staging system (Edge & Compton, 2010); 52 cases (54.7%) were stage IV. Tumor sizes were reported in 78 cases (82.1%); 41 cases (52.6%) had tumors ≥7 cm, while 37 cases (47.4%) had tumors <7 cm (Table 1). The overall survival time, calculated from the date of surgery to death, ranged from 30 to 2,593 days, with a median survival time of 377 days. Surgical interventions included both curative and palliative procedures.

High expression of cGAS, STING, and differential expression of NF-κB p65 in CCA tissues

To demonstrate the expression of cGAS-STING pathway-related molecules in the clinical CCA samples, cGAS, STING, and NF-κB were selected as the representative proteins. Protein expression was assessed by IHC staining and semi-quantitatively assessed using H-scores. The representative IHC images are presented in Fig. 3A and H-score distributions are shown in Fig. 3B. The median H-scores of cGAS and STING were 300, indicating consistently high expression across the cohort. In contrast, the median H-score for NF-κB p65 was 135, reflecting variable expression.

Kaplan-Meier survival analyses using the log-rank test showed no significant correlation between the expression levels of cGAS, STING, and NF-κB p65 and overall patient survival (Figs. S1A–S1C). However, Pearson’s chi-square test revealed that cGAS and STING expression levels were significantly higher in perihepatic CCA compared to intrahepatic CCA (Table 1, p = 0.046 and p = 0.030). Additionally, elevated NF-κB p65 was observed more frequently in tumors ≥7 cm (p = 0.043).

The non-papillary CCA is more aggressive and associated with poorer outcomes (Saranaruk et al., 2023; Zen et al., 2006). Therefore, survival analyses were conducted separately for papillary and non-papillary subtypes. In non-papillary CCA, tumor stage IV and high NF-κB p65 were correlated with shorter survival of the patients (Figs. 3C, 3D). Patients with stage IV CCA had a mean overall survival of 344 ± 96 days, compared to 627 ± 282 days for those with CCA stage I–III (p = 0.042). Similarly, patients with high NF-κB p65 expression had a significantly shorter mean survival time (257 ± 79 days) than those with low expression (638 ± 212 days; p < 0.001). In contrast, the expression levels of cGAS and STING in non-papillary CCA (Figs. S2A, S2B), as well as the expression of all three proteins in papillary CCA (Figs. S3A–S3C), were not significantly associated with patient survival. It is also noteworthy that when CCAs were classified anatomically into intrahepatic and perihilar subtypes (Figs. S4–S6 and Table S1–S4), stage IV disease in the perihilar subtype was associated with shorter survival (Fig. S4C). No other clinical parameters showed a significant correlation with patient survival (Figs. S4–S6 and Table S1–S4).

Non-papillary subtype and high NF-κB p65 are the poor prognostic factors

The multivariate Cox regression analysis indicated that both non-papillary subtype and high NF-κB p65 levels were independent poor prognostic factors of CCA patients (Table 2, HR = 1.597; 95% CI [1.028–2.480], p = 0.037; HR = 1.679; 95% CI [1.074–2.626], p = 0.023). These finding suggest that both the non-papillary subtype and NF-κB p65 overexpression contribute to poorer prognosis in CCA patients.

CGAS, STING, and RELA (NF-κB p65) are commonly overexpressed in CCA

To validate our findings, expression levels of CGAS, STING, and RELA were analyzed using publicly available transcriptomic datasets (Andersen et al., 2012; Chaisaingmongkol et al., 2017). Comparison between CCA samples and normal tissues revealed that all three genes were significantly upregulated in CCA (Fig. 4A). When the correlation between two variables was determined by Pearson correlation coefficient, a significant positive correlation was observed between CGAS and STING expression in both cohorts (r = 0.2682, p = 0.006 and r = 0.2990, p = 0.004). Additionally, RELA levels were positively correlated with STING expression in the GSE76297 cohort (r = 0.2165, p = 0.039) (Fig. 4B). However, no significant correlation was found between RELA and CGAS in either dataset.

Correlations between CGAS, STING, and RELA expression levels and specific immune cell infiltration (ICI)

To explore the relationship between the expression levels of CGAS, STING, and RELA were associated with specific ICI, the TIMER2.0 web platform was utilized. The cGAS-STING pathway plays dual roles in promoting both anti-tumor and pro-tumor immunity, which can vary depending on the cancer type and stage (Khoo & Chen, 2018; Ng et al., 2018; Zheng et al., 2020). A range of relevant immune cell types engaged in the cGAS-STING immune response were analyzed, including neutrophils (Nagata et al., 2021), macrophages (M1 and M2 subtypes) (Hu et al., 2021), myeloid dendritic (DC) cells, natural killer (NK) cells, gamma delta (γδ) T cells, CD4⁺ T and CD8⁺ T cells, regulatory T (Treg) cells, and B cells (Khoo & Chen, 2018; Liu et al., 2023; Luo et al., 2025; Ng et al., 2018; Zheng et al., 2020) (Figs. 5). The results showed positive correlations between CGAS-STING and the infiltration of neutrophils, γδ T cells, and CD8⁺ T cells. Conversely, CGAS was negatively correlated with M2 macrophages, a cell type typically associated with tumor promotion. STING and RELA levels were positively correlated with B cell infiltration. Notably, only RELA showed a significant positive correlation with immunosuppressive Tregs infiltration. All three genes, CGAS, STING, and RELA, were positively associated with DC cell infiltration. However, no significant associations were observed between their expression and the infiltration of M1 macrophages, NK cells, or CD4⁺ T cells. These findings suggest that CGAS and STING may be linked to anti-tumor immunity, whereas RELA might contribute to immunosuppressive mechanisms through its association with Treg infiltration.