The RNA-seq data and clinical information of 80 patients with UM were obtained from TCGA database (https://portal.gdc.cancer.gov/). Data analysis was performed using R ×64 4.0.4 software (https://www.rstudios.co/). Overall survival (OS) and disease-free survival (DFS) analyses of STING expression were performed using the Gene Expression Profiling Interactive Analysis database (http://gepia.cancer-pku.cn) (14).

A total of 20 clinical tissues were collected from patients who underwent primary enucleation at Eye & ENT Hospital of Fudan University (Shanghai, China), diagnosed with UM between January 2013 and December 2019. Among these, five were metastatic UM tissues obtained from patients who developed metastasis in the liver, lungs or other distant organs. By contrast, the other five were non-metastatic UM tissues obtained from patients who did not develop metastasis after more than five years of follow-up time. Retrospective data on survival and metastasis were obtained from medical records updating patients' follow-up information semi-annually. The remaining ten tissues comprised five primary UM tissues with surrounding para-UM tissues (choroid tissues), which were used to compare UM tissues and normal tissues. Written informed consent was obtained by all patients for using their tissues and information in researches. The present study was approved by The Ethics Committee of Eye & ENT Hospital of Fudan University (approval no. 2020044-1; Shanghai, China). Tissues were collected by an experienced pathologist immediately after enucleation and stored in liquid nitrogen. In the present study, none of the patients had received any treatment prior to enucleation. Clinicopathological characteristics (including age and sex distribution) are included in Table II.

The UM cell lines (Mel 202, 92.1, Mel270, Omm2.2, Omm2.3, Omm1 and Omm2.5) used in the present experiments were kindly provided by WuXi AppTec (https://www.wuxiapptec.com/zh-cn). THP-1 (Human Monocytic Cell Line), 293T (Human Embryonic Kidney Cell Line) and ARPE-19 (Adult Retinal Pigment Epithelial Cell Line) were purchased from the American Type Culture Collection. The UM cell lines, THP-1 and ARPE-19 were maintained in Roswell Park Memorial Institute 1640 (RPMI-1640) medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin (PS) and 293T in Dulbecco's Modified Eagle Medium (DMEM; all from Gibco; Thermo Fisher Scientific, Inc.) with 10% FBS and 1% PS. All the cell lines were cultured in a humidified incubator at 37°C under 5% CO₂. These procedures were conducted in accordance with the highest ethical standards and best laboratory practices to ensure the validity and reliability of the findings.

Primary antibodies targeting STING (cat no. 13647), p38-MAPK (cat. no. 8690), phosphorylated p38-MAPK (cat. no. 4511) and GAPDH (cat. no. 2118) were obtained from Cell Signaling Technology, Inc. The Anti-Rabbit Secondary antibody (cat. no. DM-001) Detection Module for Wes was purchased from ProteinSimple. The compound SB203580 (cat. no. S1076) was purchased from Selleck Chemicals. Wes Separation kits (12–230 kDa; cat. no. SM-W004; ProteinSimple) were used to perform our protein immunoassay procedures.

Short hairpin RNA (shRNA) sequences (Genomeditech) were designed to downregulate STING expression, named shSTING, and the scramble shRNA control was named Scramble. The shRNA sequence targeting human STING and the scramble shRNA control sequences were as follows: 5′-TCTCAAGAGAAATCCGTGCGGA-3′ (shSTING), and 5′-TTCTCCGAACGTGTCACGT-3′ (Scramble). The double strands of shRNA were inserted into lentiviral vector through the pGMLV-SC5-PURO RNAi packaging plasmid. The concentrations of the two packaged lentiviral vectors were 5×10⁸ TU/ml. Human full-length STING cDNA was cloned into the expression plasmid pHBLV-CMV-MCS-3FLAG-EF1-ZsGreen-T2A-PURO by Hanbio Biotechnology Co. Ltd. to upregulate STING (STING⁺), and the empty pHBLV-PURO lentiviral vector was used as negative control (Control). The concentrations of the two packaged lentiviral vectors were 4.5×10⁸ TU/ml and 3×10⁸ TU/ml, respectively.

UM cells mixed with targeted or negative control lentiviral vectors were seeded in 24-well-plates (1.6×10⁵ cells vs. 3.2×10⁶ TU lentiviral vectors per well), incubated in a humidified incubator at 37°C under 5% CO₂ for 4~6 h until cell attachment, and then replaced with fresh mediums. Polybrene Reagent (cat. no. GM-040901; Genomeditech) was used for transfection with the concentration of 5 µg/ml. Stable cells with STING downregulated or upregulated were selected using puromycin with the concentration of 2 µg/ml. The transfected cells were used for further experiments after 48 h incubation.

Proteins were extracted from UM cells or tissues using RIPA buffer and a phosphatase inhibitor, followed using the Wes Separation (Wes) system according to the manufacturer's protocol. The Wes system is a next-generation hand-free capillary immunoassay platform combining protein separation with sensitive chemiluminescence and fluorescence immunodetection. This mature technology has been applied widely in vaccines, biopharmaceutical purification and other proteins-related researches (15–17).

For immunodetection, a Master Mix solution, sample buffer, ladder solution, and a chemiluminescence mixture were prepared. Next, the protein samples were mixed with the sample buffer and Master Mix and denatured at 95°C for 5 min. The primary antibodies were diluted in 1:50~200. Ladder solution, protein samples, primary antibodies and secondary antibodies, chemiluminescence mixtures, and wash buffer were sequentially added to the reaction plate on-ice operation. The reaction plate and capillary tubes were placed into an automated analyzer that automatically loaded the proteins for separation and immunodetection in each capillary tube. Finally, immunoreactive protein bands were visualized according to chemiluminescence signals and quantified based on the signal intensities, using Compass Simple Western software (ProteinSimple).

Cell proliferation was assessed using Cell Counting Kit-8 (CCK-8; Dojindo Laboratories, Inc.) according to the manufacturer's instructions. A total of 2.5×10³ UM cells were seeded into 96-well plates before transfection and treatment. 10 µl of CCK-8 reagent was added in each well and incubated at 37°C for 2 h. Cell viability was measured at 450 nm at 0, 24, 48 and 72 h after transfection and drug treatment using microplate reader (Tecan Group, Ltd.).

In brief, hydrated Transwell chambers (pore size, 8-µm, Corning, Inc.) were coated with Matrigel by RPMI-1640 medium in the cell incubator at 37°C for 2 h. UM cells (1×10⁵) were seeded into each Transwell chambers in RPMI-1640 medium without FBS. The Transwell chambers were immersed in 24-well plates containing RPMI-1640 medium with 10% FBS and the plates were incubated at 37°C for 24 h. The cells were fixed in chambers using 4% paraformaldehyde at room temperature for 20 min and stained with 1% crystal violet at room temperature for 15 min. The cells attached to the upper surface of the chambers were wiped out, and cells that invaded through the pores were counted in six randomly selected fields (magnification, ×100) using the light microscope (Leica Microsystems GmbH). The UM cells (5×10⁵) were placed into six-well plates and cultured for 24 h. A 200-µl pipette was then used to scrape the cell layer in a straight line, creating a scratch. The cells were washed twice with RPMI-1640 medium without FBS, then cultured in RPMI-1640 medium supplemented with 1% FBS. Images were captured after 48 h, and the cell migration rate was compared with that of the initial scratch.

In the Transwell assays and wound healing experiments, although four UM cell lines were treated in different time and repeated experiments were performed, the backgrounds of the images compared at the same time/cell lines were consistent whether in Transwell assays or wound healing experiments.

To compare the differences between two groups of data, the unpaired Student's t-test was employed. MEDCAL software was used to identify the cutoff value of STING expression through receiver operating characteristic (ROC) curve analysis. The Kaplan-Meier test (followed by log-rank test) was used to analyze OS and DFS. The correlation between the clinical information of patients with UM and STING expression was assessed using the chi-square test. All statistical analyses were performed using IBM SPSS Statistics software (version 25.0; IBM Corp.). *P<0.05 was considered to indicate a statistically significant significance.

A total of 80 mRNA-seq datasets and the corresponding clinical information of patients with UM were retrieved from the TCGA database. An area under the curve was generated based on STING expression to predict the 5-year survival of patients. The area under the ROC curve was 0.787, indicating that expression of STING was a favorable predictor of survival in patients with UM. The cutoff value for distinguishing between high and low STING expression was determined to be 10.693 Fragments per kilobase of transcript per million mapped reads (Fig. 1A and Table I). Survival analysis of patients with UM demonstrated that those with higher STING expression had significantly shorter overall and disease-free survival times than those with lower STING expression (Fig. 1B and C; P<0.01). Furthermore, the chi-square test was conducted to evaluate the relationship between STING expression and clinical characteristics. The results revealed a significant association between high risks of metastasis and higher expression of STING, including the histological types of epithelioid cells, higher UM clinical stage, and survival with tumors (Table II; P<0.05). These findings suggested that upregulation of STING expression indicates a poor prognosis in patients with UM.

To determine the expression levels of STING in UM, a Wes Separation protein immunoassay was used to detect STING in UM cell lines, UM tissues and para-UM tissues (choroid tissues). In UM cell lines, STING exhibited variable expression levels relative to those in THP-1, 293T and ARPE-19 cells (Fig. 2A). The present findings also revealed that STING in UM tissues was significantly higher in UM tissues than in para-UM tissues, suggesting that STING may be involved in the development of UM (Fig. 2B; P<0.05). However, STING protein expression levels did not significantly differ between metastatic and non-metastatic UM tissues (Fig. 2C; P>0.05).

Transwell experiments with Matrigel and wound healing experiments were performed to evaluate the invasion and migration ability. Loss-of-function experiments were carried out on Mel270 and Mel202 cells, which exhibit high STING expression, while gain-of-function experiments were carried out on Omm2.3 and Omm2.5 cells, with low STING expression levels. First, the shRNA targeting STING (shSTING) was transfected into Mel270 and Mel202 cells to downregulate STING with 55~66 and 42~59% reduction respectively, compared with the Scramble groups (Fig. S1A; P<0.001). Transwell experiments and wound healing experiments demonstrated that the downregulation of STING in Mel270 and Mel202 significantly inhibited their invasive and migratory abilities (Fig. 3A-D; P<0.05). In parallel, the STING cDNA (STING⁺) was transfected into Omm2.3 and Omm2.5 to overexpress STING with 29~32-fold and 15~18-fold overexpression rate respectively (Fig. S1B; P<0.001), compared with the Control groups. The experiment assay revealed that overexpression of STING in Omm2.3 and Omm2.5 significantly promoted their invasive and migratory abilities (Fig. 3E-H; P<0.05). However, changes in STING expression did not significantly affect UM cell proliferation (Fig. S2A-D; P>0.05).

The roles of MAPKs signaling and BAP1 in UM cell proliferation and metastasis have been extensively investigated (18,19). Studies have shown that STING promotes the development of tumors through canonical and non-canonical nuclear factor-kappa B (NF-κB) pathways as well as p21 (13,20,21). To gain insights into the mechanism of STING-induced UM invasion and migration, Wes separations were performed to detect the expression of MAPKs (p38-MAPK, extracellular regulated protein kinases/ERK, c-Jun N-terminal kinase/JNK), BAP1, NF-κB (p65, p100/52), and p21 in UM cells. The present results demonstrated that STING downregulation and overexpression consistently modulated the levels of phosphorylated p38-MAPK (Fig. 4A), whereas the expression of ERK, JNK, BAP1, NF-κB (p65, p100/52) and p21 remained unchanged. The phosphorylation of p38-MAPK decreased when STING was downregulated, but increased upon STING overexpression. Furthermore, Transwell and wound healing experiments revealed that SB203580, a potent p38-MAPK inhibitor (at 10 µM), effectively reversed STING-enhanced cell invasion and migration (Fig. 4B-E; P<0.05). Notably, p38-MAPK inhibition had no significant effect on UM cell proliferation, irrespective of STING expression levels (Fig. S3A-D; P>0.05).