GBC specimens were collected from patients at Shengjing Hospital of China Medical University (Shenyang, China) from October 20, 2011 to December 10, 2020. Patients diagnosed with GBC and confirmed by pathologists according to the American Joint Committee on Cancer (AJCC) cancer staging system (8th edition) (26) and the 2019 (5th edition) WHO classification of tumours of the digestive system (27), who underwent surgery as the primary treatment and with comprehensive clinical and follow-up data, were included. Patients with preoperative chemotherapy, other concurrent malignant tumors, unsuccessful follow-up or those whose survival data cannot be obtained were excluded from this study. A total of 89 patients were selected from our patient database. Among these patients, 37 were males (42%) and 52 were females (58%). The mean age was 62.34±10.19 years, ranging from 30 to 89 years. This research arm of the present study was approved by the Ethics Committee of Shengjing Hospital of China Medical University (approval no. 2019PS036K). Written consent was obtained from all the patients involved in the present study. Each patient's age, sex, differentiation, Nevin stage (28), tumor invasion, regional lymph node metastasis, distant metastasis and histological stage were obtained.

The final date of the last follow-up was November 1, 2021. OS was defined as the interval either between the initial diagnosis and mortality or between the initial diagnosis and the date of the last observation for the surviving patients. Data were censored at the date of the last follow-up for any living patients.

HHLA2 expression in the GBC specimens was measured by immunohistochemistry. The GBC specimens were fixed in a 10% formalin solution for 24 h at room temperature, before they were embedded in paraffin, sliced into 4 µm-thick sections and placed onto glass slides. They were then dehydrated in an ascending ethanol gradient and cleared with xylene. The slices were treated with 95°C Sodium Citrate Antigen Retrieval Solution (cat. no. C1031; Beijing Solarbio Science & Technology Co., Ltd.) for 10 min. The slices were then treated with 3% H₂O₂ at 37°C for 10 min and blocked in 3% BSA (Beyotime Institute of Biotechnology) solution for 30 min at room temperature. Immunohistochemical staining were subsequently performed. Rabbit anti-HHLA2 monoclonal antibodies (1:100, cat. no. ab214327; Abcam) were used as the primary antibody and incubated at 4°C for 10 h, whereas HRP-conjugated goat anti-rabbit IgG was used as the secondary antibody (1:100, cat. no. ab288151; Abcam) and incubated at room temperature for 30 min. Diaminobenzidine (cat. no. DA1010; Beijing Solarbio Science & Technology Co., Ltd.) incubating for 5 min at room temperature was used to develop color reaction. Paraffin sections were counterstained with Haematoxylin (Beijing Solarbio Science & Technology Co., Ltd.) lasting for 2 min at room temperature.

Immunohistochemical staining results were assessed by two different pathologists using light microscopy (magnifications, ×200 and ×400; Olympus Corporation). In cases where there was disagreement, the results were evaluated again. If a consensus could not be reached, another pathologist participated in the evaluation. All pathologists were blinded to the patient's clinical data. In summary, ≥ two pathologists must agree with each result. Immunostaining in the tumor tissues was semi-quantified using the ImageJ software (version 1.52; National Institutes of Health), for the staining intensity. The staining intensity was divided into four levels as follows: i) Negative, 0; ii) low positive, 1; iii) positive, 2; and iv) high positive, 3. In total, five high-power fields (magnification, ×400) of each slice were randomly selected to judge the corresponding percentage of positive cells. The H-score of each case was measured by multiplying the staining intensity by the corresponding percentage of positive cells (range from 0 to 168). To associate the relationship between HHLA2 and patient outcome, the existence of tumor classifications should be revealed. Since H-score was determined to be continuous variables, the optimal cut-off value was determined using X-tile software (version 3.6.1; https://medicine.yale.edu/lab/rimm/research/software/) to divide the specimens into HHLA2-high group and the HHLA2-low group. The X-tile program performed a single, global assessment of every possible method of dividing the population. In the assessment, each H-score value of all the specimens was used as a cut-off value to divide the specimens into two groups. Subsequently, every χ² value of each grouping grouped by different H-score values were calculated using the Kaplan-Meier method respectively to find the highest χ² value, which was 7.16 (ranging from 0.40 to 7.16). The grouping with a χ² value of 7.16 was regarded as the optimal grouping, where the cut-off value of the optimal grouping was determined as the optimal cut-off value, which was 80 (29). Therefore, specimens with H-score >80 were included in the HHLA2-high group and specimens with H-score ≤80 specimens were included in the HHLA2-low group.

The GBC cell line GBC-SD (cat. no. CSTR:19375 .09.3101HUMTCHu16) was purchased from The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences. The cells were cultured in RPMI1640 medium (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.) and were treated with or without TGF-β1 (PeproTech, Inc). The TGF-β1 cytokine was dissolved in PBS containing 5% Trehalose as storage solution at a concentration of 20 µg/ml. The storage solution of TGF-β1 was added into the culture medium to reach a concentration of 10 ng/ml for treatment for 48 h at 37°C, whilst the control group received the same amount of vehicle treatment (PBS containing 5% Trehalose). For culturing, 5% CO₂ and 37°C was applied for the cell lines.

The cells were lysed with RIPA buffer (Beyotime Institute of Biotechnology) for 2 h on ice with a protein inhibitor cocktail (Beyotime Institute of Biotechnology). This mixture was then centrifuged at 12,000 × g for 30 min at 4°C, before the supernatant was collected and the protein concentration was measured using a BCA kit (Beyotime Institute of Biotechnology). In total, 5X loading buffer was added to the protein solution at a 1:4 ratio and heated at 95°C for 5 min. After SDS-PAGE in 5% stacking gel and in 10% separating gel (15 µg protein per lane) and transfer onto polyvinyl difluoride membranes (Thermo Fisher Scientific, Inc.), before blocked with 5% skimmed milk powder solution (diluted in Tris-buffered saline containing 0.1% Tween-20) for 2 h under room temperature. After blocking, the membranes were incubated with the corresponding primary antibodies against GAPDH (1:1,000; cat. no. ab181602; Abcam), vimentin (1:1,000; Abcam; cat. no. ab92547), α-smooth muscle actin (SMA; 1:1,000; Abcam; cat. no. ab124964), HHLA2 (1:1,000; Abcam; cat. no. ab214327), E-cadherin (1:1,000; Abcam; cat. no. ab40772), N-cadherin (1:1,000; Abcam; cat. no. ab18203) and Collagen I (Col-I; 1:1,000; Abcam; cat. no. ab34710) for 10 h at 4 °C. After incubating with the primary antibodies, the membranes were incubated with HRP-conjugated goat anti-rabbit IgG secondary antibodies (1:5,000; cat. no. ab6721; Abcam) for 1 h at room temperature. Protein expression levels were detected by Chemistar™ High-sig ECL Western Blotting substrate (cat. no. 180-5001; Tanon Science and Technology Co., Ltd.). The western blotting results were quantified using the ImageJ software (version 1.52; National Institutes of Health).

Total RNA was isolated using RNAiso Plus (Takara Bio, Inc.) according to the manufacturer's protocol and reverse transcribed into cDNA using the PrimeScript™ RT Master Mix kit (Takara Bio, Inc.). The temperature protocol was 37°C for 15 min, 85°C for 5 sec and 4°C for 1 h. qPCR was performed using the TB Green™ Premix Ex Taq™ II (Takara Bio, Inc.) and a Roche 96 instrument (Roche Diagnostics) to measure mRNA expression. The thermocycling conditions were as follows: Initial denaturation at 95°C for 30 sec; followed by 40 cycles at 95°C for 5 sec and 60°C for 30 sec; Relative mRNA expression levels were determined using the 2^−ΔΔCq method (30), where the experiments were repeated three times. GAPDH was used as the internal control. The GAPDH primers were purchased from Guangzhou RiboBio Co., Ltd. The sequences for GAPDH primers were: Forward, 5′-GAA CGG GAA GCT CAC TGG-3′, and reverse, 5′-GCC TGC TTC ACC ACC TTC T-3′, The sequences of the H19 primers were H19-forward, 5′-CGT GAC AAG CAG GAC ATG ACA-3′ and reverse, 5′-CCA TAG TGT GCC GAC TCC G-3′.

Cultured cells were first fixed in 4% paraformaldehyde at room temperature for 30 min, before they were blocked with 5% BSA (Beyotime Institute of Biotechnology) at room temperature for 1 h and permeabilized with 0.1% Triton X-100 at room temperature for 15 min. Next, the cells were incubated with 1:200 diluted primary antibodies against vimentin (cat. no. ab92547; Abcam), α-smooth muscle actin (cat. no. ab124964; Abcam), HHLA2 (cat. no. ab214327; Abcam), E-cadherin (cat. no. ab40772; Abcam), N-cadherin (cat. no. ab18203; Abcam) and Col-I (cat. no. ab34710; Abcam) for 10 h at 4°C. After washing three times with PBS containing 0.1% Tween-20 (PBST), the cells were incubated in 1:200 diluted Goat anti-Rabbit IgG (H+L) Highly Crossed-Adsorbed Secondary Antibody, Alexa Fluor™ 488 (cat. no. A-11034; Thermo Fisher Scientific, Inc.) for 1 h at room temperature, protected from light. DAPI (Thermo Fisher Scientific, Inc.) at the concentration of 300 nM was used to stain nuclei for 5 min at room temperature. Protein expression was observed and recorded by fluorescence microscopy (magnification, ×200; Olympus Corporation).

Lentiviruses encoding the short hairpin (sh)RNA silencing HHLA2 and non-target shRNA, the HHLA2 (GenBank accession no. NM_007072) or H19 (GenBank accession no. NR_ (NR_002196) overexpression vector and empty control lentiviral vector were designed and synthesized by Shanghai GeneChem Co., Ltd. The overexpression vector was sent for sequencing and designated GV348 [Ubi-MCS-SV40-Puromycin]. The knockdown vector was sent for sequencing and designated GV112 [hU6-MCS-CMV-Puromycin]. In brief, the 20 µg GV348 vectors were co-transfected with 15 µg pHelper 1.0 (envelope plasmid) and 10 µg pHelper 2.0 (packaging plasmid) in the 2nd generation transfection system into 293T cells (The Cell Bank of Type Culture Collection of the Chinese Academy of Sciences) cultured in DMEM (Hyclone; Cytiva) with 10% FBS (Hyclone; Cytiva) at 37°C with 5% CO₂. After 48 h culturing at 37°C, the supernatant of the 293T cells were collected and the lentivirus was concentrated through centrifugation of 25,000 g at 4°C for 2 h. Subsequently, the GBC-SD cell lines were plated into six-well plates until reaching 80% confluence, before the lentivirus was added and co-cultured with the cells for 24 h at 37°C (multiplicity of infection, 10). The medium was then replaced prior to another 48 h of incubation in the 5% CO₂ and 37°C incubator. Subsequently, 25 µg/ml puromycin treatment for 2 weeks at 37°C was used to select for the stably transfected cell lines before the knockdown efficiency was confirmed by western blotting or RT-qPCR after the cell line reached a confluence of 80%. In the rescue experiments, HHLA2-shRNA or NC-shRNA was transfected into H19-overexpressing GBC-SD cells. In total, 24 h after transfection, the cells were harvested for subsequent experiments. The target sequence of shRNA for HHLA2 was 5′-ATC CCG ATT CTC ATG GAA CAA-3′, whereas the target sequence of NC shRNA was 5′-TTC TCC GAA CGT GTC ACG T-3′.

The GBC-SD cells were seeded into six-well plates, and cultured until 100% confluence. A 20-µl sterile yellow plastic pipette tip was then used to scratch the cell monolayer. After 48 h of incubation with or without 10 ng/ml TGF-β1 at 37°C with serum-free RPMI1640 medium, the scratches were imaged by light microscopy (magnification, ×200) and analysed by ImageJ software (version 1.52; National Institutes of Health). Migration area was calculated by multiplying the migration distance (scratch width observed at 0 h-scratch width observed at 48 h) with the width of field of view (500 µm). Relative migration area was calculated by dividing the migration area by the migration area of each control group.

Transwell (Corning, Inc.) chambers were fist pre-coated at 37°C for 30 min with Matrigel (cat. no. 356234; BD Biosciences) that was diluted at 1:8 with serum-free RPMI1640 medium. In total of 2×10⁴ GBC-SD cells suspended with serum-free RPMI-1640 medium with or without 10 ng/ml TGF-β1 were seeded into the upper surfaces of each Transwell chambers, whilst the lower chambers were added with RPMI-1640 medium with 10% FBS. while the lower chambers were filled with culture medium with 10% FBS. After 48 h invasion at 37°C, the cells that had not passed through the chamber membrane were gently removed with a wet cotton ball. The membranes were then fixed with 4% paraformaldehyde at room temperature for 30 min and stained with 0.1% crystal violet at room temperature for 15 min and photographed under a light microscope (magnification, ×200). In total, five high-power fields (magnification, ×200) of each membrane were randomly selected for calculating the cell numbers by ImageJ software (version 1.52; National Institutes of Health).

An EdU assay was performed to evaluate the proliferation levels of the cells using Cell-Light Edu Apollo 488 In Vitro Kit (cat. no. C10310-3; Guangzhou RiboBio Co., Ltd.). Briefly, GBC-SD cells were labelled with 50 µM EdU (100 µl/well) for 2 h in a 5% CO₂ and 37°C incubator. After fixation with 4% paraformaldehyde at room temperature for 30 min, the cells were permeabilized with 0.5% Triton X-100 at room temperature for 10 min. Next, the cells were incubated with 1X Apollo staining solution of the EdU kit at room temperature for 30 min before the nuclei were stained with DAPI (Thermo Fisher Scientific, Inc.) at the concentration of 300 nM for 5 min at room temperature. The cells were observed and imaged by fluorescence microscopy (magnification, ×200).

The mouse xenograft experimental protocol was approved by the Ethics Committee of Shengjing Hospital of China Medical University (approval no. 2019PS325K) and performed in accordance with the Laboratory Animal Guideline for Ethical Review of Animal Welfare (31). All mice were housed in an animal room at a constant temperature of 25°C with 30-40% humidity, under a 12-h light/dark cycle and had free access to food and water.

A total of 24 6-week-old female BALB/c nude mice (weight, 19-22 g) were used to establish the GBC xenograft models. The mice were randomly divided into the HHLA2 overexpression group, NC overexpression group, HHLA2-shRNA group and NC-shRNA group, with 6 mice in each group. After the cells in each group reached a confluence of 80-90%, they were digested with 0.25% trypsin. Next, the cells were resuspended in PBS to form a cellular suspension at a density of 1×10⁶ cells in 100 µl. Under isoflurane inhalation anaesthesia (1-2%), each mouse was gradually and slowly inoculated subcutaneously with 100 µl this cell suspension in the right armpit. The health and behaviour of the mice were monitored once a day to determine if there were difficulties eating or drinking, symptoms of pain or long-term problems without recovery. If the tumor volume was found to be >2,000 mm³, the animal would be euthanized as a humane endpoint. In total, 18 days after inoculation, all mice were sacrificed by cervical dislocation after isoflurane inhalation (5% for 5 min). Cardiac arrest was then used to verify death by pulse palpation. Tumor tissues were collected for measurement and analysis of the volume and weight. The maximum tumor volume observed was 1,680 mm³ and the maximum tumour diameter was 15 mm.

All count data were analysed using the χ² test, continuity correction χ² test or Fisher's exact test. Survival differences were compared using the Kaplan-Meier method and log-rank test. Multivariate and univariate analyses were performed using a Cox stepwise proportional hazard model. Data for experiments performed in triplicate are expressed as the mean ± standard deviation. Comparisons between two groups were made using the unpaired Student's t-test. Differences among groups were tested using one-way analysis of variance (ANOVA) and Tukey's post hoc test. SPSS software version 26.0 (IBM Corp.) or GraphPad Prism 8.0 (GraphPad Software, Inc.) was used to perform the statistical analysis. P<0.05 was considered to indicate a statistically significant difference.

The basic clinical characteristics of 89 patients with GBC are summarized in Table I. In total, 37 (42%) of the patients with GBC had a lower Nevin staging of I-III, whereas 58% (n=52) had a high grade (IV-V). According to the AJCC 8th edition cancer staging system, the tumour invasions were classified into two groups, namely the low group (stage I-II; 36 cases; 40%) and the high group (stage III-IV; 53 cases; 60%). Tumour invasions were also classified into two groups, where the low group included T1 and T2 (44 cases; 49%) and the high group included T3 and T4 (45 cases; 51%). In terms of regional lymph node metastasis, 31 (34%) patients had regional lymph node metastasis, whilst 46 (52%) patients did not. For distant metastasis, 28 (31%) patients had distant metastasis (to the liver, intestine and peritoneum), whereas 61 (69%) patients did not. At the end of the follow-up period, 50 patients had succumbed to the disease, where the median follow-up duration was 70 months (range, 1 to 114).

To detect the HHLA2 expression spectrum in the GBC tissues, immunohistochemistry was performed. As shown in Fig. 1A, which shows a cross section of the GBC tissue, HHLA2 was expressed in the membrane and/or cytoplasm in GBC tissues and corresponding adjacent noncancerous tissues. The intense staining was predominately seen in apical surfaces of the papillary epithelium and luminal surfaces of the glands (black arrows). Compared with tissues with low HHLA2 expression, HHLA2 high expression tissue showed higher staining intensity and percentage of positively staining cells. Positive HHLA2 staining was observed in all 89 cases (100%). To evaluate the association between HHLA2 expression and various clinicopathological parameters of GBC progression, the cases were subdivided into two groups: HHLA2-high group (45 cases, 51%) and the HHLA2-low group (44 cases, 49%; Fig. 1A and B). HHLA2 expression was found to be positively associated with tumour progression (Fig. 1; Table I). Specifically, 32 cases in the HHLA2-high group reached Nevin stage IV-V, accounting for 71% (Fig. 1C). In addition, 33 cases in the HHLA2-high group reached AJCC stage III-IV, accounting for 73% (Fig. 1D), whilst 28 cases in the HHLA2-high group exhibited tumor invasion T3-T4, accounting for 62% (Fig. 1E). In terms of regional lymph node metastasis, 20 cases in the HHLA2-high group were demonstrated to be positive, accounting for 54% (Fig. 1F). According to Table I, higher HHLA2 expression was significantly associated with Nevin stage, AJCC stage, tumour invasion and regional lymph node metastasis.

Subsequently, univariate analysis of OS was performed to analyse the clinical prognosis of the patients with GBC. The OS of the HHLA2-high group was found to be shorter compared with that in the HHLA2-low group (hazard ratio=2.15; 95% CI=1.20-3.83; P=0.01), where patients with GBC and higher HHLA2 expression tended to have significantly shorter survival times (Fig. 1G).

To further explore the function of HHLA2 in the GBCs, the GBC-SD cell line were transfected with the HHLA2 plasmid for overexpression. The efficiency of this HHLA2 overexpression in GBC-SD cells was verified by western blot analysis (Fig. 2A and B). Subsequently, the expression of EMT markers α-SMA, vimentin, N-cadherin and Col-I were all significantly increased, whereas the expression of E-cadherin was significantly decreased, in HHLA2-overexpressing GBC-SD cells compared with those in the HHLA2-NC cells (Fig. 2C and D). The relative expression levels of these markers were then observed through immunofluorescence staining, which yielded similar observations compared with those from western blotting. These results suggest that HHLA2 can promote EMT in GBC (Fig. 2E).

Changes in the migratory ability of GBC-SD cells after HHLA2 overexpression was next measured using wound-healing assays. Overexpression of HHLA2 led to significantly more potent migratory capabilities compared with those in the HHLA2-NC group (Fig. 3A and B). The invasive abilities of these HHLA2-overexpressing cells were next explored through Transwell assays. The results demonstrated a significantly higher invasive ability in the HHLA2-overexpressing cell line compared with that in the HHLA2-NC group (Fig. 3C and D). An EdU assay was used to investigate the proliferation of GBC-SD cells, which revealed that there was a significant increase in the proliferation of the HHLA2-overexpressing cell line compared with that in the HHLA2-NC group (Fig. 3E and F). These results suggest that HHLA2 can mediate oncogenic functions in GBC in vitro.

To determine whether HHLA2 can regulate the tumorigenic ability of GBC in vivo, a xenograft mouse tumor model was established. HHLA2 is not expressed in the mouse genome (14), negating the requirement for measuring the protein or mRNA expression levels in these samples prior to the establishment of this xenograft model. The GBC cell line stably overexpressing HHLA2 and the negative control group were injected into the left flanks of six nude mice in each group. After 18 days, compared with those in the negative control group, tumors in the HHLA2 overexpression group exhibited a significantly larger volume and were heavier in weight (Fig. 3G-I). Taken together, these results suggest that HHLA2 can promote the tumorigenic ability of GBC in vivo.

To determine if HHLA2 has the potential to become a therapeutic target in GBC, GBC-SD cells were transfected with HHLA2-shRNA to knock down HHLA2 expression in GBC-SD cells. The knockdown efficiency was validated by western blotting (Fig. 4A and B). According to western blotting and immunofluorescence staining data, HHLA2 knockdown led to significant decreases in the expression of Col-I, N-cadherin, α-SMA and vimentin but significant increases in the expression of E-cadherin in GBC-SD cells (Fig. 4C-E). TGF-β1-induced EMT in vitro model was next established (Fig. S1A and B). There was a significant increase in the expression of HHLA2 in the GBC cells (Fig. S1A and B), which was reversed by transfection with HHLA2-shRNA. In addition, transfection with HHLA2-shRNA significantly reduced the expression of the EMT biomarkers compared with that in the TGF + sh-NC group (Fig. S1C-E). These results suggest that downregulation of HHLA2 can inhibit EMT in GBC both alone and/or in the presence of TGF-β1.

The effects of HHLA2 knockdown on GBC-SD migration, invasion and proliferation was subsequently assessed. Results of the wound healing assay revealed that HHLA2-shRNA transfection significantly decreased the migratory ability of GBC-SD cells both alone (Fig. 5A and B) and in the presence of TGF-β1 (Fig. 6A and B). Transwell assay results demonstrated that HHLA2 knockdown in GBC-SD cells significantly inhibited cell invasion (Fig. 5C and D) both alone and in the presence of TGF-β1 (Fig. 6C and D). In addition, results of the EdU assay showed significantly lower proliferation levels in the HHLA2-shRNA group compared with those in the sh-NC group (Fig. 5E and F). In the TGF-β1-treated in vitro model, knocking down HHLA2 expression was found to significantly reduce the proliferation of GBC-SD cells (Fig. 6E and F).

GBC-SD stably transfected with HHLA2-shRNA was then used to establish a xenograft model to further investigate its potential role in vivo. After 18 days, the tumors in the HHLA2-shRNA group exhibited a significantly light in weight and smaller volume (Fig. 5G-I). These observations suggest that HHLA2 can serve as a therapeutic target for inhibiting the tumorigenicity of GBC.

A previous study reported the oncogenic role of H19 in GBC, showing that H19 can promote EMT, migration and invasion in GBC cells (25). After stably transfected with H19 pc-DNA, GBC-SD cells were found with significantly higher expression levels of Col-I, N-cadherin, α-SMA and vimentin but lower expression levels of E-cadherin compared with those in the the H19-NC group (Fig. S2A and B). The H19 overexpression group also showed higher migration and invasion capabilities compared with those in the H19-NC group (Fig. S2C-F). However, the mechanism underlying this require further exploration. Therefore, a GBC cell line stably overexpressing H19 was constructed. The efficiency of H19 overexpression was verified using reverse transcription-quantitative PCR (Fig. 7A). According to the western blotting results, it was found that H19 overexpression significantly increased HHLA2 expression in the GBC-SD cells compared with that in the H19-NC group (Fig. 7B and C).

To explore the possible association between H19 and HHLA2 in EMT, GBC-SD cells stably overexpressing H19 were transiently transfected with HHLA2-shRNA. Both western blotting and immunofluorescence revealed marked downregulation of HHLA2 expression and the expression of EMT markers in the HHLA2 knockdown group, namely that of α-SMA, vimentin, N-cadherin and Col-I (Fig. 7D-F). By contrast, HHLA2 knockdown significantly increased the expression of E-cadherin (Fig. 7D-F). These results suggest that HHLA2 is a key factor in H19-induced GBC-EMT.

Following the observation that HHLA2 is a key factor in H19-induced EMT in GBC-SD, additional assays were performed to investigate the role of HHLA2 in H19 overexpression-induced migration, invasion and proliferation in vitro. Results from wound healing assay showed that cells with HHLA2 expression knocked down exhibited significantly lower migratory ability compared with that in the H19 + sh-NC group (Fig. 8A and B). Subsequently, results from the Transwell assay revealed that the invasive ability of cells transfected with HHLA2-shRNA was significantly decreased (Fig. 8C and D). In addition, results of the EdU assay suggested that knockdown of HHLA2 led to a significant reduction of GBC-SD cell proliferation (Fig. 8E and F).