Overexpressed ITGA2 promotes malignant tumor aggression by up-regulating PD-L1 expression through the activation of the STAT3 signaling pathway

The PANC-1, HepG2, SGC-7901, and MDA-MB-231 cell lines were purchased from ATCC. The Panc02 cell line was acquired from Tong Pai Technology (Shanghai, China). PANC-1, HepG2, and MDA-MB-231 cell lines were cultured in DMEM (Thermo Fisher Scientific, USA) supplemented with 10% FBS (HyClone, USA) at 37 °C in a 5% CO₂ incubator. SGC-7901 cells were cultured in RPMI 1640 containing 10% FBS under similar conditions.

The sh-Control and sh-ITGA2s were procured from Sigma-Aldrich. Lipofectamine 3000 (Invitrogen, USA) and Opti-MEM media (Invitrogen, USA) were used for the transfection reactions; lipofectamine 3000 was used to transfect 293 T cells to shRNA plasmids and viral packaging plasmids (pVSV-G and pEXQV). 24 h after transfection, the medium was replaced with fresh DMEM containing 10% FBS and 1 mM sodium pyruvate, and 48 h post-transfection, the virus culture medium was collected and added to the PANC-1, HepG2, SGC-7901, and MDA-MB-231 cells supplemented with 12 μg/ml of polybrene. 24 h after infection, the infected cells were selected with 10 μg/ml of puromycin. The shRNA sequence information is provided in the Additional file 1: Table S2.

Written informed consent was obtained from patients before surgery, as described previously [23]. The use of human tissue was approved by the Ethics Committee of Tongji Medical College, HUST, China.

Cell lysates were prepared using a radioimmunoprecipitation assay (RIPA) buffer (Thermo Fisher Scientific, USA) in the presence of a protease inhibitor cocktail (Sigma-Aldrich, USA) and Halt phosphatase inhibitor cocktail (Thermo Scientific, USA). Protein concentration was determined using a protein quantification kit (Sigma-Aldrich, USA) to ensure equal amounts of total protein were loaded in each well of SDS-PAGE gels. The protein was transferred to PVDF membranes (Pierce Biotechnology, USA) eventually and blocked with 5% not-fat milk for 1 h at room temperature (RT), then incubated with the specific primary antibody at 4 °C overnight or at room temperature (RT) for 2 h and washed with PBS for 3 times, followed by another incubation with secondary antibodies (BOSTER, USA). Finally, after washing with PBS for 3 times, the membranes were exposed to X-ray films using ECL detection reagents (Thermo Fisher Scientific, USA).

Total RNA was prepared using a Trizol reagent (Invitrogen, 15,596,026, USA). RNA samples (1 μg) were reverse-transcribed using a PrimeScript™ RT reagent Kit (TAKARA, RR047A, JPN). Quantitative real-time PCR was performed using a TB Green™ Fast qPCR Mix kit (TAKARA, RR430A, JPN). The sequences of the primers used for qRT-PCR are shown in the Additional file 1: Table S1. Values represent the averages of three technical replicates from at least three independent experiments (biological replicates).

Tissue microarray slides were purchased from Outdo Biobank (Shanghai, China) (HPan-Ade060CD-01). IHC analysis was performed with ITGA2 (Abcam, 1:5000), PD-L1 (CST, 1: 1000), and Phospho-STAT3-Y705 (Absin, 1:2000) antibodies to determine their protein expression levels. Two independent pathologists, who were uninformed about the patient data and histopathological features of the samples, were responsible for reviewing and scoring the degree of immunostaining separately. Staining intensity scoring was done, as described previously [24].

The colony formation assay was performed to examine the biological effect of ITGA2 on tumor cell survival. 500 Tumor cells infected with sh-Control or sh-ITGA2s, pcDNA3.1, or flag-ITGA2 plasmids were plated in six-well plates, and colony formation assays were performed and photographed after 7–14 days of plating.

The transwell assay was employed to evaluate cell invasion using 24-well Corning Costar inserts with 8-μm pores precoated with Matrigel (BD, United States; diluted 1: 8) for 6 h in an incubator. To begin with, 1 × 10⁴ cancer cells were planted into upper invasion chambers without FBS, and DMEM containing 30% FBS was introduced to the lower chambers. After a 12 to 24-h culture, cells were fixed on the insert membranes using methanol and stained with a Crystal Violet Staining Solution (Solarbio, China). The invading cells were photographed and evaluated under a microscope at five fields per well.

Transwell assay was employed again to examine cell migration using 24-well Corning Costar inserts with 8-μm pores without Matrigel-coating. Initially, 1 × 10⁴ cancer cells were planted into upper chambers, and DMEM containing 30% FBS was placed in the lower chambers. After a 12 to 24-h culture, the cells were fixed on the insert membranes using methanol and stained with a Crystal Violet Staining Solution (Solarbio, China). The migrated cells were photographed and assessed under a microscope at five fields per well.

The proliferation ability of tumor cells was assessed using (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2- (4-sulfophenyl)-2H-tetrazolium) (MTS reagent) (Abcam, ab197010, USA). Briefly, 1000 cells were introduced into 96-well plates with 100 μl DMEM containing 10% FBS and treated with serial small molecular inhibitors under different concentration gradients. 20 μl of an MTS reagent was added to each well three hours before the end of the incubation period following the manufacturer’s instructions. The absorbance of each well was detected at 490 nm with a microplate reader.

PANC-1, HepG2, SGC-7901, and MDA-MB-231 cells infected with sh-Control or sh-ITGA2s were prepared and stained for flow cytometry with the following antibodies: isotype APC anti-human IgG Fc Antibody (Biolegend, clone HP6017, USA) or APC anti-human CD274 antibody (Biolegend, clone 29E.2A3, USA) or for 30 mins at 4 °C. The cells were then washed with PBS three times, resuspended in 150 μl staining buffer, and analyzed for flow cytometry.

For flow cytometry analysis of the mouse tissue samples, single-cell suspensions were prepared and stained for flow cytometry with the following antibodies: APC conjugated CD45 antibody (Biolegend, 103,112, USA); FITC conjugated CD4 antibody (Biolegend, 100,510, USA); PE-conjugated CD8 antibody (Biolegend, 100,708, USA); APC conjugated CD11b antibody (Biolegend,101,212, USA); and FITC conjugated Gr1 antibody (Biolegend, 108,406, USA). Flow cytometry was performed on BD FACSCelesta (BD Biosciences, USA), and the data were analyzed using FlowJo.

GEPIA (http://​gepia.​cancer-pku.​cn/​) and Human Protein Atlas cancer databases (https://​www.​proteinatlas.​org/​) were mined to predict the ITGA2 differential expression level in cancerous and healthy groups. The survival analyses, performed by GEPIA for hypothesis evaluation, used a log-rank test based on gene expression levels. Gene correlation analyses conducted by Pearson correlation statistics were carried out with GEPIA for all the given sets of GTEx and TCGA expression data.

Athymic nude (nu/nu) mice (4–5 weeks old, male) were purchased from Vitalriver (Beijing, China). 5 × 10⁶ PANC-1 human pancreatic cancer cells infected with sh-Control or sh-ITGA2 lentivirus were dispersed in a 100 μl solution of PBS and then inoculated subcutaneously into the left dorsal of the nude mice. Tumor sizes were measured with a digital Vernier caliper every two days for a total of 21 days. Tumor volumes were calculated using the formula: tumor volume (mm³) = (L x W²)/2. The animals were sacrificed on day 21 or when tumor volume reached 1000 mm³. The experimental procedures with nude mice were conducted per the guidelines approved by our local ethics committee (Tongji Medical College, HUST, China).

All experimental procedures with nude mice were conducted per the guidelines approved by the local ethics committee (Tongji Medical College, HUST, China). C57BL/6 mice (6–8 weeks male, Vitalriver, Beijing, China) were subcutaneously injected at the right flank with 5 × 10⁶ Panc02 cells infected with sh-Control or sh-ITGA2 #1 lentivirus and diluted in 100 μl solution (PBS and Matrixgel, 1:1 ratio). Tumor lengths and widths were measured every day with a digital caliper, and tumor volumes were calculated using the formula: (L x W²)/2. When the tumor size reached approximately 50 mm³, mice with analogous types of tumors were divided randomly into two groups and injected intraperitoneally with 200 μg of anti-PD-1 antibody (BioXcell, Clone RMP1–14, USA) or 200 μg of IgG (BioXcell, Clone 2A3, USA) (given at days 0, 3, 6). Mice were euthanized once the tumor volumes reached 200 mm³, and their tumors were harvested.

A 0.5–2.0 μg primary antibody was added to the cell lysate sample and incubated on ice overnight. The next day, 20–50 μL of Protein A beads (CST, #9863, USA) was added to the cell lysate and primary antibody mixture, and the content was gently rotating incubated at 4 °C for 3 h. The next day, the beads were washed at least six times with lysate buffer on ice, and then subjected to western blotting analysis.

After analyzing the mRNA expression level of ITGA2 in several cancers and non-tumor tissues using the GEPIA web tool, we found that the mRNA expression level of ITGA2 was significantly up-regulated in pancreatic adenocarcinoma (PAAD) and stomach adenocarcinoma (STAD) (Fig. 1)a.

Since the differential expression of ITGA2 was more significant in pancreatic cancer compared with non-tumor tissue, we paid more attention to exploring the biological role of ITGA2 in PDAC. The protein expression levels of ITGA2 in PDAC specimens were examined. Immunohistochemistry (IHC) based on a cohort of patients (healthy pancreatic specimens: n = 25, PDAC TMA specimens: n = 31) was performed using the tissue microarray approach (TMA). The IHC staining score was evaluated, as described previously [23], and the results showed that ITGA2 proteins increased profoundly in pancreatic cancer tissues compared with non-tumor pancreatic tissues (Fig. 1)b,c. Consistent with IHC, comparative analysis with matched non-tumor pancreatic tissues demonstrated that ITGA2 was overexpressed in twelve primary PDAC samples (Fig. 1)d,e.

To confirm the clinical relevance of ITGA2 expression levels on survival rate in malignant tumors further, the overall survival (OS) and disease-free survival (DFS) of several cancer patients related to ITGA2 expression were examined using the GEPIA web tool and The Human Protein Atlas web tool. The results revealed that the high expression of ITGA2 was significantly associated with unfavorable survival rates in PAAD and liver hepatocellular carcinoma (LIHC) patients, although the outcome was more evident in PDAC patients (Fig. 1f-h and Additional file 1: S1a-c). These findings suggest that ITGA2 is up-regulated in PAAD and STAD and is related to poor prognoses in PAAD and LIHC patients.

Since up-regulated ITGA2 was a poor prognostic biomarker for cancer, we explored its biological role in cancerous tumors further. Following the knockdown of ITGA2 using short Harpine RNA (shRNA) in PANC-1, HepG2, SGC-7901, and MDA-MB-231 cell lines (Fig. 2)a,b, the MTS and colony formation assays demonstrated that silencing ITGA2 inhibited the cancer cell growth (Fig. 2)c,d. Additionally, the transwell assay showed that the migration and invasion abilities of cancer cells decreased markedly after ITGA2 down-regulation (Fig. 2e and Additional file 1: S2). Our results indicate that knocking down ITGA2 could inhibit cancer aggression in vitro.

To explore the role of overexpressed ITGA2 in cancer progression further, PANC-1, HepG2, SGC-7901, and MDA-MB-231 cell lines with stably overexpressed ITGA2 were established (Fig. 3)a,b. Colony formation and MTS assays indicated that overexpressed ITGA2 enhanced the growth capability of PANC-1 cells, HepG2 cells, SGC-7901 cells, and MDA-MB-231 cells (Fig. 3)c,d. Moreover, the transwell assay showed that migration and invasion abilities were also up-regulated in ITGA2-overexpressing cancer cells (Fig. 3e and Additional file 1: S3). These data suggest that the overexpression of ITGA2 could promote the aggression ability of cancer cells in vitro.

Given that ITGA2 levels regulated the progression of cancer cells in vitro, we studied the tumor-growth-promoting effect of ITGA2 in pancreatic cancer in vivo. PANC-1 cells infected with sh-Control, sh-ITGA2, or sh-ITGA2 + ITGA2 were harvested for RT-PCR analysis, Western blot analysis, colony formation assay, and MTS assay (Fig. 4)a,d. The colony formation and MTS assays revealed that the inhibition of ITGA2 significantly slowed down tumor growth, whereas ITGA2 overexpression promoted cancer growth in PANC-1 cells (Fig. 4)c,d.

PANC-1 cells under the same conditions as above were subcutaneously injected into the left flank of nude mice for the xenograft assay. As Fig. 4e-g shows, tumors formed in ITGA2-silenced PANC-1 cells were smaller and lighter than those in ITGA2 expressed cells, but the body weights were not significantly different between the different treatments (Additional file 1: Fig. S4a). In turn, tumors formed after rescuing ITGA2 were more massive than tumors from ITGA2-knocked down cells. Moreover, Ki-67 staining of the implanted tumors further showed that cell proliferation ability correlated significantly with the expression level of ITGA2 (Fig. 4h). Therefore, ITGA2 could promote tumor growth in pancreatic cancer in vivo.

Because ITGA2 is critical for promoting cancer cell progression in vivo and in vitro, we employed RNA-seq analysis to identify the underlying mechanism of how ITGA2 regulates carcinogenesis. Our results identified 427 up-regulated and 365 down-regulated genes after knocking down ITGA2 in PANC-1 cells (Fig. 5)a,b. The pathway analysis of the 40 most significantly regulated genes indicated that ITGA2 repression occurred in the PD-1 checkpoint pathway in cancer cells. Interestingly, the expression of PD-L1 was down-regulated by ITGA2 inhibition (Fig. 5)c. Consistent with the finding from RNA-seq analysis, ITGA2 silencing by two independent sh-RNAs decreased PD-L1 expression in pancreatic, breast, gastric, and liver cancer cells (Fig. 5)d,f. Also, the expression of PD-L1 was up-regulated at both the mRNA and protein levels in cancer cells with overexpressed ITGA2 (Fig. 5g,h). Per all these results, ITGA2 transcriptionally regulates PD-L1 expression in various types of cancer cells.

To determine the clinical relationship between ITGA2 and PD-L1 in human cancer specimens, IHC analysis of the protein expression of ITGA2 and PD-L1 was conducted in a cohort of PDAC patients (n = 31). The result showed that there was a positive correlation between the protein levels of ITGA2 and PD-L1 expression in PDAC specimens (Spearman correlation r = 0.6352, P < 0.001) (Fig. 6a,c). Furthermore, using the GEPIA web tool, we found that the mRNA expression levels of ITGA2 correlated positively with PD-L1 levels in several malignant tumors, including pancreatic, liver, gastric, and breast cancers (Fig. 6d). Therefore, our data suggest that the level of ITGA2 in cancer specimens has a positive correlation with PD-L1 expression.

As demonstrated above, ITGA2 regulates the expression of PD-L1 in solid cancer cells. So, we assessed the same phenomenon in vivo. Panc02 cells, the murine pancreatic cancer cell lines, were infected with sh-Control or sh-Itga2. In agreement with previous findings, Western blot and RT-PCR analyses showed that Itga2 silencing resulted in decreasing expression of Pd-l1 in Panc02 cells (Fig. 7a,b). After subcutaneous injections of Panc02 cells into the left flank of C57BL immune-proficient mice, the mice were treated with or without anti-PD-1 antibody according to the procedure shown in Fig. 7c. Consistent with the inhibition of cell proliferation by knocking down ITGA2 in human PDAC cells (Fig. 2 and Fig. 4), Itga2 repression inhibited tumor growth in murine Panc02 cells in vivo (Fig. 7d,e). Due to the reduction of Pd-l1 expression after Itga2 suppression, anti-PD-1 antibody treatment manifested a stronger anti-tumor effect in the Itga2 knockdown group by slowing down tumor growth and improving the survival rate of tumor-bearing mice (Fig. 7d,e). The body weights, though, were not significantly different between the different treatments (Additional file 1: Fig. S4b). Flow cytometry analysis of tumors excised from the mice at the end of treatment revealed that the populations of tumor-infiltrated CD45⁺CD8⁺ T-cells and CD45⁺CD4⁺ T-cells increased after Itga2 knockdown, while the population of CD11b⁺Gr1⁺ myeloid cells decreased (Fig. 7f). Meanwhile, the combined therapy of anti-PD-1 antibody and Itga2-knockdown increased the populations of tumor-infiltrated CD45⁺CD8⁺ T-cells and CD45⁺CD4⁺ T-cells further but reduced that of myeloid-infiltrated CD11b⁺Gr1⁺ cells in tumors (Fig. 7f). Using the TIMER web tool [25] to study the immune infiltration level also showed that the expression of ITGA2 correlated positively with dendritic cell infiltration levels and negatively with CD4+ T cell infiltration levels (Fig. 7g). These results suggest that blocking ITGA2 could inhibit cancer-associated immunosuppression and improve the efficacy of immune checkpoint blockade therapy through the downregulation of PD-L1 in vivo.

Given that ITGA2 promoted PD-L1 expression in cancer cells, we wanted to identify the underlying mechanism of the process. Interestingly,we detected drug sensitivity (IC50 values) from PANC-1 cells to some cancer-related pathway small inhibitors in the sh-Control and sh-ITGA2 groups (Fig. 8a). The results showed that the JAK and STAT3 inhibitors increased viability loss in ITGA2-knockdown cells further (Fig. 8a), indicating that ITGA2 might be involved in modulating the STAT3-related signaling pathway in cells.

To determine the relationship between ITGA2 and the STAT3 signaling pathway, we, first of all, examined the mRNA expression level of STAT3 in cancer cells infected with sh-Control or sh-ITGA2. The result revealed that there was no noticeable change in the mRNA expression of STAT3 (Fig. 8b), which pointed to ITGA2 not regulating the expression of STAT3 at the transcriptional level. Per the immunoprecipitation assay (IP), ITGA2 interacted with STAT3 in various types of cancer cells. (Fig. 8c). Furthermore, Western blot analysis established that there was no regulation of STAT3 expression at the protein level (Fig. 8d). Consequently, we could not link the role of ITGA2 in regulating PD-L1 expression to the regulation of STAT3 expression.

L.L. Bu et al. reported that the expression of PD-L1 was regulated by p-STAT3 in a dose-dependent manner [26]. Herein, in our next step, we hypothesized that p-STAT3 was the key protein that mediated PD-L1 expression induced by ITGA2. To test this hypothesis, we examined the changes in phospho-STAT3 in cancer cells. We found that knocking down ITGA2 suppressed the phosphorylation level of STAT3 (Fig. 8d). Similarly, the overexpression of ITGA2 increased the phosphorylation level of STAT3 and protein level of PD-L1 in cancer cells (Fig. 8e), and these effects were diminished by STAT3 inhibitor treatment (Fig. 8f). More importantly, the activation of the STAT3 pathway while knocking down ITGA2 also upregulated the expression level of PD-L1 (Additional file 1: Fig. S5). Overall, our findings suggest that ITGA2 enhances STAT3 phosphorylation by serving as the origin of the transcription and translation of PD-L1 in cancer cells.