1. Expression of ITGβ8 in patients with NSCLC
We analysed ITGβ8 expression levels across cancers in TCGA database. ITGβ8 was upregulated across most cancer types, including Cholangiocarcinoma (CHOL), colon adenocarcinoma (COAD), Esophageal carcinoma (ESCA), Glioblastoma multiforme (GBM), Head and Neck squamous cell carcinoma (HNSC), Kidney renal papillary cell carcinoma (KIRP), Acute Myeloid Leukaemia (LAML), Brain Lower Grade Glioma (LGG), Ovarian serous cystadenocarcinoma (OV), Pancreatic adenocarcinoma (PAAD), Rectum adenocarcinoma (READ), Stomach adenocarcinoma (STAD), Thyroid carcinoma (THCA), Thymoma (THYM), UCEC (Uterine Corpus Endometrial Carcinoma), LUAD (Lung adenocarcinoma) and LUSC (Lung squamous cell carcinoma) (Fig. 1A). More specifically, we analysed the expression levels of ITGβ8 in NSCLC tissues and adjacent normal tissues using the TCGA database (Fig. 1B). The results showed that ITGβ8 expression levels were significantly higher in LUAD and LUSC than in normal tissues. Next, to further confirm the expression of ITGβ8 in tumour tissue, we conducted experiments at the cellular level. Western blotting was used to analyse the expression levels of ITGβ8 in human standard bronchial epithelial cell lines (HBE) and human NSCLC cell lines (A549, H1299, H520, H1993). The results showed that the expression levels of ITGβ8 in A549, H1299, H520, and H1993 cells were higher than that in HBE cells (Fig. 1C). Moreover, we confirmed this result by qPCR. The expression level of ITGβ8 in A549, H1299, H520, and H1993 cells was upregulated (Fig. 1D). *p < 0.05; **p < 0.01, ***p < 0.001, ****p < 0.0001.
Using the HPA database, we further explored the protein expression level of ITGβ8. Figure 2 shows the results of the IHC image analysis. The protein expression level was higher in both LUAD and LUSC tissues than in normal lung tissues (Fig. 2A-C). Then, to further investigate the specific expression of ITGβ8, we performed IHC staining to assess the expression of ITGβ8 in 100 NSCLC samples. Images of representative samples were taken at 100× and 400× magnification (Fig. 2D). Samples with brown-yellow staining represented the positive expression of ITGβ8 cell. And the protein expression of ITGβ8 was mostly concentrated in the cell membrane. Then the staining of ITGβ8 was scored according to the intensity of staining and the proportion of positive cells. The results after scoring showed that among 100 NSCLC patients, there were 64 samples with high expression of ITGβ8. Then, we collected the clinicopathological characteristics of patients with negative and positive ITGβ8 expression (Table 1). The results show that ITGβ8-positive expression was positively correlated with lymphatic metastasis (P = 0.005) and TNM stage (P = 0.001). However, sex, age, differentiation, pathological type, and tumour size were not correlated with ITGβ8 expression.
Table 1
Association of ITGβ8 expression with clinicopathological features in NSCLC
Variables | | ITGβ8 | P value |
| Numbers | negative | positive | |
Gender | | | | 0.142 |
Male | 78 | 31 | 47 | |
Female | 22 | 5 | 17 | |
Age (years) | | | | 0.701 |
≤ 55 | 47 | 16 | 31 | |
> 55 | 53 | 20 | 33 | |
Differentiation | | | | 0.541 |
Well/moderate | 65 | 22 | 43 | |
poor | 35 | 14 | 21 | |
Pathological type | | | | 0.285 |
Adenocarcinoma | 54 | 22 | 32 | |
Squamous cell carcinoma | 46 | 14 | 32 | |
T classification | | | | 0.297 |
T1-2 | 57 | 23 | 34 | |
T3-4 | 43 | 13 | 30 | |
Lymph node metastasis classification | | | | 0.005* |
Negative | 35 | 19 | 16 | |
Positive | 65 | 17 | 48 | |
TNM stage | | | | 0.001* |
I-III | 42 | 23 | 19 | |
IV | 58 | 13 | 45 | |
*p<0.01 |
2. The expression of ITGβ8 in immune subtypes of NSCLC
From relevant research, immune subtypes are also related to the efficacy of immune checkpoint inhibitors. Immune subtypes are classified into six categories (C1-C6) according to somatic variation, microenvironment, and survival environment [26]. We thus explored different immune subtypes using the TISIDB database. ITGβ8 expression varied considerably in different immune subtypes (Fig. 3A, B). In LUAD, ITGβ8 expression was downregulated in type C3 (inflammatory) and upregulated in type C6 (TGF-β dominant) (P = 2.79e-04). In LUSC, it was significantly downregulated in type C3 and upregulated in types C4 (lymphocyte deplete) and C6 (P = 9.92e-04). Type C3 and C6 represent the longest and shortest prognosis, respectively.
3. Relationship between tumor-associated immune cells and ITGβ8 expression in NSCLC
The tumour immune microenvironment (TME) is an important factor affecting the effect of immunotherapy [27]. Among them, the infiltration of immune cells is an important factor affecting the degree of response of TME to immunotherapy. So, we focused on 24 immune infiltrating cells that are highly correlated with the TME in NSCLC (Fig. 4A-B). Notably, in LUAD, ITGβ8 expression is positive related to the presence of many immune cells, including aDCs, macrophages, NK cells, TFH cells and Th1 cells. Of these, macrophages, NK cells had a significant correlation (p < 0.0001). Negative related to CD8 T cells and Th17 cells (p < 0.0001). In LUSC, ITGβ8 expression is only positive related to the Tcm cells. And Negative related to almost all immune cells, including T cells, B cells, aDCs, CD8 T cells, cytotoxic cells, DCs, NK CD56bright cells, pDCs, T helper cells, Th1 cells, Th17 cells, Th2 cells and Tregs. In these cells, T cells, CD8 T cells, Cytotoxic cells and pDCs are more strongly correlated with ITGβ8 in LUSC (p < 0.0001).
4. Correlations between ITGβ8 expression and tumour mutational burden and Immune checkpoint genes in NSCLC
It has been established that TMB is linked to the cancer immunotherapeutic response and prognosis [28]. In this study, we assessed this relationship using the TCGA database. The scatter plot shows that TMB and ITGβ8 expression are negatively correlated with each other in LUAD (R=-0.088, P = 0.048). In LUSC, ITGβ8 expression was negatively correlated with TMB, but this relationship was not statistically significant (R=-0.025, P = 0.593) (Fig. 5A, B). These results shown that ITGβ8 may has little correlation with TMB in NSCLC.
Through the action of immune checkpoint genes, immune escape occurs in tumours [29]. To further explore the association between ITGβ8 expression and the effect of ICI treatment in NSCLC, we utilized the TCGA database to analyse the correlation of ITGβ8 expression and the expression of immune checkpoint genes (Fig. 6). In NSCLC, ITGβ8 expression was positively related to HHLA2 (R = 0.210, P < 0.001), CTLA4 (R = 0.140, P = 0.001), CD274 (R = 0.310, P < 0.001), HAVCR2 (R = 0.230, P < 0.001), PDCD1LG2 (R = 0.240, P < 0.001), SIGLE200CT15 (R = 0.360, P < 0.001), TIGIT (R = 0.160, P < 0.001), and IDO1 (R = 0.200, P < 0.001) levels, which are important immune checkpoint genes.
5. Significant pathways influenced by ITGβ8 in NSCLC
According to the |cor| >0.3 & P value < 0.05, select the relevant co-expressed genes, respectively. We separately selected LUAD and LUSC the top ten differentials ITGβ8-related genes from the enrichment analysis results to generate bubble plots for KEGG, GO-BP, GO-CC, and GO-MF (Fig. 7,8). In LUAD, enrichment analysis of the KEGG pathway showed that these co-expressed genes were significantly enriched in signaling pathways, including MAPK signaling pathway, TNF signaling pathway, etc. GO enrichment analysis revealed that these co-expressed genes were significantly enriched during focal adhesion, recycling endosome, etc. In LUSC, KEGG pathway showed that enrichment of Endocytosis, Hippo signaling pathways, ect. GO enrichment analysis include focal adhesion, developmental cell growth, Adherens juncton, etc. The detailed KEGG and GO enrichment analysis results are shown in Table S1 and S2.
6. The prognostic value of ITGβ8 expression in NSCLC
Because of the meaningful results above, we further investigated the impact of ITGβ8 expression on the prognosis of NSCLC patients. We performed Kaplan-Meier survival analysis using the databases and NSCLC specimens (Fig. 9). We discovered that patients with high ITGβ8 expression levels had a shorter OS in NSCLC than that of patients with low ITGβ8 expression levels from TCGA databases (P = 0.00019) (Fig. 9A). Of note, the survival analysis based on our NSCLC specimens showed the same results: the patient with high expression of ITGβ8 had an unfavourable OS (P = 0.024) (Fig. 9B). Furthermore, we wanted to determine whether there was a link between ITGβ8 expression and NSCLC patient prognosis after receiving ICI therapy. Then, we analysed the PFS and ORR of 30 patients with NSCLC who had received immunotherapy in our hospital. They had received first-line treatment with one of the following anti-PD-1 agents, including sintilimab (Cinda Biopharmaceuticals, China), camrelizumab (Jiangsu Hengrui Medicine, China), or tislelizumab (BeiGene, China). Characteristics of 30 patients enrolled in the study are summarized in Table 2. All patients were diagnosed as stage IV. Patients with high ITGβ8 expression experienced were significantly shorter survival compared with low ITGβ8 expression in patients receiving immunotherapy(P = 0.013) (Fig. 9C). The objective response rate was 66.7% for high ITGβ8 expression compared with 52.4% for low ITGβ8 expression.
Table 2
Patient’s Characteristics with ICI treatment
Characteristic | No. (%) of Patients (n = 30) |
Gender | |
Male | 25 |
Female | 5 |
Age (years) | |
≤ 55 | 13 |
> 55 | 17 |
Differentiation | |
Well/moderate | 14 |
poor | 16 |
Pathological type | |
Adenocarcinoma | 18 |
Squamous cell carcinoma | 12 |
T classification | |
T1-2 | 14 |
T3-4 | 16 |
Lymph node metastasis | |
Negative | 4 |
Positive | 26 |
ITGβ8 | |
Negative | 9 |
Positive | 21 |
We further evaluated and confirmed the prognostic value of ITGβ8 expression in NSCLC. We collected the clinical information of NSCLC patients, including age, sex, pathological type, TNM stage, lymph nodes, T stage, differentiation and expression levels of ITGβ8, together to perform Cox regression analysis (Table 3). The findings of the univariate analysis revealed that TNM stage (P = 0.000) and lymph nodes (P = 0.034) were risk factors for NSCLC patient OS and that high expression levels of ITGβ8 were associated with a shorter survival time (HR = 3.478, P = 0.001). Other clinicopathological characteristics were not linked with patient survival time. Furthermore, multivariate analysis revealed that ITGβ8 expression (P = 0.030) and TNM stage (P = 0.002) were independent prognostic variables in patients with NSCLC.
Table 3
Univariate and multivariate Cox regression analyses of clinicopathological and ITGβ8 expression
Variables | Univariate analysis | Multivariate analysis |
| HR (95% CI) | P value | HR (95% CI) | P value |
Age | 1.125(0.558, 2.026) | 0.782 | | |
Gender | 0.647(0.213, 1.243) | 0.562 | | |
Pathological type | 0.344(0.123,1.148) | 0.096 | | |
TNM Stage | 2.967(1.356, 2.573) | 0.000** | 1.677(1.255,2.654) | 0.002** |
Lymph Node | 1.526(1.316, 2.148) | 0.034* | | 0.672 |
T Stage | 1.257(0.835, 1.530) | 0.794 | | |
Differentiation | 0.459(0.850, 1438) | 0.589 | | |
ITGβ8 | 3.472(1.897, 6.326) | 0.001** | 2.847(1.467,5.368) | 0.030* |
CI, confidence interval; OS, overall survival; HR, hazard ratio; *p<0.05; **p<0.01.