ITGA5 is an independent prognostic biomarker and potential therapeutic target for laryngeal squamous cell carcinoma

Abstract Background Integrin α5 (ITGA5) was involved in a variety of cancers. However, the role of ITGA5 in laryngeal squamous cell carcinoma (LSCC) remains unknown. Methods The expression of ITGA5 and the corresponding clinicopathological parameters of LSCC patients the TCGA database. Five datasets (GSE51985, GSE59102, GSE84957, GSE27020, and GSE65858) were downloaded from the GEO database as validation sets. Kaplan–Meier plotter, Cox regression analysis, and nomogram were performed to determine the prognostic value of ITGA5 in LSCC. GO, KEGG, and GSEA were used to explore the underlying biological functions of ITGA5 in LSCC. The algorithms ESTIMATE and CIBERSORT were adopted to evaluate the association between ITGA5 and the infiltration of the immune cells. The algorithm pRRophetic was used to estimate the response to chemotherapeutic drugs. Results The expression of ITGA5 was higher in the LSCC samples and linked to poor overall survival and recurrence‐free survival. Further, the Cox regression analysis confirmed that high expression of ITGA5 was an independent unfavorable prognostic factor. The predictive performance of nomogram based on the expression of ITGA5 was accurate and practical. The functional enrichment analysis confirmed that ITGA5 was related to the construction of the components and structures of the extracellular matrix. Finally, patients with high ITGA5 expression were more likely to benefit from docetaxel and gemcitabine. Conclusion The expression of ITGA5 was elevated in the LSCC and was a predictor for prognosis and chemotherapeutic response in LSCC patients.

made in the comprehensive management of patients with LSCC, the 5-year overall survival (OS) rate had reduced over the past 50 years, owing to the insidious symptoms and lack of effective diagnostic methods for early stages. 2 Approximately 60% of the patients with LSCC are diagnosed at an advanced stage or even with metastasis of the lymph node, and have had a poor prognosis. 3 Recently, more and more evidence has demonstrated that the prognosis and chemotherapeutic efficacy of patients with tumor was largely affected by the infiltration of the immune cells. [4][5][6][7][8] Therefore, the identification of novel potential biomarkers and their correlation with immune infiltration will potentially improve outcomes and direct individualized treatment for patients with LSCC.
Integrins, consisting of α and β subunits, are heterodimeric transmembrane proteins that function as surface adhesion receptors in different cells or cells and extracellular matrix communication. 9 Integrin α5 (ITGA5) is a member of the integrin alpha chain family and is located on chromosome 12q13.13 in Homo sapiens. 10 ITGA5 mainly recognizes its specific ligand through the formation of an α5β1 heterodimer and combines with integrin β1 (TGFβ1). 11 Increasing evidence has proved that ITGA5 is highly expressed in multiple cancers and is related to the progressions of tumor, such as cell proliferation, migration, and invasion. For instance, ITGA5 was upregulated in pancreatic ductal adenocarcinoma. Silencing ITGA5 could suppress the differentiation of human pancreatic stellate cells and reduce desmoplasia. 9 In oral cancer, ITGA5 facilitates cell growth, migration, and invasion by epithelial-mesenchymal transition. 12 According to Yu et al.,11 ITGA5 was overexpressed in colorectal cancer and promoted cell proliferation and tumorigenesis and reduced cell apoptosis. Additionally, ITGA5 was shown to maintain chemotherapeutic resistance 9,13 and cancer cell stemness. 14 Russano et al. 15 reported that ITGA5 acts as a mediator of breast-to-bone metastasis and might improve the efficacy of volociximab/M200 in patients with breast cancer along with bone metastasis. These findings implied that ITGA5 may be a potential therapeutic target for tumor angiogenesis, lymphangiogenesis, and metastasis. However, the expression and potential functions of ITGA5 in the progression of LSCC remain largely unknown.
In this study, we explored the expression and prognostic value of ITGA5 in patients with LSCC, and then performed functional enrichment analysis. More importantly, we revealed the relationship between the expression of ITGA5 and tumor immune microenvironment (TME) as well as the infiltration of the immune cells in LSCC. Our results revealed that ITGA5 was highly expressed in LSCC and might appear as an independent prognostic factor and therapeutic target for LSCC.

| Data preparation
The Cancer Genome Atlas (TCGA) database (https://portal.gdc. cancer.gov/) was used to download the ITGA5 mRNA expression (FPKM form) and the corresponding clinicopathological parameters of 111 LSCC tissues and 12 normal tissues. GSE51985 (including 10 LSCC tissues and 10 adjacent normal tissues), GSE59102 (including 29 LSCC tissues and 13 normal tissues), and GSE84957 (including 9 LSCC tissues and 9 adjacent normal tissues), and the corresponding platform annotation files were downloaded from the

| The expression of ITGA5 and its association with the clinical characteristics of patients with LSCC
"Limma" R package was applied to estimate the expression of ITGA5 between LSCC and non-tumor samples in the TCGA cohort.
Further, the association between the expression of ITGA5 and the clinical characteristics was also explored in patients with LSCC.

| Prognostic analysis of ITGA5 in patients with LSCC
According to the optimal cut-off point by routing algorithms, the OS and the RFS curves were plotted by the Kaplan-Meier (K-M) method, and the difference in the survival rates between high-ITGA5 and low-ITGA5 groups was assessed using the log-rank test. Univariate and multivariate Cox regression analyses were used to evaluate the prognostic value of ITGA5 in LSCC. Further, a nomogram was constructed based on the expression of ITGA5 and the clinical features to predict the survival probability. The 1-, 3-, and 5-years calibration curves and receiver operating characteristic (ROC) curves were used to assess the prediction performance of the nomogram.

| Functional enrichment analysis
Patients with LSCC from the TCGA were separated into the high-ITGA5 and the low-ITGA5 groups according to their median value.
The Gene Set Enrichment Analysis (GSEA, version 4.1.0) was performed in our previous study. 16 The c2.cp.kegg.v7.2.symbols.gmt was utilized as reference gene sets. Pathways with false discovery rate (FDR) p < 0.05 were considered statistically enriched.
Differentially expressed genes (DEGs) between the high-ITGA5 and the low-ITGA5 groups were identified with FDR p < 0.05 and

Unknown 13
Radiation therapy

Unknown 23
Targeted therapy

| TME and the infiltration of the immune cells
The ESTIMATE algorithm was used to calculate the immuneScore, stro-malScore, and estimate score of the LSCC samples. 17 The CIBERSORT algorithm was adopted to quantify the absolute abundance of 22 tumor-infiltrating immune cell types in the LSCC samples. 18 Then, the difference in the TME score and the infiltration of the immune cells between the high-ITGA5 and the low-ITGA5 groups were analyzed.

| Chemotherapeutic drug sensitivity analysis
In this study, we demonstrated whether ITGA5 could influence the clinical response to treatment and could be considered as a potential biomarker for drug screening. The half-maximum inhibitory concentration (IC50) was estimated using the "pRRophetic" package 19 to predict the chemotherapeutic response of four commonly used chemotherapeutic drugs (paclitaxel, gemcitabine, docetaxel, and cisplatin) in the high-ITGA5 and the low-ITGA5 groups.

| Statistical analysis
The R statistical language (version 4.

| Expression of ITGA5 in LSCC
By data mining using the TCGA database, ITGA5 was determined to be elevated in LSCC compared with the normal tissues ( Figure 1A,  Figure 1B, Figure 1C, p < 0.001], and GSE84957 [ Figure 1D, p = 0.001]). Subsequently, we analyzed the correlations between the expression of ITGA5 and the pathological parameters in LSCC (Figure 2A). Both Chi square test ( Figure 2B, p = 0.017) and Wilcoxon signed rank test ( Figure 2C, p = 0.005) showed that the dead patients had an elevated expression of ITGA5 than the alive patients. However, no correlation between the expression of ITGA5 and other clinical characteristics was found.

| ITGA5 is associated with poor prognosis of patients with LSCC
The K-M curves and the log-rank test showed that patients in the high-ITGA5 group had worse OS than the patients in the low-  Table S1) between the high-ITGA5 and the low-ITGA5 groups. The results of the GO analysis ( Figure 6C, D) and the KEGG analysis ( Figure 6E, F) confirmed that the DEGs were related to the construction of the ECM components and structures.

| Association between ITGA5 and TME as well as the infiltration of the immune cells in LSCC
According to the TME analysis, using the ESTIMATE algorithm, the results showed that patients with high-ITGA5 had elevated stromalScore than patients with low-ITGA (Figure 7, p < 0.01), indicating that proportional to M0 macrophages ( Figure 8B) and resting CD4 memory T cells ( Figure 8C) as well as resting natural killer (NK) cells ( Figure 8D), and inversely proportional to CD8 T cells ( Figure 8E), follicular helper T cells ( Figure 8F) as well as plasma cells ( Figure 8G).

| Chemosensitivity analysis
To explore the potential role of ITGA5 in the clinical treatment strategy, we estimated the IC50 of four common chemotherapeutic drugs recommended for laryngeal cancer therapy. Patients with low expression of ITGA5 had higher IC50 of docetaxel ( Figure 9A) and gemcitabine ( Figure 9B), indicating that patients with high expression of ITGA5 were more likely to respond to these two chemotherapeutic drugs. However, there was no difference in IC50 between the high-ITGA5 and the low-ITGA5 groups for cisplatin ( Figure 9C) and paclitaxel ( Figure 9D) therapies.

| DISCUSS ION
Laryngeal squamous cell carcinoma is the second most common malignant tumor of the upper respiratory system as well as the head and neck. 20 The current treatment for LSCC is a comprehensive protocol that is based on surgical operation combined with chemotherapy, radiotherapy, targeted therapy, and immunotherapy. 21 However, LSCC is a heterogeneous disease, resulting in the treatment outcome to be significantly different. Therefore, more biomarkers for prediction and risk stratification are needed urgently for patients with LSCC. ITGA5 is an important member of the integrin family, which acts as an adhesive receptor for the ECM proteins.
Recent research indicates that ITGA5 is overexpressed in various types of tumors and promotes proliferation, migration, invasion, and metastasis of cancer cells. 14,[22][23][24] However, the role of ITGA5 in LSCC remains unclear. Hence, we investigated the expression profile and clinical significance of ITGA5 in LSCC by analyzing the TCGA and the GEO databases.
In this study, we combined results from the TCGA database with three independent datasets from the GEO database as validation cohorts, and confirmed that the expression levels of ITGA5 were upregulated in LSCC, which was consistent with the findings of previous studies in other cancers. 12,24 After the analysis of the clinicopathological data, it was found that the expression level of ITGA5 was correlated with the survival status of patients with LSCC. Breuksch et al. also indicated ITGA5 as a potent biomarker to discriminate the prognosis of patients with renal cell carcinoma. 25 Further, we evaluated the prognostic value of ITGA5 in LSCC. Both TCGA cohort and GSE65858 dataset showed that high expression of ITGA5 was strongly associated with worse OS in patients with LSCC. Regional recurrence was an important factor associated with the prognosis of patients with LSCC. 26 Interestingly, we found the expression level of ITGA5 was significantly associated with poor RFS in the TCGA cohort, which was validated by GSE27020 and GSE65858 datasets.
Moreover, the univariate and multivariate Cox regression analyses demonstrated that ITGA5 was an independent predictive factor for death and recurrence in patients with LSCC. Collectively, all the above results strongly supported that the expression of ITGA5 is a perfect prognostic biomarker for LSCC. Therefore, we developed a nomogram based on the expression of ITGA5, which can be used to precisely predict the survival rate of patients with LSCC. Calibration curves and ROC curves confirmed the accuracy of the nomogram prediction, which was superior to any single factor. However, in the near future, further studies will be needed to validate the nomogram using a larger sample size.
To further explore the potential mechanisms that might be me- Conversely, previous studies have suggested that the infiltration of the CD8 T cells was positively correlated to improve prognosis and survival in several cancers, including melanoma, 31 lung adenocarcinoma, 32 and cutaneous angiosarcoma. 33  In conclusion, our research was the first to identify ITGA5 as an independent potential of OS and an RFS predictor for patients with LSCC. We also revealed the potential biological functions and

CO N FLI C T O F I NTE R E S T
The authors declare that there are no financial or other conflicts of interest associated with this study.

AUTH O R CO NTR I B UTI O N S
The research was designed by CZ and HD. The data analysis from the public database was made by CZ, ZW, MT, JZ, and YS. The data were analyzed by CZ, ZS, and YS. CZ supervised the writing of the draft. The manuscript was reviewed by all the authors.