Prognostic value and therapeutic potential of IAP family in head and neck squamous cell carcinoma

Aberrant expression of the IAP family members in HNSCC

Using the TIMER2.0 database, we examined IAP transcript levels to determine if there were any variations in their expression levels between normal and tumour samples. Notably, the mRNA expression levels of XIAP, BIRC3, BIRC2, NAIP, and BIRC5-7 were markedly upregulated in HNSCC tissues. However, BIRC8 expression showed no statistically significant differences (Figure 1A). The University of Alabama at Birmingham Cancer Data Analysis (UALCAN) portal (https://ualcan.path.uab.edu) provides easy access to precalculated gene and protein expression based on tumour subgroups. This comprehensive web resource was employed to delve deeper into the analysis of the IAP family genes’ mRNA expression. In comparison to normal tissues, HNSCC tissues exhibited a significant upregulation of all IAPs except BIRC8 (P < 0.05) (Figure 1B).

We verified these findings by delving deeper into the HPA database for immunohistochemistry results from the IAP family. Normal tissues had lower protein levels of NAIP, BIRC2/3/5/6, and XIAP compared to HNSCC (Figure 2A–2F). Our previous findings on the IAP family mRNA expression levels are corroborated by these results.

Correlation of IAP family expression with clinicopathologic features of HNSCC

Next, we aimed to determine whether the IAP family members’ expression levels were linked to staging, grading, and lymph node metastasis of HNSCC. Seven IAP family genes (except BIRC8) exhibited elevated mRNA levels in the tumour stage 1–4 subgroups relative to normal tissue, according to the findings. NAIP, BIRC2/5, and XIAP expression were related to the individual cancer stages of HNSCC (Figure 3A). The BIRC2, XIAP, and BIRC5 mRNA expression levels were markedly linked to the individual HNSCC cancer grades as shown in Figure 3B. We found a variable degree of correlation between NAIP, BIRC2/3/5/6, and XIAP, and lymph node metastasis in HNSCC (Figure 3C). Based on these results, IAPs (particularly BIRC2/5 and XIAP) contribute to the onset and progression of HNSCC.

Prognostic value of IAP gene family in patients with HNSCC

IAP family members’ predictive significance was evaluated in patients with HNSCC depending on their mRNA expression. This analysis was performed via the Kaplan–Meier (KM) plotter database. To ascertain the clinical prognostic outcome, survival analysis was performed, involving both recurrence-free survival (RFS) and overall survival (OS). According to the analysis, elevated mRNA levels of BIRC5 [OS: hazard ratio (HR) = 1.33 (1.01–1.76), P = 0.041] and BIRC8 [OS: HR = 1.39 (1.05–1.84), P = 0.021] were linked to a dismal OS in patients with HNSCC (Figure 4A). BIRC2 upregulation was linked to unfavorable OS; Still, the variation did not reach statistical significance. Increased expression of NAIP [RFS: HR = 2.14 (1.01–4.55), P = 0.043], BIRC2 [RFS: HR = 3.81 (1.53–9.49), P = 0.0021], and BIRC7 [RFS: HR = 2.09 (0.99–4.43), P = 0.048] was related to unfavorable RFS in HNSCC patients (Figure 4B).

Genetic alteration and functional analysis of the IAP family in HNSCC individuals

Mutations are integral to cancer development, and some of these mutations can create specific vulnerabilities [18]. We investigated the genetic alterations in individual IAP family members using the cBioPortal dataset. Alterations were seen in all eight members of the IAP family among patients diagnosed with HNSCC, with respective alteration rates of 5, 14, 9, 11, 4, 14, 3, and 2 %. Additionally, mRNA alterations constituted the most prevalent forms of aberrations within the IAP family, with amplifications and deep deletions ranking second and third, respectively (Figure 5A). Next, cBioPortal was utilised to identify co-expressed genes with a threshold of |log2 fold-change| ≥ 0.5 and P < 0.05 (Supplementary Table 1), and Cytoscape v3.9.0 was utilised to produce a visualisation of the co-expression network of crucial genes that are related to the IAP family (Figure 5B). Metascape was utilised to evaluate the biological activities of the IAP members and their co-expressed genes by means of Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis and Gene Ontology (GO) annotation. Using KEGG pathway analysis, the co-expressed genes were examined for their involvement in arachidonic acid metabolism, cytokine-cytokine receptor interactions, and the NF-kappa B signalling pathway (Figure 5C). The GO analysis illustrated the co-expression of genes primarily linked to the sarcomere, epidermal development, and cytokine-mediated signalling pathways (Figure 5D). Epidermal development, muscular contraction, and cytokine-mediated signalling pathways comprised the majority of the functions of these genes, according to bioprocess analysis. (Figure 5E). Additionally, the molecular function analysis revealed the involvement of these genes primarily in the structural components of the epidermis of the skin, serine peptidase activity, and triglyceride lipase activity (Figure 5F). The examination of cellular components demonstrated that these genes had a high degree of association with the cornified envelope and sarcomere (Figure 5G).

Association of expression of IAP family members with immune infiltration in HNSCC

The Tumour Immune Estimation Resource (TIMER) was employed to determine the link between immune cell infiltration and the IAP gene family (Figure 6). The results illustrated that NAIP expression was substantially linked to B-cell (cor = 0.314, P < 0.05), CD8⁺ T-cell (cor = 0.445, P < 0.05), CD4⁺ T-cell (cor = 0.657, P < 0.05), macrophage (cor = 0.576, P < 0.05), neutrophil (cor = 0.717, P < 0.05), and dendritic cell (cor = 0.675, P < 0.05) infiltration. There was a strong link between BIRC2 mRNA expression and the infiltrating levels of CD4⁺ T-cells (cor = 0.271, P < 0.05), macrophages (cor = 0.119, P < 0.05), neutrophils (cor = 0.235, P < 0.05), and dendritic cells (DCs) (cor = 0.199, P < 0.05). Significant associations were observed between BIRC3 expression and B-cell (cor = 0.356, P < 0.05), CD8⁺ T-cell (cor = 0.326, P < 0.05), CD4⁺ T-cell (cor = 0.432, P < 0.05), macrophage (cor = 0.342, P < 0.05), neutrophil (cor = 0.451, P < 0.05), and DCs (cor = 0.479, P < 0.05) infiltrates. Significant associations were observed between the levels of B-cells (cor = 0.116, P < 0.05), CD8⁺ T-cells (cor = 0.099, P < 0.05), CD4⁺ T-cells (cor = 0.422, P < 0.05), macrophages (cor = 0.188, P < 0.05), neutrophils (cor = 0.197, P < 0.05), and DCs (cor = 0.270, P < 0.05). There was a strong link between BIRC6 mRNA expression and the infiltrating levels of B-cells (cor = 0.169, P < 0.05), CD8⁺ T-cells (cor = 0.153, P < 0.05), CD4⁺ T-cells (cor = 0.474, P < 0.05), macrophages (cor = 0.287, P < 0.05), neutrophils (cor = 0.252, P < 0.05), and DCs (cor = 0.365, P < 0.05). Considerable associations were observed between BIRC7 expression and B-cell (cor = 0.225, P < 0.05), CD8⁺ T-cell (cor = 0.186, P < 0.05), CD4⁺ T-cell (cor = 0.280, P < 0.05), macrophage (cor = 0.426, P < 0.05), neutrophil (cor = 0.249, P < 0.05), and DCs (cor = 0.348, P < 0.05) infiltrates. The level of BIRC8 mRNA was substantially linked to B-cell (cor = 0.135, P < 0.05), CD8⁺ T-cell (cor = 0.183, P < 0.05), CD4⁺ T-cell (cor = 0.127, P < 0.05), macrophage (cor = 0.140, P < 0.05), neutrophil (cor = 0.118, P < 0.05), and DC (cor = 0.125, P < 0.05) infiltrates. However, immune cell infiltration did not exhibit a link to BIRC5 expression. In the HNSCC tumour microenvironment (TME), our data indicate that the IAP family members, including BIRC2, may affect immune response.

The marker types present in HNSCC were examined by the TIMER database, including DCs, CD8+ T lymphocytes, neutrophils, T helper 1 (Th1) cells, and tumour-associated macrophages (TAM). This study aimed to examine immune cell-IAP family expression relationships (Table 1). A strong association exists between the levels of immune cells and members of the IAP family, except for BIRC5. Moreover, NAIP was shown to be strongly linked to CD8+ T lymphocytes, B, and TAMs, M2 macrophages, neutrophils, DCs, natural killer cells (NK), Th1, Th2, Tfh, Th17, regulatory T cells (Tregs), exhausted T cells, and monocytes. Additionally, BIRC2 exhibited a good correlation with TAMs, Tregs, M2 macrophages, DCs, Th1, Th2, and monocytes. BIRC3 demonstrated a robust correlation with CD8+ T lymphocytes, B cells, TAM, M1/M2 macrophages, monocytes, DCs, NK cells, exhausted T cells, Th1 cells, regulatory T cells, Th2 cells, T follicular helper cells, Th17 cells, and neutrophils. XIAP exhibited a significant association with B cells, T cells, TAM, M1/M2 macrophages, neutrophils, DCs, NK cells, Th1 cells, Th2 cells, T follicular helper cells, Th17 cells, regulatory T cells, exhausted T cells, and monocytes. Moreover, BIRC6 demonstrated a positive relationship with TAMs, M1/M2 macrophages, neutrophils, DCs, NK cells, Th1 cells, Th2 cells, monocytes, exhausted T cells, Th17 cells, regulatory T cells, and T follicular helper cells. There is a significant connection between BIRC7 and CD8+ T lymphocytes, T cells, TAMs, exhausted T cells, neutrophils, DCs, B cells, NK, Th1, Th2, Tfh, Th17, Tregs, M1/M2 macrophages, and monocytes. Additionally, BIRC8 displays a strong correlation with T cells, CD8+ T lymphocytes, B cells, TAMs, DC, NK, and Th1 cells. These observations indicate that the IAP family members may have contributed to the immune infiltration in HNSCC.

The investigation revealed that BIRC2 might have a key role in promoting HNSCC progression. Additional experiments were carried out to validate the BIRC2 differential expression and its relationship with molecules linked to immune cells to elucidate its involvement in HNSCC. The BIRC2 expression in HNSCC samples was studied by qRT-PCR, and 12 HNSCC tissue specimens exhibited a much higher relative expression level of BIRC2 than corresponding nearby non-tumour tissue samples, according to the findings (P < 0.05; Figure 7A). Furthermore, we found a correlation between STAT1 and BIRC2, which is a marker gene for Th1 cells [19]. Relative to adjoining normal tissues, 12 HNSCC samples had a substantially higher relative expression level of STAT1, according to the qRT-PCR data (P < 0.05; Figure 7B). Lastly, we observed a strong positive correlation (R = 0.7086, P < 0.05; Figure 7C) after comparing the expression of BIRC2 and STAT1 in the 12 HNSCC samples.

The TME is the internal environment upon which tumour cells depend for survival. It comprises tumour cells, immune cells, chemokines, cytokines, and various other factors. TME is intimately linked with tumour growth, metastasis, immune evasion, and other behaviors. The main functions of Th1 cells are cellular immunity and inflammation. This includes activating other immune cells, like macrophages, B cells, and cytotoxic CD8+ T lymphocytes (CTL), promoting the targeted killing of infected cells, tumour cells, and abnormal cells, as well as inducing cell lysis and other effector functions. The IAP family may regulate macrophage polarisation in HNSCC patients, as certain members have a significant association with M2 macrophage markers. Consequently, the prognostic relevance of IAP family members in HNSCC patients with macrophage enrichment was explored. The findings indicate that in HNSCC patients, increased expression of BIRC2/5 negatively impacted the OS when there was macrophage enrichment (Figure 8A). In HNSCC patients with macrophage enrichment, higher expression of NAIP and BIRC2 was associated with poorer RFS (Figure 8B). These findings suggest that BIRC2 has a vital function in the infiltrating levels of the immune cells and the TME in HNSCC.

It was further determined whether aberrant IAP family expression affects immunotherapy response in HNSCC. As shown in Figure 9A, BIRC2 expression increased in anti-PD-1/CTLA-4 non-responders in the Riaz cohort (P = 0.021). The Riaz cohort’s area under the receiver operating characteristic curve (AUC) was 0.658, implying that BIRC2 may effectively distinguish between individuals who respond to anti-PD-1/CTLA-4 and those who do not (Figure 9B). This study established BIRC2 as a promising immunotherapeutic target for HNSCC treatment.

Patients’ tissue samples

Twelve pairs of paraffin-embedded archival specimens of HNSCC and corresponding adjacent normal tissue samples were prospectively collected from Xiangya Hospital (Changsha, People’s Republic of China). The cohort included diverse HNSCC subtypes, specifically, four pairs of oral squamous cell carcinoma, one pair of nasopharyngeal squamous cell carcinoma, three pairs of tongue carcinoma, and four pairs of laryngeal carcinoma, along with their corresponding adjacent tissues.

Exclusion and inclusion criteria

Inclusion and exclusion criteria were applied to identify eligible patients for this study. The inclusion criteria included: (1) histopathological confirmation of diagnosis; (2) no prior treatment with chemotherapy, radiation, or immunotherapy before resection; (3) the availability of complete data on clinical and pathological characteristics. The accompanying table offers a summary of the clinical parameters of the patients who were recruited. Exclusion criteria encompassed post-operative administration of alternative treatments, vital organ dysfunction, and the presence of secondary tumours in other organs.

Quantitative real-time polymerase chain reaction (qRT-PCR)

RNA extraction, amplification, and qRT-PCR were executed using established protocols. The thermocycling protocol included an initial denaturation at 95° C for 30 seconds, followed by 40 cycles of annealing and extension phases at 60° C and 72° C, respectively, each lasting 30 seconds. The sequences of the primers utilised in qRT-PCR are highlighted in Table 2.

TIMER2.0

TIMER2.0 (https://cistrome.shinyapps.io/timer/) encompasses three primary components: immunisation, exploration, and estimation. This interactive tool enables the analysis of gene-immune infiltrating cell correlations, expression comparisons of genes between normal and malignant cells in a variety of cancers, and other functionalities, offering user-friendly interactive visualisations to facilitate data exploration [37]. The present work utilised the TIMER2.0 database to ascertain the expression patterns of IAP family members. Specifically, we evaluated the expression levels of these genes in 44 normal and 520 HNSCC samples. Additionally, using TIMER, we assessed the link between IAP family members’ mRNA expression and immune infiltrating cells in HNSCC, including CD4+ and CD8+ T lymphocytes, dendritic cells (DCs), neutrophils, macrophages, and B cells. The correlation analysis employed Spearman’s algorithm while adjusting for tumour purity. A P value < 0.05 was set as the significant threshold.

GEPIA2

With a vast collection of 198,619 isoforms and covering 84 cancer subtypes, GEPIA2 (Gene Expression Profiling Interactive Analysis 2) provides an advanced platform for quantifying gene expression at the transcript level. It facilitates specific cancer subtype analysis and enables comparisons across different subtypes [38]. We utilised the GEPIA2 database to investigate the expression patterns of IAP family members within 44 normal samples and 520 samples of HNSCC.

Human protein atlas (HPA)

The HPA portal (http://www.proteinatlas.org) serves as a valuable platform for a plethora of biomedical research projects, providing comprehensive protein expression data using immunohistochemistry (IHC) across different cancer types [39]. We implemented an in-depth evaluation of the protein expression profiles of IAP family members in both normal and HNSCC tissues utilising IHC techniques.

UALCAN

UALCAN (https://ualcan.path.uab.edu) is an accessible and easily usable tool that offers readily available, pre-computed data for promoter DNA methylation status, protein/gene expression, and Kaplan-Meier (KM) survival analyses, organised by tumour subgroups [40]. Our stratified analysis focused on key patient-specific variables like cancer stage, tumour grade, and nodal status. Statistical significance was determined by the Student’s t-test, with a significance threshold of P < 0.05.

Kaplan–Meier plotter database

The KM plotter database (https://kmplot.com) was utilised as a valuable resource for conducting a meta-analysis, integrating clinical prognostic data with gene expression data to detect and confirm survival-related molecular markers [41]. We utilised the KM plotting approach to explore the link between RFS and OS with the IAP family members in HNSCC patients. The KM analysis was evaluated using the log-rank test, and a P < 0.05 was set as the significance threshold.

CBioPortal

The cBio Cancer Genomics Portal (http://cbioportal.org) allows users to interactively explore multidimensional cancer genomic information on a publicly available platform [41, 42]. We leveraged the capabilities of cBioPortal to obtain a dataset comprising 523 patients diagnosed with HNSCC. Subsequently, the IAPs were analysed using this dataset for co-expression and gene alterations.

STRING

The STRING database (https://cn.string-db.org/) serves as a comprehensive resource offering information regarding predicted and known protein-protein associations across a wide range of organisms. Physical interactions and functional linkages are both included in this comprehensive database, along with confidence levels indicating their reliability [43]. In our study, we harnessed the capabilities of the STRING database to evaluate the correlations among IAP genes.

Cytoscape

Cytoscape (http://cytoscape.org) is a publicly accessible program that aims to build biomolecular interaction networks that can be easily integrated with other molecular states and high-throughput expression data [44]. Within the scope of this study, Cytoscape was utilised to enable the functional integration of 156 frequently altered genes belonging to the IAP family, which were identified through screening the cBioPortal database. The degree values, reflecting the interactions between these proteins, were visually represented by the size of the nodes in the network, with larger circles denoting proteins exhibiting higher degrees of connectivity.

Metascape

Membership search, gene annotation, interactome analysis, and functional enrichment, are all synergistically integrated into Metascape (https://metascape.org), an all-inclusive and integrated portal. It harnesses the collective knowledge from over 40 independent databases to provide a robust analytical framework [45]. In this study, we utilised GO and KEGG pathway enrichment analyses to ascertain the functional implications of the IAP family.

Biomarker Exploration of Solid Tumours (BEST)

The study’s findings were validated by utilising the BEST portal, an established and reputable platform for comprehensive biomarker analysis in the field of oncology research (https://rookieutopia.com/app_direct/BEST/). Patients with HNSCC were studied using BEST to determine whether there was a link between the IAPs and responsiveness to immunotherapy as well as the prognostic outcomes.

RNA isolation from the Formalin-Fixed and Paraffin-Embedded (FFPE) samples

FFPE colon cancer or normal tissue samples were subjected to deparaffinisation using xylene. Total RNA isolation from these samples was performed using the AmoyDx® FFPE RNA Extraction Kit for total RNA (Cat. # 8.02.0019, AmoyDx, Xiamen, P. R. China).

Statistical analysis

Survival analysis was conducted via the log-rank test, while Spearman’s correlation was employed to evaluate the connection between the IAP family and indicators of immune cell type and immunological infiltration. The two independent samples were contrasted utilising Student’s t-test. P < 0.05 was set as the significance criterion.

Data availability statement

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.