Potential roles of Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4) in head and neck squamous cell carcinoma

Abstract

BACKGROUND:

It is of great clinical significance to discover novel biomarkers for neck squamous cell carcinoma (HNSCC) treatments. We discovered a potential cancer-related gene, Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4), that can be a biomarker for HNSCC.

METHODS:

We access multiple open databases and analyzed bulk mRNA-sequencing, protein staining, and single-cell mRNA-sequencing data of HNSCC and investigated the diagnostic and prognostic value of CNIH4 in HNSCC. The potential association between CNIH4 and the immune microenvironment of HNSCC was also estimated.

RESULTS:

CNIH4 was significantly up-regulated in HNSCC compared with non-cancer tissues. Higher CNIH4 resulted in a shorter overall survival time and we further constructed a survival nomogram for clinical applications. 2012 and 421 genes were identified as positive and negative differentially expressed genes of CNIH4 in HNSCC respectively. These genes were mostly mapped to “Cell cycle”, “DNA replicate”, “Cytokine-cytokine receptor interaction” KEGG pathways. Functions associated with CNIH4 were “stemness”, “cell cycle”, and “DNA repair” in single-cell data. CNIH4 potentially affected immune cell infiltration levels and cancer immune therapy.

CONCLUSION:

CNIH4 is a potential diagnostic and prognostic biomarker associated with cancer stemness and immunity in HNSCC.

Keywords

CNIH4 HNSCC cancer stemness;immunity

List of abbreviation

CNIH4: Cornichon Family AMPA Receptor Auxiliary Protein 4

HNSCC: head and neck squamous cell carcinoma

TCGA: The Cancer Genome Atlas

TIDE: Tumor Immune Dysfunction and Exclusion

ICB: immune checkpoint blockade

OCLR: one-class logistic regression machine learning algorithm

TISCH: Tumor Immune Single-cell Hub

UMAP: Dimensionality reduction via Uniform Mani- fold Approximation and Projection

1. Introduction

As one of the most influential cancer types in the world, head and neck squamous cell carcinoma (HNSCC) results in five hundred thousand new cancer cases each year and causes approximately three hundred thousand deaths [1, 2]. The most common clinical management of HNSCC includes chemotherapy, radiotherapy, and surgery [3]. Over the past ten years, although several novel therapeutic methods for HNSCC patients have been developed, the HNSCC-related death number is still high. The challenges in the early diagnosis and precise prognosis of HNSCC patients prevented the treatments of this disease. The delay of an in-time therapy and the following lymph node metastasis lead to a much higher death rate [4, 5]. Therefore, the discovery of novel biomarkers of HNSCC for clinical HNSCC diagnosis and prognosis is urgent in the field.

Our preliminary screening revealed a potential cancer-related gene, Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4), which encoded a protein interacting with newly synthesized G protein-coupled receptors (GPCR) and regulates their export from the endoplasmic reticulum [6]. CNIH4 has been reported to play potential roles in cancer. A study revealed that CNIH4 could promote colon cancer metastases [7], while in breast cancer, CNIH4 was found to be one of the cancer marker genes [8]. CNIH4 was also identified as one of the genes in the five gene signatures for prognostic of hepatocellular carcinoma patients [9]. However, the clinical values of CNIH4 in many other cancer types remain largely unknown and the roles of CNIH4 in HNSCC required further investigation.

Here, we investigated the value of CNIH4 as a diagnostic and prognostic biomarker for HNSCC. We also proposed that CNIH4 is a regulator for the cancer cell biology and tumor immune microenvironment in HNSCC. This study provided a clinical molecular marker for the improvement of clinical HNSCC treatment.

2. Methods

2.1 RNA-seq and clinical data acquisition

Tissue samples expression data were downloaded from TCGA (The Cancer Genome Atlas) HNSCC cohort in January 2020 following the rules and policies.

2.2 RNA-seq data analysis

R foundation for statistical computing (2020) version 4.0.3 and the ggplot (v3.3.2) was used for the expression and survival analysis and the plotting.

2.3 Oncomine analysis

The expression and copy number of CNIH4 were also analyzed using the ONCOMINE [10], a cancer microarray database and integrated data-mining platform. All the data mining and differential analyses were done with the ONCOMINE platform. Detailed analysis methods can be found in this paper [10].

2.4 Immune cell analysis

The immune cell infiltration levels were calculated using the xCell algorithms [11]. R foundation for statistical computing (2020) version 4.0.3 and the ggplot (v3.3.2) was used for the plotting.

2.5 Stemness analysis

The one-class logistic regression machine learning algorithm (OCLR) [12] algorithm was used to calculate the mRNAsi (stemness index based on mRNA expression) score for the evaluation of cancer stemness.

2.6 Immune therapy analysis

Potential immune checkpoint blockade (ICB) response prediction was calculated using the Tumor Immune Dysfunction and Exclusion (TIDE) [13] algorithm with the TCGA HNSCC cohort.

2.7 Immunohistochemistry staining

The representative images of CNIH4 were accessed from HPA (Human Protein Atlas) [14]. HPA044268 antibody was used for protein staining. The sample details and the general pathological annotations and results were provided by HPA.

2.8 Single-cell sequencing data acquisition and analysis

The CancerSEA [15] was used to analyze the functional correlation to the CNIH4 expression. Single-cell datasets GSE103322 [16] (5902 single cells from 18 patients with oral cavity tumors, including five matched pairs of primary tumors and lymph node metastases) were analyzed. The TISCH (Tumor Immune Single-cell Hub) [17] was used to analyze the expression of CNIH4 in immune cells and malignant cells. Data sets GSE103322 [16] and GSE139324 [18] (130,721 single cells from peripheral and intra-tumoral CD45 $+$ immune populations from 26 HPV $-$ and HPV $+$ HNSCC patients.) were analyzed. The UMAP (Dimensionality reduction via Uniform Manifold Approximation and Projection) [19] method was used to display the data. Cell-type annotations were provided by TISCH.

2.9 Statistical significance

Results were considered statistically significant at $P<$ 0.05 (*).

3. Results

3.1 The expression of CNIH4 in HNSCC

In this study, we compared the CNIH4 expression in HNSCC and non-cancer tissues. Data suggested that the level of CNIH4 in cancer was up-regulated compared with non-cancer tissues (Fig. 1A1). In high-grade HNSCC, tissues had a significantly higher expression of CNIH4 than low-grade tissues (Fig. 1A2). HNSCC had a significantly higher expression of CNIH4 in male over female patients (Fig. 1A3). The pTNM IV had significantly higher CNIH4 than pTNM I, but the other comparison had no significant difference (Fig. 1A4). Results showed no significance among pT stagings or pN stagings (Fig. 1A5–6). HNSCC had a slightly higher expression of CNIH4 in white over black (Fig. 1A7) and in smoke patients over non-smoke patients (Fig. 1A8). The limit in pack years for smokers is 20 pack years. In addition, the cancer-noncancer paired comparison also supported the conclusion that CNIH4 was up-regulated in cancer over non-cancer (Fig. 1B). The clinical relevance of the high (50–100%, red) and low (0–50%, blue) CNIH4 groups were also analyzed and displayed in a table (Table 1). To observe the up-regulating protein level of CNIH4 in HNSCC, we accessed and observed the protein expression of CNIH4 in tissue staining. Representative images revealed that HNSCC tissues had a higher staining signal of CNIH4 over normal oral mucosa (Fig. S1). However, these are just representative images of the staining for future reference only, thus, we do not have any statistical analysis and quantitative analysis to show the difference. Moreover, to further validated the up-regulation of CNIH4 in HNSCC, we explored the expression and copy number of CNIH4 in HNSCC and normal tissues across 27 data sets. The expression or copy number of CNIH4 in HNSCC was up-regulated in normal tissues in 20 of the 27 data sets included. The overall statistical analysis demonstrated that the mRNA expression or copy number of CNIH4 in HNSCC was up-regulated compared with normal tissues. (Fig. S2).

Figure 1.

The expression of CNIH4 in HNSCC. (A1) The mRNA expression of CNIH4 in HNSCC and normal head and neck tissues. TCGA and GETx data were plotted. (A2–A8) The mRNA expression of CNIH4 in HNSCC of different clinicopathological groups. (B) The mRNA expression of CNIH4 in HNSCC and normal head and neck tissues. TCGA paired data were plotted, and cancer and non-cancer samples were linked. Paired $t$ -test was used to analyze the difference.

Table 1

Distribution of HNSCC patients with different clinicopathological variables in CNIH4 high and low groups

Term	AIMP1 high	AIMP1 low	$P$ -value
Alive	122	162
Dead	129	89	$<$ 0.01
Mean (SD)	61.9 (11)	60.2 (12.7)
Median [min, max]	61 [26, 87]	60 [19, 90]	0.103
Female	61	73
Male	190	178	0.19
Asian	3	7
Black	31	16
White	206	222	0.074
T1	16	18
T2	65	79
T3	65	68
T4	16	9
T4a	78	74
T4b	2	1
TX	9	2	0.21
N0	121	120
N1	38	43
N2	9	10
N2a	10	8
N2b	35	41
N2c	21	20
N3	6	1
NX	11	8	0.644
M0	234	243
M1	3	2
MX	14	6	0.168
I	8	17
II	39	42
III	42	48
IVA	151	139
IVB	10	3
IVC	1	2	0.381
G1	19	43
G2	146	154
G3	73	46
GX	11	5
G4		2	$<$ 0.01
Metastasis	11	8
Primary	6	3
Recurrence	24	15	0.904
Non-smoking	44	67
Smoking	203	178	0.015
Non-radiation	29	33
Radiation	66	55	0.401
Neoadjuvant	7	3
No neoadjuvant	244	248	0.338
Chemotherapy	83	76

3.2 Prognostic value of CNIH4 in HNSCC

To investigate the association between CNIH4 and the survival of HNSCC patients, we conducted a Kaplan-Meier plotting and log-rank test to compare the overall survival of the CNIH4 high (50–100%, red) and low (0–50%, blue) patients. Results showed that the Hazard ratio (HR) of CNIH4 high was 1.5 (95% CI $=$ 1.152–1.983) and the median survival time for the CNIH4 high and low patients was 3 and 5.2 respectively (Fig. 2A). These data demonstrated that CNIH4 potentially affected the overall survival of HNSCC patients. To develop the prognostic value, we calculated and constructed a clinical survival nomogram for HNSCC patients. CNIH4 expression, age, gender, grade, and race were taken into the calculation as factors. Univariate Cox regression analysis results identified the expression of CNIH4 and age as survival-associated factors (Fig. 2B). Multivariate Cox regression supported that the expression of CNIH4 and age provided independent prognostic power in the prediction of overall survival (Fig. 2C). Based on the Cox regressions, a nomogram including the expression of CNIH4 and age as factors was established, with a C-index of 0.602 (Fig. 2D). The prediction results of the nomogram calibration curves were generally consistent with patients’ observations (Fig. 2E).

Figure 2.

The prognostic value of the CNIH4 in HNSCC. (A) Kaplan-Meier plot of the high (50–100%, red) and low (0–50%, blue) CNIH4 groups. Log-rank analysis results were shown. (B) The overall survival univariate Cox regression of HNSCC. (C) The overall survival multivariate Cox regression of HNSCC. (D) Nomogram for the overall survival of HNSCC. (E) Calibration plots of the nomogram.

3.3 CNIH4 associated genes enrichment analysis.

To investigate the potential functional association of CNIH4 in HNSCC, we analyzed differentially expressed genes (DEGs) between CNIH4 high (75–100%) and low (0–25%) groups. The analysis identified 2012 positive DEGs and 421 negative DEGs of CNIH4 in HNSCC (Fig. 3A–B). These DEGs were further investigated in the GO terminologies and the KEGG pathways enrichment analysis. The positive DEGs were found associated with the “Cell cycle”, while the negative DEGs were associated with the “Cytokine-cytokine receptor interaction” KEGG pathways. In GO enrichment, positive DEGs were associated with “organelle fission”, “nuclear division”, “mitotic nuclear division”, “chromosome segregation”, and “DNA replication”, while negative DEGs were associated with “epidermis development”, “skin development”, “epidermal cell differentiation”, “keratinization”, and “keratinocyte differentiation” (Fig. 3C). Other enriched terms were immune-associated, such as “T cell activity”. Generally, these DEGs were found associated with cancer growth and immune cells, indicating that CNIH4 might affect HNSCC stemness and immunity.

Figure 3.

The differentially expressed genes (DEGs) enrichment analysis. (A) Volcano plots of CNIH4 DEGs. (B) Heat map and hierarchical clustering analysis of the DEGs. (C) The GO and the KEGG pathways enrichment analysis of CNIH4 DEGs in HNSCC.

3.4 CNIH4 was associated with stemness of HNSCC

Figure 4.

Association of CNIH4 and stemness. (A) Correlations between CNIH4 expression and functional scores in HNSCC single-cell datasets. The GSE103322 data ( $n=$ 2105) were accessed and analyzed using the CancerSEA. The coefficients and p-value were presented. ${}^{***}p<$ 0.001; ${}^{**}p<$ 0.01; ${}^{*}p<$ 0.05. (B) Scatter plot of the significant correlations between CNIH4 expression and functional states. Grey points were not considered to compute the correlations. (C) The distribution of mRNAsi scores in TCGA HNSCC CNIH4 high (75–100%) and low (0–25%) groups. The OCLR algorithm was used to calculate the mRNAsi scores for the evaluation of stemness.

To further study the functional association of CNIH4 in HNSCC, we accessed and analyzed an HNSCC single-cell mRNA-sequencing data set, which included 2150 single HNSCC cells. The correlations of CNIH4 and cancer functional scores were calculated. The calculation showed that 12 functional scores were significantly correlated with the expression of CNIH4. Specifically, the expression of CNIH4 was positively correlated with scores of stemness, cell cycle, DNA repair, invasion, and proliferation with coefficients of 0.26, 0.23, 0.13, 0.06, and 0.05 respectively, and was negatively correlated with angiogenesis, quiescence, metastasis, hypoxia, inflammation, DNAdamage, and differentiation with coefficients of $-$ 0.16, $-$ 0.16, $-$ 0.10, $-$ 0.10, $-$ 0.08, $-$ 0.07, and $-$ 0.06 (Fig. 4A–B). Although these correlations were significant, none of these correlations were over 0.4, thus, none of these correlations were strong. Yet, these correlations can still provide a hint of the potential effects of CNIH4 on HNSCC. Because stemness and cell cycle were correlated with the level of CNIH4, we suggested that CNIH4 was closely associated with cancer stemness in HNSCC. Thus, to further demonstrate the cancer stemness association of CNIH4, we calculated the stemness of HNSCC from TCGA bulk data and compared the CNIH4 high (75–100%) and low (0–25%) groups. The bulk data suggested that the CNIH4 high group was associated with a higher cancer stemness compared with the CNIH4 low group (Fig. 4C).

3.5 CNIH4 was associated with immunity regulation of HNSCC

Figure 5.

The expression of CNIH4 in different types of cells in HNSCC. The data were accessed and analyzed using the TISCH. (A) The average expression of CNIH4 in different cell types across datasets (Heatmap). (B1–B2) HNSCC_GSE103322. (C1–C2) HNSCC_GSE139324. The gene expression level was displayed using the UMAP method.

Figure 6.

The effect of CNIH4 on the immunity in HNSCC. (A) Comparison of immune cell infiltration levels of high CNIH4 (75–100%) vs. low CNIH4 (0–25%) groups. The xCell algorism was used to calculate the levels of immune cell infiltration levels. The significance was calculated using the Wilcox test. ${}^{*}p<$ 0.05, ${}^{**}p<$ 0.01, ${}^{***}p<$ 0.001. (B) The expression of four immune checkpoint genes in the high CNIH4 (75-100%) and low CNIH4 (0–25%) groups. (C) Immune checkpoint blockade (ICB) analysis. CNIH4 low (0–25%) and high (75–100%) groups were compared. Potential ICB response was predicted using the TIDE algorithm.

Another point found by the enrichment study was the potential association of CNIH4 and immune cells in HNSCC. To estimate this association, we analyzed the distribution of CNIH4 in different cell types in HNSCC. Single-cell seq data suggested that, in HNSCC, malignant cells expressed much higher CNIH4 than immune cells (Fig. 5), indicating that the expression of CNIH4 in HNSCC bulk data was mainly dependent on the expression of CNIH4 in malignant cells. In addition, we analyzed the immune cell infiltration levels using TCGA bulk data. The immune cell infiltration levels between the CNIH4 high (75–100%) and low (0–25%) groups were compared. Results showed that immune cells up-regulated in the CNIH4 high group included common lymphoid progenitor and T cell CD4 $+$ Th2. However, the CNIH4 high group was significantly lower at levels of T cell CD4 $+$ central memory, Endothelial cell, Myeloid dendritic cell activated, Myeloid dendritic cell, Plasmacytoid dendritic cell, T cell CD8 $+$ , T cell CD8 $+$ central memory, B cell plasma, B cell, B cell memory, T cell CD4 $+$ naïve, Class-switched memory B cell, Monocyte, Neutrophil, Mast cell, and T cell CD4 $+$ effector memory (Fig. 6A). These results suggested that CNIH4 was negatively associated with infiltration levels of most immune cells and reduced immune score and stroma score in HNSCC. Besides, we calculated the correlation between the expression of CNIH4 and four common immune checkpoints, CTLA4, LAG3, PDCD1, and TIGIT, and we found that the expression of CNIH4 was negatively correlated with the expression of all the four immune checkpoints (Fig. 6B). To further evaluate the potential effects of CNIH4 on clinical immune therapy, ICB responses of CNIH4 low (0–25%) and high (75–100%) groups were calculated and compared. The calculation demonstrated that the CNIH4 low group had a significantly higher TIDE score compared with the CNIH4 high group (Fig. 6C bottom), and only 28 out of 98 (22.22%) HNSCC patients were predicted to respond to the ICB treatment in the low CNIH4 group, while 84 out of 126 (66.67%) were predicted to respond to the ICB treatment in the high CNIH4 group (Fig. 6C top).

4. Discussion

This study reported that CNIH4 was overexpressed in HNSCC compared to normal tissues. Data revealed that this overexpression might result from the increased gene copy number in HNSCC, but other factors such as transcriptional factors or methylation can also contribute to the overexpression of CNIH4 in cancer. In addition, we found that the expression of CNIH4 was higher in the higher grade HNSCC, hence, we proposed that CNIH4 was a potential prognostic biomolecule in HNSCC. A previous study reported that CNIH4 was upregulated in colon cancer [7], which was consistent with our initial proposal that cancer tissues might express higher CNIH4 than their corresponding normal tissues. Therefore, we believe that CNIH4 has a similar cancer-noncancer expression pattern and role in different tumor types. However, to date, the expression cancer-noncancer pattern of CNIH4 in most other cancer types has not been studied.

Figure 1 and Table 1 are different in comparison and statistical analysis. Figure 1 compared expression levels in different groups and used a $T$ -test type analysis. Table 1 compared the distribution of patients within the expression group and used chi-square distribution analysis. Therefore, they might come out with different $p$ -value. If the significance is not very remarkable, they might not agree with each other on the significance. We provided both of the comparisons because they together provided a comprehensive view for the readers. In addition, some of the groups are not in Fig. 1 because they might have a too low case number, but we provided these data in table1, such as the M0-M1 comparison.

One of the most striking findings of this study was that CNIH4 expression potentially affects the overall survival of HNSCC patients. The Cox regression analysis indicated that CNIH4 was a risky factor for HNSCC patients and was independent of patients’ age, gender, grade, and race. Although a previous study suggested that CNIH4 was valuable in the prognosis of hepatocellular carcinoma patients [9], the prognostic value of CNIH4 for HNSCC patients has never been reported. We suggested that most of the poor survival was caused by stemness or immune alteration or other aspects rather than by metastasis. Based on Table 1 M0-M1 analysis, most of the patients analyzed were at M0 which meant no distant cancer spread has been found. In this study, we constructed a nomogram to demonstrate the potential clinical practice of CNIH4 for HNSCC prognosis. Hence, our analysis provided a biomarker candidate for clinical HNSCC diagnosis and prognosis.

Although our results showed that a higher expression of CNIH4 was associated with worse survival of HNSCC patients, it was not clear what roles CNIH4 played in HNSCC. Previously, a study reported that the protein secretion modulator TMED9 drives CNIH4 signaling to inhibit TMED3-WNT-TCF, thereby promoting the migration of colon cancer [7]. Yet, our results showed a potential role of CNIH4 in cancer growth. In the present study, the enrichment results showed that the CNIH4 level was positively associated with stemness-related terms, such as “cell cycle” and “DNA replication”. Cancer proliferation is largely dependent on cancer stem cells [20]. The OCLR score also demonstrated that a higher CNIH4 expression might result in a higher stemness of HNSCC. In addition, in the single-cell analysis, stemness, cell cycle, and DNA replicate were the three most CNIH4-correlated states with considerable large coefficients. The single-cell sequencing gets about 20,000 gene expressions. CNIH4 does not have the highest correlation. But this study focused on CNIH4, not the other genes, hence, although other genes might have an even higher correlation, we did not mention them. In the correlation analysis, we only look at one gene, CNIH4. These single-cell data further supported CNIH4 as a biomarker for cancer stemness in HNSCC. The association between CNIH4 and cancer stemness might account for the association between the high expression of CNIH4 and severe HNSCC. Further functional validation of the stemness association of CNIH4 in HNSCC cells is required in the future.

Another interesting finding in the enrichment study was that CNIH4 might be involved in the immunity in HNSCC. These results raised our interest in the effect of CNIH4 in tumor immune therapy. Immunotherapies such as immune checkpoint therapy (ICT), tumor vaccines, immune adaptive therapy, and immunomodulators have been applied for the treatment of many cancer types [21]. However, for HNSCC, more than 70% of patients who received programmed death 1 (PD-1) or programmed death-ligand 1 (PD-L1) blockade did not benefit from immunotherapy [22, 23]. These disappointing results urged a better understanding of the complex biological mechanism underlying the immune response within the HNSCC microenvironment. Therefore, it is of great significance to explore novel immunotherapy targets or biomarkers for the development of immunotherapy regimens in the clinic. CNIH4, as demonstrated in this study, is a potential predictive biomarker for the ICB response. Therefore, we are looking forward to the development of CNIH4 as a clinically applicable index with a clinical referential significance for HNSCC immunotherapy.

The levels of CNIH4 in single-cell data were analyzed in this study. The data set GSE103322 included expression data of 5902 single cells from 18 patients with oral HNSCC, including cells of primary tumors and lymph node metastases tumors, while GSE139324 included 130,721 single cells from peripheral and intra-tumoral CD45 $+$ sorted immune cells from 26 HNSCC patients. The biological relevance of using these two datasets was that these data sets enabled us to compare the gene expression levels in malignant cells and immune cells around them. The single-cell data we analyzed provided insight into the relative expression level of CNIH4 in cancer cells and immune cells.

Our results showed that HNSCC tumor cells expressed a much higher level of CNIH4 than immune cells. A previous study has reported that CNIH4 interacted with newly synthesized G protein-coupled receptors (GPCR) and controlled their export from the endoplasmic reticulum. The role of GPCRs in HNSCC and their clinical relevance to the disease have been extensively explored [24]. GPCR was found to activate the epidermal growth factor receptor (EGFR) and mediate EGFR-independent signaling pathways to promote the growth of HNSCC [25]. We proposed that the potential cancer-promoting effect of CNIH4 in HNSCC might be associated with GPCR and its relative signaling pathways, including pathways that potentially regulated cancer stemness. While in immune cells, regarding the low expression level of CNIH4, we suggested that CNIH4 might not be functional in most immune cells.

In addition, for the interpretation of our immune infiltration analysis, these single-cell expression analysis results provided evidence that a higher level of CNIH4 in HNSCC resulted from higher levels of CNIH4 in tumor cells rather than in immune cells. Therefore, the immune infiltration analysis revealed the effect of CNIH4 expression in cancer cells on the immune microenvironment. Results showed that CNIH4 was negatively associated with most of the immune cell infiltration levels. The only two positive associated immune cell types were common lymphoid progenitor and T cell CD4 $+$ Th2. The response of immune therapy is dependent on the immune cell infiltration level and the expression of immune checkpoint molecules. Immune checkpoints have modulatory effects on immune cells [26]. By these immune checkpoint molecules, cancer cells can escape from the attack of immune cells. CNIH4 expression was negatively correlated with immune checkpoint molecules CTLA4, LAG3, PDCD1, and TIGIT, inferring that CNIH4 might potentially facilitate immune escape. On the other hand, we also analyzed the TIDE score, which evaluated immune escape using a set of marker genes. The TIDE score indicated the dysfunction of tumor-infiltrating cytotoxic T lymphocytes (CTL) and the exclusion of CTL by immunosuppressive factors [13]. A high TIDE score means a worse response to the ICB treatment. Our analysis showed that CNIH4 was associated with the TIDE score, indicating that HNSCC with a higher level of CNIH4 might be easier to escape from cancer immunity and not respond to ICB. Therefore, we proposed that CNIH4 was a sign of ICB-responsive HNSCC.

In conclusion, this study demonstrated that CNIH4 is a diagnostic and prognostic biomarker candidate for HNSCC patients and can potentially affect the cancer stemness and tumor immune microenvironment of HNSCC cells.

Ethics approval and consent to participate

The authors confirm that all methods were performed following the relevant guidelines and regulations.

Consent for publication

All the authors have given consent for publication.

Availability of data and materials

The datasets analyzed during the current study are available in the TCGA and GEO repositories (GSE103322 and GSE139324).

Author contributions

Conception: Hengrui Liu.

Interpretation or analysis of data: Hengrui Liu and Yixue Li.

Preparation of the manuscript: Hengrui Liu and Yixue Li.

Revision for important intellectual content: Hengrui Liu.

Supervision: Hengrui Liu.

Author information

The corresponding author of this study, Hengrui Liu, is a Principal Investigator at Tianjin Yinuo Biomedical Co., Ltd.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-220143.

sj-docx-1-cbm-10.3233_CBM-220143.docx - Supplemental material

Supplemental material, sj-docx-1-cbm-10.3233_CBM-220143.docx

Footnotes

Acknowledgments

We thank Yi Liu for the technical support in the bioinformatics study. We thank Weifen Chen, Zongxiong Liu, Yaqi Yang, Xinming Li, Ruihua Wang, Yinghua Wang, Siying Wang, Pinghui Zeng, and Shengjie Wang for their help in the study.

Conflict of interest

The authors claimed that there is no conflict of interest.

References

Siegel

R.L.

Miller

K.D.

and Jemal

, Cancer statistics, 2019, CA Cancer J Clin 69 (2019), 7–34.

Wingo

P.A.

Cardinez

C.J.

Landis

S.H.

Greenlee

R.T.

Ries

L.A.

Anderson

R.N.

and Thun

M.J.

, Long-term trends in cancer mortality in the United States, 1930–1998, Cancer 97 (2003), 3133–275.

Miller

K.D.

Nogueira

Mariotto

A.B.

Rowland

J.H.

Yabroff

K.R.

Alfano

C.M.

Jemal

Kramer

J.L.

and Siegel

R.L.

, Cancer treatment and survivorship statistics, 2019, CA Cancer J Clin 69 (2019), 363–385.

Duprez

Berwouts

De Neve

Bonte

Boterberg

Deron

Huvenne

Rottey

and Mareel

, Distant metastases in head and neck cancer, Head Neck 39 (2017), 1733–1743.

Chiesa-Estomba

C.M.

Lechien

J.R.

Ayad

Calvo-Henriquez

Gonzalez-Garcia

J.A.

Sistiaga-Suarez

J.A.

Dequanter

Fakhry

Melesse

and Piazza

, Clinical and histopathological risk factors for distant metastasis in head and neck cancer patients, Acta Otorhinolaryngol Ital 41 (2021), 6–17.

Sauvageau

Rochdi

M.D.

Oueslati

Hamdan

F.F.

Percherancier

Simpson

J.C.

Pepperkok

and Bouvier

, CNIH4 interacts with newly synthesized GPCR and controls their export from the endoplasmic reticulum, Traffic 15 (2014), 383–400.

Mishra

Bernal

Silvano

Anand

and Ruiz

I.A.A.

, The protein secretion modulator TMED9 drives CNIH4/TGFα/GLI signaling opposing TMED3-WNT-TCF to promote colon cancer metastases, Oncogene 38 (2019), 5817–5837.

Varisli

, Gene Expression Signatures of Ductal Breast Carcinoma Shows Differentially Expression of Cell Cycle, Proliferation and Apoptosis Related Genes, World Journal of Research and Review 3, 262904.

Wang

Pan

Guo

Luo

Tang

Yang

Zhang

Luo

and Pan

, A novel five-gene signature predicts overall survival of patients with hepatocellular carcinoma, Cancer Medicine 10 (2021), 3808–3821.

10.

Rhodes

D.R.

Shanker

Deshpande

Varambally

Ghosh

Barrette

Pandey

and Chinnaiyan

A.M.

, ONCOMINE: a cancer microarray database and integrated data-mining platform, Neoplasia 6 (2004), 1–6.

11.

Aran

and Butte

A.J.

, xCell: digitally portraying the tissue cellular heterogeneity landscape, Genome Biol 18 (2017), 220.

12.

Malta

T.M.

Sokolov

Gentles

A.J.

Burzykowski

Poisson

Weinstein

J.N.

Kamińska

Huelsken

Omberg

Gevaert

Colaprico

Czerwińska

Mazurek

Mishra

Heyn

Krasnitz

Godwin

A.K.

Lazar

A.J.

Cancer Genome Atlas Research

Stuart

J.M.

Hoadley

K.A.

Laird

P.W.

Noushmehr

and Wiznerowicz

, Machine Learning Identifies Stemness Features Associated with Oncogenic Dedifferentiation, Cell 173 (2018), 338–354.e15.

13.

Jiang

Pan

Sahu

Traugh

Liu

Freeman

G.J.

Brown

M.A.

Wucherpfennig

K.W.

and Liu

X.S.

, Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response, Nat Med 24 (2018), 1550–1558.

14.

Pontén

Jirström

and Uhlen

, The Human Protein Atlas – a tool for pathology, The Journal of Pathology: A Journal of the Pathological Society of Great Britain and Ireland 216 (2008), 387–393.

15.

Yuan

Yan

Zhang

Liu

Deng

Liao

Luo

Yan

Long

Shi

Zhao

Xiao

and Li

, CancerSEA: a cancer single-cell state atlas, Nucleic Acids Res 47 (2019), D900–d908.

16.

Puram

S.V.

Tirosh

Parikh

A.S.

Patel

A.P.

Yizhak

Gillespie

Rodman

Luo

C.L.

Mroz

E.A.

Emerick

K.S.

Deschler

D.G.

Varvares

M.A.

Mylvaganam

Rozenblatt-Rosen

Rocco

J.W.

Faquin

W.C.

Lin

D.T.

Regev

and Bernstein

B.E.

, Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer, Cell 171 (2017), 1611–1624.e24.

17.

Sun

Wang

Han

Dong

Zheng

Shi

Wang

Ren

Sun

Yan

Zhang

and Wang

, TISCH: a comprehensive web resource enabling interactive single-cell transcriptome visualization of tumor microenvironment, Nucleic Acids Res 49 (2021), D1420–D1430.

18.

Cillo

A.R.

Kürten

C.H.L.

Tabib

Onkar

Wang

Liu

Duvvuri

Kim

Soose

R.J.

Oesterreich

Chen

Lafyatis

Bruno

T.C.

Ferris

R.L.

and Vignali

D.A.A.

, Immune Landscape of Viral- and Carcinogen-Driven Head and Neck Cancer, Immunity 52 (2020), 183–199.e9.

19.

McInnes

Healy

and Melville

, Umap: Uniform manifold approximation and projection for dimension reduction, arXiv preprint arXiv:1802.03426 (2018).

20.

Liu

, A Prospective for the Potential Effect of Local Anesthetics on Stem-Like Cells in Colon Cancer, Biomedical Journal of Scientific & Technical Research 25 (2020), 18927–18930.

21.

Chen

Lan

Wang

Yuan

Xiao

Zhang

Gong

Xiao

and Cao

, Analysis of Autophagy-Related Signatures Identified Two Distinct Subtypes for Evaluating the Tumor Immune Microenvironment and Predicting Prognosis in Ovarian Cancer, Frontiers in Oncology 11 (2021), 616133–616133.

22.

Burtness

Harrington

K.J.

Greil

Soulières

Tahara

de Castro

, Jr. Psyrri

Basté

Neupane

Bratland

Å.

Fuereder

Hughes

B.G.M.

Mesía

Ngamphaiboon

Rordorf

Wan Ishak

W.Z.

Hong

R.L.

González Mendoza

Roy

Zhang

Gumuscu

Cheng

J.D.

Jin

and Rischin

, Pembrolizumab alone or with chemotherapy versus cetuximab with chemotherapy for recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-048): a randomised, open-label, phase 3 study, Lancet 394 (2019), 1915–1928.

23.

Ferris

R.L.

Blumenschein

, Jr. Fayette

Guigay

Colevas

A.D.

Licitra

Harrington

Kasper

Vokes

E.E.

Even

Worden

Saba

N.F.

Iglesias Docampo

L.C.

Haddad

Rordorf

Kiyota

Tahara

Monga

Lynch

Geese

W.J.

Kopit

Shaw

J.W.

and Gillison

M.L.

, Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck, N Engl J Med 375 (2016), 1856–1867.

24.

Bhola

N.E.

Thomas

S.M.

Freilino

Joyce

Sahu

Maxwell

Argiris

Seethala

and Grandis

J.R.

, Targeting GPCR-mediated p70S6K activity may improve head and neck cancer response to cetuximab, Clin Cancer Res 17 (2011), 4996–5004.

25.

Bhola

N.E.

Freilino

M.L.

Joyce

S.C.

Sen

Thomas

S.M.

Sahu

Cassell

Chen

C.S.

and Grandis

J.R.

, Antitumor mechanisms of targeting the PDK1 pathway in head and neck cancer, Mol Cancer Ther 11 (2012), 1236–46.

26.

Chan

H.L.

and Chen

, Immune Checkpoint Inhibitors: Basics and Challenges, Curr Med Chem 26 (2019), 3009–3025.