Potential Roles of Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4) in Head and Neck Squamous Cell Carcinoma

Background The diagnosis and prognosis of neck squamous cell carcinoma (HNSC) is a challenge for clinical HNSC management, thus, the investigation of molecular biomarkers of HNSC is urgent. We hypothesized that Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4) is a biomarker for HNSC. Methods We analyzed mRNA seq data, protein staining data, and single-cell expression data of HNSC from open databases and evaluated the diagnostic and prognostic value of CNIH4, and investigated the association of CNIH4 to HNSC cancer biology and immunity. Results CNIH4 was expressed higher and have higher copy number in HNSC compared to normal tissues. CNIH4 was associated with worse overall survival of HNSC patients. A survival nomogram was constructed. 2012 and 421genes were identi�ed as positively and negatively associated with CNIH4 respectively, and they were enriched in “Cell cycle”, “DNA replicate”, “Cytokine−cytokine receptor interaction”, etc. CNIH4 was positively correlated with “stemness”, “cell cycle”, and “DNA repair” in single-cell data. CNIH4 was potentially associated with changes in multiple immune cell in�ltration and cancer immune escape. Conclusion CNIH4 is a diagnostic and prognostic biomarker for HNSC patients and can potentially affect the cancer stemness and tumor immune microenvironment of HNSC cells.


Introduction
Head and neck squamous cell carcinoma (HNSC) is one of the most common cancer types in the world.
There are more than ve hundred thousand new HNSC cases occur in the world, which leads to about 3 hundred thousand deaths each year [1,2].Conventional management for HNSC includes surgery, radiotherapy, and chemotherapy [3].Although a number of studies have been done over the past decades, the therapy of HNSC remains largely unsatisfactory and the overall survival rate of HNSC patients over the past several years remains low.The diagnosis and prognosis of HNSC are a challenge for clinical HNSC treatment.A late diagnosis or imprecise prognosis might lead to the underestimate of cancer and if left untreated or undertreated, patients might develop lymph node metastasis that results in a much lower survival rate [4,5].Therefore, the investigation of molecular biomarkers of HNSC for diagnosis and prognosis is urgent.
Cornichon Family AMPA Receptor Auxiliary Protein 4 (CNIH4) is a protein-coding gene that was reported to interact with newly synthesized GPCR and regulated their export from the endoplasmic reticulum [6].In cancer, CNIH4 was supposed to have potential effects.Studies showed that CNIH4 was found in colon cancer and promoted colon cancer metastases [7].It was found to be one of the differential expressed genes in breast cancer [8].CNIH4 was also proposed as one of the ve genes used for signature prediction of overall survival of patients with hepatocellular carcinoma [9].However, the clinical application of CNIH4 in many other cancer types remains largely unknown and the role of CNIH4 in cancers required further investigation.
Preliminary screening results showed that CNIH4 is a regulatory molecule for HNSC.In this study, we tested CNIH4 as a prognostic gene for HNSC patients.We also proposed that CNIH4 is a regulator for the cancer cell biology and tumor immune microenvironment in HNSC.This study provided a novel molecular marker for the improvement of clinical HNSC treatment.

RNA-seq and clinical data acquisition
Bioinformatic data in this study were obtained from TCGA (The Cancer Genome Atlas) in January 2020.The acquisition and the application of these data followed the guidelines and policies of TCGA.

RNA-seq data analysis
R foundation for statistical computing (2020) version 4.0.3 and ggplot (v3.3.2) was used for the analysis.
The immune cell in ltration levels were analyzed using xCell algorithms.The one-class logistic regression machine learning algorithm (OCLR) [10] algorithm was used to calculate the mRNAsi for the evaluation of stemness.Potential immune checkpoint blockade (ICB) response was predicted using the Tumor Immune Dysfunction and Exclusion (TIDE) [11] algorithm.

Immunohistochemistry staining
The images of CNIH4 protein staining were downloaded from HPA (Human Protein Atlas).Antibody HPA044268 was used for the staining.

Single-cell sequencing data acquisition and analysis
The CancerSEA [12] was used to analyze the functional states and their correlation with CNIH4 expression.Single-cell datasets GSE103322 [13] were analyzed.The TISCH (Tumor Immune Single-cell Hub) [14]was used to analyze the expression of CNIH4 in immune cells and malignant cells.Data sets GSE103322 [13] and GSE139324 [15] were analyzed.The UMAP (Dimensionality reduction via Uniform Manifold Approximation and Projection) [16] method was used to display the data.Cell-type annotations were provided by TISCH.

Statistical signi cance
The statistical signi cance was de ned by a P-value of over 0.05.

The expression of CNIH4 in HNSC
Firstly, we analyzed the CNIH4 expression in HNSC.Results showed that the expression of CNIH4 in HNSC was signi cantly higher than that in head and neck normal tissues (Fig. 1A1).We also found that higher grade HNSC expressed signi cantly higher CNIH4 than lower grade HNSC (Fig. 1A2).Samples from male patients had signi cantly higher CNIH4 than those from female patients (Fig. 1A3).The pTNM IV had signi cantly higher CNIH4 than pTNM I, but the other comparison had no signi cant differences (Fig. 1A4).There was no signi cant difference among different pT stagings or different pN stagings (Fig. 1A5-6).Compared to the white, black patients had a slightly higher expression of CNIH4 (Fig. A7).Compared to the patients who smoke, patients who do not smoke expressed a slightly lower CNIH4 (Fig. A8).Besides, we also compared paired cancer-noncancer samples from the same patient, paired ttest showed that CNIH4 was overexpressed in cancer (Fig. 1B).Clinical information of the high (50-100%, red) and low (0-50%, blue) CNIH4 groups were listed in Table 1.To further con rm CNIH4 protein was overexpressed in HNSC compared to paired normal tissue, we compared protein staining results.Results showed that, compared to the normal oral mucosa, and HNSC tissues had a higher staining intensity of CNIH4 (Fig. 2).Thus, CNIH4 was a diagnostic marker of HNSC.In addition, to investigate the reason for overexpression of CNIH4 in HNSC, we compared the copy number of CNIH4 in HNSC and normal tissues in 27 data sets.The overall statistical analysis revealed that the copy number of CNIH4 in HNSC was signi cantly higher than that in normal tissues.Speci cally, the copy number of CNIH4 in HNSC was signi cantly higher than normal tissues in 20 of the 27 data sets (Fig. 3).Therefore, we suggested that the overexpression of CNIH4 in HNSC resulted from the higher gene copy number of CNIH4 in HNSC.

Survival Prognostic analysis of CNIH4 gene in HNSC.
To evaluate the prognostic power of CNIH4 for HNSC patients, we analyzed the association of CNIH4 expression and the overall survival of HNSC patients.We conducted a Kaplan-Meier survival analysis and log-rank test to compare the survival of the high (50-100%, red) and low (0-50%, blue) CNIH4 groups.
The HR for the high CNIH4 group was 1.5 (95%IC = 1.152-1.983)and the median time for high and low CNIH4 groups were 3 and 5.2 respectively (Fig. 4A).Therefore, we suggested that CNIH4 was a potential impact factor for the overall survival of HNSC.To apply the CNIH4 expression to the clinical prognosis of HNSC, we constructed a survival nomogram.Variables CNIH4 expression, age, gender, grade, and race were analyzed.Univariate Cox regression analysis results showed that, among these ve variables, only CNIH4 expression and age were signi cantly associated with survival (Fig. 4B).Multivariate Cox regression showed that CNIH4 expression and age were independent factors for HNSC patient survival (Fig. 4C).Therefore, based on the Cox regression results, a nomogram including variables CNIH4 expression and age was constructed for the prediction of 1-, 2-, 3-, 5-year survival for HNSC patients.The C-index of the nomogram was 0.602 (Fig. 4D).The prediction results of the nomogram calibration curves were consistent with all patients' observation results (Fig. 4E).These analyses demonstrated the clinical value of CNIH4 for HNSC prognosis.

CNIH4 associated genes enrichment analysis.
To explore the potential mechanisms involved in the effect of the CNIH4 gene on HNSC, we identi ed differentially expressed genes (DEGs) between CNIH4 high (75-100%) and low (0-25%) groups.We set cutoff values of 1.5 and 0.01 for fold change and p-value respectively.Results showed that 2012 and 421 genes were identi ed as DEGs positively and negatively associated with CNIH4 in HNSC respectively (Fig. 5AB).These genes were further enriched in GO terminologies and KEGG pathways.Results of KEGG enrichment showed that genes positively associated with CNIH4 were most enriched in "Cell cycle", while genes negatively associated with CNIH4 were most enriched in "Cytokine − cytokine receptor interaction".In terms of GO enrichment, genes positively associated with CNIH4 were highly enriched in "organelle ssion", "nuclear division", "mitotic nuclear division", "chromosome segregation", and "DNA replication", while genes negatively associated with CNIH4 were highly enriched in "epidermis development", "skin development", "epidermal cell differentiation", "keratinization", and "keratinocyte differentiation" (Fig. 5C).Other less enriched terms included immune-associated terms, such as "T cell activity".Because these genes were enriched in multiple terms that were associated with cancer stemness and immune cells, these results inferred that CNIH4 might play a role in cancer stemness and regulation of immunity in HNSC.

CNIH4 was associated with stemness of HNSC
To further explore the potential role of CNIH4 in HNSC, we investigated a single-cell data set of HNSC.We analyzed the correlation of CNIH4 expression and cancer functional state scores of 2150 single HNSC cells and identi ed 12 functional states that signi cantly correlated with CNIH4 (p < 0.05).Results showed that CNIH4 was positively correlated with stemness, cell cycle, DNA repair, invasion, and proliferation with coe cients of 0.26, 0.23, 0.13, 0.06, and 0.05 respectively.On the other hand, CNIH4 was negatively correlated with angiogenesis, quiescence, metastasis, hypoxia, in ammation, DNAdamage, and differentiation with coe cients of -0.16, -0.16, -0.10, -0.10, -0.08, -0.07, and − 0.06 (Fig. 6A).Detailed data were shown in Fig. 6B.As the correlations of stemness and cell cycle were striking, we further demonstrate the association of CNIH4 and cancer stemness.We applied OCLR to compared the stemness of CNIH4 high (75-100%) and low (0-25%) groups.Results showed that the CNIH4 high group had signi cantly higher stemness than that of the low groups, indicating that CNIH4 might upregulate stemness of HNSC.These results were consistent with the enrichment results that CNIH4 was associated with "cell cycle" and "DNA replicate".

CNIH4 was associated with immunity regulation of HNSC
To study the potential role of CNIH4 in the immunity of HNSC, we rst analyzed the distribution of CNIH4 in different cell fractions in HNSC.Single-cell seq data sets HNSC_GSE103322 and HNSC_GSE139324 were analyzed.Results showed that HNSC malignant cells expressed much higher CNIH4 than immune cells (Fig. 7).Therefore, we suggested that the expression level of CNIH4 in HNSC samples was mainly dependent on the expression of CNIH4 in tumor cells.Furthermore, we calculated the immune cell in ltration score of TCGA data using the xCell algorithms.We compared immune cell in ltration levels between CNIH4 high (75-100%) and low (0-25%) groups.In detail, compared to the CNIH4 low group, the CNIH4 high group was only signi cantly higher in the levels of Common lymphoid progenitor and T cell CD4 + Th2.However, CNIH4 high group was signi cantly 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.These results suggested that CNIH4 was negatively associated with in ltration levels of most immune cells and might negatively regulate immune and stroma in HNSC.Apart from that, we analyzed the correlation of CNIH4 and four immune checkpoints, including CTLA4, LAG3, PDCD1, and TIGIT.Results showed that CNIH4 expression was negatively correlated with CTLA4, LAG3, PDCD1, and TIGIT expression (Fig. 8-B).To further explore the value of CNIH4 for clinical immune therapy, we compared ICB respond of CNIH4 low (0-25%) and high (75-100%) groups.Potential ICB response was predicted using the TIDE algorithm.Results showed that the CNIH4 low group had a signi cantly higher TIDE score than the CNIH4 high group (Fig. 8 bottom).The calculation predicted that only 28 out of 98 (22.22%)HNSC patients responded to ICB treatment in the low CNIH4 group, while 84 out of 126 (66.67%) respond to ICB treatment in the high CNIH4 group (Fig. 8 top).

Discussion
This is the rst study reported that CNIH4 was overexpressed in HNSC compared to normal squamous epithelial tissues.We also revealed that this overexpression was, at least partly, resulted from the increased gene copy number in HNSC, but other factors such as transcriptional factor or methylation can also contribute to the overexpression of CNIH4 in cancer.We also found that the expression of CNIH4 is associated with higher grade HNSC.Thus, our data supported that CNIH4 was a potential diagnostic and prognostic biomolecule in HNSC.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.Thus, we believe that CNIH4 has similar expression patterns and effects in different tumor types.However, to date, the expression pattern of CNIH4 in most other cancer types has not been studied.
One of the most striking ndings of this study was that CNIH4 expression potentially affects the overall survival of HNSC patients.The Cox regression analysis indicated that CNIH4 was a risky factor for HNSC patients and it was independent of patients' age, gender, grade, and race.Although a previous study suggested CNIH4 was valuable in the prognosis of hepatocellular carcinoma patients [9], the clinical prognostic value of CNIH4 for HNSC patients has never been reported previously.In this study, we constructed a nomogram to demonstrate the clinical value of CNIH4 for HNSC prognosis.Hence, our analysis provided a potential biomarker for clinical HNSC diagnosis and prognosis.
Although our results showed that CNIH4 was associated with worse survivals of HNSC patients and CNIH4 expression might increase during the development of HNSC, it was not clear what role CNIH4 played in HNSC.Previously, a study reported that the protein secretion modulator TMED9 drives CNIH4 signaling to inhibit TMED3-WNT-TCF, thereby promoted the migration of colon cancer [7].In the present study, the enrichment results showed that CNIH4 was positively associated with some stemness-related terms, such as "cell cycle" and "DNA replication".Cancer proliferation is largely dependent on the cancer stem cells [17].The OCLR score also demonstrated that high CNIH4 might result in a higher stemness of HNSC.In addition, in the single-cell analysis, stemness and cell cycle and DNA replicate were the three most CNIH4-correlated states with considerable large coe cients.These single-cell data further supported CNIH4 as a biomarker of cancer stemness in HNSC and accounted for the association of CNIH4 and more severe HNSC.Further functional validation of the stemness of CNIH4 in HNSC cells is required in the future.
Another interesting nding in the enrichment study was that CNIH4 might be involved in immunity in HNSC.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 in many cancers [18].Tumor immune single-cell data showed that HNSC tumor cells expressed a much higher level of CNIH4 than immune cells.These results indicated that a higher level of CNIH4 in HNSC samples of TCGA should result from higher levels of CNIH4 in tumor cells.The immune in ltration analysis revealed that CNIH4 was negatively associated with most of the immune cell in ltration levels.The only two positive associated immune cell types were common lymphoid progenitor and T cell CD4 + Th2.The response of immune therapy is supposed to be dependent on the immune cell in ltration level and the expression of immune checkpoint molecules.Immune checkpoints have modulatory effects on immunity [19].By these immune checkpoint molecules, cancer cells can escape from the immunity.CNIH4 expression was negatively correlated with immune checkpoint molecules CTLA4, LAG3, PDCD1, and TIGIT, inferring that CNIH4 might promote the immune escape.On the other hand, we also analyzed the TIDE score, which evaluating immune escape using a set of marker genes.TIDE score indicated the dysfunction of tumor-in ltrating cytotoxic T lymphocytes (CTL) and the exclusion of CTL by immunosuppressive factors [11].Patients with a high TIDE score have worse response to the ICB.Our analysis showed that CNIH4 was associated with a TIDE score, indicating that high CNIH4 HNSC might be easier to escape from cancer immunity and not respond to ICB.These results suggested that patients with higher CNIH4 were more likely to respond to immunotherapy.Therefore, we proposed that CNIH4 was a sign of ICB-responsive HNSC.
In conclusion, this study demonstrated that CNIH4 is a diagnostic and prognostic biomarker for HNSC patients and can potentially affect the cancer stemness and tumor immune microenvironment of HNSC cells.

Declarations Figures
The expression of CNIH4 in HNSC.A1.The expression of CNIH4 in HNSC and normal head and neck tissues.TCGA and GETx data were plotted.A2-8.CNIH4 expression levels in HNSC of different clinicopathological groups.B. The expression of CNIH4 in HNSC and normal head and neck tissues.TCGA paired data were plotted, cancer and non-cancer samples were linked.Paired t-test was used to analyze the difference.

Figure 2 Protein
Figure 2

Figure 3 DNA
Figure 3

Figure 4 The
Figure 4

Figure
Figure

Figure 7 The
Figure 7

Table 1
Distribution of HNSC patients with different clinicopathological variables in CNIH4 high and low groups.