Identification of a Novel Prognostic Model for Patients with Head and Neck Squamous Cell Carcinoma Based on Autophagy-related Genes

doi:10.21203/rs.3.rs-1663506/v1

PDF

Research Article

Identification of a Novel Prognostic Model for Patients with Head and Neck Squamous Cell Carcinoma Based on Autophagy-related Genes

https://doi.org/10.21203/rs.3.rs-1663506/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background. Head and neck squamous cell carcinoma (HNSCC) is one of the most common cancers in the world. Experimental evidence indicates that autophagy-related genes (ARGs) are associated with the initiation and progression of HNSCC. However, it's not well defined whether ARGs have the predictive ability of prognostic significance in patients with HNSCC. In this study, we aimed to identify critical ARGs associated with the prognosis of HNSCC, and establish a prognostic signature for the risk assessment.

Methods. Differential expression and functional enrichment analysis were conducted to analyze the datasets from The Cancer Genome Atlas (TCGA) and the Human Autophagy Database (HADb) with R software. The prognosis-related ARGs were screened by univariate Cox regression and LASSO regression analysis. The multivariate Cox regression analysis were applied to identify the prognosis-related ARGs signature of HNSCC.

Results. We established the prognostic signatures based on 9 ARGs by multivariate Cox regression analysis for the overall survival (OS) analysis. The following risk scoring was based on this 9-gene prognostic signature. Kaplan-Meier curves showed that high-risk score group had worse survival than low-risk score group. ROC curves revealed that the risk signature was a predictive indicator in the outcomes of patients with HNSCC. Moreover, nomogram in the prognosis of HNSCC patients confirmed the findings that the predicted probabilities and OS rates at 1-, 3- and 5- years were relatively consistent.

Conclusion. In this study, we identified a novel ARG-based prediction signature, which could predict the prognosis of patients with HNSCC.

TCGA

HNSCC

survival

autophagy-related genes

prognostic signature

Originating from epithelial cells of the oral tongue, cavity, pharynx, nasal cavity, and larynx, head and neck squamous cell carcinoma (HNSCC) is one of the most common and lethal cancers, accounting for 5–10% of all human cancers [1, 2]. Despite advanced treatments by surgery, radiotherapy, and combined chemotherapy [3, 4], the 5-year survival rate for HNSCC remains ~ 60%, mainly due to the local recurrence and distant metastases [5, 6]. Numerous studies have identified the risk factors that induce HNSCC, including human papillomavirus (HPV) infection, smoking, alcohol abuse, and a variety of environmental factors and genetic mutations [7, 8]. Current studies on HNSCC are enriched in molecular targets exploration and novel treatment strategies [9]. However, to our knowledge, the prognostic prediction of HNSCC is still limited. Therefore, the pressing question remains developing specific and sensitive biomarkers for prognostic prediction.

Autophagy is a catabolic process in mammalian cells that responds to extra- and intra-cellular stress. In autophagy, cellular components are degraded in lysosomes to provide the material and energy sources [10, 11]. Emerging evidence has established that abnormalities in autophagy are closely associated with a variety of human diseases, including cancers [12]. Autophagy plays a dual role in regulating the fate of cancer cells, both promoting and suppressing tumorigenesis in different types and stages of cancer [13, 14]. Regulation of autophagy-related genes (ARGs) is thought to be the central mechanism by which autophagy exerts its function in human cancers [15, 16]. Therefore, the expression profiles of ARGs can provide the opportunity to build prognostic models for human cancers, including HNSCC.

In this study, we investigated the ARGs in HNSCC by exploring TCGA (The Cancer Genome Atlas) database, and applied interrogated bioinformatics approaches to screen genes and identified prognostic signatures associated with the survival of HNSCC patients. Our work described the prognostic values of ARGs in HNSCC, and provided novels targets for the improvement of HNSCC prognosis.

Data acquisition

A total of 232 autophagy-related genes (ARGs) were obtained from the Human Autophagy Database (HADb, http://www.autophagy.lu/index.html), an online database of a complete set of human coding genes related to autophagy. The RNA-seq data of tumor samples of HNSCC were obtained from TCGA portal (https://tcga-data.nci.nih.gov/tcga/), which contains 502 tumor samples and 44 adjacent non-tumor samples. The clinical characteristics and survival information of HNSCC patients were also collected.

Screening of differentially expressed ARGs and enrichment analysis

Data analysis of differential expression between tumor tissue samples and their corresponding normal tissue samples downloaded from TCGA was performed using the “limma” package in R (version 4.0.3). Differential genes were obtained with a threshold of |log2(FC)| > 2 and a P value < 0.05. The differential genes were intersected with autophagy genes to obtain differentially expressed ARGs for subsequent analysis. For gene functional analysis, the Gene Ontology (GO) analysis were conducted using the "clusterProfiler" package to identify the main biological properties of differentially expressed ARGs. The results of the annotation analysis were then visualized using the "ggplot2" package for GO enrichment plots. Gene Ontology (GO) analysis consists of three main categories, namely biological processes (BP), molecular functions (MF), and cellular components (CC).

Construction of prognostic model

To establish the prognostic model using ARGs in HNSCC, the univariate Cox regression analysis was performed to screen for ARGs significantly associated with overall survival (OS) under a threshold of P < 0.05. Next, LASSO Cox regression analysis and multivariate Cox regression analysis were performed to establish the ARGs prognostic signature. The prognostic value of the ARGs prognostic signature was assessed by risk scores. The Kaplan-Meier survival curve was plotted to assess the efficacy of the risk signature, comparing the high-risk score group and low-risk score group based on the predictive characteristics. The predictive value of the prognostic signature was then assessed using the areas under the curve (AUC) of the receiver-operator characteristic (ROC) curve by the "timeROC" package in R software. A nomogram was created with the “rms” package and the “survival” package to predict the OS for HNSCC patients. The calibration curves of the nomogram were plotted to evaluate the gap between the nomogram and actual incidence.

Statistical analysis

All statistical analysis in this work were performed by the R software (version 4.0.3). Results were considered to be of statistical significance with the p value < 0.05.

Identification of differentially expressed ARGs in HNSCC.

As a first step toward understanding the prognostic value of autophagy in HNSCC, we screened the differentially expressed genes related to autophagy in HNSCC. As presented by the Venn diagram, among 10,342 differentiated expression genes in HNSCC, there were 104 autophagy-related genes (ARGs), including 69 up-regulated and 35 down-regulated genes (Fig. 1A).

To further figure out the biological functions of these 104 genes, we performed Gene Ontology (GO) enrichment analysis. Results showed that the differentially expressed ARGs were mainly enriched in biological processes (BP) of macroautophagy/autophagy (Fig. 1B). In cellular components (CC), the top enriched terms were autophagosome assembly (Fig. 1C). In molecular functions (MF), the top ranked function terms were cysteine-type endopeptidase activities and ubiquitin-like protein ligase bindings (Fig. 1D).

Construction of the prognostic models using ARGs in HNSCC.

We assessed the prognostic value of above 104 ARGs using univariate Cox regression, and found 21 ARGs to be significantly associated with the prognosis of HNSCC patients. Based on these 21 ARGs, we used LASSO Cox regression analysis (“lars” package and “glmnet” package in R software) to screen out 14 prognosis-related genes. Next, we employed the multivariate Cox regression analysis to finally identify 9 most prognostically valuable ARGs, including ATG4A, RAF1, BAG3, RAB11A, MAP2K7, STK11, ATIC, FADD and USP10. The risk scores were then calculated based on the gene expression levels and the associated Cox regression coefficients, as risk score = expr MAP2K7*-0.1020 + expr STK11*-0.0592 + expr FADD*0.0109 + expr USP10*0.0248 + expr RAB11A* 0.0149 + expr RAF1*-0.0496 + expr ATIC *0.0328 + expr ATG4A*0.0989 + expr BAG3*0.0087 (Fig. 2A-B).

Then we classified HNSCC patients into high- and low-risk score groups based on the median of the OS-related prediction signature. The Kaplan-Meier survival curves clearly showed that the low-risk group had lower mortality and better prognosis than that in the high-risk score group (p < 0.001) (Fig. 2C). Additionally, the ROC curves for the overall survival (OS) were applied to reveal the predictive performance of the 9-gene based risk signature. The AUC values of ROC for 1-, 3-, and 5- years were 0.685, 0.78, and 0.749, respectively, indicating an excellent predictive value of the signature (Fig. 2D).

Identification of the ARGs prognostic signature in HNSCC.

To further evaluate the ARGs prognostic signature, we established the risk factor association map by the "ggplot2" package and the "pheatmap" package in R software, and analyzed the distribution of risk scores, survival time and correlated ARG expression for HNSCC patients. As shown in the association map, the risk scores of patients increased in parallel with decreased survival time and increased numbers of death (Fig. 3A-B).

Next, we analyzed the expression profiles of the 9 ARGs including ATG4A, RAF1, BAG3, RAB11A, MAP2K7, STK11, ATIC, FADD and USP10. It’s noted that MAP2K7, STK11 and RAF1 were highly expressed in the group of low-risk scores, whereas genes including RAB11A, FADD, BAG3 and USP10 were in the group of high-risk scores (Fig. 3C).

To further confirm the effect of the 9 ARGs on the prognosis of HNSCC, we performed Kaplan-Meier logarithm test to assess the OS difference between the different ARGs expressions. Results showed that low expression levels of MAP2K7, STK11 and RAF1 were tightly correlated with the poor survival time in HNSCC patients, whereas low expression levels of RAB11A, FADD, BAG3 and USP10 were closely associated with better prognosis (Fig. 4). All the survival analysis results were in accordance with the gene expression patterns with different risk-scores.

Establishment of the nomogram prognostic model in HNSCC.

To provide better prognosis for HNSCC patients based on our risk score model, we conducted a nomogram to predict the OS of 1-, 3-, and 5-years using five prognostic factors, including gender, stage, grade, age, and risk scores (Fig. 5). The calibration curves showed that the predicted calibration curves were close to the standard curves, suggesting that the nomogram could be reliable in the prediction of OS in HNSCC patients.

HNSCC represents the 6th most prevalent cancer worldwide, with malignant behaviors and low survival rates [17]. One of the major reasons for the poor prognosis of HNSCC was due to the fact that ~ 25% of patients developed tumor recurrence within 5 years after primary diagnosis [18]. However, clinical diagnosis of HNSCC is often made at an intermediate or even late stage, which inevitably leads to a poor prognosis. Therefore, there is an urgent need to identify reliable and effective biomarkers in the diagnosis, treatment, and prognostic assessment of HNSCC.

Currently, clinical and fundamental studies have demonstrated a variety of risk factors and gene mutation/pathways in the initiation and progression of HNSCC [19]. But the prognostic determinant of HNSCC is still limited. In this study, we developed a risk score model using 9 autophagy-related genes to predict the prognosis of HNSCC patients. The risk scores based on prognostic models could be applied clinically to provide convenient and better prognostic monitoring.

At the molecular level, experimental evidence has highlighted the importance of autophagy in HNSCC. Autophagy is an essential catabolic process in cells, which degrades dysfunctional/impaired organelles and macromolecules, and thus maintains the cellular homeostasis [20]. There are accumulating evidence suggesting the role of autophagy-mediated cell survival or progression in HNSCC [21]. In the early stages of tumor formation, autophagy is considered to be anti-cancer and protect normal cells from transforming into malignant tumor cells. When tumor progressed, the autophagy in tumor cells may protect tumor cells from stress-induced death, and contribute to radio- and chemo-resistance. Therefore, autophagy itself is considered as a double-edged sword in the development of HNSCC. Here, we focused on the differentially expressed ARGs in HNSCC, and established a prognostic model using 9 selected ARGs. We noted that several reports have identified prognostic models in HNSCC using immune-related genes [22] and pyroptosis-related genes [23]. Furthermore, another pilot studies also demonstrated the ARG signature in the prediction of prognosis of patients with HNSCC [24–27]. Our work supplements current understandings of the prognostic signature of HNSCC, and reveals that autophagy/ARGs may play a role in the development of HNSCC.

It’s noted that most of our identified ARGs (ATG4A, RAF1, BAG3, RAB11A, MAP2K7, STK11, ATIC, FADD and USP10) had been reported to be involved in HNSCC by either bioinformatic or experimental studies. For example, mutations of STK11 gene produced functionally inactive proteins, and contributed to the development of HNSCC [28]. FADD, an adaptor molecule for death receptor-mediated apoptosis, regulates cell proliferation and has a role in predicting long-term outcomes in patients with HNSCC [29, 30]. Our study provides integrative evidence to demonstrate that the 9 ARGs identified had abilities to predict the prognosis of HNSCC. Of note, the 9 genes were divided into two subgroups, one is negative associated with risk scores (e.g., MAP2K7, STK11 and RAF1) and the other is positive associated with risk scores (e.g., RAB11A, FADD, BAG3 and USP10). This notion was consistent with the OS curves of these ARGs in HNSCC.

There are some limitations in this study. We have not provided experimental evidence to further verify our results. Moreover, future prospective studies with large-sized cohorts and studies are needed to confirm the predictive effect and clinical value of our model.

In conclusion, we performed comprehensive bioinformatics analysis to investigate the role and prognostic value of ARGs for HNSCC. We identified a risk signature based on 9 ARGs that can classify HNSCC patients into subgroups with different prognosis. Our study provides a novel insight into the autophagic status of HNSCC, and affords the potential application of this signature in clinical predictions.

Ethical declarations

All gene expression datasets were collected from TCGA database. This study was approved by the Ethical Committee in the West China Hospital of Stomatology of Sichuan University and conducted according to the principles expressed in the Declaration of Helsinki.

Authors’ contribution

Study design: Xuefeng Zhang and Wanchun Yang. Data retrieve: Xuefeng Zhang. Statistical analysis: Xuefeng Zhang. Result interpretation: Xuefeng Zhang and Wanchun Yang. Writing: Xuefeng Zhang and Wanchun Yang. Funding acquisition: Xuefeng Zhang.

Conflict of interest

The authors have declared no conflicts of interest.

Availability of Data and Materials

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Funding

This work was supported by The National Natural Science Foundation of China (Grant Nos. 82002877). This fund contributed to the design of the study, collection, analysis, and interpretation of data, and writing the manuscript.

Johnson DE, Burtness B, Leemans CR, Lui VWY, Bauman JE, Grandis JR: Head and neck squamous cell carcinoma. Nat Rev Dis Primers 2020, 6(1):92.
Marur S, Forastiere AA: Head and Neck Squamous Cell Carcinoma: Update on Epidemiology, Diagnosis, and Treatment. Mayo Clin Proc 2016, 91(3):386–396.
Gougis P, Moreau Bachelard C, Kamal M, Gan HK, Borcoman E, Torossian N, Bieche I, Le Tourneau C: Clinical Development of Molecular Targeted Therapy in Head and Neck Squamous Cell Carcinoma. JNCI Cancer Spectr 2019, 3(4):pkz055.
Polverini PJ, D'Silva NJ, Lei YL: Precision Therapy of Head and Neck Squamous Cell Carcinoma. J Dent Res 2018, 97(6):614–621.
Citron F, Armenia J, Franchin G, Polesel J, Talamini R, D'Andrea S, Sulfaro S, Croce CM, Klement W, Otasek D et al: An Integrated Approach Identifies Mediators of Local Recurrence in Head and Neck Squamous Carcinoma. Clin Cancer Res 2017, 23(14):3769–3780.
Duprez F, Berwouts D, De Neve W, Bonte K, Boterberg T, Deron P, Huvenne W, Rottey S, Mareel M: Distant metastases in head and neck cancer. Head Neck 2017, 39(9):1733–1743.
Gingerich MA, Smith JD, Michmerhuizen NL, Ludwig M, Devenport S, Matovina C, Brenner C, Chinn SB: Comprehensive review of genetic factors contributing to head and neck squamous cell carcinoma development in low-risk, nontraditional patients. Head Neck 2018, 40(5):943–954.
Benson E, Li R, Eisele D, Fakhry C: The clinical impact of HPV tumor status upon head and neck squamous cell carcinomas. Oral Oncol 2014, 50(6):565–574.
Leemans CR, Braakhuis BJ, Brakenhoff RH: The molecular biology of head and neck cancer. Nat Rev Cancer 2011, 11(1):9–22.
Klionsky DJ, Petroni G, Amaravadi RK, Baehrecke EH, Ballabio A, Boya P, Bravo-San Pedro JM, Cadwell K, Cecconi F, Choi AMK et al: Autophagy in major human diseases. EMBO J 2021, 40(19):e108863.
Ktistakis NT, Tooze SA: Digesting the Expanding Mechanisms of Autophagy. Trends Cell Biol 2016, 26(8):624–635.
Levy JMM, Towers CG, Thorburn A: Targeting autophagy in cancer. Nat Rev Cancer 2017, 17(9):528–542.
Miller DR, Thorburn A: Autophagy and organelle homeostasis in cancer. Dev Cell 2021, 56(7):906–918.
Guo JY, Xia B, White E: Autophagy-mediated tumor promotion. Cell 2013, 155(6):1216–1219.
Levine B, Kroemer G: Biological Functions of Autophagy Genes: A Disease Perspective. Cell 2019, 176(1–2):11–42.
Li X, He S, Ma B: Autophagy and autophagy-related proteins in cancer. Mol Cancer 2020, 19(1):12.
Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A: Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin 2018, 68(6):394–424.
Chuang SC, Scelo G, Tonita JM, Tamaro S, Jonasson JG, Kliewer EV, Hemminki K, Weiderpass E, Pukkala E, Tracey E et al: Risk of second primary cancer among patients with head and neck cancers: A pooled analysis of 13 cancer registries. Int J Cancer 2008, 123(10):2390–2396.
Blise KE, Sivagnanam S, Banik GL, Coussens LM, Goecks J: Single-cell spatial architectures associated with clinical outcome in head and neck squamous cell carcinoma. NPJ Precis Oncol 2022, 6(1):10.
Mizushima N: A brief history of autophagy from cell biology to physiology and disease. Nat Cell Biol 2018, 20(5):521–527.
Sannigrahi MK, Singh V, Sharma R, Panda NK, Khullar M: Role of autophagy in head and neck cancer and therapeutic resistance. Oral Dis 2015, 21(3):283–291.
Jiang P, Li Y, Xu Z, He S: A signature of 17 immune-related gene pairs predicts prognosis and immune status in HNSCC patients. Transl Oncol 2021, 14(1):100924.
Shen Y, Li X, Wang D, Zhang L, Li X, Xia T, Shang X, Yang X, Su L, Fan X: Novel prognostic model established for patients with head and neck squamous cell carcinoma based on pyroptosis-related genes. Transl Oncol 2021, 14(12):101233.
Li Q, Wang J, Meng X, Chen W, Feng J, Mao J: Identification of autophagy-related gene and lncRNA signatures in the prognosis of HNSCC. Oral Dis 2021.
Ren Z, Zhang L, Ding W, Luo Y, Shi Z, Shrestha B, Kan X, Zhang Z, Ding J, He H et al: Development and validation of a novel survival model for head and neck squamous cell carcinoma based on autophagy-related genes. Genomics 2021, 113(1 Pt 2):1166–1175.
Feng H, Zhong L, Yang X, Wan Q, Pei X, Wang J: Development and validation of prognostic index based on autophagy-related genes in patient with head and neck squamous cell carcinoma. Cell Death Discov 2020, 6:59.
Li C, Wu ZH, Yuan K: Autophagy-Related Signature for Head and Neck Squamous Cell Carcinoma. Dis Markers 2020, 2020:8899337.
Qiu W, Schonleben F, Thaker HM, Goggins M, Su GH: A novel mutation of STK11/LKB1 gene leads to the loss of cell growth inhibition in head and neck squamous cell carcinoma. Oncogene 2006, 25(20):2937–2942.
Pattje WJ, Melchers LJ, Slagter-Menkema L, Mastik MF, Schrijvers ML, Gibcus JH, Kluin PM, Hoegen-Chouvalova O, van der Laan BF, Roodenburg JL et al: FADD expression is associated with regional and distant metastasis in squamous cell carcinoma of the head and neck. Histopathology 2013, 63(2):263–270.
Rasamny JJ, Allak A, Krook KA, Jo VY, Policarpio-Nicolas ML, Sumner HM, Moskaluk CA, Frierson HF, Jr., Jameson MJ: Cyclin D1 and FADD as biomarkers in head and neck squamous cell carcinoma. Otolaryngol Head Neck Surg 2012, 146(6):923–931.

No competing interests reported.

PDF

Reviewers invited by journal
15 Jun, 2022
Editor assigned by journal
10 Jun, 2022
Editor invited by journal
23 May, 2022
Submission checks completed at journal
23 May, 2022
First submitted to journal
16 May, 2022

You are reading this latest preprint version

Identification of a Novel Prognostic Model for Patients with Head and Neck Squamous Cell Carcinoma Based on Autophagy-related Genes

Status:

Version 1

Abstract

Figures

Introduction

Materials & Methods

Data acquisition

Screening of differentially expressed ARGs and enrichment analysis

Construction of prognostic model

Statistical analysis

Results

Discussion

Conclusion

Declarations

References

Competing Interests

Status:

Version 1