Sage Journals: Discover world-class research

Abstract

PURPOSE:

Aberrant DNA methylation plays a crucial role in oral carcinogenesis. Our previous study demonstrated hypermethylation of DAPK1, LRPPRC, RAB6C, and ZNF471 promoters in patients with tongue squamous cell carcinoma compared with normal samples. Methylation profiling using salivary DNA is considered a non-invasive alternative to tissue samples. Hence, the present study tested the DNA methylation status of these four promoters as indicators of oral cancer progression.

METHODS:

We performed the bisulfite-based targeted next-generation sequencing of four candidate genes in saliva and tissue DNA from normal, premalignant, and squamous cell carcinoma subjects. The clinicopathological association, diagnostic, and prognostic utility of aberrant DNA methylation were evaluated using the TCGA-HNSCC dataset. Using the Xgboost algorithm and logistic regression, CpG sites were prioritized, and Receiver Operating Characteristic was generated. By Log-rank test and Kaplan-Meier (KM) curves, an association between methylation and overall survival (OS), disease-free interval (DFI), and progression-free interval (PFI) were computed.

RESULTS:

We identified all four genes as significantly hypermethylated in premalignant and malignant samples compared with normal samples. The methylation levels were comparable between saliva and tissue samples with an r-value of 0.6297 to 0.8023 and 0.7823 to 0.9419 between premalignant tissue vs. saliva and OC vs. saliva, respectively. We identified an inverse correlation between DAPK1, LRPPRC, RAB6C, and ZNF471 promoter methylation with their expression. A classifier of 8 differentially methylated CpG sites belonging to DAPK1, RAB6C, and ZNF471 promoters was constructed, showing an AUC of 0.984 to differentiate tumors from normal samples. The differential methylation status of DAPK1, LRPPRC, and ZNF71 promoters was prognostically important. Abnormal expression of all four genes was associated with immune infiltration.

CONCLUSIONS:

Thus, methylation analysis of these candidate CpG sites from saliva can be helpful as a non-invasive tool for the clinical management of OC.

Keywords

DNA hypermethylation saliva non-invasive marker oral squamous cell carcinoma bisulfite targeted next-generation sequencing prognosis

1. Introduction

Oral cancer (OC) is a major public health problem that accounts for 1.8% of all cancer deaths [1]. Lip and oral cancer accounted for 377,713 new cases and 177,757 deaths in 2020 [1]. The use of tobacco in smoking, chewing pan, and alcohol consumption are the risk factors for OC [2]. Epidemiological studies causally associated human papillomavirus (HPV) infection with OC [3]. Despite the availability of advanced early detection tools, OC is detected late, contributing to the high treatment cost, adverse clinical outcomes, and high mortality rate [4]. Identifying the early molecular changes may improve the clinical management of OC by providing markers for early diagnosis, prognosis, and therapy response.

Aberrant DNA methylation is linked to features that underlie many phenotypic changes in OC pathogenesis that could serve as potential markers for the detection and prognosis [5, 6, 7]. Quantifying methylation differences at specific sites could delineate a clinically significant threshold to distinguish the disease states in an evolving tumor and establish signature epigenetic marks for early detection [8]. Methylation profiling using DNA isolated from buccal swabs, oral rinse, or saliva suggested the usefulness of these samples as a non-invasive alternative to tissue samples [9, 10]. Further, salivary methylation profiling was demonstrated as a non-invasive and reliable alternative for tissue biopsy samples for the clinical management of OC [10].

Previously we have reported that the CpG site belonging to Death-Associated Protein Kinase 1 (DAPK1), Leucine-Rich Pentatricopeptide Repeat Containing (LRPPRC), Ras-related Protein Rab-6C (RAB6C), and Zinc Finger Protein 471 (ZNF471) promoters are hypermethylated in a cohort of tongue squamous cell carcinoma patients when compared with normal samples [8]. Herein, the promoter DNA methylation status of four candidate genes was evaluated using saliva and tissues collected from the normal, premalignant lesion, and OC patients. The findings were confirmed by analyzing the methylation levels of candidate genes from The Cancer Genome Atlas head and neck squamous cell carcinoma (TCGA-HNSCC) dataset [11]. We next tested these candidate genes’ diagnostic and prognostic utility in OC. We observed that the methylation status of four genes in salivary DNA was comparable to those of the premalignant and malignant tumor tissue specimens. We identified that aberrant methylation of four genes was (i) linked with high sensitivity and specificity to differentiate tumors from normal samples and (ii) potential prognosticator for OC. Our study showed an association between the methylation status of DAPK1, LRPPRC, RAB6C, and ZNF471 gene promoter as surrogate non-invasive markers for the early diagnosis of malignant transformation of the potentially malignant lesion to OC.

2. Materials and methods

2.1 Sample collection

The Institution Ethics Committee approved the study. The samples were collected only after written consent from the participants. Biopsy specimens from normal, premalignant, and squamous cell carcinoma subjects ( $n=$ 20 each) were collected as per the standard clinical procedure. Whole mouth unstimulated saliva from the corresponding subjects was collected by the drooling method after rinsing the mouth with saline. Saliva (2 mL) was gathered together in a sterile sample collector, transported in ice, and immediately used for DNA extraction. Relevant demographic data and informed consent were obtained. A senior oral pathologist provided the histopathological diagnosis of cases for all the cases. The patients who consented to the study were new cases without any history of earlier treatment. The adjacent normal tissue from the same cancer patient was used as a control.

2.2 DNA extraction

QIAamp ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}\textregistered}$ DNA Blood and tissue Mini Kit (Qiagen, Germany) was used for genomic DNA from saliva and tissue biopsy samples. The DNA quality and concentration were evaluated by agarose gel electrophoresis and spectrophotometer and stored at a $-$ 20 ${}^{\circ}$ C refrigerator until used.

2.3 Bisulfite conversion and PCR

Bisulfite conversion of DNA (0.5 to 1 $\mu$ g) was carried out by EZ DNA methylation kit (Zymo Research, USA) and subjected to gene-specific PCR for the promoter region of the four candidate genes. The PCR cycling condition included 95 ${}^{\circ}$ C-5 min; 34 cycles of 95 ${}^{\circ}$ C-30 s, respective annealing temperature- 45 s, and 72 ${}^{\circ}$ C-1 min; and extension of 72 ${}^{\circ}$ C-5 min. Supplementary Table 1 contains details of primer sequence, annealing temperature, and amplicon size [8]. The PCR products purified using HighPrep™PCR (MAGBIO, USA) were subjected to next-generation targeted bisulfite sequencing [12, 13].

2.4 Next-generation sequencing

We used NGS to identify the differently methylated CpG sites within a given amplicon, as published previously [12, 13]. Ion Xpress™Plus Fragment Library Kit was used for library preparation from purified PCR products. PCR products of four genes/samples were pooled, fragmented using Ion Shear™Plus Reagents, purified using Agencourt AMPure XP beads (Beckman Coulter, USA), and ligated to an Ion P1 adapter. The individual samples were barcoded with the Ion Xpress barcode. Following purification by Agencourt AMPure XP beads, 330bp target library (200base read length) was selected with 2% agarose gel (E-Gel SizeSelect™), amplified using Ion Xpress™Plus Fragment Library Kit. The quality and quantity of the amplified library were analyzed using the High Sensitivity DNA kit and Qubit ${}^{\@setsize{\scriptsize}{8pt}{\viipt}{\@viipt}\textregistered}$ dsDNA HS Assay Kit in Bioanalyzer (Agilent Technologies, USA) and Qubit fluorometer, respectively. 100pM of each pooled library was used for emulsion PCR and enriched using Ion PGM™Hi-Q™OT2 reagent in Ion OneTouch™2 system. Templates were enriched using Ion OneTouch™ES, and sequencing was carried out using 316v2 Chips in the Ion Personal Genome Machine (PGM) Sequencer. The NGS reagents were procured from Thermo Fisher Scientific, USA, and used per the manufacturer’s instruction.

2.5 Analysis of NGS data

The NGS data was processed, and differentially methylated CpG sites were identified, as published earlier [12, 13]. The FastQC files were used for analyzing the quality of raw reads, followed by their alignment with reference sequence using NextGenMap [14]. BiQAnalyzer HT software (http://biq-analyzer-ht.bioinf.mpi-inf.mpg.de/) was used for scoring the extent of methylation at each CpG site [15].

2.6 In silico analysis

The methylation status of four candidate genes in the TCGA-HNSCC dataset was analyzed using DNMIVD (http://119.3.41.228/dnmivd/index/) [16]. Genes were considered as hyper and hypomethylated: beta difference $>$ 0.2 & adjusted $p$ -value $<=$ 0.05, Hypomethylation: beta difference $<$ $-$ 0.2 & adjusted $p$ -value $<=$ 0.05, respectively. The Pearson correlation between promoter methylation and FPKM was computed. The association between methylation and clinicopathological characteristics such as stage, race, gender, grade, smoking status, nodal metastasis, and p53 status were carried out using the TCGA-HNSCC dataset using the UALCAN online tool [17].

2.7 Diagnostic model construction

As published earlier, a diagnostic model was constructed using the DNMIVD tool [16]. A cluster map, Receiver Operating Characteristic (ROC) curve, and Area Under Curve (AUC) was constructed to differentiate tumor from normal samples. Feature importance scores for individual CpGs were computed using the xgboost algorithm [16]. The logistic regression model was implemented to generate the diagnostic model and predict AUC via ROC curve construction.

2.8 Prognostic model construction

As published earlier, the prognostic model was constructed using the DNMIVD tool [16]. The samples were partitioned into low and high methylation groups using default parameters. The Log-rank test and Kaplan-Meier (KM) curves were generated to predict the overall survival (OS), disease-free interval (DFI), and progression-free interval (PFI).

2.9 Immune infiltration analysis

We have used the TIMER (http://timer.cistrome.org/) tool to comprehensively explore the association between DAPK1, LRPPRC, RAB6C, and ZNF471 with immune infiltration and abundance [18]. The analysis included the association between the expression of these four genes with six types of immune cells.

2.10 Statistical analysis

The average methylation was calculated by the sum of the methylation levels of all the CpG sites analyzed divided by the total number of CpG sites multiplied by 100. GraphPad Prism (version 7) was used for statistical analyses. The clinical significance of hypermethylation was evaluated by Fisher’s two-tailed tests and paired $t$ -tests. A $p$ -value of $<$ 0.05 was considered statistically significant. The heatmap was generated using the pheatmap tool under the R statistical environment (https://cran.r-project.org/web/packages/pheatmap/index.html).

Figure 1.

Heatmap representing methylation levels of (a) DAPK1, (b) LRPPRC, (c) RAB6C, and (d) ZNF471 by bisulfite-based targeted next-generation sequencing in normal, premalignant, OSCC ( $n=$ 20, each).

3. Results

3.1 Demographic characteristics of the study population

The median age of premalignant, OC and normal subjects was 56 $\pm$ 12.2 years, 60 $\pm$ 12.3 years, and 58 $\pm$ 8.7 years. Nearly 90% of patients in the precancerous group and 69.5% of oral cancer subjects were males. While 35% of precancerous patients and 25% of cancerous subjects had the habit of smoking tobacco, 80% of precancerous subjects and 70% of OC patients were using smokeless tobacco. Further, 70% of patients in the precancerous group had more than one habit of either chewing tobacco and alcohol consumption, or smoking tobacco and alcohol consumption, or both chewing and smoking tobacco. The remaining cases presented a history of either consuming alcohol or smoking tobacco. Few oral squamous cell carcinoma (OSCC) cases (27.7%) presented mixed habits of consuming alcohol and smoking. However, 72.2% of subjects were linked with the habit of either chewing tobacco or smoking. Histopathological diagnosis of Oral potentially malignant disorders (OPMD) was categorized into mild – 30%, moderate – 45%, and severe – 25% dysplasia, whereas all the OC cases were graded as moderately differentiated squamous cell carcinoma (Supplementary Table 2).

3.2 DNA Methylation status of candidate genes in normal, OPMD, and OSCC cases

The level of DNA methylation of the four gene promoters in normal, premalignant, and OC samples were tested by NGS on DNA extracted from saliva and tissue samples. The heat map representing the methylation levels of DAPK1, LRPPRC, RAB6C, and ZNF471 is shown in Fig. 1. The methylation frequency of all the four-gene promoters was significantly higher in premalignant and OC compared with controls samples (Supplementary Table 3, $p<$ 0.05). At least 1 out of 4 genes tested showed hypermethylated promoters in premalignant and OSCC (in both saliva and tissue) compared to healthy control samples. For example, compared to normal samples, the DAPK1 promoter showed hypermethylation in premalignant tissue, premalignant saliva, OSCC tissue, and OSCC saliva. However, the methylation extent was not significantly different between premalignant and OSCC samples. DAPK1 promoter methylation was higher in OSCC tissue than in OSCC saliva. The LRPPRC methylation was not significantly different between normal and premalignant tissue samples. However, compared to normal samples, we observed hypermethylation of the LRPPRC promoter in premalignant saliva, OSCC tissue, and OSCC saliva. RAB6C promoter methylation was higher in both premalignant (tissue and saliva) and OSCC (tissue and saliva) samples than in normal samples. RAB6C promoter methylation was higher in OSCC samples compared to premalignant samples. For ZNF471, premalignant and OSCC samples showed higher methylation than normal samples. ZNF471 methylation was higher in premalignant saliva when compared with OSCC samples(Supplementary Table-3).

Figure 2.

The promoter methylation analysis of four candidate genes showed significant hypermethylation of DAPK1, LRPPRC, and ZNF471 in premalignant and OSCC samples compared to healthy normal control samples ( $p<$ 0.05). However, in saliva, RAB6C hypermethylation was observed in premalignant and OSCC but not in tissue samples.

The average methylation level for any individual gene studied was higher in OPMD and OSCC (tissue and saliva) compared to normal samples (Supplementary Table 3). Figure 2 represents the average promoter methylation of the four candidate genes in tissue and saliva collected from normal, premalignant, and tumor samples. DAPK1, LRPPRC, and ZNF471 promoters showed significant hypermethylation in premalignant and OSCC samples compared to healthy normal controls (Fig. 2). RAB6C promoters showed hypermethylation in premalignant, and OSCC saliva but not in tissue samples (Fig. 2). Individual CpG site analysis showed significantly higher promoter methylation for the gene promoters analyzed in OPMD and OC samples than in normal samples (Fig. 3).

Figure 3.

The individual CpG sites of four candidate genes showed significant differential methylation as analyzed by two-way ANOVA analysis ( $p<$ 0.05). The individual CpG sites were analyzed by taking the mean value for each CpG site across the sample.

3.3 The methylation level of candidate genes in saliva from corresponding subjects

Within a given individual, the site-specific methylation levels were positively correlated in both saliva and tissue samples as analyzed by Pearson’s correlation analysis(Supplementary Table 4, $<$ 0.0001). The r-value among premalignant tissue vs. saliva and OC vs. saliva ranged from 0.6297 to 0.8023 and 0.7823 to 0.9419, respectively. Overall, (i) the methylation level in saliva and tissue samples was comparable in both OPMD and OSCC categories, and (ii) all the four genes tested in the present study were hypermethylated in OPMD and OC samples.

3.4 DAPK1, LRPPRC, RAB6C, and ZNF471 promoters are abnormally methylated in TCGA- HNSC dataset

Since our sample number was small, we analyzed the methylation status of four gene promoters in the oral cancer samples (normal $=$ 34, tumor $=$ 318) using data extracted from the TCGA-HNSCC dataset. The heat map of average promoter methylation of all the genes across all the samples is shown in Fig. 4a. The clustering analysis differentiated the normal and tumor samples (Fig. 4a). Tumor samples showed significant hypermethylation of DAPK1, LRPPRC, RAB6C, and ZNF471 promoters compared to normal samples (Fig. 4b). The effect of promoter methylation on the expression of DAPK1, LRPPRC, RAB6C, and ZNF471 in the TCGA-HNSCC oral cancer cohort were analyzed by Pearson correlation analysis using the Shiny Methylation Analysis Resource Tool [19]. Our study showed a significant inverse correlation between the extent of promoter methylation of DAPK1, LRPPRC, RAB6C, and ZNF471 with its expression in the TCGA-HNSCC cohort. The samples with high methylation showed significant downregulation of the corresponding genes(Supplementary Fig. 1).

Figure 4.

The promoter methylation status of four candidate genes in the oral cancer samples in the TCGA-HNSCC dataset (a) The heat map showing the methylation level of 34 normal and 318 oral cancer samples. The normal and tumor samples showed distinct methylation patterns for all the four tested genes (b) Dot plot representing the methylation status of four genes in 34 normal and 318 oral cancer samples. DAPK1, LRPPRC, RAB6C, and ZNF471 were significantly hypermethylated in oral cancer samples compared with normal samples.

3.5 Differential methylation status of four genes correlated with clinicopathological characteristics in OC

We next tested the association between promoter methylation of four test genes with clinicopathological characteristics such as stage, grade, race, gender, smoking, metastasis, and p53 status in the TCGA-HNSC dataset(Supplementary Table 5). The methylation levels of DAPK1, LRPRC, and ZNF471 were significantly different between normal vs. stage 1 and stage 4 cancers. However, the RAB6C promoter was significantly differentially methylated between normal vs. stage 2 and stage 4 cancers. All four-gene promoters’ differential methylation status was significantly different in Caucasians and African Americans compared with normal. More interestingly, among the cancer patients, DAPK1 methylation was significantly different between male and female patients. All four genes were significantly differentially methylated between normal vs. grade 1 to grade 4 patients. More specifically, DAPK1 showed methylation differences for Grade 1vs. Grade 3, Grade2 vs. Grade3, Grade3 vs. Grade4- tumors. RAB6C methylation could differentiate Grade1 vs. Grade2, Grade1 vs. Grade4, and Grade2 vs. Grade4 tumors. ZNF471 methylation was significantly different between Grade1 vs. Grade 2 tumors.

Methylation of LRPPRC and RAB6C was significantly different between Non-Smoker vs. Smoker. DAPK1 methylation was significantly different between Non-Smoker vs. Reformed Smoker1. ZNF471 methylation was significantly different between Smoker and Reformed Smoker1. LRPPRC and RAB6C showed significant differential methylation in Smoker vs. Reformed Smoker 2. For nodal metastasis, the methylation level was significantly different from LRPPRC and ZNF471 for N0 vs. N2. Furthermore, the methylation status of DAPK1 and LRPPRC were significantly different between TP53 Mutant vs. TP53 Non-Mutant OC.

Figure 5.

The methylation of 8 CpG sites can distinguish tumor from normal samples (a) Clustered map of 8 differentially methylated CpG sites DAPK1, RAB6C, and ZNF471 promoter. (b) The Xgboost and logistic regression analysis shortlisted 8 out of 33 CpG sites with diagnostic importance. (c) The methylation status of 8 CpG sites showed a ROC of 0.984.

3.6 Diagnostic significance of DAPK1, LRPPRC, RAB6C, and ZNF471 promote DNA methylation

We next investigated the diagnostic significance of the promoter methylation status of 4 genes for diagnosing HNSCC using the TCGA-HNSC dataset. First, the clinically relevant differentially methylated CpG sites were computed using the xgboost algorithm to build a diagnostic model (Fig. 5a and b). Figure 5a represents the cluster map of 8 hypermethylated CpG sites. Figure 5b represents the key CpG sites with the potential to differentiate tumors from normal samples identified via the xgboost algorithm. The xgboost algorithm and logistic regression analysis of 8 hypermethylated CpG sites belonged to DAPK1, RAB6C, and ZNF471 promoter differentiated tumors from normal samples with an AUC of 0.984 (Fig. 5c).

3.7 Differential methylation of DAPK1, LRPPRC, and ZNF71 are prognostically important

Using the Log-rank test, we analyzed the prognostic significance of differential methylation of four candidate gene promoters in HNSCC using the TCGA-HNSCC dataset(Supplementary Fig. 2). Among the four genes tested, the hypermethylation of ZNF471 was associated with poor OS. The methylation status of DAPK1, LRPPRC, RAB6C, and ZNF471 did not show any significant association with PFI. However, the methylation status of DAPK1 and LRPPRC genes was associated with DFI.

3.8 DAPK1, LRPPRC, RAB6C, and ZNF471 expression correlated with immune infiltration

The TIMER tool was used for immune infiltration. In HNSC, all four gene expressions positively and negatively correlated with CD4 $+$ T cells, neutrophils, and CD8 $+$ T cells, respectively(Supplementary Fig. 3). For example, DAPK1, LRPPRC, RAB6C, and ZNF471 expressions positively correlated with CD4 $+$ T cell and neutrophils enrichment with immune infiltration level. DAPK1, LRPPRC, RAB6C, and ZNF471 expression negatively correlated with CD8 $+$ T cell abundance with the immune infiltration level.

4. Discussion

OC is among India’s critical public health problems and many other countries [1]. Cancer detection at an advanced stage and lack of reliable prognostic markers contribute to OC’s high incidence and mortality [2, 4, 20]. Changes in the DNA methylation status have been shown to occur at an early stage of tumor development, highlighting the importance of epigenetic modification in OC’s initiation, development, and progression [5]. Hence, assessing the extent of methylation in the CpG-rich regions and functional characterization of downstream targets may unravel the molecular events facilitating carcinogenesis. Besides, such changes may be helpful in diagnosis, prognosis, and improved management of OC.

This study systematically analyzed the relationship of four gene promoter methylation in normal, premalignant, and SCC using saliva and tissue samples, followed by an independent confirmation of our findings using TCGA-HNSCC datasets. We demonstrated that all the genes promoter tested were hypermethylated in our samples. The presence of hypermethylated CpG sites within ZNF47, RAB6C, LRPPRC, and DAPK1 promoters in the premalignant condition suggests that the methylation profiling of these sites could be helpful in an early diagnosis of OC. ROC analysis demonstrated that the methylation of 8 CpG belonging to ZNF471, RAB6C, and DAPK1 could help distinguish tumors from normal samples. The differential promoter methylation of four genes also correlated with immune infiltrations. Collectively, our data demonstrated that DAPK1, LRPPRC, RAB6C, and ZNF471 are differentially methylated between normal, premalignant, and tumor samples and that abnormal methylation may be helpful as a diagnostic and prognostic indicator in OC.

The present study is the first to describe the hypermethylation of ZNF471, RAB6C, and LRPPRC in saliva. A high positive correlation of DNA methylation profile between saliva and tissue implies the feasibility of saliva as a reliable alternative to oral tissues for methylation analysis in a clinical setting. Our study supports the previous studies supporting saliva as an alternative to tissue for methylation profiling in OC [10, 21, 22]. Hansen and co-workers showed that the response rate of participants to provide saliva was significantly higher (72%) than blood samples (31%) [21]. This is particularly beneficial to patients in clinical settings where salivary methylation profiling could be an adjunct diagnostic tool for screening, monitoring, and patient follow-up.

DAPK1, located at 9q21.33, encodes for a 160kD tumor suppressor protein that is downregulated in OC via hypermethylation [23]. It is a calcium-calmodulin-dependent serine-threonine kinase linked with pathways such as cell survival, type 1 apoptosis, and type II autophagy [24]. Downregulation of DAPK1 induces cell proliferation and growth in multiple cancer, suggesting it is a tumor growth regulator gene [25, 26, 27]. Inactivation of DAPK1 by hypermethylation is reported in metastatic progression in multiple cancer types [28, 29]. Tumor-specific DAPK1 hypermethylation is reported using DNA isolated from saliva in HNSCC [30]. A study by Dmitry and co-workers showed that salivary methylation profiling for RASSF1A, DAPK1 and p16 genes showed a tumor-specific accuracy of 81% in HNSCC [30]. A comprehensive meta-analysis showed that the aberrant methylation of DAPK1 promotes in HNSCC [23]. Our study showed that differential methylation of DAPK1 can help differentiate grading the tumor samples. Interestingly, the promoter methylation of DAPK1 was also significantly different between p53 wild type vs. p53 mutant tumor samples.

ZNF471, located at 19q13.43, encodes for a 73kD protein proposed to possess DNA-binding transcription factor activity and is frequently hypermethylated and silenced in a few cancers [8, 12, 31, 32, 33]. We have previously reported the hypermethylation of ZNF471 for the first time in SCCT [8]. The TCGA-HNSCC data set analysis demonstrated a strong negative correlation between promoter methylation and ZNF471 expression, suggesting that ZNF471 may be a methylation-regulated gene in OC. Besides, hypermethylation of ZNF471 also correlated with poor OS. ZNF471 methylation profiling could help differentiate smoker vs. reformed smoker 1 and N0 vs. N2 nodal metastasis. Hence, more detailed functional and mechanistic investigations are required to understand the contribution of ZNF471 in OC.

LRPPRC located at 2p21 encodes for a 157.9kD leucine-rich protein having multiple pentatricopeptide repeats is suggested to take part in transcriptional regulation of nuclear and mitochondrial encoded genes [34]. It is predicted to contain a mitochondrial targeting sequence at its N-terminal and is predominantly localized to mitochondria [35]. Increased LRPPRC expression is reported in lung, gastric, and prostate cancer and correlated with disease progression [35, 36, 37, 38]. Inhibition of LRPPRC has been shown to inhibit growth and invasion with concomitant induction of apoptosis [38]. LRPPRC is an overexpressed gene in gastric cancer, and its forced downregulation reversed therapy resistance [33]. LRPPRC is overexpressed in retinoblastoma and promotes migration, invasion, and glycolysis via autophagy suppression 3̧9. LRPPRC promoted cisplatin resistance in lung cancer [36]. Our study reported the hypermethylation of LRPPRC in saliva and OC samples. LRPPRC hypermethylation did not influence OS and PFI. However, LRPPRC hypermethylation prolonged the DFI in TCGA-HNSCC dataset. More detailed mRNA or protein level expression studies are needed to understand the role of LRPPRC hypermethylation in prolonging the survival of HNSCC patients. Towards this, its contribution to OSCC requires further studies. A comprehensive mechanistic studies are needed to understand the role of LRPPRC in OC.

RAB6C is located in 2q21.1 and encodes for a 28.35kD protein detected as hypermethylated in our samples and the TCGA-HNSCC dataset. In breast cancer, hypermethylation and downregulation of RAB6C are linked with the induction of multidrug resistance [40]. RAB6C is an inhibitor of migration, invasion, and metastasis in breast cancer [41]. The ER $+$ /PR $-$ breast cancer patients with low RAB6C benefited from treatment with tamoxifen [42]. We have previously reported the hypermethylation of RAB6C promoter in SCC of tongue and cervical cancer [8, 12]. However, the molecular mechanisms and signaling pathways regulated by RAB6C in OC are still elusive.

With the progress in immunotherapy and success with immune checkpoint inhibitors in cancer, researchers are increasingly focusing on the role of abnormal expression of the genes in immune infiltration and cancer progression [43]. We used TIMER online database to find the possible association of the four candidate genes with immune infiltration. Our analysis showed that altered expression of all four genes in HNSCC was correlated with CD4 $+$ T cell, CD8 $+$ Tcell, and neutrophil infiltrations. The impact of tumor infiltration of CD4 $+$ , CD8 $+$ T cells, and neutrophils on the outcome of oral cancer patients is unclear. However, some reports on the correlation between tumor infiltration by various immune cells with OC development and progression are already reported [44]. The neutrophils present in the tumor microenvironment are shown as both promoters and suppressors of tumor progression [45]. Imbalances in the neutrophil functions in the oral cavity may contribute to immune evasion to facilitate carcinogenesis and tumor progression [46]. A study showed the association between neutrophils with immune infiltration and tumor progression [47]. An increase in CD4 T cells (Tregs) cells in the tumor microenvironment have been correlated with poor prognosis in oral cancer [48]. Another study showed that the density of tumor-infiltrating CD8 ${}^{+}$ T-cells is an independent risk factor in oral cancer cases [49]. High infiltrates of CD4 $+$ , and CD8 $+$ T cells are reported from the invasive margins and tumor centers in gingivo-buccal squamous cell carcinoma [50]. Hence, more detailed mechanistic studies are needed to understand the molecular details of abnormal methylation and expression of four candidate genes with immune infiltration and subsequent role in OC.

Our study had several limitations. First, the methylation analysis was performed on only a small amount of saliva. Second, we did not perform the expression analysis of the four-candidate gene. Third, detailed functional and mechanistic studies are needed to understand the biological effect of these candidate genes. Finally, many clinical correlation analyses used TCGA-HNSCC data sets via freely available online tools. However, it is worth noting that the associations between aberrant methylation of DAPK1, LRPPRC, RAB6C, and ZNF471 with the development and progression of OC and salivary DNA methylation may be helpful for early diagnosis of OC.

5. Conclusion

The utility of saliva as an alternative tissue source for molecular studies and biomarker discovery is gaining attention. In our study of OPMD and OC subjects, the salivary DNA methylation pattern of the candidate gene promoters was comparable to tumor tissue specimens, thus demonstrating the potential to be used as a surrogate non-invasive marker for the early diagnosis of OC. Our study can potentially establish DAPK1, LRPPRC, RAB6C, and ZNF471 gene promoter methylation analysis as potential biosignatures in a large OSCC clinical cohort to fully establish its diagnostic utility. Although the sample size is small, our prospective study uses a saliva sample, i.e., a non-invasive technique to explore the methylation signatures. Future studies must validate our findings in a larger cohort of patients. Detailed studies are also required to investigate the biological function and cell signaling pathways regulated by these genes in the pathophysiology of OC.

Statements and declarations

The authors declare no conflict of interest. All the authors have read and approved the final draft of the manuscript.

Author contributions

Conception: SPK and AK.

Interpretation Or Analysis Of Data: SC, VKV, SG, RR, and SM.

Preparation Of The Manuscript: SPK, RR, and AK.

Revision For Important Intellectual Content: RR, SPK, and AK.

Supervision: SC, SPK, and AK.

Supplementary data

The supplementary files are available to download from http://dx.doi.org/10.3233/CBM-220028.

sj-pdf-1-cbm-10.3233_CBM-220028.pdf - Supplemental material

Supplemental material, sj-pdf-1-cbm-10.3233_CBM-220028.pdf

Footnotes

Acknowledgments

This study received funding from the Royal College of Surgeons of Edinburgh under the Small Research Pump Priming Grants scheme (SRG/16/099) awarded to Adarsh Kudva. We acknowledge Dr. TMA Pai Structured Ph.D. fellowship, Manipal Academy of Higher Education (MAHE), for fellowships. All the authors thank the Technology Information Forecasting and Assessment Council (TIFAC)-Core in Pharmacogenomics at MAHE, Manipal, Fund for Improvement of S&T Infrastructure (FIST), and Karnataka Fund for Infrastructure Strengthening in Science and Technology (K-FIST), Government of Karnataka.

References

Sung

Ferlay

Siegel

R.L.

Laversanne

Soerjomataram

Jemal

and Bray

, Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries, CA Cancer J Clin 71 (2021), 209–249.

Ram

Sarkar

Kumar

Konwar

Bhatt

M.L.

and Mohammad

, Oral cancer: Risk factors and molecular pathogenesis, J Maxillofac Oral Surg 10 (2011), 132–137.

Chaitanya

N.C.

Allam

N.S.

Gandhi Babu

D.B.

Waghray

Badam

R.K.

and Lavanya

, Systematic meta-analysis on association of human papilloma virus and oral cancer, J Cancer Res Ther 12 (2016), 969–974.

Johnson

D.E.

Burtness

Leemans

C.R.

Lui

V.W.Y.

Bauman

J.E.

and Grandis

J.R.

, Head and neck squamous cell carcinoma, Nat Rev Dis Primers 6 (2020), 92.

Flausino

C.S.

Daniel

F.I.

and Modolo

, DNA methylation in oral squamous cell carcinoma: From its role in carcinogenesis to potential inhibitor drugs, Crit Rev Oncol Hematol 164 (2021), 103399.

Shenoy

U.S.

Adiga

Kabekkodu

S.P.

Hunter

K.D.

and Radhakrishnan

, Molecular implications of HOX genes targeting multiple signaling pathways in cancer, Cell Biol Toxicol 38 (2022), 1–30.

Khongsti

Lamare

F.A.

Shunyu

N.B.

Ghosh

Maitra

and Ghosh

, Whole genome DNA methylation profiling of oral cancer in ethnic population of Meghalaya, North East India reveals novel genes, Genomics 110 (2018), 112–123.

Bhat

Kabekkodu

S.P.

Jayaprakash

Radhakrishnan

Ray

and Satyamoorthy

, Gene promoter-associated CpG island hypermethylation in squamous cell carcinoma of the tongue, Virchows Arch 470 (2017), 445–454.

Gissi

D.B.

Tarsitano

Gabusi

Rossi

Attardo

Lenzi

Marchetti

Montebugnoli

Foschini

M.P.

and Morandi

, 13-gene DNA methylation analysis from oral brushing: A promising non invasive tool in the follow-up of oral cancer patients, J Clin Med 8 (2019).

10.

Rapado-Gonzalez

Lopez-Cedrun

J.L.

Lopez-Lopez

Rodriguez-Ces

A.M.

and Suarez-Cunqueiro

M.M.

, Saliva gene promoter hypermethylation as a biomarker in oral cancer, J Clin Med 10 (2021).

11.

Cancer Genome Atlas

, Comprehensive genomic characterization of head and neck squamous cell carcinomas, Nature 517 (2015), 576–582.

12.

Bhat

Kabekkodu

S.P.

Varghese

V.K.

Chakrabarty

Mallya

S.P.

Rotti

Pandey

Kushtagi

and Satyamoorthy

, Aberrant gene-specific DNA methylation signature analysis in cervical cancer, Tumour Biol 39 (2017), 1010428317694573.

13.

Bhat

Mallya

Varghese

V.K.

Jayaram

Chakrabarty

Joshi

K.S.

Nesari

T.M.

and Satyamoorthy

, DNA methylation detection at single base resolution using targeted next generation bisulfite sequencing and cross validation using capillary sequencing, Gene 594 (2016), 259–267.

14.

Sedlazeck

F.J.

Rescheneder

and von Haeseler

, NextGenMap: Fast and accurate read mapping in highly polymorphic genomes, Bioinformatics 29 (2013), 2790–2791.

15.

Lutsik

Feuerbach

Arand

Lengauer

Walter

and Bock

, BiQ Analyzer HT: Locus-specific analysis of DNA methylation by high-throughput bisulfite sequencing, Nucleic Acids Res 39 (2011), W551–W556.

16.

Ding

Chen

Feng

Chen

Guo

and Shi

, DNMIVD: DNA methylation interactive visualization database, Nucleic Acids Res 48 (2020), D856–D862.

17.

Chandrashekar

D.S.

Bashel

Balasubramanya

S.A.H.

Creighton

C.J.

Ponce-Rodriguez

Chakravarthi

and Varambally

, UALCAN: A portal for facilitating tumor subgroup gene expression and survival analyses, Neoplasia 19 (2017), 649–658.

18.

Zeng

Cohen

Chen

and Liu

X.S.

, TIMER2.0 for analysis of tumor-infiltrating immune cells, Nucleic Acids Res 48 (2020), W509–W514.

19.

and Lu

, The SMART app: An interactive web application for comprehensive DNA methylation analysis and visualization, Epigenetics Chromatin 12 (2019), 71.

20.

Rivera

, Essentials of oral cancer, Int J Clin Exp Pathol 8 (2015), 11884–11894.

21.

Hansen

T.V.

Simonsen

M.K.

Nielsen

F.C.

and Hundrup

Y.A.

, Collection of blood, saliva, and buccal cell samples in a pilot study on the Danish nurse cohort: Comparison of the response rate and quality of genomic DNA, Cancer Epidemiol Biomarkers Prev 16 (2007), 2072–2076.

22.

Langie

S.A.S.

Moisse

Declerck

Koppen

Godderis

Vanden Berghe

Drury

and De Boever

, Salivary DNA methylation profiling: Aspects to consider for biomarker identification, Basic Clin Pharmacol Toxicol 121(Suppl 3) (2017), 93–101.

23.

Jayaprakash

Varghese

V.K.

Bellampalli

Radhakrishnan

Ray

Kabekkodu

S.P.

and Satyamoorthy

, Hypermethylation of Death-Associated Protein Kinase (DAPK1) and its association with oral carcinogenesis – An experimental and meta-analysis study, Arch Oral Biol 80 (2017), 117–129.

24.

Farag

A.K.

and Roh

E.J.

, Death-associated protein kinase (DAPK) family modulators: Current and future therapeutic outcomes, Med Res Rev 39 (2019), 349–385.

25.

Yuan

Shu

Chen

Liu

Zhou

Liu

Tang

Zhang

and Shu

, Downregulation of DAPK1 promotes the stemness of cancer stem cells and EMT process by activating ZEB1 in colorectal cancer, J Mol Med (Berl) 97 (2019), 89–102.

26.

Y.H.

Chou

T.F.

Young

Hsieh

F.Y.

Pan

H.Y.

S.T.

Brown

S.B.

Chen

R.H.

Kimchi

and Lai

M.Z.

, Tumor suppressor death-associated protein kinase 1 inhibits necroptosis by p38 MAPK activation, Cell Death Dis 11 (2020), 305.

27.

Elbadawy

Usui

Yamawaki

and Sasaki

, Novel functions of death-associated protein kinases through mitogen-activated protein kinase-related signals, Int J Mol Sci 19 (2018).

28.

Martinez-Glez

Franco-Hernandez

Gonzalez-Gomez

Isla

De Campos

J.M.

Vaquero

Gutierrez

Casartelli

and Rey

J.A.

, DAPK1 promoter hypermethylaiton in brain metastases and peripheral blood, Neoplasma 54 (2007), 123–126.

29.

Steinmann

Kunze

Hampel

Eckstein

Bertram Bramsen

Muenzner

J.K.

Carle

Ndreshkjana

Kemenes

Gasparini

Friedrich

Andersen

Geppert

Wang

Eyupoglu

Bauerle

Hartmann

and Schneider-Stock

, DAPK1 loss triggers tumor invasion in colorectal tumor cells, Cell Death Dis 10 (2019), 895.

30.

Ovchinnikov

D.A.

Cooper

M.A.

Pandit

Coman

W.B.

Cooper-White

J.J.

Keith

Wolvetang

E.J.

Slowey

P.D.

and Punyadeera

, Tumor-suppressor gene promoter hypermethylation in saliva of head and neck cancer patients, Transl Oncol 5 (2012), 321–326.

31.

Bhat

Kabekkodu

S.P.

Adiga

Fernandes

Shukla

Bhandari

Pandey

Sharan

and Satyamoorthy

, ZNF471 modulates EMT and functions as methylation regulated tumor suppressor with diagnostic and prognostic significance in cervical cancer, Cell Biol Toxicol 37 (2021), 731–749.

32.

Tao

Luo

Tang

Zhou

Feng

Qiu

Putti

T.C.

Xiang

Tao

and Ren

, The tumor suppressor Zinc finger protein 471 suppresses breast cancer growth and metastasis through inhibiting AKT and Wnt/beta-catenin signaling, Clin Epigenetics 12 (2020), 173.

33.

Cao

Wang

Zhang

Wong

K.C.

Nakatsu

Wang

Wong

and Yu

, Zinc-finger protein 471 suppresses gastric cancer through transcriptionally repressing downstream oncogenic PLS3 and TFAP2A, Oncogene 37 (2018), 3601–3616.

34.

Cui

Wang

Ren

Zhang

and Zhang

, LRPPRC: A multifunctional protein involved in energy metabolism and human disease, Front Physiol 10 (2019), 595.

35.

Tang

Xiong

Zhang

Xia

and Liu

, Mitochondria-associated protein LRPPRC exerts cardioprotective effects against doxorubicin-induced toxicity, potentially via inhibition of ROS accumulation, Exp Ther Med 20 (2020), 3837–3845.

36.

Cui

Jin

and Zhang

, LRPPRC contributes to the cisplatin resistance of lung cancer cells by regulating MDR1 expression, Oncol Rep 45 (2021).

37.

Zheng

Zhou

Liu

Chen

Liang

Wang

and Fan

, The significance of LRPPRC overexpression in gastric cancer, Med Oncol 31 (2014), 818.

38.

Zhou

Zhang

Hou

and Zhang

, Downregulation of LRPPRC induces apoptosis in prostate cancer cells through the mitochondria-mediated pathway, Cancer Biother Radiopharm 29 (2014), 345–350.

39.

Song

Chen

Y.Y.

Wang

Liu

K.C.

Tan

and Zou

, LRPPRC regulates metastasis and glycolysis by modulating autophagy and the ROS/HIF1-alpha pathway in retinoblastoma, Mol Ther Oncolytics 22 (2021), 582–591.

40.

Tian

Jurukovski

Wang

X.P.

Kaplan

M.H.

and Xu

, Epigenetic regulation of WTH3 in primary and cultured drug-resistant breast cancer cells, Cancer Res 65 (2005), 10024–10031.

41.

Gan

Zuo

Wang

Min

Wang

and Lv

, Expression of WTH3 in breast cancer tissue and the effects on the biological behavior of breast cancer cells, Exp Ther Med 10 (2015), 154–158.

42.

Fohlin

Bekkhus

Sandstrom

Fornander

Nordenskjold

Carstensen

and Stal

, Low RAB6C expression is a predictor of tamoxifen benefit in estrogen receptor-positive/progesterone receptor-negative breast cancer, Mol Clin Oncol 12 (2020), 415–420.

43.

Esfahani

Roudaia

Buhlaiga

Del Rincon

S.V.

Papneja

and Miller

W.H.

Jr., A review of cancer immunotherapy: From the past, to the present, to the future, Curr Oncol 27 (2020), S87–S97.

44.

Fang

Liu

Chen

Wang

Lui

V.W.Y.

Xia

Cheng

and Wang

, Prognostic significance of tumor infiltrating immune cells in oral squamous cell carcinoma, BMC Cancer 17 (2017), 375.

45.

Singel

K.L.

and Segal

B.H.

, Neutrophils in the tumor microenvironment: Trying to heal the wound that cannot heal, Immunol Rev 273 (2016), 329–343.

46.

Domnich

Riedesel

Pylaeva

Kurten

C.H.L.

Buer

Lang

and Jablonska

, Oral neutrophils: Underestimated players in oral cancer, Front Immunol 11 (2020), 565683.

47.

Wang

Feng

Wang

Liu

Xiang

Sun

Zhang

Liu

and Wei

, Neutrophils infiltration in the tongue squamous cell carcinoma and its correlation with CEACAM1 expression on tumor cells, PLoS One 9 (2014), e89991.

48.

O’Higgins

Ward

F.J.

and Abu Eid

, Deciphering the Role of Regulatory CD4 T Cells in Oral and Oropharyngeal Cancer: A Systematic Review, Front Oncol 8 (2018), 442.

49.

Shimizu

Hiratsuka

Koike

Tsuchihashi

Sonoda

Ogi

Miyakawa

Kobayashi

Kaneko

Igarashi

Hasegawa

and Miyazaki

, Tumor-infiltrating CD8(+) T-cell density is an independent prognostic marker for oral squamous cell carcinoma, Cancer Med 8 (2019), 80–93.

50.

Mukherjee

Bag

Chakraborty

Dey

Roy

Jain

Roy

Soong

Majumder

P.P.

and Dutt

, Density of CD3+ and CD8+ cells in gingivo-buccal oral squamous cell carcinoma is associated with lymph node metastases and survival, PLoS One 15 (2020), e0242058.

Salivary DNA methylation markers for cancer of oral cavity

Abstract

PURPOSE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Sample collection

2.2 DNA extraction

2.3 Bisulfite conversion and PCR

2.4 Next-generation sequencing

2.5 Analysis of NGS data

2.6 In silico analysis

2.7 Diagnostic model construction

2.8 Prognostic model construction

2.9 Immune infiltration analysis

2.10 Statistical analysis

3.1 Demographic characteristics of the study population

3.2 DNA Methylation status of candidate genes in normal, OPMD, and OSCC cases

3.4 DAPK1, LRPPRC, RAB6C, and ZNF471 promoters are abnormally methylated in TCGA- HNSC dataset

3.7 Differential methylation of DAPK1, LRPPRC, and ZNF71 are prognostically important

3.8 DAPK1, LRPPRC, RAB6C, and ZNF471 expression correlated with immune infiltration

4. Discussion

5. Conclusion

Statements and declarations

Author contributions

Supplementary data

sj-pdf-1-cbm-10.3233_CBM-220028.pdf - Supplemental material

Footnotes

Acknowledgments

References