TERT promoter mutation positive oral cavity carcinomas, a clinically and genetically distinct subgroup of head and neck squamous cell carcinomas

The importance of TERT promoter (pTERT) mutation of oral cavity squamous cell carcinoma (OCSCC) with clinical features and genetic alterations are not well recognized.

7][8] Genetic mutation studies have focused on young-onset OCSCC, with TP53, CDKN2A, CASP8, NOTCH1, and FAT1 being reported as driver gene mutations. 9omprehensive genomic profiling tests for HNSCC have been developed based on the Cancer Genome Atlas (TCGA) project in the United States, 10 but real-world studies in Asia remain scarce (China, 11 Taiwan, 12 and Japan 13 ).To overcome this situation, we conducted a retrospective observational study at the Center for Cancer Genomics and Advanced Therapeutics (C-CAT), which has obtained consent to be used after the comprehensive cancer genome profiling test that is covered by insurance in Japan since 2019.Additionally, we used cbioportal.org 14 to analyze the MSK MetTropism data 15 for HNSCC gene profiles, and provide additional evidence for the conclusions drawn from the C-CAT data.Furthermore, to corroborate the findings of these comprehensive genetic databases, we examined specific gene mutations and their clinical significance from oral cancer tissues at our institution.

| Multi-institutional observational study
This was a real-world study that retrospectively examined the clinical and genetic mutation information of HNSCC registered in the C-CAT database.Comprehensive genomic profiling has been covered by insurance in Japan since 2019.The C-CAT is an organization established within the National Cancer Center (NCC) based on national policy to collect, analyze, and provide information on genomic medicine nationwide, and to use this information appropriately.This study received ethics approval at our institution (No. 2021341G, approved 2/17/2022) and submitted our research to the Information Utilization Review Board of C-CAT, which approved our research and permitted the secondary use of data.
The genomic information used in this study is as follows: < Genomic information: Genetic alterations described in test results < clinical information > Basic patient information: sex, age, cancer type, date of the first visit, patient background, pathological diagnosis, smoking history (presence/absence), alcohol drinking history, ECOG PS, multiple cancers (presence/absence/activity), cancer type information (presence/absence of metastasis) date of last known alive, date of death, and cause of death.

| Comprehensive genomic test
The two test methods of choice for comprehensive genomic profiling in Japan are the FoundationOne ® CDx Cancer Genome Profile (F-One) and OncoGuide™ NCC Oncopanel System.The F-One panel was based only on a tumor specimen and searched for mutations in 324 genes (Table S1).Since the number of analyzed genes should be the same within the study cases and the promoter mutation of TERT (pTERT mutation) was not analyzed in the NCC Oncopanel System, only those tested in F-One were included in the analysis in the present study.
All gene mutational information obtained from the C-CAT data consisted of nonsynonymous mutations with amino acid changes and copy number changes.Of the genetic mutations obtained, those that yielded variant allele frequencies of 10% or greater were listed as mutated genes.We visualized the gene mutation profile using the OncoPrinter platform at cbioportal.org (accessed August 9, 2022).Simultaneously, significant combinations of cooccurrence gene information and mutually exclusive gene pairs were identified.

| MSK MetTropism data validation
Nguyen et al.'s 14 comprehensive gene profiling data from MSK MetTropism were downloaded via cbioportal.org.Age, gender, primary site, TMB, metastatic site, observation period, and prognosis were among the clinical data that were gathered as a result of the data used in the Nguyen et al. report.Alcohol and tobacco use were not clinical indicators that were downloaded for this study.Only HNSCC cases with both information on survival and genetic profiles were included.

| In-house OCSCC cohort
To corroborate the findings of the above mentioned large database, we included in our analysis OCSCC patients whom had undergone curative surgery without prior treatment during the 9-year period from 2012 to 2020 and whose postoperative course was observed for at least 3 years.Fresh-frozen carcinoma specimens were stored after collecting informed consent.We decided to search for C250T (c.1-146C>T) and C228T (c.1-124C>T) mutations in the pTERT by digital PCR.This retrospective data analysis has institutional IRB approval (#G2904-( 9)).
Probes used were TERT C228T_113 Assay (dHsaEXD72405942: Bio-Rad Laboratories, CA) and TERT C250T_113 Assay (dHsaEXD46675715: Bio-Rad Laboratories CA), and the Reaction Mix was prepared with ddPCR Supermix for Probes (Bio-Rad Laboratories, CA), 80 mM EDTA, CviQI, and 20 ng of DNA was input per well.After adjusting the Reaction Mix, a droplet was created using Automated Droplet F I G U R E 1 Study flow chart.This study included 260 cases of squamous cell carcinoma of the head and neck with prognostic information from 32 894 cases of solid tumors accumulated in the C-CAT database at the end of July 2022 and with gene mutational information retrieved by Foundation One.As a validation, 407 head and neck squamous cell carcinoma cases with prognostic information from MSK-MeTropism data were downloaded from cbioporatl.org and compared.To evaluate the significance of mTERT in oral squamous cell carcinoma, 40 cases in which TERT promoter mutations were examined by digital PCR from tumor tissue were also included in the study.Generator.The droplets were subjected to 50 cycles of PCR using a T100 Thermal Cycler, and the fluorescence signal of each droplet was measured using a QX200 Droplet Reader after PCR.For digital PCR, RIKEN Genesis (RIKEN GENESIS CO., LTD., Tokyo, Japan) was commissioned based on the above experimental design.

| Statistical analysis
The endpoint for overall survival was the date of last certified survival or the date of death, and the start date of observation was the date of initial diagnosis or the date of biopsy of the primary tumor.For in-house OCSCC cohort, time to first recurrence or death after surgery (plus postoperative treatment) were considered as RFS events.The χ 2 test or Fischer's exact test was used for itemby-item comparisons.A volcano plot was made using a log risk ratio with a bottom of 2 and a negative log p-value with a bottom of 10 to show the comparison of many gene-by-gene mutation frequency comparisons.The log-rank test was used to compare life expectancy using a survival curve based on the Kaplan-Meier method.R and its command input platform ezR 16 were used as analysis packages.Kaleidagraph ver 5.0 was used for graph visualization.
The C-CAT data generated in this study are available in the article and its supplementary data files and MSK MetTropism data are available on cbioportal.org.

| Extraction of genomic data on HNSCC and its clinical profile
As of July 31, 2022, the C-CAT database contained 32 894 cases of solid cancer.Of these, 260 cases with survival analysis information and analyzed by the F-One panel were selected for this analysis.Validation data indicate that 25 775 cases of solid cancer were present in the MSK MetTropism database as of October 29, 2022.Of these, 407 cases were registered as HNSCC using survival analysis.Furthermore, 40 patients with OCSCC who underwent resection for curative purposes and were appropriately treated with postoperative radiation therapy were included in the study.(Figure 1).Table 1 shows the clinical variables of the 260 patients from C-CAT (A), 407 patients from MSK MetTropism (B), and 40 patients from in-house OCSCC cohort (C).C-CAT and MSK MetTropism cohorts are similar in sex ratio, age, and primary site distribution, although there are racial differences; the in-house data is limited to OCSCC, and the majority of the current cohort is male (85%).

| Visualization of obtained gene mutation profile by OncoPrinter
From C-CAT data, in total, 144 gene alterations with a frequency of 3% or higher were found in the 260 cases included in this study (Table S2).Finally, we could extract 3628 genetic variants, including 1854 missense mutations, 1248 copy number variations, 320 nonsense mutations, 129 frameshift mutations, and 77 promoter mutations.Mutational information was also stored, and putative drivers of unknown significance were selected using OncoPrinter.In this study, 26 genes were identified, which were found in >10% of the selected cases.Age, tumor mutation burden (TMB), and primary site were also visualized along with alterations.The most frequently altered gene in the C-CAT database was TP53 (66%).The second most frequent mutation was CDKN2A (51%), which included copy number alteration.The third most frequent mutation was TERT (29%) (Figure 2A).Mutations in pTERT and SOX2, PRKCL, TERC, KLHL6, BCL6, PIK3CA, and MAP3K13 were identified as mutually exclusive (Table S3).
In this study, 16 genes were discovered from MSK-MetTropism data that were present in more than 9% of the chosen cases.TP53 was the gene that had the most changes (54%).TERT (29%) had the second-most mutations.CDKN2A (51%), which had a copy number change, was the third most frequent mutation (Figure 2B).

| Description of pTERT mutation and their clinical presentation
The pTERT mutation was compared with clinical and genetic parameters to clarify the characteristics of this mutation (Figure 3A-L).From C-CAT data, pTERT mutation subgroups in patients with HNSCC are younger, more frequent in females, and have lower alcohol consumption, lower smoking history, and lower second primary carcinoma than wildtype mutations.The majority of OCSCC cases had the pTERT mutation.The study also identifies common genetic alterations in TP53, HRAS, CASP8, GNAS, and MEN1, whereas BCL6, MAP3K13, KEAP1, FGF12, and CDKN2B alterations were less common.The mutation sites of pTERT mutation lesion were mostly located in hotspots (c.-124C>T and c.-146C>T).Mutational distributions are presented in Table S4.From MSK-MetTropism data, TERT mutations were discovered in a total of 145 cases, of which 10 were missense mutations, 1 was a splice mutation, and 134 cases were promoter mutations.In four of the cases, duplicate mutations were discovered; for the MSK-MetTropism data, all of these mutations were examined as TERT alteration.No significant difference in age distribution between TERT alteration and TERT wild-type F I G U R E 2 Mutational signature.(A) Gene mutational information of 260 cases was visualized using the OncoPrinter platform on cbioportal.org.The visualized genes were limited to those with mutation frequencies exceeding 10%, and more genes were sorted by mutation frequency.Each vertical column shows the clinical and genetic information of a single patient, sorted in order of gene frequency, to visualize the genetic mutation information.TERT promoter mutations occur predominantly in oral cancer.(B) As validation, the MSK-MeTropism data were similarly sorted in order of highest gene mutation frequency with more than 9% mutations and visualized using OncoPrinter.As with the CCAT data, most of the TERT mutations were promoter mutations, which had the second highest mutation frequency.[Color figure can be viewed at wileyonlinelibrary.com] Comparison of TERT mutations with clinical indicators and genetic variants for C-CAT and MSK-MetTropism data.From C-CAT data, TERT promoter (pTERT) mutation subgroups were found to be younger than the wildtype subgroup (p = 0.004, A).Of the 20 patients, 12 (60%) under 40 years of age were positive for pTERT mutation.Most OCSCC also had pTERT mutation (41/65; 63%, B). pTERT mutation subgroups were more frequent in females (p < 0.001, C), had lower alcohol consumption (p = 0.016, D), a lower smoking history (p = 0.001, E), and lower second primary carcinoma (p < 0.05, F) than the wildtype mutations.Based on genetic alterations, of the 144 gene alterations with a frequency of ≥3%, TP53, HRAS, CASP8, GNAS, and MEN1 alterations were common, whereas BCL6, MAP3K13, KEAP1, FGF12, and CDKN2B alterations were less common (G).From MSK-MetTropism data, the age distribution in the TERT alteration and TERT wild-type subgroups did not differ significantly (H).TERT alteration was positive in 15  subgroups.Most OCSCCs and patients under the age of 40 had the TERT alteration, and females and Asians were more frequently found in the TERT alteration subgroups.The study also identifies common genetic alterations in TP53, FAT1, CDKN2A, CASP8, HGF, PTPRD, and CD274, and SOX2, DCUN1D1, FOXA1, and PIK3CA alterations were less frequent.

| Clinical significance of pTERT mutation among OCSCC
Since OCSCC has an enriched pTERT mutated population, we looked at the significance of pTERT mutation in OCSCC.In this study, we combined data from CCAT and MSK-MetTropism with information on age, gender, and genetic mutation (Figure 4A).It was discovered that the pTERT mutated subgroups were younger than the wild-type subgroup.In total, 21 (77.8%) of the 27 patients <40 years of age tested positive for the pTERT mutation.Females made up more pTERT mutation subgroups than males did ( p = 0.008, Figure 4B).According to genetic changes, TP53, CDKN2A, CD274, CASP8, HRAS, NOTCH1, and PDCD1LG2 changes were frequent, while RB1 and PTEN changes were less frequent (Figure 4C).

| Clinical significance of pTERT mutation on in-house cohort
The significance of pTERT mutations in in-house 40 cases was discussed.We divided the study into F I G U R E 4 Integration of C-CAT and MSK-MeTropism data, analysis of the data for oral primary cancers only, and visualization of age, sex, and gene mutation frequency.Compared with TERT wild-type oral cavity cancers, TERT-altered oral cavity cancers tended to be younger (A) and more prevalent in females (B).Additionally, oral cavity cancers with TERT alterations tended to have higher levels of TP53, CDKN2A, CD274, CASP8, and HRAS mutations.PTEN and RB1 were typically more prevalent in the wildtype (C).[Color figure can be viewed at wileyonlinelibrary.com] hotspot C250T and C228T mutations and wildtype.First, pTERT mutation were found in 20 (50%) of the 40 cases in this study.and C250T and C228T mutations were mutually exclusive.
As shown in Figure 5A-O, pTERT was more common in younger patients and in tongue cancer, while absent in the floor of the mouth.pTERT mutation was also more prominent in T4 cases and associated with a lower likelihood of heavy alcohol consumption, smoking history, and second primary carcinoma.There were no significant differences in age, gender, oral subsite, T-classification, and N-classification in the presence of pTERT mutation.Additionally, in the univariate analysis, pTERT mutated cases tended to have a better prognosis than wildtype cases for overall survival.
We further investigated the mutually exclusive C250T and C228T mutations: of the 20 pTERT mutations, 7 had C250T mutations and 13 had C228T mutations.The F I G U R E 5 TERT promoter (pTERT) mutation in in-house oral cavity squamous cell carcinoma cohort.There were no significant differences in the distribution of age (A), gender (B), oral subsite (C), T-classification (D), and N-classification (E) in the presence of pTERT mutation; however pTERT mutation were prominent in younger patients (<40 years) and more common in tongue specimen, whereas they were absent in floor of mouth specimen.pTERT mutation were more prominent in T4 cases, and pTERT mutated cases were less likely to have heavy alcohol consumption (F), smoking history (G), and second primary carcinoma (H).In univariate analysis, pTERT mutated cases tended to have a better prognosis than wildtype cases for overall survival (p = 0.06, I).Of the 20 mTERT40 cases, 7 had C250T mutations and 13 had C228T mutations.The C250T mutation was associated with a higher age group (J) than the C228T mutation and there was a superior trend in the overall survival rate (K).In terms of site of initial recurrence, neither the C228T mutation nor the C250T mutation caused cervical lymph node recurrence, and all cases of cervical lymph node recurrence were TERT wildtype cases (L).C250T mutation was statistically significantly superior in recurrence-free survival (M).The C-CAT (N) and MSK MetTropism data (O) for the C250T and C228T mutations, respectively, showed that C250T had a higher survival curve than C228T in both data sets, although the difference was not significant.[Color figure can be viewed at wileyonlinelibrary.com] C250T mutation was associated with a higher age, had a superior trend in overall survival rate, and was statistically significant superior in recurrence-free survival.Neither the C228T mutation nor the C250T mutation caused cervical lymph node recurrence, and all cases of cervical lymph node recurrence were TERT wildtype cases.The C-CAT and MSK MetTropism data for the C250T and C228T mutations showed that C250T had a higher survival curve than C228T in both datasets, However, the difference was not significant.

| DISCUSSION
The present study investigated the genomic data and clinical profiles of HNSCC using data from two large cancer databases and an in-house cohort of patients with OCSCC.One of the notable findings of this study was the high frequency of pTERT mutation, which encodes the telomerase reverse transcriptase.In the C-CAT database, the pTERT mutation was found in 29% of cases, while in the MSK-MetTropism data, pTERT mutation were identified in 35% cases.pTERT mutation was associated with several clinical characteristics, including younger age, female sex, lower alcohol consumption, and lower smoking history.Our findings might indicate that this mutation also functions as a prognostic factor.pTERT mutation have not been searched in TCGA database, and we believe that this is an important genetic mutation in OCSCC that was overlooked until now.
To the best of our knowledge, this is the first nationwide report of a gene mutation profile with survival analysis of HNSCC in Asia.Yao et al. 17 reported that TERT promoter mutations are characteristic of OCSCC, have a poor local prognosis, and are an independent prognostic factor for TP53 mutations in 117 patients with HNSCC, which were examined using MSK-IMPACT.Ghantous et al. 18 also extracted 159 oral cancers from TCGA data, focused on the amplification of TERT mRNA, and proposed that it is a prognostic factor in oral cancer.Yilmaz et al. 19 also examined 189 head and neck cancer samples from their institution, wherein they found that TERT promoter mutations were common in OCSCC and that the patients were young, the proportion of women was high, and they were less likely to drink alcohol and smoke.The present report is consistent with previous studies.On the other hand, Chang et al. 20 in their review of 201 cases of OCSCC in Taiwanese patients reported that these mutations were present in 64.7% of OCSCC, but did not show a correlation with age or gender as in the present study.Betel nut chewing showed a correlation with TERT promoter mutations, thereby indicating that the exposure factor of betel nut chewing probably had an effect.In Japan, betel nut is not common.It is interesting to note that pTERT mutation cancers do not often occur on the floor of the month in in-house cohort, which may provide insight into OCSCC carcinogenesis.pTERT mutations are known to amplify the expression of TERT protein in tissues, 21 and a review by Gonz alez-Moles et al. 22 stated that TERT protein expression may be a better factor than gene mutation, but TERT mutation is present in the promoter, a finding that is missed by conventional coding region sequencing.We have not found a report that comprehensively describes the relationship between pTERT mutations, their mRNA expression, and TERT protein immunostaining in oral cancer.A recently published systematic review 23 on pTERT mutations and HNSCC does not mention this point.
HNSCC with pTERT mutation is a group of OCSCC patients who are young, are not heavy smokers and drinkers, and have been increasing in recent years, making them a socially important disease group.In this report, we show that the OCSCC with pTERT mutation subgroup is more likely to be TP53, CDKN2A, HRAS, CASP8, CD274, and PDCD1LG2 altered.The recent development of an inhibitor of HRAS, tipifarnib, 24 has already produced positive results in a phase 2 clinical trial.PD-L1 and PD-L2 are the proteins that are expressed by CD274 and PDCD1LG2.Both genes' increased copy numbers could play a significant role in tumor immune escape.Intriguingly, oral cavity cancer with a pTERT mutation is associated with changes in the CD274 and PDCD1LG2 genes, which may indicate the effectiveness of immune checkpoint inhibitors. 25ne drawback of this study is that the C-CAT and MSK-MetTropism data only contain information on gene mutations and does not contain information on expression.Nevertheless, the comprehensive genomic profiling has already been established in Asia, and we have also discovered the importance of pTERT mutations and validated in MSK-MetTropism data.The fact that this disease can be screened by simple methods such as TERT promoter sequencing establishes a new disease concept of pTERT mutations (or TERT positivity) in OCSCC from conventional HNSCC, just as HPV-OPSCC is considered a distinct disease concept from HNSCC.However, the clinical question of why pTERT mutations occur predominantly in oral cancers remains unclear.

T A B L E 1
Patient clinical indicators.