Abstract
Laryngeal cancer disproportionately affects more African-Americans than European-Americans. Here, we analyze the genome-wide somatic point mutations from the tumors of 13 African-Americans and 57 European-Americans from TCGA to differentiate between environmental and ancestrally-inherited factors. The mean number of mutations were different between African-Americans (151.31) and European-Americans (277.63). Other differences in the overall mutational landscape between African-American and European-American were also found. The frequency of C>A, and C>G were significantly different between the two populations (p-value<0.05). Context nucleotide signatures for some mutation types significantly differ between these two populations. Thus, the context nucleotide signatures along with other factors could be related to the observed mutational landscapes differences between two races. Finally, we show that mutated genes associated with these mutational differences differ between the two populations. Thus, at the molecular level, race appears to be a factor in the progression of laryngeal cancer with ancestral genomic signatures best explaining these differences.
Keywords: Laryngeal cancer, cancer genomics, African-Americans, European-Americans, mutational landscapes, context nucleotides
1.0. INTRODUCTION
Laryngeal cancer afflicts approximately 12,000 new individuals in the United States each year[1, 2] with different incidence and survival rates across ethnic groups [1]. This particular cancer type affects more African-American (Afr-Amr) individuals than European-Americans (Eur-Amr) [1] and the five year survival rate for Afr-Amr with laryngeal cancer is consistently lower than that for Eur-Amr [1]. While socio-economic factors and life styles are associated with the higher incidence and lower survival rates among Afr-Amr [3], we have shown that the contribution of an individual's genetics cannot be ignored [4].
The major risk factors for laryngeal cancer are tobacco smoke and alcohol consumption [5, 6]. Pro-carcinogens found in tobacco smoke are absorbed by cells, metabolized to form active carcinogens, and subsequently excreted from the body following detoxification [7]. If the active carcinogens are not excreted from the cell, the carcinogenic compounds may bind to and ultimately damage DNA [7]. The effect of alcohol with tobacco is synergistic; it is hypothesized that alcohol accelerates the absorption and action of tobacco-based carcinogens [8]. Defects in the enzyme activity or metabolic pathway of tobacco metabolism may lead to the accumulation of tobacco carcinogens in the body and increase the risk of tumor progression. Higher levels of nicotine and cotinine (the major nicotine-based metabolite that contributes to cancer development) have been reported in Afr-Amr compared to individuals of European descent, irrespective of smoking levels [9-12]. In addition, reduced metabolic clearance of nicotine to cotinine and decreased excretions of nicotine and cotinine have been observed in Afr-Amr, relative to Caucasians, for similar cigarette consumption [11, 12]. Genetic studies have identified gene variants associated with reduced rates of nicotine metabolism in populations with significant African descent [13-15]. African-ancestry related genetic variants associated with susceptibility to cancer chemotherapeutic agents have also been demonstrated [16]. In addition, genetic variants associated with increased risk for head and neck cancers in patients of African descent have also been revealed by meta-analysis [17]. These evidences suggest the possible role of genetic ancestry, together with other non-genetic factors, in increased laryngeal cancer risk and poor survival rate among Afr-Amr. Nevertheless, genome-wide analysis to address the disparity issues in laryngeal cancer has not been conducted and genome level analyses are warranted to understand the molecular basis of cancer disparity.
The recent advancements in sequencing technologies has enabled researchers to analyze the whole genome/ exomes of tumor and matched normal samples of several cancer types including laryngeal cancer [18-21]. Nonetheless, the potential baseline effect of population (racial/ethnic) level genetic variation in laryngeal cancer has not been examined. This study compares the distribution of de novo point mutations that have developed in laryngeal cancers among Afr-Amr and Eur-Amr patients to gain insight into the genetic basis of racial disparities in laryngeal cancers.
2.0. MATERIALS AND METHODS
2.1. Data source
Level-2 mutation calls based on whole exome sequence data analyses for Head and Neck Squamous Cell Carcinoma were obtained from The Cancer Genome Atlas (TCGA) (http://cancergenome.nih.gov/). These publically available mutation data were manually curated by the TCGA experts.
2.2. Laryngeal cancer data
Dataset-1: We stratified clinical patient data along with their unique IDs for laryngeal cancer from the TCGA portal based on race: Afr-Amr (n=18) and Eur-Amr (n=91). Of these, only 13 Afr-Amr and 57 Eur-Amr had data in the publicly available TCGA database and we included all 13 Afr-Amr and 57 Eur-Amr patients in Dataset-1.
Dataset-2: Afr-Amr patients possessed a number of common characteristics: current smoker or current reformed smoker; clinical stage III or stage IV; and all were less than 70 years of age. We matched 36 Eur-Amr with similar characteristics to the Afr-Amr patients: all 13 Afr-Amr and 36 Eur-Amr patients represent Dataset-2.
We retrieved the somatic mutations specific for each individual from TCGA data using custom perl/shell scripts. The clinical data contained patient ID and other metadata. We used patient IDs to retrieve corresponding mutations from the TCGA dataset.
2.3. Statistical analyses
On Dataset-1 and Dataset-2, respectively, we conducted a Mann-Whitney U test to compare the differences in each potential factors such as age and pack years between Afr-Amr and Eur-Amr patients. The effects of race, age, smoking status, pack years, and the number of years smoked on the number of mutations of 13 Afr-Amr and 36 Eur-Amr patients (Dataset-2) were studied using multiple linear regression models. We sequentially removed non-significant covariates at a 5% significance level and the final fitted regression model with significant covariates against mutational load (i.e., the number of mutations) was plotted using R.
The following analyses were carried out for all two datasets. We compared the distribution of individual somatic mutations from each individual's tumor for each population. Each mutation was classified as transitions (Ti) (substitution of purine to purine or pyrimidine to pyrimidine) and transversions (Tv) (substitution of purine to pyrimidine or vice-versa), their frequencies were estimated for each individual, and distributions were plotted for each group. Ti and Tv were further classified into all six possible mutational changes, C>T, C>A, C>G, T>A, T>G and T>C, and transitional frequencies were estimated for each individual. A Mann-Whitney U test at a 5% significance level was employed to compare the differences in the: (a) number of mutations, and (b) frequency of mutations for each mutation type between Afr-Amr and Eur-Amr patients. We used custom perl/ shell/R scripts for mutation estimates and STATA 10.0 (Stata Corp, College Station, TX) was used for the Mann-Whitney U test.
2.4. Context nucleotide signatures
We studied the context nucleotide signatures for somatic point mutations in Dataset-2 as this dataset contains matched samples for potential risk factors associated with laryngeal cancer. The development of point mutations highly depends on the localized, or contextual, neighborhood sequence that they are located in [22, 23]. Context nucleotides for a mutation are adjacent nucleotides of that mutation (i.e., nucleotides that exist 3’ and 5’ adjacent to the point mutation) and studying the context nucleotide signatures may explain the genomics factor associated with observed mutational landscape differences. We used the Bioconductor package, SomaticSignatures [24] to analyze the context nucleotide signatures in Afr-Amr and Eur-Amr patients. In total, we analyzed 96 context nucleotide signatures (as there are 16 possible combinations of the four nucleotides (A,T,G,C) at the 5′ and 3′ end of each of 6 possible mutation types.). Differences in the frequency of each context nucleotide signature between these two ethnic groups were assessed statistically using a Mann-Whitney U test.
2.5. Significantly differently mutated genes
We created a list of genes mutated in one or more Afr-Amr or Eur-Amr patients (dataset-2) and the differences in frequency of patients with mutations between Afr-Amr and Eur-Amr groups were studied using chi-square test in R. In addition, we obtained a list of 44 cancer driver genes for Head and Neck Squamous Cell Carcinoma (HNSCC) from the Broad GDAC Firehose (gdac.broadinstitute.org). The Broad Institute has identified these cancer driver genes using HNSCC dataset of TCGA. We analyzed the frequency of patients with mutations in these driver genes in Afr-Amr and Eur-Amr groups and the differences between the two populations were examined by chi-square test for homogeneity with continuity correction in R.
3.0. RESULTS
3.1. Laryngeal cancer samples
Summary statistics of age and pack years for Afr-Amr and Eur-Amr patients for each dataset are given in Table 1. The age and number of pack years were not significantly different between Afr-Amr and Eur-Amr patients in Dataset-1 and Dataset-2. Other clinical characteristics of patients in Dataset-2 are given in Supplementary Table 1.
Table 1.
Clinical data summary statistics for Afr-Amr (AA) and Eur-Amr (EA) laryngeal cancer patients and significant somatic point mutational differences observed between Afr-Amr and Eur-Amr patients, across two data sets. Q1: first quartile; Q3: third quartile; M: median; P-val: P value.
Set | Samples matched for | Age | Pack years | # Mutations | C>A (%) | C>G (%) | ||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Q1 | M | Q3 | Q1 | M | Q3 | M | P-val | M | P-val | M | P-val | |||
1 | 13 AA | None | 54.7 | 60.9 | 63.8 | 34.0 | 40.0 | 60.0 | 150 | 0.063 | 16.5 | 0.162 | 11.3 | 0.01* |
57 EA | 56.3 | 61.9 | 67.9 | 40.0 | 60.0 | 80.0 | 186 | 20.0 | 15.0 | |||||
2 | 13 AA | Smoking status, Stage III/IV, Age below 70 | 54.7 | 60.9 | 63.8 | 34.0 | 40.0 | 60.0 | 150 | 0.056 | 16.5 | 0.037* | 11.3 | 0.03* |
36 EA | 53.0 | 60.1 | 65.3 | 40.0 | 60.0 | 80.0 | 213 | 23.8 | 14.7 |
3.2. Dataset-1: Somatic point mutations and mutational landscapes
The distributions of the number of somatic point mutations per sample for Afr-Amr and Eur-Amr were different (Figure 1A). Specifically, Eur-Amr possessed more point mutations compared to Afr-Amr. The number of mutations ranged from 46 to 1,026 with a mean of 277.63 and a median of 186 for Eur-Amr whereas the number of mutations varied from 29 to 313 with a mean of 151.31 and a median of 150 for Afr-Amr patients. At a significance level of 5%, the medians of the number of mutations in Afr-Amr and Eur-Amr were not significantly different (Table 1; P=0.063).
Figure 1.
Mutational distributions for two filtered cohort datasets. Distribution of the number of somatic point mutations for Dataset-1(A), and Dataset-2 (B). Distributions of frequencies of transitions (Ti) and transversions (Tv) for Dataset-1(C), and Dataset-2 (D). Mutational landscapes observed for African-Americans (AA) and European-American (EA) patients in Dataset-1(E), and Dataset-2 (F). Distribution of frequencies of mutation types for African-Americans and European-American patients in Dataset-1(G), and Dataset-2 (H).
We classified the somatic point mutations into transitions (Ti) and transversions (Tv) and the medians of Ti and Tv frequencies from Afr-Amr and Eur-Amr patients were found to be significantly different between these two racial groups (Figure 1C; P=0.0454). In particular, Afr-Amr patients had a higher proportion of Ti (median=53.11; Q1= 47.18; Q3= 67.01) than Tv (median=46.89; Q1= 32.99; Q3= 52.82). In contrast, Eur-Amr patients had higher Tv proportions (median=50.6; Q1= 42.52; Q3= 57.92) compared to Ti (median=49.4; Q1= 42.08; Q3= 57.48).
The somatic mutations were further categorized into each of the six possible nucleotide changes. The tumors from Afr-Amr and Eur-Amr patients have distinct mutational landscapes (Figure 1E). Of these six possible mutations, only the frequency of somatic mutation type C>G was found to vary significantly between these two populations (Figure 1G and Table 1; P=0.01).
3.3. Dataset-2: Somatic point mutations and mutational landscapes
Eur-Amr patients harbored a higher number of point mutations in the tumor (median= 213; Q1= 129; Q3= 383) compared to Afr-Amr patients (median= 150; Q1= 117; Q3= 177), as we observed in Dataset-1. At a significance level of 5%, the numbers of mutations from Afr-Amr and Eur-Amr patients were not significantly different (Table 1; P= 0.0557). The distribution of the number of mutations per sample for each patient group is shown in Figure 1B. To study the relative effects of race, age, smoking status, the number of years of smoking and pack years on mutation load, we applied multiple linear regressions. Smoking status, the number of years of smoking and pack years were discarded from the regression model at a 5% significance level. After removing such non-significant covariates, the effects of race and age on the number of mutations remained significant. The final fitted model and the ANOVA table are provided in Supplementary Table 2. As age increases, there is significantly more mutational burden in each population (P=0.026) and the increasing rates, that is the effects of age, in Afr-Amr and Eur-Amr are the same (Figure 2; Supplementary Table 2). The number of mutations in Eur-Amr was larger on average (by 151.91 mutations) than in Afr-Amr when ages were matched (Figure 2; Supplementary Table 2).
Figure 2.
Relationship between the number of mutations and age of Eur-Amr (EA; ○) and of Afr-Amr (AA;×) patients, with fitted multiple regression line.
When somatic point mutations were classified as transitions (Ti) and transversions (Tv), we found significantly different distributions of Ti and Tv frequencies between Afr-Amr and Eur-Amr patients (Figure 1D; p=0.0425). Afr-Amr patients had a higher proportion of Ti (median=53.11; Q1= 47.18; Q3= 67.01) than Tv (median=46.89; Q1= 32.99; Q3= 52.82). Eur-Amr patients had higher Tv proportions (median=52.99; Q1= 42.64; Q3= 58.88) compared to Ti (median=47.01; Q1= 41.12; Q3= 57.36).
The somatic mutations were further categorized into each of the six possible nucleotide changes. Figure 1F shows that the tumors from Afr-Amr and Eur-Amr patients have distinct mutational landscapes and Figure 1H shows different distributions of frequencies by mutation type. The frequencies of somatic mutation types, C>A and C>G, were found to vary significantly between these two populations (Table 1; P=0.0372 for C>A; P=0.0297 for C>G). Tumors from Afr-Amr patients harbor lower C>A (median= 16.49; Q1= 13.79; Q3= 26.50) and C>G (median= 11.28; Q1= 8.25; Q3= 16.51) somatic mutations compared to Eur-Amr patients (for C>A: median= 23.79; Q1= 18.21; Q3= 32.90; for C>G: median= 14.68; Q1= 12.66; Q3= 17.72).
3.4. Context nucleotide signatures
The development of somatic mutations depends on the contextual neighborhood of the mutation loci.[22, 23] We investigated whether Afr-Amr patients in Dataset-2 possess different context nucleotide signatures at each somatic mutation loci compared to Eur-Amr patients in Dataset-2, thus favoring different mutational landscapes. Figure 3 shows different context nucleotide patterns for Afr-Amr and Eur-Amr patients with p-values < 0.05. Some context nucleotide frequencies significantly differed between Afr-Amr and Eur-Amr patients. Mutation types C>G and C>A whose frequency distributions are different have four and two, respectively, significantly different context nucleotide patterns between Afr-Amr and Eur-Amr. Among mutation types, C>G showed the most different patterns. Mutation T>A does not have any significantly different patterns and others have 1 or 2 different patterns.
Figure 3.
Fraction of context nucleotide signatures for all six mutation types based on TCGA patient data from African-American and European-American populations.
3.5. Significantly differently mutated genes
We obtained a list of genes mutated in one or more Afr-Amr or Eur-Amr patients (Dataset-2). We found 576 genes were mutated in both Afr-Amr and Eur-Amr patients while 37 genes and 1731 genes were mutated only in Afr-Amr and Eur-Amr patients, respectively. Thus, we analyzed a list of 2,344 (576+37+1731) genes mutated in one or more Afr-Amr or Eur-Amr patients. Of these 2,344 genes, six genes, RUNX1T1, TTN, NAV3, PIK3CA, KIAA1033, and ZMYM6, were significantly differently mutated between Afr-Amr and Eur-Amr patients (Table 2). We also obtained a list of 44 driver genes identified in HNSCC and compared the frequencies of patients with mutations in these driver genes in our dataset. The results are shown in Figure 4. Several driver genes were mutated in different frequencies in Afr-Amr as compared to Eur-Amr. The Eur-Amr patients had mutations in 29 of 44 driver genes while Afr-Amr had mutations in 20 of 44 driver genes. Importantly, a driver gene, PIK3CA was significantly differently mutated between Afr-Amr and Eur-Amr (Table 2).
Table 2.
Genes significantly differing in frequency between African-American (Afr-Amr) and European-American (Eur-Amr) laryngeal cancer patients. Asterick (*) indicates known cancer driver gene from HNSCC.
Gene | Afr-Amr (N=13) | Eur-Amr (N=36) | P-value |
---|---|---|---|
RUNX1T1 | 5 (38.46%) | 3 (8.33%) | 0.037 |
TTN | 5 (38.46%) | 30 (83.33%) | 0.007 |
NAV3 | 0 (0%) | 13 (36.11%) | 0.031 |
PIK3CA* | 0 (0%) | 12 (33.33%) | 0.043 |
KIAA1033 | 3 (23.08%) | 0 (0%) | 0.021 |
ZMYM6 | 3 (23.08%) | 0 (0%) | 0.021 |
Figure 4.
Frequency of alterations of HNSCC driver genes in our study samples (Dataset-2). Genes mutated in significantly different proportions between Afr-Amr and Eur-Amr is denoted by an asterisk, *.
4.0. DISCUSSION
Disparities have been reported in the incidence rate of laryngeal cancer between Afr-Amr and Eur-Amr [1]. However, our knowledge of the differences of laryngeal cancer at the genetic level between populations is limited. The goal of this study is to understand differences in the mutational landscape of tumors between Afr-Amr and Eur-Amr laryngeal cancer patients. The sample size, its large sample space (i.e., whole exomes), and our risk-matched comparison provided us with sufficient power to identify inherent and significant differences between the cancer genomes of Afr-Amr and Eur-Amr patients.
4.1. Effect of race on mutational landscapes
In this study, we grouped and analyzed two sequentially-filtered datasets. While the first dataset was generated to identify overall mutational landscape differences between Afr-Amr and Eur-Amr patients, Dataset-2 was matched for smoking status, and clinical stage III/ IV to test the effect of race after adjusting for smoking status and late stage cancer. If the mutational burdens are independent of race, we expect to see more or less similar mutational landscapes for Afr-Amr and Eur-Amr in our two datasets. However, we observed different mutational landscapes between Afr-Amr and Eur-Amr patients. The results from these two datasets are congruent for most of the observations (Figure 1 and Table 1) and reveal that ancestral origin is an important factor associated with mutation burdens. In each of the analyses, the number of somatic mutations and the frequencies of Ti and Tv were dissimilar between Afr-Amr and Eur-Amr patients. While the differences of the number of mutations were close to significant or not significant, Ti and Tv frequencies varied significantly (p-value < 0.05) between Afr-Amr and Eur-Amr patients in all datasets. In addition, the somatic mutational landscapes of Afr-Amr patients were different from that of Eur-Amr patients.
We also studied the effect of race, age, smoking status, number of pack years, and number of years smoked on mutational load on Dataset-2 using multiple linear regression models. We sequentially removed non-significant factors from regression models. Smoking status, the number of years smoked and pack years did not have significant effects on mutation burden. This result also supports that we removed the effect of known risk factors of laryngeal cancers such as smoking levels and studied the effect of race on mutational landscape using Dataset-2. After removing all non-significant factors, race and age showed significant effects on mutation load (P=0.02 for race, and P=0.049 for age). Thus analyses suggest that race and age are significant factors related to the number of mutations. As we found from other analyses in different datasets, Eur-Amr patients carry about 151 more mutations on average than Afr-Amr while controlling for the age of patients. Our results support race as a potential factor associated with mutational burdens.
Our Dataset-2 provided stronger power to analyze the mutational landscapes. Dataset-2 showed a significant differences in the frequency of C>A and C>G mutation types between Afr-Amr and Eur-Amr patients. However, Dataset-1 did not show a significant difference in the frequency of C>A mutations between Afr-Amr and Eur-Amr patients. The C>A mutations are related to tobacco smoke [25, 26]. In Dataset-1, we did not match the patients for smoking status with Afr-Amr and Eur-Amr patients possessing broad ranges of smoking histories. In fact, the Eur-Amr cohort in Dataset-1 included current smokers (49.1%), current reformed smokers for ≤15 years (34.5%), current reformed smokers for > 15 years (12.73%), and life-long non-smokers (3.6%) whereas the Afr-Amr cohort included current smokers (61.5%), current reformed smokers for ≤15 years (30.8%), and current reformed smokers for > 15 years (7.8%). These differences in smoking history between Afr-Amr and Eur-Amr patients in Dataset-1 may have obscured the significance of race on C>A mutations. Nevertheless, when the patients were matched for smoking status in Dataset-2, we found a significant difference in the frequency of C>A mutations between the two populations.
While socio-economic, behavior, and lifestyle factors are known to be associated with the higher incidence and lower survival rates of laryngeal cancer in Afr-Amr [3], host factors may also play an important role. We attempted to study the effect of ancestry on mutational burden in laryngeal cancer in our study and findings suggest that genetic ancestry could also be associated with the cancer disparity. Each genome-wide sample contained, on average, hundreds of new mutations providing ample power for our comparisons. In each of the datasets, we observed a higher number of somatic point mutations in Eur-Amr as compared to Afr-Amr patients, despite the fact that Afr-Amr and Eur-Amr patients have a similar smoking history in Dataset-2. In contrast, earlier studies showed a direct correlation between number of mutations and smoking rate in tobacco-related cancers [27, 28]. If smoking is the only major factor for mutation burden, we expect to see a similar number of mutations in Afr-Amr and Eur-Amr patients in Dataset-2, however, we found a lower mutational load in Afr-Amr patients, adding more evidence for the effect of race on mutations. Since this study only investigates the mutational landscape differences of Afr-Amr vs Eur-Amr patients, we could not study the effect of mutations or the functional impact of these mutations on these populations. These and other molecular mechanisms explaining the lower number of mutations in Afr-Amr patients need to be further explored.
To delve deeper into the reason for the disproportionate mutational landscapes between these two populations, we further analyzed the context nucleotide signatures (+/− 1 bp) of each mutation types in Afr-Amr and Eur-Amr patients (Dataset-2). Our analyses reveal that the Afr-Amr and Eur-Amr populations have different frequencies of context nucleotide signatures (Figure-3). Thus, ancestral genomic signatures may play a key role in the observed mutational landscape differences. In addition to these signatures, other genomic factors also might have influenced the mutational landscapes. For instance, population-specific driver mutations and other mutational burdens in the tumor cell could have also attributed to the different mutational patterns observed between Afr-Amr and Eur-Amr patients. However, at present, it is not possible for us to study the influence of driver mutations and mutational burdens on the mutational landscape of these two populations since the raw data are not yet available.
4.2. Differentially mutated genes
This study also shows differences in the distribution of mutated genes between Afr-Amr and Eur-Amr patients (Dataset-2). Several of 44 known HNSCC cancer driver genes were mutated in different frequencies between Afr-Amr and Eur-Amr patients. Eur-Amr patients had mutations in 29 of 44 driver genes, whereas Afr-Amr had mutations in 20 of 44 driver genes. Importantly, Afr-Amr did not have any mutation in a driver gene, PIK3CA whereas many Eur-Amr had mutations in this gene and the difference is significant (P=0.043). The analyses of 2,344 genes reveal Afr-Amr had mutations in significantly limited number of genes compared to Eur-Amr (613 vs 2,307; chi-square test P<2.2e-16). While none of the Afr-Amr patients had mutations in PIK3CA, a higher frequency of Afr-Amr patients had mutations in RUNX1T1, KIAA1033, and ZMYM6 as compared to Eur-Amr (P<0.05) (Table 2). The gene, RUNX1T1is known to be a key gene associated with acute myeloid leukemia [29]. In addition, KIAA1033, and ZMYM6 are important for cellular function and cell morphology. Thus, mutations in these genes may be associated with HNSCC progression or survival and needs to be studied further. However, RUNX1T1, KIAA1033, and ZMYM6 were not identified as cancer driver genes in HNSCC. A recent study based on data from 279 TCGA HNSCC patients also showed PIK3CA as a potential driver gene but the genes, RUNX1T1, KIAA1033, and ZMYM6 were not identified as driver genes [21]. The limited inclusion of Afr-Amr patients might have severely biased cancer gene discovery in TCGA cohort and the analyses of more black patients could provide a different list of genes (the TCGA HNSCC cohort currently has ~390 Eur-Amr and ~40 Afr-Amr patients). These observations suggest Afr-Amr may have a different set of frequently-mutated genes that could be associated with a higher risk for laryngeal cancer. Either more Afr-Amr samples or population-specific analyses are needed to identify additional candidate genes from this population.
5.0. CONCLUSION
This study demonstrates the power of a genomics approach to study the etiology of health disparities by analyzing, for the first time, the effect of race on mutational signatures in laryngeal cancer. If laryngeal cancer has a common genetic basis in all populations, we should find a similar mutational burden in Afr-Amr and Eur-Amr patients. Our results reveal different mutation loads between Afr-Amr and Eur-Amr patients. Afr-Amr patients may experience different genetic burdens in addition to other difficulties arising from non-biological factors which make them more prone to HNSCC. Thus, our results support a molecular disparity in laryngeal cancer between Eur-Amr and Afr-Amr as suggested by our mutational landscape analyses. Our study also suggests that genetic background may play an important role in mutational development and cancer progression and highlights the need for more data and specific analyses to identify the driver mutations in Afr-Amr patients. In addition, if we could compare the effects of driver mutations on laryngeal cancer risk and outcome, we may better address why Afr-Amr are at a higher risk for laryngeal cancer than Eur-Amr. More research into the effect size of the genomic level variations found in the laryngeal cancers of Afr-Amr and Eur-Amr as well as genome- environment interactions will further help us to understand the associated baseline risk disparity observed between these two populations.
Supplementary Material
Highlights.
The mutational landscapes of laryngeal cancer differ between African-American and European-American patients.
Tumors from African-American patients have a lower proportion of G>T changes, a tobacco related mutation, compared to European Americans.
Study reveals differences in frequently mutated genes between African-American and European-American patients.
At the molecular level, race appears to be a factor associated with the risk disparity observed between African-American and European-American for laryngeal cancer.
ACKNOWLEDGMENTS
Authors thank Dr. Bhawna Dubey and Dr. Sudhir Kumar from, respectively, Fox Chase Cancer Center and Temple University, for their valuable comments on data analyses and the manuscript. We thank Dr. Kara Maxwell, University of Pennsylvania for critical review of the MS. We also thank Dr. J. Robert Beck from Fox Chase Cancer Center for supporting this study. We gratefully acknowledge the efforts of TCGA network for making the data available for research use. This work was supported in part by grants RSG-14-033-01-CPPB from the American Cancer Society and CA006927 from the National Cancer Institute, and by an appropriation from the Commonwealth of Pennsylvania. We also acknowledge the Russian Government Program of competitive growth of Kazan Federal University for the partial funding support to IS. The results shown here are in whole or part based upon data generated by the TCGA Research Network: http://cancergenome.nih.gov/.
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Conflict of interest
The authors declare no conflict of interest.
References
- 1.Howlader N NA, Krapcho M, Garshell J, Miller D, Altekruse SF, Kosary CL, Yu M, Ruhl J, Tatalovich Z, Mariotto A, Lewis DR, Chen HS, Feuer EJ, Cronin KA. SEER Cancer Statistics Review, 1975-2011. National Cancer Institute; Bethesda, MD, Bethesda, MD: 2013. [Google Scholar]
- 2.Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. CA: a cancer journal for clinicians. 2014;64:9–29. doi: 10.3322/caac.21208. [DOI] [PubMed] [Google Scholar]
- 3.Ghafoor A, Jemal A, Cokkinides V, Cardinez C, Murray T, Samuels A, Thun MJ. Cancer statistics for African Americans. CA: a cancer journal for clinicians. 2002;52:326–341. doi: 10.3322/canjclin.52.6.326. [DOI] [PubMed] [Google Scholar]
- 4.Ragin CC, Langevin SM, Marzouk M, Grandis J, Taioli E. Determinants of head and neck cancer survival by race. Head & neck. 2011;33:1092–1098. doi: 10.1002/hed.21584. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Menvielle G, Luce D, Goldberg P, Bugel I, Leclerc A. Smoking, alcohol drinking and cancer risk for various sites of the larynx and hypopharynx. A case-control study in France. Eur J Cancer Prev. 2004;13:165–172. doi: 10.1097/01.cej.0000130017.93310.76. [DOI] [PubMed] [Google Scholar]
- 6.Ragin CCR, Modugno F, Gollin SM. The epidemiology and risk factors of head and neck cancer: a focus on human papillomavirus. J Dent Res. 2007;86:104–114. doi: 10.1177/154405910708600202. [DOI] [PubMed] [Google Scholar]
- 7.Benowitz NL, Hukkanen J, Jacob P., 3rd Nicotine chemistry, metabolism, kinetics and biomarkers. Handbook of experimental pharmacology. 2009:29–60. doi: 10.1007/978-3-540-69248-5_2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.McCoy GD, Wynder EL. Etiological and preventive implications in alcohol carcinogenesis. Cancer research. 1979;39:2844–2850. [PubMed] [Google Scholar]
- 9.Kandel DB, Hu MC, Schaffran C, Udry JR, Benowitz NL. Urine nicotine metabolites and smoking behavior in a multiracial/multiethnic national sample of young adults. American journal of epidemiology. 2007;165:901–910. doi: 10.1093/aje/kwm010. [DOI] [PubMed] [Google Scholar]
- 10.Berg JZ, Mason J, Boettcher AJ, Hatsukami DK, Murphy SE. Nicotine metabolism in African Americans and European Americans: variation in glucuronidation by ethnicity and UGT2B10 haplotype. The Journal of pharmacology and experimental therapeutics. 2010;332:202–209. doi: 10.1124/jpet.109.159855. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Benowitz NL, Perez-Stable EJ, Fong I, Modin G, Herrera B, Jacob P., 3rd Ethnic differences in N-glucuronidation of nicotine and cotinine. The Journal of pharmacology and experimental therapeutics. 1999;291:1196–1203. [PubMed] [Google Scholar]
- 12.Perez-Stable EJ, Herrera B, Jacob P, 3rd, Benowitz NL. Nicotine metabolism and intake in black and white smokers. JAMA : the journal of the American Medical Association. 1998;280:152–156. doi: 10.1001/jama.280.2.152. [DOI] [PubMed] [Google Scholar]
- 13.Ho MK, Mwenifumbo JC, Zhao B, Gillam EM, Tyndale RF. A novel CYP2A6 allele, CYP2A6*23, impairs enzyme function in vitro and in vivo and decreases smoking in a population of Black-African descent. Pharmacogenetics and genomics. 2008;18:67–75. doi: 10.1097/FPC.0b013e3282f3606e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mwenifumbo JC, Al Koudsi N, Ho MK, Zhou Q, Hoffmann EB, Sellers EM, Tyndale RF. Novel and established CYP2A6 alleles impair in vivo nicotine metabolism in a population of Black African descent. Human mutation. 2008;29:679–688. doi: 10.1002/humu.20698. [DOI] [PubMed] [Google Scholar]
- 15.Mwenifumbo JC, Zhou Q, Benowitz NL, Sellers EM, Tyndale RF. New CYP2A6 gene deletion and conversion variants in a population of Black African descent. Pharmacogenomics. 2010;11:189–198. doi: 10.2217/pgs.09.144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Wheeler HE, Gorsic LK, Welsh M, Stark AL, Gamazon ER, Cox NJ, Dolan ME. Genome-wide local ancestry approach identifies genes and variants associated with chemotherapeutic susceptibility in African Americans. PloS one. 2011;6:e21920. doi: 10.1371/journal.pone.0021920. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Cadoni G, Boccia S, Petrelli L, Di Giannantonio P, Arzani D, Giorgio A, De Feo E, Pandolfini M, Galli P, Paludetti G, Ricciardi G. A review of genetic epidemiology of head and neck cancer related to polymorphisms in metabolic genes, cell cycle control and alcohol metabolism. Acta otorhinolaryngologica Italica : organo ufficiale della Societa italiana di otorinolaringologia e chirurgia cervico-facciale. 2012;32:1–11. [PMC free article] [PubMed] [Google Scholar]
- 18.Kandoth C, McLellan MD, Vandin F, Ye K, Niu B, Lu C, Xie M, Zhang Q, McMichael JF, Wyczalkowski MA, Leiserson MD, Miller CA, Welch JS, Walter MJ, Wendl MC, Ley TJ, Wilson RK, Raphael BJ, Ding L. Mutational landscape and significance across 12 major cancer types. Nature. 2013;502:333–339. doi: 10.1038/nature12634. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Stransky N, Egloff AM, Tward AD, Kostic AD, Cibulskis K, Sivachenko A, Kryukov GV, Lawrence MS, Sougnez C, McKenna A, Shefler E, Ramos AH, Stojanov P, Carter SL, Voet D, Cortes ML, Auclair D, Berger MF, Saksena G, Guiducci C, Onofrio RC, Parkin M, Romkes M, Weissfeld JL, Seethala RR, Wang L, Rangel-Escareno C, Fernandez-Lopez JC, Hidalgo-Miranda A, Melendez-Zajgla J, Winckler W, Ardlie K, Gabriel SB, Meyerson M, Lander ES, Getz G, Golub TR, Garraway LA, Grandis JR. The mutational landscape of head and neck squamous cell carcinoma. Science. 2011;333:1157–1160. doi: 10.1126/science.1208130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Agrawal N, Frederick MJ, Pickering CR, Bettegowda C, Chang K, Li RJ, Fakhry C, Xie TX, Zhang J, Wang J, Zhang N, El-Naggar AK, Jasser SA, Weinstein JN, Trevino L, Drummond JA, Muzny DM, Wu Y, Wood LD, Hruban RH, Westra WH, Koch WM, Califano JA, Gibbs RA, Sidransky D, Vogelstein B, Velculescu VE, Papadopoulos N, Wheeler DA, Kinzler KW, Myers JN. Exome sequencing of head and neck squamous cell carcinoma reveals inactivating mutations in NOTCH1. Science. 2011;333:1154–1157. doi: 10.1126/science.1206923. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.N. Cancer Genome Atlas, Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature. 2015;517:576–582. doi: 10.1038/nature14129. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Krawczak M, Ball EV, Cooper DN. Neighboring-nucleotide effects on the rates of germ-line single-base-pair substitution in human genes. American journal of human genetics. 1998;63:474–488. doi: 10.1086/301965. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Malyarchuk BA. [The role of nucleotide context in the induction of mutations in human mitochondrial DNA genes] Genetika. 2005;41:385–390. [PubMed] [Google Scholar]
- 24.Gehring JS, Fischer B, Lawrence M, Huber W. SomaticSignatures: Inferring Mutational Signatures from Single Nucleotide Variants. Bioinformatics. 2015 doi: 10.1093/bioinformatics/btv408. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Hecht SS. Tobacco smoke carcinogens and lung cancer. Journal of the National Cancer Institute. 1999;91:1194–1210. doi: 10.1093/jnci/91.14.1194. [DOI] [PubMed] [Google Scholar]
- 26.Cooper CS. Smoking, lung cancers and their TP53 mutations. Mutagenesis. 2002;17:279–280. doi: 10.1093/mutage/17.4.279. [DOI] [PubMed] [Google Scholar]
- 27.Govindan R, Ding L, Griffith M, Subramanian J, Dees ND, Kanchi KL, Maher CA, Fulton R, Fulton L, Wallis J, Chen K, Walker J, McDonald S, Bose R, Ornitz D, Xiong D, You M, Dooling DJ, Watson M, Mardis ER, Wilson RK. Genomic landscape of non-small cell lung cancer in smokers and never-smokers. Cell. 2012;150:1121–1134. doi: 10.1016/j.cell.2012.08.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Pfeifer GP, Denissenko MF, Olivier M, Tretyakova N, Hecht SS, Hainaut P. Tobacco smoke carcinogens, DNA damage and p53 mutations in smoking-associated cancers. Oncogene. 2002;21:7435–7451. doi: 10.1038/sj.onc.1205803. [DOI] [PubMed] [Google Scholar]
- 29.Micol JB, Duployez N, Boissel N, Petit A, Geffroy S, Nibourel O, Lacombe C, Lapillonne H, Etancelin P, Figeac M, Renneville A, Castaigne S, Leverger G, Ifrah N, Dombret H, Preudhomme C, Abdel-Wahab O, Jourdan E. Frequent ASXL2 mutations in acute myeloid leukemia patients with t(8;21)/RUNX1-RUNX1T1 chromosomal translocations. Blood. 2014;124:1445–1449. doi: 10.1182/blood-2014-04-571018. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.