Prevalence of Aflatoxin-Associated TP53R249S Mutation in Hepatocellular Carcinoma in Hispanics in South Texas

Liver cancer is the second leading cause of cancer-related mortality worldwide with an estimated 745,000 deaths in 2012 (1, 2). In the United States, in contrast to the 25% decline in overall cancer mortality from 1991 to 2014, deaths from liver cancer have increased at the highest rate of all cancer sites and liver cancer incidence has increased sharply, second only to thyroid cancer (3). There are significant geographic and ethnic variations in the incidence of hepatocellular carcinoma (HCC), the major form of liver cancers, with the highest rates observed in Hispanics in South Texas (4, 5). The main risk factor of HCC is liver cirrhosis. Other risk factors contributing to HCC include chronic hepatitis B or C virus (HBV, HCV) infection, alcohol abuse, nonalcoholic steatohepatitis (NASH) and aflatoxin exposure (6). We previously reported that the prevalence of cirrhosis in Hispanics in South Texas is 0.94%, which is 4-fold higher than the national prevalence (7). Risk factors independently associated with cirrhosis in this population are central obesity, diabetes, HCV, and alcohol with a remarkable population attributable fraction of 65.3% from central obesity. In 20% of the cases, no known risk factor was identified (7).

In this article, we explored the possibility that aflatoxin exposure could be a contributing risk factor for HCC in Hispanics in South Texas. Aflatoxin is a mycotoxin produced by the common fungi Aspergillus flavus and Aspergillus parasiticus and is classified as group 1 carcinogen by the World Health Organization (8). Aflatoxin B1 (AFB1) is the most carcinogenic type of aflatoxin. AFB1 consumption and HCC are epidemiologically linked in much of the developing world, including Southeast Asia, China, and sub-Saharan Africa (9). The highly mutagenic AFB1-DNA adduct induces primarily G:C→T:A mutations and the molecular hallmark of AFB1 exposure and HCC risk is a specific mutation at codon 249 of the TP53 gene (10). This mutation is a single-base substitution at the third base of codon 249 (AGG to AGT), which replaces an arginine by a serine (R249S) (11). In regions with high aflatoxin exposure such as Gambia and Qidong in China, TP53R249S mutation has been detected in 35% to 61% of HCC tumors (12–14). In moderate aflatoxin exposure regions such as Thailand, TP53R249S mutation is present in 8% to 27% of HCC (15, 16). This mutation is detected in 19% of HCCs in Mexico and 16% of HCCs in Brazil, suggesting a medium exposure of aflatoxin in these countries (17, 18). In low aflatoxin exposure regions such as in Europe and the United States, TP53R249S mutation was only rarely found if at all, and patients with that mutation were often immigrants from Asia or Africa (19). TP53R249S mutation can also be detected in circulating cell-free DNA (cfDNA), with good correlation with aflatoxin exposure. TP53R249S mutation has been detected in cfDNA from 40% of HCC patients in Qidong, from 36% to 53% of HCC patients in Gambia and from 26% to 34% of HCC patients in Thailand (20–24). The TP53R249S mutation was not detected in cfDNA from European patients with HCC and was never measured in cfDNA from patients with HCC in the United States (24). Aflatoxin exposure is not believed to play any role in hepatocarcinogenesis in the United States. However, aflatoxin contamination has been reported in foods from Texas, particularly after periods of drought (25), and in an early study, AFB1 levels have been detected in HCC patients treated at MD Anderson Cancer Center in Houston (26). Although 1% of the general U.S. population had detectable levels of AFB1–albumin adduct in years following the severe drought of 1998, a study performed among Hispanics in Bexar County, Texas, showed that AFB1–albumin adduct was detected in 21% of the participants with a median level almost 5-fold higher than that detected in the U.S. population (27). A recent study in South Texas reported that HCC patients had higher serum and urine aflatoxin levels than matched controls (28). Therefore, because pockets of exposed persons have been reported in South Texas, in particular among Hispanics, and because of the high incidence of HCC in this population, we aimed to determine the prevalence of the TP53R249S mutation in HCCs from Hispanics in South Texas, particularly in counties with high liver cancer rates, and addressed the role aflatoxin exposure may play in the etiology of HCC in this population.

The study was approved by the Institutional Review Boards of all collaborating institutions. Formalin-fixed paraffin-embedded (FFPE) HCCs of 41 Hispanic patients were collected at the University of Texas Medical Branch, Galveston, and at Doctors Hospital at Renaissance, Edinburg (Texas). The demographic and clinical parameters of the 41 Hispanic patients with HCC are described in Supplementary Table S1. DNA and RNA extraction was performed from the tumor areas and from distant nontumoral liver, using QIAamp DNA FFPE Tissue Kit (Qiagen) and High Pure FFPET RNA Isolation Kit (Roche). Plasma samples from 218 histologically confirmed HCC patients were collected between 2002 and 2010, prior to treatment at MD Anderson Cancer Center (Houston, TX). The demographic and clinical parameters of the 218 HCC patients are described in Supplementary Table S2. Plasma samples were also collected from 96 participants of the Cameron County Hispanic Cohort (CCHC) with aspartate transaminase (AST) to platelet ratio index (APRI) scores ≥1, indicative of the presence of cirrhosis/advanced fibrosis. The demographic and clinical parameters of the 96 CCHC subjects were described previously (7). All these subjects were enrolled in 4 Texas counties with high rates of liver cancer. The 2014 liver cancer mortality rates in all 4 counties were significantly higher than the statewide rate of 8.74_/100,000, ranging from 9.51_/100,000 to 14.30_/100,000. Dietary information was obtained from a survey administered to CCHC participants (n = 2,606) and showed that 56% of Mexican Americans in Cameron/Webb are consuming corn tortillas at least once a day, including 19.1% consuming 3 or more corn tortillas daily. cfDNA was extracted from all plasma samples (500 μL) using QIAamp Circulating Nucleic Acid Kit (Qiagen). FFPE tumor DNA and cfDNA samples were quantified using Qubit Fluorometer and dsDNA High-Sensitivity Assay Kit (Thermo Fisher Scientific). Their quality was assessed using a fragment analyzer and high sensitivity Genomic DNA Analysis Kit (Advanced Analytical Technologies). RNA samples were assessed for quantity using Qubit Fluorometer with RNA High-Sensitivity Assay Kit (Thermo Fisher Scientific) and for quality using Tapestation with High-Sensitivity RNA Kit (Agilent Technologies).

TP53R249S mutation was detected using the QX200 droplet digital PCR (ddPCR) system (Bio-Rad Laboratories) and the following assays: dHsaCP2000088 for wild-type (wt) TP53 and dHsaCP2000087 for TP53R249S allele. Mutant and wt TP53 alleles were differentiated by the fluorophores attached to the probes, with HEX fluorescence for wt TP53 alleles and FAM fluorescence for mutant TP53R249S alleles. Four microliters of DNA was used in each reaction. Thermocycling conditions were: 95°C for 10 minutes, followed by 40 cycles of 94°C for 30 seconds and 55°C for 1 minute, followed by 98°C for 10 minutes. The sealed plates were then placed in the droplet reader for detection of complete ddPCR reactions in individual droplets. The data were analyzed using QuantaSoft software (Bio-Rad Laboratories, Inc.). Samples with a R249S allele fraction ≥0.1% and with at least 2 visible mutant signals were considered positive for the mutation as described previously (29, 30).

Exon 7 of TP53 was amplified with two rounds of PCR as described previously (24). Both PCR reactions involved a 15-minute hotstart DNA polymerase (KAPA Biosystems) activation at 95°C, 40 cycles at 94°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds, followed by a final 5-minute extension at 72°C. The pairs of primers were: F1-CTTGCCACAGGTCTCCCCAA and R1-AGGGGTCAGCGGCAAGCAGA; F2-AGGCGCACTGGCCTCATCTT and R2-TGTGCAGGGTGGCAAGTGGC. The PCR products were then digested by HaeIII restriction endonuclease (Promega Corporation) and separated on 3% agarose gel stained with ethidium bromide. PLC/PRF/5 and HepG2 hepatoma cell lines were used as positive and negative controls for TP53R249S mutation, respectively (31).

Barcoded, Illumina-compatible stranded total RNA libraries were prepared using the TruSeq Stranded Total RNA Sample Preparation Kit (Illumina). DNase I–treated RNA (250 ng) was depleted of cytoplasmic and mitochondrial ribosomal RNA using Ribo-Zero Gold (Illumina). After purification, the RNA was fragmented using divalent cations, and double-stranded cDNA was synthesized using random primers. The ends of the resulting double-stranded cDNA fragments were repaired, 5′-phosphorylated, 3′-A tailed and Illumina-specific indexed adapters were ligated. The products were purified and enriched with 12 cycles of PCR to create the final cDNA library. The libraries were quantified using the Qubit dsDNA HS Assay Kit and assessed for size distribution using Agilent Tapestation (Agilent Technologies), then multiplexed using 6–7 libraries per pool. Library pools were quantified by qPCR and sequenced on a HiSeq4000 sequencer using 75-bp paired-end format. Sequence files were generated in FASTQ format, and reads were mapped to human genome 19 and then aligned by TopHat2 (32). Htseq-count was used to generate gene read counts for each sample, and R package “DESeq2” was used to normalize the data. Genes with zero counts across all samples were removed. Feature-by-feature linear models were used to adjust for batch effects. Gene expression profiles were compared with feature-by-feature t tests, using a beta-uniform mixture (BUM) models to fit the resulting distribution of P values and allow for estimation of the FDR.

Liver HCC [The Cancer Genomic Atlas (TCGA)] was chosen on the cBioPortal online platform (http://www.cbioportal.org/; refs. 33, 34). The 373 sequenced HCC tumors were selected as the patient set. We queried for samples with TP53R249S mutation and identified 11 patients with TP53R249S mutation. Clinical and demographic information of all 373 HCC patients were downloaded from the cBioPortal website. Overall survival and disease-free survival were examined to compare the prognosis of the patients with TP53R249S mutation with those without that mutation. The results are displayed as Kaplan–Meier plots with P values from a log-rank test. Using the cBioPortal Enrichment module, we retrieved mRNA expression data from all HCC tumors and compared genes from tumors with TP53R249S mutant to nonmutant tumors using Student t tests.

Statistical difference between each group was assessed by Student t tests for continuous variables and by Fisher exact tests for categorical variables using GraphPad 6.0 software and R Version 3.3. A value of P < 0.05 was considered statistically significant.

We obtained FFPE tumor samples from 41 Hispanic patients with HCC treated at two institutions in South Texas. Demographic and clinical parameters of these 41 patients are summarized in Supplementary Table S1. These patients were predominantly male (82.9%; 34/41), 37.1% were obese, and 58.3% had diabetes. These HCCs were associated with HCV (75%) or NAFLD/NASH (15%), and 61.3% of these patients had underlying cirrhosis. The distribution of well, moderately, and poorly differentiated tumors was 38.9%, 41.7%, and 19.4%, respectively. Following DNA extraction, we measured TP53R249S mutation using ddPCR in these 41 HCCs. Droplets positive for mutant alleles, positive for wt alleles, and double-positive for both mutant and wt alleles were clearly separated as shown in Fig. 1. Black, blue, green, and orange dots represent empty droplets, TP53R249S-positive droplets, wt DNA-positive droplets, and double-positive wt and TP53R249S droplets, respectively. The assay was first validated using HepG2 hepatoma cell line, not harboring the TP53R249S mutation and PLC/PRF/5 hepatoma cell line, harboring the TP53R249S mutation (Fig. 1A–C). TP53R249S mutation was detected in 3 of the 41 HCCs (7.3%; HCC-T1, HCC-T17, and HCC-T41), with 33.5%, 66.5%, and 62.5% mutant allele fractions, respectively (Fig. 1D–F). TP53R249S mutation was not detected in nontumor-adjacent liver. The presence of TP53R249S mutation in these 3 tumors was further confirmed by RFLP (Supplementary Fig. S1). The assay was again validated using HepG2 and PLC/PRF/5 cell lines (Supplementary Fig. S1). After digestion by restriction enzyme HaeIII, two bands (92 and 66 bp) bands are generated in wt samples, such as HepG2, whereas samples harboring the TP53R249S mutation, in which the restriction site is destroyed, yield only one band of 158 bp as observed for PLC/PRF/5 (Supplementary Fig. S1). For all 3 HCC tumors identified positive for TP53R249S mutation by ddPCR, the PCR products resulted following restriction enzyme digestion, in a mixture of uncleaved 158-bp fragment and cleaved 92 and 66 bp fragments, confirming the presence of TP53R249S mutation in the HCC-T1, HCC-T17, and HCC-T41 DNA samples (Supplementary Fig. S1). There was a very good agreement observed between the ratio of uncleaved to cleaved fragments detected by RFLP and the mutant allele fractions measured by ddPCR for all 3 tumors. No difference in gender, the presence of diabetes, HCV, NAFLD/NASH, cirrhosis or obesity, tumor differentiation, and site of recruitment was identified between the HCCs with TP53R249S mutation and those wt for that mutation (Table 1). All subjects with mutated TP53R249S were obese (100%) compared with 33% for those subjects without TP53R249S mutation, and 66.7% of them had multiple tumors compared with 19.20% for those subjects without TP53R249S mutation. The age average of the patients with TP53R249S mutation was younger (mean age, 55.7) than the age average of the patients without TP53R249S mutation (mean age, 64.1).

We then measured TP53R249S mutation in cfDNA isolated from plasma collected from 218 patients with HCC, seeking care at MD Anderson Cancer Center. The ethnicity/race distribution among these 218 HCC patients was 54.6% (119) Non-Hispanic White, 16.1% (35) Hispanic, 4.6% (10) Asian, 7.8% (17) Black, and 17% (37) unknown. The demographic and clinical parameters of these 218 HCC patients are shown in Supplementary Table S2. The majority of the patients were male (72%). Among them, 37.9% were obese, 40.8% had diabetes, 32.5% were positive for HCV, 20.9% were positive for HBV, and 10.9% had NAFLD/NASH. Underlying cirrhosis was present in 54.5% of the patients, with 19.1% alcoholic cirrhosis and 35.4% nonalcoholic cirrhosis. No known risk factor (HCV, HBV, NAFLD/NASH, alcohol) could be identified in 7% of the patients. The distribution of well, moderately, and poorly differentiated tumors is 39.6%, 34.3%, and 26.1%, respectively. Among the 218 cfDNAs analyzed by ddPCR, 4 were positive for TP53R249S mutation (HCC-P177, HCC-P182, HCC-P199, and HCC-P207), with mutant fractions of 36.7%, 17.6%, 1.2%, and 44.4%, respectively (Fig. 2). The presence of the TP53R249S mutation in these 4 cfDNA samples was further confirmed by RFLP (Supplementary Fig. S2). Here again, there was a very good agreement observed between the ratio of uncleaved to cleaved fragments detected by RFLP and the mutant allele fractions measured by ddPCR for all 4 cfDNA samples. There was no difference in gender, etiology, alpha-fetoprotein levels, degree of differentiation, presence of cirrhosis, child score, or tumor stage among patients with TP53R249S mutation compared with those negative for this mutation (Table 2). There was a trend of younger age for those with TP53R249S mutation compared with those without mutation (55.8 vs. 65.1) and of shorter overall survival in patients with TP53R249S mutation (Supplementary Fig. S3). There was a highly significant difference in ethnicity/race distribution between patients with TP53R249S mutation and those negative for this mutation (P = 0.002). The TP53R249S mutation was indeed only detected in Hispanics and Asians, while Blacks, non-Hispanic Whites, and patients with unknown ethnicity were all negative for TP53R249S. The frequency of TP53R249S mutation in cfDNA from Hispanic HCC patients was 5.7%. Among the Hispanic HCC patients with TP53R249S mutation, one was born in Mexico, while the other was born in the United States. The other patients harboring the TP53R249S mutation were born in Asia.

Because TP53R249S mutation has been reported in subjects with cirrhosis in regions of high AFB1 exposure, we also measured TP53R249S mutation in plasma cfDNA of Hispanics in South Texas, with AST to APRI ≥1, predictive of the presence of cirrhosis or advanced fibrosis. To that end, we interrogated a community-based Hispanic cohort at the U.S.–Mexico border, the CCHC. We previously identified 102 CCHC subjects with APRI ≥1. These subjects were more likely to have diabetes and HCV, to present with higher body mass index (BMI), waist circumference, fasting triglyceride, glucose, and insulin levels than controls. The detailed clinical and demographic features of these subjects have been published previously (7). TP53R249S mutation was measured in cfDNA of 96 of these 102 subjects but was not detected in any of these samples.

To determine whether TP53R249S mutation is associated with a specific tumor transcriptomic signature, we generated gene expression profiles using RNA sequencing (RNA-Seq) on the same HCC tumors from Hispanics analyzed for TP53R249S mutation. We then analyzed the gene expression profiles according to the presence or absence of TP53R249S mutation. Using P < 0.05 and fold change ≥1.5, we identified 239 upregulated and 194 downregulated genes in HCCs with TP53R249S mutation compared with HCCs without TP53R249S mutation (Supplementary Table S3). The largest expression changes were observed for solute carrier family 6 member 11 (SLC6A11), seizure-related 6 homolog-like 2 (SEZ6L2), epiplakin 1 (EPPK1), solute carrier family 2 member 14 (SLC2A14), transmembrane protein 82 (TMEM82) and IL17 receptor E (IL17RE), all overexpressed in samples with TP53R249S mutation compared with samples without mutation, and for eukaryotic translation elongation factor 1, alpha-2 (EE1A2), homolog of odd Oz 2 (ODZ2), serum amyloid A2 (SAA2), and glutathione S-transferase theta 1 (GSTT1), all underexpressed in samples with TP53R249S mutation. Ingenuity Pathway Analysis (IPA) of all 433 genes identified cell-cycle control of chromosomal replication and role of BRCA1 in DNA damage response, as the top canonical pathways affected by TP53R249S mutation status (P = 1.10 × 10⁻⁹ and P = 8.85 × 10⁻⁵, respectively) and E2f, E2F4, as the top upstream regulators affected by TP53R249S mutation status (P = 1.62 × 10⁻¹² and 4.39 × 10⁻¹¹, respectively; Table 3).

To compare our results of TP53R249S mutation in HCC in Hispanics in South Texas to other HCC population cohorts, we interrogated TCGA data using the cBioPortal for Cancer Genomics (http://www.cbioportal.org/). TP53R249S mutation was detected in 11 of the 373 HCC tumors in TCGA (2.9%). There was no difference in gender, presence of cirrhosis, viral hepatitis or NAFLD/NASH, or child score among patients with TP53R249S mutation compared with those negative for this mutation (Supplementary Table S4). Patients with TP53R249S mutation had early onset of disease with an average age of diagnosis at 48.8 compared with 59.9 for those without this mutation (P = 0.006; Supplementary Table S4). A significant difference in ethnicity/race distribution in patients harboring TP53R249S mutation was also observed, with TP53R249S mutation detected in Asians (72.7%), Blacks (18.2%), and Hispanics patients (9.1%) but not in non-Hispanic Whites (P = 0.004). Patients harboring TP53R249S mutation were more likely to have stage II (40%) and III (50%) disease than those not harboring the mutation (24.1% and 23.8%, respectively, P = 0.036). Kaplan–Meier plot analysis showed that the presence of TP53R249S mutations was significantly associated with shorter overall survival and short disease-free survival (P = 0.008 and 0.001, respectively; Supplementary Fig. S4A). We also used the cBioPortal for Cancer Genomics (http://www.cbioportal.org/) to further assess the impact of TP53R249S mutation on gene expression in HCCs in TCGA. IPA analysis of genes differentially expressed in tumors with TP53R249S mutation, identified as for the HCCs in Hispanics in South Texas, the same top canonical pathways, namely cell-cycle control of chromosomal replication and role of BRCA1 in DNA damage response (P = 3.55 × 10⁻²⁶ and 3.09 × 10⁻¹⁴) and the same top upstream regulators, namely E2F4 and E2F1 (P = 6.12 × 10⁻⁶¹ and 4.29 × 10⁻⁴¹; Supplementary Fig. S4B).

The aim of this study was to determine whether aflatoxin dietary exposure plays a role in the etiology of HCC in Hispanics in South Texas. To that end, we measured a hallmark mutation induced by aflatoxin exposure in HCC, namely TP53R249S. Because this mutation was previously detected in both tumor tissue and plasma cfDNA of patients with HCC but also patients with liver cirrhosis (23, 35), we measured this mutation in HCC tumors from Hispanics in South Texas, in plasma cfDNA from patients with HCC, seeking care at MD Anderson Cancer Center and in plasma cfDNA from Hispanics with advanced fibrosis or cirrhosis, participants of the community-based CCHC. All subjects were enrolled in Texas counties with very high rates of liver cancer (36, 37). The national mortality rate of liver cancer in 2014 was 6.81_/100,000 but reached 8.74_/100,000 in Texas. The samples used in our study were collected in counties along the U.S.–Mexico border (Cameron and Webb counties) and along the Gulf Coast (Galveston and Harris counties). The 2014 liver cancer mortality rates in all 4 counties were significantly higher than the statewide rate, ranging from 9.51_/100,000 to 14.30_/100,000. All 4 counties were among the top 17% of Texas counties with the highest rates of liver cancer. Therefore, the tumor samples and other biological materials analyzed in this study were indeed from hotspots for liver cancer in Texas. From the dietary surveys in CCHC and in the National Health and Nutrition Examination Survey (NHANES) 2005–2006 (38), we extracted data on daily consumption of grain and of corn tortillas in Mexican Americans in CCHC (n = 2606) and in Mexican Americans nationwide (n = 1397). Although the overall consumption of grain products was similar for both groups (87% and 94%, respectively), Mexican Americans in Cameron/Webb counties in Texas consumed significantly more corn tortillas than Mexican American nationwide, with 56% versus 20% consuming corn tortillas at least once a day, respectively. A total of 19.1% of Mexican Americans in Cameron/Webb counties consumed 3 or more corn tortillas daily.

The technologies we used for the detection of TP53R249S mutation were RFLPs and ddPCR. Although RFLP is a wildly used technique for the detection of TP53R249S mutation based on the modification of a restriction site by the mutation (14, 24, 35), ddPCR is a novel highly sensitive and robust technology for detection of rare mutations using massive sample partitioning and fluorescence-based detection (39, 40). We observed a good agreement between RFLP and ddPCR results.

Overall, we found TP53R249S mutation present in 7.3% (3/41) of Hispanic HCC tumors collected from two different sites in South Texas. This is higher, although not significantly, than the overall prevalence found in TCGA (11/373, 2.9%), even among Hispanics (1/18, 5.5%). We also reported for the first time, the detection of TP53R249S mutation in plasma cfDNA in 218 HCC patients. Although, indeed, such studies have been done in other countries (13, 22, 41), this has never been done in the HCC population in the United States. We found a significant difference in ethnicity/race distribution of patients with TP53R249S mutation (P < 0.0001). This mutation was only detected in Hispanics and Asians. The frequency of TP53R249S mutation in cfDNA from Hispanic HCC patients was 5.7%. Of importance, TP53R249S mutation was found in a Hispanic born in the United States, suggesting that although the role of aflatoxin exposure in HCC among Hispanic in South Texas is low, we cannot exclude it.

Patients harboring TP53R249S mutation were likely to be younger (55.7 vs. 65.0, P = 0.039). This result was further confirmed in HCCs from TCGA and is in agreement with a study performed in Thailand, a country with moderate aflatoxin exposure (15). Patients with TP53R249S mutation also presented with a significantly reduced overall survival and disease-free survival. TP53R249S mutation has been shown to be a prognostic marker in a Chinese cohort with HCC patients from high and moderate aflatoxin exposure areas (19).

A limitation of the study is the lack of examination of the correlation with actual aflatoxin dietary exposure in those individuals with Tp53R249S. Future studies should include the measurements in the same individuals, of plasma biomarkers of covalent adduction to DNA or protein or urinary AFB1-DNA repair products together with the measurement of Tp53R249S in plasma cfDNA. Such studies would be particularly valuable in prospective cohorts with serial biospecimens, allowing for a time-to-event correlation analysis between these biomarkers.

Integrative analysis of TP53R249S mutation status and RNA-Seq transcriptomic data from both the HCC tumors from Hispanics in South Texas and the TCGA HCCs, identified cell-cycle control of chromosome replication and role of BRCA1 in DNA damage response as the top two canonical pathways in HCCs with TP53R249S mutation. Several genes overexpressed in HCCs with TP53R249S mutation are known to be associated with poor survival in HCC and other human cancers. These include SLC2A14 and IL17RE (42). EPPK1 serves as a useful marker for hepatic and pancreatic progenitor cells (43, 44). Among the genes underexpressed in HCCs with TP53R249S mutation, null genotype of GSTT1 increases the risk of lung cancer, prostate cancer, and HCC (45). BRCA1 and E2F1 mRNA expression was increased in tumors bearing TP53R249S mutation in both datasets. Overexpression of BRCA1 has been reported in HCC and shown to correlate with mesenchymal-like feature and chemoresistance (46). Members of the E2F family, E2F1 and E2F4, were identified as upstream regulators of the genes associated with TP53R249S mutation. E2F family of transcription factors plays vital roles in cell proliferation, apoptosis, differentiation, senescence, DNA damage response, and DNA repair (47). In HCC, E2F1 has both been shown to have proapoptotic and antiapoptotic roles, in addition to proliferative effects (48). Copy number gain of E2F1 resulted in dosage-dependent spontaneous HCC in mice, suggesting a direct and cell-autonomous role for E2F in HCC (49). Microsatellite instability and mutations of E2F4 commonly occur in HCC and may play an important role in hepatocarcinogenesis (50). Our results demonstrate for the first time an association between TP53R249S mutation with E2F network, suggesting that aflatoxin exposure could promote HCC onset through E2F. Our results also demonstrate for the first time an association between TP53R249S mutation with BRCA1 pathway. Whether BRCA1/BRCA2 alterations could increase susceptibility to HCC onset in the context of aflatoxin dietary exposure should be further evaluated.

In conclusion, TP53R249S mutation was detected in 5.7% to 7.3% of Hispanic patients with HCC in South Texas. This mutation was associated with development of HCC at a young age and worse prognosis. TP53R249S was however not detected in cfDNA from Hispanics in South Texas with cirrhosis or advanced fibrosis. Therefore, aflatoxin exposure may contribute to a small number of HCCs in Hispanics in South Texas, but the detection of TP53R249S mutation in plasma cfDNA is not a promising risk predictor marker in subjects with cirrhosis or advanced fibrosis in this population.

No potential conflicts of interest were disclosed.

Conception and design: S.P. Fisher-Hoch, L. Beretta

Development of methodology: J. Jiao

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J. Jiao, Y. Ye, J.L. Almeda, M.M. Betancourt-Garcia, H.L. Stevenson, J.B. McCormick, S.P. Fisher-Hoch

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J. Jiao, W. Niu, Y. Wang, K. Baggerly, G.P. Watt

Writing, review, and/or revision of the manuscript: J. Jiao, Y. Wang, Y. Ye, X. Wu, R.A. Forse, H.L. Stevenson, G.P. Watt, J.B. McCormick, S.P. Fisher-Hoch, L. Beretta

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D. Davenport, M.M. Betancourt-Garcia, R.A. Forse, J.B. McCormick, L. Beretta

Study supervision: R.A. Forse, S.P. Fisher-Hoch, L. Beretta

We would like to thank Haidee Chancoco from the Biospecimen Extraction Facility of MD Anderson for help with RNA/DNA extraction and Denaha J Doss from the Sequencing and Microarray Facility for help with ddPCR. We would like to thank the Center for Translational and Public Health Genomics at University of Texas MD Anderson Cancer Center (XW), and Dr. Yang Deng for biospecimen and associated clinical information.

The study was supported by the NIH/NCI under award number P30CA016672 and use of the Sequencing and Microarray Facility, by UTHealth Innovation for Cancer Prevention Research Training Program Post-doctoral Fellowship (Cancer Prevention and Research Institute of Texas grant # RP160015, to J. Jiao), by the Trans-Texas HCC Study and a Start-Up fund from The University of Texas MD Anderson Cancer Center (to L. Beretta).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.