Search

Supplementary Tables 1-6. Supplemental Table 1. Primers used in Ion Torrent sequencing of HPV. Supplemental Table 2. Predicted somatic mutations in PIK3CA discovered by targeted sequencing of cervical tumors from Mexico, Venezuela, and Guatemala. Supplementary Table 3. Predicted somatic mutations in 5 candidate genes discovered by targeted sequencing of cervical tumors. Supplementary Table 4. Association of HPV type and PIK3CA mutation. Supplementary Table 5. PIK3CA mutations in histologic subtype in this and other published studies Supplementary Table 6. Fig.3A_B_C raw data

Supplementary Figures 1-11. Supplementary Figure 1. Sample and analyses diagram. Supplementary Figure 2. Aggregate mutation spectrum of Guatemalan cervical tumors. Supplementary Figure 3. Highly amplified regions of chromosome (Chr.) 2, 6 and 10 in tumor A4. Supplementary Figure 4. Chromosome breakpoints in cervical tumors. Supplementary Figure 5. PIK3CA mutation and pathology. Supplementary Figure 6. Co-occurring mutations in the PIK3CA and PTEN genes. Supplementary Figure 7. Sanger sequence concordance of mutant read percentage. Supplementary Figure 8. Mutation Fraction of PIK3CA mutations. Supplementary Figure 9. Age range associations with PIK3CA mutation and HPV type. Supplementary Figure 10. PIK3CA and HPV E6 and E7 mRNA expression. Supplementary Figure 11. Model of cervical cancer progression.

Detecting differentially expressed genes under different biological conditions is crucial to characterize mechanisms of cancer development and identifying determinants of patient outcome. High throughput technologies such as RNA sequencing and mass spectrometry-based proteomics have been widely used to identify differentially expressed genes (DEG), on a transcript and on a peptide basis, respectively. However, both transcription and translation of genes provide information about gene expression. Thus, leveraging both RNA and protein expression data could potentially produce more accurate results. Various statistical tools have been developed to tackle the differential expression problem for a single platform, such as edgeR, DESeq2, etc. However, a tool that integrates both transcriptome and proteomics data for differential expression analysis has not yet been developed. Meta-analysis can potentially increase statistical power and generate a more consensus conclusion by combining multiple datasets. Here we present DE-Meta, a new tool developed using R, which performs combined meta-analysis of RNA-seq and MS-based proteomics on matched tumor specimens. To characterize the performance of our method using real world data, we utilized a published lung cancer dataset and set a truth standard of DEG between known two expression subtypes: terminal respiratory unit (TRU) and proximal proliferative (PP). Briefly, we restricted to 7,458 common genes detected by RNA-seq and MS-based proteomics. We then, randomly selected tumors as the truth sample set, n=20 TRU and n=20 PP tumors. With this truth sample set, DEGs were calculated separately by RNA or by protein, using DESeq2 and regression models. The union of these DEGs defined the truth DEG set. Using independent tumors (n=10 TRU and n=10 PP), we applied DE-meta to detect DEGs by RNA, by protein or joint RNA and protein data. Through receiver operating characteristic curve analysis, we show that the DE-meta joint model performed better than the single platform models (joint AUC=0.651, RNA AUC=0.610, protein AUC=0.623). Focusing genes relevant to tumor biology, we found that the DE-meta joint model called an 169 cancer-relevant DEGs, such as MSH2, CD4, WDR3, PIK3R1, and RPS6KA3. which were not detected by the RNA or protein models, at the same false discovery rate threshold of 0.05 (Benjamini-Hochberg procedure). In summary, our new DE-meta tool, by joint analysis of RNA-seq and MS-based proteomics datasets, yields more accurate and biologically-relevant results. The views expressed in this abstract are solely of the authors and do not reflect the official policy of the Departments of Army/Navy/Air Force, Department of Defense, USUHS, HJF, or U.S. Government. Citation Format: Xijun Zhang, Clifton L. Dalgard, Matthew D. Wilkerson. DE-meta: Revealing tumor gene expression by meta-analysis of RNASeq and proteomics datasets [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 3168.

Introduction: DNA damage repair genes (DDRGs) play a critical role in genomic stability and their dysfunction contributes to mutagenesis in several cancer types. In prostate cancer (CaP) emerging data provide potential roles of DDR pathways in aggressive disease. However, the association with disease progression and therapeutic stratification based on inherited mutations of DDRGs remains to be defined in African American (AA) CaP patients. Our objective was to genomically profile all known annotated DDRGs in AA and Caucasian American (CA) CaP patient to determine whether DDRG germline variation status can refine patient stratification for targeted therapeutic options. Methods: Germline mutations in all DDRGs (N=276) was evaluated by whole genome sequence (WGS) analysis of archived blood DNA samples from 600 CaP patients (300 AA and 300 CA) who underwent primary treatment at Walter Reed National Military Medical Center. The WGS mean coverage exceeded 37x. Principal Component Analysis (PCA) was used infer axes of genetic variation within AA men and examine individual and population clustering to predict ancestry of each sample using the Peddy program. Variant frequencies in CPDR CaP patients were compared to variant frequencies available from the Exome Aggregation Consortium (ExAC) control cases with no CaP by Fisher's Exact Test, using false discovery rate adjusted p-values. Results: Interrogation of the complete known DDRG set of 276 genes revealed several known and novel mutations in this cohort. The Pathogenic/likely pathogenic (P/LP) variant carrier rate was higher than reported before (23.5%) in both AA and CA patients. However, the analysis revealed that more than 2/3 of the identified 47 DDRGs with P/LP mutation were different between AA and CA patients. Unlike in CA patients, several RAD family genes (RAD51, RAD54L, RAD54B), PMS2, and BRCA1 were among the most frequently mutated DDRGs in AA patients, but not in CA patients. The most frequent (over 1% carrier frequency) and potentially targetable type of mutations were independently validated by ddPCR. These genes are part of targetable DDRG pathways (homologous recombination and mismatch repair), suggesting that targeted therapy could potentially benefit AA patients. AA men harbor more potentially targetable DDRG germline mutations (over 10%) than CA men which may contribute to addressing CaP disparity. Germline mutations in any of the DDRG genes was associated with shorter time to BCR (Kaplan-Meier analysis, log rank p value 0.044) in AA patients, but not in CA patients. Conclusion: Our findings highlight distinct racial differences in DDRGs and addresses the clinical utility by targeted therapy across AA and CA men. The percentage of patients with DDRG germline variation is of suitable threshold (23%) to consider early genetic testing for them in both AA and CA patients. Citation Format: Indu Kohaar, Xijun Zhang, Shyh-Han Tan, Darryl Nousome, Kevin Babcock, Lakshmi Ravindranath, Gauthaman Sukumar, Elisa Mcgrath-Martinez, John Rosenberger, Camille Alba, Amina Ali, Denise Young, Yongmei Chen, Jennifer Cullen, Inger Rosner, Isabella Sesterhenn, Albert Dobi, Gregory Chesnut, Clesson Turner, Clifton Dalgard, Matthew Wilkerson, Shiv Srivastava, Gyorgy Petrovics. Germline mutation landscape of all DNA repair genes in African American prostate cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2074.

Lung adenocarcinoma is a highly lethal tumor that displays extensive molecular heterogeneity of which deep characterization may drive therapeutic development and improve clinical outcomes. Through the Applied Proteogenomics Organizational Learning and Outcomes (APOLLO) research network, we utilized five molecular profiling technologies (DNA whole genome sequencing, RNA sequencing, total and phospho-proteomics by mass spectrometry, and reverse phase protein arrays [RPPA]) to characterize a longitudinally-annotated cohort of 87 lung adenocarcinomas. Through whole genome sequencing, we identified molecular signatures from patterns of somatic SNVs, indels, and large structural alterations that stratified tumors into three groups associated with patient smoke exposures. We also identified TP53, EGFR, KRAS, and STK11 as recurrently mutated genes, which together represent 80% of the cohort, in addition to genes mutated in smaller cohort subsets (e.g. RBM10), fusion genes, and pathogenic germline variants. To characterize tumor proteomes, we quantified >7,000 proteins and >10,000 phosphopeptides by mass spectrometry and >300 species by RPPA. Matched RNAs and proteins were typically positively correlated across samples (median ρ = 0.49). Through quantitative trait loci analyses, we identified genes whose RNA and protein expression levels were significantly modified by somatic mutations. We then classified tumors into RNA expression subtypes and found coordinated proteogenomic alterations and distinct clinical associations: terminal respiratory unit subtype – EGFR mutations and RNA/protein overexpression, acinar histology, non-smokers; proximal-proliferative subtype – STK11 mutations and RNA/protein underexpression, high smoking signature; proximal-inflammatory subtype – high tumor mutational burden. We also identified phospho-peptide signatures associated with these subtypes, including downregulation of CDK1/2 targets in terminal respiratory unit tumors. Protein co-expression network analysis discovered biologically-diverse pathway activities of the RNA expression subtypes. To interrogate somatic mutations in the context of molecular pathways, we projected DNA alterations onto known interaction networks and identified four subtypes with markedly distinct proteomic and microenvironment characteristics. Finally, several molecular characteristics were found to significantly predict patient outcomes, including RNA expression subtype classification against metastasis-free survival. Thus, our integrative, proteogenomic characterization of lung adenocarcinoma uncovered novel tumor biology and identified potential molecular markers for predicting patient outcomes. The views expressed in this abstract are solely of the authors and do not reflect the official policy of the Departments of Army/Navy/Air Force, Department of Defense, USUHS, HJF, or U.S. Government. Citation Format: Anthony R. Soltis, Nicholas W. Bateman, Thomas P. Conrads, Clifton L. Dalgard, Hai Hu, Teri J. Franks, Jianfang Liu, Daoud Meerzaman, Emanuel F. Petricoin, Qingrong Chen, Chunhua Yan, Xijun Zhang, Clesson E. Turner, The APOLLO Research Network, Craig D. Shriver, Christopher A. Moskaluk, Robert F. Browning, Matthew D. Wilkerson. Comprehensive proteogenomic analysis and classification of lung adenocarcinoma [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5893.

Background: DNA damage repair genes (DDRGs) play a critical role in protecting genome integrity and have been implicated in several cancer types. In the context of prostate cancer (CaP) emerging data provide potential role of this pathway in aggressive disease. It has also been recently demonstrated that PARP inhibitors can extend overall survival in metastatic patients with DDRG mutations. However, the association and therapeutic stratification based on inherited mutations of DDRGs remains to be defined in African American (AA) CaP patients. Our objective was to assess the frequency and association with disease aggressiveness of all known DDRGs in blood derived genomic DNAs of AA and Caucasian American (CA) CaP patients archived at the DOD Center for Prostate Disease Research and assess how this information can refine patient stratification for specific targeted therapeutic options. Method: Germline mutations in all DDRGs was evaluated by whole genome sequence (WGS) analysis of archived blood DNA samples from 600 CaP patients (300 AA and 300 CA) who underwent primary treatment at Walter Reed National Military Medical Center (WRNMMC) over the past 20 years. These patients had equal access Department of Defense healthcare system with up to 20 years of follow-up time. Following quality control steps to assess DNA quantity by Qubit assay and DNA quality by Bioanalyzer assay, DNA samples were used to generate PCR-free libraries for WGS using the NovaSeq (Illumina) platform. Out of the 600 libraries we generated, 14 dropped out, achieving a success rate of 97.6%. The successful libraries had an excellent quality based on DNA library metrics (yield and fragment length). Whole genome sequencing depth of the samples exceeded 37x on average and about 4 million SNPs were identified in the samples. Patient genotypes were projected onto principal components from reference populations. Sample ancestries were predicted by using the “Peddy” program, which uses a machine learning model trained on individuals of diverse ancestries from the 1000 Genomes Project reference panel. Due to mismatched ancestry, 33 samples were excluded from further analyses. An additional 17 samples were excluded due to higher than minimal noise level based on analyses using “ContEst” tool from Broad GATK package and Illumina noise percent values. Results: Interrogation of an inclusive DDRG set of 180 genes predicted to have non-silent effects on the protein sequence (e.g. missense, nonsense, frameshift) for germline mutations in this cohort revealed several known and novel mutations. The analysis has not been finalized yet. Mutation data will be assessed for association with the extensive available clinical and pathological data, including disease progression (metastasis), family history and African ancestry. Novel mutations, anticipated especially in the understudied AA context, will be analyzed for functional impact. Citation Format: Kevin Babcock, Xijun Zhang, Matthew Wilkerson, Clifton L. Dalgard, Shyh-Han Tan, Lakshmi Ravindranath, Yongmei Chen, Jennifer Cullen, Shiv Srivastava, Inger L. Rosner, Gyorgy Petrovics. Defining germline mutations of DNA damage repair genes in African American prostate cancer patients [abstract]. In: Proceedings of the Twelfth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2019 Sep 20-23; San Francisco, CA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2020;29(6 Suppl_2):Abstract nr B075.

Background: Lung squamous cell carcinoma (SqCC) arises in the epithelial layer of the bronchial airway and is often preceded by the development of premalignant lesions. However, not all premalignant lesions progress to lung SqCC and many will regress spontaneously. Understanding the somatic alterations and molecular subtypes associated with progression will allow us to identify biomarkers for early detection and develop therapeutic strategies for disease prevention and interception. Methods: Biopsies were obtained from high-risk smokers undergoing lung cancer screening by auto-fluorescence bronchoscopy and CT at the Roswell Park Cancer Institute. For each subject, multiple sites were sampled repeatedly over time. One biopsy from each region was sent for pathological review while another biopsy was taken for molecular studies. Whole-exome sequencing (WES) was performed at Uniform Services University to 120x coverage and RNA-seq was performed at Boston University School of Medicine. Results: The median number of somatic mutations across all premalignant lesions that underwent DNA-seq (150 biopsies from 20 subjects) was 0.45 per megabase and displayed a modest association with histological grade (p=0.05). The most frequently mutated known lung cancer genes included NOTCH1 (14%), TP53 (6%), FAT1 (3%), PIK3CA (2%), KRAS ( Citation Format: Joshua D. Campbell, Xijun Zhang, Catalina Perdomo, Sarah Mazzilli, Yaron Geshalter, Samjot S. Dhillon, Gang Liu, Sherry Zhang, Hanqiao Liu, Jessica Vick, Christopher Moy, Stefano Monti, Evan Johnson, Matthew Meyerson, Matthew Wilkerson, Clifton Dalgard, Suso Platero, Chris Stevenson, Marc Lenburg, Mary Reid, Jennifer Beane, Avrum Spira. Genomic characterization of premalignant lung squamous cell carcinoma lesions [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 3248.

Background: Accurate detection of somatic mutations is critical for informing targeted therapy options. Prevalent non-cancer cell admixture complicates this detection in breast cancer. Conventional mutation detection relies on DNA sequencing; however in prior work, we demonstrated that combining RNA and DNA sequencing increases mutation signal strength, or mutant allele fraction (MAF). The ratio of RNA MAF versus DNA MAF (RNA:DNA MAF) was greatest in low purity breast tumors. We hypothesized that this elevation is biologically driven and would be conserved in a second, distinct tissue specimen of the same tumors. Here, we compare mutation characteristics between two tissue blocks in a cohort of breast tumors (n = 8) to evaluate possible preservation of RNA versus DNA mutation signal throughout the tumor. Methods: We selected four high purity and four low purity breast tumors (“Block1”) from The Cancer Genome Atlas (TCGA) cohort and associated ABSOLUTE purity analysis. For these tumors, we acquired a second tissue block (“Block2”) not analyzed by TCGA, cut sections, analyzed sections by H&E stains, and extracted nucleic acids. Whole genome DNA sequencing and mRNA sequencing was performed for Block2 specimens using Illumina X and NextSeq 500 sequencers, respectively. Somatic mutations in Block2 were detected using UNCeqR and compared to published UNCeqR somatic mutations from TCGA. We then evaluate MAF characteristics in the entire TCGA breast tumor cohort (n = 695). Results: Tumor purity estimates, determined by histology and by sequencing, were reduced in Block2 of the low purity tumor set versus the high purity tumor set, consistent with Block1 analysis. Molecular properties of genome-wide gene expression and somatic DNA copy number were highly similar between block-mated specimens (p < 0.01). We then identified expressed mutations present in Block1 and Block2 of the same tumor and compared the MAFs on these common mutations. DNA MAF and RNA MAF were each significantly correlated between Block1 and Block2 (p < 1e-12 in both cases). The average RNA:DNA MAF was 2.5 for the cohort, indicating that RNA mutation signal is greater than DNA in general. In Block2 specimens, the RNA:DNA MAFs were significantly greater in the low purity tumor set than the high purity tumor set (mean 2.7 versus 2.1, p < 6e-5), reflecting the same trend observed in Block1 specimens. Analyzing the entire TCGA cohort, RNA:DNA MAF was positively correlated with proliferation pathway gene expression (p < 3e-16 ) and was greatest in the Basal subtype versus other subtypes (p < 2e-9). Conclusion: Mutant allele fraction both of DNA and of RNA was conserved across breast tumor subsections. Low purity and basal subtype breast tumors had elevated RNA:DNA MAF supporting a relationship to underlying biology and identifying classes of tumors with pronounced benefit for DNA and RNA integrated mutation analysis. Citation Format: Jerez Te, Coralie Viollet, Xijun Zhang, Jatinder Singh, Jeffrey A. Hooke, Harvey B. Pollard, Hai Hu, Craig D. Shriver, Clifton L. Dalgard, Matthew D. Wilkerson. Reproducible elevation of RNA versus DNA mutation signal in low purity breast tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 5403. doi:10.1158/1538-7445.AM2017-5403

Background: Lung squamous cell carcinoma (SqCC) arises in the epithelial layer of the bronchial airway and is often preceded by the development of premalignant lesions. However, not all premalignant lesions progress to lung SqCC and many regress without therapeutic intervention. Understanding the somatic alterations that contribute to progression of premalignant lesions in the airway will allow us to identify biomarkers for early detection and develop therapeutic strategies for early intervention. Methods: Airway biopsies were obtained from high-risk smokers undergoing lung cancer screening by auto-fluorescence bronchoscopy and chest CT at the Roswell Park Cancer Institute. For each subject (n=30), multiple premalignant lesions were sampled repeatedly over time (n=144 samples). One biopsy from each region was sent for pathological review while another biopsy was taken for molecular studies. DNA was also isolated from the blood or cytologically normal bronchial brushings to serve as a matched normal control. Exome capture was performed using the Illumina TruSeq Rapid Exome kit and sequenced to a mean depth of coverage of 120x at Uniform Services University and Walter Reed National Military Medical Center. Results: The median number of somatic mutations across all premalignant lesions was 0.73 per megabase (range: 0.10 - 9.8 per Mb) and displayed a modest association with histological grade (p=0.07). The most frequently mutated lung cancer genes included KMT2C (12%), NOTCH1 (11%), FAT1 (6%), TP53 (5%), and CDKN2A (7/Mb) were taken from adjacent sites over two time points in the same individual with a prior history of lung squamous cell carcinoma. These lesions had a significantly overlapping set of mutations including FAT1 indicating a common evolutionary ancestor. Conclusions: The somatic alterations observed in known cancer genes such as TP53, KMT2C, NOTCH1, and FAT1 may be among the earliest driver events in lung SqCC development and may be useful as biomarkers for early detection as well as targets for lung cancer interception. Citation Format: Joshua Campbell, Xijun Zhang, Samjot S. Dhillon, Catalina Perdomo, Sarah Mazzilli, Yaron Geshalter, Gang Liu, Sherry Zhang, Hanqiao Lin, Jessica Vick, Christopher Moy, Stefano Monti, Evan Johnson, Matthew Meyerson, Steven Dubinett, Suso Platero, Matthew Wilkerson, Clifton Dalgard, Marc Lenburg, Mary Reid, Jennifer Beane, Avrum Spira. The genomic landscape of premalignant lung squamous cell carcinoma lesions [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3259. doi:10.1158/1538-7445.AM2017-3259

Osteosarcoma, the most common primary bone malignancy, has a bimodal age distribution, with a primary peak in adolescence and a smaller peak in the elderly. The etiologic contribution of germline genetic variation to osteosarcoma is not well-understood. It occurs at higher than expected frequencies in individuals with the Li-Fraumeni syndrome (LFS) cancer predisposition syndrome. Approximately 70% of classic LFS families have germline TP53 mutations. Two previous studies reported that 3% of young osteosarcoma cases ( We determined the prevalence of germline TP53 mutations in 765 unselected osteosarcoma cases. DNA was extracted from blood and TP53 sequenced using custom Ampliseq panels. Variants were validated with Sanger sequencing. The IARC germline TP53 database was used to identify TP53 mutations reported in families with LFS (LFS-associated mutations). Variants were considered “likely LFS-associated mutations” if absent from publically available databases (ESP and 1,000 Genomes Project) and predicted non-functional or deleterious using in silico algorithms. Variants were considered “rare exonic variants” if their minor allele frequency (MAF) was There were 32 LFS-associated or rare TP53 variants in 62 osteosarcoma cases. The frequency of cases with an LFS or likely LFS-associated mutation and/or rare exonic variant was 8.1%. Notably, all 32 TP53 variants were present in cases A logistic regression case-case analysis identified a novel significant association between a rare TP53 variant, rs1800372 (p.R213R), and metastasis at diagnosis in cases of European ancestry (odds ratio [OR] 4.27, 95% CI 1.2-15.5, P = 0.026). We additionally confirmed that a common exonic variant, rs1042522 (p.P72R), was significantly associated with osteosarcoma risk (OR 1.22, 95% CI 1.1-1.4, P = 0.0098) and poorer survival (HR 1.35, 95% CI 1.00-1.83, P = 0.048). Our data suggest that genetic susceptibility to young onset osteosarcoma is distinct from adult onset osteosarcoma. The high TP53 mutation prevalence we identified in osteosarcoma cases aged Citation Format: Lisa J. Mirabello, Meredith Yeager, Phuong L. Mai, Julie Gastier-Foster, Richard Gorlick, Chand Khanna, Ana Patiño-Garcia, Luis Sierrasesúmaga, Fernando Lecanda, Irene L. Andrulis, Jay S. Wunder, Nalan Gokgoz, Donald A. Barkauskas, Xijun Zhang, Aurelie Vogt, Kristine Jones, Joseph F. Boland, Stephen J. Chanock, Sharon A. Savage. High prevalence of germline TP53 mutations in young osteosarcoma cases. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5574. doi:10.1158/1538-7445.AM2015-5574