Search

Acute Myeloid Leukemia (AML) affects 20,000 patients in the US annually with a five-year survival rate of approximately 25%. One reason for the low survival rate is the high prevalence of clonal evolution that gives rise to heterogeneous sub-populations of leukemic cells with diverse mutation spectra, which eventually leads to disease relapse. This genetic heterogeneity drives the activation of complex signaling pathways that is reflected at the protein level. This diversity makes it difficult to treat AML with targeted therapy, requiring custom patient treatment protocols tailored to each individual’s leukemia. Toward this end, the Beat AML research program prospectively collected genomic and transcriptomic data from over 1000 AML patients and carried out ex vivo drug sensitivity assays to identify genomic signatures that could predict patient-specific drug responses. However, there are inherent weaknesses in using only genetic and transcriptomic measurements as surrogates of drug response, particularly the absence of direct information about phosphorylation-mediated signal transduction. As a member of the Clinical Proteomic Tumor Analysis Consortium, we have extended the molecular characterization of this cohort by collecting proteomic and phosphoproteomic measurements from a subset of these patient samples (38 in total) to evaluate the hypothesis that proteomic signatures can improve the ability to predict response to 26 drugs in AML ex vivo samples. In this work we describe our systematic, multi-omic approach to evaluate proteomic signatures of drug response and compare protein levels to other markers of drug response such as mutational patterns. We explore the nuances of this approach using two drugs that target key pathways activated in AML: quizartinib (FLT3) and trametinib (Ras/MEK), and show how patient-derived signatures can be interpreted biologically and validated in cell lines. In conclusion, this pilot study demonstrates strong promise for proteomics-based patient stratification to assess drug sensitivity in AML.

Acute Myeloid Leukemia (AML) affects 20,000 patients in the US annually with a five-year survival rate of approximately 25%. One reason for the low survival rate is the high prevalence of clonal evolution that gives rise to heterogeneous sub-populations of leukemic cells with diverse mutation spectra, which eventually leads to disease relapse. This genetic heterogeneity drives the activation of complex signaling pathways that is reflected at the protein level. This diversity makes it difficult to treat AML with targeted therapy, requiring custom patient treatment protocols tailored to each individual’s leukemia. Toward this end, the Beat AML research program prospectively collected genomic and transcriptomic data from over 1000 AML patients and carried out ex vivo drug sensitivity assays to identify genomic signatures that could predict patient-specific drug responses. However, there are inherent weaknesses in using only genetic and transcriptomic measurements as surrogates of drug response, particularly the absence of direct information about phosphorylation-mediated signal transduction. As a member of the Clinical Proteomic Tumor Analysis Consortium, we have extended the molecular characterization of this cohort by collecting proteomic and phosphoproteomic measurements from a subset of these patient samples to evaluate the hypothesis that proteomic signatures can improve the ability to predict drug response in AML patients. In this work we describe our systematic, multi-omic approach to evaluate proteomic signatures of drug response and compare protein levels to other markers of drug response such as mutational patterns. We explore the nuances of this approach using two drugs that target key pathways activated in AML: quizartinib (FLT3) and trametinib (Ras/MEK), and show how patient-derived signatures can be interpreted biologically and validated in cell lines. In conclusion, this pilot study demonstrates strong promise for proteomics-based patient stratification to assess drug sensitivity in AML.

Introduction and study purpose: Prostate cancer (PCa) is the leading type of newly diagnosed cancer and second leading cause of cancer death in American men. Significant racial differences in PCa incidence and mortality have been long-reported in the U.S., with Black men demonstrating a disproportionately higher burden of disease compared to White men. Though some studies conducted in equal access health care settings have revealed equivalent PCa characteristics at diagnosis and comparable PCa outcomes for White and Black men. Identification of early predictors of prostate disease aggressiveness is essential to avoid over-treatment of clinically favorably disease that is unlikely to progress. This study examines the independent and joint roles of biological and social determinants of health (SDOHs) in predicting prostate disease aggressiveness in a racially diverse cohort of men undergoing biopsy for suspicion of PCa. Methodology: A retrospective cohort study was conducted at the University Hospitals Seidman Cancer Center in Cleveland, Ohio for the period January 1, 2005-May 2022. A retrospective chart review was conducted to identify eligible men, including White and Black men undergoing transrectal ultrasound-guided biopsy whose biopsy results included the following 3 groups: (i) negative biopsy with a history of 1+ negative biopsy (-ies) and no history of prostate cancer; (ii) biopsy-detected PCa of Gleason sum 6; and (iii) biopsy-detected PCa with metastases at initial cancer detection. Detailed clinical and biological specimen data were abstracted during retrospective chart review, and quantitative proteomics analysis of biopsy tissues is ongoing, using mass spectrometry, at the Pacific Northwest National Laboratory (PNNL). Multi-level modeling will be conducted to predict prostate disease aggressiveness as a function of protein biomarker candidates in combination with census block group-level SDOHs. Results: This study has achieved its initial goal of assembling a study cohort of 300 eligible men with available biopsy tissue, necessary clinical information, and full street address for assigning census regions. To date, proteomics data have been generated on the first 60 men (“Discovery cohort”) and SDOHs metrics have been linked to the first 180 men. Proteomic profiles and SDOH data will be examined jointly with SDOHs to predict patient clinical phenotype (i.e., metastatic, biopsy Gleason 6, biopsy negative). Conclusions: Identifying new markers of cancer aggressiveness within and across race has the potential to tailor disease management, and identify patients who are likely to benefit from earlier, timely treatment interventions, which will ultimately improve prostate cancer outcomes and quality of life. Predictive modeling that jointly considers biological determinants of cancer disparities in conjunction with neighborhood-level social factors have the potential to elucidate who can be safely spared repeat biopsies, as well as how to optimize cancer care. Citation Format: Jennifer Cullen, Tao Liu, Karin Rodland, Ranjini Ghosh, Anood Alfamy, Julia Payne, Kirsten Eom, Sweta Balaji, Nicole Diaz, Holly Hartman, Annie Liang Zhang, Daniel Spratt, Jonathan Shoag, Erika Trapl, Lee Ponsky, Gregory MacLennan. Biological and neighborhood-level social determinants of prostate disease aggressiveness among Black and White men undergoing prostate biopsy [abstract]. In: Proceedings of the 15th AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2022 Sep 16-19; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2022;31(1 Suppl):Abstract nr C088.

High-grade serous cancer (HGSC), the most prevalent histotype of ovarian cancer, has the lowest survival rates and is the leading cause of gynecological cancer-related deaths in the developed world. HGSC is characterized by the presence of nearly universal TP53 mutations, diverse and widespread chromosomal instability, and a general shortage of targetable driving mutations. The major clinical needs in HGSC include the identification of alternative therapeutic targets and an improved understanding of the mechanisms driving chromosomal heterogeneity. We prospectively collected 83 new ovarian HGSC samples and 20 normal precursor tissue samples (Fallopian tube, FT) for deep characterization at the genomic, transcriptomic, proteomic, and phosphoproteomic levels. We were also able to obtain cytobrush samples of Fallopian tube epithelium and provide the first extensive comparison of HGSC and FT proteomes and phosphoproteomes that used surgical specimens rather than cell lines. The ability to compare tumor and normal precursor tissues provided new biological insights regarding the role of increased proliferation in promoting replication stress in a DNA repair-deficient background, potentially explaining the high degree of chromosomal instability associated with HGSC. Our collection protocols were specifically designed to minimize the effect of ischemia on protein phosphorylation, which has been identified as a significant confounding variable. Thus, we were able to identify novel signaling interactions that would have been lost in the background of ischemic stress, most notably the activation of CDK4 and CDK7. Widespread increases in protein phosphorylation and pathway activation in tumors compared to normal tissues, particularly in the proliferation-associated CDK/RB and AURKA pathways that are targetable by FDA-approved inhibitors, provide a rational basis for the use of these therapeutics in ovarian cancer. Phosphosite-specific analysis of CDKs and FANCD was consistent with an elevated response to replication stress, potentially induced by proliferative signals. Additionally, pathway-level observations from the retrospective TCGA cohort were conserved in the prospective cohort, including histone acetylation marks associated with homologous repair deficiency. In addition to confirming key features of HGSC from previous proteogenomic study, analysis of prospectively collected HGSC samples provided significant new insights stemming from direct comparisons to normal precursor tissues; deep phosphoproteomics revealed new insights regarding the potential role of proliferation-induced replication stress in promoting the characteristic chromosomal instability of HGSC and suggested novel therapeutic targets for use in precision medicine trials. Citation Format: Jason McDermott, Osama Arshad, Vladislav Petyuk, Yi Fu, Marina Gritsenko, Chia-Feng Tsai, Samuel Payne, Mathangi Thiagarajan, Christopher Kinsinger, Ana Robles, Emily Boja, Henry Rodriguez, Daniel Chan, Bing Zhang, Zhen Zhang, Li Ding, Richard Smith, Tao Liu, Karin Rodland. Proteogenomic characterization reveals mitotic kinase and replication stress implicated in ovarian high-grade serous cancer [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 5123.

Mass spectrometry (MS)-based proteomics has great potential for overcoming the limitations of antibody-based immunoassays for antibody-independent, comprehensive, and quantitative proteomic analysis of single cells. Indeed, recent advances in nanoscale sample preparation have enabled effective processing of single cells. In particular, the concept of using boosting/carrier channels in isobaric labeling to increase the sensitivity in MS detection (e.g., our recent boosting to amplify signal with isobaric labeling (BASIL) approach) has also been increasingly used for quantitative proteomic analysis of small-sized samples including single cells. However, the full potential of such boosting/carrier approaches has not been significantly explored, nor has the resulting quantitation quality been carefully evaluated. Herein, we have systematically evaluated and optimized the BASIL approach, originally developed for quantifying phosphorylation in small number of cells, for highly effective analysis of single cells. This improved, intelligent BASIL (iBASIL) approach enables comprehensive and quantitative single-cell proteomics analysis by carefully controlling the boosting-to-sample ratio and optimizing MS automatic gain control and ion injection time settings. Using standard LC-MS platforms, iBASIL enabled precise quantification of 800-1,200 proteins from single FACS-isolated cells (depending on the cell type). Moreover, coupling iBASIL to nanoscale fractionation enabled identification of >3,000 proteins and reliable quantification of >2,000 proteins while readily separating single cells from 3 different acute myeloid leukemia cell lines. We believe iBASIL has broad utility for comprehensive and precise quantitative single-cell proteomics for systems biology and biomedical research, as well as for comprehensive proteomic analysis of size-limited clinical specimens that cannot be readily accessed by regular proteomics platforms. It also has the potential to be adapted for phosphoproteome analysis of as few as 100 cells. Citation Format: Chia-Feng Tsai, Rui Zhao, Ronald Moore, Sarah Williams, Kendall Schultz, Paul Piehowski, Ljiljana Pasa-Tolic, Karin Rodland, Richard Smith, Tujin Shi, Ying Zhu, Ying Zhu, Tao Liu. An intelligent boosting to amplify signal with isobaric labeling (iBASIL) strategy for precise quantitative single-cell proteomics analysis [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 2856.

The Department of Defense Serum Repository (DODSR) represents a unique resource enabling longitudinal studies of cancer risk, progression, and response to therapy. The DODSR was initiated in 1989, initially as a means for HIV surveillance, and is comprised of serum samples from active and reserve military personnel drawn at enlistment and annually or biennially throughout the service members' participation in the Military Health System, accompanied by the service member’s electronic health records. This enables the identification of serum samples for all service members with a particular diagnosis, such as Head and Neck Squamous Cell Carcinoma (HNSCC), a cancer significantly represented in the military population with substantial impact on military operations. For HNSCC, 175 cases diagnosed between 2003 and 2013 were identified, as well as 175 corresponding healthy controls matched for age at the time of diagnosis, gender, and ethnicity. Serum samples drawn at the time of diagnosis, 2 and 4 years prior to diagnosis, and 2 years after diagnosis were retrieved and analyzed by targeted mass spectrometry analysis using a panel of 148 qualified selected reaction monitoring (SRM) assays selected from the curated literature and discussions with military oncologists at the Murtha Cancer Center. Initial analyses of 212 samples (38 cases and 38 matched controls, each with 2-4 longitudinal samples) from the total 978 samples available focused on comparisons of pre-diagnosis and at diagnosis serum proteins. Preliminary results identified proteins that were differentially abundant between serum samples taken from HNSCC cases at the time of diagnosis, compared to age-matched controls. Proteins such as A2GL displayed a trend of gradually increasing abundance as a function of time to diagnosis, in HNSC cases, suggesting behavior that may be proven useful for early detection of HNSCC, prior to overt symptoms. These preliminary results suggest that the DODSR can be used to provide serum samples of sufficiently high quality for extensive proteomic analysis, and form cohorts sufficiently large to enable meaningful statistics. Access to longitudinal samples representing years prior to diagnosis enables the identification of serum factors that predict the risk of cancer and may also enable early diagnosis. Access to post-treatment serum samples accompanied with data indicating the clinical response to treatment enables the development of algorithms predicting response to treatment and may eventually contribute to the selection of appropriate therapies. Citation Format: Ju Yeon Lee, Tujin Shi, Vladislav Petyuk, Athena Schepmoes, Thomas Fillmore, Wayne Cardoni, George Coppit, Joseph Goodman, Shiv Srivastava, Craig Shriver, Tao Liu, Karin Rodland. Use of longitudinal samples from the DOD serum repository to identify candidate biomarkers for HNSCC [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 440.

Introduction: Although many (~40%) of the screen-detected prostate cancers are indolent (Gleason Score Material and Methods: To identify a panel of protein markers that could predict prostate cancer progression, we developed ultra-sensitive, high-pressure, high-resolution separations coupled with intelligent selection and multiplexing-selected reaction monitoring (PRISM-SRM) assays for 52 protein markers. Candidate protein markers were identified and selected from existing prostate cancer genomics data sets and validated lists of known prostate cancer drivers. The PRISM-SRM assays used heavy isotope-labeled synthetic peptides as internal standards for quantitative proteomics analysis. Study comprised of a prostate cancer patient cohort with organ confined primary tumors (N=338) presenting following post-surgery features: 53 (15.7%) metastatic progression, 124 (36.7%) biochemical recurrence (BCR), and 161 (47.6%) no progression after more than ten years of follow-up after radical prostatectomy. Index tumor region for each case was scraped from representative 10-m sections of formalin-fixed paraffin embedded (FFPE)-whole-mounted prostatectomy specimens and processed for PRISM-SRM analysis. Results: Overall, PRISM-SRM analysis of the FFPE tissue samples enabled the detection of 42 (80.8%) out of 52 biomarker candidates; in comparison regular LC-SRM without the front-end chromatographic enrichment could detect only 21 (40.4%) of these candidates at the protein level. Kruskal-Wallis testing was used for statistical evaluation of the PRISM-SRM results and comparison of relative protein levels between the “no progression”, BCR and “metastatic progression” groups. Several prostate differentiation/androgen receptor signaling related proteins (FOLH1, PSA and NCOA) and tumor progression-related proteins (TGFB1, CCND1 and SPRC) had significantly different expression levels between the three groups, and showed initial promise in predicting progression to invasive cancer, BCR and metastasis. Conclusion: Ultra-sensitive targeted proteomics can be used to select and verify performance of early prognostic markers based on the analysis of prostatectomy specimens. The top performing markers appeared able to predict progression from organ confined cancer to BCR and metastasis. Citation Format: Yuqian Gao, Yi-Ting Wang, Hui Wang, Denise Young, Jennifer Cullen, Yingjie Song, Yongmei Chen, Athena Schepmoes, Gyorgy Petrovics, Thomas Fillmore, Tujin Shi, Wei-Jun Qian, Richard Smith, Sudhir Srivastava, Jacob Kagan, Albert Dobi, Inger Rosner, Karin Rodland, Isabell Sesterhenn, Shiv Srivastava, Tao Liu. Identification of candidate biomarkers for aggressive prostate cancer using targeted proteomics and FFPE tissue samples with outcomes data [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 3165.

Mass spectrometry based targeted proteomics such as selected reaction monitoring (SRM) provides an effective, antibody-independent strategy for sensitive, specific and multiplexed verification of genomics biomarker candidates at the protein level. In order to identify a panel of proteins with the potential to predict prostate cancer progression, we have selected 52 protein candidates from existing prostate cancer genomics data sets and validated cancer drivers, and performed quantitative proteomics analysis in tissue samples using the highly sensitive PRISM (high-pressure, high-resolution separations coupled with intelligent selection and multiplexing)-SRM approach. PRISM-SRM assays have been developed for the 52 prostate cancer biomarker candidates including: prostate cancer relevant genes and common cancer drivers. One set of 105 formalin-fixed paraffin-embedded (FFPE) whole mount prostate tissue specimens were analyzed using PRISM-SRM with heavy isotope-labeled synthetic peptides as internal standards: 20 primary tumors from patients showing metastatic progression, 37 primary tumors from patients who showed biochemical recurrence (BCR), and 48 primary tumors from patients with no BCR or metastatic progression after more than ten years of follow-up after radical prostatectomy. Overall, PRISM-SRM analyses of the FFPE tissue samples enabled the detection of 42 out of 52 biomarker candidates; in comparison regular LC-SRM without the front-end chromatographic enrichment could detect only 21 of these candidates at the protein level. Kruskal-Wallis test of the PRISM-SRM results provided a statistical evaluation of comparison of relative protein levels among the “no progression”, BCR and “metastatic progression” groups. Prostate differentiation/AR signaling related proteins (FOLH1, PSA and NCOA) or tumor progression (TGFB1, CCND1 and SPRC) were significantly different between the three groups. These promising biomarker candidates for early detection of aggressive prostate cancer are being further evaluated, individually and in panels, in an independent cohort of 234 samples for their potential prognostic applications. In summary, PRISM-SRM provides a highly sensitive method for quantification and rapid screening of promising cancer biomarker candidates defined by multiomics platforms. Citation Format: Yuqian Gao, Hui Wang, Jennifer Cullen, Yongmei Chen, Athena Schepmoes, Gyorgy Petrovics, Thomas Fillmore, Tujin Shi, Wei-Jun Qian, Richard Smith, Sudhir Srivastava, Jacob Kagan, Albert Dobi, Karin Rodland, Shiv Srivastava, Tao Liu. Selection of candidate biomarkers for aggressive prostate cancer based on targeted proteomics [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 2573.

Summary: The National Cancer Institute (NCI) Clinical Proteomic Tumor Analysis Consortium is applying the latest generation of proteomic technologies to genomically annotated tumors from The Cancer Genome Atlas (TCGA) program, a joint initiative of the NCI and the National Human Genome Research Institute. By providing a fully integrated accounting of DNA, RNA, and protein abnormalities in individual tumors, these datasets will illuminate the complex relationship between genomic abnormalities and cancer phenotypes, thus producing biologic insights as well as a wave of novel candidate biomarkers and therapeutic targets amenable to verification using targeted mass spectrometry methods. Cancer Discov; 3(10); 1108–12. ©2013 AACR.