Estimating mono- and bi-phasic regression parameters using a mixture piecewise linear Bayesian hierarchical model

The dynamics of tumor burden, secreted proteins or other biomarkers over time, is often used to evaluate the effectiveness of therapy and to predict outcomes for patients. Many methods have been proposed to investigate longitudinal trends to better characterize patients and to understand disease progression. However, most approaches assume a homogeneous patient population and a uniform response trajectory over time and across patients. Here, we present a mixture piecewise linear Bayesian hierarchical model, which takes into account both population heterogeneity and nonlinear relationships between biomarkers and time. Simulation results show that our method was able to classify subjects according to their patterns of treatment response with greater than 80% accuracy in the three scenarios tested. We then applied our model to a large randomized controlled phase III clinical trial of multiple myeloma patients. Analysis results suggest that the longitudinal tumor burden trajectories in multiple myeloma patients are heterogeneous and nonlinear, even among patients assigned to the same treatment cohort. In addition, between cohorts, there are distinct differences in terms of the regression parameters and the distributions among categories in the mixture. Those results imply that longitudinal data from clinical trials may harbor unobserved subgroups and nonlinear relationships; accounting for both may be important for analyzing longitudinal data.

Abstract LB-280: The landscape of somatic microsatellite indels across cancer: detection and identification of driver events

Microsatellites (MSs) are tracts of variable-length repeats of short DNA motifs that are abundant in the human genome and exhibit high rates of mutations in the form of insertions or deletions of the repeated motif (MS indels). Despite their prevalence, the contribution of somatic MS indels to cancer is largely unexplored due to difficulties in detecting them and assessing their significance. Here, we present a comprehensive analysis of MS indels across 20 tumor types. We characterize the overall MS indel landscape and detect genes with candidate driver MS indel events. We present two novel tools: MSMuTect for accurate detection of somatic MS indels and MSMutSig for identifying candidate cancer genes containing events at higher frequency than expected by chance. We observe high variability of the frequency of MS indels across tumors and demonstrate that the number and pattern of MS indels can accurately distinguish microsatellite stable (MSS) tumors from tumors with microsatellite instability (MSI). Applying MSMutSig across 6,788 tumors from 20 different tumor types identified 7 genes with significant MS indel hotspots: ACVR2A, RNF43, DOCK3, MSH3, ESRP1, PRDM2 and JAK1. In the four genes that have been previously implicated in cancer (ACVR2A, RNF43, JAK1 and MSH3), we identified previously unreported MS indels events. Three of the genes with significant loci - DOCK3, PRDM2 and ESRP1- had not been previously listed as cancer genes. MS indels in DOCK3, a negative regulator of the WNT pathway, were mutually exclusive with mutations in CTNNB1. MS indels in ESRP1, an RNA processing gene, correlated with alternative splicing of FGFR2, an event associated with the epithelial-to-mesenchymal transition. Overall, our comprehensive analysis of somatic MS indels across cancer highlights their importance, particularly in

MSI tumors, significantly contributes to the ongoing global efforts to detect cancer genes, and may improve classification of patients into clinically-relevant subgroups.

Citation Format: Yosef E. Maruvka, Kent W. Mouw, Rosa Karlic, Rasanna Parasuraman, Atanas Kamburov, Paz Polak, Nicholas J. Haradhvala, Julian M. Hess, Esther Rheinbay, Yehuda Brody, Lior Z. Braunstein, Alan D’Andrea, Michael S. Lawrence, Adam Bass, Andre Bernards, Franziska Michor, Gad Getz. The landscape of somatic microsatellite indels across cancer: detection and identification of driver events [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 LB-280. doi:10.1158/1538-7445.AM2017-LB-280

Pharmacokinetics and Drug Interactions Determine Optimum Combination Strategies in Computational Models of Cancer Evolution

The identification of optimal drug administration schedules to battle the emergence of resistance is a major challenge in cancer research. The existence of a multitude of resistance mechanisms necessitates administering drugs in combination, significantly complicating the endeavor of predicting the evolutionary dynamics of cancers and optimal intervention strategies. A thorough understanding of the important determinants of cancer evolution under combination therapies is therefore crucial for correctly predicting treatment outcomes. Here we developed the first computational strategy to explore pharmacokinetic and drug interaction effects in evolutionary models of cancer progression, a crucial step towards making clinically relevant predictions. We found that incorporating these phenomena into our multiscale stochastic modeling framework significantly changes the optimum drug administration schedules identified, often predicting nonintuitive strategies for combination therapies. We applied our approach to an ongoing phase Ib clinical trial (TATTON) administering AZD9291 and selumetinib to EGFR-mutant lung cancer patients. Our results suggest that the schedules used in the three trial arms have almost identical efficacies, but slight modifications in the dosing frequencies of the two drugs can significantly increase tumor cell eradication. Interestingly, we also predict that drug concentrations lower than the MTD are as efficacious, suggesting that lowering the total amount of drug administered could lower toxicities while not compromising on the effectiveness of the drugs. Our approach highlights the fact that quantitative knowledge of pharmacokinetic, drug interaction, and evolutionary processes is essential for identifying best intervention strategies. Our method is applicable to diverse cancer and treatment types and allows for a rational design of clinical trials. Cancer Res; 77(14); 3908-21. ©2017 AACR.

Mathematical Modeling Links Pregnancy-Associated Changes and Breast Cancer Risk

Recent debate has concentrated on the contribution of bad luck to cancer development. The tight correlation between the number of tissue-specific stem cell divisions and cancer risk of the same tissue suggests that bad luck has an important role to play in tumor development, but the full extent of this contribution remains an open question. Improved understanding of the interplay between extrinsic and intrinsic factors at the molecular level is one promising route to identifying the limits on extrinsic control of tumor initiation, which is highly relevant to cancer prevention. Here, we use a simple mathematical model to show that recent data on the variation in numbers of breast epithelial cells with progenitor features due to pregnancy are sufficient to explain the known protective effect of full-term pregnancy in early adulthood for estrogen receptor-positive (ER⁺) breast cancer later in life. Our work provides a mechanism for this previously ill-understood effect and illuminates the complex influence of extrinsic factors at the molecular level in breast cancer. These findings represent an important contribution to the ongoing research into the role of bad luck in human tumorigenesis. Cancer Res; 77(11); 2800-9. ©2017 AACR.

Distance in cancer gene expression from stem cells predicts patient survival

The degree of histologic cellular differentiation of a cancer has been associated with prognosis but is subjectively assessed. We hypothesized that information about tumor differentiation of individual cancers could be derived objectively from cancer gene expression data, and would allow creation of a cancer phylogenetic framework that would correlate with clinical, histologic and molecular characteristics of the cancers, as well as predict prognosis. Here we utilized mRNA expression data from 4,413 patient samples with 7 diverse cancer histologies to explore the utility of ordering samples by their distance in gene expression from that of stem cells. A differentiation baseline was obtained by including expression data of human embryonic stem cells (hESC) and human mesenchymal stem cells (hMSC) for solid tumors, and of hESC and CD34+ cells for liquid tumors. We found that the correlation distance (the degree of similarity) between the gene expression profile of a tumor sample and that of stem cells orients cancers in a clinically coherent fashion. For all histologies analyzed (including carcinomas, sarcomas, and hematologic malignancies), patients with cancers with gene expression patterns most similar to that of stem cells had poorer overall survival. We also found that the genes in all undifferentiated cancers of diverse histologies that were most differentially expressed were associated with up-regulation of specific oncogenes and down-regulation of specific tumor suppressor genes. Thus, a stem cell-oriented phylogeny of cancers allows for the derivation of a novel cancer gene expression signature found in all undifferentiated forms of diverse cancer histologies, that is competitive in predicting overall survival in cancer patients compared to previously published prediction models, and is coherent in that gene expression was associated with up-regulation of specific oncogenes and down-regulation of specific tumor suppressor genes associated with regulation of the multicellular state.

Phase 1 study of twice weekly pulse dose and daily low-dose erlotinib as initial treatment for patients with EGFR-mutant lung cancers

Background

Patients with EGFR-mutant lung cancers treated with EGFR tyrosine kinase inhibitors (TKIs) develop clinical resistance, most commonly with acquisition of EGFR T790M. Evolutionary modeling suggests that a schedule of twice weekly pulse and daily low-dose erlotinib may delay emergence of EGFR T790M. Pulse dose erlotinib has superior central nervous system (CNS) penetration and may result in superior CNS disease control.

Methods

We evaluated toxicity, pharmacokinetics, and efficacy of twice weekly pulse and daily low-dose erlotinib. We assessed six escalating pulse doses of erlotinib.

Results

We enrolled 34 patients; 11 patients (32%) had brain metastases at study entry. We observed 3 dose-limiting toxicities in dose escalation: transaminitis, mucositis, and rash. The MTD was erlotinib 1200 mg days 1-2 and 50 mg days 3-7 weekly. The most frequent toxicities (any grade) were rash, diarrhea, nausea, fatigue, and mucositis. 1 complete and 24 partial responses were observed (74%, 95% CI 60-84%). Median progression-free survival was 9.9 months (95% CI 5.8-15.4 months). No patient had progression of an untreated CNS metastasis or developed a new CNS lesion while on study (0%, 95% CI 0-13%). Of the 18 patients with biopsies at progression, EGFR T790M was identified in 78% (95% CI 54-91%).

Conclusion

This is the first clinical implementation of an anti-cancer TKI regimen combining pulse and daily low-dose administration. This evolutionary modeling-based dosing schedule was well-tolerated but did not improve progression-free survival or prevent emergence of EGFR T790M, likely due to insufficient peak serum concentrations of erlotinib. This dosing schedule prevented progression of untreated or any new central nervous system metastases in all patients.

Proceedings of the third international molecular pathological epidemiology (MPE) meeting

Molecular pathological epidemiology (MPE) is a transdisciplinary and relatively new scientific discipline that integrates theory, methods, and resources from epidemiology, pathology, biostatistics, bioinformatics, and computational biology. The underlying objective of MPE research is to better understand the etiology and progression of complex and heterogeneous human diseases with the goal of informing prevention and treatment efforts in population health and clinical medicine. Although MPE research has been commonly applied to investigating breast, lung, and colorectal cancers, its methodology can be used to study most diseases. Recent successes in MPE studies include: (1) the development of new statistical methods to address etiologic heterogeneity; (2) the enhancement of causal inference; (3) the identification of previously unknown exposure-subtype disease associations; and (4) better understanding of the role of lifestyle/behavioral factors on modifying prognosis according to disease subtype. Central challenges to MPE include the relative lack of transdisciplinary experts, educational programs, and forums to discuss issues related to the advancement of the field. To address these challenges, highlight recent successes in the field, and identify new opportunities, a series of MPE meetings have been held at the Dana-Farber Cancer Institute in Boston, MA. Herein, we share the proceedings of the Third International MPE Meeting, held in May 2016 and attended by 150 scientists from 17 countries. Special topics included integration of MPE with immunology and health disparity research. This meeting series will continue to provide an impetus to foster further transdisciplinary integration of divergent scientific fields.

Clonal Hematopoiesis Associated With Adverse Outcomes After Autologous Stem-Cell Transplantation for Lymphoma

Purpose Clonal hematopoiesis of indeterminate potential (CHIP) is an age-related condition characterized by somatic mutations in the blood of otherwise healthy adults. We hypothesized that in patients undergoing autologous stem-cell transplantation (ASCT) for lymphoma, CHIP at the time of ASCT would be associated with an increased risk of myelodysplastic syndrome and acute myeloid leukemia, collectively termed therapy-related myeloid neoplasm (TMN), and other adverse outcomes. Methods We performed whole-exome sequencing on pre- and post-ASCT samples from 12 patients who developed TMN after autologous transplantation for Hodgkin lymphoma or non-Hodgkin lymphoma and targeted sequencing on cryopreserved aliquots of autologous stem-cell products from 401 patients who underwent ASCT for non-Hodgkin lymphoma between 2003 and 2010. We assessed the effect of CHIP at the time of ASCT on subsequent outcomes, including TMN, cause-specific mortality, and overall survival. Results For six of 12 patients in the exome sequencing cohort, mutations found in the TMN specimen were also detectable in the pre-ASCT specimen. In the targeted sequencing cohort, 120 patients (29.9%) had CHIP at the time of ASCT, which was associated with an increased rate of TMN (10-year cumulative incidence, 14.1% v 4.3% for those with and without CHIP, respectively; P = .002). Patients with CHIP had significantly inferior overall survival compared with those without CHIP (10-year overall survival, 30.4% v 60.9%, respectively; P < .001), including increased risk of death from TMN and cardiovascular disease. Conclusion In patients undergoing ASCT for lymphoma, CHIP at the time of transplantation is associated with inferior survival and increased risk of TMN.

Probabilistic Modeling of Reprogramming to Induced Pluripotent Stem Cells

Reprogramming of somatic cells to induced pluripotent stem cells (iPSCs) is typically an inefficient and asynchronous process. A variety of technological efforts have been made to accelerate and/or synchronize this process. To define a unified framework to study and compare the dynamics of reprogramming under different conditions, we developed an in silico analysis platform based on mathematical modeling. Our approach takes into account the variability in experimental results stemming from probabilistic growth and death of cells and potentially heterogeneous reprogramming rates. We suggest that reprogramming driven by the Yamanaka factors alone is a more heterogeneous process, possibly due to cell-specific reprogramming rates, which could be homogenized by the addition of additional factors. We validated our approach using publicly available reprogramming datasets, including data on early reprogramming dynamics as well as cell count data, and thus we demonstrated the general utility and predictive power of our methodology for investigating reprogramming and other cell fate change systems.

A computational strategy to adjust for copy number in tumor Hi-C data

MOTIVATION

The Hi-C technology was designed to decode the three-dimensional conformation of the genome. Despite progress towards more and more accurate contact maps, several systematic biases have been demonstrated to affect the resulting data matrix. Here we report a new source of bias that can arise in tumor Hi-C data, which is related to the copy number of genomic DNA. To address this bias, we designed a chromosome-adjusted iterative correction method called caICB. Our caICB correction method leads to significant improvements when compared with the original iterative correction in terms of eliminating copy number bias.

AVAILABILITY AND IMPLEMENTATION

The method is available at https://bitbucket.org/mthjwu/hicapp CONTACT: michor@jimmy.harvard.eduSupplementary information: Supplementary data are available at Bioinformatics online.

Mathematical modeling of erythrocyte chimerism informs genetic intervention strategies for sickle cell disease

Recent advances in gene therapy and genome-engineering technologies offer the opportunity to correct sickle cell disease (SCD), a heritable disorder caused by a point mutation in the β-globin gene. The developmental switch from fetal γ-globin to adult β-globin is governed in part by the transcription factor (TF) BCL11A. This TF has been proposed as a therapeutic target for reactivation of γ-globin and concomitant reduction of β-sickle globin. In this and other approaches, genetic alteration of a portion of the hematopoietic stem cell (HSC) compartment leads to a mixture of sickling and corrected red blood cells (RBCs) in periphery. To reverse the sickling phenotype, a certain proportion of corrected RBCs is necessary; the degree of HSC alteration required to achieve a desired fraction of corrected RBCs remains unknown. To address this issue, we developed a mathematical model describing aging and survival of sickle-susceptible and normal RBCs; the former can have a selective survival advantage leading to their overrepresentation. We identified the level of bone marrow chimerism required for successful stem cell-based gene therapies in SCD. Our findings were further informed using an experimental mouse model, where we transplanted mixtures of Berkeley SCD and normal murine bone marrow cells to establish chimeric grafts in murine hosts. Our integrative theoretical and experimental approach identifies the target frequency of HSC alterations required for effective treatment of sickling syndromes in humans. Our work replaces episodic observations of such target frequencies with a mathematical modeling framework that covers a large and continuous spectrum of chimerism conditions. Am. J. Hematol. 91:931-937, 2016. © 2016 Wiley Periodicals, Inc.

Punctuated copy number evolution and clonal stasis in triple-negative breast cancer

Aneuploidy is a hallmark of breast cancer; however, our knowledge of how these complex genomic rearrangements evolve during tumorigenesis is limited. In this study we developed a highly multiplexed single-nucleus-sequencing method to investigate copy number evolution in triple-negative breast cancer patients. We sequenced 1000 single cells from 12 patients and identified 1–3 major clonal subpopulations in each tumor that shared a common evolutionary lineage. We also identified a minor subpopulation of non-clonal cells that were classified as: 1) metastable, 2) pseudo-diploid, or 3) chromazemic. Phylogenetic analysis and mathematical modeling suggest that these data are unlikely to be explained by the gradual accumulation of copy number events over time. In contrast, our data challenge the paradigm of gradual evolution, showing that the majority of copy number aberrations are acquired at the earliest stages of tumor evolution, in short punctuated bursts, followed by stable clonal expansions that form the tumor mass.

ABCG2 regulates self-renewal and stem cell marker expression but not tumorigenicity or radiation resistance of glioma cells

Glioma cells with stem cell traits are thought to be responsible for tumor maintenance and therapeutic failure. Such cells can be enriched based on their inherent drug efflux capability mediated by the ABC transporter ABCG2 using the side population assay, and their characteristics include increased self-renewal, high stem cell marker expression and high tumorigenic capacity in vivo. Here, we show that ABCG2 can actively drive expression of stem cell markers and self-renewal in glioma cells. Stem cell markers and self-renewal was enriched in cells with high ABCG2 activity, and could be specifically inhibited by pharmacological and genetic ABCG2 inhibition. Importantly, despite regulating these key characteristics of stem-like tumor cells, ABCG2 activity did not affect radiation resistance or tumorigenicity in vivo. ABCG2 effects were Notch-independent and mediated by diverse mechanisms including the transcription factor Mef. Our data demonstrate that characteristics of tumor stem cells are separable, and highlight ABCG2 as a potential driver of glioma stemness.

Improving Cancer Treatment via Mathematical Modeling: Surmounting the Challenges Is Worth the Effort

Drug delivery schedules are key factors in the efficacy of cancer therapies, and mathematical modeling of population dynamics and treatment responses can be applied to identify better drug administration regimes as well as provide mechanistic insights. To capitalize on the promise of this approach, the cancer field must meet the challenges of moving this type of work into clinics.

Myeloma Cell Dynamics in Response to Treatment Supports a Model of Hierarchical Differentiation and Clonal Evolution

PURPOSE

Since the pioneering work of Salmon and Durie, quantitative measures of tumor burden in multiple myeloma have been used to make clinical predictions and model tumor growth. However, such quantitative analyses have not yet been performed on large datasets from trials using modern chemotherapy regimens.

EXPERIMENTAL DESIGN

We analyzed a large set of tumor response data from three randomized controlled trials of bortezomib-based chemotherapy regimens (total sample size n = 1,469 patients) to establish and validate a novel mathematical model of multiple myeloma cell dynamics.

RESULTS

Treatment dynamics in newly diagnosed patients were most consistent with a model postulating two tumor cell subpopulations, "progenitor cells" and "differentiated cells." Differential treatment responses were observed with significant tumoricidal effects on differentiated cells and less clear effects on progenitor cells. We validated this model using a second trial of newly diagnosed patients and a third trial of refractory patients. When applying our model to data of relapsed patients, we found that a hybrid model incorporating both a differentiation hierarchy and clonal evolution best explains the response patterns.

CONCLUSIONS

The clinical data, together with mathematical modeling, suggest that bortezomib-based therapy exerts a selection pressure on myeloma cells that can shape the disease phenotype, thereby generating further inter-patient variability. This model may be a useful tool for improving our understanding of disease biology and the response to chemotherapy regimens. Clin Cancer Res; 22(16); 4206-14. ©2016 AACR.

Abstract PR05: The effect of chemotherapy on HER2+ breast cancer heterogeneity measured by STAR-FISH: Detection of PIK3CA mutation and HER2 amplification at single-cell level in situ

Current therapies in HER2-positive breast cancer are effective in only a subset of cases and part of the resistance is attributed to single nucleotide mutation H1047R in PIK3CA. Conventional PIK3CA mutation detection methods require isolation of DNA from the tumor bulk, which requires relatively large amount of tissue and may not detect mutations in rare cancer cells.

We developed a novel method, Specific-To-Allele PCR-FISH (STAR-FISH), which allows for in situ detection of point mutation and gene amplification at single cell level. The assay consists of in situ PCR steps with mutation specific primers, followed by hybridization of a fluorescently labeled DNA probe homologous to PCR primer overhangs and probes for genomic regions of interest. The STAR-FISH signals present in intact formalin-fixed paraffin embedded (FFPE) samples are imaged and quantified in each individual nucleus within a tissue, with false discovery rate at 0.001, which facilitates identification of sub-populations of cells with different genetic makeup. The method was validated against FACS, immunofluorescence, droplet digital PCR, and MassArray; high correlation of the results was observed (R2=0.901 -0.9037, p&lt;0.001).

We applied STAR-FISH for PIK3CA hot-spot mutation and HER2 amplification to FFPE samples of HER2 positive breast tumors from 22 patients. For each case a chemotherapy naïve core needle biopsy and a post-neoadjuvant chemotherapy sample upon tumor resection were collected. STAR-FISH analysis was performed on 3-5 regions of each sample, to account for intratumor heterogeneity. Long-term patient survival data after adjuvant treatment, mostly with trastuzumab, were available for all the patients.

High-sensitivity of STAR-FISH allowed us to detect rare single cells carrying PIK3CA mutation in most of the pre-treatment samples. After adjuvant chemotherapy the frequency of these cells was significantly increased. Since the STAR-FISH signals are quantified in each individual nucleus, subpopulations of cells with PIK3CA mutation or HER2 amplification or both features can be distinguished. Based on frequencies of cells within each of these subpopulations we calculated Shannon diversity index for each pre- and post-chemotherapy sample. The index was significantly increased after treatment. However, only topologic and not overall changes in diversity predicted poor long-term survival of the patients.

In addition to analyzing the frequency of cells with PIK3CA mutation, HER2 amplification or both changes, STAR-FISH also assesses the spatial distribution of genetically distinct subtypes. We have found that cells with PIK3CA mutation, irrespective of their HER2 status, are much more dispersed within tumors after neaodjuvant chemotherapy, whereas cells with HER2 amplification and wild-type PIK3CA cluster together. These results suggest that PIK3CA mutant cells are more migratory and invasive, in agreement with prior studies of cell lines and animal models.

STAR-FISH provides a unique view into genetic intratumor heterogeneity since thousands of cells within different regions of a single tumor biopsy can be analyzed within their tissue environment. Application of this novel in situ method allowed us to detect rare cells with PIK3CA mutation, pre-existing in the majority of treatment-naïve tumors and increasing in frequency after neoadjuvant chemotherapy. Moreover, STAR-FISH data revealed the correlation of chemotherapy-induced changes in intratumor heterogeneity with long-term survival of HER2+ breast cancer patients and support the significance of tumor diversity in situ analyses.

Citation Format: Michalina Janiszewska, Lin Liu, Vanessa Almendro, Yanan Kuang, Cloud Paweletz, Britta Weigelt, Rita A. Sakr, Tari A. King, Sarat Chandarlapaty, Jorge S. Reis-Filho, Ariella B. Hanker, Carlos L. Arteaga, Park So Yeon, Franziska Michor, Kornelia Polyak. The effect of chemotherapy on HER2+ breast cancer heterogeneity measured by STAR-FISH: Detection of PIK3CA mutation and HER2 amplification at single-cell level in situ. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Breast Cancer Research; Oct 17-20, 2015; Bellevue, WA. Philadelphia (PA): AACR; Mol Cancer Res 2016;14(2_Suppl):Abstract nr PR05.

The mathematics of cancer: integrating quantitative models

Mathematical modelling approaches have become increasingly abundant in cancer research. The complexity of cancer is well suited to quantitative approaches as it provides challenges and opportunities for new developments. In turn, mathematical modelling contributes to cancer research by helping to elucidate mechanisms and by providing quantitative predictions that can be validated. The recent expansion of quantitative models addresses many questions regarding tumour initiation, progression and metastases as well as intra-tumour heterogeneity, treatment responses and resistance. Mathematical models can complement experimental and clinical studies, but also challenge current paradigms, redefine our understanding of mechanisms driving tumorigenesis and shape future research in cancer biology.

Abstract A21: Proteomic profiling to elucidate intratumoral heterogeneity and cancer evolution in lung cancer

Tumors often display a high degree of intratumoral heterogeneity as manifested by dynamic changes in gene expression, protein expression, and on gross examination of histology, among many other features. Clinically, this underlying heterogeneity can drive tumor evolution and progression towards a more aggressive neoplastic state and a worse prognosis for patients; therefore, identifying the diverse composition of a tumor for early risk stratification is of critical importance. To elucidate intratumoral heterogeneity and intracellular hierarchy in a novel manner, we first conducted a low-cost quantitative proteomics analysis using MALDI-TOF mass spectrometry on over 1900 samples from different histological regions of individual tumors from 35 lung cancer patients, as well as from 3 mesenchymal stem cell samples. The histologies identified were acinar, basal cells, bronchial epithelium, lepidic, complex gland, micropapillary, near tumor normal, normal alveolar, papillary, papillary lepidic, papillary mucinous, and solid. Patient-specific information including survival status, sex, age, smoking status, SUV by FDG-PET scan, tumor size, EGFR, KRAS, and ERCC1 mutation status, among other variables was obtained. We then compared the proteomes derived from each tumor to the stem cell proteomes, and using computational strategies, mapped the distance of each histological sample from the mesenchymal stem cell state; using clustering techniques, we organized the major histological subtypes into a phylogenetic tree from stem cells to normal lung. We hypothesized that by applying and improving upon map of tumor evolution based on the distance of each individual histological sample from a stem cell state. Apart from liquid tumors, there have thus far been limited studies on the prognostic significance of different subclones in solid tumors, and therefore we treated each histological sample as a subclone within each patient. We also aimed to identify survival-associated subclones and prognostic molecular signatures across combinations of subclones. Identifying these subclones may provide insight into malignant micrometastases to other organs. Using co-expression network analysis, we further pinpointed distinctive significantly dysregulated co-regulatory protein networks within each histological subtype. Based on these networks, we sought to identify important hub proteins within each histology. Ultimately, using proteomic profiling in solid tumors can be a novel approach in functionally characterizing intratumoral heterogeneity, and may allow for a more robust analysis of the diverse molecular expression of single tumor samples. Our results may help inform the field of targeted broad-scale proteomics profiling for therapeutic use.

Citation Format: Charlotte Lee, Hua-Jun Wu, Andre L. Moreira, Erin H. Seeley, Callee Walsh, Robert J. Downey, Franziska Michor. Proteomic profiling to elucidate intratumoral heterogeneity and cancer evolution in lung cancer. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr A21.

Abstract A82: Prognostic relevance and genomic profile of circulating tumor cells in multiple myeloma

Introduction: Genomic sequencing of tumor cells obtained from the bone marrow (BM) of patients with multiple myeloma (MM) has demonstrated significant clonal heterogeneity. However, it could be envisioned that such clonal diversity may be even higher since the pattern of BM infiltration in MM is typically patchy. In addition, BM biopsies cannot be repeated multiple times during the course of therapy, indicating a need for less invasive methods to genomically characterize MM patients. In this study, we aimed to determine the overall applicability of performing genomic characterization of MM patients non-invasively using circulating tumor cells (CTC).

Methods: We performed CTC enumeration using multi-parameter flow cytometry (MFC) in 50 newly-diagnosed patients with symptomatic MM who were prospectively enrolled on the Spanish clinical trial PETHEMA/GEM2010MAS65 as well as 64 patients with MM with relapsed disease or in remission/on maintenance therapy seen at the Dana-Farber Cancer Institute. For sequencing studies, we obtained 8 samples of newly-diagnosed untreated patients. We sequenced BM clonal PCs and CTCs up to 200x, and germline cells up to 50x. Whole genome amplification (WGA) was performed for CTCs, and two independent libraries were sequenced up to 100x for each duplicate. Only single nucleotide variants (SNVs) shared in both parallel WGA libraries were used.

Results: Using sensitive MFC, we showed that CTCs were detectable in 40/50 (80%) newly-diagnosed MM patients, and in 71/130 (55%) of multiple sequential samples from patients with relapsed disease or in remission/on maintenance. Nineteen of the 40 newly-diagnosed cases displaying PB CTCs had relapsed (median TTP of 31 months); by contrast, only 1 of the 10 patients with undetectable CTCs has relapsed (median TTP not reached; P = .08). Afterward, increasing CTC counts in sequential PB samples from patients with relapsed disease or in remission/on maintenance therapy were associated with poor overall survival (P = .01), indicating that both the absolute numbers of CTCs and trend of CTC are predictive of outcome in MM.

After demonstrating that CTCs can be readily detected in the majority of MM patients, we then determined the mutational profile of CTCs and compared it to that of patient-paired BM clonal PCs. We identified a median of 223 and 118 SNVs in BM clonal PCs and CTCs, respectively. The concordance of somatic variants found in matched BM clonal PCs and CTCs was of 79%. Noteworthy, upon investigating specific mutations implicated in MM (eg. KRAS, NRAS, BRAF) a total of 18 non-synonymous SNVs (NS-SNVs) in 13 genes were identified in our cohort, and most of these NS-SNVs were simultaneously detected in matched BM clonal PCs and CTCs. That notwithstanding, we also identified several unique mutations present in CTC or BM clonal PCs; of those, up to 39 NS-SNV were identified as CTC specific, and 6 NS-SNVs in 4 genes (CR1, DPY19L2, TMPRSS13, HBG1) were detected in multiple patient samples. A significant concordance for the pattern of copy number variations (CNVs) between matched BM and PB tumor cells was also observed.

Conclusion: This study defines a new role for CTCs in the prognostic and molecular profiling of MM patients, and provides the rational for an integrated flow-molecular algorithm to detect CTCs in PB and identify candidate patients for noninvasive genomic characterization to predict outcomes.

Citation Format: Yuji Mishima, Bruno Paiva, Jiantao Shi, Mira Massoud, Salomon Manier, Adriana Perilla-Glen, Yosra Aljawai, Satoshi Takagi, Daisy Huynh, Daisy Huynh, Aldo Roccaro, Antonio Sacco, Diego Alignani, Maria-Victoria Mateos, Joan Blade, Juan-Jose Lahuerta, Paul Richardson, Jacob Laubach, Robert Schlossman, Kenneth Anderson, Nikhil Munshi, Felipe Prosper, Jesus San Miguel, Franziska Michor, Irene M. Ghobrial. Prognostic relevance and genomic profile of circulating tumor cells in multiple myeloma. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2015 Nov 5-9; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2015;14(12 Suppl 2):Abstract nr A82.

Abstract IA05: Evolution of glioblastoma subtypes

To understand the relationships between the non-GCIMP glioblastoma (GBM) subgroups, we performed mathematical modeling to predict the temporal sequence of driver events during tumorigenesis. The most common order of evolutionary events is 1) chromosome (chr) 7 gain and chr10 loss, followed by 2) CDKN2A loss and/or TP53 mutation, and 3) alterations canonical for specific subtypes. We then developed a computational methodology to identify drivers of broad copy number changes, identifying PDGFA (chr7) and PTEN (chr10) as driving initial nondisjunction events. These predictions were validated using mouse modeling, showing that PDGFA is sufficient to induce proneural-like gliomas and that additional NF1 loss converts proneural to the mesenchymal subtype. We further performed integrated genomic and epigenetic profiling of patient-derived cell lines to identify further genes driving subtype transformation as well as alterations in the three-dimensional architecture of the GBM genome during subtype changes. Our computational methodologies can be applied to study the evolutionary dynamics of other cancer types.

Abstract PR09: Extracellular stiffness cues drive spatial reorganization of the genome to globally constrain RNA abundance

The stiffness of the extracellular matrix (ECM) drives mechanosignaling that regulates tissue development and malignancy. We previously showed that a stiff ECM disrupts tissue organization and enhances malignant progression by inducing cell invasion and migration. However, the specific transcriptional and molecular events in which mechanotransduction directs these phenotypes are not well understood. To clarify this process, we used a combination of genome-scale approaches to monitor changes in gene expression and protein abundance as a function of acinar morphogenesis and tissue homeostasis in three dimensional extracellular matrix hydrogels with tunable stiffness. Elevated ECM stiffness perturbed tissue homeostasis and reverted the transcriptional phenotype of differentiated mammary acini to resemble that observed in rapidly proliferating nonpolarized mammary cell aggregates. These findings suggest that tissue tension induces cellular changes that directly reflect higher-order tissue organizational states. We found that these changes involve the spatial rearrangement of peripheral chromatin, and that the expression levels of multiple histone deacetylases increase in organized tissues concurrently with elevated nuclear heterochromatin content, an effect that is abrogated in rigid ECM conditions. We support these observations by mapping mechanoresponsive peripheral heterochromatin elements via ChIPseq, enabling us to directly identify dynamic regions containing genes whose transcriptional activity is responsive to mechanical cues. Finally, using a combination of genomic, imaging, and molecular biology techniques we demonstrated that ECM compliance and tissue organization significantly influences global RNA abundance. Notably, this model presents formidable conceptual and practical challenges for the interpretation of genomic data. Collectively, this work indicates that tissue organization is critically dependent on the cellular mechanical environment, which qualitatively and quantitatively shapes the epigenetic and transcriptional landscape by mechanisms that have not yet been elucidated.

Note: This abstract was not presented at the conference.

Citation Format: Russell Bainer, Yoshihiro Yui, Shannon Mumenthaler, Parag Mallick, Lin Liu, Hua-Jun Wu, Ondrej Podlaha, Franziska Michor, Jan Liphardt, Jonathan Licht, Valerie Weaver. Extracellular stiffness cues drive spatial reorganization of the genome to globally constrain RNA abundance. [abstract]. In: Proceedings of the AACR Special Conference on Computational and Systems Biology of Cancer; Feb 8-11 2015; San Francisco, CA. Philadelphia (PA): AACR; Cancer Res 2015;75(22 Suppl 2):Abstract nr PR09.

In situ single-cell analysis identifies heterogeneity for PIK3CA mutation and HER2 amplification in HER2-positive breast cancer

Detection of minor, genetically distinct subpopulations within tumors is a key challenge in cancer genomics. Here we report STAR-FISH (specific-to-allele PCR-FISH), a novel method for the combined detection of single-nucleotide and copy number alterations in single cells in intact archived tissues. Using this method, we assessed the clinical impact of changes in the frequency and topology of PIK3CA mutation and HER2 (ERBB2) amplification within HER2-positive breast cancer during neoadjuvant therapy. We found that these two genetic events are not always present in the same cells. Chemotherapy selects for PIK3CA-mutant cells, a minor subpopulation in nearly all treatment-naive samples, and modulates genetic diversity within tumors. Treatment-associated changes in the spatial distribution of cellular genetic diversity correlated with poor long-term outcome following adjuvant therapy with trastuzumab. Our findings support the use of in situ single cell-based methods in cancer genomics and imply that chemotherapy before HER2-targeted therapy may promote treatment resistance.