Jessie M. English

Bias-Corrected Targeted Next-Generation Sequencing for Rapid, Multiplexed Detection of Actionable Alterations in Cell-Free DNA from Advanced Lung Cancer Patients

Purpose: Tumor genotyping is a powerful tool for guiding non–small cell lung cancer (NSCLC) care; however, comprehensive tumor genotyping can be logistically cumbersome. To facilitate genotyping, we developed a next-generation sequencing (NGS) assay using a desktop sequencer to detect actionable mutations and rearrangements in cell-free plasma DNA (cfDNA). Experimental Design: An NGS panel was developed targeting 11 driver oncogenes found in NSCLC. Targeted NGS was performed using a novel methodology that maximizes on-target reads, and minimizes artifact, and was validated on DNA dilutions derived from cell lines. Plasma NGS was then blindly performed on 48 patients with advanced, progressive NSCLC and a known tumor genotype, and explored in two patients with incomplete tumor genotyping. Results: NGS could identify mutations present in DNA dilutions at ≥0.4% allelic frequency with 100% sensitivity/specificity. Plasma NGS detected a broad range of driver and resistance mutations, including ALK, ROS1, and RET rearrangements, HER2 insertions, and MET amplification, with 100% specificity. Sensitivity was 77% across 62 known driver and resistance mutations from the 48 cases; in 29 cases with common EGFR and KRAS mutations, sensitivity was similar to droplet digital PCR. In two cases with incomplete tumor genotyping, plasma NGS rapidly identified a novel EGFR exon 19 deletion and a missed case of MET amplification. Conclusions: Blinded to tumor genotype, this plasma NGS approach detected a broad range of targetable genomic alterations in NSCLC with no false positives including complex mutations like rearrangements and unexpected resistance mutations such as EGFR C797S. Through use of widely available vacutainers and a desktop sequencing platform, this assay has the potential to be implemented broadly for patient care and translational research. Clin Cancer Res; 22(4); 915–22. ©2015 AACR. See related commentary by Tsui and Berger, p. 790

Cytotoxic T Cells in PD-L1–Positive Malignant Pleural Mesotheliomas Are Counterbalanced by Distinct Immunosuppressive Factors

In malignant pleural mesothelioma, immunohistochemical expression of PD-L1 does not accurately predict whether patients respond to treatment with PD-1 pathway inhibitors. Comprehensive immunoprofiling by flow cytometry uncovered immunophenotypes that improve our understanding of response and resistance to checkpoint blockade. PD-L1 immunohistochemical staining does not always predict whether a cancer will respond to treatment with PD-1 inhibitors. We sought to characterize immune cell infiltrates and the expression of T-cell inhibitor markers in PD-L1–positive and PD-L1–negative malignant pleural mesothelioma samples. We developed a method for immune cell phenotyping using flow cytometry on solid tumors that have been dissociated into single-cell suspensions and applied this technique to analyze 43 resected malignant pleural mesothelioma specimens. Compared with PD-L1–negative tumors, PD-L1–positive tumors had significantly more infiltrating CD45+ immune cells, a significantly higher proportion of infiltrating CD3+ T cells, and a significantly higher percentage of CD3+ cells displaying the activated HLA-DR+/CD38+ phenotype. PD-L1–positive tumors also had a significantly higher proportion of proliferating CD8+ T cells, a higher fraction of FOXP3+/CD4+ Tregs, and increased expression of PD-1 and TIM-3 on CD4+ and CD8+ T cells. Double-positive PD-1+/TIM-3+ CD8+ T cells were more commonly found on PD-L1–positive tumors. Compared with epithelioid tumors, sarcomatoid and biphasic mesothelioma samples were significantly more likely to be PD-L1 positive and showed more infiltration with CD3+ T cells and PD-1+/TIM-3+ CD8+ T cells. Immunologic phenotypes in mesothelioma differ based on PD-L1 status and histologic subtype. Successful incorporation of comprehensive immune profiling by flow cytometry into prospective clinical trials could refine our ability to predict which patients will respond to specific immune checkpoint blockade strategies. Cancer Immunol Res; 4(12); 1038–48. ©2016 AACR.

Establishment of patient-derived tumor xenograft models of epithelial ovarian cancer for preclinical evaluation of novel therapeutics

Purpose: Ovarian cancer is the leading cause of death from gynecologic malignancy in the United States, with high rates of recurrence and eventual resistance to cytotoxic chemotherapy. Model systems that allow for accurate and reproducible target discovery and validation are needed to support further drug development in this disease. Experimental Design: Clinically annotated patient-derived xenograft (PDX) models were generated from tumor cells isolated from the ascites or pleural fluid of patients undergoing clinical procedures. Models were characterized by IHC and by molecular analyses. Each PDX was luciferized to allow for reproducible in vivo assessment of intraperitoneal tumor burden by bioluminescence imaging (BLI). Plasma assays for CA125 and human LINE-1 were developed as secondary tests of in vivo disease burden. Results: Fourteen clinically annotated and molecularly characterized luciferized ovarian PDX models were generated. Luciferized PDX models retain fidelity to both the nonluciferized PDX and the original patient tumor, as demonstrated by IHC, array CGH, and targeted and whole-exome sequencing analyses. Models demonstrated diversity in specific genetic alterations and activation of PI3K signaling pathway members. Response of luciferized PDX models to standard-of-care therapy could be reproducibly monitored by BLI or plasma markers. Conclusions: We describe the establishment of a collection of 14 clinically annotated and molecularly characterized luciferized ovarian PDX models in which orthotopic tumor burden in the intraperitoneal space can be followed by standard and reproducible methods. This collection is well suited as a platform for proof-of-concept efficacy and biomarker studies and for validation of novel therapeutic strategies in ovarian cancer. Clin Cancer Res; 23(5); 1263–73. ©2016 AACR.

Protein kinase inhibitors: breakthrough medicines and the next generation

In this issue of Expert Opinion on Investigational Drugs, several protein kinases families and pathways underlying cancer and other diseases are reviewed and several small molecule inhibitors that are in clinical trials are further described. Highlights of these reviews and drug evaluations are summarized in this editorial.

Fine needle aspirate flow cytometric phenotyping characterizes immunosuppressive nature of the mesothelioma microenvironment

With the emergence of checkpoint blockade and other immunotherapeutic drugs, and the growing adoption of smaller, more flexible adaptive clinical trial designs, there is an unmet need to develop diagnostics that can rapidly immunophenotype patient tumors. The ability to longitudinally profile the tumor immune infiltrate in response to immunotherapy also presents a window of opportunity to illuminate mechanisms of resistance. We have developed a fine needle aspirate biopsy (FNA) platform to perform immune profiling on thoracic malignancies. Matching peripheral blood, bulk resected tumor, and FNA were analyzed from 13 mesothelioma patients. FNA samples yielded greater numbers of viable cells when compared to core needle biopsies. Cell numbers were adequate to perform flow cytometric analyses on T cell lineage, T cell activation and inhibitory receptor expression, and myeloid immunosuppressive checkpoint markers. FNA samples were representative of the tumor as a whole as assessed by head-to-head comparison to single cell suspensions of dissociated whole tumor. Parallel analysis of matched patient blood enabled us to establish quality assurance criteria to determine the accuracy of FNA procedures to sample tumor tissue. FNA biopsies provide a diagnostic to rapidly phenotype the tumor immune microenvironment that may be of great relevance to clinical trials.

Abstract 3157: Serial droplet digital PCR (ddPCR) of plasma cell-free DNA (cfDNA) as pharmacodynamic (PD) biomarker in Phase 1 clinical trials for patients (pts) with KRAS mutant non-small cell lung cancer (NSCLC)

Introduction: Phase 1 clinical trials of novel therapeutics have historically focused on toxicity, but increasingly are doubling as efficacy studies in biomarker-enriched populations. Given the small sample sizes (∼3-6 patients per dose), response on imaging may be a coarse marker of therapeutic effect. Here we piloted serial ddPCR of plasma cfDNA as a PD marker in a phase I combination study of a MEK inhibitor and a CDK 4/6 inhibitor in patients with RAS mutated cancers. Methods / Results: Twenty-five pts with RAS-mutated cancer (incl. 17 patients with KRAS-mutant NSCLC) have been enrolled to date in a phase I dose escalation trial of the MEK inhibitor PD-0325901 with the CDK4/6 inhibitor palbociclib (NCT02022982). Plasma for cfDNA genotyping was collected at baseline prior to therapy and at the beginning of cycle 2. Plasma genotyping for KRAS G12X mutations was performed using a validated and highly quantitative droplet digital PCR assay. Pts were enrolled in 5 dose level cohorts ranging from 75 mg palbociclib daily (3 weeks on, 1 week off) with 2 mg PD-0325901 BID (3 weeks on 1week off) to 125 mg palbociclib daily with 8 mg PD-0325901 BID (Table). KRAS mutations were detected in 14/24 pts at baseline (59%, median 1402 copies/mL plasma, range: 11-93000), consistent with the previously reported sensitivity of 64%. A second blood draw at cycle 2 was obtained for all 14 pts. A positive plasma response, defined as decrease of KRAS G12X mutants from first to second dose, was observed in 6 pts (range -6% - -100%) with the most plasma responders (n = 4 pts) at the maximum administered dose. At lower administered doses, there was a median increase in plasma KRAS mutant levels. Conclusions: Increasing dose levels resulted in more consistent decreases in KRAS mutation in cfDNA, consistent with a dose-dependent pharmacodynamic effect.These results highlight the potential value of serial plasma ddPCR as a PD marker in early phase clinical trials. Citation Format: Cloud P. Paweletz, Geoffrey R. Oxnard, Nora Feeney, John F. Hilton, Leena Gandhi, Khanh T. Do, Adrienne Anderson, Andrew Wolanski, Alexander Tejeda, Jessie M. English, Paul T. Kirschmeier, Pasi A. Janne, Geoffrey I. Shapiro. Serial droplet digital PCR (ddPCR) of plasma cell-free DNA (cfDNA) as pharmacodynamic (PD) biomarker in Phase 1 clinical trials for patients (pts) with KRAS mutant non-small cell lung cancer (NSCLC). [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3157.

Abstract B99: Immune profiling of NSCLC tumors and matching normal lung samples by multicolor flow cytometry

With the recent advancements in cancer immunotherapy, understanding of tumor microenvironment becomes crucial to progressing oncology research. Standard method for tumor immune characterization today is traditional IHC technique with 1-2 color staining which presents certain limitations. In order to simultaneously identify immune infiltrate cellular components and parallel it with the analysis of expression of activation/suppression markers in various subpopulations, we developed a multicolor flow cytometry pipeline for freshly resected tumors. Thirty one individual biomarkers were used in several antibody staining panels with suffficientdesigned redundancy to allow crossover and address intra-experiment variability. Aliquots of the samples were frozen and stained on separate days to test for inter-assay variability. We performed flow cytometry analysis of 23 NSCLC tumors (17 adenocarcinoma, 1 adenosquamous and 5 squamous carcinoma subtypes) and matcheding normal lung tissue. Unsupervised clustering ofby 70 parameters, including percentages of cellular subtypes as well as modulator marker expression, showed distinct segregation of normal and tumor tissue samples. The immune infiltrate in resected tumors exhibited significantly decreased CD66b+ granulocytes and CD56+ NK cells (especially cytotoxic CD56+CD16+ NK cells) and highly increased CD3+ T and CD19+ B cells when compared to flow analysis of normal lung tissue. Adeno and squamous carcinomas were not segregated into separate groups by clustering analysis but rather joined together to form 3 subgroups defined by exhaustion/activation marker expression. Grouping the tumor samples by the expression of the clinically relevant PD-L1 marker in immune cells indicated that “high PD-L1” tumors tend to have highly activated T cells populations. This finding correlates with the latest views on high PD-1/PD-L1 expressing immune cells being “negatively regulated” rather than “exhausted” as the result of their extensive interaction with cancer cells. As we continue to characterize the immune profile of NSCLC tumors and paired normal lung, in conjunction with genetic and clinical information, we aim to further understanding molecular and clinical correlates that influence the tumor microenvironment and to ultimately of the mechanisms of immune response to overcome both innate and acquired resistance to immune therapy agents. Citation Format: Elena Ivanova, Robert E. Jones, Mark M. Awad, Mark A. Bittinger, Meghana M. Kulkarni, Shohei Koyama, Christina G. Almonte, Abigail A. Santos, Jessie M. English, Julianne Barlow, William G. Richards, Peter S. Hammerman, Scott J. Rodig, Raphael Bueno, Kwok-Kin Wong. Immune profiling of NSCLC tumors and matching normal lung samples by multicolor flow cytometry. [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 B99.

Abstract A69: An epigenetic-focused CRISPR/Cas9 screen to identify regulators of IFNγ-induced PD-L1 expression

PD-L1 (B7-H1, CD274) is a clinically validated immuno-oncology target, which is often over-expressed on the surface of tumor cells. PD-L1 binds to PD-1 expressed on T cells generating an immunosuppressive signaling response that limits T cell activation and facilitates immune evasion. The tumor microenvironment often recruits immune cells that produce a number of secreted factors, including IFNγ, a potent inducer of PD-L1 expression on tumor cells. Blocking IFNγ-induced PD-L1 expression with small molecules could be a potential alternative to antibody-based PD-L1/PD-1 blockade. Recent studies on human patient samples indicate that the level of PD-L1 expression on tumor cells is inversely related to the level of DNA methylation at the PD-L1 promoter (Gettinger et al, 2015), suggesting that PD-L1 expression is epigenetically silenced in tumor cells with low PD-L1 expression. Therefore, cytokines that induce PD-L1 expression on tumor cells, such as IFNγ, may regulate the activity of epigenetic silencing factors at the PD-L1 promoter, and identification of these epigenetic factors could provide novel therapeutic targets to block PD-L1 expression on tumor cells. CRISPR/Cas9 gene editing has recently emerged as a powerful technology for phenotypic screening. To identify potential epigenetic regulators of IFNγ-induced PD-L1 on tumor cells, several murine tumor cell lines were treated with IFNγ and PD-L1 expression was monitored by flow cytometry. The ovarian cancer cell line ID8 demonstrated high PD-L1 expression following IFNγ stimulation, and stable Cas9 expressing clones were generated. A high expressing Cas9 clone was selected for follow-up transduction with a sgRNA library targeting >350 known epigenetic factors, plus numerous positive and negative controls. Library transduced cells were evaluated in two separate screening streams: i) an enrichment screen to identify genes regulating IFNγ-induced PD-L1 expression, and ii) a depletion screen to identify genes essential for the growth of ID8 tumor cells. For the enrichment screen, sgRNA transduced cells were treated with IFNγ, and FACS was performed to collect cells with low PD-L1 expression i.e. cells refractory to IFNγ-induced PD-L1 expression. Genomic DNA was then isolated from the sorted cells and sgRNA sequences were quantified by next-generation sequencing (NGS). As an important validation of the FACS-based screening format, the most highly enriched sgRNAs in the low PD-L1 population were PD-L1 itself, and the canonical mediators of IFNγ signaling JAK1/2 and STAT1. For the depletion screen, sgRNA transduced cells were cultured for up to 14 days, with cell pellets collected on day 0, 3, 7 and 14 for NGS quantification of sgRNAs. All positive and negative controls scored as expected and several epigenetic factors were strongly depleted indicating an essential role for ID8 cell growth. In summary, CRISPR/Cas9 gene editing is a powerful screening technology for the identification of factors essential for cell growth, and when paired with FACS, is a useful methodology to identify factors regulating expression of immune checkpoint molecules. Gettinger et al, Journal of Clinical Oncology , 2015 ASCO Annual Meeting, Vol 33, No 15 (May 20 supplement), 2015: 3015 Citation Format: Troy A. Luster, Meghana Kulkarni, Erica Fitzpatrick, Lauren Badalucco, Jessie English, Kwok-Kin Wong, Mark Bittinger. An epigenetic-focused CRISPR/Cas9 screen to identify regulators of IFNγ-induced PD-L1 expression. [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 A69.

Abstract 928: Bias-corrected next generation sequencing of cell free DNA for detection of actionable mutations and rearrangements in advanced non-small cell lung cancer

Genotype-directed targeted therapies are transforming the care of advanced cancer. However, having adequate tumor tissue for comprehensive genotyping remains a challenge for many patients. We recently reported on a new assay for noninvasive plasma genotyping using droplet digital PCR (ddPCR) of cell free DNA (cfDNA) (Oxnard et al., CCR, 2014). While appealing in its simplicity, ddPCR assays cannot detect complex genomic alterations such as chromosomal rearrangements and cannot be massively multiplexed; a key requirement for efficiently testing for the increasing number of rare but targetable genomic alterations in non-small cell lung cancer (NSCLC). Here we report on a novel, targeted, bias-corrected, and quantitative Next Generation Sequencing (NGS) solution for analyses of cfDNA. Our plasma NGS assay is focused on targetable driver mutations and rearrangements in lung cancer targeting 23 exons, 9 introns covering ∼17kb for 10 driver genes. Rather than using traditional hybrid capture, our assay integrates efficient fragment cloning, guided target retrieval, and molecular annotation of library sequences, which reduces off-target reads and improves signal-to-noise ratio. Performance of this assay panel was optimized in two iterations of probe design using targeted sequencing results of fifteen annotated NSCLC-derived cell lines. We were able to detect hotspot base changes and indels with 100% sensitivity and specificity at mutant allele frequencies of ∼1% and with 92.3% specificity at allele frequencies less than ≤0.1%. An automated pipeline for mutation calling correctly identified rare alleles with a sensitivity and specificity for point mutations and indels of >90% at allele frequencies of ≤0.2%. We then piloted the technology using cfDNA from a patient with newly diagnosed advanced ALK-rearranged NSCLC and two NSCLC patients with micro-duplication/insertions in Her2neu. Blinded to the true tumor sequences, our NGS platform detected the correct ALK rearrangement (EML4-ALK v2) in 2 separate reads as well as the correct Her2neu insertions (2311-2322dup and 2332-2340 dup) in circulating DNA. We further show clinical translation to plasma from a cohort of 20 lung cancer patients that harbored rare driver genetic events, including ALK, ROS1 and RET rearrangements by tissue sequencing. We are the first to sequence chromosomal rearrangements in plasma without prior knowledge of tissue genotyping results. The potential for this NGS-based plasma genotyping assay to allow for the noninvasive, multiplexed detection of complex, targetable genomic alterations including rearrangements and rare mutations represents a significant improvement over existing methods. Citation Format: Cloud P. Paweletz, Chris K. Raymond, Adrian G. Sacher, Yanan Kuang, Allison O9Connell, Lee Lim, Mark Li, Chris Armour, Jessie M. English, Paul T. Kirschmeier, Pasi A. Janne, Geoffrey Oxnard. Bias-corrected next generation sequencing of cell free DNA for detection of actionable mutations and rearrangements in advanced non-small cell lung cancer. [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 928. doi:10.1158/1538-7445.AM2015-928

Abstract 1202: Patient derived xenograft model platform of high grade serous ovarian cancer supporting discovery of targeted therapies and biomarkers

Proceedings: AACR Annual Meeting 2014; April 5-9, 2014; San Diego, CA High grade serous ovarian carcinoma (HGSOC) is the most lethal gynecologic malignancy in the United States. It is the most common form of ovarian cancer but early detection and effective therapeutics remain elusive and are a major unmet medical need. The current standard of care for ovarian cancer is platinum based therapy, without consideration of histological subtypes. While many patients exhibit an initial response to this standard therapy, few achieve long term remissions or cures. A major limitation in advancing therapeutic development has been a lack of experimental animal models that accurately represent the spectrum of HGSOC, are amenable to assess efficacy of novel therapeutics, are predictive of patient responses in the clinic and can identify patient subsets that will benefit from specific targeted therapeutics. To address this gap, we generated a platform of clinically relevant, early passage, orthotopic HGSOC patient-derived xenograft (PDX) models. These models were established by implanting mice intraperitoneally with fresh human ovarian cancer cells purified from operative or paracentesis samples under an IRB-approved protocol. Implanted mice were sacrificed after showing signs of abdominal distension and ascites development. Fresh mouse ascites-derived ovarian cancer cells were luciferized ex vivo by sophisticated lentivirus based methodologies to allow for non-invasive methods of tumor burden measurement. Luciferized PDX models were further expanded, banked and utilized for drug efficacy and biomarker evaluation studies. Surrogate biomarkers such as plasma CA-125 measured by BioScale AMMP method and circulating human cfDNA by LINE-1 qPCR have been validated and qualified in these models to support secondary methods of tumor burden evaluation. A diverse library of clinically-annotated, HGSOC PDX models has been established that encompass both treatment naive and treatment refractory HGSOC. Consistent with clinical disease, these orthotopic PDX models exhibit diffusely disseminated peritoneal disease with tissue infiltration to the omentum, ovaries, pancreas and spleen; along with distended abdomens full with ascites. Immunohistochemistry for ovarian specific markers including PAX8 and CK7 were analyzed in these tissues and shown to maintain fidelity to patient material through serial passages in mice. In vivo drug sensitivity studies with platinum and platinum/taxane doublet regimens recapitulate response as expected in HGSOC and, in a model of chemotherapy-resistant disease, demonstrate expected resistance as predicted by the clinical history. A clinically relevant HGSOC PDX model platform, amenable for testing efficacy of novel therapeutics and delineating responder ID has been established. Data will be presented with ongoing additional characterization of this platform. Citation Format: Sangeetha S. Palakurthi, Joyce F. Liu, Qing Zeng, Shan Zhou, Elena Ivanova, Cloud Paweletz, John Murgo, Justin Evangelista, Jennifer Curtis, Huiying Piao, Prafulla Gokhale, Jessie M. English, Paul Kirschmeier, Kwok-Kin Wong, Ursula A. Matulonis, Ronny Drapkin. Patient derived xenograft model platform of high grade serous ovarian cancer supporting discovery of targeted therapies and biomarkers. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 1202. doi:10.1158/1538-7445.AM2014-1202