Joint Conference Research Articles

Abstract A24: Using single-cell RNA-seq approaches to decipher heterogeneity in autochthonous mouse models of small cell lung cancer

Abstract Small cell lung cancer (SCLC) is one of the most aggressive cancer types, and patients in clinic usually (>60% cases) present with metastasis. Current therapies for SCLC have not changed from 1980, and they include a combination therapy of cisplatin and etoposide. Most patients relapse after initial response, and the 5-year survival for extensive stage SCLC is 2%. One possible explanation for the failure of conventional and targeted therapies in cancers is the cellular heterogeneity that exists within tumors. Thus, understanding phenotypic heterogeneity at the single-cell level can be leveraged to predict mechanisms of resistance, which enables the design of effective combination therapies. In this project, we used as a model system an autochthonous mouse model of human SCLC, in which we deleted p53, RB, and p130. We collected multiple primary tumors, circulating tumor cells (CTC), and lymph node (proximal site) and liver (distant site) metastases from the same mouse, and across different mice. These samples were collected at the stage that should correspond to the limited/extensive stage of human SCLC, which is most commonly seen in clinic. Then we used single-cell RNA-seq (sc-RNAseq) methods to define and understand transcriptional heterogeneity in these samples. Furthermore, we aimed to understand the evolutionary relationships between primary tumors, CTCs, lymph node and liver metastases using different computational approaches such as diffusion maps. Further defining and functionally annotating transcriptional heterogeneity will help us better understand the disease, and find the new Achilles heel for targeting. Note: This abstract was not presented at the conference. Citation Format: Nemanja Despot Marjanovic, Sheng Rong Ng, Aviv Regev, Tyler Jacks. Using single-cell RNA-seq approaches to decipher heterogeneity in autochthonous mouse models of small cell lung cancer [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A24.

Abstract B12: Centrosome clustering inhibition as a novel strategy to sensitize non-small cell lung cancer to radiation treatment and immunotherapy

Abstract The centrosome is a microtubule-organizing center that plays an important role during M-phase and cell division where it organizes the two poles of the bipolar microtubule spindle apparatus from which chromosomes are segregated. In contrast to most normal cells, the vast majority of lung cancers contain extra copies of centrosomes, and a large body of circumstantial evidence links extra centrosomes to the development of cancer. Cells with supernumerary centrosomes form multipolar mitotic spindles, which, if not corrected, lead to lethal multipolar divisions or “mitotic catastrophe.” To overcome this, lung cancer cells cluster their centrosomes into two spindle poles, enabling tumor cells to survive. Thus, inhibition of centrosome clustering (CCi) can force cancer cells into multipolar divisions, lead to chromosomal instability, make cancer cells more sensitive to radiation treatment, and eventually result in the selective killing of cancer cells. At the same time, this approach should spare healthy tissues with normal centrosome numbers, resulting in a favorable therapeutic index. To examine the relationship between CCi and radiation-treatment sensitivity, we first treated three NSCLC cell lines (A549, H460 and H358) and one immortalized primary human bronchial epithelial cell (HBEC) with griseofulvin—a centrosomal clustering inhibitor. Immunofluorescent staining for pericentrin and alpha-tubulin, we observed a significant increase in the number of cells with multipolar spindles in NSCLC cells compared to HBECs (40-80% vs. 15%). Next, we examined the ability of griseofulvin to radiosensitize these cells to radiation treatment. By using clonogenic assays, we saw significantly less clonogenic potential of NSCLC cells following radiation treatment after exposure to 30 uM of griseofulvin, but griseofulvin did not radiosensitize HBECs. To further confirm our results, we genetically targeted KIFC1, a member of kinesin-14 family of motor proteins that have been shown to be essential for clustering centrosomes in cancer cells, but are not required for division in normal cells. In our experiments, knockdown of KIFC1 expression with siRNA also significantly sensitized NSCLC cells to radiation treatment. Finally, when we treated the NSCLC cells with griseofulvin, we saw a significant increase in the formation of micronuclei in cancer cells. Recently, it has been shown that micronuclei are important sources of immunostimulatory DNA via activation of the cGAS-STING pathway. Therefore, inhibition of CC could potentially augment immunogenicity and increase efficiency of immunotherapy (IMT). In conclusion, inhibition of CC resulted in the formation of multipolar spindle in NSCLCs and further sensitized NSCLCs to radiation treatment. CC is not typically required for normal cells, and thus CCi may be a more specific and perhaps less toxic therapy for NSCLCs. The potential of CCi to enhance IMT is under investigation. Citation Format: Hailun Wang, Katriana Nugent, Matthew Ballew, Ghali Lemtiri-Chlieh, Natasha Raman, Michelle Levine, Andrew Holland, Phuoc Tran. Centrosome clustering inhibition as a novel strategy to sensitize non-small cell lung cancer to radiation treatment and immunotherapy [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr B12.

Abstract A16: Sodium-dependent glucose transporter 2 is a novel diagnostic and therapeutic target for early-stage lung adenocarcinoma

Abstract Lung cancer claims approximately 160,000 lives in the United States every year, and lung adenocarcinoma (LADC) is the most frequent type. Early diagnosis is crucial; the National Lung Screening Trial showed a 20% reduction in mortality in high-risk individuals screened by computed tomography (CT). However, CT has low specificity in early-stage LADC. In vivo detection of glucose uptake by FDG PET is a standard tool for lung cancer staging, but has low sensitivity in early-stage LADC. Here, we demonstrate that GLUT transporters, responsible for FDG uptake, are expressed in advanced, poorly differentiated LADC, whereas early LADC utilizes a different glucose transport system, the sodium-dependent glucose transporter 2 (SGLT2), that is not detected by FDG. This can be imaged with the novel tracer Me4FDG. We investigated the expression of two different glucose transporters, GLUT1 and SGLT2, in a cohort of 56 human LADC specimens, in patient-derived xenografts, and in a KrasG12D-driven, p53-null genetically engineered mouse model of LADC. We observed a switch in the modality of glucose transport during lung carcinogenesis: SGLT2 was highly expressed in premalignant lesions and well-differentiated LADC, whereas GLUT1 was upregulated in advanced, poorly differentiated lesions. This pattern was observed both in human samples and in murine models. This observation led us to hypothesize that early-stage LADCs are often negative on FDG PET because this imaging modality does not detect the activity of SGLT2, which is expressed in early lesions. Therefore, we performed PET imaging with the tracer Me4FDG, which measures SGLT2 activity, in our mouse model, and observed that Me4FDG accumulated in small nodules that were negative with FDG. We confirmed the functionality of SGLT2 in human LADC by Me4FDG PET in patient-derived xenografts. Finally, we tested the hypothesis that sodium-dependent glucose transport is a critical metabolic supply strategy in the early stages of lung adenocarcinoma development. We performed therapeutic trials in genetically engineered mouse models with the FDA-approved SGLT2 inhibitor canagliflozin, which slowed down tumor progression and prolonged mouse survival, by specifically reducing the proliferation rate of premalignant lesions. In conclusion, Me4FDG PET is a novel diagnostic test to image lung premalignancy and early-stage LADC. We propose that Me4FDG-PET will identify the patients who are likely to respond to SGLT2 inhibition, as a cancer interception strategy to prevent the progression of premalignancy to invasive disease, by blocking the supply of glucose required for proliferation of premalignant lesions. Citation Format: Claudio Scafoglio, Sean T. Bailey, Gihad Abdelhady, Jie Liu, Jane Yanagawa, Aditya Shirali, Dean W. Wallace, Jorge R. Barrio, Ernest M. Wright, Tonya Walser, Steven M. Dubinett, David B. Shackelford. Sodium-dependent glucose transporter 2 is a novel diagnostic and therapeutic target for early-stage lung adenocarcinoma [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A16.

Abstract IA38: Progress and unanswered questions in EGFR mutation-positive lung cancer

Abstract A summary of the major advances in management of EGFR mutation-positive lung cancer will be provided, along with a discussion of missing pieces of information that could improve the care of patients. Citation Format: Lecia V. Sequist. Progress and unanswered questions in EGFR mutation-positive lung cancer [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr IA38.

Abstract A29: The CUTO panel of patient-derived NSCLC cell lines reveals unique molecular characteristics and responses to targeted therapies

Abstract Cell lines generated from human tumors have been an invaluable tool in dissecting the underlying molecular mechanisms of cancer biology and for cancer drug development. The large library of original human lung cancer cell lines established by Minna and Gazdar, though instrumental in investigating various aspect of lung cancer biology, lacks examples of some of the diverse mutations responsible for this cancer. The CUTO (Colorado University Thoracic Oncology) cell line series, initiated in 2011, had the goal of curating a new panel of cells derived from NSCLC patients with distinct oncogenic drivers. By generating multiple, unique cell lines representing each oncogene driver subset, we believe that these lines would allow study of interpatient variability that underlie differential patient responses and duration of response to targeted therapies. The current CUTO panel consist of 33 cell lines with either 1) gene rearrangements in ALK, RET, ROS1 or NTRK1; 2) activating mutations in the ERBB gene family including exon 20 insertions in EGFR and ERBB2 (HER2) as well as rare mutations in EGFR; and 3) inactivation of the NF1 gene either as a concurrent mutation with other oncogenic alterations or as the sole driver mutation. For cells harboring gene rearrangements we have amassed four EML4-ALK lines, two KIF5B-RET cell lines, five ROS1 fusion lines (3 CD74-ROS1, 1 SDC4-ROS1, and 1 TPM3-ROS1), and one MPRIP-NTRK1 cell line. Our ERBB mutant lines include one cell line with HER2 exon 20 insertion, three lines with different exon 20 insertions in EGFR, and two EGFR mutant cell lines with compound mutations. We present our characterization of these cell lines in terms of their proliferation profiles and molecular signaling pathways in the presence of targeted inhibition. The derivation and characterization of these cell lines have facilitated the study of cell signaling in oncogene-driven cancer, have helped identify new resistance mechanisms, and have facilitated drug development for rare oncogenes such as NTRK1 gene fusions and EGFR exon 20 insertions. We expect that this growing library of cell lines will continue to further our understanding of oncogenic-driven tumor biology and provide mechanistic insight towards the development of novel therapeutics and drug combinations. Citation Format: Anh Tuan Le, Adriana Estrada-Bernal, Laura Schubert, Andrea Doak, Nan Chen, Kurtis Davies, Aria Vaishnavi, Mary Jackson, Vignesh Narayana, Kimi Kondo, John Mitchell, Michael Weyant, Tom Purcell, Paul Bunn, Ross Camidge, Janet Freeman-Daly, Robert Doebele. The CUTO panel of patient-derived NSCLC cell lines reveals unique molecular characteristics and responses to targeted therapies [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A29.

Abstract IA05: Early steps in lung oncogenesis

Abstract Lung cancer is the biggest cancer killer of men and women, and there has been no improvement in survival over the last 50 years. To reduce the mortality of lung cancer, we need to both detect and treat it earlier, ideally at a preinvasive cancer stage before progression into full-blown invasive cancer. This would allow us to initiate treatment in a timely manner, increasing the chances of cure and saving lives. Lung cancers arise from preinvasive lesions but, interestingly, not all these lesions progress to cancer. We believe that understanding the differences between the lesions that progress to invasive cancer and those that disappear will direct us to the key proteins and genes that are mis-expressed, or inappropriately switched on or off, that lead to invasive cancer formation. My talk will outline studies that shed new understanding on how lung cancer develops and identify novel biological markers that accurately predict whether a lesion will eventually grow into a cancer. This, I hope, will enable targeting potential therapies to those patients with lesions that are likely to progress and identify the key genes and proteins associated with cancer development, thereby opening the possibility of developing chemoprophylactic therapeutic approaches to prohibit cancer development. Citation Format: Samuel M. Janes. Early steps in lung oncogenesis [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr IA05.

Abstract A22: Establishing a deep cfDNA methylation sequencing-based signature for noninvasive early-stage lung cancer diagnosis

Abstract Although mortality rate has been greatly reduced by low-dose computed tomography (LDCT) screening of high-risk patients, the benefit is compromised by the high false-positive rate, thus necessitating the development of biomarkers differentiating benign and malignant nodules. The role of aberrant DNA methylation in the process of tumorigenesis both at individual genes and a genome-wide scale has been well elucidated. It occurs very early in cancer development, and thus is capable of serving as a diagnostic biomarker. In this prospective study, we evaluated the feasibility and performance of assessing cfDNA methylation status as an early diagnostic marker for lung cancer. We developed a novel targeted bisulfite sequencing assay called ELSA and designed a methylation panel based on TCGA data on 838 lung cancer patients and 74 normal controls. We then interrogated the methylation status of 7,572 regions consisting of about 100,000 CpG sites from plasma samples of 163 newly recruited Chinese patients with early-stage NSCLC and 41 with benign nodules using our panel. A region-based methylation statistic was used in the classification model. We constructed diagnostic classification models using elastic net and random forest based on 3,000 DMRs selected by logistic regression between benign and malignant nodules using FFPE samples. Within this cohort, 54 patients had malignant tumor (39 stage I, 6 stage II and 9 stage IIIa) and 30 patients had benign nodules. We performed 5-fold cross-validation with 20 time repeats to gain a robust estimation of model performance, achieving a sensitivity of 92.6%, specificity of 100% and AUC of 98.1%. We further refined the DMRs selected above using 68 healthy baseline plasmas, and subsequently constructed a weighting decision tree model for plasma samples obtained from a separate cohort consisting of 163 patients with malignant tumor (127 stage I, 36 stage II and above) and 41 patients with hamartoma, chronic inflammation, granuloma, or other lung benign lesions, achieving a sensitivity of 84.0%, specificity of 85.7% and AUC of 90.7% based on 170 markers. We further analyzed PPVs according to stage and tumor size. PPVs for patients with stage I, II, and III diseases are 82.4%, 84.6%, and 94.4%, respectively; and for tumor diameter less than 1 cm, between 1-2 cm, and greater than 2 cm are 78.5%, 87.1%, and 88.2%, respectively. Citation Format: Zhihong Zhang, Naixin Liang, Wanglong Deng, Chenyang Wang, Tao Zheng, Wenjun Li, Jian Hu, Jinh Su, Han Han-Zhang, Bingsi Li, Hao Liu, Shanqing Li. Establishing a deep cfDNA methylation sequencing-based signature for noninvasive early-stage lung cancer diagnosis [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A22.

Abstract PR08: Functional characterization and evolutionary reconstruction of small cell lung cancer transformation of EGFR-mutant lung adenocarcinomas

Abstract Purpose: Recently, we performed a whole-genome sequencing (WGS) study of serially acquired tumors and demonstrated that an early, complete inactivation of RB1 and TP53 is common among patients with EGFR-mutant lung adenocarcinoma (LADC) whose disease was transformed into small cell lung cancer (SCLC). To further understand this phenomenon, here we functionally characterized a patient-derived cell line of transformed SCLC and related in vitro models. We also deepened our genome analysis focusing on the complex rearrangements to probe their evolutionary paths in detail. Method: mRNA sequencing was performed for SNU-2962A cells and their transcriptome was analyzed with published mRNA sequencing datasets of primary LADCs (n = 87), and SCLCs (n = 81). A cell viability screening of SNU-2962A cells was performed to understand their therapeutic vulnerability using a kinome-targeting siRNA library. We also characterized isogenic RB1/TP53-knockout LADC cell lines with EGFR mutation (PC9, HCC-827 and HCC-4006). In addition, we reconstructed complex genomic rearrangements and analyzed them in relationship with whole-genome duplication to investigate their time points during clonal evolution. Results: The SNU-2962A cell line was established from the pleural effusion of the patient LC1. These cells exhibited adherent morphology in vitro and strongly expressed neuroendocrine markers. A hierarchical clustering of SNU-2962A cells with the published LADCs and SCLCs showed complete clustering with SCLCs. This cell line strongly expressed NEUROD1 and MYC, which was consistent with the variant-type SCLC. Kinome siRNA library screening showed a significantly reduced viability of SNU-2962A cells with knockdown of PI3K-MTOR pathway genes (PIK3CA or MTOR). Pharmacologic inhibition of this pathway using PI3K or AKT inhibitors showed moderate sensitivity. We tried to induce transformation into SCLCs using three different EGFR-mutant LADC cell lines by knockout of both RB1 and TP53 as well as knockin of PIK3CA E545K mutations using CRISPR, but we did not observe any meaningful evidence of neuroendocrine differentiation. We further analyzed our WGS data of 11 tumors from five patients, and found that many large-scaled genomic events such as whole-genome duplication, chromothripsis, or kataegis are frequently shared by LADCs and subsequent SCLCs, indicating their early occurrence. MYC amplification is frequently selected in transformed SCLCs. We also found a somatic deletion of BIM gene in LC3 patient, and this deletion was shared by the early LADC and the late SCLC. This may confer a dysfunctional apoptosis upon EGFR inhibition, which could predispose the early emergence of resistant tumor in this patient. Conclusion: Although a complete inactivation of both RB1 and TP53 is necessary for transformation from LADC into SCLC, it is not sufficient. Phylogenic reconstruction of large structural variations in our WGS dataset further clarified the mutational complexity of the common ancestor clones. This abstract is also being presented as Poster A23. Citation Format: June-Koo Lee, Seongyeol Park, Joon Kim, Tae Min Kim, Young Seok Ju. Functional characterization and evolutionary reconstruction of small cell lung cancer transformation of EGFR-mutant lung adenocarcinomas [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr PR08.

Abstract IA06: Molecular mechanisms of lung cancer development: Between metabolic reprogramming and genomic instability in the field of cancerization

Abstract Airway epithelial cells are prone to assault by risk factors and considered to be the primary cell type involved in the field of cancerization. To investigate risk-associated changes in the bronchial epithelium proteome that may offer new insights into molecular pathogenesis of lung cancer, proteins were identified in the airway epithelial cells of bronchial brushings specimens from risk-stratified individuals by shotgun proteomics. Differential expression of selected proteins was validated by parallel reaction monitoring mass spectrometry in an independent set of individual bronchial brushings. We identified over 300 proteins with a significant trend in expression. Pathway analysis revealed enrichment of carbohydrate metabolic enzymes in high-risk individuals. Glucose consumption and lactate production were increased in human bronchial epithelial BEAS2B cells treated with cigarette smoke condensate for seven months. Increased lipid biosynthetic capacity and net reductive carboxylation were revealed by metabolic flux analyses of [U-13C5] glutamine in this in vitro model, suggesting profound metabolic reprogramming in the airway epithelium of high-risk individuals. These results also pointed towards a fundamental role of glutamine and cysteine metabolism in lung cancer progression. Two amino acid transporters, specifically SLC1A5 and SLC7A11, captured our attention for their high level of expression in lung cancer. These transporters are associated with worse survival in lung cancer. A small-molecular inhibitor of SLC1A5 prevents tumor progression in nude mice, and a glutaminase inhibitor CB-839 exhibits radiosensitization to lung cancer cells, an effect that may be mediated by decreased GSH production. Taken together, these studies provide a rationale for new chemopreventive strategies, selection of patients for surveillance programs, and determining how metabolic reprogramming may contribute genomic instability a requirement to cancer development. Further investigation into how glutamine and cystine metabolism contribute to lung cancer development and progression may lead to novel early-interception strategies in lung cancer. This work is supported by NCI CA186145, CA152662, and DoD W81XWH-11-2-0161. Citation Format: Jamshed Rahman, Xiangming Ji, Chirayu Patel, Jamey Young, Pierre P. Massion. Molecular mechanisms of lung cancer development: Between metabolic reprogramming and genomic instability in the field of cancerization [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr IA06.

Abstract IA12: Developing precision medicine-based new lung cancer therapeutics

Abstract We have used a “chemistry first” approach to discover druggable acquired vulnerabilities that have been acquired in the pathogenesis of non-small cell lung cancer (NSCLC). We screened chemical libraries (~200,000 compounds) for chemical toxins that killed subsets of NSCLC but not normal human lung epithelial cells (HBECs). We first screened a panel of 12 NSCLC lines that represented a variety of known oncogenotypes and identified chemicals with large Z scores and appropriate properties, including re-supply, chemistry, and reproducible drug response phenotypes, and from this narrowed down a list of 202 chemicals and 18 drugs with known targeting (we called our “Precision Oncology Probe” set, POPS). These and a panel of 30 clinically available drugs, targeted therapies, and drug combinations, already in use or in trials for NSCLC treatment, were then tested on a panel of 96 NSCLC lines for their drug response phenotypes in 12-point dose response curves. This information was analyzed using scanning ranked KS (Kolmogorov–Smirnov) and elastic net biostatistics approaches to identify molecular biomarkers (mutations, mRNA expression, copy number variation, protein expression, and metabolomics) that could predict for sensitivity or resistance to a particular chemical toxin or treatment regimen. From this we have discovered that our approach identifies already known molecular biomarker drug sensitivities (e.g., EGFR mutations and EGFR TK inhibitors); many clinically available chemotherapy agents have molecular biomarkers predicting preclinical model drug responses; the POP set of chemical toxins provides novel drug-response phenotype patterns in the large NSCLC panel different from those found with clinically available agents including a therapeutic window; many of the POP toxins only hit a small % (~5%) of the NSCLC panel but the POP set as a whole provides “coverage” of the entire NSCLC panel; there are simple, one- or two-component molecular biomarkers (mutations, mRNA expression) that predict responses to the different chemical toxins in the NSCLC panel; and that the molecular biomarkers provide some information on the targets and pathways involved in response to the chemical toxins. Thus, we have identified a group of chemical toxins with selectivity for subsets of NSCLC and associated tumor molecular biomarkers to facilitate their development for precision medicine and also, in some cases, information on the targets and pathways interdicted by these chemical compounds. In addition, we have discovered NSCLC predictive biomarkers for clinically available agents. University of Texas SPORE in Lung Cancer (P50CA70907), NCI CTD2N (U01 CA176284), and CPRIT Grants. Citation Format: John D. Minna, Elizabeth McMillan, Luc Girard, Michael Peyton, Kenneth Huffman, Dhruba Deb, Paul Yenerall, Amit Das, Longshan Li, Maithili Dalvi, Boning Gao, Yang Xie, Yonghao Yu, Suzie Hight, Rachel Vaden, Caroline Diep, Michael Roth, Bruce Posner, John MacMillan, Ralph Deberardinis, David Wheeler, John V. Heymach, Ignacio I. Wistuba, Adi Gazdar, Michael White. Developing precision medicine-based new lung cancer therapeutics [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr IA12.

Abstract IA02: The role of aneuploidy during tumorigenesis

Abstract Since Boveri’s postulates in 1914, aneuploidy has been recognized as a hallmark of human cancer; however, no comprehensive theory exists to explain its patterns and its role in tumor evolution. We have developed Tumor Suppressor and Oncogene (TUSON) Explorer, a computational method for the prediction of tumor suppressor genes (TSG) and oncogenes (OG) through the analysis of the pattern of somatic mutations in cancer. By integrating our prediction of TSGs and OGs with information of copy number alterations, we found that the distribution on chromosomes and the potency of TSGs and OGs can predict the frequency of arm-level and chromosome-level deletions and amplification (Davoli et al., 2013). We propose that it is the decrease or increase in the gene dosage of TSG and OG genes in deletion and amplification to determine the recurrent patterns of aneuploidy selected during tumor evolution. More recently, we have performed gene expression analysis on tumor samples, comparing tumors with different levels of aneuploidy (Davoli et al., 2017). We have found that tumors with high levels of aneuploidy show increased markers of proliferation as well as a strong decrease in infiltrating immune cells, especially CD8+ T cells. Notably, we show that aneuploidy predicts patients’ survival in two clinical trails of immunotherapy in cancer patients. Compared to neoantigen load, aneuploidy represents an independent and stronger predictor of survival. A combined score including aneuploidy and neoantigen load allows superior prediction of survival after immunotherapy. We are now investigating the mechanism by which aneuploidy affects cancer immune evasion. I will present novel data suggesting that aneuploidy may play a role in inhibiting and dysregulating antigen processing and presentation in tumor cells. Overall, our data indicate that aneuploidy is a driver event in human tumorigenesis. Citation Format: Teresa Davo4li. The role of aneuploidy during tumorigenesis [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr IA02.

Abstract B09: Evaluating transcription factor networks as targets for the treatment of small cell lung cancer

Abstract Achaete-scute homolog 1 (ASCL1) is a transcription factor that is highly expressed in most small cell lung cancer (SCLC) tumors and cell lines and is critical for their proliferation. Inhibiting ASCL1 transcriptional activity may thus be a valid therapeutic strategy for SCLC. However, transcription factors (TFs) have been historically difficult to target. Given that TFs belong to complex regulatory networks, we propose to identify and target multiple components within an ASCL1-transcriptional network rather than ASCL1 alone. We have previously detected ASCL1-bound genomic regions associated with the active chromatin epigenetic mark, H3K27Ac in SCLC cells. Here we report the identification of an ASCL1-containing transcriptional network in SCLC cells and evaluate its targeting by the transcriptional inhibitor, mithramycin. We conducted chromatin immunoprecipitation (ChIP) by immunoprecipitating the H3K27Ac chromatin mark from NCI-H2107 and NCI-H69 SCLCs (5X107 cells) nuclear lysates with10 μg H3K27Ac antibodies (Abcam-Ab4729), and sequenced ChIP libraries on an Illumina High-Seq2000. Further characterization using siRNA transfections, immunoblotting and MTS-cell proliferation assays were conducted following standard procedures. We found 82 overlapping TF genes, including lineage-specific TF genes for ASCL1, Forkhead box protein A2 (FOXA2), and nuclear factor 1 B-type (NFIB), associated with the active chromatin epigenetic marks H3K27Ac in SCLC cell lines. This suggests that ASCL1, FOXA2, and NFIB may belong to the same transcriptional network. siRNA-mediated knockdown of ASCL1 leads to decreased FOXA2 levels, but it has no effect on NFIB levels, implicating ASCL1 as a transcriptional regulator of FOXA2 but not NFIB. Mithramycin stops cell proliferation and reduces ASCL1, FOXA2 but not NFIB protein levels. In conclusion, our studies identify the beginning of a transcriptional network necessary for SCLC proliferation in which ASCL1 regulates FOXA2. Targeting multiple TFs within a network with mithramycin can stop cell proliferation and as such suggests a therapeutic strategy for SCLC. Because of the importance of NFIB in mediating metastatic behavior of SCLC, these data indicate that approaches other than targeting ASCL1 and use of mithramycin will be needed to target NFIB. Citation Format: Karine Pozo, Rahul K. Kollipara, John D. Minna, Adi F. Gazdar, Jane E. Johnson. Evaluating transcription factor networks as targets for the treatment of small cell lung cancer [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr B09.

Abstract IA11: Gene networks, airway smoke injury, and cancer risk

Abstract Background: Fifteen percent of smokers and former smokers develop lung cancer. This might be chance; an alternative is that some are at greater risk because of different individual responses to airway smoke injury. Several groups have investigated gene expression in normal airway epithelium as a readout of smoke injury and as a stratifier for risk. Questions include, but are not limited to: i) Can nasal epithelium provide a useful and accessible surrogate for bronchial? ii) Is there an inherited component to the injury-related risk? iii) What components of the injury response contribute most to increased risk in current and in former smokers? iv) What is the distribution of injury response in the healthy current and former smoker populations, rather than symptomatic patients, and how does this relate to risk? Results and Work in Progress: We, like others, have based our analysis on gene expression in normal bronchial and nasal epithelium. We use transcription factor-centered gene regulatory networks, derived from normal bronchial and nasal epithelium, to identify the overlying cellular regulatory mechanisms that differ between injury responses. We identify transcriptional “master regulators” (MRs), the activity of which differs between cancer and benign, and confirm the overlap of MRs between nasal and bronchial epithelium. A preliminary pathway analysis suggests the involvement of these MRs in oxidative stress response, DNA repair, immune responses, and developmental processes. Using data from the TRICL consortium, we show that the regulons of a subset of these MRs are enriched in genes located at lung GWAS loci. This suggests that the observed gene expression differences between cancer and benign are in part causal, related to inherited variation rather than secondary to the cancer. Using samples from the Cambridge Bioresource volunteers, we are analyzing data from nasal samples to compare healthy current and former smokers with healthy never-smokers, and so obtain a “smoke injury response.” We will use this to describe the distribution of injury response in the healthy smoker/former smoker populations. Our aim is to test what component of risk-related gene response may be apparent in this healthy asymptomatic smoker population rather than at the “later” symptomatic stages of hospital investigation, and to find which genes or regulatory processes contribute most to this. The results may have implications both for risk prediction and possible interventions in these populations. My talk will be a “work in progress” aimed at stimulating discussion about questions and approaches, rather than a presentation of finished results. Citation Format: Bruce A.J. Ponder. Gene networks, airway smoke injury, and cancer risk [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr IA11.

Abstract A02: Oncogenic drivers of lung cancer induce production of CCL5 and recruitment of regulatory T-cells

Abstract Lung cancer development is driven by the expression of mutant oncogenes, with EGFR and KRAS being the most frequent mutations in lung adenocarcinoma. However, additional factors may influence lung tumor development and progression, including the balance of antitumor immune effector cells and pro-tumorigenic immune suppressor cells within the lung and lung tumor microenvironment. Tumor cells can evade immune attack by producing cytokines that recruit immune modulatory cells, such as regulatory T cells (Tregs), that promote a localized immune suppressive environment We hypothesized that oncogene signaling regulates the production of cytokines by tumor cells at the earliest stages of transformation that can recruit immune suppressive cells and promote lung tumour development. We used CIBERSORT-based analysis of gene expression data to quantify 22 different immune cell types from over 300 human lung adenocarcinomas and 100 matched normal lung tissues. We found that Tregs were significantly enriched in early-stage lung adenocarcinoma tumors compared to matched normal tissue from the same patient, and validated these findings with immunohistochemistry staining of lung sections. To identify cytokines that could recruit Tregs early in lung tumorigenesis, we used normal cells expressing doxycycline-inducible wild-type EGFR, mutant EGFRL858R or mutant KRASG12V. Secreted cytokines were quantified using a multiplex LUMINEX assay with subsequent validation by ELISA. Induction of EGFRL858R and KRASG12V expression in normal cells rapidly increased the production of CCL5 (RANTES), as did expression of wild-type EGFR in the presence of exogenous EGF. To elucidate the mechanism of oncogene-driven CCL5 secretion, we treated lung cancer cells harboring EGFR or KRAS mutations with a MEK inhibitor (trametinib) to disrupt oncogenic signaling downstream of EGFR and KRAS. In KRAS mutant lung cancer cells, treatment with trametinib decreased CCL5 production and inhibited both ERK and AKT signaling. To determine if oncogene-driven cytokines could induce migration of Tregs ex vivo, we used a trans-well assay with conditioned media from cells expressing doxycycline inducible EGFRL858R or KRASG12V. Conditioned media from EGFRL858R and KRASG12V-expressing cells induced Treg migration, which was mitigated by the addition of an anti-CCL5 antibody. These data indicate that oncogenic EGFR and KRAS signaling regulates expression of CCL5 in lung tumor cells, and that CCL5-mediated Treg recruitment to lung tumors may occur in early stages of lung tumor development. Therefore, targeted inhibition of CCL5, Tregs, and/or oncogenic EGFR and KRAS signaling may represent therapeutic strategies to block recruitment and function of immunosuppressive Tregs during lung tumor development. Citation Format: Elizabeth Franks, Elizabeth C. Halvorsen, Etienne Melese, Unni Arun, Jenna L. Collier, Bryant T. Harbourne, Min Hee Oh, Lam Vivian, Gerry Krystal, John C. English, Wan L. Lam, Stephen Lam, Ninan Abraham, Kevin L. Bennewith, William W. Lockwood. Oncogenic drivers of lung cancer induce production of CCL5 and recruitment of regulatory T-cells [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A02.

Abstract A26: Identification of a novel therapeutic target in lung adenocarcinoma

Abstract The reactivation of biologic signaling events that occur throughout fetal development has been observed during malignant cell transformation and tumor progression. Transcription factors are typically at the hub of these signaling events, such as NKX2-1 and several ETS transcription factors. ELF3 is an uncharacterized ETS family member that is highly expressed during fetal lung development and could play a biologic role in lung cancer based on its location within the recurrently gained chromosome 1q. We hypothesize that ELF3 is a novel oncogenic transcription factor and a therapeutic target. Multiple independent datasets encompassing 1,685 clinical samples of lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LUSC), small cell lung cancer, and nonmalignant lung tissues were analyzed to establish the frequency of ELF3 overexpression and underlying genetic mechanisms of selection. ELF3 overexpression was validated by immunohistochemistry. Associations with patient survival were tested using the log-rank method. Isogenic cell lines were established to assess oncogenic phenotypes including tumor growth in a xenograft model. Protein-protein interaction (PPI) networks were constructed around ELF3, and integrated pathway analysis was performed to decipher the signaling network disruptions resulting from ELF3 overexpression. ELF3 overexpression was frequently observed in LUAD (>2-fold: BCCA 73% TCGA 40%), but was not observed in other lung cancer subtypes. Similarly, high ELF3 expression was significantly associated with poor overall survival of LUAD patients (p<0.0001), but not LUSC patients. These clinical associations prompted further examination of ELF3 in the LUAD subtype of lung cancer. ELF3 knockdown in LUAD cell lines resulted in significantly reduced proliferation, viability, and anchorage-independent growth, demonstrating that ELF3 regulates oncogenic phenotypes. Loss of ELF3 abolished the ability of LUAD cells to establish tumors in xenograft mouse models, demonstrating the requirement of ELF3 expression for tumor growth. ELF3 overexpression is associated with remodeling of 23 direct PPI networks, resulting in loss of interaction with proteins such as NFKB1 and MYC, while forming new interactions with NKX2-1, HOXA5 and ERBB3. Pathway analysis suggests a transcriptional reprogramming from inflammatory and MAPK signallng in nonmalignant and ELF3-low tissues, to adhesion and motility pathways in transformed tissues displaying high ELF3 expression. Core pathways included cell cycle, apoptosis, WNT, and NOTCH signaling, agreeing with our cell models. While mutations in ELF3 were rare, up to 80% of LUAD patients harbored focal amplification, DNA gain, and/or promoter hypomethylation at the ELF3 locus, which resulted in transcript overexpression. We have deciphered the oncogenic role of ELF3 in LUAD. Its requirement for tumor growth in our model indicates that therapeutic targeting of ELF3 could benefit the 73% of patients who display ELF3 overexpression. Citation Format: Katey S.S. Enfield, Erin A. Marshall, Christine Anderson, Kevin W. Ng, Sara Rahmati, Zhaolin Xu, Calum E. MacAulay, Stephen Lam, William W. Lockwood, Raj Chari, Aly Karsan, Igor Jurisica, Wan L. Lam. Identification of a novel therapeutic target in lung adenocarcinoma [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A26.

Abstract A30: Gene regulatory mechanisms governing invasive mucinous adenocarcinoma of the lung (IMA)

Abstract Background: Invasive mucinous adenocarcinoma of the lung (IMA) comprises 5-10% of lung adenocarcinoma. IMA is pathologically characterized as lung tumor cells with goblet cell morphology containing abundant intracytoplasmic mucin. Recent advances using next-generation sequencing technology revealed driver oncogenes (e.g., KRAS mutation and NRG1 fusions) and a gene expression signature (e.g., NKX2-1 negative and SPDEF positive) inducing IMA. However, the genetic mechanism governing the pathogenesis of IMA is not fully understood. Here, we report the mechanism of the gene regulatory network regulating mucins and immune checkpoints in IMA. Methods: Using lentiviral vectors expressing the anti-mucous transcription factor NKX2-1 (also known as TTF-1) and the pro-mucous transcription factor SPDEF, we assessed their functions in regulating the expression of mucins and immune checkpoints in multiple lung cancer cell lines. Using ChIP-seq datasets, we further identified genome-wide binding sites of NKX2-1 and SPDEF in A549 lung carcinoma cells that carry a KRAS mutation and produce mucins. We further deleted such binding sites in the genome using CRISPR/Cas9 to assess the function of the sites. Results: NKX2-1 influenced the expression of immune checkpoints, including PD-L1, and mucin-related genes, including MUC5AC and SPDEF. Notably, NKX2-1 bound to the same noncoding region upstream of MUC5AC as SPDEF. The genomic deletion study using CRISPR/Cas9 indicated that both NKX2-1 and SPDEF bound to the enhancer region (but not the silencer region) of MUC5AC, suggesting that NKX2-1 suppresses mucin gene expression by blocking the activity of the enhancer region but not by recruiting a co-repressor complex such as HDAC to the region. Conclusion: Transcription factors NKX2-1 and SPDEF function as yin and yang to affect the pathogenesis of IMA. Citation Format: Minzhe Guo, Koichi Tomoshige, Iris Fink-Baldauf, Yutaka Maeda. Gene regulatory mechanisms governing invasive mucinous adenocarcinoma of the lung (IMA) [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A30.

Abstract A36: Establishment of patient-derived xenograft models of lung adenocarcinoma with two different EGFR mutations, L858R and exon19 deletion

Abstract Preclinical tumor models that resemble the characteristics of patients’ tumors are essential for development of novel therapy. Patient-derived tumor xenograft (PDX) models are considered to be interesting candidates. However, success rate of PDX models of lung cancer, especially adenocarcinoma, is much lower than that of other types of tumors, including colon cancer. This study was conducted to establish clinically relevant PDX models of lung cancer. We obtained fresh surgically resected specimens from 20 NSCLC patients (10 squamous cell carcinomas, 9 adenocarcinomas, and 1 large cell carcinoma) who had submitted the written informed consent for this study, which was performed in accordance with protocols approved by the IRB of Kanazawa University. We inoculated the tumor fragments subcutaneously into immunodeficient mice, NOG mice or SCID hairless outbred (SHO) mice. The tumor take rate was 40% (8/20), with higher rates for squamous cell carcinoma (60%, 6/10) than for adenocarcinoma (22%, 2/9). Interestingly, both of two adenocarcinomas that successfully developed PDX tumors had EGFR mutations, L858R and exon 19 deletion, respectively. These PDX tumors maintained histology of adenocarcinoma and EGFR mutation even after repeated in vivo passages. Furthermore, the third-generation EGFR tyrosine kinase inhibitor (EGFR-TKI) osimertinib dramatically regressed tumors in these EGFR mutated PDX models. Osimertinib treatment is ongoing in these EGFR-mutated PDX models to obtain tumors that acquired resistance to osimertinib. Our PDX models of lung adenocarcinoma with different EGFR mutations may be useful for investigating the mechanisms of resistance to EGFR-TKIs, including third-generation inhibitor osimertinib. Citation Format: Kenji Kita, Koji Fukuda, Akihiro Nishiyama, Sachiko Arai, Shinji Takeuchi, Koushiro Ohtsubo, Kaname Yamashita, Masaya Tamura, Isao Matsumoto, Seiji Yano. Establishment of patient-derived xenograft models of lung adenocarcinoma with two different EGFR mutations, L858R and exon19 deletion [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A36.

Abstract A37: Tobacco smoke increases lung adenocarcinoma risk by downregulating TGF-beta and AhR-regulated focal adhesion proteins involved in injury resolution

Abstract Background: The most prevalent type of lung cancer—lung adenocarcinoma—arises from the most distal reaches of the lung in the alveolar epithelium. The alveolar epithelium consists of two cell types: (i) thin type I alveolar epithelial cells (AT1) that facilitate gas exchange and constitute more than 90% of the alveolar surface, and (ii) cuboidal surfactant-producing type II (AT2) cells. Following injury, dynamic interactions between extracellular matrix components, integrins, focal adhesion proteins, and cytoskeletal components allow AT2 progenitor cells to migrate, spread, and transdifferentiate into AT1 cells to restore the denuded epithelial barrier. Methods: Utilizing primary human AT2 cells isolated from lung transplant remnants, our lab has established and extensively profiled an in vitro model of epithelial restitution, wherein AT2 cells differentiate over a period of six days into terminally differentiated AT1-like cells. Because cigarette smoking is the most recognized risk factor for lung adenocarcinoma (LUAD), and because the alveolar space is the primary target of tobacco smoke, we also compared the RNA-seq profiles of AT2>AT1 differentiation to those of A549 LUAD cells exposed to cigarette smoke condensate (CSC). Results: We find that CSC downregulates many factors that are upregulated in the AT2>AT1 restitution model, particularly integrins, extracellular matrix components, and novel focal adhesion proteins that are also downregulated in LUAD relative to adjacent normal tissue. We describe regulation of four such novel focal adhesion proteins, specifically by TGF-β and aryl hydrocarbon receptor (AhR)-mediated mechanisms. Conclusions: Our findings shed light on early mechanisms in the carcinogenic process, and support the notion that alveolar cells engaged in repeated cycles of injury and repair are critical actors in LUAD risk trajectories. Note: This abstract was not presented at the conference. Citation Format: Theresa Ryan Stueve, Chenchen Yang, Crystal N. Marconett, Chunli Yan, Beiyun Zhou, Zea Borok, Ite A. Laird-Offringa. Tobacco smoke increases lung adenocarcinoma risk by downregulating TGF-beta and AhR-regulated focal adhesion proteins involved in injury resolution [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A37.

Abstract A12: Non-small cell lung tumor-derived autoantibodies can distinguish benign from malignant pulmonary nodules

Abstract High-risk smokers are currently being screened with low-dose CT imaging after the National Lung Screening Trial demonstrated a 20% reduction in lung cancer mortality. Unfortunately, small pulmonary nodules are a common finding in smokers (present in 25-50%), with very few (<2.5%) nodules ultimately determined to be cancer. We observed a 4-fold increase in lung tumor-associated B cells compared to normal lung and hypothesized that lung-tumor derived autoantibodies could be useful biomarkers for malignant nodule diagnosis. We first discovered 13 IgG or IgM autoantibodies found in both non-small cell lung tumor and peripheral plasma (n=10) on a large (~17,000) protein array. These autoantibodies were predominantly absent in plasma from patients with benign nodules (n=10). In the plasma from an expanded nodule positive cohort (n=250) we validated 3/13 autoantibodies on a targeted protein array and 5/13 autoantibodies on a high-density antibody array. By using both antibody arrays (to measure antigen-autoantibody complexes) and protein arrays (to measure free autoantibody), we obtained a comprehensive picture of malignant nodule-associated autoantibodies. When we combine our top two autoantibodies (one free and one complexed with antigen) with two features from CT images, they yield an AUC = 0.92 (72% sensitivity at 90% specificity). These observations indicate that our noninvasive autoantibody biomarkers can increase the diagnostic accuracy of pulmonary nodules identified by CT imaging. Citation Format: Kristin J. Lastwika, Julia Kargl, Yuzheng Zhang, Xiaodong Zhu, David Shelley, Edward Lo, Wei Wu, Sudhakar N. Pipavath, Paul E. Kinahan, A. McGarry Houghton, Paul D. Lampe. Non-small cell lung tumor-derived autoantibodies can distinguish benign from malignant pulmonary nodules [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr A12.

Abstract B33: Expansion study of pegylated arginine deiminase (ADI-PEG20), pemetrexed, and cisplatin in patients with ASS1-deficient non-squamous non-small cell lung cancer (TRAP)

Abstract Background: Loss of the metabolic tumor suppressor argininosuccinate synthetase 1 (ASS1) sensitizes non-squamous non-small cell lung cancer (NSCLC) cells to arginine deprivation with pegylated arginine deiminase (ADI-PEG20), which also disrupts thymidine pools and potentiates pemetrexed (Pem) cytotoxicity. In the phase I dose-escalation TRAP study (NCT02029690) we showed that ADI-PEG20 with first-line PEM and cisplatin (Cis) chemotherapy (ADIPemCis) produced a 100% disease control rate (DCR) (n = 9) of ASS1-deficient thoracic cancers, with no additional toxicity (Beddowes et al., 2017). Here, we present the TRAP expansion cohort experience in non-squamous NSCLC. Methods: Good-performance (ECOG 0-1) NSCLC patients (pts) with nonresectable disease were enrolled in a phase I TRAP expansion cohort at the maximum tolerated dose (MTD) of ADIPemCis, using tumoral ASS1 loss as a selection biomarker. Pem (500mg/m2) and Cis (75mg/m2) were given every 3 weeks with weekly IM ADI-PEG20 (36mg/m2) for a maximum of 4 cycles with maintenance ADI-PEG20 or Pem in responding pts. Primary endpoint was tumor response rate (RECIST 1.1), with secondary endpoints including progression-free survival (PFS), overall survival (OS), and toxicity. We measured plasma arginine and citrulline concentrations, ADI-PEG20 antibodies, early molecular responses by 18-FLT-PET/CT, and biopsied pts on progression to explore resistance mechanisms. Results: 21 ASS1-deficient non-squamous NSCLC pts (median age 60) were enrolled out of 70 screened pts; 14% (3/21) had stable brain metastases at entry. Plasma arginine decreased with a reciprocal increase in plasma citrulline. The partial response rate was 48% (n=10/21) with a DCR of 85.7% (n=18/21). Median PFS was 2.6 months and median OS was 5.3 months. 23.8% (n=5/21) pts experienced grade 3/4 treatment-related toxicities, the most common being non-febrile neutropenia (14.3%). A partial metabolic response was detected in 57.1% (n=4/7) pts by 24 hrs of arginine deprivation using 18-FLT-PET/CT with EORTC response criteria. Upregulation of ASS1 expression was observed in 1/3 biopsies on progression. Conclusions: The ADIPemCis regimen is active in ASS1-deficient non-squamous NSCLC pts with a higher than expected response rate. However, the short survival (PFS and OS) compared with ASS1-agnostic historical controls indicates that ASS1 loss selects for a biologically more aggressive phenotype of NSCLC. Further studies are warranted exploring rationale combinations with ADIPemCis in patients with ASS1-deficient non-squamous NSCLC. Citation Format: Melissa Phillips, Teresa Szyszko, Peter Hall, Sukaina Rashid, Ramsay Khadeir, Louise Lim, Jeremy Steele, John Conibear, Paula Wells, Stephen Ellis, Xiaoxing Feng, Jim Thomson, Amanda Johnston, Bor-Wen Wu, John Bomalaski, Simon Pacey, Gary Cook, Peter Szlosarek. Expansion study of pegylated arginine deiminase (ADI-PEG20), pemetrexed, and cisplatin in patients with ASS1-deficient non-squamous non-small cell lung cancer (TRAP) [abstract]. In: Proceedings of the Fifth AACR-IASLC International Joint Conference: Lung Cancer Translational Science from the Bench to the Clinic; Jan 8-11, 2018; San Diego, CA. Philadelphia (PA): AACR; Clin Cancer Res 2018;24(17_Suppl):Abstract nr B33.