Functional Drug Sensitivity Research Articles

Abstract 6890: 2D Tumor cell culture based Ex vivo drug sensitivity screening platform for South African patient samples (leukemia, multiple myeloma, ovarian and uterine cancer) demonstrate potential for targeted cancer therapies

Abstract Resistance to standard chemotherapy is a major challenge with Leukemia (Leuk), Multiple Myeloma (MM), Ovarian Cancer (OvCa) and Uterine Cancer (UtCa) treatments, which ultimately leads to relapse. Our study objective is to establish a robust drug sensitivity screening platform that can identify drugs and drug combinations that are effective in precision medicine for individual South African patient samples (Leuk, MM, OvCa and UtCa). Given that resistance to chemotherapy treatment accounts for high mortality and morbidity, we focus on the hypothesis that several drug combinations will be required for patients to achieve better response for drug regimens beyond first line of chemotherapy. Therefore, our defined objectives were to 1) establish 2D ex vivo tumour cell culturing using South African patient samples; 2) develop a drug sensitivity screening platform through which selected single and drug combinations will be identified; 3) to provide South African cancer (OvCa, Ut Ca, CLL, MM) patients with individualized treatment options with effective drug combinations for targeted therapies. Through collaboration with, Steve Biko Hospital Pretoria, Chris Hani Baragwanath Hospital and Wits Donald Gordon Medical Centre, Johannesburg, South Africa, we performed 100 patient sample collections including Acute myeloid leukaemia (AML) (n=7), Chronic lymphocytic leukaemia (CLL) (n=4), Chronic myeloid leukaemia (CML) (n= 30), MM (n=40), OvCa (n=10), UtCa (n=4) and healthy donors (n=5). Culture setting for all patient-derived tumor cells were established through the 2D cancer cell culturing in 96 well plates. This was followed by functional drug sensitivity testing using a panel of FDA approved chemotherapy drugs. The average turnaround time from biopsy to drug sensitivity test results is currently at 14 days. We have completed ex vivo drug sensitivity screening to perform analysis on patient samples (n=80) using 30 drugs with a concentration range of 1-1000 nm. Our preliminary drug sensitivity screening studies show good response to proteosome inhibitors, melphalan, doxorubicin and kinase inhibitors. We have established a wound healing assay for OvCa and UtCa samples and demonstrated sustained cell growth for the period of 2 weeks with tumor cells growing faster than twin normal tissue. We are now at stage of performing drug sensitivity screening for OvCa and UtCa patient samples for both twin normal and tumor tissues. We have established a drug sensitivity screening platform and wound healing assay with primary patient derived cells that utilizes South African Patient samples to select effective drugs for monotherapy and also drug combinations. Citation Format: Mutsa Takundwa, Vanelle Kenmogne, Thudzelani Malise, Bernice Monchusi, Phumuzile Dube, Ekene Nweke, June Fabian, Heather Maher, Justin du Toit, Vinitha Philip-Cherian, Pascaline Fru, Adekunle Sajo, Arnold (Arrie) Mouton, Matthys C. van Aardt, Cathy Visser, Greta Dreyer, Deepak Govindaraj. 2D Tumor cell culture based Ex vivo drug sensitivity screening platform for South African patient samples (leukemia, multiple myeloma, ovarian and uterine cancer) demonstrate potential for targeted cancer therapies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 6890.

Abstract 3904: Multiplexed flow cytometry based immunophenotyping paired with functional ex vivo (MFLEX) drug profiling informs potential efficacy of therapeutic agents in acute myeloid leukemia

Abstract Introduction: Acute Myeloid Leukemia (AML) is a heterogenous hematological malignancy with a complex clonal architecture that requires in-depth immunophenotyping for disease characterization. Conventional AML ex vivo models for preclinical drug screening lack functional readouts that provide single cell level data to identify drug activity in defined AML subsets with unique immune signatures. Here, we present a novel platform that links Multiparameter Flow cytometry with high-throughput EX vivo drug screening (MFLEX) in AML patient-derived cell models to simultaneously detect phenotypic changes and predict cell-subset specific drug responses in AML. Method: We developed a 21-color high-dimensional immune profiling panel that captures AML heterogeneity by delineating primitive and mature AML blasts along with normal hematopoietic lineages. The panel also includes markers for functional measurement of apoptosis, proliferation, differentiation, and expression of BCL2 family proteins in lymphoid and myeloid cell subsets. Ex vivo culture conditions were optimized using physiologically relevant stroma-free cytokine-rich media that supports primary AML cells with minimal phenotypic shift, allowing for high throughput functional drug sensitivity testing. Timepoints and dosing conditions for AML standard of care (SoC) agents were optimized to support combination screens. Multi-parameter datasets were analysed for functional differences and visualized using dimension reduction methods. Results: We tested MFLEX on clinically annotated AML samples to understand sensitivity profiles to venetoclax and combination partners: 5-azacytidine, cytarabine, fludarabine and gilteritinib. We identified distinct drug sensitivity profiles for venetoclax in AML patients with varying leukemic differentiation states where M0/M1 subtypes were more sensitive to therapy than M4/M5. We also observed cell lineage specific sensitivities, with myeloid blast being more sensitive to venetoclax compared to autologous CD4+ and CD8+ T-cells and this observed difference was correlated with higher expression of BCL2. Our platform demonstrated that patients who previously failed venetoclax therapy could benefit from a combination with 5-azacytidine or fludarabine and this was patient-specific. We further evaluated MFLEX’s predictive value for patient clinical response by systematically comparing venetoclax sensitivity in patient-derived xenograft models in vivo vs ex vivo and found a high correlation between drug responses. Conclusion: We have established a unique, immune drug screening platform, MFLEX, that has high translational value and can be implemented in clinical trials to identify drug response patterns, novel combinations, and potential biomarkers of response for patient stratification in AML. Citation Format: Reecha Shah, Heba Wander, Michelle Hsueh, Muskan Floren, Wei Liu, Patricia Cheung, Courtney Andersen, Michael White, Nathan Kingston, Pablo Morentin Gutierrez, Lisa Drew, Giulia Fabbri, Elena Bibikova, Daniel Auclair. Multiplexed flow cytometry based immunophenotyping paired with functional ex vivo (MFLEX) drug profiling informs potential efficacy of therapeutic agents in acute myeloid leukemia [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3904.

Integrating cfDNA liquid biopsy and organoid-based drug screening reveals PI3K signaling as a promising therapeutic target in colorectal cancer

BackgroundThe current precision medicine relies on biomarkers, which are mainly obtained through next-generation sequencing (NGS). However, this model failed to find effective drugs for most cancer patients. This study tried to combine liquid biopsy with functional drug tests using organoid models to find potential drugs for cancer patients.MethodsColorectal cancer (CRC) patients were prospectively enrolled and blood samples were collected from patients before the start of treatment. Targeted deep sequencing of cfDNA samples was performed using a 14-gene panel. Gastrointestinal (GI) cancer organoids were established and PI3K and mTOR inhibitors were evaluated on organoid models.ResultsA total of 195 mutations were detected across 58 cfDNA samples. The most frequently mutated genes were KRAS, TP53, PIK3CA, and BRAF, all of which exhibited higher mutation rates than tissue biopsy. Although 81% of variants had an allele frequency of less than 1%, certain mutations in KRAS, TP53, and SMAD4 had high allele frequencies exceeding 10%. Notably, among the seven patients with high allele frequency mutations, six had metastatic tumors, indicating that a high allele frequency of ctDNA could potentially serve as a biomarker of later-stage cancer. A high rate of PIK3CA mutation (31 out of 67, or 46.3%) was discovered in CRC patients, suggesting possible tumor progression mechanisms and targeted therapy opportunities. To evaluate the value of anti PI3K strategy in GI cancer, different lines of GI cancer organoids were established. The organoids recapitulated the morphologies of the original tumors. Organoids were generally insensitive to PI3K inhibitors. However, CRC-3 and GC-4 showed response to mTOR inhibitor Everolimus, and GC-3 was sensitive to PI3Kδ inhibitor Idelalisib. The CRC organoid with a PIK3CA mutation showed greater sensitivity to the PI3K inhibitor Alpelisib than wildtype organoids, suggesting potential treatment options for the corresponding patients.ConclusionLiquid biopsy holds significant promise for improving precision treatment and tumor prognosis in colorectal cancer patients. The combination of biomarker-based drug prediction with organoid-based functional drug sensitivity assay may lead to more effective cancer treatment.

Tumor Genomic Profiling and Ex Vivo Drug Sensitivity Testing for Pediatric Leukemia and Lymphoma Patients.

To describe the frequency of use of tumor genomic profiling and functional ex vivo drug sensitivity testing in pediatric patients with hematologic malignancies at our institution, and to determine how the results affected treatment selection. A retrospective chart review was conducted to analyze the frequency of tumor genomic profiling and functional drug sensitivity screening in our institution in pediatric patients with hematologic malignancies and to ask if the results were used to direct treatment. A case series of patients for whom these testing recommendations resulted in therapeutic interventions is reported. Thirty-three patients underwent tumor genomic profiling assays, functional ex vivo testing, or both. Nineteen patients (58%) had genomic profiling assays performed alone, 3 (9%) had functional ex vivo testing performed alone, and 11 (33%) had both tests performed. Twenty-one (64%) patients had potentially actionable mutations detected by the genomic profiling assay. Seven (21%) patients received at least 1 chemotherapeutic agent in accordance with the tumor genomic profiling or functional ex vivo drug sensitivity testing results. Three (43%) of the 7 patients who were treated with testing directed therapy had a favorable treatment response (PR or CR) to treatments selected based upon results of genomic or functional ex vivo testing. This retrospective case series demonstrates that precision medicine techniques such as genomic profiling and drug sensitivity testing can positively inform treatment selection in pediatric patients with relapsed or refractory leukemia and lymphoma.

Abstract CT209: A clinical trial of cResponse, a functional assay for cancer precision medicine

Abstract Precision cancer therapy has the potential to revolutionize treatment outcome. While genomic analysis has become central to cancer personalized medicine, recent studies have not shown that it drastically improves patient survival as compared to standard drug selection. Additionally, genomic mutations may suggest several treatment protocols without elucidating which approach will yield the best clinical response. Moreover, for many drugs no genetic predictive biomarkers are available. To advance cancer precision diagnostics, we have developed cResponse, a functional drug sensitivity platform to determine individualized patient treatment regimens. Fresh patient cancer samples are taken by biopsy or resection and sectioned into 300 uM slices which when cultured in the cResponse platform demonstrate similar architecture and tissue proliferation to those found in vivo. We show that cResponse is able to preserve human cancer tissue in 3D together with its microenvironment, including endothelial and immune cells, at a high viability (>90%) with continued cell division for 7 days in over twenty types of solids tumors. Importantly, the high viability of the tissue over an extended period of time allows for a rapid genomic profiling to help prioritize drugs to be tested by cResponse, as well as the capacity to evaluate the effects of slow acting drugs such as targeted therapy. To validate the capacity of the cResponse platform to predict patient response to cancer treatment, forty two patients with different cancer types receiving diverse chemotherapeutic and targeted agents were recruited to the study and their tumors were tested using cResponse prior to initiation of treatment. Cancer types included bladder, pancreatic, lung, colorectal, breast and sarcoma. After five days of treatment ex-vivo, the samples were fixed and designated a viability score (ranging from 0 to 100) based on an algorithm composed of a panel of histological and morphological markers. The cResponse score was compared to the clinical response to treatment of 27 patients for which this data is already available. The results demonstrated that cResponse could predict the patient's response with a specificity of 82% (9/11) and a sensitivity of 93.75% (15/16) when compared to the patient's imaging results. In the future, the integration of this platform in directing anti-cancer treatment may lead to better response rate of cancer patients to therapy. Citation Format: Seth Salpeter, Vered Bar, Sara Aharon, Luba Torovsky, Adi Zundelevich, Hamutal Shachar, Hagit Shapira, Nancy Gavert, Ravid Straussman, Shay Golan, Eli Rosenbaum, Talia Golan, Raanan Berger, Zohar Dotan, Dan Leibovici, Shani Breuer, Yakir Rotenberg, Aviad Zick, Ayala Hubert, Hovav Nechushtan, Guy Neev. A clinical trial of cResponse, a functional assay for cancer precision medicine [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr CT209.

Barriers to Successful Implementation of Precision Medicine to the Treatment of Refractory AML: Lessons Learned

Tremendous genomic heterogeneity has been reported for patients with acute myeloid leukemia (AML), as well as variability in gene expression. We have developed a custom high throughput drug sensitivity (HTS) assay enabling testing of ~150 drugs and combinations, both FDA approved and investigational, including chemotherapy agents and targeted inhibitors. We enrolled advanced relapsed/refractory AML patients (pts) on two successive trials, NCT01872819 “Treatment for Relapsed / Refractory AML Based on a High Throughput Drug Sensitivity Assay” and NCT NCT02551718 “Individualized Treatment for Relapsed/Refractory Acute Leukemia Based on Chemosensitivity and Genomics/Gene Expression Data.” The first trial enrolled 16 AML pts (3-5 prior regimens) of whom 14 received single agent study directed treatment, and the second trial (2-6 prior regimens) enrolled 25 AML pts of whom 14 received study directed therapy including drug combinations. We were able to choose chemotherapy drugs for all pts that had IC50's of ≤ 0.1 µM, or regimens that demonstrated synergy for drugs in combination. We have established feasibility as defined by the study, in that HTS assay results were reported at an average of 5.6 days (range 4-10), mutation analysis for 194 genes by MyAML™ reported in 12.5 days (range 9-17) with turnaround from receipt median 8 days (range 7-12) and treatment started at 7.8 (mean), 8 (median) days (range 4-11).In the first trial, 1 pt (5prior regimens), achieved complete remission with incomplete count recovery (CRi) with the single chosen drug, and 1 achieved CR, 1 CRi with subsequent regimens of additional drugs based on screening and published combinations. Median overall survival (OS) was 88 days for all patients, range 7 to 411. For the 2nd trial, where combinations were used, there were 2 CRs, 1 CRi and 3 partial responses. Median OS was 122 days, range 19-379. Of the patients in the 1st trial who had therapy that could have been directed by mutation analysis, there were 4 with FLT3, 1 KIT, 1 IDH2, 1 TP53, 1 FLT3+KIT, 2 FLT3+ IDH2 mutations (10/12 tested). For the 2nd trial there were 2 pts with FLT3D835, 2 TP53, 2 translocations involving FLT3, 1 IDH2 + FLT3-ITD, 1 FLT3-ITD + IDH1, 1 IDH1, 1 FLT3D839, 1 FLT3-ITD + TP53, 1 IDH2, 1 BCR-ABL (13/17 tested), and one patient received imatinib for BCR-ABL, 2 sorafenib for Flt3 mutations, and 1 with a TP53 mutation received decitabine. We had a number of logistical difficulties, including very advanced disease , poor performance status, inability to obtain the drugs with the lowest IC50's or as directed by mutations due to investigational status or lack of insurance approval, and inability to treat pts who returned home to their local physicians. We observed a higher response rate in pts who received drug combinations (8) vs. single agents (1). Although these were just feasibility trials, we also tried to compare their outcomes to a historical control group. We selected 2 matches for each of our pts, total of 50. They were matched to age, length of CR1, prior HCT, prior intense salvage and prior number of total salvages. One limitation was to find controls that had received ≥3 salvages like some of our studied population. Most of the patients in our clinical trial received single drug therapy (64 vs 24%) and non-intense therapy (4 vs 44%), when compared to our historical control group. Although these were very important treatment differences, the median OS did not differ statistically between both groups (88 vs 115 days, p: 0.327).We have gathered clinical data, and data on mutation analysis, gene expression, and drug sensitivity testing, that have been analyzed to determine correlations of specific features with drug resistance. Moreover, we have the ability to confirm functional drug sensitivity to the targeted inhibitors based on mutation data. All these data can contribute to future algorithms to assist in optimizing drug choices.In conclusion, our future iterations of this approach will include 1) only drug combinations, 2) study of pts with only 2 prior regimens, and 3) earlier incorporation of targeted inhibitors such as midostaurin and the new IDH2 inhibitor, enasidenib, into combination regimens. The current salvage options for relapsed/refractory AML have dismal results, as exhibited by our historical control group; only 10% were alive at 1 year. We will therefore pursue alternative approaches for this population, including the optimization of precision medicine methodology. DisclosuresWalter:ADC Therapeutics: Research Funding; Aptevo Therapeutics: Research Funding. Patay:Invivoscribe: Employment. Carson:Invivoscribe, Inc.: Employment. Radich:Amgen: Consultancy; Gilliad: Consultancy; Novartis: Consultancy, Other: lab contracts for bid and service assays; BMS: Consultancy; Ariad: Consultancy; Pfizer: Consultancy. Becker:GlycoMimetics, Inc.: Research Funding.

Development of a multiomics and functional drug sensitivity platform to investigate cell type specific drug effects in AML.

e19013 Background: Understanding the heterogeneity of AML is necessary for developing targeted drugs and diagnostics. A key measure of heterogeneity is the variance in response to treatments. Previously, we developed an ex vivo flow cytometry drug sensitivity assay (DSA) that predicted response to treatments in myelodysplastic syndrome. Unlike bulk cell viability measures of other drug sensitivity assays, our flow cytometry assay provides single cell resolution. The assay measures a drug’s effect on the viability or functional state of specific cell types. Here we present the development of this technology for AML, with additional measurements of DNA-Seq and RNA-Seq. Using the data from this assay, we aim to characterize the heterogeneity in AML drug sensitivity and the molecular mechanisms that drive it. Methods: As an initial feasibility analysis, we assayed 1 bone marrow and 3 peripheral blood AML patient samples. For the DSA, the samples were cultured with six AML standard of care (SOC) compounds across seven doses, in addition to two combinations. The cells were stained to detect multiple cell types including tumor blasts, and drug response was measured by flow cytometry. For the multi-omics, the cells were magnetically sorted to enrich for blasts and then assayed using a targeted 400 gene DNA-Seq panel and whole bulk transcriptome RNA-Seq. For comparison with BeatAML, Pearson correlations between gene expression and venetoclax sensitivity were investigated. Results: In our drug sensitivity assay, we measured dose response curves for the six SOC compounds, for each different cell type across each sample. The dose responses had cell type specific effects, including differences in drug response between CD11b+ blasts, CD11b- blasts, and other non-blast populations. Integrating with the DNA-Seq and RNA-Seq data, known associations between ex vivo drug response and gene expression were identified with additional cell type specificity. For example, BCL2A1 expression was negatively correlated with venetoclax sensitivity in CD11b- blasts but not in CD11b+ blasts. To further corroborate, among the top 1000 genes associated with venetoclax sensitivity in BeatAML, 93.7% had concordant directionality in effect. Conclusions: Here we describe the development of an integrated ex vivo drug sensitivity assay and multi-omics dataset. The data demonstrated that ex vivo responses to compounds differ between cell types, highlighting the importance of measuring drug response in specific cell types. In addition, we demonstrated that integrating these data will provide unique insights on molecular mechanisms that affect cell type specific drug response. As we continue to expand the number of patient samples evaluated with our multi-dimensional platform, this dataset will provide insights for novel drug target discovery, biomarker development, and, in the future, informing treatment decisions.

A Functional Profiling of Microenvironmental Factors and Small Molecules Reveals Monocyte Chemoattractant Protein-1 Mediates Drug Resistance in Acute Myeloid Leukemia

Background: The treatment of acute myeloid leukemia (AML) patients is a formidable challenge, due to disease heterogeneity and the ability to acquire secondary mutations in response to treatment. As a result, persistence and expansion of residual AML clones prevents complete remission. Emerging evidence suggests that extrinsic factors from the microenvironment protect AML cells from therapy, promote disease progression and drug resistance. Previous work from our lab highlights that higher levels of proinflammatory cytokines in the microenvironment, promote AML progression. This emphasizes the rationale to implement targeted therapies that block extrinsic signaling along with tumor intrinsic changes to overcome drug resistance in AML. Methods and Results: To identify factors and pathways that promote drug sensitivity and resistance in AML, we quantified the levels of 41 cytokines and growth factors from the plasma of 350 newly diagnosed AML patient samples using a multiplex luminex assay. The same cohort were analyzed for drug response to 122 small molecule inhibitors using an ex vivo functional drug sensitivity assay. These data were integrated with the gene expression and reactome pathway data to identify differentially regulated pathways and markers for drug response. Data integration revealed that protein synthesis pathways were significantly enriched in venetoclax, (Bcl-2 inhibitor) resistant AML samples, whereas, cytokine and immune pathways such as aberrant activation of monocyte chemoattractant protein-1 (MCP-1; p= 4.9e-07) were correlated with resistance to trametinib, a MEK inhibitor. In contrast, trametinib sensitivity was associated with activation of IL-1 and TLR signaling, suggesting specific extrinsic pathways may drive specific drug response. As a proof-of-concept, we dissected the role of MCP-1 activation in mediating trametinib resistance in AML. For this, we treated primary AML cells with MCP-1 in the presence of trametinib that led to a 4-fold increase in IC50 (0.4 vs 2.4 µM). To model drug response in vitro, we generated trametinib resistant AML cell lines MOLM13 and MV4;11 (FLT3-ITD mutated) and OCI-AML2 (DNMT3A mutated) by culturing cells in trametinib continuously over 4 months. All the trametinib resistant AML cell lines showed 2-10-fold increase in MCP-1 levels in whole cell lysate and in the conditioned media compared to the parental cells. Further, a long-term exposure of AML cells to MCP-1 in the presence of trametinib conferred a growth advantage with a 4-fold increase in cell numbers in comparison to the cells cultured with trametinib alone (n=4/condition). To delineate the molecular pathways driven by MCP-1 in trametinib resistance, we performed global phosphoproteomics using AML cell lines. We identified that MCP-1 activated various pro-survival pathways such as ERK, JNK, SRC, P70, and PKC kinases as well as cell cycle regulatory proteins as early as 5 mins in trametinib resistant cells. Targeting MCP-1 receptor-CCR2 by pharmacological inhibition, significantly reduced viability of trametinib resistant AML cells by 3-fold. The treatment with CCR2 inhibitor or trametinib individually reduced p-JNK levels by 3-fold in AML cells. A combined treatment with CCR2 inhibitor and trametinib, showed a 7-fold reduction in p-JNK levels. These data reinforced the role of MCP-1 in conferring trametinib resistance and that blockade of MCP-1 signaling re-sensitizes AML cells to trametinib. Conclusion: We show that MCP-1 augments trametinib resistance in AML by triggering novel signaling pathways. These effects can be reversed by blockade of MCP-1-CCR2 axis. Such combination treatment strategy incorporating extrinsic pathways together with the intrinsic ones would aid in re-sensitizing AML cells to therapy and represent an attractive new strategy. Our functional screen identifies that these extrinsic signals modulating drug response are unique to specific targeted therapy and should be considered when designing new therapies and clinical trials. We also created a large resource with 350 primary AML samples correlating microenvironment factors and gene expression with response to 122 small molecules, many of those which are FDA approved. This resource will help identify new drug response pathways in the context of microenvironment to design novel treatment regimens for preclinical and clinical testing and to improve AML outcomes. Disclosures Tyner: Syros: Research Funding; Agios: Research Funding; Incyte: Research Funding; Seattle Genetics: Research Funding; Takeda: Research Funding; Petra: Research Funding; Janssen: Research Funding; Constellation: Research Funding; Aptose: Research Funding; Genentech: Research Funding; Gilead: Research Funding; Array: Research Funding; AstraZeneca: Research Funding.

Abstract 5942: An integrated approach reveals novel extrinsic pathways and cytokine signatures associated with drug response in acute myeloid leukemia

Abstract Background: Acute myeloid leukemia (AML) is a heterogeneous disease with a dismal 5 year survival rate below 30%. Current AML therapies have provided little improvement in achieving complete remission. This is attributed to the development of drug resistance and extrinsic factors from the microenvironment that promote AML progression. Previously, our lab demonstrated that proinflammatory cytokines from the bone marrow microenvironment such as IL-1, promote AML progression. This emphasizes the rationale to design and implement targeted therapies to block extrinsic pathways coupled with intrinsic pathways conferring drug resistance in AML. Methods: To identify factors that promote drug sensitivity and resistance in AML, we used a cohort of 350 AML patients with various genetic subtypes. We quantified levels of 41 growth factors in plasma samples using a multiplex luminex assay. The same cohort of AML samples were analyzed for drug response to 130 small molecule inhibitors in in vitro functional drug sensitivity assay, for RNA-seq gene expression, and whole exome sequencing. The data was integrated to identify differential pathways and markers for drug response. To model drug response in vitro we created a series of drug resistant AML cell lines by culturing these cells in specific drugs long-term. Results: Pathway analysis integrating drug sensitivity, transcriptome, and cytokine expression data identified distinct differentially regulated pathways for various inhibitors. We found that protein synthesis pathways were significantly enriched in sorafenib (FLT-3 inhibitor) resistant samples whereas cytokine and immune pathways (such as TLR and MCP1 signaling) were correlated with trametinib (MEK inhibitor) and venetoclax (Bcl-2 inhibitor) resistance. Specifically, venetoclax resistance correlated with augmented levels of cytokines such as IL-1 and IP-10 and growth factors such as PDGF. Further, trametinib resistance in AML samples is correlated with the upregulation of CCL11, MCP1, and TGFα but a downregulation of CCL22 levels. Accordingly, using AML cell line models, we observed an increase in MCP1 and TGFα at the transcript and secreted protein levels in trametinib resistant cell lines. We demonstrated that MCP1 stimulation activates survival pathways by activating pERK and pJNK signaling, thus reducing apoptosis in AML cells. Targeting MCP1 in combination with trametinib reverses these effects, thus offering a novel therapeutic approach to overcome drug resistance. Conclusion and perspectives: Our data suggest that distinct extrinsic pathways may regulate the response to specific targeted therapy, thus providing a strong basis to design new treatment regimens. Combination treatment incorporating extrinsic pathways would aid in sensitizing AML cells to therapy. Our study offers an integrated approach and a resource to identify such pathways for a large number of tested drugs, and narrows down functionally important pathways associated with drug response. Citation Format: Rucha V. Modak, Alisa Damnernsawad, Ted Laderas, Guanming Wu, Jeffery W. Tyner, Karin D. Rodland, Shannon K. McWeeney, Anupriya Agarwal. An integrated approach reveals novel extrinsic pathways and cytokine signatures associated with drug response in acute myeloid leukemia [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5942.

Abstract IA18: Molecular signatures of circulating tumor cells

Abstract Circulating tumor cells (CTCs) are shed from primary cancers, as well as from metastatic deposits, into the blood circulation, where they may travel either as single cells or as CTC clusters, facing both mechanical and oxidative stress. Only a small subset of CTCs remain sufficiently viable to initiate metastatic growth upon dissemination to distant organs; their detection in patients with localized invasive cancer may present an opportunity for early detection, while their analysis in advanced cases provides a “whole cell biomarker,” enabling single-cell DNA, RNA, and protein studies, as well as functional drug-sensitivity testing using cultured CTCs. We have applied microfluidic technologies to deplete normal hematopoietic cells from blood specimens of patients with cancer, enhancing the ability of unmanipulated CTCs to proliferate ex vivo and recapitulating features of metastatic-competent cancer cells. Cultured CTCs from patients with either breast cancer or melanoma are highly tumorigenic when injected orthotopically at low inoculum into immunosuppressed mice. However, they display sensitivity to ROS and appear to upregulate compensatory mechanisms that are not commonly observed in primary tumors. In addition, they may undergo prolonged dormancy when lodged within distant organs. By combining single-cell RNA seq of patient-derived CTCs together with CRISPR and other genome-wide screens for modulators of metastasis phenotypes, we have sought to identify novel vulnerabilities in these cellular initiators of metastasis. While CTCs transit only very briefly through the blood circulation, these exceptional conditions may present new therapeutic opportunities to suppress the metastatic recurrence of cancer. Citation Format: Daniel Haber, Xin Hong, Richard Ebright, Doug Micalizi, Aditya Bardia, Ryan Sullivan, Mehmet Toner, Shyamala Maheswaran. Molecular signatures of circulating tumor cells [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr IA18.

A Multi-Omic Precision Medicine Clinical Trial in Acute Leukemia

A Multi-Omic Precision Medicine Clinical Trial in Acute Leukemia

Abstract 2153: Ex vivo drug sensitivity testing of primary cells for precision cancer medicine

Abstract Introduction/Purpose: Cancer therapy is increasingly moving towards individualized care and therapy, but there are still gaps between what is known and described on the molecular level about cancers and what is applied in the clinic. In an attempt to bridge the knowledge gap, we at the Institute for Molecular Medicine Finland (FIMM) have set up an Individualized Systems Medicine program that integrates clinical information, molecular profiling and functional information about individual patients’ cancers (Pemovska et al, Cancer Discov, 2013). Central to this program is the Drug Sensitivity and Resistance Testing (DSRT) where we functionally profile the responses of primary cancer cells to a comprehensive clinical oncology and signal transduction inhibitor drug collection of 528 compounds. Methods: Acoustic dispensing platforms are integral to the success of this profiling activity. We have to date produced approximately 3000 drug sets as dose response assay ready plates. The acoustic dispensing allows for making pre-drugged single drug plate sets and/or drug combination plates within hours after sampling of the cells. The plates are also readily sent to researchers anywhere in the world for running comparable assays at other sites. The drugging reproducibility is excellent generating results with correlations of 0.98 or higher in replicate assays. We have developed in-house software solutions to aid these processes: a script for quick creation of transfer list for combination plates and automated analysis pipelines with web-based software interfaces to enable the screening biologists to analyze the screening results effectively. Results: The results of these assays are used to explore and understand cancer biology in terms of druggability, functional heterogeneity and mechanism of drug response and resistance. The profiling data can be used to stratify and position the relevance of specific drugs in different diseases and has been used to identify novel clinically relevant activities of existing and investigational drugs (see e.g. Pemovska et al, Nature, 2015). This information is further utilized to establish hypotheses on drug combinations selectively targeting individual cancers and their predictive biomarkers, which can be explored in the clinic by our clinical collaborators to guide the treatment of the individual patient. Conclusions: In summary, we describe our platform for a functional drug sensitivity testing within our individualized cancer systems medicine program, which generates consistent biological and clinically relevant data. Citation Format: Sergey G. Kuznetsov, Alexander Ianevski, Evgeny Kulessky, Karoliina Laamanen, Elina Lehtinen, Maria Nurmi, Swapnil Potdar, Jani Saarela, Katja Suomi, Laura Turunen, Krister Wennerberg, Päivi Tammela. Ex vivo drug sensitivity testing of primary cells for precision cancer medicine [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2153.

Abstract 1066: Drug resistant PDX models for testing targeted drug sensitivity

Abstract Advanced lung cancer of late stage patients is selectively sensitive to available targeted therapeutics and often develops resistance to early sensitive drug treatment, thus, the patient-derived xenograft (PDX) models are critical to evaluate drug sensitivity and resistance. We established PDX models from routine surgery tissues, as well as from tumor cells in pleural effusion specimens and from biopsies of advanced cancer patients who are refractory to treatments. Genetic profiling of the PDX lung models revealed a group of them with TKI drug-associated EGFR mutations (L858R, T790M, Exon19del, C797S). Among them, are an AZD9291-resistant sub-group with EGFR double mutations (Exon19Δ, T790M) in the clinic before AZD9291 treatment, and yet, acquired resistance to AZD9291 either through emergence C797S mutation or T790M loss in EGFR or through mutations in other pathways, providing insights into potential combo therapies. Indeed, combinations of AZD9291 with Cetuximab as well as with inhibitors targeting other signaling pathways such as MEK, are efficacious against an AZD9291-resistant lung PDX model with EGFR triple mutations (Exon19Δ, T790M, C797S). We also establish valuable lung models resistant to ALK inhibitors, Avastin, PD-1 antibody drugs, and resistant hematopoietic models. In summary, functional drug sensitivity profiling and genetic profiling with drug resistant PDX models provide insights into resistance mechanisms as well as precise therapeutic options against resistance emergence. Citation Format: Jijun Cheng, Feifei Zhang, Zhen Zhou, Yuan Long, Wenhua Xu, Shizhu Zhao, Hongkui Chen, Shun Lu, Danyi Wen. Drug resistant PDX models for testing targeted drug sensitivity [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 1066.

Comparative Analysis of Independent Ex Vivo functional Drug Screens Identifies Predictive Biomarkers of BCL-2 Inhibitor Response in AML

Introduction:Therapeutic options for patients with AML were recently expanded with FDA approval of four drugs in 2017. As their efficacy is limited in some patient subpopulations and relapse ultimately ensues, there remains an urgent need for additional treatment options tailored to well-defined patient subpopulations to achieve durable responses. Two comprehensive profiling efforts were launched to address this need-the multi-center Beat AML initiative, led by the Oregon Health & Science University (OHSU) and the AML Individualized Systems Medicine program at the Institute for Molecular Medicine Finland (FIMM).Methods:We performed a comparative analysis of the two large-scale data sets in which patient samples were subjected to whole-exome sequencing, RNA-seq, and ex vivo functional drug sensitivity screens: OHSU (121 patients and 160 drugs) and FIMM (39 patients and 480 drugs). We predicted ex vivo drug response [quantified as area under the dose-response curve (AUC)] using gene expression signatures selected with standard regression and a novel Bayesian model designed to analyze multiple data sets simultaneously. We restricted analysis to the 95 drugs in common between the two data sets.Results:The ex vivo responses (AUCs) of most drugs were positively correlated (OHSU: median Pearson correlation r across all pairwise drug comparisons=0.27; FIMM: median r=0.33). Consistently, a samples's ex vivo response to an individual drug was often correlated with the patient's Average ex vivo Drug Sensitivity (ADS), i.e., the average response across the 95 drugs (OHSU: median r across 95 drugs=0.41; FIMM: median r=0.58). Patients with a complete response to standard induction therapy had a higher ADS than those that were refractory (p=0.01). Further, patients whose ADS was in the top quartile had improved overall survival relative to those having an ADS in the bottom quartile (p<0.05). Standard regression models (LASSO and Ridge) trained on ADS and gene expression in the OHSU data set had improved ex vivo response prediction performance as assessed in the independent FIMM validation data set relative to those trained on gene expression alone (LASSO: p=2.9x10-4; Ridge: p=4.4x10-3). Overall, ex vivo drug response was relatively well predicted (LASSO: mean r across 95 drugs=0.62; Ridge: mean r=0.62).The BCL-2 inhibitor venetoclax was the only drug whose response was negatively correlated with ADS in both data sets. We hypothesized that, whereas the predictive performance of many other drugs was likely dependent on ADS, the predictive performance of venetoclax (LASSO: r=0.53, p=0.01; Ridge: r=0.63, p=1.3x10-3) reflected specific gene expression biomarkers. To identify biomarkers associated with venetoclax sensitivity, we developed an integrative Bayesian machine learning method that jointly modeled both data sets, revealing several candidate biomarkers positively (BCL2 and FLT3) or negatively (CD14, MAFB, and LRP1) correlated with venetoclax response. We assessed these biomarkers in an independent data set that profiled ex vivo response to the BCL-2/BCL-XL inhibitor navitoclax in 29 AML patients (Lee et al.). All five biomarkers were validated in the Lee data set (Fig 1).Conclusions:The two independent ex vivo functional screens were highly concordant, demonstrating the reproducibility of the assays and the opportunity for their use in the clinic. Joint analysis of the two data sets robustly identified biomarkers of drug response for BCL-2 inhibitors. Two of these biomarkers, BCL2 and the previously-reported CD14, serve as positive controls credentialing our approach. CD14, MAFB, and LRP1 are involved in monocyte differentiation. The inverse correlation of their expression with venetoclax and navitoclax response is consistent with prior reports showing that monocytic cells are resistant to BCL-2 inhibition (Kuusanmäki et al.). These biomarker panels may enable better selection of patient populations likely to respond to BCL-2 inhibition than would any one biomarker in isolation.

Cegal Protocol : Evaluation of the Feasibility of a Chemogenomic Approach to Identify Personalized Therapy for Relapse or Refractory AML Patients

Introduction :Significant improvements of leukemia free survival have been obtained over the last 30 years in AML. This progress has resulted from optimization of the same chemotherapy-based “one-size-fits-all” approach. However, even though the majority of patients eligible for intensive chemotherapy achieve complete remission (60% to 80%), the majority still relapse and ultimately die (5-year overall survival being around 20%). Recent genomic studies have revealed the heterogeneity of AML at the molecular level, stressing the need for personalized therapeutic approaches. The fundamental idea in genomic medicine is that somatic genetic alterations can be identified and matched with drugs targeting those abnormalities for a patient's benefit. However, there are not that many potent, tolerated targeted agents in the clinic that can induce significant clinical responses. Interestingly, the majority of the most effective drugs do not function by exploiting genetic mutations in individual tumors. Thus, additional functional approaches are needed to identify new drugs and assign existing drugs to larger numbers of patients with AML.From this perspective, the“chemogenomic” approach realizes an integratedanalysis of the molecular clonal abnormalities for the leukemic patients, and compares it to a functional drug sensitivity testing (DST) which can test simultaneously dozens of drugsex vivo. This approach aims to identify correlations between genomic changes and sensibilities increased in certain categories of drugs. The CEGAL project is a prospectivemonocentric study aiming to evaluate the feasibility of this personalized approach for relapse or refractory AML patientsby determining the proportion of patients whosechemogenomic results are available in less than 21 days.Methods :Patients with relapsed and/or refractory AML were eligible if no alternative therapy was available. Written consent was obtained from all patients (Study CeGAL-IPC 2014-012, EUDRACT 2014-A01209-38). Ex vivo sensitivity profiling was performed on freshly isolated AML blasts derived from patient bone marrow aspirates and peripheral blood with a panel of 78 drugs representative of the main therapeutic classeskwown to be active in AML. After 48h of incubation, cellular viability was measured with the CellTiter-Glo© Luminescent cell viability assay (Promega) kit to obtain one EC50. Next generation sequencing (NGS) was performed on a 200 genes panel with Mi-Seq (Illumina) and genomic profils by cGH-array (Agilent4x180K) were established. Personalized treatment plans were then reviewed by an institutional personalized committee, which issued final treatment recommendations.Results: A total of 48 AML patients were included (refractory=8,relapse=40) with a median age of 64 years old (range 24-81) and a median number of prior therapies of 2. More than half of the patients presented an adverse prognostic according to the 2017 ELN risk stratification (n=29). The most frequent genomic alterations were TET2 in 32.5% of the patients, ASXL1 (27.5%), DNMT3A (25%), TP53 (22.5%), SRSF2 (20%) and RUNX1 (17.5%). The median number of mutations per patient was 5,6. The main objective was completed with chemogenomic results in less than 21 days for 62.5% of the patients (n=25) with a median time of 22.4 days. Twenty four patients (67,5%, n=27) had targetable mutations and for 36 patients (90%), there was at least one drug predicted efficient. In light of the chemogenomic analysis, we found a correlation between genomic alteration and drug sensitivy for 28 patients. Thirteen patients (32.5%) received a treatment guided by the analysis whatsoever chemotherapies or targeted therapies. The 27 others didn't receive a personalized treatment either because the molecule wasn't commercially available or because the physician decided so.Conclusion : In our experience, the chemogenomic is an innovative and promising approach, allowing to assign patient-specific treatment options according to their mutational profile and to in vitro drugs' response in a median time of 3 weeks. DisclosuresNo relevant conflicts of interest to declare.

In-Vitro Drug Sensitivity Screening in Chronic Lymphocytic Leukemia (CLL) Primary Patient Samples Identifies Drug Candidates for Precision Cancer Therapy

Abstract Introduction Chronic Lymphocytic Leukemia (CLL) is the most common leukemia in adults and is currently considered incurable. Although current treatment regimens prolong life for patients, CLL eventually relapses. Efficient therapies may require a personalized approach that combines targeting cancer cells and the tumor microenvironment by restoring the patient's own anti-tumor immunity. However, a major limitation is that no efficient approach exists to identify the most effective drugs for each patient and cancer stage. With the aim to support future introduction of individualized treatment for patients, we assessed the sensitivity of CLL patient samples to several drug candidates using our in vitro functional drug sensitivity screening platform. Methods We have established novel in vitro culture settings that mimic the CLL tumor microenvironment and allow proliferation of CLL cells for 5 days. Using our unique method, we performed drug screening on 26 patient samples and 10 healthy donors against a customized, annotated library of 516 drugs including kinase inhibitors, proteasome inhibitors, B-cell pathway inhibitors and several other approved drug classes. Primary patient samples were cultured in 384 well-plates with the presence of individual drugs over a concentration range over 5 logs. Drug sensitivity was assessed using CellTiter-Glo® luminescent cell viability assay and CellTox™ green cytotoxicity assay on day 5. Drug Sensitivity Score (DSS) was then calculated for each drug using the IC50 value, slope and area under the curve (AUC). DSSs for CLL patient samples were next compared with DSS of healthy donors for the full patient sample cohort screened so far to generate a selective DSS (sDSS = DSSpatient - DSShealthy) for each patient. Drugs which have sDSS &gt;5 were considered clearly more effective for patient samples in the in vitro test system. CLL samples were assessed for sDSS using our screening data and we ranked all the drugs by their score. Results In order to find drug candidates for targeted therapies in CLL patients, we performed in vitro drug sensitivity screening on 13 IgVH unmutated and 13 IgVH mutated CLL patient samples, as well as 10 healthy donors (due to the lower number of cells healthy donors were pooled into two samples of 5 donors each). Our in vitro assay showed that proteasome inhibitors, kinase inhibitors and several approved CLL drug candidates were considered sensitive in the majority of patient samples. This included venetoclax, the Bcl-2 inhibitor ABT-737, doxorubicin, acalabrutinib, other kinase inhibitors (sunitinib, volasertib, trametinib, copanlisib) and proteasome inhibitors (carfilzomib, bortezomib). Selective drug sensitivity scores of the top 5 drugs in all patient samples (73 drugs in total) are shown in the heatmap (Figure 1). Venetoclax showed a higher sDSS score in 10 of the 20 patients with an average sDSS score of 22.3 followed by ABT-737 (Bcl-2 inhibitor) with an average sDSS score of 19.7. By performing hierarchical clustering analysis (Euclidean distance, Ward linkage method), we observed unsupervised clustering of patient samples irrespective of the IgVH mutation status. We are currently expanding the analysis by classifying the patient samples by age, sex and mutation status. Conclusion Our novel CLL culture method that allows cell proliferation along with our established functional in-vitro drug sensitivity screening platform enabled us to screen a number of patient samples and evaluate the sensitivity of a library of approved drugs and investigational drug candidates for CLL. Our analysis shows that several drugs may be effective for CLL and can be tested in drug combinations in order to identify synergistic effects. As a future perspective, we want to combine machine learning strategies with the experimental drug screening strategies to identify drug combinations and validate drug candidates by xenografting and in precision medicine clinical trials. Figure 1: Selective Drug Sensitivity Screening (sDSS) score for top 3 drugs (44 drugs) for 20 CLL patient samples. Green label is IgVH mutated CLL patient samples and blue label is IgVH unmutated patient samples. Disclosures Schjesvold: Oncopeptides: Consultancy; Abbvie: Honoraria; Novartis: Honoraria; Janssen: Consultancy, Honoraria, Research Funding; Adaptive: Consultancy; Bayer: Consultancy; Bristol Myers Squibb: Consultancy; Takeda: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Amgen: Consultancy, Honoraria, Research Funding.

Identification of precision treatment strategies for relapsed/refractory multiple myeloma by functional drug sensitivity testing.

Novel agents have increased survival of multiple myeloma (MM) patients, however high-risk and relapsed/refractory patients remain challenging to treat and their outcome is poor. To identify novel therapies and aid treatment selection for MM, we assessed the ex vivo sensitivity of 50 MM patient samples to 308 approved and investigational drugs. With the results we i) classified patients based on their ex vivo drug response profile; ii) identified and matched potential drug candidates to recurrent cytogenetic alterations; and iii) correlated ex vivo drug sensitivity to patient outcome. Based on their drug sensitivity profiles, MM patients were stratified into four distinct subgroups with varied survival outcomes. Patients with progressive disease and poor survival clustered in a drug response group exhibiting high sensitivity to signal transduction inhibitors. Del(17p) positive samples were resistant to most drugs tested with the exception of histone deacetylase and BCL2 inhibitors. Samples positive for t(4;14) were highly sensitive to immunomodulatory drugs, proteasome inhibitors and several targeted drugs. Three patients treated based on the ex vivo results showed good response to the selected treatments. Our results demonstrate that ex vivo drug testing may potentially be applied to optimize treatment selection and achieve therapeutic benefit for relapsed/refractory MM.

Novel Targets Identified By in Vitro Screening of Patient Samples with Adult Acute Lymphoblastic Leukemia

Introduction: Acute lymphoblastic leukemia (ALL) is considered a success story in pediatrics, with cure rates greater than 85%. In contrast, the estimated cure rate for adolescent and young adult (AYA) patients is estimated at 60% and is just 20-40% for adults. This is thought to be secondary to chemotherapy associated toxicities for older patients as well as different disease biology. In the AYA population, more patients carry the chemo-resistant Ph-like phenotype, with an estimated incidence of up to 27%. However, the genetic profile of adult ALL has not been well described. In addition, although the ABL specific kinase inhibitors and JAK inhibitors are predicted to be active in Ph-like ALL, there are no reports on functional drug sensitivity data in AYA or adult ALL. Our group has been collecting functional data on both newly diagnosed and relapsed AYA and adult patients since 2010; in this study we summarize the functional data obtained during this time. Materials and Methods: All clinical samples were obtained with informed consent and approval of the institutional review board at OHSU. Data obtained on both newly diagnosed and relapsed patients obtained from Jan 2010 to present was reviewed. At the time of diagnosis or at relapse, mononuclear cells were isolated from bone marrow aspirates of patients and exposed to a library of 130 small molecule inhibitors at varying concentrations. After 72 hours, cell viability was assessed via an MTS based assay and drug efficacy was assessed based on the IC50 of of each panel drug against each patient sample. The calculated IC50 for each drug was compared to the median IC50 of over 1000 patient samples and was deemed hypersensitive if the IC50 value was below the 20th percentile. A pathway was considered to be actively targeted if patients had more than one inhibitor targeting that pathway identified as hypersensitive, or a single drug was considered valid if it was hypersensitive in replicate testing.Results: We identified a total of 42 adults with either B-cell or T-cell ALL who had analyzable data; those with Ph+ ALL were excluded. Of these, 28 had B cell ALL including the MLL rearrangement and 14 had T cell ALL. For the B cell ALL group, 17 had data obtained at diagnosis and 11 at relapse. 75% of patients had at least one active inhibitor identified, with the range being from 2-25 inhibitors per patient targeting from 1-5 different pathways per patient. We found that 35% of patient samples were sensitive to a JAK inhibitor, an ABL inhibitor or both. In addition, we found overlapping sensitivities with 43% of patient samples sensitive to PI3K/AKT/mTOR inhibitors, 43% of samples were sensitive to IGF1R inhibitors, 49% were sensitive to PDGFR/VEGFR inhibitors, 18% were sensitive to p38MAPK inhibitors and 11% each were sensitive to CSF1R or FLT3 inhibitors. Of note, all of the FLT3 inhibitor sensitivities were within the MLL-rearranged subgroup. Among the T-ALL group, 10 had data obtained at diagnosis and 4 at relapse. 52% of the samples had at least one inhibitor identified, with the range being from 1-20 inhibitors targeting from 1-4 pathways per patient. Drugs targeting the following pathways were identified: JAK family (57.1%), ERBBB family (57.1%), SYK (28.5%), and HDAC, MDM2 and aurora kinase (14.3% each).Discussion: This study provides the evidence that there are multiple novel potentially targetable pathways present in AYA and adult patients with ALL. Given the poor outcomes in this group of patients, further exploration of these potentially targetable pathways is warranted. DisclosuresTyner:Agios Pharmaceuticals: Research Funding; Array Biopharma: Research Funding; Aptose Biosciences: Research Funding; AstraZeneca: Research Funding; Constellation Pharmaceuticals: Research Funding; Genentech: Research Funding; Inctye: Research Funding; Janssen Research & Development: Research Funding; Seattle Genetics: Research Funding; Takeda Pharmaceuticals: Research Funding; Leap Oncology: Consultancy. Druker:Curis: Patents & Royalties; Pfizer: Patents & Royalties; Array: Patents & Royalties; Dana-Farber Cancer Institute: Patents & Royalties: Millipore royalties via Dana-Farber Cancer Institute; Oncotide Pharmaceuticals: Research Funding; Novartis: Research Funding; BMS: Research Funding; ARIAD: Patents & Royalties: inventor royalties paid by Oregon Health & Science University for licenses, Research Funding; Roche: Consultancy; Gilead Sciences: Consultancy, Other: travel, accommodations, expenses; D3 Oncology Solutions: Consultancy; AstraZeneca: Consultancy; Ambit BioSciences: Consultancy; Agios: Honoraria; MolecularMD: Consultancy, Equity Ownership, Patents & Royalties; Lorus: Consultancy, Equity Ownership; Cylene: Consultancy, Equity Ownership; CTI: Consultancy, Equity Ownership; Pfizer: Patents & Royalties; Curis: Patents & Royalties; Array: Patents & Royalties; Dana-Farber Cancer Institute: Patents & Royalties: Millipore royalties via Dana-Farber Cancer Institute.

Abstract 5588: Functional drug sensitivity and resistance profiling of AML patient cells defines a disease-specific combination of druggable signal addictions.

Abstract Numerous underlying molecular events have been described in acute myeloid leukemia (AML), but still, the fundamental disease mechanisms are poorly understood. Several targeted therapies have been investigated for improved AML therapy but have not succeeded to date, likely due to the inability to identify the patient subgroups that are most likely to benefit. Here, we describe functional profiling of AML patient cells ex vivo with a drug sensitivity and resistance (DSRT) platform in order to distinguish disease- and patient-specific molecular vulnerabilities and individualized therapeutic strategies. The oncology drug collection covers 306 anti-cancer agents including 131 approved, 107 investigational and 68 experimental compounds. Each compound is tested in a dose response series allowing for calculation of drug sensitivity scores. The functional exploration of AML patient samples was accompanied with comprehensive molecular profiling and clinical background data to link drug sensitivities to molecular aberrations and predictive biomarkers. Comparison of the drug sensitivity profiles of 24 AML patient and 5 control samples revealed that targeted inhibitors often exhibit no to little effect in the controls and the majority of AML patient samples with only a subset of patients showing very selective responses, indicating that cancer specific vulnerabilities can be detected with the DSRT platform. Clustering of the drug responses among patient and control samples identified distinct subgroups of patients and drugs. Each of these groups could be defined by a separate drug class, implying that the linked samples were addicted to the corresponding signaling networks. Specifically, we saw select addictions to drug classes such as dasatinib/VEGFR TKIs, MEK inhibitors, rapalogs, JAK inhibitors, mTOR inhibitors in 6 (25 %), 5 (21%), 5 (21%), 5 (21%), and 6 (25%) of the patients, respectively. Together, these sensitivities highlight a set of disease-specific and individualized signaling nodes that can serve as starting points for personalized combination therapy strategies in AML. Further, integration of DSRT results with molecular and clinical profiling link some of the selective drug responses to clinically actionable markers. For example, dasatinib sensitivity was found to be predominant in AML M5 patients, suggesting that this drug is a promising therapeutic candidate for relapsed and refractory AML M5 patients. In conclusion, the DSRT platform can be used as a functional interrogation of patient cells leading to the identification of druggable vulnerabilities that can be used for personalized medicine strategies. Furthermore, insight into disease-specific signal addictions and establishment of how these are linked can aid in deconvoluting the complex molecular disease mechanisms of cancers and generate strategies for novel drug development. Citation Format: Tea Pemovska, Bhagwan Yadav, Riikka Karjalainen, Evgeny Kulesskiy, Mika Kontro, Muntasir Mamun Majumder, Laura Turunen, Ida Lindenschmidt, Anna Lehto, Jonathan Knowles, Caroline Heckman, Kimmo Porkka, Tero Aittokallio, Olli Kallioniemi, Krister Wennerberg. Functional drug sensitivity and resistance profiling of AML patient cells defines a disease-specific combination of druggable signal addictions. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 5588. doi:10.1158/1538-7445.AM2013-5588

461 Identification of Personalized Therapeutic Strategies and Associated Biomarkers in Adult Acute Myeloid Leukemia Using a Functional Drug Sensitivity and Resistance Testing Platform

461 Identification of Personalized Therapeutic Strategies and Associated Biomarkers in Adult Acute Myeloid Leukemia Using a Functional Drug Sensitivity and Resistance Testing Platform