A first-in-class selective inhibitor of EGFR and PI3K offers a single-molecule approach to targeting adaptive resistance.

Despite tremendous progress in precision oncology, adaptive resistance mechanisms limit the long-term effectiveness of molecularly targeted agents. Here we evaluated the pharmacological profile of MTX-531 that was computationally designed to selectively target two key resistance drivers, epidermal growth factor receptor and phosphatidylinositol 3-OH kinase (PI3K). MTX-531 exhibits low-nanomolar potency against both targets with a high degree of specificity predicted by cocrystal structural analyses. MTX-531 monotherapy uniformly resulted in tumor regressions of squamous head and neck patient-derived xenograft (PDX) models. The combination of MTX-531 with mitogen-activated protein kinase kinase or KRAS-G12C inhibitors led to durable regressions of BRAF-mutant or KRAS-mutant colorectal cancer PDX models, resulting in striking increases in median survival. MTX-531 is exceptionally well tolerated in mice and uniquely does not lead to the hyperglycemia commonly seen with PI3K inhibitors. Here, we show that MTX-531 acts as a weak agonist of peroxisome proliferator-activated receptor-γ, an attribute that likely mitigates hyperglycemia induced by PI3K inhibition. This unique feature of MTX-531 confers a favorable therapeutic index not typically seen with PI3K inhibitors.

Targeting DNA Repair and Survival Signaling in Diffuse Intrinsic Pontine Gliomas to Prevent Tumor Recurrence.

Therapeutic resistance remains a major obstacle to successful clinical management of diffuse intrinsic pontine glioma (DIPG), a high-grade pediatric tumor of the brain stem. In nearly all patients, available therapies fail to prevent progression. Innovative combinatorial therapies that penetrate the blood-brain barrier and lead to long-term control of tumor growth are desperately needed. We identified mechanisms of resistance to radiotherapy, the standard of care for DIPG. On the basis of these findings, we rationally designed a brain-penetrant small molecule, MTX-241F, that is a highly selective inhibitor of EGFR and PI3 kinase family members, including the DNA repair protein DNA-PK. Preliminary studies demonstrated that micromolar levels of this inhibitor can be achieved in murine brain tissue and that MTX-241F exhibits promising single-agent efficacy and radiosensitizing activity in patient-derived DIPG neurospheres. Its physiochemical properties include high exposure in the brain, indicating excellent brain penetrance. Because radiotherapy results in double-strand breaks that are repaired by homologous recombination (HR) and non-homologous DNA end joining (NHEJ), we have tested the combination of MTX-241F with an inhibitor of Ataxia Telangiectasia Mutated to achieve blockade of HR and NHEJ, respectively, with or without radiotherapy. When HR blockers were combined with MTX-241F and radiotherapy, synthetic lethality was observed, providing impetus to explore this combination in clinically relevant models of DIPG. Our data provide proof-of-concept evidence to support advanced development of MTX-241F for the treatment of DIPG. Future studies will be designed to inform rapid clinical translation to ultimately impact patients diagnosed with this devastating disease.

Loss of NF1 in Melanoma Confers Sensitivity to SYK Kinase Inhibition.

Neurofibromin 1 (NF1) loss of function (LoF) mutations are frequent in melanoma and drive hyperactivated RAS and tumor growth. NF1LoF melanoma cells, however, do not show consistent sensitivity to individual MEK, ERK, or PI3K/mTOR inhibitors. To identify more effective therapeutic strategies for treating NF1LoF melanoma, we performed a targeted kinase inhibitor screen. A tool compound named MTX-216 was highly effective in blocking NF1LoF melanoma growth in vitro and in vivo. Single-cell analysis indicated that drug-induced cytotoxicity was linked to effective cosuppression of proliferation marker Ki-67 and ribosomal protein S6 phosphorylation. The antitumor efficacy of MTX-216 was dependent on its ability to inhibit not only PI3K, its nominal target, but also SYK. MTX-216 suppressed expression of a group of genes that regulate mitochondrial electron transport chain and are associated with poor survival in patients with NF1LoF melanoma. Furthermore, combinations of inhibitors targeting either MEK or PI3K/mTOR with an independent SYK kinase inhibitor or SYK knockdown reduced the growth of NF1LoF melanoma cells. These studies provide a path to exploit SYK dependency to selectively target NF1LoF melanoma cells. SIGNIFICANCE: A kinase inhibitor screen identifies SYK as a targetable vulnerability in melanoma cells with NF1 loss of function.

Targeting ADAM17 inhibits human colorectal adenocarcinoma progression and tumor-initiating cell frequency.

ADAM17 (a disintegrin and metalloproteinase 17)/TACE (TNFα converting enzyme) has emerged as a potential therapeutic target in colorectal cancer (CRC) and other cancers, due in part to its role in regulating various tumor cell surface proteins and growth factors and cytokines in the tumor microenvironment. The emergence of MEDI3622, a highly potent and specific antibody-based ADAM17 inhibitor, has allowed testing of the concept that targeting ADAM17 may be an important new therapeutic approach for CRC patients. We demonstrate that MEDI3622 is highly efficacious on tumor growth in multiple human CRC PDX models, resulting in improved survival of animals bearing tumor xenografts. MEDI3622 was further found to impact Notch pathway activity and tumor-initiating cells. The promising preclinical activity seen here supports further clinical investigation of this treatment approach to improve therapeutic outcome for patients diagnosed with metastatic CRC, including patients with KRAS-mutant tumors for whom other therapeutic options are currently limited.

Mechanistic evaluation and transcriptional signature of a glutathione S-transferase omega 1 inhibitor.

Glutathione S-transferase omega 1 (GSTO1) is an atypical GST isoform that is overexpressed in several cancers and has been implicated in drug resistance. Currently, no small-molecule drug targeting GSTO1 is under clinical development. Here we show that silencing of GSTO1 with siRNA significantly impairs cancer cell viability, validating GSTO1 as a potential new target in oncology. We report on the development and characterization of a series of chloroacetamide-containing potent GSTO1 inhibitors. Co-crystal structures of GSTO1 with our inhibitors demonstrate covalent binding to the active site cysteine. These potent GSTO1 inhibitors suppress cancer cell growth, enhance the cytotoxic effects of cisplatin and inhibit tumour growth in colon cancer models as single agent. Bru-seq-based transcription profiling unravelled novel roles for GSTO1 in cholesterol metabolism, oxidative and endoplasmic stress responses, cytoskeleton and cell migration. Our findings demonstrate the therapeutic utility of GSTO1 inhibitors as anticancer agents and identify the novel cellular pathways under GSTO1 regulation in colorectal cancer.

Sensitivity of KRAS-Mutant Colorectal Cancers to Combination Therapy That Cotargets MEK and CDK4/6.

PURPOSE: The emerging need for rational combination treatment approaches led us to test the concept that cotargeting MEK and CDK4/6 would prove efficacious in KRAS-mutant (KRAS(mt)) colorectal cancers, where upregulated CDK4 and hyperphosphorylated retinoblastoma (RB) typify the vast majority of tumors. EXPERIMENTAL DESIGN: Initial testing was carried out in the HCT-116 tumor model, which is known to harbor a KRAS mutation. Efficacy studies were then performed with five RB(+) patient-derived colorectal xenograft models, genomically diverse with respect to KRAS, BRAF, and PIK3CA mutational status. Tolerance, efficacy, and pharmacodynamic evaluation of target modulation were evaluated in response to daily dosing with either agent alone or concurrent coadministration. RESULTS: Synergy was observed in vitro when HCT-116 cells were treated over a broad range of doses of trametinib and palbociclib. Subsequent in vivo evaluation of this model showed a higher degree of antitumor activity resulting from the combination compared to that achievable with single-agent treatment. Testing of colorectal patient-derived xenograft (PDX) models further showed that combination of trametinib and palbociclib was well tolerated and resulted in objective responses in all KRAS(mt) models tested. Stasis was observed in a KRAS/BRAF wild-type and a BRAF(mt) model. CONCLUSIONS: Combination of trametinib and palbociclib was well tolerated and highly efficacious in all three KRAS-mutant colorectal cancer PDX models tested. Promising preclinical activity seen here supports clinical evaluation of this treatment approach to improve therapeutic outcome for patients with metastatic colorectal cancer.

Cancer stem cell marker phenotypes are reversible and functionally homogeneous in a preclinical model of pancreatic cancer.

Survival rates associated with pancreatic cancer remain dismal despite advancements in detection and experimental treatment strategies. Genetically engineered mouse models of pancreatic tumorigenesis have gained considerable attention based on their ability to recapitulate key clinical features of human disease including chemotherapeutic resistance and fibrosis. However, it is unclear if transgenic systems exemplified by the Kras(G12D)/Trp53(R172H)/Pdx-1-Cre (KPC) mouse model recapitulate the functional heterogeneity of human pancreatic tumors harboring distinct cells with tumorigenic properties. To facilitate tracking of heterogeneous tumor cell populations, we incorporated a luciferase-based tag into the genetic background of the KPC mouse model. We isolated pancreatic cancer cells from multiple independent tumor lines and found that roughly 1 out of 87 cells exhibited tumorigenic capability. Notably, this frequency is significantly higher than reported for human pancreatic adenocarcinomas. Cancer stem cell (CSC) markers, including CD133, CD24, Sca-1, and functional Aldefluor activity, were unable to discriminate tumorigenic from nontumorigenic cells in syngeneic transplants. Furthermore, three-dimensional spheroid cultures originating from KPC tumors did not enrich for cells with stem-like characteristics and were not significantly more tumorigenic than cells cultured as monolayers. Additionally, we did not observe significant differences in response to gemcitabine or salinomycin in several isolated subpopulations. Taken together, these studies show that the hierarchical organization of CSCs in human disease is not recapitulated in a commonly used mouse model of pancreatic cancer and therefore provide a new view of the phenotypic and functional heterogeneity of tumor cells.