Distinct Biomarker Profiles and TCR Sequence Diversity Characterize the Response to PD-L1 Blockade in a Mouse Melanoma Model

Only a subset of patients responds to immune checkpoint blockade (ICB) in melanoma. A preclinical model recapitulating the clinical activity of ICB would provide a valuable platform for mechanistic studies. We used melanoma tumors arising from an Hgf^tg;Cdk4^R24C/R24C genetically engineered mouse (GEM) model to evaluate the efficacy of an anti-mouse PD-L1 antibody similar to the anti-human PD-L1 antibodies durvalumab and atezolizumab. Consistent with clinical observations for ICB in melanoma, anti-PD-L1 treatment elicited complete and durable response in a subset of melanoma-bearing mice. We also observed tumor growth delay or regression followed by recurrence. For early treatment assessment, we analyzed gene expression profiles, T-cell infiltration, and T-cell receptor (TCR) signatures in regressing tumors compared with tumors exhibiting no response to anti-PD-L1 treatment. We found that CD8⁺ T-cell tumor infiltration corresponded to response to treatment, and that anti-PD-L1 gene signature response indicated an increase in antigen processing and presentation, cytokine-cytokine receptor interaction, and natural killer cell-mediated cytotoxicity. TCR sequence data suggest that an anti-PD-L1-mediated melanoma regression response requires not only an expansion of the TCR repertoire that is unique to individual mice, but also tumor access to the appropriate TCRs. Thus, this melanoma model recapitulated the variable response to ICB observed in patients and exhibited biomarkers that differentiate between early response and resistance to treatment, providing a valuable platform for prediction of successful immunotherapy. IMPLICATIONS: Our melanoma model recapitulates the variable response to anti-PD-L1 observed in patients and exhibits biomarkers that characterize early antibody response, including expansion of the TCR repertoire.

Abstract A162: Continuous treatment with rociletinib in EGFR transgenic mice results in acquired resistance to both rociletinib and osimertinib and intra-“patient” tumor heterogeneity

Activating mutations in EGFR are key drivers in a subset of non-small cell lung cancer (NSCLC) patients, as well as effective predictors of response to first-generation reversible EGFR tyrosine kinase inhibitor (TKI) therapy. However, resistance develops in many patients due to acquisition of a T790M mutation in EGFR. Rociletinib and osimertinib target the T790M mutation, and have been evaluated in previously treated EGFR mutant lung cancer patients. Osimertinib recently received FDA approval for treatment of patients with EGFR T790M mutation-positive NSCLC whose disease has progressed on or after EGFR TKI therapy. However, there is limited information on the mechanisms of resistance to these drugs, as well as limited in vivo models for acquired resistance to examine signaling pathways and/or possible treatments to abrogate resistance. We treated an EGFR-L858R-T790M transgenic lung adenocarcinoma mouse model daily with rociletinib to examine sensitivity and acquired resistance. Rociletinib exhibited potent antitumor activity as a single agent, resulting in complete tumor regression and suppression of phospho-EGFR. Despite the success of treatment, after long-term dosing (4-5 months) tumors acquired resistance to the drug, as evidenced by tumor regrowth on MRI. DNA from multiple tumor samples representing four individual drug-resistant mice was sequenced, and several novel mutations in EGFR were observed that may account for the resistant outgrowth of tumors in these mice. These included an insertion and a deletion mutation as well as missense mutations. Interestingly, in some cases different mutations arose in individual tumor nodules within the same mouse, suggesting that each tumor is an individual resistant clone. All mutations were located in the kinase domain of EGFR. For example, a de novo G796C mutation was one of the potential mechanisms of resistance to rociletinib found in two isolated nodules from one mouse lung. A mutation at the same amino acid has been found in a lung cancer patient after treatment with osimertinib; moreover, this mutation was found in a cell line mutagenesis screen to study the resistance to different EGFR TKIs. We generated Ba/F3 cell lines carrying each de novo identified EGFR mutation (along with the activating L858R and T790M mutations) and compared the potency of EGFR TKIs to cell lines carrying only the activating mutations. The cells that expressed putative acquired resistance mutations proved to be resistant to both rociletinib and osimertinib in vitro, suggesting that these mutations indeed mediated acquired resistance in vivo. Tumor nodules from drug-resistant transgenic mice were transplanted into recipient mice and expanded for further evaluation. Mice with transplanted resistant tumors did not respond to rociletinib or osimertinib treatment. Pathway analysis showed that phospho-EGFR was no longer suppressed in drug-resistant tumors carrying a deletion mutation, confirming that novel mutations result in drug resistance. We discovered that resistance to third-generation EGFR inhibitors may occur through novel mutations in EGFR. Heterogeneity of mutations in different nodules within the same lung is akin to intra-patient heterogeneity and highlights the utility of mouse models for examining acquired resistance to drugs.

Citation Format: Rajaa El Meskini, Anthony Iacovelli, Alan Kulaga, Michelle Gumprecht, Myla Spencer, Lilia Ileva, Andrew Simmons, Zoë Weaver Ohler. Continuous treatment with rociletinib in EGFR transgenic mice results in acquired resistance to both rociletinib and osimertinib and intra-“patient” tumor heterogeneity [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2017 Oct 26-30; Philadelphia, PA. Philadelphia (PA): AACR; Mol Cancer Ther 2018;17(1 Suppl):Abstract nr A162.

Abstract 1479: Preclinical model of human melanoma for evaluation of targeted drug treatment and for immunotherapy validation

Malignant melanoma accounts for less than 5% of skin cancer cases, yet it represents 75% of deaths from skin cancer. The high mortality rate is due to the malignant, metastatic nature of the disease and resistance to chemotherapeutic treatments. Most mouse melanoma models have not fully recapitulated the histopathology of the disease and its metastatic nature. At NCI's Center for Advanced Preclinical Research (CAPR), we have adapted the HGF/SF; CDK4R24C transgenic mouse model to an optimized allograft transplant model for preclinical therapeutic studies in primary and metastatic melanoma. This genetically engineered mouse-derived Allograft (GDA) model recapitulates the features of the original GEM, including the epithelioid histopathology and key marker expression of human melanoma. It is an efficient and tractable tool for monitoring of both tumor growth and therapeutic responses in primary and metastatic melanoma in the context of a normal immune system. Additionally, aberrant expression of c-Met and upregulation of the downstream signaling pathway in HGF-GDA tumors is relevant for targeted therapeutics in melanoma. Thus, the model is a useful platform for evaluating therapies that target tumor cells and/or immunomodulatory pathways in intervention or adjuvant settings. Although drugs such as the c-Met inhibitor crizotinib and the MEK inhibitor trametinib were potent in cell culture, PD analyses of short-term (4-6 hour) treatment with small molecule therapies indicated that treatment incompletely suppresses the pathway in vivo compared to the corresponding primary cell line, and does not inhibit tumor growth. Therefore the HGF-GDA can be exploited to examine combination treatments that either prevent feedback activation of downstream pathway nodes in vivo, or modify alternate pathways, such as immunomodulatory targets. Hence, we are currently exploring rational combinations of oncogene-targeted therapy with immune-targeted therapy, for example, combined trametinib and anti-CTLA4 antibody treatment. In the HGF-GDA, complete response was observed in a subgroup of mice treated with anti-CTLA-4, i.e. established tumors fully regressed, yet the durable response and increased survival time (based on tumor volume) was not enhanced by concurrent treatment with trametinib. Future treatment studies will involve alternative regimens. Additionally, since metastasis, not the primary tumor, leads to progression of melanoma in patients, we have characterized a primary tumor resection model in which only metastatic disease is treated. Lung metastases develop after resection of the HGF-GDA primary tumor, which

may be treated in an intervention or adjuvant setting. Therefore we evaluated the effect on survival of the same combination therapy (trametinib and anti-CTLA-4) we previously used to treat the primary tumor, but as adjuvant treatment for the metastatic melanoma model.

Citation Format: Rajaa El Meskini, Michelle Gumprecht, Alan Kulaga, Anthony Iacovelli, Terry Van Dyke, Chi-Ping Day, Glenn Merlino, Zoe Weaver Ohler. Preclinical model of human melanoma for evaluation of targeted drug treatment and for immunotherapy validation. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 1479.

Abstract B03: Synergistic control of GBM growth by MEK and PI3kinase signaling in a RAS-driven preclinical orthotopic model for human glioblastoma multiforme

PI3K, RTK/RAS, and Rb signaling are commonly altered in human GBM. We utilized a genetically engineered mouse (GEM) model for GBM, designated “TRP,” that expresses GFAP-T121 to suppress Rb (T), the KrasG12D mutation (R), and is heterozygous for a PTEN null allele (P), to develop an efficient and tractable orthotopic mouse model for GBM treatment preclinical evaluation. The orthotopic brain tumors presented linear foci of necrosis with peudopalisading by neoplastic cells described in human GBM, and were highly proliferative, invasive, and vascular. Immunohistochemistry analysis of TRP orthotopic tumors identified markers characteristic of human GBM, and tumor progression was readily examined by serial MRI.

Both PI3K and MAPK pathway inhibitors inhibited growth of TRP GEM-derived primary tumor cells, but did not result in significant apoptosis. However, when GBM derived cells were treated with a combination of pathway inhibitors such as BKM120 (a pan-PI3K inhibitor currently in clinical trials for solid tumors) and PD0325901 (a MEK inhibitor), potency was enhanced with a substantial increase in cell death. Analysis of downstream targets revealed a synergistic effect on target downregulation in the PI3K pathway. In the orthotopic GBM model, the PI3K/MAPK targeted combination delayed tumor growth by increasing GBM cell apoptosis and resulted in significantly increased survival. Taken together, our results validate this new orthotopic model for the assessment of targeted therapeutic regimens for human GBM, and show that the feedback loops and incomplete pathway suppression, observed in vivo with targeted single agents, can be overcome with a combination treatment strategy.

Citation Format: Rajaa El Meskini, Anthony Iacovelli, Alan Kulaga, Michelle Gumprecht, Philip Martin, Maureen Baran, Deborah Householder, Terry van Dyke, Zoe Weaver Ohler. Synergistic control of GBM growth by MEK and PI3kinase signaling in a RAS-driven preclinical orthotopic model for human glioblastoma multiforme. [abstract]. In: Proceedings of the AACR Special Conference on RAS Oncogenes: From Biology to Therapy; Feb 24-27, 2014; Lake Buena Vista, FL. Philadelphia (PA): AACR; Mol Cancer Res 2014;12(12 Suppl):Abstract nr B03. doi: 10.1158/1557-3125.RASONC14-B03

A preclinical orthotopic model for glioblastoma recapitulates key features of human tumors and demonstrates sensitivity to a combination of MEK and PI3K pathway inhibitors

Current therapies for glioblastoma multiforme (GBM), the highest grade malignant brain tumor, are mostly ineffective, and better preclinical model systems are needed to increase the successful translation of drug discovery efforts into the clinic. Previous work describes a genetically engineered mouse (GEM) model that contains perturbations in the most frequently dysregulated networks in GBM (driven by RB, KRAS and/or PI3K signaling and PTEN) that induce development of Grade IV astrocytoma with properties of the human disease. Here, we developed and characterized an orthotopic mouse model derived from the GEM that retains the features of the GEM model in an immunocompetent background; however, this model is also tractable and efficient for preclinical evaluation of candidate therapeutic regimens. Orthotopic brain tumors are highly proliferative, invasive and vascular, and express histology markers characteristic of human GBM. Primary tumor cells were examined for sensitivity to chemotherapeutics and targeted drugs. PI3K and MAPK pathway inhibitors, when used as single agents, inhibited cell proliferation but did not result in significant apoptosis. However, in combination, these inhibitors resulted in a substantial increase in cell death. Moreover, these findings translated into the in vivo orthotopic model: PI3K or MAPK inhibitor treatment regimens resulted in incomplete pathway suppression and feedback loops, whereas dual treatment delayed tumor growth through increased apoptosis and decreased tumor cell proliferation. Analysis of downstream pathway components revealed a cooperative effect on target downregulation. These concordant results, together with the morphologic similarities to the human GBM disease characteristics of the model, validate it as a new platform for the evaluation of GBM treatment.

Abstract 3137: Combined inhibition of MEK and PI3 kinase signaling results in improved survival in a preclinical orthotopic model for human glioblastoma multiforme

Glioblastoma multiforme (GBM; astrocytoma grade IV) is the most frequent and aggressive brain tumor for which no effective therapy is currently available. It has been shown that PI3K, RTK/RAS, and Rb signaling are commonly altered in human GBM. We utilized a genetically engineered mouse (GEM) model for GBM, designated

“TRP,” that expresses GFAP-T121 (T; for Rb suppression), the KrasG12D mutation (R), and is heterozygous for a PTEN null allele (P), to develop an orthotopic mouse model for the preclinical evaluation of potential therapeutics for GBM treatment. Although the de novo TRP brain tumor GEM model recapitulates features of human GBM including tissue invasion, pseudopalisading necrosis and dense vascularization, the latency to tumorigenesis (4-6 months) makes its use as a preclinical model for drug screening challenging. Therefore we isolated primary GBM cells from TRP GEMs and injected cells intracranially (IC) into syngeneic mouse brains. Recipient mice developed grade IV astrocytomas and recapitulated TRP GEM tumor histopathology. The orthotopic tumors presented linear foci of necrosis with peudopalisading by neoplastic cells that are hallmarks of human GBM, and were highly proliferative, invasive, and vascular. In addition, immunohistochemistry analysis of TRP orthotopic tumors identified markers characteristic of human GBM, and tumor progression was readily examined by serial MRI.

We used primary tumor cells derived from the TRP model in cell proliferation assays and found that PI3K and MAPK pathway inhibitors used as single agents inhibited cell growth alone, but did not result in significant apoptosis. However, when cells were treated with a combination of agents such as BKM120 (a pan-PI3K inhibitor currently in clinical trials for solid tumors) and PD0325901 (a MEK inhibitor), potency was enhanced and there was a substantial increase in cell death. Analysis of downstream targets revealed a synergistic effect on target downregulation in the PI3K pathway. The drug combination was also evaluated in vivo in the orthotopic model. We found that it delayed tumor growth by increasing GBM cell apoptosis, resulting in significantly increased survival. Our results show that combined targeting of the PI3K and MAPK pathways can have a synergistic effect in glioblastoma, and that our primary cells were a valuable tool to predict the in vivo outcome. We also validated this new orthotopic model of GBM and showed that it is tractable for the assessment of potential therapeutic regimens for human GBM.

Citation Format: Rajaa El Meskini, Anthony Iacovelli, Alan Kulaga, Michelle Gumprecht, Philip Martin, Maureen Baran, Deborah Householder, Terry Van Dyke, Zoe Weaver Ohler. Combined inhibition of MEK and PI3 kinase signaling results in improved survival in a preclinical orthotopic model for human glioblastoma multiforme. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 3137. doi:10.1158/1538-7445.AM2014-3137

Abstract C11: Combined inhibition of MEK and PI3 kinase signaling results in improved survival in a preclinical orthotopic model for human glioblastoma multiforme

Glioblastoma multiforme (GBM; astrocytoma grade IV) is the most frequent and aggressive brain tumor for which no effective therapy is currently available. It has been shown that PI3K, RTK/RAS, and Rb signaling are commonly altered in human GBM. We utilized a genetically engineered mouse (GEM) model for GBM, designated “TRP,” that expresses GFAP-T121 (T; for Rb suppression), the KrasG12D mutation (R), and is heterozygous for a PTEN null allele (P), to develop an orthotopic mouse model for the preclinical evaluation of potential therapeutics for GBM treatment. Although the de novo TRP brain tumor GEM model recapitulates features of human GBM including tissue invasion, pseudopalisading necrosis, and dense vascularization, the latency to tumorigenesis (4-6 months) makes its use as a preclinical model for drug screening challenging. Therefore we isolated primary GBM cells from TRP GEMs and injected cells intracranially (IC) into syngeneic mouse brains. Recipient mice developed grade IV astrocytomas and recapitulated TRP GEM tumor histopathology. The orthotopic tumors presented linear foci of necrosis with peudopalisading by neoplastic cells that are hallmarks of human GBM, and were highly proliferative, invasive, and vascular. In addition, immunohistochemistry analysis of TRP orthotopic tumors identified markers characteristic of human GBM, and tumor progression was readily examined by serial MRI. We used primary tumor cells derived from the TRP model in cell proliferation assays and found that PI3K and MAPK pathway inhibitors used as single agents inhibited cell growth alone, but did not result in significant apoptosis. However, when cells were treated with a combination of agents such as BKM120 (a pan-PI3K inhibitor currently in clinical trials for solid tumors) and PD0325901 (a MEK inhibitor), potency was enhanced and there was a substantial increase in cell death. Analysis of downstream targets revealed a synergistic effect on target downregulation in the PI3K pathway. The drug combination was also evaluated in vivo in the orthotopic model. We found that it delayed tumor growth by increasing GBM cell apoptosis, resulting in significantly increased survival. Our results show that combined targeting of the PI3K and MAPK pathways can have a synergistic effect in glioblastoma, and that our primary cells were a valuable tool to predict the in vivo outcome. We also validated this new orthotopic model of GBM and showed that it is tractable for the assessment of potential therapeutic regimens for human GBM.

Citation Information: Mol Cancer Ther 2013;12(11 Suppl):C11.

Citation Format: Rajaa El Meskini, Anthony Iacovelli, Alan Kulaga, Michelle Gumprecht, Philip L. Martin, Maureen L. Baran, Deborah B. Householder, Terry Van Dyke, Zoe Weaver Ohler. Combined inhibition of MEK and PI3 kinase signaling results in improved survival in a preclinical orthotopic model for human glioblastoma multiforme. [abstract]. In: Proceedings of the AACR-NCI-EORTC International Conference: Molecular Targets and Cancer Therapeutics; 2013 Oct 19-23; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2013;12(11 Suppl):Abstract nr C11.

Temporal molecular and biological assessment of an erlotinib-resistant lung adenocarcinoma model reveals markers of tumor progression and treatment response

Patients with lung cancer with activating mutations in the EGF receptor (EGFR) kinase, who are treated long-term with tyrosine kinase inhibitors (TKI), often develop secondary mutations in EGFR associated with resistance. Mice engineered to develop lung adenocarcinomas driven by the human EGFR T790M resistance mutation are similarly resistant to the EGFR TKI erlotinib. By tumor volume endpoint analysis, these mouse tumors respond to BIBW 2992 (an irreversible EGFR/HER2 TKI) and rapamycin combination therapy. To correlate EGFR-driven changes in the lung with response to drug treatment, we conducted an integrative analysis of global transcriptome and metabolite profiling compared with quantitative imaging and histopathology at several time points during tumor progression and treatment. Responses to single-drug treatments were temporary, whereas combination therapy elicited a sustained response. During tumor development, metabolomic signatures indicated a shift to high anabolic activity and suppression of antitumor programs with 11 metabolites consistently present in both lung tissue and blood. Combination drug treatment reversed many of the molecular changes found in tumored lung. Data integration linking cancer signaling networks with metabolic activity identified key pathways such as glutamine and glutathione metabolism that signified response to single or dual treatments. Results from combination drug treatment suggest that metabolic transcriptional control through C-MYC and SREBP, as well as ELK1, NRF1, and NRF2, depends on both EGFR and mTORC1 signaling. Our findings establish the importance of kinetic therapeutic studies in preclinical assessment and provide in vivo evidence that TKI-mediated antiproliferative effects also manifest in specific metabolic regulation.

Abstract A42: Sensitivity and resistance to a p-AKT inhibitor in an erlotinib-sensitive model for lung cancer

Activating mutations in EGFR, such as the L858R mutation, are effective predictors of response to EGFR tyrosine kinase inhibitor (TKI) therapy. While most patients initially respond to EGFR TKIs, the majority relapse within 6-12 months, underscoring the importance of identifying therapies that enhance and prolong EGFR TKI activity in patients. Agents targeting additional nodes in the EGFR pathway have the potential to overcome such resistance but single-agent activity must be understood first. We utilized mice engineered to develop erlotinib-sensitive lung adenocarcinomas driven by the human EGFR L858R mutation to evaluate the efficacy of a p-AKT inhibitor, MK-2206, currently in clinical trials for a variety of solid tumor malignancies. MK-2206 concentration and target activity were assessed in short-term PK/PD studies in both the lungs and blood of tumor-bearing mice. MK-2206 was then evaluated in an efficacy study using a twice-weekly dosing regimen, both as a single agent and in combination with other TKIs. MK-2206 alone resulted in tumor regression or tumor growth inhibition in ∼50% of the mice, but exhibited greater efficacy in combination with other upstream pathway inhibitors. Temporal evaluation of tumor growth inhibition indicated that MK-2206 was most effective at early time points and did not produce a sustained pathway and growth response as treatment continued. Analysis of gene expression changes in mice treated with MK-2006 alone vs. MK-2206 in combination with EGFR TKIs highlighted the pathways perturbed by both drugs.

Abstract 4322: A preclinical orthotopic mouse model for human GBM: Recapitulation of features of GEM model of origin and potency of PI3K inhibitors

Glioblastoma (GBM; grade IV) is the most frequent and aggressive brain tumor for which no effective therapy is currently available. It has been shown that PI3 kinase, involving both PI3 kinase – Ras activation and PTEN inhibition, and RB signaling are commonly altered in human GBM. We utilized a GEM model expressing GFAP-T121, KrasG12D and heterozygous mutant in PTEN +/- (TRP+/-) to develop an orthotopic mouse model for the preclinical evaluation of potential therapeutics for GBM treatment. The TRP+/- GEM model recapitulates features of human GBM including tissue invasion, pseudopalisading necrosis and dense vascularisation, however the latency to tumorigenesis (4-6 months) makes its use as a preclinical model for drug screening challenging. Therefore we isolated primary tumor cells from induced TRP+/- GEMs with Grade III or Grade IV astrocytoma and injected cells intracranially (IC) into syngeneic mouse brains. Primary cells induced grade IV tumors and recapitulated TRP+/- GEM tumor histopathology. Highly proliferative, invasive, and vascular, the orthotopic grade IV tumors presented linear foci of necrosis with peudopalisading by neoplastic cells that is characteristic of human GBM. Immunohistochemistry characterization of TRP+/- orthotopic tumors revealed similar T121, GFAP and PTEN expression profile to TRP+/- GEM grade IV tumors. Other cellular markers known to be expressed in neuronal progenitors, glial cells or found to be associated with human GBM aggressiveness were assessed in the orthotopic mouse model GBM tumors in comparison with the TRP+/- GEM model. By adjusting the number of primary cells IC injected, the time period of tumor growth and assessment by MRI was increased from 3 weeks to 6 weeks allowing reasonable time for tumor growth assessment, before potential drug efficacy testing. We used primary cells of grade IV tumor origin from TRP+/- GEM to evaluate the effect of PI3 kinase pathway inhibition on cell proliferation. BEZ235, PI103 and PIK75, compounds that target the PI3 kinase pathway, are effective inhibitors of primary tumor cell proliferation with an EC50 in the 10-100 nM range. MK2206, an Akt inhibitor, was found to be a potential inhibitor of tumor cell proliferation. Although MAPK pathway activation through RB inactivation has been found to be key in triggering cell proliferation in TRP+/- GEM model, MEK inhibitors, namely AZD6244, PD0325901 and UO126 did not have a strong effect on cell proliferation. In vitro inhibition of cell proliferation assays and target validation in TRP+/- GEM primary cells are useful tools to choose compounds for orthotopic mouse efficacy studies. Preliminary in vivo mouse pharmacokinetic studies are underway to insure potential drug distribution in the brain tissue. Using the orthotopic model, we found that PI3 kinase inhibitors may be effective against brain tumors in the mouse orthotopic model for human GBM.

Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 4322. doi:10.1158/1538-7445.AM2011-4322