Abstract
The clinical utility of approved EGFR small-molecule kinase inhibitors is plagued both by toxicity against wild-type EGFR and by metastatic progression in the central nervous system, a disease sanctuary site. Here, we report the discovery and preclinical efficacy of GNS-1486 and GNS-1481, two novel small-molecule EGFR kinase inhibitors that are selective for T790M-mutant isoforms of EGFR. Both agents were effective in multiple mouse xenograft models of human lung adenocarcinoma (T790M-positive or -negative), exhibiting less activity against wild-type EGFR than existing approved EGFR kinase inhibitors (including osimertinib). In addition, GNS-1486 showed superior potency against intracranial metastasis of EGFR-mutant lung adenocarcinoma. Our results offer a preclinical proof of concept for new EGFR kinase inhibitors with the potential to improve therapeutic index and efficacy against brain metastases in patients. Cancer Res; 77(5); 1200–11. ©2017 AACR.
Introduction
Lung cancer is the leading cause of cancer mortality worldwide, with lung adenocarcinoma as the most common histologic subtype (1, 2). The clinical success of oncogene-targeted therapy in specific subsets of lung adenocarcinoma patients, such as those with activating mutations in EGFR, has heralded a new era of precision cancer medicine with great promise for improving patient survival and quality of life (3, 4–10). However, in the case of EGFR-mutant lung adenocarcinoma, both clinical toxicity due to residual activity against wild-type (WT) EGFR versus mutant EGFR and metastatic tumor progression in the central nervous system (CNS) are two remaining obstacles that limit the overall clinical impact of the current first- (gefitinib, erlotinib), second- (afatinib), and third-generation (osimertinib) EGFR tyrosine kinase inhibitors (TKI) that are FDA approved (11, 12, 13–18, 19). Importantly, lung adenocarcinoma patients with CNS metastasis have a particularly dismal prognosis, as no drug therapy has shown consistent or durable efficacy against intracranial metastasis to date (19, 20).
During the treatment of EGFR-mutant lung adenocarcinoma patients with first-generation EGFR TKIs (erlotinib, gefitinib), tumor progression often occurs via the emergence of the EGFRT790M resistance mutation (21, 22). This observation prompted the development of second- and third-generation irreversible EGFR inhibitors with activity against EGFRT790M (21, 23, 24). Some of these newer EGFR inhibitors, such as CO-1686 (rociletinib) and AZD9291 (osimertinib), exhibit increased selectivity for mutant EGFR with relative sparing of WT EGFR, as compared with earlier EGFR inhibitors, including erlotinib, gefitinib, and afatinib (12). This relative selectivity for mutant EGFR over WT EGFR can enhance the therapeutic index for EGFR inhibition in patients, potentially reducing certain toxicities that occur because of WT EGFR blockade (such as cutaneous and gastrointestinal side effects; refs. 25, 26). Although the development of CO-1686 (rociletinib) has been discontinued (in part due to less impressive clinical efficacy than initially anticipated), AZD9291 (osimertinib) is now approved for the second-line treatment of lung adenocarcinoma patients with EGFRT790M-positive disease (25, 26). Although osimertinib appears to be associated with decreased clinical toxicity (by historical comparison with first- and second-generation EGFR TKIs), side effects linked to residual activity against WT EGFR remain a clinical challenge and impair the quality of life in patients (including grade 3 adverse events occurring in ∼33% of osimertinib-treated individuals; toxicity that is consistent with the experience using osimertinib in our own clinical practices; 25–28). In addition to the clinical toxicity and quality-of-life issues, the recommended drug dose (or in some cases dose reduction or suspension) that is used as a consequence of the toxicity resulting from the submaximal selectivity for mutant EGFR over WT EGFR of the current FDA-approved EGFR TKIs can lead to incomplete (or nonsustained) target inhibition in both intracranial and extracranial tumor cells, thereby potentially contributing to the progression of metastatic tumors both within and outside of the CNS (11, 21, 26, 29).
Disease progression in the CNS, a sanctuary site, is a widespread cause of death in EGFR-mutant lung adenocarcinoma patients (19). Limited published reports show that the current approved EGFR inhibitors (including osimertinib) have documented but inconsistent and often temporary clinical efficacy against CNS metastases [abstracts: Kim D and colleagues Annals of Oncology (2014) 25 (Suppl_4): iv146-iv164. 10.1093/annonc/mdu331; Camidge DR and colleagues MINI16.04, 16th World Conference on Lung Cancer, 2015; Sequist LV and colleagues J Clin Oncol. 2014;32 (15 Suppl): abstract 8010 2014 ASCO Annual Meeting; refs. 7, 19, 24, 25, 26, 30, 31]. There remains no established and widely effective systemic treatment for CNS metastases in patients with EGFR-mutant lung adenocarcinoma, and progression of CNS metastasis has been reported and observed in our own clinical practices in patients treated with all current FDA-approved EGFR inhibitors, including osimertinib (Ahn MJ, and colleagues ESMO 2015. Abstract 3083; refs. 19, 28, 32). Thus, although recently initiated clinical trials are testing certain EGFR TKIs, such as osimertinib, in patients with CNS metastasis (e.g., NCT02736513), the CNS antitumor efficacy of the EGFR TKIs that are currently approved remains an unresolved and active area of investigation.
To address the limitations of the current approved EGFR TKIs, we conducted a drug discovery program to discover a potent, mutant-selective EGFR TKI with less WT EGFR activity and thus potentially a wider therapeutic index versus the currently approved EGFR TKIs and that also exhibits pronounced activity against intracranial EGFR-mutant lung adenocarcinoma metastasis. This discovery program has led to the identification of two novel and improved EGFR TKIs; the data provide the rationale for clinical trials that will be initiated soon testing these promising new agents in EGFR-mutant (including EGFRT790M) lung adenocarcinoma patients with intracranial and extracranial metastatic disease.
Materials and Methods
Cell culture and reagents
The human non–small cell lung cancer (NSCLC) cell lines (HCC827 and H1975) and NIH 3T3 (mouse embryonic fibroblast cell line) were obtained from the ATCC. PC-9 cells were a gift from F. Koizumi and K. Nishio (National Cancer Center Hospital, Tokyo, Japan). PC-9/GR (gefitinib-resistant cell line) and PC-9/ER (erlotinib-resistant cell line) cells were established as part of a previous study (33). Cells were cultured in RPMI1640 medium containing 10% FBS, 2 mmol/L l-glutamine, and 100 U/mL of penicillin and streptomycin, and maintained at 37°C in a humidified chamber containing 5% CO2. Osimertinib, CO-1686, and WZ4002 were purchased from Selleck Chemicals. The cell lines used were authenticated by STR analysis and confirmed to be mycoplasma free using standard methods.
Cellular viability assays
To perform the MTT assay, cells (5 × 103) were seeded in 96-well sterile plastic plates overnight and then treated with relevant agents. After 72 hours, 15 μL of MTT solution (5 mg/mL) was added to each well, and plates were incubated for 4 hours. Crystalline formazan was solubilized with 100 μL of a 10% (w/v) SDS solution for 24 hours, and then absorbance at 595 nm was read spectrophotometrically using a microplate reader. To test the colony formation assay, cells (0.1–1 × 103) were seeded in 6-well plates and then treated with relevant agents. After 10 to 14 days, the colonies were stained with crystal violet and counted. The results are representative of at least three independent experiments, and the error bars signify SDs.
Cell-free kinase assay
Cell-free kinase assays were conducted using Lance Ultra time-resolved fluorescence resonance energy transfer technology according to the manufacturer's instructions (PerkinElmer). Briefly, various concentrations of EGFR inhibitors were mixed with each enzyme (WT or mutant EGFR, Her2, and Her4), the Ulight-poly-GT peptide substrate and ATP in a kinase buffer (50 mmol/L HEPES pH 7.5, 10 mmol/L MgCl2, 1 mmol/L EFTA, 2 mmol/L DTT, and 0.01% Tween-20) in a 96-well plate. Kinase reactions were incubated at room temperature for 1 hour and then stopped by the addition of EDTA. The specific europium-labeled anti-phospho petide antibody (PerkinElmer) was added to the reaction in Lance detection buffer. The mixture was allowed to incubate for 30 minutes to allow binding of the antibody to the phosphorylated site before the plate was read. The LANCE signal was measured on an EnVision Multilabel Reader (PerkinElmer). Excitation wavelength was set at 320 nm, and emission was monitored at 615 nm (donor) and 665 nm (acceptor). The IC50 values were determined using GraphPad Prism software.
Kinase profile assay
The kinase selectivity was assessed by KinaseProfiler (Millipore) consisting of 321 kinases for GNS-1481 or 323 kinases for GNS-1486 and osimertinib at a single concentration of 1 μmol/L using ATP Km for each kinase.
Expression vectors and transfections
EGFR constructs (EGFR wild-type, EGFR del746-750, EGFR T790M/L858R, and EGFR T790M/del746-750) were purchased from Addgene. Transfections were performed using Lipofectamine 2000 (Invitrogen) in accordance with the manufacturer's instructions. Transfected cells were selected using puromycin (2 μg/mL for NIH3T3; Sigma).
Immunoblotting
Cells were lysed in buffer containing 137 mmol/L NaCl, 15 mmol/L EGTA, 0.1 mmol/L sodium orthovanadate, 15 mmol/L MgCl2, 0.1% Triton X-100, 25 mmol/L MOPS, 100 mmol/L phenylmethylsulfonyl fluoride, and 20 mmol/L leupeptin, adjusted to pH 7.2. Lysis of tumor specimens was performed using Omni Tissue Homogenizer (TH; Omni International). Antibodies specific for p-EGFR (Tyr1173), EGFR, Akt, ERK, and actin were obtained from Santa Cruz Biotechnology, and antibodies for PARP, caspase 3, p-Akt (Ser473), and p-ERK (Thr202/Tyr204) were purchased from Cell Signaling Technology. Proteins were detected with an enhanced chemiluminescence Western Blotting Kit (Amersham Biosciences), according to the manufacturer's instructions.
Animal models
To establish xenograft models, female SCID mice (18–20 g, 6 weeks of age) were purchased from Charles River Laboratories. All experimental procedures were conducted following a protocol approved by the Institutional Animal Care and Use committee of Asan Institute for Life Sciences (Seoul, Korea; 2014-12-103 and 2015-12-087). Tumors were grown by implanting cells (1–5 × 106 cells/0.1 mL) in 50% Matrigel (BD Biosciences) and subcutaneously injected into the right flank of animals. Drug treatment was started when the tumors reached a volume of 50 to 100 mm3. To measure tumor size, the length (L) and width (W) of each tumor was measured using calipers, and tumor volume (TV) was calculated as TV = (L × W2)/2. To perform the intracranial implantation of HCC827-luc cells, the human NSCLC cell line HCC827 was stably integrated with a luciferase reporter gene, RediFect lentiviral particles (PerkinElmer). Female athymic nude mice (6–8 weeks of age, Charles River Laboratories) were anesthetized with a ketamine/xylazine cocktail solution. The head of the mouse was stabilized by using a Harvard Apparatus stereotaxic head frame. After disinfection of the skin, a 1-cm midline scalp incision was made, and a burr hole (coordinates, 2.5 mm lateral and 0.5 mm posterior to the bregma) in the skull was made by using a high-speed microdrill. Cells (1–5 × 105 cells/4 μL serum-free RPMI) were injected into the right striatum using a 10 μL Hamilton syringe to deliver tumor cells to a 3.5-mm intraparenchymal depth. The burr hole in the skull was sealed with bone wax and the incision closed using Dermabond. Tumor growth was monitored and measured via bioluminescence imaging (BLI) in vivo.
Drug administration
GNS-1481 and -1486 were dissolved in NMP/PEG300 (1:9, v/v), and osimertinib and CO-1686 were dissolved in Tween-80. All drugs were given by oral intubation for the indicated times.
Bioluminescence monitoring and μCT coregistration of intracranial tumors
Intracranial tumor growth quantified by BLI was performed using an IVIS spectrum system (Caliper, PerkinElmer Company). Mice were administered by intraperitoneal injection with 150 mg/kg body weight of d-luciferin (Caliper Life Sciences) dissolved in DPBS (Gibco). Before and during imaging, mice were anesthetized by 1% isoflurane inhalation (Forane). Bioluminescent signals were acquired with open filter or emission at 620 nm using autoacquisition and field of view of 13.4 cm, and bioluminescent signals were quantified as radiance (photon/sec/cm2/sr) within a circular region of interest (ROI) using Living Image 4.4 software. To visualize the anatomic location, mice were imaged with a Quantum FX μCT system after optical imaging. μCT image used a 120 mm FOV with a 236 μm voxel size and a dose of 26 mGy per scan. Three-dimensional (3D) optical and reconstructed μCT images were automatically generated with the Living Image 4.4 Software and 3D ROIs were measured as voxels (photons/second).
Immunohistochemical staining
Each tumor was harvested at the indicated times postadministration with drugs. Resected tumors were fixed in 10% formaldehyde and embedded in paraffin. Immunohistochemical staining was done using a specific primary antibody (Ki-67; DakoCytomation), the EnVision Plus Staining Kit (DakoCytomation), and APO-Direct terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) Assay Kit (Millipore) according to the supplier's instructions. Quantitative analysis of section staining was done by counting immunopositive cells in five arbitrarily selected fields.
Statistical analysis
P values were determined with unpaired t tests between comparator groups using GraphPad software.
Results
Discovery and characterization of novel mutant-selective EGFR TKIs
In search of novel WT or mutant EGFR (Del 19, E746-A750), L858R, L858R/T790M, T790M, Del 19/T790M) inhibitors, we performed a high-throughput screen of a proprietary compound library consisting of 1,583 structurally diverse molecules. This campaign led to the identification of 30 compounds displaying activity against mutant EGFR. Structure-based drug design utilizing structure activity relationship information from the 30 hit compounds led to several potent and selective inhibitors with favorable physical properties and in vitro ADME (absorption, distribution, metabolism, and excretion). Among those 30 compounds, two novel 3-pyrazolypyrimidine compounds (GNS-1481 and GNS-1486) were identified as potent and selective irreversible EGFR inhibitors against all forms of mutant EGFR tested. GNS-1481 and GNS-1486 contain a pyrimidine-based scaffold, analogous to other third-generation EGFR kinase inhibitors (Fig. 1A and B; refs. 23, 24). GNS-1481 and GNS-1486 exhibited broad activity at low nanomolar concentrations against all EGFR mutations tested, including EGFRT790M in kinase inhibition in vitro assays (Table 1). Moreover, GNS-1481 and GNS-1486 showed substantial selectivity for mutant EGFR as compared with WT EGFR, as indicated by the selectivity index (defined as the ratio of the IC50 for WT EGFR/IC50 for the L858R-T790M double mutant, Table 1, bottom row). GNS-1486, in particular, showed superior selectivity for mutant EGFR over WT EGFR compared with the other approved EGFR inhibitors, including osimertinib (Table 1). Structural modeling studies of the thermodynamics and binding of GNS-1481 and GNS1486 to either WT or mutant EGFR suggested distinct interactions enabling a tighter association with mutant EGFR versus WT EGFR overall and improved mutant selectivity compared with osimertinib in general, consistent with the kinase profiling studies (Supplementary Fig. S1; Table 1). These findings indicating the substantial potency and mutant EGFR selectivity of the GNS compounds were further extended across a panel of approximately 320 kinases in a multikinase inhibition assay in vitro (Fig. 1C; Supplementary Table S1). Although strong activity against various forms of mutant EGFR was observed for GNS-1481 and GNS-1486 (each tested at 1 μmol/L), these agents exhibited less or minimal activity against WT EGFR or other EGFR family members, such as ErbB2 and ErbB4, when compared with osimertinib (Fig. 1C; Supplementary Table S1). Interestingly, we noted that GNS-1481 and GNS-1486, in contrast to osimertinib, exhibited in vitro activity against RET [either WT or the gatekeeper mutant V804L that is resistant to certain other RET inhibitors (34)], an oncogene in other tumor types including thyroid cancer and a distinct subset of lung adenocarcinoma driven by RET gene rearrangements (Fig. 1C; Supplementary Table S1; refs. 34, 35). Furthermore, GNS-1481 and GNS-1486 showed unique activity against MLK1 (mixed-lineage kinase 1), which can activate MEK–ERK signaling and promote RAF inhibitor resistance in melanoma (Fig. 1C; Supplementary Table S1; ref. 36). Together, these biochemical data reveal the high potency, substantial mutant EGFR selectivity, and unique target profiles of the two novel irreversible EGFR inhibitors we identified, GNS-1481 and GNS-1486.
Kinase . | GNS-1481 . | GNS-1486 . | Osimertinib . | CO-1686 . | Afatinib . | Erlotinib . |
---|---|---|---|---|---|---|
EGFR | ||||||
Wild type | 26.5 | 121.6 | 16.7 | 928.2 | 0.2 | 0.6 |
Del19 (E746-A750) | 6.6 | 20.5 | 8.6 | 216.5 | 0.2 | 0.8 |
L858R | 25.8 | 103.6 | 12.2 | 742.9 | 0.3 | 0.9 |
Double mutant (L858R-T790M) | 6.2 | 8.3 | 4.5 | 45.7 | 18.6 | 549.3 |
T790M | 3.4 | 4.2 | 2.2 | 20.5 | 1.6 | 395.2 |
Double mutant (Del19-T790M) | 4.2 | 4.3 | 3.3 | 44.1 | 7.0 | 715.6 |
Wild type/double mutant (L858R-T790M) | 4.3 | 14.7 | 3.7 | 20.3 | 0.01 | <0.01 |
Kinase . | GNS-1481 . | GNS-1486 . | Osimertinib . | CO-1686 . | Afatinib . | Erlotinib . |
---|---|---|---|---|---|---|
EGFR | ||||||
Wild type | 26.5 | 121.6 | 16.7 | 928.2 | 0.2 | 0.6 |
Del19 (E746-A750) | 6.6 | 20.5 | 8.6 | 216.5 | 0.2 | 0.8 |
L858R | 25.8 | 103.6 | 12.2 | 742.9 | 0.3 | 0.9 |
Double mutant (L858R-T790M) | 6.2 | 8.3 | 4.5 | 45.7 | 18.6 | 549.3 |
T790M | 3.4 | 4.2 | 2.2 | 20.5 | 1.6 | 395.2 |
Double mutant (Del19-T790M) | 4.2 | 4.3 | 3.3 | 44.1 | 7.0 | 715.6 |
Wild type/double mutant (L858R-T790M) | 4.3 | 14.7 | 3.7 | 20.3 | 0.01 | <0.01 |
NOTE: Shown is the IC50 in nmol/L of each drug against the kinase activity of each indicated protein. The bottom row indicates the mutant-selective index (increased proportional to the degree of mutant-selective activity, ratio calculated as the IC50 WT EGFR/L858R-T790M double mutant).
Preclinical efficacy of the novel mutant-selective EGFR kinase inhibitors GNS-1481 and GNS-1486 in vitro
We next examined the activity of both GNS-1481 and GNS-1486 on signaling and cell viability in EGFR-mutant preclinical models. We first studied the impact of treatment with each agent and with other third-generation EGFR inhibitors (osimertinib, CO-1686, and WZ4002; ref. 37) on signaling in NIH-3T3 cells engineered to express either WT or mutant EGFR (Del 19, Del 19/T790M, L858R/T790M; Fig. 2A and B). GNS-1481 and GNS-1486 had comparatively minimal impact on the levels of phosphorylated (p)-EGFR, or the downstream signaling components p-ERK and p-AKT, in WT EGFR-expressing cells, consistent with the WT EGFR-sparing activity of these agents observed in the biochemical assays (Fig. 2B; Supplementary Fig. S2A). In contrast, treatment with each agent suppressed the levels of p-EGFR, p-ERK, and p-AKT and induced cleavage of PARP (as a measure of apoptosis) in each EGFR-mutant model (Fig. 2A and B; Supplementary Fig. S2A and B). We did not find substantial in vitro activity of the GNS agents against the C797S-mutant form of EGFR that can promote resistance to other third-generation EGFR inhibitors, such as osimertinib (data not shown; ref. 27).
We next assessed the effects of GNS-1481 and GNS-1486 treatment on cell viability in human EGFR-mutant lung adenocarcinoma cell lines, including those with EGFRT790M. We found that treatment with either agent was highly effective against multiple EGFR-mutant models in vitro (including PC9 and HCC827 cells with Del 19; Fig. 3A and B; Supplementary Table S2). Moreover, GNS-1481 and GNS-1486 exhibited substantial efficacy in EGFRT790M-positive lung adenocarcinoma models in vitro, including PC-9/GR and PC-9/ER sublines with acquired resistance to gefitinib or erlotinib (33, 38), respectively, and H1975 cells that intrinsically harbor EGFRL858R/T790M (Fig. 3A and B; Supplementary Table S2). We further established the specificity of GNS-1481 and GNS-1486 efficacy for lung cancer cells with mutant EGFR by testing these agents in multiple EGFR WT lung cancer models, including A549, H460, and A341 cells. We found no significant impact of treatment with either agent on cell viability in these models (Supplementary Fig. S2C). GNS-1481 and GNS-1486 (or osimertinib) monotherapy had no significant effect on the viability of EGFR-mutant lung adenocarcinoma cells with acquired erlotinib/gefitinib resistance driven by non-EGFRT790M–mediated mechanisms, such as MET and AXL kinase upregulation, as expected (Supplementary Fig. S2D; refs. 39, 40). Together, these findings further establish the substantial and specific activity of GNS-1481 and GNS-1486 in multiple EGFRT790M (−) and (+) EGFR-mutant lung adenocarcinoma preclinical models.
Analysis of the impact of GNS-1481 and GNS-1486 treatment on key signaling components in the EGFR-mutant lung adenocarcinoma models (PC-9, PC-9/GR, PC-9/ER, and H1975) revealed that each agent suppressed the levels of p-EGFR, p-ERK, and p-AKT in each system (Fig. 3C; Supplementary Fig. S2E). Treatment with GNS-1481 and GNS-1486 was lethal in these EGFR-mutant lung adenocarcinoma cell lines, as these effects on signaling were accompanied by the induction of apoptosis upon GNS-1481 and GNS-1486 treatment, indicated by increased levels of both cleaved PARP and caspase-3 (Supplementary Fig. S2F). Together, these data indicate that GNS-1481 and GNS-1486 are novel mutant-selective EGFR inhibitors that strongly suppress mutant EGFR signaling and induce apoptosis, resulting in substantial efficacy in multiple EGFRT790M (−) and (+) EGFR-mutant lung adenocarcinoma preclinical models. Overall, we noted in these in vitro studies that the GNS agents showed superior WT EGFR-sparing properties compared with the other clinically approved EGFR TKIs and demonstrated therapeutic effects that are comparable with the other third-generation EGFR inhibitors tested (with relatively subtle differences observed across these in vitro systems and the parameters measured).
Preclinical efficacy of the novel mutant-selective EGFR kinase inhibitors GNS-1481 and GNS-1486 in vivo
Supported by our encouraging findings showing improved sparing of WT EGFR, substantial mutant EGFR-selective potency, and efficacy in vitro, we next investigated the in vivo properties of GNS-1481 and GNS-1486, including pharmacokinetic, pharmacodynamic, safety, and antitumor efficacy analysis. In vivo pharmacology and pharmacokinetic studies in mouse, rat, and dog revealed substantial exposure upon oral administration at 10 mg/kg for mouse and rat and 5 mg/kg for dog (Supplementary Fig. S3A, data not shown). This oral dosing regimen achieved plasma concentrations near or above 1 μmol/L (particularly in mouse and dog; Supplementary Fig. S3A, data not shown). Furthermore, this pharmacokinetic analysis demonstrated a half-life predictive of once-daily oral dosing for each agent (or potentially twice daily for GNS-1481 pending ongoing in vivo pharmacokinetic assessment; Supplementary Fig. S3A, data not shown). Importantly, no systemic toxicity was noted in the animals during the pharmacokinetic and dose-finding studies (data not shown).
We further tested the efficacy of GNS-1481 and GNS-1486 treatment against multiple human EGFR-mutant lung adenocarcinoma models implanted subcutaneously into mice, including models with EGFRT790M (PC-9/GR and H1975; Fig. 4A). We found that once-daily oral treatment with either agent demonstrated substantial in vivo efficacy in each model, inducing sustained tumor regressions in PC-9 and PC-9/GR tumors and initial regressions and subsequent disease control in H1975 tumors (Fig. 4A). The efficacy of GNS-1481 and GNS-1486 was similar to gefitinib in the PC9 (EGFRDel 19) tumor xenografts, and superior to gefitinib in the PC9/GR (EGFRDel19/T790M) and H1975 (EGFRL858R/T790M) tumor xenografts, as expected given that gefitinib is largely inactive against EGFRT790M-positive cancers (Fig. 4A). The antitumor efficacy of GNS-1481 and GNS-1486 occurred without substantial overt toxicity in the treated animals (Supplementary Fig. S3B, data not shown). Analysis of key pharmacodynamic biomarkers, including p-EGFR, p-AKT, and p-ERK, indicated that GNS-1481 and GNS-1486 substantially suppressed EGFR activation and downstream signaling (Fig. 4B and C). These effects of GNS-1481 and GNS-1486 treatment on signaling were accompanied by decreased proliferation (as measured by quantitative analysis of Ki-67 staining in the tumor cells by IHC) and increased apoptosis (as measured by quantitative analysis of TUNEL staining in the tumor cells by IHC; Fig. 4D and E). Moreover, we compared the activity of the GNS agents with osimertinib in two distinct EGFRT790M-positive in vivo models and, overall, found comparable antitumor and signaling effects in response to treatment with each drug (Supplementary Fig. S4A–S4D). However, we found that both GNS-1486 and GNS-1481 showed less activity against WT EGFR in the skin of treated animals in multiple different in vivo systems, compared with treatment with either erlotinib or osimertinib and at doses where p-EGFR inhibition in the tumors (H1975) in vivo was equivalent between the GNS agents and osimertinib (Supplementary Fig. S4E and S4F). Together, these data indicate that GNS-1481 and GNS-1486 show substantial mutant EGFR-selective, oral antitumor efficacy with less activity against WT EGFR in vivo than currently approved EGFR TKIs (including osimertinib), offering a potentially wider safety margin than these current EGFR TKIs while showing substantial antitumor efficacy.
Intracranial antitumor efficacy of the novel mutant-selective EGFR kinase inhibitor GNS-1486 in vivo
During the preclinical pharmacology and pharmacokinetic in vivo analysis, we noted that GNS-1481 and GNS-1486 exhibited substantial CNS penetration, with intravenous administration in rats resulting in a CNS/plasma concentration ratio of 0.53–6.15 within 2 hours of initial dosing (Supplementary Table S3). These observations suggested that GNS-1481 or GNS-1486 could potentially show efficacy against intracranial tumors in vivo. We tested this hypothesis by establishing intracranial tumors in mice using human EGFR-mutant (Del 19) lung adenocarcinoma cells engineered to stably express a luciferase reporter to enable in vivo bioluminescence-based monitoring of tumor growth (HCC827-Luc cells, Fig. 5A). Intracranial implantation of HCC827-Luc cells resulted in substantial tumor growth in the brain within 2 to 3 weeks, confirmed by both BLI and by pathologic analysis of brain sections obtained from tumor-bearing mice (Fig. 5A and B). We then treated mice with established intracranial EGFR-mutant lung adenocarcinoma with once-daily orally administered GNS-1486 and assessed CNS tumor growth in vivo, making comparison with osimertinib as the only approved third-generation EGFR inhibitor and that has shown some CNS activity in limited published reports (30, 31). Importantly, the CNS activity of osimertinib and other EGFR TKIs in patients remains under active clinical investigation and has not yet been firmly established in large patient cohorts (to our knowledge). GNS-1486 was chosen for study because of its better selectivity for mutant EGFR over WT EGFR, and its improved pharmacokinetic parameters compared with GNS-1481. We found that once-daily oral GNS-1486 substantially suppressed intracranial tumor growth in these mice, as measured both by BLI and μCT (Fig. 5C–G). In an animal with a large detectable spinal metastasis, we further noted excellent in vivo activity of GNS-1486 against both the primary brain and spinal tumors, consistent with the substantial CNS penetration demonstrated by this agent (Supplementary Fig. S5A). We found that CO-1686 (rociletinib) did not, but osimertinib did show CNS activity in the HCC827 intracranial tumor system (Fig. 6A–D, data not shown), consistent with prior data (30). However, GNS-1486 showed increased potency versus osimertinib against the CNS metastases in this in vivo system, as evidenced by the increased efficacy of GNS-1486 versus osimertinib at the 3 mg/kg/day dosing of each agent (measured by quantitative BLI analysis and comparison; P = 0.044 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib at 1 week and P = 0.007 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib at 2 weeks; Figs. 5D–F and 6B–E).
Consistent with these observations, we found that GNS-1486 treatment suppressed p-EGFR, p-AKT, and p-ERK levels and increased apoptosis as measured by both cleaved PARP and caspase-3 levels in the intracranial tumors, again with an improved potency compared with osimertinib as shown by the increased efficacy of GNS-1486 versus osimertinib in the 3 mg/kg/day treatment cohorts [p-EGFR: P < 0.0001 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib; p-Akt: P = 0.004 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib; p-Erk: P = 0.015 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib; cleaved PARP: P = 0.028 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib; cleaved caspase-3: P = 0.018 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib (Supplementary Fig. S5B)]. Superior in vivo pharmacodynamics of GNS-1486 versus osimertinib were also observed at the higher 10 mg/kg/day cohorts for certain pharmacodynamic biomarkers (p-Erk: P = 0.049 for 10 mg/kg GNS-1486 superiority vs. 10 mg/kg osimertinib; cleaved PARP: P = 0.016 for 10 mg/kg GNS-1486 superiority vs. 10 mg/kg osimertinib; cleaved caspase-3: P = 0.024 for 10 mg/kg GNS-1486 superiority vs. 10 mg/kg osimertinib; p-EGFR: P = 0.16 for 10 mg/kg GNS-1486 vs. 10 mg/kg osimertinib; p-Akt: P = 0.334 for 10 mg/kg GNS-1486 vs. 10 mg/kg osimertinib); in contrast to the 3 mg/kg/day comparison between GNS-1486 and osimertinib, these improved pharmacodynamic effects at the 10 mg/kg dose were not associated with statistically significant differences in the antitumor efficacy as measured by BLI analysis at these higher drug doses and time points in this system (P = 0.28 for 10 mg/kg GNS-1486 vs. 10 mg/kg osimertinib at 1 week; P = 0.9 for 10 mg/kg GNS-1486 vs. 10 mg/kg osimertinib at 2 weeks; Figs. 5D–F and 6B–D).
We further found that treatment with GNS-1486 substantially improved survival in mice with intracranial EGFR-mutant tumors (using the HCC827-Luc system; Fig. 6E). We again noted an improved potency of GNS-1486 compared with osimertinib in these preclinical trials, as evidenced by the efficacy outcome data in the 3 mg/kg/day treatment cohorts (P = 0.0035 for 3 mg/kg GNS-1486 superiority vs. 3 mg/kg osimertinib; P = 0.7519 for 10 mg/kg GNS-1486 vs. 10 mg/kg osimertinib). We further confirmed the substantial intracranial antitumor activity of GNS-1486 in an additional patient-derived EGFR-mutant lung adenocarcinoma model (H1975 EGFRL858R/T790M cells; Supplementary Fig. S6). In contrast, CO-1686 (rociletinib) treatment was less effective in this intracranial H1975 tumor system (Supplementary Fig. S6). Importantly, no signs of systemic or CNS toxicity were noted during treatment with GNS-1486 in the mice (data not shown), consistent with the high selectivity for mutant EGFR versus WT EGFR we observed for this agent. The data suggest that GNS-1486 exhibits superior potency against EGFR-mutant lung adenocarcinoma CNS disease versus osimertinib (and rociletinib) in these in vivo systems, offering a new highly effective therapeutic agent whose potency and selectivity for mutant EGFR provides the potential advantage of decreased clinical toxicity. Altogether, these findings establish GNS-1486 as a novel, improved, orally administered, CNS-penetrant, and mutant-selective EGFR (including EGFRT790M) inhibitor with both potent extracranial and intracranial antitumor efficacy and a wide apparent safety margin. The data provide the rationale for clinical trials, testing the safety and efficacy of GNS-1486 in EGFR-mutant (including EGFRT790M) lung adenocarcinoma patients with active CNS metastatic disease (as well as extracranial disease); these trials will be initiated soon.
Discussion
In summary, despite recent important progress in the field, the identification of a potent, more mutant EGFR-selective and CNS-active EGFR TKI with an improved safety margin could have a substantial and immediate beneficial impact on both patient outcomes and quality of life. Toward this end, our study establishes the preclinical efficacy, safety, and potential clinical utility of novel mutant-selective EGFR TKIs (GNS-1481, GNS-1486). The improved, potent mutant-EGFR selective agents that we describe here show two important advantages over the currently approved EGFR TKIs, including osimertinib: (i) less WT EGFR inhibition and, therefore, risk of toxicity while exhibiting substantial antitumor efficacy and (ii) improved potency against CNS metastasis. Given these distinct attributes, these new EGFR TKIs complement the currently approved EGFR TKIs and are now under clinical development; indeed, clinical trials testing these new agents in EGFR-mutant (including EGFRT790M) lung adenocarcinoma patients with active CNS and extracranial metastatic disease will be initiated soon. Although only these clinical trials can establish whether the greater selectivity for mutant EGFR versus WT EGFR and the increased potency against CNS disease that we observed in the preclinical systems will yield improved outcomes and safety in patients, these promising new agents have the potential to exhibit efficacy in patients with both extracranial and intracranial EGFR-mutant lung adenocarcinoma (including EGFRT790M-positive disease), with less clinical toxicity, to thereby potentially improve not only the quantity but also quality of life for patients. Our study highlights the utility of identifying highly potent and oncoprotein-selective targeted agents that show substantial activity against CNS metastasis early in the drug discovery and development process to reduce both clinical toxicity and the high burden and related mortality of CNS metastasis, as well as extracranial disease, in patients.
Disclosure of Potential Conflicts of Interest
T.G. Bivona reports receiving other commercial research support from Ignyta and is a consultant/advisory board member for Array, AstraZeneca, and Novartis. No potential conflicts of interest were disclosed by the other authors.
Authors' Contributions
Conception and design: J.K. Rho, I.Y. Lee, C.-M. Choi, J.S. Koh, J. Lee, H.-J. Song, P. Salgaonkar, Jaesang Lee, W.S. Kim, C.H. Kim, T.G. Bivona, J.C. Lee
Development of methodology: I.Y. Lee, B.-C. Suh, H.-J. Song, Jaesang Lee, D.S. Jung, D.-C. Woo, I.-J. Baek, J.-Y. Lee, J.K. Kim
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): J.K. Rho, I.Y. Lee, C.-M. Choi, J.-Y. Hur, B.-C. Suh, H.-J. Song, P. Salgaonkar, Jaesang Lee, D.S. Jung, S.-Y. Kim, I.-J. Baek, Y.H. Sung, J.K. Kim, W.S. Kim, J.S. Song
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): J.K. Rho, I.Y. Lee, Y.J. Choi, B.-C. Suh, H.-J. Song, Jungmi Lee, Jaesang Lee, D.S. Jung, D.-C. Woo, C.H. Ha, J.K. Kim, C.H. Kim, T.G. Bivona, J.C. Lee
Writing, review, and/or revision of the manuscript: J.K. Rho, I.Y. Lee, B.-C. Suh, H.-J. Song, P. Salgaonkar, Jaesang Lee, J.K. Kim, T.G. Bivona, J.C. Lee
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): J.K. Rho, I.Y. Lee, J.K. Kim, C.H. Kim, T.G. Bivona, J.C. Lee
Study supervision: J.K. Rho, C.-M. Choi, H.-J. Song, P. Salgaonkar, J.K. Kim, C.H. Kim, T.G. Bivona, J.C. Lee
Other (synthesis of small-molecule kinase inhibitors being tested in this study): P. Salgaonkar
Grant Support
This study was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF, grant 2013R1A1A2005112 to J.K. Rho and grant 2014R1A2A2A01003438 to J.C. Lee) and a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant HI15C0516). T.G. Bivona acknowledges support from NIH (NIH Director's New Innovator Award, grants DP2 CA174497 and R01 CA169338).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.