Small Molecule Inhibition of MERTK Is Efficacious in Non–Small Cell Lung Cancer Models Independent of Driver Oncogene Status

Personalized medicine—therapy against the activated molecular targets within an individual patient's tumor—has transformed the field of clinical oncology. However, few personalized therapeutics are currently available, and the majority of cancer patients still receive conventional cytotoxic regimens. In NSCLC, for example, targeted therapies against activated EGFR or ALK improve survival compared with conventional chemotherapy, but only approximately 10% of newly diagnosed patients have a tumor with an activating EGFR mutation, while just 5% have an ALK translocation (1–5). For the remaining 85% of patients, opportunities clearly exist to identify new therapeutic targets and compounds.

Our laboratory and others have investigated the role of the TAM family, consisting of the RTK's TYRO3, AXL, and MERTK, in NSCLC. AXL has been well established as a promising target in primary and therapy-resistant NSCLC (6, 7). AXL expression is detected in 48% to 93% of patient samples, correlates with an invasive phenotype in cell lines, and with lymph node status and clinical stage in patients (8, 9). In addition, AXL has been shown to mediate resistance to conventional cytotoxic chemotherapeutics, as well as to EGFR-targeted agents (8, 10, 11). MERTK is also overexpressed or ectopically expressed in a multitude of human cancers, and activates a diverse network of downstream pro-oncogenic signaling pathways (8, 12–20). In NSCLC, MERTK is overexpressed in approximately two-thirds of patient tumor samples, regardless of histology (8). In addition, activation of MERTK in NSCLC cell lines induces pro-oncogenic signaling, and genetic inhibition of MERTK using shRNA decreases this signaling, resulting in increased apoptosis, increased chemosensitivity, decreased colony formation in vitro, and inhibition of tumor formation in a murine subcutaneous xenograft model (8). Finally, a monoclonal antibody against MERTK decreases colony formation and increases chemosensitivity to carboplatin, further validating the multiple functional roles of MERTK in NSCLC (21).

As our laboratory and others have validated MERTK as a promising target in NSCLC and other cancers, we have taken the next step of developing therapies against MERTK, including a monoclonal antibody (21, 22), and small molecule TKIs (17, 23–27). We have previously published on the synthesis and structure–activity relationships of several first-generation tool compound TKIs, and here report the preclinical characterization of UNC2025, the first with sufficient pharmacokinetic properties for murine-based studies and potential clinical translation (28). UNC2025 is approximately 15-fold selective for MERTK over AXL, and MERTK is the most potently inhibited kinase, with the exception of FLT3. UNC2025 is approximately equipotent against FLT3 and MERTK. While known to play important roles in subsets of acute myeloid leukemia, where it is constitutively activated through mutation, FLT3 has not been shown to be expressed in NSCLC, and therefore effects of UNC2025 in NSCLC are likely to be MERTK dependent (29). In addition, UNC2025 is approximately 20-fold selective for MERTK over TYRO3, and has more than 700-fold selectivity for MERTK over MET, another closely related RTK to the TAM family with established roles in NSCLC. In this study, we demonstrate selectivity of UNC2025 for MERTK within the TAM family in NSCLC cells, UNC2025-mediated abrogation of MERTK-mediated pro-oncogenic downstream signaling, and functional consequences, including induction of cellular apoptosis and decreased colony formation in vitro, and inhibition of tumor growth in vivo. In addition, we demonstrate that UNC2025 is broadly efficacious in a wide variety of cell lines, independent of driver oncogene status. Specifically, UNC2025 retained potent efficacy against cell lines containing an EGFR mutation, a KRAS/NRAS mutation, an RTK fusion, or another or unknown driver oncogene. These results support further consideration of testing UNC2025 in clinical trials.

All cell lines were maintained in RPMI medium supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL) until plating for experiments. Specific culture conditions for each experiment are provided within the relevant sub-sections below. All cell line identities were confirmed by the authors by comparison of short tandem repeat profiles to publically available databases. Testing was performed at the time of acquisition of each cell line, as well as approximately annually thereafter. Validated STR profiles were not available for some of the cell lines (H157, H3122, HCC4011, SF188), but the profiles obtained did not match any other cell line. The A172, A549, G361, H226, H1299, H2009, HMCB, and U251 cell lines were purchased from the ATCC, the Colo699 cell line was purchased from the German Collection of Microorganisms and Cell Cultures (DSMZ), the SF188 cell line was obtained from the UCSF Brain Tumor Bank, and the Calu3, H157, H322, H358, H441, H460, H1650, H1975, H2126, H2228, H3122, H3255, HCC827, HCC4006, HCC4011, and LC-2/ad cell lines were obtained from Drs. John Minna and Adi Gazdar (University of Texas Southwestern Medical Center, Dallax, TX). UNC2025 and UNC1653 were synthesized as previously described (25, 27).

Cells were either lifted with 0.02% EDTA in PBS and resuspended in lysis buffer (50 mmol/L HEPES, pH 7.5, 150 mmol/L NaCl, 10 mmol/L EDTA, 10% glycerol, 1% Triton X-100, 1 mmol/L Na₃VO₄) supplemented with protease inhibitors (Complete Mini, Roche Molecular Biochemicals) or washed with PBS and scraped into lysis buffer. Total protein concentrations were determined and Western blotting was performed as previously described (8).

For Western blotting, the following antibodies were obtained from Cell Signaling Technology: pAKT (S473, Cat# 9271), AKT (Cat# 9272), pERK1/2 (T202/Y204, Cat# 9106), ERK1/2 (Cat# 9102), c-KIT(Cat# 3074), survivin (Cat# 2808), α-tubulin (Cat# 2125), and TYRO3 (Cat# 5585). Additional antibodies used for Western blotting include AXL (R&D Systems, AF154), FLT3 (Santa Cruz Biotechnology, sc-480), pMERTK (Y749, Y753, Y754, PhosphoSolutions), MERTK (Abcam ab52968), pan-TRK (Santa Cruz Biotechnology, sc-11), and pTyrosine (4G10 Platinum, Millipore 05-1050). Horseradish peroxidase (HRP)-conjugated secondary antibodies (goat anti-rabbit, Bio-Rad 170-6515; goat anti-mouse, Bio-Rad 170-6516; donkey anti-goat, Santa Cruz Biotechnology sc-2020) were used for detection of proteins using enhanced chemiluminescence.

After one hour of treatment with DMSO, UNC2025, or UNC1653 in RPMI medium supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL), cells were treated with pervanadate phosphatase inhibitor (0.12 mmol/L Na₃VO₄ in 0.002% H₂O₂ in PBS) for 5 minutes (A549, H1299, H2228) or 1 minute (Colo699), and then lysed. Lysates were incubated with antibodies against MERTK (R&D Systems, MAB8912), AXL (R&D Systems, AF154), or TYRO3 (Cell Signaling Technology, 5585) and with rec-Protein G-sepharose 4B beads (Invitrogen 10-1242) overnight. Beads were then washed twice with lysis buffer, resuspended in Laemmli buffer (62.5 mmol/L Tris–HCl, pH 6.8, 25% glycerol, 5% β-mercaptoethanol, 2% SDS, and 0.01% bromophenol blue), boiled, and resolved on polyacrylamide gels. Blots were probed for phospho-MERTK or phospho-tyrosine, stripped, and reprobed for total MERTK (Abcam, 52968), AXL (R&D Systems, AF154), or TYRO3 (Cell Signaling Technology, 5585).

Recombinant human Gas6 (rhGas6, 885-GS) was purchased from R&D Systems. Subconfluent cultures were starved in serum-free medium for 2 hours and then cultured in the presence of UNC2025 or vehicle control for an additional 2 hours before Gas6 stimulation for 10 minutes. Lysates were quantified by Bradford assay and analyzed by Western blot analysis (Fig. 2A). Alternatively, cultures were treated with UNC2025 or vehicle control for 72 hours in RPMI medium supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL), before lysis, quantification, and analysis on Western blot analysis (Fig. 2B and Supplementary Fig. S2).

A549, Colo699, H1299, and H2228 cells were treated with UNC2025, UNC1653, or DMSO vehicle control for 72 hours in RPMI medium supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL). Supernatants were collected and combined with cells after lifting with EDTA. Cells were then stained with 0.2 μmol/L YO-PRO-1-iodide (YO-PRO-1) and 1.5 μmol/L propidium iodide (PI; Invitrogen). Uptake of dyes was assessed by flow cytometry using an FC500 cytometer and CXP analysis software (Beckman Coulter).

Fifty thousand cells per well were plated in 0.35% Noble agar in RPMI, and overlaid on 0.5% Noble agar in RPMI in 6-well plates. Agar layers were overlaid with 2 mL RPMI medium supplemented with 10% FBS, penicillin (100 U/mL), and streptomycin (100 μg/mL) containing DMSO vehicle control, UNC2025, or UNC1653. This medium layer was aspirated and replaced three times per week. After 2 weeks of growth, medium was aspirated, and cells were stained with 200 μL nitrotetrazoline blue chloride (NTB; Biosynth N-8100) overnight. Colonies were then counted using a GelCount colony counter and software (Oxford Optronix).

All experiments involving animals were approved by the University of Colorado Institutional Animal Care and Use Committee (Aurora, CO). Two million H2228 cells or 5.5 million A549 cells were suspended in 100 μL PBS (with 50% Matrigel for the A549 cell line) and injected subcutaneously into the flank of athymic Nude-Foxn1^nu mice. Tumors were measured weekly with calipers, and volume was calculated using the formula Volume = πa²b/6, where “a” represents the shortest diameter measured perpendicular to the longest diameter, “b”. When tumors reached an average size of approximately 150 mm³ (Fig. 5A), 100 mm³ (Fig. 5B), or 200 mm³ (Fig. 5C), mice were randomized to groups and treated twice daily by oral gavage with saline or UNC2025 at 30 or 50 mg/kg, administered at a dose of 10 mL/kg.

Statistical analyses were performed using Prism 5 software (GraphPad Software, Inc.). An unpaired t test was used in Figs. 3B and 5A–C, a paired t test was used in Fig. 4B after normalization to vehicle-treated control cells within each replicate, and the IC₅₀ values in Fig. 4C were calculated using Segmental Linear Regression. All data are representative of at least three independent experiments.

To test the selectivity of UNC2025 within the TAM family, levels of phosphorylated MERTK, AXL, and TYRO3 were assessed in H2228, A549, Colo699, and H1299 cells after one hour of treatment with UNC2025 or UNC1653, a negative control compound that lacks activity against MERTK (ref. 27; Fig. 1A). UNC2025 potently inhibited MERTK phosphorylation at doses in the low nanomolar range, with IC₅₀ values below 50 nmol/L in all four cell lines studied (Fig. 1B). In addition, UNC2025 was selective for MERTK over the related kinases AXL and TYRO3. Phospho-AXL was expressed in the H2228, A549, and H1299 cell lines, whereas Colo699 cells did not express AXL. Phospho-TYRO3 was expressed in the H2228, H1299, and A549 cell lines, but was not reliably detected in the Colo699 cell line. Regardless of expression levels, treatment with UNC2025 at doses up to 300 nmol/L was far less potent against phospho-AXL or phospho-TYRO3 than against phospho-MERTK in any of these cell lines. Importantly, AXL and TYRO3 were two of the most potently inhibited kinases after MERTK in enzymatic assays characterizing the specificity of UNC2025 against a panel of over 300 kinases (28). Other kinases determined to have greater sensitivity to UNC2025 than AXL and TYRO by this in vitro screening assay include FLT3, c-KIT, and TRKA; however, none of these proteins were expressed in any of the four cell lines studied (Supplementary Fig. S1). All other of the 300+ kinases assayed in the selectivity profiling are predicted to be less sensitive to UNC2025 than even TYRO3, and thus MERTK appears to be the only relevant target inhibited by UNC2025 in these cell lines and at these concentrations. Although it is impossible to rule out the contribution of partial inhibition of other kinases to the effects observed with UNC2025, our extensive characterization of the translation of in vitro potency to cellular phospho-protein potency demonstrates that kinases less inhibited by UNC2025 than TYRO3 in vitro would also be expected to be less inhibited in cells, and therefore less than 50% inhibited at doses up to 300 nmol/L (25). Finally, UNC1653 did not decrease phosphorylation of MERTK at doses up to 300 nmol/L (Fig. 1C). Interestingly, total MERTK expression was increased in response to treatment with higher doses of UNC2025 in three of the four lines tested.

MERTK is known to stimulate a large network of downstream signaling proteins, most commonly through the MAPK and PI3K/AKT pathways (20). To determine whether treatment with UNC2025 abrogated these effects, cells were treated with UNC2025 or vehicle control before stimulation with the TAM-family ligand GAS6 (Fig. 2A). Decreased levels of pAKT and/or pERK1/2 were detected after UNC2025 treatment in all four cell lines studied; however, effects were cell line specific. In the H2228 cell line, UNC2025 reduced basal levels of both pAKT and pERK1/2 (lane 1 vs. lane 3). Similarly, basal pAKT levels were reduced in the A549 cell line (lane 1 vs. lane 3), whereas pERK1/2 levels were unaffected. In the Colo699 cell line, pAKT levels were unaffected; however, Gas6-induced pERK1/2 stimulation was inhibited by UNC2025 (lane 2 vs. lane 4). Basal pERK1/2 levels were undetectable in this cell line. Finally, in the H1299 cell line, both basal (lane 1 vs. lane 3) and stimulated (lane 2 vs. lane 4) pAKT and pERK1/2 were decreased by UNC2025 administration.

Unexpectedly, and in contrast with the Colo699 and H1299 cell lines, GAS6 induced a decrease of pAKT and pERK1/2 in the H2228 and A549 cell lines at the one hour time-point measured. Compensatory regulatory mechanisms are hypothesized to play a role in this effect, as has been shown to be the case in other oncogenic signaling pathways; however, specific negative regulatory pathways here remain unidentified (30). Despite these paradoxical effects of GAS6, positive effects of UNC2025 were demonstrated in all four cell lines assayed. However, long-term signaling assays were next utilized in order to identify a more conserved downstream marker of UNC2025 response. Cultures were treated with 300 nmol/L UNC2025 for 72 hours, lysed, and analyzed by Western blot analysis. Survivin, a member of the inhibitor of apoptosis protein (IAP) family, was consistently decreased after UNC2025 treatment (Fig. 2B). As survivin has been shown to be a promising drug target, we sought to determine whether survivin levels were decreased in response to UNC2025 treatment in additional solid tumor cell lines representing a variety of tumor types (31). We treated four additional NSCLC cell lines, three glioblastoma cell lines, and two melanoma cell lines with UNC2025, and examined survivin levels in whole-cell lysates. In 10 of the 13 solid tumor cell lines assessed, survivin levels were dramatically decreased. Survivin protein expression was decreased in six of the eight total NSCLC cell lines assessed, all three glioblastoma cell lines, and one of the two melanoma cell lines (Supplementary Fig. S2).

pAKT, pERK1/2, and survivin are all well validated targets in the mechanism of action of NSCLC pathogenesis, and indeed all three have therapeutics in various stages of clinical trials for lung cancer patients (32, 33). The inhibition of these targets by UNC2025 suggested mechanisms of action by which functional effects would be expected to occur, and that such functional testing was warranted.

We next hypothesized that the potent effects mediated by UNC2025 on MERTK phosphorylation and downstream oncogenic signaling would translate to a functional decrease in survival of NSCLC cells. To test this hypothesis, NSCLC cells were treated with 300 nmol/L UNC2025 or UNC1653 for 72 hours, stained with YO-PRO-1 and PI dyes, and analyzed by flow cytometry to assess apoptotic cell death. Treatment with UNC2025, but not UNC1653, significantly increased the fraction of apoptotic and dead cells in all four cell lines studied (Fig. 3). UNC2025 increased cell death from basal levels of 23.9 ± 4.2%, 13.8 ± 2.2%, 9.9 ± 1.6%, and 44.2 ± 0.1%, in the H2228, A549, Colo699, and H1299 cell lines, to 40.5 ± 3.8%, 50.3 ± 7.1%, 32.9 ± 6.2%, and 59.1 ± 3.4%, respectively (P = 0.0428, P = 0.0081, P = 0.0224, P = 0.0118). Levels of apoptosis were not significantly different from basal levels after UNC1653 treatment in any of the four NSCLC cell lines.

A soft-agar assay was utilized to determine the long-term functional effects of UNC2025 treatment on NSCLC colony formation. UNC2025, UNC1653, or vehicle control was administered three times per week. After 2 weeks of growth, colonies were stained and counted (Fig. 4A). Treatment with UNC2025, but not UNC1653, significantly reduced colony formation of all four NSCLC cell lines assayed (Fig. 4B). At a dose of 300 nmol/L UNC2025, colony formation was reduced by 99.2%, 98.1%, 98.2%, and 80.1%, in the H2228, A549, Colo699, and H1299 cell lines, respectively (P < 0.0001 for each).

Because UNC2025 robustly decreased colony formation in all four of the studied NSCLC cell lines, we sought to determine whether this trend was conserved in a larger panel of NSCLC cell lines with different driver oncogene statuses. Cell lines harboring an EGFR mutation, an RTK fusion, a KRAS or NRAS mutation, and those with another or with no known driver were assessed in the soft-agar assay. IC₅₀ values were calculated for each cell line (Fig. 4C). Driver oncogene status did not correlate with sensitivity to UNC2025, as IC₅₀'s between 50 nmol/L and 350 nmol/L were observed in cell lines of each type studied. Similarly, expression of MERTK or phospho-MERTK also did not correlate with sensitivity to UNC2025. As seen in Supplementary Fig. S3, cell lines were arranged from most to least sensitive, utilizing the IC₅₀ values from Fig. 4C, and assayed for pMERTK levels by immunoprecipitation, or total MERTK levels by Western blot analysis of whole-cell lysates. Although all cell lines express pMERTK/MERTK, levels of both variably increase and decrease from the most sensitive to least sensitive cell lines, indicating a lack of predictive value for response. This is potentially due to the activation of alternative signaling pathways present in these cell lines; however, these remain unidentified at this time.

As UNC2025 shows potent antitumor effects in vitro, and since it has the pharmacokinetic properties for in vivo administration, we determined the effects of UNC2025 in vivo in murine subcutaneous xenograft models. Mice were injected with H2228 (Fig. 5A) or A549 (Fig. 5B). NSCLC cells and tumors were established until an average size of approximately 100 cubic millimeters (A549) or 150 cubic millimeters (H2228) was reached. Mice were then treated twice daily by oral gavage with either saline vehicle or 50 mg/kg UNC2025 and tumor size was measured weekly. Tumor growth was significantly impaired in mice treated with UNC2025, as indicated by significantly smaller tumor volumes relative to mice treated with saline at every measurement point in the H2228 cell line (P = 0.0193, P = 0.0058, P = 0.0011, and P = 0.0034 after 1, 2, 3, and 4 weeks of treatment, respectively). In the A549 cell line, tumor volume between the two groups was significantly different during the final 2 weeks of measurements (P = 0.0021 and P = 0.0015, respectively). At the end of the study period in the H2228 study, tumor volume was reduced by 70.2%, from 1,183 ± 258 mm³ in saline-treated mice, to 352 ± 51 mm³ in mice that had received UNC2025. In the A549 cell line study, final tumor volume was reduced by 61.9%, from 373 ± 57 mm³ in saline-treated mice, to 143 ± 28 mm³ in UNC2025-treated mice. In addition, in a dose-reduction study, mice were again injected with A549 tumor cells, and allowed to establish until an average size of approximately 200 mm³ was reached (Fig. 5C). As before, mice were then randomized and treated with saline, 50 mg/kg UNC2025, or a lower-dose treatment arm of 30 mg/kg UNC2025. Mice were again treated twice daily orally, and tumors measured weekly. Tumor growth was nearly completely inhibited in both treatment arms of this study, and there was no significant difference between the doses of UNC2025. To further analyze the mechanism behind this effect, a pharmacokinetic study assaying concentrations of UNC2025 in the plasma and tumors of mice post-treatment was utilized. Mice harboring established A549 tumors were treated twice daily for 3 days with saline or UNC2025 at 30 mg/kg or 50 mg/kg, in order to establish a steady-state level. Plasma and tumors were harvested one or 12 hours after the final treatment, to capture peak and trough concentrations of UNC2025. Specimens were analyzed by LC-MS/MS for compound concentration (Supplementary Table S1). Importantly, trough concentrations of 161 and 50 nmol/L were found in plasma of mice treated with 50 mg/kg and 30 mg/kg UNC2025, respectively, well within the 50 to 300 nmol/L range utilized within the functional assays in this manuscript. In light of this, it is somewhat unsurprising that these doses are equally efficacious in Fig. 5C, as both maintain trough levels above 50 nmol/L; however, 30 mg/kg may be nearing the minimum effective dose, as doses lower than 50 nmol/L are relatively ineffective in in vitro assays previously shown.

MERTK was initially cloned from a B-lymphoblastoid expression library and its role in a wide variety of blood cancers has subsequently been extensively characterized (7, 34). More recently, it has become increasingly appreciated that MERTK plays important roles in solid tumors as well. In NSCLC in particular, MERTK is aberrantly expressed in approximately two-thirds of patient tumors, and its activation results in increased downstream pro-oncogenic signaling. Genetic inhibition of MERTK by shRNA in NSCLC cells inhibits this signaling network, resulting in induction of apoptosis, decreased colony formation in vitro, enhanced chemosensitivity, and decreased tumor formation in vivo in a murine model (8). MERTK was further validated as an important target in NSCLC by the finding that Mer590, a monoclonal antibody targeting MERTK, decreases pro-oncogenic signaling, increases apoptosis, increases chemosensitivity, and decreases colony formation in NSCLC cell lines (21). Importantly, two of the cell lines used in the current study, A549 and Colo699, were used extensively in the shRNA target validation and Mer590 monoclonal antibody studies, respectively. The consistency between the effects observed in these studies using three different mechanisms of MERTK inhibition provides strong support that the observed phenotypes are indeed MERTK dependent.

MERTK has been extensively validated as a potential therapeutic target not only in NSCLC, but also in a multitude of other cancers. For this reason, our group recently began pursuing the synthesis of novel small molecule tyrosine kinase inhibitors targeting MERTK. We have generated a series of novel compounds representing the first-in-class MERTK-targeted small molecule inhibitors (17, 23–28). UNC2025 is the first compound of these inhibitors with pharmacokinetic properties suitable for in vivo study and potential clinical translation. Here, we have investigated the utility of UNC2025 in NSCLC and have demonstrated that UNC2025 is selective for MERTK over-related TAM-family members AXL and TYRO3, inhibits downstream signaling through AKT, ERK1/2, and survivin, induces apoptotic cell death, decreases colony formation in vitro, and decreases tumor formation in vivo in murine xenograft models. Importantly, sensitivity to UNC2025 did not depend on driver oncogene status. Development of MERTK-selective TKIs for clinical application may therefore provide a molecularly targeted treatment option for patients without known oncogenic mutations or with a wide variety of alternative driver mutations.

One mechanism by which MERTK inhibition may mediate antitumor activity is inhibition of survivin expression. We have previously demonstrated decreased total levels of survivin in response to shRNA-mediated MERTK inhibition, and here confirm this finding with UNC2025-mediated MERTK inhibition (8). Survivin plays important roles in several critical aspects of oncogenesis, where it functions to promote tumor cell survival, proliferation, metastasis, and chemoresistance, and can also contribute to tumor angiogenesis (31). In addition, survivin is an extremely cancer-specific target, as it is expressed widely in a multitude of cancer types, but not in the vast majority of normal adult tissues. These qualities have made survivin a top target for new therapy development, and several novel therapies are currently in clinical development. However, early results have not been as impressive as hoped, and additional strategies to target survivin, including potentially via UNC2025-mediated MERTK inhibition, are needed. In addition, as MERTK-targeted therapeutics transition to the clinic, pharmacodynamic biomarkers indicating response of the tumor to treatment will become a necessity. Survivin has previously been used as a tumor biomarker due to its tumor-specific expression. We have demonstrated decreased survivin levels following treatment with UNC2025 in a wide variety of solid tumor cell lines, including the majority of NSCLC lines assayed. Decreased survivin expression is a more consistent indicator of UNC2025-mediated MERTK inhibition than changes in phospho-AKT and phospho-ERK levels, and is more easily measured than phospho-MERTK. It is conceivable that survivin levels could be determined pre- and post-UNC2025 administration as an early marker of tumor response to facilitate clinical decisions about therapeutic strategy.

In addition to MERTK, important roles for AXL have also been characterized in NSCLC (6, 8). AXL expression has been correlated with lymph node status and clinical stage in patient tumor samples and can mediate resistance to EGFR-targeted therapy (9–11). Several TKIs against AXL are in various stages clinical development (7). The data presented here suggest that continued development of MERTK-selective TKIs may allow a second TAM-family member to be targeted clinically. In addition, dual inhibition of both MERTK and AXL represents another potential therapeutic strategy and may be more efficacious than inhibition of either single TAM-family member.

UNC2025 is a novel TKI that is selective for MERTK within the TAM family, inhibits MERTK's downstream signaling network, induces apoptotic cell death and inhibits colony formation in tumor cell cultures independent of driver oncogene status, and decreases tumor growth in vivo in murine models. These promising effects mediated by UNC2025 in preclinical studies demonstrate that MERTK-targeted TKIs deserve further investigation and optimization toward clinical trials.

W. Zhang has ownership interest in a UNC2025 patent: Preparation of pyrrolopyrimidine compounds for the treatment of cancer. D. DeRyckere has ownership interest (including patents) in Meryx, Inc. X. Wang has ownership interest (including patents) in Meryx, Inc. and in a patent for UNC. S.V. Frye is the President of Meryx Inc, has ownership interest (including patents) in Meryx Inc and as inventor on UNC patent, and is a consultant/advisory board member for Meryx, Inc. H.S. Earp III is a founder of and has ownership interest (including patents) in Meryx, Inc. D.K. Graham reports receiving a commercial research grant from Merck and has ownership interest (including patents) in Meryx. No potential conflicts of interest were disclosed by the other authors.

Conception and design: C.T. Cummings, W. Zhang, K.D. Davies, X. Wang, H.S. Earp III, D.K. Graham

Development of methodology: W. Zhang, K.D. Davies, D. Zhang

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): C.T. Cummings, W. Zhang, K.D. Davies, G.D. Kirkpatrick, D. Zhang, X. Wang

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): C.T. Cummings, W. Zhang, K.D. Davies, D. Zhang, X. Wang, H.S. Earp III, D.K. Graham

Writing, review, and/or revision of the manuscript: C.T. Cummings, K.D. Davies, G.D. Kirkpatrick, D. DeRyckere, X. Wang, S.V. Frye, H.S. Earp III, D.K. Graham

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): D. Zhang, X. Wang

Study supervision: D. DeRyckere, D.K. Graham

Other (supplied UNC2025): S.V. Frye

The authors thank Christopher Porter, Karen Helm, Christine Childs, Lester Acosta, Kristina Heide, and Erin Erdahl in the University of Colorado Cancer Center Flow Cytometry Core.

This work was funded in part by a research grant from Merck & Co., Inc. D.K. Graham was supported by a grant from the NIH (RO1CA137078). K.D. Davies was supported by a grant from the NIH (SPORE in Lung Cancer, grant P50CA058187), and the Lung Cancer Colorado Fund.

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.