ONO-7475, a Novel AXL Inhibitor, Suppresses the Adaptive Resistance to Initial EGFR-TKI Treatment in EGFR-Mutated Non–Small Cell Lung Cancer

EGFR tyrosine kinase inhibitors (EGFR-TKI) are effective for the treatment of non–small cell lung cancer (NSCLC) harboring EGFR mutations, such as the exon 19 deletion and the L858R mutation in EGFR. Of them, the first-generation EGFR-TKIs gefitinib and erlotinib and the second-generation afatinib and dacomitinib have been approved for untreated EGFR-mutated advanced NSCLC patients in countries such as the United States and Japan (1–4). However, almost all EGFR-positive NSCLC patients ultimately acquired resistance to EGFR-TKIs after approximately 1 year. The most commonly acquired resistance mechanism, known as the secondary resistant mutation, exon20 T790M mutation in EGFR, is detected in half of the EGFR-positive patients (5). The third-generation EGFR-TKI osimertinib showed promising efficacy for overcoming EGFR T790M mutation-positive NSCLC in the AURA2 and AURA3 studies (6, 7). However, an optimal therapeutic strategy using molecularly targeted agents for treating NSCLC patients without the EGFR T790M mutation after acquired resistance to initial EGFR-TKIs is not available. Current phase III clinical trials demonstrated that the treatment with osimertinib or the novel second-generation EGFR-TKI dacomitinib showed better outcomes than that of first-generation EGFR-TKIs at the first-line setting for advanced EGFR-mutated NSCLC patients (8, 9). However, the intervention with EGFR-TKI treatment for NSCLC cells with EGFR-activating mutations has facilitated tumor evolution and leads to acquired resistance to such EGFR-TKIs, resulting in the lack of an optimal therapeutic strategy with molecularly targeted therapy after the acquired resistance to osimertinib. In addition, approximately 20% of EGFR-mutated NSCLC patients show intrinsic resistance to the new generation EGFR-TKIs dacomitinib and osimertinib. Therefore, the initial therapeutic intervention plays a crucial role in the survival of NSCLC patients with EGFR mutations.

Anexelekto (AXL) is a tyrosine kinase receptor belonging to the TAM family of proteins. AXL is usually expressed in epithelial and mesenchymal cells, as well as in breast, pancreatic, lung, and bone marrow cancers (10–14). AXL in malignant tumors plays a pivotal role in contributing to proliferation, migration, survival, and the epithelial-to-mesenchymal transition (EMT); its overexpression is consequently correlated with poor prognosis in several cancers (15–19). The activation of AXL signaling in tumors is associated with acquired resistance to several targeted molecular therapy drugs and chemotherapeutic agents (15, 20–22). Preclinical studies showed that the Gas6–AXL axis induced acquired resistance to the EGFR-TKI erlotinib or osimertinib in EGFR-mutated NSCLC cells, and that the addition of the AXL inhibitor in combination could overcome this resistance (20–22). In melanoma harboring NRAS and BRAF mutations, AXL inhibition has a pivotal role in overcoming intrinsic or acquired resistance to BRAF/MEK inhibitors (23). We previously revealed the role of the AXL pathway in the intrinsic resistance to osimertinib and the emergence of osimertinib-tolerant cells in EGFR-mutated NSCLC cells (24). Hence, the regulation of the AXL pathway is expected to be a promising therapeutic strategy against malignant progress in several cancers.

Several AXL inhibitors are being developed in clinical trials. Of them, multitargeted kinase inhibitors, such as cabozantinib, crizotinib, and sunitinib, are approved for advanced cancers, whereas bosutinib is applied for refractory tumors that are resistant to other RTK inhibitors. Recent clinical studies show that AXL inhibitors are promising combination partners (25). Of them, ONO-7475 is a novel inhibitor of the TAM receptor tyrosine kinase family that has been shown to inhibit the phosphorylation of AXL and Mer, and to suppress the growth of acute myeloid leukemia with FLT3 mutations (26, 27) and solid tumors (28, 29). A phase I clinical trial of ONO-7475 is currently under way for acute leukemia and advanced or metastatic solid tumors, although the results remain unpublished (ClinicalTrials.gov identifier NCT03176277 and NCT03730337).

In this study, we investigated the best therapeutic strategy to combat AXL-induced tolerance to EGFR-TKIs using the novel AXL inhibitor ONO-7475 in combination with osimertinib or dacomitinib in AXL-overexpressing EGFR-mutated NSCLC cells.

Ten human NSCLC cell lines with mutations in EGFR were utilized. As previously described (24), HCC4011 and H3255 were generously provided by Dr. David P. Carbone (Ohio State University Comprehensive Cancer Center, Columbus, OH) and Dr. John D. Minna (University of Texas Southwestern Medical Center, Dallas, TX), respectively. The H1975 human lung adenocarcinoma cell line with the EGFR-L858R/T790M double mutation was kindly provided by Dr. Yoshitaka Sekido (Aichi Cancer Center Research Institute, Japan) and Dr. John D. Minna. The human cell lines HCC827 and HCC4006 were purchased from the American Type Culture Collection, and the PC-9 cell line was obtained from RIKEN Cell Bank. The PC-9KGR cells, which contain deletions in EGFR exon 19 and the T790M mutation, were developed from PC-9 cells by stepwise exposure to gefitinib with limiting dilution analysis (30). The PC-9GXR cells, which contain deletions in EGFR exon 19 and the T790M mutation, were established at Kanazawa University (Kanazawa, Japan) from PC-9 cell xenograft tumors in nude mice that had acquired resistance to gefitinib. The PC-9AR1 cells and HCC827 OR1 cells, which both contain deletions in EGFR exon 19, were developed from PC-9 cells or HCC827 cells, respectively, by stepwise exposure to afatinib or osimertinib, respectively, using the limiting dilution method. All cell lines were maintained in Roswell Park Memorial Institute (RPMI) 1640 medium (GIBCO) supplemented with 10% fetal bovine serum (FBS), penicillin (100 U/mL), and streptomycin (50 g/mL) in a humidified CO₂ incubator at 37°C. All cells were passaged for less than 3 months before being renewed with frozen, early-passage stocks. Cells were regularly screened for Mycoplasma using a MycoAlert Mycoplasma Detection Kit (Lonza). Cell lines were authenticated by DNA fingerprinting. Patient-derived xenograft (PDX) from a 44-year-old woman with EGFR L858R-positive lung cancer (TM0199) was purchased from The Jackson Laboratory. Female NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG) mice ages 6–8 weeks were implanted with tumor fragments at passage 7 at The Jackson Laboratory. The mice were transferred to the animal facility at Kanazawa University and the tumors were resected. The study protocol was approved by the Ethics Committee on the Use of Laboratory Animals and the Advanced Science Research Center, Kanazawa University, Kanazawa, Japan. Osimertinib and dacomitinib were obtained from Selleckchem. ONO-7475 was supplied by Ono Pharmaceutical Co., Ltd.

The IC₅₀ value of ONO-7475 against recombinant human AXL was determined using the Off-chip Mobility Shift Assay (MSA; Carna Biosciences, Inc.), according to the manufacturer's instructions. The ATP concentration in the assay was set at the Km value of AXL for ATP (32 μmol/L). The IC₅₀ value was calculated from the concentration versus % inhibition curve by fitting a four-parameter logistic curve to the data.

For cell-based AXL inhibition, IC₅₀ value of ONO-7475 was determined using the ACD cell–based tyrosine kinase assay (Advanced Cellular Dynamics, Inc.). Recombinant human AXL-expressing mouse Ba/F3 cells, which depend on AXL kinase activity for survival and proliferation, were exposed to ONO-7475. After 48 hours, viable cells were analyzed for ATP concentration using CellTiter-Glo Luminescent Cell Viability Assay. IC₅₀ value was determined using nonlinear regression analysis.

Inhibitory activity of 100 nmol/L ONO-7475 against 60 tyrosine kinases was assessed using the ACD cell–based tyrosine kinase assay (Advanced Cellular Dynamics, Inc.). For the 52 kinases showing < 50% inhibition, the ratio of IC₅₀ for each kinase to IC₅₀ for AXL was considered to be >100. For the remaining 8 kinases (Mer, TYRO3, TrkB, EphB2, PDGFRA, TRKa, TRKc, and FLT3), further studies were conducted at various concentrations of ONO-7475, and IC₅₀ values were determined. The ratio of IC₅₀ for each kinase to IC₅₀ for AXL (0.7 nmol/L) was calculated.

Protein aliquots of 25 μg each were resolved by SDS polyacrylamide gel electrophoresis (Bio-Rad), as previously described (24, 31). Electrophoresed protein samples were transferred to polyvinylidene difluoride membranes (Bio-Rad). After washing three times, the membranes were incubated with blotting-grade blocker (Bio-Rad) for 1 hour at room temperature and overnight at 4 °C with primary antibodies to p-AXL (Tyr702), t-AXL, p-EGFR (Tyr1068), p-AKT (Ser473), t-AKT, p-p70S6K (Thr389), t-p70S6K, cleaved PARP, β-actin (13E5) (1:1,000 dilution; Cell Signaling Technology), p-Erk1/2 (Thr202/Tyr204), t-Erk1/2, and t-EGFR (1:1,000 dilution, R&D Systems).

After washing three times, the membranes were incubated for 1 hour at room temperature with HRP-conjugated species-specific secondary antibody. Immunoreactive bands were visualized using SuperSignal West Dura Extended Duration Substrate Enhanced Chemiluminescent Substrate (Pierce Biotechnology). Each experiment was independently performed at least three times.

Cell viability was determined using the MTT dye reduction method, as previously described (24, 31). Briefly, tumor cells (2–3 × 10³ cells/100 μL/well) in RPMI-1640 medium supplemented with 10% FBS were plated in 96-well plates and cultured with the indicated compound for 72 hours. After culturing, 50 μg of MTT solution (2 mg/mL, Sigma) was added to each well. Plates were incubated for 2 hours, the medium was removed, and the dark blue crystals in each well were dissolved in 100 μL of dimethyl sulfoxide (DMSO). Absorbance was measured with a microplate reader at a test wavelength of 550 nm and a reference wavelength of 630 nm. The percentage of growth was determined relative to untreated controls. Experiments were repeated at least three times with triplicate samples. As another cell viability assay, cells were treated with DMSO, EGFR-TKI osimertinib or dacomitinib, ONO-7475, or a combination for 7 and 15 days where the drugs were replenished every 72 hours. The plates were stained with crystal violet and visually examined. A plate representative of three independent experiments is shown.

For the wound-healing assay, the PC-9 and HCC4011 cells (1 × 10⁵ cells) were seeded and incubated for 24 hours at 37°C. After achieving confluence, the cellular layer in each plate was scratched using a plastic pipette tip. The migration of the cells at the edge of the scratch was analyzed at 0 and 24 hours after treatment with the EGFR-TKIs osimertinib or dacomitinib, ONO-7475, or a combination, when microscopic images of the cells were captured.

Duplexed Silencer Select siRNAs against AXL (s1845 and s1846; Invitrogen) were transfected into cells using Lipofectamine RNAiMAX (Invitrogen), according to the manufacturer's instructions. In all experiments, Silencer Select Negative Control no.1 siRNA (Invitrogen) was used as the scrambled control, as previously described (24). Knockdown of AXL was confirmed by Western blotting. Each sample was tested in triplicate in three independent assays.

pWPXL plasmids expressing empty vector (Vector) and human AXL (AXL-WT) were purchased from Addgene, Inc. (32). X-tremeGene HP DNA Transfection Reagent (Roche) was used to transfect HCC827 cells with Vector or AXL-WT plasmid, following the manufacturer's instructions. Whole-cell lysates (25 μg) were analyzed by Western blot.

Cells (2 × 10⁵) were cultured in RPMI1640 medium with 10% FBS for 24 hours, washed with PBS, and incubated for 48 hours in 2 mL of the same medium. The culture medium was harvested and centrifuged, and the supernatant was stored at −70°C until analysis, as previously described (24, 31). Level of Gas6 was determined with Human Gas6 DuoSet ELISA kit (R&D Systems), according to the manufacturer's protocols. All culture supernatants were tested twice. Color intensity was measured at 450 nm using a spectrophotometric plate reader. Concentrations of growth factors were determined from standard curves.

Suspensions of 5 × 10⁶ cells were injected subcutaneously into the flanks of 5-week-old male mice with severe combined immunodeficiency (SCID) obtained from Clea Japan, as previously described (24). Once the mean tumor volume reached approximately 100–200 mm³, 5 or 6 mice each were injected with the PC-9 and PC-9KGR cell line–derived xenografts (CDX) at the time of initial treatment or at the time of sequential therapy. Drugs were administered 7 days a week by oral gavage and the body weight and general condition of the mice were monitored daily. Tumors were measured twice weekly using calipers, and their volumes were calculated as width² × length/2. Approval was obtained from the institutional review board at University Hospital, Kyoto Prefectural University of Medicine for a study using mice (approval no. M29-529). According to institutional guidelines, surgery was performed after the animals were anesthetized with sodium pentobarbital, and efforts were made to minimize animal suffering.

Genomic DNA was extracted from frozen samples using a nucleospin tissue kit (Macherey-Nagel) according to the manufacturer's instructions. Concentrations of the extracted DNAs were measured using a Qubit 2.0 fluorometer (Thermo Fisher Scientific). The extracted DNAs were stored at −20°C until use.

The PNA-LNA PCR clamp reaction was carried out using the LightCycler 480 II (Roche) as previously reported (33). Quantitative PCR was then carried out with a 30-second hold at 95°C followed by 45 cycles at 95°C for 3 seconds, and 62°C for 30 seconds.

Specimens of EGFR-positive tumors were obtained from 7 lung adenocarcinoma patients hospitalized at University Hospital, Kyoto Prefectural University of Medicine (Kyoto, Japan) prior to osimertinib treatment. Rebiopsy specimens of tumors containing EGFR-activating mutations with the T790M mutation, after acquired resistance to initial treatments with the EGFR-TKIs gefitinib, erlotinib, afatinib and prior to treatment with osimertinib, were obtained from 28 lung adenocarcinoma patients hospitalized at the Kanazawa University Hospital (Kanazawa, Japan), Japanese Red Cross Kyoto Daiichi Hospital (Kyoto, Japan), Niigata University Hospital (Niigata, Japan), Nagasaki University Hospital (Nagasaki, Japan), or Japanese Red Cross Nagasaki Genbaku Hospital (Nagasaki, Japan). The study was conducted according to the Declaration of Helsinki. All patients participated in the Institutional Review Board of University Hospital, Kyoto Prefectural University of Medicine (approval no. ERB-C-812 and ERB-C-1349-2) and each hospital-approved study, and provided written informed consent.

Formalin-fixed, paraffin-embedded tissue sections (4-μm-thick) were deparaffinized. Antigen was retrieved by microwaving the tissue sections in 10 mmol/L citrate buffer (pH 6.0). Proliferating cells were detected by incubating the tissue sections with Ki-67 antibody (Clone MIB-1; DAKO Corp), as previously described (24). Cell apoptosis was quantitated using the terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick-end labeling (TUNEL) method, according to the manufacturer's instructions. Based on expression patterns, tumor cells in tissue specimens were separately evaluated for expression of AXL using an anti-AXL antibody (1:200; goat polyclonal, R&D Systems). Because immunohistochemical studies have shown that AXL is present primarily in the cytoplasm of cells and that its staining varies in intensity, we quantified its expression as negative (0), weak (1+), moderate (2+), and strong (3+) compared with vascular endothelial cells as an internal control (34). After incubation of the specimens with the secondary antibody and treatment using the Vectastain ABC Kit (Vector Laboratories), peroxidase activity was visualized using 3,3′-diaminobenzidine (DAB) as a chromogen. The sections were counterstained with hematoxylin.

The five areas containing the highest numbers of positively stained cells within each section were selected for histologic quantitation using light microscopy at a 400-fold magnification, as previously described (24, 31).

Data from the MTT assays and tumor progression of xenografts were expressed as means ± standard deviation (SD) and as means ± standard error, respectively. The statistical significance of differences was analyzed using one-way ANOVA and Spearman rank correlations. Progression-free survival (PFS) and 95% confidence intervals (CI) were determined using the Kaplan–Meier method and compared using the log-rank test. Hazard ratios (HR) of clinical variables for PFS were determined using a univariate Cox proportional hazards model. All statistical analyses were performed using GraphPad Prism Ver. 7.0 (GraphPad Software, Inc.), with a two-sided P-value less than 0.05 being considered statistically significant.

We first tested 10 EGFR-mutated NSCLC cell lines and one PDX tumor, which were divided into those with high levels of AXL expression and those with low levels of AXL expression. Phosphorylated EGFR tended to be higher in cells with lower relative AXL expression levels, as we previously described (ref. 24; Fig. 1A). The chemical structure of ONO-7475 is shown in Fig. 1B. IC₅₀ values of ONO-7475 against recombinant human AXL were 0.414 and 0.7 nmol/L using Off-chip MSA and ACD cell–based tyrosine kinase assay, respectively (Supplementary Fig. S1A). Ratio of IC₅₀ for each kinase to that for AXL (0.7 nmol/L) was calculated (Supplementary Fig. S1B).

We examined the potency of the AXL inhibitor ONO-7475 on the sensitivity of EGFR-mutated NSCLC cells to the EGFR-TKIs osimertinib or dacomitinib. Two different doses with ONO-7475 for 72 hours increased the sensitivity to osimertinib and dacomitinib and reduced viability of high AXL-expressing PC-9 and HCC4011 cells, but not of low-AXL-expressing HCC827 cells (Fig. 1C; Supplementary Fig. S2). In addition, ONO-7475 enhanced osimertinib efficacy on the viability of cell lines PC-9, PC-9KGR, HCC4011, and H1975, all of which express high levels of AXL, but had only a marginal effect on the viability of cell lines HCC827, HCC4006, and H3255, which express low levels of AXL. Similar results were observed in the treatment with dacomitinib (Fig. 1D). In addition, the continuous cotreatment of EGFR-TKI osimertinib or dacomitinib and ONO-7475 in PC-9 and HCC4011 cells reduced the cell viability for 15 days of incubation, but did not affect viability of low-AXL–expressing HCC827 cells (Fig. 1E). To prove further evidence that AXL is an important mediator of the therapeutic action of osimertinib on EGFR-mutated NSCLC cells, we surmised that enforced expression of AXL in low-AXL–expressing cells might suppress their sensitivity to osimertinib. To test this hypothesis, we transfected HCC827 cells with Vector or AXL-WT. Western blot showed an increase in total AXL protein in the AXL-WT–expressing cells (Supplementary Fig. S3A), which reduced their sensitivity to osimertinib, as expected. Moreover, ONO-7475 in combination with osimertinib restored viability of AXL-WT–transfected cells (Supplementary Fig. S3B).

To elucidate the underlying mechanisms of the combined therapy of ONO-7475 and the EGFR-TKIs osimertinib or dacomitinib, we investigated protein expression by Western blot. The combination of osimertinib and ONO-7475 for 4 hours markedly inhibited the phosphorylation of AXL, AKT, and p70S6K compared with treatment of the high-AXL–expressing cell lines treated with osimertinib alone (Fig. 1F). The combined use of ONO-7475 with osimertinib for 48 hours increased cleaved PARP in PC-9 and HCC4011 cells compared with the treatment with osimertinib alone. These findings were also observed in the dacomitinib treatment. Moreover, the cotreatment of ONO-7475 with osimertinib on PC-9 and HCC4011 cells inhibited cell migration for 24-hour incubation, compared with osimertinib alone (Supplementary Fig. S4).

These results showed that cell sensitivity to the EGFR-TKIs osimertinib or dacomitinib is enhanced by the combined treatment with AXL inhibitor ONO-7475, resulting in reduced viability and migration, and induced apoptosis of high-AXL–expressing EGFR-mutated NSCLC cell lines, but not with cells expressing low levels of AXL.

Drug-tolerant (DT) cells are reported as a small subpopulation of cells with remarkably reduced sensitivity to targeted drugs. They are generated within several days to several weeks of exposure to target drugs (35). Thus, we isolated DT cells from PC-9, HCC4011, PC-9KGR, and H1975 cells after 9 days of exposure with high doses of osimertinib (3 μmol/L) or dacomitinib (1 μmol/L). As expected, these DT cells were resistant to the EGFR-TKIs osimertinib or dacomitinib, compared with their parental cells (Fig. 2A; Supplementary Fig. S5A). We next examined the expression level of AXL, and its ligand Gas6, in both parental cells and DT cells. The DT cells from osimertinib and dacomitinib expressed high level of total AXL compared with parental cells in high-AXL–expressing cells, but not low-AXL–expressing cells (Fig. 2B; Supplementary Fig. S6). Production of Gas6 could not be detected in culture medium of either parental cells or osimertinib-DT cells isolated from PC-9 and HCC4011 cells, indicating that AXL activation in DT cells is independent of ligand stimulation (Supplementary Table 1). ONO-7475 treatment decreased the viability of DT cells, but not that of the parental PC-9, HCC4011, PC-9KGR, or H1975 cells (Fig. 2C; Supplementary Figs. S5B and S7A). In addition, knockdown of AXL decreased the viability of DT cells compared with scrambled control (Supplementary Fig. S7B and S7C).

Western blot analysis showed that although osimertinib and dacomitinib slightly inhibited the phosphorylation of EGFR in the DT cells, ONO-7475 suppressed the phosphorylation of AKT, p70S6K, ERK, as well as EGFR (Fig. 2D). Moreover, the continuous treatment of PC-9DT and HCC4011DT cells with ONO-7475 inhibited the viability of these DT cells for 7 and 15 days of incubation (Fig. 2E and F).

These results indicated that activated AXL played a critical role in the emergence and the maintenance of DT cells by treatment with the EGFR-TKIs osimertinib or dacomitinib, and that ONO-7475 could suppress the formation of these DT cells though AKT and ERK signals.

We evaluated the effect of ONO-7475 and osimertinib in a CDX model using high-AXL–expressing PC-9KGR cells, which has an exon 19 deletion and the exon21-T790M mutation in EGFR. Mice were continuously administered osimertinib alone, ONO-7475 alone, or a combination of the 2 drugs 7 days a week by oral gavage until day 29. Treatment with osimertinib alone caused tumor regression within 1 week, but the tumors reappeared within 3 weeks, indicating that recurrence was induced by acquired resistance. In contrast, treatment with ONO-7475 alone had little effect on the tumor growth. The combined initial treatment with osimertinib and ONO-7475 caused tumor regression compared with osimertinib alone, and the size of tumors was maintained for 4 weeks (Fig. 3A). No apparent adverse events, including weight loss, were observed during these treatments (Supplementary Fig. S8A).

To further elucidate the best therapeutic strategy of ONO-7475 plus osimertinib, we next evaluated the effect of the combination in a CDX model of PC-9 cells, using two different treatment schedules: (i) during the initial phase and (ii) during the acquired resistance phase with osimertinib. For the initial phase studies, mice were treated as above with PC-9KGR cells (Fig. 3B). The results were similar, except that the combined initial treatment with osimertinib and ONO-7475 caused tumor regression within 1 week, and the size of the regressed tumors was maintained for 7 weeks. The number of Ki-67–positive proliferating tumor cells was significantly lower in the combination-treated tumors than in osimertinib-treated tumors derived from PC-9 cells, whereas the number of TUNEL-positive tumor cells was not significantly different between the two (Fig. 3C and D, Supplementary Fig. S9A and S9B). In the PC-9 cell–derived tumors, expressions of phosphorylated AKT and p70S6K were inhibited by a combination of ONO-7475 with osimertinib, compared with osimertinib alone (Fig. 3E).

These results indicated that combined treatment with osimertinib and ONO-7475 during the initial phase prevented the growth of high-AXL–expressing EGFR-mutated NSCLC cells with or without the EGFR-T790M mutation in vivo. For the acquired resistance phase studies, the PC-9 tumor–bearing mice were initially treated with osimertinib alone. The size of tumors gradually increased, despite the continued osimertinib treatment, indicating that the tumor cells had acquired resistance to osimertinib. On day 29, the mice were randomly divided into two groups. One group was continuously treated with osimertinib alone and the other was additionally treated by a combination with ONO-7475. The tumors treated with osimertinib alone grew persistently until day 57. Tumor growth when treated with the combination of osimertinib with ONO-7475 from the acquired resistance phase was slightly slower than those treated with osimertinib alone until day 47. From days 47 to 57, tumor growth was similar to osimertinib alone (Fig. 3B). No apparent adverse events, including weight loss, were observed during these treatments (Supplementary Fig. S8B). Overall, the results from the rapid recurrence model using PC-9 cells indicated that the therapeutic intervention of combined treatment with osimertinib and ONO-7475 prevents more tumor regrowth at the initial phase than at the osimertinib-resistant phase, indicating the crucial role of initial therapeutic intervention for high-AXL–expressing EGFR-mutated NSCLC cells.

We harvested subcutaneous tumors from mice on day 29 after osimertinib treatment and cultured them in vitro. The expanded tumor cells were named PC-9OR1 (Fig. 4A). We first examined the EGFR mutation status of both PC-9 and PC-9OR1 cells. Both the PC-9 and PC-9OR1 cells possessed deletions in the EGFR exon 19 gene; however, the C797S EGFR mutation reported as the osimertinib resistance-inducing mutation, was not detected in both cell lines (Supplementary Table S2). Although the growth of PC-9OR1 cells was slightly faster than that of PC-9 cells, both cell lines grew at a constant rate in vitro (Fig. 4B). To evaluate the role of AXL activation between the initial and acquired resistance phases of osimertinib treatment, we assessed PC-9OR1 cells compared with PC-9 cells in subsequent experiments. The expression of AXL and phosphorylation of AKT, but not phosphorylation of EGFR and ERK, increased in PC-9OR1 cells compared with PC-9 cells by Western blot (Fig. 4C). Next, we examined the drug sensitivity of PC-9OR1 cells in vitro. As expected, PC-9OR1 cells were resistant to osimertinib compared with PC-9 cells (Fig. 4D; IC₅₀ PC-9OR1, 1424 nmol/L; IC₅₀ PC-9, 2 nmol/L). Importantly, consistent with in vivo experiments, PC-9OR1 cells were also less sensitive to osimertinib in combination with ONO-7475 than PC-9 cells, even though PC-9OR1 cells have higher AXL expression than PC-9 cells (Fig. 4E). Western blot analysis showed that osimertinib inhibited phosphorylation of EGFR and ERK in both cell lines, whereas the EGFR downstream molecule AKT was activated in PC-9OR1 cells, but not in PC-9 cells. These findings suggest that osimertinib failed to inhibit AKT phosphorylation in PC-9OR1 cells, resulting in acquired osimertinib resistance. Moreover, the combination of ONO-7475 and osimertinib did not completely inhibit AKT phosphorylation in PC-9OR1 cells compared with PC-9 cells (Fig. 4F).

To characterize DT cells from the acquired resistance phase, we tested the growth of PC-9OR1 DT cells selected by osimertinib treatment. PC-9OR1 DT cells were less sensitive to ONO-7475 compared with the PC-9 DT cells, suggesting that DT cells from the osimertinib-resistant phase reduced AXL-dependent cell viability (Fig. 4G).

These results showed that the survival of osimertinib-acquired resistant cells was independent of the AXL pathway. Cells could not recover sensitivity to osimertinib by treatment with a combination of an AXL inhibitor, suggesting the progression of drug tolerance with clonal evolution during osimertinib treatment.

To assess the role of AXL expression in osimertinib-treated tumors, we performed a retrospective study in patients with TKI-naïve EGFR-mutant NSCLC (N = 7). Pretreatment expression of AXL in the cytoplasm of tumor cells was evaluated using IHC staining, and scored as high (3+), intermediate (2+), low (1+), or negative (0). High AXL protein expression was detected in 1 (14.3%), intermediate in 0 (0%), low in 3 (42.9%), and negative in 3 (42.9%) of the 7 tumors. We next assessed whether AXL expression in tumors could serve as a negative predictor for treatment with osimertinib. The response to osimertinib was high (83.3%) and low (0%) in patients with AXL expression scores of 0 to 1+ and 3+, respectively. Additionally, the average tumor shrinkage rate relative to baseline was high (53.3%) and low (−19%) in osimertinib-treated patients with AXL expression scores of 0 to 1+ and 3+, respectively (Fig. 5A and B).

To assess the role of AXL expression in tumors after acquired resistance to EGFR-TKIs with osimertinib treatment, we next performed a retrospective study in patients with EGFR-mutant tumors containing the T790M mutation (N = 28). Time to progression with osimertinib ranged from 0.3 to 28 months. Among 28 tumors obtained from 28 patients with the EGFR-T790M mutation, high AXL protein expression was detected in 2 (7.1%), intermediate in 7 (25.0%), low in 19 (67.9%), and negative in 0 (0%). The response rate to osimertinib for the patients with AXL expression scores of 0 to 1+ was low (62.3%), whereas for those patients with AXL expression scores of 2+ to 3+, the response rate to osimertinib was relatively higher (77.8%; P = 0.67). PFS with osimertinib treatment was not significantly altered in patients with low and high AXL expression (P = 0.111; Supplementary Fig. S10A and S10B).

These findings show that AXL expression in tumors may be a good predictor of osimertinib sensitivity, although the sample size was small. In contrast, after acquired resistance to initial EGFR-TKIs, AXL expression in tumors may not be a good predictor for the acquired resistance to osimertinib. It appears to be difficult to select promising populations for combined therapy with osimertinib and an AXL inhibitor in EGFR-mutated tumors obtained from the acquired resistance phase.

Acquired resistance to first- and second-generation EGFR-TKIs is caused by various mechanisms, such as gatekeeper mutations such as the EGFR-T790M second-site mutation, activation of an alternate pathway, activation of EGFR downstream signals, transformation to small cell lung cancer, and the EMT (36). Several clinical trials targeting the bypass signals related to drug resistance have been conducted in EGFR-mutated NSCLC patients with acquired resistance to initial EGFR-TKIs. Of these, several clinical trials with an AXL inhibitor have been ongoing for EGFR-mutated NSCLC patients at the acquired resistance setting (37, 38). However, previous clinical studies demonstrated that the outcomes are limited for NSCLC patients with acquired resistance to EGFR-TKIs (39–41). These observations suggest that maintenance of residual tumor cells during EGFR-TKI treatment may be accelerated by various molecular mechanisms, such as minor subpopulations with a resistance mechanism, the tumor microenvironment, and the reversible DT state (42), which ultimately leads to tumor heterogeneity and drug resistance.

Currently, DT cells are thought to play a pivotal role in the progression of tumor heterogeneity because these cells are considered to eventually enhance tumor recurrence (43). However, it is not completely understood how to prevent the development of DT cells and tumor heterogeneity, which is related to epigenetic changes in treated cells (44). Previous studies demonstrated that a small subpopulation of reversibly DT cells maintain viability through the activation of the IGF1 receptor, aurora kinase A, and GPX4 (35, 45, 46). However, a clinical study using the combination of an IGF1 receptor inhibitor plus the EGFR-TKI erlotinib in unselected patients with lung cancer failed to demonstrate clinical benefits (47), although preclinical studies showed that the combination reduced these DT cells in EGFR-mutated NSCLC (35). These preclinical and clinical observations suggest that a predictive biomarker for the emergence of DT cells needs to be developed for the clinical setting.

We previously reported that addition of an AXL inhibitor NPS1034 during the initial phase of osimertinib treatment reduced tumor growth in AXL-expressing EGFR-mutated NSCLC cells. In addition, in analysis of clinical specimens, high AXL expression in the pretreatment of EGFR tumors correlates with a poor response to initial EGFR-TKIs, including osimertinib, suggesting the high AXL expression in tumors may be a promising negative biomarker to detect the response to EGFR-TKIs (24). However, in cases of refractory EGFR-mutated tumors to initial EGFR-TKIs, a significant correlation was not observed between those with AXL expression and clinical outcomes with osimertinib. Additionally, CDX mice models and cell line–based analysis showed that the combination of ONO-7475 and osimertinib is more effective at intervention of the initial phase than at the osimertinib-acquired resistance phase in high-AXL–expressing EGFR-mutated NSCLC cells. These results indicate that in high-AXL–expressing tumors, intervention with initial EGFR-TKIs may not be sufficient to combat the resistance to initial EGFR-TKIs by inhibition of the AXL-dependent pathway. Interestingly, PC-9OR1 expressed more AXL proteins compared with the pretreatment PC-9 and their DT cells, whereas the DT cells from PC-9OR1 showed less sensitivity to ONO-7475 than that of PC-9 DT cells, indicating that the intervention of EGFR-TKIs expanded the diversity of DT mechanisms in addition to AXL activation and led to insensitivity to an AXL inhibitor. In contrast, our results showed that EGFR-mutated NSCLC cells harboring the T790M mutation are also sensitive to combined therapy of ONO-7475 and osimertinib at the initial therapeutic intervention, suggesting that the combination of ONO-7475 and osimertinib is effective for EGFR-T790M–mutated NSCLC cells, without the progression of tumor heterogeneity. Collectively, our findings support the idea that therapeutic intervention with initial EGFR-TKI leads to clonal evolution and facilitates tumor heterogeneity in EGFR-mutated NSCLC cells. Therefore, an AXL-targeted therapy to conquer DT mechanisms at the initial phase might be promising to reduce the emergence of tumor heterogeneity compared with those at the resistance phase of initial treatment in AXL-overexpressing EGFR-mutated NSCLC cells.

In summary, a novel AXL inhibitor ONO-7475 suppresses the emergence and maintenance of cells tolerant to the initial EGFR-TKI osimertinib or dacomitinib in AXL-overexpressing EGFR-mutated NSCLC cells, suggesting that the combination of ONO-7475 and osimertinib is highly potent for the initial treatment phase. Further clinical investigations are warranted for the development of novel strategies using an AXL inhibitor at the initial therapeutic setting to overcome the tolerance to the EGFR-TKIs osimertinib and dacomitinib in AXL-overexpressing EGFR-mutated NSCLC patients.

T. Yamada reports receiving commercial research grants from the Takeda Science Foundation, the Cancer Research Institute of Kanazawa University, and JSPS KAKENHI, and reports receiving other commercial research support from Ono Pharmaceutical, Chugai Pharmaceutical, and Takeda Pharmaceutical. S. Watanabe reports receiving speakers bureau honoraria from Lilly, Pfizer, Novartis Pharma, AstraZeneca, Chugai Pharma, Bristol-Myers, Boehringer Ingelheim, Ono Pharmaceuticals, Daiichi Sankyo, and Taiho Pharmaceuticals. T. Kikuchi reports receiving commercial research grants from Chugai Pharma, Boehringer Ingelheim, Eli Lilly, MSD, Taiho Pharmaceutical, Ono Pharmaceutical, and AstraZeneca, and reports receiving speakers bureau honoraria from Chugai Pharma, Boehringer Ingelheim, Eli Lilly, MSD, Bristol-Myers Squibb, Pfizer Japan, Ono Pharmaceutical, Novartis Pharma, and AstraZeneca. K. Tanaka is a paid employee of Ono Pharmaceutical. R. Kozaki is a paid employee of Ono Pharmaceutical. S. Yano reports receiving commercial research grants from Chugai and Boehringer Ingelheim and reports receiving speakers bureau honoraria from Chugai, AstraZeneca, and Boehringer Ingelheim. K. Takayama is a paid employee of Ono Pharmaceutical; reports receiving commercial research grants from Ono Pharmaceutical, Chugai-Roche, and Boehringer Ingelheim; and reports receiving speakers bureau honoraria from Ono Pharmaceutical, AstraZeneca, Chugai-Roche, Boehringer Ingelheim, and Pfizer. No potential conflicts of interest were disclosed by the other authors.

Conception and design: T. Yamada, K. Takayama

Development of methodology: N. Okura, T. Yamada, H. Taniguchi

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N. Okura, N. Nishioka, T. Yamada, H. Taniguchi, S. Watanabe, T. Kikuchi, S. Shiotsu, J. Uchino, M. Horinaka, T. Sakai, S. Yano

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N. Okura, N. Nishioka, T. Yamada, A. Yoshimura, S H. Uehara

Writing, review, and/or revision of the manuscript: N. Okura, N. Nishioka, T. Yamada

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): N. Okura, N. Nishioka, H. Taniguchi, K. Tanimura, Y. Katayama, S. Shiotsu, T. Kitazaki, A. Nishiyama, M. Iwasaku, Y. Kaneko, M. Horinaka, K. Tanaka, R. Kozaki

Study supervision: T. Yamada, K. Takayama

This study was supported by a research grant for developing innovative cancer chemotherapy from a Grant for Lung Cancer Research, founded by the Japan Lung Cancer Society (to T. Yamada), a research grant from the Takeda Science Foundation (to T. Yamada), research grants for Joint Research with the Cancer Research Institute of Kanazawa University (to T. Yamada), and grants from JSPS KAKENHI (grant number 19K08608; to T. Yamada). We appreciate the gifts of the cell lines HCC4011 and H3255, provided by Dr. David P. Carbone (The Ohio State University Comprehensive Cancer Center, Columbus, OH) and Dr. John D. Minna (University of Texas Southwestern Medical Center), and the H1975 cells kindly provided by Dr. Yoshitaka Sekido (Aichi Cancer Center Research Institute, Japan) and Dr. John D. Minna (University of Texas Southwestern Medical Center).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.