Targeting AXL and mTOR Pathway Overcomes Primary and Acquired Resistance to WEE1 Inhibition in Small-Cell Lung Cancer

Abstract

Purpose: Drugs targeting DNA repair and cell-cycle checkpoints have emerged as promising therapies for small-cell lung cancer (SCLC). Among these, the WEE1 inhibitor AZD1775 has shown clinical activity in a subset of SCLC patients, but resistance is common. Understanding primary and acquired resistance mechanisms will be critical for developing effective WEE1 inhibitor combinations.Experimental Design: AZD1775 sensitivity in SCLC cell lines was correlated with baseline expression level of 200 total or phosphorylated proteins measured by reverse-phase protein array (RPPA) to identify predictive markers of primary resistance. We further established AZD1775 acquired resistance models to identify mechanism of acquired resistance. Combination regimens were tested to overcome primary and acquired resistance to AZD1775 in in vitro and in vivo SCLC models.Results: High-throughput proteomic profiling demonstrate that SCLC models with primary resistance to AZD1775 express high levels of AXL and phosphorylated S6 and that WEE1/AXL or WEE1/mTOR inhibitor combinations overcome resistance in vitro and in vivo Furthermore, AXL, independently and via mTOR, activates the ERK pathway, leading to recruitment and activation of another G2-checkpoint protein, CHK1. AZD1775 acquired resistance models demonstrated upregulation of AXL, pS6, and MET, and resistance was overcome with the addition of AXL (TP0903), dual-AXL/MET (cabozantinib), or mTOR (RAD001) inhibitors.Conclusions: AXL promotes resistance to WEE1 inhibition via downstream mTOR signaling and resulting activation of a parallel DNA damage repair pathway, CHK1. These findings suggest rational combinations to enhance the clinical efficacy of AZD1775, which is currently in clinical trials for SCLC and other malignancies. Clin Cancer Res; 23(20); 6239-53. ©2017 AACR.

Figure 1

WEE1 is overexpressed in SCLC, and genetic knockdown of WEE1 induces DNA damage and causes apoptosis in SCLC cell lines. A, RNA sequencing analysis showing the gene expression profile of WEE1 in 68 SCLC samples and 26 normal lung tissue samples. SCLC samples had significantly higher (P < 0.0001; ANOVA) WEE1 gene expression levels than normal lung tissue, with a fold change of 4.23 (left). RPPA analysis showing protein expression of WEE1 in 63 SCLC and 114 NSCLC cell lines. SCLC cells had 1.66-fold higher WEE1 expression than NSCLC cells (P < 0.0001; right). See also Supplementary Fig. S1A. B, Quantitative reverse transcriptase PCR analysis showing the efficiency of siRNA-mediated knockdown of WEE1 in four human SCLC cell lines: H82, H1836, H526, and H865. Parental cells (no siRNA; Con) and scramble shRNA cells (SCR) were used as controls in each case. GAPDH was used as the reference in this analysis. C, Western blot analysis confirming the knockdown efficiency of WEE1, expression of phosphorylated γH2AX, and phosphorylated histone H3 (PH3) levels in total protein lysates from the parental, scramble, and WEE1-knockdown (KD) cell lines H82, H1836, H526, and H865, 24 hours after transfection. Actin was used as the loading control. See also Supplementary Fig. S1B and S1C. D, Annexin V–propidium iodide–based flow cytometry 48 hours after transfection showing significantly increased apoptosis in the WEE1 KD cell lines H82, H1836, and H526. There was no significant induction of apoptosis in WEE1 KD H865 cells. E, Cell viability in response to treatment with AZD1775 in a panel of nine human SCLC cell lines, demonstrating that SCLC in vitro models have a range of sensitivity to single-agent AZD1775; about 30% of the cell lines were resistant to the drug (red box; IC₅₀ > 100 nmol/L). The genetic profile information for each cell line is presented in Supplementary Table S1. F, Single-agent treatment with AZD1775 (100 nmol/L) for 24 hours significantly increased apoptosis in H82 and H526 but did not increase apoptosis in H865 and H1417 cells. G, Flow cytometry–based cell-cycle analysis demonstrated that treatment with AZD1775 (100 nmol/L) for 48 hours caused sub-G₁ accumulation and increased G₂–M arrest in cells showing initial sensitivity to single-agent AZD1775 (H82). However, there was no significant change between the control and AZD1775-treated cell-cycle signatures in cells showing de novo resistance to WEE1 targeting (H865). See also Supplementary Fig. S1D. In all panels, data are derived from three independent experiments conducted in triplicate (error bars indicate SEM). The P values were calculated using the Student t test.^**, P < 0.01;^***, P < 0.001.

Figure 2

WEE1 inhibition is synergistic with temozolomide (TMZ) independent of MGMT expression or initial sensitivity to WEE1 targeting in SCLC cell lines. A, Dose–response curve of human SCLC cell lines (n = 10) treated with temozolomide (red), AZD1775 (green), and AZD1775+ temozolomide (black), from a 5-day CellTiter Glo assay. Doses started at 3.1 μmol/L and were serially diluted at 3-fold dilutions for five dose ranges. See also Supplementary Fig. S2A. B, Difference in area under the curve (ΔAUC) analysis demonstrating the degree of synergy of AZD1775+ temozolomide in human SCLC cell lines (n = 10). A negative value of ΔAUC suggests a more than additive effect (synergistic) and a positive value of ΔAUC suggest a less than additive effect (antagonistic). The bars are color-coded according the expression of MGMT in these cell lines, demonstrating that the combination is synergistic in all cell lines tested irrespective of MGMT expression. C, Two de novo AZD1775-sensitive cell lines (H82 and H1836) and two de novo AZD1775-resistant cell lines (H865 and H1930) were treated with single-agent AZD1775 (100 nmol/L) or siRNA WEE1 knockdown (KD) alone or combined with temozolomide (1 μmol/L) for 24 hours and apoptosis was assessed by Annexin–propidium iodide FACS assay. Data are derived from three independent experiments conducted in triplicate (error bars indicate SEM). The P values were calculated using the Student t test:^**, P < 0.01;^***, P < 0.001. See also Supplementary Fig. S2B and S2C. D, Scramble (SCR), H82, H865, and H1836 cells were treated with AZD1775 (100 nmol/L) and temozolomide (1 μmol/L) for 24 hours; WEE1 KD cells were treated with just temozolomide (1 μmol/L) for 24 hours. The cells were lysed at 24 hours and subjected to Western blot analysis for phosphorylated WEE1 (pWEE1_S642), pCDC2_Y15 (WEE1 downstream effector), γH2AX (DNA damage marker), PH3 (mitotic marker), cleaved caspase-3 (apoptotic marker), and actin (loading control). E, SCR and WEE1 KD H82, H865, and H1836 cells were treated as described in D and subjected to Western blot analysis for DNA repair markers pATM_S1981, MRE11, RAD51, and E2F1. Actin was used as a loading control.

Figure 3

Increased basal expression of AXL and activated mTOR pathway is associated with primary AZD1775 resistance. A, Spearman correlation of differential expression of 195 reverse-phase protein array (RPPA) markers and sensitivity [half-maximal inhibitory concentration (IC₅₀)] to AZD1775. The top markers of sensitivity are marked with arrows and color-coded according to the P values. B, Heatmap of RPPA markers significantly correlated with AZD1775 response in human SCLC cell lines (P < 0.05). In the top index, RPPA marker expression is denoted by red (high) and blue (low); cell lines with relatively low AZD1775 IC₅₀ values are marked in green and those with relatively high IC₅₀ values are marked in red. C, Box plot of AXL and pS6_S240/244 protein expression correlated with dichotomized IC₅₀ of AZD1775, as determined by t test. See also Supplementary Fig. S3A. D, Three AZD1775-sensitive (IC₅₀ < 30 nmol/L) cell lines (H1836, H82, and H1048) and three AZD1775-resistant (IC₅₀ > 100 nmol/L) cell lines (H1417, H865, and H1930) were treated with AZD1775 (100 nmol/L) for 24 hours and assessed for AXL expression by Western blot analysis. E, Bimodality of AXL expression in SCLC patient samples [n = 110; bimodality index (BI) = 1.27]. F, Scatter plot showing the correlation of AXL with WEE1 expression in SCLC cell lines (n = 69). The P value was calculated by Spearman correlation. G, Two AZD1775-sensitive H82 and H1836) and two AZD1775-resistant (H865 and H1930) cell lines treated with AZD1775 (100 nmol/L) for 24 hours were subjected to Western blot analysis for pWEE1_S642, total and phosphorylated pCDC2_Y15, total and phosphorylated pAKT_S473, total and phosphorylated pS6_S240/244, and actin as a loading control. See also Supplementary Fig. S3B and S3C.

Figure 4

Cotargeting AXL and mTOR overcomes de novo AZD1775 resistance in SCLC in vitro and in vivo. A, Viability (5-day assay) of AZD1775-resistant cell lines (H865, H1930, and H1417) treated with different concentrations of AZD1775 with or without the AXL inhibitor TP0903 (10 nmol/L, 20 nmol/L, 40 nmol/L, and 80 nmol/L). Data are derived from three independent experiments conducted in triplicate (error bars indicate SEM). The P values were calculated using the Student t test:^**, P < 0.01;^***, P < 0.001. B, AZD1775-resistant cells (H865, H1930, and H1417) were treated with AZD1775 (100 nmol/L), TP0903 (20 nmol/L), or both for 24 hours and apoptosis was measured by Annexin-Propidium Iodide–based FACS analysis. Data are derived from three independent experiments conducted in triplicate (error bars, SEM). The P values were calculated using the Student t test:^***, P < 0.001. C, Viability (in a 5-day assay) of control (con), scramble (SCR), and AXL knockdown (AXL KD) cells after treatment with single-agent AZD1775. The P values were calculated using the Student t test:^***, P < 0.001. See also Supplementary Fig. S4A. D, H865, H1930, and H1417 cells were treated with AZD1775 (100 nmol/L) and/or TP0903 (20 nmol/L) for 24 hours and then subjected to Western blot analysis for expression of pAXL_Y702, AKT, pAKT_S473, S6, S6_S240/244, and actin (as a loading control). E, Viability (5-day assay) of AZD1775-resistant cell lines (H865, H1930, and H1417) treated with different concentrations of AZD1775 with or without the mTOR inhibitor RAD001 (0.03 μmol/L, 0.1 μmol/L, 0.3 μmol/L, and 1 μmol/L). Data are derived from three independent experiments conducted in triplicate (error bars indicate SEM). The P values were calculated using the Student t test:^***, P < 0.001. F, AZD1775-resistant cells (H865, H1930, and H1417) were treated with AZD1775 (100 nmol/L), RAD001 (100 nmol/L), or the combination for 24 hours and apoptosis was measured by Annexin–Propidium iodide–based FACS analysis. Data are derived from three independent experiments conducted in triplicate (error bars, SEM). The P values were calculated using the Student t test:^***, P < 0.001. G, Viability (in a 5-day assay) of control, SCR, and WEE1 KD cells after treatment with single-agent RAD001. The P values were calculated using the Student t test:^***, P < 0.001. See also Supplementary Fig. S4B. H, Cells were treated as described in B and then subjected to Western blot analysis for expression of p-mTOR_S2448, pWEE1_S642, AKT, pAKT_S473, S6, S6_S240/244, and actin (loading control). I, Individual tumor volume changes in H865 xenograft mice treated with vehicle, AZD1775 (60 mg/kg 5/7, every day, orally), TP0903 (50 mg/kg, 5/7, every day, orally), or combination (n = 10 per group). Significant differences between the combination and AZD1775 alone are displayed. The P values were calculated using the Student t test. J, Survival of mice treated with vehicle, AZD1775, TP0903, or the combination (n = 10 per group). The P value was established by the Mantel–Cox test. K, Individual tumor volume changes in H865 xenograft mice treated with vehicle, AZD1775 (60 mg/kg, 5/7, every day, orally), RAD001 (5 mg/kg, 2/7, every day, orally) or combination (n = 10 per group). Significant differences between the combination and AZD1775 alone are displayed. The P values were calculated using the Student t test. L, Survival of mice treated with vehicle, AZD1775, RAD001, or combination (n = 10 per group). The P value was established by the Mantel–Cox test.^***, P < 0.001. See also Supplementary Fig. S5.

Figure 5

AXL/mTOR signaling activates other DNA repair proteins via the ERK/p90RSK pathway. A, RPPA analysis corresponding to the indicated mTOR pathway, ERK/p90RSK, and DNA repair markers in AZD1775-resistant SCLC cells, which were treated with vehicle (DMSO), 100 nmol/L AZD1775, 20 nmol/L TP0903, or the combination for 24 hours in triplicate (false discovery rate < 0.05). See also Supplementary Fig. S6. B, AZD1775-resistant cells (H865, H1930, and H1417) were treated with AZD1775 (100 nmol/L), RAD001 (100 nmol/L), or the combination for 24 hours and then subjected to Western blot analysis for expression of total ERK1/2, pERK1/2_T202/204, S6_S235/236 total p90RSK, phospho-p90RSK_T359/S363, and actin (loading control). C, H865 and H1930 cells were treated with AZD1775 (100 nmol/L) and TP0903 (20 nmol/L) for 24 hours and then the nuclear and cytoplasmic lysates were subjected to Western blot analysis for expression of pCHK1_S280, actin (cytoplasmic loading control), and PCNA (nuclear loading control). D, AZD1775-resistant cells (H865, H1930, and H1417) were treated with AZD1775 (100 nmol/L), LY2606368 (30 nmol/L), or the combination for 24 hours and apoptosis was measured by Annexin–propidium iodide–based FACS analysis. Data are derived from three independent experiments conducted in triplicate (error bars indicate SEM). The P values were calculated using the Student t test:^***, P < 0.001.

Figure 6

Acquired AZD1775 resistance in SCLC models can be reversed. A, Heatmap and B, boxplots of the comparative RPPA analysis of AZD1775-sensitive (parental) cells and cells with acquired resistance to AZD1775 (AZD-Res) for H524, H1048, and H1836 SCLC cell lines. Overexpression of AXL, MET, and S6_S240/244 and abrogation of WEE1 expression was observed in the resistant models. FC indicates fold change. C–E, Cell viability assay of H524 AZD-Res, H1048 AZD-Res, and H1836 AZD-Res treated with single-agent AZD1775 (red lines), single-agent (green lines) AXL inhibitor TP0903 (C), single-agent mTOR inhibitor RAD001 (D), single-agent dual AXL/MET inhibitor cabozantinib (E) and combinations (black lines). The combinations showed superior effects compared with single agents alone. F, Schematic summarizing the proposed mechanism by which AXL drives primary and acquired resistance to AZD1775 in SCLC and possible treatment combinations to overcome this resistance. Increased basal expression of AXL and sustained activation of the mTOR pathway leads to activation of the ERK/p90RSK pathway and finally to recruitment of a parallel DNA repair mechanism by CHK1 activation. The combination of AZD1775 inhibition with AXL, mTOR, or CHK1 targeting would abrogate this resistance pathway and result in enhanced efficacy of AZD1775 in SCLC models. AXL/MET and mTOR is overexpressed in cells with acquired resistance, and targeting AXL and MET with a dual inhibitor shows enhanced response in these models. See also Supplementary Fig. S7.