Adaptive resistance to lorlatinib via EGFR signaling in ALK-rearranged lung cancer

Abstract

Anaplastic lymphoma kinase (ALK)-tyrosine kinase inhibitors rarely elicit complete responses in patients with advanced ALK-rearranged non-small cell lung cancer (NSCLC), as a small population of tumor cells survives due to adaptive resistance. Therefore, we focused on the mechanisms underlying adaptive resistance to lorlatinib and therapeutic strategies required to overcome them. We found that epidermal growth factor receptor (EGFR) signaling was involved in the adaptive resistance to lorlatinib in ALK-rearranged NSCLC, activation of which was induced by heparin-binding EGF-like growth factor production via c-Jun activation. EGFR inhibition halted ALK-rearranged lung cancer cell proliferation by enhancing ALK inhibition-induced apoptosis via suppression of Bcl-xL. Xenograft models showed that the combination of EGFR inhibitor and lorlatinib considerably suppressed tumor regrowth following cessation of these treatments. This study provides new insights regarding tumor evolution due to EGFR signaling after lorlatinib treatment and the development of combined therapeutic strategies for ALK-rearranged lung cancer.

Fig. 1. EGFR signaling plays a pivotal role in the adaptive resistance of ALK-rearranged NSCLC cells.

a A925L and H2228 cells were visualized after crystal violet staining following 9 days of treatment with the indicated concentrations of lorlatinib. The drugs were replenished every 72 h. b Human phospho-RTK array analysis of parental cells and DT cells in A925L and H2228 cells. c, d Effect of a combination of lorlatinib (100 nmol/L) and knockdown of 56 receptor tyrosine kinases (RTKs) from the Silencer® Select human kinase siRNA library V4 on the viability of A925L cells was assessed using MTT assays. The 56 RTKs are rank-ordered from highest to lowest according to their inhibitory effect on the viability of A925L cells relative to nonspecific control siRNA. The effects of the top 10 genes indicated by a red line are further shown in d. e A925L and H2228 cells treated with nonspecific control siRNA or EGFR-specific siRNAs were incubated with or without alectinib (100 nmol/L), brigatinib (100 nmol/L), or lorlatinib (100 nmol/L) for 72 h and cell viability was detected using MTT assays. *P < 0.05 (one-way ANOVA). f Western blotting of A925L and H2228 cells treated with nonspecific control siRNA or EGFR-specific siRNA and incubated with or without lorlatinib (100 nmol/L) for 4 h. Data are represented as mean ± S.D.

Fig. 2. EGFR activation by lorlatinib treatment is induced by exogenous overexpression of HB-EGF.

a Western blotting of A925L and H2228 cells treated with lorlatinib (100 nmol/L) for indicated durations. b Scatterplot of EGFR ligand expression analyzed using microarrays in either A925L or H2228 DT cells vs. parent cells. c qPCR of heparin-binding EGF-like growth factor (HB-EGF) in A925L and H2228 parent cells and cells treated with lorlatinib (100 nmol/L) for 48 h. *P < 0.01 (unpaired t-tests). d A925L and H2228 cells with nonspecific control siRNA or HB-EGF-specific siRNAs were treated for 72 h with or without lorlatinib (100 nmol/L). Cell growth was determined using MTT assays. *P < 0.05 (two-way ANOVA). e A925L and H2228 cells with nonspecific control siRNA or EGFR-specific siRNAs were treated for 72 h with or without lorlatinib (100 nmol/L) and/or HB-EGF (50 ng/mL, 10 ng/mL, respectively). Cell growth was determined using MTT assays. *P < 0.01 (two-way ANOVA). f Western blotting of A925L and H2228 cells treated with nonspecific control siRNA or EGFR-specific siRNA and incubated with or without lorlatinib (100 nmol/L) and/or HB-EGF (50 ng/mL, 10 ng/mL, respectively) for 4 h. Data are represented as mean ± S.D.

Fig. 3. EGFR activation is induced by JNK/c-Jun axis activation via suppression of DUSP4 and DUSP6.

a Scatterplot of genes in the KEGG_ERBB_SIGNALING_PATHWAY gene set in either A925L or H2228 DT cells vs. that in parent cells. b Human phospho-kinase array analysis of parental A925L and H2228 cells, as well as cells treated with lorlatinib (100 nmol/L) for 48 h. c Western blotting of A925L and H2228 cells treated with lorlatinib (100 nmol/L) for the indicated durations. d A925L and H2228 cells were incubated with nonspecific control siRNA or JUN-specific siRNA and lysed, and the indicated proteins were detected using western blotting. e qPCR of heparin-binding EGF-like growth factor (HB-EGF) in parent A925L and H2228 cells and cells incubated with nonspecific control siRNA or JUN-specific siRNA. *P < 0.01 (unpaired t-tests). f Western blotting of A925L and H2228 cells treated with lorlatinib (100 nmol/L) for the indicated durations. g A925L and H2228 cells were incubated with nonspecific control siRNA and DUSP4 or DUSP6-specific siRNA and lysed, and the indicated proteins were detected using western blotting. h Schematic diagram showing the mechanisms of adaptive resistance, including activation of EGFR signaling via endogenous HB-EGF stimulation, the feedback loop of which was induced by the JNK/c-Jun axis, in ALK-rearranged NSCLC cells. Data are represented as mean ± S.D.

Fig. 4. EGFR inhibitor sensitized ALK-rearranged lung cancer cells expressing high levels of EGFR to lorlatinib.

a A925L, H2228, and JFCR-278 cells were incubated with lorlatinib (100 nmol/L), erlotinib (100 nmol/L), or a combination of lorlatinib and erlotinib for 72 h. Cell growth was determined using MTT assays. *P < 0.05 (one-way ANOVA). b Western blotting of A925L, H2228, and JFCR-278 cells treated with lorlatinib (100 nmol/L), erlotinib (100 nmol/L), or a combination of lorlatinib and erlotinib for 4 h. c A925L and H2228 cells were visualized using crystal violet staining following 9 days of treatment with lorlatinib (100 nmol/L), erlotinib (100 nmol/L), or combination of lorlatinib and erlotinib. d Western blotting of ALK-rearranged NSCLC parental cells. e Quantitative determination of the inhibition of cell viability of high- and low-EGFR-expressing ALK-rearranged NSCLC cells treated with lorlatinib in the presence or absence of erlotinib. Cell growth was determined using MTT assays. *P < 0.05 (paired Student’s t-test). Data are represented as mean ± S.D.

Fig. 5. Combination therapy with lorlatinib and erlotinib inhibits cell proliferation by inducing apoptosis via suppression of Bcl-xL.

a Apoptotic cell percentages of A925L and H2228 cells, which were double-stained with annexin V and propidium iodide, were detected using flow cytometry following treatment with lorlatinib (100 nmol/L), erlotinib (100 nmol/L), or a combination of lorlatinib and erlotinib for 48 h. *P < 0.05 (one-way ANOVA). b A925L and H2228 cells were incubated with lorlatinib (100 nmol/L), erlotinib (100 nmol/L), or a combination of lorlatinib and erlotinib for 96 h, and the indicated proteins were detected using western blotting. c A925L and H2228 cells treated with nonspecific control siRNA or Bcl-xL-specific siRNAs were incubated with or without lorlatinib (100 nmol/L) for 72 h. Cell growth was determined using MTT assays. *P < 0.01 (two-way ANOVA). d A925L and H2228 cells were incubated with lorlatinib (100 nmol/L), a combination of lorlatinib and buparlisib (100 nmol/L), or a combination of lorlatinib and trametinib (100 nM) for 96 h, and the indicated proteins were detected using western blotting. e Western blotting of parental cells and DT cells generated from A925L and H2228 cells. Data are represented as mean ± S.D.

Fig. 6. Combination with EGFR inhibitor and ALK-TKIs prevented the emergenceof ALK-rearranged NSCLC in vivo.

a A925L xenografts were treated with a vehicle (control), lorlatinib (5 mg/kg), erlotinib (25 mg/kg), or lorlatinib plus erlotinib (n = 6) for 15 days daily via oral gavage. Thereafter, treatment was discontinued, and tumor regrowth was evaluated on day 40. b H2228 xenografts were treated daily with a vehicle (control), lorlatinib (1.5 mg/kg), erlotinib (25 mg/kg), or lorlatinib plus erlotinib (n = 6) for 7 days via oral gavage. Thereafter, the treatment was discontinued, and tumor regrowth was evaluated on day 40. Tumor volume was measured from the start of treatment and is shown as mean ± SEM. *P < 0.01 (one-way ANOVA). Representative immunohistochemical staining images of A925L (c) and H2228 (d) xenografts with antibodies to human Ki-67 and TUNEL assays. Bar, 50 μm. A925L cell line-derived xenograft (CDX) tumors were treated daily with a vehicle (control), lorlatinib (5 mg/kg), erlotinib (25 mg/kg), or lorlatinib plus erlotinib (n = 6) for four days via oral gavage. H2228 CDX tumors were treated daily with a vehicle (control), lorlatinib (1.5 mg/kg), erlotinib (25 mg/kg), or lorlatinib plus erlotinib (n = 6) for four days via oral gavage. Proliferative and apoptotic cells were quantified by their Ki-67-positive proliferation index (percentage of Ki-67-positive cells) and TUNEL assays in A925L (e) and H2228 (f) cells. *P < 0.05 (one-way ANOVA). Data are represented as mean ± S.D. g qPCR of heparin-binding EGF-like growth factor (HB-EGF) in A925L and H2228 CDX tumors treated with a vehicle (control), lorlatinib, erlotinib, or a combination of lorlatinib and erlotinib for 4 days. *P < 0.05 (one-way ANOVA). Data are represented as mean ± S.D.