BMX inhibition overcomes small cell lung cancer chemoresistance by stabilizing E2F1 via ERK1/2-Cyclin D1/CDK4/6 axis

Abstract

Chemotherapy resistance remains a critical bottleneck limiting its clinical efficacy in small cell lung cancer (SCLC), with its core mechanisms and targeted intervention strategies urgently requiring breakthroughs. Our study revealed that the BMX (bone marrow tyrosine kinase on chromosome X)-E2F1 (E2F transcription factor 1) axis is a pivotal regulator of chemoresistance in SCLC. Synchronous upregulation of phosphorylated BMX (Tyr566) and E2F1 was observed in SCLC tissues and cells. Mechanistically, BMX stabilized E2F1 via the ERK1/2 (extracellular signal-regulated kinase 1/2)-Cyclin D1/CDK4/6 (cyclin-dependent kinase 4/6) signaling axis, phosphorylating E2F1 at Ser332/337 and inhibiting its degradation via the ubiquitin-proteasome pathway. Inhibition or knockdown of BMX reduced E2F1 stability, promoting its degradation and reversing chemoresistance. E2F1 knockdown decreased the expression of genes associated with cell cycle regulation, migration, invasion, and DNA repair, further sensitizing chemoresistant SCLC cells to cisplatin. We also discovered IHMT-15137, a potent and selective BMX inhibitor. In vitro studies using SCLC patient-derived cells (PDCs)/patient-derived organoids (PDOs) and chemoresistant cell lines revealed that IHMT-15137, combined with cisplatin, synergistically induced cell cycle arrest, apoptosis, and DNA damage while suppressing cell migration and invasion. In vivo xenograft models demonstrated that the combination significantly inhibited tumor growth without causing significant toxicity. Our findings reveal the molecular mechanisms of SCLC chemoresistance and suggest potential therapeutic strategies targeting the BMX-E2F1 axis to overcome this challenge.

Introduction

Small cell lung cancer (SCLC) is a highly aggressive neuroendocrine carcinoma that accounts for ~15% of all lung cancer cases, and approximately two-thirds of patients present with extensive-stage SCLC (ES-SCLC).¹ Unlike non-small cell lung cancer (NSCLC), which has undergone significant progress in targeted therapies and immunotherapies, SCLC treatment has remained largely unchanged over the past few decades. Platinum-based chemotherapy (cisplatin or carboplatin) combined with etoposide remains the standard treatment for SCLC. Although this regimen is highly effective in inducing initial remission, relapse is almost inevitable, and recurrent disease is typically resistant to subsequent therapies, resulting in a dismal 5-year survival rate of only ~7%.² Genomic analyses of SCLC have revealed frequent deletions of the tumor-suppressor genes tumor protein p53 (TP53) and retinoblastoma 1 (RB1), yet actionable genetic alterations remain elusive.¹ While recent research has identified distinct biological subtypes of SCLC on the basis of transcription factor expression profiles, current clinical treatment approaches remain uniform across all subtypes.³ Although immunotherapy has shown promise in treating various cancers, its efficacy in SCLC remains limited. The combination of immune checkpoint inhibitors (ICIs) with chemotherapy has only modestly improved survival outcomes for ES-SCLC patients, with a median overall survival (OS) of 12–15 months and a progression-free survival (PFS) of less than 6 months.^4,5 With ~250,000 new SCLC cases diagnosed and ~200,000 deaths occurring annually worldwide,⁶ there is an urgent need for more effective therapies to improve patient outcomes and quality of life.

E2F transcription factor 1 (E2F1), a pivotal transcription factor essential for cell growth, DNA repair, and differentiation,⁷ is tightly regulated in normal cells but dysregulated and hyperactive in most human cancers. The role of E2F1 in cancer is context-dependent and dichotomous.⁸ For example, E2F1 gene knockout mice develop genital tract sarcomas, lung adenocarcinomas, and lymphomas,⁹ suggesting a tumor-suppressor function. Conversely, E2F1 overexpression is associated with poor prognosis in breast, lung, colon, prostate, and bladder cancers, making it a potential prognostic biomarker. In these cancers, elevated E2F1 levels are correlated with reduced OS and disease-free survival (DFS).¹⁰ In SCLC, RB1 loss leads to E2F1 upregulation in ~95% of tissue samples, and this upregulation is closely linked to tumor invasion and metastasis, indicating that E2F1 dysregulation plays a crucial role in SCLC carcinogenesis.¹¹ Recent studies have shown that E2F1 promotes SCLC invasion and metastasis by regulating matrix metalloproteinase (MMP) expression and facilitates epithelial-mesenchymal transition (EMT) through the modulation of zinc finger E-box-binding homeobox 2 (ZEB2) gene expression.^12,13 Consequently, targeting E2F1 hyperactivity represents a promising therapeutic strategy.

Despite its therapeutic potential, drug development targeting E2F1 remains in the exploratory phase, with no E2F1-specific drugs approved for clinical use. Given that E2F1 is a transcription factor, small-molecule inhibitor development strategies have focused primarily on blocking the E2F1-DNA interaction. For example, HLM006474, a small-molecule pan-E2F inhibitor, has demonstrated tumor growth inhibitory effects by blocking E2F1-DNA binding in preclinical studies but has yet to enter clinical trials.¹⁴ Other approaches, such as E2F1-targeted protein degraders (E2F-PROTACs),¹⁵ oligonucleotide decoys,^16,17 peptides that prevent E2F1-DP dimerization,^18,19 and E2F1-specific siRNAs,²⁰ have shown promise but remain in the basic research or preclinical stages. The complexity of E2F1 functions, the selectivity between normal cells and cancer cells, and the lack of biomarkers all limit the ability of drugs that directly target E2F1 to be developed. Therefore, deciphering the upstream and downstream regulatory networks of E2F1 in cancers such as SCLC, indirectly targeting E2F1, and using combination therapies may be key to breakthroughs.

Bone marrow tyrosine kinase on chromosome X (BMX, also known as Etk) has emerged as a promising therapeutic target for SCLC. Previous studies have shown that BMX overexpression contributes to chemoresistance in multiple cancer types, including hepatocellular carcinoma,²¹ prostate cancer,^22,23 colorectal cancer,²⁴ gastric cancer,²⁵ breast cancer,²⁶ acute myeloid leukemia²⁷, and SCLC,²⁸ posing major challenges in cancer treatment. In recent years, the association between BMX and SCLC chemoresistance has gained increasing recognition. Guo’s group from Southern Medical University was among the first to report the involvement of BMX in SCLC chemoresistance,^28–32 which is consistent with findings from other research teams.³³ They reported significant BMX upregulation in chemoresistant SCLC patient tumor tissues. BMX kinase regulates the phosphoinositide 3-kinase (PI3K)/signal transducer and activator of transcription 3 (STAT3) signaling pathway, apoptosis, autophagy, and EMT, making its depletion a potential strategy to overcome chemoresistance in SCLC. However, despite the preliminary confirmation of the role of the BMX kinase in SCLC chemoresistance, the underlying molecular mechanisms and regulatory networks remain unclear, and more experimental evidence is needed to validate the efficacy of BMX small-molecule inhibitors against SCLC chemoresistance.

In the present study, we revealed concurrent aberrant BMX kinase activation and E2F1 upregulation in chemoresistant SCLC tissues. Our findings demonstrate that BMX kinase regulates E2F1 protein stability and nuclear localization through the ERK1/2-Cyclin D1/CDK4/6 axis, protecting E2F1 from ubiquitin-proteasome-mediated degradation. Dysregulated and hyperactive E2F1 promotes the expression of its target genes involved in cell proliferation, DNA repair, migration, and invasion, ultimately driving SCLC chemoresistance. To address this, we developed IHMT-15137, a potent and selective BMX inhibitor that effectively resensitized chemoresistant SCLC cells to chemotherapy by targeting the BMX-E2F1 axis. We further confirmed its therapeutic efficacy in xenograft tumor models and SCLC patient-derived cells (PDCs) and patient-derived organoids (PDOs), providing new insights into SCLC chemoresistance and offering a novel therapeutic strategy and experimental basis for overcoming this challenge by targeting the upstream BMX-E2F1 axis.

Results

Concurrent upregulation of BMX phosphorylation (Tyr566) and E2F1 is correlated with SCLC pathogenesis and chemoresistance

SCLC is characterized by its highly aggressive progression and rapid development of chemotherapeutic resistance, posing major treatment challenges. Previous investigations have reported substantial upregulation of E2F1 and BMX in SCLC tissues, underscoring their critical roles in SCLC carcinogenesis and the development of chemoresistance.^11,29,31 Our immunohistochemical analysis of tumor specimens from 100 SCLC patients and adjacent normal lung tissues from 15 lung cancer controls (the available clinical information is summarized in Supplementary Table 1) revealed pronounced aberrant activation of BMX kinase and upregulation of E2F1 (Fig. 1a, b and Supplementary Fig. 1). These results suggest that elevated levels of p-BMX (Tyr566) and E2F1 contribute to SCLC pathogenesis, potentially driving the aggressive phenotype of SCLC.

Fig. 1.

Positive correlation between E2F1 and p-BMX (Tyr566) expression in SCLC. a Representative immunohistochemical staining of p-BMX (Tyr566) and E2F1 proteins in tumor and adjacent normal lung tissues from SCLC patients. b Quantitative analysis of p-BMX (Tyr566) and E2F1 expression in immunohistochemical images of 15 adjacent normal lung tissues and 100 SCLC tumor tissues. c Correlation analysis between p-BMX (Tyr566) and E2F1 expression in 100 SCLC patients. d Western blot analysis of p-BMX (Tyr566), BMX, and E2F1 in six SCLC cell lines (H526, H889, DMS114, H196, H446, and H69) and one normal bronchial epithelial cell line (BEAS-2B). e Western blot analysis of p-BMX (Tyr566), BMX, and E2F1 in six SCLC cell lines after 4 h of treatment with chemotherapy drugs (ADM-2 μM, CDDP-20 μM, or VP16-40 μM). f Western blot analysis of p-BMX (Tyr566), BMX, and E2F1 in chemoresistant SCLC cells (H446DDPR and H69AR) and their parental cells (H446 and H69) (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

We then validated the clinical relevance of BMX and E2F1 by performing survival analyses in multiple publicly available SCLC cohorts.

In the cBioPortal Small Cell Lung Cancer (U Cologne, 2015) study (study ID: sclc_ucologne_2015) and the GEO GSE60052 cohort, higher BMX expression correlated with shorter overall survival (OS) and poorer prognosis. E2F1 was associated with an unfavorable OS trend in sclc_ucologne_2015 (not significant) but significantly shorter OS with higher E2F1 in GEO GSE281525 (Supplementary Fig. 2). Within each cohort, patients were dichotomized into high- and low-expression groups based on the median expression value of BMX or E2F1. These external cohort analyses confirmed their potential as SCLC biomarkers and therapeutic targets.

Furthermore, comparative analysis of SCLC cell lines (the characteristics and subtypes of the cell lines used are summarized in Supplementary Table 2 and Supplementary Fig. 3) against the normal bronchial epithelial cell line BEAS-2B revealed significant upregulation of p-BMX (Tyr566) and E2F1 (Fig. 1d), further supporting their potential role in SCLC pathogenesis. To explore the involvement of BMX and E2F1 in chemotherapy resistance, we analyzed their expression profiles in SCLC cells treated with standard chemotherapeutic agents, including cisplatin (CDDP), etoposide (VP16), and adriamycin (ADM). Notably, drug treatment induced the synchronous upregulation of both E2F1 protein levels and BMX phosphorylation at Tyr566 (Fig. 1e), indicating a coordinated response to chemotherapeutic stress. Additionally, in chemotherapy-resistant SCLC cell lines, we observed a marked increase in p-BMX (Tyr566) and E2F1 levels compared with their drug-sensitive counterparts (Fig. 1f). Collectively, these findings demonstrate the synchronous upregulation of E2F1 and p-BMX (Tyr566) in both chemotherapy-treated and resistant SCLC cells, suggesting a functional interplay between E2F1 and BMX kinase activity that contributes to the development of chemoresistance.

To further investigate the relationship between p-BMX (Tyr566) and E2F1, a correlation analysis was conducted on the immunohistochemical data of 100 SCLC patient samples. Statistical analysis revealed a significant positive correlation between p-BMX (Tyr566) and E2F1 protein levels (Fig. 1c). Taken together, our clinical and experimental data provide compelling evidence that the coordinated activation of the p-BMX (Tyr566)/E2F1 axis represents a novel molecular mechanism contributing to SCLC pathogenesis and chemotherapy resistance, offering potential therapeutic targets for improved treatment in SCLC patients.

BMX confers chemoresistance in SCLC by protecting E2F1 from ubiquitin-proteasome-mediated degradation

Building on previous research, our study systematically explored the role of BMX in SCLC. Lentivirus-mediated manipulation of BMX expression in H446 cells demonstrated that BMX promotes cell proliferation, migration, and invasion (Supplementary Fig. 4). In chemosensitivity studies, BMX activation was induced by chemotherapy, and BMX knockdown sensitized cells to cisplatin by lowering the GI₅₀, increasing apoptosis, and disrupting DNA repair, whereas BMX overexpression had the opposite effects (Supplementary Fig. 5 and 6). In the chemoresistant H446DDPR and H69AR cell lines (characterized in Supplementary Fig. 7), BMX signaling was hyperactive, and its knockdown significantly reversed cisplatin resistance (Supplementary Fig. 8), suggesting that BMX is a key regulator of SCLC progression and chemoresistance.

Given the synchronous upregulation of p-BMX (Tyr566) and E2F1 in SCLC tissues and cells, we investigated their functional relationship. Gene perturbation experiments revealed that the knockdown or overexpression of E2F1 did not affect the total protein or phosphorylation levels of BMX in SCLC cells (Supplementary Fig. 9). In contrast, BMX knockdown in chemoresistant cells reduced E2F1 protein levels, whereas BMX overexpression increased E2F1 protein levels (Fig. 2a–d). qRT-PCR analysis revealed that BMX did not affect E2F1 mRNA expression (Fig. 2e), and the BMX inhibitor ibrutinib similarly reduced E2F1 protein levels without affecting mRNA levels (Fig. 2f, g), indicating post-transcriptional regulation.

Fig. 2.

BMX regulates E2F1 stability through the ubiquitin-proteasome pathway in SCLC cells. a, b Validation of lentiviral-mediated BMX knockdown (shBMX#1, shBMX#2) in H446DDPR and H69AR cells via qRT-PCR and Western blotting. The protein levels of E2F1 were assessed by Western blotting following BMX knockdown. c, d Confirmation of the efficiency of BMX overexpression via lentiviral infection in H446 and H69 cells via qRT-PCR and Western blotting. The protein levels of E2F1 were also assessed by Western blotting in H446 and H69 cells after BMX overexpression. e E2F1 mRNA expression was assessed via qRT-PCR after BMX knockdown in H446DDPR and H69AR cells or BMX overexpression in H446 and H69 cells. f, g Assessment of E2F1 protein and mRNA levels in H446DDPR and H69AR cells after ibrutinib treatment (0, 0.1, 0.3, or 1 μM) for 24 h. h Half-life analysis of E2F1 in chemoresistant cells (H446DDPR and H69AR) and their parental cells (H446 and H69). i, j Half-life assessment of E2F1 after BMX knockdown in H446DDPR and H69AR cells or BMX overexpression in H446 and H69 cells. k Western blot analysis of the effect of ibrutinib (pretreatment with 1 μM, 8 h) on the half-life of E2F1 in H446DDPR and H69AR cells. The cells were treated with cycloheximide (CHX) (H446DDPR/H446: 20 μg/ml; H69AR/H69: 50 μg/ml) to inhibit protein synthesis, and protein degradation was monitored over time. The degradation curves of E2F1, quantified from the Western blot results, are shown on the right. l, m Effects of MG132 (0.5 μM, 24 h) or CQ (1 μM, 24 h) on E2F1 protein levels after BMX knockdown in H446DDPR and H69AR cells. n, o Effects of MG132 or CQ on E2F1 protein levels after ibrutinib treatment. p Poly-ubiquitination levels of E2F1 in chemoresistant SCLC cells and their parental cells. The cell lysates were immunoprecipitated with an anti-E2F1 antibody, and the ubiquitination level of E2F1 was analyzed with an anti-ubiquitin antibody. Then, the ubiquitin, E2F1, and GAPDH protein levels were detected via Western blotting. q, r Assessment of E2F1 poly-ubiquitination in H446DDPR and H69AR cells after BMX knockdown or BMX overexpression in H446 and H69 cells. Data were presented as mean ± SEM (n = 3; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

Using cycloheximide (CHX), an inhibitor of protein synthesis, we assessed E2F1 protein stability in SCLC cells. In chemoresistant cells with high BMX activity, E2F1 was notably more stable than it was in parental cells (Fig. 2h). Furthermore, BMX knockdown or inhibition substantially shortened the E2F1 half-life, whereas BMX overexpression extended it (Fig. 2i–k), suggesting that BMX stabilizes E2F1 by inhibiting its proteolytic degradation.

Next, to determine the specific degradation mechanism, we treated cells with the proteasome inhibitor MG132 or the autophagy inhibitor chloroquine (CQ). MG132 treatment, but not CQ treatment, rescued E2F1 degradation induced by BMX knockdown or ibrutinib treatment (Fig. 2l–o), indicating the involvement of the ubiquitin-proteasome pathway. Immunoprecipitation assays further revealed that the level of E2F1 ubiquitination was significantly lower in chemoresistant cells than in parental cells (Fig. 2p). Similarly, BMX knockdown significantly increased E2F1 ubiquitination, whereas BMX overexpression decreased E2F1 ubiquitination (Fig. 2q, r).

Overall, our findings establish that BMX stabilizes the E2F1 protein in chemoresistant SCLC cells by inhibiting its ubiquitin-proteasome-mediated degradation, revealing a previously unrecognized mechanism underlying E2F1-mediated chemoresistance in SCLC.

BMX stabilizes E2F1 via the ERK1/2-Cyclin D1/CDK4/6 signaling pathway to promote chemoresistance in SCLC

To elucidate the mechanism by which BMX stabilizes E2F1, we initially explored whether direct physical interaction was involved. Co-immunoprecipitation (co-IP) assays in H446 cells expressing FLAG-tagged BMX or E2F1, with or without cisplatin treatment, revealed that endogenous E2F1 did not coprecipitate with BMX-FLAG, nor did endogenous BMX coprecipitate with E2F1-FLAG (Supplementary Fig. 10). These findings indicate that BMX-mediated E2F1 stabilization occurs independent of direct binding.

E2F1 stability and function are tightly regulated by multiple posttranslational modifications, including phosphorylation, acetylation, and ubiquitination.⁸ Among these, phosphorylation at Ser332 and Ser337 is particularly pivotal, as phosphorylation at these sites by the Cyclin D1/CDK4/6 complex enhances the DNA-binding affinity, stability, and transcriptional activity of E2F1 while preventing its interaction with retinoblastoma protein (Rb).³⁴ In contrast, ubiquitination targets E2F1 for proteasomal degradation.³⁵ The balance between phosphorylation and ubiquitination is essential for maintaining the role of E2F1 in cell cycle regulation.

Notably, the ERK1/2 pathway can upregulate Cyclin D1 expression to activate CDK4/6,³⁶ which then phosphorylates Rb to release E2F1 while directly phosphorylating E2F1 at Ser332 and Ser337.^8,37 Given that ERK1/2, p38, and JNK are known downstream effectors of BMX,³⁸ we hypothesized that BMX regulates E2F1 stability via the ERK1/2-Cyclin D1/CDK4/6 axis.

To test this hypothesis, we first compared pathway activation patterns between chemoresistant and parental SCLC cells. Western blot analysis revealed elevated phosphorylation levels of BMX (Tyr566) and ERK1/2 (Thr202/Tyr204), along with increased protein levels of Cyclin D1, p-E2F1 (Ser332 and Ser337), and total E2F1, in chemoresistant cells compared with their parental counterparts (Fig. 3a). RNA-seq analysis further confirmed significant Cyclin D1 upregulation (adjusted P < 0.05, |log₂Fold Change | >1.0) in chemoresistant cells (volcano plot, Fig. 3b), as validated by qRT-PCR in cisplatin-treated H69/H69AR cells (Supplementary Fig. 11).

Fig. 3.

BMX regulates E2F1 stability via the ERK1/2-Cyclin D1/CDK4/6 axis. a Comparative Western blot analysis of p-BMX (Tyr566), BMX, p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 levels between chemoresistant SCLC cells (H446DDPR and H69AR) and their parental cells (H446 and H69). b Volcano plot of RNA-seq data showing differentially expressed genes (DEGs) between cisplatin-treated (10 μM, 48 h) H69AR and H69 cells from RNA-seq analysis. Red dots represent upregulated genes, blue dots represent downregulated genes, and gray dots represent genes that were not differentially expressed (P < 0.05, |log₂Fold Change|>1.0). c, d Effects on p-BMX (Tyr566), BMX, p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 after BMX knockdown in H446DDPR and H69AR cells or BMX overexpression in H446 and H69 cells. e Dose-response effects of ibrutinib (0, 0.1, 0.3, or 1 μM; 24 h treatment) on p-BMX (Tyr566), BMX, p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 in H446DDPR and H69AR cells. f Poly-ubiquitination assessment of E2F1 in H446DDPR and H69AR cells after ibrutinib treatment (0, 0.1,0.3,1 μM) for 24 h. g Western blot analysis of p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-Rb (Ser807/811), Rb, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 in H446DDPR and H69AR cells after treatment of ERK1/2 inhibitor ulixertinib (0, 1, 3,10 μM) for 24 h. h Poly-ubiquitination assessment of E2F1 in H446DDPR and H69AR cells after treatment of ulixertinib (0, 1, 3,10 μM) for 24 h. i Western blot analysis of p-Rb (Ser807/811), Rb, p-ERK1/2 (Thr202/Tyr204), ERK1/2, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 in H446DDPR and H69AR cells after 24-h treatment of CDK4/6 inhibitor palbociclib (0, 1, 3,10 μM). j Assessment of E2F1 poly-ubiquitination in H446DDPR and H69AR cells after 24 h of palbociclib (0, 1, 3, or 10 μM) treatment

Genetic manipulation of BMX expression in SCLC cells revealed that genetic knockdown of BMX in chemoresistant cells markedly decreased p-ERK1/2 (Thr202/Tyr204) levels, Cyclin D1 expression, and both p-E2F1 (Ser332/337) and total E2F1 protein levels (Fig. 3c). Conversely, BMX overexpression in parental cells increased ERK1/2 phosphorylation (Thr202/Tyr204) and Cyclin D1 levels and upregulated p-E2F1 (Ser332/337) and total E2F1 (Fig. 3d), demonstrating the role of BMX as an upstream regulator of the ERK1/2-Cyclin D1/CDK4/6-E2F1 signaling axis.

Pharmacological inhibition experiments further validated this mechanism, showing that treatment with the BMX inhibitor ibrutinib, the ERK1/2 inhibitor ulixertinib, or the CDK4/6 inhibitor palbociclib dose-dependently reduced E2F1 phosphorylation at both Ser332 and Ser337 (Fig. 3e, g, i). Concurrently, E2F1 ubiquitination increased (Fig. 3f, h, j), leading to accelerated E2F1 proteasomal degradation. These findings demonstrate that BMX stabilizes E2F1 through activation of the ERK1/2-Cyclin D1/CDK4/6 axis, which suppresses E2F1 ubiquitination by maintaining its phosphorylation at Ser332/337.

We further investigated the role of BMX kinase activity in SCLC chemoresistance and demonstrated that CRISPR-Cas9-mediated BMX knockout in H446DDPR cells increased cisplatin sensitivity, which was reversed by wild-type BMX (BMX WT) reconstitution but not by the kinase-dead BMX K421R mutant (Supplementary Fig. 12). Rescue experiments confirmed that BMX knockdown increased cisplatin sensitivity in H446DDPR cells, which was counteracted by E2F1 overexpression, whereas BMX overexpression increased resistance, which could be overcome by E2F1 knockdown (Fig. 4). These results collectively establish that BMX kinase activity is essential for maintaining cisplatin resistance in SCLC, primarily because of its role in stabilizing E2F1.

Fig. 4.

The BMX-E2F1 axis promotes chemoresistance in SCLC. a Validation of E2F1 overexpression (FLAG-E2F1) following BMX knockdown (shBMX#1, shBMX#2) in H446DDPR cells. b Effect on cisplatin-induced apoptosis (cleaved PARP and cleaved caspase-3) in H446DDPR cells after E2F1 overexpression following BMX knockdown. c Effects of BMX knockdown and subsequent E2F1 restoration on the cisplatin-mediated inhibition of H446DDPR cell proliferation, as assessed by an EdU incorporation assay. d Quantification of the data in (c). e The sensitivity of H446DDPR cells to cisplatin was assessed via a CTG assay after BMX knockdown and subsequent restoration of E2F1 expression. f Confirmation of E2F1 knockdown (shE2F1#1, shE2F1#2) following BMX overexpression in H446 cells via Western blotting. g Effect on cisplatin-induced apoptosis (cleaved PARP and cleaved caspase-3) in H446 cells after E2F1 knockdown following BMX overexpression. h Effects of BMX overexpression and subsequent E2F1 knockdown on the cisplatin-mediated inhibition of H446 cell proliferation, as assessed by an EdU incorporation assay. i Quantification of the data in (h). j Assessment of the sensitivity of H446 cells to cisplatin via a CTG assay after E2F1 knockdown following BMX overexpression. Data were presented as mean ± SEM (n = 3; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

BMX kinase orchestrates E2F1 nuclear translocation to promote chemoresistance in SCLC

To elucidate the functional consequences of elevated E2F1 levels induced by activated BMX kinase in SCLC chemoresistance, we performed gain- and loss-of-function experiments using lentiviral vectors. In chemoresistant H446DDPR and H69AR cells, we achieved efficient E2F1 knockdown, whereas in parental H446 and H69 cells, we established stable E2F1 overexpression, with qRT-PCR and Western blot analyses confirming successful modulation of E2F1 at both the mRNA and protein levels (Fig. 5a, b).

Fig. 5.

E2F1 modulates chemoresistance in SCLC. a, b Validation of lentiviral-mediated E2F1 knockdown in H446DDPR and H69AR cells and E2F1 overexpression in H446 and H69 cells by Western blotting and qRT-PCR, respectively. c, d Assessment of the sensitivity of SCLC cells to cisplatin via a CTG assay after E2F1 knockdown in H446DDPR and H69AR cells or E2F1 overexpression in H446 and H69 cells. e, f Anti-proliferative effect of cisplatin on H446DDPR cells after E2F1 knockdown or on H446 cells after E2F1 overexpression, as determined via a colony formation assay. The quantification data are shown on the right. g, h Effects on the cisplatin-mediated inhibition of H446DDPR and H69AR cell proliferation after E2F1 knockdown or on H446 and H69 cell proliferation after E2F1 overexpression, as determined via an EdU incorporation assay. The quantification data are shown on the right. i, j Effects of E2F1 knockdown in H446DDPR and H69AR cells or overexpression in H446 and H69 cells on cisplatin-induced DNA damage (γ-H2AX) and apoptosis (cleaved PARP and cleaved caspase-3). Data were presented as mean ± SEM ( n= 3; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

Upon cisplatin treatment, CellTiter-Glo (CTG) assays demonstrated that E2F1 knockdown restored cisplatin sensitivity in H446DDPR and H69AR cells (Fig. 5c), whereas E2F1 overexpression decreased sensitivity in H446 and H69 cells (Fig. 5d). Colony formation (Fig. 5e, f) and EdU incorporation assays (Fig. 5g, h) further validated these findings. Evaluation of apoptosis and the DNA damage response in cisplatin-treated cells revealed that E2F1 knockdown in resistant cells significantly increased apoptosis, as indicated by increased levels of cleaved PARP and cleaved caspase-3 (Fig. 5i). Conversely, E2F1 overexpression reduced apoptosis in parental cells (Fig. 5j). Additionally, E2F1 knockdown increased cisplatin-induced γ-H2AX levels (Fig. 5i), whereas E2F1 overexpression reduced γ-H2AX accumulation (Fig. 5j), indicating that E2F1 plays a role in DNA damage repair. These results collectively demonstrate that E2F1 contributes to SCLC chemoresistance by increasing cell proliferation, suppressing apoptosis, and facilitating DNA repair.

Given the transcriptional regulatory function of E2F1 in chemoresistance,⁸ we investigated its subcellular localization. Nuclear-cytoplasmic fractionation experiments revealed that chemoresistant SCLC cells with excessive activation of BMX presented significantly increased nuclear E2F1 accumulation compared with parental cells (Supplementary Fig. 13a). BMX knockdown in resistant cells decreased nuclear E2F1 levels (Supplementary Fig. 13b), whereas BMX overexpression in parental cells increased nuclear accumulation (Supplementary Fig. 13c).

To assess the functional impact of E2F1 nuclear translocation on transcriptional regulation, we performed qRT-PCR analysis. In chemoresistant SCLC cell lines, BMX or E2F1 knockdown significantly downregulated the expression of E2F1 target genes involved in cell proliferation (PCNA, CCNE1, and CCNA2), DNA repair (CHEK1 and BRCA1), and metastasis (CDH2 and VIM) (Supplementary Fig. 13d, e). Conversely, BMX or E2F1 overexpression in parental cells upregulated these genes (Supplementary Fig. 13f, g). CUT&RUN-qPCR analysis revealed that E2F1 overexpression increased its enrichment at these gene promoters in H69/H446 cells. Compared with parental H69/H446 cells, chemoresistant H69AR/H446DDPR cells also had greater E2F1 occupancy, consistent with their elevated E2F1 expression (Supplementary Fig. 14).

Taken together, our findings demonstrate that BMX-dependent E2F1 nuclear translocation activates a transcriptional program that drives chemoresistance in SCLC cells. Through the coordinated regulation of ‌cell cycle progression, DNA repair, and metastatic pathways, the BMX-E2F1 axis has emerged as a critical determinant of the chemotherapy response in SCLC, highlighting its potential for overcoming treatment resistance.

Identification of IHMT-15137 as a potent and selective BMX inhibitor

BMX kinase plays crucial roles in tumor progression, metastasis, and cancer chemoresistance, making it an attractive therapeutic target for overcoming chemoresistance in cancer treatment.³⁹ Targeting BMX with specific inhibitors has shown promise in preclinical studies as a potential strategy for treating SCLC with chemoresistance.³⁰

To identify a potent and selective BMX inhibitor, we applied a multistep screening method: virtual screening of ~6000 in-house compounds yielded 1102 candidates, which were then evaluated via high-throughput CTG assays in BMX-dependent TEL-BMX-BaF3 cells.⁴⁰ Compounds displaying comparable inhibition effects were selected for secondary validation; cytotoxic compounds were excluded via WT-BaF3 viability tests. After further characterization of the top hits, IHMT-15137 (with unsaturated acrylamide, suggesting irreversible binding; Supplementary Fig. 15) emerged as a leading candidate. To verify the predicted irreversible binding mode, we synthesized its reversible analog IHMT-15138 (acrylamide saturated with propionamide; Fig. 6a). ADP-Glo™ biochemical assays demonstrated that IHMT-15137 exhibited inhibitory activity against purified BMX kinase, with an IC₅₀ value of 26.97 ± 2.45 nM, which was slightly weaker than that of ibrutinib (3.85 ± 0.26 nM), a dual inhibitor of BTK/BMX. In contrast, IHMT-15138 showed significantly reduced inhibitory activity against BMX, with an IC₅₀ greater than 2500 nM (Fig. 6b), indicating the irreversible binding mechanism of IHMT-15137.

Fig. 6.

Characterization of IHMT-15137 as a potent and selective BMX inhibitor. a Chemical structure of IHMT-15137 and its reversible analog IHMT-15138. b Determination of the IC₅₀ values of IHMT-15137, IHMT-15138 and ibrutinib against the BMX protein via the ADP-Glo kinase assay. c Anti-proliferative activity of IHMT-15137, IHMT-15138 and ibrutinib against TEL-BaF3-BMX cells determined by the CTG assay. d CETSA analysis of the effects of IHMT-15137 on BMX protein stability in H446DDPR and H69AR cell lysates: temperature- and dose-dependent stabilization of BMX by IHMT-15137. The right panels show the results of the quantitative analysis of the BMX protein levels from the Western blots. e DARTS assay confirmation of the direct binding of IHMT-15137 to BMX by assessing the stability of BMX against pronase-mediated proteolysis in H446DDPR and H69AR cell lysates. f Inhibitory effects of IHMT-15137 and IHMT-15138 on BMX phosphorylation (Tyr566) in TEL-BaF3-BMX cells after 4 h of treatment, as determined by Western blotting. g Inhibitory effects of 12 h of IHMT-15137 treatment on BMX phosphorylation (Tyr566) and its downstream signaling pathways in H446DDPR and H69AR cells. h Molecular docking analysis of the irreversible binding mode of IHMT-15137 with BMX (PDB ID: 3SXR). i Effect of IHMT-15137 on Biotin-PEG2-C2-iodoacetamide labeling of BMX via a competitive biotin-iodoacetamide assay. Data were presented as mean ± SEM (n= 3)

Using TEL-BMX-BaF3 cells whose proliferation depends on BMX kinase activity, we evaluated the growth inhibitory activity of IHMT-15137 and IHMT-15138. IHMT-15137 displayed potent anti-proliferative activity (GI₅₀ = 0.83 ± 0.09 nM) but was less potent than ibrutinib (GI₅₀ = 0.19 ± 0.01 nM). In contrast, IHMT-15138 showed substantially diminished activity, with a GI₅₀ of 3267 ± 0.27 nM (Fig. 6c), further supporting the irreversible BMX inhibition of IHMT-15137.

A temperature- and dose-dependent cellular thermal shift assay (CETSA) demonstrated that IHMT-15137 significantly stabilized the BMX protein (Fig. 6d). These results suggest the direct binding of IHMT-15137 to BMX under the experimental conditions. Additionally, the drug affinity responsive target stability (DARTS) assay, which is based on the principle that small-molecule compounds stabilize target proteins upon binding and thereby increase their resistance to proteolysis, revealed that IHMT-15137 protected BMX from proteolysis in a dose-dependent manner in cell lysates from H446DDPR and H69AR cells after incubation with IHMT-15137 for 4 h (Fig. 6e).

Moreover, we assessed BMX phosphorylation levels in TEL-BMX-BaF3 cells after 4 h of treatment with IHMT-15137 or IHMT-15138. The results showed that IHMT-15137 effectively inhibited the phosphorylation of BMX (Tyr566) in a dose-dependent manner, whereas IHMT-15138 had no significant effect on BMX phosphorylation (Tyr566) in these cells (Fig. 6f). We then investigated the effects of IHMT-15137 treatment on BMX and its downstream signaling pathways in chemoresistant SCLC cells, H446DDPR and H69AR. Treatment with IHMT-15137 resulted in a dose-dependent decrease in the phosphorylation levels of BMX (Tyr566), ERK1/2 (Thr202/Tyr204), and STAT3 (Tyr705) in both cell lines (Fig. 6g).

Kinase selectivity profiling of IHMT-15137 was performed using DiscoverX’s KINOMEScan platform against a panel of 468 kinases at 1 μM. TreeSpot analysis demonstrated that IHMT-15137 is a highly selective inhibitor with particularly strong binding to BMX kinase (Supplementary Fig. 16 and Supplementary Table 3). Since KINOMEScan is a binding-based assay, the selectivity of IHMT-15137 was further evaluated via a panel of kinase-dependent BaF3 isogenic cell lines. Notably, IHMT-15137 exhibited selective and potent anti-proliferative activity specifically against TEL-BMX-BaF3 cells (GI₅₀ = 0.83 ± 0.09 nM) but significantly reduced activity against both parental BaF3 cells (GI₅₀ = 6189 ± 228 nM) and other isogenic BaF3 cell lines (e.g., BLK-, EGFR-, and JAK3-dependent; GI₅₀ > 2000 nM). In contrast, ibrutinib demonstrated potent activity against multiple kinases, including TEL-BMX-BaF3 (GI₅₀ = 0.19 ± 0.01 nM), TEL-BLK-BaF3 (GI₅₀ = 0.34 ± 0.04 nM), and TEL-EGFR-BaF3 (GI₅₀ = 1.01 ± 0.03 nM) (Supplementary Fig. 17a–c). To further validate selectivity, we compared their inhibitory potency against BTK, a primary target of ibrutinib, via Western blot analysis. In the BTK-activated monocytic leukemia cell line MOLM-13, ibrutinib had a potent inhibitory effect on BTK phosphorylation (Tyr223) (EC₅₀ < 1 nM), whereas the activity of IHMT-15137 was much weaker (EC₅₀ = 23.85 nM) (Supplementary Fig. 17d).

A subsequent CTG assay revealed that IHMT-15137 had negligible cytotoxicity toward normal human lung epithelial BEAS-2B cells, whereas ibrutinib induced significant cell death at concentrations above 3 µM (Supplementary Fig. 17e). This effect is potentially attributable to ibrutinib’s broader kinase inhibition profile.

Molecular docking analysis of IHMT-15137 into the BMX crystal structure (PDB: 3SXR) (Fig. 6h) revealed that the compound occupies the ATP-binding pocket. The pyrazole core of IHMT-15137 forms three hydrogen bonds with Ile492 and Glu490 in the hinge region of BMX, whereas the adjacent amide group forms an additional hydrogen bond with Ser553. The trifluoromethyl and acrylamide groups extend toward the solvent-exposed region, adopting a U-shaped conformation that facilitates stable covalent bond formation with Cys496. Competitive biotin-iodoacetamide assays further demonstrated that IHMT-15137 dose-dependently inhibited probe labeling of BMX, supporting the covalent engagement of Cys496 (Fig. 6i).

In summary, IHMT-15137 is a potent and selective BMX inhibitor that is proposed to covalently bind BMX Cys496, making it a potential pharmacological tool for subsequent mechanistic studies and therapeutic development.

Targeting the BMX-E2F1 axis with IHMT-15137 overcomes chemoresistance in SCLC cells in vitro

Chemotherapy resistance remains a formidable challenge in SCLC treatment. Our previous studies have revealed that the BMX-E2F1 signaling axis is critically involved in regulating the proliferation, migration, invasion, and chemotherapeutic resistance of SCLC cells. On the basis of these findings, we propose that targeting BMX is a promising therapeutic strategy to combat SCLC progression.

To test this hypothesis, we treated chemoresistant H446DDPR and H69AR cells with IHMT-15137, and CTG assays revealed a dose-dependent cytotoxic effect of IHMT-15137 on both cell lines (Supplementary Fig. 18a). The results of the proliferation, colony formation, and EdU incorporation assays collectively revealed that IHMT-15137 significantly suppressed cell growth (Supplementary Fig. 18b–d). Importantly, flow cytometric analysis revealed that IHMT-15137 induced G2-phase cell cycle arrest (Supplementary Fig. 18e). Western blot analysis revealed that 24 hours of IHMT-15137 treatment significantly increased the levels of cleaved PARP and cleaved caspase-3, confirming the induction of apoptosis. Moreover, treatment with IHMT-15137 led to dose-dependent increases in γ-H2AX levels, suggesting impaired DNA repair (Supplementary Fig. 18f). Wound healing, transwell migration and invasion assays demonstrated that IHMT-15137 potently inhibited cell migration and invasion (Supplementary Fig. 18g, h).

We next evaluated whether IHMT-15137 could overcome chemoresistance in SCLC. Compared with single-agent treatment or control treatment, the combination of cisplatin with IHMT-15137 significantly inhibited the viability and proliferation of chemoresistant cells, as evidenced by cell survival, growth curve, colony formation, and EdU incorporation assays (Supplementary Fig. 19a–d). Mechanistically, cell cycle analysis revealed that the combination treatment increased the proportion of cells arrested in the G2/M phase while reducing the G0/G1 phase cell population (Supplementary Fig. 19e). Western blot analysis further revealed enhanced apoptosis, as evidenced by elevated levels of cleaved PARP and cleaved caspase-3 and exacerbated DNA damage, as indicated by higher γ-H2AX levels than those in response to cisplatin alone (Supplementary Fig. 19f). Importantly, the combination treatment also more effectively inhibited both cell migration and invasion (Supplementary Fig. 19g, h).

To validate the role of the BMX-ERK1/2-Cyclin D1/CDK4/6-E2F1 axis in SCLC chemoresistance, we performed a cell cycle-related antibody microarray analysis. Our results revealed that Cyclin D1, CDK4/6, and E2F1 protein levels were markedly elevated in both cisplatin-treated and chemoresistant H446DDPR cells, whereas IHMT-15137 treatment effectively reversed these alterations (Supplementary Fig. 19i). Further mechanistic studies demonstrated that BMX inhibition by IHMT-15137 decreased E2F1 protein levels without affecting mRNA expression, reduced E2F1 phosphorylation at Ser332 and Ser337, increased E2F1 ubiquitination, and eventually promoted E2F1 degradation in chemoresistant SCLC cells (Supplementary Fig. 20).

To confirm the on-target effects of IHMT-15137, we generated BMX-knockout H446DDPR cells and reconstituted them with either wild-type BMX or a kinase-dead mutant. CTG assays revealed that IHMT-15137 enhanced cisplatin sensitivity in cells with wild-type BMX but not in either BMX-knockout cells or cells expressing the kinase-dead mutant BMX K421R (Supplementary Fig. 21a–d). This BMX-dependent activity was further corroborated by Western blot analysis, which revealed that IHMT-15137 potentiated cisplatin-induced apoptosis only in cells expressing functional BMX (Supplementary Fig. 21e).

To clarify the hierarchy of the BMX-ERK1/2-Cyclin D1/CDK4/6-E2F1 cascade, we conducted time-course and rescue experiments. IHMT-15137 (1 μM) reduced ordered signaling in resistant cells: early loss of p-BMX (Tyr566) was followed by decreased p-ERK1/2 (Thr202/Tyr204), a subsequent decrease in Cyclin D1, and finally reduced p-E2F1 (Ser332/337) and total E2F1. Furthermore, BMX knockdown or IHMT-15137 treatment reduced p-ERK1/2 (Thr202/Tyr204), Cyclin D1, and E2F1 phosphorylation (Ser332/337), whereas forced E2F1 overexpression restored E2F1 levels and phosphorylation without reversing the inhibition of p-ERK1/2 (Thr202/Tyr204), or Cyclin D1, confirming that E2F1 acts as a downstream effector (Supplementary Fig. 22).

In summary, our findings demonstrate that IHMT-15137 effectively overcomes chemoresistance in SCLC through multiple mechanisms, including suppressing cell proliferation, inducing apoptosis, inhibiting metastasis, and impairing DNA repair. These effects are mediated through disruption of the BMX-ERK1/2-Cyclin D1/CDK4/6-E2F1 signaling axis, highlighting the BMX-E2F1 axis as a promising therapeutic target for SCLC treatment.

Targeting BMX with IHMT-15137 overcomes chemoresistance in SCLC

To quantify the synergy between IHMT-15137 and cisplatin, we used a dose-response matrix assay in H446DDPR/H69AR cells via the CTG method and analyzed the data with SynergyFinder 3.0. Consistently low combination index (CI) values (<1) confirmed a synergistic, rather than additive, interaction between the two agents (Supplementary Fig. 23). To translate preclinical findings into relevant clinical outcomes, we evaluated the efficacy of IHMT-15137, a selective BMX inhibitor, using PDCs (C23084 and LUC22009) from two SCLC patients (Supplementary Table 4). These cells were isolated after PD following first-line etoposide/platinum chemotherapy and thus represent a model of multidrug resistance.

When combined with cisplatin, IHMT-15137 synergistically enhanced cytotoxicity, as demonstrated by a significant reduction in cell viability in the CTG assay (Fig. 7a). Moreover, flow cytometric analysis revealed significant cell cycle arrest in the S phase (Fig. 7b). Furthermore, growth curve, colony formation, and EdU incorporation assays demonstrated profound suppression of tumor cell proliferation, with remarkable inhibition of colony formation and markedly decreased DNA synthesis (Fig. 7c–e). In addition, Western blot analysis revealed enhanced apoptosis (cleaved PARP and cleaved caspase-3) and significantly increased DNA damage (γ-H2AX) in the treated groups compared with the monotherapies (Fig. 7f).

Fig. 7.

IHMT-15137 overcomes cisplatin resistance in SCLC PDC models. a CTG assay showing effects on cisplatin-mediated proliferation inhibition in two SCLC PDC models (C23084 and LUC22009) after treatment with DMSO or IHMT-15137 (3 μM). b Cell cycle distribution analysis of C23084 and LUC22009 cells treated with DMSO, cisplatin (5 μM), IHMT-15137 (3 μM), or a combination of both. c Cell proliferation curves of C23084 and LUC22009 cells treated with DMSO, cisplatin (5 μM), IHMT-15137 (1 μM), or a combination of both. d An EdU incorporation assay was used to evaluate the anti-proliferative effects of DMSO, cisplatin (5 μM), IHMT-15137 (3 μM) or their combination for 24 h in C23084 and LUC22009 cells. The quantification of the results is shown on the right. e Anti-proliferative ability of DMSO, cisplatin (3 μM), IHMT-15137 (3 μM), or their combination on C23084 and LUC22009 cells was detected via a colony formation assay for 14 days. The quantification of the results is shown on the right. f Western blot analysis of DNA damage (γ-H2AX) and apoptosis (cleaved PARP and cleaved caspase-3) in C23084 and LUC22009 cells treated with DMSO, cisplatin (30 μM), IHMT-15137 (10 μM), or their combination for 24 h. g Western blot analysis of p-BMX (Tyr566), BMX, p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 in C23084 and LUC22009 cells treated with DMSO, CDDP (30 μM), IHMT-15137 (3 μM), or their combination for 24 h. h Immunoprecipitation assay analyzing E2F1 ubiquitination levels after 24-h treatment with DMSO, CDDP (30 μM), and IHMT-15137 (3 μM) or their combination in C23084 and LUC22009 cells. Data were presented as mean ± SEM (n = 3; ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

We further investigated the mechanism by which BMX regulates E2F1 in these primary cells. Cisplatin treatment alone increased the phosphorylation of BMX (Tyr566), ERK1/2 (Thr202/Tyr204), and E2F1 (Ser332/Ser337), along with the upregulation of Cyclin D1 and E2F1. However, cotreatment with IHMT-15137 effectively reversed these changes (Fig. 7g). Moreover, ubiquitination assays revealed that the combination of cisplatin and IHMT-15137 significantly increased E2F1 ubiquitination, accelerating E2F1 degradation (Fig. 7h). Overall, these results indicate that IHMT-15137 potentiates the effects of cisplatin by modulating the ERK1/2-Cyclin D1/CDK4/6 signaling axis and promoting the degradation of E2F1.

Platinum-etoposide remains the standard therapy for SCLC. In chemoresistant SCLC cells, the triple combination (CDDP + VP16 + IHMT-15137) reduced proliferation (EdU) and viability (CTG) compared with either CDDP-VP16 alone or IHMT-15137 monotherapy and elevated the levels of cleaved PARP and cleaved caspase-3 (apoptosis; Supplementary Fig. 24), confirming that IHMT-15137 potentiates the effect of CDDP-VP16 in vitro.

Before proceeding to in vivo studies, we evaluated the pharmacokinetics of IHMT-15137 in mice after intravenous (i.v.), intraperitoneal (i.p.), and oral (p.o.) administration. The compound showed negligible oral absorption; following intraperitoneal injection (10 mg/kg), an acceptable AUC_0‑t (1014 ± 85 ng·h/mL) and bioavailability (F = 81.6%) were achieved. Therefore, the i.p. route was selected for the in vivo experiments (Supplementary Table 5).

We then validated the therapeutic potential of IHMT-15137 in vivo via chemoresistant SCLC xenograft models (H69AR and H446DDPR). The combination of CDDP + VP16 with IHMT-15137 resulted in significant tumor growth inhibition (TGI) (60.1% in H69AR; 69.0% in H446DDPR) compared with the monotherapies (Fig. 8b–d, h–j), without causing significant body weight loss (Fig. 8a, g). Furthermore, Western blot analysis of tumor tissues revealed that CDDP + VP16 increased p-BMX (Tyr566), p-ERK1/2 (Thr202/Tyr204), Cyclin D1, and p-E2F1 (Ser332/337) levels, which were suppressed by IHMT-15137 cotreatment (Fig. 8e, k). Compared with vehicle, IHMT-15137 alone induced only a modest increase in the levels of the apoptosis markers cleaved PARP and cleaved caspase-3, whereas the combination of IHMT-15137 with cisplatin+VP16 led to a marked increase in the level of apoptosis compared with that resulting from chemotherapy alone (Supplementary Fig. 25). Additionally, immunohistochemistry (IHC) confirmed enhanced apoptosis (TUNEL) and proliferation inhibition (Ki-67) with the combination therapy (Fig. 8f, l).

Fig. 8.

IHMT-15137 overcomes chemoresistance in SCLC xenograft models. Female NCG mice bearing established H69AR or H446DDPR tumor xenografts were treated with vehicle, IHMT-15137, CDDP + VP16, or their combination via intraperitoneally (i.p.) injection. a, g Body weight changes in the four groups, with the initial body weight set to 100%. b, h Tumor size changes in each group during the 16-day treatment period. c, i Comparison of final tumor weights in H69AR and H446DDPR xenograft mice after treatment. d, j Representative images of tumors from each group after treatment. e, k Western blot analysis demonstrated the effects of vehicle, CDDP + VP16, IHMT-15137, and combination treatment on p-BMX (Tyr566), BMX, p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-E2F1 (Ser332), p-E2F1 (Ser337), and E2F1 expression in tumor tissues following the 16-day treatment period. f, l Histopathological analysis of H&E, Ki-67, and TUNEL staining of tumor tissue sections. The quantification of the results is shown on the right. Data were presented as mean ± SEM (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

Although IHMT-15137 retains BTK inhibitory activity, our data revealed that both total BTK and p-BTK (Tyr223) are essentially undetectable in H69/H69AR and H446/H446DDPR cells, suggesting that BTK is unlikely to contribute to the observed phenotypes in our SCLC models (Supplementary Fig. 26). Given that BTK inhibition is associated with increased bleeding risk,^41,42 we utilized a tail bleeding assay to investigate the effects of IHMT-15137 on the hemostatic system. The compound had a negligible effect on bleeding time, suggesting an acceptable safety profile for long-term use. In contrast, ibrutinib significantly prolonged the bleeding time under identical experimental conditions. A 21-day subacute toxicity study demonstrated that IHMT-15137 did not affect major organs, such as the heart, liver, spleen, lungs, or kidneys. We also evaluated its effects on blood biochemistry indices and hematological parameters. Overall, our results suggested that IHMT-15137 displays low toxicity and a relatively promising safety profile (Supplementary Fig. 27 and Supplementary Table 6).

Finally, we evaluated the antitumor efficacy in both the SCLC PDO and PDCX models. We obtained primary SCLC patient samples and successfully established four independent PDO models. The clinicopathologic characteristics of the patients whose SCLC samples were donated are listed in Supplementary Table 4. Compared with either agent alone, the combination of IHMT-15137 with CDDP resulted in a more pronounced reduction in organoid size and viability (Fig. 9a–c). We also established a PDCX model using SCLC PDC cells (C23084). Consistent with our previous data in H69AR and H446DDPR xenograft models, the combination of CDDP + VP16 with IHMT-15137 significantly inhibited tumor growth (59.8%) compared with that of monotherapies, without significant body weight loss (Fig. 9d–g). Western blot analysis of tumor tissues revealed that IHMT-15137 suppressed the CDDP + VP16-induced increases in p-BMX (Tyr566), p-ERK1/2 (Thr202/Tyr204), Cyclin D1, and p-E2F1 (Ser332/337), and the combination therapy increased apoptosis (Fig. 9h). Additionally, IHC confirmed that combination therapy enhanced apoptosis (TUNEL) and inhibited proliferation (Ki-67) (Fig. 9i).

Fig. 9.

The combination of IHMT-15137 with cisplatin has synergistic antitumor effects on SCLC PDOs and PDCX models. a Representative brightfield images of four PDO models following 5 days of treatment with vehicle, IHMT-15137, cisplatin, or the combination. b Quantification of the PDO diameters corresponding to the treatment groups described in (a). c CTG assay evaluating the viability of the four PDO models after 5 days of treatment with vehicle, IHMT-15137, cisplatin, or the combination. d–i Female NCG mice bearing established SCLC PDCXs were treated with vehicle, IHMT-15137, CDDP + VP16, or their combination via intraperitoneally (i.p.) injection. d Body weight changes in the four groups, with the initial body weight set as 100%. e Tumor growth curves during the 28-day treatment. f Representative photographs of excised tumors from each group after treatment. g Comparison of final tumor weights. h Western blot analysis of tumor lysates for p-BMX (Tyr566), BMX, p-ERK1/2 (Thr202/Tyr204), ERK1/2, Cyclin D1, p-E2F1 (Ser332), p-E2F1 (Ser337), E2F1, PARP, caspase-3, and cleaved caspase-3. i Histopathological analysis of H&E, Ki-67, and TUNEL staining of tumor tissue sections. Data were presented as mean ± SEM (ns, P > 0.05; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001)

Collectively, these findings from clinically relevant primary cells and in vivo models demonstrate that IHMT-15137 effectively overcomes chemoresistance in SCLC. Through its multifaceted mechanism of suppressing cell proliferation, inducing apoptosis, impairing DNA repair, and modulating the BMX-E2F1 signaling axis, IHMT-15137 represents a promising therapeutic strategy for improving treatment outcomes in SCLC patients.

Discussion

This study reveals the pivotal role of the BMX-E2F1 axis in SCLC chemoresistance and demonstrates the therapeutic potential of targeting this axis with the selective BMX inhibitor IHMT-15137. Our findings elucidate the molecular mechanisms underlying SCLC chemoresistance and offer novel therapeutic strategies for this aggressive malignancy.

The synchronous upregulation of p-BMX (Tyr566) and E2F1 in SCLC tissues, cells, and in response to chemotherapy aligns with emerging evidence highlighting the importance of transcription factor-kinase interactions in cancer progression.⁴³ While prior studies reported high BMX protein expression in SCLC patient tissues,^29,31 our research provides novel insights into the functional state of BMX by specifically characterizing its phosphorylation status. Additionally, by introducing inactivating mutations into the BMX kinase domain, we demonstrated the essential role of BMX kinase activity in SCLC chemoresistance, offering a more in-depth mechanistic understanding that extends beyond protein expression analysis. Our findings reveal that the BMX-E2F1 axis is a critical regulator of SCLC chemoresistance. BMX stabilizes E2F1 through the ERK1/2-Cyclin D1/CDK4/6 signaling cascade, thereby preventing E2F1 degradation via the ubiquitin-proteasome pathway. Once stabilized, E2F1 activates a transcriptional program that simultaneously promotes cell proliferation, suppresses apoptosis, enhances DNA repair, and drives metastasis (Fig. 10). These findings establish that targeting upstream regulators to precisely modulate E2F1 represents a promising therapeutic strategy for overcoming chemoresistance in SCLC.

Fig. 10.

Schematic model of how BMX kinase inhibition overcomes the chemoresistance of SCLC through regulating E2F1 stability via the ERK1/2-Cyclin D1/CDK4/6 axis. BMX is significantly upregulated in chemotherapeutic and chemoresistant SCLC cells. This activation induces downstream phosphorylation of ERK1/2 (Thr202/Tyr204), which subsequently increases cyclin D1 expression and enhances the activity of the cyclin D1/CDK4/6 complex, which then phosphorylates E2F1 at Ser332 and Ser337, stabilizing E2F1 by blocking its degradation via the ubiquitin-proteasome system. Phosphorylation of E2F1 at these sites facilitates its nuclear import and enhances its transcriptional activity, thereby promoting the expression of genes involved in cell cycle regulation, DNA damage repair, migration, and invasion, including CCNA2, CCNE1, PCNA, BRCA1, CHEK1, VIM, and CDH2. These molecular changes collectively promote tumor cell proliferation, migration, invasion, and resistance to apoptosis, ultimately driving chemotherapy resistance. The selective BMX inhibitor IHMT-15137 reverses this resistance mechanism by destabilizing E2F1, inhibiting downstream resistance mechanisms, and restoring cisplatin sensitivity in chemoresistant SCLC cells. This model highlights the potential of targeting the BMX-E2F1 signaling axis to overcome chemotherapy resistance in SCLC. This image was created with Microsoft PowerPoint

In this study, E2F1 is regulated by upstream kinases, including BMX, ERK1/2, and CDK4/6, and inhibitors targeting these kinases have been shown to be effective in overcoming chemoresistance in SCLC cell lines. Notably, a key distinction exists among these kinases: while ERK1/2 and CDK4/6 are expressed in both cancerous and normal cells,^44,45 BMX kinase is expressed at relatively low levels in most normal tissues.⁴⁶ This unique expression pattern suggests that the development of a selective BMX kinase inhibitor could offer a wider therapeutic window than targeting more broadly expressed kinases. By selectively inhibiting BMX, IHMT-15137 disrupts the BMX-E2F1 axis at its upstream node. Preclinical studies using multiple SCLC PDC and PDO models, together with a PDCX model, have demonstrated that IHMT-15137 synergizes with chemotherapy, producing enhanced antitumor effects through multiple mechanisms, including inducing cell cycle arrest, promoting apoptosis, and impairing DNA repair, leading to a substantial reduction in tumor growth. Our findings establish the BMX-E2F1 axis as a viable therapeutic target for SCLC.

To clarify BMX activation in SCLC, we first analyzed BMX expression. BMX transcripts were upregulated in a subset of SCLC cell lines and were further elevated in the chemoresistant H69AR and H446DDPR cells, consistent with their higher BMX protein levels. Mechanistically, BMX activation is tightly linked to PI3K signaling.⁴⁷ Treatment with the PI3K inhibitor GSK1059615 (but not the AKT inhibitor MK2206) partially reduced p-BMX (Tyr566) and E2F1 levels. We also observed increased Src activation in chemoresistant cells, which correlated with increased p-BMX (Tyr566), supporting a PI3K/Src-driven mechanism for BMX upregulation/activation in SCLC, especially in chemoresistant states (Supplementary Fig. 28). To explore how BMX activates ERK1/2, we tested for a direct BMX-ERK1/2 interaction via co-IP in 293T cells expressing FLAG-BMX. No coprecipitation of endogenous ERK1/2 was detected (Supplementary Fig. 29), ruling out direct binding. As previously reported,^38,47 BMX likely acts as an upstream regulator of ERK1/2 through indirect cascades (e.g., PLCγ2-PKC and adapter-TAK1), driving ERK1/2 phosphorylation.

However, several limitations must be acknowledged. First, while our study focused primarily on preclinical models, translation to clinical applications requires further investigation. Second, the long-term effects of targeting the BMX-E2F1 axis, including potential compensatory mechanisms and chemoresistance development, remain to be fully elucidated. Third, while IHMT-15137 shows high selectivity against BMX, its potential off-target effects in complex biological systems require further investigation. Fourth, the poor oral bioavailability and limited blood-brain barrier permeability of the compound, as observed in preliminary pharmacokinetic (PK) studies (Supplementary Tables 5 and 7), may limit its clinical application, especially in SCLC patients with brain metastases, necessitating further chemical optimization to obtain orally bioavailable and blood-brain barrier-permeable BMX inhibitors.

Additionally, little is known about how BMX might affect other modifications of E2F1, such as methylation and acetylation,^7,48 which are also critical for E2F1 activity. More studies are needed to clarify whether BMX influences these modifications and their role in maintaining E2F1 stability in chemoresistant SCLC cells. Clinically, the positive correlation between p-BMX (Tyr566) and E2F1 expression in SCLC tissues suggests that these markers could serve as potential biomarkers for chemotherapy resistance and poor prognosis. However, further clinical validation is needed to determine their potential for patient stratification and predictive value for the response to chemotherapy. Another critical aspect is the potential crosstalk between the BMX-E2F1 axis and other signaling pathways in SCLC. Given that SCLC is characterized by a complex network of dysregulated signaling,⁴⁹ the BMX-E2F1 axis may interact with pathways such as the PI3K/AKT and STAT3 pathways. Preliminary data from our chemoresistant SCLC models revealed that PI3K/Src signaling drives BMX activation, whereas the canonical PI3K/AKT cascade is dispensable for the BMX-E2F1 axis (Supplementary Fig. 28). Notably, long-term BMX inhibition in H69AR/H446DDPR xenografts did not induce compensatory STAT3 or AKT activation in chemoresistant tumor tissues (Supplementary Fig. 30). A detailed understanding of these interactions is critical for developing effective therapies, and future work should focus on delineating the relevant pathways.

Although preclinical studies suggest that BMX targeting (e.g., via ibrutinib, a dual BTK/BMX inhibitor) may overcome chemoresistance in SCLC, ibrutinib has never been evaluated in SCLC-specific trials. Its development has focused on hematological malignancies (consistent with FDA approvals for CLL and MCL), and even its limited solid-tumor trials (e.g., NCT02321540 and NCT02403271) have focused primarily on NSCLC, with stalled progress. Key shortcomings explain this outcome and underscore the need for new BMX inhibitors: (1) Ibrutinib potently inhibits BTK but only moderately inhibits BMX; its off-target effects (e.g., on ITK and EGFR) induce toxicities that prevent dose escalation for sustained BMX suppression; (2) Suboptimal drug exposure in SCLC lesions and a lack of predictive biomarkers have led to the enrollment of unselected patients, diminishing treatment efficacy. Consequently, next-generation BMX inhibitors—with increased selectivity, improved tumor penetration, and the ability to serve as companion biomarkers—are essential for effectively targeting BMX-driven SCLC and overcoming resistance.

In conclusion, our study established the BMX-E2F1 axis as a pivotal player in SCLC chemoresistance. The discovery of IHMT-15137 provides a promising starting point for targeting this axis. Addressing the study’s limitations and fully characterizing the axis’s interactions with other pathways will be crucial for advancing SCLC treatment. This research lays the foundation for the development of innovative therapies that could significantly improve outcomes for SCLC patients.

Materials and methods

Cell lines

Human small cell lung cancer cell lines (H526, H889, DMS114, H196, H446, and H69), a bronchial epithelial cell line (BEAS-2B), and the AML cell line (MOLM-13) were obtained from YuChi Biology (Shanghai, China). H69AR cells were kindly provided by Professor Linlang Guo’s laboratory at ZhuJiang Hospital, Southern Medical University. H446DDPR cells were acquired through long-term cisplatin exposure via a concentration gradient method in our laboratory.

The DMS114, H196, H446, H69, H446DDPR and MOLM-13 cell lines were maintained in RPMI-1640 medium (Corning, Midland, NY, USA) supplemented with 10% FBS (VivaCell, Shanghai, China) and 1% penicillin/streptomycin (v/v). H526, H889, and H69AR cells were cultured in RPMI-1640 medium (Corning, Midland, NY, USA) supplemented with 20% FBS and 1% penicillin/streptomycin (v/v). BEAS-2B cells were cultured in BEGM (Lonza) supplemented with 10% FBS and 1% penicillin/streptomycin (v/v). Transgenic BaF3 cells were cultured in RPMI-1640 medium supplemented with 10% FBS and 1% penicillin/streptomycin (v/v).

The two SCLC PDCs (C23084 and LUC22009) and four PDOs (C23084, LUC240506, and LUC23148 and LUC240765) were provided by Hefei PreceDo Pharmaceuticals Co. Ltd. (Hefei, Anhui, China), with all culture procedures approved by the Medical Ethics Review Committee of the Hefei Institutes of Physical Science, Chinese Academy of Sciences. PDCs and PDOs were cultured in lung cancer-specific medium (PRS-LCM-CR) and 3D lung cancer-specific medium (PRS-LCM-3D), respectively, both of which were purchased from Hefei PreceDo Pharmaceuticals Co. Ltd. (Hefei, Anhui, China).

C23084 cells were obtained from a 40-year-old female SCLC patient. The patient’s clinical history included the following: initial EP chemotherapy in December 2022 (etoposide 120 mg on days 1–3 plus carboplatin 500 mg on day 1); between February and May 2023, she completed four cycles of modified EP chemotherapy (etoposide 120 mg on days 1–3 plus carboplatin 400 mg on day 1) combined with camrelizumab immunotherapy (200 mg per cycle); and in July 2023, she received intrathoracic therapy comprising bevacizumab (300 mg) and cisplatin (initial dose of 30 mg followed by 20 mg).

LUC22009 cells were isolated from a 55-year-old male patient diagnosed with stage IVB SCLC (cT4N2M1e; with pulmonary hilar lymph node and brain metastases). The diagnosis was made on 2022-10-27 following symptoms of headache and seizures; 3 months prior, facial swelling and dizziness prompted a chest CT, which revealed a right upper lobe mass, and a subsequent bronchoscopic biopsy confirmed SCLC. Detailed information regarding chemo/immunotherapy (including dosages and cycles) is unavailable.

Additional clinical and pathological information for the other patients is provided in Supplementary Table 4.

Antibodies

The following antibodies were obtained from Cell Signaling Technology (Danvers, MA, USA): BMX (#24773), Caspase-3 (#14220), PARP (#9542), ubiquitin (#3936), p44/42 MAPK (#4695), phospho-p44/42 MAPK (#4370), STAT3 (# 12640), phospho-STAT3 (Tyr705) (#9145), Cyclin D1 (#55506), Rb (#9313), phospho-Rb (Ser807/811) (#8516), Akt (#4691), phospho-Akt (Ser473) (#4060), phospho-Akt (Thr308) (#4056), phospho-Src (Tyr416) (#6943 T), and Src (#2109). Phospho-E2F1(Ser332) (#abs128824) and phospho-E2F1(Ser337) (#abs106305) were purchased from Absin (Shanghai, China). β-actin antibody (#HC201-02) was purchased from TransGen Biotech (Beijing, China). The anti-FLAG antibody (#F2555-100UL) was obtained from Sigma (Danvers, MA, USA). Phospho-BMX (pY566) (#PAG5-104921) was purchased from Invitrogen (Thermo Fisher Scientific, Waltham, MA, USA). E2F1 (#66515-1-Ig), GAPDH (#60004-1-Ig), and Tubulin (#66031-1-Ig) antibodies were purchased from Proteintech. The secondary antibodies (anti-rabbit IgG) (#7074) and (anti-mouse IgG) (#7076) conjugated with horseradish peroxidase were also obtained from Cell Signaling Technology. All the antibodies were diluted according to the manufacturers’ instructions and used for Western blotting experiments.

Chemical reagents

IHMT-15137 and IHMT-15138 were synthesized and characterized in-house; ibrutinib (CAS #936563-96-1), adriamycin (CAS #23214-92-8), cisplatin (CAS #15663-27-1), etoposide (CAS #33419-42-0), ulixertinib (CAS #869886-67-9), palbociclib (CAS #571190-30-2), MG132 (CAS #133407-82-6), chloroquine (CAS #54-05-7), and cycloheximide (CAS #66-81-9) were purchased from MedChemExpress (MCE, Monmouth Junction, NJ, USA). Paclitaxel (CAS #33069-62-4), vincristine sulfate (CAS #2068-78-2), daunorubicin hydrochloride (CAS #23541-50-6), carboplatin (CAS # 41575-94-4), 5-FU (CAS # 2155827-07-7), and mitoxantrone (CAS # 65271-80-9) were purchased from TargetMol (Shanghai, China).

SDS-PAGE and immunoblot analysis

Tissues and cells were lysed with RIPA buffer (50 mM Tris-HCl, pH 7.4; 150 mM NaCl; 1% NP-40; 0.5% sodium deoxycholate; 0.1% SDS) supplemented with protease and phosphatase inhibitors for 20 min at 4 °C. Lysates were centrifuged at 13,000 rpm for 10 min at 4 °C. Protein concentrations were determined with the BCA protein assay. Proteins were separated by SDS-PAGE, transferred onto a nitrocellulose (NC) membrane, and then blocked with 5% non-fat dry milk in 1× TBST (Tris-buffered saline with 0.1% Tween 20) for 1 h at room temperature. Membranes were incubated with the appropriate primary antibodies at the dilution recommended by the manufacturer overnight at 4 °C. After three washes with 1× TBST for 10 min each, membranes were incubated with HRP-conjugated anti-rabbit or anti-mouse secondary antibodies (1:3000 dilution for each) for 1 h at room temperature. After three additional washes with 1× TBST, protein bands were visualized using Immobilon™ Western Chemiluminescent HRP Substrate (Millipore Sigma, Burlington, MA, USA; Cat. No. WBKLS0500) according to the manufacturer’s instructions.

Cytotoxicity assessment by CellTiter-Glo® assay

Cells were seeded in 96-well plates at 3000 cells per well in 100 μL of complete growth medium and allowed to adhere for 12 h. Compounds were serially diluted in complete growth medium and added to cells at the indicated concentrations. After 72 h of treatment, 15 μL of CellTiter-Glo® reagent (Promega, USA) was added to each well according to the manufacturer’s instructions. Luminescence was measured using an EnVision multimode plate reader (PerkinElmer, USA). Data were normalized to the DMSO control.

5-ethynyl-2’-deoxyuridine (EdU) staining assay

The EdU staining assay was performed using the BeyoClick™ EdU Cell Proliferation Kit with Alexa Fluor 594 (Beyotime, Shanghai, China; Cat. No. C0078S) according to the manufacturer’s instructions. Briefly, cells were seeded into 12-well plates at a density of 3 × 10⁵ cells per well. After treatment, cells were incubated with 10 μM EdU at 37 °C for 4 h. The medium was removed, and the cells were fixed with 4% paraformaldehyde for 15 min, then permeabilized with 0.3% Triton X-100 for another 15 min. After washing with 1× PBS, the cells were incubated with 1× Click reaction cocktail for 2 h in the dark, and the nuclei were stained with Hoechst 33342 for 30 min. Images were captured using a fluorescence microscope.

Covalent docking study

CovalentDock,⁵⁰ a computational tool built upon Autodock, was implemented to simulate the covalent binding interaction between IHMT-15137 and BMX. The crystal structure of human BMX kinase used for this docking study was obtained from the Protein Data Bank (PDB ID 3SXR, chain A) and was prepared using PyMOL, including removing water molecules and adding hydrogens. Energy minimization was then performed to avoid local collision. The ligand was set to form a covalent bond with the cysteine residue (Cys496). A grid box of 60 × 60 × 60 points centering on the coordinate of 22.97, 21.52, and 16.71 was implemented, which encloses the entire binding pocket. All remaining docking parameters followed the default protocol of the software.

Real-time quantitative PCR

Total RNA from cells was isolated using the RNeasy Kit (TIANGEN; Beijing, China; Cat. No. DP430). A total of 0.5 to 1 μg of RNA was reverse transcribed into cDNA using the ToloScript™ All-in-One RT Easy Mix (Tolo Biotech, Shanghai, China; Cat. No. 22107). Quantitative PCR was performed using the Hieff® qPCR SYBR® Green Master Mix (No Rox) (Yeasen Biotechnology, Shanghai, China; Cat. No. 11201ES08). The primer sequences used are listed in Supplementary Table 8. The relative expression levels of target genes were calculated using the 2^-ΔΔCt method, with GAPDH as the internal reference gene. Data were normalized to the control group.

CETSA assay

H446DDPR or H69AR cells (5 × 10⁶ cells) were lysed in 1 mL ice-cold lysis buffer (50 mM Tris-HCl, pH 7.5; 20 mM NaCl) containing a protease inhibitor cocktail. Lysis was achieved by three cycles of freezing-thawing (1 min freezing in liquid nitrogen and 3 min thawing at 37 °C), followed by centrifugation at 13,000 rpm for 10 min at 4 °C to obtain soluble lysates.

For temperature-dependent treatment, the lysates were incubated with either IHMT‑15137 (1 μM) or DMSO (vehicle control) at 4 °C for 4 h with gentle rotation. Subsequently, 50 μL aliquots of the lysate were heated to the indicated temperatures (50–62 °C) for 3 min using a ProFlex™ PCR thermal cycler (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA). After heating, the samples were allowed to cool to room temperature for 3 min and then centrifuged at 13,000 rpm for 20 min at 4 °C to remove aggregated proteins.

For dose-dependent treatment, aliquots (100 μL) of the lysates were treated with different concentrations of IHMT-15137 (0–10 μM) and incubated at 4 °C for 4 h with gentle rotation. After treatment, samples were subjected to thermal challenge in a PCR thermal cycler, with experimental groups heated to either 55.3 °C (H446DDPR cells) or 59.3 °C (H69AR cells) and control groups maintained at 37 °C for exactly 3 min. Following the heat treatment, samples were cooled at room temperature for 3 min and centrifuged at 13,000 rpm for 20 min at 4 °C to remove any aggregated proteins.

The remaining soluble fractions were mixed with 5× SDS loading buffer, boiled at 95 °C for 7 min, and then analyzed by Western blotting to detect BMX protein. The shift in the thermal melting profile of BMX protein was used to evaluate the binding affinity of IHMT-15137 to BMX.

DARTS assay

Cells were lysed in NP-40 lysis buffer (20 mM Tris-HCl, pH 8.0; 150 mM NaCl; 2 mM EDTA; 1% NP-40; 10% glycerol) with protease and phosphatase inhibitors for 3 h at 4 °C. Lysates were centrifuged at 13,000 rpm for 10 min at 4 °C, and the supernatants were collected. Protein concentrations were determined with the BCA protein assay. IHMT-15137 or vehicle (DMSO) was added to cell lysates and incubated for 3 h. Then the lysates were incubated with Pronase (Sigma-Aldrich, St. Louis, MO, USA; Cat. No. 10165921001) at a dose of 30 ng/μg protein for 15 min. The 5× SDS loading buffer was diluted to 1× with the protein samples. The samples were boiled at 95 °C for 7 min. The protection of BMX protein from proteolysis by IHMT-15137 was evaluated by comparing the protein levels in treated and untreated samples via Western blotting.

Colony formation assay

Cells (2000 cells/well) were seeded in six-well plates and allowed to adhere for 24 h before treatment with the indicated compounds at the specified concentrations. Medium containing compounds was refreshed every 3–4 days. Cells were incubated for 14–21 days in a humidified incubator at 37 °C with 5% CO₂. Colonies were stained with 1% crystal violet, washed with 1× PBS, air-dried, photographed, and counted.

Cell cycle analysis and flow cytometry (FACS)

Cells were seeded in six-well plates at a density of 1 × 10⁵ cells/well and cultured overnight. The following day, they were treated with either vehicle (DMSO) or test compounds for 24 or 48 h. After treatment, cells were harvested by trypsinization without EDTA (for adherent cells) or direct centrifugation (for suspension cells), washed once with 1× PBS, and fixed in ice-cold 75% ethanol at 4 °C overnight (≥12 h). The fixed cells were then washed twice with PBS to remove residual ethanol and stained with 0.5 mL PI/RNase staining buffer for 15 min at room temperature in the dark. Cell cycle distribution was analyzed using a FACScan™ flow cytometer (BD Biosciences, San Jose, CA, USA). Data were processed using ModFit 5.0 for cell cycle phase quantification (G0/G1, S, G2/M) and further analyzed with GraphPad Prism for statistical comparisons.

Competitive biotin-iodoacetamide assay

Cells were lysed with NP-40 lysis buffer supplemented with protease inhibitor cocktail at 4 °C for 30 min and centrifuged at 13,000 rpm for 10 min at 4 °C. The supernatants were collected and then incubated with IHMT-15137 for 4 h at 4 °C. Subsequently, the samples were incubated with 8 mM Biotin‑PEG2‑C2‑iodoacetamide (TargetMol, Shanghai, China; Cat. No. T17565) for 4 h at room temperature in the dark. The biotin-PEG2-C2-iodoacetamide-labeled proteins were isolated from the total protein pool using streptavidin magnetic beads (30 μL bead slurry per sample) for 4 h at 4 °C in the dark. After washing, the bound proteins were eluted from the beads by boiling in 30 μL of 2× SDS loading buffer for 10 min. The BMX protein levels were then detected by Western blotting.

Biochemical assay

BMX enzymatic inhibition assays of IHMT-15137, IHMT-15138, and ibrutinib were carried out by Hefei PreceDo Pharmaceuticals Co., Ltd. (Hefei, Anhui, China).

SCLC xenograft tumor model

Six-week-old female NCG mice were purchased from GemPharmatech Co. Ltd. (Nanjing, China). All animals were housed in a specific pathogen-free (SPF) facility and handled according to the animal care regulations of the Hefei Institutes of Physical Science, Chinese Academy of Sciences. All in vivo studies were approved by the Hefei Institutes of Physical Science Ethics Committee, Chinese Academy of Sciences (Approval No. DWLL(E)-2024-31). For tumor implantation, five million cells suspended in culture medium were mixed 1:1 with Matrigel® (BD Biosciences, San Jose, CA, USA; Cat. No. 354234) and injected into the subcutaneous space on the right flank of NCG mice. The animals were then randomized into treatment groups of five mice each for efficacy studies. Intraperitoneal (i.p.) dosing was initiated when tumors reached 200 mm³. Body weight and tumor volume were measured every other day after treatment. The tumor volume was calculated as follows: tumor volume (mm³) = [(W² × L)/2], in which width (W) was defined as the smaller of the two measurements and length (L) was defined as the larger of the two measurements.

21-day subacute toxicity study

A 21-day subacute toxicity study was conducted in C57BL/6 mice (n = 6 per group; three males and three females). Mice were dosed once daily for 21 consecutive days with vehicle, IHMT-15137 (100 or 300 mg/kg/day), or ibrutinib (100 or 300 mg/kg/day). All animals were housed in a specific pathogen-free (SPF) facility and handled in accordance with the animal care regulations of the Hefei Institutes of Physical Science, Chinese Academy of Sciences. All in vivo studies were approved by the Hefei Institutes of Physical Science Ethics Committee, Chinese Academy of Sciences (Approval No. DWLL(E)-2025-10).

Hemostatic tolerability was evaluated in a dedicated subset using the tail bleeding time (TBT) assay. Briefly, the tail was transected at a predefined site, and the bleeding stump was immersed in pre-warmed (37 °C) saline. The time to cessation of bleeding was recorded, and blood loss was quantified from the collection medium, following published guidance for murine TBT procedures.⁵¹

A parallel subset was monitored for body weight through day 21 and then euthanized for terminal evaluation. Whole blood was collected for complete blood count (CBC) and serum biochemistry (ALT, AST, ALB, and T-BIL). Major organs (heart, liver, spleen, lung, and kidney) were harvested for weighing and hematoxylin and eosin (H&E) histopathology.

Pharmacokinetics study

Male ICR mice were purchased from Vital River (Beijing, China) and housed in a specific pathogen-free (SPF) facility. All animal procedures were approved by the Ethics Committee of the Hefei Institutes of Physical Science, Chinese Academy of Sciences (Approval No. DWLL(E)-2025-39). IHMT-15137 was administered intraperitoneally (i.p.; 10 mg/kg), intravenously (i.v.; 1 mg/kg), or orally (p.o.; 10 mg/kg) (n = 3 per group). Compound concentrations in plasma and brain samples were quantified by LC-MS/MS. Pharmacokinetic parameters were calculated from individual animal data using noncompartmental analysis in Phoenix WinNonlin (v8.1).

PDCX study

The PDCX model was established using PDC C23084 cells. Six-week-old female NCG mice were purchased from GemPharmatech Co., Ltd. (Nanjing, China). All animals were housed in a specific pathogen-free (SPF) facility and handled according to the animal care regulations of Hefei Institutes of Physical Science, Chinese Academy of Sciences. All in vivo studies were approved by the Hefei Institutes of Physical Science Ethics Committee, Chinese Academy of Sciences (Approval No. DWLL(E)-2025-24). For tumor implantation, ten million cells suspended in culture medium were mixed 1:1 with Matrigel® (BD Biosciences, San Jose, CA, USA; Cat. No. 354234) and injected into the subcutaneous space on the right flank of NCG mice. Mice were monitored twice a week for tumor growth. When tumors reached 500–800 mm³, they were excised, minced, and re-implanted into new mice to passage the model. P2 animals were used for the efficacy study. Intraperitoneal (i.p.) dosing was initiated when tumors reached 200 mm³. Body weight and tumor volume were measured every other day after treatment. The tumor volume was calculated as follows: tumor volume (mm³) = [(W² × L)/2], in which width (W) was defined as the smaller of the two measurements and length (L) was defined as the larger of the two measurements.

Statistical analysis

All the data were presented as the means ± SEM, and the statistical analyses were performed via one-way ANOVA. GI₅₀ and EC₅₀ values were calculated via Prism 9.0 (GraphPad Software, San Diego, CA, USA) via the normalized dose-response curve for inhibition (variable slope model).

Author contributions

T.W., S.Q., C.S., C.W., Q.L., C.H., J.H.: methodology, investigation, formal analysis, validation, and visualization. A.W. and J.L.: project administration; T.W. and Z.Q.: writing—original draft. J.L., Z.Q., W.W., and Q.L.: conceptualization and writing—review & editing; Q.L.: supervision, resources, and funding acquisition. All the authors have read and approved the article.

Data availability

The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive (Genomics, Proteomics & Bioinformatics 2025) in National Genomics Data Center (Nucleic Acids Res 2025), China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA-Human: HRA016275) that are publicly accessible at https://ngdc.cncb.ac.cn/gsa-human. All other data supporting the findings of this study are included in the article and its Supplementary Materials.

Competing interests

The authors declare no competing interests.

Supplementary information

The online version contains supplementary material available at 10.1038/s41392-026-02644-1.