Luc Girard

ISL1 “High” SCLC Exhibits DNA Damage Following Lurbinectedin Treatment. Quantification of percentage of DNA in tail by alkaline comet assay for H209 treated with DMSO, 0.1 nanomolar lurbinectedin, 1 nanomolar lurbinectedin, sterile water, 1 μM cisplatin, 10 μM cisplatin for 24 hours H209 treated with 100 μM H2O2 served as a positive control. Dots indicate the means of three independent experiments, each with >50 comets scored. ****, P < 0.0001. Representative immunofluorescence microscopy images of Vista Green DNA Dye staining are shown in green color. Original magnification 10X.

Heatmap of Global Changes in SCLC Gene Expression Following Lurbinectedin Treatment. LogFC for 83 downregulated and 14 upregulated genes in SCLC cell lines following 24 hours lurbinectedin treatment, regardless of ISL1 expression. All P values for genes mapped are <0.01.

Heatmap of Differentially Expressed Genes Between ISL1 “High” and ISL1 “Low” SCLC Cell Lines Following Lurbinectedin Treatment. Select genes differentially expressed between ISL1 “high” lines (H209 and H187) and ISL1 “low” lines (H82 and SHP77) following 24 hours lurbinectedin treatment are mapped by log fold change (LogFC). Changes in MYC paralog expression are additionally mapped. *, P < 0.05, **, P < 0.01, ***, P < 0.001, ****, P < 0.0001.

Baseline Characteristics, Smoking Status, and Platinum Sensitivity of Patients Enrolled on Study. Table depicting baseline characteristics, including age at diagnosis, sex, race, stage at diagnosis by the Veterans Affairs classification, LS-SCLC, limited-stage SCLC, ES-SCLC, extensive-stage SCLC, smoking history at diagnosis, eastern cooperative oncology group (ECOG) performance status, platinum-sensitivity with chemotherapy-free interval (CTFI), and number of cycles of lurbinectedin received.

Excel File for Supp Table S3 (raw data)

Heatmap of Differential Gene Expression for SCLC Stratified by ISL1 Expression. Sixty-two (62) sequenced lung cancers were stratified into two cohorts by ISL1 expression (Transcripts per million, TPM 6) and queried for differentially expressed genes between cohorts. TPM for ISL1 and 8 significantly differentially expressed genes between cohorts are mapped. All P values for genes mapped are <0.05.

Expression of SLFN11, Previously Identified SCLC Subtypes, RNA Polymerase II Subunits, and Novel Targets Differentially Expressed Between Lurbinectedin Cohorts by Tissue-based Proteomics. Table of log2 ratios and P values showing change in protein expression between resistant (Res, 8 lurbinectedin cycles received) patient cohorts (Res/Sens). Novel genes identified to be differentially expressed between lurbinectedin sensitive and resistant cohorts included ISL1, SOX5, SIX1, and SIX4. SLFN11 and previously identified SCLC subtypes were not significantly different between cohorts, although ASCL1 bordered on significance (P = 0.0689).

Correlation Analysis of ISL1 with Neuroendocrine (NE) Subtypes and Lurbinectedin Response in SCLC. A) Spearman correlation analysis for cBioPortal SCLC mRNA expression Z-Scores between ISL1 and different SCLC NE subtypes (ASCL1, NeuroD1, and POU2F3). Correlation analysis (B-C) comparing ISL1 delta CT mean and lurbinectedin IC50 values for SCLC Cell Lines. A) Graph omitting outlier H526 for Spearman correlation analysis (r = 0.86, P = 0.0023). B) Graph of ASCL1+ cell lines only for Spearman correlation analysis (r = 0.71, P = 0.1361).

Small cell lung cancer (SCLC) is a recalcitrant thoracic malignancy known for acquired chemoresistance, early metastatic spread, and poor overall survival. Lurbinectedin, a DNA minor groove alkylating agent, provides durable efficacy in a minority. Predictive biomarkers for lurbinectedin are needed. Patients with relapsed SCLC who received lurbinectedin (n = 16) were classified by cycles received, including eight durable responders defined as ≥8 cycles (average, 14.75 cycles; median progression-free survival, 9.8 months). Pretreatment specimens were analyzed by immunohistochemistry (IHC) for Schlafen-11 (SLFN11) and tandem mass tag–labeled expression proteomics. Top candidates were confirmed by IHC and functionally validated in SCLC cell lines. SLFN11 failed to predict lurbinectedin response (P = 0.40). Proteomics highlighted a primitive neuroendocrine (NE) pathway [Islet-1 (ISL1), SOX5, SIX1, SIX4]. ISL1 expression significantly correlated with lurbinectedin response (r = 0.65; P = 0.0351). IHC confirmed that lurbinectedin reduced ISL1 after treatment. Lurbinectedin preferentially induced DNA damage in ISL1-“high” SCLC (P < 0.0001) without causing NE subtype switching. RNA sequencing showed downregulation of ISL1, RBMS3, ASCL1, SOX5, and SIX1 and upregulation of ATF3. ISL1-“high” SCLC demonstrated cellular dependency on ISL1; ISL1 knockdown reduced lurbinectedin sensitivity. L-MYC positively regulated ISL1, whereas ISL1 positively regulated ASCL1 and SOX5. This is the first comprehensive investigation of predictive biomarkers for lurbinectedin. Proteomics identified ISL1 as defining a novel SCLC subtype with enhanced lurbinectedin sensitivity. ISL1 serves as both a predictive biomarker and a functional dependency, as evidenced by its essentiality for cell survival and loss following treatment. Prospective studies using ISL1 as a predictive biomarker for lurbinectedin are planned.

Differentially Expressed Proteins Associated with ISL1 Expression from Tissue-based Proteomics. List of differentially expressed proteins (n = 142) that are associated with ISL1 expression (defined as log2 transformed ion intensities, ISL1 “low” 13.2) in the patient samples from tissue-based proteomics, selected based on the threshold of |log2 ratio| >0.9 and P < 0.05; red denotes log2ratio >0.9; blue denotes <-0.9.