ONECUT2 reprograms neuroendocrine fate and is an actionable therapeutic target in small cell lung cancer

Abstract

Small cell lung cancer (SCLC) is a highly aggressive malignancy with extremely poor prognosis. SCLC cells exhibit high plasticity and can progress from neuroendocrine (NE) to non-NE phenotypes. This dynamic evolution promotes treatment resistance and relapses, representing a challenge for targeted therapies in this elusive disease. Here we identify the transcription factor ONECUT2 (OC2) as a driver of plasticity in SCLC, leading to non-NE transcriptional states. OC2 is highly expressed in SCLC tumors compared to normal lung tissue and its expression is associated with heightened clinical stage and lymph node metastasis. We show that OC2 is a repressor of ASCL1, the NE master regulator transcription factor. In addition, OC2 upregulates non-NE programs through activation of c-MYC and Notch signaling. We also demonstrate that OC2 is required for growth and survival of SCLC cells and that it can be targeted with a small molecule inhibitor that acts synergistically with the standard combination of cisplatin and etoposide, providing a novel therapeutic strategy for OC2 active SCLC tumors.

Keywords: ONECUT2; Phenotypic plasticity; SCLC; Therapeutic target; Tumor heterogeneity.

Fig.1

Association between OC2 expression and clinicopathological features in patients with SCLC. A OC2 mRNA expression (TPM normalized) in normal lung tissue compared to SCLC from the GSE60052 cohort (Jiang et al. 2016). The boxes show the 25–75^th percentile range, and the center line is the median. Whiskers extend from the minimum and maximum values. P-values were obtained from Wilcoxon two-tailed rank-sum test. B, C Representative IHC images (B) and quantification (C) of OC2 expression in benign (left) and SCLC tissue (right). OC2, brown. Scale bar, 50 µm. D, E Association of stage (D) and lymph node metastasis (E) with OC2. For (C-E) boxplots of intensity levels of nuclear OC2 expression assessed by IHC using a SCLC TMA are shown. The boxes show the 25–75^th percentile range, and the center line is the median. Whiskers extend from the minimum and maximum values. P-values were obtained from Wilcoxon two-tailed rank-sum test. F Association of stage with OC2 mRNA expression (arbitrary units) from SCLC tumors from the Lung Cancer Transcriptome Atlas (LCTA, in prep) cohort. G Association of lymph node metastasis with OC2 mRNA expression (FPKM, Fragments Per Kilobase per Million mapped fragments) from SCLC tumors from the U Cologne cohort (George et al. 2015). Nx = N1, N2 and N3 grouped. For (F, G) the boxes show the 25–75^th percentile range, and the center line is the median. Whiskers show 1.5 times the interquartile range (IQR) from the 25^th or 75^th percentile values. For (F), P-values were obtained from unpaired two-tailed Student´s t-test. For (G), Wilcoxon two-tailed rank-sum test was performed. H Correlation between OC2 mRNA expression (FPKM) and age of patients at diagnosis in the U Cologne cohort (George et al. 2015). I OC2 log2 (tpm + 1) mRNA expression in different lung NE neoplasms from the combined lungNENomics cohort. P-values were obtained from Wilcoxon two-tailed rank-sum test. J Pearson correlation between OC2 and ASCL1 (left) and NEUROD1 (right) mRNA expression in the lungNENomics cohort. For every comparison, P < 0.05 was considered significant. *P < 0.05, **P < 0.01

Fig. 2

OC2 modulates plasticity from NE to non-NE phenotypes in the DMS53 cell line. A Heatmap showing DEGs (adjusted P-value < 0.001) after OC2 enforced expression in DMS53 cells analyzed by hierarchical clustering. Three independent RNA-Seq experiments were performed per condition. B Plot showing upregulated (red bars) and downregulated (blue bars) SCLC subtype-specific gene signatures (Ireland et al. 2020) after OC2 enforced expression. C Immunohistochemical staining of SYP in OC2 overexpressing DMS53 cells. The boxes show the 25–75^th percentile range, and the center line is the median. Whiskers extend from the minimum and maximum values. P-values were obtained from Wilcoxon two-tailed rank-sum test. D, E mRNA (D) and protein levels (E) of OC2, ASCL1, NEUROD1 and YAP1 after OC2 enforced expression in DMS53 cells. For (D) qRT-PCR results were normalized using β-actin. The mean + SEM from three independent experiments is shown. Unpaired two-tailed Student’s t-test, *P < 0.05, **P < 0.01. For (E) representative blots from three independent experiments are shown. F Plot showing upregulated (red bars) non-NE gene signatures and downregulated (blue bars) NE gene signatures (Cai et al. ; Zhang et al. 2018)

Fig. 3

Inducible expression of OC2 in NCI-H510 cells leads to a phenotypic transdifferentiation towards a non-NE subtype. A Scheme of the experimental protocol of induction/off-phase/induction of OC2 expression in the NCI-H510 cell line. Cells were cultured with 50 ng/mL doxycycline (dox) for 5 days (first induction). Subsequently, the medium was replaced with complete RPMI for 6 days (off-phase). Finally, doxycycline was added again for 5 days (second induction). B Heatmap showing DEGs (adjusted P-value < 0.001) during the first induction (samples 1A and 1B), off-phase (samples 2A and 2B), and second induction (samples 3A and 3B) analyzed by hierarchical clustering. Three independent RNA-Seq experiments were performed per condition. C) Venn diagram of the overlapping DEGs during the induction/off-phase/induction experiment. D Plots showing upregulated (purple bars) and downregulated (green bars) SCLC subtype-specific gene signatures (Ireland et al. 2020) in NCI-H510 cells. E Immunoblot showing OC2 and ASCL1 expression during the induction/off-phase/induction experiment. F YAP1, ASCL1 and NEUROD1 mRNA expression after constitutive OC2 overexpression in NCI-H510 cells. Results were normalized using β-actin. The mean + SEM from at least three independent experiments is shown. G MSigDB Hallmark Gene Sets enriched in OC2 upregulated (purple bars) and downregulated (green bars) genes in OC2 overexpressing NCI-H510 and DMS53 cells

Fig. 4

OC2 induction alters NE transdifferentiation drivers. A GSEA plots showing enrichment of Notch signaling signatures (Vilimas et al. ; Milacic et al. 2024) in NCI-H510 and DMS53 cells after OC2 inducible and constitutive overexpression, respectively. Three independent RNA-Seq experiments were performed. B REST mRNA expression after OC2 enforced expression in the DMS53 and NCI-H510 cell lines. The mean + SEM from three independent experiments is shown. Unpaired two-tailed Student’s t-test, **P < 0.01. Results were normalized using β-actin. C Cell proliferation curve showing OC2 induced vs. control NCI-H510 cells. Proliferation was calculated relative to T0 number of cells. The values shown are the mean ± SEM from five independent experiments. D Heatmap showing DEPs (P-value < 0.05) after OC2 induction in NCI-H510 cells. E Heatmap of transcription factors generated with Metascape employing the TRRUST method using DEPs between control and OC2 overexpressing NCI-H510 cells. F c-MYC and l-MYC mRNA expression after OC2 constitutive overexpression in the NCI-H510 cell line. The mean + SEM from four independent experiments is shown. Unpaired two-tailed Student’s t-test, **P < 0.01. Results were normalized using β-actin. G c-MYC protein expression levels detected by western blot after OC2 overexpression in the NCI-H510 and DMS53 cell lines. β-actin was used as control. H Protein networks and their cellular localization obtained in response to OC2 induction in the NCI-H510 cell line. Networks are arranged in the form of nodes (proteins) and lines (biological relationships between nodes). The proteins in green are overexpressed and those in red are repressed. The proteins in white do not belong to our dataset but are included in the database. In orange, the program predicts protein activity activation, and in blue, inhibition. The solid lines connecting the nodes represent direct interactions, while the dashed lines represent indirect interactions. The color of the lines indicates activation if orange, inhibition if blue, inconsistent findings if yellow, and unpredictable effects if gray. The shape of the nodes denotes the protein function: enzymes (diamond), transcriptional regulators (oval), kinases (inverted triangle), peptidases (horizontally oriented diamond), nuclear receptors (rectangle), transporters (trapezoid), others (circle). Image obtained from Ingenuity

Fig. 5

OC2 depletion decreases cell viability and induces apoptosis of a subset of SCLC cell lines. A OC2 protein expression levels detected by western blot after OC2 silencing with shRNAs. β-actin was used as control. B Heatmap showing DEGs (adjusted P-value < 0.001) after OC2 depletion in NCI-H510 cells. C MSigDB Hallmark Gene Sets upregulated (blue bars) and downregulated (grey bars) in OC2 depleted NCI-H510 cells. D Cell death caused by OC2 silencing measured by flow cytometry. The graph shows the mean + SEM from three independent experiments. Unpaired two-tailed Student´s t-test, * = P < 0.05, ** = P < 0.01. E Dose–response curves (left) and IC50 values (right) for compound CSRM617 in mouse and human SCLC cell lines after 48 h treatment with CSRM617. The values shown are the mean ± SEM from three independent experiments. F SynergyFinder software was used to calculate the synergy score according to the HSA model. Positive (red) or negative (green) synergy scores indicate synergistic and antagonistic effects, respectively. The shown values are the mean from three independent experiments