Christian Joerg Braun

Adaptation to PRMT5 inhibition involves upregulation of splicing factors. A, Scheme of RNA-seq experiment: DanG and PSN1 cells were treated for 72 hours with 20 nmol/L JNJ-64619178 or the vehicle control. RNA-seq data were analyzed by GSEA. B, Venn diagram of upregulated GO-BP gene sets in JNJ-64619178–treated DanG and PSN1 cells. C, GSEA (GeneTrail 3.2) results of GO-BP signatures in JNJ-64619178–treated vs. DMSO control DanG and PSN1 cells. Depicted are the top 14 enriched gene sets sorted by the lowest mean adjusted P value. Enrichment scores are shown on the x-axis. D, Venn diagram of JNJ-64619178 upregulated genes with a gene score >4 (GeneTrail 3.2) in the GO-BP mRNA metabolic process gene set. E and F, Heatmap of the 15 genes (D) upregulated in DanG (E) and PSN1 (F) cells by JNJ-64619178 treatment. G, Western Blot of RBM39 expression in PSN1 cells treated with PRMT5i JNJ-64619178 (4 nmol/L) or GSK-3326595 (200 nmol/L) or DMSO as the vehicle control for 24 and 48 hours. β-Actin was used as the loading control. One of three representative biological replicates is shown. H, Quantification of RBM39 expression (G) after 48 hours (n = 3). The intensity of antibody signal of RBM39 was normalized to the loading control. I and J, Western Blot of RBM39 protein expression in PSN1 (I) and BxPc3 (J) cells treated with increasing dosages of MRTX1719 for 48 hours. β-Actin was used as the loading control, and one of two (PSN1) or three (BxPc3) biological replicates is shown. K–N, Top, Western blot of RBM39 protein expression in PSN1 (K), DanG (L), MiaPaCa2 (M), and BxPc3 (N) cells treated with increasing dosages of TNG462 for 72 hours. β-Actin was used as the loading control. Bottom, quantification of three independent experiments. K, Friedman test P = 0.0009, *Dunn multiple-comparison test: P < 0.05; (L) Friedman test P = 0.015; (M) Friedman test P = 0.026; (N) Friedman test P = 0.038. [A, Created in BioRender. Schneider, G. (2026) https://BioRender.com/8or8uh9.]

RBM39 knockdown increased MRTX1719 responsiveness. A–C, PSN1 (A), MiaPaCa2 (B), and BxPc3 (C) cells were transfected with the indicated esiRNAs. Left, Western blot of RBM39 expression 72 hours after the transfection. β-Actin was used as the loading control. Right, quantification of RBM39 expression of (A) n = 6, (B) n = 5, and (C) n = 3 biological replicates. D–F, PSN1 (D), MiaPaCa2 (E), and BxPc3 (F) were transfected with the indicated siRNAs and treated after 24 hours with different dosages of MRTX1719 as indicated for additional 72 hours. Left, viability normalized to cells treated with the vehicle control and esiFLUC-transfected cells is depicted. Right, areas under the dose–response curves (AUC) were calculated and compared. Statistical significance was assessed using a two-tailed paired t test; *, P < 0.05 or the exact P value is indicated.

Activity of indisulam in PDAC models. A, Scheme of the mechanism of indisulam-dependent degradation of RBM39. B, Western blot of RBM39 expression of PSN1 cells treated with 0, 200, and 400 nmol/L indisulam for 24 (left) and 48 (right) hours. β-Actin was used as the loading control. One of three biological replicates is shown. C and D, Western blot of RBM39 expression of MiaPaCa2 (C) and DanG (D) cells treated with 0, 200, and 400 nmol/L indisulam for 24 hours. β-Actin was used as the loading control. One of three biological replicates is shown. E, Correlation analysis of PRMT5 (y-axis) and RBM39 (x-axis) protein expression using CPTAC PDAC proteomics data. For statistical analysis, the Pearson correlation coefficient, P value, and FDR are indicated. F, Venn diagram of the top 20 GO-BP gene sets connected to RBM39 and/or PRMT5 protein expression. Gene sets were determined via LinkInterpreter of LinkedOmics. G, Bar graph showing the commonly enriched GO-BP gene sets linked to RBM39 and PRMT5 determined in F. The normalized enrichment score (NES) is shown on the x-axis. H, Correlation analysis of indisulam sensitivity and DCAF15 expression in pancreatic cancer cell lines using DepMap data. Spearman’s r = −0.41, P = 0.02. I, GSEA (GeneTrail3.2) of Hallmark (left) and GO-BP (right) in indisulam-sensitive PDAC cell lines. The enrichment score is shown on the x-axis, and the q values are color-coded. [A, Created in BioRender. Schneider, G. (2026) https://BioRender.com/rzri1up.]

Combined PRMT5 and RBM39 inhibition induced apoptosis. A, Cell-cycle distribution of PSN1 cells treated with the vehicle control (DMSO), 4 nmol/L JNJ-64619178, 200 nmol/L indisulam, or the combination of 4 nmol/L JNJ-64619178 and 200 nmol/L indisulam for 48 (left) and 72 hours (right). Cell-cycle distribution was determined by FACS of PI-stained cells (n = 3). Repeated measure one-way ANOVA with correction for multiple-testing according to the Tukey test. The P value is indicated or *, P < 0.05. B and C, Caspase-3/7 Assay of PSN1 cells treated with the indicated dosages of indisulam and JNJ-64619178 (B) or GSK-3326595 (C) for 48, 72, and 96 hours. The fold activation of caspase-3/7 is color-coded (n = 3). D, Panc-1 caspase reporter cells were treated with 8 nmol/L JNJ-64619178 and/or 400 nmol/L indisulam, or DMSO was used as the vehicle control. The green fluorescence intensity (caspase activation) normalized to cell count, DMSO control, and t = 0 is shown over time (left) at 96 hours (middle) and 120 hours (right). Ninety-six hours: Friedman test, P < 0.0001; 120 hours: Friedman test, P < 0.0001. The P value after Dunn multiple-comparison test is indicated: **, P < 0.01; ***, P < 0.001, n = 3. E, PSN1 caspase reporter cells were treated with PRMT5i (8 nmol/L JNJ-64619178 or 100 nmol/L MRTX1719) and/or 400 nmol/L indisulam. DMSO was used as the vehicle control. The green fluorescence intensity (caspase activation) normalized to cell count, DMSO control, and t = 0 is shown over time (left) at 96 hours (middle) and 120 hours (right). Ninety-six hours: Friedman test, P < 0.0001; 120 hours: Friedman test, P < 0.0001. Dunn multiple-comparison test: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001, n = 3.

Synergistic response to PRMT5 and RBM39 inhibitors. A, Clonogenic assays of PSN1 cells treated with the indicated dosages of indisulam and JNJ-64619178 (above) or GSK-3326595 (below) for 9 days. One of three biological replicates is shown. B, Most synergistic area scores according to the ZIP, Loewe, and HSA models for the combination of indisulam with PRMT5 inhibition [JNJ-64619178 (left) or GSK-3326595 (right)] in four PDAC cell lines: PSN1, PaTu-8988T, HPAC, and Panc-1. Response values are from the corresponding clonogenic growth assays (n = 3) as determined in (A), and concentrations were adapted to the responsiveness of each cell line. C, Synergy distribution plots showing the ZIP synergy score of PSN1 (left) and MiaPaCa2 (right) cells treated with MRTX1719 and indisulam as indicated. Values are based on clonogenic growth assays (n = 3). D, Scheme of the generation of PDCLs. PDX, patient-derived xenotransplant. E, ZIP, Loewe, and HAS most synergistic area scores of six GoeCDX cell lines treated with JNJ-64619178 and indisulam. The experiment corresponds to at least three independent biological replicates for every cell line. F, Clonogenic assay of GoeCDX4 cells with the indicated dosages of indisulam and JNJ-64619178. One of six biological replicates is shown. G, Western blot of PRMT5 expression in GoeCDX cells. HSP90 was used as the loading control. One of three biological replicates is shown. H, Quantification of G. I, Scheme of the experimental setup of the ISRA. PSN1 cells were seeded at a low density and treated with the individual GI80 value of the indicated drugs and drug combinations (8 nmol/L JNJ-64619178, 800 nmol/L indisulam, 800 nmol/L GSK-3326595, 200 nmol/L GSK-3326595, and 400 nmol/L indisulam or 4 nmol/L JNJ-64619178 and 70 nmol/L indisulam; DMSO was used as the control). The medium with fresh therapies was changed weekly, and confluence was measured weekly for 10 weeks. Wells with a confluence more than 50% were considered resistant. J, Results of the ISRA plotted as a Kaplan–Meier curve Statistical analysis was performed with a log-rank test: ****, P < 0.0001. The results from three biological replicates (3 × 30 wells per condition) are shown. K, ISRA outgrown cell populations from the indicated conditions were further cultured in medium containing the respective drug. RBM39 protein levels were assessed by Western blotting. β-Actin served as a loading control. L, Quantification of the RBM39/β-actin ratio in control (n = 3 outgrown wells) and PRMT5i (n = 3 GSK3326595 and n = 2 JNJ-64619178 outgrown wells). Mann–Whitney test: *, P < 0.05. M, Cell populations that grew out in the control ISRA, GSK-3326595 ISRA, and indisulam ISRA were treated with vehicle (DMSO), 800 nmol/L indisulam, or 800 nmol/L GSK-3326595. Relative clonogenic growth was assessed after 7 days. The bar graph shows the percentage inhibition of clonogenic growth normalized to DMSO-treated controls, and each dot represents one population derived from a single well. For each biological replicate of the ISRA (J), three populations were analyzed, yielding a total of n = 9 population tested in one biological replicate. Statistical analysis was performed using the Kruskal–Wallis test with Dunn multiple-comparison test: **, P < 0.01. [D, Created in BioRender. Schneider, G. (2026) https://BioRender.com/0w6dd14; I, Created in BioRender. Schneider, G. (2026) https://BioRender.com/xsbxjh0.]

PRMT5 and RBM39 inhibition regulate DNA repair and metabolic pathways. A, Workflow of the RNA-seq approach: PSN1 cells were treated with DMSO as the vehicle control, 4 nmol/L JNJ-64619178, 200 nmol/L indisulam, or the combination of 4 nmol/L JNJ-64619178 and 200 nmol/L indisulam for 48 hours, mRNA was isolated, and mRNA sequencing with subsequent GSEA was performed. The experiment was conducted as four biological replicates. B, Venn diagram showing downregulated GO-BP gene sets in PSN1 cells; all treatments were compared with the vehicle control in individual GSEAs. Only gene sets with an FDR < 0.05 were considered significant and used for the analysis. C, 24 GO-BP gene sets downregulated exclusively in the combination treatment are indicated. The value of the normalized enrichment score (NES) is color-coded. D, Number and composition of significant changed alternative splicing (AS) events observed upon monotherapy or combination treatment of JNJ-64619178 and/or indisulam. E, Percent spliced-in (ΔPSI) distribution of different AS events identified following monotherapy or combination treatment with JNJ-64619178 and/or indisulam. P values based on the Wilcoxon test are indicated. F, PSN1 cells were either treated with DMSO control or 4 nmol/L JNJ-64619178 and 200 nmol/L indisulam for 96 hours followed by mass spectrometry determination of the proteome. G, Volcano plot of deregulated proteins in PSN1 cells treated with the combination therapy [see (F)] vs. control. Proteins with an adjusted P value < 0.05 were colored. Proteins involved in the DNA repair pathway are highlighted lilac, whereas proteins involved in the splicing pathway are highlighted pink. H, ΔPSI distribution of different AS events in the DNA damage signaling and repair pathway genes ATM, FAN1, and FANCL observed upon monotherapy or the combination treatment with JNJ-64619178 and/or indisulam. I, Results of the γH2A.X phosphorylation assay: PSN1 cells were treated with 50 nmol/L MRTX1719 and/or 240 nmol/L indisulam for 48 hours (left) or 72 hours (right). For negative control, cells were left vehicle-treated; for positive control, cells were irradiated with 2 Gy for 30 minutes. The fold increase of γH2A.X phosphorylation is shown on the y-axis. Statistical analysis was performed using a repeated-measures one-way ANOVA with correction for multiple-testing according to the Tukey test: **, P < 0.01; *, P < 0.05, n = 3. A3SS, alternative 3′ splice site; A5SS, alternative 5′ splice site; MSE, mutually exclusive exon; RI, retained intron; SI, skipped exon. [A, Created in BioRender. Schneider, G. (2026) https://BioRender.com/fwl2ir3; F, Created in BioRender. Schneider, G. (2026) https://BioRender.com/z84e701.]