Exploring RNA biology in pseudomyxoma peritonei uncovers splicing dysregulation as a novel, targetable molecular vulnerability

Abstract

Pseudomyxoma peritonei (PMP) is a rare neoplasm coursing with uncontrollable mucus accumulation, with a high relapse rate. RNA biology processes have emerged as new players in cancer development and progression, nevertheless their role in PMP remains unknown. In this study, we aimed to examine RNA-regulatory machineries in PMP and their potential contribution to this disease progression. We analyzed 62 splicing-related genes, 27 RNA exosome and 21 nonsense-mediated decay genes, in a cohort of 29 patients using a microfluidic array, comparing tumor and control/reference tissues, together with external RNA-seq and proteomic data. Our results revealed a profound dysregulation of key components, which correlated to relevant clinical parameters and enabled to distinguish between tumor and control tissues. In vitro splicing inhibition using Pladienolide-B, as well as the modulation of specific splicing factors, reduced aggressiveness parameters, enhanced the effect of clinically used drugs, and revealed a strong correlation between dysregulated genes and key cancer-related genes. This inhibition also affected mucin secretion and mucin variants production. Collectively, our findings provide the first evidence for dysregulation of the genes of pivotal RNA-regulatory processes in PMP, implying that these targetable mechanisms may be functionally altered and play a role in the disease. Hence, a thorough understanding of its RNA biology could aid in the discovery of new clinically actionable vulnerabilities in this rare disease.

Fig. 1. Splicing, RNA exosome and Nonsense-mediated decay machineries dysregulation in PMP.

A RNA expression level of each machinery individually as Log₂ of the Fold-Change Tumor/Control. In the upper panel the expression of all genes is represented showing upregulated (colored in red) and downregulated genes (colored in blue) as the Log₂ of the Fold-Change of Tumor/Control (n = 11 controls; n = 18 tumors); in the graph below are represented the statistically significant genes altered as violin plots as Log₂ of normalized copy number, showing the median expression and the first and third quartiles (thin lines). T-test or Mann–Whitney U test were used to compare means between the two groups. Asterisks represent significant differences: *p < 0.05; **p < 0.01; ****p < 0.0001. B PLS-DA and VIP-Score of all significantly altered genes. PLS-DA shows two different clusters: control tissue (“0”, green) and PMP tissue (“1”, red); in the right panel is the score of each of these genes (top 15) contribution to the discrimination model, VIP-Score, and the mean expression of each gene in the control, “0”, and PMP, “1”, groups (“high expression” in red, “low expression” in blue).

Fig. 2. Protein distribution in PMP and control tissue of some relevant splicing factors.

Representative IHC images (10X) showing the distribution of HNRNPK, MBNL1 and PTBP1 in FFPE control samples (appendix) and in LG and HG PMP. Zoomed images of the regions framed in blue (20X) and pink (40X) are shown in the bottom panel.

Fig. 3. Splicing machinery, RNA exosome complex and NMD genes are associated to PMP and allow to discriminate between tumor and control samples.

A Clinical correlations of all genes measured. Relevant clinical parameters in PMP as OS, RFP, Ki67, p53 levels, PCI, CK7 presence and morbidity were associated to some splicing, RNA exosome and NMD genes expression. These correlations are represented in the graph as the p-value in blue (for splicing machinery), green (RNA exosome complex), and orange (NMD) being the symbol bigger as the p-value is smaller. The upward triangle represents direct correlations, the inverted triangle represents inverse correlations. Logrank test, Pearson or Spearman correlation and t-test or Mann–Whitney U test were used for survival and relapse analysis, Ki67, p53 and PCI, and CK7 positivity and morbidity, respectively. B Best ROC curves for each machinery. In the splicing machinery (blue), RNU4ATAC was the best gene to discriminate between PMP and control tissue; RBM7 was the best among the RNA exosome (green) and SMG8 in the NMD (orange).

Fig. 4. Cell cycle and proliferation, immunomodulation and inflammation and angiogenesis genes expression in PMP and its association to splicing genes expression.

A Significantly altered cancer-related genes in PMP. CDKN2A, CDKN2D, CDK4, CDK6, ATM, CXCL3, CCL5, CXCR2 and KDR expression are represented as Log₂ of normalized copy number in the control (white violin, n = 11) and the PMP (colored violin, n = 18) tissues. T-test or Mann–Whitney U test were used and asterisks represent significant differences: *p < 0.05; **p < 0.01; *** p < 0.001. B Significant correlations found between these cancer-related altered genes and splicing-related dysregulated genes. Significant correlations are shown together with Pearson or Spearman correlation coefficient (r) and p value in each case.

Fig. 5. In vitro experiments inhibiting splicing process using Pladienolide-B in the PMP cell line model N14A.

A Cell viability assay using Pladienolide-B at 10^-10M, 10^-9M, and 10^-8M. The graph shows cell viability percentage respect to the control of N14A in crescent doses of Pladienolide-B at the different times of the experiment (24, 48 and 72 h). The dotted line at Y = 100 represents the control (vehicle). B Cell viability assay using Pladienolide-B (10^-9 M) in combination with mitomycin-C (4 µg/mL) and cisplatin (0.5 × 10^-3 M). The graph shows the percentage of cell viability relative to the control (vehicle, dotted line at Y = 100) in N14A cells under different treatments: Pladienolide-B, mitomycin-C, cisplatin, Pladienolide-B + mitomycin-C, and Pladienolide-B + cisplatin at 24, 48, and 72 h. C Cell migration using Pladienolide-B at 10^-9M in the wound-healing assay. The left panel represents the quantification of the area of the scratch in the cells treated respect to the control in percentage in N14A cell line after 16 h. The dotted line at Y = 100 represents control. In the right panel are the pictures of the scratch at 16 h for the control (up) and the treatment (down). D Turbidity assay using Pladienolide-B at 10^-9M. OD at 900 nm is represented for the control (white box) and treated (colored box) cells. E RNA expression level of mucin 2 splicing variants using Pladienolide-B at 10^-9M. The RNA expression levels of all MUC2 isoforms, MUC2-204, MUC2-205 and MUC2-206 are represented in the control and treated cells. Data are represented as mean ± SEM, n = 3 in viability, migration and qPCR assays and n = 6 in turbidity assay. For cell viability, mean differences were assessed using ANOVA, with post hoc Tukey’s test to compare between group pairs. In the rest of the cases, t-test or Mann–Whitney U test were used. Asterisks represent significant differences compared to the control: *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; symbol plus represents significant differences between the different treatments: Pladienolide-B vs Pladienolide-B + mitomycin-C; Pladienolide-B vs Pladienolide-B + cisplatin; mitomycin-C vs Pladienolide-B + mitomycin-C; cisplatin vs Pladienolide-B + cisplatin; +p < 0.05; ++p < 0.01; ++++p < 0.0001. F MUC-205 and MUC-206 splice variants relative expression in N14A. PSI (percent spliced-in) is represented for MUC-205 and MUC-206 splice variants in the control and after treatment (Pladienolide-B at 10^-9M). The structure of each mucin 2 variant is represented in the right scheme. Pd: Pladienolide-B.

Fig. 6. In vitro experiments following modulation of key splicing factors: overexpression of HNRNPK and MBNL1 (1 µg/mL each) and silencing of PTBP1 (75 nM).

A Validation of each modulation at the RNA level by qPCR. The graphs show the expression levels of each splicing factor relative to the control (CMV6 plasmid for overexpression; scramble for silencing) 24 h post-transfection. Expression was normalized using GAPDH as the housekeeping gene. The dotted line at Y = 100 represents the control. B Cell viability assay after each transfection. Cell viability rate is shown as percentage of the control (dotted line at Y = 100) at the different times (24, 48, 72 h). C Cancer-related genes altered after modulation of HNRNPK, MBNL1, and PTBP1, measured by qPCR. Gene expression is presented as Log₂ of the Fold-Change modulation/control. Upregulated genes are shown in red, downregulated genes in blue. D RNA expression level of mucin 2 splicing variants after HNRNPK, MBNL1 and PTBP1 modulation. Expression levels of MUC2 isoforms (MUC2-204, MUC2-205, and MUC2-206) are shown in the control transfection (CMV6 plasmid for HNRNPK and MBNL1, or scramble for PTBP1), normalized to GAPDH expression. Data are presented as mean ± SEM, n = 3. For cell viability, differences were assessed using ANOVA with post hoc Tukey’s test for pairwise comparisons. For other cases, t-test or Mann–Whitney U test were used. Asterisks represent significant differences compared to the control: *p < 0.05; **p < 0.01; *** p < 0.001; ****p < 0.0001. E Relative expression of MUC2-205 and MUC2-206 splice variants in N14A after modulation of HNRNPK, MBNL1, and PTBP1. PSI (percent spliced-in) for MUC2-205 and MUC2-206 splice variants are shown for control and different modulations.