Caspases compromise SLU7 and UPF1 stability and NMD activity during hepatocarcinogenesis

Abstract

Background & aims: The homeostasis of the cellular transcriptome depends on transcription and splicing mechanisms. Moreover, the fidelity of gene expression, essential to preserve cellular identity and function is secured by different quality control mechanisms including nonsense-mediated RNA decay (NMD). In this context, alternative splicing is coupled to NMD, and several alterations in these mechanisms leading to the accumulation of aberrant gene isoforms are known to be involved in human disease including cancer.

Methods: RNA sequencing, western blotting, qPCR and co-immunoprecipitation were performed in multiple silenced culture cell lines (replicates n ≥4), primary hepatocytes and samples of animal models (Jo2, APAP, Mdr2 ^-/- mice, n ≥3).

Results: Here we show that in animal models of liver injury and in human HCC (TCGA, non-tumoral = 50 vs. HCC = 374), the process of NMD is inhibited. Moreover, we demonstrate that the splicing factor SLU7 interacts with and preserves the levels of the NMD effector UPF1, and that SLU7 is required for correct NMD. Our previous findings demonstrated that SLU7 expression is reduced in the diseased liver, contributing to hepatocellular dedifferentiation and genome instability during disease progression. Here we build on this by providing evidence that caspases activated during liver damage are responsible for the cleavage and degradation of SLU7.

Conclusions: Here we identify the downregulation of UPF1 and the inhibition of NMD as a new molecular pathway contributing to the malignant reshaping of the liver transcriptome. Moreover, and importantly, we uncover caspase activation as the mechanism responsible for the downregulation of SLU7 expression during liver disease progression, which is a new link between apoptosis and hepatocarcinogenesis.

Impact and implications: The mechanisms involved in reshaping the hepatocellular transcriptome and thereby driving the progressive loss of cell identity and function in liver disease are not completely understood. In this context, we provide evidence on the impairment of a key mRNA surveillance mechanism known as nonsense-mediated mRNA decay (NMD). Mechanistically, we uncover a novel role for the splicing factor SLU7 in the regulation of NMD, including its ability to interact and preserve the levels of the key NMD factor UPF1. Moreover, we demonstrate that the activation of caspases during liver damage mediates SLU7 and UPF1 protein degradation and NMD inhibition. Our findings identify potential new markers of liver disease progression, and SLU7 as a novel therapeutic target to prevent the functional decay of the chronically injured organ.

Keywords: apoptosis; hepatocellular carcinoma; liver damage; splicing; transcriptome.

Graphical abstract

Fig. 1

Intron retained mRNAs, NMD targets and mRNAs encoding NMD factors accumulate in HCC and upon SLU7 silencing. (A) Transcript level Kallisto and SLEUTH genome-wide RNA sequencing analysis in SLU7-silenced (siSLU7) PLC/PRF/5 cells compared to control siGL cells. (B) Enrichment plot of the NMD gene set after GSEA of siSLU7 vs. siGL PLC/PRF/5 cells. (C) PCR analysis of SRSF1, SRSF2 and SRSF3 mRNA isoforms in siGL and siSLU7 PLC/PRF/5 cells. Positive control: Cells treated with the protein synthesis inhibitor CHX. The accumulation of PTC-containing isoforms, NMD targets, is indicated. (D) Real-time PCR relative expression of SLU7 and different NMD targets in siSLU7 PLC/PRF/5 vs. siGL cells established as 100%. (E) Transcript level Kallisto and SLEUTH genome-wide RNA sequencing analysis of HCC and NT liver samples in the TCGA database. (F) Enrichment plot of the NMD gene set after GSEA of HCC and NT liver samples in the TCGA database. CHX, cycloheximide; GSEA, gene set-enrichment analysis; NMD, nonsense-mediated RNA decay; NT, non-tumoral; PTC, premature termination codon.

Fig. 2

SLU7 regulates NMD and interacts with UPF1. (A) Left panel: luciferase plasmid system used to study NMD activity. Renilla construct is fused at the 3’ end to the human WT β-globin gene or a mutated β-globin carrying a PTC substitution at position 39 (NS39) that is degraded by NMD. Right panel: PCR amplification of β-globin and control (pCtrol) plasmid mRNAs in siGL and siSLU7 PLC/PRF/5 cells transfected with WT and NS39 plasmids. (B) Renilla activity of siGL and siSLU7 PLC/PRF/5 and HepG2 cells transfected with WT and NS39 β-globin plasmids. (C) Molecular pathways revealed by KEGG enrichment analysis associated with the proteins co-immunoprecipitated with SLU7 and identified by mass spectrometry. (D) Venn diagram comparing the top 145 proteins interacting with SLU7 with the 98 proteins included in KEGG category 03015 corresponding to NMD proteins. (E) UPF1 and SLU7 western blot after immunoprecipitation in PLC/PRF/5, HepG2, Hep3B and mouse liver protein extracts with anti-SLU7 antibody or control IgG. Input is shown. (F) SLU7 and UPF1 western blot after immunoprecipitation of PLC/PRF/5, H358 and HCT116 protein extracts treated or not with RNase A with anti-SLU7 antibody. Input is shown. Data are means±SEM. ∗p <0.05, ∗∗p <0.01 Mann-Whitney test. MS, Mass Spectrometry; NMD, nonsense-mediated RNA decay; PTC, premature termination codon; WT, wild-type.

Fig. 3

SLU7 is required to preserve UPF1 protein stability and NMD function. (A) Western blots of UPF1, SLU7 and Actin (loading control) in PLC/PRF/5, HepG2, H358 and HCT116 cells 48 h after transfection with siGL or siSLU7. Lower graph: quantification of UPF1 protein expression in several experiments. (B) SLU7 and UPF1 mRNAs real-time PCR quantification in the experiments used in (A). (C) NMD targets SRSF3-ISO2 and TBL2 real-time PCR quantification in PLC/PRF/5 cells 48 h after transfection with siGL or siSLU7 and an empty control plasmid (pCtrol) or the same plasmid over-expressing GFP-UPF1 (pUPF1). Lower panels: Western blot of UPF1, SLU7 and Actin (loading control) in the same transfected PLC/PRF/5 cells. The arrowhead indicates over-expressed GFP-UPF1. Data are means ± SEM. ∗p <0.05, ∗∗p <0.01, ∗∗∗p <0.001, ∗∗∗∗p <0.0001 Mann-Whitney test. NMD, nonsense-mediated RNA decay.

Fig. 4

Oxidative stress and caspase activation are implicated in SLU7 knockdown-mediated UPF1 instability and NMD inhibition. (A) UPF1, SLU7 and Actin (loading control) western blots in PLC/PRF/5 cells transfected with siGL or siSLU7 in the absence or presence of the antioxidant NAC. Lower graph: quantification of UPF1 protein in two independent experiments. (B) Renilla activity of siGL and siSLU7 PLC/PRF/5 cells transfected with WT and NS39 β-globin plasmids and treated or not with NAC. (C) PCR analysis of SRSF1 and SRSF2 mRNA isoforms in siGL and siSLU7 PLC/PRF/5 cells treated as in (A). (D) Real-time PCR analysis of different NMD targets in siGL and siSLU7 PLC/PRF/5 cells treated as in (A). (E) UPF1, SLU7, PARP and Actin (loading control) western blots in PLC/PRF/5 cells transfected with siGL or siSLU7 in the absence or presence of the pan-caspase inhibitor zVAD-fmk. The graph on the right represents the quantification of UPF1 protein in two independent experiments. (F) Real-time PCR analysis of SLU7 and different NMD targets in siGL and siSLU7 PLC/PRF/5 cells treated or not with zVAD-fmk. Data are means ± SEM. ∗p <0.05, ∗∗p <0.01 Mann-Whitney test. NAC, N-acetylcysteine; NMD, nonsense-mediated RNA decay; PTC, premature termination codon; WT, wild-type.

Fig. 5

UPF1 and SLU7 are substrates of caspases. (A) UPF1, SLU7, PARP and Actin (loading control) western blots in HeLa cells treated overnight with two different concentrations of STS in the absence or presence of the pan-caspase inhibitor zVAD-fmk. (B) SLU7, UPF1, PARP and Actin (loading control) Western blots in PLC/PRF/5 cells treated with STS, AA2 or CIS in the absence or presence of zVAD-fmk. (C) Western blot of GFP-SLU7 and SLU7 in HeLa transfected cells after treatment with cisplatin in the absence or presence of zVAD-fmk. (D) SLU7, PARP and Tubulin (loading control) Western blots in HeLa or HepaRG cells transfected with a plasmid over-expressing SLU7 (pSLU7) or a D7N mutant isoform and treatment with cisplatin or APAP respectively. AA2, apoptosis activator; APAP, acetaminophen; CIS, cisplatin; STS, staurosporin.

Fig. 6

In vivo hepatic UPF1 and SLU7 protein downregulation and NMD inhibition in mouse models of liver damage and hepatocarcinogenesis. (A) UPF1, SLU7, PARP, cleavage caspase 3 and Actin (loading control) western blots in control mouse liver and livers of mice 5 h after Fas agonistic Jo2 antibody intraperitoneal administration, without or with previous injection of zVAD. (B) Real-time PCR analysis of NMD targets in the livers described in (A). (C) UPF1, SLU7, CHOP, ATF4 and GAPDH (loading control) western blots in the liver of control (C) and 6-hour APAP-treated Slu7^+/+ and Slu7^+/- mice. Graphs represent real-time PCR levels of mRNA NMD targets in the livers of APAP-treated mice. (D) UPF1, SLU7 and GAPDH (loading control) western blots in the liver of 4.5- and 17-month-old control mice (Mdr2^+/+) and Mdr2 knockout mice (Mdr2^-/-). Samples from tumors (T) detected in the 17-month-old Mdr2^-/- mice were analyzed together with the NT liver. Graphs represents protein quantification data. Data are means ± SEM. ∗p <0.05, ∗∗p <0.01. Mann-Whitney test. APAP, acetaminophen; m, month; NMD, nonsense-mediated RNA decay; NT, non-tumoral; T, tumor.