Loss of the methylarginine reader function of SND1 confers resistance to hepatocellular carcinoma

Abstract

Staphylococcal nuclease Tudor domain containing 1 (SND1) protein is an oncogene that 'reads' methylarginine marks through its Tudor domain. Specifically, it recognizes methylation marks deposited by protein arginine methyltransferase 5 (PRMT5), which is also known to promote tumorigenesis. Although SND1 can drive hepatocellular carcinoma (HCC), it is unclear whether the SND1 Tudor domain is needed to promote HCC. We sought to identify the biological role of the SND1 Tudor domain in normal and tumorigenic settings by developing two genetically engineered SND1 mouse models, an Snd1 knockout (Snd1 KO) and an Snd1 Tudor domain-mutated (Snd1 KI) mouse, whose mutant SND1 can no longer recognize PRMT5-catalyzed methylarginine marks. Quantitative PCR analysis of normal, KO, and KI liver samples revealed a role for the SND1 Tudor domain in regulating the expression of genes encoding major acute phase proteins, which could provide mechanistic insight into SND1 function in a tumor setting. Prior studies indicated that ectopic overexpression of SND1 in the mouse liver dramatically accelerates the development of diethylnitrosamine (DEN)-induced HCC. Thus, we tested the combined effects of DEN and SND1 loss or mutation on the development of HCC. We found that both Snd1 KO and Snd1 KI mice were partially protected against malignant tumor development following exposure to DEN. These results support the development of small molecule inhibitors that target the SND1 Tudor domain or the use of upstream PRMT5 inhibitors, as novel treatments for HCC.

Keywords: HCC; SND1; arginine methylation; liver cancer; tudor domain.

Figure 1.. Generation and characterization of SND1 KO mice.

(A) Snd1 KO mouse design. CRISPR/Cas9 editing resulted in the loss of a single nucleotide (G) in codon 90 causing a frame shift, and introducing a premature stop codon in the third exon, which encodes the first SN-domain. The DNA and amino acid sequence changes are highlighted in red. (B) Western blot analysis of protein isolated from MEFs derived from a Snd1 KO heterozygous cross. Homozygous null MEFs are indicated as KO, homozygous wild-type as WT and heterozygous MEFs as HE. Top panel shows SND1, bottom panel depicts the β-actin loading control. (C) Western blot analysis of protein isolated from immortalized WT and KO MEFs. Two different anti-SND1 antibodies were used, obtained from Active Motif and Bethyl, respectively. Bottom panels depicts the β-actin loading control.

Figure 2.. Generation and characterization of SND1 KI mice.

(A) Peptide pulldown of recombinant GST-Tudor domain of SND1, wither of WT or harboring the Y766L mutation. The pulldown assay was performed using SDMA methylated (Rme2s) biotinylated glycine-arginine-rich (GAR) peptide. Equal peptide loading was validated by Ponceau S staining. (B) 293T cells were transiently transfected with GFP, GFP-SND1 (GFP-WT) and GFP-SND1 mutant (GFP-Y766L). Immunoprecipitation was performed with an anti-GFP antibody, and Western analysis was performed for SmB/B’ and GFP. (C) Snd1 KI mouse design. CRISPR/Cas9 editing was used to introduce a CTA codon in place of a TAC codon to create a Y766L substitution as highlighted in red. The nucleotide change generated a AvrII restriction enzyme site (CCTAGG in bold) which was used for subsequent genotyping. (D) Peptide pulldown from primary MEFs of WT and Snd1 KI mice (Y766L). Protein extracts from the indicated MEFs were used in a pulldown assay with SDMA methylated (Rme2s) and unmethylated (Rme0) biotinylated glycine-arginine-rich (GAR) peptides. The associated SND1 (upper panel) and the biotinylated peptide loading control were detected by western blot using anti-SND1 (SND1) antibodies and HRP-conjugated streptavidin (αBiotin).

Figure 3.. SND1 KO and KI mice are phenotypically distinct but regulate the expression of APP genes similarly.

(A) Litter size of Snd1 KO and Snd1 KI homozygous crossed mice. Points records the number of pups per litter. Pups were counted on the day of birth. n = 21, 30 and 25 for WT, KO, and KI, respectively. (B) The relative body mass of 4-week-old Snd1 KO and Snd1 KI mice resulting from heterozygous crosses. n = WT (3m, 5f), KO (3m, 5f) and KI (5m, 5f). (C) RT-qPCR of selected acute phase proteins genes, performed in triplicate for 3 WT, 2 KO, and 3 KI independent biological replicates for each genotype. Statistical t-test, two-tailed unpaired, P-value * P < 0.5; ** P < 0.01; *** P < 0.001; **** P < 0.0001.

Figure 4.. Snd1 KO and KI is hepatoprotective against DEN-induced HCC.

(A) (Top) Schematic of diethylnitrosamine (DEN) injection schedule. Only male mice were used for DEN injection studies. n = WT (28); KO (10) and KI (10). All mice survived to nine-month termination. (Bottom) Representative images of whole liver with gallbladder (scale bar = 1 cm). Ventral and dorsal respective to mouse orientation. (B) Ratio of observed pathology findings from liver lobe sections representing all 240 lobes from 48 mice (n = 50 lobes from each of KO and KI and 140 from WT). Hepatocellular carcinoma and adenocarcinoma were binned together as carcinoma. (C) Percent tumor area per total section area of all tumor types from the livers in B, obtained using ImageScope and plotted as the ratio of tumor to liver area. (D) Number of tumor foci per lobe section from the livers in b. Sections with hepatocellular hypertrophy had no foci. Right medial (RM), right lateral (RL), left medial (LM), left lateral (LL), caudate (Caud). (E) Ratio of liver plus gallbladder mass to full mouse from the livers in b. Mouse liver mass was obtained post sacrifice, prior to further manipulation. Statistical t-test, two-tailed unpaired, P-value * P < 0.05; **** P < 0.0001.