PRMT5 activates AKT via methylation to promote tumor metastasis

Abstract

Protein arginine methyltransferase 5 (PRMT5) is the primary methyltransferase generating symmetric-dimethyl-arginine marks on histone and non-histone proteins. PRMT5 dysregulation is implicated in multiple oncogenic processes. Here, we report that PRMT5-mediated methylation of protein kinase B (AKT) is required for its subsequent phosphorylation at Thr308 and Ser473. Moreover, pharmacologic or genetic inhibition of PRMT5 abolishes AKT1 arginine 15 methylation, thereby preventing AKT1 translocation to the plasma membrane and subsequent recruitment of its upstream activating kinases PDK1 and mTOR2. We show that PRMT5/AKT signaling controls the expression of the epithelial-mesenchymal-transition transcription factors ZEB1, SNAIL, and TWIST1. PRMT5 inhibition significantly attenuates primary tumor growth and broadly blocks metastasis in multiple organs in xenograft tumor models of high-risk neuroblastoma. Collectively, our results suggest that PRMT5 inhibition augments anti-AKT or other downstream targeted therapeutics in high-risk metastatic cancers.

Fig. 1. Overexpression of PRMT5 is associated with high-risk neuroblastoma and poor patient survival and is required for neuroblastoma cell proliferation in vitro.

a The RNA expression of PRMT5 in increasing grades of neuroblastoma from the GSE49711 RNA-seq data series (Stage 1, n = 122, Stage 2, n = 91, Stage 3, n = 64, Stage 4, n = 184). Kruskal–Wallis with Dunn’s multiple comparisons test was used to determine p values, which were corrected with the Benjamini–Hochberg method.). b–d PRMT5 expression levels (left) and patient survival probability (right) in stage 3 and stage 4 neuroblastoma patients in Kocak (b, Stage 3, n = 91, Stage 4, n = 214), SEQC (c, Stage 3, n = 63, Stage 4, n = 183), and Versteeg (d, Stage 3, n = 13, Stage 4, n = 40) databases. P values were calculated with a two-sided Wilcoxon rank-sum test for boxplots (left) and p values were calculated with a log-rank test for survival curves (right). a–d Boxplot center represents mean, the box represents SD, and whiskers represent minimum and maximum. e The efficacy of PRMT5 small molecule inhibitor GSK3203591 (GSK591) was determined by MTS assay in CHLA20, NGP, and SK-N-BE (2) cells (n = 3). The IC50 value was determined by nonlinear regression (curve fit) using log₁₀ (inhibitor) versus response (three parameters) model in GraphPad Prism. f Cell viability was measured by MTS assay in a scramble or shRNA targeting PRMT5 in NGP (left) and SK-N-BE (2) cells (right) (n = 4). g Cleaved caspase-3 levels in neuroblastoma cell lines treated with increasing doses of GSK591. h Hoechst 33342 staining in cells treated with DMSO or 100 nM GSK591. Scale bars, 100 μm. i Apoptosis was measured by caspase-3/7 staining in CHLA20 cells treated with DMSO or GSK591. j Quantification of caspase-3/7 positive cells (n = 6). f, j p values were indicated by a two-tailed unpaired Student’s t-test using Microsoft Excel; error bars represent SD (e–j). e–i Representative results from three independent experiments. Uncropped immunoblots are provided in the Source Data file.

Fig. 2. Inhibiting PRMT5 methyltransferase activity attenuates primary tumor growth and metastasis.

a Schematic diagram of PRMT5 inhibitor GSK3326595 (GSK595) in vivo study in the kidney renal capsule implantation xenograft model. b Tumor mass of CHLA20 iRFP720-Luc xenograft tumors from mice treated with vehicle (n = 9) or GSK595 (n = 10) (top) and representative in vivo bioluminescent images and quantification of tumor (bottom). Tumor mass was calculated by subtracting the weight of the normal kidney from the weight of the kidney that was implanted with tumor cells. Representative ex vivo images of the tumor by light microscope (c) or bioluminescence (d). e Tumor mass of NGP iRFP720-Luc xenograft tumors from mice treated with vehicle (n = 7) or GSK595 (n = 9) (top) and representative in vivo bioluminescent images and quantification of tumor (bottom). Representative ex vivo images of the tumor by light microscope (f) or bioluminescence (g). h Representative ex vivo bioluminescent images of liver from mice bearing CHLA 20 or NGP xenograft tumors treated with vehicle or GSK595. i Representative charts of FACS analysis of iRFP720 positive human neuroblastoma cells in hepatocytes isolated from the whole liver from CHLA20 xenografted mice, NC negative control. j Quantification of the percentage of iRFP720 positive tumor cells determined by flow cytometry in hepatocytes isolated from the whole liver from CHLA20 xenografted mice. Vehicle, n = 9, GSK595, n = 7, NC, n = 4. p values were indicated by a two-tailed unpaired Student’s t-test using Microsoft Excel; error bars represent SD (b, e, j).

Fig. 3. PRMT5 inhibition impairs AKT signaling.

a Proteomics-based pathway screening by IPAD platform (immuno-paired-antibody detection assay) for the expression or modification of key proteins involved in more than 20 signaling pathways. Signals were normalized to internal levels of GAPDH and beta Tubulin. Heatmap showed differences between DMSO and GSK591 groups based on the average value from two independent experiments (n = 2 biologically independent samples). The color scale represented log2 fold changes over DMSO treatment. b Immunoblots showing phosphorylation of AKT, and its downstream targets phospho-GSK3α and phospho-GSK3β in CHLA20 and NGP cells treated with DMSO or increasing doses of GSK591. c Western blots of AKT phosphorylation and AKT downstream targets phospho-GSK3α and phospho-GSK3β in the presence and absence of PRMT5 in NGP or BE2 cells harboring scramble or shPRMT5. d Levels of phosphorylated AKT, GSK3α, and GSK3β with or without EGF stimulation in DMSO and GSK591-treated CHLA20 and NGP cells. e Activated AKT and its targets phospho-GSK3α and phospho-GSK3β were detected by Western blotting in xenograft tumors from vehicle or GSK595-treated mice. b–d Representative results from three independent experiments. Uncropped immunoblots are provided in the Source Data file.

Fig. 4. PRMT5 methylates AKT.

a PRMT5/AKT interaction captured by co-immunoprecipitation (co-IP). The lysate was immunoprecipitated with anti-PRMT5 antibody followed by immunoblotting with anti-AKT antibody in CHLA20 and NGP cells (top), and the reciprocal co-immunoprecipitation was shown in the middle. BRG1 and bait protein PRMT5 were shown as positive controls for the PRMT5 IP, whereas GSK3β and bait protein AKT served as positive controls for AKT IP. The input was used as internal controls (bottom). b Immunoprecipitation of AKT1 followed by a Western blotting analysis of symmetric dimethylarginine (SDMA) of AKT1, phosphorylation of AKT1 on Thr308 and Ser473 in DMSO or GSK591-treated CHLA20 and NGP cells. c AKT1 phosphorylation was detected by immunoblotting in a scramble or PRMT5 knockdown cells when forced expressing a wild-type PRMT5 or an enzymatic deficient form of PRMT5. d Analysis of AKT1 phosphorylation in DMSO or GSK591-treated CHLA20 and NGP cells expressing PRMT5 wild type or enzyme dead mutant. e In vitro methylation assay showing the methylation of AKT1 wild type and R15K mutant by recombinant PRMT5/MEP50 (top), Ponceau S staining of the membrane showing equal loading of each sample (middle), and Western blotting analysis showing an equal amount of HA-tagged proteins pulled down by anti-HA beads (bottom). f Analysis of SDMA and phosphorylation of AKT1 wild type or R15K mutant in CHLA20 and NGP cells with (right) or without (left) EGF stimulation. Cells were transfected with AKT1 wild type or R15K mutant. In the case of EGF stimulation, 24 h post-transfection, cells were serum-starved overnight and then treated with 10 ng/mL EGF for 15 min before harvest. a–d, f Representative results from three independent experiments. e Representative results from two independent biological samples. Uncropped immunoblots are provided in the Source Data file.

Fig. 5. PRMT5-mediated AKT1-R15 methylation is required for its activation.

a Western blotting analysis of the distribution of activated AKT1 in cytosolic or membrane fractions from DMSO or GSK591-treated CHLA20 cells in the absence or presence of EGF. b The presence of activated AKT1 in cytosolic and membrane fractions from scramble or PRMT5 knockdown cells with or without EGF stimulation. c Colocalization of AKT1 wild type or R15K mutant with plasma membrane by immunofluorescence visualized under confocal microscopy in CHLA 20 cells in the absence or presence of EGF. Scale bars, 100 μm. d The binding to phosphatidylserine (PS) of AKT1 wild type or R15K mutant was examined by the presence of HA-tagged protein in the eluate or flowthrough fraction after incubation with phosphatidylserine coated agarose beads by Western blotting. Top, elution from PS beads; middle, flowthrough after incubation; bottom, HA-tagged protein eluted from anti-HA beads. e The association of AKT1 wild type or R15K mutant with PDK1 or mTORC2 was analyzed by Western blotting in CHLA20 with or without EGF stimulation. Exogenous AKT1 wild type or R15K mutant was pulldown by HA beads, and the precipitants were analyzed by Western blotting for PDK1 and SIN1. f The interaction of AKT1 with PDK1 and mTORC2 was measured by immunoprecipitation of endogenous AKT1 followed by Western blotting against PDK1 and SIN1 in SK-N-BE(2) and NGP cells with or without PRMT5 knockdown. g The recruitment of PDK1 and mTORC2 was examined in DMSO or GSK591-treated CHLA20 and NGP cells. a–g Representative results from three independent experiments. Uncropped immunoblots are provided in the Source Data file.

Fig. 6. PRMT5/AKT regulates the EMT program.

a Immunoblots showing the expression of TWIST1, SNAIL, and ZEB1 in neuroblastoma cells treated with DMSO or increasing doses of GSK591 in CHLA20 and NGP cells. b Analysis of TWIST1, SNAIL, and ZEB1 protein levels in the scramble and shPRMT5 cells. c The expression of EMT transcription factor TWIST1, SNAIL, and ZEB1 was examined in xenograft tumors from mice treated with vehicle or GSK595 (n = 4). d Representative images of DMSO or GSK591-treated cells migrated to the cleared space (wound) after 24 h. Scale bars, 100 μm. e Quantification of in vitro cell migration assay (n = 6). The migration area was determined by measuring the total area of the wound using the ImageJ software. f Representative images of DMSO or GSK591-treated cells invaded to ECM coated membrane in transwell invasion assay. Scale bars, 100 μm. g Percentage of invasive cells normalized by cell numbers in the non-ECM coated 12-well plate using ImageJ (n = 6). h Protein levels of TWIST1 and SNAIL were analyzed in DMSO or GSK591-treated cells transfected with a vector or constitutively activated AKT1. i The protein levels of TWIST1 and SNAIL in cells overexpressing PRMT5 wild type or enzyme activity deficient mutant by Western blotting. j Representative images of CHLA20 cells overexpressing vector, wild type PRMT5, or an enzymatic deficient form of PRMT5 invaded to ECM coated membrane in the transwell invasion assay (left). Scale bars, 100 μm. Percentage of invasive cells normalized by cell numbers in the non-ECM coated 12-well plate using ImageJ (n = 6) (right). e, g p values were calculated by two-tailed unpaired Student’s t-test using Microsoft Excel. j p values were determined using one-way ANOVA with Tukey’s multiple comparisons test. Error bars represent SD. a, b, d–j Representative results from three independent experiments and the results shown are from a representative experiment. Uncropped immunoblots are provided in the Source Data file.

Fig. 7. PRMT5 depletion effectively blocks tumor cell metastasis to the liver and lung.

a Bioluminescent imaging of mice tail vein injected SK-N-BE (2) shPRMT5 cells with or without doxycycline treatment (n = 6). b Bioluminescent (upper) and fluorescent imaging (lower) of livers harvested from mice described in a. c Bioluminescent (upper) and fluorescent imaging (lower) of lungs harvested from mice described in a. d FACS analysis of iRFP720+ human neuroblastoma cells in the liver from mice with or without doxycycline treatment (n = 6). e FACS analysis of iRFP720+ human neuroblastoma cells in the lung from mice with or without doxycycline treatment (n = 6). d, e p values were determined by a two-tailed unpaired Student’s t-test using Microsoft Excel. Error bars represent SD.

Fig. 8. Graphic summary of the mechanism by which PRMT5 regulates AKT activation to promote metastasis.

AKT is activated in a cascade of events. Upstream stimuli, such as growth factors and cytokines, induce the production of phosphatidylinositol (3,4,5) trisphosphates (PIP3) by phosphoinositide 3-kinase (PI3K). These phospholipids serve as plasma membrane docking sites for AKT and PDK1 at pleckstrin-homology (PH) domains. PDK1 phosphorylates AKT at Thr308 at the plasma membrane which partially activates AKT, whereas AKT is fully activated after phosphorylation at Ser473 by mTORC2. PRMT5 methylates AKT1 on R15 in the PH domain, by which it promotes AKT1 association with the plasma membrane and subsequent phosphorylation by PDK1 and mTORC2. Downstream of AKT signaling, PRMT5 increases the expression of EMT transcription factors, such as SNAIL, ZEB1, and TWIST1 augmenting the EMT program to promote tumor metastasis.