PRMT5 supports multiple oncogenic pathways in mantle cell lymphoma

Abstract

Mantle cell lymphoma (MCL) is an incurable B-cell malignancy with an overall poor prognosis, particularly for patients that progress on targeted therapies. Novel, more durable treatment options are needed for patients with MCL. Protein arginine methyltransferase 5 (PRMT5) is overexpressed in MCL and plays an important oncogenic role in this disease via epigenetic and posttranslational modification of cell cycle regulators, DNA repair genes, components of prosurvival pathways, and RNA splicing regulators. The mechanism of targeting PRMT5 in MCL remains incompletely characterized. Here, we report on the antitumor activity of PRMT5 inhibition in MCL using integrated transcriptomics of in vitro and in vivo models of MCL. Treatment with a selective small-molecule inhibitor of PRMT5, PRT-382, led to growth arrest and cell death and provided a therapeutic benefit in xenografts derived from patients with MCL. Transcriptional reprograming upon PRMT5 inhibition led to restored regulatory activity of the cell cycle (p-RB/E2F), apoptotic cell death (p53-dependent/p53-independent), and activation of negative regulators of B-cell receptor-PI3K/AKT signaling (PHLDA3, PTPROt, and PIK3IP1). We propose pharmacologic inhibition of PRMT5 for patients with relapsed/refractory MCL and identify MTAP/CDKN2A deletion and wild-type TP53 as biomarkers that predict a favorable response. Selective targeting of PRMT5 has significant activity in preclinical models of MCL and warrants further investigation in clinical trials.

Graphical abstract

Figure 1.

PRMT5 inhibition drives potent anti-MCL activity in vitro. (A) Eight MCL cell lines were treated with increasing doses of the PRMT5 inhibitor PRT-382 for 9 days. Cell viability was measured by annexin-V/propidium iodide (PI) staining and flow cytometry. (B-C) Jeko-1, Z-138, and Granta-519 were treated for 6 days with increasing doses of PRT-382. The viable cell count was measured with count beads by gating on annexin-V/PI–negative cells. (D-E) CCMCL1, Jeko-1, and Z-138 were treated with PRT-382 for 3 days at the indicated doses. Protein lysates were obtained and immunoblotted for H4R3me2s, SDMA, and ADMA. β-Actin was used as a loading control. Protein expression densitometry values normalized to β-actin and relative to dimethyl sulfoxide (DMSO) control. Dots represent individual replicate experiments. (F) CCMCL1, Jeko-1, Granta-519, and Z-138 were treated with PRT-382 at day 9 IC50 values. Protein lysates were collected at the indicated time points and immunoblotted for total SDMA with β-actin, α-tubulin, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) used as loading controls. Protein expression densitometry values for total SDMA normalized to loading control and relative to the no treatment time point 0 in lane 1. (G) Primary patient samples were cocultured with a CD40-ligand overexpressing murine fibroblast cell line and stimulated with cytokines. Cells were treated with PRT-382 for 6 days and viability and growth was measured by near-IR live/dead stain and gating on human CD19⁺/CD5⁺ MCL with flow cytometry. (H) Primary patient sample 1972 was treated with PRT-382 for 2 days at the indicated doses. Protein lysates were immunoblotted for total SDMA with β-actin used as a loading control.

Figure 2.

PRMT5 inhibition is therapeutic in xenograft models derived from patients with MCL. (A-C) PDX-DA. (A) Experimental design of PDX-DA. Treatment with VC (n = 5), ibrutinib (0.16 mg/mL in drinking water; n = 5), or PRT-382 (10 mg/kg 4D3D; n = 5) was started 26 days after engraftment. In the crossover treatment group (n = 7), PRT-382 was added at day 33 and ibrutinib was discontinued on day 45. (B) Circulating disease in peripheral blood of mice was quantified by staining for the percentage of human CD19⁺/CD5⁺ cells detected by flow cytometry. (C) Survival was assessed by Kaplan-Meier survival analysis. (D-J) PDX-AA. (D) Experimental design of PDX-AA. Treatment with VC (n = 5), ibrutinib (0.21 mg/mL in drinking water; n = 5), or PRT-382 (10 mg/kg 4D3D; n = 5) was started on day 25 after engraftment. (E) Circulating disease in peripheral blood of mice was quantified by staining for the percentage of human CD19⁺/CD5⁺ cells detected by flow cytometry. (F) Survival was assessed by Kaplan-Meier survival analysis. (G) Average spleen weight in grams after 1 and 2 weeks of treatment and at ERC. (H) Cell proliferation was assessed by staining for Ki67 in spleens of mice treated in vivo for 1 and 2 weeks. (I) Spleens of mice treated in vivo were harvested after 2 weeks of treatment and human CD19⁺ cells were isolated and immunoblotted for H4R3me2s, SDMA, and ADMA. β-Actin was used as a loading control. (J) Protein expression densitometry values normalized to β-actin and relative to VC sample in lane 1. Each dot represents an individual mouse.

Figure 3.

MTAP expression defines vulnerability to PRMT5 inhibition in MCL. (A) Baseline mRNA expression was analyzed by RT-qPCR. Data are shown relative to resting B cells and normalized to ACTB as a loading control. Data are biological duplicate experiment completed in technical quadruplicate. (B) Baseline PRMT5 mRNA expression is significantly higher in MCL cell lines and samples from patients with MCL (n = 15) compared with resting or activated normal B cells (n = 3). (C) Baseline protein expression was analyzed by immunoblot. (D) Pearson two-tailed, XY correlation between normalized protein expression values. (E-F) Simple linear regression between normalized protein expression and/or PRT-382 IC50 values, apart from cell line REC-1. (G-I) Jeko-1 and Mino were transfected with empty vector or shRNA specific to MTAP. KD was validated with RT-qPCR. ShRNA control or MTAP-transfected cell lines were treated with DMSO control or PRT-382 for 2 days. (H) Lysates were then immunoblotted using the indicated antibodies. (I) Viability and growth were measured after 5 days of treatment using annexin-V/PI and flow cytometry. (J) Western blot validation of MTAP KI and expression of SDMA upon treatment of either empty vector or MTAP-KI–transduced cells with 2 different doses of PRT-382. (K) WST-1 proliferation assay showing proliferation of Granta-519 cells transduced with either empty vector or MTAP overexpression construct in presence of 500 nM PRT-382. (L) Summary of mutations (TP53, MTAP, and ATM) and PRT-382 sensitivity in 8 MCL cell lines, 2 PDXs, and 5 primary patient samples used in this study.

Figure 3.

MTAP expression defines vulnerability to PRMT5 inhibition in MCL. (A) Baseline mRNA expression was analyzed by RT-qPCR. Data are shown relative to resting B cells and normalized to ACTB as a loading control. Data are biological duplicate experiment completed in technical quadruplicate. (B) Baseline PRMT5 mRNA expression is significantly higher in MCL cell lines and samples from patients with MCL (n = 15) compared with resting or activated normal B cells (n = 3). (C) Baseline protein expression was analyzed by immunoblot. (D) Pearson two-tailed, XY correlation between normalized protein expression values. (E-F) Simple linear regression between normalized protein expression and/or PRT-382 IC50 values, apart from cell line REC-1. (G-I) Jeko-1 and Mino were transfected with empty vector or shRNA specific to MTAP. KD was validated with RT-qPCR. ShRNA control or MTAP-transfected cell lines were treated with DMSO control or PRT-382 for 2 days. (H) Lysates were then immunoblotted using the indicated antibodies. (I) Viability and growth were measured after 5 days of treatment using annexin-V/PI and flow cytometry. (J) Western blot validation of MTAP KI and expression of SDMA upon treatment of either empty vector or MTAP-KI–transduced cells with 2 different doses of PRT-382. (K) WST-1 proliferation assay showing proliferation of Granta-519 cells transduced with either empty vector or MTAP overexpression construct in presence of 500 nM PRT-382. (L) Summary of mutations (TP53, MTAP, and ATM) and PRT-382 sensitivity in 8 MCL cell lines, 2 PDXs, and 5 primary patient samples used in this study.

Figure 4.

PRMT5 inhibition drives transcriptional reprogramming in MCL. In PDX-AA, spleens of mice treated in vivo for 2 weeks were harvested and human CD19⁺ cells were isolated and subjected to RNA sequencing. (A) Principal component analysis (PCA) of RNA sequencing data shows global differences in the transcriptome between treatment groups log tracks per million (TPM); SD > 0.9). Each dot represents and individual mouse. (B) Venn diagram depicting the number of down or up DEGs in mice treated with ibrutinib or PRT-382 in comparison to VC (q < 0.01; |log2FC| > 0.58). (C) Heatmap of the top 50 DEGs by q-value (q < 0.01; |log2FC| > 0.58). GSEA of PRT-382 in comparison with the VC, showing (D) key enrichment plots and (E) normalized enrichment scores (NES) of the top HALLMARK and REACTOME gene sets from Molecular Signatures Database (MsigDB). (F) GSEA of the top HALLMARK gene sets significantly enriched across 4 cell lines (Z-138, CCMCL1, SP53, and REC-1) treated with day 9 IC50 PRT-382 for 6 days in comparison with DMSO control.

Figure 5.

Inhibition of PRMT5 attenuates E2F1signaling in MCL. (A) Log2 fold change (FC) of DEGs in the RNA sequencing data set of PDX-AA cells of mice treated in vivo with PRT-382 in comparison with VC. (B) FC in mRNA expression of HALLMARK E2F targets analyzed by RT-qPCR in PDX (DA and AA) cells treated in vivo and in primary patient samples (7949 and 1972) treated for 3 days in vitro. Data are normalized to GAPDH and shown relative to control. PDX data are in biological triplicate (n = 3 per group) apart from VC in PDX-DA, which is a technical duplicate (n = 1 VC mouse). Primary patient samples are in biological duplicates. (C) FC in mRNA expression of HALLMARK E2F targets analyzed by RT-qPCR in Z-138 and Granta-519 treated for 6 days in vitro with the indicated doses of PRT-382. Data are normalized to ACTB and shown relative to DMSO control. (D) E2F1 was immunoprecipitated from protein lysates of Z-138 treated with PRT-382 for 4 days and patient 1972 treated with PRT-382 for 3 days. Lysates were immunoblotted for E2F1, RB, and SDMA. β-Actin was used as a loading control for the western blot (WB) input lysates. E2F1 was used as a loading control for eluted bound antigen IP lysates. (E) Z-138 and Granta-519 were treated with the indicated doses of PRT-382 for 6 days. Protein lysates were immunoblotted for p-RB(s780), CHK1, CCNA2, CDK1, and TK1, with α-tubulin used as a loading control. Protein densitometry values were normalized to α-tubulin and shown relative to 0 nM DMSO control. (F) Immunoblot of CDK1 and CCNA2 in protein lysates from PDX-AA treated with PRT-382 or ibrutinib for 2 weeks in vivo. Protein densitometry values normalized to β-actin and shown relative to VC in lane 1. (G) CSFE (carboxyfluorescein succinimidyl ester) staining of viable cells shows decreased cell replication in PRT-382–treated cells compared with DMSO control after 9 days.

Figure 6.

The inhibition of PRMT5 leads to cell cycle arrest with p53-dependent and -independent apoptotic cell death. (A) Log2FC mRNA expression analyzed by RT-qPCR in PDX-DA and PDX-AA cells treated in vivo for 2 weeks. Data are normalized to GAPDH and shown relative to VC. Data are in biological triplicates (n = 3 per group) apart from VC in the PDX-DA, which is a technical duplicate. (B) Log2FC mRNA expression analyzed by RT-qPCR in 8 MCL cell lines treated with day 9 IC50 values of PRT-382 for 6 days. Data are normalized to ACTB and shown relative to the DMSO control. (n = 3 or more biological replicates). (C) MCL cell lines were treated with 300 nM PRT-382 for 4 or 6 days, and protein lysates were immunoblotted for p53 and p21 with β-actin as a loading control. Protein densitometry values of p21 shown normalized to β-actin and relative to 0 nM DMSO control. (D) Cells were synchronized by a 24-hour incubation with 1 μM palbociclib before washing out and then treatment with PRT-382 for 5 days. Cells were stained with intracellular PI to measure DNA content by flow cytometry. The percentage of cells in G1, S, or G2 was calculated using the Michael H. Fox algorithm in the Kaluza analysis software (n = 3). (E) Granta-519 cells were treated with 50 nM PRT-382. Cells were harvested at the indicated time points and protein lysates were immunoblotted. β-Actin and GAPDH were used as loading controls. (F-G) Granta-519 and Z-138 were transfected with empty vector or shRNA specific to TP53. Cell lines were treated with DMSO control or PRT-382 then (F) immunoblotted and (G) stained with annexin-V/PI to measure the percentage of viable cells. (H) Jeko-1 cells were treated with 350 and 500 nM PRT-382. Cells were harvested at the indicated time points and protein lysates were immunoblotted. β-Actin and α-tubulin were used as loading controls. (I) Primary patient 1972 was treated with PRT-382 for 3 days, and protein lysates were then immunoblotted for protein expression with the indicated antibodies.

Figure 6.

The inhibition of PRMT5 leads to cell cycle arrest with p53-dependent and -independent apoptotic cell death. (A) Log2FC mRNA expression analyzed by RT-qPCR in PDX-DA and PDX-AA cells treated in vivo for 2 weeks. Data are normalized to GAPDH and shown relative to VC. Data are in biological triplicates (n = 3 per group) apart from VC in the PDX-DA, which is a technical duplicate. (B) Log2FC mRNA expression analyzed by RT-qPCR in 8 MCL cell lines treated with day 9 IC50 values of PRT-382 for 6 days. Data are normalized to ACTB and shown relative to the DMSO control. (n = 3 or more biological replicates). (C) MCL cell lines were treated with 300 nM PRT-382 for 4 or 6 days, and protein lysates were immunoblotted for p53 and p21 with β-actin as a loading control. Protein densitometry values of p21 shown normalized to β-actin and relative to 0 nM DMSO control. (D) Cells were synchronized by a 24-hour incubation with 1 μM palbociclib before washing out and then treatment with PRT-382 for 5 days. Cells were stained with intracellular PI to measure DNA content by flow cytometry. The percentage of cells in G1, S, or G2 was calculated using the Michael H. Fox algorithm in the Kaluza analysis software (n = 3). (E) Granta-519 cells were treated with 50 nM PRT-382. Cells were harvested at the indicated time points and protein lysates were immunoblotted. β-Actin and GAPDH were used as loading controls. (F-G) Granta-519 and Z-138 were transfected with empty vector or shRNA specific to TP53. Cell lines were treated with DMSO control or PRT-382 then (F) immunoblotted and (G) stained with annexin-V/PI to measure the percentage of viable cells. (H) Jeko-1 cells were treated with 350 and 500 nM PRT-382. Cells were harvested at the indicated time points and protein lysates were immunoblotted. β-Actin and α-tubulin were used as loading controls. (I) Primary patient 1972 was treated with PRT-382 for 3 days, and protein lysates were then immunoblotted for protein expression with the indicated antibodies.

Figure 7.

Inhibition of PRMT5 activates negative regulators of BCR/PI3K/AKT signaling. (A) PTPRO and PIK3IP1 mRNA expression measured by RT-qPCR in MCL samples (n = 10 total samples: cell lines [n = 3], PDX [n = 2], primary patient samples [n = 5], activated normal B cells [n = 5], and resting normal B cells [n=7]). Data show normalized fold expression relative to resting normal B cells (n = 7) using ACTB as internal control. (B) ChIP assay was performed with antibody H3R8me2s crosslinked magnetic beads. Retained DNA was amplified by qPCR with PTPRO and PIK3IP1 TaqMan primers. Fold enrichment was calculated relative to the input sample (PI). (C) Log2FC mRNA expression measured by RT-qPCR in PDX-DA and PDX-AA cells treated in vivo. Data are shown relative to VC treatment with ACTB as internal control. Data are in biological triplicates (n = 3 per group) apart from VC in PDX-DA is a technical duplicate. (D) Log2FC mRNA expression measured by RT-qPCR in primary patients with MCL (7949 and 1972), treated for 3 days with 500 nM PRT-382 in vitro. Data are normalized to GAPDH and shown relative to VC. Data are biological triplicates or duplicates. (E) PTPROt protein expression evaluated by immunohistochemistry (IHC) staining in spleens obtained from PDX-AA animals treated in vivo with PRT-382 or VC. Quantification of IHC PTPROt-stained slides showing the proportion of pixels positive for staining. Each dot represents an individual mouse. (F) Jeko-1 cells were treated with DMSO control or 500 nM PRT-382 for 5 days then evaluated for protein expression of PTPROt by confocal microscopy. The total number of cells counted and the percentage of cells positive for fluorescent immunolabeling of PTPROt is shown. (G) The indicated MCL cell lines were treated with day 9 IC50 values of PRT-382, and protein lysates were collected at day 6 and immunoblotted for PIK3IP1 with β-tubulin as a loading control. Protein expression densitometry values shown normalized to loading control and relative to 0 nM DMSO control–treated cells. (H) Jeko-1 cells were treated with 500 nM PRT-382 and Granta-519 cells were treated with 50 nM PRT-382 out to 8 days. Protein lysates were immunoblotted for PHLDA3, p-AKT(s473), and AKT with α-tubulin and GAPDH as loading controls.