Sulforaphane inhibits PRMT5 and MEP50 function to suppress the mesothelioma cancer cell phenotype

Abstract

Mesothelioma is a highly aggressive cancer of the mesothelial lining that is caused by exposure to asbestos. Surgical resection followed by chemotherapy is the current treatment strategy, but this is marginally successful and leads to drug resistant disease. We are interested in factors that maintain the aggressive mesothelioma cancer phenotype as therapy targets. Protein arginine methyltransferase 5 (PRMT5) functions in concert with the MEP50 cofactor to catalyze symmetric dimethylation of key arginine resides in histones 3 and 4 which modifies the chromatin environment to alter tumor suppressor and oncogene expression and enhance cancer cell survival. Our studies show that PRMT5 or MEP50 loss reduces H4R3me2s formation and that this is associated with reduced cancer cell spheroid formation, invasion and migration. Treatment with sulforaphane (SFN), a diet-derived anti-cancer agent, reduces PRMT5/MEP50 level and H4R3me2s formation, and suppresses the cancer phenotype. We further show that SFN treatment reduces PRMT5 and MEP50 levels and that this reduction is required for SFN suppression of the cancer phenotype. SFN treatment also reduces tumor formation which is associated with reduced PRMT5/MEP50 expression and activity. These findings suggest that SFN may be a useful mesothelioma treatment agent that operates, at least in part, via suppression of PRMT5/MEP50 function.

Introduction

Mesothelioma is an aggressive and lethal cancer of the mesothelial lining of the pleura and peritoneum that develops in response to asbestos exposure ^1,2. The accepted treatment for mesothelioma is surgical removal of the tumor followed by chemotherapy ³. However, this strategy is not highly efficient, as aggressive and drug-resistant cancer subsequently develops ² leading to extremely poor post-therapy survival ⁴. An extensive literature indicates that cancer stem cells maintain tumors and mediate drug resistance ^5-7. We have identified mesothelioma cancer stem cells (MCS cells) ^8,9 that comprise 0.15% of tumor cell population ⁸. These cells are highly aggressive and drive vigorous spheroid formation, invasion, migration and tumor formation as compared to non-stem cancer cells ^8,9. Our goal is to identify proteins that are important for survival of MCS cells and non-stem cancer cells as cancer prevention and therapy targets.

PRMT5 is a type II methyltransferase that functions in multiple protein complexes ¹⁰. It symmetrically dimethylates histones H3R8 and H4R3 in tumor suppressor genes leading to transcriptional gene silencing ^10-12. PRMT5 functions in conjunction with methylosome protein 50 (MEP50), a WD repeat (tryptophan-aspartic acid)-containing protein cofactor, that is necessary for PRMT5 activity ^13,14. Elevated MEP50 and PRMT5 expression is associated with cancer development ^11,12,15-18. For example, the PRMT5/MEP50 complex has an important regulatory role in human epidermal keratinocytes ^19,20 and elevated expression drives cell proliferation and tumor formation in epidermal squamous cell carcinoma ^19-21.

Sulforaphane (1-isothiocyanato-4-(methylsulfinyl) butane, SFN) is a promising cancer prevention and therapy agent that is abundant in cruciferous vegetables ²² that is highly bioavailable and has no known side effects ^23-26. SFN has been shown to impact epigenetic mechanisms, including PRMT5/MEP50 function. For example, SFN suppresses PRMT5 and MEP50 function in squamous cell carcinoma to reduce cancer cell survival and tumor formation ²¹.

Although it has been suggested that PRMT5 may be a suitable treatment target in mesothelioma ²⁷ no studies have been performed. In the present study, we examine the ability of PRMT5 and MEP50 to maintain mesothelioma cancer cell function. We show that PRMT5/MEP50 knockdown reduces cell proliferation, invasion, migration, and spheroid formation. We further show that SFN treatment reduces PRMT5 and MEP50 levels and that this reduction is required for SFN suppression of the cancer phenotype. Moreover, SFN treatment of tumor xenografts reduces PRMT5/MEP50 level, H4R3me2s formation and tumor growth, suggesting that these proteins are important in vivo targets. These studies suggest that the PRMT5/MEP50 complex is required for optimal mesothelioma cell tumor formation, that SFN interferes with PRMT5/MEP50 function to reduce tumor formation and that SFN may be a treatment agent for mesothelioma.

Materials and Methods

Chemicals and Reagents

RPMI1640 medium (27519003), L-glutamine, penicillin-streptomycin and 0.25% trypsin-EDTA were purchased from Gibco (Gaithersburg, MD). Growth medium for monolayer and spheroid growth is RPMI1640 supplemented with penicillin-streptomycin and 10% fetal bovine serum (FBS, 19B416). Sulforaphane (SFN), 1-isothiocyanato-4-(methylsulfinyl) butane (S8044), was purchased from LKT laboratories (St. Paul, MN) and dissolved in dimethyl sulfoxide and stored at −20 °C. BD Matrigel matrix (354234) and BD BioCoat Millicell inserts (353097) were purchased from BD Bioscience (San Jose, CA). Hoechst (33258) stain and Alexa 555-conjugated goat anti-rabbit IgG (A21430) were purchased from Invitrogen (Carlsbad, CA). Control- (D-001206-13-20), PRMT5- (M-015817-02-0005), and WDR77/MEP50-siRNA (M-006895-01-0005) were obtained from Dharmacon (Lafayette, CO). Mouse monoclonal anti-PRMT5 (sc-376937) was purchased from Santa Cruz Biotechnology (Dallas, TX), and rabbit monoclonal anti-MEP50 (D56B8) was purchased from Cell Signaling Technology (Danvers, MA). Rabbit anti-H4R3me2s (ab5823) and anti-H4 (ab10158) were purchased from Abcam (Cambridge, UK). Mouse monoclonal anti-β-actin (A5441) and lactacystin were purchased from Sigma (St. Louis, MO). Peroxidase-conjugated anti-mouse IgG (NXA931) and anti-rabbit IgG (NA934V) were obtained from GE Healthcare (Chicago, IL). The PRMT5 or MEP50 inserts were cloned into pcDNA3 to produce the pcDNA3-hMEP50-FLAG and pcDNA3-cMYC-hPRMT5 expression vectors. The PRMT5 inhibitor, GSK-3326595, was obtained from Chemietek (Indianapolis, IN, CT-GSK332). Meso-1 and NCI-Meso-17 cells are derived from peritoneal and plural mesothelioma respectively ^9,28. Mycoplasma testing is performed when new cell stocks are thawed for use.

Immunology methods

Cultured cells and tumor samples were harvested, washed and dissolved in Laemmli sample buffer (0.0625 M Tris-HCl, pH 7.5, 10% glycerol, 5% SDS, 5% β-mercaptoethanol). Equivalent amounts of protein were electrophoresed on denaturing 12% polyacrylamide gels and transferred to nitrocellulose membrane for immunoblot. Primary antibody binding was visualized using ECL chemiluminescence detection reagent (RPN2106) obtained from Amersham (Little Chalfont, UK). For immunofluorescence detection of H4R3me2s, paraffin-embedded tumor slices were incubated overnight at 4 °C with rabbit anti-H4R3me2s primary antibody (1:500) followed by a 2 h incubation with Alexa 555-conjugated goat anti-rabbit IgG secondary antibody (1:400). The fluorescence signal was detected using an inverted confocal microscope following Hoescht counterstain.

Quantitative qRT-PCR

Cells were treated with 0 or 20 μM SFN for 48 h and total RNA was isolated using Illustra RNAspin Mini Kit (25-0500-71) from GE Healthcare (Chicago, IL) and 1 μg of RNA was used for cDNA synthesis. mRNA level was measured using the Light Cycler 480 SYBR Green I Master mix (Roche Diagnostics, Basel, CH). PRMT5 and MEP50 mRNA levels were normalized using cyclophilin A mRNA level. The gene specific primers for mRNA detection include: PRMT5 (forward, 5′-TGA GGC CCA GTT TGA GAT GCC TTA; reverse, 5’-AGT AGC CGG CAA AGC CAT GTA GTA), MEP50 (forward, 5′-TTG CTC AGC AGG TGG TAC TGA GTT; reverse, 5′-AAT CTG TGA TGC TGG CTT GGG ACA), and cyclophilin A (forward: 5’-CAT CTG CAC TGC CAA GAC TGA, reverse: 5’-TTC ATG CCT TCT TTC ACT TTGC) ²¹.

Electroporation

Cells (1.2 million cells/electroporation) were harvested with trypsin, washed with phosphate buffered saline and resuspended in 100 μl of nucleofection solution (VPD-1002) containing 1.5 or 3 μg of siRNA or 3 μg of expression vector. The solution was gently mixed and electroporated using the T-018 setting on the AMAXA Nucleofactor II electroporator (Basel, CH). Cells were electroporated twice with siRNA for knockdown experiments and single electroporated with plasmid for protein overexpression experiments ^29,30. After electroporation, the cells were resuspended in pre-warmed growth medium (500 μl) and transferred to a 10 cm dish containing 10 ml of growth medium. After recovery, these cells were plated for assay of proliferation, spheroid formation, invasion and migration.

Biological assays

To cell proliferation assays, cells were plated at 20,000 cells per well in a 6-well dish in triplicate and counted every 24 h using a hemocytometer. To measure spheroid formation, monolayer cultures of cells were harvested and plated at 40,000 cells per 9.5 cm² well in ultra-low attachment dishes in serum-supplemented growth medium and spheroid formation was monitored from 0 - 3 days. Mesothelioma cells require serum-supplemented media to form spheroids ⁸. For invasion assay, Matrigel was diluted in 0.01 M Tris-HCl/0.7%NaCl to 250 μg/ml and 120 μl was aliquoted atop the membrane in each BD BioCoat Millicell chamber. Cells were seeded at 25,000 cells/chamber in growth media supplemented with 1% FBS. Medium containing 10% FBS was added to the lower well chamber and cells permitted to migrate for 0 - 24 h at 37 C. Invading cells were visualized by fixing the membrane with 4% paraformaldehyde, staining with Hoechst (diluted 1:2000) before cell detection using an inverted fluorescent microscope. For migration, 250,000 - 300,000 cells were plated at confluent density in 24-well dishes in growth medium. After 24 h the confluent monolayers were wounded using a 200 μl micropipette tip and closure of the wound was monitored from 0 - 24 h. Statistics were performed on triplicate samples using the students t-test.

Tumor xenograft growth assays

Spheroid derived cells were resuspended in phosphate buffered saline containing 30% matrigel and 100 μl (3 million cells) were injected subcutaneously into each front flank of five NOD scid IL2 receptor gamma chain knockout (NSG) mice (5 mice x 2 tumor/mouse = 10 tumors/treatment group) using a 26.5 gauge needle. When tumors were first palpable, at 9 weeks post-injection, treatment was initiated with 0 or 25 μmoles/dose SFN delivered by oral gavage three times per week (M/W/F). Tumor growth was monitored from weeks 9 to 12 using calipers and tumor size was reported as volume = 4/3π x (diameter/2)³. The tumors were harvested at 12 weeks photographed and samples were harvested to prepare extracts for immunoblot and immunostaining. Animal studies were reviewed and approved by the Institutional Animal Care and Use Committee and followed standard international practices for treatment of animals.

Results

SFN suppresses PRMT5/MEP50 function and attenuates the cancer cell phenotype

Mesothelioma cells are highly aggressive and display vigorous spheroid formation, invasion and migration ^8,9. To assess the role of PRMT5/MEP50 in maintaining this phenotype, we examine the impact of loss of these proteins. Meso-1 cells, derived from peritoneal mesothelioma ⁹, were electroporated with control-, PRMT5-, MEP50- or PRMT5/MEP50-siRNA. As shown in Fig. 1A, loss of PRMT5 or PRMT5/MEP50 is associated with a marked reduction in H4R3me2s formation, while loss of MEP50 results in a partial reduction. PRMT5 and PRMT5/MEP50 loss is associated with reduced cell proliferation, spheroid formation, invasion and migration (Fig. 1B/C/D/E). MEP50 knockdown reduces spheroid formation and invasion but is less effective at suppressing cell proliferation and migration (Fig. 1B/C/D/E). We attribute the lesser response to the fact that MEP50 is only partially reduced. These findings suggest that PRMT5/MEP50 function is required to maintain the aggressive cancer cell phenotype. To provide additional evidence of a role for PRMT5, we show that treatment with GSK-3326595, a PRMT5 inhibitor, reduces H4R3me2s formation and cell proliferation (Fig. 1F/G).

Fig. 1.

PRMT5, MEP50 and SFN impact on the Meso-1 cell cancer phenotype. A/B/C/D/E Meso-1 cells were treated with control-, PRMT5-, MEP50-, or combined PRMT5/MEP50-siRNA and then plated for immunoblot, proliferation, spheroid formation, invasion and migration assays. F/G Meso-1 cells were treated with 0 or 100 μM GSK-3326595, a PRMT5 inhibitor, and PRMT5 activity and cell growth were monitored. H/I/J/K/L Cells were treated with 0 or 20 μM SFN for 48 h and extracts were prepared to monitor epitope levels and assays were performed to monitor the impact on proliferation, spheroid formation, invasion and migration. The values are mean ± SEM and the asterisks indicate a significant reduction relative to control, n = 3, p < 0.001 except for panel C where n = 3, p < 0.03. In this and all other figures the spheroid and proliferation images are 40X and the invasion and migration images are 100X magnification. The bar in panel J = 100 microns.

We next examined the impact of SFN on PRMT5/MEP50 level and activity. As shown in Fig. 1H, treating with SFN markedly reduces PRMT5 and MEP50 level, and this is associated with reduced H4R3me2s formation. Moreover, as shown in Fig. 1I/J/K/L, the SFN dependent loss of PRMT5/MEP50 is associated with reduced cell proliferation, spheroid formation, invasion and migration.

To assure that these responses can be generalized, we tested the role of PRMT5/MEP50 in NCI-Meso-17 cells which are derived from pleural mesothelioma ²⁸. Fig. 2A shows that PRMT5 and MEP50 levels are reduced in cells treated with the corresponding siRNAs, and that H4R3me2s formation is markedly reduced in PRMT5 knockdown cells. Moreover, loss of PRMT5 or MEP50 is associated with reduced spheroid formation, invasion and migration (Fig. 2B/C/D). In addition, treatment with PRMT5 inhibitor reduces H4R3me2s and cell proliferation (Fig. 2E/F).

Fig. 2.

PRMT5, MEP50 and SFN impact on the NCI-Meso-17 cell cancer phenotype. A NCI-Meso-17 cells were treated for 48 h with control-, PRMT5-, MEP50- or PRMT5/MEP50-siRNA and protein levels were monitored. B/C/D NCI-Meso-17 cells were treated with 3 μg of control-, PRMT5-, MEP50- or PRMT5/MEP50-siRNA and then plated for spheroid formation, invasion and migration assays. E/F NCI-Meso-17 cells were seeded for proliferation assay and treated with 0 or 100 μM GSK-3326595, a PRMT5 inhibitor, for 0 - 72 h. G/H/I NCI-Meso-17 cells were treated with 0 or 20 μM SFN for 48 h and extracts were prepared to monitor the level of the indicated proteins. Assays were performed to measure SFN impact on invasion and migration. The values are mean ± SEM and the asterisks indicate a significant reduction relative to control, n = 3, p < 0.001, except for panel B where n = 3, p < 0.05. and panel F where n = 3, p < 0.01.

We next monitored the impact of SFN treatment on PRMT5/MEP50 function. As shown in Fig. 2G, SFN treatment reduces PRMT5/MEP50 and H4R3me2s formation and this is associated with reduced invasion and migration (Fig. 2H/I). Thus, SFN suppresses PRMT5/MEP50 level and H4R3me2s formation in both Meso-1 and NCI-Meso-17 cells.

PRMT5/MEP50 loss is required for SFN action

SFN has multiple actions in cancer cells ^31-35. Therefore, it is important to determine if loss of PRMT5/MEP50 function is required for SFN suppression of the cancer phenotype. We therefore treated cells with SFN in the presence or absence of forced PRMT5 and MEP50 expression. Fig. 3A shows the levels of PRMT5 and MEP50 following electroporation of Meso-1 cells with PRMT5 and MEP50 expression plasmids. The increase in PRMT5 is readily detected but MEP50 is less obvious. Fig. 3B/C/D show that forced expression of PRMT5 or MEP50 antagonizes SFN suppression of spheroid formation, invasion and migration. Figs. 3E/F/G/H show a similar pattern of response for NCI-Meso-17 cells. Thus, PRMT5 and MEP50 overexpression increases H4R3me2s formation and this antagonizes SFN suppression of the cancer phenotype, suggesting that PRMT5/MEP50 loss is required for SFN action.

Fig. 3.

Reduced PRMT5/MEP50 activity is required for SFN suppression of the cancer cell phenotype. A Meso-1 cells were electroporated with 3 μg empty vector (EV), PRMT5, MEP50 or PRMT5/MEP50 expression vectors and after 48 h PRMT5, MEP50, and H4R3me2s levels were measured by immunoblot. B/C/D Control and PRMT5, MEP50 and PRMT5/MEP50 over producing cells were treated with 0 or 20 μM SFN and the impact on spheroid formation, invasion and migration was monitored. E NCI-Meso-17 cells were electroporated with 3 μg empty vector (EV), PRMT5, MEP50 or PRMT5/MEP50 expression vectors and after 48 h PRMT5, MEP50, and H4R3me2s levels were measured by immunoblot. F/G/H Control and PRMT5, MEP50 and PRMT5/MEP50 over producing NCI-Meso-17 cells were treated with 0 or 20 μM SFN and the impact on spheroid formation, invasion and migration was monitored. The values are mean ± SEM and the single asterisks indicate a significant reduction relative to control and double asterisks indicate a significant increase relative to the SFN treated group, n = 3, p < 0.001.

SFN regulation of PRMT5/MEP50

Our findings show that SFN treatment markedly reduces PRMT5 and MEP50 levels in Meso-1 and NCI-Meso-17 cells and that this is associated with reduced H4R3me2s formation (Figs. 1H and 2G). The loss of these epigenetic regulators could be due to proteolytic degradation or an alteration in mRNA level. To assess the role of proteasome degradation, we treated cells with SFN in the presence of 0 or 1 μM lactacystin, a proteasome inhibitor. As shown in Fig. 4A, lactacystin treatment does not reverse the SFN suppression of PRMT5 or MEP50 in Meso-1 or NCI-Meso-17 cells. We next examined the impact of SFN treatment on PRMT5/MEP50 mRNA level. Fig. 4B shows that SFN treatment reduces PRMT5 and MEP50 mRNA in both cell types. This suggests that SFN suppression of PRMT5/MEP50 level may be due to reduced transcription or enhanced RNA degradation.

Fig. 4.

SFN reduces PRMT5 and MEP50 mRNA level. A Meso-1 and NCI-Meso-17 cells were treated with 0 or 20 μM SFN in the presence or absence of lactacystin and after 48 h extracts were prepared to detect levels of PRMT5 and MEP50. B Meso-1 and NCI-Meso-17 cells were treated for 48 h with 0 or 20 μM SFN and total RNA was prepared for qRT-PCR quantification of PRMT5 and MEP50 mRNA. The values are mean ± SEM and the asterisks indicate a significant reduction compared to control, n = 3, p < 0.001.

SFN suppresses PRMT5/MEP50 signaling in tumors

We were particularly interested to determine if SFN regulates PRMT5 and MEP50 level and activity in tumors. To test this, we produced xenograft tumors in NSG mice and then treated with SFN. The tumors are palpable at 9 weeks post-cell injection at which time SFN treatment is initiated until the time of tumor harvest at 12 weeks. These studies show that SFN treatment markedly reduces tumor formation (Fig. 5A) and that this is associated with reduced PRMT5 and MEP50 mRNA and protein (Fig. 5B/C) and reduced H4R3me2s formation as measured by immunoblot (Fig. 5C) and immunostaining (Fig. 5D).

Fig. 5.

SFN suppresses PRMT5/MEP50 activity and tumor formation. A Spheroid-derived cells (3 million) were injected subcutaneously in NSG mice and at 9 weeks, when tumors were first palatable, treatment was initiated with 0 or 20 μmoles/dose SFN three times per week (M/W/F) by oral gavage and at 12 weeks the tumors were harvested. The tumor volumes are mean ± SEM (n = 10 tumors, 2 tumors/mouse) and the asterisks indicate a significant reduction in tumor size compared to control (p < 0.001). B Tumor RNA was prepared and PRMT5 and MEP50 mRNA levels were determined by qRT-PCR. The values are mean ± SEM and the asterisks indicate a significant reduction compared to control, n = 5, p < 0.001. C Extracts were prepared from two independent (12 wk) tumor sets for immunoblot detection of PRMT5/MEP50 and H4R3me2s. D Tumor sections were prepared for H/E staining, detection of DAPI-stained nuclei, and H4R3me2s distribution was detected by immunostaining with rabbit anti-H4R3me2s. E Schematic of SFN regulation of PRMT5/MEP50 function. PRMT5 or MEP50 knockdown reduces cancer cell spheroid formation, invasion and migration, indicating that PRMT5/MEP50 act to maintain the cancer phenotype. This schematic illustrates that SFN suppresses PRMT5/MEP50 level and activity as a mechanism to reduce spheroid formation, invasion, migration, and tumor formation.

Discussion

PRMT5 and MEP50

Modulation of chromatin structure involves chromatin-remodeling factors that post-translationally modify histones to control transcription factor access to DNA ³⁶. These regulators include histone-modifying enzymes, ATP-dependent chromatin remodelers and DNA-modifying enzymes. Lysine acetylation and methylation are important modifications ³⁶. Protein arginine methyltransferases (PRMTs) methylate specific histone arginine residues to modulate gene expression and regulate a host of processes including signaling, differentiation, apoptosis and tumorigenesis ^37,38. Type I PRMTs (PRMT1-4, PRMT6 and PRMT8) asymmetrically dimethylate arginine while type II enzymes (PRMT5, PRMT7 and PRMT9) symmetrically dimethylate arginine ³⁶.

Among these enzymes PRMT5 is often elevated in cancer cells and has been shown to have an important enabling role in cancer ³⁶ where it regulates signaling pathways that modulate cell death and malignant transformation ^21,39-42. PRMT5 catalyzes symmetric dimethylation of H4R3me2s and H3R8me2s ⁴³ which results in closure of the surrounding chromatin to facilitate gene silencing. The H4R3me2s mark is recognized by the DNA methyltransferase DNMT3a which binds and then methylates neighboring CpG dinucleotides to further enforce repression ⁴⁴. PRMT5 is unique as it forms a complex with methylosome protein 50 (MEP50) to form the active enzymatic complex. This octameric complex consists of four MEP50 and four PRMT5 subunits ^13,43.

PRMT5/MEP50 are required to maintain the cancer cell phenotype

Although PRMT5 has been mentioned as a potential epigenetic target in mesothelioma ²⁷ no studies have examined its role in mesothelioma disease models. MCS cells comprise a small subpopulation (0.15%) of highly aggressive cells in mesothelioma tumors ⁸. We produced a highly enriched population of MCS cells by growth on ultra-low attachment plates ⁸ and examined the impact on the spheroid formation process. These studies show that PRMT5 and MEP50 function to maintain spheroid formation, as PRMT5 or MEP50 knockdown reduces spheroid formation. We also show that PRMT5 and MEP50 knockdown reduces Meso-1 (peritoneal) and NCI-Meso-17 (pleural) cell proliferation, invasion and migration and that this is associated with reduced H4R3me2s formation. These findings indicate that PRMT5/MEP50 function is required to maintain the aggressive cancer stem cell and non-stem cancer cell phenotypes.

SFN treatment reduces PRMT5 level and activity, and tumor formation

SFN is a promising diet-derived cancer prevention and therapy agent that is present at high levels in cruciferous vegetables ³¹. SFN is readily ingested, has minimal side actions and displays high bioavailability in mice and humans ^23,25,45. We show that SFN suppresses PRMT5 and MEP50 level, and H4R3me2s formation, in peritoneal (Meso-1) and pleural (NCI-Meso-17) mesothelioma cancer cells. Previous studies show that PRMT5 level can be controlled by microRNA ^12,46 and proteasome ²¹ related mechanisms. Our present studies show that SFN treatment reduces PRMT5 and MEP50 mRNA levels in Meso-1 cells and NCI-Meso-17 cells. These finding are consistent with a previous report showing that SFN suppresses PRMT5 and MEP50 mRNA content in squamous cell carcinoma ²¹. Additional studies will be necessary to determine if the mRNA loss is due to reduced transcription or enhanced mRNA degradation. SFN has also been reported to reduce PRMT5 and MEP50 levels via proteasome degradation in squamous cell carcinoma ²¹. Our present studies show that treatment of mesothelioma cancer cells with proteasome inhibitor does not antagonize the SFN-dependent loss of PRMT5/MEP50, suggesting that the proteasome does not have a role in these models. On balance, these findings suggest that SFN suppression of PRMT5 and MEP50 mRNA is an important mechanism regulating the level of these proteins.

SFN, like other diet-derived anti-cancer agents, modulates multiple intracellular signaling events ^31-35. For this reason, it is important to determine if PRMT5/MEP50 loss is essential for SFN suppression of the cancer cell phenotype. To test this, PRMT5 and MEP50 were overexpressed in cells challenged with SFN. PRMT5 or MEP50 overexpression antagonized SFN suppression of spheroid formation, invasion and migration, suggesting that loss of PRMT5/MEP5 function is an important event that is required for optimal SFN suppression of the cancer phenotype.

To demonstrate clinical relevance, we examined the SFN impact on PRMT5 and MEP50 function in tumors. We note that Meso-1 cells used for injection were grown as spheroids on ultra-low attachment plates to enrich for the MCS cell population ⁸. This study shows that SFN treatment can reduce PRMT5 and MEP50 function, H4R3me2s formation and tumor formation of this aggressive cell population. Thus, the tumor studies reflect the response in cultured cells and suggest that the PRMT5/MEP50 complex is required for aggressive tumor formation and that SFN reduction of this activity contributes to the SFN suppression of tumor formation.

Our findings suggest that PRMT5 and MEP50 are required for optimal maintenance of an aggressive cancer phenotype, and that SFN suppression of PRMT5/MEP50 level and/or activity can attenuate the phenotype and reduce tumor formation. At present, we do not know which downstream proteins are targeted by PRMT5/MEP50 to maintain the aggressive cancer phenotype. However, PRMT5 is known to enhance cancer cell survival and function by increasing the level of proteins that enhance cell proliferation, by suppressing tumor suppressor expression and by enhancing oncogene expression ^47-50. Moreover, although we have studied the impact of SFN on PRMT5/MEP50 regulation of histone demethylation, PRMT5 is also known to directly methylate and alter the activity of cytosolic proteins ⁴⁷. Thus, ongoing studies will be necessary to understand the targets responsible for the PRMT5-dependent maintenance of the aggressive mesothelioma cancer cell phenotype.