PRMT5 inhibition induces pro‐inflammatory macrophage polarization and increased hepatic triglyceride levels without affecting atherosclerosis in mice

Abstract Protein arginine methyltransferase 5 (PRMT5) controls inflammation and metabolism through modulation of histone methylation and gene transcription. Given the important role of inflammation and metabolism in atherosclerotic cardiovascular disease, here we examined the role of PRMT5 in atherosclerosis using the specific PRMT5 inhibitor GSK3326595. Cultured thioglycollate‐elicited peritoneal macrophages were exposed to GSK3326595 or DMSO control and stimulated with either 1 ng/mL LPS or 100 ng/mL interferon‐gamma for 24 h. Furthermore, male low‐density lipoprotein (LDL) receptor knockout mice were fed an atherogenic Western‐type diet and injected intraperitoneally 3×/week with a low dose of 5 mg/kg GSK3326595 or solvent control for 9 weeks. In vitro, GSK3326595 primed peritoneal macrophages to interferon‐gamma‐induced M1 polarization, as evidenced by an increased M1/M2 gene marker ratio. In contrast, no difference was found in the protein expression of iNOS (M1 marker) and ARG1 (M2 marker) in peritoneal macrophages of GSK3326595‐treated mice. Also no change in the T cell activation state or the susceptibility to atherosclerosis was detected. However, chronic GSK3326595 treatment did activate genes involved in hepatic fatty acid acquisition, i.e. SREBF1, FASN, and CD36 (+59%, +124%, and +67%, respectively; p < 0.05) and significantly increased hepatic triglyceride levels (+50%; p < 0.05). PRMT5 inhibition by low‐dose GSK3326595 treatment does not affect the inflammatory state or atherosclerosis susceptibility of Western‐type diet‐fed LDL receptor knockout mice, while it induces hepatic triglyceride accumulation. Severe side effects in liver, i.e. development of non‐alcoholic fatty liver disease, should thus be taken into account upon chronic treatment with this PRMT5 inhibitor.


| INTRODUC TI ON
The type II protein arginine methyltransferase (PRMT) family member PRMT5 controls cell proliferation, inflammation, and metabolism through modulation of histone methylation and gene transcription. 1 Due to its central role in these essential processes, PRMT5 has been associated with a variety of diseases. As reviewed by Xiao et al., PRMT5 is highly expressed in different types of cancer where it contributes to enhanced tumour cell proliferation and invasion. 2 In addition, PRMT5 protects the HIV-1 accessory protein viral protein R from proteasomal degradation to support HIV replication 3 and regulates the secretion of pro-inflammatory cytokines such as interleukin-6 (IL-6) and interferon (IFN)-gamma inducible protein-10 (IP-10/CXCL10) by macrophages in endometriosis. 4 Furthermore, Webb et al. have shown that PRMT5 is upregulated upon memory T cell reactivation and essential for T cell activation and expansion in delayed-type hypersensitivity and experimental autoimmune encephalomyelitis. 5 Interestingly, studies by Tan et al. have suggested that changes in PRMT5 functionality may also be relevant in the context of atherosclerotic cardiovascular disease, a progressive pathology that includes macrophage-driven inflammation secondary to dyslipidemia-induced cholesterol accumulation within the arterial wall. 6 More specifically, Tan et al. showed that the relative mRNA expression levels of PRMT5 in peripheral blood cells are reduced with increasing coronary artery disease severity. 7 We therefore hypothesize that a potential causal relationship may exist between PRMT5 activity, inflammation, and atherosclerosis development. To test this hypothesis, in the current study we determined the effect of treatment with the selective PRMT5 inhibitor GSK3326595 on (1) the macrophage response to pro-inflammatory stimuli in vitro and ex vivo and (2) atherosclerosis development in hypercholesterolemic low-density lipoprotein (LDL) receptor knockout mice in vivo.

| In vitro experiment with thioglycollateelicited peritoneal macrophages
C57BL/6 Three C57BL/6 wild-type female mice received an intraperitoneal injection with 1 mL of 3% thioglycollate to elicit macrophage recruitment to the peritoneal cavity. After 5 days, mice were sacrificed and the peritoneal cavity flushed with PBS. The collected peritoneal macrophages were pooled together and plated in 12-well plates at a concentration of one million cells per ml in Dulbecco's Modified Eagle Medium (DMEM), containing 10% fetal bovine serum, penicillin/streptomycin and L-glutamine. After allowing the cells to attach to the culture plate for 4 h (5% CO 2 and 36°C), cells were washed with PBS, and exposed to 0.1% DMSO or 100 μM GSK3326595 dissolved in 0.1 DMSO% and in the presence of PBS as control, 1 ng/mL LPS (L3024-5MG, Merck), or 100 ng/ mL IFN-gamma (I4777-1MG, Merck). Each condition was tested in 6 independent wells (n = 6 replicates). After 12 h, cell media were collected for cytokine detection by ELISA. Cells were lysed in guanidine thiocyanate (GTC) (L-15809, Fisher Scientific) for further mRNA expression analysis.

| Gene expression analysis through real-time quantitative PCR
Total RNA was isolated according to Chomczynski and Sacchi. 8 The concentration of the obtained RNA was determined using a Nanodrop Spectrophotometer (Nanodrop Technologies). cDNA was synthesized from the RNA using Maxima H Minus reverse transcriptase (Thermo Scientific, Cat. no. EP0751). Analysis of gene expression was accomplished through an ABI PRISM 7500 machine (Applied Biosystems) using SYBR Green technology. Β-actin, ribosomal protein L27 (RPL27), acidic ribosomal phosphoprotein P0 (36B4) and peptidylpropyl isomerase A (PPIA) were used as housekeeping genes. The sequences of the primers used can be found in

| Cytokine enzyme-linked immunosorbent assay (ELISA)
For measurement of the concentrations of the pro-inflammatory cytokine IP-10 in the cell culture medium, we used the Mouse CXCL10/ IP-10/CRG-2 DuoSet ELISA kit (Cat. No. DY466-0) from R&D systems. Absorbances were measured at 450 nm and 570 nm.

| Experimental animals
Ten-week-old male LDL receptor knockout mice (on a C57BL6/J background), obtained from The Jackson Laboratory and bred at Gorlaeus Laboratories, were fed a Western-type diet containing 0.25% cholesterol and 15% cocoa butter (SDS, Sussex, UK) to induce the development of atherosclerotic lesions. The mice were randomly allocated to two different treatment groups receiving intraperitoneal injections of either the control solvent DMSO (100 μL 10% DMSO in PBS, N = 12) or GSK3326595 (0.125 mg in 100 μL 10% DMSO in PBS, ~5 mg/kg based on body weight at start of the experiment, N = 15) three times per week for altogether 9 weeks. As Gerhart et al. showed that an in vivo treatment dosage of 4.2 mg/kg GSK3326595 is able to reduce the synthesis of the PRMT5 product symmetric dimethylarginine by >80%, 9 we chose 5 mg/kg as our treatment dosage. For sacrifice, mice were anaesthetised through a subcutaneous injection with 100-150 μL of a ketamine (100 mg/kg), xylazine (12.5 mg/kg), and atropine (125 μg/kg) mixture. Subsequently, orbital blood was collected and a whole-body perfusion was performed using PBS. Organs were excised and parts were fixed for 24 h in a 3.7% formalin solution for subsequent histological analysis or stored at −20°C for biochemical analysis.

| Ex vivo experiment with peritoneal macrophages
Upon sacrifice of the LDL receptor knockout mice, peritoneal cells were collected from both the DMSO and GSK3326595-treated groups. These were subsequently plated in 12-well plates in DMEM, containing 10% fetal bovine serum, penicillin/streptomycin, and Lglutamine. After allowing the cells to attach to the culture plate for 4 h at 5% CO 2 and 37°C, cells were exposed to PBS as control or to 100 ng/mL IFN-gamma for ex vivo activation. After 12 h, cell medium was collected for cytokine detection by ELISA and cells were lysed in GTC for further mRNA expression analysis.

| Flow cytometry
Immunostaining was performed on single cell suspensions derived from the blood, spleen, and peritoneum of LDL receptor knockout mice treated with GSK3326595 or DMSO for 9 weeks. Single cell suspensions were obtained by filtering the samples/organs through a 70 μm cell strainer using PBS. Cells were stained using the appropriate fluorochrome-labelled antibodies ( gating strategy for CD4+ and CD8+ T cells in the blood, peritoneal cavity, and spleen is shown in the Appendix S1 ( Figure S1).

Flow cytometric analysis was performed on a Beckman Coulter
Cytoflex S and the acquired data were analysed using FlowJo software.

| Plasma lipid measurements
Plasma specimens from LDL receptor mice were isolated from or-   M1 macrophage gene markers TNF-alpha, iNOS, IL-1beta, MHC-II, and IP-10 ( Figure 1A) were on average lower than those of the antiinflammatory/resolving macrophage marker genes ARG1 and CD206

| Tissue lipid extraction and quantification
( Figure 1C). Treatment with GSK3326595 for 24 h was not associated with a change in the baseline macrophage phenotype. As expected, subsequent exposure of the thioglycollate-elicited macrophages to LPS and IFN-gamma induced a shift in the M1/M2 gene marker ratio towards a more pro-inflammatory phenotype characterized by high TNF-alpha, iNOS, and IP-10 expression levels. Importantly, relative gene expression levels of iNOS (+207%, p < 0.05) and IP-10 (+73%, p < 0.05) were even higher in cells pre-treated with GSK3326595 than in those treated with solvent control after IFN-gamma stimulation, but not after LPS exposure ( Figure 1A). As a result, GSK3326595 treatment was associated with a greater rise in the overall M1/M2 gene marker ratio (p < 0.001; Figure 1D) and IP-10 protein secretion or their basal iNOS and ARG1 protein expression levels ( Figure 2B).
Accordingly, although CD206 mRNA expression levels were reduced by GSK3326595 treatment (−39%; p < 0.01; Figure 2C), gene expression levels of other M2 marker genes as well as M1 markers were, in general, not different between the cultured peritoneal macrophage populations ( Figure 2C). However, peritoneal macrophages isolated from the GSK3326595-treated mice did display an exacerbated polarization response upon exposure to IFN-gamma, resulting in a 2.3fold higher M1/M2 ratio (p < 0.01; Figure 2D). This latter effect could be attributed to elevated mRNA expression levels of IP-10 (+73%; p < 0.01), in line with our in vitro findings, as well as a reduction in CD206 transcript levels (−41%; p < 0.01) ( Figure 2C).

| Treatment of LDL receptor knockout mice with GSK3326595 does not change T cell numbers or phenotype
Given that PRMT5 has also been suggested to play a role in T cell differentiation and survival 5,13,14 a potential effect of our chronic GSK3326595 treatment on T cell subsets was examined by means of flow cytometry. PRMT5 inhibition did not significantly impact total CD4 + helper and CD8 + cytotoxic T cell numbers in any of the compartments studied (data not shown). Staining for the cell surface markers CD62L and CD44 revealed that both the CD4 + and CD8 + T cells were also distributed equally between the two treatment groups over their respective naïve (CD44 − CD62L + ), effector (CD44 + CD62L − ), and memory (CD44 + CD62L + ) subclasses ( Figure 3).

| Treatment of LDL receptor knockout mice with GSK3326595 does not alter atherosclerosis susceptibility
Sections of the aortic root were stained with Oil red O for neutral lipids to identify atherosclerotic lesions. As can be seen from the representative images in Figure 4A,  Figure 4B). MOMA-2 staining showed that lesions of both groups of mice were equally rich in macrophages (4 ± 1 × 10 4 μm 2 for GSK3326595-treated mice versus 5 ± 1 × 10 4 μm 2 for control-treated mice; Figure 4C). Trichrome staining further verified a similar lesion composition, with collagen areas of ~7 × 10 4 μm 2 in both groups of mice ( Figure 4D). Liu et al. have shown that PRMT5 can enhance the activity of the lipogenic transcription factor SREBP-1a. 15 In addition, recent findings by Wang et al. have suggested that PRMT5 is involved in the control of LDL uptake by hepatocytes. 16 Given that disturbances in lipid metabolism are the driving force in the initiation of atherosclerotic lesion development, we investigated a potential effect of GSK3326595 treatment on the body's lipid status. Weight development was not influenced by GSK3326595 treatment (data not shown). As a result, average body weights at sacrifice, after 9 weeks of Western-type diet feeding, were not significantly different between the two treatment groups ( Figure 5A). In accordance with a similar body weight development, no differences were found in the triglyceride content of perigonadal white adipose tissue and subcutaneous brown fat depots ( Figure 5D-F). 17,18 Strikingly, an on average 50% higher (p < 0.05) hepatic triglyceride content was detected in GSK3326595-treated mice as compared to control-treated mice ( Figure 5D-G). The apparent change in fatty liver development upon the Western-type diet challenge was not associated with an altered inflammation state of the liver as judged from the unchanged hepatic relative mRNA expression levels of the general macrophage marker CD68 and the macrophage-derived cytokine CCL2 ( Figure 6A,B).

| GSK2236595 treatment increases hepatic triglyceride levels without changing the hyperlipidemia extent in LDL receptor knockout mice
Furthermore, no concomitant change in the hyperlipidemia extent, that is plasma total cholesterol and triglyceride levels, was observed ( Figure 5B,C).

| GSK2236595 treatment activates genes involved in fatty acid acquisition in livers of LDL receptor knockout mice
Gene expression analysis on liver specimens was performed to potentially uncover the mechanism behind the rise in hepatic and PPARα, ACOX1, and CPT1 that mediate lipolysis and fatty acid oxidation, respectively ( Figure 6E-K). In contrast, PRMT5 inhibition was associated with a significant increase in the expression of genes involved in fatty acid acquisition. More specifically, relative mRNA expression levels of the lipogenic transcription factor SREBF1 and the basolateral fatty acid transporter CD36 were respectively 59% higher (p < 0.05; Figure 6L) and 67% higher (p < 0.05; Figure 6M) in livers from GSK3326595-treated mice than in those of controls. In addition, a significant increase in fatty acid synthase (FASN) transcript levels was detected in response to GSK3326595 treatment (+124%; p < 0.05; Figure 6N). PRMT5 inhibition did not impact relative mRNA expression levels of the lipogenic genes ACC and SCD1 ( Figure 6O,P).   Our study also showed that the basal activation state of macrophages in the peritoneum was not changed by GSK3326595, nor were the T cell populations in several T cell-rich compartments such as blood, spleen, and peritoneum. We therefore assume that the null effect on atherosclerosis susceptibility is due to the inability of GSK3326595 to alter the systemic inflammatory state.
The fact that we did not observe any effect of PRMT5 inhibition on T cell activation is unexpected, since a previous study has indicated that blocking PRMT5 using the inhibitor CMP5 21 severely suppressed the CD4 + CD44 + memory T cell differentiation in the spleen of a murine model. 5 In addition, T cell-specific deletion of PRMT5 leads to peripheral T cell lymphopenia. 14 We chose 5 mg/ kg as our treatment dosage since Gerhart et al. 22 showed that a dosage of 4.2 mg/kg GSK3326595 is sufficient to reduce SDMA production by >80% in vivo. However, it has been found that a daily dose of >50 mg/kg is required to inhibit PRMT5-mediated cancer growth in mice. 22 As such, it can be hypothesized that our 10-fold lower dose is also too low to functionally inhibit PRMT5 in leukocytes, leading to the overall unchanged inflammation state after chronic GSK3326595 treatment.
Even though the inflammation state was not affected, the lowdose GSK3326595 treatment for 9 weeks did cause a pharmacological effect in liver. More specifically, GSK3326595 treatment increased hepatic triglyceride accumulation in our LDL receptor knockout mice in response to Western-type diet feeding. A common risk factor for hepatic steatosis and cardiovascular disease is hypertriglyceridemia. 23  Important to note is that increased CD36 and FASN expression has shown to be related to increased hepatic steatosis. 24 Hepatic steatosis arises from an imbalance between acquisition by uptake of fatty acids and de novo lipogenesis and triglyceride removal by fatty acid oxidation and the secretion of triglyceride-rich lipoproteins such as VLDL. 24 As CD36 and FASN are both involved in these processes, the upregulation of these genes could contribute to the observed increase in liver triglyceride levels. In conclusion, we have shown that PRMT5 inhibition by chronic low dose GSK3326595 treatment does not affect atherosclerosis development and lesion composition in Western-type diet-fed LDL receptor knockout mice, while it does induce hepatic triglyceride accumulation. Notably, GSK3326595 is currently being tested as anti-cancer drug therapy in both preclinical mice studies and clinical phase I/II studies. [26][27][28] In light of our current findings, it will be important to consider that long-term pharmacological inhibition of PRMT5 to treat cancer may cause severe side effects in liver, i.e.
induce non-alcoholic fatty liver disease.

ACK N O WLE D G E M ENTS
The authors thank Mireia N.A. Bernabé Kleijn and Yousra Abbasi for technical assistance.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors confirm that there are no known conflicts of interest associated with this publication and there has been no significant financial support for this work that could have influenced its outcome.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.