Characterizing the Role of Glycogen Synthase Kinase-3α/β in Macrophage Polarization and the Regulation of Pro-Atherogenic Pathways in Cultured Ldlr-/- Macrophages

The molecular and cellular mechanisms that link cardiovascular risk factors to the initiation and progression of atherosclerosis are not understood. Recent findings from our laboratory indicate that endoplasmic reticulum (ER) stress signaling through glycogen synthase kinase (GSK)-3α/β induces pro-atherosclerotic pathways. The objective of this study was to define the specific roles of GSK3α and GSK3β in the activation of pro-atherogenic processes in macrophages. Bone marrow derived macrophages (BMDM) were isolated from low-density lipoprotein receptor knockout (Ldlr-/-) mice and Ldlr-/- mice with myeloid deficiency of GSK3α and/or GSK3β. M1 and M2 macrophages were used to examine functions relevant to the development of atherosclerosis, including polarization, inflammatory response, cell viability, lipid accumulation, migration, and metabolism. GSK3α deficiency impairs M1 macrophage polarization, and reduces the inflammatory response and lipid accumulation, but increases macrophage mobility/migration. GSK3β deficiency promotes M1 macrophage polarization, which further increases the inflammatory response and lipid accumulation, but decreases macrophage migration. Macrophages deficient in both GSK3α and GSK3β exhibit increased cell viability, proliferation, and metabolism. These studies begin to delineate the specific roles of GSK3α and GSK3β in macrophage polarization and function. These data suggest that myeloid cell GSK3α signaling regulates M1 macrophage polarization and pro-atherogenic functions to promote atherosclerosis development.


INTRODUCTION
Cardiovascular disease (CVD) is the leading cause of death in the world today (1) and atherosclerosis is a major underlying cause of CVD. Macrophages are centrally involved in every stage of the development of atherosclerosis, and they are the main cellular component within the atherosclerotic lesion (2,3). Atherosclerosis initiates when endothelial cells (EC) respond to injury, which mediates the attachment and infiltration of monocytes. Monocytes invade the subintima and differentiate into macrophages. These macrophages take up modified-LDL particles and become foam cells, which form fatty streaks in the artery wall. Macrophage/foam cell apoptosis leads to the establishment of a necrotic core, which is a key feature of unstable plaques that are prone to rupture. Lesion rupture triggers atherothrombosis and can occlude the artery. This can lead to acute cardiovascular complications (myocardial infarction or stroke) and potentially death. The underlying molecular mechanisms that regulate macrophage function during the development of atherosclerosis are not completely understood.
Macrophages can be polarized into many different subtypes that have distinct characteristics and functions. The extreme phenotypes are pro-inflammatory (M1) macrophages and antiinflammatory (M2) macrophages. M1 macrophages can be induced by exposure to T helper type 1 (Th1) cell products, such as interferon (IFN)-g, or microbial products, such as lipopolysaccharide (LPS) (4). In contrast, M2 antiinflammatory macrophages can be induced by exposure to T helper type 2 (Th2) cell products, including interleukin (IL)-4 or tumor growth factor (TGF)-b. M1 macrophages produce proinflammatory cytokines (TNFa, IL-1b) and are believed to promote atherosclerotic lesion development and complexity (5,6), whereas M2 macrophages produce anti-inflammatory cytokines (IL-10) and have tissue remodeling properties (7,8). Other macrophage subtypes have been identified, including M ox , M hem , and M4 (9). The roles and functions of these macrophages are less well understood. Macrophages are directly involved in a variety of processes during atherosclerosis including polarization, foam cell formation, apoptosis, cell viability/ proliferation, and migration. The mechanism(s) and cellular signals that regulate macrophage polarization and other functions that contribute to the development of atherosclerosis are still unclear.
Glycogen synthase kinase 3 (GSK3) is a serine/threonine kinase that plays an important role in many cellular pathways that regulate metabolism and viability. GSK3 has been linked to several disorders and diseases, including cancer (10), bipolar mood disorder (11), diabetes (12), and Alzheimer's disease (13). There are two main forms of GSK3 in mammals: GSK3a (51 kDa) and GSK3b (47 kDa), as well as the splice variant of GSK3b, GSK3b2 (14). Isoforms GSK3a and GSK3b are 98% homologous in the kinase domain and are expressed ubiquitously (15). GSK3a/b is predominantly located in the cytoplasm, endoplasmic reticulum (ER), and nucleus (16). GSK3a/b is a constitutively active kinase and its activity is directly regulated (inhibited) by the insulin and Wnt signaling pathways (17,18). A study from our lab has shown that the presence of ER stress in Thp-1 derived macrophages activates the protein kinase R-like ER kinase (PERK) signaling branch of the unfolded protein response (UPR) to promote GSK3a/b activity (19). Recent evidence suggests that GSK3a and GSK3b have distinct functions (20)(21)(22). Whole-body GSK3a-deficient mice are viable and develop normally, while GSK3b deletion is embryonically lethal (23,24). GSK3a and b play unique and independent roles in skeletal muscle cell insulin signaling (25)(26)(27), cardiomyocyte development and proliferation (23), and Th cell polarization (28). Recent studies support a role for GSK3a/b in atherosclerosis. Results from our lab suggest that myeloid deletion of GSK3a, pharmacological mitigation of ER stress (by 4 phenylbutyrate), or inhibition of GSK3a/b (by valproate) attenuates the progression of atherosclerosis (29)(30)(31)(32). Together, these results suggest a role for myeloid-specific GSK3a in the development of atherosclerosis. However, the specific roles of myeloid GSK3a and GSK3b in macrophage polarization and other pro-atherogenic functions are not known.
In this study we have isolated bone marrow derived macrophages from myeloid cell-specific GSK3a and/or GSK3bdeficient Ldlr -/mice to characterize the roles of GSK3a and GSK3b in specific cellular functions. Our results demonstrate that GSK3a and GSK3b play distinctive roles in defining macrophage phenotype and regulating atherogenic responses.

Mouse Models
Myeloid-specific GSK3a-and/or GSK3b-deficient mice were created in an Ldlr -/genetic background, as previously described (30). Myeloid specific GSK3a knockout mice (Ldlr −/− LyzMCre +/-GSK3a fl/fl or LMaKO), myeloid-specific GSK3b knockout mice (Ldlr −/− LyzMCre +/-GSK3b fl/fl or LMbKO), and myeloid-specific GSK3a and b knockout mice (Ldlr −/− LyzMCre +/-GSK3a fl/fl GSK3b fl/fl or LMabKO) were utilized in this study. Ldlr −/-GSK3a fl/fl GSK3b fl/fl (Labfl/fl) mice were used as controls. All experimental mice had unlimited access to food and water and were maintained on a 12-hour light/dark cycle. All animal experiments were conducted with pre-approval of the McMaster University Animal Research Ethics Board. All experiments conform with the guidelines and regulation of the Canadian Council on Animal Care.

Bone Marrow-Derived Macrophage Isolation and Polarization
At the age of 8-10 weeks, tibias and femurs were harvested and bone marrow was collected from Labfl/fl (control), LMaKO, LMbKO, and LMabKO mice using a 70 mm nylon mesh passing through the medullary cavity. Cells were resuspended in Dulbecco's Modified Eagle Medium (DMEM) containing 15% (v/v) fetal bovine serum, 100 IU/ml penicillin and 100 mg/ml streptomycin, 1X MEM non-essential amino acids, and 20 ng/ml macrophage colony stimulating factor (MCSF, Cell Signaling). Cells were counted using a hemocytometer and 5 x 10 6 cells were seeded onto a 10 cm plate containing 10 ml of medium. After six days in a humidified 37°C incubator (5% CO2), cells were washed twice with warm, sterile 1X Dulbecco's phosphatebuffered saline (DPBS) without calcium or magnesium. Macrophages were detached using accutase (Cedarlane), and cells were counted with a hemocytometer and replated for subsequent experiments. Subsets of cells were polarized to M1 macrophages by exposure to 10 ng/ml lipopolysaccharide (LPS), or M2 macrophages by exposure to 10 ng/ml IL-4 for 24 hours or left unstimulated as M0 macrophages (30,33). Culture media was collected after the 24 hours of treatment (UT, LPS or IL-4) and cytokine and chemokine levels were quantified using the pro-inflammatory focused multiplexing LASER Bead Assay ([Mouse Cytokine Array/Chemokine Array 31-Plex (MD31), Eve Technologies, Calgary, AB].

Gene Expression
BMDM were seeded onto 12-well tissue culture plates at a density of 4 x 10 5 cells/well in 1 ml medium and polarized as described above. Total RNA was isolated using TRIzol ® Reagent (Invitrogen), as previously dscribed (19,30). Purified total RNA was resuspended in DNase/RNase-free water and RNA concentration and purity were determined using a Nanodrop spectrophotometer (Thermo Fisher Scientific). DNA was prepared from 1 mg of total RNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantitative real-time reverse transcription-polymerase chain reaction (RT-PCR) was performed using 1 ml of resulting cDNA, 12.5 ml SensiFAST SYBR-Rox (Thermo Fisher Scientific), 1.25 ml of forward and reverse primers (500 nM, IDT) (Supplementary Table I), and 8 ml of RNase-water in a total volume of 24 ml/well. The following conditions were used to amplify cDNA: 10-minute hold at 95°C, followed by 40 cycles consisting of a 15-second melt at 95°C, followed by 1-minute annealing at 60°C. Relative quantitative analysis (2-ddCt) was performed by normalizing data to the b-actin reference gene.

Characterization of Macrophages
BMDM were cultured and resuspended in lysis buffer (4x SDS PAGE sample buffer). Cell extracts were fractionated by SDS-PAGE and transferred to a polyvinylidene difluoride (PVDF) membrane (Bio-Rad). The membranes were blocked with 5% non-fat milk in TBST (10 mM Tris, pH 8.0, 150 mM NaCl, 0.5% Tween 20) for 45min and then incubated overnight with primary antibody against GSK3a/b (1:1000, Cell signaling) or b-Actin Blots were washed with TBST three times for 5 min each and developed with the ECL system (Millipore). Images were captured using a Molecular Imager ChemiDoc XRS+ (Bio-Rad).

AlamarBlue Cell Viability Assay
BMDM were cultured in 96 well tissue culture plates at a density of 1 x 10 5 cells/well/100 ml medium. Cells were polarized as described above and then treated with 10 mM thapsigargin (Tg), or 10 mg/ml tunicamycin (Tm), or left untreated for 24 hours. Cell viability was determined using the alamarBlue ™ assay (Biorad). Cells were washed and alamarBlue ™ reagent (Bio-Rad) was added. Absorbance was determined at 570 nm (reduction) and 600 nm (oxidation) to calculate cell viability.

Immunofluorescent Staining
BMDM were seeded onto 8 chamber slides (Thermo Fisher Scientific) at a density of 1 X 10 5 cells/200ul/chamber and incubated at 37°C. Immunostaining was performed as previously described (19,30). Cells were washed with 1x PBS, fixed with 4% paraformaldehyde (PFA) for 15 min. Cells were permeabilized with 0.5% Triton X-100 for 5 min and then incubated in blocking solution (3% goat serum, 0.5% BSA, 1X PBS) for 1hr. Primary antibodies against the proliferation marker Ki67 (1:200, Abcam), NF-kB p65 (1:50, Santa Cruz), NLRP3 (1:50, Abcam), or CCR7 (1:100, Abcam) were added. After 24 hours, cells were washed and then incubated with secondary antibodies, Alexa Fluor 488 goat anti-mouse IgG (1:250, Thermo Fisher Scientific) or Alexa Fluor 488 goat anti-rabbit IgG (1:250, Thermo Fisher Scientific), for 2 hrs. Separate slides of cells were stained with pre-immune IgG instead of primary antibodies to control for non-specific staining (Supplementary Figure I). Cells were washed and stained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, 1:5000, Invitrogen). The slides were mounted using Fluoromount Aqueous Mounting Medium (Sigma) and stored at 4°C in the dark. Images of the stained sections were collected using a Leitz LABORLUX S microscope connected to a DP71 Olympus camera. ImageJ 1.52q software was used to quantify immunofluorescent staining. For each experimental group of cells four biological replicates were analyzed. For each replicate, a minimum of four images were captured, each containing approximately 200 cells. The stained area over background as well as cell number were quantified using ImageJ 1.52q software. Data from each image of a biological replicate (four images) were combined providing a stained area per cell with a minimum of 800 cells. Data are presented as average stained area per cell from four biological replicates.

Oil Red O Staining
Chamber slides with 8 chambers (wells) were used for the lipid accumulation assay. Cells were plated in 8 chamber slides (Thermo Fisher Scientific) at a density of 1.5 X 10 5 cells/200ul/ chamber and incubated at 37°C for 24 hrs in DMEM. Oil Red O stock solution was prepared by dissolving 2.5 g Oil Red O powder (Sigma) in 500 ml isopropanol (100%) in a water bath at 70°C for 10 min. The stock solution was filtered while it was warm. Before staining, the working solution was prepared by diluting stock solution 3:2 within ddH 2 O. Cells were fixed with 4% PFA for 15 min. and permeabilized with 0.5% Triton X-100 for 5 min. Cells were stained with filtered oil red O working solution at 37°C for 15-20 min. Cells were washed in 60% isopropanol for 15-30 sec. and PBS 2 times, then stained with DAPI (1:5000, Invitrogen) for 2-5 min. The cells were mounted by using Fluoromount Aqueous Mounting Medium (Sigma) and stored at 4°C in the dark. Lipid content was visualized by a bright-field or fluorescent microscope (Olympus BX41 microscope connected to a DP71 Olympus camera) and quantified. For each experimental group of cells, four biological replicates were analyzed. For each replicate, four images were captured, each containing x approximately 200 cells. The stained area over background as well as cell number were quantified using ImageJ 1.52q software. Data from each image of a biological replicate (four images) were combined providing a stained area per cell with a minimum of 800 cells. Data shown are average stained areas per cell from four biological replicates.

Migration Assay
Cells were seeded at a density of 0.8 X 10 5 cells/200ul/insert onto Transwell inserts (pore size of 3 µm, Corning Costar) that were pre-coated with rat tail collagen I (4mg/ml, Millipore). Cells were added in the upper chamber and incubated at 37°C for 1 hr in serum-free media. These filter inserts were placed in wells containing the serum-free media with 0.5ug/ml chemokine ligand 19 (CCL19, R&D Systems) and incubated at 37°C. After 4 hrs, inserts with the cells were removed and washed with 1x PBS. Cells were fixed with 4% PFA for 15 min. After washing with 1x PBS, cells were stained with DAPI (1:5000, Invitrogen) for 2-5 min. Filters were then rinsed twice with 1x PBS. The cells on the upper surface, that had not migrated, were removed by carefully scraping with a cotton swab. Migrated cells, on the lower surface, were visualized and quantified using a fluorescent microscope (Olympus BX41 microscope connected to a DP71 Olympus camera, with 4x objective). For each experimental group of cells four biological replicates were analyzed. For each replicate, a minimum of five images were captured. The number of cells number was quantified using ImageJ 1.52q software. Data from each image of a biological replicate (five images) were combined providing a total number of cells migrated. Data shown are average percentage migrated cells from four biological replicates.

Extracellular Flux Analysis
BMDM were plated in Seahorse XF24 plate (Agilent) at a density of 40,000/well in DMEM and cultured for 24 hours before the medium was replaced with a fresh DMEM. The assay was performed as described previously (34). One hour before the assay, media were exchanged for XF24 media. Mito Stress assay was performed by sequential addition of Oligomycin (inhibitor of ATP synthesis), carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, uncoupling agent), and rotenone/ antimycin A (inhibitors of complex I and complex III of the respiratory chain, respectively) were diluted into XF24 media and loaded into the accompanying cartridge to achieve final concentrations of 1 mM, 2 mM, and 0.5 mM, respectively. Injections of the drugs into the medium occurred at the time points specified. Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were monitored using a Seahorse Bioscience XF24 Extracellular Flux Analyzer (Agilent). Data are analyzed using Wave Desktop Software.

Statistical Analysis
All statistical analyses were performed using GraphPad Prism 8 software. Data were analyzed by one-or two-way ANOVA, followed by the Bonferroni multiple comparison test between all groups. Error bars represent the standard error of the mean (SEM). For all experiments, a p-value lower than 0.05 was considered statistically significant. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

RESULTS
Characterization of Myeloid Cell-Specific GSK3a and/or GSK3b Deficiency in Ldlr −/− Mice Myeloid GSK3a and/or GSK3b deletion did not significantly alter the number of monocytes or other cell types in whole blood (Supplementary Table II). To determine the effect of GSK3a and/or GSK3b deficiency on macrophage function, bone marrow was harvested from 8-10-week-old Ldlr -/mice with myeloidspecific GSK3a and/or GSK3b deficiency. Bone marrow was cultured in the presence of MCSF to induce macrophage differentiation. Subsets of BMDM were exposed to 10 ng/ml LPS or 10 ng/ml IL-4 for 24 hrs to polarized them to M1 and M2, respectively. Myeloid cell-specific deficiency of GSK3a and/or GSK3b was confirmed by RT-PCR and immunoblot. Polarization of BMDM to either M1 (LPS) or M2 (IL4) resulted in a significant decrease in gene expression of GSK3a and GSK3b in both control and specific knockout macrophages ( Figures 1A, B). There was no detectable compensation in the gene expression ( Figures 1A, B To determine the efficiency of MCSF-induced differentiation, the macrophage (CD11b+ F4/80+ cells) number was assessed by flow cytometry. Results show that over 95% of cells were F4/80+ and CD11b+ ( Figure 1E and Supplementary Figure III). Our gating method was able to capture all of CD11b+ and F4/80+ cells in one box. The results show that bone marrow from LMaKO, LMbKO, and LMabKO mice is not significantly different than Labfl/fl (control) bone marrow in terms of its ability to differentiate into macrophages ( Figure 1F). This suggests that GSK3a and/or b deficiency does not affect bone marrow cell number or the efficiency of bone marrow differentiation into M0 macrophages.

Lipid Accumulation Is Impaired in GSK3a-Deficient Macrophages and Increased in GSK3b-Deficient Macrophages
Macrophages are endocytotic cells that readily take up lipoprotein particles and cell debris. To determine the roles of GSK3a and GSK3b in lipid accumulation, BMDM were isolated from Labfl/fl (control), LMaKO, LMbKO, and LMabKO mice and cultured for 24 hrs. Cells were stained with Oil Red O and analyzed for lipid accumulation by quantifying ORO-stained area. The results indicate that M1 macrophages accumulate significantly more lipid compared to M2 macrophages ( Figures 4A, B and Supplementary Figure VIIA). In addition, LMaKO macrophages have a decreased tendency to accumulate lipids while LMbKO and LMabKO macrophages have an increased ability to accumulate lipids ( Figure 4B).
We next investigated the effect of GSK3a and/or GSK3b deletion on gene expression of scavenger receptor (SR)-A and CD36, two genes encoding proteins involved in lipid uptake ( Figures 4C, D), and ATP-binding cassette transporter (ABCA1) and ATP binding cassette subfamily G member 1 (ABCG1), two genes encoding proteins involved in lipid efflux ( Figures 4E, F). Results show that there is an increase in gene expression of (SR)-A, CD36, and ABCA1 in LMbKO and LMabKO M1 macrophages. LMabKO M1 macrophages displayed increased gene expression of ABCG1. LMbKO and LMabKO M2 macrophages show a significant increase in (SR)-A and CD36 expression that does not affect actual lipid accumulation in these experiments. There is no difference in gene expression of these markers in LMaKO macrophages, relative to controls. Analysis of other genes involved in lipid accumulation, including SR-B1, liver X receptor alpha (LXRa), and lecithin-cholesterol acyltransferase (LCAT), revealed no significant differences in expression in LMaKO, LMbKO or LMabKO macrophages compared to control (within the same treatment group) (Supplementary Figure VIIB-VIID). The effect of GSK3a and/or GSK3b deficiency on lipid biosynthesis was determined by analyzing the gene expression of 3-hydroxy 3methylglutaryl-CoA (HMG-CoA) and fatty acid synthase (FAS). Results show that LMbKO and LMabKO macrophages have significantly reduced gene expression of HMG-CoA and FAS (Supplementary Figures VIIE, VIIF). Together, these findings suggest that the presence of GSK3a promotes lipid uptake and accumulation and GSK3b impedes lipid accumulation in M1 macrophages.

GSK3a and GSK3b Play Redundant Roles in Cell Viability and Proliferation
We next determined the effect of GSK3a and/or GSK3b deficiency on macrophage viability and function. BMDM were challenged with ER stress inducing agents, tunicamycin (Tm) or thapsigargin (Tg). Cell viability (% alamarBlue reduction) was measured using an alamarBlue cell viability assay ( Figure 5A). As expected, Labfl/fl (control) M1 and M2 macrophages show a decrease in alamarBlue reduction when treated with Tm or Tg. The alamarBlue reduction in LMaKO and LMbKO macrophages was not significantly different than the Labfl/fl (control). The alamarBlue reduction was significantly increased in both M1 and M2 macrophages that were deficient in both GSK3a and GSK3b. These results suggest that GSK3a and GSK3b play redundant roles in regulating the metabolic activity of macrophages.
Immunofluorescence staining was used to analyze the effects of GSK3a and/or GSK3b deficiency on macrophage proliferation markers. BMDM from Labfl/fl (control), LMaKO, LMbKO, and LMabKO mice were fixed and immunostained with an antibody against the proliferation marker Ki67 and DAPI ( Figure 5B) and quantified ( Figure 5C and Supplementary Figure VIIIA). The percentage of Ki67 positive cells was significantly increased in M2 macrophages in comparison to M1 macrophages ( Figure 5C). LMaKO, LMbKO, and LMabKO M2 macrophages had significantly increased expression of Ki67, compared to Labfl/fl controls. There was no significant difference found in the gene expression of proliferation marker cMyc (Supplementary Figure VIIIB). These results suggest that GSK3a and GSK3b suppress M2 macrophage proliferation.

Migration of M1 Macrophages Increase With GSK3a Deficiency and Decreases With GSK3b Deficiency
Macrophage migration was determined using a transwell plate assay. BMDM isolated from Labfl/fl (control), LMaKO, LMbKO, and LMabKO mice were induced to migrate toward the chemokine, CCL19. As expected, M1 macrophages showed increased migratory activity towards CCL19, compared to M2 macrophages ( Figure 6A). The results indicate that LMaKO macrophages had a greater tendency to migrate, whereas LMbKO and LMabKO macrophages showed decreased migration.
To investigate the underlying mechanism, the effect of GSK3a and/or GSK3b deficiency on C-C chemokine receptor type 7 (CCR7) expression was examined (Figures 6B-D). M1 and M2 macrophages have similar levels of CCR7 protein (Figures 6B, C) and gene expression ( Figure 6D). GSK3a-deficient macrophages displayed increased expression of CCR7 in M1 macrophages ( Figures 6C, D). LMbKO and LMabKO macrophages displayed no difference in expression of CCR7 in M1 or M2 macrophages ( Figures 6C, D). These results suggest that GSK3a actively suppresses the expression of CCR7 expression. Analysis of other factors involved in macrophage migration, including sphingosine-1-phosphate receptor (S1PR) 1, S1PR3, and macrophage migration inhibitory factor (MIF), revealed no significant differences in gene expression (Supplementary Figure IX). Together these results suggest that GSK3a inhibits migration and GSK3b induces migration of M1 macrophages towards CCL19.
Analyzer XF24. The XF Cell Mito Stress analysis was used to measure mitochondrial activity. BMDM from Labfl/fl (control), LMaKO, LMbKO, and LMabKO mice were cultured for 24 hrs after which mitochondrial activity was measured ( Figure 7). As expected, M2 macrophages showed increased oxygen consumption rate (OCR) in comparison to M1 macrophages (Supplementary Figure XA) and M1 macrophages showed increased extracellular acidification rate (ECAR) in comparison to M2 macrophages (Supplementary Figure XB). Results suggest that LMabKO macrophages have an increased OCR (Figures 7A, C) and ECAR (Figures 7E, F) compared to Labfl/fl (control) in both M1 and M2 macrophages. LMabKO macrophages show a significant increase in basal, ATP-linked, spare respiratory capacity, and maximal OCR (Figures 7B, D). LMaKO and LMbKO macrophages showed no change in OCR in both M1 and M2 macrophages (Figures 7A, C). Together these results suggest that GSK3a and GSK3b play redundant roles in regulating metabolic activity.

DISCUSSION
GSK3a and GSK3b are highly homologous, constitutively active kinases that are expressed in most cells including macrophages. GSK3a and GSK3b function within central signal transduction pathways that regulate cell viability and metabolism and over 100 putative substrates have been identified (17). Dysregulation of GSK3a/b has been implicated in several metabolic disorders. Recent evidence suggests that GSK3a and GSK3b play distinct roles in sperm motility and fertility (35), amyloid production in the brain (13), cortical development (36), atherosclerosis development (37), and acute myeloid leukemia development (38). However, therapeutic targeting has been limited by our lack of understanding of homolog-specific functions as well as the lack of isoform-specific inhibitors. Here, we delineate the specific roles of GSK3a and GSK3b in macrophage polarization and pro-atherogenic functions. We show that myeloid GSK3a and GSK3b have distinct effects in the regulation of macrophage phenotype, inflammatory response, lipid accumulation, and migration ( Figure 8). Conversely, GSK3a and GSK3b appear to play complementary, redundant roles in macrophage viability, proliferation, and metabolism ( Figure 8).
The mechanisms regulating macrophage polarization involve specific signaling pathways. In M1 polarization, LPS and IFN-g interact with cell surface receptors to promote the activation of NF-kB and hypoxia-inducible factor 1-alpha (HIF1a), which induce expression of pro-inflammatory cytokines (TNFa, IL-1b) and iNOS (38). Our results suggest that GSK3a-deficient macrophages exhibit impaired LPS-induced gene expression of iNOS, STAT1 and CD38, resulting in impaired M1 polarization. This is consistent with our previous findings showing that GSK3a deletion suppresses M1 polarization in macrophages within the atherosclerotic lesion (29,30). GSK3b-deficient macrophages have increased LPS-induced gene expression of iNOS and STAT1. GSK3b-deficient macrophages also have decreased IL-4-induced Arg1, Fizz1 and Ym1 expression, resulting in impaired M2 polarization. Together, these data are consistent with a pro-inflammatory role for GSK3a, and an antiinflammatory role for GSK3b during macrophage polarization. Previous reports have suggested both pro-and antiinflammatory roles for GSK3b. Specifically, GSK3b has been shown to negatively regulate TLR4-mediated pro-inflammatory cytokine (IFN-b) production (39). In neonatal mouse cardiomyocytes and heart tissue culture, inhibition of GSK3b with chemical and genetic inhibitors enhanced LPS-induced proinflammatory cytokine expression (40). GSK3b has been shown to be capable of both activating and inhibiting NF-kB (41). GSK3b-deficient embryos have reduced NF-kB activity (24). Signaling through the ANO-PI3K-Akt pathway, which inhibits GSK3b, leads to increased binding of the cAMP response element-binding protein (CREB) to nuclear coactivator CREBbinding protein (CBP), resulting in suppression of the binding of NF-kB p65 to CBP (42). Another study suggests that inactivation of GSK3b via PI3K-Akt signaling pathway enhanced NF-kB activity, which leads to subsequent production of pro-inflammatory cytokines (43). Much less is known about the ability of GSK3a to modulate inflammation.
Our result suggests that GSK3a-deficient macrophages exhibit impaired LPS-induced expression of NF-kB. GSK3bdeficient macrophages have increased LPS-induced expression of NF-kB and NLRP3. This result is consistent with previous reports showing the inactivation of GSK3b enhanced NF-kB activity and increases the inflammatory response (44). Furthermore, our results show that GSK3a deficiency decrease the production of pro-inflammatory cytokines/chemokines (IL-1a, IL-1b, TNFa, IL-6, IL-12p70, MCP-1, MIP-2, KC) and GSK3b deficiency increase the production of pro-inflammatory cytokines/chemokines (IL-1a, IL-1b, MCP-1, MIG). This suggest that GSK3a positively-and GSK3b negatively-regulate NF-kB to induce pro-inflammatory response in LPS treated (M1) macrophages. Together, these results suggest that GSK3a and GSK3b function to regulate M1 and M2 polarization, with GSK3a promoting the pro-inflammatory response and GSK3b supporting the anti-inflammatory response in macrophages. Macrophages endocytose modified-LDL particles through the action of scavenger receptors (SR-A and CD36) and become foam cells. Consistent with previous reports (45), we found that M1 macrophages accumulate significantly more lipid compared to M2 macrophages. Furthermore, we show that lipid accumulation in M1 macrophages is reduced with GSK3a deficiency and increased with GSK3b deficiency. To determine the mechanism(s) underlying these observations we assessed the expression levels of genes involved in lipid uptake, export, and biosynthesis (46)(47)(48)(49). We found increased expression of genes encoding SR A and CD36 in GSK3b-and GSK3ab-deficient macrophages. This suggests that there is an overall increase in lipid transport in GSK3b-deficient macrophages that results in increased lipid accumulation. Furthermore, we found increased gene expression of ABCA1 and ABCG1 in GSK3b-and GSK3ab double-deficient macrophages. This suggests that the lipid accumulation leads to increased expression of genes involved in lipid efflux. Together, these findings suggest that GSK3a promotes lipid accumulation and GSK3b impedes lipid accumulation in M1 macrophages. The observation that GSK3b-deficient and GSK3ab double-deficient macrophages have similar phenotypes may indicate that GSK3a is regulated by GSK3b.
Our previously published findings have suggested that cardiovascular risk factors promote atherogenesis by a mechanism involving ER stress-induced activation of GSK3a/b (50,51). This pathway promotes lipid accumulation and apoptosis of macrophage/foam cells in vitro (19). We investigated the effect of ER stress on macrophage viability in GSK3a-or GSK3b-deficient cells. We found that myeloid deficiency of either the GSK3a or GSK3b isoform did not have any effect on cell viability. Deficiency of both isoforms significantly increased cell viability in both M1 and M2 macrophages. These results suggest that GSK3a and GSK3b are both required to facilitate ER stress-induced cell death pathways.
It has recently become evident that lesional macrophage proliferation may play an important role in atherosclerotic plaque development (52,53). The complete understanding of the molecular and cellular mechanisms that regulate macrophage proliferation in atherosclerotic lesions is still unknown. Our results show that there were significantly more Ki67 positive cells when macrophages were polarized to an M2 phenotype relative to an M1 phenotype. A previous study has shown that exposure to Th2 cytokines stimulates adipose tissue macrophage (M2) proliferation and inhibits M1 proliferation (54). Deficiency of GSK3a and/or GSK3b significantly increases the % of Ki67 positive cells. These results suggest that GSK3a and/or GSK3b suppress M2 macrophage proliferation.
To effectively carry out their phagocytic role within the artery wall, macrophages are required to migrate toward chemotactic signals within the lesional environment. Macrophage migration is known to be regulated by several factors including intracellular cholesterol content (55)(56)(57). We investigated the effect of GSK3a and/or b deficiency on CCL19-CCR7 stimulated migration using a transwell assay (58). Our results show that GSK3a-deficient macrophages are more proficient at migrating, whereas GSK3band GSK3ab double-deficient macrophages are less proficient at migrating, towards a CCL19 chemokine signal. To determine if migration ability was regulated at the level of chemokine receptor expression, we determined the expression of CCR7 in these GSK3a/b-deficient macrophages (58). As expected, we found that both M1 and M2 macrophages have similar CCR7 expression levels. GSK3a-deficient macrophages, but not GSK3b-deficient macrophages, had increased expression of CCR7 in M1 macrophages. Together, these results suggest that GSK3a inhibits the migration of M1 macrophages towards CCL19 by regulating CCR7 expression.
M1 macrophages are known to have enhanced glycolytic metabolism and reduced mitochondrial activity, whereas M2 macrophages rely upon mitochondrial oxidative phosphorylation (OXPHOS). The effect of metabolic changes in macrophages on the development of atherosclerotic plaque is poorly defined. Recent evidence suggests that a high degree of anaerobic glycolysis (59) and mitochondrial oxidative stress (60) in macrophages play an important role in the development of advanced atherosclerosis lesions. Our results indicate that deficiency of both GSK3a and GSK3b increases the mitochondrial activity (OCR) in both M1 and M2 macrophages. These results suggest that GSK3a and GSK3b play a redundant role in the mitochondrial activity of macrophages.
It is important to note that both the control BMDM and the GSK3a and/or GSK3b-deficient BMDM utilized in these experiments come from Ldlr-deficient mouse strains. In addition to modulating circulating lipid levels, the Ldlr is known to indirectly effect inflammatory responses and cellular metabolism. At the present time, we do not know how the presence or absence of the Ldlr might affect these results.
In summary, GSK3a and GSK3b play distinct, and often opposing roles in the signaling pathway of atherogenic functions. Consistent with previous findings, myeloid GSK3a signaling appears to play an important pro-atherogenic role, inducing M1 macrophage polarization and activating pro-atherogenic pathways to accelerate plaque development. Further investigations are needed to delineate the downstream substrates and pathways through which GSK3a acts and to explore the potential for targeting GSK3a with isoform specific small molecule inhibitors as an anti-atherogenic therapy.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
GW and SP conceived and designed the study. SP conducted and analyzed all experiments except Extracellular Flux assay.