Coronary artery tissues were obtained in accordance with the McGill University Health Centre guidelines. This study was approved by the Research Ethics Board of McGill University Health Centre and Montreal Heart Institute. We obtained written informed consent from all participants involved in this study. Healthy and diseased coronary arteries were obtained from heart transplants or autopsies between 2009 and 2012 at the McGill University Health Centre and Montreal Heart Institute. Fifty-eight coronary artery plaques were collected for examination. A detailed clinical history of the tissues can be found in Table 1. Tissues were either snap-frozen in liquid nitrogen or fixed in formalin and embedded in paraffin blocks.

Total RNAs were extracted from snap-frozen normal arteries (N = 10) and coronary plaques (N = 14) using the TRIzol (Invitrogen)/RNeasy Mini Kit (QIAGEN) combining protocol as previously described (24). Briefly, the first-strand cDNAs were synthesized using 1 µg of total RNAs with the iScript™ cDNA Synthesis Kit (Bio-Rad). The Real-Time qPCR was done using Advanced qPCR mastermix with SUPERGREEN Hi-ROX (Wisent Bioproducts) on the StepOnePlus™ Real-Time PCR System (Applied Biosystems™). The data were analyzed using PrimePCR Analysis (Bio-Rad), and the normalized relative mRNA expression was calculated against GAPDH expression. Sequences of gene-specific primers are listed in Supplementary Table S1.

A total of 58 patient donor samples were used in the immunohistochemistry studies. The paraffin-embedded tissue blocks of coronary arteries were cut into 4 μm sections using a microtome, six serial sections (the first one of all five consecutive sections) were examined per lesion (for each antiserum) at ∼20 μm intervals, and a total of six images were analyzed per section for histological scoring. The sections were incubated for 1 h in 10% normal goat serum/phosphate buffered saline (PBS) solution and then incubated overnight with primary antibodies in 0.1% bovine serum albumin (BSA)/PBS solution in humid chambers at 4°C. The following primary antibodies and concentrations were used: mouse antihuman Wnt3a (Novus, at 1/100), mouse antihuman Wnt5a (Sigma Aldrich, at 1/100), rabbit antihuman Wnt5b (Abcam ab115563, 1/500), α-smooth muscle actin (αSMA) (Cedarlane, at 1/500), and cluster of differentiation 68 (CD68) (Sigma Aldrich, at 1:2,000). Secondary biotinylated goat antirabbit or antimouse IgG (Vector Labs, BA1000, BA9200 at 1:200) was applied, followed by the use of the Vectastain ABC Elite Kit (Vector Laboratories, Burlingame, CA), according to manufacturer protocols. Immunostaining was visualized by using a 1xDAB/H₂O₂ solution, subsequently counterstained with hematoxylin, and mounted with Permount (Sigma Aldrich). Immunostaining without primary antibodies or immunoabsorption with the respective antigens was performed by using negative controls. The serial sections were used for colocalization studies. Normal control studies were performed on normal sections of the coronary arteries.

Coronary plaques were semiquantitatively scored in a blinded manner for overall immunostaining based on the intensity and distribution throughout the plaque (Supplementary Table S2). The plaques were also assessed to determine the overall levels of calcification, fibrosis, lipids, remodeling, and inflammation (25). The relationship between Wnt3a, Wnt5a, and Wnt5b immunostaining and coronary plaque histopathological features was also assessed as previously described.

Total proteins were extracted from tissues and cultured cells using a protein extraction buffer (8M Urea, 10 mM Tris, 1% SDS, pH 7.5). Briefly, frozen tissue was ground in liquid nitrogen, transferred into the extraction buffer, and then homogenized using a handheld homogenizer on ice. The cultured cells were washed with 1x PBS, scraped into a collection tube, and homogenized by passing them through a 22G needle 10 times. The extracted samples were spanned for 20 min at 4°C, and supernatants were collected; protein concentration was assayed using Protein DC reagents (Bio-Rad).

The total protein samples were separated on 10% SDS-PAGE and transferred into a PVDF membrane. The blot was blocked with 5% non-fat dry milk (BIO-RAD, #170-6404) in 1xTBST and incubated with primary antibodies overnight at 4°C in a blocking buffer. The following primary antibodies and concentrations were used: mouse antihuman Wnt3a (Novus, at 1/1,000), mouse antihuman Wnt5a (Sigma Aldrich, at 1/1,000), and rabbit antihuman Wnt5b (Abcam ab115563, 1/1,000). Secondary HRP-conjugated antibodies [Goat antimouse/Rabbit/Rat IgG (H + L) cross-absorbed, ThermoFisher, cat# G21040/G21234/A18870] were applied accordingly with primary antibodies. The SuperSignal™ West Pico PLUS Chemiluminescent Substrate (ThermoFisher, cat#34580) was applied and chemiluminescence images were taken and analyzed using the Chemidoc MP System (Bio-Rad).

The cell lines presented in this study were normal primary human aortic SMC lines (ThermoFisher Scientific) that were cultured in Medium-231 including the Smooth Muscle Cell Differentiation Supplement (ThermoFisher Scientific) in the presence of 200 ng/mL of Wnt3a, Wnt5a, and Wnt5b (R&D Systems). Cells from passages 3 to 5 were used throughout the study.

Cell migration was assessed by using 24-well Boyden chamber plates with an 8 µm pore size insert. Serum-free media were added to the bottom chamber and cells were seeded and treated with recombinant Wnt ligands in the top chamber for 24 h. The cells were washed with PBS, fixed with 4% paraformaldehyde, and stained with 0.05% crystal violet. Inserts were air-dried and the number of migrated cells was counted under a microscope.

CellROX™ Green Reagent (ThermoFisher Scientific) was used to measure mitochondrial superoxide, and the Amplex™ Red Hydrogen Peroxide/Peroxidase Assay Kit (ThermoFisher Scientific) was used to measure extracellular hydrogen peroxide, in accordance with the manufacturer's instructions. The cells were seeded in a 96-well black plate and treated with Wnt ligands for 2 h. Fluorescence intensity was measured using a TECAN Infinite® M200 Microplate reader (Tecan Group Ltd.) and normalized to total protein concentration.

Alizarin Red S staining was used to visualize the calcium deposition and calcification nodule formation. Briefly, SMCs were treated with or without 200 ng/mL Wnt proteins in an osteogenic medium containing 1 mM phosphate (OSM) for 10 days with fresh medium changed every 2–3 days. On the day of experimentation, the cells were fixed using 4% buffered formaldehyde and incubated with an Alizarin Red S stain solution (40 mM, pH4.2) for 20 min at room temperature. Calcification nodules were observed under microscopy, and phase-contrast images were taken with inverted microscopy. Calcium-bound Alizarin Red S was extracted by using a 10% cetylpyridinium chloride (Sigma) solution at room temperature for 60 min. Extracts were spanned for 5 min at 12,000 g and the supernatant was used for OD405mm reading on a spectrophotometer (BioTek). The cells were further washed with ddH₂O and scraped into a 1.5-mL centrifuge tube in a cell lysis buffer (20 mM Tris.HCl pH7.4 with 2% SDS, 0.2M glycine) and heated at 80°C for 60 min occasionally with a vortex. The total protein content was assayed as described above and used for Alizarin Red S normalization. For comparison, we also measured calcium deposition using the Arsenazo III protocol following acid calcium extraction and observed data similar to those of Alizarin Red S quantification.

SMCs were seeded in triplicates in 96-well culture plates and treated with 200 µg/mL of oxidized LDL (OxLDL) with or without 200 ng/mL Wnt proteins for 7 days in Medium-231 with fresh medium changes every 2–3 days. The OxLDL-treated cells for 7 days were used as controls. Total lipid was extracted using a 300 µL hexane: isopropanol (3:2 v/v) solution by incubating for 30 min on ice and was air-dried at 50°C. Total and free cholesterol were assayed using the Amplex Red Cholesterol Assay Kit (ThermoFisher Scientific). The lipid extract was reconstituted in a 120 µL 1xreaction buffer, 50 µL each was used for total and free cholesterol assay, and the cholesterol ester amount was calculated by subtracting free cholesterol from total cholesterol. The cells were lysed with 0.2N NaOH and protein concentration was assayed using the DC protein assay reagent (Bio-Rad) and used for normalization. A cholesterol standard curve was created by diluting a 2 mg/mL (5.17 mM) cholesterol reference standard into a 1x Reaction Buffer solution to produce cholesterol concentrations of 0–8 μg/mL (0 to ∼20 μM).

SMCs were seeded in triplicates in 96-well culture plates and treated with 200 µg/mL of OxLDL with or without 200 ng/mL Wnt proteins for 48 h in Medium-231. The OxLDL-treated cells for 48 h were used as controls. On the day of experimentation, the treated cells were loaded with 25 μM TopFluor Cholesterol (Avanti Polar Lipids, Alabaster, Ala) overnight. Cholesterol efflux was assayed using 10 μg/mL APOA1 after 6 h. Cholesterol efflux was calculated by dividing the measured cholesterol in the media by the measured cholesterol in the cell and media, multiplied by 100%.

An unpaired t-test was used to assess statistically significant differences in Wnt mRNA expression between diseased and normal arteries. A multivariate analysis was used to test the association of Wnt3a, Wnt5a, and Wnt5b immunohistochemical staining with the histological (overall levels of calcification, fibrosis, remodeling, inflammation, lipid score) and clinical parameters. Student’s t-test or one-way ANOVA, followed by Dunnett's multiple comparisons test, was used to assess differences between groups in the semiquantitative immunostaining experiments and in all in vitro assays. All values are presented as mean ± SEM. Statistical analyses were done using R (The R Project for Statistical Computing, Vienna, Austria), and all graphs were generated using GraphPad Prism 8 software. A score of P < 0.05 was considered statistically significant for all analyses.

There was a significant increase in the mRNA expression of Wnt3a and Wnt5b in coronary plaques compared with normal tissues (Figure 1A, P < 0.05). There was a significant increase in the protein expression of Wnt3a and Wnt5a in coronary plaques compared with normal control tissues (Figures 1B, C, P < 0.05). Although we observed an increase in Wnt5b protein expression, it did not reach statistical significance (P = 0.06). An analysis of Wnt receptor and coreceptor mRNA expression revealed an increased expression of FZD2, FZD9, and FZD10 in coronary plaques compared with normal tissues.

Negative control sections immunostained with non-immune serum or reabsorbed with the respective antigens showed no immunoreactivity (Figure 2A). Faint immunoreactivity for Wnt3a was seen in the medial and intimal layers of normal coronary artery tissues with no evidence of plaque development (Figure 2B; 1.59 ± 0.19). Diseased coronary arteries widely expressed intense Wnt3a immunoreactivity, especially around calcified lesions (3.34 ± 0.31, P < 0.0001) and in the fibrous cap (3.13 ± 0.27, P = 0.0002) (Figure 2E). Primarily, Wnt3a was evident in myointimal cells (Figures 2E, H) and infiltrating macrophages (Figures 2I, L) as demonstrated by colocalization studies with a-SMC actin and CD68. A semiquantitative analysis of Wnt3a immunostaining revealed significantly elevated immunostaining in the diseased sections of coronary arteries compared with histologically healthy sections of the same artery (Supplementary Table S3, P < 0.0001). In tissues with severe remodeling, Wnt3a showed a moderately elevated expression (2.90 ± 0.23, P = 0.0022). Wnt3a was also abundantly expressed in infiltrating macrophages around the lipid core (3.28 ± 0.27, P < 0.0001). There was also an apparent presence of Wnt3a expression in the endothelial cells of microvessels (2.59 ± 0.26, P = 0.0346).

Weak immunoreactivity for Wnt5a was seen in the medial and intimal layers of normal coronary artery tissues with no evidence of plaque development (Figure 2C). Focal intense Wnt5a immunostaining was seen in coronary plaques, mainly around the lesion (Figure 2F) and in myointimal cells (Figures 2F, H). No significant Wnt5a staining was found in infiltrating foam cells (Figures 2J, L).

Weak immunoreactivity for Wnt5b was seen in the medial and intimal layers of normal coronary artery tissues with no evidence of plaque development (Figure 2D; 1.56 ± 0.17). A stronger Wnt5b immunostaining was seen around the lesion and in the fibrous cap (Figure 2G). Intense Wnt5b staining was found in myointimal cells (Figures 2G, H) and infiltrating macrophages (Figures 2K, L) around the lipid core. Semiquantitative analyses of the immunohistochemical data revealed significantly greater immunostaining for Wnt5b in diseased coronary arteries compared with normal coronary arteries (Supplementary Table S3, P = 0.0022). For example, compared with histologically normal coronary arteries, Wnt5b immunostaining was elevated in the areas of fibrosis (2.70 ± 0.29, P = 0.0055), remodeling (2.60 ± 0.30, P = 0.0143), and infiltrating macrophages (2.90 ± 0.22, P = 0.0007).

There was a significantly higher scoring for Wnt3a in coronary plaques from male patients compared with female patients (Supplementary Table S4, P = 0.0015). No significant differences were found for Wnt5a or Wnt5b histological scoring and other clinical parameters.

SMCs treated with Wnt recombinant proteins reduced the expression of SMC markers alpha-smooth muscle cell actin and tropomyosin 1 (TPM1) (Figure 3A). Furthermore, all Wnt proteins were found to increase the expression of the osteoblast marker bone morphogenetic protein 2 (BMP2) (P < 0.0001), with the greatest fold change seen for Wnt3a. The latter was also found to increase the expression of macrophage marker CD68 (P < 0.0001), along with Wnt5a (P < 0.05) and Wnt5b (P < 0.0001). Interestingly, only the non-canonical ligands Wnt5a and Wnt5b induced significant oxidative stress in vSMCs (P < 0.05, Figure 3B), with only Wnt5b showing a significant generation of mitochondrial superoxides (P < 0.05, Figure 3C). Treatment of vSMCs with Wnt3a, Wnt5a, and Wnt5b resulted in a profound increase in migratory ability (Figure 3D), with Wnt3a having the largest effect (P < 0.01). SMCs treated with Wnt5b showed significant calcium deposition (P < 0.05), with no changes seen in cells treated with Wnt3a or Wnt5a (Figures 3E, F).

SMCs treated with OxLDL and Wnt3a, Wnt5a, or Wnt5b showed an increased cholesterol ester to total cholesterol ratio, compared with cells treated with OxLDL alone (P < 0.05, P < 0.05, and P < 0.01, respectively, Figure 4A). Furthermore, OxLDL-treated vSMCs, in conjunction with non-canonical Wnt5a or Wnt5b, significantly reduced apolipoprotein A1 (APOA1)-dependent cholesterol efflux, with no changes seen for Wnt3a (P < 0.05, Figure 4B).

Non-canonical Wnt5a and Wnt5b reduced ATP binding cassette subfamily A member 1 (ABCA1) mRNA expression, with no changes seen for canonical Wnt3a (P < 0.0001, P < 0.001, respectively, Figure 4C). All proteins were found to significantly reduce the expression of plasma membrane receptor proteins, low-density lipoprotein receptor–related protein 5 (LRP5) (P < 0.001) and LRP6 (P < 0.001). There was also an increased expression of low-density lipoprotein receptor (LDLR) mRNA after treatment with all Wnt proteins (P < 0.001), while only Wnt3a and Wnt5b increased the expression of oxidized low-density lipoprotein receptor 1 (OLR1) mRNA. In addition, all Wnt proteins significantly reduced the expression of the cholesterol transport ras-related protein rab-18 (RAB18), which is involved in lipid transport. Lastly, Wnt3a (P < 0.0001) and Wnt5a (P < 0.0001) increased the expression of the cholesterol synthesis gene, sterol regulatory element-binding transcription factor 1 (SREBF1), while all Wnt proteins increased the expression of SREBF2 (P < 0.0001). A summary of these findings is shown in Supplementary Figure S1.