Wnt16 Promotes Vascular Smooth Muscle Contractile Phenotype and Function via Taz (Wwtr1) Activation in Male LDLR−/− Mice

Abstract

Wnt16 is expressed in bone and arteries, and maintains bone mass in mice and humans, but its role in cardiovascular physiology is unknown. We show that Wnt16 protein accumulates in murine and human vascular smooth muscle (VSM). WNT16 genotypes that convey risk for bone frailty also convey risk for cardiovascular events in the Dallas Heart Study. Murine Wnt16 deficiency, which causes postnatal bone loss, also reduced systolic blood pressure. Electron microscopy demonstrated abnormal VSM mitochondrial morphology in Wnt16-null mice, with reductions in mitochondrial respiration. Following angiotensin-II (AngII) infusion, thoracic ascending aorta (TAA) dilatation was greater in Wnt16−/− vs Wnt16+/+ mice (LDLR−/− background). Acta2 (vascular smooth muscle alpha actin) deficiency has been shown to impair contractile phenotype and worsen TAA aneurysm with concomitant reductions in blood pressure. Wnt16 deficiency reduced expression of Acta2, SM22 (transgelin), and other contractile genes, and reduced VSM contraction induced by TGFβ. Acta2 and SM22 proteins were reduced in Wnt16−/− VSM as was Ankrd1, a prototypic contractile target of Yap1 and Taz activation via TEA domain (TEAD)-directed transcription. Wnt16−/− VSM exhibited reduced nuclear Taz and Yap1 protein accumulation. SiRNA targeting Wnt16 or Taz, but not Yap1, phenocopied Wnt16 deficiency, and Taz siRNA inhibited contractile gene upregulation by Wnt16. Wnt16 incubation stimulated mitochondrial respiration and contraction (reversed by verteporfin, a Yap/Taz inhibitor). SiRNA targeting Taz inhibitors Ccm2 and Lats1/2 mimicked Wnt16 treatment. Wnt16 stimulated Taz binding to Acta2 chromatin and H3K4me3 methylation. TEAD cognates in the Acta2 promoter conveyed transcriptional responses to Wnt16 and Taz. Wnt16 regulates cardiovascular physiology and VSM contractile phenotype, mediated via Taz signaling.

Cardiovascular diseases remain the leading causes of death for both men and women (1). Endocrine- and metabolism-directed pharmacotherapies (2), with advanced percutaneous intervention technologies (3), have significantly improved cardiovascular disease outcomes. However, treatment of certain cardiovascular disorders—including aortic valve (4) and thoracic aortic diseases (5)—remain poorly responsive to medical management. Either insufficient mechanical integrity (aortic root dilation, aneurysm with rupture or dissection) or excessive mechanical stiffness of thoracic aorta (6) and aortic valve (7) compromise the finely tuned tissue compliance necessary to sustain healthy conduit artery function (8). Paracrine Wnt (wingless-type MMTV integration site family member) signals have emerged as important in regulation of extracellular matrix metabolism and mineralization of aorta and aortic valve (9, 10). This is similar to actions of Wnt in bone, where contributions are well recognized (11). Intriguingly, many clinical examples exist wherein genetic, endocrine, or metabolic disorders produce concomitant vascular and bone disease (7, 12-14). This includes genetic disorders such as Marfan, Loeys-Dietz, Ehlers-Danlos, and certain osteoporosis imperfecta syndromes (1, 15-17); these disorders increase risk for bone loss as well as thoracic ascending aorta (TAA) aneurysm. The roles for Wnt signaling in vascular tissue remodeling, including TAA aneurysm, are poorly understood.

Vertebrate Wnts are a family of 19 secreted polypeptides that convey paracrine signals, mediated via heteromeric cell-surface receptors (Frizzled, low-density lipoprotein receptor related [LRP], and ROR/RYK proteins) with impressive combinatorial complexity (9, 18). Mani and colleagues first identified the Wnt co-receptor LRP6 as important in control of the coupled bone-vascular disease risks, revealed by a private LRP6(R611C) mutation in an Iranian family afflicted with precocious osteoporosis and cardiovascular disease (12). His group (19) and ours (20, 21) have demonstrated the role of LRP6 in limiting pathological VSM phenotypic modulation; this is mediated in part through LRP6 inhibition of specific noncanonical Wnt pathways (20, 21) and preservation of TCF7L2- directed, homeostatic canonical Wnt programs (19). However, the roles of specific Wnt ligands in the pathogenesis or mitigation of bone-vascular disease biology (7) remain poorly studied.

Wnt16 is an enigmatic member of this family of proteins, first identified as important in the accrual of fracture-resistant bone mass and structure in both mice (22) and humans (23). Importantly, Wnt16 has no overt impact on skeletal morphogenesis or mortality during murine development or early postnatal life; however, with aging, Wnt16-null mice develop spontaneous fractures, most notable in hindlimbs, due to thinning of cortical bone (22). Wnt16 supports bone-forming osteoblast function and inhibits bone-resorbing osteoclast activities that remodel postnatal bone to regulate skeletal mass and geometry—and thus bone strength (22). In our systematic analyses of LRP6-regulated arterial calcification, we noted that Wnt16 is expressed in aortic VSM (9). However, the role of Wnt16 in arterial physiology and cardiovascular disease is unknown.

We hypothesized that Wnt16 may play a role in postnatal arterial physiology in a mode analogous to its role in postnatal skeletal physiology, remodeling, and strength. Therefore, we studied the impact of Wnt16 signaling on VSM differentiation (24, 25) and cardiovascular disease physiology, including angiotensin-II (AngII)-dependent aneurysmal remodeling (26). We show that Wnt16 protein is expressed in both mouse and human arterial VSM, and that aortic Wnt16 expression is increased by AngII-induced arterial remodeling. In the Dallas Heart Study (27), we identify WNT16 nonsynonymous coding single nucleotide polymorphisms (cSNPs) shown previously to reduce bone mass and increase fracture risk (23, 28) in humans also convey increased risk for composite cardiovascular disease events (stroke, myocardial infarction, other atherosclerotic cardiovascular disease, heart failure, cardiovascular death). Furthermore, we identify that Wnt16 supports the arterial VSM contractile phenotype including Acta2 (VSM alpha actin) expression and mitochondrial respiration as necessary for VSM contraction (29, 30)—and that Wnt16 serves to limit aneurysmal remodeling in the thoracic ascending aorta (TAA) (29, 30). Mechanistically, we show that Wnt16 promotes the VSM contractile phenotype in part via a Taz (also known as Wwtr1), a transcriptional co-adapter inhibited by Lats (large tumor suppressor kinase, also known as Warts) kinases (31) and a component of a recently described alternative Wnt signaling relay (32) Thus, in addition to previously known roles in bone (22, 23), Wnt16 signaling participates in postnatal arterial physiology, function, and structure.

Materials and Methods

Biochemical, Tissue Culture Supplies, and Reagents

Immunoreagents are listed in Table 1, and all reagents are listed in Supplementary Table S1 (33). Tissue culture reagents were purchased from Thermo Fisher Scientific. Molecular biology reagents, enzymes, biochemicals, and synthetic oligodeoxynucleotides were purchased from Thermo Fisher, Sigma-Aldrich, and other vendors as indicated in Supplementary Table S1 (33). Luciferase enzyme assay reagents were purchased from Promega. Commercially available antibodies, TaqMan gene expression assays (Life Technologies), and small interfering RNA (siRNA)/RNA interference (RNAi) reagents used for knockdown (Dharmacon Horizon Discovery) in primary VSM are listed in the Supplementary Table S1 (33), along with primers used for chromatin immunoprecipitation (ChIP) assay analyses. Plasmids generated by the Towler laboratory, purchased from commercial sources, or provided as gifts via Addgene, are presented in the Supplementary Table S1 (33).

Table 1.

Immunoreagents

Protein target/immunogen	Vendor	Catalog number; RRID
Wnt16 antibody polyclonal rabbit anti-mouse/human Wnt16	Novus Biologicals	NBP1-86403; RRID AB 11034126
Alexa Fluor 594 donkey anti-rabbit IgG (L + H)	Abcam	ab150064; RRID AB 2734146
CD31 AlexaFluor488 conjugated rat monoclonal anti-mouse CD31	Novus Biologicals	NB100-1642 488; RRID: AB 11009953
Yap1 + Taz ab for Western blot (Taz ChIP antibody below)	Cell Signaling Technology	8418S; RRID: AB 10950494
GAPDH antibody	Santa Cruz	SC-20357; RRID: AB 641107
Histone H3 antibody	Abcam	ab1791; RRID: AB 302613
CCM2 antibody	ProteinTech Group	26270-1-ap; RRID: AB 2880454
Wnt16 rabpAb	Novus Biologicals	NBP1-86403; RRID; AB 11034126
Acta2 rabpab	ProteinTech Group	14395-1-ap; RRID: AB 2223009
Sm22 / Transgelin Antibody	Abcam	ab14106; RRID: AB 443021
Vinculin Antibody	BosterBio	A01207-1; RRID: AB 3075459
RPS21 antibody	ProteinTech Group	16946-1-AP; RRID: AB 2180351
ERAB ab	ProteinTech Group	10648-1-AP; RRID: AB 2264270
Alkaline phosphatase–conjugated Goat Anti-Rabbit IgG Antibody	Vector Laboratories	AP-1000; RRID AB 2336194
Wnt16 primary ab, rabbit anti-mouse/human Wnt16	Novus	NBP1-86403; RRID: AB 11034126
Secondary Alexa Fluor 488 goat anti-rabbit IgG (L + H)	Thermo Fisher	A-11070; RRID: AB 2534114
Anti-Smooth muscle actin mAb -Cy3 conjugated (mouse anti-human/mouse/multiple species)	Sigma-Aldrich	C6198; RRID: AB 476856
Anti-Smooth muscle actin mAb-FITC conjugated (mouse anti-human/mouse/multiple species)	Sigma-Aldrich	F3777; RRID AB 476977
Alexa Fluor 594 donkey anti-rabbit IgG (L + H)	Abcam	ab150064; RRID AB 2734146
CD11b magnetic microbeads	Miltenyi	130-049-601; RRID: AB 2927377
H3K27me3 ChIP Antibody	Diagenode	C15410069; RRID AB 2814977
H3K4me3 ChIP Antibody	Diagenode	C15410069; RRID: AB 2924768
Taz(Wwtr1) ChiP Antibody (E8E9G)	Cell Signaling Technology	#83669; RRID: AB 2800026
Normal rabbit IgG control ChIP Antibody	Cell Signaling Technology	#2729; RRID AB 1031062
Aldosterone ELISA Kit	Enzo Life Sciences	ADI900-173; RRID: AB 3075458
C3 ELISA	Abcam	ab263884; RRID AB 3075457
CD11b magnetic microbeads	Miltenyi	130-049-601; RRID AB 2927377

Transgenic Mice, Dietary and AngII Challenge, Pulse Wave Analysis, Primary Aortic VSM Cultures and Bone Marrow Monocyte Cultures, Subcellular Fractionation, and Mass Spectrometric Characterization of Cellular and Vesicular Protein Fractions

All procedures for studying mice were approved by the University of Texas (UT) Southwestern Institutional Animal Care and Use Committee. The Wnt16−/− mice (Wnt16tm2a(EUCOMM)Wtsi/Mmucd, RRID:MMRRC_037042-UCD), were obtained from the Mutant Mouse Resource and Research Center (MMRRC) at University of California at Davis, donated to the MMRRC by Ramiro Ramirez-Solis, PhD, Wellcome Trust Sanger Institute. Mice were generated at the Sanger Institute as part of the Baylor College of Medicine, Sanger Institute, and MRC Harwell (BaSH) Consortium for the NIH Common Fund program for Knockout Mouse Production and Cryopreservation (1U42RR033192-01) and Knockout Mouse Phenotyping (1U54HG006348-01). Deploying methods described previously (20, 34), Wnt16−/− mice were backcrossed onto the low-density lipoprotein receptor (LDLR)−/− background (35) for study. The LDLR−/− (35) (strain #002077), Myh11-Cre (36) (strain #007742), and Taz(fl/fl);Yap1(fl/fl) (37) (strain #030532) transgenic mice were obtained from The Jackson Laboratory. Genotyping primers are listed in Supplementary Table S1 (33). All experimental animals and controls used to generate primary aortic VSM cells were male sibling cohorts (males susceptible to diet-induced diabetes and arteriosclerosis) (38, 39).

For studies involving challenge Western high fat diet (HFD; TD.88137, Envigo) to induce arterial calcification, male siblings of the indicated genotypes were fed HFD beginning at 2 months of age (20, 40). Aortic pulse wave analysis (PWV), an index of arterial stiffness (41), was measured in the thoracoabdominal aorta using an Indus Doppler Flow Velocity System (Indus Instruments, Webster, TX) outfitted with 2 20-MHz transducers using the method of Hinton, Reddy, and colleagues (41). Tribromoethanol (TBE)-anesthetized mice were monitored on an Indus Rodent Surgery platform (warming pad set at 35-40 °C), with one transducer positioned at the aortic arch, and the second at the abdominal aorta below the takeoff of the renal arteries with caliper-measured distance between probes as detailed (41). Foot-to-foot Doppler assessments of aortic pulse wave velocity (PWV) (41, 42) were determined 3 times for each animal, then values averaged to provide the PWV value for that individual. Tail cuff systolic blood pressures were also obtained under the above light TBE anesthesia with warming, using the CODA Monitor Noninvasive Blood Pressure System as detailed (43). Aortic primary VSM cells were isolated the proximal thoracic aortas from genetically engineered mice as detailed previously (20, 21, 34, 44) using a timed digestion protocol with type 1 collagenase (Worthington Biochemical Corporation, cat. # LS004149, 1 mg/mL), DNase I (Sigma-Aldrich, D5025, 60 U/mL), elastase (Worthington Biochemical cat. #LS002279, 0.8 mg/mL), and hyaluronidase (Sigma-Aldrich, H3506, 0.5 mg/mL) in DMEM; 2X penicillin/streptomycin (200 IU/mL and 200 ug/mL, respectively) and fungizone (2.5 µg/mL) were included in initial isolation and passage to inhibit microbial growth. For routine experiments, pools of cells isolated from 4 to 8 male animals per genotype were expanded with passage on collagen in growth media as described previously (20, 21, 34, 44). Primary bone marrow monocytes were prepared using Miltenyi CD11b magnetic microbeads with 40 ng/mL Csf1 precisely following the manufacturer's instructions (cat. #130-049-601; RRID: AB 2927377. Results were confirmed in at least 2 independent primary culture preparations (range 2-6), with n = 3 to 4 replicates per condition as indicated.

Aneurysm Induction and Ex Vivo Aorta Image Analyses

Implantation of Alzet osmotic minipumps to deliver angiotensin-II (AngII) for 4 weeks to induce aortic aneurysms compared to saline drug-free controls. All procedures for studying mice including this survival surgery protocol were reviewed and approved by the UT Southwestern Institutional Animal Care and Use Committee. The protocol with ex vivo aortic image analyses follows that of Cassis et al and Daugherty et al (26, 45). Lyophilized human angiotensin-II (AngII; Sigma #A9525) was prepared in sterile saline in a laminar flow hood, with concentration adjusted to deliver 1000 ng/kg/min with a pump rate of 0.25 uL/h in a 250 uL Alzet pump (model #2004). Male mice (5-6 months old) of the indicated genotypes underwent general anesthesia with isoflurane (1%-3%) and were placed on a 35 to 40 °C warming pad (Indus Instruments, Surgical Monitor). Following shaving of the flank with a handheld, battery-powered electric razor and skin sterilization (betadine, 70% ethanol), a sterile scalpel incision over the right shoulder blade was opened with hemostat subcutaneous tunneling to create space for placement of the sterile minipump. The incision was closed with 4-0 nylon, 0.5 cc of warmed sterile 37 °C saline administered subcutaneously for hydration on the contralateral side, and analgesia achieved with meloxicam 5 mg/kg s.c. once daily for 2 days. Mice recovered with warming, monitored every 15 minutes until awake. Mice were observed twice a day during the first 2 days after surgery, and daily once thereafter. Sutures were removed 10 to 14 days after surgery. Following 4 weeks of infusion animals were euthanized as above, hearts and aortas removed en bloc and adventitial fat trimmed under dissecting microscope (Olympus Model #SZ-ET), and high-resolution aortic images captured (Leica Z16 APO apochromatic optical zoom system with Leica MC170 HD high definition camera) and aneurysmal dilatation digitally quantified (FIJI / ImageJ (46)) as detailed by Daugherty and colleagues (45). A set of 5 to 7 animals were used per treatment and genotype.

Protein Extraction and Proteomic Analyses

For subcellular fractions of Wnt16−/−;LDLR−/− and control LDLR−/− VSM, confluent monolayers in 10-cm culture dishes were processed as before (20) using Pierce Subcellular Protein Fractionation Kit for Cultured Cells (Cat #78840) according to the manufacturer's protocol. Liquid chromatography–tandem mass spectrometry (LC-MS/MS) techniques deployed to identify and quantify proteins in extracts of the microdissected ascending aorta and arch as we have detailed previously (20, 44) (n = 3-5 independent aorta extract replicates per genotype). Briefly, following euthanasia by exsanguination via cardiac puncture under heavy (300 mg/kg) TBE general anesthesia, the heart and thoracic aorta (ascending aorta, aortic arch, descending thoracic aorta) were removed en bloc. The ascending aorta and arch to the left carotid was then microdissected and each aorta individually homogenized in 500 uL of 1.25 × Laemmli buffer supplemented with Sigma Protease inhibitor cocktail, using a Qiagen TissueLyser II with 5 mm stainless steel bearings. Equal amounts of ascending aorta + arch protein extract was partially resolved by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) as before (44), stained protein identified and gel cut, extensively de-stained and washed to remove SDS, and gel pieces with proteins reduced and alkylated with DTT (20 mM) and iodoacetamide (27.5 mM). A 0.1 µg/µL solution of trypsin in 50 mM triethylammonium bicarbonate was then added to completely cover the gel, allowed to sit on ice, and then 50 µL of 50 mM triethylammonium bicarbonate was added and the gel pieces were digested overnight (Pierce). Following solid-phase extraction cleanup with an Oasis HLB microelution plate (Waters), the resulting peptides were reconstituted in 10 uL of 2% (v/v) acetonitrile (ACN) and 0.1% trifluoroacetic acid in water. Two uL of this were injected onto an Orbitrap Fusion Lumos mass spectrometer (Thermo Electron) coupled to an Ultimate 3000 RSLC-Nano liquid chromatography systems (Dionex). Samples were injected onto a 75 μm i.d., 75-cm long EasySpray column (Thermo), and eluted with a gradient from 1% to 28% buffer B over 90 minutes. Buffer A contained 2% (v/v) ACN and 0.1% formic acid in water, and buffer B contained 80% (v/v) ACN, 10% (v/v) trifluoroethanol, and 0.1% formic acid in water. The mass spectrometer operated in positive ion mode. Mass spectrometer scans were acquired at 120 000 resolution in the Orbitrap and up to 10 MS/MS spectra were obtained in the ion trap for each full spectrum acquired using higher-energy collisional dissociation for ions with charges 2 to 7. Dynamic exclusion was set for 25 seconds after an ion was selected for fragmentation. As before (44), raw MS data files were analyzed using Proteome Discoverer v2.4 (Thermo), with peptide identification performed using Sequest HT searching against the mouse protein database from UniProt. Fragment and precursor tolerances of 10 ppm and 0.6 Da were specified, and 3 missed cleavages were allowed. Carbamidomethylation of Cys was set as a fixed modification, with oxidation of Met set as a variable modification. The false discovery rate cutoff was 1% for all peptides. Label-free peptide spectral counts from LC-MS/MS spectral data (44) were then used to quantify peptides from identified proteins in Wnt16−/−;LDLR−/− vs LDLR−/− control samples. The peak intensities for all peptides matched to a protein were summed, and these values were used to quantify peptides (Supplementary Table S2 (33)).

RNA Analysis

Total RNA of aortic primary cells was isolated using RNeasy Mini kit (Qiagen), whereas total RNA of whole aorta was isolated using RNeasy Lipid Tissue Mini kit after homogenization with Qiagen TissueLyser II, using 5 mm stainless steel bearings at 30 Hz for 5 minutes, in QIAzol Lysis Reagent and chloroform extraction, as described previously (20, 34, 44). Messenger RNA was quantified by real-time fluorescence reverse-transcriptase polymerase chain reaction (RT-PCR) as previously described (20) using commercially available, inventoried Taqman Probes (Applied Biosystems, Foster City, CA). Relative mRNA levels were normalized to 18S ribosomal RNA or housekeeping gene (glyceraldehyde 3-phosphate dehydrogenase [GAPDH]) accumulation as indicated.

Plasmids, siRNA, and Transfection

The murine 4 kb Acta2 LUC (firefly luciferase reporter gene) and 1.3 kb C3LUC promoter–luciferase reporter constructs were prepared in pGL3-Basic (Promega) using methods detailed previous (44, 47), using genomic DNA as a template. Amplimer sequences are provided in Supplementary Table S1 (33). Numbering of the Acta2 promoter is based upon the phylogenetically conserved CACCAC initiator region identified by Owens and colleagues (48) (+1 underlined above as transcription start site, TSS; base pair 34 252 619 on mouse chromosome 19 per UCSC Genome Browser), placing the TATAA box (48) at nucleotides −31 to −27, CArG (C-A/T rich- G binding element CC(A/T)6GG) box B (48) at −124 to −116, and the intronic CArG box (49) at +2131 to +2140 relative to the TSS. The complement C3 promoter initiator region CACCCC (base pair 57 228 081 on mouse chromosome 17 per UCSC Genome Browser) places the TATAA box at nucleotides −41 to −37 relative to the C3 TSS. Site-directed modifications of parental plasmids listed in Supplementary Table ST1 (33), including the Acta2LUC and C3LUC promoter mutants described below, were generated by using the In-Fusion HD Cloning kit (cat. # 639649) according to the protocol provided by Clontech (PT5162-1). All promoter fragments and point mutants functionally analyzed were authenticated by DNA sequencing. The pCMV-based expression vector for human Taz (also known as Wwtr1) was obtained from Origene (TruORF Gold). CMV-Renilla (E2261) was obtained from Promega. The remaining plasmids were purchased either from Origene or from Addgene as indicated. For all plasmid transfections, HEK293 T cells (ATCC; CRL-3216) or A7r5 rat aortic smooth muscle cells (ATCC; CRL-1444) were seeded the day before at 65 000 cells/well (A7r5) or 10 000 cells / well (HEK293 T) in 12 well culture plates, and transient transfections and reporter assays carried out as detailed previously (44), with Qiagen Polyfect. All transfections were performed in independent triplicates, each measured twice, and independent experiments performed at least twice. As before (21, 44), for siRNA “knockdown” transfections aortic primary cells were seeded at 100 000 cells/well in type I collagen- coated 12 well culture plates the day before transfection. Control siRNA or targeted siRNA (60 nM total siRNA; siRNAs listed in Supplementary Table S1 (33)) at the indicated concentrations were incubated with Lipofectamine RNAiMAX reagent (3 uL/well, ThermoFisher Invitrogen, cat. #13778-150) in Opti-MEM for 20 minutes. During incubation, cells were washed once with DMEM medium and fresh DMEM growth medium containing 10% fetal bovine serum (FBS) was added (0.9 mL/well). At the end of incubation, siRNA mixture (100 uL/well) was added in triplicate to cells. As indicated, cells were harvested for either RNA isolation for quantitative RT-PCR or protein extraction for Western blot as indicated 3 days later.

Western Blot Analyses, Immunohistochemistry, and Chromatin Immunoprecipitation Assays

Western blot analyses were performed as before (21, 44) using either whole cell extracts or extracts subjected to subcellular fractionation as indicated. For whole cell extracts, cell cultures were twice rinsed with 1X TBS and extracted with modified RIPA buffer containing 10 mM Tris HCl, pH 8.0, 1 mM EDTA, 140 mM NaCl, 1% Triton X-100 and 0.5% NP40 plus cocktails of protease inhibitors and phosphatase inhibitors (Sigma). Subcellular fractions, confluent cells in 10 cm culture dishes were processed using Pierce Subcellular Protein Fractionation Kit for Cultured Cells (Cat # 78840) according to the manufacturer's protocol. Extracts and fractions containing equal protein mass were subjected to SDS-PAGE and Western blotting was performed as described previously (21, 44), with immune complexes visualization by enhanced chemiluminescence. Gel bands intensities were quantified as described previously (44) using Fiji ImageJ (46) for Mac OS X. Immunolocalization was carried out for Wnt16 (RRID: AB 11034126) using methods we have described previously (20, 21, 44) using a Leica TCS SP8 Confocal Microscope and the addition of 5% normal donkey serum to the blocking admixture. Five-micron frozen sections of aorta and kidney were prepared from our murine cohorts, and 10-micron frozen sections of peripheral human artery and arteriole were purchased from a commercially available tissue repository (Origene). The antibodies utilized in Western blot analysis and immunohistochemistry are listed in Table 1 with RRIDs. Taz chromatin immunoprecipitation (ChIP) (RRID: AB 2800026) was carried out following our modification of the previously reported fast ChIP method of Nelson et al (50) as we have detailed (20, 44), but using 15 sonication cycles (30 seconds on, 30 seconds off, Diagenode Bioruptor Pico Sonication System) with cooling prior to clearing and immunoprecipitation. One pg of mouse genomic DNA was assessed in parallel aliquots as a standard to normalize quantitation between each plate and experiment. H3K4me3 methylation ChIP (RRID: AB 2924768) utilized the True ChIP-Seq kit protocol from Diagenode. Normal polyclonal rabbit serum (RRID: AB 1031062) was used as a control. ChIP antibodies and primer pairs are presented in Table 1 and Supplementary Table S1, respectively (33). Cycling conditions were (94 °C × 10 minutes, then 40 cycles of 94 °C × 30 seconds, 60 °C × 30 seconds, 68 °C × 30 seconds), using a BioRAD CFX Connect Real-Time PCR Detection System with Sybr Green dye (20).

Floating Matrix Model Cell Contraction Assays

The floating matrix model cell contraction assay (51) (Cell CytoSelect™ 48-Well Cell Contraction Assay Kit, Cell Biolabs Inc.) was carried out per the manufacturer's protocol (product manual CBA-5021), using collagen as the culture matrix, seeding 140 000 primary VSM cells per well (n = 5-6 independent replicates per genotype and treatment) prepared from Wnt16−/−;LDLR−/− and LDLR−/− control (Wnt16+/+) aortas as described above. The following day, cultures were incubated with 10 ng/mL recombinant human transforming growth factor β1 (rhTGF-β1) in 10% fetal bovine serum in DMEM (25 mM glucose), and 48 hours later floating collagen gel diameters were digitally imaged and analyzed using FIJI/ImageJ (46)– quantifying the area of each contracted gel seeded with either LDLR−/− vs Wnt16−/−;LDLR−/− primary VSM as indicated. Contractile responses to 400 ng/mL recombinant Wnt16 treatment were carried out using the same protocol, implementing graded verteporfin concentrations (0 uM, 0.5 uM, 2 uM) in DMSO vehicle to inhibit Yap/Taz signaling (52, 53). N = 5-6 independent experimental replicates per treatment.

Transmission Electron Microscopy

Analysis of ascending aorta tunica media of Wnt16−/−;LDLR−/− and LDLR−/− control (Wnt16+/+) mice by transmission electron microscopy was carried out via the UT Southwestern Electron Microscopy Core Facility (Dr. Kate Luby-Phelps, Director). Briefly, under general anesthesia Wnt16−/−;LDLR−/− and LDLR−/− control mice were perfused with 4% paraformaldehyde and 1% glutaraldehyde in 0.1 M cacodylate buffer pH 7.4, and the ascending aortas were isolated as described above. Following overnight fixation in 2.5% glutaraldehyde in cacodylate, specimens were rinsed again and secondary fixation achieved with 1% OsO₄ and 0.8% K₃Fe (CN)₆ in 0.1 M cacodylate pH 7.4 for 1.5 hours at room temperature. After rinsing 3 times with distilled water, samples were prestained en bloc in 4% uranyl acetate in 50% EtOH for 2 hours. Specimens were then dehydrated through graded EtOH, transitioned in propylene oxide, and sequentially infiltrated with (a) EMBed-812:propylene oxide (1:2), then (b) 100% EMBed-812, and polymerized in embedding molds in a 60 °C oven overnight. Cross-sections of ascending aorta (70 nm, Ultracut 7 ultramicrotome, Leica) from Wnt16−/−;LDLR−/− and Wnt16+/+;LDLR−/− control mice were prepared and collected onto copper grids and grid stained with 2% aqueous uranyl acetate:lead citrate. Samples were then imaged on a JEOL 1400 Plus transmission electron microscope equipped with a LaB6 source operating at 120 kV, capturing images with a BioSprint 16 M camera (AMT Imaging; exposure 1500 msec × 1 std. frame, with Gain 2/Bin 1/Gamma 1.00, no sharpening and normal contrast).

Seahorse Oximetry Mito Stress Testing

Mitochondrial respiratory function (54) in intact primary VSM cultures was carried out using a Seahorse XF Flux Analyzer (Agilent) (55). Primary VSM per plated at 10 000 cells/well in rat tail collagen–coated XF FluxPaks. Per the manufacturer's XF Mito Stress Test protocol (Agilent #103010-100), analysis was carried out with 25 mM glucose / 1 mM pyruvate / 6 mM glutamine in phenol red–free minimal DMEM excess base media (Agilent #102353-100), assessing oxygen consumption rate at baseline and following sequential oligomycin (1 uM), FCCP (1 uM) and rotenone + antimycin A (1 uM each) to permit quantitation of basal and maximal respiration, subtracting the measured mean background of nonmitochondrial respiration. Spare mitochondrial respiratory capacity is the difference between maximal and basal oxygen consumption rate measurements (54). Three time points were measured and the values averaged at basal, proton leak, maximal, and nonmitochondrial respiration states (54). Oxygen consumption rate data are normalized to the DNA content per well as an index of cell number (55). A minimum of n = 3 independent samples were assayed for each genotype and treatment as indicated, with all experiments independently performed at least twice.

Dallas Heart Study and Analysis of WNT16 Genotype Association With Cardiovascular Disease Events

The association of WNT16 nonsynonymous cSNP variants (rs2908004 and rs2707466) with cardiovascular events was tested among participants of the Dallas Heart Study–1, for whom DNA samples were available and who had consented to long-term follow-up. The Dallas Heart Study is a population-based probability sample of Dallas County, with deliberate oversampling of Black residents, as described previously (27). The Dallas Heart Study was approved by the UT Southwestern Institutional Review Board and all participants provided written informed consent. The primary endpoint was a composite cardiovascular disease outcome, defined as a first nonfatal myocardial infarction, nonfatal stroke, coronary revascularization (percutaneous coronary intervention or coronary artery bypass grafting), peripheral revascularization, hospitalization for heart failure or atrial fibrillation, or death from cardiovascular causes. Nonfatal events were captured through annual follow-up calls to study participants and through examination of hospital admission records. All endpoints were adjudicated by 2 cardiologists who were unaware of genotype status or other study variables. The National Death Index was used to determine vital status for all participants through December 31, 2016. Death from cardiovascular causes was defined according to the International Classification of Diseases, Tenth Revision, codes I00 to I99. Genomic DNA was extracted from circulating leukocytes as previously described (56). Genotypes for these WNT16 rs2908004 and rs2707466 variants, and FAMC3 rs7776725 (23, 57), were extracted from Illumina Exome chip genotyping (56). Quality control procedures were described previously (56). Genotypes were in Hardy-Weinberg equilibrium (P > .05 using exact test). Genetic ancestry was determined using principal component analysis (56). Genotypes were coded assuming an additive genetic model (0 for reference allele homozygotes, 1 for heterozygotes, and 2 for alternate allele homozygotes). The associations between WNT16 rs2908004 and rs2707466 cSNP variants and incident cardiovascular disease events were tested using Cox proportional hazards models adjusted for age, sex, ancestry, systolic blood pressure, and body mass index. Analyses were performed using R statistical software version 3.6.3.

Statistical Analyses for Preclinical Studies

All statistical analyses were performed with 3 to 6 independent replicates per group as indicated. Statistical analyses were performed using GraphPad Prism 8.0 software, implementing 2-tailed standard parametric methods after Shapiro-Wilk normality testing, or with nonparametric methods when indicated. Graphic data are presented as mean ± standard error of the mean (SEM). When one-way ANOVA was applied, Holm-Sidak post hoc testing was implemented to test for significance between groups and correct for multiple comparisons. Functional analyses identifying aortic LIM (Lin11, Isl1, and MEC3 tandem zinc finger) domain proteins regulated by Wnt16 deficiency in proteomic data sets were carried out using DAVID 6.8 (Database for Annotation, Visualization, and Integrated Discovery) (58), deploying the Benjamini-Hochberg correction (59) of F-tests to control false discovery rate.

Results

Wnt16 Is Expressed in Arterial VSM, and WNT16 Genotypes in Dallas Heart Study Convey Cardiovascular Disease Risk

In prior studies of smooth muscle–specific Lrp6 knockout mice, we noted significant accumulation of Wnt16 mRNA in aorta (20), consistent with observations of Nurminskaya (60). Systematic survey of tissue RNA from C57BL/6 mice revealed Wnt16 mRNA accumulated to highest levels in pinna (skin and elastic cartilage), aorta, and bone (femur), with lower levels in kidney, atria, liver, and skeletal muscle (Fig. 1A). Moreover, in response to AngII infusion in male LDLR−/− mice—a model of atherosclerosis and aneurysmal remodeling (26, 61)—Wnt16 was significantly upregulated 2-fold (Holm-Sidak adjusted P = .04), along with strong uptrends in Wnt4 and Wnt7b mRNA abundance (Fig. 1B). Immunofluorescence staining confirmed that Wnt16 protein was also expressed in arterial smooth muscle of ascending aorta and coronary artery (Fig. 1C and 1D; RRID: AB 11034126) and in descending thoracic aorta (Fig. 1E, left) and renal artery (not shown). In plaques of LDLR−/− mice fed Western high fat diet (HFD), faint Wnt16 protein signal was also observed in the atheroma cap (Fig. 1D, right panel). No significant Wnt16 immunofluorescence staining above background was observed in Wnt16−/− mice, confirming the specificity of the antibody (Fig. 1E, right panel). Immunofluorescence and RT-qPCR confirmed WNT16 protein and mRNA expression in human arterial smooth muscle as well (Supplementary Fig. S1 (33)).

Figure 1.

Wnt16 is expressed in murine arterial smooth muscle. (A) Relative Wnt16 mRNA accumulation in murine tissues. Note prominent Wnt16 expression in aorta, bone (femur), and skin/pinna. (B) Wnt16 mRNA accumulation in aorta was upregulated by AngII infusion. (C) Wnt16 protein (red) accumulation localized in VSM of the tunica media in coronary arteries (asterisks) and adjacent ascending aorta (arrows) by confocal immunofluorescence (bar = 100 microns). (D) No signal was observed in the absence of Wnt16 primary antibody (left panel). Right panel, some Wnt16 signal was noted over fibrous caps of atheroma in LDLR−/− mice fed Western HFD (arrowheads; bar = 100 microns). (E) Medial VSM Wnt16 immunoreactivity (red) was absent in Wnt16 knockout mice. Arrows, thoracic descending (left) and ascending (right) aorta. Asterisk, coronary arteries. Bar = 250 microns.

To ascertain whether WNT16 contributes to cardiovascular disease in humans, we examined genotype- phenotype relationships in the Dallas Heart Study I (56). WNT16 nonsynonymous cSNPs (rs2908004, rs2707466) encoding WNT16(G72R) and WNT16(T253I), 2 alleles associated with reduced bone mineral density and fracture in adults (23) but not children (62), significantly increased risk for composite cardiovascular disease events in human subjects (stroke, MI, heart failure, cardiovascular death; Table 2). These WNT16 cSNPs are in strong linkage disequilibrium (R² > 0.9) (57, 63), and both alleles conveyed hazard ratio for compositive cardiovascular disease events of ∼1.25 (adjusted P-values = .021 and .029 respectively) with adjustment for age, sex, ancestry, systolic blood pressure, and body mass index (Table 2). Of note, no association was identified between the intronic SNP FAMC3_rs7776725 (64)—present in the FAMC3 gene immediately adjacent to the WNT16 on chromosome 7—and composite cardiovascular events (hazard ratio 0.99, adjusted P = .9). Thus, Wnt16 is expressed in mouse and human arterial VSM, is dynamically regulated in response to AngII-induced vascular remodeling in mice, and nonsynonymous cSNPs in WNT16 correlate with cardiovascular disease in humans.

Table 2.

WNT16 genotype–composite cardiovascular event phenotype relationship in the Dallas Heart Study

SNP	Event	Amino acid change if nscSNP	N events	N total	N carriers	N events among carriers	Hazard ratiob (95% CI)	P value
WNT16_rs2908004	Composite Cardiovasculara	WNT16 G72R	310	2940	1572	144	1.25 (1.03-1.51)	.021
WNT16_rs2707466	Composite Cardiovasculara	WNT16 T2 53I	310	2941	1623	150	1.23 (1.02-1.48)	.029
FAMC3_rs7776725	Composite Cardiovasculara	None (intronic)	310	2941	1453	160	0.99 (0.83-1.18)	.9

WNT16 rs2908004 and rs707466 increase the risk for low bone mass and fracture and exhibit strong linkage disequilibrium (R² > 0.95). See text for details.

Abbreviations: nscSNP, nonsynonymous coding single nucleotide polymorphism; SNP, single nucleotide polymorphism.

^a Stroke, myocardial infarction, other atherosclerotic cardiovascular disease, heart failure, cardiovascular death.

^b Adjusted for age, sex, ancestry, body mass index, and systolic blood pressure.

Wnt16 Deficiency Reduces Aortic Pulse Wave Velocity and Systolic Blood Pressure, Causes Abnormal Aortic Mitochondrial Morphology, and Increases Aneurysmal Dilatation in Ascending Aorta Following AngII Infusion

To assess if Wnt16 deficiency contributes to cardiovascular phenotype in mice, we bred Wnt16−/− mice (see “Materials and Methods”) onto the LDLR−/− background (65). The LDLR−/− mouse is a well-validated background for studying multiple cardiovascular disease phenotypes, including diet-induced atherosclerosis, arterial calcification, and aneurysm formation (66, 67)—the latter elicited by AngII infusion (45). Initial phenotyping first identified that systolic blood pressure was reduced significantly in Wnt16−/−;LDLR−/− vs Wnt16+/+;LDLR−/− control mice (hence LDLR−/−; Fig. 2A), along with strong trend (P = .07) for lower aortic pulse wave velocity (PWV) (Fig. 2B)—an index of arterial stiffening (68, 69). Serum aldosterone concentrations were not reduced but significantly increased, consistent with reflex activation of the renin-angiotensin-aldosterone axis with reduced blood pressure (Fig. 2C). Transmission electron microscopy of ascending aorta tissue revealed abnormal aortic VSM morphology in these Wnt16−/− mice, most prominent in the inner tunica media (Fig. 2D). Intriguingly, higher power magnification identified mitochondrial morphology to be abnormal in Wnt16−/− aortic VSM (Fig. 2E). As compared to mitochondria of the LDLR−/− control mice that possess distinct bar-like cristae (red arrows), the mitochondria of aortic VSM of Wnt16−/−;LDLR−/− mice (yellow arrows) appeared swollen—exhibiting a more homogeneous, electron-dense mitochondrial matrix with indistinct cristae (Fig. 2E). We next challenged LDLR−/− and Wnt16−/−;LDLR−/− with Western high fat diet (HFD) to induce atherosclerotic calcification (20). No significant differences were observed in the net accumulation of aortic calcium between the 2 genotypes after 3 months of HFD. However, after 10 months of HFD challenge, there was a nonsignificant trend (P = 0.07, Holm-Sidak corrected for multiple comparisons) for a 40% reduction in aortic calcium accrual in Wnt16−/−;LDLR−/− cohorts (Supplementary Fig. S2) (15), consistent with the impact of lower systolic blood pressure on reducing cardiovascular calcification risk (70). Serum cholesterol concentration did not differ by genotype (Supplementary Fig. S2 (33)).

Figure 2.

Wnt16-deficient mice exhibit reduced systolic blood pressure in the setting of elevated serum aldosterone, with abnormal aortic morphology by electron microscopy. (A) Wnt16 deficiency reduced systolic blood pressure (SBP). (B) Aortic PWV, an index of arterial stiffness, was reduced in a strong trend (P = .07) in Wnt16-deficient mice, consistent with reduced SBP. (C) Serum aldosterone concentrations were increased with Wnt16 deficiency in the setting of lower SBP. (D) Representative electron micrographs reveal a disorganized morphology in the tunica media of ascending aorta in Wnt16−/−;LDLR−/− vs LDLR−/− control mice. Scalar bar, 10 microns. (E) Representative electron micrographs demonstrated abnormal VSM mitochondrial morphology in ascending aortas of Wnt16−/− mice. As compared to mitochondria of LDLR−/− control mice with distinct bar-like cristae (red arrows), mitochondria in aortic VSM of Wnt16−/−;LDLR−/− mice (yellow arrows) appear swollen and exhibit a more homogeneous electron-dense mitochondrial matrix with indistinct cristae. Scalar bar, 1 micron.

We then assessed the impact of Wnt16 deficiency on AngII-dependent aneurysmal remodeling (26). AngII infusion, in combination with Western HFD feeding, induced aortic aneurysmal dilatation in both thoracic and abdominal segments (45) as expected (Supplementary Fig. S3A (33)) with elastin fragmentation, ulceration and dissection (Supplementary Fig. S3B and S3C (33)). The extent of HFD-induced hypercholesterolemia did not differ with Wnt16 deficiency or AngII infusion (Supplementary Fig. S3D (33)). Ex vivo image analysis and quantification (45) demonstrated that the diameter of the thoracic ascending aorta (TAA, or Zone 0) (71), was increased 60% (58) with HFD and AngII infusion (hence AngII + HFD) in LDLR−/− mice (Fig. 3A). Remarkably, TAA aneurysmal diameter was increased by an additional 20% (17) in Wnt16−/−;LDLR−/− cohorts (Fig. 3A). By contrast, no differences existed in TAA diameter between genotypes at baseline without AngII + HFD challenge (Fig. 3A) or after 10 months of HFD challenge alone (data not shown). Of note, with AngII challenge this percentage increase and absolute magnitude of significant change in ascending aorta diameter has been observed previously in other published models (72, 73), including aneurysmal disease of the aortic root (72, 73). With AngII + HFD challenge, strong trends for larger aortic diameters in Wnt16−/− mice were also observed in the aortic arch between innominate and left carotid arteries (Zone 1; P = .07) and the widest abdominal aorta segment (Supplementary Fig. S4 (33)). Pulse wave velocity (PWV) was once again lower in Wnt16−/−;LDLR−/− vs LDLR−/− mice even following challenge with AngII + HFD (Fig. 3B). Histological quantitation of thoracic aorta revealed no Wnt16-dependent increases in elastin fragmentation (Supplementary Fig. S5A; (33) quantitatively diminished) or reduced wall thickness (Supplementary Fig. S5B (33)) to explain the reductions in aortic PWV (Fig. 3B) in Wnt16−/− cohorts. In our gene expression, aortic digital imaging, histology, immunofluorescence, and proteomic studies, a total of 53 animals were studied. The overall mortality observed due to aortic rupture (23%) was equivalent in Wnt16−/−;LDLR−/− mice (n = 7 of 27) and Wnt16+/+;LDLR−/− control mice (n = 6 of 26) infused with AngII. Thus, aortic Wnt16 expression, induced in response to AngII infusion, serves to limit aneurysmal dilatation in the ascending thoracic aorta in a setting of reduced baseline blood pressures and vascular stiffness.

Figure 3.

Wnt16 deficiency increases aneurysmal dilatation of thoracic ascending aorta (TAA) in response to AngII infusion. (A) Aneurysmal dilatation of ascending aorta in responses to AngII infusion was increased in Wnt16−/− mice. Note that Wnt16 deficiency did not impact baseline TAA diameter. (B) The increased in aortic stiffness induced by AngII, assessed by aortic PWV, was reduced in Wnt16−/−;LDLR−/− vs LDLR−/− control mice. Note that similar reductions in aortic PWV were observed with Wnt16 deficiency in the absence of AngII (Fig. 2B).

Wnt16 Deficiency Reduces Aortic VSM Contractile Protein Accumulation and Gene Expression, Decreases Contractile Function, and Reduces Mitochondrial Respiratory Capacity

TAA aneurysmal dilation in response to AngII infusion was greater in Wnt16-null mice in the setting of reduced basal systolic blood pressure. This seemingly paradoxical relationship has been reported previously for genetic lesions impairing Acta2, Myh11 (vascular smooth muscle myosin heavy chain 11), and Tgfβ function—genes sustaining the VSM contractile phenotype as necessary to mitigate TAA aneurysm risk (74, 75). Milewicz and colleagues demonstrated that Acta2−/− mice also exhibit significantly lower baseline systolic blood pressures (their Fig. 6F in reference (29)) and increased thoracic aortic aneurysmal remodeling following AngII infusion (29, 74, 76) as we observe in Wnt16−/−;LDLR−/− mice (Fig. 2). Thus, we hypothesized that the reduced systolic blood pressure, reduced PWV, and increased aneurysmal remodeling of Wnt16-null mice arises due to altered VSM contractile differentiation and function. Therefore, we examined the impact of Wnt16 deficiency on Acta2 expression and other genes characteristic of the VSM contractile phenotype. As shown in Fig. 4, Acta2 gene expression was significantly decreased in thoracic aorta VSM isolated from Wnt16−/−;LDLR−/− mice as compared with the LDLR−/− controls. The VSM contractile genes Myh11(77) and Myocardin (Myocd) (78) were also concomitantly reduced. By contrast, Sca1 (stem cell antigen 1, also known as Ly6a), a marker of the de-differentiated mural VSM phenotype (79, 80) and expanded in response to atherosclerotic injury was increased concomitantly (Fig. 4). C3 and C4b, complement genes coregulated with Sca1 with VSM de-differentiation (79) and implicated in TAA aneurysm (81, 82), were also upregulated with Wnt16 deficiency. Wnt16 was reduced as expected in Wnt16−/− mice, and Tgfbr1 and Tgfbr2 were not significantly affected (Fig. 4). Whole cell extract proteomic analyses (Fig. 5A) and Western blot analyses (Fig. 5B and 5C) confirmed that contractile Acta2 (RRID: AB 2223009) and SM22 (RRID: AB 443021) protein levels were reduced significantly in Wnt16−/− VSM cultures (Fig. 5). Conversely, ELISA confirmed concomitant increases in C3 complement protein accumulation in conditioned media from Wnt16-null aortic VSM cultures (Fig. 5D)—consistent with upregulation of C3 mRNA observed with Wnt16 deficiency (Fig. 4).

Figure 6.

Wnt16 deficiency reduces basal and maximal respiration and spare respiratory capacity of aortic VSM as assessed by Seahorse oximetry. (A) Basal and maximal respiration (oxygen consumption rate; OCR) were reduced in Wnt16−/−;LDLR−/− VSM as compared to LDLR−/− control VSM. Glycolysis, evident in extracellular acidification rate (ECAR), was also reduced with Wnt16 deficiency (not shown). Data in panels A and B are presented as the average of 3 independent experiments, with N = 3 per genotype per experiment, with a total of N = 9 independent replicates studied per genotype. (B) Spare respiratory capacity, an age-related index of cellular resilience, was reduced with Wnt16 deficiency. N = 9 per genotype, with the average of 3 independent experiments of 3 per genotype presented. (C) A representative Mito Stress Test OCR profile from one experiment with N = 3 individual replicates per genotype, and 3 measurements per time point (mean +/− SEM at each timepoint).

Figure 4.

Wnt16 deficiency reduces expression of the VSM contractile phenotype, with concomitant upregulation of the de-differentiation markers Sca1, C3, and C4b. Primary VSM cultures exhibited lower levels of Acta2, Myh11, and Myocd mRNAs—markers of contractile VSM. Sca1, complement C3 and C4b—characteristic of dedifferentiated VSM—were increased. Tgfbr1 and Tgfbr2 mRNAs were not significantly impacted. Wnt16 was significantly reduced as expected in Wnt16 knockout aortic VSM.

Figure 5.

Wnt16 deficiency reduces aortic VSM contractile protein accumulation, increases complement C3 protein production, and reduces TGF-beta induced contraction. (A) Label-free quantitation of protein accumulation by LC-MS/MS in primary aortic VSM cultures demonstrated reduced Acta2, SM22, and Myh11 protein with Wnt16 deficiency. Rps21 is shown as a housekeeping protein control. Panels B and C, Western blot analyses confirmed reductions in Acta2 and SM22 (transgelin) protein accumulation, quantified from digital image analyses, normalized to loading controls (Rps21 and vinculin). Vinculin was used as a loading control for the SM22 Western blot since its molecular weight is close to that of Rps21. (D) Complement C3 protein, measured by ELISA, was increased in the conditioned media of Wnt16-null aortic VSM, paralleling the increase in C3 mRNA (Fig. 4). (E) Wnt16 deficiency reduced aortic VSM contraction in response to TGF-beta treatment in floating gel assays as quantified by digital image analysis. *P < .05; **P < .01; ***P < .001; ****P < .0001 vs comparator following one-way ANOVA (P = 2 × 10⁻⁶) and Holm-Sidak post hoc testing as indicated.

We then assessed the contractile activity of Wnt16-deficient aortic VSM in response to TGFβ, a robust stimulus for myofibroblast and VSM contraction (83), in the floating matrix gel contraction assay (83). As shown in Fig. 5E, no significant basal differences exist in floating gels seeded with either LDLR−/− or Wnt16−/−;LDLR−/− VSM. TGFβ1-induced contraction significantly reduced gel area at 48 hours by 45% in LDLR−/− VSM (ANOVA P = 2 × 10⁻⁶, Holm-Sidak P = 3 × 10⁻⁶; Fig. 5E) However, the magnitude of TGFβ1-induced contraction was significantly reduced to 21% in Wnt16−/−;LDLR−/− VSM, with an area that was 33% larger than TGFβ-stimulated LDLR−/− control cells (Holm-Sidak P = .016; Fig. 5E).

Mitochondrial dysfunction has emerged as a contributor to worsening thoracic aortic aneurysmal remodeling, mediated in part via impaired energetics required for the contractile phenotype(30). Moreover, both mitochondrial morphology (Fig. 2E) and contractile phenotype (Fig. 5) were abnormal in aortic VSM from Wnt16−/− mice. Thus, we hypothesized that Wnt16 deficiency might also impair VSM mitochondrial function. Therefore, we measured the mitochondria respiration of Wnt16−/− VSM using Seahorse oximetry (84). As compared to LDLR−/− VSM, the Wnt16−/−;LDLR−/− aortic VSM exhibited significantly reduced basal and maximal oxygen consumption in the mitochondrial stress test (Fig. 6A). Moreover, VSM spare respiratory capacity, a key index of mitochondrial dysfunction with aging (84, 85) and phenotypic modulation with arterial remodeling (86), was also reduced with Wnt16 deficiency (Fig. 6B and 6C). Thus, Wnt16 deficiency reduces aortic VSM contractile protein accumulation and gene expression, decreases contractile function, and reduces mitochondrial respiratory capacity.

Wnt16 Incubation Increases Elaboration of the Aortic VSM Contractile Phenotype and Mitochondrial Respiratory Capacity

To further confirm the regulation of the VSM contractile phenotype by Wnt16, we examined the in vitro and in vivo responses to incubation with exogenous recombinant Wnt16 protein. In vitro, Wnt16 increased contractile markers including Acta2, SM22, Cnn1 (calponin1), and Smoothelin (Smtn), and upregulated Elastin (Eln) accumulation but suppressed Sca1, C3, and C4b mRNA accumulation (Fig. 7A). These responses are reciprocal to genetic Wnt16 deficiency (Fig. 4). Wnt16 incubation did not impact Axin2 expression (Fig. 7A), suggesting that genomic regulation was not via the canonical Wnt cascade (87). Moreover, Wnt16 incubation directly stimulated thoracic aortic VSM contraction in floating gel contraction assays to a level equivalent to that of TGFβ1 (Fig. 7B), confirming the importance of Wnt16 signaling on contraction function. Furthermore, Wnt16 incubation increased maximal respiration and spare respiratory capacity (54) in aortic VSM (Fig. 7C) as predicted from our studies of Wnt16-null VSM.

Figure 7.

Wnt16 treatment increases contractile genes, suppresses markers of de-differentiation, induces contraction, and increases respiration in primary aortic VSM. (A) Relative mRNA accumulation was quantified by real-time quantitative RT-PCR and normalized to 18S. Note that Wnt16 dose-dependently increased accumulation of Acta2, SM22, and other contractile VSM phenotypic markers, while decreasing Sca1, C3, and C4b accumulation. Axin2 was not induced, indicating the absence of canonical Wnt signal activation. (B) Like TGFbeta1, Wnt16 (8 nM) incubation significantly induced VSM contraction in floating gel assays. Tgb1 and Wnt16 actions were not additive. (C) Mitochondrial spare respiratory capacity, an age-related index of cellular resilience, was increased by Wnt16 treatment of aortic VSM, achieved by enhanced maximal respiration. All experiments were repeated at least twice. *P < .05; **P < .01; ***P  < .001; ****P < .0001 vs no treatment following one-way ANOVA and Holm-Sidak post hoc testing.

We previously demonstrated that Wnt3a upregulated SM22 expression part via canonical Wnt/beta-catenin signals (88). Thus, we thus compared gene expression responses elicited by Wnt16 vs Wnt3a and Wnt5a, prototypic agonists of canonical Wnt/beta-catenin and noncanonical Wnt/Ca++ pathways, respectively (9, 18). Wnt16 genomic responses are not fully recapitulated by either Wnt3a or Wnt5a incubation (Fig. 8A; 400 ng/mL or 10 nM for all incubations). For example, while Wnt3a suppressed C3 and augmented contractile gene expression including Acta2 and SM22 (88), neither Wnt3a nor Wnt5a upregulated Elastin (Fig. 8A). Moreover, while Wnt3a upregulated the senescence marker Cdkn2a/p16 (89) and prototypic Wnt modulators Sfrp1 (90) and Sfrp2 (91), Wnt16 downregulated all 3 of these genes (Fig. 8A). The senescence associated secretory phenotype (SASP) (92, 93) encompasses (a) the Tgf-beta family member Gdf15—an aging-related mitokine that demarcates the mitochondrial integrated stress response (93, 94)—along with Bmp2 and Tgfb1; and (b) multifunctional cytokines, including IL6 and Vegfc, that have proved to be important in adaptive tissue responses in bone (95) and aorta (96). As shown in Fig. 8B, incubation of VSM with Wnt16—but not Wnt3a or Wnt5a—suppressed the senescence and mitochondrial stress biomarker Gdf15, while increasing Tgfb1, IL6, and Vefc gene expression. By contrast, Wnt3a uniquely upregulated Bmp2. Since monocytes are major contributors to the SASP with tissue senescence (92), we went on to study the impact of Wnt16 on primary bone marrow monocytes. As shown in Fig. 8C, in monocytes, Wnt16 incubation downregulated expression of the canonical senescence marker Cdkn1a/p21 (92), and again suppressed Gdf15. Interestingly, Mif—a pivotal SASP cytokine (92)—was significantly suppressed by Wnt16 in monocytes, but with little response in VSM. Although IL1b was upregulated, Mmp9 (matrix metalloproteinase 9) and Cxcl10—IL1b genomic targets in the SASP and vascular inflammation (93, 97)—were suppressed by Wnt16 incubation. Thus, Wnt16 incubation regulates the contractile aortic VSM phenotype as predicted from our studies of Wnt16−/− mice, in a manner overlapping yet distinct from the prototypic canonical (Wnt3a) and noncanonical (Wnt5a) Wnt agonists (9, 18). Moreover, Wnt16 actions can shape key components of the SASP in both VSM and monocytes.

Figure 8.

Wnt ligands differentially regulate gene expression, including contractile, de-differentiation, and senescence markers. (A) Wnt16 treatment upregulated expression of the contractile phenotype. Like Wnt16, Wnt3a also upregulated contractile gene expression and suppressed C3. However, Wnt3a did not reproduce the Elastin mRNA induction observed with Wnt16 treatment. Furthermore, while Wnt16 suppressed Cdkn2a mRNA, Wnt3a increased expression. Differential regulation was also observed for Sfrp1 and Sfrp2. However, the angiotensin receptor message Agtr1a was suppressed by both Wnt16 and Wnt3a, indicating overlapping yet distinct actions. Holm-Sidak post hoc testing was used to correct for multiple comparisons following significance by one-way ANOVA. (B) In primary VSM, Wnt16 differentially regulated key members of the SASP transcriptome, suppressing Gdf15 but increasing expression of Tgfb1 with little impact on Bmp2 or Mif. Wnt3a preferentially induced Bmp2 expression, while IL6 and Vegfc were upregulated by Wnt16. (C) In monocytes, Wnt16 downregulated the canonical senescence gene Cdkn1a encoding p21, and suppressed both Gdf15 and Mif. Although IL1b was upregulated, Mmp9 and Cxcl10 –IL1b genomic targets in the SASP—were suppressed by Wnt16 treatment.

Wnt16 Deficiency Reduces Aortic Accumulation of the Yap/Taz Target Ankrd1 and Reduces Nuclear Accumulation of Yap and Taz (Wwtr1)

To better understand the mechanisms whereby Wnt16 deficiency might impact aneurysmal remodeling, we performed proteomic analysis of thoracic aortic proteins (TAA and arch) extracted from LDLR−/− and Wnt16−/−;LDLR−/− mice subjected to AngII-mediated aneurysmal remodeling. As before (44), we implemented LC-MS/MS with label-free quantitation of tryptic peptide fragments generated from thoracic aortic protein extracts to identify and characterize protein changes (n = 3 / genotype). Of the thoracic aorta proteins significantly altered with Wnt16 deficiency, 3 salient features emerged. First, Ankrd1 (ankyrin repeat domain 1), a protein necessary for myofibroblast contractile responses (98) and prototypic target of Yap/Taz (Wwtr1) transcription (99), was downregulated significantly with Wnt16 deficiency (Fig. 9A, Supplementary Table ST2 (33)). By contrast, the monocyte/macrophage metalloproteinase Mmp12 was upregulated significantly with Wnt16 deficiency (Fig. 9A), along with multiple members of the Lim domain protein family that characterize actin-dependent mechanosensation and focal adhesion regulation (100) (Fig. 9A, and Supplementary Fig. S6 (33)). Knockdown of Wnt16 in primary aortic VSM confirmed downregulation of Ankrd1 along with Acta2 and Myh11 with concomitant upregulation of Sca1 (Fig. 9B); Wnt16 siRNA acutely recapitulated the actions of genetic Wnt16 deficiency we observed above (Fig. 4). Conversely, Wnt16 incubation dose-dependently increased Ankrd1 and Ctgf (connective tissue growth factor, also known as CCN2, a Yap1/Taz target) mRNA (Fig. 9C), the latter also a direct target of Yap/Taz signaling (99) like Anrkd1. Wnt3a and Wnt5a incubation could not recapitulate the actions of Wnt16 on Anrkd1 and Ctgf expression (Fig. 9C), confirming the distinct features of Wnt16 actions in arterial VSM. Moreover, in vivo intraperitoneal dosing with recombinant Wnt16b upregulated aortic accumulation of Ankrd1 and VSM contractile programs including Acta2, Myh11, Cnn1, and Smoothelin (Smtn) mRNAs, without increasing Axin2 (canonical Wnt target) or alkaline phosphatase (TNAP) expression (Fig. 9D).

Figure 9.

Wnt16 supports VSM Ankrd1 expression, a prototypic contractile target of Yap1/Taz activation. (A) Label-free quantitation of peptide fragments by LC-MS/MS was used to characterize proteins in aortic extracts from LDLR and Wnt16−/−;LDLR−/− mice following AngII + HFD challenge. N = 3 individual mice per genotype. Aortic Ankrd1 protein was profoundly reduced in aortic extracts of Wnt16-null mice. Mmp12 accumulation was concomitantly upregulated, along with multiple members of the Lim domain protein family (right panel). The 40S ribosomal subunit Rsp21 is shown as a housekeeping control protein. (B) Like genetic Wnt16 deficiency, siRNA targeting Wnt16 in aortic VSM cultures acutely downregulated Ankrd1, Acta2, and other contractile phenotypic marker mRNAs, while increasing Sca1 accumulation. (C) Wnt16 incubation of cultured primary aortic VSM dose-dependently increased Anrkd1 and Ctgf gene expression. Note that this Wnt16 response is not recapitulated by Wnt3a or Wnt5a incubation. Holm-Sidak post hoc testing was used to correct for multiple comparisons following significance by one-way ANOVA. (D) Intraperitoneal in vivo dosing with recombinant human Wnt16 (4 ug daily for 2 days, n = 3 animals per treatment) increased aortic expression of Ankrd1 and VSM contractile genes (Acta2, Myh11, Cnn1, Smtn). Note that canonical Wnt target genes Axin2 and TNAP were not induced by Wnt16 dosing.

As a prototypic Yap/Taz target (99, 101-103), Ankrd1 also plays an important role in the contractile phenotype of wound myofibroblasts (98). Therefore, due to the strong genetic connection between the contractile VSM phenotype and ascending aortic aneurysm risk (74)—and the impact of Wnt16 deficiency and treatment on contractile phenotype we observed above (Fig. 3)—we focused on VSM Yap-Taz signaling. Yap1 and Taz (Wwtr1) are transcriptional co-adapters that shuttle between the cytoplasm and nucleus, inhibited by Lats1/Lats2-mediated phosphorylation that limits nuclear accumulation via complexes with cytoplasmic 14-3-3 proteins (31, 101, 104). Since (a) Wnt16 deficiency reduced aortic VSM Ankrd1; and (b) Wnt16 treatment upregulated both Ankrd1 and Ctgf expression—both targets of Yap/Taz signaling (99, 105)—we hypothesized that Yap-Taz nuclear accumulation (31) would be reduced in Wnt16-deficient VSM. To directly assess this, we prepared cytoplasmic and nuclear protein extracts from LDLR−/− and Wnt16−/−;LDLR−/− VSM, performing Western blot for Yap and Taz (Wwtr1) (RRID: AB 10950494) to assess their cytoplasmic-nuclear distribution. As shown in Fig. 10 the nuclear accumulation of Yap and Taz (Wwtr1) was decreased significantly, while cytoplasmic accumulation was concomitantly and significantly increased, with Wnt16 deficiency. Thus, Wnt16 deficiency perturbed the VSM cytoplasmic-nuclear distribution of Yap and Taz, reducing nuclear accumulation as necessary for transcriptional activation.

Figure 10.

Wnt16 deficiency reduces nuclear and increases cytoplasmic accumulation of Yap1 and Taz (Wwtr1) in cultured primary aortic VSM. Cytosolic and nuclear fractions were prepared from cultured aortic VSM cells from LDLR−/− and Wnt16−/−;LDLR−/− mice, then analyzed by Western blot for Yap and Taz (Wwtr1) protein accumulation. The antibody deployed recognizes both proteins (see validation in Fig. 11). Left panel, Western blot analyses. Right panel, quantitation of Yap1 and Taz by digital image analysis with Fiji/Image J, with normalization to GAPDH (cytosol) and Histone H3 (nucleus).

Knockdown of Taz (Wwtr1) Phenocopies the Impact of wnt16 Deficiency on the VSM Contractile Phenotype, and Inhibits Wnt16-Stimulated Acta2 and Ankrd1 Expression

The canonical Wnt target Axin2 was not regulated in aorta by Wnt16 treatment (Fig. 9). Since the Yap/Taz pathway has emerged as a unique type of alternative, noncanonical Wnt signaling (32), we studied the contributions of Yap/Taz to VSM Wnt16 actions. Thus, we examined the impact of verteporfin—an inhibitor of Yap/Taz signaling (52)—and specific Yap1 vs Taz siRNAs on Wnt16-regulated VSM contractile function and phenotype. Wnt16-induced arterial VSM contraction in the floating gel assay was reversed by treatment with verteporfin in a dose-dependent manner (Fig. 11A). Furthermore, siRNA targeting Taz (Wwtr1) preferentially downregulated Ankrd1 and the contractile phenotype (Acta2, Cnn1) while upregulating Sca1 and complement C3 expression (Fig. 11B)—phenotyping actions of genetic Wnt16 deficiency (Fig. 4) and Wnt16 siRNA (Fig. 9B). Yap1 siRNA did not recapitulate Wnt16 deficiency or Taz siRNA effects. Intriguingly, Yap1 siRNA preferentially upregulated the elaboration of Ccl2 (CC chemokine ligand 2), while Taz siRNA did not (Fig. 11B), again demonstrating the unique genomic biology of Taz and Yap1 (101). Combinatorial addition of Yap1 siRNA to Taz siRNA did not further reduce the VSM contractile phenotype (Fig. 11C). Moreover, siRNA targeting Taz alone is sufficient to inhibit Wnt16-dependent upregulation of Ankrd1, Acta2, and Cnn1 and mitigate Wnt16 suppression of C3 (Fig. 11D). Western blot analyses confirmed efficient knockdown of Yap1 and Taz proteins following RNAi in primary VSM (Fig. 11E). We crossed Taz(fl/fl);Yap1(fl/fl)(37) and Myh11-Cre (36) transgenic mice to generate conditionally delete both Taz alleles and 1 Yap1 allele in the VSM lineage to study the impact of deficiency on mitochondrial respiration. Primary aortic VSM from Myh11-Cre:Taz(fl/fl);Yap1(fl/+) mice exhibited reduced maximal respiration and spare mitochondrial reserve capacity vs Taz(fl/fl);Yap1(fl/+) controls (Fig. 11F), once again phenocopying the impact of Wnt16 deficiency on aortic VSM respiration (Fig. 6 above).

Figure 11.

Inhibition of VSM Taz signaling reduces Wnt16-induced contraction, inhibits Ankrd1 and contractile gene expression, and reduces respiration. (A) Wnt16-induced VSM contraction is reversed by treatment with the Yap/Taz inhibitor verteporfin. *P < .05; **P < .01; ***P  < .001; ****P < .0001 vs no treatment following one-way ANOVA (P = 1.5 × 10 ⁻⁷) and Holm-Sidak post hoc testing. (B) Note that siRNA targeting Taz reduced Ankrd1, Acta2, and Cnn1 while increasing Sca1 and complement C3 accumulation, mimicking key features of Wnt16 genetic deficiency and knockdown. Yap1 siRNA did not, but selectively increased Ccl2 expression. (C) Addition of Yap1 siRNA to Taz siRNA did not enhance inhibition of Ankrd1, Acta2, or Myh11. (D) RNAi targeting Taz profoundly inhibited Wnt16 induction of Ankrd1 and Acta2 expression in VSM, and antagonized Wnt16 suppression of C3. (E) Western blot analysis confirmed Taz (Wwtr1) and Yap1 protein depletion by RNAi in transfected primary aortic VSM. The siRNA targeting the LIM domain protein Fhl1 was included as an additional negative control. The same antibody recognizing both Yap1 and Taz was as deployed in Fig. 10. (F) Conditional deletion of both Taz alleles and 1 Yap1 allele in Myh11-Cre:Taz(fl/fl):Yap(fl/+) VSM reduced maximal respiration (left panel) and mitochondrial spare respiratory capacity (right panel), phenocopying the impact of Wnt16 deficiency (Fig. 6).

Yap1 and Taz (Wwtr1) are components of the Hippo pathway, negatively regulated by Mst1/2 (also known as Stk4/Stk3, Hippo orthologs) and Lats1/Lats2 kinase activities that control Yap/Taz cytoplasmic-nuclear shuttling(31). While inputs other than Mst1/2 can activate Lats1/Lats2, Lats1 and Lats 2 are the direct Yap-Taz kinases that inhibit Yap-Taz function by enabling cytoplasmic sequestration via 14-3-3 proteins(31). RNAi targeting Lats1 + Lats2 upregulated Ankrd1, Ctgf, Acta2, Cnn1, and Elastin while reducing Sca1 in aortic VSM (Fig. 12A), phenocopying the impact of Wnt16 activation (Fig. 4) and consistent with the known Lats1/2 inhibition of Taz signaling (31). By contrast, siRNA targeting the Hippo kinases Mst1 + Mst2 had little effect on the contractile phenotype or Ankrd1 expression (Fig. 12B). Thus, another regulatory relay must control Lats1/2 activity as relevant to the contractile phenotype dependent upon downstream of Taz function in VSM.

Figure 12.

RNAi targeting Lats1/2, but not Mst1/2, upregulates expression of the VSM contractile phenotype. (A) RNAi targeting Lats1 + Lats2 induced Ankrd1, Ctgf, Acta2, and other VSM phenotype markers, phenocopying effects of Wnt16 treatment. RNAi targeting Lats1/2, but not Mst1/2, upregulated expression of the VSM contractile phenotype. (B) Unlike RNAi targeting Lats1/2, RNAi targeting the Hippo kinases Mst1 (Stk4) +Mst2 (Stk3) did not downregulate the VSM contractile phenotype and does not phenocopy Wnt16 treatment.

Cavernous malformation (Ccm) proteins are components of a second upstream regulatory arm in Lats1/2 activation independent of Mst1/2 (106), and recently shown to inhibit Yap-Taz signaling in fibroblasts by restraining force transmission via focal adhesions (106). In Western blot analyses (Fig. 13A) Ccm2 protein (RRID: AB 2880454) was upregulated in primary aortic VSM cultures from Wnt16−/−;LDLR−/− vs LDLR−/− control mice, confirming preliminary insights gleaned from proteomic inquiry (data not shown). Moreover, as shown in Fig. 13B, siRNA targeting Ccm2 increased elaboration of VSM contractile gene expression phenocopying the effect of Lats1/2 siRNA (Fig. 12A) and Wnt16 incubation (Figs. 7A, 8, and 9C). Efficient Ccm2 protein depletion by siRNA was again confirmed by Western blot (Fig. 13C). Thus, Taz (Wwtr1) is a component of an alternative (32) noncanonical Wnt16 signaling relay that preserves the contractile phenotype of VSM, regulated in part via inhibitory Ccm2–Lats1/2 relays.

Figure 13.

Expression of Ccm2, an upstream regulator of Lats1/2, is increased with Wnt16 deficiency, and siRNA targeting Ccm2 phenocopies key features of Wnt16 treatment. (A) Western blot analysis confirmed upregulation of the Ccm2 platform protein in Wnt16 deficient VSM. (B) RNAi targeting Ccm2 phenocopies key features of Wnt16 treatment. Note that siRNA targeting Ccm2, but not Ccm3, upregulated the VSM contractile phenotype (Acta2, Myh11, Cnn1, SM22) and Elastin expression while suppressing C3. Ccm2 siRNA did not alter Ccm1 or Ccm3 expression. (C) Western blot analysis validated efficient siRNA-directed Ccm2 protein depletion in primary VSM. *P < .05; **P < .01; ***P < .001; ****P < .0001 vs comparator following one-way ANOVA and Holm-Sidak post hoc testing.

Taz(Wwtr1) and Wnt16 Reciprocally Regulate the Acta2 and C3 Promoters, Dependent Upon Intact TEAD Binding Cognates

The 4 kb Acta2 promoter fragment −1.1 kb to +2.7 kb—encompassing the proximal promoter, first exon, and first intron—drives sustained expression in arterial smooth muscle (49, 107) as well as transient expression in early skeletal osteoprogenitors (107). Taz functions as a transcriptional co-adapter for TEAD DNA binding proteins (101, 104) that recognize the cognate 5′-A/TGGAATG-3′ (5′-CATTCCA/T-3′ on the complementary strand). Analysis of the 4 kb Acta2 promoter revealed 4 TEAD sites predicted to convey response to Taz (Fig. 14A), 2 within the proximal promoter region just upstream of the transcription start site (TSS), in addition to the intronic serum response factor (SRF) CArG element important for VSM Acta2 expression (49). Taz (Wwtr1) is a coactivator of Tead1, with Tead1 required for elaboration of the VSM phenotype during development (108). Therefore, we cloned the Acta2 promoter region −1.4 kb to +2.7 kb upstream of the luciferase (LUC) reporter gene for functional analysis of transcriptional regulation by Taz (Wwtr1) and Tead1. Co-expression of Taz (Wwtr1) and Tead1 in HEK 293 cells synergistically activated transcription driven by the 4 kb Acta2 promoter (luciferase reporter; Fig. 14B) with this synergy augmented by co-expression of SRF (Fig. 14C). Dipyrimidine mutations of TEAD elements 1, 2, or 3 reduced basal Acta2 promoter activity, while mutation of TEAD element 4 (TTMUT4) did not (Fig. 14D). Mutation of any one of the 4 TEAD elements significantly reduced Taz + Tead1 activation (Fig. 14D). Moreover, ablation of the TEAD element most proximal to the TSS, that is, TT2MUT, profoundly inhibited basal and Wnt16-activated 4 kb Acta2 promoter activity (Fig. 14E). The smaller 1.8 kb Acta2 promoter fragment encompassing the 2 TEAD elements proximal to the TSS was sufficient to convey Wnt16 responsiveness (Fig. 14F). In this minimal promoter context, the transcriptional response to recombinant Wnt16 was completely lost by dinucleotide mutations (AGGAATG → ATTAATG) in either of the 2 TEAD cognates upstream of the Acta2 TSS (Fig. 14F), with mutation of the upstream element (TT1MUT) more consequential to basal activity. Furthermore, chromatin immunoprecipitation (ChIP) demonstrated that Taz (RRID: AB 2800026) associates with the promoter region of Acta2 chromatin in VSM and is significantly increased by recombinant Wnt16 (Fig. 14G). By contrast, little Taz is associated with the TNF promoter chromatin (negative control) and is not altered by Wnt16 (Fig. 14H). Furthermore, H3K4me3 histone methylation (RRID: AB 2924768), a mark of activated transcription (109), is also significantly increased over Acta2 promoter chromatin following Wnt16 treatment of primary aortic VSM (Fig. 14I).

Figure 14.

Wnt16 and Taz activate transcription driven by the Acta2 promoter. (A) The Acta2 promoter contains 4 TEAD elements, shown to convey Yap1/Taz transcription in other promoters. The GG → TT dipyrimidine substitutions that mutate these cognates are indicated. (B) Co-expression of Taz and Tead1 synergistically activated the Acta2 promoter in HEK293 T cells. (C) Taz + Tead1 synergistically augmented SRF activation of Acta2 transcription. (D) Dipyrimidine mutations of TEAD elements 1, 2 or 3 reduced basal promoter activity, while mutation of TEAD element 4 (TTMUT4) did not. Mutation of any one of the 4 TEAD elements significantly reduced Taz + Tead1 activation. (E) Dipyrimidine mutation of the TEAD element most proximal to the transcription start site (TSS) profoundly inhibited basal and Wnt16-activated 4 kb Acta2 promoter activity. (F) The shorter 1.8 kb Acta2 promoter fragment encompassing the 2 TEAD elements surrounding the TSS was sufficient to convey Wnt16 responsiveness. In this minimal promoter fragment, dipyrimidine mutation of either of these 2 TEAD cognates reduced basal activity and abrogates Wnt16 responsiveness. Panels G and H, Taz (Wwtr1)-specific ChIP assay identified that recombinant Wnt16 increased Taz association with Acta2 promoter chromatin (Panel G), but not TNF promoter chromatin (Panel H) in primary aortic VSM cultures. Note differences in y-axis scales in panels G and H, adjusted in panel H to facilitate visualization of TNF data. (I) Active H3K4me3 histone methylation was also selectively increased on the Acta2 promoter, but not the TNF promoter, following Wnt16 incubation. ChIP primers anneal to DNA adjacent to each promoter's TSS. *P < .05; **P < .01; ***P < .001; ****P < .0001 vs comparator following one-way ANOVA and Holm-Sidak post hoc testing. Abbreviation: ns, not significant.

Because Wnt16 deficiency and Taz (Wwtr1) deficiency both upregulated complement C3 expression, we also studied the mouse C3 promoter. Inspection of the proximal C3 promoter revealed a single 5′- AGGAATG −3′ TEAD cognate just downstream of the TSS (nucleotides +21 to +27, bottom strand) within the 5′-UTR (Fig. 15A). Wnt16 treatment or Taz expression suppressed 1.3 kb FC3 proximal promoter activity (Fig. 15B), consistent with changes in the C3 protein (ELISA; Fig. 5D) and mRNA we noted above (Fig. 4-6). Dinucleotide mutation (TTMUT; AGGAATG → AAAAATG) of this TEAD cognate in the 1.3 kb C3 promoter reduced basal activity and abrogated the Wnt16- and Taz -dependent inhibitory responses (Fig. 15C). Thus, Wnt16 supports the VSM contractile phenotype--including Acta2, SM22, and Myh11 protein accumulation, Acta2 gene transcription, and VSM contractile function, mediated in part via Taz signaling relays (Fig. 16, working model).

Figure 15.

Wnt16 and Taz suppress transcription driven by the C3 promoter. (A) The murine 1.3 kb promoter possesses a single TEAD element in the 5'-UTR just downstream of the transcription start site. (B) Recombinant Wnt16 suppressed C3 promoter activity (luciferase reporter) in A7r5 cells, consistent with Wnt16 actions that suppressed both C3 mRNA and protein (ELISA) in primary aortic VSM (see above). (C) Co-expression of Taz suppressed the C3 promoter. Suppression was dependent upon the intact TEAD element in the C3 5′-UTR. *P < .05; **P < .01; ***P < .001; ****P < .0001 vs comparator following one-way ANOVA and Holm-Sidak post hoc testing. Abbreviation: ns, not significant.

Figure 16.

Working model of Wnt16 actions in VSM. In aortic VSM, Wnt16 signals support the contractile phenotype via the nonconical alternative Wnt pathway, activating Taz (Wwtr1) transcriptional relays that concomitantly suppress VSM de-differentiation programs (complement, Sca1). VSM Wnt16-Taz signaling maintains and augments mitochondrial respiratory capacity and contraction as well. In toto, these actions help preserve conduit artery function and structure, including AngII-mediated aneurysmal remodeling of the thoracic ascending aorta (TAA). See text for details.

Discussion

Wnt signaling was initially discovered due to the vital roles of Wnt ligands in tissue morphogenesis, epithelial malignancies, and bone homeostasis (18). However, Wnt relays have emerged as important regulators of cardiovascular physiology as well (9, 110), best characterized in atherosclerotic calcification (20, 111) and calcific aortic valve disease (112). To date, most data have emphasized the negative impact of excessive cardiovascular Wnt signaling (110), namely, the prosclerotic programs activated by noncanonical Wnt signals when left unchecked by the moderating influence of homeostatic canonical Wnt programs (19, 20). Specific noncanonical Wnt ligands such as Wnt5a/b, Wnt11, and Wnt7b have been identified as contributing to cardiovascular calcification in response to metabolic or genetic risk (20, 112). However, the protective roles of specific Wnt ligands in preservation of cardiovascular health and function have been largely unexplored.

In this study, we identify for the first time a protective role of a specific Wnt in cardiovascular disease physiology. We identify that Wnt16, shown previously to be important in postnatal bone strength and homeostasis (22), is expressed in arterial smooth muscle cells. Our data reveal that Wnt16 genetically contributes to cardiovascular disease physiology, both in humans and in a preclinical disease model of AngII-induced aneurysmal remodeling. We show that Wnt16 supports the contractile VSM phenotype, mediated in part via a noncanonical Wnt16-Taz signaling relay. As Milewicz and colleagues have highlighted, the thoracic ascending aorta (TAA) is remarkably sensitive to aneurysmal dilation in response to heritable mutations that globally impair the VSM contractile phenotype, including Acta2 (29, 74). TAA aneurysm and dissections predominate with Acta2 variants, with a lower abdominal aortic aneurysm risk even though mutations in Acta2 are globally present (29, 74). This predilection may relate to the mechanical forces experienced by the TAA. Importantly, with murine Acta2 deficiency the baseline systolic blood pressure is reduced prior to aneurysm induction (29)—a phenotype we also observed in Wnt16−/− mice. Similar results have also been reported by Tellides and colleagues in their study of TAA aneurysm Tsc1 null mice (73). It was this phenotype—namely, the convergence of increased TAA aneurysmal dilatation in the setting of reduced baseline systolic blood pressure—that prompted our studies of Wnt16 and the contractile VSM phenotype.

Very recent data from Daoud et al (113) and Zhang and colleagues (114) have also identified a critical role for Yap/Taz signaling in maintenance of the VSM contractile phenotype, consistent with our studies of Wnt16−/− mice. Our work extends these recent reports, providing not only a novel therapeutic strategy via specific Wnt16-mediated activation but also mechanistic insights highlighting unique contributions of VSM Taz vs Yap1 signaling in aortic physiology. In unpublished observations, we have confirmed increased adventitial accumulation of CD11b+ cells of the myeloid lineage, consistent with the increased Mmp12 protein we discovered in proteomic studies of Wnt16-null aortas. While we focus upon Wnt16-regulated VSM phenotype in this study, we anticipate that future studies of Wnt16-regulated VSM-myeloid cell interactions will emerge as important in arterial disease physiology—akin to the osteoblast-osteoclast interactions perturbed by Wnt16 deficiency in osteoporosis (22). Since Yap1 siRNA uniquely induced the myeloid chemotactic factor Ccl2, we speculate that Wnt16 regulation of VSM Yap1 activity may mediate VSM-monocyte interactions in the vessel wall. It is interesting to note that, unlike Yap1, Taz can uniquely homodimerize to activate dyads of adjacent TEAD elements (101), a configuration we observed in the Wnt16- and Taz-responsive Acta2 promoter (Fig. 14). The unique roles of VSM Taz vs Yap1 in aortic disease biology remain to be fully investigated.

The consequences of Wnt16 deficiency in bone only become apparent during postnatal aging (22, 23, 62). Saul, Khosla, and colleagues recently established Wnt16 expression in a molecular signature of cellular senescence that is shared across multiple tissues including bone (92). The reduction in VSM mitochondrial spare respiratory capacity we identify with Wnt16 deficiency is also a prominent feature of senescence (54, 115, 116). Moreover, Yap/Taz signaling in stromal cells declines with age, and sustained Yap/Taz expression delays the cellular SASP of inflammaging (117). AngII induces precocious VSM senescence via Cdkn2a (p16) mechanisms (118), and our data revealed that Wnt16 suppresses VSM Cdkn2a expression. Thus, in addition to promoting VSM contractile phenotype, Wnt16 may also restrain VSM senescence in part by activating Yap/Taz relays as relevant to aneurysmal remodeling (reviewed in reference (119). Moreover, downregulation of the mitochondrial stress mitokine Gdf15 (93, 94) in both VSM and monocyte lineages with Wnt16 treatment suggests that mitochondrial dysfunction—first evident in the aberrant mitochondrial morphology and respiration we noted in Wnt16-null arterial VSM—might also arise in monocytes with reduced Wnt16 signaling tone. In aging skeletal muscle, alterations in mitochondrial morphology with altered Gdf15 mitokine expression suggest abnormal mitochondrial quality control (120). Whether Wnt16 regulates mitochondrial quality control remains to be addressed. In addition, the relative contributions of Wnt16 signaling relays in VSM vs monocyte lineages to arterial remodeling have yet to be established.

It is interesting to note that antibody-mediated inhibition of sclerostin, an antagonist of canonical Wnt actions with great utility in the treatment of severe osteoporosis, has potentially negative actions in the cardiovascular system (121, 122). Use of anti-sclerostin therapy is contraindicated in patients with recent cardiovascular disease events because of these concerns. Given the homeostatic role of the Wnt16-Taz relay in mitigating arterial remodeling and preserving the contractile VSM phenotype—and modifying certain features of senescence—sclerostin inhibition may perturb arterial remodeling and plaque stability by sustaining a canonical Wnt signaling bias. During arterial remodeling and repair, such a monotonic Wnt signaling bias toward canonical actions may have negative consequences in high-risk patients with pre-existing cardiovascular disease (121, 122). More sophisticated, personalized stratification of cardiovascular risk in our aging patient populations, also at highest risk for severe osteoporosis and its devastating consequences, may improve risk-benefit assessment and health outcomes by guiding osteoporosis pharmacotherapy (9).

Bulk RNA-seq data as collated by the Bgee curated interspecies expression atlas (123), and reverse transcription–PCR (124), have documented significant WNT16 expression in the human cardiovascular system as consistent with our data. However, the curated transcriptomics of the Human Protein Atlas has not captured WNT16 expression in cardiovascular tissue or bone (125). This problem has been observed for other Wnt genes including WNT1, as Lee and colleagues discussed in their detailed human and murine studies of WNT1 and Wnt1 in osteogenesis imperfecta (126, 127); differences in the relative accumulation of Wnt transcripts may exist between humans and mice, even though molecular genetics reveals similar biological roles and relevance in both species (126, 127).

There are limitations to our study, with many additional questions yet to be investigated. Our preclinical studies challenged Wnt16−/−mice with AngII as a mechanism to accelerate aneurysm and vascular disease. However, the composite cardiovascular endpoint (stroke, myocardial infarction, cardiovascular death, other atherosclerotic disease events, and heart failure) of the Dallas Heart Study also encompassed nonaneurysmal cardiovascular diseases significantly impacted by WNT16 genotype. The role of Wnt16 in other VSM-dependent cardiovascular injury responses, including neointima formation and plaque stability, remains to be explored. Cellular senescence as regulated by Wnt16-Yap/Taz relays may represent a common mechanistic underpinning. While the basic framework of the Wnt16-Taz signaling relay as relevant to the VSM phenotype has emerged (Fig. 16), key upstream components are as yet unknown. The Wnt16-regulated receptor complexes conveying responses in VSM have yet to be identified. The unique features of Taz vs Yap1 (101, 128, 129) signaling points to the importance of identifying these receptor complexes. While our focus on contractile phenotype experimentally directed our attention to Taz-dependent transcription, there is no doubt that regulation of Yap1 actions will also be important, as Zhang et al have highlighted (114). The relationship between the multiple aortic LIM domain proteins we observe to be upregulated with Wnt16 deficiency and the impaired VSM contractile phenotype is also unknown. Given our data, it is tempting to speculate that these changes relate to altered mechanosensation of tensed F-actin, mediated by LIM domain proteins (100). Since (a) focal adhesion biology is central to actin-dependent contractile force generation (106); and (b) acute Wnt16 depletion, Wnt16 genetic alteration, and Wnt16 treatment all regulated the contractile phenotype, future studies will examine Wnt16-regulated focal adhesion biochemistry as necessary for force generation. The discovery that Ccm proteins are critical in focal adhesion-dependent control of Yap1-Taz activation (106), relevant to mechanosensation and cell senescence (117), gives further impetus to pursue this line of investigation. Nevertheless, our studies provide the first genetic and molecular evidence, with novel mechanistic insight, of a role for Wnt16 signaling in arterial physiology as relevant to cardiovascular health. Wnt16 regulates cardiovascular physiology and VSM contractile phenotype, mediated in part via Taz signaling. As such, strategies that augment VSM Wnt16—Yap/Taz relays may help mitigate clinical cardiovascular frailty and risk, including TAA aneurysms, that afflict our aging patient population (13).

Abbreviations

ACN

acetonitrile

Acta2

vascular smooth muscle alpha actin

AngII

angiotensin-II

Ankrd1

ankyrin repeat domain 1 (a Yap1/Taz target)

C3

complement protein C3

C4b

complement protein C4b

CArG

C-A/T rich- G binding element CC(A/T)6GG

Ccl2

CC chemokine ligand 2 (also known as monocyte chemoattractant protein 1)

CCM

cerebral cavernous malformation protein

ChIP

chromatin immunoprecipitation

cSNP

coding single nucleotide polymorphism

Ctgf

connective tissue growth factor (also known as CCN2 a Yap1/Taz target)

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

HFD

Western high fat diet

Lats

large tumor suppressor kinase (also known as Warts)

LC-MS/MS

tandem liquid chromatography–mass spectrometry

LDLR

low-density lipoprotein receptor

LIM

Lin11, Isl1, and MEC3 tandem zinc finger domain protein

LRP

LDLR-related protein

LUC

firefly luciferase reporter gene

Mmp

matrix metalloproteinase

Mst

mammalian sterile 20-like kinase (also known as Stk or Hippo)

Myh11

vascular smooth muscle myosin heavy chain 11

PWV

pulse wave velocity

RNAi

RNA interference

RT-PCR

reverse-transcriptase polymerase chain reaction

SASP

senescence associated secretory phenotype

Sca1

stem cell antigen 1 (also known as Ly6a)

SDS-PAGE

sodium dodecyl sulfate–polyacrylamide gel electrophoresis

siRNA

small interfering RNA

SM22

transgelin

Smtn

Smoothelin

SRF

serum response factor

TAA

thoracic ascending aorta

Taz

transcription factor Wwtr1

TBE

tribromoethanol

TEAD

TEA domain (TEF1 and abaA) transcription factor, binds DNA element AGGAATG

TGF

transforming growth factor

TNAP

tissue non-specific alkaline phosphatase

TNF

tumor necrosis factor

TSS

Transcriptional start site

UT

University of Texas

VSM

vascular smooth muscle

Wnt

wingless-type MMTV integration site family member

Wwtr1

WW domain containing transcription factor 1, (also known as Taz)

Yap1

Yes1 associated transcriptional regulator

Funding

Supported by National Institutes of Health Grants R01HL069229 and R03TR004144, the J.D. and Maggie E. Wilson Endowed Chair, and the Louis V. Avioli Professorship to D.A.T.

Disclosures

The authors have declared that no conflict of interests exists, including industry relationships.

Data Availability

All data sets are available either as presented in the manuscript, the data repository supplement to this manuscript (33), or from the corresponding author upon reasonable request. Proteomics data sets will be available upon acceptance of the manuscript.