WISP-1 drives bone formation at the expense of fat formation in human perivascular stem cells

The vascular wall within adipose tissue is a source of mesenchymal progenitors, referred to as perivascular stem/stromal cells (PSC). PSC are isolated via fluorescence activated cell sorting (FACS), and defined as a bipartite population of pericytes and adventitial progenitor cells (APCs). Those factors that promote the differentiation of PSC into bone or fat cell types are not well understood. Here, we observed high expression of WISP-1 among human PSC in vivo, after purification, and upon transplantation in a bone defect. Next, modulation of WISP-1 expression was performed, using WISP-1 overexpression, WISP-1 protein, or WISP-1 siRNA. Results demonstrated that WISP-1 is expressed in the perivascular niche, and high expression is maintained after purification of PSC, and upon transplantation in a bone microenvironment. In vitro studies demonstrate that WISP-1 has pro-osteogenic/anti-adipocytic effects in human PSC, and that regulation of BMP signaling activity may underlie these effects. In summary, our results demonstrate the importance of the matricellular protein WISP-1 in regulation of the differentiation of human stem cell types within the perivascular niche. WISP-1 signaling upregulation may be of future benefit in cell therapy mediated bone tissue engineering, for the healing of bone defects or other orthopaedic applications.

in a perivascular location. WISP-1 is better known to be expressed in osteoprogenitor cells, either during skeletal development or fracture repair 13 . CCN family members all have roles in osteochondral cell specification, although the relative importance for bone or cartilage differentiation differs between family members [14][15][16] . Mechanistically, WISP-1 exerts complex and incompletely understood effects on both canonical Wnt and BMP (Bone morphogenetic protein) signaling in order to specify MSC lineage determination and osteogenic differentiation 13,[17][18][19] . For example, at the extracellular surface of the MSC, WISP-1 binds to BMP2 to enhance BMP2 binding to BMPR1/2, resulting in Smad1/5/8 phosphorylation and canonical BMP signaling activation 18 . Recent studies have also found WISP-1 to functionally 'de-repress' canonical Wnt signaling, by blocking Sclerostin (SOST) binding to LRP5 19 . The exact mechanism by which WISP-1 blocks SOST/LRP5 binding is not yet known. As well, recent studies have elucidated important roles for WISP-1 in bone maintenance. Mice with global Wisp-1 deficiency display a low bone mass phenotype, with reduced cortical and trabecular bone, reduced osteoprogenitor cell differentiation, increased osteoclast activity, and increased sensitivity to ovariectomy induced bone loss 19 . Conversely, Wisp-1 overexpression driven by the Col1a1 promoter leads to a high bone mass phenotype 18 . In aggregate, WISP-1 is a novel pro-osteogenic secreted matricellular protein that enhances both Wnt and BMP signaling. These observations led us to examine the localization and function of WISP-1 within the perivascular niche and in human PSC.

WISP-1 localization to the perivascular niche.
To confirm the biologic relevance of WISP-1 in PSC biology, we first returned to the in vivo residence of PSC, in the perivascular niche of human adipose tissue. By immunohistochemical detection in human adipose tissue, WISP-1 showed a strong perivascular localization in both the perivascular cells and associated vascular wall connective tissue (Fig. 1). WISP-1 immunoreactivity of intermediate to strong intensity was present in vessels of all calibers within adipose tissue, including arterioles, capillaries, venules and larger veins. In contrast, inconspicuous staining of adipocytes or other tissue elements was seen. These findings were supported by the Human Protein Atlas (proteinatlas.org), which confirmed predominant WISP-1 immunoreactivity in the vascular wall in sections of adipose tissue. Also published within the Human Protein Atlas was high WISP-1 immunoreactivity in the tunica media and adventitia of thick walled arteries (CAB012205, Soft tissue 2, proteinatlas.org), which not present for analysis on our study material. Thus, WISP-1 protein is highly expressed across vessel types in human adipose tissue, in a distribution that includes both presumptive microvascular pericytes and adventitial progenitor cells (APCs).
Finally, FACS purified, uncultured PSC were applied to a previously validated lumbar spinal fusion model in athymic rats, and WISP-1 expression was interrogated during the process of in vivo bone formation 20 (Fig. 2C-J). In this xenotransplant model, and as we have previously reported, defined human PSC cell numbers (1.5 × 10 6 human cells) when applied in a demineralized bone matrix scaffold to the lateral aspects of the rat lumbar spine induces a 100% incidence of spinal fusion after a 4 week period 20 . Spinal fusion segments were treated with or without human PSC (see Table 1 for treatment group allocation). The H&E appearance of PSC treated spinal fusion segments demonstrates interconnected bone trabeculae which incorporate the devitalized bone graft scaffold (DBM, demineralized bone matrix) (Fig. 2C,E). Higher magnification demonstrates easily discernable bone-lining osteoblasts within PSC treated samples (arrowheads, Fig. 2E), as well as vascular and bone marrow elements between bony trabeculae. In contrast, spinal fusion segments treated without PSC showed DBM particles predominantly embedded in fibrous tissue and with only sporadic new bone formation (Fig. 2D). At higher magnification, control treated sections showed inconspicuous bone lining cells were present on the edges of DBM particles (arrowheads), which lacked the cuboidal morphology of synthetically active osteoblasts. WISP-1 immunolocalization was next performed on representative PSC-or control-treated spinal fusion segments ( Fig. 2G-J). In PSC-treated specimens, strong and diffuse WISP-1 immunoreactivity was seen in bone lining osteoblasts (Fig. 2G,I), as well as foci of endochondral ossification (not shown). In contrast, weak to intermediate and more focal immunostaining for WISP-1 was identified in control sections, predominantly in inflammatory cells and stromal fibroblasts (Fig. 2H,J). In aggregate, high WISP-1 expression was observed among cells of the perivascular niche of adipose tissue, both in situ, after FACS purification, and upon in vivo orthopaedic application.   (Fig. 3A). The consequences of WISP-1 knockdown on human PSC osteogenic differentiation were next assessed ( Fig. 3B-D). Consistent with prior reports in other cell types, gene expression for markers of osteogenic differentiation were significantly reduced among PSC with WISP-1 knockdown (Fig. 3B). This included a reduction in the transcription factor runt-related transcription factor 2 (RUNX2, 58.4% reduction), the enzyme alkaline phosphatase (ALP, 41.3% reduction), the principal matrix protein of bone collagen type I alpha I chain (COL1A1, 27.0% reduction), as well as a non-significant reduction in the terminal differentiation marker osteocalcin (OCN, 57.2% reduction) (Fig. 3B). In agreement with gene expression studies, staining and photometric quantification of ALP enzymatic activity was significantly reduced with WISP-1 knockdown (66.5% reduction, d 12 of differentiation) (Fig. 3C). Bone nodule deposition was assessed by Alizarin Red (AR) staining and photometric quantification, and likewise showed a robust inhibition of bone nodules with WISP-1 knockdown (62.1% reduction, d 12 of differentiation) (Fig. 3D). As mentioned, WISP-1 is known to positively regulate both canonical BMP and canonical Wnt signaling in the process of osteogenic differentiation. In the context of WISP-1 knockdown in human PSC, gene markers of BMP and Wnt signaling activity were next assessed (Fig. 3E). Three days post-transfection, the BMP2 responsive element, inhibitor of DNA binding 1 (ID1) was significantly reduced in expression under WISP-1 siRNA treatment conditions (44.1% reduction). In contrast, transcript abundance for the canonical Wnt signaling marker AXIN2 (Axis Inhibition Protein 2) was not significantly affected by WISP-1 knockdown. These results suggested that at least in the context of WISP-1 knockdown in human PSC osteogenic differentiation, differential BMP signaling may represent the dominant effect. To further pursue this, recombinant BMP2 (rBMP2) was added to osteogenic differentiation medium (ODM), with or without WISP-1 knockdown (Fig. 3F). As expected, rBMP2 treatment under control conditions increased transcripts for osteogenic markers, including a 33.2-119.2% increase in RUNX2, ALP and OCN at 3 days differentiation (blue bars, Fig. 3F). WISP1 knockdown again significantly reduced all osteogenic gene markers (red bars, Fig. 3F). Moreover, WISP-1 knockdown significantly reversed rBMP2 induction of osteogenic marker expression (purple bars, Fig. 3F).

WISP-1 gain of function and PSC differentiation.
The effects of gain-of-function in WISP-1 signaling were next assessed, using either WISP-1 overexpression or recombinant WISP-1 protein (Fig. 4). First, PSC were transfected using a WISP-1 expression plasmid and the consequences on osteogenesis and adipogenesis were examined (Fig. 4A-D). As expected, PSC transfection led to a significant upregulation in the production of WISP-1 transcripts which was maintained across three timepoints of osteogenic differentiation (days 6, 9, and 15) (Fig. 4A). WISP-1 overexpression in PSC under osteogenic differentiation conditions led to an increase in osteogenic gene markers, including RUNX2 (119.1% increase), ALP (15.1% increase), and a non-significant trend toward increased OCN expression (Fig. 4B). WISP-1 overexpression was next examined in PSC under adipogenic differentiation conditions (Fig. 4C,D). WISP-1 overexpression led to a significant reduction in adipogenic gene markers, including PPARγ (42.0% reduction) and a trend towards reduced CEBPα expression (Fig. 4C). Oil Red O staining likewise showed a significant reduction in the intracellular accumulation of lipid with WISP-1 overexpression after 15 days differentiation (Fig. 4D).

Discussion
Here, we have elucidated a new role of WISP-1 signaling within stromal progenitor cells of the perivascular niche of human adipose tissue. WISP-1 has been previously defined by its role as a bone matricellular protein and in its positive regulation of osteoblastogenic differentiation in other osteoblastic cell types 13,19 . However, prior studies have described a role for WISP-1 in the vascular wall, specifically in vascular smooth muscle cell proliferation and migration 21 . Importantly, vascular smooth muscle cells represent a distinct and terminally differentiated cell, and lack the multipotentiality of PSC. WISP-1 has also been implicated in the promotion of angiogenesis, especially in tumor associated angiogenesis [22][23][24] . In the current study, we found enrichment of WISP-1 within PSC both in situ, after purification, and after in vivo application in a bone defect microenvironment. This led us to examine the modulation of PSC differentiation by WISP-1 signaling.
WISP-1 has previously been shown to induce osteogenic differentiation by multiple mechanisms, including association with biglycan (BGN) in the matrix of mineralizing tissues 25 , activation of BMP signaling 18 , and potentiation of canonical Wnt signaling 13 . Our examination of PSC in vitro suggested a role for WISP-1 in the osteogenic differentiation of PSC by regulation of BMP activity. Previous examination of the effects of WISP-1 on BMP2 activity by Ono et al. showed that a direct interaction between BMP2 and WISP-1 is essential for the enhancement of canonical BMP signaling, and that neutralizing antibodies that block integrin α 5 β 1 mitigated this interaction 18 . An independent study by Kohaha et al. found that BMP2 and WISP-1 had synergistic effects in an ectopic bone formation model in mice 17 . In the current study, we observed that the knockdown of WISP-1 led to a marked decrease in transcripts of the BMP responsive element ID1. Moreover WISP-1 knockdown antagonized Scientific RepoRts | (2018) 8:15618 | DOI:10.1038/s41598-018-34143-x the pro-osteogenic effects of BMP2, and recombinant WISP-1 synergized with rBMP2 with the promotion of PSC osteogenesis.
The effects of WISP-1 on adipogenic differentiation are less well understood. Recent studies have found that human ASC express and secrete WISP-1 during in vitro adipogenesis, and that changes in adiposity alter WISP-1 expression 26 . As a regulator of both Wnt and BMP signaling, it would be anticipated that the role of WISP-1 in adipogenic differentiation is complex. Our results of the anti-adipocytic effects of WISP-1 are an independent confirmation of a recent report by Ferrand et al. who observed that WISP-1 is a negative regulator of adipogenesis 27 . Here, they observed that WISP-1 expression is reduced during adipogenesis, that adipocytic cell lines demonstrate a significant reduction in adipogenesis with WISP-1 overexpression associated with decreased PPARγ transcriptional activity, and that WISP-1 knockdown induces preadipocyte differentiation 27 . The current study confirms these same anti-adipocytic effects of WISP-1 in human multipotent mesenchymal cells.
In summary, our experiments showed that the matricellular protein WISP-1 is highly expressed in PSC in the vascular wall of human adipose tissue. Our results demonstrate the importance of WISP-1 in the positive regulation of osteogenesis and negative regulation of adipogenesis in human stem cell types within the perivascular niche. Thus, manipulation of WISP-1 signaling may have broad and important therapeutic implications for In the context of bone tissue engineering, a cellular therapy augmented with WISP-1 protein may be used to ossify bones with more success and rapidity than an unstimulated cell type. Conversely, antagonism of WISP-1 or downstream signaling may be of future benefit when an adipocytic cell fate is most desirable in soft tissue augmentation or reconstruction.

Materials and Methods
Perivascular stem/stromal (PSC) cell isolation. PSC were isolated from human subcutaneous adipose tissue via fluorescence activated cell sorting (FACS). Human lipoaspirate was obtained from healthy adult donors under IRB approval at UCLA and JHU with a waiver of informed consent, and was stored for less than 48 h at 4 °C before processing. The SVF of human lipoaspirate was obtained by collagenase digestion. Briefly, lipoaspirate was diluted with an equal volume of phosphate-buffered saline (PBS) and digested with Dulbecco's modified Eagle's medium (DMEM) containing 3.5% bovine serum albumin (Sigma-Aldrich, St. Louis) and 1 mg/ml type II collagenase for 70 min under agitation at 37 °C. Adipocytes were separated and removed by centrifugation. The cell pellet was resuspended in red blood cell lysis buffer (155 mM NH 4 Cl, 10 mM KHCO3, and 0.1 mM EDTA) and incubated at room temperature. After centrifugation, cells were resuspended in PBS and filtered at 70 μm. The resulting SVF was further processed for cell sorting, using a mixture of the following directly conjugated antibodies: anti-CD34-allophycocyanin (1:100; BD Pharmingen, San Diego, CA), anti-CD45-allophycocyanin-cyanin 7 (1:100; BD Pharmingen), anti-CD146-fluorescein isothiocyanate (1:100; Bio-Rad, Hercules, CA), and anti-CD31-allophycocyanin-cyanin 7 (1:100, Bio Legend, San Diego, CA (summary of antibodies presented in Table 2). All incubations were performed at 4 °C for 15 min. The solution was then passed through a 70-μm cell filter and then run on a FACS Diva 8.0.1 cell sorter (BD Biosciences). In this manner, microvessel pericytes (CD146+CD34−CD45−CD31−) and adventitial cells (CD34+CD146−CD45−CD31−) were isolated and combined to constitute the PSC population. Sorted PSC were either snap frozen for RNA isolation, applied in a rat spinal fusion model, or culture expanded for in vitro studies. For in vitro expansion, cells were cultured at 37 °C in a humidified atmosphere containing 95% air and 5% CO 2 . PSC were cultured in DMEM, 20% fetal bovine serum (FBS), 1% penicillin/streptomycin. Medium was changed every three d unless otherwise noted.
Osteogenic differentiation assays. Assays for PSC differentiation are adapted from our prior publications 28,29 . Osteogenic differentiation medium (ODM) was constituted with 10 mM β-glycerophosphate and 50 μM ascorbic acid in DMEM + 20% FBS. In select experiments, ODM was supplemented with recombinant WISP-1 (200 ng/mL) (ab50041, Abcam, Cambridge MA) or BMP2 (50 ng/mL, Medtronic), and protein supplementation was changed every three d of differentiation. In select experiments, WISP-1 gene expression was manipulated by siRNA or overexpression by plasmid using appropriate controls (see below).
Alkaline phosphatase staining was performed using the Leukocyte Alkaline Phosphatase Kit (Sigma-Aldrich). Briefly, cells were seeded in 24 well plates at a density of 1 × 10 4 cells/well. Cells were cultured under osteogenic differentiation conditions for 12 d prior to staining. Cells were then washed with PBS and fixed with formalin for 10 min at room temperature. Following fixation, cells were stained using Leukocyte Alkaline Phosphatase Kit (Sigma-Aldrich) according to the manufacturer's protocol. Cells were incubated in alkaline phosphatase stain for 15 min at 37 °C, then washed with PBS. Cells were allowed to dry and pictures were taken at 100x magnification using Olympus IX71 inverted system microscope (Olympus, Cypress, CA). Relative staining was quantified using Adobe Photoshop CC2015.
For the detection of mineralization, cells were seeded in growth medium in 24 well plates at a density of 1 × 10 4 cells/well. 24 h after seeding cells, basal medium was replaced with ODM in triplicates per each treatment for 12 d. Cells were washed with PBS and fixed with 4% paraformaldehyde. Following fixation, cells were stained with 2% alizarin red (Sigma-Aldrich, Saint Louis, Missouri) at room temperature for 15 min then washed with deionized water and allowed to dry. Pictures were taken at 100x magnification using Olympus IX71 inverted system microscope (Olympus, Cypress, CA). In order to quantify bone nodule deposition, 10% v/v acetic acid was added and cells were incubated at room temperature for 30 min with shaking. Cells were then scraped from the wells and vortexed for 30 s. Next, cells were overlaid with mineral oil and heated to 85 °C for 10 min. Briefly, cells were cooled on ice for 5 min then centrifuged at 20,000 × g for 15 min. 10% ammonium hydroxide was added to adjust the pH to between 4.1 and 4.15. Absorbance was measured in triplicates at 405 nm in 96 well plates using Epoch microspectrophotometer (Bio-Tek, Winooski, VT).
Adipogenic differentiation assays. PSC were seeded in six well plates at a density of 1 × 10 5   Red O powder with 100 ml of 99% isopropanol. Stock solution was diluted 3:2 stock solution: deionized water and allowed to sit at room temperature for 10 min. The working solution was then filtered by gravity filtration. After fixation by formalin, cells were washed with water and 60% isopropanol was added to wells for 5 min before staining. Oil Red O working solution was then added to each well and incubated at 37 °C for 30 min. Following incubation, the cells were washed with tap water images were taken using Q capture software.

Ribonucleic acid (RNA) isolation and quantitative real-time polymerase chain reaction (qRT-PCR).
Gene expression was assayed by quantitative RT-PCR, based on our previous methods 29,30 . Primers sequences are shown in Table 3. Timepoints for specific gene expression include 3-15 d of osteogenic or adipogenic differentiation, with additional details provided in the accompanying figure legends. Briefly, total RNA was extracted using RNEasy Kit (Qiagen, Santa Clarita, CA). 1 μg of total RNA from each sample was subjected to first-strand complementary deoxyribonucleic acid (cDNA) synthesis using the SuperScript III Reverse-Transcriptase Kit (Life Technologies) to a final volume of 20 μL. The reverse transcription reaction was performed at 65 °C for 5 min, followed by 50 °C for 50 min and 85 °C for 5 min. For qRT-PCR, the reaction was performed using 2 × SYBR green RT-PCR master mix and an ABI PRISM 7300 qRT-PCR system instrument (Applied Biosystems, Foster City, CA). qRT-PCR was performed using 96 well optical plates at 95 °C for 10 min, followed by 40 cycles at 95 °C for 15 s, and at 60 °C for 60 s. The relative quantification of gene expression was performed using a Comparative CT method according to the manufacturer's protocol and was normalized to the expression levels of ACTB in each sample.

Small interfering RNA (siRNA) and transfection. Knockdown of WISP-1 was performed using
Silencer Select chemically synthesized siRNA (Thermo Fisher Scientific, catalog number: 4392420; S16873). Cells were seeded in 6 well plates at a density of 4 × 10 4 /well. At 50% confluence, basal medium was replaced with antibiotic-free basal medium. Transfection was performed using X-tremeGENE siRNA Transfection Reagent (Sigma-Aldrich) and 150 pM WISP-1 siRNA or scramble siRNA diluted in minimal essential medium (Opti-MEM). For the confirmation of siRNA efficiency: six h post-transfection, medium was replaced with basal medium and the efficiency of the knockdown was validated using qRT-PCR. For alizarin red (AR) and alkaline phosphatase (ALP) staining: Six h post-transfection, medium was replaced with ODM and cells were cultured under osteogenic differentiation conditions for up to 12 d.  20 . Animals were housed and experiments were performed in accordance with guidelines of the Chancellor's Animal Research Committee of the Office for Protection of Research Subjects at the University of California, Los Angeles. All animals were treated with postoperative medications of buprenorphine for 48 h and trimethoprim/sulfamethoxazole for 10 d, for pain management and prevention of infection, respectively. Rats were anesthetized using isoflurane (5% induction, 2-3% maintenance). Posterior midline incisions were made over the caudal portion of  the lumbar spine, and two separate fascial incisions were made 4 mm bilaterally from the midline. Blunt muscle splitting technique was used lateral to the facet joints to expose the transverse processes of L4 and L5 lumbar spines. The processes were then decorticated using a low speed burr under regular irrigation with sterile saline solution to cool the decortication site and maintain a clean surface for implantation. Next, the treatment material was delivered via a scaffold, implanted between the transverse processes bilaterally into the paraspinal muscle bed. Finally, the fasciae and skin were each closed using a simple continuous technique with 4-0 Vicryl sutures (Ethicon Endo-Surgery, Blue Ash, OH). Rats were sacrificed 4 weeks postsurgery via CO 2 overdose, and the spines were harvested for analysis. Spinal fusion with PSC treatment was confirmed by biomechanical testing and radiographic analysis as previously reported 20 . Samples were next decalcified using 19% EDTA and embedded in paraffin. Immunohistochemical staining was performed with primary antibodies against WISP-1 (1:100, ab155654, Abcam, Cambridge, MA) using the ABC (Vector Laboratories, Burlingame, CA) method. Immunohistochemistry was performed after paraffin sections were deparaffinized, rehydrated, rinsed, and incubated with 3% H 2 O 2 for 20 min. All sections were then blocked with 0.1% bovine serum albumin in PBS for 1 h. At a dilution of 1:100, primary antibodies were added to each section and incubated at 37 °C for 1 h at 4 °C. Images were obtained on an Olympus (Center Valley, PA) BX51 microscope.
Statistical analysis. All results were expressed as mean ± standard deviation (SD). Statistical analyses were performed using the SPSS16.0 software. All data were normally distributed. Student's t test was used for two-group comparisons, and one-way ANOVA test was used for comparisons of 3 or more groups, followed by Tukey's post hoc test. Differences were considered significant when *P < 0.05 and *P < 0.01.

Data Availability
The datasets generated or analyzed during the current study are available from the corresponding author on reasonable request.