Endocannabinoids increase human adipose stem cell differentiation and growth factor secretion in vitro

Adipose stem cells (ASCs) possess the capacity to proliferate, to differentiate into various cells types, and they are able to secrete growth factors. These characteristics are supposed to contribute to their potential for regenerative medicine approaches. In order to advance the therapeutic effects of ASCs, different modulatory procedures have been examined. In this context, the endocannabinoid system (ECS) represents an interesting possibility, since the increased availability of cannabinoids and the underlying molecular pathways of the ECS are of relevance for the development of new regenerative strategies. The effects of the endocannabinoids anandamide (AEA) and 2‐arachidonoylglycerol (2‐AG) were investigated on ASC metabolic activity, quantified by PrestoBlue conversion, and cell numbers, evaluated by crystal violet staining. enzyme‐linked immunosorbent assay (ELISA) measures were performed to determine cytokine release, and differentiation was assessed by specific labeling techniques. AEA increased the metabolic activity, while 2‐AG decreased it in a concentration dependent manner. AEA significantly enhanced OilRed O staining after adipogenic differentiation by over 100%, and both compounds significantly increased cresolphthalein staining after osteogenic differentiation. By contrast, they did not affect sphere diameter or safranin O staining after chondrogenic differentiation. Both substances significantly increased the release of insulin‐like growth factor‐1 and hepatocyte growth factor, while only AEA enhanced transforming growth factor‐β secretion. The results demonstrated that stimulating the ECS exerted significant effects on the biology of ASCs. Exposure to endocannabinoids modulated viability, induced release of regenerative growth factors, and promoted adipogenic and osteogenic differentiation. Our findings could be of specific relevance in ASC based therapies for regenerative medicine.


| INTRODUCTION
Advances in understanding the interaction of biomaterials, growth factors, and progenitor cells contribute to the emerging field of regenerative medicine. In plastic surgery, regenerative medicine promises alternative solutions to classic reconstructive techniques by harnessing the endogenous reparative resources. Indications for such concepts encompass congenital defects, aging or trauma with acute and chronic wounds. Adipose stem cells (ASCs) have been identified suitable for therapeutic application due to their abundance, easy harvest and high regenerative potential (Zuk et al., 2001). These cells have shown an improved outcome in wound healing, as evidenced by their efficacy in case studies of burn wound healing and scarring (Conde-Green et al., 2016), and in pilot studies focusing on ischemic (Lee et al., 2012) and skin cancer-related wounds (Rigotti et al., 2007). Also, phase II randomized clinical trials evidenced the safety of ASC therapy and its positive effects on healing of chronic leg ulcers (Zollino et al., 2019), as well as complex perianal fistulae (Garcia-Olmo et al., 2009).
Although the paracrine activity of ASCs and their ability for multilineage differentiation is unquestioned, the anticipated merits of ASCs have not completely met the initial expectations in the clinical scenario (Patrikoski, Mannerstrom, & Miettinen, 2019).
When translating ASCs into clinical practice, the still missing precise comprehension of the ASC biology is one reason for high attrition rates (Arrowsmith & Miller, 2013). To fully exercise their desired regenerative effects at the recipient site, ASCs have to survive, proliferate, release soluble factors or differentiate into distinct cell lineages. In the context of wound healing, several attempts to direct the regenerative capacities and to increase the survival of ASCs have been suggested, including 3-D scaffolds, hypoxia and the supplementation of bioactive proteins (Li & Guo, 2018). Possibilities to modify ASC characteristics in vivo or in vitro are the genetical manipulation or pharmacological conditioning (Seo, Shin, & Kim, 2019). A new promising approach could target on the endocannabinoid system (ECS). The term ECS derives from the Asian hemp Cannabis sativa, which has been used as medical in ancient times, i.a., for the treatment of wounds (Butrica, 2002). The ECS is an endogenous system that consists of synthesizing and degrading enzymes of the endocannabinoids anandamide (AEA) and 2-arachidonoylglycerol (2-AG), and the cannabinoid receptors. It is well-known for its modulatory effects on central nervous system processes, but endocannabinoid signaling is also observed in various nonneuronal tissues, including adipose (Maccarrone et al., 2015).
In a recent study, we investigated the effects of specific activation of the two cannabinoid receptors, CB1 and CB2, of ASCs (Ruhl, Karthaus, Kim, & Beier, 2020). We found that CB2 activation promotes proliferation, while CB1 ligation enhances adipogenesis and chondrogenesis, and agonists of both receptors induce growth factor secretion. The effects of endocannabinoid exposure, which bind to further receptors besides CB1 and CB2, and are thus more comparable to the medical cannabinoids, have not been investigated.
In order to elucidate the role of the ECS in ASC viability, release of soluble factors and differentiation, we exposed ASCs to the endocannabinoids in vitro. Results of this study may unravel potential strategies to: 1. Precondition patients before fat grafting/ASC harvest, 2. Condition ASCs within harvested fat grafts/stromal vascular fraction (SVF) or 3. Condition cultured ASCs for tissue engineering purposes, with the aim to increase the regenerative potential of ASCs.

| Materials
PrestoBlue, fetal bovine serum (FBS), ITS 1 premix, high glucose medium (4.5 g/L), low glucose medium (1 g/L) and Dulbecco's and used for ASC isolation as described previously (Ruhl, Storti, & Pallua, 2017). The donors were informed about the cellular use from their tissue and gave informed consent. The study protocol was approved by the regional ethics committee (Ethics Committee of the RWTH Aachen University; EK163/07).
ASCs were expanded in proliferation medium (DMEM with 0.1% bFGF, 10% FBS, 1% P/S), until reaching confluence of ∼90%. Experiments were performed with cells from passages two to four seeded at a density of 3 � 10 4 cells per cm 2 for experiments targeting viability, RUHL ET AL.
For chondrogenic differentiation, 10 5 cells were seeded per cm 2 in 2D-culture or used for pellett culture in eppendorf tubes (Ruhl & Beier, 2019). The cells were exposed to chondrogenic medium for 5 days in 2D or for 21 days in 3D. To ensure complete adherence of the cells, cannabinoids were freshly added to the media three days after seeding, except for the 3D chondrogenic differentiation protocol, and during each media exchange.

| Crystal violet assay
After seven days, ASCs were stained with 0.1% crystal violet solution as described earlier (Ruhl, Kim, & Beier, 2018). The staining solution was removed and crystal violet was washed out with 33% acetic acid.
The samples were transferred to an optical plate and absorbance was quantified at 620 nm in a FLUOstar Optima microplate reader (BMG Labtech).

| PrestoBlue assay
Metabolic activity was measured by PrestoBlue (Invitrogen Corporation) following the manufacturer's instructions. After 1 h of incubation (10% PrestoBlue) at 37°C, 5% CO2 in a humidified atmosphere, 100 μl of the medium were carefully transferred into a 96-well plate and fluorescence was measured in triplets at wavelength of 590 nm (excitation 540 nm).

| Growth factor determination by enzymelinked immunosorbent assay
After seven days of stimulus exposure, the cell supernatant of each well was collected to determine the concentrations of vascular endothelial growth factor (VEGF), the insulin-like growth factor-1 (IGF-1), epidermal growth factor (EGF), hepatocyte growth factor (HGF), and the transforming growth factor-β1 (TGF-β1) by enzymelinked immunosorbent assay ELISA Duo-Sets (R&D Systems) following the manufacturer's instructions. Extinction was measured in duplicates per well.

| OilRed O staining
Adipogenesis-induced cells were washed in PBS, fixed for 20 min in 4% PFA and stained with OilRed O for 15 min at RT, as described earlier (Ruhl et al., 2017). To determine adipogenesis, adsorbed OilRed O was washed out with 100% isopropanol. Absorbance was measured in triplets at 540 nm.

| Chondrogenic tissue analysis and safranin staining
Chondrogenic spheres from 3D-culture were embedded in tissue freezing medium (Jung, Leica Instruments GmbH), cryosectioned (Leica) into 35 µm slices and stained with Alcian blue using the PAS staining kit (Merck), following the manufacturer's instructions.
Quantification of sphere diameters were performed using the free software ImageJ (Wayne Rasband, Institutes of Health). The chondrogenic monolayers from 2D-culture experiments were stained with safranin O. After 30 min, safranin O was washed out in isopropanol.
The samples were transferred to an optical plate and absorbance was quantified at 540 nm.

| Analyses and statistics
Data of experiments were grouped and analyzed for normality using the Kolmogorov-Smirnov test. Normally distributed data were expressed as means þ standard error of the mean (SEM),

| AEA increased and 2-AG decreased metabolic activity
AEA at 1 µM (optical density [OD] mean ¼ 0.54 � 0,014) and 3 µM (OD mean ¼ 0.53 � 0,015) had no effect, while cell numbers were significantly reduced at concentrations ≥10 µM (OD mean ¼ 0.47 � 0,017), when compared to the Veh (OD mean ¼ 0.55 � 0,016; Figure 1A). On the other hand, the metabolic activity measures of ASCs correlated inversely relative to increasing AEA concentrations. AEA ≥10 µM (OD mean ¼ 8309.9 � 280.7) significantly increased the PrestoBlue conversion compared to Veh (OD mean ¼ 6829.1 � 396.3; Figure 1A). The impact of 2-AG was comparable but not identical to the AEA effects ( Figure 1B). 2-AG dose-dependently decreased the number of ASCs, with a significant effect at concentrations ≥10 µM (OD mean ¼ 0.525 In contrast to AEA, exposure to 2-AG did not increase metabolic activity of the cells but decreased Figure 1B). Although the highest levels of metabolic activity were measured at ≥10 μM for AEA, these concentrations significantly reduced the cell numbers. A comparable effect was found for 2-AG. Therefore, both compounds were applied at 3 µm for investigating cell differentiation.

| AEA and 2-AG increased the release of growth factors
The level of EGF was below the detection limit (data not shown).

| AEA promoted adipogenic differentiation
For examination of adipogenesis, ASCs were cultured in proliferation medium ( Figure 3A) or in adipogenic differentiation medium containing insulin ( Figure 3B). Basal levels of OilRed O staining were higher when the cells were cultured in differentiation medium (OD mean ¼ 0.075 � 0.003) than in proliferation medium (OD mean ¼ 0.036 � 0.004). In both approaches, exposure to 3 µM AEA significantly increased lipid staining (U ¼ 4.6, N Veh ¼ 42, about 100%, while 2-AG did not produce an effect ( Figure 3A-B).

| Endocannabinoids regulate viability of ASCs
Since the detection of the cannabinoid receptors on the plasma membranes of ASCs, their influence on cellular vitality has been -93 investigated using exocannabinoids. Most of these studies were performed with rodent cells and showed that blocking the CB1 receptor inhibits proliferation (Bellocchio, Cervino, Vicennati, Pasquali, & Pagotto, 2008), which could not be observed in human ASCs (Ruhl et al., 2020). By contrast, we found that both AEA and 2-AG decreased cell numbers at increasing concentrations. Antiproliferative effects and cell death mechanisms through cannabinoids have been observed before, but mostly in different cancer cells (Calvaruso, Pellerito, Notaro, & Giuliano, 2012). Since the cell count has been performed only once at the end of our experiment, the rate of the cell growth cannot be seen. This would have been important in order to differentiate between a cytotoxic and an anti-proliferative effect of endocannabinoid exposure. Thus, an additional, independent method to exclude cytotoxicity is recommended, for example, G6PDor LDH-release assay from the supernatants.
We found that specific activation of CB1 decreases, whereas ligation to CB2 increases cell numbers (Ruhl et al., 2020). Since both endocannabinoids show affinity to CB1 as well as to CB2, it is possible that at increasing concentrations their activity at CB1 superimposed their CB2 effects, which in turn decreased cell numbers. Furthermore, although both endocannabinoids inhibited ASC growth, AEA increased but 2-AG decreased the metabolic activity. This surprises since both ligands bind to CB1, and it has been described that CB1 activation increases intracytoplasmic cAMP as a molecular signal involved in regulation of energy homeostasis (Ravnskjaer, Madiraju, & Montminy, 2016). Thus, the dissimilar effects of both receptor ligands could be based on varying binding affinities to CB1 but also to non-CB receptors. AEA is considered a high affinity agonist with high CB1 selectivity,

-
lipolysis. The capacity of adipocytes to recycle glycerol is limited, but glycerol is a main substrate for hepatic gluconeogenesis (Rotondo et al., 2017). Fatty acid amide hydrolase also catalyzes the hydrolysis of 2-AG, but it is mainly responsible for the cleavage of AEA into arachidonic acid and ethanolamine (Maccarrone, 2017). The latter one can be converted into acetyl-CoA, acetate and ATP, which can be used in a variety of metabolic processes and would account for our findings on the increased PrestoBlue conversion (Zhou et al., 2017).

| Endocannabinoids modulate cytokine release of ASCs
Activation of CB1 and CB2 by specific receptor ligands induces the release of growth factors on a comparable level as found for endocannabinoids in the present study (Ruhl et al., 2020). HGF regulates cell growth and mobilization of several cell types, including epithelial and endothelial cells, confirming its contribution to epithelial repair and neovascularization (Conway, Price, Harding, & Jiang, 2006). Imbalances in activation and deactivation of the HGF pathway are important pathogenetic factors in chronic wounds (Behm, Babilas, Landthaler, & Schreml, 2012). Therefore, the topical application of HGF or its increase through ECS activation on secreting cells could be a potential therapy in wound healing. IGF-1 is wound induced in animals and humans (Gartner, Benson, & Caldwell, 1992). IGF-1 promotes wound healing by multiple mechanisms, while its levels are decreased in wounds with low regenerative potential, such as diabetic wounds (Blakytny & Jude, 2006). There are many studies indicating that the addition of exogenous IGF-1 accelerates wound healing in rodent animal models, principally by dampening the local inflammatory response and promoting re-epithelialization (Emmerson et al., 2012). Thus, an increased release of IGF-1 by ECS stimulated ASCs could be of benefit for cell driven therapy in wound healing. Admittedly, we tested the highest levels of endocannabinoid exposure, which did not reduce cell numbers. It is very unlikely that these concentra-

| Endocannabinoids increase differentiation capabilities of ASCs
Exposure to the exocannabinoids WIN55,212-2 and Δ 9 -THC as well as to AEA increases adipogenic differentiation in murine cells (Bellocchio et al., 2008;Teixeira et al., 2010). The amplification in adipogenesis is accompanied by an up-regulation of the endocannabinoid receptor CB1 and of PPAR-γ, which indicates the associated receptor participation (Karaliota, Siafaka-Kapadai, Gontinou, Psarra, & Mavri-Vavayanni, 2009). In a recent study on human ASCs, we confirmed these earlier findings, namely that ligation to the CB1 induces adipogenic differentiation (Ruhl et al., 2020). Another study on human MSCs reported that AEA promotes adipogenesis, while 2-AG has no such effect (Ahn et al., 2015). This is in accordance with our findings on ASCs. AEA induced adipogenesis when the cells were kept in proliferation medium, and it promoted adipogenic differentiation at co-exposure with insulin as initiator of adipogenesis. Cannabinoids increase glucose uptake via CB1 in ASCs, which introduces them as insulin-mimetic substances (Gasperi et al., 2007). Since AEA is considered a high affinity agonist with strong CB1 selectivity, whereas 2-AG is a moderate affinity CB1 and CB2 agonist (Di Marzo & De Petrocellis, 2012), these binding characteristics could explain the varying findings between both ligands.
All key components of the ECS were found in bone tissue, and AEA and 2-AG are present in this tissue at levels similar to those found in the brain (Bab, Ofek, Tam, Rehnelt, & Zimmer, 2008). This indicates a fundamental role of the ECS in bone formation. In the present study, AEA and 2-AG increased the level of calcium deposition. We could recently show that activating the CB2 receptor increases osteogenic differentiation in ASCs (Ruhl et al., 2020), which represents a target for both endocannabinoids. However, only AEA increased the enzyme activity of ALP, which is important for the mineralization of bone cells and tissue. This suggests that 2-AG affected osteogenesis differently from AEA. The role of the ECS in chondrogenesis has been less well investigated. The treatment of rat MSCs with Δ 9 -THC enhances chondrogenic differentiation by increasing the expression of collagen and proteoglycan (Gowran et al., 2010). In ASCs, the exocannabinoid WIN55, 212-2 increases chondrogenic differentiation (Ruhl et al., 2020), but endocannabinoids did not exert an effect in the present study. It is possible that the absence of any effect could reflect ligand interactions with non-CB receptors, which inhibited their chondrogenic activity.

| CONCLUSION
ASCs represent a cell population that holds great promise for regenerative medicine. Although there is a deficit in number of randomized controlled trials and quantitative analysis supporting the efficacy of ASCs, their therapeutic effects in regenerative medicine are promising. Furthermore, these cells express an ECS, which is ascribed significance in tissue regeneration (Wang et al., 2016).
The present study suggests the ECS as a novel candidate for enhancing the regenerative capacity of ASCs, including for applications in wound repair. Although more precise assays are required to make conclusive statements on their value on wound healing, for example, ASC migration in scratch assay or co-culture with keratinocytes. AEA and 2-AG increased the release of growth factors and promoted their differentiation capabilities. Translated into the clinical situation, a chronic low dose of medical cannabis applied to a patient prior fat grafting may be a possible way to modulate ASCs for regenerative approaches (Figure 6, green arrows). Alternatively, in vitro cultured ASCs may be stimulated by cannabinoids and injected in a second procedure ( Figure 6, red arrow). Approved cannabinoid substances are Sativex, the medicinal THC that is an under-the-tongue spray, as well as its synthetic forms, that is, nabilone, dronabinol, marinol, and Syndros. By interacting with the ECS, they are effective in pain patients who are resistant to other pharmacological interventions, like in rheumatoid arthritis pain, cancer pain and also in central and peripheral neuropathic (Russo, 2008;Tanasescu & Constantinescu, 2013). Furthermore, they are used in the treatment of chemotherapy induced nausea and vomiting (Gerra et al., 2010). A possible effect of these substances on ASCs in vivo has not been experimentally investigated. Therefore, the next experimental step would be an animal study to investigate the effect of continuous ECS stimulation on ASC parameters after isolation.

ACKNOWLEDGMENTS
The authors thank Andrea Dresen for helpful support during technical performances. The authors received no specific funding for this work.
Open access funding enabled and organized by Projekt DEAL.

F I G U R E 6
Theoretical design of ECS stimulated ASC harvest for cell based therapy on injury (green arrows), or cannabinoid conditioning of cultured ASCs for engineered tissue reconstruction (red arrows) ASC, Adipose stem cell; ECS, endocannabinoid system [Colour figure can be viewed at wileyonlinelibrary.com]