Inhibition of FAAH confers increased stem cell migration via PPARα.

Regenerative activity in tissues of mesenchymal origin depends on the migratory potential of mesenchymal stem cells (MSCs). The present study focused on inhibitors of the enzyme fatty acid amide hydrolase (FAAH), which catalyzes the degradation of endocannabinoids (anandamide, 2-arachidonoylglycerol) and endocannabinoid-like substances (N-oleoylethanolamine, N-palmitoylethanolamine). Boyden chamber assays, the FAAH inhibitors, URB597 and arachidonoyl serotonin (AA-5HT), were found to increase the migration of human adipose-derived MSCs. LC-MS analyses revealed increased levels of all four aforementioned FAAH substrates in MSCs incubated with either FAAH inhibitor. Following addition to MSCs, all FAAH substrates mimicked the promigratory action of FAAH inhibitors. Promigratory effects of FAAH inhibitors and substrates were causally linked to activation of p42/44 MAPKs, as well as to cytosol-to-nucleus translocation of the transcription factor, PPARα. Whereas PPARα activation by FAAH inhibitors and substrates became reversed upon inhibition of p42/44 MAPK activation, a blockade of PPARα left p42/44 MAPK phosphorylation unaltered. Collectively, these data demonstrate FAAH inhibitors and substrates to cause p42/44 MAPK phosphorylation, which subsequently activates PPARα to confer increased migration of MSCs. This novel pathway may be involved in regenerative effects of endocannabinoids whose degradation could be a target of pharmacological intervention by FAAH inhibitors.


Cell culture
Cultivation of primary human MSCs was performed as described recently ( 33 ). For isolation of human MSCs, subcutaneous adipose samples were acquired by liposuction. The donors had been informed about the use of cells from their tissue for scientifi c purposes and had given informed consent. Adipose tissue was digested in collagenase NB4 standard grade solution (6 mg/ml) with gentle agitation for 30 min at 37°C. Afterwards, the digested adipose tissue was fi ltered through a 100 m nylon mesh [BD Falcon TM cell strainer (BD Biosciences, Heidelberg, Germany)], washed in 10 ml PBS with 10% (v/v) FCS, and allowed to sediment . The liquid fraction was removed and fi ltered through a 40 m nylon mesh [BD Falcon TM cell strainer (BD Biosciences)]. Both fractions were centrifuged at 400 g for 10 min. The supernatant in both fractions was removed and the pellets were combined, resuspended in PBS containing 10% (v/v) FCS, and centrifuged at 400 g for 5 min. All centrifugation steps were performed at room temperature. Cells were subsequently cultured in DMEM containing 10% (v/v) FCS, 100 U/ml penicillin, and 100 g/ml streptomycin for 24 h. From these primary cultures, MSCs were isolated via their characteristic expression of CD34 surface antigen, using the Dynabeads ® CD34 positive isolation kit (Life Technologies GmbH) according to the manufacturer's instructions.
All experiments were conducted with cells from passage 4. MSCs were seeded at a density of 2 × 10 4  ). The only exception was migration assays (see Migration assay) with 1 × 10 5 cells seeded per insert of a 24-well format Boyden chamber.
All incubations were performed in serum-free DMEM. Test substances were dissolved in ethanol or DMSO and diluted with PBS to yield fi nal concentrations of 0.1% (v/v) ethanol (for AEA, OEA, and PEA) or 0.1% (v/v) DMSO (for AA-5HT, 2-AG, AM-251, AM-630, capsazepine, GW6471, PD98059, SB203580, and URB597), if not otherwise specifi ed. WY-14643 was dissolved in DMSO and diluted with PBS to yield a fi nal concentration of 0.5% (v/v) DMSO. PBS containing the respective concentration of ethanol and/or DMSO was used as vehicle control.

Migration assay
The effect of test substances on the transmigration of MSCs was determined using a modifi ed Boyden chamber assay according to the manufacturer's instructions (BD Biosciences, Becton Dickinson GmbH, Heidelberg, Germany) as described recently ( 33 ). In this assay, cellular motility is monitored by migration through transwell inserts with pores (8 m pore size) toward a chemoattractant. In brief, cells were seeded onto the upper side of the transwell inserts in serum-free DMEM and immediately treated with test substances or vehicles for the indicated times. DMEM (high glucose, GlutaMAX™) containing 10% (v/v) FCS (22)(23)(24)(25). As a consequence of these fi ndings, cannabinoids have gained interest as potential options for the treatment of osteodegenerative disorders such as osteoporosis ( 24,(26)(27)(28) or neurodegenerative diseases such as Parkinson's disease ( 29 ). Among a variety of cellular features, previous investigations proved cannabinoids to increase tissue healing properties via induction of cellular migration at different sites such as blood vessels ( 30 ), corneal tissue ( 31 ), and colonic tissue ( 32 ). In this context, a recent study from our group found the nonpsychoactive phytocannabinoid, cannabidiol, to enhance the migratory potential of human mesenchymal stem cells (MSCs) ( 33 ). It is noteworthy that the pluripotent MSCs contribute to healing processes of injured tissues such as myocardium ( 34,35 ), injured spots of the eye ( 36 ), and bone tissue ( 37,38 ). Thus endocannabinoids and endocannabinoid-like substances, as well as enzymes that degrade these substances, are discussed as an endogenous system that regulates the differentiation of neural tissue (39)(40)(41), modulates embryogenesis ( 42 ), and coordinates remodeling of bone tissue ( 43 ).
Regarding the molecular mechanism underlying the promigratory effect of cannabinoids on cells of mesenchymal origin, CB 2 receptor activation has been shown to increase the migration of murine osteoblast-like cells ( 44 ). Moreover, activation of p42/44 MAPK, an essential intracellular trigger of migratory behavior of stem cells ( 45,46 ), has recently been shown to confer the promigratory effect of cannabidiol on MSCs ( 33 ). However, the impact of the endocannabinoids, AEA and 2-AG, or the endocannabinoid-like substances, OEA and PEA, on the migratory action of MSCs is largely unknown.
The present study therefore investigates the impact of [3-(3-carbamoylphenyl)phenyl] N -cyclohexylcarbamate (URB597), an irreversible carbamate-based FAAH inhibitor (47)(48)(49), and N -arachidonoyl serotonin (AA-5HT), an endogenous molecule with FAAH inhibitory activity ( 50 ), as well as of the FAAH substrates, AEA, 2-AG, OEA, and PEA, on the migration of human MSCs and the mechanism underlying a potential response. Here we present evidence for a promigratory action of these substances involving initial phosphorylation of p42/44 MAPK and subsequent activation of PPAR ␣ .
Proteins were separated on a 10% SDS polyacrylamide gel. Following transfer to nitrocellulose (Carl Roth GmbH) and blocking of the membranes with 5% milk powder, blots were probed with specifi c antibodies raised to ␤ -actin (loading control; Sigma), p42/44 MAPK, or phospho-p42/44 MAPK (Cell Signaling Technology, Leiden, The Netherlands). Subsequently, membranes were probed with horseradish peroxidase-conjugated Fab-specifi c anti-mouse (for detection of ␤ -actin) or anti-rabbit IgG (for detection of p42/44 MAPK and phospho-p42/44 MAPK, both secondary antibodies from Cell Signaling Technology). The following antibody dilutions were used: 1:10,000 ( ␤ -actin), 1:1,000 (p42/44 MAPK and phospho-p42/44 MAPK). Antibody binding was visualized using chemiluminescence solution containing 1.25 mM luminol, 200 M para-cumaric acid, 0.09% (v/v) H 2 O 2 , and 0.0072% (v/v) DMSO in 100 mM Tris-HCl (pH 8.5). Densitometric analysis of band intensities was done by optical scanning and quantifying using the Quantity One 1-D analysis software (Bio-Rad Laboratories GmbH). Activation of p42/44 MAPK was calculated by normalizing densitometric values of blots exposed to phospho-specifi c antibodies to densitometric values of the bands obtained from blots exposed to the antibodies against the nonphosphorylated p42/44 MAPK. ␤ -actin was used to visualize equal protein loading.

LC-MS analyses
MSCs were seeded in Petri dishes with a 10 cm diameter and grown at 37°C in DMEM (high glucose, GlutaMAX™) supplemented with 10% (v/v) FCS, 100 U/ml penicillin, and 100 g/ml streptomycin. For one sample (vehicle or FAAH inhibitor) per donor, eight Petri dishes were used. After 24 h, the medium was replaced by fresh DMEM containing 10% (v/v) FCS and antibiotics and incubated for another 24 h. Subsequently, the cells were washed once with PBS and incubated for 3 h in serum-free DMEM. Thereafter cells were incubated in serum-free DMEM with vehicle, 10 M URB597, or 10 M AA-5TH for another 6 h. Subsequently, cells were harvested by scraping and cell pellets obtained after centrifugation (10 min, 2,000 g , 4°C) were frozen in liquid nitrogen and stored at Ϫ 80°C prior to analysis. For the determination of endocannabinoids, cell pellets were further resuspended in 1 ml of 20 mM Tris-HCl buffer (pH 6.8) spiked with 20 ng/ml of AEA-d 8 and lysed using a Sonopols U-tip sonifi er (Bandelin, Berlin, Germany) three times with a 15 × 5 s pulse at 75% power followed by a 60 s pause. The lysates were transferred to ice-cold screw-capped glass tubes. In parallel with standard solutions, samples were extracted and analyzed as described recently ( 51 ). Briefl y, extracted samples (30-60 l) were analyzed on a Waters HPLC 2695 separation module using a Multospher 120 C18 column 125 × 2 mm, 5 m particle size (CS-Chromatographie Service GmbH, Langerwehe, Germany) coupled with a guard column (20 × 2 mm, 5 m particle size). Endocannabinoids and endocannabinoid-like substances were resolved using mobile phase A (water containing 0.2% formic acid) and mobile phase B [acetonitrile/2-propanol (60:40, v/v) containing 0.2% formic acid] at a fl ow rate of 0.15 ml/min. The elution scheme was as follows: linear increase of the mobile phase B from 65% to 80% in 10 min, isocratic at 80% of phase B in 3 min and linearly to 100% phase B in the following 6 min. Finally, the system was reequilibrated at 35% phase A over 4 min. The HPLC column effl uent was introduced into a Micromass Quatro Micro TM API mass spectrometer (Waters, Milford, MA) and analyzed using electrospray ionization in the positive mode and a single ion monitoring modus: m/z 300.8 for PEA, m/z 326.8 for OEA, m/z 348.8 for AEA, m/z 379.8 for 2-AG, and m/z 356.8 for the internal standard (AEA-d 8 ). The mass spectrometer and source parameters were set up as follows: capillary voltage, 3.5 kV; cone voltage, 20 and 24 V for AEA/AEA-d 8 /2-AG and PEA/OEA, respectively; source temperature, 120°C; desolvation temperature, 350°C; fl ow was used as chemoattractant in the companion plate. Following incubation at 37°C and 5% CO 2 atmosphere for 6 h (except timecourse experiments in Fig. 1D, E ), the nonmigrated cells on the upper surface of the inserts were removed with a cotton swab. For calculation of migration, the viability of the migrated cells on the lower side of the transwell insert was determined by the colorimetric 4-[3-(4-iodophenyl)-2-(4-nitrophenyl)-2H-5-tetrazolio]-1.6-benzene disulfonate (WST-1) test (Roche Applied Science, Mannheim, Germany). WST-1 tests were likewise performed in 96-well plates to exclude the possibility that promigratory actions of test substances were unspecifi c effects caused by increased cellular viability ( Table 1 ).

Scratch wound healing assay
To visualize migration of cells, human MSCs were seeded into 24-well plates and allowed to grow for 24 h in DMEM containing 10% (v/v) FCS. A 100 l plastic pipette tip was used for gently scratching the cell monolayer to create a cell-free area. Subsequently, cells were washed extensively with PBS to remove cellular debris. Thereafter, cells were incubated with URB597, AA-5HT, or vehicle in serum-free DMEM. Wound closure was monitored after 24 h (URB597) and 12 h (AA-5HT), respectively. Images of marked regions along the wounded area were obtained using an inverted microscope.

Fluorescence microscopy
MSCs were seeded in 4-well culture slides [BD Falcon™ Cul-tureSlides (BD Biosciences)] at a density of 1 × 10 5 cells/well and grown overnight at 37°C in DMEM (high glucose, GlutaMAX™) supplemented with 10% (v/v) FCS, 100 U/ml penicillin, and 100 g/ml streptomycin. After 24 h, the medium was changed and after an additional 24 h, cells were stimulated with the indicated test substances in serum-free DMEM. Test substances were prepared as described above (see Cell culture). Following incubation, cells were washed three times and fi xed using 4% (v/v) paraformaldehyde. Fixation was performed for 1 h at room temperature or overnight at 4°C. Subsequently, cells were washed three times and blocked using a PBS solution containing 5% (v/v) FCS and 0.3% (v/v) Triton ® X-100 for 1 h at room temperature. Following another washing step, cells were probed with an anti-PPAR ␣ antibody (Abcam, Bristol, UK) that was diluted 1:500 in PBS containing 1% (v/v) FCS and 0.3% (v/v) Triton ® X-100 for 1 h at room temperature. Thereafter, cells washed with PBS were probed using an Alexa Fluor ® 488 dye-conjugated goat antimouse IgG antibody (Life Technology Corporation, Darmstadt, Germany) at a dilution 1:1,000. This PBS solution additionally contained the DNA binding fl uorescent dye bisbenzimide (4 g/ml), 1% (v/v) FCS, and 0.3% (v/v) Triton ® X-100. Antibody binding and nuclei detection were visualized by fl uorescence microscopy using an Axioscope 1.A from Zeiss (Jena, Germany). Determination of nuclear PPAR ␣ was carried out by merging nuclear regions (bisbenzimide-stained areas) and Alexa Fluor ® 488-labeled PPAR ␣ . Analysis of light intensities and quantifi cation was achieved using the Zeiss Zen Pro 2012 analysis software.
In addition, the promigratory action of both FAAH inhibitors was confi rmed using the scratch wound healing assay. Accordingly, a 24 h (URB597) or 12 h (AA-5HT) incubation of MSCs with 10 M of the respective FAAH inhibitor was associated with enhanced wound closure, as compared with vehicle-treated cells ( Fig. 1E ).
To rule out the possibility that enhanced migration of MSCs in response to URB597 and AA-5HT was an unspecifi c effect due to increased cellular viability, WST-1 tests were performed using a similar treatment protocol (see Migration assay). Interestingly, URB597-and AA-5HT-treated MSCs exhibited a tendency toward a decrease, rather than an increase, of viability, with cells exposed to URB597 even exhibiting a signifi cant loss of viability. However, the loss of viability did not fall below a value of approximately 85% of the vehicle controls (100%) ( Table 1 , fi rst two columns, left hand side).

Involvement of p42/44 MAPK in the promigratory effect of FAAH inhibitors
As demonstrated recently, p42/44 MAPK activation plays a crucial role in the signaling cascade conferring enhanced migration of MSCs ( 33 ). Based on this fi nding, URB597 and AA-5HT were tested for phoshorylation of p42/44 MAPK in Western blot analyses. Both FAAH inhibitors caused a time-dependent activation of p42/44 MAPK with URB597 yielding a signifi cant peak phosphorylation following a 2 h incubation period ( Fig. 2A ), whereas MSCs exposed to AA-5HT already showed signifi cant p42/44 MAPK activation after 30 min ( Fig. 2B ). MSCs exhibited a concentrationdependent activation of p42/44 MAPK when treated with URB597 for 2 h ( Fig. 2C ) or with AA-5HT for 1 h ( Fig. 2D ). rate of desolvation gas, 700 l/h. Dwell and delay times were 0.05 and 0.1 s, respectively. All instrument parameters for the monitored analytes were tuned by injecting standard solutions at a concentration of 100 ng/ml at 10 l/min fl ow rate by a syringe pump. The data were acquired using MassLynx software version 4.1 (Micromass Ltd., Manchester, UK). Upon quantitation, the signals obtained for each analyte were normalized to the amount of internal standard observed in the corresponding sample. No effect of additives (URB597 or AA-5HT) was observed during preparation of calibration curves for each of the standards used. Finally, an aliquot of each lysate (10 l) was used for quantifi cation of total protein using the bicinchoninic acid assay (Pierce ® , Thermo Fisher Scientifi c).

Statistics
Comparisons between groups were performed with Student's 2-tailed t -test or with one-way ANOVA plus post hoc Bonferroni or Dunnett test using GraphPad Prism 5.04 (GraphPad Software, Inc., San Diego, CA). P values of less than 0.05 were considered signifi cant.

Promigratory impact of FAAH inhibitors on MSCs
The effect of the FAAH inhibitors, URB597 and AA-5HT, on the migration of MSCs was determined using Boyden chambers. Time-course experiments revealed the promigratory effect of 10 M URB597 to yield statistical significance following a 2 h incubation period ( Fig. 1A ), while 10 M AA-5HT had already reached signifi cance after a 1 h incubation ( Fig. 1B ). Following a 6 h incubation with either FAAH inhibitor, the migration of MSCs was increased in a concentration-dependent manner ( Fig. 1C, D ).

Increased migration and p42/44 MAPK activation by exogenously added AEA, 2-AG, OEA, and PEA
Based on the detected elevation of endocannabinoids and endocannabinoid-like substances in FAAH inhibitortreated MSCs, a potential promigratory and p42/44 MAPKactivating effect of exogenously added AEA, 2-AG, OEA, and PEA was quantifi ed next using Boyden chamber assays and Western blots, respectively.
Analysis of AEA, 2-AG, OEA, and PEA tested in a range of 0.1-10 M at an incubation time of 6 h revealed a concentration-dependent promigratory impact on MSCs that was likewise abrogated in the presence of PD98059 for all test substances ( Fig. 4A-D , left panels). Increased migration in response to endocannabinoids and endocannabinoidlike substances was associated with profound activation of p42/44 MAPK following a 2 h incubation with test substances at 10 M ( Fig. 4A-D , right panels). According to the indicated densitometric analyses, p42/44 MAPK activation by test substances used at 1 M yielded means ± SEMs of 140 ± 18% (AEA), 150 ± 19% (OEA), and 138 ± 21% (PEA), as compared with the respective vehicle . Noteably, activation of p42/44 MAPK in response to 2-AG appeared as a threshold effect.
To rule out the possibility that a modulation of viability may confer an unspecifi c impact on the migratory behavior of MSCs, AEA, 2-AG, OEA, and PEA were evaluated for their infl uence on viability following a 6 h incubation period. According to Table 1 (four right columns), none of the FAAH substrates yielded a statistically signifi cant alteration of viability. Interestingly, 2-AG at 10 M elicited an increase of viability to 126.9% of control, and OEA at 10 M increased viability to 115.7% of control. However, both effects did not reach statistical signifi cance.

Involvement of cannabinoid receptors and TRPV1 in FAAH inhibitor-and endocannabinoid-induced migration
In a recent study from our group, the cannabinoid receptors, CB 1 and CB 2 , as well as TRPV1, were detected by Western blot analyses of membrane fractions from human MSCs ( 33 ). To investigate a probable involvement of these cannabinoid-activated receptors in the observed promigratory action of FAAH inhibitors and endocannabinoids, MSCs were pretreated with antagonists to CB 1 (AM-251), CB 2 (AM-630), and TRPV1 (capsazepine) for 1 h. Receptor antagonists were used at a fi nal concentration of 1 M that has previously been demonstrated to be suffi cient to modulate In an attempt to evaluate a causal link between the enhanced migration by FAAH inhibitors and their ability to induce p42/44 MAPK activation, the inhibitor of p42/44 MAPK activation, PD98059, was tested for its involvement in the promigratory impact of both FAAH inhibitors. To this end, MSCs were treated with vehicle or 10 M URB597 and AA-5HT in the presence or absence of PD98059. Inhibition of p42/44 MAPK activation by PD98059 was associated with a signifi cant reduction of migration caused by URB597 ( Fig. 2E ) and AA-5HT ( Fig. 2F ). In order to test a second member of the MAPK family for a possible involvement in the promigratory impact on MSCs, the p38 MAPK inhibitor, SB203580, was included in these experiments. However, inhibition of p38 MAPK left the promigratory effect of both FAAH inhibitors virtually unaltered ( Fig. 2E, F ).

FAAH inhibitors confer increased levels of AEA, 2-AG, OEA, and PEA in MSCs
To evaluate the inhibitory action of both FAAH inhibitors on the activity of FAAH that has been found to be expressed in MSCs from different donors as assessed by Western blot analyses (data not shown), MSCs were treated with URB597 and AA-5HT (both at 10 M) for 6 h, respectively, and subsequently analyzed for FAAH substrates by LC-MS. As shown in Fig. 3 , all of the analyzed FAAH substrates (AEA, 2-AG, OEA, and PEA) were found at higher intracellular concentrations when MSCs were treated with URB597 or AA-5HT, as compared with the respective vehicle. In experiments with cells from three different donors, AA-5HT was shown to exhibit the highest mean of AEA in MSCs, although the variation of response was most pronounced, as compared with the other evaluations. On the other hand, LC-MS analyses of cells from four different donors yielded higher levels of the endocannabinoid-like substances, OEA and PEA, in cells treated with URB597, as compared with AA-5HT. Notably, URB597 elicited only a weak increase of 2-AG, which is mainly catabolized by MAGL ( 4 ). The extraction procedure and HPLC conditions were designed to minimize undesirable isomerization of 2-AG ( 51 ). For the analysis shown in Fig.  3 , only 2-AG was considered.
Collectively, these experiments suggest that FAAH degrades both endocannabinoids and endocannabinoid-like substances in MSCs, with all analyzed lipid mediators being increased upon inhibition of FAAH by URB597 and AA-5HT, respectively. tion, the promigratory effect of both FAAH inhibitors appeared to be fully reversed in the presence of a combination of antagonists to both CB 1 and CB 2 ( Table 2 , rows 1-10). Furthermore, the TRPV1 antagonist, capsazepine, receptor activity ( 33,52 ). URB597-and AA-5HT-induced migration was signifi gantly abrogated in the presence of AM-630, suggesting the CB 2 receptor to functionally contribute to the observed effects ( Table 2 , rows 1-10 ). In addi- Here, the TRPV1 antagonist, capsazepine, was additionally tested for its impact on the promigratory effect of both fatty acid amides, as was also done with the other test substances (see Table 2 ). As shown in Fig. 5C, D , capsazepine led to an inhibition of the OEA-and PEA-induced migration of MSCs. Moreover, the concentration-dependent induction of migration by both FAAH inhibitors ( Fig. 1C, D ) and FAAH substrates ( Fig. 4A-D , left panels) was mimicked by the synthetic PPAR ␣ agonist, WY-14643 ( Fig. 5E ).
Infl uence of p42/44 MAPK inhibition on PPAR ␣ activation by URB597, AA-5HT, AEA, 2-AG, OEA, and PEA In an attempt to confi rm an activation of PPAR ␣ by FAAH inhibitors and substrates, fl uorescence microscopy using a PPAR ␣ antibody visualized by a Alexa Fluor ® 488 dye-tagged secondary antibody and a DNA binding fl uorescent dye, bisbenzimide, was performed next to detect the nuclear accumulation of PPAR ␣ . Noteably, subcellular distribution of PPAR ␣ toward nuclear accumulation has been described as a reliable marker of PPAR ␣ activation ( 53 ). To provide evidence for a causal relationship between PPAR ␣ activation and p42/44 MAPK phosphorylation, analysis of nuclear accumulation of PPAR ␣ in the presence or absence of the p42/44 MAPK upstream inhibitor, PD98059, was carried out. According to Fig. 6A , 10 M URB597 caused an increase of PPAR ␣ in the nuclei of MSCs that was significantly inhibited by PD98059, thereby confi rming an involvement of p42/44 MAPK activation in this response. As a control, the PPAR ␣ antagonist, GW6471, also conferred an inhibition of the cytosol-to-nucleus translocation of PPAR ␣ . Comparable results were found in experiments using MSCs that were treated with 10 M AA-5HT for 2 h ( Fig. 6B ).
In order to evaluate whether the FAAH substrates that are upregulated upon treatment of MSCs with URB597 and AA-5HT also cause a p42/44 MAPK-dependent PPAR ␣ activation, AEA, 2-AG, OEA, and PEA were tested using the same experimental setting. In fact, 10 M AEA ( Fig. 7A ) and 2-AG ( Fig. 7B ) induced a cytosol-to-nuclear translocation of PPAR ␣ that was likewise inhibited by upstream inhibition of p42/44 MAPK activation with PD98059 and by the PPAR ␣ antagonist. Similar results were further proven for OEA ( Fig. 7C ) and PEA ( Fig. 7D ).

Involvement of PPAR ␣ in the promigratory effect of URB597, AA-5HT, AEA, 2-AG, OEA, and PEA
To elucidate a possible contribution of PPAR ␣ to the promigratory action of URB597 and AA-5HT, the impact of the selective PPAR ␣ antagonist, GW6471, on migration of MSCs using Boyden chamber assays was analyzed next. Increases of MSC migration caused by both FAAH inhibitors were completely reversed by the PPAR ␣ antagonist, GW6471 ( Fig. 5A ). In line with this fi nding, the endocannabinoids, AEA and 2-AG, were likewise found to elicit their inductive effect on MSC migration in a GW6471-sensitive manner ( Fig. 5B ). Similar results could be proven for the endocannabinoid-like substances OEA ( Fig. 5C ) and PEA ( Fig. 5D ).  inhibitors (URB597 and AA-5HT) and FAAH substrates (i.e., the endocannabinoids, AEA and 2-AG, as well as the endocannabinoid-like substances, OEA and PEA) was completely reversed by PD98059, an inhibitor of p42/44 MAPK activation. In line with this notion, the increase of migration by either substance was accompanied by significant elevations of p42/44 MAPK phosphorylation. The inhibitory impact of FAAH inhibitors on their target enzyme was proven by LC-MS analysis that revealed increased levels of all four FAAH substrates in MSCs as a response to both FAAH inhibitors. Second, inhibitor as well as fl uorescence microscopy approaches indicate an activation of the transcription factor, PPAR ␣ , to be involved in the promigratory response of MSCs upon treatment with FAAH inhibitors and their substrates. These fi ndings were corroborated by use of the PPAR ␣ agonist, WY-14643, that mimicked a concentration-dependent stimulation of migration, as well as a nuclear accumulation of PPAR ␣ under the same experimental conditions. Third, evidence for a causal relationship between p42/44 MAPK phosphorylation and subsequent PPAR ␣ activation was provided by experiments showing PD98059 to inhibit an increase of PPAR ␣ in the nuclei of MSCs treated with URB597 and AA-5HT as well as FAAH substrates, thus confi rming an involvement of p42/44 MAPK in PPAR ␣ activation.
Our data are in agreement with reports demonstrating the promigratory effect of cannabinoids on human corneal epithelial ( 31 ) and embryonic kidney cells ( 54 ) to be mediated via p42/44 MAPK activation. Likewise, the nonpsychoactive phytocannabinoid, cannabidiol, was recently found to confer increased migration of human MSCs via a pathway involving p42/44 MAPK ( 33 ).
PPARs represent phosphoproteins whose transcriptional activity can be modulated via phosphorylation by MAPKs [for review see ( 55 )]. The inhibition of PPAR ␣ activation by blockade of the p42/44 MAPK pathway, as shown here, is corroborated by several other studies. For example, insulin was previously demonstrated to confer increased PPAR ␣ activity via prior activation of p42/44 MAPK ( 56 ) with the insulin-induced transactivation being due to phosphorylation of two serines (positions 12, 21 ) in the A/B domain of human PPAR ␣ ( 57 ).
In the present study, treatment of MSCs with the PPAR ␣ inhibitor, GW6471, did not signifi cantly interfere with p42/44 MAPK activation by FAAH inhibitors and substrates, thereby excluding phosphorylation of p42/44 MAPK by the test substances to occur downstream to PPAR ␣ activation. This control experiment was initiated on the basis of spare, but existing, data indicating PPARs to control kinase-linked receptor signaling on different levels, such as by interfering with the phosphorylation cascade or with transcription factor binding on DNA [for review see ( 58 )].
In agreement with the PPAR ␣ -mediated effects of FAAH inhibitors and substrates reported here, several other studies confi rmed PPAR ␣ -dependent effects for either group of substances. First evidence of cannabinoid interaction with PPAR ␣ was published by Kozak et al. ( 12 ), who showed that 15-hydroxyeicosatetraenoic acid glyceryl ester, a p42/44 MAPK activation. As expected, these experiments revealed the URB597-induced p42/44 MAPK activation to be insensitive to preincubation with the PPAR ␣ antagonist, GW6471 ( Fig. 8A ). AA-5HT-induced p42/44 MAPK activation in MSCs was slightly inhibited by GW6471, but did not yield statistical signifi cance ( Fig. 8B ). Finally and in agreement with this data, Western blot experiments with MSCs treated with AEA, 2-AG, OEA, and PEA revealed no impact of GW6471 on the activation of p42/44 MAPK by the four FAAH substrates ( Fig. 8C , left panel). GW6471 did not signifi cantly infl uence basal p42/44 MAPK phosphorylation in the absence of test substance ( Fig. 8C , right panel).

DISCUSSION
The present study provides fi rst-time proof for induction of migration of MSCs by FAAH inhibitors, endocannabinoids, and endocannabinoid-like substances. As the underlying pathway, a phosphorylation of p42/44 MAPK followed by activation of PPAR ␣ was identifi ed ( Fig. 9 ).
There are several lines of evidence supporting this mechanism. First, the promigratory action of both FAAH  effects of FAAH inhibitors in animal experiments, including enhancement of memory acquisition ( 20 ), antinociception ( 17,21 ), and blockade of nicotine reward and relapse ( 18 ), have been associated with activation of PPAR ␣ .
Remarkably, our experiments with the inhibitor of p42/44 MAPK activation, PD98059, revealed the phosphorylation of p42/44 MAPK as a crucial regulator of PPAR ␣ activation for all tested FAAH inhibitors and substrates . Thus, besides a direct PPAR ␣ binding, as previously shown by Sun et al. metabolite of 2-AG generated upon metabolism via 15-lipoxygenase, acts as a PPAR ␣ agonist. Thereafter, OEA was demonstrated to confer appetite suppression and weight loss ( 10 ) and lipolysis ( 11 ), as well as neuroprotection ( 15 ), through activation of PPAR ␣ . Likewise, PEA was reported to exert anti-infl ammatory and analgesic effects via PPAR ␣ ( 13 ). AEA, an endocannabinoid with PPAR ␣ binding and transcription activity ( 16 ), was recently reported to mediate relaxation of ophthalmic arteries in a PPAR ␣ -dependent manner ( 14 ). In consequence, various been shown to increase the migration of murine osteoblastlike cells ( 44 ). Moreover, the nonpsychoactive cannabinoid, cannabidiol, was recently demonstrated to mediate increased migration of MSCs via activation of the CB 2 receptor ( 32 ). In line with this notion, an inhibitor approach in the present study yielded both FAAH inhibitors and CB receptor-active endocannabinoids (AEA and 2-AG) to confer increased migration, at least in part, via activation of CB 2 . In addition, evidence was provided for a contribution of TRPV1 in increased migration by all FAAH inhibitors and FAAH substrates tested, which is in agreement with the reported TRPV1 activation by AEA ( 8 ), 2-AG ( 9 ), OEA ( 6 ), and PEA ( 7 ). Accordingly, the TRPV1 antagonist, capsazepine, blocked the promigratory actions in a partial (2-AG) and complete manner (URB597, AA-5HT, AEA, OEA, and PEA), respectively. However, interpretation of this data is hampered by the fact that capsazepine even diminished basal migration of MSCs, which is in agreement with the data published by Schmuhl et al. ( 33 ).
Another aspect that merits special consideration is that the levels of FAAH substrates were measured in cell lysates and were therefore necessarily indicated as picomoles per milligram of cellular protein. Thus, it was not possible to compare endogenous levels to the concentrations of exogenously added AEA, 2-AG, OEA, and PEA.
( 16 ) using a cis -parinaric acid-based ligand-binding system and a transient transfection system, additional mechanisms of PPAR ␣ activation appear feasible.
Apparently contradictory results have been published on the infl uence of PPAR ␣ on migration. In support of the data presented here, an increase of migration by PPAR ␣ activation was observed in human circulating angiogenic cells ( 59 ) and monocytic U937 cells ( 60 ). Interestingly, in the latter publication, the same PPAR ␣ agonist as that used in the present study (WY-14643) conferred increased migration of human monocytes, while a PPAR ␥ agonist elicited the opposite effect. However, in other systems, including human airway smooth muscle cells ( 61 ) and human umbilical vein endothelial cells ( 62 ), PPAR ␣ activators have been reported to decrease migration. These results suggest that PPAR ␣ regulates migration in a cell-and stimulus-dependent manner with mechanisms still being poorly understood. Referring to the mechanism underlying PPAR ␣ -induced migration, a recent study found a reversal of PPAR ␣ -induced migration of circulating angiogenic cells by an Akt inhibitor ( 59 ).
Regarding the receptors conferring promigratory effects of cannabinoids on cells of the mesenchymal niches, CB 2 receptor activation by the CB 2 selective agonist, HU308, has neuropathic pain, and, in the case of URB597, signs of osteoarthritic pain [for review see ( 2 )]. According to the data of the present study, endocannabinoids and endocannabinoid-like substances, whose concentrations may be pharmacologically enhanced by FAAH inhibitors, may impact the migratory action of stem cells of the mesenchymal niches in vivo, thereby contributing to tissue regeneration.
Collectively, this is the fi rst study to provide evidence for and insights into the promigratory action of inhibitors of FAAH and its substrates on human MSCs. The fi ndings may offer a novel fi eld of future in vivo investigations aimed at the regenerative effects of FAAH inhibitors which, according to animal experiments, already proved to be effective at ameliorating signs of acute, infl ammatory, visceral, and    9. Inhibition of FAAH by URB597 and AA-5HT causes increased levels of endocannabinoids and endocannabinoid-like substances in MSCs. As a consequence, these FAAH substrates induce activation of p42/44 MAPK that subsequently enhances PPAR ␣ translocation to the nucleus thereby conferring increased cellular migration .