Ceramide kinase is required for a normal eicosanoid response and the subsequent orderly migration of fibroblasts[S]

In these studies, the role of ceramide-1-phosphate (C1P) in the wound-healing process was investigated. Specifically, fibroblasts isolated from mice with the known anabolic enzyme for C1P, ceramide kinase (CERK), ablated (CERK−/− mice) and their wild-type littermates (CERK+/+) were subjected to in vitro wound-healing assays. Simulation of mechanical trauma of a wound by scratching a monolayer of fibroblasts from CERK+/+ mice demonstrated steadily increasing levels of arachidonic acid in a time-dependent manner in stark contrast to CERK−/− fibroblasts. This observed difference was reflected in scratch-induced eicosanoid levels. Similar, but somewhat less intense, changes were observed in a more complex system utilizing skin biopsies obtained from CERK-null mice. Importantly, C1P levels increased during the early stages of human wound healing correlating with the transition from the inflammatory stage to the peak of the fibroplasia stage (e.g., proliferation and migration of fibroblasts). Finally, the loss of proper eicosanoid response translated into an abnormal migration pattern for the fibroblasts isolated from CERK−/−. As the proper migration of fibroblasts is one of the necessary steps of wound healing, these studies demonstrate a novel requirement for the CERK-derived C1P in the proper healing response of wounds.


EXPERIMENTAL PROCEDURES
Materials DMEM , RPMI, FBS, and penicillin/streptomycin (100 U/ml penicillin G sodium and 100 g/ml streptomycin sulfate) were obtained from Invitrogen Life Technologies (Carlsbad, CA). The HPLC used was a Shimadzu Prominence LC-20-AD system, and the mass spectrometer was a 4000 QTRAP® from ABSciex. Prior to mass spectrometric analysis, lipids were separated by reversephase chromatography using a Phenomenex Kinetex 2.6 m C18 100A 50 × 2.1 mm reverse-phase HPLC column (Torrance, CA). HPLC-grade methanol, HPLC-grade chloroform, and American Chemical Society-grade formic acid (EMD Chemicals) were purchased from VWR (Bridgeport, NJ).

Isolation of mouse embryonic fi broblasts
Primary mouse embryonic fi broblasts (MEFs) were isolated from 13-or 14-day pregnant wild-type (CERK +/+ ) or knockout (CERK Ϫ / Ϫ ) females in BALB/c genetic background as described previously ( 16 ). The harvested MEFs were cultured in high glucose DMEM (Invitrogen) supplemented with 20% FBS (Invitrogen) and 2% penicillin/streptomycin (BioWhittaker) at standard incubation conditions. These cells were either used at the primary stage or passaged every 3 days for 20 serial passages to obtain immortalized MEFs.

Scratch-induced mechanical trauma of fi broblasts
Primary or immortalized MEFs (2 × 10 6 ) obtained from CERKnull mice and their wild-type littermates were plated on 100 mm tissue culture plates in DMEM supplemented with 10% FBS and 2% penicillin/streptomycin. The cells were allowed to adhere overnight under standard incubation conditions. Following overnight incubation, the medium was changed from full serum to 2% serum, and a single scratch was made on the monolayer along the diameter of the plate with a 200 l pipette tip. The migration of the fi broblasts into the cleared area was monitored via video microscopy.

Steady-state AA labeling of fi broblasts
Primary or immortalized MEFs (2 × 10 6 ) obtained from CERKnull mice and their wild-type littermates were plated on 100 mm tissue culture plates in DMEM supplemented with 10% FBS and 2 % penicillin/streptomycin and labeled with 0.25 Ci/ml 3 H AA as previously described ( 17 ). The resultant monolayer obtained the next morning was rested for 2 h in medium containing 2% FBS and 2% penicillin/streptomycin. The monolayer was then subjected to a scratch with a 200 l pipette tip. Medium was collected at the indicated time points, and liberated AA was measured via scintillation.

Sphingolipid and eicosanoid analysis
For sphingolipid extraction, cells (1 × 10 6 ) were harvested using a modifi ed Bligh-Dyer protocol. Briefl y, the plates were placed on ice, and the medium was transferred to another tube and used in the quantitative and qualitative analysis of eicosanoids as detailed subsequently. For sphingolipid analysis, cells were washed twice with ice-cold PBS and harvested by scraping in 200 l of PBS followed by sonication to obtain a homogenous mixture. Lipids were extracted from the remaining cells using a modifi ed Bligh and Dyer method and analyzed as described by Wijesinghe et al. ( 18 ). Briefl y, to 200 l of the cells in PBS, 1.5 ml of 2:1 methanol-chloroform was added. The samples were "spiked" with 500 pmol of d 18:1/12:0 C1P, sphingomyelin, ceramide, and monohexosylceramide as the internal standard (Avanti). The mixture was sonicated to disperse the cell clumps and incubated for 6 h at 48°C. Following incubation, the extracts were transferred in this stage leads to a healed wound with ‫ف‬ 80% the strength of the original skin.
Fibroblasts are one of the most important groups of cells taking part in the wound-healing process. During the proliferation stage of wound healing, the fi broblasts in the surrounding tissue are stimulated to proliferate followed by migration into the wound. Fibroblasts provide the necessary extracellular matrix promoting the migration of additional cell types necessary for the completion of the wound-healing process. Thereafter, the fi broblasts alter their phenotype into myofi broblasts. Contraction of the myofi broblasts induces the wound edges to conjoin, minimizing the exposed surface area ( 1,2 ). While not as predominant as in rodents, wound contraction still accounts for up to 88% of the mechanism toward wound closure in humans ( 1 ) indicating the relevance of these cells in human wound healing. Inhibition or loss of fi broblast migration leads to these cells remaining at the margins and thus causes abnormal wound healing ( 3 ). On the other hand, if there is increased, disorganized, or random migration of fi broblasts, the deposition of collagen would be uneven and unregulated, leading to hypertrophic scar formation or healed wounds with compromised strength (4)(5)(6)(7).
Several published studies demonstrate the importance of eicosanoids, a class of infl ammatory lipids, in the migration of fi broblasts (7)(8)(9)(10)(11). Green et al. ( 8 ) demonstrated that products of 5-lipoxygenase and cyclooxygenase (COX) are important in the regulation of wound closure in NIH/3T3 cells, while Kohyama et al. ( 9 ) demonstrated that prostacyclin analogs inhibit the migration of fi broblasts. Rieger et al. ( 7 ) demonstrated that 5-and 12-hydroxyeicosatetranoic acid (HETE) cause a dose-dependent increase in the chemotaxis of fi broblasts. Su et al. ( 11 ) demonstrated that nonsteroidal anti-infl ammatory drugs result in excessive scar formation due to enhanced fi broblast migration and proliferation. These studies demonstrate an important role for eicosanoids in the migration and proliferation of fi broblasts.
Group IVA cytosolic phospholipase A 2 (cPLA 2 ␣ ) is one of the major phospholipases activated in response to mechanical insults, which leads to the production of arachidonic acid (AA). AA is metabolized enzymatically to various eicosanoids (prostaglandins, leukotrienes, thromboxanes, etc.). Ceramide-1-phosphate (C1P) is a potent and specifi c activator of cPLA 2 ␣ , and our published reports demonstrate the requirement for ceramide kinase (CERK)-derived C1P in the release of AA in response to infl ammatory agonists (12)(13)(14)(15)(16). Here, we explore the hypothesis that CERK-derived C1P regulates the migration of fi broblasts due to its regulatory role in eicosanoid synthesis. We demonstrate that genetic loss of CERK severely affects the ability of fi broblasts to synthesize eicosanoids in response to mechanical injury as well as in wound biopsies. This compromised eicosanoid cascade resulted in the migration of fi broblasts in a random pattern. The requirement of C1P for orderly fi broblast migration correlates with the observed peak production of C1P in the fi broplasia stage of human wound healing. Collectively, these fi ndings demonstrate that the presence of CERK-derived C1P is required for the proper eicosanoid response and migration of fi broblasts into a wound site. serum (Cellgro). Actin was labeled by staining with Alexa Fluor 555 phalloidin (Invitrogen) at 1:40 in PBS, and DNA was stained with 300 nM 4',6-diamidino-2-phenylindole, dilactate (Invitrogen) for 5-10 min in distilled water. Coverslips were then mounted on microslides (Fisher) using 20-30 l polyvinyl alcohol mounting medium with DABCO antifade (Fluka) and allowed to dry overnight at room temperature in the dark. Cell morphology was determined by visualizing cells using a point scanning laser confocal microscope (LSM 510 META).

Migration analysis of fi broblasts
Cells were seeded into 35 mm × 10 mm cell culture dishes (Corning) and allowed to grow until confl uency. Initiation of migration was achieved by scratching confl uent cells with a p10 pipette tip. Motility was measured on a Ziess Axiovert 200M microscope mounted with a Perkin Elmer Ultraview ERS enclosed in a heated chamber (37°C) with 5% CO 2 injection. Images were acquired every 10 min for at least 15 h. After image acquisition, Volocity software was used to determine percent wound closure at designated time points and to analyze single cell tracks over time. Cell tracking data included average velocity and meandering index (calculated for each track by measuring the displacement of the cell from origin at time of observation and dividing by the track length), and t -test was used to determine signifi cance.

Rescue of CERK ؊ / ؊ fi broblasts via exogenous addition of eicosanoids
Immortalized MEFs (2 × 10 6 ) obtained from CERK-null mice and their wild-type littermates were plated on 100 mm tissue culture plates in DMEM supplemented with 10% FBS and 2% penicillin/streptomycin. The cells were allowed to adhere overnight under standard incubation conditions. The resultant monolayer obtained the next morning was rested for 2 h in medium containing 2% FBS and 2% penicillin/streptomycin. The monolayer was then subjected to a scratch with a 200 l pipette tip, and the medium was supplemented either with ethanol (vehicle) or the relevant defi cient eicosanoids (0.9 ng/ml AA, 0.3 ng/ml PGE 2 , 0.5 ng/ml 6-keto PGF 1 ␣ , 0.44 ng/ml 5-HETE, 0.5 ng/ml 11-HETE, 1.0 ng/ml 12-HETE, and 1.1 ng/ml 15-HETE). The migration of the fi broblasts into the cleared area was monitored photographically, recorded at 3 h, and the percent closure of the scratched area was calculated with the aid of ImageJ and Microsoft Excel.

Human in vivo wound analysis
Using alcohol and Betadine, the site of implantation was sterilized and anesthetized using 3 cc lidocaine (1%) without epinephrine. Five 6.0 cm high-porosity polytetrafl uoroethelene (PTFE; Custom Profi le Extrusions, Tempe, AZ) tubes were implanted subcutaneously into the inner aspect of the upper arms of healthy volunteer subjects. Standardized placement was made by a 5.5 cm cannulation of the subcutaneous tissue in a proximal direction. Using a sterile 14 gauge trochar containing PTFE tubing, the skin was punctured, and the trochar was inserted subcutaneously arising through the skin 5.5-6.0 cm away. The trochar was then removed, and the proximal and distal ends of the PTFE tubing were sutured to the skin using a single 5.0 nylon suture. A punch biopsy (10 mm) was taken from near the surgical site, and the dissected dermal portion was stored in 10% formalin and used for comparison in the lipid analyses (day 0). The implantation sites and the punch biopsy sites were covered with antibiotic ointment and a transparent surgical dressing. On days 3, 5, 7, and 14, one PTFE tube was removed and stored in 10% formalin. Lipid analysis was carried out via HPLC ESI-MS/MS. to a new glass tube, dried down, and reconstituted in methanol (600 l) by sonicating and incubating at 48°C for 15 min. The reconstitution in methanol and incubating at 48°C is a new addition to our previously published method ( 18 ) and was incorporated into the existing method to ensure proper solubilization of the long-chain sphingolipids. The lipid extract thus obtained contained insignifi cant levels of proteins as measured by the Bradford assay (data not shown) and was used in the analysis of the sphingolipids C1P, ceramide, sphingomyelin, and monohexosylceramide. The lipids were separated using a Kinetix C18 column (50 × 2.1 mm, 2.6 µm; Phenomenex) on a Prominence HPLC system (Shimadzu) and eluted using a linear gradient (solvent A, 58:41:1 CH 3 OH/water/HCOOH 5 mm ammonium formate; solvent B, 99:1 CH 3 OH/HCOOH 5 mm ammonium formate, 20-100% B in 3.5 min and at 100% B for 4.5 min at a fl ow rate of 0.4 ml/min at 60°C). ESI-MS/MS using an AB Sciex 4000 QTRAP® instrument (Applied Biosystems, MDS Sciex) was used to detect C1P, ceramide, sphingomyelin, and monohexosylceramide under positive ionization.
Eicosanoids were analyzed from culture medium as we described previously (19)(20). Briefl y , to 4 ml of medium, 10% methanol (400 l) and glacial acetic acid (20 l) were added. The samples were spiked with internal standard mixture (100 l) containing the following deuterated eicosanoids (100 pg/ l, 10 ng total): Strata-X SPE columns (Phenomenex) were washed with methanol (2 ml) and then distilled water (2 ml ). The samples were applied to the column. Thereafter, the sample vials were rinsed with 5% methanol (2 ml), which was then used to wash the columns. Finally, the eicosanoids were eluted with isopropanol (2 ml). The eluent was dried under vacuum, and the samples were reconstituted in 50:50 ethanol-distilled water (100 l) for LC/MS/MS analysis. The lipid extract thus obtained contained insignifi cant levels of proteins as measured by the Bradford assay (data not shown), and the reconstituted eicosanoids were analyzed via HPLC ESI-MS/MS. A 30 min reversed-phase LC method utilizing a Kinetex C18 column (100 × 2.1 mm, 2.6 µm) was used to separate the eicosanoids at a fl ow rate of 200 µl/min at 50°C. The column was equilibrated with 100% solvent A [acetonitrile-water-formic acid (40:60:0.02, v/v/v)] for 5 min, and then 10 µl of sample was injected. The 100% solvent A was used for the fi rst minute of elution. Solvent B [acetonitrile-isopropanol (50:50, v/v)] was increased in a linear gradient to 25% solvent B to 3 min, to 45% until 11 min, to 60% until 13 min, to 75% until 18 min, and to 100% until 20 min. The 100% solvent B was held until 25 min, decreased to 0% in a linear gradient until 26 min, and then held until 30 min. The eicosanoids were then analyzed using a tandem quadrupole mass spectrometer (AB Sciex 4000 QTRAP®, Applied Biosystems) via multiple-reaction monitoring in negative-ion mode. Eicosanoids were monitored using analyte specifi c precursor → product multiple reaction monitoring pairs, which can be found in supplementary Table I. The mass spectrometer parameters used were as follows: curtain gas: 30; CAD: high; ion spray voltage: Ϫ 3,500 V; temperature: 500°C; gas 1: 40; gas 2: 60; declustering potential, collision energy, and cell exit potential vary per transition.

Genetic ablation of CERK results in decreased levels of C1P in fi broblasts
In order to ascertain whether CERK ablation has an effect on the intracellular C1P concentrations, we investigated the levels of C1P in MEFs in the presence and absence of mechanical trauma. As anticipated, decreased levels of the major chain length of C1P, d 18:1/16:0 , as well as that of d 18:1/24:1 ( Fig. 1 ), were observed in the CERK Ϫ / Ϫ cells, confi rming the importance of CERK in the synthesis of C1P. Additional decreases were also observed for d 18:1/22:0 C1P, although this decrease is dependent on culture conditions. Interestingly, d 18:1/14:0 and d 18:1/18:0 C1P were found to be elevated in the CERK Ϫ / Ϫ cells. Neither the wild-type nor the CERK-null fibroblasts demonstrated additional changes to the endogenous C1P content upon induction of mechanical trauma ( Fig. 1 , inset). These data demonstrate that a signifi cant portion of the d 18:1/16:0 and d 18:1/24:1 C1P is CERK derived, but the change in the total levels of C1P is negligible ( Fig. 1 , inset).

CERK knockout mouse and animal welfare assurances
Breeding pairs for the CERK +/+ and CERK Ϫ / Ϫ counterparts were obtained from Novartis Pharma as a gracious gift from Dr. Frederic Bornancin. All cells derived from these mice were genetically verifi ed for each experiment as described by Bornancin and coworkers ( 21

Institutional Review Board assurance of human studies
All human studies were carried out under the approval of the Institutional Review Board (IRB) of VCU-School of Medicine (IRB number 11087 ) and written informed consent was obtained from all participants.

Statistical analysis
Data are expressed as mean ± SD. A Student's t -test was utilized when comparing two independent groups against each other for statistical signifi cance. Where statistical comparison of four or more independent groups was required, a one-way ANOVA with Tukey's post hoc method was applied. In these instances, a Levene's test was used to confi rm equality of variances among the  ) from both CERK +/+ and CERK Ϫ / Ϫ backgrounds were seeded onto 100 mm tissue culture plates and allowed to grow to confl uency overnight under standard incubation conditions in medium containing [ 3 H] AA as described in Experimental Procedures. Following the removal of unlabeled AA, the cells were fi rst rested for 2 h in medium containing 2% serum, and the monolayers were either left unharmed as controls or subjected to scratch-induced mechanical trauma as described in Experimental Procedures. The medium was collected at the indicated time points and counted via scintillation. The data are presented as nanograms of AA released per milliliter of medium and are the average of three different experiments ± SD. A one-way ANOVA followed by a Tukey's post hoc test was used to assess statistical signifi cance of the AA released by the different groups. The analysis revealed signifi cant differences in AA release between wounded CERK +/+ and CERK Ϫ / Ϫ fi broblasts (* P < 0.05, ** P < 0.01). ( 16 ). Thus, CERK-derived C1P may play a role in cPLA 2 ␣ activation in response to mechanical trauma known to induce infl ammation. In order to investigate this possibility, we examined the AA release of MEFs in response to scratch-induced mechanical trauma. Over a 4 h period, wild-type MEFs demonstrated increasing basal AA in contrast Genetic ablation of CERK inhibits the ability of fi broblasts to release AA in response to mechanical trauma Previously, our laboratory demonstrated that C1P produced by CERK was a necessary cofactor in the A23187-and interleukin 1 ␤ -induced liberation of cPLA 2 ␣ -dependent AA discount this possibility, we compared the expression levels of cPLA 2 ␣ between wild-type and the CERK-null fi broblasts. In agreement with our recent publication ( 20 ), there was no signifi cant difference in the expression level of cPLA 2 ␣ between the two genotypes (data not shown). Thus, the genetic loss of CERK adversely affects the AA release in response to mechanical trauma without affecting the expression of cPLA 2 ␣ .
to CERK Ϫ / Ϫ MEFs ( Fig. 2 ). Upon induction of mechanical trauma by scratching the monolayer, the wild-type MEFs demonstrated a robust increase of liberated AA, while CERK Ϫ / Ϫ MEFs demonstrated a minor and insignifi cant increase in AA release ( Fig. 2 ). The possibility existed that the subdued basal and mechanical trauma-induced AA release by fi broblasts was due to a decrease in the expression of cPLA 2 ␣ . In order to

ESI-MS/MS analysis of CERK
Ϫ / Ϫ MEFs demonstrated a reduction in both basal and induced AA release compared with CERK +/+ MEFs, ( Fig. 3A ) cross-validating the two experimental approaches. Additional lipidomic investigation demonstrated that the decrease in AA release in CERK Ϫ / Ϫ MEFs translated to reduced synthesis of the prostaglandins PGE 2 and PGF 2 ␣ , and the prostacyclin metabolite 6-keto PGF 1 ␣ ( Fig. 3B-D ). Additionally, the substrate limitation was also observed to affect the lipoxygenase products 5-, 11-, 12-, and CERK is required for eicosanoid biosynthesis in response to mechanical insult AA is the precursor of all eicosanoids, and the production of AA is the initial rate-limiting step of eicosanoid synthesis. In order to investigate whether the observed decrease of the release of AA in response to mechanical trauma translated to the level of eicosanoids, we utilized a lipidomic approach using HPLC ESI-MS/MS. Congruent with our fi ndings for the release of steady-state labeled [ 3 H] AA ( Fig. 2 ), HPLC  ) from both CERK +/+ and CERK Ϫ / Ϫ backgrounds were seeded onto 100 mm tissue culture plates and allowed to grow to confl uency overnight under standard incubator conditions. Following overnight incubation, the cells were rested for 2 h in medium containing 2% serum. The monolayer was then subjected to scratch-induced mechanical trauma with a 200 l pipette tip, and the medium was supplemented either with ethanol (vehicle) or the relevant defi cient eicosanoids as described in Experimental Procedures. The migration of the fi broblasts to the cleared area was photographically recorded at 3 h postscratch. B-E: Cells were seeded into 35 mm cell culture dishes and allowed to grow to confl uency and were rested for 2 h in medium containing 2% serum. The monolayer was then subjected to scratchinduced mechanical trauma cells with a p10 pipette tip, and the medium was supplemented either with ethanol (vehicle) or the relevant defi cient eicosanoids as described in Experimental Procedures. Cell motility was tracked as described in Experimental Procedures using CERK-derived C1P plays a central role in the proper migration of fi broblasts to a wound site.

fi broblasts can be rescued via the addition of exogenous eicosanoids
Our studies have demonstrated that CERK-derived C1P is an important component in the migration of fi broblasts induced by wounding. In order to determine whether CERKderived C1P regulates fi broblast migration via eicosanoid biosynthesis, we performed "rescue" experiments. Specifically, the eicosanoids downregulated in the CERK Ϫ / Ϫ cells and skin were exogenously reintroduced into the culture medium to the same concentration experimentally determined for wild-type fi broblasts. The data demonstrate that while the HETEs did not have any effect on the rate of migration of the CERK-null fi broblasts, reintroduction of AA and the COX-derived eicosanoids, PGE 2 and 6-keto PGF 1 ␣ , induced a signifi cant decrease (i.e., "rescued") in the migration of CERK Ϫ / Ϫ MEFs at the early time points of linear migration versus migration time points approaching confl uency ( Fig. 6A , B ). To further investigate this phenomenon, wild-type and CERK-null fi broblasts subjected to surface wounding were monitored by video microscopy in the presence and absence AA, PGE 2 , and 6-keto PGF 1 ␣ . These studies again demonstrated a "rescue" of the rate of wound closure by these agents in CERK-null fi broblasts at early time points of linear migration ( Fig. 6B ). Furthermore, during this early rescue period, both the average velocity and the meandering index of the CERK Ϫ / Ϫ were completely "rescued" ( Fig. 6C-E ). Hence, the enhanced migration and dysregulated polarity of CERK-null fi broblasts is due to the decreased production of COX-2-derived eicosanoids.

C1P generation in human wounds correlates with the infl ammatory and proliferative stages of wound healing
In order to extend our fi ndings regarding the importance of C1P in fi broblast migration to wound healing in humans, we investigated changes in the levels of this bioactive lipid in a human acute wound model ( Fig. 7 ). The levels of C1P were observed to increase from day 0 through day 5 before decreasing through day 7 to day 14 ( Fig. 7 ). These data indicate a steady increase in C1P through the infl ammatory phase reaching a maximum centered on the peak of fi broplasia during the proliferative phase. Hence, there is a tight correlation between the levels of C1P and the migration stage of fi broblast, further implicating an important role for this bioactive lipid in fi broplasia during wound healing.

DISCUSSION
In this study, our laboratory identifi ed a role for CERKderived C1P in the wound-healing response of fi broblasts. Specifi cally, we found that removal of CERK-derived C1P 15-HETEs ( Fig. 3E-H ). Taken together, the data indicate that CERK ablation affects both basal and mechanical trauma-induced induction of AA release, which translates into subsequent eicosanoid synthesis. Thus, CERK-derived C1P plays an important and necessary role in the biosynthesis of eicosanoids in response to the wounding of fi broblasts.

CERK-derived C1P is necessary for the full eicosanoid response of mouse skin to mechanical injury
In order to extend our fi ndings on eicosanoid production to a more complex model, we investigated the synthesis of these lipids by the skin of CERK Ϫ / Ϫ and CERK +/+ mice in response to mechanical injury. Punch biopsies (10 mm full thickness) were obtained from the depilated dorsum of wild-type and CERK-null mice immediately after being euthanized and were transferred to culture medium. The eicosanoids liberated into the medium were quantifi ed over time (2 h and 4 h) via HPLC ESI-MS/MS. Signifi cant reduction was observed in AA, PGE 2 , 5-HETE, 11-HETE, and 15-HETE in the CERK-null mice ( Fig. 4A , B , E , F , H ) compared with the injured skin of their wild-type littermates at 4 h postwounding. In contrast, no signifi cant effects on PGF 2 ␣ , 6-keto PGF 1 ␣ , and 12-HETE were observed between the two groups ( Fig. 4C, D, G ). Thus, the synthesis of specifi c eicosanoids in response to injury of the skin of CERK-null mice was severely affected compared with that of the wild-type mice, demonstrating a predominant role for CERK-derived C1P in the initial wound response of the skin.

CERK-null fi broblasts demonstrate abnormal migration to a wound site
Multiple studies demonstrate the importance of eicosanoids in fi broblast migration and proliferation (7,9,11), an important aspect of the proliferation stage (fi broplasia stage) and the subsequent remodeling phases of wound healing ( 22 ). Our fi ndings that MEFs and skin derived from CERK Ϫ / Ϫ mice have impaired eicosanoid response to wounding raised the possibility of altered migration properties as well. Indeed, the CERK Ϫ / Ϫ MEFs demonstrated an activated fi broblast phenotype compared with CERK +/+ as shown by the presence of more stress fi bers by actin staining ( Fig. 5A , B ). In order to directly investigate the role of CERK-derived C1P in migration, a monolayer of MEFs was again subjected to scratch-induced mechanical trauma followed by video microscopic monitoring. Analysis of the individual cell tracks revealed that the CERK-null fi broblasts migrate in a highly random pattern in contrast to their wildtype counterparts ( Fig. 5C, D ) as measured by the meandering index ( Fig. 5E, F , lower right panels). Hence, CERK Ϫ / Ϫ MEFs have a signifi cant loss in migration polarity. However, the CERK-null fi broblasts demonstrated an overall increase in migration, thereby covering the wound area faster compared with CERK wild-type fi broblasts ( Fig. 5E, F , lower left panels). Regardless, these data demonstrate that  via genetic ablation led to a muted AA release and eicosanoid biosynthesis of fi broblasts in response to scratchinduced mechanical trauma. We also demonstrated that this effect resulted in changes to the rate and the pattern of migration of fi broblasts. Thus, for the fi rst time, we have found a link between the CERK-derived C1P and polarity/ migratory process of fi broblasts. Our study also strengthens the argument for a major role played by CERK-derived C1P in the mediation of important eicosanoid responses. For example, CERK-null mice were reported as mainly aphenotypic ( 23 ). Furthermore, genetic ablation of CERK was observed to not affect the pathways of cPLA 2 ␣ activation in certain cell types such as peritoneal macrophages ( 21 ).The inference of these fi ndings was that CERK is not central to eicosanoid synthesis. However, in the absence of this CERK-derived pool of C1P (d 18:1/16:0 and d 18:1/24:1 ), a developmental adaptation is observed via the upregulation of the levels of d 18:1/14:0 and d 18:1/18:0 C1P through an alternate/unknown route of synthesis as we recently reported ( 20 ). Possible alternate mechanisms for anabolism of C1P include an unidentifi ed CERK, a potential mammalian enzyme with sphingomylenase D activity that hydrolyzes choline from sphingomyelin to generate C1P, or an enzyme with acylase activity that generates C1P via acylation of sphingosine-1-phosphate. However, this adaptation is not suffi cient to overcome the negative effects of the loss of CERK on eicosanoid biosynthesis in all ex vivo cells. Specifi cally, fi broblasts have lost a majority of their ability to respond to crucial stimuli such as mechanical trauma. Hence, CERK-derived C1P appears to play a central role in this eicosanoid response, and this role cannot be completely replaced via increases in non-CERK-derived species of C1P. Furthermore, the importance of CERK-derived C1P in the cPLA 2 ␣ -mediated AA response should be evaluated for the specifi c cell type and the stimulation under investigation and should not be discounted based on the phenotype of a limited number of cell types such as macrophages.
Our data demonstrating impaired ability of fi broblasts to induce a robust eicosanoid response also have implications for the other cells involved in the wound-healing process. For example, numerous reports have demonstrated an interaction between fi broblasts and keratinocytes during wound healing ( 24 ). Indeed, fi broblast-generated PGE 2 is known to infl uence keratinocyte proliferation ( 25 ). Furthermore, when culturing keratinocytes, 3T3 fi broblasts treated with mitomycin c are sometimes used as a feeder line to provide the required growth factors including eicosanoids, indicating the relevance for the keratinocyte-fibroblast interaction ( 26 ). Thus, any disruptions in the eicosanoid generation potential of the fi broblasts will most likely affect the keratinocyte proliferation, compromising the epithelialization step of the wound-healing process. Thus, it is likely the CERK is also necessary for the induction of wound epithelialization by keratinocytes, and its loss may lead to nonhealing wounds. Indeed, during the revision of this manuscript, a study was published by Kim et al. ( 27 ) demonstrating a relationship between C1P and the migration of multipotent stromal cells and endothelial progenitor cells with similar implications for tissue regeneration. Our results are in accord with their fi ndings in regard to the C1P requirement for fi broblast migration, but differ in that the effect of C1P in fi broblasts is to generate COX-2-derived eicosanoids important for driving migration and cell polarity in an autocrine and paracrine manner. Disease states due to the disruptions in the migration of fi broblasts are not limited to wound healing. There are many other fi broproliferative disorders that are the result of excessive accumulation and infl ux of fi broblasts and subsequent collagen deposition. These include pulmonary fi brosis, systemic sclerosis, liver cirrhosis, cardiovascular diseases, progressive kidney diseases, and macular degeneration, as well as infl uencing cancer metastasis and accelerating graft rejection ( 28 ). Many of the fi broblasts isolated from fi broproliferative disorders demonstrate an increased migratory capacity ( 4 ), proliferation, and increased collagen synthesis that is associated with a diminished ability to produce PGE 2 ( 6 ). Interestingly, we also observed a decrease in C1P toward the end of the fi broplasias stage of wound healing in humans, which would coincide with a decrease in PGE 2 and induction of collagen synthesis by fi broblasts. Furthermore, the dysregulated migratory and proliferative phenotypes observed in the CERK-null fi broblasts is quite similar to the phenotype of fi broblasts isolated from the fi brotic lung diseases ( 6 ). The decrease in PGE 2 observed in CERK Ϫ / Ϫ cells is of particular importance as the phenotype of the fi broblasts obtained closely mimicked those isolated from fi broproliferative disorders, suggesting a possible role for CERK in those disorders (e.g., excessive scar and keloid formation following wounding). Indeed, this study also shows that In conclusion, we have demonstrated that the loss of CERK results in a decrease of C1P production, which results in signifi cant decreases in many of the eicosanoids produced by fi broblasts. In turn, this decrease in COX-2derived eicosanoids resulted in the enhanced but random migration of fi broblasts (loss of cell polarity). Our studies in complex wound-healing systems in humans and mice further corroborate this role for CERK-derived C1P in this paradigm. Therefore, this study has implicated the modulation of CERK, and thus C1P, in the enhancement of wound healing as well as ameliorating fi broproliferative disorders.
the migration of fi broblasts is slower but very orderly under high PGE 2 concentration. Thus, while our in vitro wound-healing assay demonstrated faster closure of the simulated wound via enhanced but random migration of CERK-null fi broblasts, in the context of fi broproliferative disorders, loss of CERK functionality may be a detrimental event. Furthermore, uneven migration of these cells can also be expected to result in uneven collagen deposition with consequent effects toward the tensile strength of the healed wound. Hence, cellular C1P content may be an important factor in determining the fi broproliferative phenotype as well as tensile strength of healed wounds, and further studies are warranted to investigate this new paradigm.
While the "rescue" of the migration abnormalities of CERK Ϫ / Ϫ with AA and PGE 2 is consistent with published data and fi ndings, a full "rescue" by 6-keto PGF 1 ␣ was quite unexpected. This eicosanoid is the stable nonenzymatic hydrolysis product of prostacyclin and until now was considered to be biologically inactive. As such, it was included in our assays as a negative control. However, repeated analysis by two different laboratories indicates a role for 6-keto PGF 1 ␣ in the migration of MEFs. Because very little is known regarding the role of eicosanoids in fi broblast migration and general wound healing, these data suggest that future studies should be undertaken to understand what specifi c role this eicosanoid metabolite plays in this process.
Abnormal migration can be due to defi ciencies in one or more factors including eicosanoid signaling. Studies carried out by White et al. ( 5 ) demonstrated that lung fi broblast migration is inhibited by PGE 2 via an E-prostanoid 2 receptor-mediated mechanism that involved increased cAMP production and subsequent phosphatase and tensin homolog (PTEN) activation via decreased tyrosine phosphorylation of PTEN. Indeed, PTEN counteracts the activities of phosphatidylinositol 3-kinase via the dephosphorylation of phosphatidylinositol (3,4,5)-trisphosphate, and the loss of PTEN activity was identifi ed as a cause of enhanced migration of murine fi broblasts ( 29 ). The rescue of the migration phenotype of CERK-null fi broblasts by exogenous addition of PGE 2 may be attributed to the reestablishment of this signaling pathway. Because AA is the direct precursor of PGE 2 , the observed reversal of the enhanced migration phenotype by AA is likely due to the conversion of AA into PGE 2 . However, this proposed mechanism does not explain the rescue by 6-keto PGF 1 ␣ . There is a possibility that this metabolite of prostacyclin is being converted to a more biologically active form, which is responsible for the observed partial rescue. Studies carried out by Wong et al. ( 30 ) have demonstrated that 6-keto PGF 1 ␣ can be converted into 6-keto prostaglandin E 1 with similar properties to prostacyclin. The receptor for prostacyclin, IP also leads to the elevation of cAMP via activation of adenylate cyclase ( 9 ) and as such can be expected to inhibit fi broblast migration as described by Kohyama et al. ( 9 ) through the same pathway. Thus, the observed abnormalities in the migration of CERK-null fi broblasts that is "rescued" by the exogenous addition of eicosanoids is hypothesized to occur via the adenylate cyclase-PTEN axis of signaling.