COMMENTARYSee related article on page 880In£ammatory Modulation andWound

Repair

Je¡rey M. Davidson and Matthew D. BreyerVanderbilt University School of Medicine and the Department of Veterans A¡airs Medical Center, Nashville,Tennessee, USA

The widespread use and development of new nonsteroidal anti-in£ammatories (NSAIDs) has been driven, at least in part, by thedesire to avoid collateral e¡ects on the gastric mucosa and othertissues. The question then arises as to whether either nonselectiveCOX (cyclooxygenase) inhibitors or selective COX inhibitorsmight have a relationship to healing in the cutaneous compart-ment. Several previous studies have addressed these concepts(Muscara et al, 2000; Baatar et al, 2002; Berenguer et al, 2002; Lau-lederkind et al, 2002; Muller-Decker et al, 2002;Wilgus et al, 2003).The availability of selective inhibitors of both key enzymes inprostanoid biosynthesis, COX-1 and COX-2, as well as the crea-tion of transgenic mouse strains lacking expression of COX-1,has allowed the authors of the paper by Kampfer et al to ask ^ ina thorough and precise fashion ^ whether and how either COX-1or COX-2 and their principal products, PGE2 and/or PGD2,impinge on the wound healing process in mice.This study places a somewhat di¡erent perspective on the

question of the role of cyclooygenases in wound repair. Previousreports have shown that COX-2 is induced in the epidermis andother cells at the wound site (Abd-El-Aleem et al, 2001; Muller-Decker et al, 1998; Futagami et al, 2002). COX-2 inhibition re-sulted in a decrease in PGE2 and resurfacing in a gastric ulcermodel (Berenguer et al, 2002), but it had little or no e¡ect in arat sponge implant (Muscara et al, 2000). Two wound healingstudies show COX-2 inhibitors a¡ect cutaneous repair (Futagamiet al, 2002;Wilgus et al, 2003), and one report shows no signi¢cante¡ect of either COX-1 or COX-2 inhibition (Muller-Decker et al,2002). However, in the present study, systemic administration ofthe selective COX-2 inhibitor, celecoxib, to mouse excisionalwounds did not a¡ect either PG synthase mRNA or PGE2/D2levels. Despite the strong induction of COX-2 at the wound site,PGE2/D2 concentrations within the wound transiently droppedin parallel with COX-1 expression and activity. Furthermore, itseems clear that the principal source of prostaglandin synthaseexpression and PGE2/D2 shifted to the wound margin whereCOX-1 levels were unperturbed. One could presume that the in-crease in COX-2 resulted in production of other (non-PGE2/D2)prostanoids, such as thromboxane A2. Regardless of the syntheticpathways a¡ected by COX-2, wound healing in the presentstudy was una¡ected by pharmacologic inhibition of COX-2 ac-tivity. This result is not in complete agreement with other inves-tigations (Futagami et al, 2002), but the models have not beenidentical.The current studies provide novel information regarding the

signi¢cance of COX-1 expression in wound healing. The authorsobserved transient down-regulation of COX-1, con¢nement ofits expression to the wound periphery, and dramatic reduction ofboth PG synthase and PGE2/D2 levels by either COX-1 selectiveor nonselective inhibitors. This inhibition was accompanied byreduced epithelialization and loss of granulation tissue. Impor-tantly, the e¡ects of COX-1 inhibition were ameliorated by directapplication of PGE2. COX-1 inhibition also brought about in-

creased levels of TNF-a, a cytokine that would be expected toimpact the repair process negatively. The prediction of the studyis that selective COX-2 inhibitors are not likely to impair cuta-neous wound healing. Based on gastric and esophageal injurystudies, however, COX-2 inhibitors can have an impact on mu-cosal resurfacing. In the present excisional model, COX-2 maypredominantly act at the level of epithelial di¡erentiation ratherthan proliferation.While COX-2 has grabbed the spotlight over the past decade,

the functional signi¢cance of the highly expressed, constitutiveisoform, COX-1, has remained unclear. For example, epithelialdevelopment in the kidney is dependent on COX-2 but not onCOX-1 (Komho¡ et al, 2000). The present ¢ndings add substan-tially to our developing understanding that COX-1 likely servescritical roles, not only for the maintenance of epithelial integritybut healing as well. Data in the GI tract suggest the importanceof both COX-1 and COX-2 in preventing gastrointestinal ulcera-tion. These studies now suggest an even more central role forCOX-1 in epidermal healing.The relationship between excess in£ammation and wound

healing is a double-edged sword: chronic in£ammation, whichis typically characterized by excess and persistent neutrophil ac-tivity, can forestall the normal healing sequence, while acute in-£ammation appears to promote a cytokine repertoire thatexaggerates several aspects of repair including epidermal hyper-plasia, excess angiogenesis, and ¢brosis. Mononuclear cells andtheir derivatives are thought to coordinate these processes, butother cell types can certainly be involved.With respect to the in-£uence of in£ammation on wound quality, a recent reportof reduced scarring in murine excisions treated with a topicalCOX-2 inhibitor presents a di¡erent perspective: Suppression ofwound in£ammation by COX-2 inhibition can reduce theextent of granulation/scar tissue without compromising tensileproperties (Wilgus et al, 2003). If these ¢ndings are to be trans-lated and borne out in human studies, it may evolve that selectiveanti-in£ammatory agents can be ¢ne-tuned to suppress the im-balances of in£ammation and consequently lead to a more idealhealing response. Based on these preceding considerations, wemay be hopeful that selective COX-2 inhibitors rather thanNSAIDs may provide improved wound healing in the postopera-tive setting.

REFERENCES

Abd-El-Aleem SA, Ferguson MW, Appleton I, Bhowmick A, McCollum CN, Ire-land GW: Expression of cyclooxygenase isoforms in normal human skin andchronic venous ulcers. J Pathol 195:616^623, 2001

Baatar D, Jones MK, Pai R, et al: Selective cyclooxygenase-2 blocker delays healingof esophageal ulcers in rats and inhibits ulceration-triggered c-Met/hepatocytegrowth factor receptor induction and extracellular signal-regulated kinase 2activation. AmJ Pathol 160:963^972, 2002

0022-202X/03/$15.00 � Copyright r 2003 by The Society for Investigative Dermatology, Inc.

xi

Berenguer B, Alarcon de la Lastra C, Moreno FJ, Martin MJ: Chronic gastric ulcerhealing in rats subjected to selective and non-selective cyclooxygenase-2 inhi-bitors. Eur J Pharmacol 442:125^135, 2002

Futagami A, Ishizaki M, FukudaY, Kawana S,Yamanaka N:Wound healing involvesinduction of cyclooxygenase-2 expression in rat skin. Lab Invest 82:1503^1513,2002

Komho¡ M,Wang JL, Cheng HF, Langenbach R, McKanna JA, Harris RC, BreyerMD: Cyclooxygenase-2-selective inhibitors impair glomerulogenesis and renalcortical development. Kidney Int 57:414^422, 2000

Laulederkind SJ,Thompson-Jaeger S, Goorha S, et al: Both constitutive and inducibleprostaglandin H synthase a¡ect dermal wound healing in mice. Lab Invest82:919^927, 2002

Muller-Decker K, Hirschner W, Marks F, Furstenberger G: The e¡ects of cyclooxy-genase isozyme inhibition on incisional wound healing in mouse skin. J InvestDermatol 119:1189^1195, 2002

Muller-Decker K, Scholz K, Neufang G, Marks F, Furstenberger G: Localization ofprostaglandin-H synthase-1 and -2 in mouse skin: Implications for cutaneousfunction. Exp Cell Res 242:84^91, 1998

Muscara MN, McKnightW, Asfaha S,Wallace JL:Wound collagen deposition in ratse¡ects of an NO-NSAID and a selective COX-2 inhibitor. Br J Pharmacol129:681^686, 2000

Wilgus TA,Vodovotz Y,Vittadini E, Clubbs EA, Oberyszyn TM: Reduction of scarformation in full-thickness wounds with topical celecoxib treatment.WoundRepair Regen 11:25^34, 2003

xii DAVIDSON & BREYER THE JOURNAL OF INVESTIGATIVE DERMATOLOGY