LJ\~y~!{ ]~y~~~~~~~,

healing: -F~~o^ o~~~~i /

!

heterogeneity,y~~~~v~~: n~~x ~~n~~~~~~o~~

effecis~~ . i !!4r°~~~( l~~~ ~~~~(~~!( l~y~~~r~rl ~~ . ~°~^ .'~~

^~~ ~~~~~}y~~~~~ molecular ~='~° . . .~~~~ ...~~ . .y~" .=° .~~ . =. .

" .~^~~

"_~.

~~~~ f~~~~ ~~~~~y~'r ~~!^~~full-thickness

~~~'~~ skin ~A~~~!~~~~~~`^~ . . the .~~~~~=U

mU~°!~~

~ . " ""~.~= . ^~~~

FRANK AR&OLD,8M, FRCS-, CHIYU JIA, MD-, CHUFA HE, MD-; GEORGE W

DPHIL(Oxor)", BIRGITCARBOW, fhDb,, WOLJGANG ME}ER-#NGOLD, FhDb: DAJ0BADJFR,FOX; DAVID C. WEST, PhDd

The exfnzce!!ulan matrix macromolecule, hyaluronan, is thought to modulate wound healing . However, themolecular size of hyaluronan and contaminating associated proteins may be important determinants of theseeffects, We have examined the results of seven daily topical treatments of full-thickness skin wounds in pigs withu!hopurehyo!umnonofdefined molecular size . High molecular weight hyo!umnon(> 1000 kd) enhanced, whereaslow molecular weight hyaluronan decreased, the rate of early wound contraction as compared with intermediatehyo!umnun (molecular weight = 100 kd) and saline solution controls . Fracture strength at 21 days was reduced byhighond!nfernnediuferno!eou!orvv*!ghthyo!urononbufnofby!ovvnno!ecu!orvveighfhyo!uxonun .VVoundperfuuion.measured by means of o scanning laser-Doppler technique as o noninvasive indicator of anglogenesis, showeddoprexaion by hQh and intermediate molecular weight hyo!uronon on day 3, but all forms of hyaluronan causedelevated blood flow on day 7 . The architecture of granulation tissue in this wet healing model was highly organized,bufnogross h!nfo!ugic differences were seen because of treatment Different molecular species of hyaluronan havedifferenfio!effeofoonoontrochon'angiogeneoiu .ondfheevo!ufionufwoundofrengfh .VVh*rehyo!uronaniouu*daso treatment or vehicle for wounds, !foprecise composition should be specified . (WOUND REP REG 1995;3:299-310)

Hyaluronan is u repeating dime of glucuronic acidand .It iopresent bmthe earlystages of adult wound healing but persists through-out fetal repair, and it may play u part in the "scar-less" nature of healing before birth.' Exogenous

From the Wound Healing Institute, Churchill Hospital.Oxford,o X7C Biornateriols Unit, St. Mary's College,University of London, Londonc Department ofImmunology, University of Liverpool, LiverpooldUnited Kingdom; and 8eiemoorfAG, Hombuqg,Germany.*

Presented in part at the First Joint Meeting of the WoundHealing Society and the European Tissue RepairSociety, Amsterdam, The Netherlands, August22-25, 1993 .

Reprint requests: Frank Amold, 84 FRCS, WoundHealing Institute, Deportment of Dermatology,Churchill Hospital, Heocinghnn, Oxford DX37LJ,United Kingdom.

Copyright @ 1995 hby The Wound Healing Society./067c/927195 $&00+0 3611/67501

hyaluronanhas been found to enhance indexes ofbeal-corneal,3 and tympanic mem-

brane

and skin incisons,' and toreduce adhe-sions after tendon repair6 and abdominal surgery inexperimental models . Flyaluronan was reported toincrease tympanic membrane closure' and to reducethe rate of abdominal dehiscence after surgery forobesity iuhuman beings«modiowidely used iooph-thalmic surgery . It is also used as a carrier for otherwound healing agents and in cosmetic formulations .

Most forms of byulnronao are highly heteroge-neous with respect tnsize andprotein contamination .Wehave previously shown that ul'

oamobardeeof5to 25 disaccharides (but not larger molecules of6yaloronau)ure angiugenioiubiouaoayvandspongewound models" and promote endothelial proliferation

uu Opposing effects were ob-tained with

Other workers have [bnodthat

caneither accelerate orinhibit aogiogeneaio in viru, de-

2999

300 ARNOLD ET AL .

pending on the dose of hyaluronan and its frequencyof application."-15 Degradation of endogenoushyaluronan by adding hyaluronidase alters fetal re-pair ; angiogenesis and fibroplasia are enhanced . 16 Thusthe size of hyaluronan molecules within wounds mayhave an important influence on healing. The situa-tion is further complicated by the known contamina-tion of commercially available hyaluronan with pro-teins" (which have been shown to have independentactions on wound-related cells") .

The objective of the present experiments was todetermine whetherultrapurified hyaluronan modifiesthe healing of acute full- thickness skin wounds in ananimal model and whether the molecular weight ofthe hyaluronan determines the effect . We have exam-ined wound closure rates, tensile strength, and anindex of angiogenesis derived from scanning laser-Doppler imaging. The pig was chosen for these stud-ies because the anatomy and behavior of the skin isgenerally held to closely resemble that ofhumanskin."

MATERIALSAND METHODS

Hyaluronic acid solutionsMacromolecular (molecular weight [MW] approxi-mately 1,000,000) and intermediate (MW approxi-mately 100,000) hyaluronan with a protein concen-tration ofless than 0 .2% were supplied as dried powderby Pronova Biopolymers (Drammen, Norway). Molecu-lar sizes were confirmed by means of intrinsic viscos-ity, laser light scattering, and gel permeation chro-matography (TSK PW5000 column, NaCl 0.1 mol/Lbuffered with phosphate 0.1 mol/L, pH 5 .0 ; Biorad,Hemel Hempstead, United Kingdom) . A preparationoflow MW hyaluronan (approximately 10 kd) waspro-duced from high MW hyaluronan by hyaluronidasetreatment followed by gel filtration on aSephadex G50column (Pharmacia, Milton Keynes, U.K.), calibratedby reference to sodium dodecylsulfate-polyacrylamidegel electrophoresis, as previously described (Figure 1,A and B; reference 9) . The mitogenicity of the prepa-ration for bovine aortic endothelial cells (Figure 1, C)and angiogenic activity on the chick chorio-allantoicmembrane (not shown) were confirmed .

Hyaluronan preparations were dissolved overnightat a concentration of 1 mg/ml in sterile normal salinesolution at 4° C . A control solution of physiologic saline without additive was also made . Volumes of 11ml for each pair of pigs were made up not more than48 hours before wounding and commencement oftreat-ment and used within 1 week. They were sterilized by

WOUNDREPAIRANDREGENERATIONJULY-SEPTEMBER 1995

filtration through a 0.22 ~tm filter, stored, and refrig-erated as aliquots suitable for a single day's treatment .

WoundsFemale large White pigs (initial weight 22 to 26 kg)from the stock maintained by JSR Farms Ltd(Banbury, U.K.) were used for all experiments. Ani-mals were admitted to the research facility 1 weekbefore surgery for acclimatization, penned separately,and maintained on pig chow and tap water ad libi-tum. They were anaesthetized with nitrous oxide, oxy-gen, and halothane . Postwounding analgesia was pro-vided by a single intramuscular dose ofbuprenorphine .All animals were humanely cared for in compliancewith national regulations (UK Home Office AnimalInspectorate) .

While the animal was anesthetized, the flankswere shaved and depilated with Immac (Reckitt &Coleman, Hull, U.K.), which was then carefullywashed off. This procedure was essential to ensureadherence of dressings ; the effects of healing (if any)were equal for all wounds. The positions of woundswere marked with an indelible ink pen by means of athin aluminium template which conformed to the cur-vature of the animal's flank (Figure 2, A) . Four regis-ter marks were tattooed on the skin with India ink ata diagonal distance of 6 mm from each wound corner .The skin was cleaned with aqueous chlorhexidine ; 2 x2cm square full-thickness wounds were carefully madedown to, while preserving, the deep fascia . Hemostasiswas achieved by gentle pressure for up to 30 seconds.All wounds were separated by a minimum of 4 cm.

Wounds were traced onto sterile acetate film, andthe position of register marks recorded . These imageswere quantified with the use ofa digitizing pad coupledto a computer . They were also imaged with scanninglaser-Doppler imaging (described later) before treat-ment and dressing . Groups of wounds receiving par-ticular treatments were randomized between animalsto eliminate the known effects of position on rate ofwound closure.

Treatment and dressingEach row ofthree wounds received identical treatment(200 Igg of high, intermediate, or low MW hyaluronanor no additive in 0.2 ml of normal saline solution, deliv-ered by amicropipet). Each rowwas individually treatedwith coded solutions in a blinded fashion then care-fully sealed with a transparent, vapor-permeable, sol-ute-impermeable film dressing (Cutifilm, BeiersdorfAG,Hamburg, Germany) to prevent cross-contamination .The film edges were sealed with plastic tape (Sleek ;

WOUND REPAIRAND REGENERATIONVOL . 3, NO . 3

Figure 1 Preparation of low MW hyaluronan (HA) : A, Sephadex G-50 column chromatography of hyaluronandigested by Streptomyces hyaluronidase . A 5 x 70 cm column was prepared according to the manufacturer'sinstructions and eluted with acetic acid 0.1 mol/L 20 ml fractions were collected . B, Sodium dodecylsulfate-polyacrylamide gel electrophoresis of hyaluronan digest : lanes 1 to 3 are standards of >16, 5 to 20, and 3 to 9disaccharide units, respectively . Lane4is the whole hyaluronidase digest, lane 5 is fractions 35to 65 of the Sephadexcolumn eluate . C, Effects of ultrafiltered low MW preparations of hyaluronan on proliferation of bovine aorticendothelial cells as measured by tritiated thymidine incorporation . For methods, see reference 11 .

Smith and Nephew, York, U.K.), and the flanks wereenclosed in a tubular netting jacket (Surgifix; Smithand Nephew) which was sutured into place with six 3-0 Prolene sutures (Ethicon, Inc ., Somerville, N.J.) . Nosuture was placed nearer than 8 cm from any wound.The sequence of marking, tattooing, wounding, anddressing is shown in Figure 2.

On each subsequent day of treatment, pigs wereanesthetized as before, and all dressings were care-fully removed. Skin between wounds was lightlycleaned with normal saline solution and dried . Thewound surfaces were gently syringed with saline so-lution to remove residual solution, with care beingtaken to prevent cross- contamination . Acetate woundtracings were made as before . The same treatments

ARNOLD ET AL .

301

and dressings were applied as on day 0. No analgesiawas required . Four animals were killed on day 7 . Theother four were maintained for an additional 2 weeks,with dressings changed as before but without furthertreatment, on days 11, 14, and 18 .

Laser-Doppler imagingOn days 0, 3, and 7 laser-Doppler scans of each woundand surrounding skin were performed . The movementof red blood cells backscatters Doppler-shifted lightwhen illuminated by a monochromatic (laser) source .The intensity of this signal, transduced to an electricpotential difference, is proportional to the number ofmoving red cells within the field of examination andto their velocities (i.e ., flow2°) . However, skin vascu-

302 ARNOLD ET AL.

Figure 2 Wounding and dressing of pigs. A, Marking and tattooing with template . B, Dressing ; final appearance ofpig wounds dressed as described in the Methods section .

larity in general21 and wound vascularity in particu-lar, show marked spatial heterogeneity. Conventionallaser-Doppler instruments, which examine an area ofapproximately 1 mm' are ill-suited for the analysis ofwound blood flow . We therefore used a new scanninglaser-Doppler device 22 to image wound blood flow .

This instrument uses step motors to direct a beamof coherent light over the region of interest and toharvest backscattered light. As used by us, flow withineach 1 mm2 site in a 40 x 40 mm array is sampled for10 intervals of 10 msec in sequence, corrected for back-ground noise, averaged, and stored by a 30836-basedIBMcompatible microcomputer. The software suppliedpermits a color-coded display of flow, as well as lim-ited data analysis . The scanner and a representativesequence of scans is shown in Figure 3.

WOUNDREPAIR ANDREGENERATIONJULY-SEPTEMBER 7995

Laser-Doppler scanning has not previously beenused to assess vascularity during the process of tissuerepair . However, conventional single-point laser-Doppler flow measurements correlate well with increasesin vessel counts in the present wound model.23 Fur-thermore, single-point and scanning laser-Dopplersystems give highly concordant results.24

Because it is impossible to identically align thescanner head in three dimensions with respect to thesurface of all wounds, registration of scans wasachieved by the placing of small black paper squares(approximately 7 mm) such that their wound-centralcorner corresponded with the site of the tattoo . Thesefour sites, which appear black on the printouts, wereused as reference points for subsequent analysis, whichwas carried out as follows.

WOUND REPAIR AND REGENERATIONVOL . 3, NO . 3

Figure 3 Laser-Doppler perfusion imaging technique . A, Instrumentation . B, Images of a wound on days 0, 3, and7 after surgery . (Note the picture frame appearance on day 3 (82) and wound contraction on day 7 (84 .) C,Perfusion of wound center (days 0, 3, 7) displayed as histograms . Color scale indicates percentage of maximumblood flow within field of view in six equal steps, used in B and C .

The wound center was identified as the intersec-tion ofdiagonal lines connecting reference points . A 5x 5 pixel (25 mm2) region of interest was selected andcentered on this point . The mean analog flow valuefor this area was measured in volts with the use ofadditional software supplied with the scanner. Thisvalue was corrected by subtraction ofbackground skinflow, which was calculated as the average offour sepa-

ARNOLD ET AL .

303

rate 25-pixel regions as distant from the wound ascould be chosen within the scan .

Wound histologic evaluationOn days 7 and 21 for the respective groups ofanimals,each wound and surrounding skin was excised downto the underlying fascia while the animal was anes-thetized . Portions were fixed in formalin, embedded

304 ARNOLD ET AL .

Figure 4 Woundclosure (area versus time) curves for wounds treated with different MW hyaluronan . A, Control. B,Low MW hyaluronan . C, Intermediate MW hyaluronan . D, High MW hyaluronan . E, Composite figure for woundstreated with low and high MW hyaluronan, F, Square root of area versus time for control wounds .

in paraffin, sectioned at 5 gm thickness, and stainedwith hematoxylin and eosin and Masson's trichrome .Other wound portions were preserved in paraformal-dehyde for electron microscopy and frozen. One full-length sample ofwound and surrounding normal skinfrom the middle of the wound (width 1 to 3 mm) was

WOUNDREPAIRANDREGENERATIONJULY-SEPTEMBER 1995

cut in a dumbbell shape and pinned out on cork be-fore freezing in liquid nitrogen for tensiometry .

TensiometryThe storage of frozen tissue for subsequent tensiom-etry offers significant advantages when many other

WOUND REPAIRAND REGENERATIONVOL . 3, NO . 3

parameters must also be measured. It has been usedfor investigations of cartilage and ligaments26,26 andrecently for pig skin wounds." Freezing was shownnot to cause significant changes in the biomechani-cal properties of any of these tissues . Fracture loadof the central portion ofwounds at 21 days was mea-sured with a universal tensile testing machine(Instron model 1122; Instron Ltd, High Wycombe,U.K.) at a constant cross-head speed of 5 min/min .Wounds were slowly thawed to room temperature tostandardize tissue damage caused by the ice to wa-ter transition ; their width was measured by a mi-crometer. Load extension curves were obtained upto the time of fracture . The fracture load was calcu-lated as load per millimeter ofwidth (in Newtons permillimeter).

Statistical methodsA two-tailed t-test was used for all analyses . Ap value< 0.05 was accepted as significant .

RESULTSThe application of hyaluronan in solution under anocclusive dressing corresponds closely to the "wetwound healing model" developed by Erikkson et a1 .,26which has been shown to be satisfactory for the studyof skin repair in pigs and in human beings. The pro-tective method we used was successful in preventingwound trauma and loss of the active agent, whereasthe extremely well-organized nature of the granula-

ARNOLD ET AL . 305

Figure 5 Blood flow changes between day 3andday 7 as a function oftreatment and corrected for background;values are given in perfusion units (mean±standard error of the mean). MW, Molecular weight .

tion tissue supports the use ofwet wound healing inexperimental and clinical studies . We detected nomorbidity as a result of treatment . Maceperiwound skin was seen in some animals after day14, presumably as a result of the effect of prolongedocclusion. We were unable to distinguish differencesbetween groups ofwounds subjected to different treat-ments at any stage while blinded as to their natureduring the experimental period .

Rates of wound closure

The kinetics of closure of full-thickness skin woundhave been carefully analyzed in rats . 29 We found a simi-lar pattern in our model, although the duration of thevarious phases was altered because of differences inthe size of the wounds and in the species used. Imme-diately after surgery, wound area increased, presum-ably because ofrelaxation oftension in the surround-ing skin . A plateau phase then occurred, during whichsufficient granulation tissue was generated to permitactive contraction . This phase was followed by an ex-ponential decrease in wound area because of a combi-nation of contraction and epithelialization .

The rates of wound closure (area of granulationtissue remaining exposed) for different treatments areshown in Figure 4, A to D. A composite plot of thearea ofgranulation tissue for wounds treated with highand low MW hyaluronan is shown in Figure 4, E. Ingeneral, the relaxation of skin after wound excisionproduced a modest (approximately 10%) immediateincrease in wound area. Thereafter, wounds contracted

306 ARNOLD ET AL .

Figure 6 Histologic appearance of wound sections stained with hematoxylin and eosin. A, Wound edge ;hematoxylin and eosin (bar=200gm). B, Tangential section (bar= 100gm).

in all treatments from days 0 to 2 . Over days 2 to 3,areas remained approximately constant. By day 4,small differences were seen as a function oftreatment .Wounds treated with low MW hyaluronan showed anincrease in area over days 2 to 3 of 0.16 cm2 (p < 0.05 ;two tailed t-test) . In contrast, over days 2 to 4, highMW hyaluronan produced a significant decrease inarea (0 .12 cm2 ) as compared with the mean of all othertreatments . By day 6, all hyaluronate treatments (low,

WOUNDREPAIRAND REGENERATIONJULY-SEPTEMBER 1995

intermediate, and high MW agents) showed a reduc-tion in area relative to the saline solution control of0.182 cm (p < 0 .05) .

When a plot was made ofthe square root ofwoundarea versus time, this relationship was found to beapproximately linear over days 5 to 11 . This relationship is seen in Figure 4, F, in the case of controlwounds . A similar relationship was also seen for allhyaluronan treatments (data not shown) .

WOUND REPAIR AND REGENERATIONVOL, 3, NO . 3

Figure 7 Tensile testing of treated wound tissue . Breaking strain(± standard error of the mean) of wounds on day 21 as a functionof prior treatment with hyaluronan . Control versus high molecularweight (MW) hyaluronan : p=0.05; control versus intermediate MWhyaluronan : p= 0.01 ; control versus low MW hyaluronan : p= Notsignificant; low MW versus high MW : p=0.054 ; low versus interme-diate MW : p=0.07 .

Blood flow measurementsImmediately after wounding, the vessels of the deepfascia became visible as an increase in flow withinthewound bed. On day 3, a "picture frame" appear-ance of enhanced vascularity was observed at thewound periphery. By day 7, the wound bed hadfilled completely with new granulation tissue andintense flow was seen across the whole wound . Insome wounds, a thin, barely visible layer of sloughimpeded the transmission of light, resulting in afalse negative image ofblood flow . We have recentlymade similar findings in patients with venous ul-cers .

When the wound-central blood flow (calculatedas described in the Methods section) was averagedfor all wounds receiving a particular treatment, ahigher flow was found in wounds receiving normalsaline solution or low MW hyaluronan than thosetreated with intermediate or high MW hyaluronanon day 3. By day 7, all groups of wounds treated withhyaluronan hadhigher flow values than controls (Fig-ure 5) .

The scanning laser-Doppler technique does notdistinguish between vessel growth andvasodilatationas causes of increased blood flow . Differentiation between these phenomena will require a comparison ofperfusion values with vascular morphometry. Despitethis limitation, we suggest that scanning laser-Dop-pler imaging offers a new and valuable method forexamining relative blood flow in superficial granula-tion tissue .

ARNOLD ET AL.

307

Histologic findings and tensiometryRepresentative wound histologic characteristics areshown in Figure 6. No obvious major differences wereobserved between the histologic characteristics ofwounds as a function of treatment . The appearancesof pig granulation tissue were similar to those previ-ously described.30 At the wound center, a precise ver-tical orientation of the vessels was seen, whereas ves-sels entering the wound horizontally from the edgeswere curved toward the surface . Fibroblasts were ori-ented tangentially to the epithelial plane. Previousworkers have noted a similar architectural arrange-ment.23,30

The fracture load of wound samples showed a sig-nificant correlation with treatment . We observed atrend in strength, which declined as such : control >low MW hyaluronan > intermediate MW hyaluronan> high MW hyaluronan . The difference between thefirst and last treatments was significant (p < 0.05),whereas those between controls and low MWhyaluronan groups and between intermediate or highMW hyaluronan groups were not (Figure 7) .

DISCUSSIONDaily application of ultrapure macromolecularhyaluronan altered several parameters ofhealing . Theplateau phase of closure was slightly shortened, per-fusion of the wound bed was reduced on day 3 (but noton day 7), and the breaking strength of the scar wassignificantly less . Intermediate MW hyaluronan hadsimilar effects on early blood flow and on tensilestrength but did not alter the kinetics of early con-traction . Low MW hyaluronan had no effect on earlyperfusion or eventual breaking strength but initiallydelayed contraction. All three molecular forms ofhyaluronan increased both perfusion and contractionof wounds by day 7. These effects are summarized inFigure 8 .

The presence of active hyaluronan-associated pro-teins described by Burd et al . l' cannot be completelyruled out in these experiments . However, thehyaluronan used was purified and defined for MW toa greater extent than in any previous study. We con-clude that hyaluronan has a true biologic effect ontissue repair and that its actions are dependant onmolecular size .

Interpretation of these results is complicated bythe fact that degradation ofexogenous hyaluronan mayoccur within the wound. Data are limited on the stateof hyaluronan in granulation tissue . The MW ofhyaluronan derived from human hypertrophic scars

308 ARNOLD ET AL.

Figure 8 Summary of the effects of hyaluronan (HA) . MW, Molecular weight .

is greater than that found in normal ones." In fetalhealing, the total amount of hyaluronan in woundsdeclines at a slower rate than in adult wounds.' Re-ceptor-mediated uptake ofhyaluronan and subsequentintracellular breakdown occurs in fibroblasts '32,33 butthe existence of a functional extracellular wound hy-aluronidase remains unproven . The wound environ-ment contains a spectrum of molecular forms ofhyaluronan which varies during tissue repair. The fateof hyaluronan has been explored in a rat sponge im-plant model (West et al ., unpublished data). In bothcontrol wounds and those treated with exogenous lowMWhyaluronan, the size ofhyaluronan molecules de-creased with the age of the wound. A correlation be-tween reduction in hyaluronan size and concentra-tion and the onset of angiogenesis was found in thissystem .

Although hyaluronan is the simplest of all gly-cosaminoglycans, its role in healing is poorly under-stood . Hyaluronan can "organize" water in its vicin-

WOUNDREPAIR ANDREGENERATIONJULY-SEPTEMBER 1995

ity, create preferential pathways for ion flow, and regu-late motility of proteins . 34 It binds to other woundmacromolecules and receptors on endothelial cells,wound-derived fibroblasts, and leukocytes.. ._ .. Recep-tor binding can facilitate cell adhesion and migration,modulate proliferation, or alter synthetic activity . Dif-ferent molecular sizes ofhyaluronan presumably com-pete for receptor occupancy with differential effectson granulation tissue .

These uncertainties make it difficult to explain,at a cellular level, the changes in wound closure, break-ing strength, and blood flow which we observed . Thesechanges could result directly from the actions ofhyaluronan on fibroblasts and endothelial cells, alter-ing contractility, synthesis, and cross-linking of col-lagen and angiogenesis, respectively . In addition,modulation of angiogenesis could have secondary ef-fects on fibroplasia and vice versa . Vessel growth (andits component cellular stages) is clearly modulated byhyaluronan in a size-dependant fashion. In the pres-

WOUND REPAIR AND REGENERATIONVOL . 3, NO . 3

ence of hyaluronan, fibroblasts in collagen lattices al-ter their shape and may increase matrix synthesis,"but the influence ofMW on these effects is unknown.Also, evidence exists that extracellular hyaluronanmay directly modulate collagen cross-linking . 4o

Irrespective of these considerations, hyaluronanis being developed as a treatment for chronic woundsandas avehicle for other active agents . Do thepresentexperiments support the use of such treatments? Thevalidity of generalizations from acute experimentalwounds which heal normally to chronically impairedhuman healing is extremely limited. In general, it iseasier to improve defective repair than to acceleratenormal healing.41 For example, no angiogenic agenthas yet shown any effect on thehealing ofnormal skingrafts, but low MW hyaluronan increases the rate ofvascularization of grafts impaired by repeated free-thaw cycles .42,43

Our results do not directly predict the utility ofhyaluronan as a wound healing agent or the woundtypes, if any, which are most likely to respond. However, we suggest that the purity and size composition

are important determinants of biologic effects of

hyaluronan . The precise nature of hyaluronan usedin healing experiments and clinical trials should bespecified.

ACKNOWLEDGMENTSWe are grateful for the assistance and advice of

Dr . John Hopewell, Dr . Margaret Hughes, TedHarding, Malcolm Joyce, and Michael Arnold and for

the financial support of Beiersdorf AG in this study.

REFERENCES1. Adzick NS, Longaker MT . Scarless fetal healing. Therapeutic

implications . Ann Surg 1992;215 :3-7 .2. Wiig ME, Amiel D, Vande BergJ, Kitabayashi L, Harwood FL,

Arfors KE . The early effect ofhigh molecularweight hyaluronan(hyaluronic acid) on anterior cruciate ligament healing: an ex-perimental study in rabbits . J Orthop Res 1990 ;8 :425-34 .

3. Reed DB, Mannis MJ, Hills JF, Johnson CA . Corneal epithelialhealing after penetrating keratoplasty using topical Healonversus balanced salt solution . Ophthal Surg 1987;18:52-8.

4. Hellstrom S, Bloom GD, Berghe L, Stenfors LE, Soderberg O.A comparison of hyaluronan and fibronectin in the healing oftympanic membrane perforations . Eur Arch Otorhinolaryngol1991 ;248 :2305-9.

5. Abatangelo G, Martelli M, Vecchia P. Healingofhyaluronic acidenriched wounds . J Surg Res 1983;35:410-6 .

6. Amiel D, Ishizue K, Billings E, Wiig M, Vande Berg J, AkesonWH,Gelberman R. Hyaluronan in flexor tendon repair . J HandSurgAm 1989;14:837-43 .

7. Rivas Lacarte MP, Casasin T, Pumarola F, Alonso A. An alter-native treatment for the reduction oftympanic membrane per-

ARNOLD ET AL .

309

forations : sodium hyaluronate . A double blind study. ActaOtolaryngol (Stockh) 1990;110 :110-4 .

8. Trabucchi E, Foschi D, Marazzi M, Radaelli E, Lucianetti A,Rizzitelli E, Baratti C, Mariscotti C, Malgeri C, Montorsi W.Prevention ofwound dehiscence in severely obese patientswithjejunoileal bypass : the role of hyaluronic acid . Pharma-therapeutica 1988;5 :23-39 .

9. West DC, Hampson I, Arnold F, KumarS. Angiogenesis inducedby degradation products of hyaluronic acid . Science 1985;228 :1324-6 .

10. Fan TPD, Hu DE, Smither RL, Gresham GA. Further studieson angiogenesis in a rat sponge model. In : Steiner R, et al ., edi-tors . Angiogenesis . Key principles : science, technology, medi-cine . Basle: Springer-Verlag, 1992 :308-14 .

11. West DC, KumarS. The effect ofhyaluronate and its oligosac-charides on endothelial proliferation and monolayer integrity .Exp Cell Res 1989;183:179-96 .

12. Lebel L, Gerdin B. Sodium hyaldronate increases vascular in-growth in the rabbit ear chamber. Int J Exp Pathol 1991 ;72:111-8.

13. Dvorak HIT, Harvey VS, Estrella P, Brown LF, McDonagh J,Dvorak AM. Fibrin containing gels induce angiogenesis . Impli-cations for tumor stroma generation and wound healing. LabInvest 1987 ;57:673-86 .

14 . Fournier N, Doillon CJ . Invitro angiogenesis in fibrin matricescontaining fibronectin or hyaluronic acid . Cell Biol Int Rep1992;16:1251-63 .

15 . King SR, Hickerson WL, Proctor KG . Beneficial actions of ex-ogenous hyaluronic acid onwound healing . Surgery 1991 ;109 :76-84 .

'16. Mast BA, Haynes JH, Krummel TM, Diegelmann RF, Cohen

IK . In vivo degradation of fetal wound hyaluronic acid resultsin increased fibroplasia, collagen deposition, and neovascular-ization . Plast Reconstr Surg 1992;89:503-9 .

17 . Burd DA, Siebert JW, Ehrlich HP, Garg HG. Human skin andpostburn scar hyaluronan: demonstration of the association withcollagen and other proteins . Matrix 1989;9 :322-7 .

18 . Burd DA, Greco RM, Regauer S, Longaker MT, Siebert JW,Garg HG. Hyaluronan and wound healing: a new perspective .Br J Plast Surg 1991;44:579-84 .

19 . Meyer W, Schwarz R, Neurad K. The skin of domestic mam-mals as a model for the human skin, with special reference tothe domestic pig. Curr Probl Dermatol 1978;7 :39-52 .

20 . Bonner RF, Nossal R. Principles of laser-Doppler flowmetry.In : Hepherd AP, Oberg PA, editors . Laser-Doppler flowmetry.Boston : Kluwer Academic, 1990 .

21 . Tenland T, Salerud EG, Nilsson GE, ObergPA . Spatial and tem-poral variations in human skin blood flow . Int J Microcirc ClinExp 1983 ;2 :81-90 .

22 . Wardell K, Jakobsson A, Nilsson G. Laser Doppler perfusionimaging by dynamic light scattering. IEEE Trans Biomed Eng1993 ;40:309-16 .

23. Lydon M, Hutchinson JJ, Rippon M, Johnson C, deSouza N,Scudder C, Ryan TJ, Cherry G. Dissolution ofwound coagulumand promotion of granulation tissue under Duoderm. Wounds1989 ;1 :95-106 .

24. Harrison DK, Abbott NC, Beck JS, McCollum PT . A prelimi-nary assessment oflaser-Doppler perfusion imagingin humanskin using the tuberculin reaction as a model. Physiol Meas1993 ;14:241-52 .

25. Kempson GE, Spivey CJ, Swanson SAV, Freeman MAR. Pat-terns of cartilage stiffness on normal and degenerate humanfemoral heads. J Biomech 1971 ;4 :597-609 .

26. ViidikA, Sandqvist L, Magi M. Influence ofpostmortal storageon tensile strength characteristics and histology ofrabbit liga-ments. Acta Orthopaedica Scand 1965;79 (Suppl):1-113 .

27. Baker MR. Late radiation effects in radiotherapy: changes in

O ARNOLD ET AL .

thebiomechanical properties ofnormal skin and surgically pro-duced lesions after Xirradiation measured in vivo and in vitro[thesis] . Oxford, United Kingdom: Oxford University, 1993 .

28 . Breuing K, Eriksson E, Liu P, Miller D. Healingofpartial thick-ness porcine wounds in a liquid environment. J Surg Res1991;52:50-8.

29 . McGrath MH, SimonRH . Wound geometry and the kinetics ofwound contraction . Plast Reconstr Surg 1983;72:66-72.

30 . Dyson M, Young S, Pendle L, WebsterDF, Lang S. Comparisonof the effects of moist and dry conditions on dermal repair. JInvest Dermatol 1988;91:434-9 .

31 . Ueno N, Chakrabarti B, Garg HG. Hyaluronic acid of humanskin and post-burn scar: heterogeneity in primary structure andmolecularweight . Biochem Int 1992;26:787-96 .

32 . Messadi DV, Bertolami CN . CD44 and hyaluronan expression inhuman cutaneous scarfibroblasts . Am J Path 1993;142:1041-9.

33 . Ruggiero SL, Bertolami CN, Bronson RE, Damiani PJ . Hyalu-ronidase activity ofrabbit skin wound granulation tissue fibro-blasts . JDent Res 1987 ;66:1283-7 .

34 . LaurentTC, FraserRE . Hyaluronan. FASEB J 1992 ;6:2397-404 .35 . Weigel PH, Guller GM. LeBouef RD . A model for the role of

hyaluronic acid and fibrin in the early events during the in-flammatory response and wound healing. J Theor Biol1986;119 :219-34 .

36 . West DC . Hyaluronan receptors on human endothelial cells-

WOUNDREPAIRANDREGENERATIONJULY-SEPTEMBER 1995

the effect of endothelium . In: Catravas JD, Callow AD, RyanUS, editors . Vascular endothelium : physiologic basis of clinicalproblems II . NewYork : Plenum Press, 1993 :209-10 .

37. Goldstein LA, Zhou DFH, Picker LJ, Mintry CN, Bargatze RF,Ding JF, Butcher CB . Ahuman lymphocyte homing receptor,the hermes antigen, is related to cartilage proteoglycan coreand link proteins . Cell 1989 ; 56:1063-72 .

38. Adolph VR, Bleacher JC, Dillon PW, Krummel TM . Hyaluronateuptake and CD44 activity in fetal and adult fibroblasts . J SurgRes 1993;54:328-30 .

39. Doillon CJ, Silver FH, Olson RM, Kamath CY, Berg RA . Fibro-blast and epidermal cell-type I collagen interactions : cell cul-ture and human studies . Scanning Microsc 1988 ;2 :985-92 .

40. Turley EA, Eriksson CA, Tucker RP . The retention and ultra-structural appearances ofvarious extracellular matrix moleculesincorporated into three-dimensionalhydrated collagen lattices .Exp Cell Res 1990;187:243-8 .

41 . Arnold F, West DC.Angiogenesis in wound healing. PharmacolTher 1992;52:407-22 .

42. Arnold F, West DC, KumarS. Wound healing: the effect ofmac-rophage and tumour derived angiogenesis factors on skin graftvascularisation. Br J Exp Path 1987;68:569-74 .

43. Lees VC, West DC . Angiogenesis in delayed revascularizationmodel is accelerated by low molecular weight hyaluronic acidfragments [Abstract] . WOUNDREPREG 1993 ;1 ;113 .