Clin Type procollagen marker severity scarring after ... · Type 1 procollagenas amarkerofseverity of scarring after sternotomy: effects oftopical corticosteroids YRiaz, HTCook,AWangoo,BGlenville,

8J Clin Pathol 1994;47:892-899

Type 1 procollagen as a marker of severity ofscarring after sternotomy: effects of topicalcorticosteroids

Y Riaz, H T Cook, A Wangoo, B Glenville, R J Shaw

AbstractAims-To determine whether the abun-dance of newly formed collagen in heal-ing surgical wounds correlated with scarseverity, and whether topical applicationof steroid cream reduced new collagenformation in patients who have under-gone median sternotomy.Methods-Thirty three patients six weeksafter sternotomy, and 12 controls werestudied. Scars were photographed, andbiopsy specimens from scars at sitestreated or untreated with topical corti-costeroids (clobetasol proprionate 0.5%)were examined using immunohisto-chemical staining for type 1 procollagen(PCP 1) and transforming growth factorIi (TGF-/J), and in situ hybridisation fortype 1 procollagen messenger RNA(mRNA).Results-The degree of hypertrophy ofthe scar and the abundance of PCP 1immunostaining were ranked indepen-dently, blind, and a correlation betweenthese two variables was observed (r =0-604, p < 0.001). The PCP 1 immuno-staining was accompanied by a greatabundance of PCP 1 mRNA and only aslight increase in TGF-,/ immunostain-ing, when compared with normal skin ormature scars. Following the applicationof topical corticosteroids, for either 48hours or twice daily for seven days, therewas no reduction in PCP 1 immunostain-ing nor the abundance ofPCP 1 mRNA.Conclusions-These data suggest thatthe extent of new collagen formation asassessed by PCP 1 immunohisto-chemistry may be a useful marker ofthe exuberance of the scarring processfollowing sternotomy, and that topicalcorticosteroids are ineffective in reducingthis component of the fibrotic response.

(i Clin Pathol 1994;47:892-899)

After surgery, optimal healing combines rapiddevelopment of a scar with high tensilestrength but this is minimally disfiguring.Hypertrophic, or in extreme cases keloidscars, are an overexuberent response to skininjury. This often occurs after sternotomy,"2causing discomfort, pruritis, and occasionallyfurther plastic surgery.2 A variety of differentapproaches have been used in an attempt toprevent the development of hypertrophicscars, but these have rarely been linked to

individual components of the pathophysiologyof hypertrophic scar formation. The develop-ment of a scar involves the influx of inflamma-tory cells into the wound, and the productionof growth factors for fibroblasts, such as trans-forming growth factor ,B (TGF-fl). In responseto these, fibroblasts migrate into the woundand synthesise collagen.45 Initially, type IIIcollagen is the principal collagen synthesised,but this is gradually replaced by type Icollagen.36 Later, collagen remodelling andcross-linking take place.7 This process isexaggerated in hypertrophic scars wherethere is evidence of excessive collagen de-position.458-10 Type I collagen is a triple helixcomposed of two a, and one a2 chains."Because type I procollagen carboxyl andamino terminal domains are proteolyticallyremoved during collagen secretion,45 antibod-ies to these domains have been used to stainfibroblasts synthesising type I collagen.'2 Theuse of immunocytochemistry with antibodiesto the amino or carboxyl terminal ends ofeach a chain of procollagen peptide 1 (PCP 1)and in situ hybridisation has allowed individ-ual fibroblasts synthesising new collagen to beidentified.

In this study, we examined sternal scars sixweeks after sternotomy to seek evidence ofnew collagen formation at this stage of estab-lished scar formation, and to determine if theextent of new collagen formation correlateswith the macroscopic appearance of the scar.We asked if topical corticosteroids, agentslong thought to prevent progression of keloidscars,"3 and now recognised to reduce collagenproduction in vitro,'4 have the ability toreduce the biochemical markers of new colla-gen production in these healing surgicalwounds. Furthermore, the abundance of thegrowth factor TGF-fl in the scars was alsomeasured.

MethodsThirty three patients were recruited four toeight weeks after cardiothoracic surgeryinvolving a stemotomy. A further eightpatients were recruited before surgery, to pro-vide samples of healthy skin. Four patientswho had had surgery over a year ago were alsorecruited to provide samples of mature scartissue. Informed written consent wasobtained. Approval for this study was grantedby the Parkside Health Authority EthicsCommittee.At the initial visit, the stemotomy scars

were examined to exclude any patients who

Department ofRespiratory Medicine,St Mary's HospitalMedical School,LondonY RiazA WangooR J ShawDepartment ofCardiothoracicSurgeryB GlenvilleDepartment ofHistopathologyH T CookCorrespondence to:Dr R J Shaw, Departnent ofRespiratory Medicine, StMary's Hospital MedicalSchool, Norfolk Place,London W2 1PGAccepted for publication28 February 1994

892

on March 21, 2020 by guest. P

rotected by copyright.http://jcp.bm

j.com/

J Clin P

athol: first published as 10.1136/jcp.47.10.892 on 1 October 1994. D

ownloaded from

http://jcp.bmj.com/

Type 1 procollagen as a marker of severity ofscaring after sternotomy: effects of topical corticosteroids

had had infection or dehiscence of the wound.A photograph was taken of the lower 5 cm ofthe scar with a Nikon FM camera with a 105mm Micro-Nikkor lens at 07-0 55 magnifi-cation, on Kodak Ektar 25 ASA colour film.Lighting was enhanced using a Sunpack Auto28SR Thyristor flash gun.An area 3 cm x 3 cm was marked out at

the lower end of the sternotomy scar, using apermanent marker pen. A small amountof Dermovate (clobetasol proprionate005% w/w) was gently massaged into the scarwithin this area. Eight patients had a singleapplication of steroid cream, covered with aTegaderm dressing, and were asked to return48 hours later. The other 25 patients wereasked to apply the cream, within the markedarea only, twice a day for seven days.Two samples of scar tissue were obtained

from each patient. The first was from withinthe marked area (where the steroid basedcream had been applied) and the second sam-ple was from an adjacent area of the scarwhere cream had not been applied. Thebiopsy specimens were taken under localanaesthesia (1 ml of 1% Lignocaine, in adren-aline 1 in 200 000, injected subcutaneously)using a 3 mm biopsy punch (StiefelLaboratories (UK) Ltd., Wooburn Green,Bucks, England). The steroid and non-steroidscar samples were fixed separately in formol-saline, embedded in paraffin wax, and cut intosections that were mounted onto poly-L-lysine slides. Routine haematoxylin and eosinstaining was performed to assess cellularityand the extent of scar formation.

IMMUNOHISTOCHEMISTRYSections 3-4,um thick were cut from eachbiopsy specimen, dewaxed, and treated with3% hydrogen peroxide in methanol to quenchthe activity of endogenous peroxidase. Thesections were rehydrated in phosphatebuffered saline (PBS) (pH 7-2) and weredigested with trypsin (0-5% trypsin and 0 5%chymotrypsin; Sigma) for 10 or 20 minutesbefore immunostaining for type 1 procollagenand TGF-f,, respectively. Non-specific pro-tein binding was blocked with 1 in 5 normalporcine serum (Dako Ltd., High Wycombe,Bucks) in PBS. Primary antibodies used inthis study were polyclonal rabbit anti-humanprocollagen -1 (Gift from M J Warburton, StGeorge's Hospital Medical School, London)"5and TGF-fl (British Biotechnology ProductsLtd., Oxon, England). The antibodies werediluted in 1 in 20 porcine serum in PBS. Thesections were incubated with the optimal dilu-tion of the primary antibodies (1 in 400 fortype 1 procollagen and 1 in 100 for TGF-/3 inPBS) for one hour in a humidifying chamberat room temperature. The sections werewashed with PBS and incubated with biotiny-lated goat anti-rabbit antibodies (Dako Ltd.)for 30 minutes, washed, and then incubatedwith avidin biotin peroxidase complex (DakoLtd.), washed for 15 minutes, and visualisedwith 3,3'-diaminobenzidine (DAB) substrate(Dako Ltd.). The sections were counter-stained with Mayer's haematoxylin solution

(Sigma Diagnostics, St Louis, Missouri,USA) for one minute. In control sections,specificity was ascertained when the primaryantibody step was omitted or replaced withpurified normal rabbit serum (Dako Ltd.).

IN SITU HYBRIDISATIONThe procedure for in situ hybridisation wasmodified from methods described earlier.'6Briefly, 3-4,um sections were cut from skinbiopsy specimens embedded in paraffin waxand were mounted on Vector bond (VectorLaboratories, Burlingame, California, USA)coated slides and baked overnight at 56°C.Tissue sections were dewaxed and rehydratedthrough an alcohol series to PBS. Sectionswere treated with proteinase K (BoehringerMannheim UK) 50 ug/ml for seven and a halfminutes, washed in PBS, followed by acetyla-tion with 0 1% acetic anhydride in 0 1 Mtriethanolamine to reduce non-specific inter-action between DNA probes and tissue.Slides were washed in PBS, dehydrated, andair dried and used for hybridisation on thesame day or stored at - 200C.

For type 1 procollagen, two 27 base syn-thetic DNA oligonucleotide anti-sensesequences were made to order (BritishBiotechnology UK) to be used as probes. Thesequences were (1) 5'-ACC AGT CTC ACCACG ATC ACC ACT CTT-3' [-366 to -392](2) 5'-TCC CTT AGC ACC AGT GTCTCC TTTT GCT-3' [-2229 to -2255] andwere deduced from the published DNAsequence of the human pro a (1) chain ofhuman type 1 procollagen.'7 A cocktail ofboth oligonucleotides was 3'-end labelledwith deoxyadenosine 5'-(a-_5S) thiotriphos(Amersham International plc., Bucks,England). The reaction consisted of 10 pM ofeach oligonucleotides, 8,ul of 5 x buffer(Promega Corporation, Madison, USA), 4plof terminal deoxynucleotidyl transferaseenzyme (Amersham UK), 5 pl of 35S (dATP)and 18 pl of deionised water to make a total of40 pl. The reaction was incubated at 37°C forone hour and 3' end labelled probes werepurified using NENSORB Nucleic acidpurification cartridge (NEN Du Pont (UK)Ltd) according to the manufacturer's recom-mendations. 18

Tissue sections were overlaid with 100,plof hybridisation buffer which consisted of a50% solution of formamide (Sigma), 1 xDenhardt's solution, 1 mMIl EDTA, 100pg/ml denatured herring sperm DNA, 250pg/ml yeast tRNA, 10 mM phosphate buffer,10% dextran sulphate and 4 x sodium chlo-ride/sodium citrate (SSC). Before use,hybridisation buffer was heated to boiling forfive minutes and quenched on ice, and 250pg/ml polyadenylic acid and 10 mM/l dithio-threitol was added. The 35S labelled oligonu-cleotide probes were diluted in hybridisationbuffer to give between 20-30 x 106 countsper ml. Hybridisation was performedovernight at room temperature. After incuba-tion, the sections were washed with severalchanges of 1 x SSC at 54°C for one hour.After washing, slides were air dried and

893



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Riaz, Cook, Wangoo, Glenville, Shaw

Figure 1 Variation in theseverity ofsternotomy scarssix weeks followingsurgery: (A) thin line scar;(B) mildly hypertrophicscar; (C) severelyhypertrophic scar.

4-4....

Ji

A C

dipped in Ilford K5 nuclear track emulsion

(Ilford Ltd. Mobberley, Cheshire, England).

After an exposure period of two weeks, the

sections were developed in Phenisol developer

(Ilford Ltd.), fixed in 0-3 mom/ sodium thio-

sulphate (BDH Ltd., Poole, Dorset,

England), washed in deionised water, coun-

terstained in Mayer's haematoxylin (Sigma)

and mounted.

To serve as controls, some tissue sections

were treated before hybridisation with 20

,ug/ml of ribonuclease A (Sigma, Poole,Dorset) and 80 Units/ml of ribonuclease Tiin PBS at 37°C for 30 minutes. Other controlslides were included to validate the specificityof the method. These included competition ofthe radiolabelled oligonucleotides with eitherthe unlabelled probes used at a 50-fold excess(350 ng/slide) or with a large excess (10 ,ug = 1nmol/slide) of an unlabelled "irrelevent"oligonucleotide of similar G + C composition(47%). To verify further hybridisation speci-ficity, slides from one case were probed withthe irrelevent 30-mer which had been 3' tailedwith 35S dATP.

GRADING OF SCARS AND PCP1Two observers (YR and HTC) independentlyranked the photographs and the biopsy slidesfrom each patient from 1-33, after PCP 1immunostaining, according to severity ofmacroscopic scarring and extent of PCP 1immunostaining. This ranking was performedin a coded manner, such that the photographyand immunohistology results were assessedindependently. The criteria used to assess thephotographs of the scars were width of thescar and whether the scar was flattened orraised. The immunohistochemistry slideswere analysed under the light microscope( x 340 and x 540 magnification). In the partof the study comparing biopsy specimensfrom treated and untreated skin, paired sam-ples were examined without knowledge of thecode. The observers documented both thecellular and perivascular staining and gradedaccording to the extent of staining on a scaleofO, +/-, +, + +, + + +.The results fromthe two biopsy specimens (with and withouttopical steroids) were compared and a scorewas assigned to each pair according towhether application of steroid resulted in an

894



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Type 1 procollagen as a marker of severity of scarring after sternotomy: effects of topicalcorticosteroids89

increase or a decrease in the degree of PCP 1staining.

Spearman's rank correlation coefficient'9was used to test the correlation between thedegree of PCP 1 staining and the severity ofthe scar. These differences in cellular PGP 1staining in biopsy specimens from treated anduntreated skin were analysed using theWilcoxon sign rank test for paired observa-tions.

ResultsT'here was a wide variation in the visibleappearance of stemotomy scars (fig 1). Thescar appearances included thin line scars,mildly hypertrophic scars, and severely hyper-trophic scars extending beyond the line ofincision (figs lA-C).

Haematoxylin and eosin staining showedthat the scars contained inflammatory cellsand elastic van Gieson staining for maturecollagen highlighted mature collagen fibrils.Immunohistochemical staining for PCP 1identified both pronounced intracellular stain-ing and perivascular staining (fig 2A). Thiscompared to the low abundance of PCP 1immunostaining of normnal skin (fig 2B) and

i''

Figure 2

Immunohistochemical

staining ofPCP 1 in scar

tissue: (A) six week old

scar with intracellular and

perivascular staining; (B)normal skin; (C) mature

scar. Positive PCP

staining was detected bythe brown deposition ofDAB precipitates.

i

'Vt

Iii.1I

Iit

/

I

AL/

i )I

Al

.It:'jI

i./ 1

A

.)I

B

C

£4~~ 4-'~ ~ ~ ~

- I,

X

Figure 3 PCP 1 mRNA abundance as assessed by in situhybridisation in (A) sternotomy scar six weeks aftersurgery; (B) section of six week old scar treated by RNase.

mature scar tissue (seven years after surgery)j (fig 2C). In normal "healthy" skin, PCP 1

immunostaining was restricted to perivascularregions. In early scar tissue, both cellular andperivascular PGP 1 staining were present inbiopsy specimens from all patients to varyingdegrees.

In situ hybridisation for PCP 1 messengerRNA (mRNA) revealed an increase abun-

'-'dance of this mRNA in cells in sections fromscars six weeks after surgery (fig 3A), whichwas not observed in sections treated withRNase treated from scars obtained six weeksafter surgery (fig 3B). In other experiments,the hybridisation signal was abolished by pre-hybridisation of sections with a 50-fold excessof "cold" (unlabelled) antisense probe mix-ture of the same oligonucleotides beforehybridisation. Conversely, pre-hybridisationwith an excess unlabelled irrelevent probe didnot abolish the signal. The expression of pro-collagen 1 mRNAs using in situ analysis andtheir immunohistochemical localisationrevealed consistency between mRNA andprotein expression (figs 2A and 3A).

TGF-fl protein was detected in scar tissueobtained six weeks after surgery (fig 4A). Thisstainig was both intracellular and matrixassociated. The TGF-fl imnmunostaining in sixweek old scar tissue was similar to, but slightlygreater than, that in normal skin (fig 4B), butnot as great as that observed in a model of aninflammatory skin disorder (tuberculin-heaftest) (fig 4C).When the patients were ranked both

according to the severity of the scar, asassessed by macroscopic appearance on thephotographs and the abundance of PGPimmunostaining, patients with scars which

895

i

ij

r IT'ii

II 4P ;,

f



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Riaz, Cook, Wangoo, Glenvile, Shaw

Figure 4Immunohistochemicalstaining of TGF-f in (A)sternotomy scar six weeksafter surgery; (B) normalskin; (C) biopsy specimenofa tuberculin-heaf test atfive days.

the scar where corticosteroid cream had beenapplied, and those from untreated regions ofthe scar, topical corticosteroids had no effecton PCP 1 immunostaining (figs 6A-D). In thefirst eight patients, biopsy was performed 48hours after the application of one dose of top-ical corticosteroid. This was associated with adecrease by one grade in cellular PCP 1 stain-ing in four patients and an increase in threepatients (fig 6A). Similarly, perivascular stain-ing decreased in three patients, increased intwo patients, and remained unchanged inthree patients (fig 6B).

In the second 25 patients, biopsy of bothtreated and untreated sites was performedafter seven days of twice daily application ofcorticosteroid cream. Corticosteroids had noeffect on cellular or perivascular PCP 1 stain-ing. There was a decrease in cellular stainingby one grade in nine patients, an increase byone grade in nine patients, and no change inseven patients (fig 6C). After treatment withcorticosteroids, there was a reduction by onegrade in perivascular PCP 1 staining in fivepatients, no change in the PCP 1 staining in13 patients, and increases in staining by onegrade in four patients, two grades in twopatients and three grades in one patient (fig6D). PCP 1 mRNA in situ hybridisation wasperformed in four pairs of samples on biopsyspecimens from scars with or without sevendays of priof treatment with topical cortico-steroids, and no difference in the abundanceof PCP mRNA was observed between treatedand control tissues.

looked more severe hadgreater procollagen 1 stainin0O001) (fig 5). There was a

correlation between the exte:ing and the severity of theone of the 33 patients had a

of PCP 1 immunostainingscar. Four of the 33 patientswithout intense PCP 1 immiWhen the grade of cellula

PCP 1 immunostainingbetween skin biopsy specin

Figure 5 Relationbetween ranking ofpatientsaccording to severity ofscar (vertical axis) oraccording to abundance ofPCP 1 immunostaining(horizontal axis). Patientwith most severe scarrankedfirst and patientwith least ranked 33rd,and similarly forPCP 1

immunostaining.(Statistical analysis:Speannan's rankcorrelation coefficient.)

40-

(a

C.)cni'-0 30-0

._

4)

m 20-0

CD

.0

m oX 10-

° o

U~~~~~~~~~~~~~~~~~~~~~~~~~~ao

°

a

I

*o,UUrEl

X EXEX a

el a

10 20

Increasing PCP I s

DiscussionThis study documented the biochemicalevents in sternal scars six weeks after sterno-tomy. Immunohistochemical staining revealedan increase in both intracellular and perivas-cular staining for type 1 procollagen in the

correspondingly dermis, and in situ hybridisation confirmedIg (r. = 0t604, p < that cells had an increased abundance of col-particularly good lagen mRNA. By comparison, normal healthynt ofPCP 1 stain- skin showed no evidence of cellular produc-scar in that only tion of type 1 procollagen, some perivasculargreat abundance staining of PCP 1, but a very low abundancewithout a severe of collagen mRNA. In mature scar tissuehad a severe scar seven years after surgery, there was no PCP 1unostaining. staining, suggesting that the active fibroticr and perivascular process had ceased. The degree of scarring aswas compared assessed macroscopically from photographs

iens from sites of and PCP 1 immunostaining was variable, butcorrelated one with the other. This study thusextends observations which have already iden-tified collagen mRNA in keloid tissue,2>22 asthis is the first study identifying a biochemicalcorrelate with the macroscopic severity ofscarring. This link between the biochemicalevents in the scar and the macroscopicappearances, supports the approach of devel-oping treatments against components of thefibrotic response, in that the data support the

[rs 0604 assumption that preventing new collagen for-o uIP0LJ mation is likely to reduce the severity of scar-

ring.,.| Corticosteroids offer potential as agents to

30 40 prevent excessive scarring. With their potent;taining anti-inflammatory actions, they are likely to

nl i

896

t

.!. 1.": .K I -.:

.7^ :- " ..:



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Type 1 procollagen as a marker ofseverity ofscaning after sternotomy: effects of topical corticosteroids

Figure 6 Number ofpatients demonstrating achange in the grade ofcellular (A and C) orperivascular (B and D)PCP I immunostainingafter 48 hours (A and B)or seven days (C and D)of topical corticosteroidtreatment.

B

(i)

aL)

co

0zz

cn4-'

CL

Co00z

Decreasing PCP Increasing PCPstaining with steroids staining with steroids

-3 -2 -1 0 1 2 3


u)

41)CL

Co00z

Decreasing PCPstaining with steroids

Q)4-

Co

00z

Increasing PCP Istaining with steroids

o0 _ L _-3 -2 -1 0 1 2 3


be effective in preventing the early inflamma-tory phase, and it has long been known thatsteroids prevent the formation of granulationtissue.23 When administered immediately afterinjury, indices of fibrosis are reduced. Whenapplied within 24 hours to a skin blister, corti-costeroids reduced the abundance of PCP 1production by 75% after one day and 92%after two days.'3 Corticosteroids have alsobeen shown to inhibit the transcription of col-lagen in cultured fibroblasts and there is aglucocorticoid responsive element in the 5'upstream non-coding region of the type 1 pro-collagen a gene.24 However, in clinical prac-tice corticosteroids have not been universallyeffective in the treatment of established hyper-trophic scars. Optimal treatment may requireexcision with intraoperative steroids.25 Thepresent study also suggests that in an estab-lished scar that is six weeks old, topical corti-costeroids do not appreciably reduce collagenproduction. Thus, a single application ofsteroid cream over 48 hours and repeatedapplications over seven days did not signifi-cantly alter the perivascular or cellular con-centration of procollagen 1 nor the abundanceof collagen mRNA. One possible explanationis that adequate concentrations of cortico-steroid were not absorbed. Although clobeta-sol is one of the most potent topicalcorticosteroids, the extent of skin penetrationafter topical application is poorly defined.Penetration into and through the epidermis issuggested by the work of Harding et aP6 whodocumented vasoconstriction of normal skin

following application of clobetasol proprion-ate in propylene glycol as carrier (as was thecase in the present studies). Other studiesmeasuring subcutaneous adipose blood flowin normal skin have failed to demonstrate anyeffect attributable to clobetasol.27 The evi-dence of skin penetration in dermatoses isclearer. When applied to psoriatic skin, clobe-tasol proprionate has been associated with areduction in circulating cortisol28 anddetectable plasma clobetasol concentrations.29We were unable to find published data on theabsorption into hypertrophic scars. In futurestudies injection of longer acting cortico-steroids into the scar will be assessed.Much of the published information comes

from keloid scars which by definition continueto hypertrophy 12 months or more after thewound.' We do not know what proportion ofthe scars in the present study will becomekeloid scars. However, information fromfibroblasts from keloid scars offers an alterna-tive hypothesis to explain the lack of effect oftopical steroids observed in the present study.In this hypothesis fibroblasts lose theirrequirement for exogenous growth factors andbecome relatively resistant to exogenous inhi-bition. This hypothesis is supported by anumber of pieces of evidence. Firstly, theabundance of TGF-,B in our scar biopsy speci-mens was less than that observed in biopsyspecimens of the inflammatory skin lesionoccurring five days after a tuberculin-heaftest, whereas the PCP 1 abundance was muchgreater in the six week old scars. Secondly,

897



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Riaz, Cook, Wangoo, Glenville, Shaw

Babu et alP have shown that, unlike normalfibroblasts, keloid fibroblasts do not increasetotal protein synthesis in response to TGF-fi,although under certain conditions keloidfibroblasts increase fibronectin production inresponse to TGF-fl compared with normalcells. Thirdly, others have shown that hydro-cortisone resulted in 60% inhibition of colla-gen synthesis in fibroblasts from normal scarsbut no reduction in cells from keloid scars.3'2Fourthly, keloid fibroblasts in culture arerefractory to inhibition of DNA synthesis byphorbol esters and have reduced sensitivity toprostaglandin E2." These differences betweennormal and keloid fibroblasts may be due tochanges in receptors in that keloid fibroblastshave reduced phorbol ester binding affinity,33or altered enzyme activity-for example,keloid fibroblasts have higher G6PDH activityunder oxygen saturation conditions than nor-mal cells.'4 Fifthly, at least a proportion offibroblasts from keloids have increased type 1procollagen mRNA, rates of procollagen pro-duction, and decreased rates of collagendegradation, compared with fibroblasts fromcontrol skin.'536

Therefore, early events after a wound prob-ably involve an inflammatory phase charac-terised by the presence of the inflammatorycytokines. Patients with a predisposition tokeloid formation may have mononuclear cellswith an increased capacity to produce inflam-matory cytokines such as interleukin-6,tumour necrosis factor a and interferon fl.37Growth factors are produced for fibroblasts,such as platelet derived growth factor fi andTGF-fl. Studies which have involved supple-menting the growth factors at this stage sug-gest that platelet derived growth factorcontaining the B subunit is responsible forrecruiting procollagen 1 containing fibroblastswhile TGF-fl 1 results in increased intracellularconcentrations of procollagen 1.7 3841 Theeffects of adding corticosteroids at this earlystage can also be partly overcome by additionof TGF-fl to the wound.40 Once the scar isestablished, the fibroblasts continue to beactive for a long time in the absence of aninflammatory phase or evidence of increasedgrowth factors. In our study the duration ofthe active fibrotic phase was surprisingly long.Biopsy specimens taken from patients 11months and two and a half years after ster-notomy both showed PCP 1 staining whichwere grades + + + and +, respectively.However, six and seven years after surgerythere was no PCP1 staining. As diverse pat-terns of macroscopic appearances and PCP1staining were established by six weeks, itseems that the events which determinewhether the fibroblasts in the scar are going toswitch to a keloid-like phenotype, resulting in ahypertrophic scar, occur prior to six weeks.These events may relate to the nature orextent of the inflammatory phase. It is attrac-tive to hypothesise that fibroblasts bathed inhigh concentrations of inflammatory cyto-kines and growth factors during this earlyphase remain hyperactive and thus contributeto more florid scar formation. This concept of

a fibroblast which becomes persistantly activein response to exogenous stimuli, and passesthis capacity on to subsequent generationsdespite repeated replication has been sug-gested in scleroderma and lung fibrosis.424'

This study was supported by grants from the MedicalResearch Council and the British Lung Foundation. We thankDr Foale and Mr Stanbridge (at St Mary's Hospital), andProfessor Taylor and Mr Keogh (at The HammersmithHospital), for allowing us to study their patients.

1 Harding Rains AJ, David Ritchie H. The healing and man-agement of wounds. In: Bailey and Love's A short practiceofsurgery. London: Lewis and Co Ltd, 1984:3-7.

2 Ketchum LD, Cohen ID, Masters FW. Hypertrophic scarsand keloids. Plast and Reconstruct Surg 1974;53: 140-51.

3 Kovacs EJ. Fibrogenic cytokines: The role of immunemediators in the development of scar tissue. ImmunolToday 1991;12:17-23.

4 Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA.The biosynthesis of collagen and its disorders (first oftwo parts). NEnglJ Med 1979;301:13-23.

5 Prockop DJ, Kivirikko KI, Tuderman L, Guzman NA.The biosynthesis of collagen and its disorders (second oftwo parts). NEnglJMed 1979;301:77-85.

6 Martinez-Hernandez A. Repair, regeneration and fibrosis.In: Rubin E, Farber JL, eds. Pathology. Philadelphia: J BLipincott, 1994:81-93.

7 Pierce GF, Brown D, Mustoe TA. Quantitative analysis ofinflammatory cell influx, procollagen type 1 synthesis,and collagen cross-linking in incisional wounds:Influence of PDGF-BB and TGF-Beta 1 therapy. J LabClin Med 1991;117:373-81.

8 Prockop DJ. Collagen biochemistry-design of agents toinhibit excessive accumulation of collagen during woundrepair. In: Dineen P, ed. The surgical wound.Philadelphia: Lea & Febiger, 1981:97-109.

9 Cohen IK, McCoy BJ. Keloid; Biology and treatment. In:Dineen P, ed. The surgical wound. Philadelphia: Lea &Febiger, 1981:123-31.

10 Craig RDP, Schofield JD, Jackson DS. Collagen bio-synthesis in normal and hypertrophic scars and keloid asa function of the duration of the scar. Br J Surg1975;62:741-4.

11 Stryer L. Connective tissue proteins. Biochemistry. NewYork: WH Freeman & Co., 1988:269-73.

12 McDonald JA. Broekelmann TJ, Matheke ML, Crouch E,Koo M, Kuhn C. A monoclonal antibody to the car-boxyterminal domain of procollagen type 1 visualisescollagen-synthesising fibroblasts. J Clin Invest 1986;78:1237-44.

13 Griswold ML. Effect of adrenal cortical preparations onscar hypertrophy. Plast Reconstruct Surg 1954;13:454-61.

14 Oikarinen A, Autio P, Kiistala U, Risteli L, Risteli J. A newmethod to measure type I and III collagen synthesis inhuman skin in vivo: Demonstration of decreased colla-gen synthesis after topical glucocorticoid synthesis.J Invest Dermatol 1992;98:220-5.

15 Warburton MJ, Fernes SA, Hughes CM, Sear CHJ,Rudland PS. Generation of cell types with myoepithelialand mesenchymal phenotypes during the conversion ofrat mammary tumor epithelial stem cells in to elongatedcells. JNCI 1987;78:1191-201.

16 Taylor M, Cook T, Pearson C, Risdon RA, Peart S. Reninmessenger RNA localisation in congenital mesoblasticnephroma using in situ hybridization. J Hypertension1989;7:733-40.

17 Bernard MP, Chu ML, Myers JC, Ramirez F, EikenberryEF. Nucleotide sequences of complementary deoxy-ribonucleic acids for the Proal chain of human type 1Procollagen. Statistical evaluation of structures that areconserved during evolution. Biochemistry 1983;22:5213-23.

18 Naish SJ, ed. Handbook ofimmunochemical staining methods.California: Dako Corporation, 1989.

19 Siegel S. Nonparametric statistics for the behavioural sciences.Tokyo: McGraw-Hill Book Co., 1956.

20 Lee KS, Song JY, Suh MH. Collagen mRNA expressiondetected by in situ hybridization in keloid tissue.JDernatol Sci 1991;2:316-23.

21 Peltonen J, Hsiao LI, Jaakkola S, Sollberg S, AumailleyM, Timpl R, et al. Activation of collagen gene expressionin keloids: co-localization of type I and VI collagen andtransforming growth factor beta 1 mRNA. J InvestDermatol 1991;97:240-8.

22 Sollberg S, Peltonen J, Uitto J. Combined use of in situhybridization and unlabelled antibody peroxidase anti-peroxidase methods: simultaneous detection of type Iprocollagen mRNAs and factor VIII-related antigenepitopes in keloid tissue. Lab Invest 1991;64:125-9.

23 Baker BL, Whitker WL. Interference with wound healingby local action of adrenal cortical steroids. Endocrinology1950;46:544-5 1.

24 Perez JL, Shull S, Gendimenico GJ, Capetola RJ, MazikJA, Cutroneo KR. Glucocorticoid and retinoid regula-tion of alpha 2 type 1 procollagen prometer activity. JfCell Biochem 1992;50:26-34.

898



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Type 1 procollagen as a marker of sevenity ofscamng after sternotomy: effects of topical corticosteroids

25 Tang YW. Intra- and postoperative steroid injections forkeloids and hypertrophic scars. Br J Plast Surg 1992;45:371-3.

26 Harding SM, Sohail S, Busse MJ. Percutaneous absorp-tion of clobetasol proprionate from novel ointmentand cream formulation. Clin Exp Dermatol 1985;10:13-20.

27 Peterson U, Sindrup JH, Kristensen JK. Subdermal pene-tration of topically applied clobetasol proprionate inointment base through intact human skin: lack of effecton subcutaneous adipose tissue blood flow. Jf DermatolSci 1992;3:42-5.

28 Gip L, Hamfelt A. Studies on the efficacy and adrenaleffects of Diproline ointment 0 05% and Dermovateointment 0 05% in patients with psoriasis or other der-matosis. Cutis 1984;33:215-24.

29 Hehir M, du Vivier A, Eilon L, Danie MJ, Shenoy EVB.Investigation of the pharmacokinetics of clobetasol pro-prionate and clobetasone butyrate after a single applica-tion of ointment. Clin Exp Dermatol 1983;8: 143-51.

30 Babu M, Diegelmann R, Oliver N. Keloid fibroblastsexhibit an altered response to TGF-Beta. 7 InvestDermatol 1992;99:650-5.

31 Russell JD, Russell SB, Trupin KM. Differential effects ofhydrocortisone on both growth and collagen metabolismof human fibroblasts from normal and keloid tissue. JfCell Physiol 1978;97:221-3.

32 Russell SB, Trupin JS, Myers JC, Broquist AH, Smith JC,Myles ME, et al. Differential glucocorticoid regulation ofcollagen mRNAs in human dermal fibroblasts. 7 BiolChem 1989;264:13730-5.

33 Myles ME, Russell JD, Trupin JS, Smith JC, Russell SB.Keloid fibroblasts are refractory to inhibition of DNAsynthesis by phorbol esters. Altered response is accom-panied by reduced sensitivity to prostaglandin E2 andaltered down-regulation of phorbol ester binding sites. 7Biol Chem 1992;267:9014-20.

34 Sit KH, Lau YK, Aw SE. Differential oxygen sensitivities

in G6PDH activities of cultured keloid and normal skindermis single cells. 7Dermatol 1991;18:572-9.

35 Diegelmann RF, Cohen IK, McCoy BJ. Growth kineticsand collagen synthesis of normal skin, normal scar andkeloid fibroblasts in vitro. Cell Physiol 1979;98:341-6.

36 Abergel RP, Pizzuro D, Meeker CA, Lask G, MatsuokaLY, Minor RR, et al. Biochemical composition of theconnective tissue in keloids and analysis of collagenmetabolism in keloid fibroblast cultures. J InvestDermatol 1985;84:384-90.

37 McCauley RL, Chopra V, Li YY, Herndon DN, RobsonMC. Altered cytokine production in black patients withkeloids. J Clin Immunol 1992;12:300-8.

38 Mustoe TA, Pierce GF, Thomason A, Gramates P, SpomME, Deuel TF. Accelerated healing of incisionalwounds in rats induced by transforming growth factor-beta. Science 1987;237:1333-5.

39 Pierce GF, Mustoe TA, Senior RM. In vivo incisionalwound healing augmented by platelet-derived growthfactor and recombinant c-sis gene homodimeric pro-teins. J Exp Med 1988;167:974-87.

40 Pierce GF, Mustoe TA, Linglebach J, Masakowski VR,Gramates P, Deuel TF. Transforming growth factorbeta reverses the glucocorticoid-induced wound healingdeficit in rats. Proc NadtAcad Sci USA 1989;86:2229-33.

41 Mustoe TA, Purdy J, Gramates P, Deuel TF, ThomasonA, Pierce GF. Reversal of impaired wound healing inirradiated rats by platelet-derived growth factor-BB:requirement of an active bone marrow. Am J Surg1989;158:345-50.

42 Bordin S, Page RC, Narayanan AS. Heterogeneity of nor-mal human diploid fibroblasts: isolation and characteri-zation of one phenotype. Science 1984;223:171-3.

43 Jordana M, Schulman J, McSharry C, Irving LB,Newhouse MT, Jordana G, et al. Heterogeneous prolif-erative characteristics of human adult lung fibroblastlines and clonally derived fibroblasts from control andfibrotic tissue. Am Rev Respir Dis 1988;137:579-84.

899



j.com/

J Clin P


ownloaded from

http://jcp.bmj.com/

Documents

Clin Type procollagen marker severity scarring after ... · Type 1 procollagenas amarkerofseverity of scarring after sternotomy: effects oftopical corticosteroids YRiaz, HTCook,AWangoo,BGlenville,