Br. J. Surg. 1992. Vol. 79. January,69-72 Effect of transforming growth

factor beta and basic fibroblastgrowth factor on steroid-impairedhealing intestinal wounds

vt/1P:.IrI~-.-ri

A longitudinal intestinal 11'ound model in the pig was used to assess theeffect of parenteral steroids (betamethasone 12 mg 50 kg -1 intra-

muscularly l11'ice daily) on breaking loado Steroid treatment significantlydecreased the breaking load of wounds in the ileum and colon incomparison with l1'oundsfi'om saline-treatedanimals, In afurther groupof animals receiving steroids, paired longitudinal l1'ounds l1'ereconstructed. One 1",'ound of a pair l1'as treated l1'ith a local applicationof transforming grol""th factor befa (TGF-P) (5 j1g per "'ound) or basic

fibroblasl growthfactor (5 j1gper l1.ound) in a collagen suspension. Theother l1'ound l1'as treated l1'ith a collagen suspension alone~eaf II.nul1d~treated l1,ith TGF-p

J. Slavin, J. R. Nash* andA. N. Kingsnorth

Departments o, Surgery and*Pathology, Universityo'Liverpool, UKCorrespondence lo:Mr A. N. Kingsnorth, Departmentof Surgery, University of Liverpool,PO Box 147, Liverpool L69 3BX, U K m a colla ension al lhe lllne O '.

Peptide growth factors are increasingly implicated in the controlof wound healingI. Recombinant biotechnology providesamounts of these previously rafe peptides which allows theiruse in in riro experiments and may permit their clinical use.Epidermal growth factor2, tumour necrosis factor3, trans-forming growth factor beta (TGF-P)4 and platelet-derivedgrowth factor5, applied in slow release suspensions or vehiclesto healing incisional skin wounds at the time of wounding,accelerate the gain in breaking load of these wounds. Mustoeand colleagues6 recently demonstrated acceleration of the gainin strength of healing rabbit gastrotomy wounds with topicalTGF-p in a collagen suspension. Local applications of peptidecan be shown to reverse the deleterious effects of steroidtherapy7, cytotoxic treatment8 and local radiotherapy9 insimple ski n incisional wound models. It remains unclear towhat extent these various effects are organ- or species-specific.

The failure of intestinal wounds is associated withconsiderable morbidity and mortality rates and usually occursin situations of impaired healing. Steroid impairment ofcutaneous healing is well documented1O.11. Few studies oftheeffects of steroids on healing intestinal wounds have beenpublished 12.13. One recent experimental study failed todemonstrate impairment of healing following steroid treat-mentl4. Clinical studies have failed to demonstrate anydeleterious effects on intestinal healing. The numbers of treatedpatients in any particular trial would, however, be small anddemonstrating any negative effect would be difficult 15. Theeffect of steroid therapy on healing longitudinal enterotomywounds in pigs was assessed by measurement of breaking loadoThe effect of local applications of two peptide growth factors,TGF-fJ and basic fibroblast growth factor (bFGF), each in acollagen suspension, was tested on the steroid-impaired healingof enterotomy wounds.

Pittman/Moore, Crewe, UK) was used. Biological activity ofTGF-P.and bFGF was ensured by a simple i/l riro assay utilizing Swiss 3T3fibroblasts between the tenth and 14th passage grown to confluence in96-well plates. Mitogenicity was assessed by uptake of tritiatedthymidine. The vehicle used for delivery .of peptide was 200 JIl of adilute collagen suspension (Zyderm II; Collagen Corporation, PaloAlto, California, USA) to provide a viscous preparation allowing easyapplication to wound edges that may delay release of peptide for ashort period at a si te of local application5.

DemO/l.~lr{/li{/nl?(,~leroid impairmenr l!f healinyOne group of seven pigs "as treated with betamethasone 12 mg 50 kg -I

intramuscularly twice daily, beginning 48 h before the operation andcontinuing until killing, Seven other animals received a similar volumeof saline intramuscularly. At operation animals had 6 cm anti-mesenteric enterotomy wounds constructed at standard sites (three inthe ileum and two in the colon) and were killed at 7 days. Statisticalcomparisons were bet\"een wounds in steroid- and saline-treatedanimals.

E:tfe('¡ (!f grOlI'/hfa('/orsA further 19 pigs.were treated with b~tamethasone 12 mg 50 kg-1intramuscularly twice a dar beginning 48 h before the operation andcontinuing until killing. Animals had paired wounds constructed atstandard sites. three pairs in the ileum and one in the colon (Figure 1),Each member of a pair was 30 cm from its partner. One wound in apair was treated.with growth factor in a collagen suspension. the otherwas treated with collagen suspension alone. AII control wounds in aparticular animal were compIeted before growlh factor-treated woundsto prevent contamination. A standard wounding protocol was followed:the seromuscular layer was divided with cutting diathermy until themucosa bulged. and growth factor or control solution was applied tothe wound to allow even application to the whole surface. The mucosawas then opened wilh a blade before suture. Animals were treated withonly one growth factor. In six animals the effects of saline werecompared with collagen suspension. In nine animals the effects ofTGF-fi (5 Jlg) in a collagen suspension were compared with collagenand in four animals the effects of bFG F (5 Jlg) in a collagen suspensionwere compared wilh collagen. Stalistical comparisons were betweenpaired wounds.

Wounding /1/odelFluids only were allowed for the 12 h preccding operalion. Anaesthesiawas induced with halolhane and maintained aftcr intubalion with amixture of halothane. nitrous oxide and oxygen. A single dose ofpenicillin (Triplopen': Glaxo. Greenford. UK) was given with

induction.

Materials and methods

Weanling Landrace pigs weighing between 16 and 25 kg were usedthroughout. Pigs were housed in groups of up to four and maintainedon a constant dayjnight cycle. Animals were weighed 48 h before theoperation, at the time of operation and when killed.

An intramuscular betamethasone preparalion (Belsolan' ; Coopers

Paper pre,\enled al Ihe /11ee/;ng (!f /he Surg;('a/ Re,\ear(h S(I(';e/)' he/(/

in Lant'a.\'/('r. UK (1/1 4-5 Ju/y /99/

69

~~

Effects of steroids

There were no significant changes in weight between thesteroid- and saline-treated groups. It was common to find smallintraperitoneal collections of fluid in steroid-treated animals.Steroid treatment significantly decreased the mean(s.e.m.)breaking load of ileal (160( 20) g cm -1 l'ersus 320( 32) g cm -1 )and colonic (200( 12) g cm -1 rersus 390( 33) g cm -1 ) wounds

(both P < 0.01, Wilcoxon rank sum test, Table 1 and Figure 2).There were no differences in breaking load between the threedifferent sites within the ileum or the two sites in the colon(Table 1 ). No histological differences were apparent at 7 daysbetween steroid- and saline-treated animals.

6cml

.'.'

6cml

Colon

E.tfect 01 groll,th lactorsThere were no significant differences in breaking load betweencollagen-treated wounds and saline controls (Table 2). Woundstreated with TGF -{3 had increased mean(s,e.m.) breaking loads(ileaI285(32)gcm-1 versus 180(11)gcm-l, colonic 365(59)g cm -1 versus 260(40) g cm -1) in comparison with collagen-

treated controls (Figure 3). This was significant at the 5 per centlevel for ileal wounds but just failed to reach significance withcolonic wounds. No majar differences were noted betweenwounds treated with TGF-{3 and collagen-treated controls whenhistological sections were examined.

-"30 cm.

Ileum .~

~

..'

Table 1 Breaking s/reng/h of inles/inal "'ounds in s/eroid- andsoline-/realed onimals al specific si/es in /he ileum (A. B and C) andcolon (D and E)

Figure 1 Distribution ofll'ounds in groll'lhfactor-treated animals: threepairs in the ileum and one pair in Ihe colon

Saline Steroid

310(63)290(58)360(37)320(32)

150(29)170(37)170(35)160(20)*

IleumSiteA (11 = 6)Site B (11 = 6)SiteC (11 = 6)

Mean(II=18)

ColonSiteD (11 = 6)Site E (11 = 6)Mean (11=12)

370(49)415(40)39.0(33)

210(28)185(14)200( 12)*

Values are mean(s.e.m.) expressed in g cm -'. n, Number of wounds* p < 0.01 (Wilcoxon rank sum test)

Wounds were sutured at 4 mm intervals with interrupted 3/0seromuscular polyglactin 910 (Vicry."; Ethicon Ud., Edinburgh,UK).FIuids were allowed immediately after recovery and a liquid diet wasgradually introduced. Animals were killed at 7 days. The intestine wasremoved immediately for testing and kept on ice until Ibis wascompleted. Wounds were tested beginning proximally and continuingdistally. Sutures were removed and a 2-cm ring of bowel containingthe enterotomy wound was cut. A tensiometer (model TTBM A0543;Instrom Corporation, Cantan, Massachusetts, USA) was adapted sothat hooks were used rather Iban the more traditional clampsl6. Thehooks were inserted into the lumen of each segmento The uppercrosshead was fixed and connected to a load cell; the Jower crossheadmoved at a constant strain. Each specimen was tested to disruption.A crosshead speed of 1 cm min -I was used for the initial group of

animals used to demonstrate steroid impairment of healing. In arderto shorten the testing period the crosshead speed was increased to5 cm min -I with growth factor-treated animals. A load deformation

curve was produced on a chart recorder from which {he breaking loadof each specimen was calculated. Breaking loads are expressed asgcm-l.

HislologyThe remainder of the wound was fixed in 10 per cent formal saline for48 h, processed and then orientated accurately in paraffin wax beforeembedding. Sections, 5 Jlm thick, were cut and stained withhaematoxylin and eosin or a van Gieson stain (for collagen). A singlepathologist who was unaware of the identity of individual slides thenexamined each section.

~

íEual

.!::

Elc..'-~alC~lO.,LoaJ

StatisticsResults are expressed as mean(s.e.m.) unless otherwise stated. Allstatistical comparisons are with non-parametric tests, a Wilcoxon ranksum test for unpaired data and a Wilcoxon signed rank test for paireddata.

ResultsMortality rate

Two animals were killed before 7 days; one had a volvulus ofdilated proximal ilealloops but no intra-abdominal pathologywas apparent in the other. A further six animals died before 7days. Of these, one animal had peritonitis and no particularenterotomy could be implicated; the other five animalsdemonstrated dilation and necrosis or frank volvulus of

proximal ilealloops.

Site

Figure 2 Breaking .'itrength o/ enterotom)' K"ounds in steroid- (O) and.'ia/ine-treated( O) anima/s. Va/ues are mean(s.e.m.). There were 18 i/ea/and 12 c%nic K'ound.'i in each group. .P < 0.01 ( Wi/coxon rank sum test)

70 Br. J. Surg.. Vol. 79. No. 1. January 1992

""ble 2 Effec( o/ lócaIly'app[;ed" gro~lhjaclor on breaking sl.'englh

Colon--"', ,~~-":~--c- ;-c- -

Treated Controlc .--

365(59) 260(40)(n = 6 pairs)

155(67) 185(33)(n = 4 pairs)

220(62) 165(42)(n = 4 pairs)

--,

Treated Control

285(32) 180(11)*(n=18pairs)

145(19) 200(21)*(n = 12 pairs)

165(18) 160(34)(n = 9 pairs)

iF-P

ollagen

500 *~

1100

I'eu01

.ca,c..'--;;;

g'~~..'-aJ

~

J-

~

...

~\\\\\\\\

~lO

a o ColonIleum

Site

500

1100

~

leuC\

.cc,cQ)..',;;

C\C~

/OQ)..

[!)

*~

-LI

~

cc

OiscussionWound healing in the intestine has traditionally been estimatedby the measurement of anastomotic bursting strength. Asegment of bowel with an anastomosis at its centre is distendedwith fluid or gas and the pressure at which bursting occurs istaken as a measure of intestinal strength. Unfortunately. afterthe first few days of healing. disruption characteristically occursaway from the anastomosis; this is not a reflection ofsupranormal anastomotic wound strength but rather thedecreased distensibility of the anastomotic afea. The law ofLaplace states that the tension in the wall of a distending tubeis directly related to its diameter and the tension in the bowelwall distal and proximal to an anastomosis may greatly exceedthat at the anastomotic linel7. For this reason we and a numberof otherslB,19 have preferred to measure the breaking load oflongitudinal enterotomy wounds or of segments of intestinecontaining an anastomosis, Collagen is responsible for thestrength of a wound. A parallel association has beendemonstrated between collagen concentration and woundstrength in healing rat gastrotomy incisions20. It seemsreasonable to speculate. however. that the absolute amount ofcollagen within a wound and its quality and organization must

be important factors.Mastboom and colleagues recently failed to demonstrate any

deleterious effect of steroid therapy (methylprednisolone10 mg kg-1 day-l) in rats on the healing of intestinalanastomosis as assessed by bursting pressurel4. At 7 days.however. all the segments of bowel tested burst away from theanastomosis. We have demonstrated that steroid treatment(betamethasone 24 mg 50 kg-1 day-l) significantly decreasesthe breaking load of both ileal and colonic enterotomy wounds

at 7 days.Steroids might act at a number of points in a wound healingcascade to depress collagen synthesis. Steroid treatment causesprofound monocytopenia21. Monocytes are now recognized ashaving a pivotal rol e within a wound healing cascade22. Oncepresent at a si te of healing they release peptide growth factorsthat recruit further cells and stimulate collagen and matrixproduction. In l'itro steroids can be shown to have a directeffect on fibroblasts. decreasing messenger RNA transcriptnumbers for both procollagen and matrix protein23,24.Clinically a direct depression of both collagen and matrixproduction might be responsible for the deleterious effects of

steroid therapy on fibroplasia.Basic FGF is ubiquitous and found throughout the body25.

It has chemotactic and mitogenic effects predominantly on cellsof mesenchymal origin (including fibroblasts). In ritro it is alsochemotactic and mitogenic for endothelial cells26 and in simplein rivo assays bFGF is a potent angiogenic factor'. These twoattributes of bFGF -its mitogenic effect on fibroblasts and itsangiogenic properties in ril'o -suggest it is a critical factorinvolved in the formation of granulation tissue.

Injection of bFGF into simple wound chambers increasesthe cellularity of granulation tissue2B. The infiltration ofrecombinant bFGF into healing rat incisional skin wounds onthe third day after wounding increases the rate of gain ofstrength of treated wounds29. The inhibition of healing indiabetes shows some similarities to that seen with steroidtherapy: bFGF increases the cellularity of treated sponges in

diabetic rats3o. Applications ofbFGF in a collagen suspensionto ileal enterotomy decreased the strength of treated incisionsin comparison with collagen-treated controls at 7 days. Treatedwounds appeared to be more cellular than control wounds. Wehave since confirmed that simple applications of bFGF to

o ,'.. .-:::~::::::::::~Ileum Colonb

Site

Figure 3 Breaking ,~/reng/h ~fpaired /rea/ed (O) and con/rol (O)l1'ound,\' in ,~/eroid-/rea/ed animal,~, Value,~ are mea/1(,~,e,m,), aTra/1,~formi/1g grol1'/h,fac/or be/a; 18 ileum pair,~. ,~i,'f c(/ltm pairs, b Ba,~icfibrobla,~/ grol1'/h ,fac/or; 12 ileum pair,~. ,(l/ur colo/1 pair,~, * P < 0,05

(Wilco,'ffm ,~igned rank /e,~/)

71Br. J. Surg.. Vol. 79. No.1. January 1992

collagen-treated control woundS.

300

200