B

dcn

M

tcadp(gaa

R

b

Tspiim

rmereuatpw

H

P

aS

1

Body Wall Repair Using Small Intestinal Submucosa SeededWith Cells

By Jin-Yao Lai, Pei-Yeh Chang, and Jer-Nan LinTaoyuan, Taiwan

s(sam

C

cnJr

Iis

ackground/Purpose: Prosthetic repair of large ventral ab-ominal wall defects has been associated with high compli-ation rates. This study was aimed at applying tissue engi-eering to body wall replacement.

ethods: Syngeneic Lewis rats underwent harvest of skele-al muscle specimens. Once expanded in vitro, skeletal mus-le cells or fibroblasts were suspended in a collagen gel. Allnimals underwent creation of a 2.5- � 3-cm abdominal wallefect. The defect was repaired with the cell-seeded gellaced in between 2 pieces of small intestinal submucosaSIS). The control group was repaired by SIS with acellularel. Animals were killed at different time-points for histologicnd mechanical examination. Statistical analysis was bynalysis of variance (ANOVA).

esults: Abdominal wall hernia was present in 6 of 24 fibro-last-seeded constructs (25%), 5 of 21 skeletal muscle cell-

(doi:10.1016/j.jpedsurg.2003.08.019

752 Journal

eeded constructs (23.9%), and 16 of 21 acellular grafts76.2%), respectively (P � .05). At harvest, cell-seeded con-tructs were thicker with better cellular infiltration, whereascellular grafts were thin, low in cell density, and poor inechanical resistance.

onclusions: Unlike acellular collagen matrices, engineeredellular constructs have better cell infiltration and mechani-al performance. Tissue engineering may be a viable alter-ative for body-wall replacement.Pediatr Surg 38:1752-1755. © 2003 Elsevier Inc. All rights

eserved.

NDEX WORDS: Cell culture, tissue transplantation, abdom-nal muscles, extracellular matrix, prostheses, implants, rats,urgical mesh.

HE REPAIR of large soft tissue defects, especiallyabdominal wall defects, is still a challenge for

urgeons. The ideal biomaterial for abdominal wall re-air should possess adequate strength, no hypersensitiv-ty reactions, and biocompatibility to facilitate tissuengrowth, which may help long-term maintenance ofechanical strength.Small intestinal submucosa (SIS) is an acellular

esorbable biomaterial derived from the extracellularatrix of the small intestinal submucosa that has been

valuated as an abdominal wall repair device in bothabbit and rat models. By applying the concept of tissuengineering, some investigators have reported successfulse of SIS seeded with cells for diaphragmatic repair infetal lamb model.1 The seeded cells facilitate better

issue regeneration and tissue ingrowth. This study ap-lied a similar concept in the repair of a large abdominalall defect.

From the Department of Pediatric Surgery, Chang-Gung Childrospital, Kweishan, Taoyuan, Taiwan.Presented at the 36th Annual Meeting of the Pacific Associati

ediatric Surgeons, Sydney, Australia, May 12-16, 2003.Address reprint requests to Jin-Yao Lai, MD, Department of P

tric Surgery, Chang-Gung Children’s Hospital, No. 5 Fu-Hstreet, Kweishan, Taoyuan 333, Taiwan.© 2003 Elsevier Inc. All rights reserved.0022-3468/03/3812-0013$30.00/0

MATERIALS AND METHODS

IS Preparation

Segments of fresh porcine small intestine, about 10 cm in length,ere flushed with tap water to remove the intestinal contents. After

ongitudinal splitting of the intestine segment, the seromuscular layeras removed mechanically. The decellularization process involved

haking the jejunal segments in a stirrer at 4°C with 1 L of 0.2% Triton-100 (Sigma Chemical Co, St Louis, MO) and 26.5 mmol/L ammo-ium hydroxide for 10 to 14 days. After decellularization, the jejunalegments were washed in deionized H2O for 72 hours. Acellularity wasssured by H&E staining.

uscle Biopsy

All procedures were approved by the Chang Gung Children’s Hospitalnimal Care and Use Committee. Sixty-six male syngeneic Lewis ratsith body weights between 100 and 150 g were used. The rats were

nesthetized with intraperitoneal injection of mixed ketamine (100 mg/kg)nd xylazine (10 mg/kg) solution. A 2.5- � 3-cm whole layer abdominalall defect was created. The removed specimen was prepared for cell

ulture. The defect was repaired by different cell-scaffold composites.

ell Culture

After mincing, the skeletal muscle specimens were digested withype I collagenase (2 mg/dL, Sigma) for 60 minutes in a 5% CO2

ncubator at 37°C. After trituration and passage through a 100-mmylon cell strainer (Becton Dickinson, Franklin Lake, NJ), the skeletaluscle cells were spun down and plated on 10-cm petri dishes. The

upernatant, which contained fibroblasts, was collected, spun down,nd plated the same way. Cells were cultured separately and were fedith DMEM medium (Sigma) containing 10% fetal bovine serum

S

wlwsXnsa

M

Awaawc

C

tinmsaw

en’s

on of

edi-ing

c

FBS; Sigma), glutamine, penicillin, and Fungizone (Gibco, Grand

of Pediatric Surgery, Vol 38, No 12 (December), 2003: pp 1752-1755

Island, NY), in a 95% humidified, 5% CO2 chamber at 37°C. The cellswere expanded until the desired cell numbers were obtained.

Collagen Preparation and Cell Seeding

Confluent cultures of skeletal muscle cells or fibroblasts after pas-sages 2 through 4 were detached from petri dishes with 0.01% EDTA-0.05% trypsin and then mixed with the soluble collagen (BectonDickinson) to create a cell-collagen solution. The cell density was1.5 � 106 cells/mL. The final collagen concentration in the lattices was1.5 mg/mL. A 1-mL cell-collagen solution was placed onto a 50-mmdiameter dry plastic dish to ensure that the lattice would remainattached for the 5-day culture period. This was incubated for 1 hour toallow lattice formation, and then 5 mL of fresh medium containing 10%FBS was added over the cell-collagen lattice and incubated for 5 days.The cell-collagen lattice then was mechanically released from theplastic substratum using the tip of fine forceps. The thin cell-containing

collagen lattice was placed between 2 layers of SIS as a new prosthesis.The fullthickness defect was repaired by SIS seeded with fibroblasts(FB group; n � 24), with skeletal muscle cells (SK group; n � 21), orwithout cell seeding (NC group; n � 21).

Specimen Collection and Analyses

The animals were killed by high-dose ketamine followed by cervicaldislocation at 1, 2, and 3 months after surgery. The abdominal wall wasdistended with 200 mL saline for the mechanical strength test. Anyabnormal abdominal protrusion before or after the mechanical testingwas considered fascial weakness or a hernia. The abdominal wall thenwas removed en bloc for examination. The specimen was fixed in 10%neutral buffered formalin and embedded in paraffin. Sections 6-�mthick were prepared for histologic analysis with H&E. Sections 4-�mthick were prepared for immunohistochemical analyses with mouse-derived monoclonal antibody to desmin (DAKO, Carpinteria, CA). Acommercial immunohistochemical kit was used with 3,3�-diamino-benzidine tetrahydrochoride (DAB; DAKO) as the enzyme substrate,yielding granular brown deposits for positive interpretation.

Statistical Analysis

Statistical evaluations were performed using analysis of variance(ANOVA) test, and P value �.05 was considered significant.

RESULTS

Desmin immunohistochemistry showed that approxi-mately 80% of isolated skeletal muscle cells had a

Fig 1. Gross appearance of regenerated fascia at 3 months. (A) NC group (no cell seeding); notice the visible bowel loops under the thin

fascia. (B) FB (fibroblasts seeding) group without hernia; black arrows indicate relatively normal-looking fascia without any protrusion even

after marked stretching.

Table 1. Hernia Rate Detected by Mechanical Stretching

NC Group*(n � 21)

SK Group*†(n � 21)

FB Group*†(n � 24)

1 mo 4/7 2/7 2/82 mo 5/7 1/7 2/83 mo 7/7 2/7 2/8Total 16/21 (76.2%) 5/21 (23.9%) 6/24 (25%)

Abbreviations: NC, no cell seeding; SK, skeletal muscle cell seed-ing; FB, fibroblast seeding.

*P � .005.†P � .82.

1753BODY WALL REPAIR USING SIS SEEDED WITH CELLS

myogenic phenotype. The remaining cells had a fibro-blastic appearance (desmin negative). It took about 4weeks to achieve ideal cell number for further seeding.After seeding into the collagen gel, the cells formedmultilayer cellular sheets.

The adhesion between the graft and the peritonealcontents was mild. The graft was almost transparent inthe cases of a hernia, whereas it was opaque in thenonhernia group. The cell-seeded groups (SK and FBgroups) had a thicker graft than the group without cellseeding (NC group).

Before instillation of 200 mL saline, only 10, 3, and 3in NC, FB, and SK groups, respectively, showed hernia.After abdominal distension, 16 of 21 in the NC groupsshowed an abdominal wall hernia (76.2%). In the 16cases, 4 developed at 1 month, 5 at 2 months, and 7 at 3months, respectively. Five of the 21 grafts in the SKgroup and 6 of the 24 in the FB group showed abdominalhernia after distension. The hernia rate in cell-seededgroups did not increase with time (Table 1). The herniarate in the NC group was significantly higher than ineither the FB or SK group (P � .005). However, therewas no difference in the latter 2 groups (P � .82).

All grafts in the FB and SK groups had variousdegrees of tissue regeneration, which was evidenced bygraft opacity. In the grafts with hernia, only a part of thegraft showed thinning and fascial weakness. By contrast,the NC group showed near-total degradation of the scaf-fold (Fig 1).

Histologically, SIS grafts showed different degrees ofcell infiltration over time. However, there was no truemuscle bundle visible, with scattered skeletal musclecells at 3 months postoperatively. Some calcification wasnoted, but it decreased over time.

In the NC group, the scaffold was thin and with lesscell infiltration. Both SK and FB groups had thickerfascia and better cell infiltration. Even in the cases ofhernia in the FB and SK groups, the fascia still showedbetter cell regeneration than NC groups (Fig 2).

DISCUSSION

Several synthetic materials are available for repair oflarge soft tissue defects, including knitted polypropylenemesh, polytetrafluoroethylene, PTFE plus hyaluronicacid, polyester, and polyglycolic acid.2-5 Complicationsof wound infection, bowel fistula, erosion into abdominalviscera, repair failure, and mesh extrusion have beenreported. The estimated recurrence rate approximates25%.5

Naturally derived materials, including glutaraldehydetanned bovine pericardium6 and small intestine submu-cosa,7,8 have been tried in animal models. These bioma-terials are less susceptible to infection and cause lessforeign body response.9,10 However, the lack of strength

over time is a concern for clinical applications in whichadequate tensile properties are necessary. For this reason,it is important to understand not only the biologicalresponse to degradable biomaterials, but also the ex-pected mechanical properties of the implant and replace-ment tissues over time.

In this study, it took 1 month from harvesting the cellsto graft implantation. The SIS graft without cells de-graded gradually over 3 months. With cell seeding, thegrafts showed better mechanical strength. Acellular col-lagen-based matrix alone may be in-sufficient scaffoldfor abdominal wall reconstruction, as reported by Fauzaet al.1

Although the hernia rate in this study is still high (25%in FB group and 23.9% in the SK group), the hernia ratedid not increase with time. Besides, the cell-seededgroups with hernia showed only partial fascial protru-sion. Instead, there is near-total graft weakness in the NCgroup with hernia. Failure of neo-vascularization of thecell-seeded graft and the nature of the scaffold might

Fig 2. Microscopy of tissue-engineered fascia at 3 months. (A) NC

group, only a thin construct with some cell infiltration. (B) SK (skel-

etal muscle cells seeding) group without hernia, very thick construct

with dense cells infiltration (H & E, original magnification �200).

1754 LAI, CHANG, AND LIN

play important roles. It took 7 days to detect neovascu-larization over the cell-scaffold composite by immuno-histochemistry.11 Before the establishment of a goodblood supply, the seeded cells might not proliferate andthen reinforce the scaffold. The rate of degradation ofSIS varies from 4 to 12 weeks. With time progressingand a stable blood supply developing, the new fasciagains more tensile strength from the regenerated cellsand tissue. This may explain the initially high hernia ratein the FB and SK groups even though the rate did notincrease with time.

The comparison of the FB and SK groups is anotherinteresting finding in this study group. Most of the tensilestrength of the abdominal wall comes from the fascial

layer, which is composed mostly of densely arrangedfibroblasts instead of skeletal muscle cells. Cell seedingwith skeletal muscle cells on the scaffold might regen-erate a muscularlike tissue but the new tissue did notshow any advantage over fibroblast-seeded graft in termsof hernia rate. Further quantitative measurement andlonger follow-up should be included for further studies.

Engineered cellular constructs have better cell infil-tration and mechanical performance than acellularimplants. The success of a tissue-engineered compos-ite might be determined by both the strength of thescaffold and the angiogenesis to support the new cells.Tissue engineering may be a viable alternative forbody wall replacement.

REFERENCE

1. Fauza DO, Marler JJ, Koka R, et al: Fetal Tissue Engineering:Diaphragmatic Replacement. J Pediatr Surg 36:146-151, 2001

2. Tyrell J, Siberman H, Chandrasoma P, et al: Absorbable versuspermanent mesh in abdominal operations. Surg Gynecol Obstet 168:227-232, 1989

3. Lamb JP, Vitale T, Kaminshi DL: Comparative evaluation ofsynthetic meshes used for abdominal wall replacement. Surgery 93:643-648, 1983

4. Jenkins SD, Klamer TW, Parteka JJ, et al: A comparison ofprosthetic materials used to repair abdominal wall defects. Surgery94:392-398, 1983

5. Luijendijk RW, Hop WC, van den Tol MP, et al: A comparisonof suture repair with mesh repair for incisional hernia. N Engl J Med343:392-398, 2000

6. James NL, Poole-Warren LA, Schindhelm K, et al: Comparativeevaluation of treated bovine pericardium as a xenograft for herniarepair. Biomaterials 12:801-809, 1991

7. Clarke KM, Lantz GC, Salisbury SK, et al: Intestine submucosaand polypropylene mesh for abdominal wall repair in dogs. J Surg Res60:107-114, 1996

8. Prevel CD, Eppley BL, Summerlin DJ, et al: Small intestinalsubmucosa: Use in repair of rodent abdominal wall defects. Ann PlastSurg 35:374-380, 1995

9. Hiles MC, Badylak SF, Lantz GC, et al: Mechanical properties ofxenogeneic small intestinal submucosa when used as an aortic graft inthe dog. J Biomed Mater Res 29:883-891, 1995

10. Badylak SF, Kropp B, McPherson T, et al: SIS: A rapidlyresorbable bioscaffold for augmentation cystoplasty in a dog model.Tissue Eng 4:379-387, 1998

11. Park HJ, Yoo JJ, Kershen RT, et al: Reconstitution of humancorporal smooth muscle and endothelial cells in vivo. J Urol 162:1106-1109, 1999

1755BODY WALL REPAIR USING SIS SEEDED WITH CELLS