Upward migration of cultured autologous keratinocytesin Integra� artificial skin: a preliminary report

ISABEL JONES, FRCS; S. ELIZABETH JAMES, PhD; PHILIP RUBIN, FRCS; ROBIN MARTIN, PhD

The combination of cultured autologous keratinocytes with the dermal regeneration template Integra� could offerincreased possibilities for reconstructive surgery and wound healing. A single-step application of cells, centrifugeddeep into an Integra�-like matrix at the silicone–matrix junction, has been described but might prove technicallycomplex for clinical use. We have investigated the possibility of simplifying this procedure by applying cultured cellsdirectly to the underside of the Integra�or directly to the wound bed immediately prior to grafting. The objective wasto see whether cells would migrate through the matrix in an upward direction. We tested the principle of this conceptusing a pig wound healing model. Integra� was seeded directly with cultured cells and grafted onto fresh full-thickness wounds, or unseeded Integra�was applied to freshly excised wound beds that had just been seeded withthe same number of cells. Biopsies were taken at 3, 7, 11, and 14 days. Histological sections showed that the cellsmoved through the Integra� to give a confluent surface epithelium. Direct seeding onto the Integra� was the mostefficient method. Transduction of cultured autologous keratinocytes in vitro with a MFGlacZnls retrovirus confirmedthat the epidermis was derived from the cultured autologous keratinocytes. (WOUND REP REG 2003;11:132–138)

IntegraTM artificial skin is a commercially available dermal

regeneration template for use in reconstructive surgery

and in burns.1 It is a bilayer product. The deeper layer is a

collagen/chondroitin sulfate matrix that is replaced by new

collagen synthesized by fibroblasts that grow in from the

wound bed. The template matrix serves to regulate the

in-growth of new tissue and hence minimize scarring. The

purpose of the silicone sheet, which forms the superficial

layer, is to protect the wound from infection and excessive

loss of moisture and to limit granulation tissue formation.2

Once the collagen layer has been integrated into the wound

bed, the silicone layer begins to separate spontaneously. At

this point it can be peeled away and replaced with an ultra-

thin split-thickness skin graft (SSG) of approximately

0.1 mm. Whereas an average SSG measures 0.2–0.4 mm,

due to the underlying Integra� matrix, the ultra-thin SSG

does not contract to the extent that would be expected if it

were applied directly to the wound bed. This is supported

by a long-term study.3

Cultured autologous keratinocyte (CAK ) sheets have

been proposed as an alternative form of definitive wound

CAK Cultured autologous keratinocytes

C-GAG Collagen-glycosaminoglycan

FITC Fluorescein isothiocyanate

H & E Hematoxylin & eosin

SSG Spilt-thickness skin graft

From the Blond McIndoe Research Center, Queen Vic-toria Hospital, East Grinstead, United Kingdom.

This work was first reported, in part, at the 2000 ETRSmeeting: Jones IT, James L, Rubin P, Martin R.Retroviral labeling as evidence of epidermalregeneration from autologous disaggregatedcultured keratinocytes applied beneath Inte-gra� artificial skin. Wound Rep Reg 2000;8:A420.

Current addresses: Isabel Jones: Heatherwood Hospital,Ascot, Berkshire; P.R.: Department of PlasticSurgery, Nottingham City Hospital, Sheffield;R.M.: Smith & Nephew Group Research Center,Heslington, York, United Kingdom.

Correspondence: Dr. Liz James, PhD, Blond McIndoeCenter, Queen Victoria Hospital, East Grinstead,West Sussex, RH19 3DZ, United Kingdom. Fax: 44-1342-301701; Email: liz.j@blondmcindoe.demon.co.uk.

Copyright � 2003 by the Wound Healing Society.ISSN: 1067-1927 $15.00 + 0

132

cover for Integra� in place of an ultra-thin SSG. It was

hoped that the presence of the Integra�-derived neoder-

mis would eliminate the problems of low efficiency of take,

fragility, blistering, and contraction associated with CAK

sheets applied directly onto a wound bed.4–6 However, the

few existing clinical case reports document that unpre-

dictable efficiencies of CAK take onto Integra� remain

problematic.7

The development of CAK with Integra� as a one-

stage graft presents difficulties in timing for acute burns.

The cultured cells are only available in large numbers

some 3 weeks after injury. Integra� is most commonly

employed as a skin substitute following early excision.

Bio-integration of Integra� is only approaching comple-

tion after 3 weeks, which argues strongly against a one-

stage protocol in favor of a two-stage protocol when

cultured cells are to be employed. However, Integra� and

similar substitutes are now being explored for applica-

tions other than acute burns. Elective plastic surgical

reconstruction and the treatment of chronic wounds are

techniques that could benefit from combined epidermal

skin replacement.

As early as 1989, Yannas et al.8 reported that

trypsinized whole-skin-cell suspensions (uncultured

keratinocytes and fibroblasts) centrifuged into a colla-

gen-glycosoaminoglycan (C-GAG) matrix and grafted

onto guinea pigs were able to facilitate the regener-

ation of a healthy epidermis. More recently, CAK in

suspension have been centrifuged into an Integra�-like

C-GAG matrix before grafting so that the cells come

to rest just beneath the silicone surface. An epidermis

was regenerated beneath the protective silicone layer.

This has been reported as successful in Large White

pigs, both with cells in an uncultured state9,10 and after

cell number expansion following culture and passage

of the keratinocytes in vitro.10,11 However, centrifuga-

tion is a difficult technique, particularly if it is to be

applied to large 10 · 25 cm sheets of a commercially

available C-GAG matrix such as Integra� to make it

attractive for clinical use. In addition, it is important

to determine whether cell-seeding techniques could

work with commercially available off-the-shelf products

and not only with the material from individual

laboratories.

The aim of this preliminary study was to see whether a

simplified single-stage method could be developed for

commercially available C-GAG material. We applied CAK

either directly to the wound bed beneath Integra�, or to

the under-surface of the Integra� matrix before grafting

onto Large White pigs. Genetic labeling was used to show

that cells that proliferated within the matrix were derived

from the cultured cells.

MATERIALS AND METHODSAll work was conducted in keeping with the terms of the

Animals (Scientific Procedures) Act of 1986, and the

experimental design was cognizant of the need to optimize

animal welfare. The methodology for keratinocyte culture,

retroviral labeling, animal surgery, and grafting of Integra�in Large White pigs has been described12–14. Two pigs (six

wounds per pig) were used for the first experiment, which

studied the seeding of Integra� artificial skin. Wounds

distributed on an additional three pigs were used for the

second experiment to identify the origin of the epidermal

tissue on the wound surface. Briefly, split skin grafts

measuring 2 cm · 4 cm and 200–300 lm thickness were

taken from 6-week-old pigs 10 days prior to the date set for

grafting. Keratinocytes were isolated and grown in the

culture medium Opti-MEM supplemented with 1% heat

inactivated fetal calf serum and 0.5 mM Ca2+ (all from

Invitrogen-Gibco), and in the presence of irradiated 3T3

feeder cells,15 passaging the keratinocytes three times as

80% confluence was achieved. To accomplish gene mark-

ing with lacZnls, a retroviral producer cell line was

constructed in the amphotropic cell line PT6716 using a

suspension of MFGlacZnls virus from the ectotropic cell

line TELCeB17 to transduce the PT67 cells. Irradiated

PT67MFGlacZnls (clone 8) producers were used to

transduce porcine keratinocytes by employing the virus

producer line as irradiated feeder cells in place of the 3T3

fibroblasts, as described previously.12,13 Using three rounds

of transduction on irradiated PT67MFGlacZnls feeders,

the frequency of transduction was approximately 95%.

Methods for grafting Integra� onto pigs within rigid 4-cm

diameter PTFE chambers have been described.14 The

purpose of the rigid PTFE chamber is to isolate the wound

bed from any migration of epithelial cells from the wound

margin.

Seeding IntegraTM artificial skinOn the day before the procedure, flasks of subconfluent

(80%) autologous porcine keratinocytes were given 20 ml

fresh keratinocyte growth medium. The following morning

they were trypsinized and suspended in keratinocyte

growth medium at a concentration of 10–12 · 106 per ml.

A total of 0.5 ml of solution was applied to each 4-cm

diameter wound (12.6 cm2 total area). These quantities

were chosen to deliver approximately 5 · 105 keratino-

cytes per cm2 using a volume of solution sufficient to cover

the wound without flooding it. Figure 1a shows cells being

dripped from a needleless 1 ml syringe onto the under-

surface of Integra�, which had been equilibrated in a large

volume of normal saline to remove the isopropyl alcohol in

which the product is stored. The C-GAG surface was

carefully patted dry with a sterile dressing prior to adding

WOUND REPAIR AND REGENERATIONVOL. 11, NO. 2 JONES ET AL. 133

the cell suspension, which was then rapidly absorbed into

the sponge-like matrix. Paraffin-embedded sections of this

seeded material were prepared and stained with hematox-

ylin and eosin (H & E) to monitor the penetration of cells

into the matrix.

Grafting seeded IntegraTM and harvesting woundbiopsiesSix full-thickness wounds were created on the flanks of

each of the two animals, down to muscle fascia, using

protocol 3 previously described by Grant et al.14 Unseeded

Integra� was grafted in the standard way to wounds 2 and

5. Wounds 1 and 6 received Integra� with 0.5 ml of

autologous cell suspension evenly applied to the nonsilic-

onized surface. Wounds 3 and 4 had a similar volume of

autologous cell suspension applied directly to the wound

bed prior to grafting with Integra�. Each wound was

dressed with saline-soaked swabs applied with firm

pressure within the chamber and one silver nitrate-soaked

swab around the perimeter of each chamber.

Single biopsies from each wound were taken at days

4, 7, and 11, and two were taken at day 14. Biopsies were

taken with a size 15 scalpel blade to give a section of tissue

approximately 8 mm · 8 mm. These were taken sequen-

tially from the top left, bottom left, top right, and, finally,

the center and bottom right at day 14. The tissue was

immediately embedded in the polyvinyl alcohol tissue

support medium OCT (Tissue-Tek, Sakura Finetek, The

Netherlands) and frozen ready to cut the sections. Biopsy

material was examined using frozen sections cut at 15 lm.

H & E staining was used to examine the morphology of the

tissue and a mouse monoclonal antibody to the epithelial

a b

FIGURE 1. Application of a suspension of cultured epithelial cells to the underside of Integra�artificial skin. (a) Application of a suspension of

cultured porcine epithelial cells to the underside of a 4-cm diameter disc of Integra�was made with a needleless 1 ml syringe. A volume of

0.5 ml was applied delivering 5 · 105 cells cm)2. The Integra� is supported on a gauze dressing. (b) H & E stained section through the seeded

Integra� showing the distribution of the cultured cells immediately after seeding. The silicone layer was removed from the upper surface to

facilitate the sectioning procedure. The arrow marks a silicone remnant.

a b

c d

FIGURE 2. Control wound grafted with

unseeded Integra�. H & E stained sections

through biopsies taken (a) 4 days, (b)

7 days, (c) 11 days, and (d) 14 days after

grafting full-thickness porcine wounds with

the Integra� that was not seeded with

cultured epithelial cells.

WOUND REPAIR AND REGENERATIONMARCH–APRIL 2003134 JONES ET AL.

cell marker cytokeratin 14 (gift from the Center for

Cutaneous Research, London), was used in conjunction

with a fluorescein isothiocyanate (FITC) labeled sheep

anti-mouse secondary reagent (Sigma Chemical Company,

St. Louis, MO) to identify the presence of epithelial cells.13,18

Verification of the origin of epidermal tissueTo confirm that the epithelial tissue observed in the

wounds was indeed derived from the cultured keratino-

cytes, a second series of two pigs was grafted with

Integra�, which was seeded on the underside with autol-

ogous keratinocytes. The autologous keratinocytes were

either normal cultured cells or labeled cells that had been

transduced in culture with the retroviral gene transfer

vector MFGlacZnls. This vector permanently marks the

transduced cell and all its progeny with a stable new

genetic marker (lacZnls), which is localized to the cell

nucleus.

Unlabeled cells were used on three wounds on one pig,

with retrovirally labeled cells on three wounds on a second

a b

c d

FIGURE 3. Wounds grafted with Integra�seeded with cultured autologous keratino-

cytes. H & E stained sections through

biopsies taken (a) 4 days, (b) 7 days, (c)

11 days, and (d) 14 days after grafting full-

thickness porcine wounds with Integra�seeded with cultured autologous epithelial

cells. The arrow indicates the presence of

a small cyst of epithelial cells at day 7. By

day 11, large areas of epithelial cells can

be seen growing up through the neoder-

mis. At day 14, the epithelial growth had

reached the surface of the wound.

a b

c d

FIGURE 4. Wounds grafted with Integra�onto wound beds seeded with cultured

autologous keratinocytes. H & E stained

sections through biopsies taken (a) 4 days,

(b) 7 days, (c) 11 days, and (d) 14 days

after grafting wounds with Integra�applied following seeding of the wound

bed with cultured epithelial cells. The

arrows indicate the presence of small cysts

of epithelial cells at day 11. None could be

detected on day 7. By day 14, however,

the epithelial growth was also sufficient to

have reached the surface of the wound.

WOUND REPAIR AND REGENERATIONVOL. 11, NO. 2 JONES ET AL. 135

pig. A third pig had both labeled and unlabeled wounds to

control for interpig variation. Whole-wound biopsies were

harvested and examined for cells expressing the lacZ gene

by staining with the lacZ substrate X-Gal as previously

described, resulting in blue stained cells.12

RESULTSWhen cultured cells were seeded onto the under-surface of

the Integra� to give approximately 5 · 105 keratinocytes

per cm2, the cells were readily absorbed into the matrix.

Figure 1b shows a 10-lm paraffin, H & E stained section of

a sample of Integra� that had just been seeded with

cultured porcine keratinocytes. The cells lie mainly in the

lower half of the matrix when the material is placed onto

the wound with the silicone surface uppermost. Work by

Compton’s group9 suggested that 5 · 105 keratinocytes per

cm2 were sufficient to give a confluent epithelium at

2 weeks, although their technique differed because they

centrifuged cultured keratinocytes into an Integra�-like

matrix so that they came to rest just beneath the silicone.9–11

The six full-thickness wounds created on the flanks of

each of two animals included two control wounds with no

cultured cells, two wounds with Integra�-seeded with

autologous keratinocytes, and two wounds with Integra�where the wound-bed was seeded with the autologous

keratinocytes. Figures 2, 3, and 4 show representative

images from control wounds, seeded Integra� wounds,

and wound-bed seeded wounds, respectively, at (a) 4, (b) 7,

(c) 11, and (d) 14 days. For consistency, the figures shown

are all from one animal. The replicate wounds showed

identical features. There was 100% take of Integra� on all

wounds, although some degree of ‘‘chamber shift’’ was

noted.14 Initially, there was little macroscopic difference

between those wounds that received cultured cells and

those that had not; however, silicone separation appeared

to be earlier, by 1–2 days (days 8–9) in those wounds

seeded with keratinocytes than in the control wounds

(days 9–10).

Keratinocyte seeding of Integra� and, to a lesser

extent, of the wound bed, appeared to accelerate the

remodeling of the Integra� C-GAG matrix (Figures 3 and

4, respectively) when compared with controls (Figure 2)

on the same animals. The fibers of the Integra� matrix

were less obvious in the seeded matrices even by day 7

(Figures 3b and 4b) and noticeably so by day 11 (Figures

3c and 4c) when compared with the control wounds

(Figure 2b,c).

By day 14, large areas of what appeared to be

epidermal tissue were apparent within the Integra�matrix; some of these breaking out onto the wound

surface (Figures 3d and 4d). No such tissues appeared in

the control wounds that were grafted with Integra�without cultured cells (Figure 2d).

Figure 5 (a and b) shows that the tissue growth within

the seeded Integra� wound matrix at days 7 and 11,

respectively, does contain nests of epidermal keratino-

cytes because these express cytokeratin 14. Figure 5c

confirms the epidermal identity of tissue that has reached

the surface at 14 days. Figure 5d shows a section of a

wound at 11 days in which the Integra� was seeded with

lacZnls labeled cells at 5 · 105 cm)2. The material identified

as keratinocyte in origin by FITC staining for cytokeratin

14 is clearly derived from the autologous cultured cells by

virtue of staining blue with the lacZ substrate X-Gal, and is

not a consequence of an increased growth factor flux that

stimulates a population of wound bed epithelial material.

DISCUSSIONThese studies were conducted with the aim of discovering

whether a suspension of keratinocytes applied in culture

medium on the under-surface of Integra� could proliferate

and migrate upward through the Integra� matrix to form a

confluent epithelium on the surface beneath the protective

silicone. This is a much more straightforward procedure

than the previously described centrifugation of cells

through a C-GAG matrix to enable the cells to come to

rest at the silicone–collagen interface.9,11 Two approaches

were compared. Cells were either applied directly to the

underside of the Integra� or to the wound bed immedi-

ately prior to Integra� grafting. It was noted that the cell

suspension was more easily absorbed by the Integra� than

by the wound surface. The keratinocyte solution applied to

the wound bed could be clearly seen trickling down the

wound surface, which, in this animal model, was almost

90� from horizontal when the cells were applied. It was not

feasible to have the wounds more horizontal because both

flanks were being grafted.

By using retrovirally labeled keratinocytes, these

experiments provide evidence that the keratinocytes

migrating through Integra� do originate from the cultured

keratinocytes. The accelerated remodeling of the matrix in

the presence of cultured cells suggests that the seeding of

cultured epidermal cells might influence the migration of

host cells such as fibroblasts from the underlying wound

bed into the C-GAG matrix. This could be beneficial if it

accelerated the normal process of biointegration, but

would be undesirable if it perturbed the role of the C-GAG

biomaterial matrix in minimizing scar formation. This

aspect has not been examined in the present study, which

was performed solely to determine the propensity of the

cultured cells to migrate upward through the grafted

matrix in vivo. It remains to be seen to what extent human

WOUND REPAIR AND REGENERATIONMARCH–APRIL 2003136 JONES ET AL.

keratinocytes might behave in a similar fashion in a

clinical situation; however, experiments have been repor-

ted in which human keratinocytes were seeded into

Integra�, incubated in-vitro to allow cell proliferation, and

subsequently grafted onto athymic nude mice. These

seeded matrices were also able to generate a surface

epithelium.19 Nevertheless a method such as that explored

in the present animal study, which avoids a technically

difficult centrifugation and an in vitro incubation of a

seeded C-GAG, may offer a greater prospect for clinical

applications where a one-step dermal-epidermal graft is

desirable.

c

a b

d

FIGURE 5. The identity and origin of cul-

tured epithelial cells in Integra�. The

presence of epithelial cells among the

maturing Integra� matrix was confirmed

by immunohistochemistry using FITC-labe-

led antibody to detect the presence of

cytokeratin 14. (a) The presence of small

cysts in a day 7 biopsy from a wound

grafted with Integra� seeded with cul-

tured epithelial cells. (b) Epithelial cells

spreading throughout the neodermis by

day 11. (c) By day 14, epithelium has

reached and is spreading across the

wound surface. (d) lacZnls labeled epithe-

lial cells detected by staining with the

substrate X-Gal in a day 11 biopsy of a

wound grafted with Integra� seeded with

cultured autologous lacZnls labeled kera-

tinocytes.

WOUND REPAIR AND REGENERATIONVOL. 11, NO. 2 JONES ET AL. 137

ACKNOWLEDGMENTSThe authors acknowledge gifts of Integra� Artificial Skin

from Ethicon Inc. which were made without influence or

obligation to facilitate these studies. IJ was funded by a

Royal College of Surgeons of England Surgical Research

Fellowship and by the John Byrne Surgical Fellowship.

SEJ was funded by the Band Trust. This work was

supported by the East Grinstead Medical Research Trust

UK Reg. Charity no. 258154. We thank Professor Irene

Leigh (Center for Cutaneous Research, London) for

providing us with the cytokeratin 14 monoclonal antibody

(clone LL001).

REFERENCES1. Burke JF. Observations on the development and clinical use of arti-

ficial skin—an attempt to employ regeneration rather than scar for-

mation in wound healing. Jpn J Surg 1987;17:431–8.

2. Martin R, Bevan S, Boorman J, Grant I, James SE, Jones I, Parkhouse

N, Ng R, Rubin P, Woodward B. Cultured keratinocytes: experimental

and clinical directions in the quest for tissue engineering new skin. In:

Horch RE, Munster AM, Achauer BM, editors. Cultured human kera-

tinocytes and tissue engineered skin substitutes. Stuttgart, New York:

Thieme, 2001: 107–16.

3. Sheridan RL, Hegarty M, Tompkins RG, Burke JF. Artificial skin in

massive burns—results to ten years. Eur J Plast Surg 1994;17:91–3.

4. Woodley DT, Briggaman RA, Okeefe EJ, Inman AO, Queen LL,

Gammon WR. Identification of the skin basement-membrane auto-

antigen in epidermolysis bullosa-acquisita. N Eng J Med 1984;310:

1007–13.

5. Compton CC, Gill JM, Bradford DA, Regauer S, Gallico GG, O’Connor

NE. Skin regenerated from cultured epithelial autografts on full-

thickness burn wounds from 6 days to 5 years after grafting. A light,

electron microscopic and immunohistochemical study. Lab Invest

1989;60:600–12.

6. Woodley DT, Peterson HD, Herzog SR, Stricklin GP, Burgeson RE,

Briggaman RA, Cronce DJ, O’Keefe EJ. Burn wounds resurfaced by

cultured epidermal autografts show abnormal reconstitution of

anchoring fibrils. JAMA 1988;259:2566–71.

7. Pandya AN, Woodward B, Parkhouse N. The use of cultured autolo-

gous keratinocytes with integra in the resurfacing of acute burns.

Plast Reconstr Surg 1998;102:825–8.

8. Yannas IV, Lee E, Orgill DP, Skrabut EM, Murphy GF. Synthesis and

characterization of a model extracellular matrix that induces partial

regeneration of adult mammalian skin. Proc Natl Acad Sci USA

1989;86:933–7.

9. Butler CE, Orgill DP, Yannas IV, Compton CC. Effect of keratino-

cyte seeding of collagen-glycosaminoglycan membranes on the

regeneration of skin in a porcine model. Plast Reconstr Surg

1998;101:1572–9.

10. Butler CE, Yannas IV, Compton CC, Correia CA, Orgill DP. Compar-

ison of cultured and uncultured keratinocytes seeded into a collagen-

GAG matrix for skin replacements. Br J Plast Surg 1999;52:127–32.

11. Compton CC, Butler CE, Yannas IV, Warland G, Orgill DP. Organized

skin structure is regenerated in vivo from collagen-GAG matrices

seeded with autologous keratinocytes. J Invest Dermatol

1998;110:908–16.

12. Ng RL, Woodward B, Bevan S, Green C, Martin R. Retroviral marking

identifies grafted autologous keratinocytes in porcine wounds re-

ceiving cultured epithelium. J Invest Dermatol 1997;108:457–62.

13. Bevan S, Woodward B, Ng RL, Green C, Martin R. Retroviral gene

transfer into porcine keratinocytes following improved methods of

cultivation. Burns 1997;23 (7–8):525–32.

14. Grant I, Green C, Martin R. Strategies to improve the take of com-

mercially available collagen/glycosaminoglycan wound repair

material investigated in an animal model. Burns 2001;27:699–707.

15. Rheinwald JG, Green H. Serial cultivation of strains of human epi-

dermal keratinocytes: the formation of keratinizing colonies from

single cells. Cell 1975;6:331–43.

16. Miller AD, Chen F. Retrovirus packaging cells based on 10A1 murine

leukemia virus for production of vectors that use multiple receptors

for cell entry. J Virol 1996;70:5564–71.

17. Takeuchi Y, Cosset FLC, Lachmann PJ, Okada H, Weiss RA, Collins

MKL. Type-c retrovirus inactivation by human-complement is deter-

mined by both the viral genome and the producer cell. J Virol 1994;

68:8001–7.

18. Grant I, Warwick K, Marshall J, Green C, Martin R. The co-application

of sprayed cultured autologous fibrin sealant in a porcine wound

model. Br J Plast Surg 2002;55:219–27.

19. Kremer M, Lang E, Berger AC. Evaluation of dermal-epidermal skin

equivalents (‘composite-skin’) of human keratinocytes in a collagen-

glycosaminoglycan matrix (Integra artificial skin). Br J Plast Surg

2000;53:459–65.

WOUND REPAIR AND REGENERATIONMARCH–APRIL 2003138 JONES ET AL.