Inflammatory cell subpopulations in keloid scars

British Journal of Plastic Surgery (2001), 54, 511-516

9 2001 The British Associationof Plastic Surgeons doi: 10.1054/bjps.2001.3638

9 .lrls.

JOUn.

A,

OF

P-ASTIC

SU.aS.Y

I

Inflammatory cell subpopulations in keloid scars D. E. Boyce, J. Ciampolini*, E Ruge*, M. S. C. Murison1" and K. G. Harding*

Department of Plastic Surgery, Diana, Princess of Wales Children's Hospital, Birmingham; *Wound Healing Research Unit, University Department of Surgery, University of Wales College of Medicine, Cardiff,"and ?Welsh Regional Plastic Surgery and Burns Unit, Morriston Hospital, Swansea, UK SUMMARY. The aim of this study was to investigate the contribution of lymphocytes and macrophages to keloid scarfing by morphologically characterising inflammatory cell subpopulations in keloid scars in comparison with normal skin. We took 3 mm punch biopsies from the anterior forearms of eight normal healthy volunteers. Eight keloid scars were excised using an intralesional technique. All tissue was snap frozen in liquid nitrogen and serial sections were stained with a panel of anti-inflammatory cell monoclonal antibodies. The numbers of macrophages and lymphocytes and the proportions of the subpopulations were compared. Higher numbers of both macrophages and lymphocytes were found in keloid dermis (P = 0.01 and P = 0.02, respectively (Mann-Whitney U-test)). There was no significant increase in the expression of the lymphocyte activation markers, CD25 and CD27. However, there was a significantly higher CD4 + : CD8 + (Th : Ts) ratio (P = 0.046) in keloid tissue. This suggests that an imbalance in these inflammatory cell subpopulations may contribute to keloid scarfing in man. 9 2001 The British Association of Plastic Surgeons Keywords: keloid scarring, inflammatory cells, lymphocytes, macrophages, wound healing, scar formation.

Normal healing involves complex interactions between a variety of cell types in order to restore tissue integrity. Inflammatory cells have been shown, in animal studies, to play an integral role in this process, but relevant evidence with respect to human healing remains sparse. Macrophages have been recognised to play a central role in the transition between the inflammatory phase and the subsequent phases of healing for over a quarter of a century. 1 However, little has been learned about the role of lymphocytes, which are a significant component of the mononuclear cell infiltrate. It is likely that these cells also exert a diverse range of regulatory functions on the healing process by virtue of their ability to secrete an extensive array of messenger molecules. 2 Keloid scarfing is often described as the end result of an 'over-exuberant' healing response. Much research into the pathogenesis underlying keloid scarfing has concentrated on the potential pathological role of the keloid fibroblast and extracellular matrix. 3-9 However, mononuclear cells, which 'orchestrate' the normal healing response, may also play a major role by influencing these cells via the secretion of a wide range of growth factors and cytokines. Evidence for such a role in human tissue is extremely limited. 1~ The hypothesis underlying this study is that if mononuclear cells do play a significant part in this process, then various subsets may play different roles. The aim, therefore, was to investigate the role that these cells may play in the pathogenesis of keloid scarfing by morphologically characterising mononuclear cell

subpopulations in keloid scars and comparing them with normal skin. Efforts have been specifically directed towards lymphocytic subsets that appear to be relevant to the process of wound healing. In particular, it has been demonstrated in a murine model that CD8 + T suppressor (Ts) lymphocytes may play a part in 'down-regulating' the healing response. 13 Conversely, it has been shown that the presence of CD4 + T helper (Th) lymphocytes or their cytokines 'upregulate' healing in an in vitro model. 14 We have previously shown that normal human wound healing is characterised by an initial high CD4+:CD8 + ratio, which declines as healing progresses. 12 Low levels of potentially 'down-regulatory' Ts cells in the initial stages of acute wound healing are associated with the initial attempt of the wounded tissue to close the acute defect. High levels of Ts cells are found in chronic non-healing wounds, which are characterised by a low CD4 + :CD8 + ratio. 15 In this study, possible differences in the CD4 + :CD8 + ratio between keloid and normal dermis were examined. In addition, the specific markers of lymphocyte activation, CD25 (interleukin-2 receptor) and CD27, were selectively identified.

Materials and methods

Biopsies It was not considered ethical to obtain normal skin from patients with keloid scarfing, since they may potentially develop a further keloid at the biopsy site. Therefore, after ethics committee approval, 3 mm punch biopsies were obtained (under 1% lignocaine local anaesthesia) from the forearms of eight healthy human volunteers. Eight specimens of keloid tissue were obtained intraoperatively

Presented at the British Association of Plastic Surgeons Summer Meeting/ESPRAS 3rd European Appointed Meeting, Birmingham, UK, 6 July 2000.

511

512 from eight patients undergoing therapeutic excision of their keloid scars using an intralesional technique (Figs 1 and 2). The definition used to identify a keloid scar was that originally described by Alibert in 1806: an abnormal raised area of scar tissue that invades surrounding undamaged skin) 6 Lack of scar regression at 2 years is further evidence of keloid, as opposed to hypertrophic, scarring) 6 Patient details are given in Table 1. Processing o f biopsies Specimens were snap frozen in liquid nitrogen and 6 Ixm cryostat sections were mounted on poly-L-lysine treated microscope slides. Slides were stored desiccated at -22~ for up to 14 days prior to staining. Serial sec-

British Journal of Plastic Surgery tions were fixed in dry acetone, washed three times in phosphate buffered saline, and incubated in optimal dilutions of monoclonal antibody (Dako| Ltd, High Wycombe, UK) for 30min. The antibody panel and specificities are described in Table 2, and were applied to serial sections in the order shown. They were then washed three times in phosphate buffered saline, and antibody localisation was identified by a standard streptavidin-biotin peroxidase technique (Vector | Laboratories, Peterborough, UK), with the final reaction product developed using 3,3'-diaminobenzidine. The sections were counterstained with Ehrlich's haematoxylin, dehydrated, cleared and mounted in DPX mounting medium. Identification and interpretation o f lymphocytic infiltrate The lymphocytic infiltrate of each biopsy was quantified by counting x 40 magnification views. Numbers of positive cells were identified in an orderly manner consisting of six fields in the subepithelial region and six fields directly beneath these, giving a total of 12 fields for each specimen. Results are expressed as the mean and median (interquartile range) number of positive cells per x 40 magnification field. The markers of lymphocyte activation, CD25 and CD27, are expressed as a percentage of the total number of T lymphocytes identified. Statistics Statistical comparisons were performed using Minitab | 10.51 Xtra computer software (Minitab Inc., PA, USA) using a Mann-Whitney U-test; P < 0.05 was considered to be significant. Table 1 Details of patients with keloid scars

Figure 1--Case 1. Earlobe keloid.

Case Age Sex Previous medical history 1

12

F

2 3 4 5 6

10 13 10 17 29

M M F M M

7 8

19 32

F F

Site

nil earlobe nil abdomen nil postauricular nil earlobe nil postauricular previous sternum spontaneous keloid nil earlobe nil sternum

History of keloid

2 years post-piercing 3 years post-laparotomy 3 years post-pinnaplasty 4 years post-piercing 3 years post-pinnaplasty 2 years post-excision 2 years post-piercing 4 years post-cystexcision

Table 2 Immunocytochemistry primary monoclonal antibody panel

Figure 2--Case 6. Sternal keloid.

Antigen

Cellulardistribution

CD45 CD19 CD3 CD4 CD8 CD25

all leucocytes B lymphocytes T lymphocytes T helper/inducer lymphocytes;macrophages T suppressor/cytotoxiclymphocytes activatedT lymphocytes/macrophages(interleukin-2 receptor) activatedT lymphocytes macrophages,monocytes

CD27 CD68

Inflammatory cell subpopulations in keloid scars

513

Results

In each keloid specimen there was a distinct dermal layer with normal architecture underlying the epidermis. Cells exhibiting a macrophage morphology and stained by anti-CD68 monoclonal antibody were found diffusely distributed within the keloid dermis, with no distinct pattem of organisation (Fig. 3). There were significantly higher numbers of macrophages present in keloid (mean 23.6, median 25.6 (20.6-27.9)) than in normal dermis (mean 10.8, median 10.2 (9.1-12.2)) (P=0.01) (Fig. 4). Deep to this dermal area was a much thicker abnormal area, which formed the collagen 9 bulk of the keloid. In this deeper area, further macrophages could be seen between thickened bundles of collagen, with cellular processes extending between these bundles and interdigitating with other cells.

Unlike macrophages, cells of lymphocytic morphology, small round cells with little cytoplasm and a typically round nucleus, were found preferentially in perivascular areas (Figs 5 and 6). No B lymphocytes were identified in any of the tissue specimens. With respect to T lymphocytes, there were significantly more CD3 + T lymphocytes in keloid dermis (mean 10, median 10.1 (5.5-15.30)) than in normal dermis (mean 3.1, median 3.6 (2.1-3.8)) (P = 0.02) (Fig. 7). Further analysis of this lymphocyte population using subset-specific monoclonal antibodies showed a significant difference in the T h : T s (CD4 + :CD8 +) ratio, with keloid tissue exhibiting a significantly higher ratio (mean 3.3, median 2.7 (1.8-4.7)) than normal skin (mean 1.7, median 1.6 (1.2-2.2)) (P = 0.046) (Fig. 8).

Figure

5--Keloid scar: CD3 stain for T lymphocytes(x 10).

Figure 3---Keloidscar: CD68 stain for macrophages(x 10). 36.

35

Normal Dermis

30.

30

26.

25

Keloid Dermis

9149 x

x

a

20.

2

2

15'

§

I0'

L)

6'

:$ e9

9

9a

§

10

0

5 0

Figure

dermis.

12345678

12345578

Patient Number

Patient Number

4---Numbers of macrophages identified in normal and keloid

Figure 6--High-power view of perivascular CD3+ T lymphocytes (x40).

514

British Journal of Plastic Surgery 20

Normal Dermis

20

15

g

~"

10

+

16

f~

g B r

Keloid Dermis

5

9

B

s

0 12346678

12345678

Patient N u m b e r

Patient Number

Figure 7--Numbers of T lymphocytes identified in normal and keloid dennis.

8

6-

Normal Dermis

Keloid Dermis

7-

7

6-

6

,o_

o 5

r 4

0

3-

9

2" 1

r(3 oo

4-

a

3-

9 9

9

9149

21-

0 12345578

12346678

Patient N u m b e r

Patient N u m b e r

Figure 8---CD4 + :CD8 + T lymphocyte ratio in normal and keloid dermis.

Expression of the lymphocyte activation antigens, CD25 and CD27, was relatively low, with no significant differences between normal and keloid tissue. A mean of 7.1%, median 5% (0.7-13.8%), of lymphocytes expressed the CD25 (interleukin-2 receptor) antigen in keloid tissue, compared with a mean of 14.1%, median 15.7% (4.821.4%), in normal tissue. A mean of 28.8%, median 18.9% (10.%-46.8%), of lymphocytes expressed CD27 in keloid tissue, compared with a mean of 18%, median 17.4% (0-32.2%), in normal skin.

Discussion

The inflammatory phase of healing is the first and an essential stage in the healing process. After 24-48h, according to animal experiments, the inflammatory cells are mainly mononuclear cells, comprising macrophages and lymphocytes. Macrophages are known to be essential for normal healing. 1 Their initial role is in the inflammatory and debridement phase that precedes fibroplasia, ]7 and much work has shown that they play a major role in the orchestration of subsequent dermal repair. 18 Lymphocytes also comprise a significant proportion of the

mono nuclear cell infiltrate, but much less is known of their role in the healing process. Keloids are often referred to as the result of an 'overexuberant' healing response, in which collagen deposition persists in excess of collagen lysis for an indefinite period] Much research into the pathogenesis of keloids has concentrated on the fibroblast as the 'effector' cell in the keloid healing response. However, in the normal healing process the actions of fibroblasts are 'directed' by cytokines derived mainly from macrophages but also from lymphocytes. In view of the integral role of inflammatory cells in the healing response, it is logical to assume that they may play a part in orchestrating the pathology of this 'abnormal' healing process. Immune function has already been implicated in the aetiology of keloids. Reduced production of interleukin-2, interferon-oL and interferon-',/by peripheral blood monocytes has been found in patients with a susceptibility to keloids, while increased production of tumour necrosis factor-a and interferon-[3 was also observed. 7 Autoantibodies against host fibroblasts have been found in patients with keloids, 19 and increased deposition of IgG, IgM and IgA has been demonstrated in keloid dermis. 2~ Of particular relevance to this study is original work by Martin and Muir that showed increased numbers of inflammatory cells in keloid tissue. 1~ Indeed, leucocytederived factors can have considerable effects on fibroplasia. Collagen gene expression can be induced by interleukin-l, tumour growth factor-[3 and platelet-derived growth factor, whereas interferon-',/ and prostaglandins can reduce collagen gene expression. 18'2] Tumour necrosis factor-a has been shown to decrease levels of collagen synthesis in experimental wounds. 22 Furthermore, keloid fibroblasts in vitro are much more susceptible than controis to the effects of such factors.4'6 In view of the known importance of macrophages in the healing process, it is logical to assume that these cells would be present in greater numbers in the keloid dermis. This, indeed, was the case. Furthermore, macrophages were found interdigitating with fibroblasts and other cells within the main collagenous core of the keloid. It is logical to assume that these cells have a significant influence on local fibroplasia. However, in vivo regulation of skin healing and scar formation consists of a complex interaction between a wide range of cytokines. Such cytokines potentially arise from lymphocytes as well as macrophages. It is as yet unknown how these factors work with, or against, each other, and over what temporal sequence. It is important to note that a significantly higher number of T lymphocytes were observed in keloid tissue. Lymphocytes represent a heterogeneous group with a wide variety of functions. Natural-killer cells comprise a small component of the lymphoid pool and are concerned with viral and tumour immunity. B lymphocytes are responsible for 'antibody-mediated' immunity and T lymphocytes for 'cell-mediated' immunity. T lymphocytes are, in themselves, extremely heterogeneous but can broadly be divided into a CD4 + T helper/inducer subset, which are generally regarded as 'upregulators' of the immune response, and a CD8 + T suppressor/cytotoxic subset, which generally have the opposite effect. Th lymphocytes can be further divided into three subsets: Th-0,

Inflammatory cell subpopulations in keloid scars Th-1 and T h - 2 . 23 Th-0 lymphocytes are the precursors of Th-1 and Th-2 lymphocytes, which are characterised by differing patterns of cytokine secretion. Th-1 cells synthesise interleukin-2, interferon-~ and tumour necrosis factor- B, which are not secreted by Th-2 cells. Conversely, only Th-2 cells synthesise detectable amounts of interleukin-4, interleukin-5, interleukin-6 and interleukin-13fl 4 Recent work has shown that this is an oversimplification: certain Th-2 lymphocytes can suppress the pro-inflammatory effects of Th-1 lymphocytes, 25 and certain CD8 + lymphocytes can mimic Th-1 and Th-2 type functions and play an important part in determining the pattern of cytokine production by Th lymphocytes. 26 This is a complex and exciting area of research, where much is to yet be learned, but little is known about the function of these immune cells in normal skin, let alone keloid scars. T lymphocytes are often considered to be the major source of lymphocyte-derived cytokines in an inflammatory response such as that found after tissue injury. 2 Theories regarding T-cell contributions to wound healing can be formulated by examining the in vitro effect of T-cell lymphokines on the various cells within the wound milieu. In particular, many lymphocyte-mediated cytokines are capable of modulating fibroplasia, for example tumour growth factor-J3, tumour necrosis factor-J3 and interferon-~. 27-3~ It would be reasonable to assume, therefore, that lymphocytes may play a part in the abnormal fibroplastic response that characterises keloids. In vivo data outlining the potential role of lymphocytes in wound healing has been limited to a murine model. It has been shown that total T-cell depletion using monoclonal antibodies results in decreased wound breaking strength and reparative collagen synthesis. 3~ Furthermore, selective depletion of CD4 + Th lymphocytes had no obvious effect, whereas selective depletion of CD8 + Ts lymphocytes resulted in enhanced wound healing. 3~ This led these authors to suggest that CD8 + Ts lymphocytes are responsible for a 'down-regulation' of the healing response. Conversely, Wojciak and Crossan have shown that the presence of CD4 + Th tymphocytes or their cytokines accelerates wound healing in a rat epitenon wound model, t4 These data are consistent with our previous findings in human tissue: acute wound healing is characterised by an initial high C D 4 + : C D 8 + ratio, which declines as healing progresses. 12 Low levels of potentially 'down-regulatory' Ts cells in the initial stages of acute wound healing are, therefore, associated with the initial attempt of the wounded tissue to close the acute defect, whereas the high levels in the later stages of normal healing may contribute to the 'switching off' of this process. The discovery of a consistently low CD4 § : CD8 + ratio in chronic wounds, 15 indicating high levels of potentially 'down-regulatory' Ts lymphocytes, is also consistent with this assumption, since these cells may potentiate the chronicity of the wound. Data indicating a potential role of lymphocytes in the aetiologies of keloid and hypertrophic scarring are extremely limited. ~~ The absence of B lymphocytes in normal skin and keloid scars reported by Martin and Muir ~~ and Castagnoli et al ~1 has been further verified by our work. These authors also reported significant numbers of T lymphocytes in human wounds and keloid

515 scars, and suggested that these cells may be important in wound healing and scar formation. Our study examined how specific subsets of T lymphocytes differ between normal and keloid tissue, and found a significantly higher CD4 + :CD8 + ratio in keloid dermis than in normal skin. We therefore suggest that these observations in abnormal scar tissue represent further human evidence to corroborate the murine, in vitro and previous human studies described above, i.e. that the over-exuberance of this healing response may, in part, be due to an imbalance between an 'upregulatory' CD4 + Th subset and a 'downregulatory' CD8-- Ts subset of wound lymphocytes. We would not simply infer that Th cells help the keloid-type healing response or that Ts cells suppress it, as these terms originate from the modulation of the immune response. However, we would suggest that subpopulations identified by these markers might play differing roles in the abnormal keloid healing response. In summary, we have demonstrated high numbers of macrophages and T lymphocytes within the dermis of keloid scars, together with a significantly higher T h : T s ratio than that in normal skin. These data are in accordance with previous evidence suggesting that such a ratio is associated with an 'upregulation' of the fibroplastic response. We suggest, therefore, that imbalance in these inflammatory-cell subpopulations may contribute significantly to the over-exuberant healing response characterised by these abnormal scars.

Acknowledgements The authors thank Dr K. Moore for advice, his ideas and use of his research laboratories, and Mr J. M. Porter, Consultant Plastic Surgeon, Sandwell NHS Trust, West Midlands, UK.

References 1. Leibovich SJ, Ross R. The role of the macrophage in wound repair. A study with hydrocortisone and antimacrophage serum. Am J Pathol 1975; 78: 71-100. 2. Schaffer M, Barbul A. Lymphocyte function in wound healing and following injury. Br J Surg 1998; 85: 444-60. 3. Diegelmann RF, Cohen IK, McCoy BJ. Growth kinetics and collagen synthesis of normal skin, normal scar and keloid fibroblasts in vitro. J Cell Physiol 1979; 98: 341-6. 4. Babu M, Diegelmann R, Oliver N. Keloid fibroblasts exhibit an altered response to TGF-[3. J Invest Dermatol 1992; 99: 650-5. 5. Ehrlich HP, Desmouliere A, Diegelmann RF, et al. Morphological and immunochemical differences between keloid and hypertrophic scar. Am J Pathol 1994; 145: 105-13. 6. Kikuchi K, Kadono T, Takehara K. Effects of various growth factors and histamine on cultured keloid fibroblasts. Dermatology 1995; 190: 4-8. 7. O'Sullivan ST, O'Shaughnessy M, O'Connor TPE Aetiology and management of hypertrophic scars and keloids. Ann R Coil Surg Engl 1996; 78: 168-75. 8. Bettinger DA, Yager DR, Diegelmann RF, Cohen IK. The effect of TGF-[3 on keloid fibroblast proliferation and collagen synthesis. Plast Reconstr Surg 1996; 98: 827-33. 9. Yoshimoto H, Ishihara H, Ohtsuru A et al. Overexpression of insulin-like growth factor-1 (IGF-1) receptor and the invasiveness of cultured keloid fibroblasts. Am J Pathol 1999; 154: 883-9. 10. Martin CW, Muir IE The role of lymphocytes in wound healing. Br J Plast Surg 1990; 43: 655-62. 11. Castagnoli C, Trombotto C, Ondei S e t al. Characterization of T-cell subsets infiltrating post-burn hypertrophic scar tissues. Burns 1997; 23: 565-72.

516 12. Boyce DE, Jones WD, Ruge F, Harding KG, Moore K. The role of lymphocytes in human dermal wound healing. Br J Dermatol 2000; 143: 59-65. 13. Barbul A, Breslin RJ, Woodyard JP, Wasserkrug HL, Efron G. The effect of in vivo T helper and T suppressor lymphocyte depletion on wound healing. Ann Surg 1989; 209: 479-83. 14. Wojciak B, Crossan JE The effects of T cells and their products on in vitro healing of epitenon cell microwounds. Immunology 1994; 83: 93-8. 15. Boyce DE. Macrophages and lymphocytes in human wound healing. MD thesis, University of Wales, Cardiff, UK, 1999. 16. Muir IFK. On the nature of keloid and hypertrophic scars. Br J Plast Surg 1990; 43: 61-9. 17. Jian-Ping C, Harris B, Falanga V. Recruitment of mononuclear ceils into wounded skin: mechanism and modulation. In Clinical and Experimental Approaches to Dermal and Epidermal Repair: normal and chronic wounds. New York: Wiley-Liss, 1991: 243-56. 18. Riches DWH. Macrophage involvement in wound repair, remodelling and fibrosis. In Clark RAF, ed. The Molecular and Cellular Biology of Wound Repair. New York: Plenum Press, 1996: 95-131. 19. Janssen de Limpens AM, Cormane RH. Keloids and hypertrophic scars - immunological aspects. Aesthetic Plast Surg 1982; 6: 149-52. 20. Kischer CW, Shetlar MR, Shetlar CL, Chvapil M. Immunoglobulins in hypertrophic scars and keloids. Plast Reconstr Surg 1983; 71: 821-5. 2l. Friedman DW, Boyd CD, Mackenzie JW, Norton E Olson RM, Deak SB. Regulation of collagen gene expression in keloids and hypertrophic scars. J Surg Res 1993; 55: 214-22. 22. Regan MC, Kirk SJ, Hurson M, Sodeyama M, Wasserkrug HL, Barbul A. Tumor necrosis factor-tx inhibits in vivo collagen synthesis. Surgery 1993; 113: 173-7. 23. Mosmann TR, Coffman RL. TH1 and TH2 cells: different patterns of lymphokine secretion lead to different functional properties. Annu Rev Immunol 1989; 7: 145-73. 24. Abbas AK, Murphy KM, Sher A. Functional diversity of helper T lymphocytes. Nature 1996; 383: 787-93. 25. Nicholson LB, Kuchroo VK. Manipulation of the Thl/Th2 balance in autoimmune disease. Curr Opin Immunol 1996; 8: 837-42. 26. Kemeny DM, Noble A, Holmes BJ, Diaz-Sanchez D. Immune regulation: a new role for the CD8 + T cell. Immunol Today 1994; 15: 107-10. 27. Barbul A, Regan MC. The regulatory role of T lymphocytes in wound healing. J Trauma 1990; 30: $97-100. 28. Mustoe TA, Pierce GF, Thomason A, Gramates P, Sp0m MB, Deuel TE Accelerated healing of incisional wounds in rats induced by .... a'ansforming growth factor-t3. Science 1987; 237: 1333-6.

British Journal of Plastic Surgery 29. Granstein RD, Murphy GF, Margolis RJ, Byme MH, Amento EE Gamma-interferon inhibits collagen synthesis in vivo in the mouse. J Clin Invest 1987; 79: 1254-8. 30. Peterson JM, Barbul A, Breslin RJ, Wasserkrug HL, Eft'on G. Significance of T-lymphocytes in wound healing. Surgery 1987; 102: 300-5. 31. Fishel RS, Barbul A, Beschomer WE, Wasserkrug HL, Efron G. Lymphocyte participation in wound healing. Morphologic assessment using monoclonal antibodies. Ann Surg 1987; 206: 25-9. 32. Barbul A, Regan MC. Immune involvement in wound healing. Otolaryngol Clin North Am 1995; 28: 955-68. 33. Barbnl A. Role of T cell-dependent immune system in wound healing. Progr Clin Biol Res 1988; 266: 161-75.

The Atrthors Dean Edward Boyce MD, FRCSEd, FRCS, Specialist Registrar in Plastic Surgery Department of Plastic Surgery, Diana Princess of Wales Children's Hospital, Steelhouse Lane, Birmingham B4 6NH, UK.

Jacnpo Ciampolini FRCS, Research Fellow Fiona Ruge BSc, Research Officer Keith G. Harding MB, MRCGP, FRCS, Professor of Wound Healing Research Wound Healing Research Unit, University Department of Surgery, University of Wales College of Medicine, Heath Park, Cardiff, UK. Maxwell M. S. C. Murison FRCSEd, FRCS(Plast), Consultant Plastic Surgeon Welsh Regional Plastic Surgery and Bums Unit, Morriston Hospital, Morriston, Swansea SA6 6NL, UK. Correspondence to Mr Dean Boyce, Hand Fellow, Department of Plastic Surgery, Withington Hospital, Nell Lane, West Didsbury, Manchester M20 2LS, 94. Paper received 4 February 2000. Accepted 23 April 2001, after revision. Published online 13 July 2001.