- Email: [email protected]
Myofibroblasts and apoptosis in human hypertrophic scars: The effect of interferon-α2b
Myofibroblasts and apoptosis in human hypertrophic scars: The effect of interferon-α2b Bernadette Nedelec, PhD, Heather Shankowsky, RN, Paul G. Scott, PhD, Aziz Ghahary, PhD, and Edward E. Tredget, MD, MSc, Edmonton, Alberta, Canada
Background. Hypertrophic scars (HSc) are a dermal fibroproliferative disorder that leads to considerable morbidity. Preliminary evidence suggests that interferon (IFN) may improve HSc clinically. The aims of this study were (1) to compare the cell density in HSc and in wounds that heal without the development of HSc (normotrophic scars), (2) to examine the presence of myofibroblasts and apoptosis in normotrophic and HSc scars over time, and (3) to determine if the systemic administration of IFN-α2b can induce apoptosis. Methods. Two groups of patients underwent serial tissue biopsies. Six burn patients were studied prospectively by obtaining biopsy specimens from wound granulation tissue, normal skin, post-burn HSc, and normotrophic scars (healed donor sites). A second patient group with HSc was treated with systemic IFNα2b and had biopsy material taken before, during, and after IFN therapy. The tissue was analyzed by immunohistochemical staining for α-smooth muscle actin (α-SMA) and in situ DNA fragmentation terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay for apoptosis. Results. The total numbers of fibroblasts in HSc were found to be similar to granulation tissue and twice that of normal skin and normotrophic scar. Over time the numbers of cells in HSc tissue decreased toward normal skin levels. There was a significantly higher percentage of fibroblasts staining for α-SMA in HSc as compared with normotrophic scar or normal skin obtained from the same patient (P > .05). Serial biopsy specimens of resolving HSc tissue obtained from the patients who received systemic IFN-α2b showed a general reduction in total number of fibroblasts and myofibroblasts associated with a significant increase in the percentage of apoptotic cells compared with normal dermis from the same patient. Conclusions. HSc tissues have greater numbers of fibroblasts and myofibroblasts than normal skin and normotrophic scars. As HSc remodels, the numbers of fibroblasts and myofibroblasts reduces, possibly by the induction of apoptosis. Systemic IFN-α2b may contribute to the resolution of HSc in part by the enhanced induction of apoptosis. (Surgery 2001;130:798-808.) From the Department of Surgery and the Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada
THE DEVELOPMENT OF HSc after a burn or other major trauma to the skin is a dermal fibrotic disorder where red, raised, pruritic lesions distort the skin and are often associated with the formation of contractures.1 Scar contractures reduce range of
Supported by the Firefighters’ Burn Trust Fund of the University of Alberta Hospitals, the Harold and Emilie Tucker Trust Fund of the University of Alberta Hospitals (B.N.), Alberta Heritage Foundation for Medical Research (E.E.T., Scholar), and the Medical Research Council of Canada (P.G.S., A.G., E.E.T.). Accepted for publication April 7, 2001. Reprint requests: Edward E. Tredget, MD, MSc, FRCSC, 2D3.81 WMHCS, University of Alberta Hospital, 8440-112 St, Edmonton, Alberta, Canada T6G 2B7. Copyright © 2001 by Mosby, Inc. 0039-6060/2001/$35.00 + 0 11/56/116453 doi:10.1067/msy.2001.116453
798 SURGERY
motion, resulting in delayed return to work and reintegration into society.2 The treatment of HSc and scar contractures includes conservative treatments such as corticosteroid injections, pressure therapy, splinting, and serial casting,1,3 however, these treatments are time-consuming, uncomfortable, and often incompletely successful. Surgical excision or release of HSc also is associated with a high rate of recurrence.4 Gabbiani et al5 and Ehrlich et al6 previously have identified myofibroblasts as a specialized fibroblast that is present in HSc. This cell possesses contractile properties and specific morphologic features of both the fibroblast and smooth muscle cell and commonly is defined as a fibroblast that stains positively for α-SMA.5,7 In a phase II clinical trial in burn patients suffering from HSc, systemically administered IFN-
Surgery Volume 130, Number 5 α2b has been associated with the induction of scar remodeling and maturation. 8 In vitro IFN has been shown to decrease fibroblast proliferation,9 down-regulate the production of collagen and fibronectin,10-12 and enhance the production of collagenase.13 Reduction of the rate and extent of contraction in vitro has been shown using IFN-α, IFN-γ, and IFN-β.14-16 Intralesional injections of IFN-α or IFN-γ led to a reduction in the keloid mass, 17,18 HSc, and Dupuytren’s disease nodules.19 In Dupuytren’s disease, myofibroblast staining for α-SMA was reduced in intensity and distribution after 4 weeks of intralesional IFN-γ treatment.19 Previous research has correlated the disappearance of myofibroblasts in wound healing with the induction of apoptotic cell death.20,21 In addition, the more rapid disappearance of myofibroblasts in open wounds covered with full-thickness flaps compared with a split-thickness flap has been attributed to the accelerated induction of apoptosis. 22 Conversely, a lack of induction of apoptosis may contribute to the excessive cellularity and contraction associated with HSc.23 It is our hypothesis that HSc contains excessive numbers of fibroblasts and myofibroblasts, which reduce slowly as the scars remodel and mature into normotrophic scars. One mechanism by which IFNα2b may be of therapeutic value in the treatment of fibroproliferative disorders such as HSc is through the accelerated induction of apoptosis, which has been previously demonstrated in activated T lymphocytes, squamous cell skin cancer cells, and myeloma cells.24-26 This hypothesis was tested prospectively in 2 patient groups. In 1 group, burn wounds were biopsied serially in the granulation phase, after the development of HSc, and during the remodeling phase, and compared with normal skin from the same patient, as well as the skin graft donor site, as a wound that healed without the development of HSc (normotrophic scar). A second group of burn patients in whom HSc had already developed after burn injury underwent serial biopsy specimens of normal skin and HSc before, during, and after 24 weeks of subcutaneous IFN-α2b therapy.8 Serial sections were evaluated histologically for total fibroblast cell number as well as for myofibroblasts, before measurement of apoptosis in the sections was done over time with the TUNEL assay. MATERIAL AND METHODS Tissue biopsies. Patients recruited into this study had sustained thermal injuries that necessitated skin grafts and were treated in the Firefighters’ Burn Treatment Unit of the University
Nedelec et al 799
of Alberta Hospital. Patients were recruited and tissues were sampled after informed consent had been obtained and according to the guidelines of the University of Alberta, Faculty of Medicine Research Ethics Board. The 6-mm punch biopsy specimens of wounds and normal skin were collected from 2 patient groups. In the first group, biopsy specimens of granulation tissue, normal skin, healed burn wounds, and healed donor sites were obtained from the same patient serially over time (group 1, Table I). The burn wound granulation tissue and normal skin first were biopsied while the patient was undergoing skin grafting (mean, 19 days post burn). The healed burn wound and donor site biopsy specimens were obtained when the patient returned to the outpatient clinic for follow-up. Wounds that developed red, raised, pruritic scar tissue were designated HSc, and wounds and donor sites that healed without the development of HSc were designated normotrophic scars. A total of 6 patients and 24 biopsy specimens were studied in this group. In the second group, biopsy specimens were obtained from 9 burn patients for whom HSc had developed over more than 5% of their total body surface area (TBSA) after a thermal injury and who had been recruited into a phase II clinical trial of IFNα2b (INTRON A, Schering, Kenilworth, NJ) (group 2, Table I).8 The HSc and normal tissue biopsy specimens (6-mm punch) were obtained before the subcutaneous administration of 1 × 106 units of IFN-α2b per day for a week, then 2 × 106 units 3 times/wk for 24 weeks. The definition of clinically significant HSc was established by using a previously established scar rating system, the modified Vancouver Burn Scar Assessment (VBSA).27 A scar rating of greater than 3, indicating significant abnormalities in skin pliability, erythema, scar height, and color, was considered hypertrophic. Biopsy specimens of HSc were obtained every other month and at the conclusion of therapy. A second biopsy of normal skin was taken before starting the IFN therapy and at the termination of IFN-α2b treatment.8 All biopsy specimens were placed immediately in 4% paraformaldehyde for immunohistochemical analysis. Materials. Biotin-16-2'-deoxyuridine-5'-triphosphate, proteinase K, dUTP, 4-nitroblue tetrazolium chloride, 5-bromo-4-chloro-3-indolyl phosphate, and terminal deoxynucleotidyl transferase (TdT) were obtained from Boehringer Mannheim (Indianapolis, Ind). The RNase-free DNase was obtained from Promega (Madison, Mich). Caseinblocking buffer was obtained from Cambridge Research Biochemicals (Northwich, Cheshire). Anti-α-SMA (clone 1A4), bovine serum albumin
800 Nedelec et al
Surgery November 2001
Table I. Patient demographic data
Patient Group 1 1 2 3 4 5 6 Mean ± SD Group 2 7 8 9 10 11 12 13 14 15 Mean ± SD
Time to final biopsy of normal skin and burn wound (mo)
Sex
Age
TBSA (%)*
VBSA†
M M M M M M
67 35 28 42 49 41 43.7 ± 13.4
5 18 85 3 25 40 29.3 ± 30.5
0 1 12 6 6 9 5.7 ± 4.6
2 12 12 3 4 3 6.0 ± 4.7
M M M M M M F F M
49 26 49 36 31 25 46 10 29 33.4 ± 13.0 NS
32 70 30 65 60 45 20 40 85 49.7 ± 21.5 NS
12 6 13 10 8 9 12 11 7 9.8 ± 2.4 P < .05
17 20 17 4 20 6 16 12 8 13.3 ± 6.1 P < .05
Demographic information for the 2 patient groups is listed including gender, age, TBSA, and the number of months after injury that normal skin and burn wounds were biopsied. *TBSA, Total body surface area injured from thermal injury; †VBSA, Vancouver Burn Scar Assessment clinical scar rating.
(BSA), 3, 3'-diaminobenzidine (DAB), 3-aminopropyl-triethoxy-silane (Aptex), monoclonal peroxidase-anti-peroxidase soluble complex antibodies, ExtrAvidin-Alkaline Phosphatase and normal goat serum were obtained from Sigma Chemical Company (St Louis, Mo). The bridging goat F(ab')2 fragment to mouse IgG (whole molecule) was obtained from ICN Pharmaceuticals Inc (Aurora, Ohio). Immunohistochemistry. The tissue biopsy specimens fixed in 4% paraformaldehyde were embedded in paraffin and cut into 4- to 5-µm sections and placed on Aptex-coated slides. All sections were heated to 60°C for 10 minutes then dewaxed in xylene and rehydrated. Endogenous peroxides were inactivated by treating all sections with 10% H2O2 in methanol for 6 minutes. After washing with PBS, sections were incubated for 30 minutes at room temperature (RT) in blocking agent (10% vol/vol normal goat serum, 5% BSA wt/vol in PBS) and then incubated overnight at 4°C with the appropriate antibody. The α-SMA antibody was diluted 1:100 in PBS/Tween 0.05% solution. The control specimens were incubated with plasmacytoma cell culture supernatant diluted 1:10 in PBS/Tween 0.05% solution. After warming to RT the sections were incubated with goat F(ab')2 fragment to mouse IgG (whole molecule; 1:50 dilution in PBS/Tween 0.05% solution) for 1.5 hours,
washed with PBS, then incubated with monoclonal peroxidase-anti-peroxidase soluble complex at a 1:200 dilution for 1.5 hours and washed with PBS. The colorimetric reagent was hydrogen peroxide/DAB, and all sections were counterstained with hematoxylin. The α-SMA–positive fibroblasts were counted on 10 randomly selected high-power fields per section. The total number of fibroblasts, based on morphology, also were counted on the same section, thus allowing calculation of the percentage of myofibroblasts. Smooth muscle cells associated with blood vessels and erector pili were excluded from the α-SMA positive fibroblast count and the total fibroblast count. TUNEL assay. The in situ demonstration of DNA fragmentation was performed by using the TUNEL assay as previously described.28 All sections were treated as described for immunohistochemistry up to the point of rehydration. After washing with ddH2O, the sections were incubated for 20 minutes at 37°C in 2 × sodium chloride and sodium citrate (SSC) (pH 7.0) then washed several times in ddH2O. All sections were bathed for 5 minutes at RT in 10 mmol/L Tris-HCl solution (pH 8.0) then digested for 15 minutes at 37°C with proteinase K (20 µg/mL in 10 mmol/L Tris-HCl). After washing several times in ddH2O the tissue sections were fixed for 4 minutes in 4% paraformaldehyde then
Surgery Volume 130, Number 5
washed several times with ddH2O. The positive control material was incubated in DNase buffer for 5 minutes at RT then digested with DNase (1 µg/mL) at 37°C for 30 minutes. The positive control was washed with ddH2O then all sections were incubated for 5 minutes at RT in TdT buffer (pH 7.2). All sections were incubated in reaction buffer (94.5 µl ddH2O, 7.0 µl TdT buffer, 4.4 µl Bio-16dUTP, 4.4 µl dUTP, and 1.5 µl TdT) for 1 hour at 37°C in a humidified chamber, except the negative control where the TdT was omitted. To terminate the reaction the sections were transferred to 2 × SSC for 15 minutes at RT. All sections were washed with PBS solution then incubated for 15 minutes at RT in blocking agent (1:5 casein blocking buffer, 10% vol/vol normal goat serum, and 5% BSA wt/vol in PBS). The sections then were incubated at RT for 30 minutes in Extra-Avidin phosphatase (1:250) in PBS/Tween 0.5% solution. The sections were washed with PBS then incubated with colorimetric buffer (0.1 mol/L Tris, 0.1 mol/L NaCl, 0.25 mol/L MgCl2; pH 9.5) for 5 minutes at RT. The colorimetric reagent was 4-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate. The percentage of positive cells was calculated by counting the number of positive cells in 10 randomly selected fields and comparing it with the total number of cells within a serial section stained with hematoxylin. Cells that were obviously endothelial cells or epithelial cells—such as in vessels, hair follicles, sweat glands, sebaceous glands, and the outer epithelial layer—were eliminated from both positive cell counts and total cell counts. Statistics. A 2-tailed Student paired t test was used for all comparisons of matched data obtained from 2 time points in the same patient in group 2, and a nonpaired Student t test was performed for comparisons of 2 variables between groups 1 and 2. Analysis of variance was used for the analysis where more than 2 comparisons were made as indicated in the figure legends and text. Results are reported using P value, mean, and standard deviation as indicated. RESULTS Based on clinical criteria for rating the burn scar, 4 of the 6 patients from group 1 had HSc develop (VBSA scale > 3) in burn wound sites and 2 patients did not, thus meeting the criteria for normotrophic healing, as indicated in Table I. None of the patients from group 1 had HSc in their donor sites. All of the patients in group 2 were involved in the phase II IFN trial and had extensive HSc over >5% of the TBSA as one of the inclusion criteria. There was a poor correlation between the
Nedelec et al 801
size of the original burn and the severity of HSc using the VBSA scale (r2 = .12; P = .21, n = 15). Group 1 differed from group 2 patients in that most of the former had less severe HSc based on clinical scar rating (VBSA, Table I) (5.7 ± 4.6 vs 9.9 ± 2.4; P < .05) and were enrolled in the study within the first year following burn injury (mean, 6.0 ± 4.7 months; range, 2-12 months); whereas, most of the group 2 patients were studied more than 12 months after injury (mean, 13.3 ± 6.1 months; range, 4-20 months) and failed to respond to conservative management of more severe HSc as measured by the VBSA, which was a significant clinical and statistical difference between the 2 groups. Many group 2 patients also suffered larger burn injuries, were somewhat younger, and included 2 females, one being a 10-year-old child (Table I). Although large differences exist in these demographic variables, the broad variation likely accounted for the lack of statistically significant differences between groups 1 and 2 in these specific features (Table I). In group 1, burn patients were studied prospectively for Hsc development. The total number of fibroblasts per high-power field was significantly increased in the granulation tissue biopsied prior to wound closure (mean, 19.0 ± 27.8 days; range, 475 days) compared with the normal skin (12.0 ± 2.6 vs 4.7 ± 0.5 fibroblasts/high-power field; P < .001) and in HSc as compared with the normal skin in the same patient (10.5 ± 1.0 vs 4.7 ± 0.5; P < .05). This increase in cell density was not significantly different between HSc and granulation tissue (10.5 ± 1.0 vs 12.0 ± 2.6, P > .05, not significant). In wounds that healed without HSc development, significantly lower fibroblast cell density was found and normotrophic scar cell densities were not significantly different than normal dermis (6.8 ± 0.9 vs 4.7 ± 0.5; P > .05, not significant) (Fig 1, A). When the total number of fibroblasts in HSc was compared as a function of time following burn injury, there was a progressive reduction between early (04 months post-burn), mid (5-18 months), and late (19-30 months) HSc (Fig 1, B), which was statistically significant when comparing the early phase of HSc and the later remodeled phase (P < .01). Histologically, the distribution of the α-SMA positive fibroblasts or myofibroblasts in HSc was nonuniform and occurred in the more densely populated cell regions, but was not exclusive to any particular region of HSc (Fig 2). However, fewer αSMA positive myofibroblasts generally were visible in regions where more organized collagen fibers were present. Positive staining also was seen consistently in the endothelial cells of blood vessels
802 Nedelec et al
Surgery November 2001
A
A B
B Fig 1. Total fibroblasts per high-power field differs among various types of wound healing and returns to a more normal level as the scar matures. Total fibroblasts per highpower field was averaged from 10 randomly sampled fields. A, Total fibroblasts in granulation tissue and HSc relative to normal dermis (12.0 ± 2.6 and 10.5 ± 1.0 vs 4.7 ± 0.5). Normotrophic scar did not differ significantly from normal dermis (6.8 ± 0.9). B, The gradual reduction in total fibroblasts as a function of time, where early scar is 0-4 months post-burn (13.1 ± 2.0; n = 5), mid is 5-18 months (8.5 ± 0.4; n = 25), and late is 19-30 months (7.6 ± 0.5; n = 20).
where α-SMA is characteristically expressed and in the epidermal cells, which has previously been described in epithelial cells from other sources.7 Quantitative analysis of the percentage of αSMA–positive staining myofibroblasts in the initial biopsy obtained from group 1 (patients in whom HSc developed) and group 2 (before treatment with IFN) patients demonstrated considerable variability within the HSc sections, ranging from 1% to 65%, with an overall mean of 23.2% ± 5.2% (Fig 3). However, all patients who suffered burns over more than 70% of the TBSA had more than 35% of the total fibroblast count positive for myofibroblasts. Linear regression of myofibroblast numbers compared with the size of the original burn wound revealed a significant proportion of the variability in the percentage of myofibroblasts in HSc was positively related to the size of the original wound (r2 = .60; P < .05, n = 9), but poorly correlated with the severity of clinical burn scar assessment (r2 = .05; P = .43, n = 15) or the time after injury that the biopsy was obtained (r2 = .10; P = .39, n = 9). Significantly increased numbers of myofibroblasts were seen in HSc tissues as compared with normal skin from the
C
Fig 2. Localization of α-SMA–positive fibroblasts within HSc tissue. Positive staining of fibroblasts for α-SMA was localized to regions of disorganized collagen fibrils and to a lesser extent where collagen fibers and fiber bundles were present. A, B, and C are from the same patient biopsy but different areas of the tissue, all stained for α-SMA. A, Intense fibroblast staining in the center of a nodular region where there are disorganized collagen fibrils and capillaries on the outer edge of the nodule. B, Area where the α-SMA–positive fibroblasts are arranged in parallel alignment, with the collagen fibrils aligned in a waved pattern but parallel to the fibroblasts. C, Region where more normal collagen fibers and fiber bundles are visible. Although the vessels at the center of the panel are intensely stained there are few stained fibroblasts and the tissue is markedly less densely populated (×100 magnification).
same patients who had clinically significant HSc (23.2% ± 5.2% vs 6.2% ± 1.4%; n = 13; P < .01) (Fig 3). In wounds that healed without the Hsc development (6 donor site wounds and 2 burn wounds),
Surgery Volume 130, Number 5
Fig 3. Percentage of α-SMA–positive fibroblasts in HSc relative to normotrophic scar and normal dermis. Thirteen Hsc biopsies (4 from group 1, 9 from group 2) and normal dermis from the same patient and 8 normotrophic scars (6 donor site wounds and 2 burn wounds) are shown where 10 randomly chosen high-power fields were quantitated per biopsy. The bars indicate the mean for each group. Statistically significant differences in the means for each group are depicted on the graph (ANOVA).
the percentage of myofibroblasts was significantly lower than in the HSc tissues (6.1% ± 2.4% vs 22.9% ± 5.2%; n = 8; P < .05). In 2 patients whose burns healed without developing HSc who were included in the normotrophic group, the number of α-SMA positive myofibroblasts was 1% and 13%, respectively. Comparison of the number of myofibroblasts from normotrophic healed wounds in the first group of patients with the normal dermis from both groups revealed no significant differences. In group 2 patients who had large numbers of HSc and who received IFN-α2b, a reduction in the total fibroblast cell density was seen when comparing the initial biopsy with the final biopsy after 24 weeks of IFN therapy (9.9 ± 4.4 vs 6.7 ± 2.2 fibroblasts/high-powered field; n = 8; P < .05) (Fig 4, A). Comparison of the percentage of myofibroblasts in the HSc at the time of the initial biopsy relative to the end of the IFN trial showed a statistically significant reduction in 4 of 9 patients (data not shown) and an overall reduction in the mean proportion before and after IFN treatment, but the group effect was not statistically significant (26.9% ± 6.6% vs 16.0% ± 2.8%; n = 9) (Fig 4, B). The reduction in the number of fibroblasts in the dermal extracellular matrix that occurs as wound healing proceeds through the granulation phase to immature HSc, before remodeling into the normotrophic or mature scar, may occur through the induction of apoptosis.21 As discussed
Nedelec et al 803
earlier, one mechanism by which IFN-α2b may facilitate the maturation of HSc may be the induction of apoptotic cell death, which was investigated in our study using the TUNEL assay on serial sections of the same tissue (Fig 5). Although it previously has been reported that apoptotic cells in normal rat skin generally were not seen,21 apoptotic cells were present consistently in normal dermis of human burn survivors. The proportion of apoptotic cells from all HSc biopsies was lower than normal skin in early HSc (75.1% ± 25.7%; n = 5), slightly greater than normal in 5 to 18 months after injury (110.3% ± 15.8%; P < .05, n = 24), and significantly higher later in the Hsc remodeling phase (19-30 months) (221.6% ± 31.2%; P < .05, n = 19) (Fig 6). The effect of IFN-α2b treatment on the induction of apoptosis was examined by comparing the rates of apoptosis in the pre-treatment biopsy, the peak or highest percentage of apoptotic cells following IFN-α2b treatment and the final biopsy following treatment, with the normal skin biopsy (Fig 7). Prior to IFN therapy, the HSc demonstrated a higher proportion of apoptotic cells (33.4% ± 6.5% vs 20.9% ± 4.5%; n = 9), which was again increased significantly during administration of IFN-α2b, as compared with normal dermis (43.5% ± 5.6% vs 20.9% ± 4.5%; n = 9; P < .01, analysis of variance [ANOVA]). Following IFN-α2b treatment the proportion fell towards that in normal skin (26.3% ± 4.7% vs 20.9% ± 4.5%; n = 9; not significant). DISCUSSION Although histologic analysis of human hypertrophic burn scars and wound healing models in animals collectively have suggested that HSc resembles granulation tissue in the high cellularity of the extracellular matrix, this study is the first prospective, quantitative, longitudinal analysis in human beings that provides statistically significant confirmation of the hypothesis that early immature HSc are hypercellular and that as they remodel and mature the fibroblast cell density reduces to resemble normal skin, in part, through the induction of apoptosis. Wounds that heal over shorter periods and that often are not as deep into the dermis, such as the uncomplicated skin graft donor sites, do not become hypercellular and heal with a more normal collagen fiber bundle orientation. These normotrophic scars clinically and histologically resemble normal skin from the same patient (data not shown). Thus, the persistent high fibroblast cell density in HSc as compared with normotrophic scars likely is a very important factor contributing to the pathogenesis of this fibroproliferative disorder. Functional comparisons carried out in our lab-
804 Nedelec et al
A
B Fig 4. Total fibroblasts per high-power field of HSc tissue decreased after IFN-α2b treatment. A, Total fibroblasts in HSc tissue sections from the same patients before and 6 months after systemic administration of IFN-α2b (9.9 ± 4.4 vs 6.7 ± 2.2; P < .05, n = 8, paired Student t test). B, Percentage of α-SMA–positive fibroblasts decreased after IFN-α2b treatment. The number of α-SMA–positive fibroblasts was compared in HSc tissue sections from the same patient before and 6 months after systemic administration of IFN-α2b. Individual patient comparisons demonstrated a significant reduction in the percentage of positive cells in 4 of 9 patients; however, the overall reduction did not reach a level of significance (26.9% ± 6.6% vs 16.0% ± 2.8%; n = 9).
oratory of the HSc fibroblasts as compared with the paired normal fibroblast from the same patient have shown significant differences in collagen and collagenase production, decorin synthesis, and transforming growth factor-β (TGF-β) production.1 These phenotypic differences on a per cell basis may be exaggerated in vivo where our data indicate that approximately twice as many cells are present
Surgery November 2001
in the extracellular matrix of HSc as in normal skin. For example, increases in fibroblast TGF-β mRNA and protein production in vitro and in vivo are approximately 20% greater than the normal fibroblasts.29 This difference, when combined with the understanding that HSc have approximately twice the cell density of normal skin, likely has a significant effect on the total amount of local TGF-β produced in the early stages of wound healing and may contribute greatly to the magnitude of HSc formation. Our data also have confirmed the presence of the myofibroblast in HSc, similar to previous histologic descriptions of HSc,30 and longitudinal wound healing studies in animals.31 Significantly greater numbers of myofibroblasts were present in HSc as compared with normotrophic scar and normal skin. The strong correlation with the numbers of myofibroblasts in the healing postburn HSc with the size of the original burn injury suggests they are an important component of more complicated wound healing. Previous investigations have demonstrated increased amounts of the fibrogenic growth factor TGF-β in HSc tissues locally32 and in the serum of patients examined in this study,8 which others have found induces α-SMA in vivo.33 Other cytokines and growth factors such as plateletderived growth factor and tumor necrosis factor-α are likely present at some point during the healing of burn wounds such as investigated here, but do not appear to induce α-SMA in vivo.33 The poor correlation between myofibroblast numbers, the size of the burn wound, and the scar rating system may result from limitations of the VBSA scale. The need for more objective instruments to measure scar quality and suggestions for improvements in the clinical assessment of scars has been addressed based on the outcome of these studies.34 The HSc is characterized by the presence of nodules that contain thin collagen fibrils.30,31 It previously has been suggested that fibroblasts that stain positive for α-SMA are restricted to these nodular structures.6 In our study, fibroblasts staining for α-SMA generally were concentrated in scar regions that contained poorly organized collagen fibrils and a high density of cells commonly within these nodular structures, but also in the more mechanically stressed whorls that surrounded the nodules. In both the nodular structures and the mechanically stressed regions, thin collagen fibrils surrounded the myofibroblasts. In regions where more organized collagen fiber bundles were present, fewer α-SMA–stained fibroblasts were observed. The thin collagen fibrils and more organized fiber
Nedelec et al 805
Surgery Volume 130, Number 5
A
B
C
D
Fig 5. Representative examples of apoptotic cells within HSc at various times after burn injury relative to their normal dermis. The TUNEL assay identifies apoptotic cells after colorimetric development as blue cells. A, Example of HSc in the early period after a burn injury (0-3 months). B, HSc 19 months or later after a burn injury. There are significantly more positive cells than in A. C, Example of normal dermis, which has more positive cells than the early HSc (A) but fewer than in the late HSc (B). D, Control tissue section where terminal transferase was omitted from the reaction (×250 magnification).
bundles often were present within the same tissue section. These findings suggest that the myofibroblast phenotype also may be regulated by the organization of the extracellular matrix. In our previous work, many regions of poorly organized immature collagen fibers in the HSc stained poorly for the small proteoglycan, decorin, whereas αSMA was increased. This suggests a potential recip-
rocal interrelationship of the myofibroblast with this and possibly other components of the extracellular matrix and as well with TGF-β, which is normally bound to decorin in uninjured skin and mature scar.35,36 Our data also support the hypothesis of Gabbiani and Desmouliere37 that myofibroblasts appear in HSc prior to the development of high
806 Nedelec et al
Fig 6. The percentage of apoptotic cells in HSc increases as a function of time after burn injury relative to patientmatched normal controls. The percentage of apoptotic cells when compared with normal in early HSc (0-4 months post-burn injury) (75.1% ± 25.7%; n = 5), during the mid period (5-18 months post-burn) (110.3% ± 15.8%; n = 24), and late time period (19-30 months postburn) (221.6% ± 31.2%; n = 19) (asterisk = P < .05).
levels of apoptosis and may be a differentiated fibroblast committed to apoptotic cell death. The fact that the numbers of myofibroblasts in HSc correlate well with the size of the burn injury suggests that myofibroblasts may represent a somewhat greater proportion of the fibroblasts in large or deep wounds, whereas in normotrophic scars, the smaller numbers of myofibroblasts suggest that resident fibroblasts make up a greater proportion of the extracellular matrix and lead to more normal dermal healing resembling the normal skin. The recent recognition of the importance of bone marrow-derived fibrocytes, which can populate the wound, suggests that for patients with very large and deep injuries, the source of fibroblasts may not be the same as in uncomplicated wound healing.38 Further investigation will be required to confirm this suggestion, however. In our study the measurement of apoptosis by TUNEL was associated with substantial variability due to the heterogeneous distribution of apoptotic cells in the extracellular matrix, which was overcome in part by the analysis of large numbers of sections from many patients at different times. Significant increases in apoptosis late in the Hsc remodeling phase, however, establish apoptosis as an important mechanism by which fibroblast cell numbers are normalized. Thus a delay in apoptosis may be an important factor in the prolonged and severe nature of HSc seen in many of our patients.8 Wassermann and colleagues39 have demonstrated that peripheral blood mononuclear cells from patients who develop HSc have increased levels of the proto-oncogene Bcl-2 and reduced local tissue
Surgery November 2001
Fig 7. The percentage of apoptotic cells increased significantly in interferon-treated patients relative to normal dermis. The percentage of apoptotic cells in HSc biopsies determined by TUNEL before IFN treatment (Pre Tx), at the point of maximal apoptosis during IFN treatment (peak), 4 weeks after IFN treatment, as compared with normal skin biopsies obtained before IFN treatment. The percentage of apoptotic cells after the administration of IFNα2b was significantly increased relative to normal (43.5% ± 5.6 vs 20.9% ± 4.5, P < .01, n = 9, ANOVA).
expression of the apoptosis-inducing proteins Fas and interleukin-1 converting enzyme (ICE). To date, preliminary phase II clinical trials suggest that systemic administration of IFN-α2b improves healing by promoting HSc remodeling, resulting in more supple, pliable tissue.8 Serial biopsy specimens of HSc and normal skin from the same patient over a 7-month period, after conventional therapy had failed, demonstrated a reduction in the total fibroblast cell number in HSc. This was associated with a similar reduction in the number of α-SMA positive fibroblasts and an increase in the maximal amount of apoptosis in HSc during IFN therapy. These findings may be due to a direct reduction in fibroblast proliferation as we have observed in vitro11,12 and/or the induction of apoptosis. Without a statistically significant difference between apoptosis in HSc before and after IFN therapy, however, it cannot unequivocally be established which is the contributing factor in situ. It is difficult to draw definitive conclusions from this study because of the considerable differences between patients regarding the Hsc site, the severity of the original injury, and the time post-burn prior to commencing IFN therapy. In addition, in a preliminary phase II clinical trial where a limited number of patients were recruited and studied in situ, substantial variability arose in the distribution of myofibroblasts and apoptotic cells within HSc. Our previous work in full thickness excisional wounds in guinea pigs with a double-blinded, controlled administration of IFN or its vehicle via intraperitoneal osmotic pumps afforded more rig-
Surgery Volume 130, Number 5
orous controls and demonstrated a significant increase in apoptosis in the IFN-treated group only, however, which peaked well after re-epithelialization of the healing wounds, similar to the human HSc examined here.40 Unfortunately, further investigation of interferon’s apoptotic effect in patients with HSc is limited here by the fact that in contrast to group I patients, most of the interferon-treated patients completed therapy long after their initial wounding (13.3 ± 6.1 months, Table I), at a time when the initiation of spontaneous apoptosis and reduction in the number of myofibroblasts begins to occur as evidenced in non-interferon–treated group I patients during this period (Fig 6). In the future, interferon’s potential beneficial effects in inducing apoptosis to reduce cellularity in HSc may be appreciated more easily if therapy is commenced earlier after injury, when apoptosis is delayed and less than in the normal skin (Fig 6). Based on our findings of patients in group I, however, both our human data and animal investigations have suggested that myofibroblasts within HSc or other wounds represent a differentiated fibroblast important in wound remodeling but one that already is committed to apoptosis. The myofibroblast numbers are reduced after IFN therapy. In addition, our study was unable to establish if the maximum increase in fibroblast total cell number preceded the peak in myofibroblast cell numbers in HSc, but did suggest that maximal apoptosis occurred late in the HSc phase, providing additional evidence that apoptosis is an important process in scar remodeling that leads to a reduction in overall cellularity. Previously Gabbiani and colleagues have used qualitative methods to suggest that there is a correlation between the expression of α-SMA in fibroblasts and apoptosis.22 Similar findings have been reported in animal models of liver fibrosis41 and bleomycin-induced pulmonary fibrosis.42 In the future, microinjection studies of α-SMA offer the potential to determine the exact function of this specific actin isoform and subsequently its role in HSc.43 The assessment of the potential for IFN-α2b to assist in preventing HSc and promoting its remodeling through the induction of apoptosis or alteration in myofibroblast function will further benefit from the results of the double-blind placebo-controlled trial that presently is underway. We acknowledge the assistance of D. Carmel (Department of Oral Health Sciences) in preparation of tissue sections for histologic study. REFERENCES 1. Tredget EE, Nedelec B, Scott PG, Ghahary A. Hypertrophic scars, keloids and contractures. The cellular and molecular basis for therapy. Surg Clin N Am 1997;77:701-30.
Nedelec et al 807
2. Engrav LH, Covey MH, Dutcher KD, Heimbach DM, Walkinshaw MD, Marvin JA. Impairment, time out of school, and time off from work after burns. Plast Reconstr Surg 1987;79:927-32. 3. Ridgway CL, Daugherty MB, Warden GD. Serial casting as a technique to correct burn scar contractures. J Burn Care Rehabil 1991;12:67-72. 4. Darzi MA, Chowdri NA, Kaul SK. Evaluation of various methods of treating keloids and hypertrophic scars: a 10year follow up study. Br J Plast Surg 1992;45:374-9. 5. Gabbiani G, Ryan GB, Majno G. Presence of modified fibroblasts in granulation tissue and their possible role in wound contraction. Experientia 1971;27:549-50. 6. Ehrlich HP, Desmouliere A, Diegelmann RF, Cohen K, Compton CC, Garner WL, et al. Morphological and immunochemical differences between keloid and hypertrophic scar. Am J Pathol 1994;145:105-13. 7. Skalli O, Ropraz P, Trzeciak A, Benzonana G, Gillessen D, Gabbiani G. A monoclonal antibody against α-smooth muscle actin: a new probe for smooth muscle differentiation. J Cell Biol 1986;103:2787-96. 8. Tredget EE, Shankowsky HA, Pannu R, Nedelec B, Iwashina T, Ghahary A, et al. Transforming growth factor-β in thermally injured patients with hypertrophic scars: effects of interferon α-2b. Plast Reconstr Surg 1998;102:1317-28. 9. Jimenez SA, Freundlich B, Rosenbloom J. Selective inhibition of human diploid fibroblast collagen synthesis by interferons. J Clin Invest 1984;74:1112-6. 10. Ghahary A, Shen YJ, Scott PG, Tredget EE. Expression of fibronectin messenger RNA in hypertrophic and normal dermal tissues and in vitro regulation by interferon alfa-2b. Wound Repair Regen 1993;1:166-74. 11. Harrop AR, Ghahary A, Scott PG, Forsyth N, Uji-Friedland A, Tredget EE. Regulation of collagen synthesis and mRNA expression in normal and hypertrophic scar fibroblast in vitro by interferon-γ. J Surg Research 1995;58:471-7. 12. Tredget EE, Shen YJ, Liu G, Forsyth N, Smith C, Harrop AR, et al. Regulation of collagen synthesis and messenger RNA levels in normal and hypertrophic scar fibroblasts in vitro by interferon alfa-2b. Wound Repair Regen 1993;1:156-65. 13. Ghahary A, Shen YJ, Nedelec B, Scott PG, Tredget EE. Interferons gamma and alpha-2b differentially regulate the expression of collagenase and tissue inhibitor of metalloproteinase-1 messenger RNA in human hypertrophic and normal dermal fibroblasts. Wound Repair Regen 1995;3:176-84. 14. Dans MJ, Isseroff R. Inhibition of collagen lattice contraction by pentoxifylline and interferon-alpha, -beta, and -gamma. J Invest Dermatol 1994;102:118-21. 15. Nedelec B, Shen YJ, Ghahary A, Scott PG, Tredget EE. The effect of interferon-α2b on the expression of cytoskeletal proteins in an in vitro model of wound contraction. J Lab Clin Med 1995;126:474-84. 16. Sahara K, Kucukcelebi A, Ko F, Phillips L, Robson M. Suppression of in vitro proliferative scar fibroblast contraction by interferon alfa-2b. Wound Repair Regen 1993;1:22-7. 17. Berman B, Duncan MR. Short-term keloid treatment in vivo with human interferon alfa-2b results in a selective and persistent normalization of keloidal fibroblast collagen, glycosaminoglycan and collagenase production in vitro. J Am Acad Dermatol 1989;21:694-702. 18. Granstein RD, Rook A, Flotte TJ, Haas A, Gallo RL, Jaffe HS, et al. A controlled trial of intralesional recombinant interferon-γ in the treatment of keloidal scarring. Arch Dermatol 1990;126:1295-302. 19. Pittet B, Rubbia-Brandt L, Desmouliere A, Sappino AP, Roggero P, Guerret S, et al. Effect of γ-interferon on the
808 Nedelec et al
20.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30. 31.
32.
clinical and biologic evolution of hypertrophic scars and Dupuytren’s disease: an open pilot study. Plast Reconstr Surg 1994;93:1224-35. Darby I, Skalli O, Gabbiani G. Alpha-smooth muscle actin is transiently expressed by myofibroblasts during experimental wound healing. Lab Invest 1990;63:21-9. Desmouliere A, Redard M, Darby I, Gabbiani G. Apoptosis mediates the decrease in cellularity during the transition between granulation tissue and scar. Am J Pathol 1995;146:56-66. Garbin S, Pittet B, Montandon D, Gabbiani G, Desmouliere A. Covering by a flap induces apoptosis of granulation tissue myofibroblasts and vascular cells. Wound Repair Regen 1996;4:244-51. Kischer CW, Pindur J, Krasovich P, Kischer E. Characteristics of granulation tissue which promote hypertrophic scarring. Scanning Microsc 1990;4:877-88. Dao T, Ariyasu T, Holan V, Minowada J. Natural human interferon-α augments apoptosis in activated T cell line. Cell Immunol 1994;155:304-11. Egle A, Villunger A, Kos M, Bock G, Gruber J, Auer B, et al. Modulation of Apo-1/Fas (CD95)-induced programmed cell death in myeloma cells by interferon-alpha 2. Eur J Immunol 1996;26:3119-26. Rodriguez-Villanueva J, McDonnell TJ. Induction of apoptotic cell death in non-melanoma skin cancer by interferonα. Int J Cancer 1995;61:110-4. Sullivan T, Smith J, Kermode J, McIver E, Courtemanche DJ. Rating the burn scar. J Burn Care Rehabil 1990;11:256-60. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 1992;119:493-501. Tredget EE, Wang R, Shen Q, Scott PG, Ghahary A. Transforming growth factor-beta mRNA and protein in hypertrophic scar tissues and fibroblasts: antagonized by interferon-α and interferon-γ in vitro and in vivo. J Interferon Cytokine Res 2000;20:143-51. Blackburn WR, Cosman C. Histologic basis of keloid and hypertrophic scar differentiation. Arch Pathol 1966;82:65-71. Linares HA, Kischer CW, Dobrkovsky M, Larson DL. The histiotypic organization of the hypertrophic scar in humans. J Invest Dermatol 1972;59:323-31. Ghahary A, Shen YJ, Scott PG, Tredget EE. Enhanced expression of mRNA for transforming growth factor-β, type I and type III procollagen in human post-burn hypertrophic scar tissues. J Lab Clin Med 1993;122:465-73.
Surgery November 2001 33. Desmouliere A, Geinoz A, Gabbiani F, Gabbiani G. Transforming growth factor β1 induces α-smooth muscle actin expression in granulation tissue myofibroblast and in quiescent and growing cultured fibroblasts. J Cell Biol 1993;122:103-11. 34. Nedelec B, Shankowsky HA, Tredget EE. Rating the resolving hypertrophic scar: comparison of the Vancouver scar scale and scar volume. J Burn Care Rehabil 2000;21:205-12. 35. Scott PG, Dodd CM, Tredget EE, Ghahary A, Rahemtulla F. Immunohistochemical localization of the proteoglycans decorin, biglycan and versican and transforming growth factor-β in human post-burn hypertrophic and mature scars. Histopathology 1995;26:423-31. 36. Scott PG, Dodd CM, Tredget EE, Ghahary A, Rahemtulla F. Chemical characterization and quantification of proteoglycans in human post-burn hypertrophic and mature scars. Clin Sci 1996;90:417-25. 37. Desmouliere A, Badid C, Bochaton-Piallat ML, Gabbiani G. Apoptosis during wound healing, fibrocontractive diseases and vascular wall injury. Int J Biochem Cell Biol 1997;29:1930. 38. Chesney J, Metz C, Stavitsky AB, Bacher M, Bucala R. Regulated production of type I collagen and inflammatory cytokines by peripheral blood fibrocytes. J Immunol 1998;160:419-25. 39. Wassermann RJ, Polo M, Smith P, Wang X, Ko F, Robson MC. Differential production of apoptosis-modulating proteins in patients with hypertrophic burn scar. J Surg Res 1998;75:74-80. 40. Nedelec B, Dodd C, Scott PG, Ghahary A, Tredget EE. The effect of interferon-α2b on guinea pig wound closure and the expression of cytoskeletal proteins in vivo. Wound Repair Regen 1998;6:202-12. 41. Iredale JP, Benyon RC, Pickering J, McCullen M, Northrop M, Pawley S, et al. Mechanisms of spontaneous resolution of rat liver fibrosis: Hepatic stellate cell apoptosis and reduced hepatic expression of metalloproteinase inhibitors. J Clin Invest 1998;102:538-49. 42. Shang K, Rekhter MD, Gordon D, Phan SH. Myofibroblast and their role in lung collagen gene expression during pulmonary fibrosis. Am J Pathol 1994;145:114-35. 43. Mounier N, Desmouliere A, Gabbiani B. Subcutaneous tissue fibroblast transfected with muscle and nonmuscle actins: a good in vitro model to study fibroblastic cell plasticity. Wound Repair Regen 1999;7:45-52.
CORRECTION In the article "The role of iNOS in wound healing" (Surgery 2001;130:225-9), two authors were not noted. The byline should be as follows: Han Ping Shi, MD, PhD, Daniel Most, MD, David T. Efron, MD, U. Tantry, MD, M. H. Fischel, MD, and Adrian Barbul, MD, FACS
Background. Hypertrophic scars (HSc) are a dermal fibroproliferative disorder that leads to considerable morbidity. Preliminary evidence suggests that interferon (IFN) may improve HSc clinically. The aims of this study were (1) to compare the cell density in HSc and in wounds that heal without the development of HSc (normotrophic scars), (2) to examine the presence of myofibroblasts and apoptosis in normotrophic and HSc scars over time, and (3) to determine if the systemic administration of IFN-α2b can induce apoptosis. Methods. Two groups of patients underwent serial tissue biopsies. Six burn patients were studied prospectively by obtaining biopsy specimens from wound granulation tissue, normal skin, post-burn HSc, and normotrophic scars (healed donor sites). A second patient group with HSc was treated with systemic IFNα2b and had biopsy material taken before, during, and after IFN therapy. The tissue was analyzed by immunohistochemical staining for α-smooth muscle actin (α-SMA) and in situ DNA fragmentation terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay for apoptosis. Results. The total numbers of fibroblasts in HSc were found to be similar to granulation tissue and twice that of normal skin and normotrophic scar. Over time the numbers of cells in HSc tissue decreased toward normal skin levels. There was a significantly higher percentage of fibroblasts staining for α-SMA in HSc as compared with normotrophic scar or normal skin obtained from the same patient (P > .05). Serial biopsy specimens of resolving HSc tissue obtained from the patients who received systemic IFN-α2b showed a general reduction in total number of fibroblasts and myofibroblasts associated with a significant increase in the percentage of apoptotic cells compared with normal dermis from the same patient. Conclusions. HSc tissues have greater numbers of fibroblasts and myofibroblasts than normal skin and normotrophic scars. As HSc remodels, the numbers of fibroblasts and myofibroblasts reduces, possibly by the induction of apoptosis. Systemic IFN-α2b may contribute to the resolution of HSc in part by the enhanced induction of apoptosis. (Surgery 2001;130:798-808.) From the Department of Surgery and the Department of Biochemistry, University of Alberta, Edmonton, Alberta, Canada
THE DEVELOPMENT OF HSc after a burn or other major trauma to the skin is a dermal fibrotic disorder where red, raised, pruritic lesions distort the skin and are often associated with the formation of contractures.1 Scar contractures reduce range of
Supported by the Firefighters’ Burn Trust Fund of the University of Alberta Hospitals, the Harold and Emilie Tucker Trust Fund of the University of Alberta Hospitals (B.N.), Alberta Heritage Foundation for Medical Research (E.E.T., Scholar), and the Medical Research Council of Canada (P.G.S., A.G., E.E.T.). Accepted for publication April 7, 2001. Reprint requests: Edward E. Tredget, MD, MSc, FRCSC, 2D3.81 WMHCS, University of Alberta Hospital, 8440-112 St, Edmonton, Alberta, Canada T6G 2B7. Copyright © 2001 by Mosby, Inc. 0039-6060/2001/$35.00 + 0 11/56/116453 doi:10.1067/msy.2001.116453
798 SURGERY
motion, resulting in delayed return to work and reintegration into society.2 The treatment of HSc and scar contractures includes conservative treatments such as corticosteroid injections, pressure therapy, splinting, and serial casting,1,3 however, these treatments are time-consuming, uncomfortable, and often incompletely successful. Surgical excision or release of HSc also is associated with a high rate of recurrence.4 Gabbiani et al5 and Ehrlich et al6 previously have identified myofibroblasts as a specialized fibroblast that is present in HSc. This cell possesses contractile properties and specific morphologic features of both the fibroblast and smooth muscle cell and commonly is defined as a fibroblast that stains positively for α-SMA.5,7 In a phase II clinical trial in burn patients suffering from HSc, systemically administered IFN-
Surgery Volume 130, Number 5 α2b has been associated with the induction of scar remodeling and maturation. 8 In vitro IFN has been shown to decrease fibroblast proliferation,9 down-regulate the production of collagen and fibronectin,10-12 and enhance the production of collagenase.13 Reduction of the rate and extent of contraction in vitro has been shown using IFN-α, IFN-γ, and IFN-β.14-16 Intralesional injections of IFN-α or IFN-γ led to a reduction in the keloid mass, 17,18 HSc, and Dupuytren’s disease nodules.19 In Dupuytren’s disease, myofibroblast staining for α-SMA was reduced in intensity and distribution after 4 weeks of intralesional IFN-γ treatment.19 Previous research has correlated the disappearance of myofibroblasts in wound healing with the induction of apoptotic cell death.20,21 In addition, the more rapid disappearance of myofibroblasts in open wounds covered with full-thickness flaps compared with a split-thickness flap has been attributed to the accelerated induction of apoptosis. 22 Conversely, a lack of induction of apoptosis may contribute to the excessive cellularity and contraction associated with HSc.23 It is our hypothesis that HSc contains excessive numbers of fibroblasts and myofibroblasts, which reduce slowly as the scars remodel and mature into normotrophic scars. One mechanism by which IFNα2b may be of therapeutic value in the treatment of fibroproliferative disorders such as HSc is through the accelerated induction of apoptosis, which has been previously demonstrated in activated T lymphocytes, squamous cell skin cancer cells, and myeloma cells.24-26 This hypothesis was tested prospectively in 2 patient groups. In 1 group, burn wounds were biopsied serially in the granulation phase, after the development of HSc, and during the remodeling phase, and compared with normal skin from the same patient, as well as the skin graft donor site, as a wound that healed without the development of HSc (normotrophic scar). A second group of burn patients in whom HSc had already developed after burn injury underwent serial biopsy specimens of normal skin and HSc before, during, and after 24 weeks of subcutaneous IFN-α2b therapy.8 Serial sections were evaluated histologically for total fibroblast cell number as well as for myofibroblasts, before measurement of apoptosis in the sections was done over time with the TUNEL assay. MATERIAL AND METHODS Tissue biopsies. Patients recruited into this study had sustained thermal injuries that necessitated skin grafts and were treated in the Firefighters’ Burn Treatment Unit of the University
Nedelec et al 799
of Alberta Hospital. Patients were recruited and tissues were sampled after informed consent had been obtained and according to the guidelines of the University of Alberta, Faculty of Medicine Research Ethics Board. The 6-mm punch biopsy specimens of wounds and normal skin were collected from 2 patient groups. In the first group, biopsy specimens of granulation tissue, normal skin, healed burn wounds, and healed donor sites were obtained from the same patient serially over time (group 1, Table I). The burn wound granulation tissue and normal skin first were biopsied while the patient was undergoing skin grafting (mean, 19 days post burn). The healed burn wound and donor site biopsy specimens were obtained when the patient returned to the outpatient clinic for follow-up. Wounds that developed red, raised, pruritic scar tissue were designated HSc, and wounds and donor sites that healed without the development of HSc were designated normotrophic scars. A total of 6 patients and 24 biopsy specimens were studied in this group. In the second group, biopsy specimens were obtained from 9 burn patients for whom HSc had developed over more than 5% of their total body surface area (TBSA) after a thermal injury and who had been recruited into a phase II clinical trial of IFNα2b (INTRON A, Schering, Kenilworth, NJ) (group 2, Table I).8 The HSc and normal tissue biopsy specimens (6-mm punch) were obtained before the subcutaneous administration of 1 × 106 units of IFN-α2b per day for a week, then 2 × 106 units 3 times/wk for 24 weeks. The definition of clinically significant HSc was established by using a previously established scar rating system, the modified Vancouver Burn Scar Assessment (VBSA).27 A scar rating of greater than 3, indicating significant abnormalities in skin pliability, erythema, scar height, and color, was considered hypertrophic. Biopsy specimens of HSc were obtained every other month and at the conclusion of therapy. A second biopsy of normal skin was taken before starting the IFN therapy and at the termination of IFN-α2b treatment.8 All biopsy specimens were placed immediately in 4% paraformaldehyde for immunohistochemical analysis. Materials. Biotin-16-2'-deoxyuridine-5'-triphosphate, proteinase K, dUTP, 4-nitroblue tetrazolium chloride, 5-bromo-4-chloro-3-indolyl phosphate, and terminal deoxynucleotidyl transferase (TdT) were obtained from Boehringer Mannheim (Indianapolis, Ind). The RNase-free DNase was obtained from Promega (Madison, Mich). Caseinblocking buffer was obtained from Cambridge Research Biochemicals (Northwich, Cheshire). Anti-α-SMA (clone 1A4), bovine serum albumin
800 Nedelec et al
Surgery November 2001
Table I. Patient demographic data
Patient Group 1 1 2 3 4 5 6 Mean ± SD Group 2 7 8 9 10 11 12 13 14 15 Mean ± SD
Time to final biopsy of normal skin and burn wound (mo)
Sex
Age
TBSA (%)*
VBSA†
M M M M M M
67 35 28 42 49 41 43.7 ± 13.4
5 18 85 3 25 40 29.3 ± 30.5
0 1 12 6 6 9 5.7 ± 4.6
2 12 12 3 4 3 6.0 ± 4.7
M M M M M M F F M
49 26 49 36 31 25 46 10 29 33.4 ± 13.0 NS
32 70 30 65 60 45 20 40 85 49.7 ± 21.5 NS
12 6 13 10 8 9 12 11 7 9.8 ± 2.4 P < .05
17 20 17 4 20 6 16 12 8 13.3 ± 6.1 P < .05
Demographic information for the 2 patient groups is listed including gender, age, TBSA, and the number of months after injury that normal skin and burn wounds were biopsied. *TBSA, Total body surface area injured from thermal injury; †VBSA, Vancouver Burn Scar Assessment clinical scar rating.
(BSA), 3, 3'-diaminobenzidine (DAB), 3-aminopropyl-triethoxy-silane (Aptex), monoclonal peroxidase-anti-peroxidase soluble complex antibodies, ExtrAvidin-Alkaline Phosphatase and normal goat serum were obtained from Sigma Chemical Company (St Louis, Mo). The bridging goat F(ab')2 fragment to mouse IgG (whole molecule) was obtained from ICN Pharmaceuticals Inc (Aurora, Ohio). Immunohistochemistry. The tissue biopsy specimens fixed in 4% paraformaldehyde were embedded in paraffin and cut into 4- to 5-µm sections and placed on Aptex-coated slides. All sections were heated to 60°C for 10 minutes then dewaxed in xylene and rehydrated. Endogenous peroxides were inactivated by treating all sections with 10% H2O2 in methanol for 6 minutes. After washing with PBS, sections were incubated for 30 minutes at room temperature (RT) in blocking agent (10% vol/vol normal goat serum, 5% BSA wt/vol in PBS) and then incubated overnight at 4°C with the appropriate antibody. The α-SMA antibody was diluted 1:100 in PBS/Tween 0.05% solution. The control specimens were incubated with plasmacytoma cell culture supernatant diluted 1:10 in PBS/Tween 0.05% solution. After warming to RT the sections were incubated with goat F(ab')2 fragment to mouse IgG (whole molecule; 1:50 dilution in PBS/Tween 0.05% solution) for 1.5 hours,
washed with PBS, then incubated with monoclonal peroxidase-anti-peroxidase soluble complex at a 1:200 dilution for 1.5 hours and washed with PBS. The colorimetric reagent was hydrogen peroxide/DAB, and all sections were counterstained with hematoxylin. The α-SMA–positive fibroblasts were counted on 10 randomly selected high-power fields per section. The total number of fibroblasts, based on morphology, also were counted on the same section, thus allowing calculation of the percentage of myofibroblasts. Smooth muscle cells associated with blood vessels and erector pili were excluded from the α-SMA positive fibroblast count and the total fibroblast count. TUNEL assay. The in situ demonstration of DNA fragmentation was performed by using the TUNEL assay as previously described.28 All sections were treated as described for immunohistochemistry up to the point of rehydration. After washing with ddH2O, the sections were incubated for 20 minutes at 37°C in 2 × sodium chloride and sodium citrate (SSC) (pH 7.0) then washed several times in ddH2O. All sections were bathed for 5 minutes at RT in 10 mmol/L Tris-HCl solution (pH 8.0) then digested for 15 minutes at 37°C with proteinase K (20 µg/mL in 10 mmol/L Tris-HCl). After washing several times in ddH2O the tissue sections were fixed for 4 minutes in 4% paraformaldehyde then
Surgery Volume 130, Number 5
washed several times with ddH2O. The positive control material was incubated in DNase buffer for 5 minutes at RT then digested with DNase (1 µg/mL) at 37°C for 30 minutes. The positive control was washed with ddH2O then all sections were incubated for 5 minutes at RT in TdT buffer (pH 7.2). All sections were incubated in reaction buffer (94.5 µl ddH2O, 7.0 µl TdT buffer, 4.4 µl Bio-16dUTP, 4.4 µl dUTP, and 1.5 µl TdT) for 1 hour at 37°C in a humidified chamber, except the negative control where the TdT was omitted. To terminate the reaction the sections were transferred to 2 × SSC for 15 minutes at RT. All sections were washed with PBS solution then incubated for 15 minutes at RT in blocking agent (1:5 casein blocking buffer, 10% vol/vol normal goat serum, and 5% BSA wt/vol in PBS). The sections then were incubated at RT for 30 minutes in Extra-Avidin phosphatase (1:250) in PBS/Tween 0.5% solution. The sections were washed with PBS then incubated with colorimetric buffer (0.1 mol/L Tris, 0.1 mol/L NaCl, 0.25 mol/L MgCl2; pH 9.5) for 5 minutes at RT. The colorimetric reagent was 4-nitroblue tetrazolium chloride and 5-bromo-4-chloro-3-indolyl phosphate. The percentage of positive cells was calculated by counting the number of positive cells in 10 randomly selected fields and comparing it with the total number of cells within a serial section stained with hematoxylin. Cells that were obviously endothelial cells or epithelial cells—such as in vessels, hair follicles, sweat glands, sebaceous glands, and the outer epithelial layer—were eliminated from both positive cell counts and total cell counts. Statistics. A 2-tailed Student paired t test was used for all comparisons of matched data obtained from 2 time points in the same patient in group 2, and a nonpaired Student t test was performed for comparisons of 2 variables between groups 1 and 2. Analysis of variance was used for the analysis where more than 2 comparisons were made as indicated in the figure legends and text. Results are reported using P value, mean, and standard deviation as indicated. RESULTS Based on clinical criteria for rating the burn scar, 4 of the 6 patients from group 1 had HSc develop (VBSA scale > 3) in burn wound sites and 2 patients did not, thus meeting the criteria for normotrophic healing, as indicated in Table I. None of the patients from group 1 had HSc in their donor sites. All of the patients in group 2 were involved in the phase II IFN trial and had extensive HSc over >5% of the TBSA as one of the inclusion criteria. There was a poor correlation between the
Nedelec et al 801
size of the original burn and the severity of HSc using the VBSA scale (r2 = .12; P = .21, n = 15). Group 1 differed from group 2 patients in that most of the former had less severe HSc based on clinical scar rating (VBSA, Table I) (5.7 ± 4.6 vs 9.9 ± 2.4; P < .05) and were enrolled in the study within the first year following burn injury (mean, 6.0 ± 4.7 months; range, 2-12 months); whereas, most of the group 2 patients were studied more than 12 months after injury (mean, 13.3 ± 6.1 months; range, 4-20 months) and failed to respond to conservative management of more severe HSc as measured by the VBSA, which was a significant clinical and statistical difference between the 2 groups. Many group 2 patients also suffered larger burn injuries, were somewhat younger, and included 2 females, one being a 10-year-old child (Table I). Although large differences exist in these demographic variables, the broad variation likely accounted for the lack of statistically significant differences between groups 1 and 2 in these specific features (Table I). In group 1, burn patients were studied prospectively for Hsc development. The total number of fibroblasts per high-power field was significantly increased in the granulation tissue biopsied prior to wound closure (mean, 19.0 ± 27.8 days; range, 475 days) compared with the normal skin (12.0 ± 2.6 vs 4.7 ± 0.5 fibroblasts/high-power field; P < .001) and in HSc as compared with the normal skin in the same patient (10.5 ± 1.0 vs 4.7 ± 0.5; P < .05). This increase in cell density was not significantly different between HSc and granulation tissue (10.5 ± 1.0 vs 12.0 ± 2.6, P > .05, not significant). In wounds that healed without HSc development, significantly lower fibroblast cell density was found and normotrophic scar cell densities were not significantly different than normal dermis (6.8 ± 0.9 vs 4.7 ± 0.5; P > .05, not significant) (Fig 1, A). When the total number of fibroblasts in HSc was compared as a function of time following burn injury, there was a progressive reduction between early (04 months post-burn), mid (5-18 months), and late (19-30 months) HSc (Fig 1, B), which was statistically significant when comparing the early phase of HSc and the later remodeled phase (P < .01). Histologically, the distribution of the α-SMA positive fibroblasts or myofibroblasts in HSc was nonuniform and occurred in the more densely populated cell regions, but was not exclusive to any particular region of HSc (Fig 2). However, fewer αSMA positive myofibroblasts generally were visible in regions where more organized collagen fibers were present. Positive staining also was seen consistently in the endothelial cells of blood vessels
802 Nedelec et al
Surgery November 2001
A
A B
B Fig 1. Total fibroblasts per high-power field differs among various types of wound healing and returns to a more normal level as the scar matures. Total fibroblasts per highpower field was averaged from 10 randomly sampled fields. A, Total fibroblasts in granulation tissue and HSc relative to normal dermis (12.0 ± 2.6 and 10.5 ± 1.0 vs 4.7 ± 0.5). Normotrophic scar did not differ significantly from normal dermis (6.8 ± 0.9). B, The gradual reduction in total fibroblasts as a function of time, where early scar is 0-4 months post-burn (13.1 ± 2.0; n = 5), mid is 5-18 months (8.5 ± 0.4; n = 25), and late is 19-30 months (7.6 ± 0.5; n = 20).
where α-SMA is characteristically expressed and in the epidermal cells, which has previously been described in epithelial cells from other sources.7 Quantitative analysis of the percentage of αSMA–positive staining myofibroblasts in the initial biopsy obtained from group 1 (patients in whom HSc developed) and group 2 (before treatment with IFN) patients demonstrated considerable variability within the HSc sections, ranging from 1% to 65%, with an overall mean of 23.2% ± 5.2% (Fig 3). However, all patients who suffered burns over more than 70% of the TBSA had more than 35% of the total fibroblast count positive for myofibroblasts. Linear regression of myofibroblast numbers compared with the size of the original burn wound revealed a significant proportion of the variability in the percentage of myofibroblasts in HSc was positively related to the size of the original wound (r2 = .60; P < .05, n = 9), but poorly correlated with the severity of clinical burn scar assessment (r2 = .05; P = .43, n = 15) or the time after injury that the biopsy was obtained (r2 = .10; P = .39, n = 9). Significantly increased numbers of myofibroblasts were seen in HSc tissues as compared with normal skin from the
C
Fig 2. Localization of α-SMA–positive fibroblasts within HSc tissue. Positive staining of fibroblasts for α-SMA was localized to regions of disorganized collagen fibrils and to a lesser extent where collagen fibers and fiber bundles were present. A, B, and C are from the same patient biopsy but different areas of the tissue, all stained for α-SMA. A, Intense fibroblast staining in the center of a nodular region where there are disorganized collagen fibrils and capillaries on the outer edge of the nodule. B, Area where the α-SMA–positive fibroblasts are arranged in parallel alignment, with the collagen fibrils aligned in a waved pattern but parallel to the fibroblasts. C, Region where more normal collagen fibers and fiber bundles are visible. Although the vessels at the center of the panel are intensely stained there are few stained fibroblasts and the tissue is markedly less densely populated (×100 magnification).
same patients who had clinically significant HSc (23.2% ± 5.2% vs 6.2% ± 1.4%; n = 13; P < .01) (Fig 3). In wounds that healed without the Hsc development (6 donor site wounds and 2 burn wounds),
Surgery Volume 130, Number 5
Fig 3. Percentage of α-SMA–positive fibroblasts in HSc relative to normotrophic scar and normal dermis. Thirteen Hsc biopsies (4 from group 1, 9 from group 2) and normal dermis from the same patient and 8 normotrophic scars (6 donor site wounds and 2 burn wounds) are shown where 10 randomly chosen high-power fields were quantitated per biopsy. The bars indicate the mean for each group. Statistically significant differences in the means for each group are depicted on the graph (ANOVA).
the percentage of myofibroblasts was significantly lower than in the HSc tissues (6.1% ± 2.4% vs 22.9% ± 5.2%; n = 8; P < .05). In 2 patients whose burns healed without developing HSc who were included in the normotrophic group, the number of α-SMA positive myofibroblasts was 1% and 13%, respectively. Comparison of the number of myofibroblasts from normotrophic healed wounds in the first group of patients with the normal dermis from both groups revealed no significant differences. In group 2 patients who had large numbers of HSc and who received IFN-α2b, a reduction in the total fibroblast cell density was seen when comparing the initial biopsy with the final biopsy after 24 weeks of IFN therapy (9.9 ± 4.4 vs 6.7 ± 2.2 fibroblasts/high-powered field; n = 8; P < .05) (Fig 4, A). Comparison of the percentage of myofibroblasts in the HSc at the time of the initial biopsy relative to the end of the IFN trial showed a statistically significant reduction in 4 of 9 patients (data not shown) and an overall reduction in the mean proportion before and after IFN treatment, but the group effect was not statistically significant (26.9% ± 6.6% vs 16.0% ± 2.8%; n = 9) (Fig 4, B). The reduction in the number of fibroblasts in the dermal extracellular matrix that occurs as wound healing proceeds through the granulation phase to immature HSc, before remodeling into the normotrophic or mature scar, may occur through the induction of apoptosis.21 As discussed
Nedelec et al 803
earlier, one mechanism by which IFN-α2b may facilitate the maturation of HSc may be the induction of apoptotic cell death, which was investigated in our study using the TUNEL assay on serial sections of the same tissue (Fig 5). Although it previously has been reported that apoptotic cells in normal rat skin generally were not seen,21 apoptotic cells were present consistently in normal dermis of human burn survivors. The proportion of apoptotic cells from all HSc biopsies was lower than normal skin in early HSc (75.1% ± 25.7%; n = 5), slightly greater than normal in 5 to 18 months after injury (110.3% ± 15.8%; P < .05, n = 24), and significantly higher later in the Hsc remodeling phase (19-30 months) (221.6% ± 31.2%; P < .05, n = 19) (Fig 6). The effect of IFN-α2b treatment on the induction of apoptosis was examined by comparing the rates of apoptosis in the pre-treatment biopsy, the peak or highest percentage of apoptotic cells following IFN-α2b treatment and the final biopsy following treatment, with the normal skin biopsy (Fig 7). Prior to IFN therapy, the HSc demonstrated a higher proportion of apoptotic cells (33.4% ± 6.5% vs 20.9% ± 4.5%; n = 9), which was again increased significantly during administration of IFN-α2b, as compared with normal dermis (43.5% ± 5.6% vs 20.9% ± 4.5%; n = 9; P < .01, analysis of variance [ANOVA]). Following IFN-α2b treatment the proportion fell towards that in normal skin (26.3% ± 4.7% vs 20.9% ± 4.5%; n = 9; not significant). DISCUSSION Although histologic analysis of human hypertrophic burn scars and wound healing models in animals collectively have suggested that HSc resembles granulation tissue in the high cellularity of the extracellular matrix, this study is the first prospective, quantitative, longitudinal analysis in human beings that provides statistically significant confirmation of the hypothesis that early immature HSc are hypercellular and that as they remodel and mature the fibroblast cell density reduces to resemble normal skin, in part, through the induction of apoptosis. Wounds that heal over shorter periods and that often are not as deep into the dermis, such as the uncomplicated skin graft donor sites, do not become hypercellular and heal with a more normal collagen fiber bundle orientation. These normotrophic scars clinically and histologically resemble normal skin from the same patient (data not shown). Thus, the persistent high fibroblast cell density in HSc as compared with normotrophic scars likely is a very important factor contributing to the pathogenesis of this fibroproliferative disorder. Functional comparisons carried out in our lab-
804 Nedelec et al
A
B Fig 4. Total fibroblasts per high-power field of HSc tissue decreased after IFN-α2b treatment. A, Total fibroblasts in HSc tissue sections from the same patients before and 6 months after systemic administration of IFN-α2b (9.9 ± 4.4 vs 6.7 ± 2.2; P < .05, n = 8, paired Student t test). B, Percentage of α-SMA–positive fibroblasts decreased after IFN-α2b treatment. The number of α-SMA–positive fibroblasts was compared in HSc tissue sections from the same patient before and 6 months after systemic administration of IFN-α2b. Individual patient comparisons demonstrated a significant reduction in the percentage of positive cells in 4 of 9 patients; however, the overall reduction did not reach a level of significance (26.9% ± 6.6% vs 16.0% ± 2.8%; n = 9).
oratory of the HSc fibroblasts as compared with the paired normal fibroblast from the same patient have shown significant differences in collagen and collagenase production, decorin synthesis, and transforming growth factor-β (TGF-β) production.1 These phenotypic differences on a per cell basis may be exaggerated in vivo where our data indicate that approximately twice as many cells are present
Surgery November 2001
in the extracellular matrix of HSc as in normal skin. For example, increases in fibroblast TGF-β mRNA and protein production in vitro and in vivo are approximately 20% greater than the normal fibroblasts.29 This difference, when combined with the understanding that HSc have approximately twice the cell density of normal skin, likely has a significant effect on the total amount of local TGF-β produced in the early stages of wound healing and may contribute greatly to the magnitude of HSc formation. Our data also have confirmed the presence of the myofibroblast in HSc, similar to previous histologic descriptions of HSc,30 and longitudinal wound healing studies in animals.31 Significantly greater numbers of myofibroblasts were present in HSc as compared with normotrophic scar and normal skin. The strong correlation with the numbers of myofibroblasts in the healing postburn HSc with the size of the original burn injury suggests they are an important component of more complicated wound healing. Previous investigations have demonstrated increased amounts of the fibrogenic growth factor TGF-β in HSc tissues locally32 and in the serum of patients examined in this study,8 which others have found induces α-SMA in vivo.33 Other cytokines and growth factors such as plateletderived growth factor and tumor necrosis factor-α are likely present at some point during the healing of burn wounds such as investigated here, but do not appear to induce α-SMA in vivo.33 The poor correlation between myofibroblast numbers, the size of the burn wound, and the scar rating system may result from limitations of the VBSA scale. The need for more objective instruments to measure scar quality and suggestions for improvements in the clinical assessment of scars has been addressed based on the outcome of these studies.34 The HSc is characterized by the presence of nodules that contain thin collagen fibrils.30,31 It previously has been suggested that fibroblasts that stain positive for α-SMA are restricted to these nodular structures.6 In our study, fibroblasts staining for α-SMA generally were concentrated in scar regions that contained poorly organized collagen fibrils and a high density of cells commonly within these nodular structures, but also in the more mechanically stressed whorls that surrounded the nodules. In both the nodular structures and the mechanically stressed regions, thin collagen fibrils surrounded the myofibroblasts. In regions where more organized collagen fiber bundles were present, fewer α-SMA–stained fibroblasts were observed. The thin collagen fibrils and more organized fiber
Nedelec et al 805
Surgery Volume 130, Number 5
A
B
C
D
Fig 5. Representative examples of apoptotic cells within HSc at various times after burn injury relative to their normal dermis. The TUNEL assay identifies apoptotic cells after colorimetric development as blue cells. A, Example of HSc in the early period after a burn injury (0-3 months). B, HSc 19 months or later after a burn injury. There are significantly more positive cells than in A. C, Example of normal dermis, which has more positive cells than the early HSc (A) but fewer than in the late HSc (B). D, Control tissue section where terminal transferase was omitted from the reaction (×250 magnification).
bundles often were present within the same tissue section. These findings suggest that the myofibroblast phenotype also may be regulated by the organization of the extracellular matrix. In our previous work, many regions of poorly organized immature collagen fibers in the HSc stained poorly for the small proteoglycan, decorin, whereas αSMA was increased. This suggests a potential recip-
rocal interrelationship of the myofibroblast with this and possibly other components of the extracellular matrix and as well with TGF-β, which is normally bound to decorin in uninjured skin and mature scar.35,36 Our data also support the hypothesis of Gabbiani and Desmouliere37 that myofibroblasts appear in HSc prior to the development of high
806 Nedelec et al
Fig 6. The percentage of apoptotic cells in HSc increases as a function of time after burn injury relative to patientmatched normal controls. The percentage of apoptotic cells when compared with normal in early HSc (0-4 months post-burn injury) (75.1% ± 25.7%; n = 5), during the mid period (5-18 months post-burn) (110.3% ± 15.8%; n = 24), and late time period (19-30 months postburn) (221.6% ± 31.2%; n = 19) (asterisk = P < .05).
levels of apoptosis and may be a differentiated fibroblast committed to apoptotic cell death. The fact that the numbers of myofibroblasts in HSc correlate well with the size of the burn injury suggests that myofibroblasts may represent a somewhat greater proportion of the fibroblasts in large or deep wounds, whereas in normotrophic scars, the smaller numbers of myofibroblasts suggest that resident fibroblasts make up a greater proportion of the extracellular matrix and lead to more normal dermal healing resembling the normal skin. The recent recognition of the importance of bone marrow-derived fibrocytes, which can populate the wound, suggests that for patients with very large and deep injuries, the source of fibroblasts may not be the same as in uncomplicated wound healing.38 Further investigation will be required to confirm this suggestion, however. In our study the measurement of apoptosis by TUNEL was associated with substantial variability due to the heterogeneous distribution of apoptotic cells in the extracellular matrix, which was overcome in part by the analysis of large numbers of sections from many patients at different times. Significant increases in apoptosis late in the Hsc remodeling phase, however, establish apoptosis as an important mechanism by which fibroblast cell numbers are normalized. Thus a delay in apoptosis may be an important factor in the prolonged and severe nature of HSc seen in many of our patients.8 Wassermann and colleagues39 have demonstrated that peripheral blood mononuclear cells from patients who develop HSc have increased levels of the proto-oncogene Bcl-2 and reduced local tissue
Surgery November 2001
Fig 7. The percentage of apoptotic cells increased significantly in interferon-treated patients relative to normal dermis. The percentage of apoptotic cells in HSc biopsies determined by TUNEL before IFN treatment (Pre Tx), at the point of maximal apoptosis during IFN treatment (peak), 4 weeks after IFN treatment, as compared with normal skin biopsies obtained before IFN treatment. The percentage of apoptotic cells after the administration of IFNα2b was significantly increased relative to normal (43.5% ± 5.6 vs 20.9% ± 4.5, P < .01, n = 9, ANOVA).
expression of the apoptosis-inducing proteins Fas and interleukin-1 converting enzyme (ICE). To date, preliminary phase II clinical trials suggest that systemic administration of IFN-α2b improves healing by promoting HSc remodeling, resulting in more supple, pliable tissue.8 Serial biopsy specimens of HSc and normal skin from the same patient over a 7-month period, after conventional therapy had failed, demonstrated a reduction in the total fibroblast cell number in HSc. This was associated with a similar reduction in the number of α-SMA positive fibroblasts and an increase in the maximal amount of apoptosis in HSc during IFN therapy. These findings may be due to a direct reduction in fibroblast proliferation as we have observed in vitro11,12 and/or the induction of apoptosis. Without a statistically significant difference between apoptosis in HSc before and after IFN therapy, however, it cannot unequivocally be established which is the contributing factor in situ. It is difficult to draw definitive conclusions from this study because of the considerable differences between patients regarding the Hsc site, the severity of the original injury, and the time post-burn prior to commencing IFN therapy. In addition, in a preliminary phase II clinical trial where a limited number of patients were recruited and studied in situ, substantial variability arose in the distribution of myofibroblasts and apoptotic cells within HSc. Our previous work in full thickness excisional wounds in guinea pigs with a double-blinded, controlled administration of IFN or its vehicle via intraperitoneal osmotic pumps afforded more rig-
Surgery Volume 130, Number 5
orous controls and demonstrated a significant increase in apoptosis in the IFN-treated group only, however, which peaked well after re-epithelialization of the healing wounds, similar to the human HSc examined here.40 Unfortunately, further investigation of interferon’s apoptotic effect in patients with HSc is limited here by the fact that in contrast to group I patients, most of the interferon-treated patients completed therapy long after their initial wounding (13.3 ± 6.1 months, Table I), at a time when the initiation of spontaneous apoptosis and reduction in the number of myofibroblasts begins to occur as evidenced in non-interferon–treated group I patients during this period (Fig 6). In the future, interferon’s potential beneficial effects in inducing apoptosis to reduce cellularity in HSc may be appreciated more easily if therapy is commenced earlier after injury, when apoptosis is delayed and less than in the normal skin (Fig 6). Based on our findings of patients in group I, however, both our human data and animal investigations have suggested that myofibroblasts within HSc or other wounds represent a differentiated fibroblast important in wound remodeling but one that already is committed to apoptosis. The myofibroblast numbers are reduced after IFN therapy. In addition, our study was unable to establish if the maximum increase in fibroblast total cell number preceded the peak in myofibroblast cell numbers in HSc, but did suggest that maximal apoptosis occurred late in the HSc phase, providing additional evidence that apoptosis is an important process in scar remodeling that leads to a reduction in overall cellularity. Previously Gabbiani and colleagues have used qualitative methods to suggest that there is a correlation between the expression of α-SMA in fibroblasts and apoptosis.22 Similar findings have been reported in animal models of liver fibrosis41 and bleomycin-induced pulmonary fibrosis.42 In the future, microinjection studies of α-SMA offer the potential to determine the exact function of this specific actin isoform and subsequently its role in HSc.43 The assessment of the potential for IFN-α2b to assist in preventing HSc and promoting its remodeling through the induction of apoptosis or alteration in myofibroblast function will further benefit from the results of the double-blind placebo-controlled trial that presently is underway. We acknowledge the assistance of D. Carmel (Department of Oral Health Sciences) in preparation of tissue sections for histologic study. REFERENCES 1. Tredget EE, Nedelec B, Scott PG, Ghahary A. Hypertrophic scars, keloids and contractures. The cellular and molecular basis for therapy. Surg Clin N Am 1997;77:701-30.
Nedelec et al 807
2. Engrav LH, Covey MH, Dutcher KD, Heimbach DM, Walkinshaw MD, Marvin JA. Impairment, time out of school, and time off from work after burns. Plast Reconstr Surg 1987;79:927-32. 3. Ridgway CL, Daugherty MB, Warden GD. Serial casting as a technique to correct burn scar contractures. J Burn Care Rehabil 1991;12:67-72. 4. Darzi MA, Chowdri NA, Kaul SK. Evaluation of various methods of treating keloids and hypertrophic scars: a 10year follow up study. Br J Plast Surg 1992;45:374-9. 5. Gabbiani G, Ryan GB, Majno G. Presence of modified fibroblasts in granulation tissue and their possible role in wound contraction. Experientia 1971;27:549-50. 6. Ehrlich HP, Desmouliere A, Diegelmann RF, Cohen K, Compton CC, Garner WL, et al. Morphological and immunochemical differences between keloid and hypertrophic scar. Am J Pathol 1994;145:105-13. 7. Skalli O, Ropraz P, Trzeciak A, Benzonana G, Gillessen D, Gabbiani G. A monoclonal antibody against α-smooth muscle actin: a new probe for smooth muscle differentiation. J Cell Biol 1986;103:2787-96. 8. Tredget EE, Shankowsky HA, Pannu R, Nedelec B, Iwashina T, Ghahary A, et al. Transforming growth factor-β in thermally injured patients with hypertrophic scars: effects of interferon α-2b. Plast Reconstr Surg 1998;102:1317-28. 9. Jimenez SA, Freundlich B, Rosenbloom J. Selective inhibition of human diploid fibroblast collagen synthesis by interferons. J Clin Invest 1984;74:1112-6. 10. Ghahary A, Shen YJ, Scott PG, Tredget EE. Expression of fibronectin messenger RNA in hypertrophic and normal dermal tissues and in vitro regulation by interferon alfa-2b. Wound Repair Regen 1993;1:166-74. 11. Harrop AR, Ghahary A, Scott PG, Forsyth N, Uji-Friedland A, Tredget EE. Regulation of collagen synthesis and mRNA expression in normal and hypertrophic scar fibroblast in vitro by interferon-γ. J Surg Research 1995;58:471-7. 12. Tredget EE, Shen YJ, Liu G, Forsyth N, Smith C, Harrop AR, et al. Regulation of collagen synthesis and messenger RNA levels in normal and hypertrophic scar fibroblasts in vitro by interferon alfa-2b. Wound Repair Regen 1993;1:156-65. 13. Ghahary A, Shen YJ, Nedelec B, Scott PG, Tredget EE. Interferons gamma and alpha-2b differentially regulate the expression of collagenase and tissue inhibitor of metalloproteinase-1 messenger RNA in human hypertrophic and normal dermal fibroblasts. Wound Repair Regen 1995;3:176-84. 14. Dans MJ, Isseroff R. Inhibition of collagen lattice contraction by pentoxifylline and interferon-alpha, -beta, and -gamma. J Invest Dermatol 1994;102:118-21. 15. Nedelec B, Shen YJ, Ghahary A, Scott PG, Tredget EE. The effect of interferon-α2b on the expression of cytoskeletal proteins in an in vitro model of wound contraction. J Lab Clin Med 1995;126:474-84. 16. Sahara K, Kucukcelebi A, Ko F, Phillips L, Robson M. Suppression of in vitro proliferative scar fibroblast contraction by interferon alfa-2b. Wound Repair Regen 1993;1:22-7. 17. Berman B, Duncan MR. Short-term keloid treatment in vivo with human interferon alfa-2b results in a selective and persistent normalization of keloidal fibroblast collagen, glycosaminoglycan and collagenase production in vitro. J Am Acad Dermatol 1989;21:694-702. 18. Granstein RD, Rook A, Flotte TJ, Haas A, Gallo RL, Jaffe HS, et al. A controlled trial of intralesional recombinant interferon-γ in the treatment of keloidal scarring. Arch Dermatol 1990;126:1295-302. 19. Pittet B, Rubbia-Brandt L, Desmouliere A, Sappino AP, Roggero P, Guerret S, et al. Effect of γ-interferon on the
808 Nedelec et al
20.
21.
22.
23.
24.
25.
26.
27. 28.
29.
30. 31.
32.
clinical and biologic evolution of hypertrophic scars and Dupuytren’s disease: an open pilot study. Plast Reconstr Surg 1994;93:1224-35. Darby I, Skalli O, Gabbiani G. Alpha-smooth muscle actin is transiently expressed by myofibroblasts during experimental wound healing. Lab Invest 1990;63:21-9. Desmouliere A, Redard M, Darby I, Gabbiani G. Apoptosis mediates the decrease in cellularity during the transition between granulation tissue and scar. Am J Pathol 1995;146:56-66. Garbin S, Pittet B, Montandon D, Gabbiani G, Desmouliere A. Covering by a flap induces apoptosis of granulation tissue myofibroblasts and vascular cells. Wound Repair Regen 1996;4:244-51. Kischer CW, Pindur J, Krasovich P, Kischer E. Characteristics of granulation tissue which promote hypertrophic scarring. Scanning Microsc 1990;4:877-88. Dao T, Ariyasu T, Holan V, Minowada J. Natural human interferon-α augments apoptosis in activated T cell line. Cell Immunol 1994;155:304-11. Egle A, Villunger A, Kos M, Bock G, Gruber J, Auer B, et al. Modulation of Apo-1/Fas (CD95)-induced programmed cell death in myeloma cells by interferon-alpha 2. Eur J Immunol 1996;26:3119-26. Rodriguez-Villanueva J, McDonnell TJ. Induction of apoptotic cell death in non-melanoma skin cancer by interferonα. Int J Cancer 1995;61:110-4. Sullivan T, Smith J, Kermode J, McIver E, Courtemanche DJ. Rating the burn scar. J Burn Care Rehabil 1990;11:256-60. Gavrieli Y, Sherman Y, Ben-Sasson SA. Identification of programmed cell death in situ via specific labeling of nuclear DNA fragmentation. J Cell Biol 1992;119:493-501. Tredget EE, Wang R, Shen Q, Scott PG, Ghahary A. Transforming growth factor-beta mRNA and protein in hypertrophic scar tissues and fibroblasts: antagonized by interferon-α and interferon-γ in vitro and in vivo. J Interferon Cytokine Res 2000;20:143-51. Blackburn WR, Cosman C. Histologic basis of keloid and hypertrophic scar differentiation. Arch Pathol 1966;82:65-71. Linares HA, Kischer CW, Dobrkovsky M, Larson DL. The histiotypic organization of the hypertrophic scar in humans. J Invest Dermatol 1972;59:323-31. Ghahary A, Shen YJ, Scott PG, Tredget EE. Enhanced expression of mRNA for transforming growth factor-β, type I and type III procollagen in human post-burn hypertrophic scar tissues. J Lab Clin Med 1993;122:465-73.
Surgery November 2001 33. Desmouliere A, Geinoz A, Gabbiani F, Gabbiani G. Transforming growth factor β1 induces α-smooth muscle actin expression in granulation tissue myofibroblast and in quiescent and growing cultured fibroblasts. J Cell Biol 1993;122:103-11. 34. Nedelec B, Shankowsky HA, Tredget EE. Rating the resolving hypertrophic scar: comparison of the Vancouver scar scale and scar volume. J Burn Care Rehabil 2000;21:205-12. 35. Scott PG, Dodd CM, Tredget EE, Ghahary A, Rahemtulla F. Immunohistochemical localization of the proteoglycans decorin, biglycan and versican and transforming growth factor-β in human post-burn hypertrophic and mature scars. Histopathology 1995;26:423-31. 36. Scott PG, Dodd CM, Tredget EE, Ghahary A, Rahemtulla F. Chemical characterization and quantification of proteoglycans in human post-burn hypertrophic and mature scars. Clin Sci 1996;90:417-25. 37. Desmouliere A, Badid C, Bochaton-Piallat ML, Gabbiani G. Apoptosis during wound healing, fibrocontractive diseases and vascular wall injury. Int J Biochem Cell Biol 1997;29:1930. 38. Chesney J, Metz C, Stavitsky AB, Bacher M, Bucala R. Regulated production of type I collagen and inflammatory cytokines by peripheral blood fibrocytes. J Immunol 1998;160:419-25. 39. Wassermann RJ, Polo M, Smith P, Wang X, Ko F, Robson MC. Differential production of apoptosis-modulating proteins in patients with hypertrophic burn scar. J Surg Res 1998;75:74-80. 40. Nedelec B, Dodd C, Scott PG, Ghahary A, Tredget EE. The effect of interferon-α2b on guinea pig wound closure and the expression of cytoskeletal proteins in vivo. Wound Repair Regen 1998;6:202-12. 41. Iredale JP, Benyon RC, Pickering J, McCullen M, Northrop M, Pawley S, et al. Mechanisms of spontaneous resolution of rat liver fibrosis: Hepatic stellate cell apoptosis and reduced hepatic expression of metalloproteinase inhibitors. J Clin Invest 1998;102:538-49. 42. Shang K, Rekhter MD, Gordon D, Phan SH. Myofibroblast and their role in lung collagen gene expression during pulmonary fibrosis. Am J Pathol 1994;145:114-35. 43. Mounier N, Desmouliere A, Gabbiani B. Subcutaneous tissue fibroblast transfected with muscle and nonmuscle actins: a good in vitro model to study fibroblastic cell plasticity. Wound Repair Regen 1999;7:45-52.
CORRECTION In the article "The role of iNOS in wound healing" (Surgery 2001;130:225-9), two authors were not noted. The byline should be as follows: Han Ping Shi, MD, PhD, Daniel Most, MD, David T. Efron, MD, U. Tantry, MD, M. H. Fischel, MD, and Adrian Barbul, MD, FACS