Results

HTS fibroblasts fail to undergo apoptosis in response to collagen contraction/remodelling cues

In 1998, Fluck et al published the finding that fibroblasts derived from normal skin dermis, when seeded into three-dimensional- (3D-) collagen gels and allowed to contract and therefore remodel those gels, were induced to undergo apoptosis. We therefore determined whether fibroblasts derived from different scar types would be equally susceptible to such a phenomenon. Since the degree to which the cells proliferate in collagen gels differs between anchored and contractile gels, we titrated the amount of serum used to supplement the medium down to that which kept the cells quiescent but healthy (1%—not shown). All further experimentation was performed in this minimal growth medium in order to ensure that any differences seen in cell number were as a result of cell death and not different proliferation rates. NS cells seeded into collagen gels in minimal medium, when allowed to contract the gels, underwent significant cell death (35%–45% depending on cell strain, p<0.001), reaching a maximum at approximately 3–4 d (Figure 1a). This drop in viable cell number was not seen in gels that remained anchored. In contrast, cells derived from HTS failed to undergo cell death under either circumstance. The results shown are representative of those obtained for all NS (n=8) and HTS (n=6) fibroblast strains tested. The typical appearance of cells within the gels after 4 d is shown in Figure 1b–e: both NS cells (B) and HTS cells (D) in anchored collagen gels are seen to be distributed randomly, being mainly bipolar but with occasional polydendritic cells present. A large proportion of the NS cells in contractile collagen gels (C) are rounded up (black arrows), exhibiting dense compacted chromatin typical of that seen in apoptotic cells; the remaining cells appear multistellate and occasionally bipolar. In contrast, very few HTS cells in contractile collagen gels were rounded up (E—black arrow), being mainly multistellate in appearance (white arrow), with the extensive network of dendrite-like structures often extending outside the plane of focus. A terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling technique confirmed that the cell death seen was as a result of apoptosis. A typical example of this is shown in Figure 2: (A) demonstrating end-labelled nuclei of NS fibroblasts harvested from a 4-d-old contractile gel, whereas (B) shows the unlabelled nuclei typical of HTS cells harvested from a 4-d-old contractile gel. Interestingly, the ability of either cell type to contract collagen gels appears equivalent, with both NS and HTS cells contracting the gels by approximately 55% of the original gel area by 4 d (Figure 3).

Figure 1.

Collagen contraction/remodelling-induced cell death. Fibroblasts derived from either normal scars (NS) or hypertrophic scars (HTS) were seeded into collagen gels and incubated for 3 d, after which time they were either maintained anchored or released and allowed to contract the gel for a 4-d period. At daily time points, the cells were harvested and the number of apoptotic cells was assessed through viable staining and nick end-labelling of apoptotic nuclei. Typical results are presented in (A), which shows a graph of the percentage of cells undergoing apoptosis versus time: NS cells in anchored gels (open circles), NS cells in contractile gels (closed circles), HTS cells in anchored gels (open triangles), and HTS cells in contractile gels (closed triangles). Each point represents the mean of triplicate gels, with the error bars being the standard deviation from the mean. Micrographs B–E illustrate the typical cellular morphology seen in each type of gel (formal saline fixed and hematoxylin stained): (B) NS, anchored; (C) NS, contractile; (D) HTS, anchored; (E) HTS, contractile. Black arrows indicate rounded-up dying cells and the white arrow indicates live fibroblasts within the HTS cell-containing contractile gel. Scale bar=50 m.

Full figure and legend (161K)

Figure 2.

Confirmation of apoptosis. Typical terminal deoxynucleotidyl transferase end-labelling of apoptotic nuclei (green fluorescence) with red (propidium iodide) counterstaining of cells harvested from collagen gels that had been allowed to contract for 4 d: normal scar cells (A) or hypertrophic scar cells (B). Scale bar=25 m.

Full figure and legend (4K)

Figure 3.

Degree of collagen contraction. Typical examples of collagen gels that had been allowed to contract for 4 d that contained either normal scar (NS) (A) or hypertrophic scar (HTS) (B) cells. (C) shows the mean percentage contraction of triplicate gels calculated as the mean percent reduction in diameter of the gels from the original (measured as an average of five random diameters of each gel) and the results of a standard Student's t test.

Full figure and legend (58K)

Apoptosis defect specific to that induced by collagen-gel contraction

To determine whether this inability of HTS cells to undergo apoptosis is a general phenomenon or specific to this particular induction signal, we examined whether HTS cells were sensitive to a variety of different methods of inducing apoptosis, including ethanol, doxorubicin, and camptothecin. The mean cell death for n=3 different cell strains of each scar-tissue type are plotted in Figure 4d (ethanol) and 4E (doxorubicin). HTS fibroblasts demonstrated response curves identical (p values>6) to NS cells irrespective of the apoptotic stimuli tested. Cells treated with 10% ethanol in serum-free medium (SFM) show considerable rounding up and death (Figure 4b), which steadily increased over time until few cells remained alive after 8 h. Treatment with 1 g per mL of either doxorubicin (Figure 4c) or camptothecin (not shown) in normal growth medium (containing 10% serum) resulted in cell death that steadily increased over time, reaching a maximum at 24 h (Figure 4e). In all cases, cell death was attributed to apoptosis by nick end-labelling (not shown).

Figure 4.

Chemically induced apoptosis. Micrographs illustrating hypertrophic scar (HTS) cell death induced by standard inducers of cellular apoptosis. (A) Appearance of cells in normal medium, (B) induction of apoptosis after 7 h in 10% ethanol, and (C) apoptosis of cells after 24 h treatment with 1 g per mL doxorubicin. Scale bar=100 m. (D, E) Graphical representation of cell death over time induced by 10% ethanol and 1 g per mL doxorubicin, respectively. Data plotted are the mean of n=4 cell strains of either normal scar (NS) (circles) or HTS (triangles), with error bars being the standard deviation from the mean.

Full figure and legend (57K)

Gels inhabited by HTS cells are harder to break down compared with gels inhabited by NS cells

Under the quiescent conditions in which we performed our collagen-contraction experiments, the major differences between anchored versus contractile collagen gels that might influence the induction of apoptosis are either mechanical (altering cellular stress/tension) or related to collagen breakdown/synthesis. Either of these putative mechanisms could be affected by changes in the post-translational modification of newly laid down collagen, and there is increasing evidence of biochemical overmodification of ECM in conditions linked with fibrosis. We therefore examined whether cells derived from the different scars differed in their ability to modify the collagen gel.

Changes in modification of the collagen by the two cell types were initially examined using collagenase D digestion. Sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis of digestion products showed little significant difference in protein bands over 50 kD in size. Typically, the majority of differences were seen in breakdown products of less than 50 kD (Figure 5). In gels conditioned for 3-d the only obvious difference is that the NS cell-gel product of approximately 10 kD (arrow, lane 1) is absent in the HTS cell-gel products (lane 2). After 7 d, the differences become more marked, with many bands either not present, reduced in density, or appearing to exhibit an altered mobility in HTS cell gels (arrows, lane 3 vs lane 4). A plot of the 7-d samples clearly shows the disappearance or reduction of bands at approximately 48, 39, 35, 33, 21, 20, 19, and 10 kD. These results appear to suggest modification of the collagen gel by HTS cells in such a way as to dramatically change the products of collagenase digestion.

Figure 5.

Collagenase D digestion of cell-conditioned collagen. Collagen gels seeded with fibroblasts derived from either normal scar (NS) or hypertrophic scar (HTS) were cultured for the whole time course of the collagen contraction-induced apoptosis experiment (7 d). Gels were harvested either at day 3 (the point at which the gels are normally freed to allow contraction) or day 7 (i.e. 4 d after contraction began and the point at which apoptosis is usually maximal). The gels were digested using collagenase D and the digestion products were analyzed on 12% sodium dodecylsulfate-polyacrylamide gel electrophoresis gel under non-reducing conditions. (A) A typical gel: Lanes 1 (day 3 specimen) and 3 (day 7) are NS cell-conditioned gels; lanes 2 (day 3) and 4 (day 7) are HTS cell-conditioned gels. The arrows indicate bands that have altered between NS and HTS lanes of the same time point. (B) A plot of the density of lane 3 (dashed line) and 4 (solid line) versus molecular weight. The arrows indicate major changes in banding patterns between the NS and HTS cell-conditioned gels.

Full figure and legend (102K)

We therefore went on to examine the susceptibility of such gels to MMP-2, an enzyme responsible for the production of smaller breakdown products during collagen remodelling and active during cutaneous wound healing. The proteolytic products of MMP-2-digested cell-conditioned collagen were then analyzed via SDS-PAGE. The typical appearances of the different collagen chains are seen in undigested acid-soluble collagen samples from both NS and HTS cell-conditioned gels (Figure 6a, lanes 1 and 2, respectively); 1 (130 kD), 2 (120 kD), and (270—300 kD), with the chains being just discernible at the top of the gel. For the MMP-2-digested collagen samples, both the and chains of the NS cell-conditioned gel have been completely digested, whereas the chain band (130 kD) of HTS cell-conditioned collagen is still quite obvious (Figure 6b, lanes 1 and 2). The digested NS sample does show a strong band of higher mobility at approximately 100 kD that may represent a breakdown product of the collagen digestion. Both gel types exhibited large amounts of breakdown products of less than 37 kD, but their banding patterns were too indistinct to determine any reproducible differences between the two gels (not shown).

Figure 6.

Matrix metalloproteinase-2 (MMP-2) digestion of cell-conditioned collagen. Acid-soluble fractions of cell-conditioned collagen gels were heat denatured, digested with MMP-2 for 24 h, and analyzed using a standard non-reducing sodium dodecylsulfate-polyacrylamide gel electrophoresis 7.5% gel. Collagen samples before digestion are shown in A, lanes 1 and 2—normal scar (NS) and hypertrophic scar (HTS) cell-conditioned collagen, respectively, showing collagen bands , , and chains. MMP-2-digested collagen samples are in (B) lanes 1 and 2—NS and HTS cell-conditioned collagen, respectively.

Full figure and legend (62K)

Tissue transglutaminase is overexpressed by HTS cells in vivo and in 3D-collagen gels but not in monolayer culture

One of the enzymes known to be responsible for biochemical cross-linking and thus stabilization of ECM is tissue transglutaminase. Immunohistochemical staining of tissue samples taken from the scars from which the cell cultures were derived corroborated the previous report (^{Dolynchuk, 1996}) that mature HTS exhibit high expression of tissue transglutaminase on fibroblastic cells within the scar matrix, whereas NS of a similar age do not (Figure 7a, b). To determine whether if this was also the case in cultured HTS cells, compared with NS cells immunohistochemical staining was attempted on monolayer cultures, but with little success, with tissue transglutaminase being barely detectable in either type of scar fibroblasts cultured on plastic. When these cells are cultured in 3D-collagen gels, however, tTgase expression is detectible in both cell types (Figure 7c, d), with a particularly marked upregulation in cells derived from HTS. This was confirmed using SDS-PAGE and western blotting (not shown).

Figure 7.

Tissue transglutaminase expression and modification of collagen. Immunohistochemical staining for tissue transglutaminase: (A) normal scar (NS) tissue>6 mo maturity—dotted arrow indicates staining in blood vessel; solid arrow indicates staining in the basal layer of the epidermis, (B) chronic hypertrophic scar (HTS) tissue (>6 mo maturity)—dotted arrow indicates staining in blood vessels; solid arrows indicate staining of fibroblastic cells within the scar matrix, (C) NS cells seeded into collagen gel and cultured for 7 d–arrow indicates weak staining of occasional cells, with the vast majority of cells having undetectable staining, (D) HTS cells seeded into collagen gel and cultured for 7 d—showing that the majority of cells exhibit strong staining for this enzyme. Scale bar=50 m. (E) shows 12% sodium dodecylsulfate-polyacrylamide gel electrophoresis gel analysis typical of collagenase-D-digested (40 min) collagen gels, which had been seeded with NS fibroblasts and cultured for 7 d in the absence (control) or presence of 0.0144 U per mL of tTgase: Lane 1 (untreated control) and 2 (tTgase-treated). The arrows indicate the main bands that have altered between untreated and tTgase-treated samples. (F) A plot of the density of lane 1 (dashed line) and 2 (solid line) versus molecular weight. The arrows indicate major changes in banding patterns between the untreated and tTgase-treated collagen gels.

Full figure and legend (118K)

Tissue tranglutaminase treatment of collagen inhabited by NS cells mimics the pattern of collagenase breakdown products of HTS cell-conditioned collagen

Collagen gels seeded with NS cells and cultured anchored for 7 d in the absence or presence of 0.0144 U per mL tissue transglutaminase were subject to 40 min collagenase D digestion and the breakdown products were analyzed. Similar to the comparison of HTS- and NS cell-conditioned gels, there was no discernible difference in digestion product bands over 50 kD in size, with the majority of differences being seen in breakdown products of less than 50 kD. A representative example is shown in Figure 7e: Lane 1 is the untreated NS gel and lane 2 is the exogenous tTgase-treated NS gel. The tTgase treatment produced a markedly different banding pattern (the major differences are indicated with arrows). This proteolytic profile appears similar to that seen for HTS cell-conditioned gels (Figure 5). A plot of lane density against molecular weight clearly shows the disappearance or reduction of bands at approximately 44, 39, 36, 34, and a triplet centering around 21 kD (Figure 7f). These results suggest that addition of exogenous tTgase dramatically changes the products of collagenase digestion in such a way as to mimic that of collagen gels conditioned by HTS cells.

Inhibition of transglutaminase activity in HTS cells normalizes their behavior in response to collagen contraction/remodelling

The naturally occurring polyamine known as putrescine, or 1,4-diaminobutane (1,4-DAB), is known to inhibit the cross-linking activity of tissue transglutaminase in vitro and in vivo. We therefore attempted to modulate the overactivity of tTgase seen in HTS cells back toward more normal levels using 1,4-DAB and to examine what effect this had on the inability of these cells to undergo collagen-contraction/remodelling-induced apoptosis (CrIA). Initially, we assayed the effect of a range of 1,4-DAB concentrations in minimal growth medium on the proliferation of fibroblasts in monolayer culture over a period of 10 d, in order to assess whether 1,4-DAB had any mitogenic or toxic effects in the absence of a 3D-collagenous environment. Typically, the only difference was found at the highest concentration (500 g per mL) where the cell number was significantly lower (p<0.01) than that of the media-only control (at day 10 only—Figure 8a). A 1,4-DAB concentration of 250 g per mL was therefore chosen for further experimentation.

Figure 8.

Inhibition of tissue transglutaminase activity in hypertrophic scar (HTS) cell-containing collagen gels. The effects of 1,4-diaminobutane (1,4-DAB) on the cell number of fibroblasts cultured in monolayer in minimal growth medium were titrated over a 10-d period (A). ^* indicates the concentration at which the cell number was significantly (p<0.01) lower than that of the media-only control (at day 10 only). HTS fibroblasts seeded into collagen gels were incubated for 3 d either in the presence or absence of 250 g per mL 1,4-DAB, after which time, they were either maintained anchored or allowed to contract the gel in these media for a further 4-d period. At day 4 (post-release), the cells were harvested from the gels using collagenase D and the number of apoptotic cells was assessed through viable staining and nick end-labelling of apoptotic nuclei. Typical results are presented in (B), which shows a graph of the percentage of cells undergoing apoptosis: anchored gels (solid bars) and contractile gels (striped bars). Each point represents the mean of triplicate gels, with the error bars being the standard error of the mean (^*p<0.001).

Full figure and legend (42K)

Fibroblasts derived from HTS seeded into collagen gels were incubated for 3 d either in the presence or absence of 250 g per mL 1,4-DAB, after which time they were either maintained anchored or allowed to contract the gel in these media for a further 4-d period, whereupon cell death was assessed. The experiments were repeated at least three times with n=3 cell lines. Typical results plotted in Figure 8b show that addition of 1,4-DAB to the medium during the course of the experiment had a profound effect on the ability of collagen contraction to induce apoptosis (p<0.001) of HTS cells, with the level of death (47%Figure 8b) being similar to that seen with NS cells (Figure 1a).

Treatment of collagen gels with exogenous transglutaminase inhibits apoptosis of NS cells in response to collagen contraction/remodelling

Polyamines are pleiotropic molecules that could potentially affect cell death in many different ways other than solely through the inhibition of the action of tTgase. We therefore used an alternative approach and examined whether an increase in tTgase activity in collagen gels inhabited by NS fibroblasts would engender resistance to CrIA in these "normal" cells. Again, we first determined whether tTgase had any effect (either mitogenic or toxic) on the cells in the absence of a 3D-collagenous environment, by performing growth curves using monolayer cultures of NS fibroblasts. The titration curve at each time point is plotted in Figure 9a and shows that none of the concentrations of tTgase used significantly affected the growth or death of these cells. A tTgase concentration of 0.0144 U per mL was used in all further experimentation and is of the same order of concentration as that used by^{Murthy et al (1991)} to cross-link fibrinogen.

Figure 9.

Treatment of normal scar (NS) cell-containing collagen gels with tissue transglutaminase. The effects of added exogenous tissue transglutaminase on the cell number of fibroblasts cultured in monolayer in minimal growth medium were titrated over a 10-d period—no significant difference was seen in the cell number over the whole titration range tested. NS fibroblasts seeded into collagen gels were incubated for 3 d either in the presence or absence of 0.0144 U per mL tissue transglutaminase, after which time, they were either maintained anchored or allowed to contract the gel in these media for a further 4-d period. At day 4 (post-release), the cells were harvested from the gels using collagenase D and the number of apoptotic cells assessed through viable staining and nick end-labelling of apoptotic nuclei. Typical results are presented in (B), which shows a graph of the percentage of cells undergoing apoptosis: anchored gels (solid bars) and contractile gels (striped bars). Each point represents the mean of triplicate gels, with the error bars being the standard error of the mean. (^*p<0.001).

Full figure and legend (43K)

Fibroblasts derived from NS seeded into collagen gels were incubated for 3 d either in the presence or absence of 0.0144 U per mL tissue transglutaminase, after which time they were either maintained anchored or allowed to contract the gel in these media for a further 4-d period, after which cell death was assessed. Experiments were repeated at least three times with n=3 cell lines. Typical results plotted in Figure 9b show that, in contrast to the untreated gels (p<0.001), NS cells in contractile collagen gels in the presence of added exogenous tissue transglutaminase did not undergo apoptosis (p>0.8).

Discussion

Recently, a role for fibrillar collagen has been indicated in the induction of apoptosis of fibroblasts (^{Fluck et al, 1998;Grinnell et al, 1999}), thus establishing an experimental stimulus of apoptosis with relevance to the healing wound. Considering the long-hypothesized possibility that the hypercellular nature (and potentially the whole pathology) of hypertrophic scarring is caused by a failure of apoptosis, we therefore investigated whether a reduced susceptibility of fibroblasts to this particular form of apoptosis might underlie the pathology of this condition.

We have demonstrated that cell lines derived from chronic (>6-mo old) HTS are resistant to this CrIA, whereas fibroblasts derived from NS of similar maturity mimic dermal fibroblasts and are susceptible to CrIA. These findings highlight distinct patterns of cellular behavior exhibited by these cells indicating intrinsic cellular differences rather than purely environmental cues in the pathology of hypertrophic scarring. Furthermore, we have shown that a general defect in the machinery of apoptosis is not indicated, but rather something more particular to the specific form of apoptosis induction going on within a contractile 3D-collagen structure. These findings are not only in accordance with the hypothesis that the hypercellular characteristic of hypertrophic scarring is because of a failure of cellular apoptosis but also supports the proposition that collagen is involved in the mechanism that switches off the fibroproliferative phase of wound healing.

We further postulated that this phenomenon of apoptotic resistance might be caused by the matrix of HTS being overmodified in some way, either via intracellular post-translational modification that goes on during biosynthesis or through extracellular biochemical modification, since chronic fibrotic tissue (both skin and liver) is known to be excessively cross-linked (^{Ricard-Blum et al, 1993,1996,1998;Dolynchuk, 1996;Hirota et al, 2003}). Overmodification of ECM would considerably change its biochemical properties and how cells interact with it, potentially affecting the mechanical properties of the matrix and thus the tensional load put upon the inhabiting cells. Excessive cross-linking of ECM might also result in hidden or uncovered cellular binding sites within the matrix proteins themselves, or make available or sequester growth factors or the like. Furthermore, this type of modification would affect the quality of enzymatic cleavage, potentially making the matrix more refractive to enzymatic breakdown.

We demonstrated that HTS cells biochemically modify collagen in such a way as to change the products of proteolytic cleavage that result from digestion with a strong collagenase (Collagenase D). In addition, HTS cell modification of collagen causes resistance to enzymatic cleavage by MMP-2, a collagenase highly active during cutaneous wound healing. These results are analogous to recent findings in a model of chronic dermatitis induced in rabbit ears (^{Hirota et al, 2003}) and are of particular interest considering work by^{Buckley et al (1999)}, who implicated potential breakdown products of the ECM in the initiation of apoptosis in both lymphocytic cells and fibroblasts. Small soluble peptides containing the RGD amino acid motif were reported to directly induce apoptosis through an integrin-independent mechanism by direct binding to and activation of pro-caspase-3 as apposed to the alternative mechanism of inducing apoptosis (termed anoikis) through blocking integrin-mediated cell-matrix contact (^{Frisch and Ruoslahti, 1997}). It is possible that this overmodification and thus stabilization of collagen by HTS cells might reduce the production of small soluble RGD-containing peptides that occurs on remodelling, thereby removing a potential trigger for apoptosis.

The stabilization of newly formed ECM is thought to be largely accomplished through the actions of tissue transglutaminase (or transglutaminase II) (^{Ricard-Blum et al, 1996;Grenard et al, 2001}). The activity of this enzyme is markedly increased during wound healing, substantially effecting both the breaking strength of wounds and the solubility of ECM proteins (^{Bowness et al, 1987b,1988;Dolynchuk et al, 1994;Haroon et al, 1999}). Tissue transglutaminase cross-links proteins through the formation of isopeptide bonds between reactive -glutaminyl groups in certain specific proteins and the -amino group of lysine in other proteins. Some of the components of early granulation tissue, type III collagen along with its aminopropeptides, have been shown to be particularly good substrates for transglutaminase (^{Bowness et al, 1987a,1987b}), and elevated levels of these have been detected in both HTS (^{Weber et al, 1978}) and related fibrotic conditions such as Dupuytrens disease (^{Brickley-Parsons et al, 1981}). Since the affected fascia of Dupuytrens disease is also known to have increased transglutaminase levels (^{Dolynchuk et al, 1991}), and increased extracellular activity of this enzyme has been detected in chronic HTS tissue (^{Dolynchuk, 1996}), it seems feasible that the transglutaminase cross-linking of ECM might be expected to be excessive in these conditions.

We have not only demonstrated high expression levels of tissue transglutaminase in chronic HTS tissue, corroborating the previous report (^{Dolynchuk, 1996}), but also in HTS cells in vitro, but only when cultured within 3D-collagen gels and not in monolayer culture. This apparent matrix dependence of the expression, stabilization or extracellular localization of tissue transglutaminase is in contrast to a recent report comparing high tTgase-expressing pre-senescent, senescent, and immortalized dermal fibroblasts, which reported no dependence on culture environment (^{Stephens et al, 2004}), and thus requires further investigation. In support of the theory that the overexpression of cell surface tTgase might explain the change in resistance to enzymatic digestion, treatment of NS cell-seeded collagen gels with exogenous tTgase was shown to alter the products of proteolytic breakdown in such a way as to mimic that of HTS cell-seeded gels. We went on to manipulate the activity of tissue transglutaminase in cell-containing collagen gels and found that inhibition of tTgase activity in collagen gels containing HTS cells resulted in the normalization of their behavior; allowing induction of apoptosis on gel contraction, whereas boosting tTgase activity in gels containing NS cells completely abrogated CrIA. These findings establish a role for tTgase in the protection from CrIA and suggest that the prolonged overexpression of this enzyme by HTS fibroblasts may be the basis of their pathology.

A therapeutic use of 1,4-DAB as an inhibitor of transglutaminase activity in HTS has been considered previously. A phase II clinical study was set up in human patients with HTS to primarily determine safety but also the efficacy of a topical cream containing 0.8% wt/vol 1,4-DAB (Fibrostat Procyon Biopharma, Inc., Dorval, Canada) (^{Dolynchuk et al, 1996}). The trial was set up as a double-blind crossover study over two successive 4-wk periods, which showed that Fibrostat^® treatment resulted in a significant improvement in scar appearance irrespective of the order given. Further work used HPLC analysis to quantitate the (-glutamyl) lysine cross-linking of ECM proteins and found that these were significantly reduced in Fibrostat^®-treated HTS tissue compared with vehicle-only control-treated scar (^{Dolynchuk, 1996}), thus supporting a possible therapeutic role for topical 1,4-DAB in the treatment of HTS. Our results suggest that the mechanism of therapeutic action could be via allowing the HTS fibroblasts to finally undergo apoptosis, thus reducing hypercellularity of the scars and speeding their regression.

How tissue transglutaminase activity might contribute toward the sensitivity of a cell to the specific form of apoptosis induced by collagen contraction is a question that still needs addressing. This enzyme is a highly multifunctional molecule exhibiting a plethora of new and diverse bioactivities, a comprehensive review of which, including the cell structure-stabilizing role of intracellular transglutaminase during apoptosis, is given by^{Griffin et al (2002)}. Indeed, tTgase is unique in that it has multiple enzymatic functions. Apart from its cross-linking activity, it has long been known that tTgase also directly binds guanosine triphosphate (GTP)/guanosine diphosphate and undergoes a GTPase cycle, thus allowing this enzyme to function directly in signal transduction (^{Nakaoka et al, 1994;Mian et al, 1995}). In addition, tTgase has also recently been reported to have intrinsic kinase activity, phosphorylating insulin-like growth factor binding protein-3 (^{Mishra and Murphy, 2004}). Each of any of these three distinct enzymatic activities exhibited by tTgase could affect apoptosis in multiple different ways. Our findings that 1,4-DAB, which specifically inhibits the cross-linking activity of extracellular tTgase, and added exogenous tTgase both affect susceptibility to CrIA however, suggests that this phenomenon is strictly affected by the cross-linking activity of extracellular (presumably cell membrane-bound) tTgase and not through this enzyme's alternative functions within the cell. Interestingly, CrIA must not therefore involve fibroblast spreading, motility, nor their ability to form stress fibers or focal adhesions, since all of these cellular functions have recently been found to be dependent on the G-protein function of tTgase rather than its cross-linking ability (^{Stephens et al, 2004}). Our observation of equivalent contraction by HTS and NS cell-inhabited collagen gels would appear to suggest that contraction is not affected by overactivity of tTgase and that CrIA does not necessarily follow contraction. Whether contraction in high tTgase activity gels differs in a wholly qualitative way however is unknown.

Since cell surface transglutaminase is well known for its stabilization of ECM and its ability to reduce the solubility or degradability of these proteins, it thus has the potential to reduce the production of small RGD-containing peptides that might play a role in the induction of apoptosis as mentioned previously. Alternative roles for this particular enzymatic function, however, are being uncovered, which also have the potential to affect cell survival. Whether these play a role during collagen contraction is currently under investigation. Transglutaminase has been implicated in the activation of transforming growth factor-1 (TGF-1) via cross-linking of latent TGF--binding protein-1 (^{Nunes et al, 1997;Verderio et al, 1999}). TGF--mediated apoptosis has been implicated in the formation and homeostasis of many different tissues including the immune system (particularly B- and T-cell homeostasis), liver, prostate, uterus, breast, nervous system (^{Schuster and Krieglstein, 2002}). More recently, the importance of TGF--induced apoptosis has been demonstrated during tissue formation and remodelling that takes place on cutaneous wound healing. Here, using a TGF-/Scid double-knockout mouse to remove any exacerbated inflammatory phase found in TGF-1-only knockout mice, a substantial delay in each of the major phases of wound healing was found. This included a delayed and reduced onset of apoptosis localized in the granulation tissue, normally found occurring just beneath the advancing edge of migrating epithelium (^{Crowe et al, 2000}). It is possible therefore that perturbation of the TGF- system or its related co-factors by prolonging or overexpressing tTgase during the contraction phase of wound healing may determine whether apoptosis will take place. Furthermore, irrespective of any putative effects on apoptosis, the potential involvement of tTgase in activation of such an important profibrotic protein as TGF- combined with the overexpression of this enzyme by HTS has obvious implications for the pathology of this disease.

Finally, a number of ECM proteins are known to be bound and cross-linked by tTgase; among those molecules with the highest binding affinity is fibronectin (^{Gaudry et al, 1999;Akimov et al, 2000}), a molecule known to give survival signals to cells via integrin interactions. In recent years, Griffin's group has found that fibronectin-bound tTgase supports cell adhesion in its own capacity and in an integrin-independent manner, which inhibits anoikis (^{Verderio et al, 2003}). They reported that cells bound to the fibronectin/tTgase complex did not undergo anoikis when treated with the small soluble RGD-containing peptides that cause anoikis of cells bound to fibronectin. The role of fibronectin in tTgase-related protection of HTS cells against CrIA is currently under investigation.

The matter of exactly how tTgase is involved in the pathogenesis of pathological scarring is still unresolved but clearly merits further study. Our results do suggest that the continued development of inhibitors of tTgase as therapeutic anti-scarring agents may be a worthy endeavor and that the application of these might be particularly effective as a prophylactic measure in the weeks following successful wound closure. The exact mechanisms via which apoptosis is induced in these gels also remains uncertain, whether they be through changes in mechanical forces, interactions of cell receptors with ECM molecules, interactions of ECM with other effectors of cellular behavior, or via ECM remodelling. What is clear is that this failing of pathological cells in this particular respect not only has clinical and therapeutic implications but also may be a useful tool by which to unravel the exact mechanics of apoptosis induction under normal wound healing circumstances.

Methods

Cell culture

Fibroblast cultures were initiated from explants of redundant scar tissue from a variety of body sites left over from elective surgical procedures. All patients gave informed consent; the study has been approved by the West Herts Hospitals Trust Ethics Committee and has been conducted according to the Declaration of Helsinki principles. Cultures were initiated from eight NS (3:5 ratio of male to female; mean age 26.5 y, range 8–85; scar maturity>6 mo) and six HTS–(4:2 ratio of male to female; mean age 18.3 y, range 6–51; scar maturity>6 mo). Cells were maintained in Dulbecco's modified Eagles medium (DMEM) supplemented with 10% fetal calf serum (FCS), L-glutamine, penicillin/streptamycin. Cultures were initiated in the following manner: the central area of obvious scar tissue was dissected for culture initiation to avoid contaminating cultures with surrounding normal stroma. The selected tissue was finely minced with iris scissors and placed onto the culture flask surface. Normal growth medium was then gently added. Fibroblasts were seen to rapidly grow out from the tissue and cover the plastic surface area; these adherent cells were harvested by trypsinization and passaged. Cells of up to passage number 6 were used for experimentation.

3D-collagen gels

Collagen was prepared from rat tail tendons by a modification of the extraction method of^{Bornstein (1958)}. Hydrated collagen gels were prepared by buffering the acidic collagen solution by adding 2% (vol/vol) of 11% NaHCO₃ and 1.4% 1 M N-2-hydroxyethylpiperazine-N-2-ethane-sulfonic acid (pH 8) on ice and neutralizing the pH by adding 1 M NaOH dropwise. This mixture was further supplemented with 10 minimal essential medium (Gibco Invitrogen, Ltd., Paisley, UK), and finally, cells were added to give a concentration of 2 10⁵ cells per mL. This cell-containing collagenous mixture was then plated into six-well plates (1.5 mL per well) and allowed to set at 37°C over several hours. A minimal growth medium (DMEM containing 1% FCS) was then added to the gels and they were cultured for 3 d before further experimentation. After this time, the media were refreshed and the "contractile" gels were freed from the plastic surface using the blunt end of a scalpel and allowed to float freely in the medium. The "anchored" gels were maintained attached to the plastic surface. Both gel types were further cultured for 4 d when they were harvested for photography and further analysis.

Gel solubilization and cell recovery

Collagen gels were washed twice in phosphate-buffered saline (PBS) and the gel was solubilized by incubation at 37°C with 1 mL per gel collagenase D (0.5 mg per mL; Roche Diagnostics, Ltd, Lewes, UK) in PBS containing 5 mg per mL bovine serum albumin. After approximately 30–40 min, when all the gel had solubilized the released cells were harvested by centrifugation and standard viability staining (trypan blue) was performed or cells were air-dried onto glass slides for further analysis.

Terminal deoxynucleotidyl transferase-mediated dUTP nick end-labelling of apoptotic cells

Identification of apoptotic cells was achieved using an adapted protocol of the APO-BRDU kit (BD Pharmingen, San Diego, California). Briefly, cell suspensions on glass slides were fixed for 10 min in ice-cold methanol and air-dried. DNA-labelling solution (made up as directed by kit; containing Br-dUTP, TdT enzyme and reaction buffer) was added to the slides and incubated for 1 h at 37°C. The slides were then washed twice with rinse buffer (kit component) and a diluted antibody solution of fluorescein-labelled anti-BrdU-added and incubated in the dark for 30 min at room temperature (RT). Finally, a propidium iodide/Rnase solution (kit component) was added to the slide and incubated for a further 30 min in the dark at RT. Slides were then mounted in an anti-fade mountant and visualized using fluorescence microscopy.

Chemical induction of apoptosis

Attempts were made to induce apoptosis of the cells using a variety of alternative methods. Ethanol toxicity was used to induce apoptosis by placing the cells into 10% (vol/vol) ethanol in SFM. Cells were observed microscopically over an 8 h period for signs of toxicity (rounding up and detaching from surface), and viable cell counts were performed. Two cytotoxic drugs were also used to induce apoptosis: doxorubicin and camptothecin. Both were used over a range of concentration levels (not shown—optimum apoptosis was achieved at 1 g per mL), and the cell death was quantified over a 2-d time course. Any death seen was confirmed as occurring via apoptosis using nick end-labelling.

Extraction and purification of cell-conditioned collagen

Anchored collagen gels were set up containing either NS cells or HTS cells and cultured over the same time course used for the apoptosis assays. Gels were harvested at days 3 and 7 for collagen extraction or digestion. Extraction of the acid-soluble and -insoluble components was performed as follows: gels were loosened and washed twice (with inversion for 10 min each) with a 10 volume of PBS containing 0.1 M ethylenediaminetetraacetic acid (EDTA) and 10 M phenyl–methylsulfonyl fluoride. Four like gels were pooled for each cell type and drained of wash buffer. Three milliliters of 0.5 M acetic acid was then added and the gels were left at 4°C for 3 d to extract the acid-soluble collagen. The samples were then centrifuged at 14,000 g (30 min at 4°C) and separated into acid-soluble (the supernatant) and acid-insoluble (the pellet) portions. These were stored frozen until further use. Alternatively, for direct digestion with a collagenase preparation that had high collagenase but low tryptic activity, two like gels were washed twice with PBS containing calcium and magnesium, and then placed in 0.5 mg per mL collagenase D in PBS (+Mg+Ca) and incubated at 37°C for 40 min (10 min over that required to completely digest the gel—as judged visually). The digested gel solution was immediately centrifuged (4000 g for 5 min) to remove cellular debris and stored frozen until protein analysis by SDS-PAGE.

Collagen digestion with MMP-2

Acid-soluble collagen preparations were neutralized by adding approximately 2:1 vol:vol of 1 M Tris-HCl (pH 8.8), heat denatured at 60°C for 15 min, and then 20 L of these preps (at approximately 1 mg per mL collagen) were digested in the presence of 1 mM CaCl₂ with 10 L of pre-activated MMP-2 (100 mU per mL—R&D Systems Abingdon, UK) at 37°C for 24 h. The digested collagen was analyzed using SDS-PAGE.

Immunohistochemistry

All tissue samples were processed routinely, embedded in wax, and 4 m sections were taken for staining. Antigen retrieval was performed by microwaving slides in 10 mM Tris EDTA (pH 9.0) buffer at full power (800 W) for 15 min, and incubating for a further 15 min in the hot buffer. The slides were then immediately washed in cold tapwater for 2–3 min and then tTBS (2 drops Tween-20 in 500 mL of Tris-buffered saline (TBS)). All further washes used tTBS. Four drops of streptavidin blocking solution (Streptavidin Biotin Blocking Kit, Vector Labs, Burlingame, California) were added to 1 mL each of normal horse serum blocking reagent (Vector Labs). This was then diluted 1:5 in Chemmate antibody diluent (DakoCytomation, Ltd., Ely, UK) and 100 L was applied to each slide and incubated at RT for 30 min. This was then replaced with primary antibody, anti-transglutaminase antibody (Transglutaminase II Ab-3, Neomarkers, Lab Vision, California) diluted 1:200 in chemmate diluent containing four drops of biotin block (Vector kit) per mL of antibody mix. The slides were then incubated either for 1 h at RT or overnight at 4°C. After washing thoroughly with tTBS, the slides were further incubated for 30 min at RT with a 1:100 dilution of biotinylated horse anti-mouse antibody (Vector Labs) followed by an incubation with a 1:200 dilution of streptavidin alkaline phosphatase (Vector Labs). Antibody binding was visualized using the Vector phosphatase substrate kit (Red), made up as per kit instructions plus one drop of levamisole block for endogenous alkaline phosphatase activity (Vector Labs) per 5 mL of substrate solution. Slides were counterstained with Harris's hematoxylin, dehydrated, cleared, and mounted.

Electrophoresis

Collagenase D-digested fractions were subjected to SDS-PAGE on 10% or 12% gels and under non-reducing conditions. MMP-2-digested collagen was subjected to SDS-PAGE on 7.5% gels under non-reducing conditions. Total protein was visualized using Coomassie Brilliant Blue R250 followed by Silver staining (ProteoSilver² Silver stain kit—Sigma, Gillingham, UK).

Colorimetric analysis of cell number

A colorimetric assay of cell number was used that is based on cellular uptake and staining with crystal violet dye. Briefly, 96-well plates were seeded with cells in minimal medium and the cells were allowed to attach overnight. The following day, the media were changed to the test media (a range of seven different concentrations for each active in triplicate along with a triplicate set of minimal media alone for each active tested) and the media were refreshed twice weekly as necessary. A "time zero" set of triplicates was harvested for assay at this time point to give a starting point reading by which to judge later ones. Replica plates were set up in this way for each time point (4, 7, and 11 d). At each time point, the cells were fixed and stained (0.5% crystal violet, 5% formal saline, 50% ethanol, 0.85% NaCl) for 10 min at RT and washed 3 with PBS. The dye was then eluted using solvent (33% acetic acid) and the optical density (OD) was read at 540 nm using a plate-reading spectrophotometer. The OD was corrected for background, and the data were then plotted as a titration curve.

Manipulation of tissue transglutaminase activity in 3D-collagen gels

Inhibition of endogenous tissue transglutaminase activity was achieved by adding various amounts of the polyamine, putrescine or 1,4-DAB (provided by Procyon Biopharma Inc., Dorval, Canada), to the culture medium. 1,4-DAB is a naturally occurring polyamine that binds to the acyl-enzyme intermediate and specifically inhibits the intended protein substrate from cross-linking, forming an amine adduct instead. Transglutaminase activity was increased in gels by adding exogenous tissue transglutaminase (Sigma) to the culture medium.

References

1. Akasaka, Y, Fujita, K, Ishikawa, Y, et al: Detection of apoptosis in keloids and a comparative study on apoptosis between keloids, hypertrophic scars, normal healed flat scars, and dermatofibroma. Wound Repair Regen 2001 9: 501–506, 10.1046/j.1524-475x.2001.00501.x | Article | PubMed | ISI | ChemPort |
2. Akasaka, Y, Ishikawa, Y, Ono, I, et al: Enhanced expression of caspase-3 in hypertrophic scars and keloid: Induction of caspase-3 and apoptosis in keloid fibroblasts in vitro. Lab Invest 2000 80: 345–357,  | Article | PubMed | ISI | ChemPort |
3. Akimov, SS, Krylov, D, Fleischman, LF, Belkin, AM: Tissue transglutaminase is an integrin-binding adhesion coreceptor for fibronectin. J Cell Biol 2000 148: 825–838, 10.1083/jcb.148.4.825 | Article | PubMed | ISI | ChemPort |
4. Bornstein, MB: Reconstituted rattail collagen used as substrate for tissue cultures on coverslips in Maximow slides and roller tubes. Lab Invest 1958 7: 134–137,  | PubMed | ISI | ChemPort |
5. Bowness, JM, Folk, JE, Timpl, R: Identification of a substrate site for liver transglutaminase on the aminopropeptide of type III collagen. J Biol Chem 1987a 262: 1022–1024,  | PubMed | ISI | ChemPort |
6. Bowness, JM, Henteleff, H, Dolynchuk, KN: Components of increased labelling with putrescine and fucose during healing of skin wounds. Connect Tissue Res 1987b 16: 57–70,  | ISI | ChemPort |
7. Bowness, JM, Tarr, AH, Wong, T: Increased transglutaminase activity during skin wound healing in rats. Biochim Biophys Acta 1988 967: 234–240,  | PubMed | ISI | ChemPort |
8. Brickley-Parsons, D, Glimcher, MJ, Smith, RJ, Albin, R, Adams, JP: Biochemical changes in the collagen of the palmar fascia in patients with Dupuytren's disease. J Bone Joint Surg Am 1981 63: 787–797,  | PubMed | ChemPort |
9. Buckley, CD, Pilling, D, Henriquez, NV: RGD peptides induce apoptosis by direct caspase-3 activation [see comments]. Nature 1999 397: 534–539, 10.1038/17409 | Article | PubMed | ISI | ChemPort |
10. Chodon, T, Sugihara, T, Igawa, HH, Funayama, E, Furukawa, H: Keloid-derived fibroblasts are refractory to Fas-mediated apoptosis and neutralization of autocrine transforming growth factor-beta1 can abrogate this resistance. Am J Pathol 2000 157: 1661–1669,  | PubMed | ISI | ChemPort |
11. Crowe, MJ, Doetschman, T, Greenhalgh, DG: Delayed wound healing in immunodeficient TGF-beta 1 knockout mice. J Invest Dermatol 2000 115: 3–11,  | Article | PubMed | ISI | ChemPort |
12. 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,  | PubMed | ISI | ChemPort |
13. Dolynchuk, KN: Inhibition of tissue transglutaminase and epsilon (gamma–glutamyl) lysine cross-linking in human hypertrophic scar. Wound Repair Regen 1996 4: 16–20, 10.1046/j.1524-475X.1996.40105.x | Article |
14. Dolynchuk, KN, Bendor-Samuel, R, Bowness, JM: Effect of putrescine on tissue transglutaminase activity in wounds: Decreased breaking strength and increased matrix fucoprotein solubility. Plast Reconstr Surg 1994 93: 567–573,  | PubMed | ISI | ChemPort |
15. Dolynchuk, KN, Pettigrew, NM: Transglutaminase levels in Dupuytren's disease. J Hand Surg [Am] 1991 16: 787–790,
16. Dolynchuk, KN, Ziesmann, M, Serletti, JM: Topical putrescine (Fibrostat) in treatment of hypertrophic scars: Phase II study. Plast Reconstr Surg 1996 97: 117–123, (discussion 124–115) | PubMed | ISI | ChemPort |
17. Fluck, J, Querfeld, C, Cremer, A, Niland, S, Krieg, T, Sollberg, S: Normal human primary fibroblasts undergo apoptosis in three-dimensional contractile collagen gels. J Invest Dermatol 1998 110: 153–157, 10.1046/j.1523-1747.1998.00095.x | Article | PubMed | ISI | ChemPort |
18. Frisch, SM, Ruoslahti, E: Integrins and anoikis. Curr Opin Cell Biol 1997 9: 701–706, 10.1016/S0955-0674(97)80124-X | Article | PubMed | ISI | ChemPort |
19. Fukamizu, H, Grinnell, F: Spatial organization of extracellular matrix and fibroblast activity: Effects of serum, transforming growth factor beta, and fibronectin. Exp Cell Res 1990 190: 276–282, 10.1016/0014-4827(90)90197-I | Article | PubMed | ISI | ChemPort |
20. Gaudry, CA, Verderio, E, Jones, RA, Smith, C, Griffin, M: Tissue transglutaminase is an important player at the surface of human endothelial cells: Evidence for its externalization and its colocalization with the beta(1) integrin. Exp Cell Res 1999 252: 104–113, 10.1006/excr.1999.4633 | Article | PubMed | ISI | ChemPort |
21. Grenard, P, Bresson-Hadni, S, El Alaoui, S, Chevallier, M, Vuitton, DA, Ricard-Blum, S: Transglutaminase-mediated cross-linking is involved in the stabilization of extracellular matrix in human liver fibrosis. J Hepatol 2001 35: 367–375, 10.1016/S0168-8278(01)00135-0 | Article | PubMed | ISI | ChemPort |
22. Griffin, M, Casadio, R, Bergamini, CM: Transglutaminases: Nature's biological glues. Biochem J 2002 368: 377–396, 10.1042/BJ20021234 | Article | PubMed | ISI | ChemPort |
23. Grinnell, F, Zhu, M, Carlson, MA, Abrams, JM: Release of mechanical tension triggers apoptosis of human fibroblasts in a model of regressing granulation tissue. Exp Cell Res 1999 248: 608–619, 10.1006/excr.1999.4440 | Article | PubMed | ISI | ChemPort |
24. Haroon, ZA, Hettasch, JM, Lai, TS, Dewhirst, MW, Greenberg, CS: Tissue transglutaminase is expressed, active, and directly involved in rat dermal wound healing and angiogenesis. Faseb J 1999 13: 1787–1795,  | PubMed | ISI | ChemPort |
25. Hirota, A, Ebihara, T, Kusubata, M, et al: Collagen of chronically inflamed skin is over-modified and upregulates secretion of matrix metalloproteinase 2 and matrix-degrading enzymes by endothelial cells and fibroblasts. J Invest Dermatol 2003 121: 1317–1325, 10.1111/j.1523-1747.2003.12637.x | Article | PubMed | ISI | ChemPort |
26. Kono, T, Tanii, T, Furukawa, M, et al: Cell cycle analysis of human dermal fibroblasts cultured on or in hydrated type I collagen lattices. Arch Dermatol Res 1990 282: 258–262, 10.1007/BF00371646 | Article | PubMed | ISI | ChemPort |
27. Mian, S, el Alaoui, S, Lawry, J, Gentile, V, Davies, PJ, Griffin, M: The importance of the GTP-binding protein tissue transglutaminase in the regulation of cell cycle progression. FEBS Lett 1995 370: 27–31, 10.1016/0014-5793(95)00782-5 | Article | PubMed | ISI | ChemPort |
28. Mishra, S, Murphy, LJ: Tissue transglutaminase has intrinsic kinase activity: Identification of transglutaminase 2 as an insulin-like growth factor-binding protein-3 kinase. J Biol Chem 2004 279: 23863–23868, 10.1074/jbc.M311919200 | Article | PubMed | ISI | ChemPort |
29. Moulin, V, Larochelle, S, Langlois, C, Thibault, I, Lopez-Valle, CA, Roy, M: Normal skin wound and hypertrophic scar myofibroblasts have differential responses to apoptotic inductors. J Cell Physiol 2004 198: 350–358, 10.1002/jcp.10415 | Article | PubMed | ISI | ChemPort |
30. Murthy, SN, Wilson, J, Guy, SL, Lorand, L: Intramolecular crosslinking of monomeric fibrinogen by tissue transglutaminase. Proc Natl Acad Sci USA 1991 88: 10601–10604,  | PubMed | ChemPort |
31. Nakagawa, S, Pawelek, P, Grinnell, F: Extracellular matrix organization modulates fibroblast growth and growth factor responsiveness. Exp Cell Res 1989a 182: 572–582, 10.1016/0014-4827(89)90260-7 | Article | ISI | ChemPort |
32. Nakagawa, S, Pawelek, P, Grinnell, F: Long-term culture of fibroblasts in contracted collagen gels: Effects on cell growth and biosynthetic activity. J Invest Dermatol 1989b 93: 792–798, 10.1111/1523-1747.ep12284425 | Article | PubMed | ISI | ChemPort |
33. Nakaoka, H, Perez, DM, Baek, KJ, et al: Gh: A GTP-binding protein with transglutaminase activity and receptor signaling function. Science 1994 264: 1593–1596,  | PubMed | ISI | ChemPort |
34. Nishiyama, T, Tsunenaga, M, Nakayama, Y, Adachi, E, Hayashi, T: Growth rate of human fibroblasts is repressed by the culture within reconstituted collagen matrix but not by the culture on the matrix. Matrix 1989 9: 193–199,  | PubMed | ISI | ChemPort |
35. Nunes, I, Gleizes, PE, Metz, CN, Rifkin, DB: Latent transforming growth factor-beta binding protein domains involved in activation and transglutaminase-dependent cross-linking of latent transforming growth factor-beta. J Cell Biol 1997 136: 1151–1163, 10.1083/jcb.136.5.1151 | Article | PubMed | ISI | ChemPort |
36. Nusgens, B, Merrill, C, Lapiere, C, Bell, E: Collagen biosynthesis by cells in a tissue equivalent matrix in vitro. Coll Relat Res 1984 4: 351–363,  | PubMed | ISI | ChemPort |
37. Paye, M, Nusgens, BV, Lapiere, CM: Modulation of cellular biosynthetic activity in the retracting collagen lattice. Eur J Cell Biol 1987 45: 44–50,  | PubMed | ISI | ChemPort |
38. Ricard-Blum, S, Bresson-Hadni, S, Guerret, S, et al: Mechanism of collagen network stabilization in human irreversible granulomatous liver fibrosis. Gastroenterology 1996 111: 172–182,  | PubMed | ChemPort |
39. Ricard-Blum, S, Esterre, P, Grimaud, JA: Collagen cross-linking by pyridinoline occurs in non-reversible skin fibrosis. Cell Mol Biol (Noisy-le-grand) 1993 39: 723–727,  | PubMed | ChemPort |
40. Ricard-Blum, S, Hartmann, DJ, Esterre, P: Monitoring of extracellular matrix metabolism and cross-linking in tissue, serum and urine of patients with chromoblastomycosis, a chronic skin fibrosis. Eur J Clin Invest 1998 28: 748–754, 10.1046/j.1365-2362.1998.00335.x | Article | PubMed | ISI | ChemPort |
41. Schuster, N, Krieglstein, K: Mechanisms of TGF-beta-mediated apoptosis. Cell Tissue Res 2002 307: 1–14, 10.1007/s00441-001-0479-6 | Article | PubMed | ISI | ChemPort |
42. Stephens, P, Grenard, P, Aeschlimann, P, et al: Crosslinking and G-protein functions of transglutaminase 2 contribute differentially to fibroblast wound healing responses. J Cell Sci 2004 117: 3389–3403, 10.1242/jcs.01188 | Article | PubMed | ISI | ChemPort |
43. Teofoli, P, Barduagni, S, Ribuffo, M, Campanella, A, De Pita, O, Puddu, P: Expression of Bcl-2, p53, c-jun and c-fos protooncogenes in keloids and hypertrophic scars. J Dermatol Sci 1999 22: 31–37, 10.1016/S0923-1811(99)00040-7 | Article | PubMed | ISI | ChemPort |
44. Unemori, EN, Werb, Z: Reorganization of polymerized actin: A possible trigger for induction of procollagenase in fibroblasts cultured in and on collagen gels. J Cell Biol 1986 103: 1021–1031, 10.1083/jcb.103.3.1021 | Article | PubMed | ISI | ChemPort |
45. Verderio, E, Gaudry, C, Gross, S, Smith, C, Downes, S, Griffin, M: Regulation of cell surface tissue transglutaminase: Effects on matrix storage of latent transforming growth factor-beta binding protein-1. J Histochem Cytochem 1999 47: 1417–1432,  | PubMed | ISI | ChemPort |
46. Verderio, EA, Telci, D, Okoye, A, Melino, G, Griffin, M: A novel RGD-independent cel adhesion pathway mediated by fibronectin-bound tissue transglutaminase rescues cells from anoikis. J Biol Chem 2003 278: 42604–42614, 10.1074/jbc.M303303200 | Article | PubMed | ISI | ChemPort |
47. 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, 10.1006/jsre.1998.5267 | Article | PubMed | ISI | ChemPort |
48. Weber, L, Meigel, WN, Spier, W: Collagen polymorphism in pathologic human scars. Arch Dermatol Res 1978 261: 63–71, 10.1007/BF00455376 | Article | PubMed | ISI | ChemPort |

Acknowledgments

This work was funded by The Restoration of Appearance and Function Trust (registered charity No. 299811), The Garfield Weston Foundation, The Alan & Babette Sainsbury Charitable Fund, The Childwick Trust, and The Alan Gaynor Memorial Fellowship. The authors are grateful to Professor Roy Sanders for his support and advice, to Dr Julian Dye for discussions, and to Procyon Biopharma for producing and supplying 1,4-DAB.