Recessive dystrophic epidermolysis bullosa. Evidence for increased collagenase as a genetic characteristic in cell culture.

Fibroblast cultures from patients with recessive dystrophic epidermolysis bullosa (RDEB) demonstrated an increased capacity to synthesize and secrete collagenase. This phenotypic trait appeared to distinguish RDEB from other genetically distinct forms of epidermolysis bullosa. The finding of increased collagenase may be a specific manifestation of these cells in that prototypic lysosomal and cytoplasmic enzymes were present in approximately normal concentrations. In addition, this trait persisted through many cell passages, suggesting that the property was genetically determined. The elevated concentrations of immunoreactive collagenase in fibroblast cultures of patients with RDEB reflected those previously observed in vivo (4) and support the concept of a pathogenetic role for the enzyme in the blistering phenomenon. In three of the cell lines, the increase in enzyme protein occurred in association with a structurally defective enzyme. The data suggest that this may be a characteristic of all RDEB cells.

Scleroderma: increased biosynthesis of triple-helical type I and type III procollagens associated with unaltered expression of collagenase by skin fibroblasts in culture.

To assess potential abnormalities in collagen metabolism in systemic scleroderma, skin fibroblast lines from patients with this disease were established and compared to control cell lines derived from healthy subjects. For studies on the biosynthesis of procollagen, the cells were incubated with [(14)C]proline in a medium supplemented with ascorbic acid and beta-aminopropionitrile, and the synthesis of nondialyzable [(14)C]hydroxyproline, in relation to DNA or cell protein, was taken as an index of procollagen formation. Five of eight scleroderma fibroblast cell lines demonstrated procollagen biosynthesis rates significantly higher than the controls, and the mean rate of procollagen synthesis by scleroderma fibroblasts was about twice that of the control cells. Control experiments demonstrated that the specific activity of the intracellular free proline was not different in scleroderma and control fibroblasts, and the mean population doubling times of the scleroderma and the control fibroblast cell lines were the same. The relative synthesis of the genetically distinct procollagens was examined by isolating type I and type III procollagens from the cell culture medium using DEAE-cellulose chromatography. The ratios of type I/III procollagens in scleroderma cell lines did not differ from the controls. The helical stability of the collagenous portion of type I and type III procollagens, estimated by the resistance of (14)C-collagen to limited proteolytic digestion with pepsin under nondenaturing conditions, was the same in both scleroderma and control cultures. The capacity of the cells to synthesize enzymatically active and immunologically reacting collagenase was also studied; no marked differences in these parameters could be observed. The results suggest that cultured skin fibroblasts from patients with scleroderma demonstrate a metabolic abnormality expressed as increased synthesis of type I and type III procollagens in a normal ratio. This abnormality may play a role in the excessive accumulation of collagen in the skin and other organs affected in scleroderma.

Studies on purified rheumatoid synovial collagenase in vitro and in vivo.

Rheumatoid synovial collagenase obtained from culture medium can be separated by Sephadex gel filtration into two peaks of enzyme activity. These have been designated as fast-moving and slow-moving rheumatoid synovial collagenases on the basis of their electrophoretic mobility on polyacrylamide gels. The slow-moving rheumatoid synovial collagenase has been highly purified by affinity chromatography on collagen conjugated to Sepharose and used to prepare a monospecific anti-synovial collagenase antiserum. The antiserum against rheumatoid synovial collagenase has permitted the demonstration of immunoreactive collagenase in extracts of rheumatoid synovial tissue that have no detectable enzymatic activity. Collagenase has also been detected immunologically in enzymatically inactive culture medium from the first 24 hr of culture. Recovery of collagenase activity appears to be related to the chromatographic separation of the enzyme from serum antiproteases. The demonstration of collagenase in vivo in rheumatoid synovium adds further support for the concept that the enzyme is present in tissue at levels that are of significance in the pathogenesis of rheumatoid arthritis. In addition, rheumatoid synovial collagenase and human skin collagenase show complete immunologic identity when reacted with monospecific antiserum prepared against either of these purified enzymes, indicating that organ specificity between these two human collagenases is unlikely.

Neutral metalloproteinases produced by human mononuclear phagocytes. Enzyme profile, regulation, and expression during cellular development.

Mononuclear phagocytes are developmentally and functionally complex cells that play critical roles in extracellular matrix remodeling. We hypothesized that differentiated mononuclear phagocytes, typified by alveolar macrophages, use a spectrum of metalloproteinases to degrade various matrix macromolecules. To test this hypothesis, we have evaluated synthesis and secretion of four metalloproteinases (interstitial collagenase, stromelysin, 72-kD type IV collagenase, and 92-kD type IV collagenase) by human mononuclear phagocytes with regard to (a) the effect of cellular differentiation, (b) regulation of secretion, and (c) comparisons/contrasts with a prototype metalloproteinase-secretory cell, the human fibroblast. We found that regulated secretion of greater quantities and a wider spectrum of metalloenzymes correlated with a more differentiated cellular phenotype. As extreme examples, the 92-kD type IV collagenase was released by peripheral blood monocytes and uninduced U937 monocyte-like cells, whereas stromelysin was secreted only by lipopolysaccharide-stimulated alveolar macrophages. Macrophage production of interstitial collagenase, stromelysin, and 72-kD type IV collagenase was approximately 20%, 10%, and 1-2%, respectively, of that from equal numbers of fibroblasts; secretion of the 92-kD type IV collagenase was not shared by fibroblasts. This work confirms the potential of macrophages to directly degrade extracellular matrix via secreted metalloproteinases in a manner that differs both qualitatively and quantitatively from that of fibroblasts. Moreover, varying regulation of metalloenzyme synthesis, evidenced by distinct patterns of basal and stimulated secretion during differentiation, can be studied at a molecular level in this model system.

Enhanced collagenase production by fibroblasts derived from human basal cell carcinomas.

Fibroblast cultures derived from human basal cell carcinomas demonstrated an increased capacity to synthesize and secrete collagenase. Although the levels of collagenase were up to 8-fold greater than those of normal control cell lines, this phenotypic trait was not permanent and was expressed only for a few passages following primary explanation. The basal cell carcinoma fibroblast collagenase was secreted as a proenzyme. The kinetics of activation and the catalytic efficiency of the basal cell carcinoma fibroblast enzyme were equal to control collagenase, indicating that increased activity was due to increased synthesis of enzyme protein. Increased synthesis of collagenase was not due either to altered cell growth or to an overall increase in protein synthesis. Furthermore, synthesis of another major protein, of another major protein, collagen, was not enhanced. The data suggest that the tumors may have stimulated adjacent fibroblasts to produce more collagenase which is of importance in tumor invasion.

Comparison of cleavage site specificity of gelatinases A and B using collagenous peptides.

The gelatinases (type IV collagenases) are members of the matrix metalloproteinase family that not only have a high degree of structural homology but are known to be nearly identical in their digestion profile against macromolecular substrates. We have shown previously that the preferred cleavage sites in the hydrolysis of type I gelatin, catalyzed by gelatinase A (72 kDa type IV collagenase), are bracketed by hydroxyproline in the P5 and P5' positions. In this report, a kinetic investigation using a series of collagenous dodecylpeptides in which the P5 and P5' hydroxyprolines were systematically varied and used as substrates for recombinant human gelatinase A, we show that replacement with either proline or alanine always resulted in increased Km. In contrast, substitution of the hydroxylated amino acids tyrosine and serine at P5 and P5' reduced the Km significantly, indicating that the hydroxyl moiety of the hydroxyproline is the functional group responsible for favorable enzyme-substrate affinity. This was shown by the kcat/Km ratio, which was doubled by the substitution of serine in that site. Cleavage of the same series of dodecylpeptides by recombinant human gelatinase B (92 kDa type IV collagenase) showed a very different kinetic profile for which no patterns were discernible. In subsequent comparisons of the two enzymes, it was found that gelatinase B cleaved the thiopeptolide substrate AcProLeuGly-S-LeuGly-OC2H5 at double the velocity of gelatinase A. In contrast, gelatinase A digested type I gelatin about 2.5-times faster than gelatinase B. SDS-PAGE analysis of gelatin cleavage products showed different patterns of product peptides for each enzyme. Further comparisons of the proteinases using synthetic peptide substrates with variations in size and in substituents at the P2' site again showed marked kinetic differences. Although these two matrix metalloproteinases seem similar in that they are both gelatinolytic and can degrade a nearly identical battery of macromolecular matrix components including type IV collagen, it is clear from these results that they are very different enzymatically. Since the regulatory portions of gelatinases A and B differ markedly, it has been assumed that the enzymes serve the same function, but respond to different stimuli. The differences in substrate specificity described herein suggest that their proposed physiological roles may require reevaluation.

Evidence for mammalian collagenases as zinc ion metalloenzymes.

Collagenases (EC 3.4.24.3) from human skin, rat skin and rat uterus were inhibited by the chelating agents EDTA, 1,10-phenanthroline and tetraethylene pentamine in the presence of excess Ca2+, suggesting that a second metal ion participates in the activity of the enzyme. Collagenase inhibition by 1,10-phenanthroline could be both prevented and reversed by a number of transition metal ions, specifically Zn2+, Co2+, Fe2+ and Cu2+. However, Zn2+ is effective in five-fold lower molar concentrations (1-10(-4) M) than the other ions. Furthermore, Zn2+ was the only ion tested able to prevent and reverse the inhibition of collagenase by EDTA in the presence of excess Ca2+. Atomic absorption analysis of purified collagenase for Zn2+ showed that Zn2+ was present in the enzyme preparations, and that the metal co-purifies with collagenase during column chromatography.

Protein tyrosine phosphorylation in signalling pathways leading to the activation of gelatinase A: activation of gelatinase A by treatment with the protein tyrosine phosphatase inhibitor sodium orthovanadate.

Fibroblasts in monolayer culture secrete gelatinase A (MMP2; 72 kDa type IV collagenase) only in its proenzyme form. Unlike other secreted matrix metalloproteinases, progelatinase A is refractory to activation by serine proteinases. Disparate agents, including monensin, cytochalasin D, and concanavalin A, have been found to mediate the activation of gelatinase A zymogen secreted by fibroblast monolayers. Our finding that monensin-mediated activation can be reversed by the protein tyrosine kinase inhibitor genistein (Li et al., Experimental Cell Research 232 (1997) 332) prompted us to investigate the effect of the specific inhibitor of protein tyrosine phosphatases, sodium orthovanadate, on progelatinase A activation. Treatment of fibroblast monolayers with orthovanadate also results in the secretion of activated gelatinase A. This activation is dose- and time-dependent, requires protein synthesis, and is associated with cell membranes. Vanadate-mediated activation does not occur in the presence of herbimycin A, a protein tyrosine kinase inhibitor. As with progelatinase activation mediated by monensin, concanavalin A, and cytochalasin D, orthovanadate treatment results in increased synthesis of the membrane proteinase MT1-MMP, that can catalyze the activation of progelatinase A. Protein tyrosine kinase inhibitors are able to prevent the increase of MT1-MMP mRNA, as shown by Northern blot and RT-PCR. In addition, orthovanadate potentiates the effects of monensin and concanavalin A. While treatment with monensin or concanavalin A result only in an increase of the putative activator MT1-MMP, orthovanadate also reduces the production of the specific inhibitor TIMP-2. These experiments implicate protein tyrosine phosphorylation in the signal transduction pathways which lead to the activation of progelatinase A.

Hormonal interactions in mammalian collagenase regulation. Comparative studies in human skin and rat uterus.

The production of collagenase by human skin explants in culture is prevented by 10(-8) M dexamethasone, 5 . 10(-4) M dibutyryl cyclic AMP, or 2.5 . 10(-3) M theophylline. Decreases in collagenase activity are paralleled by reductions in the degradation of explant collagen during the culture period. Progesterone, which effectively inhibits collagenase production in rat uterine explant cultures, has no effect on human skin explants. The inhibition by cyclic AMP is nucleotide specific. When partially inhibitory concentrations of dexamethasone and dibutyryl cyclic AMP, or dexamethasone and theophylline, are added to culture medium together, the resultant inhibition is that predicted by additivity. Synergistic inhibition, as observed in rat uterus between progesterone and dibutyryl cyclic AMP, fails to occur. Dexamethasone inhibits the production of collagenase by cultured explants of rat uterus, with complete inhibition occurring at 10(-7) M steroid. Synergism between glucocorticoids and dibutyryl cyclic AMP or between dexamethasone and progesterone could not be demonstrated in the uterine culture system. These results suggest the existence of three regulatory systems for the control of collagenase production in mammalian tissues, and that cooperativity between systems may occur on a tissue-specific basis.

Glucocorticoid modulation of collagenase expression in human skin fibroblast cultures. Evidence for pre-translational inhibition.

Glucocorticoids inhibit collagenase accumulation in the medium of human skin explant cultures. To examine the mechanism for this process, skin fibroblasts were placed in serum-free medium containing various steroids. Dexamethasone produced a dose-dependent inhibition of trypsin-activatable collagenase in the culture medium with maximal inhibition of approx. 85% at 10(-6) M. Dexamethasone failed to inhibit collagenase activity directly. The decrease in activity in the medium was paralleled by a decrease in immunoreactive protein, suggesting inhibition of enzyme synthesis. The specificity of the effect was shown in two ways. At 10(-6) M steroid, only dexamethasone and hydrocortisone were inhibitory; estradiol, progesterone and testosterone produced less than 10% inhibition. In biosynthetic studies, exposure to 10(-7) M dexamethasone for 24 h produced approx. 50% inhibition of collagenase synthesis but caused no greater than 10% inhibition of total protein synthesis. The T1/2 for achieving the effect was approx. 16 h after initial exposure to dexamethasone. These kinetics were parallel to the inhibition caused by actinomycin D and cordycepin, two inhibitors of transcription, but were longer than that caused by cycloheximide (T 1/2 less than 3 h). To examine this process, cells were cultured in the presence or absence of 10(-6) M dexamethasone prior to harvesting mRNA for cell-free translation. In each case the inhibition or enzyme activity in the intact cells was paralleled by a reduction in translatable collagenase mRNA from the same cells. At the same time, there was no significant inhibition of total protein translation by the steroid. These data suggest that glucocorticoids regulate collagenase synthesis at a pre-translational level, possibly through inhibition of transcription.

Leukocyte and lymphocyte cyclic AMP responses in atopic eczema.

Lymphocytes from subjects with mild and severe atopic eczema were compared with normal control subjects in regard to their cAMP (3';5'-cyclic adenosine monophosphate) responses to a variety of stimulatory agents. Individuals in the severe eczema group were shown to have a significant diminution in their unstimulated lymphocyte cAMP levels and absolute cAMP responses to 0.5 mM theophylline, 0.5 mM theophyline + 1 micronM epinephrine, 10 mM isoproternol, 1 mM isoproterenol, 10 mM salbutamol, and 3 micron M PGE1. Individuals with mild eczema had a reduced response to 0.5 mM theophylline. The severe eczema groups also differed in a number of these responses from a group of 5 subjects with severe psoriasis. Mixed leukocyte cAMP responses to 10 mM isoproterenol also were examined and found to be diminished in individuals with eczema.

Human skin fibroblasts in culture: procollagen synthesis in the presence of sera from normal human subjects and from patients with dermal fibroses.

Various dermal fibrotic conditions, such as progressive systemic sclerosis, localized morphea and familial cutaneous collagenoma, are characterized by excessive deposition of collagen in the skin. In the present study, we examined the possibility that a circulating serum factor(s) is responsible for increased collagen production in these diseases. The effects of human serum on the synthesis of procollagen were examined by incubating normal human dermal fibroblasts with [3H]proline and varying concentrations of dialyzed heat-inactivated serum. The synthesis of procollagen was measured as formation of nondialyzable [3H]hydroxyproline and collagenase-digestible 3H]polypeptides. In the absence of serum little procollagen was formed but the synthesis was markedly stimulated by the addition of normal serum in a concentration-dependent manner. THe ratio of genetically distinct 3H-procollagens of type I and type III, assayed by DEAE-cellulose chromatography and SDS-polyacrylamide gel electrophoresis after limited pepsin proteolysis, was unaffected by the addition of serum. Thus, normal human serum contains a nondialyzable factor(s) which stimulates the synthesis of procollagens type I and type III equally. Sera from 5 patients with progressive systemic sclerosis, 3 with localized scleroderma, and 2 with familial cutaneous collagenoma were also tested. Sera from these patients failed to stimulate 3H-procollagen production more than sera from healthy age-matched controls. Therefore, no increased quantities or qualitatively aberrant factors were shown to be present in the sera of these patients.

Lipoid proteinosis: in vivo and in vitro evidence for a lysosomal storage disease.

Tissue and cultured fibroblasts derived from one patient with the classical findings of lipoid proteinosis have been used to examine pathologic mechanisms in the disease. Ultrastructural examination of the skin revealed not only extracellular deposits of finely granular, moderately electron dense material, but in addition the dermal fibroblasts characteristically demonstrated marked cytoplasmic vacuolization. Phase contrast microscopy of the cultured skin fibroblasts also showed strikingly abnormal cells with many inclusions, which by electron microscopy were delimited by a single membrane. Membranous lamellar material was also increased in these cells. Biochemical analysis of the fibroblasts revealed a 3- to 4-fold elevation in intracellular hexuronic acid. These morphologic and biochemical findings suggest certain similarities with known storage diseases and support the postulate that lipoid proteinosis may represent a lysosomal storage disease.

Regulation of gelatinase in human skin organ cultures by glucocorticoids.

Hydrocortisone and dexamethasone prevent the appearance of gelatinase in serum-free explant cultures of normal human skin. Hydrocortisone inhibits maximally at 10(-6) M and dexamethasone at 10(-8) M in culture medium. Glucocorticoids at these concentrations do not cause a generalized decrease in protein synthesis; thus the effect on gelatinase shows specificity. The reduction in gelatinase activity caused by dexamethasone can be overcome in the presence of dexamethasone 21-mesylate, a glucocorticoid antagonist that binds irreversibly to the cytoplasmic steroid receptor. These data suggest that the enzymes of collagen degradation, collagenase and gelatinase, may be coregulated.

Retinoic acid inhibition of collagenase and gelatinase expression in human skin fibroblast cultures. Evidence for a dual mechanism.

Human skin fibroblast cultures have been employed to study the effects of a variety of vitamin A analogues (retinoids) on the expression of two enzymes involved in collagen degradation in the skin, collagenase and a gelatinolytic protease. In normal and recessive dystrophic epidermolysis bullosa fibroblast cultures, retinoic acid compounds were effective inhibitors of the accumulation of both enzymes in the culture medium with half-maximal inhibitions occurring at 0.25-1 microM for collagenase and at 3-6 microM for the gelatinolytic protease. Various retinoids exhibited differing degrees of inhibitory actions, so that at a 1 microM concentration, relative inhibitions were: 13-cis-retinoic acid greater than all-trans-retinoic acid greater than aromatic retinoid (Ro 10-9359) much greater than retinol. The retinoic acid-mediated decrease in collagenase activity was accompanied by a parallel decrease in immunoreactive collagenase protein, suggesting that the retinoic acids were acting to inhibit synthesis of the enzyme. However, an additional effect of these agents was encountered. Although the retinoids themselves had no direct collagenase inhibitory action, medium derived from cultures maintained in these retinoids showed direct inhibitory capacity which was dependent both on the concentration of retinoic acid and on the length of time in culture. The results suggest that the retinoic acids modulate collagenase in vitro by two mechanisms: by decreasing the synthesis of enzyme protein and by modulating the expression of an inhibitory molecule.

Growth stimulation of human keratinocytes by tissue inhibitor of metalloproteinases.

Human recombinant tissue inhibitor of metalloproteinases (rTIMP) at 0.2-4.6 microM was found to stimulate the growth of normal human keratinocytes, in primary cultures on a plastic support, and to markedly increase their growth on a tridimensional culture system, the skin equivalent, as shown by histology, DNA measurements, and planimetry. In contrast, rTIMP had no effect on the growth of normal human fibroblasts. The growth of keratinocytes on extracellular matrix components produced by keratinocytes cultured in the presence or absence of rTIMP was similar, suggesting that rTIMP does not stimulate keratinocyte growth by modifying either the quantity or the composition of the extracellular matrix deposited. rTIMP was labeled with 125iodine in order to study its interaction with keratinocytes in culture. Binding of (125I) rTIMP to keratinocytes was found to be temperature and time dependent. Under steady-state conditions at 22 degrees C, one class of specific rTIMP binding sites was identified with KD of 8.7 nM and 135,000 sites/cell. Such findings are in keeping with the known potentiating effect of TIMP on erythroid precursors, and indicate that this protein has at least two distinct activities.

Stimulation of skin fibroblast collagenase production by a cytokine derived from basal cell carcinomas.

Our previous studies of human basal cell carcinomas (BCC) revealed increased skin collagenase in vivo. Immunocytochemically the collagenase was localized to adjacent stroma, not to the tumor cells. When grown in culture, skin fibroblasts derived from tumor stroma showed a 3- to 4-fold increase in collagenase for the first 10-14 mean population doublings, after which collagenase expression reverted to control levels. These studies suggested that tumors stimulated adjacent fibroblasts to produce more collagenase. In the present study we sought direct evidence for epithelial-stromal interaction in this neoplasm. Under dissecting microscopy tumor islands were freed of stroma, homogenized, sonicated, and centrifuged to remove insoluble tissue. Tumor extracts were incubated with monolayer cultures of normal human skin fibroblasts to assess their effect on collagenase synthesis in these target cells. Culturing the fibroblasts for 24 h in the presence of individual BCC extracts resulted in a 1.6- to 3-fold increase in trypsin-activatable collagenase in the culture medium. This was paralleled by an equal increase in immunoreactive protein, suggesting enhanced enzyme synthesis. There was no change in the activity per immunoreactive protein, indicating a catalytically unaltered enzyme. Gel filtration of pooled BCC extracts showed that the stimulatory activity was contained in eluent fractions of Mr approximately 19Kd. These data suggest that BCCs elaborate a macromolecular cytokine that induces collagenase synthesis in skin fibroblasts and emphasize the importance of epithelial-stromal interactions in cutaneous tumor invasion.

Matrix metalloproteinases in blood vessel development in human fetal skin and in cutaneous tumors.

In vitro angiogenesis models suggest that new blood vessel formation requires the induction and secretion by endothelial cells of matrix metalloproteinases. These enzymes assist in the controlled proteolytic degradation of the surrounding extracellular matrix during blood vessel formation. The results of in vitro studies cannot be extrapolated directly to the process of in vivo angiogenesis because the type of matrix employed and the repertoire of enzymes secreted by cells in vivo differ dramatically from in vivo conditions. To investigate the in vivo role of matrix metalloproteinases in blood vessel development, we looked for the presence of these proteinases in endothelial cells involved in fetal angiogenesis and in neovascularization of certain invasive skin tumors using immunofluorescent staining. In fetal tissue, interstitial collagenase was present in both early microvessels developing from undifferentiated mesoderm and in microvessels involved in elongation and sprout formation from preexisting blood vessels. In aggressive skin tumors, i.e., morpheaform and recurrent basal cell carcinomas and squamous cell carcinomas, there was a marked increase in the number of collagenase-containing blood vessels, often extending into the tumor nests. Immunofluorescent staining failed to detect stromelysin, matrilysin, or gelatinase A and B (72- and 92-kDa type IV collagenases, respectively) in fetal or tumor blood vessels. These findings are consistent with the hypothesis that proteolytic degradation of the extracellular matrix is required for the formation of new blood vessels. Interstitial collagenase appears to play an important role in this process.

Matrilysin (PUMP) correlates with dermal invasion during appendageal development and cutaneous neoplasia.

Matrix-degrading metalloproteinases play a major role in tissue remodeling. Recent studies have shown that enzymes of this class are constitutively expressed primarily by stromal cells and not by epithelium. Here we present immunohistochemical evidence that matrilysin is localized within epidermal cells in developing skin and in tumor cells of cutaneous malignancies. The expression of matrilysin protein in developing fetal skin (6-15 weeks) is localized primarily to the germinative basal cell layer of fetal epidermis and early appendageal buds. The buds continue to express matrilysin during mesenchymal invasion. As development progresses (15-19 weeks) matrilysin is concentrated only in cells at the distal portion of the invading follicular and sweat gland appendageal cords. In adult skin, matrilysin was localized specifically to the outer root sheath of the hair follicles and the secretory cells of the eccrine glands but was absent in the epidermis. Nodulocystic, keratotic, adenoid basal cell carcinomas (BCCs) did not express matrilysin. In contrast, in the more aggressive morpheaform (infiltrative) BCCs and recurrent BCCs, matrilysin was localized at the tumor-stromal interface. In squamous cell carcinomas matrilysin was present in tumor cells at the stromal interface surrounding the tumor nests. The demonstration of matrilysin protein in germinal basal cells during fetal skin development and its presence in tumor cells at the stromal junction suggests that this enzyme may contribute to the proteolytic activity associated with cell-extracellular matrix interactions during appendageal development and tumor invasion.

Basement membrane zone remodeling during appendageal development in human fetal skin. The absence of type VII collagen is associated with gelatinase-A (MMP2) activity.

Epithelial cell adhesion, migration, and differentiation are controlled by interactions at the basement membrane zone (BMZ). Type VII collagen is the major collagenous component of anchoring fibrils that are essential for the attachment of the epidermis to the dermis. Gelatinase A (MMP-2) is believed to be necessary for the degradation of type VII collagen. In this study we have examined the in vivo distribution of type VII collagen and gelatinase A (Gel A) in the developing human epidermis and its appendages. At 13-15 wk of gestation a marked decrease in type VII collagen immunoreactivity was seen in the BMZ surrounding invading appendageal buds; however, type VII collagen mRNA was strongly expressed in the budding epidermal keratinocytes adjacent to the BMZ. At these stages, Gel A-positive mesenchymal-like cells were found scattered throughout the stroma with numerous Gel A-containing cells in direct contact with the developing appendageal buds. In situ zymography was used to show Gel A-activity in vivo. Gel A-mediated lysis was present at the interface between the appendageal buds and the underlying BMZ. By 20-25 wk of gestational age, immunostaining for type VII collagen protein was absent from the BMZ surrounding the distal portion of invading appendageal epithelial cords of both hair follicles and sweat glands. In contrast, type VII collagen mRNA was present in the basal keratinocytes adjacent to the BMZ surrounding the distal portion of these invading appendageal epithelial cords. At these stages Gel A-positive cells were present in the stroma directly adjacent to the distal portion of developing appendageal cords that lacked type VII collagen. In situ zymography showed zones of Gel A-mediated stromal lysis at the distal portion of developing appendageal cords. Interestingly, no differences were seen in the distribution of type IV collagen in the BMZ of both budding and resting fetal epidermis. These observations suggest that the absence of type VII collagen protein correlates directly with the presence of Gel A-activity at the BMZ. Gel A appears to play a major role in appendageal development and contributes to remodeling of the BMZ during fetal skin morphogenesis.