Search

Human skin is exposed daily to solar ultraviolet (UV) radiation. UV induces the matrix metalloproteinases collagenase, 92-kD gelatinase, and stromelysin, which degrade skin connective tissue and may contribute to premature skin aging (photoaging). Pretreatment of skin with all-trans retinoic acid (tRA) inhibits UV induction of matrix metalloproteinases. We investigated upstream signal transduction pathways and the mechanism of tRA inhibition of UV induction of matrix metalloproteinases in human skin in vivo. Exposure of human skin in vivo to low doses of UV activated EGF receptors, the GTP-binding regulatory protein p21Ras, and stimulated mitogen-activated protein (MAP) kinases, extracellular signal-regulated kinase (ERK), c-Jun amino-terminal kinase (JNK), and p38. Both JNK and p38 phosphorylated, and thereby activated transcription factors c-Jun and activating transcription factor 2 (ATF-2), which bound to the c-Jun promoter and upregulated c-Jun gene expression. Elevated c-Jun, in association with constitutively expressed c-Fos, formed increased levels of transcription factor activator protein (AP) 1, which is required for transcription of matrix metalloproteinases. Pretreatment of human skin with tRA inhibited UV induction of c-Jun protein and, consequently, AP-1. c-Jun protein inhibition occurred via a posttranscriptional mechanism, since tRA did not inhibit UV induction of c-Jun mRNA. These data demonstrate, for the first time, activation of MAP kinase pathways in humans in vivo, and reveal a novel posttranscriptional mechanism by which tRA antagonizes UV activation of AP-1 by inhibiting c-Jun protein induction. Inhibition of c-Jun induction likely contributes to the previously reported prevention by tRA of UV induction of AP-1-regulated matrix-degrading metalloproteinases in human skin.

Summary Cutaneous atrophy arising from prolonged use of potent topical corticosteroids has long been a concern. Thus, it would be advantageous to find an agent which protects against atrophy produced by corticosteroids but at the same time does not impair their anti-inflammatory effects. Recent work shows that topical all-trans retinoic acid (tretinoin) prevents skin atrophy in mice treated with topical corticosteroids, but such studies have not been performed in humans. We performed an 8-week clinical, histological and biochemical study to test the ability of tretinoin to enhance efficacy and inhibit atrophogenicity of topical corticosteroids, when used in the treatment of psoriasis. In each of 20 psoriasis patients, one plaque, and its perilesional skin, was treated once daily with betamethasone dipropionate and tretinoin 0·1%, and one plaque, and its perilesional skin, treated with once daily betamethasone dipropionate and tretinoin vehicle. There was no difference in the speed or degree of improvement in plaques treated with either the topical corticosteroid/tretinoin combination or with corticosteroid alone. Light microscopy revealed a 19% reduction in epidermal thickness, in corticosteroid-treated perilesional skin, as compared with a slight (1%) increase in corticosteroid/ tretinoin-treated perilesional areas (P= 0.067). Western blot analysis showed a 55% reduction in procollagen I aminopropeptide in perilesional skin treated with corticosteroid alone, as compared with a 45% reduction in corticosteroid/tretinoin-treated perilesional skin. These data indicate that the addition of tretinoin does not impair the efficacy of a topical corticosteroid, in the treatment of psoriasis, and partially ameliorates epidermal atrophy produced by the topical corticosteroid.

Damage to skin collagen and elastin (extracellular matrix) is the hallmark of long-term exposure to solar ultraviolet irradiation, and is believed to be responsible for the wrinkled appearance of sun-exposed skin. We report here that matrix-degrading metalloproteinase messenger RNAs, proteins and activities are induced in human skin in vivo within hours of exposure to ultraviolet-B irradiation (UVB). Induction of metalloproteinase proteins and activities occurred at UVB doses well below those that cause skin reddening. Within minutes, low-dose UVB upregulated the transcription factors AP-1 and NF-kappa B, which are known to be stimulators of metalloproteinase genes. All-trans retinoic acid, which transrepresses AP-1 (ref. 8), applied before irradiation with UVB, substantially reduced AP-1 and metalloproteinase induction. We propose that elevated metalloproteinases, resulting from activation of AP-1 and NF-kappa B by low-dose solar irradiation, degrade collagen and elastin in skin. Such damage, if imperfectly repaired, would result in solar scars, which through accumulation from a lifetime of repeated low-dose sunlight exposure could cause premature skin ageing (photoageing).

We have quantitatively assessed the relation between type I and type III procollagen precursor levels and the severity of clinical photodamage in human skin. Levels of procollagen, pN collagen (collagen without the carbroxypropeptide), and/or pC collagen (collagen without the aminopropeptide) were determined by radioimmunoassay, Western blot, and immunohistology in punch biopsy specimens from mildly and severely photodamaged forearm skin and from sunprotected underarm and buttock skin of the same subjects. Collagen precursor levels in forearm and underarm skin were expressed relative to buttock levels for comparison. In the mildly photodamaged group, collagen precursors in the forearm did not differ from those in the underarm by any measurement, except for type I collagen precursors measured by radioimmunoassay, which were reduced 16%. In severely photodamaged forearm skin, both type I and type III collagen precursor levels, measured by radioimmunoassay, were significantly reduced (approximately 40%). Western analysis revealed similar significant reductions in type I and type III collagen precursor levels in severely photodamaged forearm skin compared with the sun-protected underarm. Immunohistology localized both type I and III pN collagens predominantly to the extracellular papillary dermis. Relative staining intensities of type I and type III pN collagen were also significantly reduced in severely photodamaged forearm skin. Multiple linear regression modeling of all data demonstrated that reductions in collagen precursor levels were significantly correlated (p < 0.03) with the severity of photodamage, but not with chronologic age. These data demonstrate, by three independent methods, coordinate reductions of both type I and type III collagen precursors in photodamaged human skin, and the degree of reduction correlated with the degree of photodamage. It is likely that such changes in collagen precursors lead to reduced levels and/or altered organization of fibrillar collagen, and thus may contribute to the wrinkled appearance of photodamaged human skin.

We have determined protein levels of total and individual nuclear retinoic acid (RAR-alpha, -beta, -gamma) and retinoid X (RXR-alpha, -beta, -gamma) receptors by ligand binding, Western analysis, and gel shift assays, in adult human skin, a major retinoid-responsive tissue. Total RARs and RXRs, measured by direct binding of specific ligands, were 0.24 +/- 0.01 fmol/micrograms (n = 13) and 1.26 +/- 0.08 fmol/micrograms (n = 7), respectively. These values calculated on an average per cell basis were 1790 RARs/cell and 9400 RXRs/cell. Similar results were obtained with competitive ligand binding assays. RAR-alpha, -beta, and -gamma were each specifically immunoprecipitated, and their levels determined by ligand binding assays of supernatants and Western analysis of precipitates. RAR-gamma was the most abundant, representing 87% of RAR protein. The remaining 12-14% of RAR protein was RAR-alpha. No RAR-beta was detected. Similar immunoprecipitation studies revealed that RXR-alpha represented 90% of RXR protein expressed in human skin. No RXR-beta or RXR-gamma proteins were detected by Western blot. Supershift gel retardation with antibodies to RARs detected probe-RAR-alpha and probe-RAR-gamma complexes in a 1 to 4 ratio. No probe-RAR-beta complex was detected. With antibodies to both RAR-gamma and RXR, a double supershifted complex was formed, indicating that RAR-gamma/RXR heterodimers bound to the probe. These data demonstrate 1) protein levels of RXRs are five times greater than RARs, 2) relative protein levels of RAR and RXR family members are compatible with their previously described relative mRNA levels, and 3) RXR-alpha/RAR-gamma heterodimers are the major retinoid receptors that have the potential to regulate transcription of target genes, in adult human skin.

Phosphatidic acid (PA) induced a rapid dose-dependent increase in production of inositol phosphates in cultured adult human keratinocytes, peaking at 30 s. Natural and dioleoyl PA were equally effective, while other phospholipid classes had no effect. Lipid A was also active. Lyso-PA also induced inositol phosphate production, but contamination of the PA preparation by lyso-PA could not account for the effect of PA. The effect of PA could not be reproduced by treatment of cells with calcium ionophore. PA-induced inositol phosphate production could be inhibited (> 50%) by pre-treatment of cells with either pertussis toxin or 12-O-tetradecanoylphorbol 13-acetate, suggesting the involvement of a GTP-binding protein and a protein kinase C-mediated negative feedback mechanism. PA also stimulated release of arachidonic acid from keratinocytes. Treatment of cells with exogenous phospholipase D similarly induced inositol phosphate production in the keratinocytes. Since PA may be formed by receptor-mediated activation of phospholipase D, or by phosphorylation of diacylglycerol, the results suggest that PA may play a significant role in signalling mechanisms of human keratinocytes.

We have investigated coupling between the epidermal growth factor (EGF) receptor and the phospholipase C (PLC)/protein kinase C (PKC) signal-transduction system in normal skin fibroblasts and keratinocytes, for which EGF and transforming growth factor alpha (TGF-alpha) are mitogenic. EGF and TGF-alpha induced a rapid increase in tyrosine phosphorylation of the EGF receptor, in both fibroblasts and keratinocytes, but failed to induce tyrosine phosphorylation of PLC-gamma 1 or detectable phosphoinositide hydrolysis, as measured by two sensitive assays. In fibroblasts, EGF induced phosphatidylcholine (PC) hydrolysis, resulting in increased diacylglycerol (DAG). In contrast, in keratinocytes, there was no detectable PC hydrolysis or elevation of DAG in response to EGF or TGF-alpha. EGF and TGF-alpha activated PKC in fibroblasts, as evidenced by increased phosphorylation of a specific cellular PKC substrate (myristoylated alanine-rich C-kinase substrate, ‘MARCKS’). In keratinocytes, TGF-alpha and EGF induced only a modest increase in MARCKS protein phosphorylation. This apparent modest activation of PKC, in the absence of detectable DAG formation, may have been mediated by arachidonic acid, which was released from keratinocytes in response to TGF-alpha, and has been shown to stimulate PKC activity in vitro. These data demonstrate that (1) in dermal fibroblasts and keratinocytes, which express normal levels of EGF receptors, EGF receptor activation is not coupled to tyrosine phosphorylation of PLC-gamma 1 or PtdIns hydrolysis, suggesting that these events are not required for the mitogenic activity of EGF or TGF-alpha in these cells, (2) coupling of EGF receptor to PC hydrolysis is cell-type specific, and (3) in skin fibroblasts, DAG, formed through EGF-induced PC hydrolysis, is capable of activating PKC.

Tyrosinase activity is a key determinant of melanin production in skin. Because retinoic acid regulates tyrosinase activity in melanoma cells, we analyzed modulation of pigmentation in vivo by retinoic acid. Black and white subjects were either not treated, or treated topically for 4 d under occlusion with vehicle, retinoic acid (0.1%), or the irritant sodium lauryl sulfate (2%). In untreated skin, tyrosinase activity and melanin content were significantly greater (2.3 times, and 3.2 times, respectively) in blacks versus whites. Four days of treatment with topical retinoic acid did not alter tyrosinase activity or melanin content in black skin. In contrast, retinoic acid treatment significantly induced (2.7 times, n = 8) tyrosinase activity, compared to vehicle treatment, in white skin. Melanin content, however, remained unchanged at 4 d. In separate experiments, tyrosinase activity in white subjects (n = 25) was increased 16% (p = 0.01) in sodium lauryl sulfate – treated skin, and 77% (p = 0.0005) in retinoic acid – treated skin, compared to vehicle-treated skin. The effect of retinoic acid on tyrosinase activity could be differentiated from non-specific irritation, because tyrosinase activity in retinoic acid – treated skin was significantly greater (52%, p = 0.004) than sodium lauryl sulfate-treated skin. Similar results were obtained with the dihydroxyphenylalanine reaction done on vehicle, sodium lauryl sulfate-, and retinoic acid – treated white skin. Northern analysis (n = 6) and semi-quantitative polymerase chain reaction (n = 6) demonstrated that retinoic acid treatment did not alter tyrosinase mRNA levels in white skin. Western analysis revealed that induction of tyrosinase activity by retinoic acid also was not associated with increased tyrosinase protein content (n = 9), indicating that regulation of tyrosinase activity by retinoic acid occurs through a post-translational mechanism. These data demonstrate that low tyrosinase activity in white skin in vivio is retinoic acid inducible and high tyrosinase activity in black skin in vivo is neither further induced nor reduced by retinoic acid

A cDNA corresponding to the membrane receptor for growth hormone (GH) was amplified by polymerase chain reaction (PCR) directly from human skin. The cDNA was cloned and found to have complete sequence homology to the extracellular domain of human liver GH receptor (GH-R). Northern analysis, using the cloned GH-R as probe, revealed relatively higher levels of GH-R transcripts in cultured human dermal fibroblasts compared to cultured keratinocytes or keratome biopsies. Semi-quantitative PCR analysis indicated that the level of GH-R mRNA in cultured melanocytes was similar to that in fibroblasts. The receptor protein encoded by GH-R mRNA in fibroblasts was shown by affinity cross-linking to have an apparent M(r) of 115-120 kDa, similar to that of 3T3-F442A fibroblasts used as a control. mRNA transcripts for the major mediator of GH actions, insulin-like growth factor 1 (IGF-1), were detected by PCR in fibroblasts, melanocytes, and keratome biopsies, but not in keratinocytes. In contrast, IGF-1 receptor mRNA were abundant in cultured keratinocytes and skin biopsies, as determined by Northern analysis. IGF-1 but not GH (5-50 ng/ml) promoted clonal proliferation of cultured keratinocytes. In contrast, GH (10 ng/ml) after 5 d markedly increased fibroblast cell numbers (70%, p less than 0.009) over 0.2% serum control. These data indicate that human skin cells possess the molecular elements necessary to respond to GH and raise the possibility that GH may influence skin growth in vivo.

The BCL2 (B cell lymphoma/leukemia-2) proto-oncogene encodes a 26-kDa protein that has been localized to the inner mitochondrial membrane and that has been shown to enhance the survival of some types of hematopoietic cells. Here we show that NIH3T3 fibroblasts stably transfected with a BCL2 expression plasmid exhibit reduced dependence on competence-inducing growth factors (platelet-derived growth factor, PDGF; epidermal growth factor, EGF) for initiation of DNA synthesis. The importance of BCL2 for growth factor-induced proliferation of these cells was further confirmed by the useage of BCL2 antisense oligodeoxynucleotides. The mechanisms by which overexpression of p26 BCL2 contributes to fibroblast proliferation are unknown, but do not involve alterations in: (a) the production of inositol triphosphates (IP3), (b) PDGF-induced transient elevations in cytosolic Ca2+ ions, or (c) the activity of protein kinase C enzymes in these transfected cells. The results imply that changes in mitochondrial functions play an important role in the early stages of the cell cycle that render 3T3 cells competent to respond to the serum progression factors that stimulate entry into S-phase.

Histologic and clinical improvement of sun-exposed skin following topical treatment with retinoic acid has been reported. Daily application of retinoic acid typically results within 2-5 d in an erythematous scaling reaction, which lessens with continued usage. The cellular, immunologic, and biochemical basis of this retinoid reaction and its role in the repair of photodamaged skin are not known. To investigate the retinoid reaction in man, we have treated non-sun-exposed skin with 0.1% retinoic acid cream (Retin-A, Ortho Pharmaceutical Corporation, Raritan, NJ) under occlusion for 4 d to induce erythema and then examined changes in 1) histology, 2) expression of cell-surface molecules, 3) the enzymes and second messengers of the phospholipase C/protein kinase C signal-transduction system, 4) levels of eicosanoids, and 5) levels of interleukin-1 protein and mRNA. These parameters were chosen for measurement both because they are indicators of epidermal function and previous studies suggest they may be responsive to retinoic acid treatment. Epidermal cell growth as judged by increased epidermal thickness and mitotic figures was significantly increased in retinoic acid-treated skin compared to vehicle-treated controls. Increased numbers of CD4+ T cells accompanied by prominence of dermal dendrocytes in the papillary dermis and focal keratinocyte expression of intercellular adhesion molecule-1 were observed in retinoic acid-treated biopsies. Phosphoinositide-specific phospholipase C activity and 1,2-diacylglycerol content were also elevated in retinoic acid-treated epidermis. Protein kinase C activity was reduced by one third in both the soluble and membrane fraction, suggesting down-regulation. Surprisingly, in view of the inflammatory nature of the retinoid reaction, no increases were observed in arachidonic acid, its metabolites, interleukin-1 alpha, or interleukin-1 beta. To examine the specificity of the retinoid reaction, subjects were treated with the irritant sodium lauryl sulfate, under conditions that resulted in a reaction clinically similar to that observed with retinoic acid. The histologic alterations induced by sodium lauryl sulfate were found to be indistinguishable from those induced by retinoic acid. These data indicate that, although a wide range of cellular and molecular alterations occur in retinoic acid-treated skin, these changes may not be necessarily specific or unique for retinoic acid.

Evidence suggests that the phospholipase C/protein kinase C signal transduction system participates in the regulation of epidermal cell growth and differentiation. Psoriatic epidermis is characterized by hyperproliferation, defective differentiation, and inflammation. In this report, we have determined the activity of phospholipase C-catalyzed hydrolysis of phosphatidylinositol-4,5-bisphosphate (PIP2) and 1,2-diacylglycerol content in normal and psoriatic involved and uninvolved epidermis. 1,2-diacylglycerol is formed from phospholipase C-catalyzed hydrolysis of PIP2 and is the physiologic activator of protein kinase C. PIP2 hydrolysis was assayed in soluble and particulate fractions prepared from keratome biopsies of normal and psoriatic skin. Total lipids were extracted from normal and psoriatic epidermis and 1,2-diradylglycerol (a mixture of 1,2-diacylglycerol and 1-ether, 2-acyl-glycerol) quantitated by enzyme assay. Because 1,2-diacylglycerol is a more potent activator of protein kinase C, the relative proportions of 1,2-diacyl and 1-ether, 2-acylglycerol in uninvolved and involved psoriatic epidermis were determined. This was accomplished by separation of acetate derivatives of 1,2-diacylglycerol and 1-ether, 2-acyl-glycerol by thin layer chromatography. Soluble and membrane-associated phospholipase C-catalyzed PIP2 hydrolysis were increased 3.7 times (p less than 0.001) and 3 times (p less than 0.004), respectively, in psoriatic involved compared to uninvolved and normal epidermis. 1,2-diradylglycerol content was also significantly elevated (3 times, p less than 0.01) in psoriatic involved versus uninvolved and normal epidermis. Analysis of the acetate derivatives of 1,2-diradylglycerol in psoriatic uninvolved and involved epidermis revealed that 1,2-diacylglycerol was the major species (86% and 95%, respectively). There were no significant differences in either phospholipase C-catalyzed PIP2 hydrolysis or 1,2-diacylglycerol content between uninvolved and normal epidermis. 1,2-diacylglycerol purified from normal and involved psoriatic epidermis was capable of activating protein kinase C from normal epidermis in vitro. In epidermal slices, activation of protein kinase C by addition of 12-0-tetradecanoylphorbol-13-acetate and 1,2-diacylglycerol (1,2-dioctanoylglycerol) resulted in subsequently decreased protein kinase C activity, a process termed down-regulation. These data are consistent with the possibility that the elevation in lesional 1,2-diacylglycerol content may account, in part, for the previously reported reduction of protein kinase C activity in psoriasis (Horn, Marks, Fisher, et al: J Invest Dermatol 88:220-222, 1987).

Cyclosporine A (CsA) is efficacious in the treatment of psoriasis. Although CsA is known to inhibit T-lymphocyte proliferation in vitro, whether this is its mode of action in psoriasis is uncertain. 12-0-tetradecanoylphorbol-13-acetate (TPA) induces an inflammatory, hyperplastic response in mouse skin, with many of the biochemical and histologic aberrations that occur in psoriatic epidermis. Protein kinase C is the major cellular phorbol ester receptor, and most responses of cells to TPA are mediated by PK-C, which is directly activated by TPA. We therefore have investigated the effects of CsA on pleiotypic responses induced by TPA and whether CsA acts in vivo as a direct inhibitor of PK-C. Simultaneous application of CsA (1.7 μmol) and TPA (10 nmol) to mouse skin significantly reduced inflammatory cell infiltration and increased epidermal thickness induced by TPA treatment alone. CsA had to be applied within 30 min of TPA application in order to have a significant inhibitory effect. Optimal doses of CsA inhibited TPA-induced ODC activity, TGase activity, arachidonic acid release, and interleukin-1β (IL-1β) mRNA to the same degree (approximately 80%), despite measurement at widely different times (30 min-12 h) required to obtain maximal induction by TPA. CsA did not, however, directly inhibit activation of PK-C by TPA. These data demonstrate that CsA blocks the pleiotypic responses of mouse skin to TPA treatment involving biochemical events, inflammatory cell infiltration, and epidermal hyperplasia. The molecular site(s) of action of CsA appears to be distal to the initial activation of PK-C by TPA and clearly inhibits PK-C inducible events. Furthermore, the above data suggest that CsA may directly affect keratinocytes in vivo.

The ability of a variety of agonists to induce formation of inositol phosphates and 1,2-diacylglycerol in cultured adult human keratinocytes has been investigated. Histamine, bradykinin, and thrombin significantly stimulated formation of inositol mono-, bis-, and trisphosphate and 1,2-diacylglycerol within 5 min after addition. Aluminum fluoride also caused a dose-dependent accumulation of inositol phosphates suggesting the participation of a GTP binding protein in the regulation of phospholipase C-catalyzed phosphoinositide hydrolysis. These data demonstrate that human keratinocytes possess the capacity for phospholipase C-mediated signal transduction and suggest that this pathway may participate in the regulation of keratinocyte function.

Two different methods were used to study directly α-thrombin modulation of polyphosphoinositide breakdown in membranes prepared from Chinese hamster lung (CHL) fibroblasts. In the first one we labelled the lipid pool by incubating the intact cells with myo -[su3H]inositol prior to membrane isolation; in the other we used exogenous [ 3 H]PIP 2 with phosphatidylethanolamine (1:10) added as liposomes to freshly isolated membranes. A Ca 2+ -dependent PIP 2 and PIP phospholipase C activity was characterized by measuring the rate of formation of inositol tris- and bisphosphate. Basal phospholipase C activity was stimulated up to 3-fold by GTP or GTP-γ-S. Of the two mitogens, α-thrombin and EGF, known to stimulate DNA synthesis in Chinese hamster fibroblasts, only α-thrombin is a potent activator of PIP 2 breakdown in intact cells. Consistent with this observation, α-thrombin but not EGF potentiated GTP-γ-S-dependent phospholipase C activity in membrane preparations. These results strongly support the hypothesis that a GTP-binding protein couples α-thrombin receptor to PIP 2 hydrolysis. Because both methods used to assay phospholipase C gave identical results, we conclude that the coupling is at the level of PIP 2 -phosphodiesterase activity.

The effects of bradykinin on activation of phosphoinositide turnover, 1,2-diglyceride formation, and growth of cultured adult human keratinocytes were investigated. Keratinocytes specifically bound [3H]bradykinin with high affinity (kd = 3.4 nM) and displayed 1.5 X 10(5) binding sites/cell. Bradykinin caused a rapid dose-dependent increase in inositol trisphosphate (IP3) inositol bisphosphate, and inositol monophosphate. IP3 was maximally increased (fivefold) at 30 s and remained elevated for at least 10 min. Half maximal stimulation of IP3 formation was observed at 27 nM bradykinin. IP3 accumulation was equally elevated by bradykinin and lys-bradykinin but was not stimulated by des-Arg9-bradykinin, indicating that phospholipase C in cultured keratinocytes is coupled to B2 bradykinin receptors. Treatment of keratinocytes with active phorbol ester (TPA) caused a significant inhibition (50%) of bradykinin-induced IP3 accumulation, suggesting negative regulation of phospholipase C by protein kinase C. Bradykinin also caused a significant elevation in 1,2-diacylglycerol (DAG) content. DAG content was maximally elevated (twofold) at 1 min and remained elevated for at least 10 min. Bradykinin also caused a significant (twofold, p less than 0.02) increase in keratinocyte growth. These data demonstrate that bradykinin is a potent agonist of the phospholipase C/protein kinase C signal transduction system in cultured adult human keratinocytes and that activation of this pathway by bradykinin is associated with increased keratinocyte growth.

Influence of aluminum on mineralization during matrix–induced bone development. A model of de novo mineralization employing matrix–induced endochrondral bone formation in rats was used to study the short–term effects of aluminum on the deposition of calcium and phosphate in vivo. In experiments where systemic aluminum concentrations were elevated, the cellular processes associated with bone development appeared to be normal, if somewhat delayed, however precipitation of the mineral phase was prevented. This suggests a primary direct physical chemical effect of aluminum in vivo on calcification, as suggested by in vitro studies which demonstrate that aluminum is a potent inhibitor of calcium phosphate precipitation. Aluminum salts implanted locally with the matrix appeared to be toxic to the cellular processes leading to chondrogenesis and osteogenesis.

Treatment of cultured adult human keratinocytes with platelet activating factor (PAF) resulted in a rapid, dose dependent accumulation of inositol phosphates. Inositol trisphosphate (IP3), inositol bisphosphate (IP2) and inositol phosphate (IP) were elevated within 15 seconds of exposure to PAF (1 microM). Lyso-PAF, phosphatidylcholine (PC) and lyso-PC had no effect on levels of inositol phosphates, indicating that the effect of PAF was specific. PAF also raised cellular 1,2-diacylglycerol content (2-fold) within two minutes of addition and stimulated mobilization of arachidonic acid (AA) and release of prostaglandin E2. In contrast, PAF did not stimulate phosphoinositide turnover or AA release in cultured dermal fibroblasts. These results suggest that the inflammatory effects of PAF in human skin result, at least in part, from its ability to directly activate keratinocytes and stimulate release of pro-inflammatory eicosanoids.