Effects of glycolic acid on desquamation-regulating proteinases in human stratum corneum.

In order to investigate the mechanism of glycolic acid (GA) function in human stratum corneum, we monitored changes in cathepsin D-like (CD) and chymotrypsin-like (SCCE) proteinases for 3 weeks following topical GA application (50% w/v, pH 0.9) for 30 min to human skin. In the early phase, weakened stratum corneum cohesion in the lower layers was observed on day 2 and the amount of active CD in the upper layer of the stratum corneum was significantly decreased from 30 min until day 2, whereas that in the lower layer remained normal. In contrast, the amount of active SCCE showed no change during the experimental period. The surface pH of the stratum corneum drastically decreased to pH 2 at 30 min and slightly recovered to around pH 3 until 1 day after treatment. From 9 to 19 days, a decrease in corneocyte cell area and a remarkable long-term increase in the amount of active CD in the upper layer were observed. In an in vitro study, the activities of desquamation-regulating proteinases were shown to have remarkably increased at around pH 3, due to activation of CD at its optimal pH. These results suggest that GA functions via at least two different mechanisms, acute activation of CD in the lower layer by acidification around pH 3, along with inactivation of CD in the upper layer, and long-term enhancement of de novo CD production in the few weeks following GA treatment.

Cathepsin D, but not cathepsin E, degrades desmosomes during epidermal desquamation.

BACKGROUND: We previously reported that an ambient aspartic proteinase is crucial to desquamation of the stratum corneum at pH 5. Identification of this aspartic proteinase by using enzyme inhibitors suggested it to be cathepsin D, although we could not exclude cathepsin E. OBJECTIVES: To determine the identity of this aspartic proteinase and its distribution within the stratum corneum. METHODS: We measured enzyme activities of cathepsin D and cathepsin E in the salt and detergent extracts from callus stratum corneum, using a fluorogenic peptide as a substrate and comparing the effect of addition of Ascaris pepsin inhibitor (specific for cathepsin E) with that of pepstatin A (which inhibits both cathepsin D and cathepsin E). Both enzymes were then extracted and purified from plantar stratum corneum samples and identified by Western blotting. Immunofluorescence microscopy was used to investigate the localization of proteinases within human plantar stratum corneum sample sections. RESULTS: We found that 20% of total aspartic proteinase activity could be attributed to cathepsin E, the remainder to cathepsin D. Two subunits of cathepsin D were identified, a mature active form at 33 kDa and an intermediate active form at 48 kDa; cathepsin E was also identified at 48 kDa, although in a stained band 10-fold weaker in the immunoblot. Immunofluorescence microscopy showed the antibody to cathepsin D to be localized in the lipid envelopes of the stratum corneum, whereas that to cathepsin E stained the tissue diffusely. The labelling for cathepsin D was similar to that observed for desmosomes, and immunoelectron microscopy confirmed that cathepsin D was present on desmosomes. On the other hand, cathepsin E occurred intracellularly within the squames. CONCLUSIONS: We conclude that cathepsin D, and not cathepsin E, causes desquamation by degrading desmosomes.

Zinc-alpha(2)-glycoprotein hinders cell proliferation and reduces cdc2 expression.

Zinc-alpha(2)-glycoprotein (Znalpha(2)gp) is widely distributed in body fluids and epithelia. Its expression in stratified epithelia increases with differentiation. We previously showed that Zn alpha(2)gp has ribonuclease activity, and that squamous tumor cells grown on a matrix of Znalpha(2)gp were growth-inhibited. Here we demonstrate, both by adding Znalpha(2)gp to the culture medium and, more unequivocally, by stably transfecting SiHa cells with Znalpha(2)gp cDNA, that the introduction of Znalpha(2)gp into SiHa tumor cells reduces proliferation. In response to Znalpha(2)gp, we find an accumulation of the cell population in G(2)/M by flow cytometry, paralleling the reduction of proliferation. In order to distinguish growth inhibition by cell cycle arrest from that produced by apoptosis or differentiation, we examine by RT-PCR how Znalpha(2)gp affects the expression of genes commonly used as markers of these properties. No changes are observed for PCNA, p53, c-myc, or bcl-2. Only cdc2 expression responds to Znalpha(2)gp, with a reduction of up to over a factor of two. Cdc2 is the only cyclin-dependent kinase regulating the G(2)/M transition without redundancy and is required as a rate-limiting step in the cell cycle. Its increased expression has been directly linked to increased proliferation and decreased differentiation of advanced tumors; conversely, its downregulation by Znalpha(2)gp might hinder tumor progression. J. Cell. Biochem. Suppl. 36: 162-169, 2001.

Correlation between pretreatment levels of interferon response genes and clinical responses to an immune response modifier (Imiquimod) in genital warts.

Imiquimod (IQ) has been successfully used in treatment of genital warts. In clinical settings, patients responded well but wart reduction rates varied. Our aim was to find a correlation between clinical responses and pretreatment (constitutive) levels of genes that might be involved in the molecular action of IQ. Since IQ is a cytokine inducer, we analyzed levels of expression of genes of the JAK/STAT signaling pathway and their inhibitors as well as interferon response factors (IRFs) in pretreatment biopsy specimens from complete responders (99 to 100% wart reduction rate) versus incomplete responders (75 to 92% wart reduction rate) by reverse transcription-PCR. We found that mRNA levels of signal transducer and activator of transcription 1 (STAT1) and IRF1 were higher in complete responders than in incomplete responders. Incomplete responders expressed larger amounts of STAT3, IRF2, and protein inhibitor of activated STAT1 (PIAS1) mRNAs compared to complete responders before IQ treatment. We hypothesize that high-level expression of STAT1 and IRF1 is advantageous for a better IQ response. The observed differences in constitutive mRNA levels of these genes may be the consequence of alterations in cellular differentiation and/or variable expression of endogenous interferons. Previous in vitro studies showed that keratinocyte differentiation coordinates the balance between positive and negative signals along the JAK/STAT pathway by regulating the IRF1:IRF2 and STAT1:PIAS1 ratios and thus affecting induction of IQ-inducible genes. Specifically, differentiation supports constitutive expression of STAT1 and IRF1 mRNAs but not expression of IRF2 and PIAS1. Our data are in good agreement with studies that showed the importance of STAT1 in cytokine induction and activation of interferon-responsive genes by IQ.

Response of keratinocytes from normal and psoriatic epidermis to interferon-gamma differs in the expression of zinc-alpha(2)-glycoprotein and cathepsin D.

Psoriasis is a T cell-mediated inflammatory disease characterized by hyperproliferation and by aberrant differentiation. We found cathepsin D and zinc-alpha(2)-glycoprotein, two catalytic enzymes associated with apoptosis and desquamation, to be present in the stratum corneum of the normal epidermis but absent from the psoriatic plaque. Psoriasis is characterized by an altered response to interferon-gamma (IFN-gamma), including the induction of apoptosis in normal but not in psoriatic keratinocytes, often with opposite effects on gene expression of suprabasal proteins. We found that IFN-gamma binding and signaling were attenuated in psoriasis: The IFN-gamma receptor, the signal transducer and activator of transcription STAT-1, and the interferon regulatory factor IRF-1 were strongly up-regulated by IFN-gamma in normal keratinocytes, but not in psoriatic ones. IFN-gamma strongly up-regulated the expression of the catalytic enzymes cathepsin D and zinc-alpha(2)-glycoprotein in normal keratinocytes but down-regulated them in psoriatic ones; the reverse was true of the apoptotic suppressor bcl-2. We believe that the aberrant response to IFN-gamma plays a central role in the pathophysiology of psoriasis, particularly the disruption of apoptosis and desquamation.

Role of endogenous cathepsin D-like and chymotrypsin-like proteolysis in human epidermal desquamation.

Even though the skin surface is acidic (about pH 5), most in vitro studies on desquamation have been performed at alkaline pH. We demonstrate that the standard in vitro model system, which achieves squame shedding upon incubation of plantar stratum corneum for 1 day in an alkaline buffer that must include a chelating agent, can be extended to a more realistic model in which the incubation is for 4 days, at varying pHs from 5 to 8, without exogenous chelators. Desmoglein I from stratum corneum was degraded by the squames shed at pH 5 as well as at pH 8. Squame shedding was inhibited to varying extents by the addition of proteinase inhibitors, whose specificity suggested that the crucial enzymatic activity at pH 8 was a chymotrypsin-like serine proteinase, while a similar activity at pH 5 was accompanied by an aspartic proteinase activity of comparable strength. Four degradation peaks were observed when the insulin B chain was reacted with shed squames at pH 5. Two of these peptides were suppressed by the addition of phenylmethylsulphonyl fluoride, the other two by pepstatin A; chymostatin inhibited all four, but E-64 and leupeptin showed no effect. The implied specificity was confirmed by reacting the insulin (without squames) with the standard enzymes human liver cathepsin D and pancreatic chymotrypsin, reproducing the expected degradation products. These results suggest that epidermal desquamation at acidic pH requires two proteolytic activities, one of which is an analogue of chymotrypsin and the other of cathepsin D. Endogenous proteinases corresponding to these activities have been previously identified, namely the stratum corneum chymotryptic enzyme and the mature active form of cathepsin D.

Characterization of zinc-alpha(2)-glycoprotein as a cell adhesion molecule that inhibits the proliferation of an oral tumor cell line.

Zn-alpha(2)-glycoprotein (Znalpha(2)gp) is a soluble protein widely distributed in body fluids and glandular epithelia. We have found it to be expressed in stratified epithelia as well. Znalpha(2)gp is clinically correlated with differentiation in various epithelial tumors, including oral and epidermal tumors. We have cloned epidermal Znalpha(2)gp and report the preparation of the recombinant protein in a Baculovirus expression system. Like the native molecule, recombinant Znalpha(2)gp has RNase activity. Znalpha(2)gp functions as a matrix protein for the Tu-138 oral squamous cell carcinoma cell line. Cell attachment to Znalpha(2)gp is comparable to that for fibronectin and is inhibited by the synthetic RGD peptides RGD, RGDV, and RGDS. Attachment is also inhibited by the antibody to integrin alpha(5)beta(1) (the fibronectin receptor), but not by antibodies to integrins alpha(v)beta(3), alpha(3)beta(1), and alpha(2)beta(1). We find that the proliferation of Tu-138 cells is inhibited on a Znalpha(2)gp matrix, as compared with other matrix proteins (fibronectin, vitronectin, laminin, and collagens I and IV) on which growth resembles that on the BSA control. We believe that the role of Znalpha(2)gp in differentiation and its RNase activity are two likely suspects as agents of the inhibition of proliferation.

Zinc-alpha2-glycoprotein expression as a marker of differentiation in human oral tumors.

Zinc-alpha2-glycoprotein (Znalpha2gp) is a soluble major histocompatibility complex homolog widespread in body fluids and in glandular epithelia; the authors recently demonstrated its presence in stratified epithelia. Znalpha2gp has been associated with tumor differentiation in breast cancers and other carcinomas. We compare here its gene expression in histopathologically graded oral squamous cell carcinomas and in their perilesional normals. Znalpha2gp levels are higher in the controls than in the tumors, and higher in well-differentiated tumors than in poorly differentiated ones. Markers of oral epithelial maturation (keratin K13 and involucrin) are less simply related to tumor histology.

Local inflammation may influence oral tumor cell differentiation.

The mRNA levels of keratin K13, involucrin, protein kinase C alpha and epsilon, and interferon-gamma and its receptors were examined in biopsies from human oral squamous cell carcinomas. Expression of all the genes was elevated in the histologically more differentiated tumors, but it was at or below normal (perilesional control) levels in the poorly differentiated ones. For the same set of biopsies, we had previously shown that the well differentiated tumors expressed higher levels of T cell markers. As interferon-gamma stimulates differentiation, its secretion by inflammatory cells at the tumor site may influence the differentiation status of the tumor.

Differentiation-dependent expression of growth regulatory cytokines and their receptors in squamous cell carcinomas of the head and neck.

Cytokines, such as IFN-alpha and TNF-alpha are capable of affecting keratinocyte proliferation in the microenvironment of the tumor. Their elevated expression along with high levels of their receptor mRNAs was determined by a semiquantitative reverse transcription- polymerase chain reaction (RT-PCR) method in biopsies of head and neck squamous cell carcinomas that were established as histologically well or moderately differentiated. In contrast, tumors with poor differentiation exhibited low levels of these growth suppressive factors, although levels of their receptors were elevated. In fact, expression of these growth suppressive cytokines highly correlated with the histological status of tumors suggesting a role of these agents in growth regulation of those tumors. Apparently, growth signaling in these tumors differs in the availability of either the ligand or the receptor.

Cytokine modulation of human corneal epithelial cell ICAM-1 (CD54) expression.

To determine whether pro-inflammatory cytokines modulate intercellular adhesion molecule-1 (ICAM-1; CD54) expression on cultured primary human corneal epithelial cells (HCEs), confluent HCEs were treated with various concentrations of interferon-gamma(IFN-gamma), interleukin-1alpha(IL-1alpha), IL-1beta, IL-4, tumor necrosis factor-alpha (TNF-alpha), or combinations over time. ICAM-1 expression was measured by flow cytometry and/or a cell-based ELISA using a monoclonal mouse anti-human CD54 antibody. The apparent MW of ICAM-1 protein was determined by immunoprecipitation of biotinylated HCEs. RT-PCR was used to detect ICAM-1 RNA. The mature cell surface form of HCE ICAM-1 was approximately 110 kDa as determined by immunoprecipitation. IFN-gammaand TNF-alpha induced both dose- and time-dependent increases in ICAM-1 expression. An approximately 20-fold increase in ICAM-1 was seen at 50-100 U IFN-gamma ml-1. ICAM-1 specific mRNA accumulated approximately 4.5-fold after IFN-gammatreatment. TNF-alpha(100 U ml-1) induced a consistent approximately 6.0-fold increase in ICAM-1 expression. When IFN-gammaand TNF-alpha were mixed, at sub-optimal concentrations of each, a synergistic effect on ICAM-1 expression was not detected. Neither IL-4, IL-1alpha nor IL-1beta affected ICAM-1 expression in a consistent fashion. In summary, ICAM-1 was modulated on primary human corneal epithelial cells by the cytokines IFN-gamma and TNF-alpha in a dose- and time-dependent fashion. Cytokine modulation of corneal epithelial cell ICAM-1 during inflammation may contribute to corneal epithelial cell injury by aiding the attachment of inflammatory cells such as eosinophils which express the receptor for ICAM-1, the beta2 integrins (CD11a,b,c/CD18).

Isoforms of cathepsin D and human epidermal differentiation.

Cathepsin D is an ubiquitously expressed lysosomal aspartic proteinase, with well-determined structural and chemical properties but a less clearly defined biological role. In stratified epithelia, the chronology of cathepsin D activation and degradation can be connected with stages of cellular differentiation. We partially purified cathepsin D from human epidermis and from separated stratum corneum by standard biochemical procedures, monitored by SDS-PAGE and Western blotting, and verified its identity as to molecular mass, pH optimum, N-terminal sequencing, reactivity with the specific antibody, inhibition by pepstatin A, and specific enzyme activity. It had hemoglobin-degrading activity over the acid range, with maximum at pH 3. It also degraded bovine serum albumin, human keratins, and stratum corneum extracts at pH 4. We discerned all three isoforms of human cathepsin D (the 52 kDa proenzyme and the active forms at 48 kDa and 33 kDa) in the epidermis; both active forms were also seen in the stratum corneum, but the proenzyme was not. Gene expression of cathepsin D in epidermal keratinocytes resembled that of suprabasal structural proteins (involucrin, keratin K10, transglutaminase) in its response to the calcium switch. An antibody to the 33 kda isoform immunolocalized to the granular layer and the stratum corneum (whereas antibodies to the 48 kDa isoform have been reported to stain mainly the upper spinous and granular layers). A plausible hypothesis to harmonize these results is that cathepsin D is first expressed as the proenzyme in the upper spinous layer, is activated in the lysosomes in the granular layer to the 48 kDa form, and is degraded to the 33 kDa form in the transition zone between the granular layer and the stratum corneum. As the stratum corneum is an acid environment, with an ambient pH of approximately 4.5, cathepsin D is available and suited to contribute to desquamation.

Differentiation and cathepsin D expression in human oral tumors.

OBJECTIVES/HYPOTHESIS: This study aimed to ascertain whether cathepsin D expression could be related to the stage of differentiation of oral tumors. STUDY DESIGN: Human oral biopsies of 10 squamous cell carcinomas and of the corresponding perilesional normal tissues were used. The tumors had all been clinically graded as advanced stage but nonmetastatic; five were classified histopathologically as poorly differentiated. METHODS: The gene expression of cathepsin D and keratin K13 in the biopsies was measured by reverse transcription polymerase chain reaction. Ratios of tumor-to-control readings helped compensate for sample variability. RESULTS: Keratin K13, as a suprabasal cell marker, tended to confirm the histological grading of the tumors (but was not otherwise useful in distinguishing tumors from normal tissue). Substantial overexpression of cathepsin D was found in the poorly differentiated tumors. CONCLUSIONS: Cathepsin D overexpression is considered a prognostic indicator of metastasis. In this sample, it was also associated with dedifferentiation. Cathepsin D might serve as a valuable gauge in clinical exploration of the connection between dedifferentiation and metastasis.

Induction by interferon-gamma of its receptor varies with epithelial differentiation and cell type.

Normal keratinocytes from epidermis and from buccal mucosa were cultured to confluence in three media with graded differentiation potential and treated with interferon-gamma (IFN-gamma). RT-PCR was used to measure the gene expression of the IFN-gamma receptor (IFNGR-1), as well as the immunomodulator HLA-DR and two enzymes of the 2-5 A pathway. We have previously reported results for a number of structural and regulatory genes in the same system (and include here involucrin for comparison). In epidermal keratinocytes, the induction of IFNGR-1 was upregulated by incubation with IFN-gamma, and this increased with the differentiation potential of the culture medium. A roughly similar pattern occurred for the other genes. In mucosal keratinocytes, in contrast, IFN-gamma failed to induce expression of IFNGR-1 or the other genes. A unique characteristic of HLA-DR was that its induction by IFN-gamma was uniform, for both tissues and all media. The gene expression of the receptor IFNGR-1 appears to be the dominant factor in the sensitivity of other genes to IFN-gamma, although there are substantial disparities among them that presumably reflect functional differences. The difference between the two tissues may be linked to differentiation, as the epidermis has a much more extensive maturation pattern than the buccal mucosa. A clinical implication is a better prognosis for IFN-gamma treatment for more differentiated tumors. Indeed, a previous study has found that the maturation pattern of condylomas responding to interferon treatment resembles that of epidermis, whereas the maturation of nonresponders is more akin to that of buccal mucosa.

Zinc-alpha 2-glycoprotein has ribonuclease activity.

Zinc-alpha 2-glycoprotein (Zn alpha 2gp) is widely distributed in body fluids and in various epithelia; its gene has been completely sequenced, but its function has long remained elusive. We have found that Zn alpha 2gp has RNase activity, comparable to onconase but two orders of magnitude less than RNase A. The RNase activity of Zn alpha 2gp is characterized by maxima in pH at 7.5, in ionic strength at 50 mM NaCl, and in temperature at 60 degreesC. It is strongly inhibited by ZnCl2, but unaffected by MgCl2. It is partially inactivated (down to 20%) by the placental RNase inhibitor. On synthetic polyribonucleotide substrates, the RNase activity of Zn alpha 2gp is specific for pyrimidine residues [poly(C) and poly(U) equally] and cleaves only single-stranded RNA. For onconase, it has been demonstrated that the RNase activity depends on pyroglutamic acid (pyr 1) as the N-terminus; Zn alpha 2gp also has pyr 1, while RNase A does not. Alignment of the amino acid sequences of Zn alpha 2gp and onconase or RNase A reveals only modest matches. Despite the more substantial overall structural homology of Zn alpha 2gp to class I major histocompatibility complex proteins, Zn alpha 2gp has not been proven to be associated with the immune response and, conversely, we could not detect RNase activity in six class I HLA heavy chains.

Proteasomal RNase activity in human epidermis.

The proteasome is a cytoplasmic high-molecular-weight structure composed of several smaller protein and RNA subunits. It has been associated with non-lysosomal pathways of intracellular degradation, expressing multicatalytic proteinase activities and specific RNase activity. By standard methods, we have isolated andpartially purified proteasomes from human epidermis. We obtained the expected multiple 24-32 kDa subunits by SDS-PAGE, and evidence of RNA. Proteasomes degraded casein, as well as chromogens for t-PA and trypsin but not for chymotrypsin, these proteolytic activities overlap, but do not coincide with those observed in other organs. We found that human epidermal 28 S and 18 S rRNAs were degraded, but yeast RNA was not. By means of zymography, we demonstrated, for the first time, that RNase activity persists after dissociation of the proteasome on the gel and that it co-localizes to the same range of molecular weight subunits as the proteinase activity.

Tumor differentiation-dependent local immunity in human head and neck cancers.

Distribution of markers of local cell-mediated immunity was examined in oral tumors exhibiting different histological stages of differentiation. Using a RT-PCR-based semiquantitative technique we determined levels of Langerhans cells, CD4- and CD8-positive T-cells, macrophages/NK cells, beta2-microglobulin and IFN-gamma mRNAs from tissue biopsies. A positive correlation was found between levels of these immunological markers and the tumor differentiation stage. Since tumor differentiation may correlate with the prognosis and response to various treatment modalities, our results may be useful clinically.

Desquamin is an epidermal ribonuclease.

Desquamin is a glycoprotein that we have isolated from the upper granular layer and the stratum corneum of human epidermis; it is not ordinarily expressed in submerged cultures, whose terminal differentiation stops short of formation of these layers. The exogenous addition of desquamin to human cultured keratinocytes extended their maturation, and hematoxylin staining indicated a loss of cell nuclei. For confirmation, cultured cells were lysed in situ, and the nuclei were incubated with desquamin for several days, then stained with hematoxylin. Damage to the nuclei was evident: the nuclear inclusions remained intact, while the surrounding basophilic nuclear matrix was degraded. Desquamin was then tested directly for nuclease activity. Ribonuclease activity was determined by incubating desquamin with human epidermal total RNA and monitoring the dose-dependent disappearance of the 28S and 18S ribosomal RNA bands in an agarose/formaldehyde gel. On RNA-containing zymogels, we confirmed the RNase activity to be specific to desquamin. Using synthetic RNA homopolymers, we found the active RNase domains to be limited to cytosine residues. On the contrary, DNA was not degraded by an analogous procedure, even after strand-separation by denaturation.

Modulation by interferon-gamma of zinc-alpha 2-glycoprotein gene expression in human epithelial cell lines.

Zinc-alpha 2-glycoprotein has been detected in most body fluids, and its antibody labels the corresponding glandular epithelia. We have also detected it in human stratified epithelia (epidermis and buccal mucosa). In this study, the mRNA levels of zinc-alpha 2-glycoprotein were found to be about twice as high in epithelial cells of mucosal origin (whether normal primaries or neoplastic cell lines) as in epidermoid cells (normal epidermal primary cultures, an immortalized but non-tumorigenic epidermal cell line, and neoplastic vulvar and cervical cell lines). Interferon-gamma strongly upregulated gene expression, but substantially less in mucosal than epidermoid cells. To compare responses as a clue to the function of zinc-alpha 2-glycoprotein, we ran parallel experiments with three markers of distinct properties, all known to be induced by interferon-gamma. There was the least resemblance for involucrin, a qualitative similarity for HLA-DR, and a rather better match for 2'-5' oligoadenylate synthetase.