No abstract

Although lamellar granules are critical to the formation of the epidermal permeability barrier and are a known marker of late keratinocyte differentiation, very little is known about the physiologic regulators of lamellar granule assembly and extrusion. Ceramide glucosyltransferase (CGT), the enzyme responsible for the synthesis of lamellar granule glucosylceramides (GlcCer; the precursors of the stratum corneum ceramides), is localized to the Golgi apparatus in other cell types. We have found that CGT is induced during keratinocyte culture differentiation coincident with increased GlcCer content and the appearance of lamellar granules. In this study we show that the differentiation-related CGT induction is likely mediated at the transcriptional level. In addition, all-trans retinoic acid, a well-known inhibitor of keratinocyte differentiation, prevents the appearance of lamellar granules and decreases culture CGT activity and GlcCer content without affecting sphingomyelin or total lipid content, indicating a specific inhibition of this enzymatic pathway. These data show a direct relationship between CGT activity and epidermal differentiation, suggesting that regulation of CGT expression is a critical part of epidermal barrier generation. The differentiation dependence of CGT activity, the key role of this Golgi-localized enzyme in epidermal GlcCer synthesis, and our previous finding that ceramides are converted to GlcCer in the Golgi apparatus in keratinocyte cultures, strongly suggest a Golgi origin for lamellar granules. In contrast to CGT, the activity of the lysosomal enzymes acid lipase and glucocerebrosidase is less clearly related to epidermal differentiation and the appearance of lamellar granules, although both enzymes show striking colocalization and enrichment in a subcellular lamellar granule fraction derived from pig epidermis. Acid lipase activity in the lamellar granule fraction was found to contain primarily a small lysosomal form of the enzyme, whereas total acid lipase secreted by keratinocyte cultures was found to contain a mannose-6-phosphorylated large prelysosomal form as well as a small lysosomal form. That secreted acid lipase activity is derived from both prelysosomal and lysosomal compartments suggests there may be multiple pathways by which lysosomal enzymes are secreted from keratinocytes. The combined secretion of lipid and lysosomal enzymes from lamellar granules places these organelles in the category of "dual-function" specialized secretory vesicles described in certain other cell types. Electron microscopic images of lamellar granules show shapes consistent with cross-sections of tubules or buds from tubules in addition to vesicles. These images provide evidence for the involvement of trans-Golgi network tubules and/or buds in lamellar granule synthesis and secretion.

This paper reviews the experimental evidence for the proposal that hydrolytic enzymes are introduced into lysosomes of cultured fibroblasts only after secretion and receptormediated recapture.Key words: lysosornes, lysosomal enzymes, pinocytosis, secretion, of-biduronidase How are hydrolytic enzymes transferred from the site of their synthesis to lysosomes? In the usual concept of lysosome formation, the hydrolases proceed from rough t o smooth endoplasmic reticulum, then either t o the Golgi apparatus or t o a specialized region of the smooth endoplasmic reticulum known as GERL. There they are concentrated into small vesicles (primary lysosomes) which detach from the Golgi or from GERL. Eventually these vesicles fuse with phagocytic, pinocytic, or autophagic vacuoles which contain macromolecules t o be hydrolyzed but not the necessary enzymes. The organelles that result from the fusion -i.e., the secondary lysosomes -contain enzymes, substrates, and the appropriately acid environment for hydrolysis to take place (for reviews, see Refs. 1-4). The best evidence for the transport of the hydrolytic enzymes by way of primary lysosomes has been obtained in polymorphonuclear leukocytes. Particles which are rich in acid hydrolases and which correspond t o primary lysosomes -the azurophilic granuleshave been isolated by centrifugation and have been seen t o fuse with phagocytic vacuoles after ingestion of bacteria (2). However, the evidence is weaker for nonphagocytic mammalian cells, since primary lysosomes have not been isolated and their role in enzyme transport has been deduced from static morphological studies by electron microscopy. THE SECRETION-RECAPTURE HYPOTHESISAn alternative hypothesis would have hydrolytic enzymes secreted t o the cell exterior (perhaps through secretory vesicles), recaptured by a receptor-mediated endocytosis, and only then packaged into lysosomes (Fig. 1). Prior t o secretion, the enzymes would be equipped with a structural feature (a "recognition marker") to insure binding to the *Postdoctoral Fellow of the Arthritis Foundation