0022-2 02X/85/8fi04 -0289$02.00/0 THE JOU R NAL OF IN VESTIGATIVE DERMATOI..Q(:Y, 85:289- 294, 1985 Vol. 85. No. 4 Copyright (c:;) 1985 by The Williams & Wilkins Co. Printed in U.S.A. REPORTS

Lamellar Body-Enriched Fractions from Neonatal Mice: Preparative Techniques and Partial Characterization

STEPHEN GRAYSON, PH.D., ANNA G. JOHNSON-WINEGAR, PH.D., BRUCE U. WINTROUB, M.D., RIVKAH R. lSSEROFF, M.D., ERVIN H. EPSTEIN, JR., M .D ., AND PETER M. ELIAS, M.D. D ermatology Service, Veterans Administration M edical Center, San Francisco, Dermatology Unit, San Francisco General Hospital, and Department of Dermatology, University of California School of Medidne, S an Francisco and Da vis, Call:[o mia, U.S.A.

Several problems have frustrated the isolation of la­ water-soluble tracers [ 4,6], and t heir variations in structure (9] mellar bodies (LB) from mammalian epidermis. We ob­ and apparent lipid content [10] in essential fatty acid defi­ tained pellets enriched in intact LB by utilizing the ciency, a role for LB in t he permeability barrier has been staphylococcal epidermolytic toxin to provide intact, proposed (reviewed in [1 ,11 ,12]) . On the other hand, t he cyto­ outer epidermal sheets, by controlled homogenization in chemical [1 ,13,14] and biochemical [15] demonstration of hy­ a cell disrupter, and by passage of homogenates through drolyt ic enzyme activity, suggests that they m ay be lysosomes a graded series of nuclepore filters (Science 221:962, that could regulate desquamat ion (reviewed in [1 ,16]). 1983). Such preparations contained more intact LB than To distinguish a mong these p ossibilities the LB first must did fractions prepared by a variety of differential or be purified a nd its contents scrut inized. Recently, Freinkel and sucrose/metrizamide discontinuous centrifugation Traczyk have obtained LB-containing fractions from fetal rat methods. Initial characterization of the enzymatic con­ skin [17], which were enriched in certain sphingolipids (18] tent of this fraction revealed it to be enriched in certain a nd hydrolytic en zymes [15]. Using a new isolation scheme we hydrolytic enzymes (acid phosphatase, carboxypepti­ a lso were able to isolate a n LB-enriched fraction from neonatal dase, cathepsin B, acid lipase, sphingomyelinase, and mouse skin [19] . In this p aper, we first compare t his technique phospholipase A), but strikingly depleted in all sulfa­ to other isolation methods; second, we provide evidence t hat tases, {j'-glucuronidase, and the non-lysosomal protease, LB display a s pecific spectrum of hydrolytic enzymes; and third, plasminogen activator. Thus, LB show some properties we demonstrate t hat LB a re enriched in total lipid a nd in of lysosomes, although certain characteristic lysosomal certain lipid species that may contribute to ba rrier format ion. enzymes are strikingly absent. Lamellar body fractions contained 2-3 times more lipid per unit weight than did MATERIALS AND METHODS homogenates, and were enriched in phospholipids, free Isolation ProtocoLs sterols, and glycosphingolipids, but not in other neutral D£fferential and ,; .erose density centrifugation: One hundred ICR lipids or ceramides. In summary, whereas some of the newborn mice (Simonsen Labs, Gi lroy, California) were i11j ected with enzymes in LB could participate in the metabolism of a hi ghly purified staphylococcal epidermolytic toxin fraction [20], and LB lipid precursors to hydrophobic barrier constituents, immediately tape-stripped d ow n to the gli stening layer (:::. 3-4 strip­ others may attack intercellular constituents, ultimately ping). After the animals were kept for 2 h at 37 "C, the outer epidermis resulting in desquamation. The lipid profile of these wa s peeled ofT and pl aced in STE buffer (0.32 M sucrose, 0.01 M Tris­ organelles suggests that they deliver precursors of HCI, 10 mM ethylenediaminetetraacetic acid, EDTA, pH 7.2). The permeability barrier lipids to intercellular domains. pooled outer epidermal sheets were homogeni zed (in so me cases pre­ ceded by incubation in 0. 1% trypsin in phosphate-buffered saline (PBS)) in a Polytron PT10 (Brinkman Instruments, Palo Alto, Cali ­ Despite intens ive investigation over t he past 25 years, t he fornia) at setting #5, for about 10 s, strained through cheesecloth or a 72- mesh n ylon screen to remove wh ole cell s, g precise function of t he epidermal lamella r body (LB) still is not and ce ntri fuged at 500 for 10 min. For di ffe rential centrifugation studies, the supernatant was known (reviewed in [1)). Careful intrastructural studies have ce ntrifuged first at 1,000 g for 5 min , and each subsequent supernatant now delineated the sequence of events t hat accompany both was centrifuged at 5,000 g, 15,000 g, 40,000 g, and at 100,000 g. For intercellular secretion of LB contents [2,3] and t he subsequent sucrose density gradients, the granular cell suspension was ce nt ri fu ged remodeling of LB discs into intercellular membrane laminae at 5,000 g for 5 min, resuspended in ice-cold homogeni zation medium within t he stratum corneum [4- 6]. Based upon their apparent (HM), containing Hanks' balanced salt solution with 10% fetal calf lipid [7] and sugar [7,8] content, t heir ability to exclude injected serum , 0.1 % soybean trypsin inhibitor, 5 l'g/ml pepstatin A (Sigma, St. Loui s, Missouri) , 5 l' g/ml leupeptin (Peninsula Labs. , San Carl os, Ca lifornia), a nd 10-5 M hydrocortisone, pH 7.0, resuspended in HM Manusc ript rece ived D ecembe r 31, 1984; accepted for publication and ruptured in a Stansted cell disrupter at 5,000 psi (Stansted Fluid April 30, 1985. Power, Ltd., Stansted, Essex, U.K. ) [21]. After discarding the 1,900 g This work was supported by NIH grants AM 19098, AM 31901, AM pell et, t he 30,000 g pellet was resuspended in 3 ml HM and layered on 01195, the Uni ve rsity of California Cance r Research Co -o rdinatin g top of a discontinuous sucrose gradient composed of 4 ml 0.6 M, 5 rnl Committee, and the Medical Research Service, Veterans Administra­ 0.75 M, and 5 ml 0.9 M sucrose made up in HM. The g radi ent was tion. ce ntrifuged in a swinging bucket rotor at 72,000 g for 2 h fo ll owed by Reprint requests to: Peter M. Elias, M.D., Dermatology Service co ll ection of the follow in g fractions: (a) top of 0.6 M layer; (b) 0.6 M/ (190), Veterans Administration Medical Center, 4150 Clement Street, 0.75 M in terface; (c) 0.75 M/ 0.9 M interface; and (d) pellet. San Francisco, California 94 121. Metrizam.ide gradient: Skin from 100-200 neonatal ICR mice was Abbreviations: incubated dermis-side-downward on 0.1% co llagenase in DME H21 DHEAS: dehydroepiandrosterone med ium, co ntaining 10 mM EDTA and 20 mM HEPES buffe r, pH 6.0, HM: homoge ni zation medium for 1 h at 37"C, as previously described [1 7]. After passage t hrough LB: lame ll ar body(i es) cheesecloth, fil tered ce ll s were ruptured in the Stansted cell disrupter PA: plasminogen activator at 5,000 p.s. i. Subsequent steps were carried out as previously described PBS: phos phate-buffered saline [1 7]. TLC: thin-layer chromatography Nuclepore filter prepara tion: Our technique for the iso lation of LB

289 290 GRAYSON ET AL Vol. 85, No. 4 has been d escribed in detail before [19]. B rie f1 y, peeled outer epidermal Freinkel a nd Traczyk [29]. Incubation mixtures contained 0.1 M tau­ sheets fr om 200 lCR newborn mi ce were ge ntly homogenized in a loose­ rodeoxycholate, 5 mM CaCI2, a nd 0.05 fl l of [1 ,2- '''C]L-a-dipalmitoyl­ fitting glass homogenizer, filtered through gauze, a nd ruptured in the phosphatidylcholine (PC) in 0.05 M sodium acetate buffer, pH 4.5. A Stanst.ed cell disrupter at 5000 p.s.i. (21]. The homogenate was rapidly ra nge of 0.2- 0.6 mg of sample was added to 0.5 ml of incubation a nd successively passed through a prefilte r, foll owed b y H series of mixture. After 45-min incubation at 37"C, a n excess (10 ml) of chlo­ nuclepore fil ters of pore sizes 8.0, 3.0, 1.0, 0.8, 0.6, 0.4 (filtrate employed roform:methanol (2: 1, v/v) was added to terminate assays. Organic for biochemical studies), a nd 0.2 11m [1 9]. phases were fractionated by one-dimensional thin-layer chromatogra­ phy (TLC) in a polar lipid solvent system on plastic- backed TLC plates Enzym.e Assays (Me rck, Darmstadt, West Germany) a nd bands corresponding to free fatty acids, PC, a nd lysolecithin (see below) were cut o ut and counted, Ca rboxypeptidase: Carboxypeptidase activity was assessed by cleav­ as above. age of t he carboxy-terminus of a n ~iotensi n J (Vega, Tucson, Arizona). Proteins were determined by either a standard or modified [30] Va ri ous samples (100 11 !) we re incubated with angiotensin I (5 X w-• Lowry method. M) in a total volume of 500 fl l in 0.01. M Tris-H CI buffer, pH 7.4, 0.15 0 M NaCl for 30 min a t :37· c. Cleavage of angiotensin I to des-leu' - angiotensin I was measured b y an isocratic high-pe rforma nce li quid Lipid Extra.ction and Analysis chromatography assay, which was externa lly calibrated with t he pep­ tide markers, a ngiotensin I, a ngiotensin II, and des-leu 10-angiotens in I . Homogenates and cell fractions were extracted in Bli gh/ Dyer solu­ X a nce TLC [22]. tiOn and fractionated by TLC on 10 20-cm high-perform Cathepsin B -lihe activity: Samples of each fraction were added to plates (Merck), as previou ly reported (7]. The most nonpolar, o rcein­ potassium phosphate buffer, containing freshly a dded 8 mM cysteine positive glycosphingolipid was identified as acylglucosylceramide by a nd 4 mM EDTA, a nd incubated for 10 min at 40"C. Reactions were cochromatography of authentic material [31] , provided courtesy of Dr. stopped b y a ddition of 100 mM sodium chloroacet.ate in phosphate Harold Yardley. buffer, containing 30 rnM sodium acetate, 70 mM acetic acid, pH 4.3, Microchromatography on silica gel-coa ted quartz rods: Lipids were and activity was determined b y measurement of amount o f 7-amino-4- dried down and resuspended in wa rm chloroform:methanol (1 :2) t.o a met.hyl couma rin (Vega) in 11 g released from N-Z-L-phenylalanyi-L­ final lipid concentration of about 20 l'g/1'1. One microliter or less was arginine 4-met.hyl coumaryl-7-amide H CI (Vega) per min in a fluori­ spotted on each silica gel-coated quartz rod (Chromarod S-II, Ancal, meter, as described previously [23]. Inc., Los Osos, California), and the rod was then developed in t he Cell-a:;8ociated plasminogen act.iuator levels: Cell pellets a nd homog­ following solvent systems: To resolve neutral lipids, rods were run 2/3 enates were disrupted b y sanification for 1 5 2-s bursts (Branson soni­ of their length in n-hexane:diethyl ethe r:formic acid (80:20:1), dried for cator) in ice-cold 0.1 M Tris-H CI, pH 8.1, containing 0.5% Triton X- 40 sat 100"C, and t hen developed fu ll length in n-hexane alone. Rod 100. The ce ll extracts were t hen centrifuged f or 20 min at 3200 gat were t hen scanned from the solvent fron t through the free sterol peak 4"C and a liquots of the s upernatant in duplicate in 2 sepa rate samples leaving the origin unburned. The origin (containing all of t he g lyco­ were assayed immediately for protein content a nd plasminogen acti­ sphingolipids, ceramides, and polar lipids) was then developed full vator (PA) activity, as previously described [ 24,25]. The dishes we re length in chloroform:methanoi:H20 (50:25:3) . Heating, d evelopment, incubated at 37"C and an aliquot of t he supernatant fluid was assayed a nd data a nalysis were performed in a microchromatographic device for solubilized " 51-labeled fibrin degradation products at 2 a nd 4 h. PA (latroscan TH10-Mark lll-TLC Analyzer, Ancal Corp., Los Osos, is expressed as milliunits (Ploug) of U rokinase Reference Standa rd California), as described recently [32]. (Leo Pha rmaceutical Products, Ballerup, Denmark). Control assays of a ll ext racts in the absence of plasminogen r evealed no plasminoge n­ Electron Microscopy dependent fibrinolysis. A cid phosphatase and {1-f!lu cu.ron.idase: Acid phosphatase and {3- Samples (whole epidermis a nd pellets) were fix ed first in 2% glutar­ glucuronidase activities were assessed by standa rd colorimetric tech­ aldehyde/2% formaldehyde in 0.1 M cacodylate buffer containing 0.06% nique ( kit.s for b oth assays we re from Sigma). CaCI2, followed by 1% aqueous or cacodylate-buffe red osmium tetrox­ Aryl sulfa.ta:;e A and B: The broken cell homogena tes were homoge­ ide, or in 2% buffered osmium tetroxide with 1.0% potassium ferrocy­ nized in iced 0.014 M Tris-H CI buffer, pH 8.0, with 0.5% Miranol anide in t he dark to accentuate g lycogen a nd membrane lipids while (Mira nol H2M was a ge ne rous g ift of the Mira nol Che mical Co., Inc., diminishing the contrast of nucleoproteins and other structural pro­ Dayton, New J e rsey), using the Polytro n PT-10 with 1 or 2 10-s bursts teins [33]. After fixation, the samples were processed and embedded as at maximum speed. The activity of arylsulfa tase A a nd B was measured described previosuly [6]. by the desulfation of 0.1 M paranitrophe nylsulfate in 0.25 M Tris­ acetat.e buffer at pH 8.2 (reagents from Sigma). These mi xtures were incubated at 37"C for 2 h, and the liberated para nit rophenol was RESULTS determined by the optical density reading a t 420 nm. Steroid sulfatase: The broke n cell homoge nates were homoge nized in Comparison of Lamellar Body Fractions Obtained by Various iced 0.014 M Tris-H CI buffe r, pH 8.0, using the Polytron PT-10 as Techniques above and enzy me assays were performed in duplicate on t he Miranol­ extracted fraction b y measurement of desulfation of l"H jdehydroe­ Of the various methods employed for the isolation of LB piandrosterone s ulfate (["H]DHEAS), as previously described [26]. only the selective, sequential fi ltration technique provided or­ Enzyme activity was expressed as picomo les of ["HJDHEAS converted ganelles resembling those found in vivo. The controlled forces to benzene-soluble ["H]DHEA/ h/ mg protein, or per total number of delivered by the cell disrupter resulted in broken cells with a mino acid residues after complete acid hydrolysis. minimal damage to the limiting membrane of the LB (Fig 1). Acid lipase: Acid lipase activity was assayed using 4-methy!-umbel­ However, since LB are highly labile, the homogenization me­ liferyl-oleate (MUO, S igma) as t he substrate according to the method dium required Ca++ -free, isotonic solutions, at slightly acid pH, of Knauer a nd Wegli cki [27]. Incubations we re carried out in t riplicate to prevent loss of LB contents. "C for 60 min, with 43- to 56-l'g sample protein in 0.5 ml a liquots at 37 Whereas the successive fi ltration steps did not appear to a nd terminated by the addition of 4.5 ml of 0.2 M potassium phosphate huller, pH 7.0. Fluorescence was measured spectrophotometrically at damage LB, they did remove a large proportion of undesired excitation and emission wavele ngths of :326 and 448 nm, respectively, cellular contents [19]. In addition to LB, glycogen was present, and results we re expressed as mM released oleate/mg protein/ h. as identified by: (a) resistance to RNAase digestion, (b) en­ Sp hin.gornye lina ~e: Sphingomyelinase activity was assessed b y the hanced contrast with osmium ferrocyanide fixation [7], and (c) method of Bowser and Gray [28]. The final incubation mixture (0.5 characteristic polyhedral structure. Although glycogen was not ml ) contained I 11Ci [melhyl-"C]sphingomyelin (New England Nuclear, present in LB preparations obtained with sucrose or metriza­ Boston, M assachusetts, 51 mCi/ mM) , T riton X-100 (0.1% vjv), calcium mide gradients, other organelles and membrane fragments per­ chloride (20 mM), and sample fractions (0.1 - 0.5 mg protein) in potas­ sisted in still greater abundance, and LB survived intact (not carried out for sium acetate butTer (60 mM) , pH 4.5. Incubations were illustrated). Other vesicular structures, presumably represent­ 1 h at 37"C, and terminated by addition of 0.5 ml water a nd 2 ml chloroform:methanol (2:1, v/v). Aliquots of the aqueous phase were ing fragments of plasma membrane and smooth endoplasmic counted in a liquid scintillation spectrometer, and results expressed as reticulum, also passed through the 0.4-llm filters (Fig 2a). % phosphorykholine liberated/mg protein/ h. Although most of these vesicular fragments could be removed Phospholipase A was assayed by a modification of the method of by subsequent passage through a 0.2 11m -pore filter, some LB Oct. 1985 LAMELLAR BODY-ENRICHED FRACTIONS FROM NEONATAL MICE 291 resulting in considerable variation in enzyme a ctivities from experiment to experiment. Despite s uch variations, certain enzymes were a lways c oncent rated in the LB fraction, while others were e ither presen t in diminished quantities or virt ually absent . In t hree experiments, t he LB:homogenate (H ) ratio of acid phosphatase ranged from 2.3-22.0. When acid phosphatase activity in the LB-conta ining f ract ion was compared to not only the homogenate, but a lso to t he pooled m ateria l released from all of t he fi lters a nd prefil ters, (homogenate minus LB fraction), t he activity was less t han half t hat in t he homogenate and one-sixt h t hat in t he LB fract ion, further emphasizing t he concentration of acid phosphatase in t he L B-containing frac­ tion. Lysosomal protease activity: Both carboxypept idase activity and cathepsin B-like activity also were c oncent rated in t his fraction (Table I) . In con trast, activity o f t he non-lysosoma l protease, plasminogen a ctivato r, appeared to be evenly distrib­ uted in t he homogenate and LB fraction (Table I) . The activities of acid l ip a.~e, phospholipase A, and sphingo­ myelinase a ll were c oncent rated significan tly in LB fractions (Table II). Since lysolecit hin was formed in smaller amounts than free fatty acids, it was not possible to determine t he relative activities of phospholipase A1 a nd A2• A la rge amount of sphingomyelinase activity was released into t he supernatant I ...... _.. , _...,..... during preparation for assays (T able II), presumably due to F I.G 1. Appea rance of homoge nate immedi ately after passage release of activity during freezing and thawing as noted in prior t hrough the cell di srupter. Abundant, large ly undama ged lamellar bod· studies, and suggesting t hat this e n zyme is concentrated even ies (arrows) li e in te rspersed among keratm fil ament bundles, mem· more in LB t han indicat bran e fr agments, keratohya lin -like materi al, and other orga nelles. (Bar ed. In cont rast to acid phosphatase and lysosomal proteases, = 1 J.LID ). t he TABLE I. Activity of proteases in granular cell fractio ns F ractions Enzyme Lamell a r LB:H ratio Homogenate body Ca rboxypeptidase (p M/ mg protein )" Exp 1 152.4 463.9 2.7 Exp 2 279 .1 762.6 2.7 Exp 3&-An gio l as substrate 3.1 4. 6 1.5 An gio II as substrate 1.6 3.6 2.2 Cathepsin 8-like activity (l

TABLE III. Phospholipid and neutral lipid composition of homogenate and lamellar body-enriched fractions (lipid wt %)

Experiment 1a Experiment 2"·' Experiment 3 Fraction Homogenate LB -enriched Homogenate LB-enriched Homogenate LB-enriched Phospholipids 31.6 40.6 36.0 44.5 30.6 35.0 Neutral li pids 58.1 43.8 63.5 54.7 36 .6 38.1 Sphingolipids 9.9 19.7 ND ND 33.1 26.6d Total 100.1 100.0 99.5 99.3 100.3 99.7 "Fractionation of phospholipids and neutral li pids from this ex periment are prese nted in Table IV. Lipid:protein ratios were 1.2 and 2.8 for homogenate and LB fraction, respectively. " Includes 0.5 and 1.9% cho lesterol sulfate in homogenate and LB fraction, respectivel y. ,. Ceramides are included in neutral lipid fraction in thi s experiment.. rl Fractionation from a comparable ex periment presented in Fig 3.

TABLE IV. Phospholipid and neutral Lipid content of neonatal mouse s H p s crude homogenate:; and Lamellar body preparations

Fraction a Lipid species Homogenate LB-enriched Phospholipids 36.0 ± 1.2 44.5 ± 2.1 Phospha t.idylethanolam i ne 6.8 ± 2.4 8.8 ± 0.6 P hosphatidylse ri ne/1no sitol 2.1 ± 1.5 2.6 ± 0.5 Phosphatidy lebo! in e 16.9 ± 1.6 19.7 ± 1.8 Lyso lec ithin/sphingomyelin 10.2 ± 1.1 13.6 ± 1.9 Neutral li pids'' 63 .5 ± 1. 1 54.7 ± 2.0 n-Alkanes 8.0 ± 0.4 4. 3 ± 0.8 Squalene 2. 1 ± 0.2 1.9 ± 0. 3 Sterol / wax esters 9.1 ± 0.04 5.1 ± 0.1 Free fatty acids 10.0 ± 0.08 8.0 ± 0.5 Diglyce rides' 1.9 ± 0.7 0.9 ± 0.3 Free sterols 20.9 ± 1.7 26.8 ± 1.3 Totals 99 .5% 99.3% Differences for total phospholipid · in LB vs homogenates are statis­ tica lly significant (p < 0.01); the differences of the free sterols in these 2 frac tions are also significant ( p < 0.05). "Results expressed as lipid weight percent. ± SEM of total neutral and phospholipids frac tionated in 3 runs o n sili ca-coated quartz rods (see Materials and Methods). 1' Sphingo lipids were not frac tionated by this technique (cf Fig 3). ' Presumpti ve identification by cochromatography against authentic .standard in one solvent system only. FIG 3. Thin -layer ch romatogram of lipids extracted from homoge­ nate (H) and 0.4-llm pellet (P), run in the sphingolipid solvent system activity of both my /sulfatase A and B and steroid sulfatase was (see Materials and Methods). When equal weights of extracted lipid are always lower in the LB fraction than in homogenates (LB:H applied to each lane, both the homogenate a nd pellet demonstrate (CER) , nonpolar neutral lipids (NL), and glyco­ ratio 0.31 for arylsulfatase a nd 0.11 for steroid sulfatase). In abundant ce ram ides sphingolipids ( GSL). However, the pell et appears to be preferentially addition to aryl sulfatase A and B, the activity of another enriched in GSL, including the most nonpolar species which comigrates lysosomal hydrolase, {:1-glucuronidase, also was absent in the with authentic acy lglucosylceramides (arrow) , isolated from pig epider­ LB fraction (data not shown). mis [31], whereas the homogenate apparently is enriched in CER. S = standard. Lipid Composit£on of Homogenates and Pellets Whereas the lipid:protein ratio of homogenates ranged from 1.0- 1.5, LB-enriched fractions ge nerally contained about 3 mice [7] (Fig 3). All of the glycolipid fractions appeared to be times more lipid than protein (ratio can be used to extrapolate more prominent in LB than in homogenates, including the the data on lipid weight percent to lipid/mg protein where most nonpolar species, which cochromatographed with authen­ appropriate). Despite the overall enrichment of lipids in LB­ tic acylglucosylceramides. containing fractions, t he homogenate contained a greater pro­ portion of neutral lipids t han did the pellet (Table III). In DISCUSSION contrast, the LB-enriched fraction consistently demonstrated The unusual difficulties encountered in obtaining enriched about a 20-30% enhancement in all phospholipid species (Table fractions of LB can be attributed: first, to the unusual resistance IV) . Although the homogenate was enriched in most neutral to homogenization of the outer layers of the epidermis (pres­ lipid species, free sterols were consistently present in greater sures in excess of 5,000 p .s.i. are required to completely disrupt quantities in the LB-enriched preparation than in the homog­ granular cells [19]; second, these pressures potentially can enate (Table IV). Whereas n-alkanes were present in signifi­ damage LB; third, the LB is often enmeshed in aggregated cant quantities in both the homogenate a nd LB-enriched frac­ keratin filaments further complicating its isolation in purified tion (Table IV) , cholesterol sulfate was present in only trace form. Fourth, LB contents are rapidly secreted or lost with the quantities in both fractions (data not shown). slightest change in milieu, e.g., LB disappear from short-term Whereas ceramides (CER) were concentrated in the homog­ organ cultures (Elias, unpublished observations), and it resides enate, as were other n eutral lipids, glycosphingolipids (GSL) in tissues enriched in hydrolytic enzyme activity (e.g., [15, were concentrated in the LB-enriched fraction. Whereas the 16]), with divalent cations lying free in the cytosol [34]. Finally, CER-GSL ratio was 3.3 in homogenated, it was 1.3 in pellets LB may be primed to fuse with the plasma membrane due to (mean of 2 experiments), and revealed the same major species lysolecithin generated by increased phospholipase A activity previously identified in the stratum granulosum of neonatal (29]. Oct. 1985 LAMELLAR BODY-ENRICHED FRACTIONS FROM NEONATAL MICE 293 The isolation procedure described here circumvents many of Lipid Content of Epidermal Lamellar Bodies t hese potential problems ~1 9 ] . First,_ t ~ e _p~ri~ed sta phy~ococ~a l Al though it is possible t hat the apparent enrichment of epidermolytic toxin fractwn_s perm1~ m1 t1_atw_n of st~d1~ s w1th certain lipids in LB could be attributed to the presence of an outer epidermal prepa~at wn obtamed m V IVO, obv1atmg t?e contaminating material in the LB-enriched pellets, its lipid usual antecedent incubatiO n step. Next, the Stansted cell d!s­ content is consistent with its putative role as t he source of rupte~ can b e regulated to pressures required to just rup~ure permeability barrier lipids. First, t he LB appears to be lipid­ most granular cells, and unbroken ~ e ll s can be ~ehomo~emzed_ enriched, containing at least 2-fold more lipid per unit weight to augment y ields still further. Fmally, desp1te surv1val of than h omogenates. Second, LB are enriched in phospholipids substant ial LB t hrough t he h omoge ni~at i o n _step, most L~ a~e and free sterols. Phospholipid catabolic products (free fatty lost during subsequent lengthy ce nt nfugatwn steps. Th1s IS ac ids) and free sterols are major constituents of stratum cor­ avoided by t he nuclepore fil ter technique, which r equires only neum intercellular lipids [21 ]. Third, confirming a recent report 15-20 min, rather t han hours [19]. Other factors t hat may have in fetal rats [18], LB contain a broad array of glycosphingo­ contributed to t he relative success of this protocol include: (a) lipids. Although we did not quantitate individual glycosylcer­ maintenance of samples in iced media at all times, and (b) amides and acylglucosylceramides, t hese fractions appeared to inclusion of 10 mM EDTA in calcium- and magnesium-free be concentrated in LB fractions, consistent wit h recent work buffers and media throughout t he isolation scheme. by Wertz et a! on fetal rat LB isolates (18]. Both t he study Neonatal mouse LB are ellipsoidal organelles whose long axis cited above and this study provide strong support for the measures a little over 0.2 Jlm , but whose short axis is approxi­ hypothesis t hat LB are a rich repository of glycosphingolipids, mately 0.15 Jlm . Although the best LB morphology was ob­ and t hat delivery of these substances to intercellular domains served in t he 0.4 Jlm-pore filtrate, additional filtration through may be important for barrier formation (reviewed in [11 ,12]). 0.2 ,u.m excluded substantial, a dditional m embrane contami­ However, it s hould be noted t hat glycosphingolipids are not nants, but at some cost to morphology and total yield. yet, it present in t he outer stratum corneum (39], despite the acknowl­ must be emphasized that even the 0.2-Jlm filtrate d1d not edged barrier capabilities of t hese layers (references cited in represent a purified prep_aration. In fact, o n~y about 20%_ ofthe [12]). Presumably, loss of glycosphingolipids results from t he material in the 0.4-Jlm filtrate can be considered to be mtact, rich content of glycosidases in the outer epidermis (40] and unequivocal LB. Therefore,. the bio~h e mi_ca l studies described lamellar bodies [37], which would convert glycosphingolipids here must be interpreted w1th cautiOn, smce other uncharac­ into ceramides, analogous to the degradation of phospholipids terized cellular components are present in substantial quan­ into free fatty acids and glycerides. tities and even might be copurified with LB. Attempts to purify t he fi~a l LB-containing filtrates further by either enzyme diges­ We appreciate the s uperb assistance of D r. Mary L. Williams, tion techniques or by a subsequent density gradient technique Barbara E. Brown, Carol E. Kaempfer, J eanette Bonifas, and Deborah were not s uccessful (S. Grayson and P . M. Elias, unpublished Martinez. observations). REFERENCES Enzyme Content of LB-Enriched Preparations 1. Odl and GF, Holbrook K: The lamellar granules of epidermis. Curr Although t he number of enzymes surveyed was limited, these Probl Dermatol 9:29- 49, 1981 2. Frithiof L: Ultrastructural changes in t he plasma membrane in studies have demonstrated the presence of several, "classic " human oral epit helium. J Ult rastruct Res 32: 1- 17, 1970 lysosomal enzymes in LB-enriched preparations. We have con­ 3. H ayward AF: Ultrastructural changes in contents of membrane­ firmed Freinkel and Traczyk's demonstration of substantial coating granules after extrusion from epitheli al cell s of hamster enrichment of acid phosphatase [15]. In cheek pouch. Cell Tissue Res 187:323-331, 1978 addition, the LB 4. E lias PM, Friend DS: The permeability barrier in mammalian fraction was e nriched in proteases and a broad spectrum of epidermis. J Cell Bioi 65:180- 191, 1975 lipase activities. In preliminary cytochemical studies, we also 5. Lavker RM: Membrane coating granules: the fate of t he discharged have locali zed both acid and neutral lipase activity to LB [35], lamellae. J Ult rastruct Res 55:79- 86, 1976 further substantiating these biochemical observations. Perhaps 6. Elias PM, McNutt NS, Friend D: Membrane alterations during co rnification of mammalian squamous epi thelia: a freeze-frac· as striking as the presence of certain lysosomal enzymes was ture, tracer and thin section study. Anat Rec 189:577-593, 1977 the paucity of others, i.e., ,6-glucuronidase and aryl sulfatase A 7. Elias PM, Brown BE, Fritsch PO, Goe rke RJ, Gray GM, White and B in the LB fraction. Finally, the observation that both RJ: Locali zation and composition of lipids in neonatal mouse plasminogen activator a stratum granulosum and stratum corneum. J Invest Dermatol nd steroid sulfatase, eac~ of which is a 73:339-348, 1979 membrane-associated, but nonlysosomal, catabolic enzy me, are 8. Hayward AF, Hackemann M: Elect ron microscopy of membrane­ absent furt her underscores the localization of acid phosphatase, coating granules and a cell -surface coat in keratinized and non­ lysosomal proteases, and lipases to LB. keratinized human oral epi thelium. J Ultrastruct Res 43:205- Although t he function of certain lysosomal enzymes in LB is 219, 1973 9. E li as PM, Brown BE: The mammalian cutaneous permeabi lity unknown, these enzymes may participate in desquamation by barrier: defective barri er function in essential fatty acid defi­ attracting certain intercellular constituents, or alternatively ciency co rrelates wit h abnormal intercellular lip id composition. t hey may be atavistic, nonfunctional residues linking t he LB Lab Invest 39:574- 583, 1978 to a lysosomal past [15], or they may be copackaged coinciden­ 10. Wertz PW, Cho ES, Downing DT: Effect of essent ial fatty acid defi ciency on t he e pidermal sphingo lipids o f the rat. Biochim tally with secretory products during passage t hrough t he en­ Biophys Acta 753:350-355, 1983 doplasmic reticulum or Golgi apparatus (e.g., [36]). 11. Wertz PW, Downing DT: Glyco lipids in mammalian epidermis: LB lipases and glycosidases [37] presumably mediate t he locali zation and relationship to cornifi.cation. Science 217:1261- remodeli ng of polar lipid substra 1262, 1982 tes (e. g., phosphatidylcholine, 12. E li as PM. Epidermal li pids, barrier function, and desquamation. J sphin gomyelin glycosphigolipids) into more hydrophobic prod­ Invest Dermatol 80:44-49, 1983 ucts (i.e., ceramides, free fatty acids). Such a view is consistent 13. Wolff K, Holubar K: Odland-Korper (membrane coating granules with t he dramatic alterations in t he morphology of intercellular Kerat inosomen) als epidermale Lysosomen. Arch Klin Exp Der ~ lipids that occurs during t ransition matol 23 1:1 - 19, 1967 from t he granular to t he 14. Gonzales LF, K ranwczyk WS, Wilgram GF: Ult rastructural obser· corn ified layer ([4,6] and see below). Whereas one fu nction for vations on t he enzymatic activity of keratinosomes. J Ultrast.ruct phospholipase A in LB might be the conversion of polar lipids Res 55:203- 211 , 1976 to more hydrophobic products, a second function might be the l". Freinkel RK,T raczyk TN: Acid hydrolases of the epidermis: sub· generation of lysolecithin, which, through i ts fu sogenic prop­ cellular l ocali zation and relationship to cornification. J Invest Dermatol 80:441- 446, 1983 erties (reviewed in [38]) , could mediate the fu sion and subse­ 16. Wolff-Sc hreiner EC: T he ul trastructural cytochemistry of t he ep· quent secretion of LB contents into the intercellular spaces. idermis. Int J Dermatol 16:77- 102, 1977 294 GRAYSON ET AL Vol. 85, No.4

17. Freinkel RK, Traczyk TN: A method for p art ial purification of dermis. J Invest Dermatol 70:331-335,1978 lamell ar granules from fetal rat epidermis. J Invest Dermatol 29 . Freinkel RK, Traczyk TN: The phospholipases A of epidermis. J 77:478- 482, 1981 Invest Dermatol 74:169- 173, 1980 18. Wertz PW, Downing PT, Freinkel RJ<, Traczyk TN: Sphingolipids 30. Bradford MM: A rapid and se nsitive method for t he quantitation of the stratum co rneum and lamell ar granul es of fetal rat epi ­ of microgram quantities of protein utilizing t he principle of dermis. J Invest Dermatol 83:193- 195, 1984 protein-dye binding. Anal B iochem 72:248- 254, 1976 19. Grayson S, Johnson-Winegar AD, E lias PM: Iso lation of lamell ar 31. Gray GM, White RJ, Majer RJ: 1-(3 -0'acyl)-{j-glucosyi-N-dihy­ bodies from neonatal mouse epidermis by selective sequential droxy-pentatriacontradienoylsphingosine, a major component of fil t ration. Science 221:962- 964, 1983 the glycosylceramides of pig and human epidermis. Biochim 20 ..J ohnson AD, Spero L, Cades JS, De C icco BT: Purification and Biophys Acta 528:127- 137, 1978 characteri zation or different types or exfoli ative tox in from 32. Brown BE, Williams ML, E li as PM: Stratum corneum lipid ab­ Staphylococcus au.reu.s. Infect lmmun 24:679-684, 1979 normali ties in ichthyosis: detection by a new lipid mi croanalyt­ 21. Grayson S, E li as PM: Isolati on and lipid biochemical characteri­ ical method. Arch Dermatol 120:204.-209, 1984 zation of stratum corneum ce ll membrane complexes: implica­ 33. Karnovsky MJ: Use of ferrocyanide- reduced osmium tetroxide in tions for the cutaneous permeabili ty barrier. J Invest Dermatol electron microscopy. Proceedin gs of the 14th M eetin g o f the 78:128- 135, 1982 Am erican Society of Cell Biology, 1974, Book of Abstracts, p 146 22. Win1 roub BU, Klickstein LB, Kaempfer CE, Austen KF: A human 34. Menon GK, Grayson S, Elias PM: Visuali zation of i onic calcium neutrophil -dependent pathway for ge neration of angiotensin ll; reservoirs in mammali an epidermis: locali zation by ion-capture purification and physicochemical characteri zation of the plasma cytochemistry. J Invest Dermatol 84:508-512, 1985 protein substrate. Proc Nat! Acad Sci USA 78: 1204- 1208, 1981 35. Menon GK, Grayson S, Elias PM: Cytochemi cal and biochemical 23. Barret AJ: F luorimetric assays of cathepsin B and cathepsin H localization of acid and neutral lipase activity in epidermis. Clin with methylcoumaryl amide substrates. Biochem J 187:902-912, Res 33:156A, 1985 1980 36. Oliver C: Cytochemical l ocalization of acid phosphatase and tri­ 24. Unkeless JC, Tobia A, Ossowski L, Quigley JP, Rifkin DB, Reich metaphosphatase activities in exocrine acinar cell s. J Histochem E: An enzymatic fun ction associated wi t h transformation of Cytochem 28:78- 81, 1980 fibrobl asts by oncogenic viruses. I. Chick embryo fibroblast 37. Freinkel RK: The lipids a nd acid hydrolases of lamell ar granules. cultures transformed by avian RNA tumor viruses. J Exp Med Clin Res 30:584A, 1982 137:85-111,1973 38. Lucy JA: Mechanisms of chemi call y induced ce ll fu sion, Membrane 25. Deutsc h DG, Mertz ET: Plasminogen: purification from human Fusion, vol 5. Edited b y G Poste, GL Nicolson. Amsterdam, plasma by a ffinity chromatography. Science 170:1095-1096, 1979 North-Holl and, 1978, pp 268-304 26. E pstein EJ Jr, Leventhal ME: Steroid sulfatase of human leuko­ 39. Lampe MA, Williams ML, E li as PM: Human epidermal lipids: cytes and epidermis a nd the diagnosis of r ecessive X-linked characteri zation and modulations during differentiation. J Lipid ichthyos is. J Clin Invest 68:1404- 1410, 1981 Res 24:131- 140, 1983 27. Knauer TE, Wegli cki WB: Characteri stics of multiple forms of the 40. Nemanic MK, Whitehead JS, Elias PM: Alterations in membrane acidic t ri acylglycerol lipase(s) of canine cardiac myocytes. suga rs during epidermal differentiation: visualization with lee ­ Biochim Biophys Acta 753: 173- 185, 1983 tins and r ole of glycosidases. J Histochem Cytochem 31:887-897, 28. Bowser PA, Gray PM: Sphingomyelinase in pig and human epi- 1983