Differences Between Connective Tissue-Epithelial Junctions in Human Skin and the Anagen Hair Follicle

Differences Betw^een Connective Tissue-Epithelial Junctions in Human Skin and the Anagen Hair Follicle Michael Nutbrown and Valerie A. Randall Department of Bionicdical Sciences, University of Bradford, Bradford, U.K.

Although the ultrastructure of the dertnal-epidermal junction has been well characterized, little is known about the junctions between the dermal papilla and the surrounding epithelial cells of the hair bulb, or between the connective tissues and the epithelial cells on the outside of the hair follicle. Because the dermal papilla plays a major role in controlling the hair follicle, w^e also examined the ultrastructure of the potentially important dermal papilla-epithelial junction in normal scalp anagen follicles. The dermal-epidermal junction in skin was a trilaminar basement membrane characterized by the anchoring points of hemidesmosomes and tonofilaments in keratinocytes. In the hair follicle, the junction that separated the dermal papilla and epithelial cells was a trilaminar basement membrane, but relatively few putative anchoring points were seen. These were

T

he structure and ultrastructure of the dermoepidermal junction (DEJ) are well documented [1-4]; its composition and formation [5,6] and the roles of hemidesmosome attachment points and the various anchoring mechanisms in skin arc clearly understood [7,8]. In contrast, most descriptions of the ultrastructure of hair follicles have concentrated on the appearance of the connective tissue and epithelial cells, their topographic organization with respect to each other, and the changes in appearance due to development and differentiation [9-12]. The junctions between the connective and epithelial tissues in the human hair follicle have been neglected, except by Puccinelli et al [13], probably because the important role of the connective tissues in the regulation of hair growth was not appreciated fully at the time. The terms basement membrane and basal lamina are both used to describe the thin layer intervening between epithelial and connective tissues. Confusion exists because many biologists have used "basal lamina" to describe the structure that is now called the lamina densa [14]. The terms recommended by the International Anatomical Nomenclature Committee [15] have been adopted Manuscript received December 22, 1993,- revised July 8, 1994; accepted for publication August 12, 1994. A preliminary report of this study was presented at the New York Academy of Sciences, January 1991, Arhngton, VA. Reprint requests to Dr. M. Nutbrown, Department of Biomedical Science.s University of Bradford, Bradford BD7 lDP, UK. Abbreviations: CTEJ, connective tissuc-cpithelial junction; DPEJ, dermal papilla-epithelial junction. ,

similar to modified dermal-melanocyte junctions, in which the intercellular cytoplasmic filaments do not come together at an attachment plaque, the laminar components tend to be thinner, and the anchoring fibrils beneath the lamina densa are fewer. A trilaminar membrane also was interposed between the connective and epithelial tissues on the outside of the follicle, but nothing that resembled a hemidesmosome or any other type of anchoring structure was seen. The difference in structure of the junctional complex between skin and hair follicles probably reflects the relatively permanent state of the epidermis, compared to the dynamic processes involved during the anagen phase of the hair follicle. Key words: dermal papilla/ultrastructure. J Invest Dermatol 104: 90-94, 1995

here, i.e., basement membrane for light microscopy and basal lamina at the electron-microscope level [16]. ln transverse sections, the DEJ is seen as a basement membrane that follows an irregular, undulating line. Briggaman and Wheeler [1] described at the ultrastructural level the details of the junction, hemidesmosomes, and modified attachment points associated with melanocytes. The ultrastructure of the junctional zone between the dermal papilla and epithelial components (DPEJ) in mouse and rat hair follicles during early and late catagen also has been examined [17—19]. An additional component, the liyalin membrane, was described on the outside of the hair follicle [17]. This is an outer layer of the follicle, adjacent to the basal lainina, which is seen as two layers of orthogonally arranged collagen fibers, with the inner layer parallel to the long axis of the hair. Puccinelli et al [13] did not describe this hyalin membrane in human follicles. Because the dermal papilla is now known to play a major role in influencing the form and dynamics of the hair follicle [20,21], which probably involves regulatory substances crossing the basement membrane [22], we investigated the ultrastructure of connective tissue-epithelial junctions (CTEJ) on the outside of the hair bulb and around the dermal papilla of normal anagen hair follicles. Samples of post-auricular skin were processed as controls. MATERIALS AND METHODS Collection of Skin Samples Samples of normal human post-auricular skin were obtained from routine surgical excisions of benign lesions and plastic surgery. Samples were immersed immediately in a solution of 3% glutaraldehyde in 0.1 M Sorensen phosphate buffer at pH 7.2, cut into approximately 2-nini cubes, and transported to the lahoratory in the

0()22-202X/95/S09.50 • SSDI0022-202X(94)()0239-4 • Copyright © 1995 by The Society for Investigative Dermatology, Inc. 90

VOL. 104, NO. 1 JANUARY 1995

ULTRASTRUCTURE OF HAIR FOLLICLE JUNCTIONS

91

Figure 1. Detail of DEJ in post-auricular human skin. Tliis view shows the undulating basal lamina (BL) containing a series of well-defined hemidesmosomes joining the epidermis (above) to the demiis (below). Attachment plaques (AP) and tonofilaments (T) fomi the epithelial side of the hemidesmosome, with a subbasal dense plaque (SBDP) contained within the lamina lucida (LL). Anchoring fibrils (AF) and collagen fibers (C) can be seen subjacent to the lamina densa (LD) in the extracellular matrix of the dermis. Bar, 1 /nm.

glutaraldehyde in a sterile container. Scalp biopsy specimens were taken from the occipitoparietal region of normal scalps from six volunteers with informed consent and Ethical Committee approval at the Dermatology Department, Leeds General Infirmary. Punch biopsy specimens of 4 mm in diameter were taken under 1% Lignocaine local anesthesia and immersed immediately in 3% glutaraldehyde in 0.1 M Sorensen phosphate buffer to check autolytic changes during transport to the laboratory.

Processing of Skin Samples and Individual Hair Follicles Skin tissue was dissected further to yield 1-mni cubes, which exposed the subcutaneous DEJ. After trimming, the samples were processed for transmission electron microscopy within 4 h of excision. Scalp specimens were transferred to S.S-mm petd dishes in glutaraldehyde solution, where anagen hair follicles were microdissected from each specimen and transferred to a second dish. The follicles were cleaned of surrounding tissue and lipid, but the connective tissue sheath was left intact. Primary fixation of the dissected follicles was continued in similar 3% glutaraldehyde for a total time of 3-4 h, including transport time. After the first fixation, subseqnent processing was done in a Lynx Microscopy Tissue Processor (Leica UK Ltd.). All tissues were fixed routinely in glutaraldehyde and osmium tetroxide and embedded in Araldite CY 212 epoxy resin for electron microscopy.

Selection and Examination of Ultrathin Sections Three control samples of post-auricular skin from different donors were sectioned on a Reichert Ultracut D ultramicrotome. Silver to pale gold ultra-thin sections, i.e., about 80 nm in thickness, were examined in two blocks from one source of skin and in one from each of the others. Sections were stained with uranyl acetate and lead citrate [23] and then examined in either a JEOL 10OS or lOOCX transmission electron microscope. Because the DEJ is a continuous and ultrastructurally stable environment, only two or three sites were selected in each of the sections examined; these were considered representative of the junction as a whole. Hair follicles were much more difficult to section than skin samples hecause of the important of orienting the follicle correctly to section in the direction of hair growth and through the center of the follicle. Excess resin was trimmed carefully fi-oni the block to leave the follicle exposed on the top of a broad-based pyramid of resin. Semi-thin survey sections (thickness 350 nm) taken at frequent intervals were stained with toluidine hlue [24] and examined hy light microscopy. We adjusted the orientation of the follicle repeatedly until it was aligned, ideally with its long axis parallel to the Z axis of the microtome and perpendicular to the X and Y planes. This ideal was not always achieved because folHcIes are rarely regular in shape; if any follicle was twisted or kinked in any way, we exposed as much of the bulb and lower part of the folhcle as possible for sectioning. Longitudinal ultra-thin section.s of the lower follicle, up to about three times papilla height, were stained and examined as descrihed above. Only sections cut through the median planes of the follicle were examined. Eleven follicles from the six biopsy specimens were selected for study; useful ultra-tliin sections were cut successfully from eight. The CTEJ was examined in detail at six or seven sites in each follicle, including the apex, approximate mid-point, and base of the dermal papilla, and three corresponding sites on the outside of the follicle; if possible, an additional site was examined at about two times papilla height. These sites are marked on Fig

RESULTS The Skill DEJ The human skin DEJ resembled that described in other studies [1-3]. Hemidesmosomes were seen lining the plasma membranes of the basal keratinocytes (Fig 1), with tonofilaments

coursing back from the hemidesmosome attachment plaques into the cytoplasm of the basal cells. The lamina lucida and lamina densa were clearly seen as pale and darkly stained bands, respectively, that faithfully followed the line of the basal cell membranes. The extracellular space of the dermis was mostly filled with transversecut collagen bundles interleaved with the processes of connective tissue cells. Fine structural detail of the junctional zone and connective tissue is shown in Fig 1. The lamina lticida was evident though not always very well defined, but in the lucent space, sub-basal dense plaques were seen associated with the attachment plaques. Tonofilament connection to the attachment plaques was very marked. A slight increase in staining density of the lamina densa was evident at the points where the membrane passed beneath a sub-basal dense plaque and an attachment plaque. Anchoring fibrils were drawn out from their connection with the lamina densa at one end and inserted into the extracellular matrix of the dermis at the other; some appeared to be in the form of a loop (Fig 1). In the demiis, collagen bundles and cytoplasmic processes were interspersed with strips of longitudinally cut collagen, which displayed characteristic periodic banding. Anchoring filaments, although not readily obvious in Fig 1, could be discerned at higher magnification, but neither dermal miciofibril bundles nor anchoring plaques associated with anchoring fibrils were seen in any of the sections examined. The CTEJ in the Normal Hair Follicle At the light-microscope level, the basement membrane of the hair follicle appeared as an amorphous, pale-staining or fine fibiillar band (Fig 2). It formed a continuous layer enclosing the dermal papilla, where it formed an tuibroken barrier, and continued arotmd the outer parts of the follicle between the connective tissue sheath and the outer root sheath. Ultrastructural examination revealed a high level of organization similar to that of the DEJ. Outside of the Hair Follicle Electron microscopy of the CTEJ confirmed the smoothness of the membrane between the outer root sheath and the connective tissue sheath. At medium magnification, the CTEJ appeared as a three-layered structure (]>ositioti 1; Fig 3a), which was less undulatory than the similar structure seen in skin. The plasma membrane of cells of the outer root sheath was very well defined and was invaginated in several places by infoldings into the cell cytoplasm. The lamina lucida was a narrow but easily seen clear space between the plasma membrane and the lamina densa, with some particulate content comparable in appearance and quantity to that seen in skin. The lamina densa was a smooth band of uniform thickness which at high magnifications had a fine granular appearance. Subjacent to the lamina densa were the orthogonal layers of collagen fibers, which make up the hyalin membrane. There was no evidence of anything resembling a hemidesmosome and no tonofilaments reaching back into the cytoplasm of cells of the outer root sheath. Located at irregular intervals along the membranes were small clouds of particulate matter (Fig 3a).

92

NUTUROWN AND RANDALL

Figure 2. Section through bulb and lower part of human scalp anagen hair follicle. The dermal papilla (DP) and epithelial ti.ssues (Epi) arc separated by a basement membrane (BM), which also encloses the entire follicle. The numbers refer to positions examined in detail in the external connective tissue and DPEJ, described in the text and referred to in Fig ia-f. Bar, 50 jxm.

These clouds difFused into the cytoplasm of the epithelial cells for about 2 ix.m and appeared to stretch across the basal lamina and into the layer of longitudinally cut collagen that surrounded the follicle. At this upper region of the bulh, the orthogonal arrangement of the collagen fiber layers was particularly obvious, with the inner layer of longitudinal fibers lying parallel to the hair follicle. Similar features were seen in the CTEJ in the mid-bulb (positiott 2; Fig 36). Invaginations were more abundant and were more complex than a simple infolding of the plasma membrane; they produced very long processes that terminated far back into the cytoplasm of cells of the outer root sheath. There were some thickenings of particulate matter in the lamina densa accompanied by increases in density in the lamina lucida, but not always complemented by similar defusions in the epithehal cytoplasm, as described earlier (Fig 3a). The junctional zone was well defined, but not as well as at the level of the papilla (positiott 6). The collagen layers were less well defined; the fibers were undulatory with a variable appearance and the circular fibers were irregularly distributed. At the outer lower part of the lobe (positiott 3; Fig 3c), the two-layer collagen system ceased to exist as such, although the circular fibers still were seen to run at right angles to the axis of the follicle. Invaginations and infoldings of the plasma membrane still were observed and the lucent space in the basement membrane was obvious, but the lamina densa was very thin and difficult to discern. The DPEJ The DPEJ (Fig id-j) had a generally le.ss smooth contour than the similar junction between the outer root sheath and the connective tissue sheath, but it did not undulate as much as in the skin DEJ (compare with Figs 1, 2, 3a). The matrix cells that surrounded the dermal papilla were mainly keratinocytes and melanocytes. The means of attaching these cells to the basal lamina was similar to that seen in the dermal-melanocyte junction in skin. The structures (Fig ie) included the dense plate, similar in appearance to the hemidesmosome attachment plaque, a clear lamina lucida, and a lamina densa that was not well defined but showed some thickenings in density below the dense plates. There was no evidence of anchoring filaments passing through the lamina lucida. A true hemidesmosome structure was not seen in any of 34 ultra-thin sections of normal hair follicles containing a good view of the dermal papilla (compare with Figs 1, 3
THE JOURNAL OF INVESTIGATIVE DEIUVIATOLOGY

Near the base of the inner part of the lobe {position 4; Fig 3rf), the orthogonal arrangement of the collagen layers was broken almost completely, though the fibers tended to lie parallel to the lamina densa. The orthogonal arrangement of fibers seen on the outside of the follicle was never observed within the dermal papilla. The plasma membranes of the basal epithelial cells had a few thickenings possibly associated with attachment points to the basal lamina. The lamina densa appeared to follow the line of the epithelial plasma membrane, though there were several widenings of the lamina lucida. The lamina lucida was mostly filled with a substance similar in staining density and appearance to the cytoplasm of the epithelial cells. No invaginations similar to those seen on the outside of the follicle were seen in this area. The interface between the epithelium and connective tissue, at mid-papilla height (positioti 5), was well defined hy a slightly undulating line, which had few gaps or invaginations (Fig 3t'). There were some slight thickenings on the plasma membrane, but these were not convincing as attachment plaques. The lamina lucida clearly had an evenly distributed, fine granular content (Fig 3c). The lamina densa was seen as a very thin, well-defined line, which paralleled the course of the epithelial cell plasma membranes. Collagen fibers were distributed intermittently along the length of the junction. The fibers amassed nearest to the lamina densa seemed to be preferentially end-cut. At the top of the papilla (position 6; Fig 3/), the plasm-a membranes displayed several gaps and infoldings similar to those on the outside of the hair follicle. Thickenings of the plasma membrane and associated increases in density of the lamina densa were seen at several points along the junction (Fig 3 ^ . The lamina lucida contained an evenly distributed, fine-grained inclusion, and the lamina densa was continuous though of variable density. Collagen fibers were collected in irregularly sized clumps along the edges of the lamina densa. Anchoring fibril-like structures were attached at one end to the lamina densa of the junction and at the other end inserted into the extracellular matrix of dermal papilla (Fig 1). Concentrations of fine fibrillar material, similar to dermal microfibril bundles [IJ, also were seen in the dermal papilla extracellular matrix. Table I summarizes the observations described in the Resttlts. DISCUSSION The ultrastructure of the CTEJ has heen examined here in two main areas of the human scalp anagen follicle: on the outside between the connective tissue sheath and outer follicular tissues from the base of the follicle up to about twice the dermal papilla height, and on the inside between the dermal papilla and the epithelial tissues of the hair bulb. These junctions are the CTEJ and the DPEJ, respectively. Because our parallel control samples of skin clearly showed the normal structures reported for the human DEJ [1], our processing technique seems valid. The follicular junctions were similar to those seen in the skin; on the outside the junction appeared as a smooth line between the epithelial and connective tissue components, whereas where the junction divides the dermal papilla from the epithelial tissue, it was undulatory, though less so than in the skin. The basement membrane was resolved into the three layers of peripheral cell membranes: lamina lucida, lamina densa, and the additional layers of the hyalin memhrane reported in mouse follicles [1 7]. The mechanical connection hetween the hair foHicle and the connective tissue components was much weaker than that hetween the corresponding components in skin. Nothing was seen that corresponded to the skin hemidesmosome with tonofilaments on either the external connective tissue or dermal papilla junctions. However, the appearance of particulate clouds of matter (Fig 3c) in the regions of the lamina densa and adjacent epithelium on the outside of the follicle were reminiscent of the thickening densities reported around and below attachment plaques in the dermalmelanocyte junction []]. The increases in tissue density may be a similar kind of loose anchoring structure. Putative anchoring structures similar to those described in the skin dermal-melanocyte

VOL. 104, NO. 1 JANUARY 1995

ULTRASTRUCTURE OF HAIR FOLLICLE JUNCTIONS

93

Figure 3. Details o f the CTEJ on the outside of the follicle (a-c) and the DPEJ (d-J). See text and Fig 2 for definitions of positions 1-6. a: The junction (position I) hetween the outer root sheath (top) and the connective tissue sheath (below) shows a three-layered structure comprising the basal-cell plasma membrane (PM), lamina lucida, and lamina densa (LD). Below the junction are two orthogonal layers of longitudinal (Cl) and transverse (Ct) collagen fibers, hivaginations (hi) are seen in this part of the junction. Small clouds of particulate matter are diffused in the epithelial cell cytoplasm and across the basal lamina into the connective tissue layers, b: CTEJ at mid-hulb (position 2) between the hair matrix (above) and the connective tissue sheath (below). There are pronounced invaginations (In) in the basal-cell plasma membranes. Longitndinal (Cl) and transverse (Ct) collagen fibers persist in orthogonal layers but are not as well defined as higher in the follicle, hifnsions of particulate matter (PC) are present in the lamina densa and connective tissue regions, hut these do not persist into the epithelial tissues, r; CTEJ in the lower part of the hulh (j)ositioti 3) hetween the division zone above and the connective tissue sheath. The trilaniinar arrangement of the junction (CTHJ) is still evident and there are invaginations (In) in the plasma memhranes, but the ordered composition of the collagen layers (C) has broken down, d: DPEJ at the base of the papilla (jiosition 4). The two-layer orthogonal arrangement of collagen fibers (C), as seen on the outside of the follicle, is no longer evident. The junction is generally well defined, though the basal-cell plasma memhranes (PM) and lamina densa (LD) are slightly diffuse, e: DPEJ at niid-papilla height (position 5) has a very well-defined interface with few gaps or invaginations. Slight thickenings (A) of the plasma membranes on the hair-matrix side are not convincing as attachment plaques. Collagen fibers (C) in the dennal papilla are much sparser than in lower regions of the follicle. /.' Putative anchoring system in the DPEJ near the apex of the papilla (jtosition 6). The presumptive cortex is seen above, with the dermal papilla below. This anchoring system, which has a few similarities to a hemidesmosome, comprises a dense plate or attachment plaque (D) and a complementary thickening of the subjacent lamina densa only. Anchoring fibrils (AF) protrude into the extracellular matrix (ECM) of the dermal papilla, and niicrofihril bundles (MB) are interconnected with collagen fibers. Bins, 0.4 ;Lim {a,b,dj'); 0.8 /xni (c,e).

junction also were seen around the top of the dermal papilla (Fig 3 ^ . A similar absence of substantial anchoring components can be seen in a single micrograph of a human dermal papilla (13], though Puccinelli ct al made no comment about tliis in their paper. Hemidesmosomes were reported as absent also from the rat hair matrix basement membrane [25]. It is interesting that buUous pemphigoid antigen, which is found in association with hemidesmosomes in the DEJ [26], was not present in the DPEJ of human anagen follicles [27] or in developing hair bulbs in mouse and rat skin [28,29]. Epithelial matrix cells surrounding the dermal papilla appear to be joined only loosely to the connective-tissue components, because it is nearly impossible to obtain dermal papillae by plucking anagen hairs [30]. The lack of definite anchoring mechanisms in the hair-follicle junctions may allow flexibility of movement while the cells undergo the dynamic processes of the growth cycle. The upward progression of the matrix epithelial cells [31] and the proposed downward movement of cells of the outer root sheath during anagen [32] would be highly restricted if they were tightly bound at the epithelial-connective tissue junctions. W e observed invaginations of the plasma membranes of cells of the outer root sheath adjacent to the CTEJ (Fig 3h,c), which apparently were not described before for the human anagen hair

Table I.

Characteristics of the CTEJ in Human Anagen Hair Follicles and the DEJ Position in Hair Follicle Outside Wall

Characteristic Undulation of basement membrane Hyalin memhrane Invaginations of plasma memhrane Hemidesmosome Sub-basal dense plaque Dense plate/attachment plaque hicreased staining of lamina densa Anchoring fibrils Anchoring filaments Tonofilaments

1

Dermal Papilla DEJ

94

follicle. Parakkal [17] described the convolution of the hyalin membrane of the rat catagen follicle as causing invaginations of adjacent follicular epithelial membranes, but this is not what has happened here. Because the invaginations in the hair follicle appeared to be pinocytotic and vesicles were often concentrated in tiie adjacent areas, this may be a mechanism for the uptake (or secretion) of material. Similar pinocytotic vesicles are seen in the plasma membranes of the basal epithelial cells of skin [1]. Near the top of the dermal papilla, there were several gaps in the basal lamina and infoldings of the plasma membranes of the basal cells. These gaps or invaginations were similar to those described on the outside of the follicle and may serve a similar purpose, but little vesicular presence was noted. The differences in attachment points between the connective and epithelial tissues are obvious morphologic characteristics that distinguish the junction in skin from that of the hair follicle. Further studies, such as those by Shimizu et al [33] using immunohistochemistry, could be useful to defme the differences between skin and both types of hair-follicle junctions. It would also be interesting to investigate changes to both ultrastructure and immunomorphology during the hair cycle, as changes have been demonstrated in the amount and distribution of ba.sement membrane components [34]. In summary, these investigations into the comparative ultrastructure of normal human anagen hair-follicle bulbs have produced some interesting findings. Although follicles .showed some marked similarities in the CTEJ, there were also clear differences both on the outside of the follicle and around the dermal papilla as compared with the similar junction in skin. In particular, neither hemidesmosomes nor tonofilaments, as seen in the DEJ, were observed in any hair-follicle junctions. Anchoring fibrils appeared well organized and interconnected with extracellular matrix collagen ill the dermal papilla, but no anchoring fibrils were detected at the junction on the outside of the follicle. The lack of complexity of the junctions in the hair follicle as compared with the similar junction between dermis and epidermis is consistent with the need for motility of follicular tissues, in contrast to the relatively permanent relation between basal epithelial cells and the basement membrane. It will be interesting to examine the ultrastructure of the important DPEJ in hair follicles from patients with hair diseases, such as alopecia areata, to see whether it is altered in any way.

We would like to thank Dr. S. Maedonald Hull and Professor WJ. Cunliffe, Department of Dermatology, The General Infirmary, Leeds, and Mr. D. Sharjie, Mr. M. Timmons, and colleagues of the Plastic Surgery and Burns Unit, Department of Biomedical Sciences, University of Bradford for providing samples of skin. We acknowledge the expert technical assistance with microdissection of hairJollides of Dr. M.f. Thornton, Department of Biomedical Sciences, University of Bradford.

INFERENCES 1.

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

NUTBROWN AND I^JVNDALL

Briggaman RA, Wheeler CE: The epidennal-derni.il junction. J Invest Dermatol 65:71-84, 1975 2. Briggaman RA: Basement membrane formation and origin with special reference to skin. I-ronI Matrix Biol 'J:142-154, 1981 3. Tidman MJ, Eady I ^ J : Ultra.structural niorphometry of normal hunian dermalepidermal junction: the influence of age, sex and body region on latninar and non laminar components. J/HI/W/ Dertnatol 83:448-453, 1984 4. Eady R.AJ: Basement memhrane-interface hetween the epithelium and the dermis: structural features. Arch Dermalol 124:709-712, 1988 5. Briggaman RA, Dalldorf FG, Wheeler CE: Formation and origin of hasal lamina and anchoring fihrils in adult human skin. J Cell Biol 51:384-394, 1971

6.

Martin GR, Tinipl R: Laminin and other basement inembrane components. Ann Rev Cell Biol 3:57-85, 1987 7. Briggaman RA: Epidermal-dermal interactions in adult skin. J Iiii'csl Dermatol 79:21 S-24S, 1982 8. Stanley JR, Woodley DT, Katz SI, Martin GR: Structure .ind function of hasement memhrane. 7 Invest Dcrmalol 79:69S-72S, 1982 9. Birbeck MSC, Mercer EH: The electron microscopy of the hunian hair follicle 1. Introduction and the hair cortex. J Biophys Biochem Cytot 3/2:203-214 + Pl, 1957 10. Birbeck MSC, Mercer EH: The electron microscopy of the human hair follicle 2. The hair cuticle. J /iwji/iy.! Biochem Cytol 3/2:21 5-222+Pl, 1957 11. Birheck MSC, Mercer EH: The electron microscopy of the human hair follicle 3. The inner root sheath and trychohyaline. J Biophys Biitchem Cytol 3/2:223-. 230 + PI, 1957 12. Puccinelli VA, Caputo R, Ccccarelli B: The structure of the hunian hair follicle and hair shaft: an electron microscope study. Ciont llat Derm 108:453-498, 1967 13. Puccinelli V, Caputo R, Ceccarelli B: Changes in the basement memhrane of the papilla and wall of the hnman hair follicle. Arch Klin Exp Dermatol 233:172— 183, 1968 14. Laurie GW, Lehlond CP: Basement membrane nomenclature. Nature 313:272, 1985 15. Nomina histologica. In: Nomiiiti Aiiatoiiiica, part 2. Williams & Wilkins, Baltimore, 1983, p H6 16. Ross MH. Romrell LJ: Epithelium. In: Histology—A Text anil Alias. Williams &' Wilkins, Baltimore, 1989, pp 51-84 17. Parakkal PF: Ultrastrnctural changes of basement nienihrane during the hair growth cycle. J Cc/1 Biol 40:561-564, 1969 18. Parakkal PF: Morphogenesis of the bair follicle during eatagen. Z Zcltforsch 107:174-186. 1970 19. Sugiyama S, Takahaslii M, Kamimum M: The ultrastrncture of hair follicles in early and late catagen, with special reference to the alteration of the junctional structure hetween the dermal pajjilla and the epithelial components. J Uttrn^ struct Res 54:358-373. 1975 20. Oliver RF: The dermal papilla and the development and growth of bair. J Soc Comet Chem 22:741-755, 1971 21. Oliver RF.Jaboda CAB: The dermal p,ipilla and maintenance of hair growth. In: Rogers GE, Reis PR, Ward KA, Marshall RC (ed.s.). The Biology of Wool and Hair. Chapman & Hall, London, 1989, pp 59-61 22. Randall VA, Thornton MJ, Hamada K, Messenger AG: Mechanism of androgen action in cultured dermal papilla cells derived from hunian hair follicles with varying re.sponses to androgens in vivo. J Invest Dermatol 98:86S-91S, 1992 23. Lewis PR, Knight DP: Staining methods for sectioned material. In: Glancrt AM (ed.). Practical Methods in Electron Microscopy, Vol 5, Part I. North-Holland Elsevier, Amsterdam, 1977, pp 40-53 24. Richardson KC, Jarret L, Finke EH: Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Ttichnol 35:313-323. 1960 25. Hardy MH, Van Exan RJ, Sonstegard KS, Sweeny PR: Basal lamina changes during tissue interactions in hair follicles—an ill l'ifro study of normal dermal p.ipillae and vitamin A-induced glandular morphogenesis. J Invest Dermatol 80:27-34, 1983 26. Mutasim DF, Takahaslii Y. L.ibib RS, Anhal GJ, Pate! HP, Diaz LA: A pooi of bullons pempbigoid antigen(s) is intracellular and associated witb the hasal cell cytoskeleton-hemidesmosome complex. J Invest Dermalol 84:47-53, 1985 27. Messenger AG, Elliott K, Temple A, Randall VA: Expression of basement memhrane proteins and interstitial collagens in dermal papillae of human hair follicles. J Iiivcsl Dermatol 96:93-97, 1991 28. Westgate GE, Shaw DA, Harrap GJ, Couchman JR: Immunohistochemical localization of basement membrane components during hair follicle lnorpho.gene.sis.J/»i'ra( Dermalol 82:259-264, 1984 29. Bard S, Macouin C, Thivolet J, Sengel P: Heterogeneous distrihution of huUous pemphigoid antigen during hair development in the mouse. Arch Anat Microscop Exp 70:141-148, 1981 30. Uno I-I: Biology of hair growtb. Semin Refrod Endocrinol 4/2:131-141, 1986 31. Montagna W, Parakkal PF: The pilary apparatus. In: Montagna W , Parakkal P (eds.). The Stnlclnre anil Function of Skin. Academic Press, London, 1974, jjp 172-258 32. Reynold.s AJ, Jahoda CAB: Inductive properties of hair follicle cells. Ann NY Acad Sci 642:22b-242, 1991 33. Shimizu H, Ishida-Yamamoto A, Eady RAJ: The use of silver enhanced I nm gold probes for light and electron-microscopic localization of intracellular and extracellular antigens in skin. J Histochem Cytochem 40:883-888, 1992 34. Couchman JR, Gibson WT: Expression of basement membrane components through morphological changes in the bair growth cycle. Dcv Biol 108:290298, 1985