Hairless guinea pig skin: anatomical basis for studies of cutaneous biology

ARTICLE

The structural characteristics of the glabrous skin of normal laboratory animals such as mice, rats, guinea pigs, rabbits, dogs and non-human primates differ markedly from those of human skin. For example, these species have skin with a thinner epidermis, relatively flat dermal-epidermal junctions devoid of rete ridges [1], a loosely organized dermal structure [2, 3] and a rudimentary dermal vascular system [3]. Consequently, the reactivity of their skin to a variety of chemicals is quite different to that of human skin [4, 5].
The hairless guinea pig (HL-GP), Crl: IAF (HA) BR strain, is a mutant that was first identified, in 1978, in a Hartley guinea pig colony at Montreal's Institute Armand Frappier [6]. HL-GPs are euthymic and have pinkish, slightly wrinkled skin with few, very short vellus hairs [7]. At birth, HL-GPs are generally smaller than normal-haired guinea pigs (HD-GPs), but after the first month, they grow and gain weight at a rate comparable to HD-GPs [6]. HL-GPs remain hairless, except for areas around the nostrils and on the dorsa of the feet where continuous hair growth is seen throughout their lives [6]. However, the normal structure of HL-GP skin has not been elucidated definitively.
The aims of this study were to establish the anatomical basis of HL-GP skin for studying cutaneous biology and to examine the structural similarities to and differences from normal human skin of HL-GP skin.

Materials and methods

Materials

Six male HL-GPs, Crl: IAF (HA) BR strain, weighting 500-700 g and 6 age- and sex-matched HD-GPs, albino Hartley strain, were purchased from Charles River Laboratories (Boston, MA, USA). They were kept at 20° C, with a 12-hr light/dark cycle and fed a pelleted diet and water. A total of eight normal unexposed skin specimens was obtained from the buttocks of four male Caucasians and four male African-Americans. Their mean age was 28.6 ± 5.2 years.

Surface morphology

In order to evaluate the skin surfaces and stratum corneum of HL-GPs and HD-GPs, detergent scrubbing, stripping with D-Squame™ tape, cyanoacrylate skin surface biopsy and silicon replicas were made on the lateral body-sides of 3 HL-GPs and 3 HD-GPs. Before performing these procedures, the hairs of the HD-GPs were plucked from the areas under investigation. We prefer plucking to shaving as it causes less alteration of the interfollicular stratum corneum and leaves a smooth skin surface without stubble.

Detergent scrubbing
A 2-ml aliquot of 0.1% phosphate-buffered (pH 7.5) Triton X 100 solution was placed in a glass well (20 mm in diameter), which was pressed on the skin, and the skin surface was gently scrubbed with a teflon rod for 60 s. The resulting corneocyte suspension was collected and centrifuged at 3,000 rpm for 5 min. The pellet was re-suspended and 1-ml aliquots of the scrub suspension were applied with a micropipette to form circles 5-7 mm in diameter on glass slides and stained with 100 mul staining solution comprising 2.5% rhodamine B and 0.75% methylene blue. Part of the scrub suspension was sonicated in a water/detergent solution (9:1 v/v) using a sonicater for 60 min, centrifugated and stained as described above.

Tape stripping
Superficial layers of the stratum corneum were stripped off with D-Squame™ tape (CuDerm Corp., Dallas, TX, USA), which was pressed onto the skin for 30 s and then gently pulled off. Each piece of D-Squame™ was covered with the staining solution described above, placed on a hot-plate at 60° C for 15 min, the excess dye was rinsed off with tap water, each type was dried, placed, sticky side up, on a glass slide and mounted.

Cyanoacrylate skin surface biopsy
After stripping the skin surface with D-Squame™ tape, a drop of Krazy Glue™ was applied to the skin, pressed by applying a plastic slide to form a thin layer of approximately 2 cm² and gently pulled off a few minutes later.

Silicon replicas
Silfo™ silicon impression material (Flexico, Potters Bar, England) was mixed with catalyst and applied to a 1-cm² area of the skin surface defined by a template.

Analytical methods
The detergent scrub specimens, D-Squame™ specimens and cyanoacrylate slides were evaluated using an Olympus BH-2 light microscope (Tokyo, Japan), which was equipped with a CCD-72 Series camera (Dage-MTI, Inc., Michigan City, Indiana, USA) connected to an IP-8/AT matrix frame grabber board within a 486 Tandy PC (19 Mb RAM). The corneocytes on the D-Squames™ were measured at magnifications of x 200 to x 400 using the image analysis program analysis SIS^®, Soft-Imaging Software GmBH, W 4400 (Muenster, Germany). Each single cell on the images displayed on a NEC MultiSync 4FG monitor (Tokyo, Japan) was outlined manually with the mouse and its area and perimeter were calculated by the program. The follicular openings on each cyanoacrylate slide were counted using the touch-count mode of the program analysis SIS^® (magnification x 40). The silicon replicas were observed under a Zeiss operation microscope OPMI 1-FC at magnifications of x 9.5 and x 21.3 with a single light source at different angles. For scanning electron microscopy (SEM), specimens were mounted on aluminum stubs with double-sided gum tape coated with platinum in an ion coater and examined using a Hitachi S-700 scanning electron microscope (Tokyo, Japan).

Light microscopy and image analysis

The skin biopsy specimens were fixed with 10% formalin, dehydrated with graded ethanols and embedded in glycomethacrylate (JB4). Sections (2 mum) were cut using a Reichert-Jung 2050 microtome and stained with 0.5% toluidine blue followed by 2% basic fuchsin. For image analysis of epidermal thickness, care was taken to cut the sections perpendicular to the surface. Histometric measurements of the mean epidermal thickness and densities of microvessels, Langerhans cells and dermal interstitial cells were taken using a Southern Micro Image Analysis System (Southern Micro Instruments, Inc., Atlanta, GA, USA). The mean epidermal thickness was expressed as the epidermal area/horizontal length of epidermis [8]. The data were expressed as mean ± SE. Statistical analysis was performed using Student's t-test (two tailed). The differences were considered significant at p < 0.05.

Adenosine-S-triphosphatase (ATPase) staining

Prior to the skin biopsy, the hairs on the back of each HD-GP were removed by plucking. Biopsy specimens from 3 HL-GPs and 3 HD-GPs were incubated in 5 ml 0.5 M ammonium thiocyanate (NH₄SCN) in sodium-potassium phosphate buffer (pH 6.8) at 37° C for 20 min. Then, with the aid of a dissecting microscope, each epidermis was grasped with fine forceps and removed in one piece. The epidermal sheets were washed with physiological saline, gently teased flat on filter papers, fixed with formol sucrose buffer (pH 7.2) for at least 20 min at 0-4° C, then washed with 7.5% sucrose in 0.07 M cacodylate buffer at 0-4° C for 20 min. Subsequently, the specimens were incubated in ATPase in 0.09 M Tris-malate buffer containing 0.25% magnesium sulphate and 0.08% lead nitrate at 37° C for 45 min, rinsed thoroughly with distilled water, and developed in ammonium sulfide solution for 1 min. The stained epidermal sheets were rinsed with distilled water, placed on glass slides and mounted in glycerol.

Transmission electron microscopy (TEM)

Tissue samples were fixed with 4% buffered glutaraldehyde overnight at 4° C, washed with 0.1 M cacodylate buffer (pH 7.4), post-fixed with 2% osmium tetroxide for 2 h dehydrated with graded ethanols and then with propylene oxide and embedded in Taab Epon 812 (Marivac Ltd., Nova Scotia, Canada). Ultrathin sections were cut with a Porter-Blum MT2B ultramicrotome (Sorvall Inc., Newtown, CT), stained with uranyl acetate and bismuth subnitrate and observed in a Hitachi H-7000 electron microscope (Tokyo, Japan).

Results

Skin surface morphology

Morphology and size of corneocytes
Two populations of corneocytes were identified in both HL-GP and HD-GP skin after D-Squame™ tape stripping. One population comprised evenly pink-stained, transparent, polygonal (mostly pentagonal) cells forming sheets that were regular, one or more cell layers thick and the cell edges overlapped slightly in a honeycomb pattern. The other population consisted of single, smaller, irregular, more darkly stained and mostly fragmented or folded cells. The latter were concentrated in band-like areas corresponding to the follicular ridges. The flat pentagonal outlines of the corneocytes from the interfollicular epidermis were also observed in the cyanoacrylate skin surface biopsy specimens.
Unlike human skin, detergent scrubbing of which yields a suspension of mainly single corneocytes, the scrub suspensions of both HL-GP and HD-GP skin contained large clumps of corneocytes and sebum and were devoid of single cells. However, after sonication for 60 min, these suspensions contained mostly single cells, approximately 90% of which were corneocyte fragments that were irregularly shaped and occasionally elongated, and very few were small, dark purple-stained intact cells. No large pentagonal corneocytes were present.
Neither parakeratotic cells, "ghost cells" nor cells with any visible nuclear remnants were present among the corneocytes from the stratum corneum. The sizes, shapes and staining properties of the cells were fairly uniform. There were no significant differences between the mean areas, perimeters or diameters of the intact corneocytes obtained from HL-GPs and HD-GPs with the D-Squames™ tape.

Arrangement of follicles, follicular density and hair morphology
In SEM, the hairs of HL-GPs were thin, short, extremely irregular in diameter (moniliform) and often twisted or curled and the corneocytes of their shafts were arranged irregularly (Fig. 1A, C). The scanning electron micrograph of the mechanically depilated skin surfaces of HD-GPs revealed the remaining hairs were straight with regularly arranged cuticles (Fig. 1B, D). In contrast, the hair follicles of both HL-GPs and HD-GPs were arranged in parallel lines 450-650 mum apart. Similar numbers of follicular openings were imprinted in the cyanoacrylate skin surface biopsy specimens of HL-GPs (Fig. 1E) and HD-GPs (Fig. 1F).

Dermatoglyphics
Observation of the silicon impressions of both HL-GP and HD-GP skin, under a magnification of 20 x, revealed no dermatoglyphics comparable to those of human skin. The skin surfaces had fine, parallel wrinkles, most of which ran horizontally to the follicular ridges. The follicular ridges of HD-GPs left more pronounced imprints in the replicas than those of HL-GPs.

Histology

The epidermis of the HL-GP had no apparent rete ridge-like downy growths that interdigitated with the dermal papillae. However, small dermal papillae containing some capillaries were observed frequently. The viable epidermis comprised the stratum basale, 3 to 6 layers of stratum malpighii and 2 to 3 layers of stratum granulosum. The density of the vellus hair follicles of HL-GPs was almost same as that of the terminal hairs of HD-GPs and the sebaceous glands of the both were poorly developed (Fig. 2A, D).
The upper dermis of the HL-GP was composed of relatively loose collagen fibers, but the border between the papillary dermis and reticular dermis was obscure. Well-developed microvascular networks, distributed preferentially within the upper dermis and perifollicular connective tissue sheath, were present and a network of vertically oriented capillary loops extending into the small dermal papillae was observed (Fig. 2E). Luna's elastin staining revealed elastic fibers intertwined among the collagen fibers of the middle and lower dermis and visualization with Hale's colloidal iron revealed scanty glycosaminoglycans only around the microvessels in the upper dermis and perifollicular connective tissue.
The most striking feature of the HL-GP dermis was the huge number of interstitial cells, which were round/oval, spindle-shaped or had dendritic outlines when examined by light microscopy. These interstitial cells were randomly distributed throughout the dermis, although some were associated with microvessels or hair follicles (Fig. 2E).
ATPase staining of HL-GP epidermal sheets (Fig. 2G) revealed a similar distribution of epidermal dendritic cells to the HD-GP (Fig. 2H). The dendrites of the HL-GP seemed to be slightly thicker than those of the HD-GP.

Histometric analysis

The mean epidermal thickness of the HL-GPs was significantly higher than that of the HD-GPs (P < 0.0001) and the value of the former was similar to that of the human skin (p = 0.114; Fig. 3A).
The density of the Langerhans cells, visualized by ATPase staining, in epidermal sheets of the HL-GPs was 1,674.7 ± 120.6/mm² (mean ± standard error; SE), while that of the HD-GP was 1,418.7 ± 91.1/mm². The difference was not statistically significant.
The total number of dermal interstitial cells of the HL-GPs was significantly higher than that of both the HD-GPs (p < 0.0001) and humans (p < 0.0001). The HL-GPs had significantly more dendritic/spindle-shaped cells than both the HD-GPs and humans (p < 0.0001), whereas all 3 species had similar numbers of round/oval cells (Fig. 3B).
The number of microvessels in the upper dermis of HL-GP was 7.05 ± 0.62/0.1 mm² (mean ± SE), that of HD-GP was 5.44 ± 0.71/0.1 mm², and that of human was 7.29 ± 0.51/0.1 mm². The difference among the 3 species was not statistically significant.

Transmission electron microscopy of HL-GP skin

Ultrastructural examination revealed that the epidermis of the HL-GP comprised keratinocytes with four strata (basal cell, spinous, granular and horny layers), melanocytes in the basal layer and Langerhans cells in the spinous layer (Fig. 4A). In both HL-GP and HD-GP the spinous cells contained relatively thick bundles of tonofilaments, whereas the basal cells contained sparse, thin bundles of tonofilaments. In contrast, human basal cells contained thicker bundles of tonofilaments than spinous cells. In all 3 species, the keratinocytes were connected to each other by desmosomes. The epidermal stratum corneum of HL-GPs consisted of 2-4 electron-dense lower layers and 15-20 less electron-dense upper layers. In both types of guinea pig, the corneocytes in the lower layers were filled with dense intermediate filaments and connected tightly to each other by vestiges of desmosomes, which have been designated corneosomes, and intercellular cement materials. In all 3 species, the corneocytes in the upper layers contained relatively sparse intermediate filaments and were connected to each other by interdigitations, fewer corneosomes being present (Fig. 4B). The granular cells contained a number of membrane-coating granules and keratohyaline granules, which were characterized by variable-sized, angular, electron-dense bodies connected to some tonofilaments (Fig. 4C). Two ultrastructurally heterogeneic types of basal cell were observed. Basal cells not facing dermal papillae had non-serrated dermal-epidermal junctions and sparse tonofilaments attaching them to hemidesmosomes (Fig. 4F). In contrast, basal cells facing small dermal papillae had serrated dermal-epidermal junctions and thicker bundles of tonofilaments attaching them to hemidesmosomes (Fig. 4G). This variation of basal cells was observed in HL-GPs and humans, but not in HD-GPs.
In both species, melanocytes were located exclusively within the basal layer, displayed dendritic cytoplasm and contained melanosomes at various stages, but lacked tonofilaments and desmosomal attachments (Fig. 4D). Human melanocytes typically hung down into the superficial dermis, whereas melanocytes of both types of guinea pig did not. Langerhans cells were distributed within the mid-spinous layer and were devoid of desmosomes and tonofilaments, but contained a number of coated vesicles, Golgi apparatus and Birbeck granules (Fig. 4E). There were fewer Birbeck granules in both types of guinea pig than in humans. Occasional cells lacked Birbeck granules and appeared to be consistent with indeterminate cells observed in human skin.
Mast cells were observed predominantly in the upper dermis. Occasional mast cells of HL-GPs were intimately associated with dermal dendrocytes or dermal macrophages like those in humans, but their distribution was not always angiocentric. The mast cells in both HL-GPs (Fig. 5A) and HD-GPs (Fig. 5B) contained a small number of granules at the periphery of their scant cytoplasm. These granules were homogeneously or segmentally electron-dense (Fig. 5A, inset) and lacked scrolls and lattices, which were frequent in the mast cells of human skin.
Collagen-producing fibroblasts in both guinea pigs had abundant cytoplasm, which contained dilated rough endoplasmic reticulum and Golgi apparatus, but lacked multiple, thin, elongated cytoplasmic processes (Fig. 5C). In all species, macrophages were characterized by indented nuclei, a number of lysosomal granules, coated vesicles and vacuoles (Fig. 5D).
The microvasculature of the HL-GPs consisted of well-developed endothelial cells, pericytes, perivascular macrophages and dermal dendrocytes. The latter displayed multiple thin, elongated cytoplasmic processes that surrounded the microvessels. Most of the dermal dendrocytes contained a small amount of rough endoplasmic reticulum and a few Golgi apparatus, and also had fibronexuses (Fig. 5E, inset) and pinocytotic vesicles along their plasma membranes (Fig. 5E). These characteristics of dermal dendrocytes in HL-GPs were similar to those in humans. In contrast, the microvasculature of the HD-GPs consistently lacked perivascular monocytes/macrophages, and perivascular dermal dendrocytes showing elongated cytoplasmic processes (Fig. 5F).

Discussion

HL-GP skin has several advantages as an experimental model for studying cutaneous biology. It is a suitable size to handle and its hairless surface enables the effects of mechanical and/or chemical depilation to be avoided. Most importantly, HL-GP skin shows more structural similarities to human skin than to HD-GP skin (Table I). A thick epidermis with distinct strata, serrated/non-serrated basal keratinocytes, the occasional presence of a papillary dermis and well-developed superficial dermal microvessels with vertically oriented loops are structural features of HL-GP skin that are found routinely in human skin, but rarely, if ever, are in the skin of HD-GPs and rodents. The principal disadvantage of the HL-GP for skin research is the lack of cross-reactivity with antibodies against several human and murine immune cell markers and this could limit the usefulness of the HL-GP for immunological studies.
The outlines of the epidermal corneocytes of both HL-GP and HD-GP skin are quite similar to those of the human skin: typically, they were flat, transparent, evenly pink stained pentagons forming honeycomb-like sheets. However, we encountered no nuclear remnants in the form of central, more faintly stained circles, which were present in a high percentage of human corneocytes [9]. Our measurements of the mean projected area suggest guinea pig corneocytes are slightly bigger than human corneocytes. The mean corneocyte area of guinea pig skin varied between 1,034 and 1,415 mm², whereas that of human volar forearm skin was reported to be 800 to 1,050 mm² [10, 11]. However, the fact that human corneocyte sizes vary depending on the body site and age of the person should be taken into consideration [12]. As the hair follicle density of guinea pigs is much higher than that of humans, far more corneocytes are shed by the guinea pig infundibula than by those of the human (except for the facial skin of the latter) and this is reflected by the high percentage of small, irregular, darkly stained corneocytes on the tapes after stripping. Unlike human skin, epidermal corneocytes cannot be obtained from either strain of guinea pig by the detergent scrub method, because the skin surfaces of guinea pigs are covered by lipids that form a coating that prevents shedding of single cells. The pilo-sebaceous unit of the HL-GP is quite different from that of the human and is almost certainly an inappropriate model of the human hair follicle and sebaceous gland.
Further striking differences between HL-GP skin and human skin are the densities of dermal interstitial cells such as monocytes/macrophages, dermal dendrocytes and fibroblasts. Our study demonstrated the population of these interstitial cells in HL-GP skin was larger than those in both human and HD-GP skin. The ultrastructural features of dermal dendrocytes in HL-GPs and humans are similar, i.e. slender, branching cytoplasmic processes and electron-dense, plasma-membrane-associated plaques, designated fibronexuses [13]. These characteristics of the dermal interstitial cells of HL-GPs raise the possibility that the HL-GP will be a useful animal model for studying wound healing, granulomatous inflammation and macrophage-mediated immunological processes.
The recent identification of the human homolog of the hairless gene on human Chromosome 8p12 confirmed the clinical significance of the phenomenon of "hairlessness" in humans, which predicted on the basis of similarities between hairless mice and a congenital hair disorder characterized by atrichia with papules [14, 15]. Several mutations in hairless gene have been identified in families of this congenital hair disorder around the world [15]. In contrast, the responsible gene and underlying genetic defect has not yet identified in HL-GPs [7]. Hence, the HL-GP skin phenotype should be clearly demarcated from the hairless (hr) phenotype observed in mice and humans caused by mutations in mouse and human hairless (hr) gene, respectively.
HL-GP skin has already been utilized as an experimental animal model in studies on contact dermatitis [6, 16], photodermatology [17], pigmentation [18], dermatophytosis [19], latent herpes simplex infection [20], wound healing [21] and cutaneous pharmacology [22, 23]. However, the functional characteristics of HL-GP skin have yet to be elucidated fully. Preliminary reports indicate that i) although allergic contact dermatitis could be induced in both strains of guinea pig by dinitrochlorobenzene (DNCB), HL-GPs developed stronger reactions than HD-GPs [16] ; ii) HL-GPs reacted more strongly than HD-GPs to UVB irradiation and PUVA treatment [17] ; iii) photoallergic contact sensitization was induced more easily in HL-GPs than HD-GPs by tetrachlorosalicylanilide [17] and iv) HL-GP skin was highly responsive to inflammatory mediators such as histamine and leukotriene D4 [6]. The density of ATPase-positive epidermal Langerhans cells has been reported to be higher in HL-GPs than HD-GPs. However, in our experiment, ATPase staining of the epidermal sheets of HL-GPs and HD-GPs failed to demonstrate a significant difference between the densities of such cells. Our data raise the possibility that the well-developed superficial dermal vessels and greater number of resident interstitial cells such as monocytes/macrophages and dermal dendrocytes in HL-GPs may result in stronger allergic contact sensitivity and UV irradiation reactions than those manifested by HD-GPs.
In conclusion, the results of our precise anatomical analysis of HL-GP skin collectively suggest that HL-GP skin is more similar to human skin than to the skin of other rodents and, therefore, the HL-GP may be a useful animal for studying cutaneous biology, experimental pathology, pharmacology and toxicology. Furthermore, our data suggest that comparative studies of HL-GP and HD-GP skin will further our understanding of the functional roles of resident dermal interstitial cells in normal and diseased skin.

REFERENCES