Cell Tissue Res (1995) 280:1-10 Cell&Tissue Research

�9 Springer-Verlag 1995

Expression of the endogenous 14-kDa [3-galaetoside-binding lectin galectin in normal human skin Yoshihiro Akimoto t, Jun Hirabayashi 2, Ken-ichi Kasai 2, Hiroshi Hirano 1

1 Department of Anatomy, Kyorin University School of Medicine, Mitaka, Tokyo, 181 Japan a Department of Biological Chemistry, Faculty of Pharmaceutical Sciences, Teikyo University, Sagamiko, Kanagawa, 199-01 Japan

Received: 3 May 1994/Accepted: 19 September 1994

Abstract. The localization of an endogenous 14-kDa fl- galactoside-binding lectin (galectin) and its pattern of gene expression were examined in normal human skin by light- and electron microscopy. Under the light mi- croscope, immunostaining of 14-kDa galectin was ob- served in the cell membrane of cells in the basal and spi- nous layers of the epidermis. Galectin was also found in the Langerhans cells, as shown by double labeling using anti-14-kDa galectin and anti-CDla antibodies. In the dermis, immunostaining for the 14-kDa galectin was positive in the extracellular matrix and fibroblasts. At the electron-microscopic level of resolution, galectin was located primarily along the plasma membrane of ke- ratinocytes, and in both the cytoplasm and nucleus of Langerhans cells in the epidermis, whereas in the dermis it was detected in the extracellular matrix and in both the nucleus and cytoplasm of fibroblasts. The gene expres- sion of 14-kDa galectin was visualized by the HRP- staining method following in situ hybridization tech- niques. The expression was detected in the cytoplasm of cells in the basal and spinous layers of the epidermis; whereas, in the dermis, it was detected in the cytoplasm of fibroblasts. Moreover, SDS-polyacrylamide gel elec- trophoresis and lectin-blot analysis revealed that this ga- lectin bound to glycoproteins of approximately 17, 62, and 72 kDa in the epidermis and to those of 29, 54, and 220 kDa in the dermis. The present study indicates that 1) normal human skin produces the fl-galactoside-bind- ing 14-kDa galectin, and 2) this galectin is located in both the epidermis, particularly in the keratinocytes and Langerhans cells, and in the dermis. These results sug- gest that galectin is important for cell-cell contact and/or adhesion in the epidermis and for cell-extracellular ma- trix interaction in the dermis.

Key words: Galectin -fl-Galactoside-binding lectin - Human - Skin - Immunocytochemistry - Immunohisto- chemistry - Hybridization, in situ - Langerhans cell - Man

Correspondence to: H. Hirano

Introduction

Animal lectins are detected in a variety of animal tis- sues, and have been regarded to play important roles in cell-to-cell recognition in which glycoconjugates are in- volved (Barondes 1986; Sharon and Lis 1989). Verte- brate fl-galactoside-binding lectins (galectins), which are Ca2+-independent and require thiol reagents for mainte- nance of their sugar binding activity, are distinct from the CaZ+-dependent lectin family, and are thought to be associated closely with development, differentiation, neoplastic transformation, and metastatic progression (Nowak et al. 1977; Kobiler et al. 1978; Briles et al. 1979; Montelione et al. 1981; Raz et al. 1986; Hiraba- yashi and Kasai 1993; Barondes et al. 1994).

We previously showed immunohistochemically the existence and localization of the 14-kDa and 16-kDa ga- lectins and also their gene expression in chick embryon- ic skin both in vivo and in vitro (Hirano et al. 1988; Oda et al. 1989; Akimoto et al. 1992, 1993). These lectins were expressed mainly in the intermediate layer of the keratinized epidermis of tarsometatarsal skin of the chick embryo. Recently it was reported that the keratino- cytes of human skin also produce an endogenous fl-ga- lactoside-binding lectin (33-kDa IgE-binding protein), and that this 33-kDa galectin binds to Langerhans cells where it modulates their binding capacity for glycoform IgE (Wollenberg et al. 1993). In parallel, another endo- genous fl-galactoside-binding isolectin (MW 14kDa: 14-kDa galectin) has been found in human tissues. In fact, we isolated a 14-kDa galectin from human placenta (Hirabayashi and Kasai 1984), and showed, based on our immunological (Hirabayashi et al. 1987a) and sequence analysis (Hirabayashi et al. 1987b, 1989; Hirabayashi and Kasai 1988), that human 14-kDa galectin is quite similar to chick 14-kDa galectin. From analysis of both complete amino acid and nucleotide sequences of a full- length cDNA for the human 14-kDa galectin (Hiraba- yashi and Kasai 1988; Hirabayashi et al. 1989), this ga- lectin was found to be composed of 134 amino acid resi- dues, 56% of which are identical with chick 14-kDa ga-

lectin. The human 14-kDa galectin has some regions ho- mologous to those in carbohydrate-binding protein 35 (CBP35) and IgE-binding protein (Albrandt et al. 1987; Laing et al. 1989). Moreover, the human 14-kDa galectin shows about 90% similarity with rat 14-kDa galectin. No signal sequence exists in the initiator region of the cDNA.

Light -microscopic localization o f the 14-kDa galectin has been reported in various normal and malignant tissues including cutaneous tissue (Allen et al. 1991; Skrincosky et al. 1993). According to Allen et al. (1991), in cutaneous tissue the extracellular matrix contains abundant galectin. However, details o f the expression pattern o f the gene as well as detailed localization and the function o f the human 14-kDa galectin, especially in the epidermis, have remained as yet undetermined.

In the present study, localization and gene expression of the 14-kDa galectin in normal human skin was investi- gated by means of conventional immunocytochemist ry and the in situ hybridization technique utilizing the c D N A of 14-kDa galectin. These results were compared with those obtained in the chick embryonic skin. Moreover, the glycoproteins recognized by the 14-kDa galectin were characterized by SDS-PAGE and Western blot analysis.

Materials and methods

Tissue

Biopsies from the abdomen, breast, face, and plantar skin were performed on patients (n=10) admitted for diagnostic purpose. Skin specimens that were histologically diagnosed as normal were used. Essentially the same results were obtained from each indi- vidual irrespective of age and sex.

lmmunoblotting procedures

Skin specimens were homogenized in phosphate-buffered saline (PBS), pH 7.2, containing 0.1 M lactose and 4 mM 2-mercaptoetha- nol, dialyzed against PBS, and then mixed with Laemmli sampling solution and boiled for 3 min. The homogenates were electrophore- sed in 12.5% polyacrylamide slab gels in the presence of 1% SDS and 100 mM 2-mercaptoethanol and then transferred to nitrocellu- lose sheets (Towbin et al. 1979). The sheets were stained with rab- bit anti-14-kDa galectin antiserum, followed by horseradish peroxi- dase (HRP)-labeled goat anti-rabbit IgG. The protein band was de- tected by immersion of the sheets in 3,3'-diaminobenzidine. 4HC1 (DAB, 0.2 mg/ml)-H202 (0.005%) for 5 rain at room temperature.

Immunostaining for light-microscopic observation

Skin specimens were fixed in 4% formaldehyde in 0.1 M phos- phate buffer (pH 7.3) for 1 h at 4 ~ C. The specimens were im- mersed in 2.3 M sucrose-phosphate buffered saline (PBS), and then frozen with liquid nitrogen. Semithin frozen sections of 1-gm thickness were cut, washed with PBS, and treated for l0 min with 1% bovine serum albumin (BSA) in PBS. The sections were then incubated with rabbit antiserum raised against placental 14-kDa galectin (Hirabayashi et al. 1987b), and/or mouse anti-CDla monoclonal antibody (Oncogene Science, N.Y., USA) or normal rabbit serum for 1 h at room temperature, washed with PBS, and

Fig. 1. Immunoblot analysis of skin lysates by use of anti-14-kDa galectin antibody. Normal human skins were lysed. The proteins were electrophoresed and immunoblotted as described in

'"Materials and methods". A single 14-kDa band is seen

subsequently incubated for 1 h with the rhodamine-conjugated F(ab')2 fragment of goat anti-rabbit IgG antibody, and/or FITC- conjugated goat anti-mouse IgG and IgM (Jackson Immunore- search, West Groove, Pa., USA). Nuclei were stained with 4', 6- diamidine-2-phenylindole hydrochloride (DAPI, Boehringer, Mannheim, Germany) (Williamson and Fennel1 1974). After a wash with PBS, the specimens were mounted in 90% glycerol-0.1 M TRIS-HC1 buffer (pH 8.5) containing 0.5 mM p-phenylene di- amine, which prevents fading of fluorescence during microscopic examination (Huff et al. 1982), and observed under a Nikon mi- croscope equipped with the epifluorescence system.

To confirm the specificity of the staining, we also carried out pre-absorption and competition experiments. Some sections were preincubated with the antigen at a concentration of 50 gg/ml for 30 rain and then incubated with the antibody in the presence of the antigen (50 gg/ml).

Immunostaining for electron microscopy

For colloidal gold labeling, fixation with formaldehyde and in- fusion of sucrose were carried out in the same way as for light microscopy; then, ultrathin frozen sections were cut at -90 to

Fig. 2a-d. Immunohistochemical localization by light micro- scopy of endogenous fl-galactoside-binding 14-kDa galectin in normal human skin from abdomen. Stained with rabbit anti-galec- tin serum and visualized by immunofluorescence with rhoda- mine-conjugated goat anti-rabbit IgG (red). DNA stained with DAPI (blue). a, e Immunofluorescent images of specimens, b, d Nomarski differential interference-contrast images of the same specimen. Arrowheads indicate basal surface of epidermis. C Stratum corneum; E epidermis; D dermis. Bar: 10 gm. a Positive

staining is observed along cell membrane of basal and intermedi- ate cells of epidermis. Stratum corneum of epidermis is scarcely stained. Both extracellular matrix and fibroblasts are intensely stained in dermis. Some nuclei of fibroblasts in dermis show white fluorescence due to relatively high concentration of dyes. e Normal human skin. Preincubated with antigen (14-kDa galectin), then incubated with anti-galectin antibody in presence of antigen. Positive staining is not observed

-100 ~ C, and collected on electron-microscopic grids coated with Formvar and carbon. The sections were washed with PBS and pre- treated with 1% BSA in PBS for 10 min. After a PBS rinse, they were incubated with the anti-14-kDa galectin antiserum for 1 h or with control normal rabbit serum, washed with PBS, and incubat- ed with the colloidal gold (10 nm in diameter)-conjugated goat

anti-rabbit IgG (Janssen, Belgium) for 1 h. After another wash with PBS, the sections were refixed in 2% glutaraldehyde-0.1 M phosphate buffer, pH 7.4, and embedded in a mixture of methyl- cellulose, polyethyleneglycol, and uranyl acetate by the absorp- tion staining method (Tokuyasu 1980). The specimens were ob- served with a JEM-1200EX electron microscope.

Fig. 3a-d. lmmunofluorescence localiza- tion of 14-kDa galectin and C D l a in nor- mal human skin. Semithin frozen sections were triply stained, i.e., for 14-kDa galec- tin with rhodamine (red), for C D l a with fluorescein (green), and for nuclear DNA with DAPI (blue). Double-exposure im- ages for 14-kDa galectin and nucleus a and for C D l a and nucleus b, and triple- exposure image for 14-kDa galectin, CDla, and nucleus e are shown, d No- marski differential interference-contrast image of the same specimen. Note CD1 a- positive sites (green; short arrows), present only in Langerhans cell of skin, are identical to a part of 14-kDa galectin- positive sites (red; long arrows). Arrow- heads indicate basal surface of epidermis. Bar: 10 ~tm

Preparation of cDNA probe for galectin

A 0.5-kb PCR fragment was used as a DNA probe for in sitn hy- bridization. For the preparation, the previously described cDNA clone 302 (Hirabayashi et al. 1989) was used as a template, and annealed to 2 convergent primers, 5'-ATG TCC GCC GAA GAG CCA A-3' (forward, previously designated primer 3) and 5'-TTA TTG GAT TTG AAT TCC AGA CA-3' (reverse, primer 5). PCR was performed with a Cetus thermal cycler using a Takara (Kyoto, Japan) GeneAmp PCR reagent kit (conditions: denaturing at 94 ~ C for 1 rain, annealing at 42 ~ C for 2 min, and extension at 72 ~ C for 3 min, 25 cycles).

Sulfonation of probe DNA

The probe DNA was dissolved in 10 mM TRIS-1 mM ethylenedi- aminetetraacetic acid (EDTA) buffer solution (pH 7.6, TE buffer) at a concentration of 0.1 mg/ml, denatured by boiling for 10 rain, and chilled quickly on ice. Fifty microliters of 2 M sodium meta- bisulfite and 12.5 gl of 1 M O-methylhydroxylamine solution (pH 6.0) were added to 100 gl of prepared probe DNA, and the mix- ture was incubated for 4 h at room temperature.

thickness, placed on poly-L-lysine coated slides, and fixed with a mixture of ethanol and acetic acid. For prevention of non-specific binding, sections were acetylated with 0.25% acetic anhydride in 0.1 M triethanolamine buffer (pH 8.0) for 30 min. Before hybridization, the sections were prehybridized for 5 h at 42 ~ C in a mixture con- taining 100 gg/ml of yeast tRNA, I00 gg/ml of salmon sperm DNA, 50% deionized formamide, 0.6 M NaC1, 10raM TRIS-HC1 (pH 7.4), 1 mM EDTA, 0.02% polyvinylpyrolidone, 0.02% bovine serum albumin, and 0.02% Ficoll. The heat-denatured probe was mixed with the hybridization mixture to give a final concentration of 0.2 to 0.4 ng/gl, and applied to each slide as a 10-50 gl volume. Hybrid- ization was carried out in a humidified chamber at 42 ~ C for 12 to 18 h. The unhybridized probe was washed off in 2• sodium saline ci- trate (SSC) at room temperature for 4 h with a buffer change every 30 rain, and the slides were finally washed in 0.1x SSC for 1 h.

After incubation with 1% BSA in PBS for 10 min, anti-sulfo- nated DNA antibody (Orgenics, Yave, Israel) diluted 1:100 with 1% bovine serum albumin in PBS was applied to each slide, which was placed in a moist chamber and incubated at room tem- perature for 1 h. The slides were washed in PBS, and HRP-labeled secondary antibody was applied. Specific hybridization sites were visualized by reaction for 3-7 min with 0.02% DAB solution in PBS containing 0.005% H202.

In situ hybridization

In situ hybridization was carried out with sulfonated probes (Aki- moto et al. 1992). Skin explants were washed in PBS, and frozen im- mediately in liquid nitrogen. The frozen sections were cut at a 4-~tm

Western blot analysis of l4-kDa galectin-binding glycoproteins

By treatment of normal human skin with 25 mM EDTA for 2 h, the epidermis was separated from the dermis (Akimoto et al.

Fig. 4a-d. Ultrastructural localization of 14-kDa galectin by im- munogold labeling method of ultrathin cryosection reacted with colloidal gold-conjugated goat anti-rabbit IgG. a Intermediate cell layer of the epidermis. Arrowheads indicate colloidal gold parti- cles localized along plasma membrane. Arrows point to desmoso- mes. b Boundary between epidermis and dermis. Arrowheads in- dicate colloidal gold particles localized along plasma membrane and in intercellular space around a basal cell (BC). Reticular layer of basement membrane and extracellular matrix in the dermis (D)

are densely labeled by colloidal gold particles. BL Basal lamina, e A fibroblast and surrounding extracellular matrix in dermis. Col- loidal gold particles are localized in both nucleus and cytoplasm of fibroblast. C Collagen bundles, d A cytochemical control to check specificity of immunostaining along boundary between epi- dermis and dermis. Anti-14-kDa galectin antibody replaced with preimmune antibody. Practically no colloidal gold labeling is ob- served. N Nucleus. Bars: 0.5 gm

1991). The isolated epidermis and dermis were separately homog- enized in PBS containing 1% Triton X-100 at 4 ~ C. Each homoge- nate was centrifuged for 30 min at 20000 g, and the pellet was discarded. The supernatants were boiled for 5 min in the presence

of 1% SDS and 100 mM 2-mercaptoethanol, and a sample of each was applied to a 7.5% polyacrylamide gel. After electrophoresis, the protein bands were transferred to nitrocellulose sheets, and the sheets were incubated with 3% BSA in PBS for 2 h and subse-

2a, b). The stratum corneum was not stained. The epi- dermo-dermal junction, however, was stained, and in the dermis, both extracellular matrix and fibroblasts were also positive for the anti-14-kDa reaction. Furthermore, for simultaneous identification of both 14-kDa galectin- containing cells and Langerhans cells, double immuno- fluorescent staining was done by means of antibodies to 14-kDa galectin and CDla, the latter of which is ex- pressed only in the Langerhans cell of the skin. The CDla-positive sites were also positive for 14-kDa galec- tin (Fig. 3). That is, the Langerhans cell was also stained by anti-14-kDa galectin antibody.

When the primary antibody was omitted or replaced with the non-immunized rabbit serum, no positive stain- ing was observed. Furthermore, specific staining with the primary antibody was completely inhibited by the addition of the antigen. Fig. 2c, and d show immuno- staining of skin that had been incubated with anti-14- kDa-galectin antibody in the presence of the antigen. Positive staining was essentially absent.

Fig. 5. Ultrastructural localization of 14-kDa galectin in a Langer- hans cell by immunogold labeling of an ultrathin cryosection. LC Langerhans cell; KC keratinocyte. In Langerhans cell, neither tonofilaments nor melanin granules are observed. Langerhans cell has irregularly indented nucleus. Rectangles (a, b) indicate areas shown in Fig. 6. Bar: 1 gm

quently stained with 14-kDa galectin (10 gg/ml) in the presence of 4 mM 2-mercaptoethanol for 1 h. The sheets were washed with PBS and then treated with rabbit anti 14-kDa galectin antibody for 1 h. Next, the sheets were washed with PBS and then treated with HRP-labeled goat anti-rabbit IgG. The protein bands were detect- ed by immersion in DAB (0.2 mg/ml) -H202 (0.005%) for 10 rain. In control experiments, the nitrocellulose sheets were incubated with 14-kDa galectin in the presence of l0 mM lactose. Non-spe- cific bindings were also checked by incubation with HRP-second- ary antibody or DAB-H202 alone.

Results

Characterization of the antibody

Immunoblot analysis of skin lysates showed that the an- tibody against placental 14-kDa galectin detected the 14- kDa protein alone (Fig. 1). This indicates that the same 14-kDa galectin is expressed in both placenta and skin. The antibody did not react with 29-kDa or 33-kDa ga- lectin (Sparrow et al. 1987; Wollenberg et al. 1993).

Light-microscopic localization of the 14-kDa galectin

Immunohistochemical study revealed a specific staining pattern in the skin. There was no significant difference in staining patterns among the skin samples obtained from various body regions, whether thick or thin. In the epidermis intense staining was observed along the cell membrane of cells in the basal and spinous layers (Fig.

Electron-microscopic localization of the 14-kDa galectin

The precise location of the galectin was visualized by colloidal gold labeling at the electron-microscopic level of resolution (Figs. 4, 5, and 6). In the epidermis, the la- beling was observed along the plasma membrane and in the intercellular space (Fig. 4a, b). In the dermis the re- ticular layer of the basement membrane and extracellular matrix were densely labeled, whereas the lamina lucida and lamina densa of the basement membrane were dif- fusely labeled (Fig. 4b). Both the nucleus and cytoplasm of dermal fibroblasts were labeled (Fig. 4c). The label- ing was also observed both in the matrix of cytoplasm and in the nucleus of Langerhans cells (Fig. 6). Few col- loidal gold particles were detected in the cytochemical control experiments (Fig. 4d).

In situ hybridization

The localization of 14-kDa galectin mRNA in sections was examined by the in situ hybridization technique us- ing sulfonated cDNA as a probe. Intense hybridization signals were observed in the basal and spinous layers of the epidermis and in the dermal fibroblasts (Fig. 7). The signals were not observed in the stratum corneum of the epidermis.

Western blot analysis of 14-kDa galectin-binding glycoproteins

The 14-kDa galectin-binding glycoproteins in the epi- dermis or dermis were characterized by the analysis us- ing SDS-PAGE and Western blotting. They were visual- ized as several main bands on the nitrocellulose sheet. The 14-kDa galectin bound to the glycoproteins of ap- proximately 17, 62, and 72 kDa in the epidermis (Fig. 8a) and to those of 29, 54, and 220 kDa in the dermis

Fig. 6a, b. Enlargements of Lan- gerhans cell (rectangles) shown in Fig. 5. Langerhans cell has Bir- beck granules (arrows) in its cyto- plasm (a). Colloidal gold particles representing 14-kDa galectin are localized both in matrix of cyto- plasm (a) and in nucleus (b) of Langerhans cell. LC Langerhans cell; KC keratinocyte; N nucleus; T tonofilament. Bar: 0.5 ~m

(Fig. 8b). In control experiments, only the 14-kDa band was observed, but other bands were not observed in the case of anti-14-kDa galectin antibody and HRP-second- ary antibody treatments (data not shown). Thus, bands

other than the 14-kDa band were recognized by 14-kDa galectin. Except for the 14-kDa band, the bands de- scribed above were completely inhibited by the addition of 10 mM lactose (data not shown).

Fig. 7. In situ localization of 14-kDa galectin mRNA in section of normal human skin. Intense hybridization signals are observed in basal and intermediate cell layers of epidermis and in dermal fi- broblasts. Stratum corneum (C) is scarcely stained. D Dermis; E epidermis. Bar: 10 ~tm

Fig. 8a, b. SDS-PAGE and Western blot analysis of 14-kDa galec- tin-binding glycoproteins in epidermis or dermis of normal human skin. a 14-kDa galectin-reactive bands corresponding to 14, 17, 62, and 72 kDa glycoproteins are observed in blot of glycopro- teins from epidermis, b Bands corresponding to 14, 29, 54, and 220 kDa glycoproteins are seen for homogenate prepared from dermis

Discussion

In the present study, the localization of the 14-kDa galectin and its gene expression were examined immunohistochem- ically in normal human skin. We showed that the 14-kDa galectin exists in the epidermis as well as in the dermis.

Our light-microscopic results are consistent with those reported by Allen et al. (1991), who reported that 14-kDa galectin exists in a small amount in the epider- mis, but abundantly in the dermis. We revealed for the first time that 14-kDa galectin gene is expressed both in the epidermis and dermis of normal human skin. The lo- cation of the 14-kDa galectin is different from that of the 33-kDa galectin (IgE-binding protein; Wollenberg et al. 1993), for the former is primarily located on the cell sur- face, whereas the latter galectin exists mainly in the cy- toplasm of keratinocytes.

This study together with the previous one (Akimoto et al. 1992) enabled us to compare the localization of hu- man 14-kDa galectin with that of chick 14-kDa galectin in the skin. As our previous studies proved, human 14- kDa galectin is similar to chick 14-kDa galectin in its structure. Also, the location of 14-kDa galectin in nor- mal human skin is quite similar to that in the normal chick (Akimoto et al. 1992). However, there was a slight difference between them in the epidermis; that is, human 14-kDa galectin was expressed in both basal and inter- mediate layers of the epidermis, whereas expression of chick 14-kDa galectin was observed mainly in the inter- mediate layer of keratinized epidermis. The results of the in situ hybridization experiment were consistent with those of our immunohistochemical observations. That is, the human 14-kDa galectin mRNA was expressed in both basal and intermediate layers of the epidermis. In view of these findings, we suggest that 14-kDa galectin is produced in the human epidermis as well as in the der- mis and may be closely involved in cell-to-cell contact in the epidermis of the skin.

Changes in glycoconjugate structure have important roles in development and differentiation (Kawai et al. 1979; Barnes 1988). These changes are observed in both developing chick embryonic epidermis (Takata and Hirano 1983) and normal human skin (Ookusa et al. 1983). In the epidermis of the 13-day-old chick embryo, glycoconjugates containing fl-galactose and N-acetyl ga- lactosamine residues are not present. As the epidermis develops toward keratinization, however, these glyco- conjugates accumulate on the cell surface of the interme- diate cells. Such changes correspond chronologically to expression of galectins during epidermal differentiation (Akimoto et al. 1992 and 1993). This strongly suggests that the galectins recognize these glycoconjugates on the cell surface and are involved in cell-to-cell interaction. The site of 14-kDa galectin in normal human skin is consistent with the binding pattern of RCA, which binds to glycoconjugates containing fl-galactose and fl-galac- tose 1-4 N-acetyl glucose (Ookusa et al. 1983). Our re- sults indicate that the glycoconjugates containing galac- rose and lactosamine ~-galactose 1-4 N-acetyl glucos- amine), for which the 14-kDa galectin has affinity, accu- mulate on the cell surface as the keratinocyte migrates

and differentiates from the basal layer to the upper lay- ers of the epidermis, and that the 14-kDa galectin recog- nizes these glycoconjugates.

The 14-kDa galectin was abundantly present in the ex- tracellular matrix of the dermis. The molecular weight of fibronectin (220 kDa) coincides with one of the 14-kDa galectin-binding glycoproteins in the dermis. Fibronectin, a major component of the extracellular matrix, has poly- lactosaminoglycan in its structure (Zhu et al. 1984), which is a possible substance recognized and bound by the galectins (Oda and Kasai 1984). Galectins are also shown to recognize laminin (Woo et al. 1990; Cooper et al. 1991; Castronovo et al. 1992; Sato and Hughes 1992). There are also various glycoproteins along the plasma membrane, in the intercellular spaces of the epidermis, and in the extracellular matrix of the dermis (Holton et al. 1990; Kapprell et al. 1985; Koch et al. 1990; Kreis and Vale 1993). Studies are now in progress for identify- ing these glycoproteins recognized by the 14-kDa lectin.

Double staining with anti-14-kDa galectin and anti- C D l a revealed that the 14-kDa galectin is located not only in the keratinocytes but also in the Langerhans cells. The 14-kDa galectin has been shown to have an extensive se- quence similarity with IgE-binding protein (33-kDa galec- tin) of human skin, which is localized on the surface of the Langerhans cells and modulates their binding capacity for IgE (Wollenberg et al. 1993). However, the 14-kDa galec- tin seems to be produced by both keratinocytes and Lan- gerhans cells, whereas the 33-kDa galectin is produced by human keratinocytes and subsequently binds to the surface of Langerhans cells (Wollenberg et al. 1993). From these results, we propose that 14-kDa galectin as well as 33-kDa galectin (IgE-binding protein) may function in modulating the binding capacity of Langerhans cells for glycoconju- gates, which cells are known to be involved in the immune response of skin (Hosoi et al. 1993).

Several galectins have been isolated from human tis- sues and cells (Sparrow et al. 1987; Cherayil et al. 1989; Laing et al. 1989; Gitt et al. 1992; Barondes et al. 1994). However, the physiological functions of these galectins, aside from the function of the 33-kDa IgE-binding ga- lectin, are unclear. In the present study, in the view of the location of the 14-kDa galectin expression in human skin and chick skin, it seems reasonable to suggest that this galectin may play a role in cell-to-cell contact by binding to complementary sugars located on the cell sur- faces of adjacent keratinocytes.

Acknowledgements. The authors wish to express their appreciation to Mr. M. Fukuda, Ms. S. Matsubara, Ms. C. Kamata, Ms. M. Kanai, and Ms. T. Shibata (Kyorin University School of Medi- cine) for their technical assistance. This study was supported in part by grants-in-aid from the Ministry of Education, Science, and Culture of Japan.

References

Akimoto Y, Obinata A, Endo H, Hirano H (1991) Reconstruction of basement membrane in recombinants of epidermis and der- mis of chick embryonic skin in vitro: an electron microscopic study. Anat Rec 231:375-382

Akimoto Y, Kawakami H, Oda Y, Obinata A, Endo H, Kasai K, Hirano H (1992) Changes in expression of the endogenous/3- galactoside-binding 14-kDa lectin of chick embryonic skin during epidermal differentiation. Exp Cell Res 199:297-304

Akimoto Y, Obinata A, Hirabayashi J, Sakakura Y, Endo H, Kasai K, Hirano H (1993) Secretion of endogenous 16-kDa fl-galac- toside-binding lectin from vitamin A-pretreated chick embry- onic cultured skin. Exp Cell Res 205:251-260

Albrandt K, Orida NK, Liu F-T (1987) An IgE-binding protein with a distinctive repetitive sequence and homology with an IgG receptor. Proc Natl Acad Sci USA 84:6859-6863

Allen HJ, Sucato D, Gottstine S, Kisailus E, Nava H, Petrelli N, Castillo N, Wilson D (1991) Localization of endogenous fl-ga- lactoside-binding lectin in human cells and tissues. Tumor Biol 12:52-60

Barnes DM (1988) Orchestrating the sperm-egg summit. Science 239:1091-1092

Barondes SH (1986) Vertebrate lectins: Properties and functions. In: Liener IE, Sharon N, Goldstein IJ (eds) The Lecfins. Aca- demic Press, Orlando, pp 437--466

Barondes SH, Castronovo V, Cooper DNW, Cummings RD, Drickamer K, Feizi T, Gitt MA, Hirabayashi J, Hughes C, Ka- sai K, Leffler H, Liu F-T, Lotan R, Mercurio AM, Monsigny M, Pillai S, Poirer F, Raz A, Rigby PW, Rini JM, Wang JL (1994) Galectins: a family of animal fl-galactoside-binding lectins. Cell 76:597-598

Briles EB, Gregory W, Fletcher R Kornfeld S (1979) Vertebrate lectins: comparison of properties of fl-galactoside-binding lec- tins from tissues of calf and chicken. J Cell Biol 81:528-537

Castronovo V, Luyten F, Brule F, Sobel ME (1992) Identification of a 14-kDa laminin binding protein (HLBP14) in human mel- anoma cells that is identical to the 14-kDa galactoside binding lectin. Arch Biochem Biophys 297:132-138

Cherayil BJ, Weiner SJ, Pillai S (1989) The Mac-2 antigen is a ga- lactose-specific lectin that binds IgE. J Exp Med 170:1959- 1972

Cooper DNW, Massa SM, Barondes SH (1991) Endogenous mus- cle lectin inhibits myoblast adhesion to laminin. J Cell Biol 115:1437-1448

Gitt MA, Massa SM, Leffier H, Barondes SH (1992) Isolation and expression of a gene encoding L-14-II, a new human soluble lactose-binding lectin. J B iol Chem 267:10601-10606

Hirabayashi J, Kasai K (1984) Human placenta fl-galactoside- binding lectin. Purification and some properties. Biochem Biophys Res Commun 122:938-944

Hirabayashi J, Kasai K (1988) Complete amino acid sequence of a fl-galactoside-binding lectin from human placenta. J Biochem 104:1-4

Hirabayashi J, Kasai K (1993) The family of metazoan metal-in- dependent fl-galactoside-binding lectins: structure, function and molecular evolution. Glycobiology 3:297-304

Hirabayashi J, Oda Y, Oohara T, Yamagata T, Kasai K (1987a) Immunological cross-reactivity between fl-galactoside-bind- ing lectins of vertebrates. Possible relationship of antigenici- ty to oligomeric structure. Biochim Biophys Acta 916:321- 327

Hirabayashi J, Kawasaki H, Suzuki K, Kasai K (1987b) Complete amino acid sequence of 14 kDa fl-galactoside-binding lectin of chick embryo. J Biochem 101:775-787

Hirabayashi J, Ayaki H, Soma G, Kasai K (1989) Cloning and nu- cleotide sequence of a full-length eDNA for human 14 kDa fl- galactoside-binding lectin. Biochim Biophys Acta 1008:85-91

Hirano H, Akimoto Y, Kawakami H, Oda Y, Kasai K (1988) Lo- calization of endogenous/3-galactoside-binding lectin in em- bryonic chick epidermis. In: Ohyama M, Muramatsu T (eds) Glycoconjugates in Medicine. Professional Postgraduate Ser- vices, Tokyo, Japan, pp 8-13

Holton JL, Kenny TP, Legan PK, Collins JE, Keen JN, Sharma R, Garrod DR (1990) Desmosomal glycoproteins 2 and 3 (desmocollins) show N-terminal similarity to calcium-depen- dent cell-cell adhesion molecules. J Cell Sci 97:239-246

10

Hosoi J, Murphy GF, Egan CL, Lerner EA, Grabbe S, Asahina A, Granstein RD (1993) Regulation of Langerhans cell function by nerves containing calcitonin gene-related peptide. Nature 363:159-163

Huff JC, Weston WL, Wanda KD (1982) Enhancement of specific immunofluorescent findings with use of a para-phenylenedia- mine mounting buffer. J Invest Dermatol 78:449-450

Kapprell H-P, Cowin E Franke WW, Ponsting H (1985) Biochem- ical characterization of desmosomal proteins isolated from bo- vine muzzle epidermis: amino acid and carbohydrate compo- sition. Eur J Cell Biol 36:217-229

Kawai N, Nishiyama F, Hirano H (1979) Changes of lectin-bind- ing sites on the embryonic muscle cell surface in the develop- ing ascidian, Halocynthia aurantium. Exp Cell Res 122:293- 304

Kobiler D, Beyer EC, Barondes SH (1978) Developmentally regu- lated lectins from chick muscle, brain, and liver have similar chemical and immunological properties. Dev Biol 64:265-272

Koch PJ, Walsh M J, Schmelz M, Goldschmidt MD, Zimbelmann R, Franke WW (1990) Identification of desmoglein, a consti- tutive desmosomal glycoprotein, as a member of the cadherin family of cell adhesion molecules. Eur J Cell Biol 53:1-12

Kreis T, Vale R (1993) Guidebook to the extracellular matrix and adhesion protein. Oxford University Press, Oxford

Laing JG, Robertson MW, Gritzmacher CA, Wang JL, Liu F-T (1989) Biochemical and immunological comparison of carbo- hydrate-binding protein 35 and an IgE-binding protein. J Biol Chem 264:1907-1910

Montelione GT, Callahan S, Podleski TR (1981) Physical and chemical characterization of the major lactose-blockable lec- tin activity from fetal calf skeletal muscle. Biochim Biophys Acta 670:110-123

Nowak TP, Kobiler D, Roel LE, Barondes SH (1977) Develop- mentally regulated lectin from embryonic chick pectoral mus- cle: purification by affinity chromatography. J Biol Chem 252:6026-6030

Oda Y, Kasai K (1984) Photochemical cross-linking of fl-galacto- side-binding lectin to polylactosamino-proteoglycan of chick embryonic skin. Biochem Biophys Res Commun 123:1215- 1220

Oda Y, Ohyama Y, Obinata A, Endo H, Kasai K (1989) Endoge- nous fl-galactoside-binding lectin expression is suppressed in retinol-induced mucous metaplasia of chick embryonic epi- dermis. Exp Cell Res 182:33-43

Ookusa Y, Takata K, Nagashima M, Hirano H (1983) Distribution of glycoconjugates in normal human skin using biotinyl lec- tins and avidin-horseradish peroxidase. Histochemistry 79:1-7

Raz A, Meromsky L, Lotan R (1986) Differential expression of endogenous lectins on the surface of nontumorigenic, tumori- genic, and metastatic cells. Cancer Res 46:3667-3672

Sato S, Hughes RC (1992) Binding specificity of a baby hamster kidney lectin for H type I and II chains, polylactosamine gly- cans, and appropriately glycosylated forms of laminin and fi- bronectin. J Biol Chem 267:6983-6990

Sharon N, Lis H (1989) Lectins as cell recognition molecules. Science 246:227-234

Skrincosky DM, Allen HJ, Bernacki RJ (1993) Galaptin-mediated adhesion of human ovarian carcinoma A121 cells and detec- tion of cellular galaptin-binding glycoproteins. Cancer Res 53:2667-2675

Sparrow CE Leffier H, Barondes SH (1987) Multiple soluble fl- galactoside-binding lectins from human lung. J Biol Chem 262:7383-7390

Takata K, Hirano H (1983) Changes in soybean agglutinin (SBA) and peanut agglutinin (PNA) binding pattern in the epidermis of the developing chick embryo. Develop Growth Differ 25:299-305

Tokuyasu KT (1980) Immunocytochemistry on ultrathin frozen sections. Histochem J 12:381-403

Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyaerylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76:4350-4354

Williamson DH, Fennell DJ (1974) The use of fluorescent DNA- binding agent for detecting and separating yeast mitochondrial DNA. Methods Cell Biol 12:335-351

Wollenberg A, Salle H, Hanau D, Liu F-T, Bieber T (1993) Hu- man keratinocytes release the endogenous fl-galactoside- binding soluble lectin immunoglobulin E (IgE-binding pro- tein) which binds to Langerhans cells where it modulates their binding capacity for IgE glycoform. J Exp Med 178: 777-785

Woo H-J, Shaw LM, Messier JM, Mercurio AM (1990) The major non-integrin laminin binding protein of macrophages is identi- cal to carbohydrate binding protein 35 (Mac-2). J Biol Chem 265:7097-7099

Zhu BC-R, Fischer SF, Pande H, Calaycay J, Shively JE, Laine RA (1984) Human placental (fetal) fibronectin: increased gly- cosylation and higher protease resistance than plasma fibro- nectin. J Biol Chem 259:3962-3970