1. Department of Dermatology, Faculty of Medicine, Kyushu Univeristy,Fukuoka, Japan
2. ^*Skin Care Division, R&D Headquarters, Sunster, Asahi-machi, Takatsuki, Osaka, Japan

Correspondence: Dr Shuhei Imayama, Department of Dermatology, Faculty of Medicine, Kyushu Univeristy, Fukuoka, Japan

Received 14 August 1998; Revised 19 August 1998; Accepted 19 August 1998.

Epidermal Langerhans cells originate in the bone marrow and stay for a certain period of time in the epidermis. As antigen-presenting cells, they are believed to capture and process foreign antigens that invade into the epidermis, move to the regional lymph nodes, and then present antigenic peptides with major histocompatibility complex to specific T-cells (^{Kripke et al. 1990;Barratt-Boyes et al. 1997}). In order to immigrate into the epidermis and then emigrate towards the lymph nodes, Langerhans cells must cross the basement membrane of the epidermis and move through the connective tissue-rich underlying dermis. Recent studies have shown that 6 integrin and tumor necrosis factor- regulate the migration of epidermal Langerhans cells (^{Cumberbatch & Kimber 1992;Price et al. 1997}). We have previously shown that the epicutaneous application of haptens stimulates the secretion of matrix metalloproteinase (MMP)-9 from Langerhans cell-enriched murine epidermal cells (^{Kobayashi 1997}).

We evaluated the normal skin specimens obtained from 13 patients, aged 25–72 y, who underwent an excision of benign tumors (including seborrheic keratosis, epidermoid cyst, and so on). The specimens, after removing fat tissue, were cut into 3 3 mm pieces and incubated in Roswell Park Memorial Institute medium-1640 (GIBCO, Grand Island, NY) containing 10% fetal bovine serum (Dainippon Pharmaceutical, Osaka, Japan) and 1 10^–5 M monensin (Sigma, St. Louis, MO) for 4 h at 37°C to accumulate intracellular MMP-9. The specimens were then treated with 1 10³ U Dispase (Godo Shusei, Tokyo, Japan) per ml in minimum essential medium (GIBCO) for 5 h at 25°C to separate the epidermis from the dermal connective tissue. The split sheets of the epidermis thus obtained were fixed with acetone for 3 min at 4°C, and then incubated in the phosphate-buffered saline containing 0.4% human AB serum (Cosmo Bio, Tokyo, Japan) and 0.6% bovine serum albumin (Sigma) for 10 h at 4°C to block nonspecific binding of immunoglobulins. They were then incubated with monoclonal mouse anti-human MMP-9 antibody (Fuji Chemical, Toyama, Japan) diluted 1:250 in phosphate-buffered saline for 3 h at 25°C. Some of them were further incubated with biotin-labeled rabbit anti-mouse immunoglobulin A, G, M antibody (Nichirei, Tokyo, Japan), and then with fluorescein isothiocyanate-conjugated streptoavidin (GIBCO). The others were incubated with rhodamine-conjugated goat anti-mouse immunoglobulin G antibody (Kirkegaard & Perry Laboratories, Gaithersburg, MD), and then with fluorescein isothiocyanate-conjugated monoclonal anti-CD1a antibody (DAKO, Glostrup, Denmark). We also stained other sheets with monoclonal mouse anti-human MMP-1 and -2 antibody (Fuji Chemical) in the same way. The sheets were then observed under a confocal laser microscope (Olympus, Tokyo, Japan).

We observed numerous MMP-9 positive cells to be present in the epidermal sheets. MMP-9 was present in the granules measuring from 100 to 200 nm in diameter in the cytoplasm (Figure 1). The cells appeared to be dendritic and were located within the epidermal sheets with a relatively uniform density. Double labeling with anti-MMP-9 antibody and anti-CD1a antibody demonstrated the MMP-9 positive cells to be consistently positive for CD1a (Figure 2). These findings indicate many dendritic cells to express MMP-9 in the cytoplasm, whereas the cells also have morphologic and immunologic features of Langerhans cell in the epidermis. CD1a is the most reliable and common marker of Langerhans cell in the epidermis. MMP-1 was expressed by basal keratinocytes, but not by Langerhans cells. MMP-2 positive cells were not detected in the epidermal sheets.

Figure 1.

Single stain for MMP-9. Typical dendritic cells positive for anti-MMP-9 antibody in the epidermal sheet. MMP-9 appears to be present in the cytoplasmic granules. Scale bar: 2 m.

Full figure and legend (80K)

Figure 2.

Double stain for MMP-9 and CD1a. MMP-9-bearing cells (A) also presenting CD1a (B), whereas the distribution of MMP-9 and CD1a differs (C). Scale bar: (A, B) 10 m, (C) 1 m.

Full figure and legend (84K)

MMP are subdivided into collagenases (MMP-1, -8, -13, -18), gelatinases (MMP-2, -9), stromelysins (MMP-3, -10, -11), matrilysin (MMP-7), metalloelastase (MMP-12), and membrane-type MMP (MMP-14, -15, -16, -17), based on the primary structures and the specificity for matrices (^{Parsons et al. 1997}). Malignant neoplasms are known to produce some of the enzymes to degrade the basement membrane and the connective tissue matrices, and the activity of the enzymes thus participate in invasive growth (^{Karelina et al. 1993}). Gelatinases, such as MMP-9, can degrade gelatin, type IV collagen, elastin, etc.; which are all essential components of the matrices of the basement membrane and the dermal connective tissue.

The presence of MMP-9 in the CD1a positive cells in the epidermis after short-term organ culture supports the hypothesis that MMP-9 participates in migration through the basement membrane and the connective tissue.