JOURNAL OF PATHOLOGY, VOL.

166: 243-253 (1992)

A NOVEL LAMINA LUCIDA COMPONENT OF EPITHELIAL AND ENDOTHELIAL BASEMENT MEMBRANES DETECTED BY LH39 MONOCLONAL ANTIBODY BERNICE M. ALMEIDA*, STEPHEN J. CHALLACOMBE* ,JOHN IRENE M. LEIGH9

w. EVESON?,

COLIN G. SMITH$ AND

*Department of Oral Medicine and Pathology, U M D S , Guy's Hospital, London SEI 9RT, U.K.; ?Department of Oral Medicine, Surgery and Pathology, Bristol Dental Hospital and School, Lower Maudlin Street, Bristol BSI 2 L Y , U.K.; 1Unilever Research Laboratories, Sharnbrook Bedford MK44 I L Q , U.K.; §E.xperimental Dermatology Laboratory, The Royal London Hospital, London E l IBB, U.K. Received 10 May 1991 Accepred I0 July 1991

SUMMARY The murine monoclonal antibody, LH39 was characterized in this study and appeared to bind to a novel basement membrane epitope. This antigen was expressed in the epithelial basement membrane of human tissue derived from all three germ cell layers and in basement membranes surrounding small blood vessels within the stroma of all organs examined. LH39 antigen could be first detected in fetal skin at the dermo-epidermal junction at 7 weeks estimated gestational age but was not present in the dermal vasculature until 16 weeks. When tested against tissue from a range of lower mammalian species, LH39 antigen appeared to be primate-specific. The epithelial basement membrane zone in organotypical cultures, where there is de novo synthesis of basement membrane components, contained abundant LH39 antigen in contrast to other basement membrane components, type IV collagen, laminin, and type VII collagen. Ultrastructural localization of LH39 epitope, using immunogold electron microscopy on unfixed freshly frozen tissue, was to the lamina lucida. No cross-reactivity could be detected between LH39 and laminin, fibronectin, and collagens I, 111, IV, and V using the ELISA assay. In vitro studies with a range of proteolytic enzymes suggested that the antigen was non-collagenous in nature. LH39 precipitated a polypeptide with a molecular weight of 185 kD from extracts of metabolically labelled cultured keratinocytes, and polypeptides of 185 and 200 k D from the culture medium. The tissue distribution of LH39 antigen suggested that it may be an epitope within anchoring filaments. Potential applications of this antibody include the study of benign and malignant human vascular disorders, diseases and tumours associated with angiogenesis, epithelial neoplasms, and conditions of tissue regeneration and repair, such as wound healing. KEY

WORDS-Basement membranes, blood vessels, LH39 monoclonal antibody, immunocytochemistry.

INTRODUCTION

Basement membranes (BMs) are deposited by epithelial, endothelial, and some mesenchymal cells and are complex, highly compartmentalized structures composed of three major structural zones.' The electron-lucent lamina lucida lies immediately subjacent to the basal cell plasma membrane and contains the anchoring filaments, which are more Addressee for correspondence: Dr B. M. Almeida, Department of Oral Medicine and Pathology, UMDS, Guy's Hospital, London Bridge, SEI 9RT, U.K.

0022-3417/92/030243-11 $05.50 0 1992 by John Wiley & Sons, Ltd.

numerous in the region of the hemidesmosomes.2 Beneath the lamina lucida lies the lamina densa, which in turn lies on the lamina reticularis. Many distinct BM antigens, which show ultrastructural stratification, have been identified using polyclonal and monoclonal a n t i b ~ d i e s . ~ Major structural proteins of BMs include type IV collagen, which is thought to form the structural backbone upon which other components are assembled: laminin, which forms a second polymer network; and entactin/nidogen, which bridges type IV collagen and laminin.' Heparan sulphate proteoglycans are thought to control macromolecular permeability

244

B. M. ALMEIDA ET AL.

in capilliaries6 and osteonectin/BM-40/SPARC oesophagus, stomach, colon, thyroid, parathyroid, possibly plays a role in stabilizing matrix s t r ~ c t u r e . ~adrenal, testis, ovary, kidney, ureter, brain, periThe molecular composition and supramolecular pheral nerve, liver, spleen, thymus, and striated arrangement of BMs vary in different sites accord- muscle. Twenty skin specimens were obtained from ing to functional requiremenk8 For example, electively aborted normal human fetuses ranging in specialized filtration and diffusion properties are estimated gestational age from 7 to 16 weeks. Skin conferred by fusion of epithelial and endothelial was obtained from the following lower mammals BMs in renal glomeruli and pulmonary alveoli4 according to established guidelines for the humane Although heterogeneity in the composition of BMs use of laboratory animals: monkey, marmoset, has been demonstrated by the production of mono- guinea pig, calf, and pig. All tissue was obtained fresh, embedded in OCT clonal antibodies with selective reactivity for BMs at different sites,'-" the biochemical basis for these compound (Miles Laboratories) on cork strips, findings is not known. BMs separate different body snap-frozen in isopentane (2-methyl butane) compartments and in this capacity influence the (B.H.H. Ltd.), precooled in liquid nitrogen, and exchange of cells and macromolecules across tis- stored under liquid nitrogen (N,) at - 190°C until sues. Their influence on the spatial orientation of used. cells and on the orderly development of distinct tissue patterns is important in embryogenesi~,'~-'~ Organotypical cultures tissue regeneration and repair, and wound healing. Recombinants of sheets of human epidermal cells Components of BMs have been implicated in various disease processes such as epidermolysis and dermis were prepared,' by growing epithelial bullosa, where absence of anchoring fibrils has been cells on type I collagen gel diaphragms at 37°C in an identified,I6 and Goodpasture's syndrome, where atmosphere of 90 per cent 0, and 10 per cent CO, antibodies to the a3 chain of type IV collagen have for 24 h and then apposing dermal sheets to the been implicated in the path0gene~is.I~ BMs are undersurface of the diaphragm prior to culture for also thought to be the first barrier that a malignant 21 days. These were then orientated, embedded in tumour, such as a carcinoma, must transgress in OCT compound, snap-frozen in precooled isopentane, and stored under liquid N,. order to become infiltrative." One of the most promising approaches to the identification of new BM components is the inducSplit skin preparation tion of antibodies against extracts of BM and Fresh intact human neonatal foreskin was subsequent identification of the target antigen. Diagnostic uses for these antibodies have been surgically stripped of excess subcutaneous fat and identified.'' In the present study we describe the dermis, and 5 mm cubes were incubated in 1 M NaCl characterization and identify potential applications at 4°C for 72 h to produce an intralamina lucida of a new BM antibody to a novel lamina lucida split.,' After gentle rinsing in phosphate-buffered epitope in epithelial and endothelial basement saline (PBS), the tissue was orientated, embedded membranes. Techniques used include immunocyto- in OCT compound, snap-frozen in precooled isochemistry, immunoelectron microscopy, enzyme pentane, and stored under liquid N,. predigestion, enzyme linked immunosorbent assays (ELISA), and immunoprecipitation. Monoclonal antibody product ion MATERIALS AND METHODS Tissues The following adult human tissue was obtained at surgery: skin, buccal, lingual, gingival and labial mucosa, parotid and submandibular gland, lymph node, tonsil, Fallopian tube, endometrium, cervix, and breast. Neonatal foreskin was retrieved after elective circumcision. Tissue obtained at autopsy included myocardium, pericardium, larynx, lung,

A single cell suspension of epidermal cells was obtained by treating 2-3 mm fragments of neonatal foreskin with 0.025 per cent trypsin at 4°C for 16 h before physical stripping. Cells were lysed in 1 per cent Nonident P40 (BDH) in phosphate-buffered saline and the insoluble pellet was sonicated to prepare insoluble fractions. BALB/c mice were immunized intraperitoneally at 6-week intervals to a total of three doses with a final intravenous boost given 72 h prior to fusion. Cell hybridization was achieved using a previously

LH39 ANTIBODY TO NOVEL BASEMENT MEMBRANE ANTIGEN

described technique.22Isotyping was performed by Ouchterlony double immunodiffusion assays in agarose using antisera specific for IgM, IgA, IgG, and IgG subclasses (Bionetics). Imrnunoperoxidase staining

Specificities in adult human organs, fetal skin, skin from lower mammal species, split skin preparation, and organotypical cultures were determined using a three-stage avidin-biotin-peroxidase complex technique.23 The distribution of LH39 epitope was compared with that of other basement membrane components using antibodies to type IV collagen (Bio-nuclear services), laminin (Serotec), and type VII collagen (LH7.222). Five pm thick adjacent serial sections from the unfixed freshly frozen specimens were mounted on chrome gelatincoated glass microscope slides, dried overnight at room temperature, and fixed in precooled acetone at 4°C for 10 min. Primary antibodies were used in previously determined optimal dilutions (LH39 neat, anti-collagen IV 1:5, anti-laminin 1:20, and LH7.2 1:lo), followed by biotinylated secondary antibodies and streptavidin biotinylated horseradish peroxidase with 3',3'-diaminobenzidine tetrahydrochloride as substrate. Negative controls consisted of substitution of the primary antibody with an irrelevant antibody of the same subclass or non-immune serum at the same dilution and substitution of each layer in turn with PBS.

245

were mounted in aqueous UV-inert mountant and viewed with an Olympus BH2 microscope with an RFL fluorescence attachment and a model PMIOADS Olympus automatic photomicrographic system. An HBO 100 W super pressure mercury lamp was used with a BP-490 filter and a DM-500 0.515 dichroic mirror for blue excitation. A BP-545 filter and DM 580 + 0.590 dichroic mirror were used for green excitation.

+

Immunogold electron microscopy Indirect immunogold electron microscopy was performed on I0 pm thick unfixed cryostat sections mounted on chrome gelatin-coated melinex strips. Sections were incubated serially with neat LH39 for 60 min, biotinylated rabbit anti-mouse immunoglobulin (Dako Ltd.) diluted 1:200 for 60 min, and streptavidin 5 nm gold (Janssen) diluted 1:40 for 60 min. Colloidal gold staining was silverenhanced using equal volumes of initiator and developer (Janssen, INTENSESEM)mixed immediately before use and incubating with the sections for 20 min. Sections were fixed in 2.5 per cent glutaraldehyde, post-fixed in 2 per cent aqueous osmium tetroxide, dehydrated through graded alcohols, and embedded in TAAB resin. Ultrathin sections were counterstained with lead citrate and examined using a Jeol 100CX2 electron microscope. Controls consisted of substitution of the primary antibody with non-immune serum or an irrelevant antibody of the same subclass and substitution of each layer in turn with PBS.

Double immunojluorescence microscopy A sequential double immunofluorescence label- Enzyme digestion of tissue sections ling technique utilizing LH39, anti-collagen IV, The relative sensitivity of LH39 antigen to and anti-laminin to label basement membranes and enzyme digestion was compared with that of type Factor VIII related antigen (Dako Ltd.) and Ulex IV collagen by preincubating 5 pm thick unfixed europaeus I (Sigma) to label vascular endothelium cryostat sections of human oral mucosa with the was performed. Duplicate 5 pm adjacent serial following enzymes:23 pepsin (100 pg/ml in 0.5 M cryostat sections were fixed in precooled acetone at acetic acid in saline, 20 rnin at room temperature), 4°C for 10 rnin prior to a 60 rnin incubation with pronase (0.01 per cent in PBS, 20 rnin at room LH39 and antibodies to type IV collagen and temperature), trypsin (0.025 per cent in 0.05 M Trislaminin. Sections labelled with fluorescein conju- HCI, pH 7.6, with 0.15 M NaCl and 0.001 M CaCl,, gated Ulex europaeus I (12.5 pg/ml; 60 min) were 20 min at room temperature), crude bacterial collafirst incubated with rhodamine conjugated rabbit genase (100 units/ml in 0.025 M Tris-HC1, pH 7.4, anti-mouse immunoglobulin (1 :20, 60 min). Label- with 0.01 M CaCI,, at room temperature), lipase (250 ling with anti-Factor VIII related antigen (1: 1000, units/ml PBS, 3 h at room temperature), hyaluroni30 min) and rhodamine conjugated swine anti- dase (230 units/ml PBS, 2 h at room temperature), rabbit immunoglobulin (1:40; 30 min) was preceded chondroitinase (0.05 units/section in chondroitiby serial incubations with biotinylated sheep nase buffer, 2 h at room temperature, and neuranti-mouse immunoglobulin (1 :25; 60 min) and aminidase (1 U/ml, 30 min at room temperature). fluorescein streptavidin (1 :20; 60 min). Sections All enzymes were obtained from Sigma aside from

246

B. M. ALMEIDA ET A L .

chondroitinase (ICN) and neuraminidase (KochLight Ltd.). Sections were then stained using a standard immunoperoxidase protocol, counterstained with Gill’s haematoxylin, dehydrated, cleared, and mounted in DPX. Enzyme linked immunosorbent assays (ELISA) An indirect antibody ELISAZ4was established for the following antigenic substrates: laminin (Sigma), fibronectin (Sigma), collagens I (Southern Biotechnology Associates, Inc.), 111, IV, and V (Sigma); and was used to examine hybridoma supernatant containing LH39 monoclonal antibody. The assay consisted of serial incubations involving specific primary antibodies to each of the antigens and serially diluted LH39 monoclonal antibody, species-specific biotinylated secondary antibodies, applied in previously determined optimal dilutions, and streptavidin biotinylated horseradish peroxidase. 0-Phenylenediamine was used as substrate and optical densities were measured in an ELISA plate reader with a filter setting of 492 nm. Immunoprecipitat ion Monolayers of cultured keratinocytes grown to between 1/2 or 2/3 confluence were metabolically labelled with ,’S methionine (25 pCi/ml; 18 h). The medium was decanted into sterile containers and stored at 4°C. Cell lysis and immunoprecipitation were performed using a standard protocol.2s Cell monolayers were harvested by centrifugation and soluble extracts prepared by 30 min incubation at 4°C in lysis buffer (9.5 ml NET, 0.5 ml NP40, and 1 mM PMSF). Nuclear debris was pelleted using a Sorvall centifruge at 18 000 rpm for 60 min and the cell lysate precleared by incubating for 30 min with a I0 per cent suspension of Staphylococcus aureus cells (Staph A) (Calbiochem) which had been washed in NET (SO mM Tris, pH 7-4,150 mM NaCI, 1 mM EDTA, and 0.02 per cent NaN,) prior to resuspension in the same buffer. To 125 p1 aliquots of cell lysate diluted with an equal volume of NET buffer and 2 5 0 ~ 1of medium (in which cells had been labelled), 200p1 aliquots of LH39 (as neat hybridoma supernatant) were added and incubated for 2 h at 4°C. Simultaneously, Staph A cells were loaded with 100 pl/ml rabbit anti-mouse immunoglobulin (Dako Ltd.) and incubated for 30 min. Antigen-antibody complexes were precipitated by adding 60 pl of rabbit anti-mouse Staph A complex and incubating overnight (14 h) at 4°C. Precipitates

were washed twice in RIPA buffer [50 mM Tris, pH 7.4, 1 SO mM NaCl, 1 per cent Triton X-100, 1 per cent sodium deoxycholate, and 0.1 per cent sodium dodecyl sulphate (SDS)] and once in NET buffer. Specific immunoprecipitates were eluted by boiling for 5 min in sample buffer, pH 6.8, and analysed by SDS-PAGE using an acrylamide concentration of 7.5 per cent.26Gels were fluorographed and exposed to preflashed Kodak XAR5 film at -70°C for 3 days.*’ I4C-Labelled molecular weight standards were run in continuous lanes. As a control, an irrelevant murine antibody of the same isotype was substituted for LH39. RESULTS Tissue distribution of LH39 ( a ) Adult human tissue-A standard hybridoma technique was used to produce LH39, an IgG, monoclonal antibody. LH39 monoclonal antibody labelled only epithelial and vascular BMs. The distribution of LH39 antigen in adult human tissue is detailed in Table I. Organs expressing this antigen within BMs at epithelialkonnective junctions were derived from ectoderm, mesoderm, and endoderm. The pattern of staining with LH39 around hair follicles and sebaceous glands in skin differed from that of the other BM antigens, with distinct attenuation at the lower part of both the hair follicle and the sebaceousgland (Fig. 1). BMs aroundcapillariesand veins within the stroma of all organs examined were stained with LH39 monoclonal antibody, including those showing no labelling of the epithelialconnective tissue interface. Tissue showing no labelling of the epithelial basement membrane was derived from cardiovascular, respiratory, endocrine, gastrointestinal, genitourinary, reticuloendothelial, and nervous systems, and represented all three germ cell layers (Fig. 2). (b) Fetal tissue-LH39 antigen was detectable at the dermo-epidermal junction (DEJ) of fetal skin at 7 weeks estimated gestational age, which was the earliest fetal tissue available. It appeared in dermal vascular basement membranes only at 16 weeks estimated gestational age. The intensity of staining at the DEJ increased with the age of the fetus. The antigen was also present in amniotic epithelial basement membrane. (c) Lower mammal skin-Expression of LH39 antigen was restricted to the primate species

LH39 ANTIBODY TO NOVEL BASEMENT MEMBRANE ANTIGEN

Table I-Distribution chemistry

247

of LH39 antigen in adult human tissue demonstrated by immunocyto-

Tissues showing staining of epithelial basement membranes with LH39 antibody

Tissues showing restricted reactivity with LH39 antibody

~

Skin Dermo-epidermal junction Hair follicle epithelium Sebaceous glands Sweat gland ducts and acini Oral epithelium (multiple sites) Lingual mucosa Buccal mucosa Gingivae Salivary glands Parotid Submandibular (Ducts and acini equally) Oesophagus (upper third) Larynx Breast Fallopian tube Endometrium and ectocervix

Stomach and colon Myocardium Pericardium Lung Pituitary and thyroid Parathyroid and adrenal Ovary and testis Kidney and ureter Urinary bladder Brain Peripheral nerve Liver and spleen Lymph node Tonsil and thymus Striated muscle Prostate Epididymis

LH39 monoclonal antibody stained small blood vessels within the stroma of all organs examined

(monkey and marmoset). The distribution was similar to that demonstrated in human skin and differed from that of the lamina lucida (laminin) and two lamina densa (collagens IV and VII) basement membrane components examined (Table 11). There was no detectable staining of the BM zone in the other lower mammalian species examined. ( d ) Organotypical cultures-LH39 antigen was strongly expressed within the epithelial basement membrane zone of all sections of organotypical cultures examined and individual cultures had from 2 to 7 keratinocyte cell layers (Fig. 3). It was also detected within some basal keratinocytes. In contrast, there was weak variable expression of type IV collagen, laminin, and LH7.2 antigen. Double immunofluorescence labelling

LH39 antigen was detected within vascular basement membranes and the distribution was similar to that of type IV collagen and laminin, though less intense than laminin. However, unlike type IV collagen and laminin, LH39 only labelled endothelial BM surrounding capillaries and veins, most of which were laminated (Fig. 4). Factor VIII

related antigen and Ulex europaeus I demonstrated vascular endothelium. Ultrastructural localization On intralamina lucida separated human neonatal foreskin, LH39 staining was seen along the dermal floor (Fig. 5). By immunogold electron microscopy the distribution o f L H 3 9 antigen was predominantly within the lower lamina lucida with some labelling in the upper lamina lucida (Fig. 6). Labelling appeared most intense in the region beneath the hemidesmosomes. EfSects of enzyme digestion In contrast to type IV collagen, expression of LH39 antigen was eliminated following exposure to pepsin, as were laminin and LH7.2 antigen. There was no diminution of binding of LH39 antibody after preincubation with bacterial collagenase.

ELISA No cross-reactivity was detected between LH39 monoclonal antibody and laminin, fibronectin, and collagens I, 111, TV, and V. The collagens were all

248

B. M. ALMEIDA ET AL.

LH39 ANTIBODY TO NOVEL BASEMENT MEMBRANE ANTIGEN

derived from human tissue; laminin was prepared from the BM of rat EHS sarcoma and fibronectin was of bovine origin. Immunoprecipitation

LH39 specifically precipitated a polypeptide with a molecular mass of 185 kD from extracts of metabolically labelled cultured keratinocytes. Polypeptides with molecular masses of 185 and 200 kD were precipitated by LH39 from the medium in which the cells were incubated during metabolic labelling and which presumably contained secretory products of the cultured keratinocytes (Fig. 7). DISCUSSION Several distinct antigenic components have been identified within the lamina lucida of the basement membrane zone. Of these, the bullous pemphigoid antigen has been shown to be associated with hemidesmosomes,28 and 19-DEJ-1 with either the anchoring filaments or the sub-basal dense platc2’ On the basis of our findings, it appears that LH39 monoclonal antibody recognizes a unique antigenic epitope within the lamina lucida of basement membranes of human tissue, labelling the BM at the epithelial-connective tissue interface of organs derived from all three germ cell layers as well as capillary and venous BMs. Monoclonal antibodies to laminin have been described showing restricted tissue reactivity. Anti-laminin antibody 4E 10 labelled the subendothelial basement membrane of capillaries and veins but not of arterioles or arteries.’ However, unlike LH39, this antibody also reacted with perineurium, intraneurium, and smooth muscle, in addition to epithelial BM, in the tissue examined. Expression of bullous pemphigoid and cicatricial pemphigoid antigens is restricted to the epithelial-connective tissue interfa~e,~’ while 19-DEJ-1 antigen is present in BM surrounding smooth muscle cells29and monoclonal antibodies AA33’ and GB332 recognize epitopes only in the dermo-epidermal junction and adnexal BM. Detection of LH39 antigen in the epithelial BM of fetal skin at 7 weeks estimated gestational age corresponds to the detection, by electron

249

microscopy, of the lamina lucida and anchoring filam e n t ~ This . ~ ~ contrasts with the detection of the hemidesmosome associated bullous pemphigoid antigen at 10 weeks estimated gestational age. A subendothelial BM is not present in the embryonic vasculature; it appears in the early fetal vasculature (6G-80days) but shows features of maturation, such as multilaminated layers in the venous segment, only in the second trimester.34 The detection of LH39 antigen in second trimester fetal dermal vasculature suggests that the epitope either first appears or is only fully assembled in more mature vascular BM. LH39 antigen is the only lamina lucida component with a distinctive primate specificity. Other primate-specific monoclonal antibodies include KF-l,35AF-1, and AF-2,36which label lamina densa epitopes. Further studies will be necessary to determine whether these findings reflect the detection of a species-restricted epitope or a new antigen. The unique distribution of LH39 antigen in organotypical cultures with consistently strong expression in cultures of all ages contrasts with that reported for other components of BMs. In cultures using mouse keratinocytes, bullous pemphigoid antigen was transiently expressed up to the first 4 days and laminin and type IV collagen showed continuous linear staining only after this time.37The appearance of these epitopes in organ cultures of human skin differed in that bullous pemphigoid antigen was consistently found beneath the epibolic epithelium, though the distribution was patchy.38. Using indirect immunogold electron microscopy, LH39 antigen was localized predominantly within the lower lamina lucida, with some labelling in the upper lamina lucida. The binding specificity was concentrated to regions underlying the hemidesmosomes. This localization was confirmed using 1 M NaCl split skin, in which the antibody bound along the dermal floor. Unlike the bullous pemphigoid antigen, LH39 antigen did not appear to be concentrated within the hemidesmosomes themselves. 19DEF-1 antigen is present in the midlamina lucida2’ while laminin, which was thought to be exclusive to the lamina l ~ c i d a , ~has ’ also been demonstrated within the lamina densa using immunogold electron m i c r o s ~ o p y .We ~ ~ have been unable to localize LH39 antigen ultrastructurally in vascular BMs

Fig. 1-Immunoperoxidase staining of human skin. (a, b) Monoclonal antibody to type IV collagen which labels all basement membranes. (c, d) LH39 monoclonal antibody which labels only basement membranes at the demo-epidermal junction and around adnexae, veins, and capillaries. There is attenuated staining around the lower part of the hair follicles and sebaceous glands

250

B. M. ALMElDA ET AL.

Fig. 2-Immunoperoxidase staining of thyroid gland. (a) Monoclonal antibody to type IV collagen which stains all basement membranes. (b) LH39 monoclonal antibody which reacts only with vascular basement membranes

Table 11-Comparison of LH39 antibody reactivity in lower mammal species with other basement membrane components

Mammal Monkey Marmoset Guinea pig Pig Calf

Collagen IV

Antibody Laminin LH7.2

++* +* +-t

++* ++* +§

+++I1

-

++t

LH39

++t +++t ++t +++t

++$ ++$ -

-

+++t

+ + + =Strong staining; + + =moderate staining; + =weak staining;

-

~

= n o staining.

*All basement membranes. ?Basement membrane at dermo-epidermal junction and around hair follicles. $Basement membrane at dermo-epidermal junction, around hair follicle epithelium, and small blood vessels. $Basement membrane at dermo-epidermal junction. IIBasement membranes at dermo-epidermal junction, around hair follicle epithelium, and blood vessels.

Fig. 3-Immunoperoxidase staining of a section of an organotypical culture with LH39 monoclonal antibody showing prominent staining of the epithelial basement membrane zone

out its presence within these structures has been demonstrated using double immunofluorescence microscopy and antibodies specific for vascular endot he1ium. The sensitivity of LH39 antigen to pepsin treatment contrasted with the resistance of type IV collagen and suggests that the antigen is peripherally located, unlike type IV collagen which is more centrally located and thus less prone to enzymatic removal or degradation. Lack of diminution of expression of LH39 antigen in tissue following prolonged exposure to bacterial collagenase suggests that it is non-collagenous in nature. However, the possibility exists that the collagenous component may be masked by non-colkenous subunits Or that it may be resistant to this particular collagenase.

25 1

LH39 ANTIBODY TO NOVEL BASEMENT MEMBRANE ANTIGEN

Fig. G-Immunofluorescence labelling of human skin. (a) LH39 monoclonal antibody (and FITC) which decorates a laminated vascular BM. (b) Factor VIII related antigen (and TRITC) labels endothelial cells, in the same vessel, with a predominantly granular distribution a

185

NaC1human split skin. ImmunoPeroxidase staining Fig. 5-1 with LH39 monoclonal antibody which is concentrated at the base of the split

Fig. ~ I m m ~ n o g o electron ld microscoPY on unfixed human oral mucosa using LH39 monoclonal antibody, 5 nm gold, and silver enhancement. Gold particles are concentrated within the lower lamina lucida with some labelling of the upper lamina lucida in the region of the hemidesmosomes (arrow)

D

b

4

200

4

185

Fig. 7-Immunoprecipitation with LH39 monoclonal antibody of polypeptides extracted and secreted from cultured human keratinocytes. (a) 185 kD polypeptide precipitated from cell extract; (b) 185 and 200 kD polypeptides precipitated from cell medium

LH39 monoclonal antibody showed no crossreactivity on ELISA with basement membrane antigens laminin, fibronectin, and type IV collagen or with collagens I, 111, and V. Both laminin (BM of rat EHS sarcoma) and fibronectin (bovine) were of non-human origin. The lack of any previous evidence of restricted species specificities for fibronectin suggests that the negative results on ELISA are due to true lack of cross-reactivity rather than a reflection of species specificity. Although laminin epitopes have recently been demonstrated to show species restriction," the unique tissue distribution of LH39 antigen clearly distinguishes it. Biochemical characterization of LH39 antigen by immunoprecipitation suggests that it consists of at least two polypeptides with molecular weights of approximately 185 and 200 kD, produced by human epithelial cells. The 185 kD Polypeptide was detected only in the secreted form. Molecular

252

B. M. ALMEIDA ET AL.

heterogeneity of the bullous pemphigoid antigens has been demonstrated with antibodies from patients with bullous pemphigoid reacting with antigens of molecular weights 180-200 kD.4' The tissue distribution of LH39 antigen, however, is distinct from that demonstrated for the bullous pemphigoid antigens. On the basis of our findings, LH39 monoclonal antibody appears to define a new BM antigen present within the lamina lucida of the epithelialconnective interface of human organs derived from all three germ cell layers as well as in mature capillary and venous BMs. The epitope is primatespecific and its detection at the dermo-epidermal junction of fetal skin coincides with the ultrastructural appearance of the lamina lucida and anchoring filaments. It was only expressed in the fetal dermal vasculature during the second trimester when ultrastructurally it is assuming the features of the more mature adult dermal vasculature. This finding is substantiated by studies in pathological conditions associated with new vessel formation, where LH39 antigen is absent in newly formed blood vessels.42 The early appearance of LH39 antigen in the BM zone of organotypical cultures is comparable to the enhanced labelling seen beneath proliferating/ regenerating epithelium in pathologically altered tissue42 and suggests that LH39 antigen may be involved in epithelial cell adhesion to the extracellular matrix. Basement membranes often show characteristic variations in morphology or chemical composition in relation to cell or tissue type or to specific pathological conditions. In this study, using the new monoclonal antibody LH39 we have demonstrated the heterogeneity of BMs in a range of normal tissues. Potential applications of this antibody include the study of this heterogeneity in diseases and tumours associated with angiogenesis, epithelial neoplasms, benign and malignant human vascular disorders, and conditions of tissue regeneration and repair such as wound healing.

ACKNOWLEDGEMENTS

This work was supported by a grant from the Frances and Augustus Newman Foundation, which is acknowledged with gratitude. We thank Dr R. Millis for provision of some of the adult human tissue, Dr N. Johnson for tissue from, the cynomologus monkeys, and Dr F. Compton for assistance

with obtaining the fetal tissue. Thanks are also due to Mr A. Woodman for advice and help with technical work and photography.

REFERENCES 1. Laurie GW, Leblond CP. Basement membrane nomenclature. Nature

1985; 313: 272. 2. Eady RAJ. The basement membrane. Arch Derniatol 1988; 124: 709-7 12. 3. Fine JD. Antigenic features and structural correlates of basement membranes. Arch Dermatol1988; 124: 713-717. 4. Martinez-Hernandez A, Amenta PS. The basement membrane in pathology. Lab Invest 1983; 48:65C678. 5. Yurchenco PD, Schittny JC. MokCubdr architecture of basement membranes. FASEB J 1990; 4 1577-1 590. 6. Schleicher ED, Wagner E-M, Olgemoller B, Nerlich AG, Gerbitz KD. Characterisation and localisation of basement membrane-associated heparan sulfate proteoglycan in human tissues. Lab Invest 1989; 61: 323-332. 7. Timpl R. Structure and biological activity of basement membrane proteins. Eur J Biochem 1989; 180 487-502. 8. Form DM, Pratt BM, Madri JA. Endothelial cell proliferation during angiogenesis. In vitro modulation of basement membrane components. Lab Invest 1986; 5 5 521-530. 9. Leu F, Engvall E, Damjanov I. Heterogeneity of basement membranes of the human genitourinary tract revealed by sequential immunofluorescence staining with monoclonal antibodies to laminin. J Hutochem Cytochem 1986; 34: 483489. 10. Wan YJ, Wu TC, Chung AE, Damjanov 1. Monoclonal antibodies to laminin reveal the heterogeneity of basement membranes in the developing and adult mouse tissues. JCeN Biol1984; 9 8 971-979. 1 1 . Hessle H, Sakai LY, Hollister DW, Burgeson RE, Engvall E. Basement membrane diversity detected by monoclonal antibodies. Differentiation 1984; 2 6 49-54. 12. Hay ED. Collagen and embryonic development. In: Hay ED, ed. Cell Biology of Extracellular Matrix. New York: Plenum Press, 1982; 379409. 13. Bernfield M, Banerjee SD, Koda JE, Rapraeger AC. Remodelling of basement membrane: Morphogenesis and maturation. In: Ciba Foundation Symposium 108. Basement Membranes and Cell Movement. London: Pitman, 1984; 179-196. 14. Ekblom P, Vestwerber D, Kemler R. Cell matrix interactions and cell adhesion during development. Annu Rev CeII B i d 1986; 2: 2 7 4 7 . 15. Vracko R. Basal lamina scaffold-anatomy and significance for maintenance of orderly tissue structure. Am J Pathol1974; 77:313-346. 16. Bruckner-Tuderman L, Mitsuhashi Y, Schnyder UW, Bruckner P. Anchoring fibrils and type VII collagen are absent from skin in severe recessive dystrophic epidermolysis bullosa. J Invest Dermatol 1989; 93: 3-9. 17. Saus J, Wieslander J, Langeveld JPM, Quinones S, Hudson BG. Identification of the Goodpasture antigen as the a3(IV) chain of collagen IV. JBio/Chem 1988; 263 1337413380. 18. Broders AC. Carcinoma in situ contrasted with benign penetration epitheliumJAmMedAssoc 1932; 99: 167&1676. 19. Heagerty AHM, Kennedy AR, Leigh IM, Purkis P, Eady RAJ. Identification of an epidermal membrane defect in recessive forms of dystrophic epidermolysis bullosa by LH7.2 monoclonal antibody: use indiagnosis. Br JDermatol1986; 115: 125-131. 20. Mackenzie IC, Fusenig NE. Regeneration of organized epithelial structure. JInvest Dermatol1983; 81: 189s-194s. 21. Gammon WR, Briggaman RA, lnman AO, Queen LA, Wheeler CE. Differentiating anti-lamina lucida and anti-sublamina densa antiBMZ antibodies by indirect immunofluorescence on 1.0 M sodium chloride-separated skin. J lnvest Dermatol1984; 82: 139-144. 22. Leigh IM, Purkis PE, Bruckner-Tuderman L. LH7.2 monoclonal antibody detects type VII collagen in the sublamina densa zone of ectodermally-derived epithelia, including skin. Epithelia 1987; 1: 17-29.

LH39 ANTIBODY TO NOVEL BASEMENT MEMBRANE ANTIGEN 23. Hsu SM, Raine L, Fanger H. Use of avidin-biotin-peroxidase complex (ABC) in immunoperoxidase techniques. J Histochem Cytochem 1981; 2 9 577-580. 24. Catty D, Raykundalia C. ELISA and related enzyme immunoassays. In: Catty D, ed. Antibodies. Volume 11. A Practical Approach. Oxford: IRL Press, 1989; 97-154. 25. de Krester TA, Crumpton MJ, Bodmer JG, Bodmer WF. Demonstration of two distinct light chains in HLA-DR-associated antigens by two-dimensional gel electrophoresis. Eur J Immunol 1982; I 2 214-221. 26. Laemmli UK. Cleavage of structural proteins during the assembly of the head of the bacteriophageT4. Nature 1970;227: 68&685. 27. Bonner WM, Laskey RA. A film detection method for tritiumlabelled proteins and nucleic acids in polyacrylamide gels. Eur J Biochem 1974; 46: 83-88. 28. Mutasim DF, Takahashi Y, Labib RS, Anhalt GJ, Patel HP, Diar LA. A pool of bullous pemphigoid antigen(s) is intracellular and associated with basal cell cytoskeleton-hemidesmosome complex. J Invest Dermatol1985; &1:47-53. 29. Fine JD, Horiguchi Y, Jester J, Couchman JR. Detection and partial characterization of a midlamina lucida-hemidesmosome associated antigen (19-DEJ-I) present within human skin. J Invest Dermatol 1989; 9 2 825-830. 30. Fine JD. Cicatricial pemphigoid, bullous pemphigoid and epidermolysis bullosa acquisita antigens: differences in organ and species specificity and localization in chemically separated human skin of three basement membrane antigens. Collagen Re1 Res 1985; 5 369-377. 3 I . Verrando P, Ortonne JP, Pautrat G, Hsi BL, Yeb CJ. Identification of a 37 kilodalton protein at the epidermal basement membrane by an antiserum to human amnion. JInvest Dermatol1986; 87: 190-196. 32. Heagerty AHM, Kennedy AR, Eady RAJ, e l al. GB3 monoclonal antibody for diagnosis of junctional epidermolysis bullosd. Lancet 1986; 1: 860.

253

33. Fine JD, Smith LT, Holbrook KA, Katz SI. The appearance of four basement membrane zone antigens in developing human fetal skin. J Invest Dermatoll984; 83: 66-69. 34. Johnson CL, Holbrook KA. Development of human embryonic and fetal dermal vasculature. J Invest Dermatol1989; 93: IOa-17s. 35. Breathnach S , Fox P, Neises GR, Stanley JR, Katz SI. A unique squamous epithelial basement membrane antigen defined by a monoclonal antibody (KF-I). J Invest Dermaroll983; 8 0 392-395. 36. Goldsmith LA, Briggdman R. Monoclonal antibodies to anchoring fibrils for the diagnosis of epidermolysis bullosa. J Invest Dermnfof 1983; 81: 464466. 37. Bohnert A, Hornung J, Mackenzie IC, Fusenig NE. Epithelial mesenchymal interactions control basement membrane production and differentiation in cultured and transplanted mouse keratinocytes. Cell Tissue Res 1986; 244: 413429. 38. Hinter H, Fritsch PO, Foidart JM, Sting1 G. Katr SI. Expression of basement membrane zone antigens at the dermo-epibolic junction in organ cultures of human skin. J Inve.yt DermatolI980;; 7 4 200-204. 39. Foidart JM, Bere EW, Yaar M, el a/. Distribution and immunoelectron microscopic localization of laminin, a noncollagenous basement membrane glycoprotein. Lab Invest 1980; 42: 336342. 40. Abrahamson DR. Post-embedding colloidal gold immunolocalization of laminin to the lamina rara interna, lamina densa, and lamina rara externa of renal glomerular basement membranes. J Hisrochem Cylochem 1986; 34: 847-853. 41. Labib RS, Anahalt GJ, Patel HP, Mustasim DF, Diaz LA. Molecular heterogeneity of the bullous pemphigoid antigen as detected by immunoblotting. JImmunol1986; 136 1231-1235. 42. Almeida BM, Cballacombe SJ, Eveson JE, Morgan PR, Leigh IM. A study of the distribution of LH39 basement membrane epitope in the tumour stroma of oral squamous cell carcinomas. In preparation.