Connective Tissue Research
ISSN: 0300-8207 (Print) 1607-8438 (Online) Journal homepage: http://www.tandfonline.com/loi/icts20
Distribution of Type Viii Collagen in Tissues: An Immunohistochemical Study Reinhold Kittelberger, Paul F. Davis, David W. Flynn & Nicholas S. Greenhill To cite this article: Reinhold Kittelberger, Paul F. Davis, David W. Flynn & Nicholas S. Greenhill (1990) Distribution of Type Viii Collagen in Tissues: An Immunohistochemical Study, Connective Tissue Research, 24:3-4, 303-318, DOI: 10.3109/03008209009152157 To link to this article: http://dx.doi.org/10.3109/03008209009152157
Published online: 07 Jul 2009.
Submit your article to this journal
Article views: 22
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=icts20 Download by: [Monash University Library]
Date: 17 April 2016, At: 04:09
Connective Tissue Research, 1990, Vol. 24, pp. 303-318 Reprints available directly from the publisher Photocopying permitted by license only 0 1990 Gordon and Breach, Science Publishers, Inc. Printed in the United States of America
DISTRIBUTION OF TYPE VIII COLL-GEN IN TISSUES: AN IMMUNOHISTOCHEMICAL STUDY
Downloaded by [Monash University Library] at 04:09 17 April 2016
REINHOLD KITTELBERGER,t PAUL E DAVIS,$ DAVID W. FLYNN,§ and NICHOLAS S . GREENHILLII fDepartment of Medicine, Wellington School of Medicine, Wellington South, New Zealand tMalaghan Institute of Medical Research, Wellington School of Medicine, Box 7060, Wellington South, New Zealand §Department of Pathology, Wellington School of Medicine, Private Box 7343, Wellington South, New Zealand IlDepartment of Chemistry, Victoria University of Wellington, Private Box 600, Wellington, New Zealand (Received June 23, 1989; in revised form October 25, 1989; in finalform November 2, 1989)
S p e VIU collagen was first detected as a secretion product of diverse endothelial cell cultures, including those derived from aorta, arteries and veins. Initial studies of its tissue distribution (using a monoclonal antibody) showed it to be present in a restricted number of tissues and failed to find it in the vasculature. Recently, type VIII collagen was shown (using monospecific polyclonal antibodies) to be a component of large blood vessels with predominant localization in the subendothelium. We applied an improved immunofluorescence technique using these antibodies to define the tissue distribution of type VIII collagen. We show that it is a component of arterioles and venules in muscle, heart, kidney, spleen, liver and lung and is also found in connective tissue layers around hair follicles, around nerve bundles in muscle, in the dura of the optic nerve, in cornea and sclera, and in the perichondrium of cartilaginous tissues. This collagen variant appears to have a wider distribution than originally assumed. It is a macromolecular component of the subendothelium, possibly a constituent of the vascular intimal basement membrane. KEYWORDS: blood vessels, collagen, eriochrome black T, immunofluorescence, tissue distribution, type VIII
INTRODUCTION Collagens are the major macromolecular constituents of the extracellular matrix, representing a family of genetically distinct molecules. An increasing number of new collagen variants have been discovered during the last two decades in addition to the most abundant representative: type I collagen. 1 At least twelve distinct collagen types have been characterized so far and it is probable that more will be found. According to a classification scheme introduced by Miller,1 type VIII collagen belongs to the group 2 molecules, with a-chains larger than M , 95,000 Da and triple-helical regions interrupted by globular domains. The so-called “cassette model” of type VIII collagen proposes a molecule composed of three, possibly identical, triple-helical cassettes (M,50,000 Da), two interconnecting noncollagenous sequences and one or two C- or N-terminal short segments of about M , 303
Downloaded by [Monash University Library] at 04:09 17 April 2016
304
R. KITTELBERGER er al.
10,000 Da.2 The whole molecule is composed of three a-chains of M, 180,000 Da (EC1). Although these a-chains, as well as fragments of them (M,125,000 Da (EC2) and 100,000 Da (EC3)) have been found in the media of certain cell cultures, and a pepsin-resistant fragment of M, 50,000 Da has been purified from corneal tissue,3 an intact tissue form of type VIII collagen has yet to be isolated. 5 p e VIII collagen was originally discovered by Sage et aZ.4 in the culture media of bovine aortic endothelial cell cultures and thus named endothelial collagen.4 Subsequently they found5 that most endothelial cells that had been established in culture, and several tumor cell lines secreted type VIII collagen. Its presence in the culture medium of corneal endothelial cells, derived from Descemet’s membrane, and in a number of vascular endothelial cell cultures of aortic, arterial and venous origin6 suggested it might be found in intact corneal and vascular tissues. It was immunolocalized in Descemet’s membrane and in the corneal stroma’ using an anti-serum, produced to a mixture of type VIII and type V collagen. However, a later study using a monoclonal anti-type VIII antibody, located the protein only in Descemet’s membrane and not in the corneal stroma.7 The same study demonstrated a restricted distribution of type VIII collagen in specialized extracellular matrices of fetal tissues, such as sclera, perichondria, periostia, cartilage growth plate and uncalcified calvarial bone. Surprisingly, no staining of the vascular tree was observed. Recently, however we have successfully used affinity-purified polyclonal antibodies to demonstrate type VIII collagen in vascular tissues,s where it is predominantly located in the subendothelial region. These discrepancies prompted us to determine the distribution of type VIII collagen in other tissues using the affinity-purified polyclonal antibodies. We report the distribution in a number of tissues of adult animals.
MATERIALS AND METHODS Materials Reagents were obtained as follows: Monospecific antiserum to collagen type IV from Bioscience Products AG, Emmembrucke, Switzerland; FITC-conjugated anti-rabbit antiserum from Miles Laboratories, Naperville IL, USA; O.C.T. compound from Lab-Tek Products, Naperville, IL, USA; 30% hydrogen-peroxide solution from Merck, Darmstadt, West Germany; Pepsin (porcine stomach mucosa, 1950 unitdmg), hyaluronidase (bovine testes, 1000 unitdmg), poly-L-lysine (M,> 300,000 Da), lbeen 20, eriochrome black T (C.I. 14645), p-phenylenediamine and 4-chloro-1-naphthol from Sigma, St. Louis MO, USA. Anti-rabbit (donkey) IgG and preformed streptavidin-biotin-peroxidasecomplex were purchased from Amersham, Little Chalfont, UK. Isolation of 5 p e VIII Collagen and Production of Antibodies A mixture of type VIII and type V collagen, as well as the pure M,50,000 Da fragment of type VIII collagen were isolated and purified from ovine Descemet’s membrane following the method of Kapoor et aZ.3 Affinity-purified monospecific antibodies to the pepsin-resistant fragment of type VIII collagen were produced as described previously.8 These antibodies were tested for their specificity to ELISA against the type VIII collagen fragment, bovine and human type I
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIIl COLLAGEN
305
collagen, ovine and human types 111, IV and V collagen, human fibronectin and murine laminin. The microtiter plates (Nunc, Roskilde, Denmark) were coated with 10p,g/well of the collagens or 1Fg/well of glycoproteins in 0.1M carbonate buffer, pH 8.3,O.5M NaCl at 4°C overnight. After washing with PBS, 0.05% Tween 20, 3 times for 3 minutes at room temperature, the plates have been incubated for 30 min at room temperature with horseradish peroxidase-conjugated goat anti rabbit Ig (Sigma, St Louis, MO; lOOFl/well, diluted 1500) and then washed again. Substrate incubation was performed for 15 min with lOOp,l/well of 41mM disodiumhydrogen phosphate, 24mM citric acid, pH 5.3, 2.2mM o-phenylene diamine, 3.5mM hydrogenperoxide at room temperature. The reaction was stopped by adding lOOp,l of 2.5M sulfuric acid into each well. Reading of the plates was done in a Titertek ELISA reader (Flow Laboratories, Irvine, Scotland) at 492 nm. Titers were defined as the intersection of the extrapolated linear part of the dilution curve to the x-axis on a semilogarithmic plot. The affinity-purified antibodies showed titers around 1:1000 against type VIII collagen but virtually no reactivity against other collagens and the glycoproteins. The protein concentrations of the antibody solutions were typically around 0.2 mg/ml, as determined by the absorbance at 280 nm. Further characterization of the antibody was performed by immunoblotting (as described below) of the gelelectrophoretically separated mixture of type V and type VIII collagen. The M, 5OkDa band (type VIII collagen fragment) reacted strongly with the antibody, but not with an antibody to type VI collagen.
Immunofluorescence Microscopy Tissue samples as fresh as possible were obtained from the local abbatoir. They were embedded in O.C.T. compound and quick frozen in liquid nitrogen within 30 minutes of acquisition. Cryosections of 5-8 p,m thickness were cut, mounted on poly-L-lysine-coated glass slides,g and air dried. For immunostaining the recently described immunofluorescence techniques10 was used. Briefly it involves: Incubation of the tissue sections with primary antibody (type VIII collagen antibodies are applied undiluted, type IV collagen antibodies 1:50) in phosphate-buffered saline (PBS = 50 mM sodiumlpotassium phosphate, 0.15 M sodium chloride, pH 7.4), containing 1% bovine serum albumin (BSA) at 4°C overnight. The sections are washed three times (each of three minutes) with PBS. Incubation with FITC-conjugated goat anti-rabbit serum (diluted 1:40 in PBS-1% BSA) at room temperature for 30 minutes. Then follow 3 X 3 minute washes with PBS. Staining for 5 minutes at room temperature with 0.3% eriochrome black T in PBS followed by washing twice (3 minutes each) with PBS. Mounting in glyceroYPBS (9:1), containing 0.1% p-phenylenediamine. 11 Some sections were preincubated with 2000, 200, 20 or 2 units/ml pepsin in 0.1M acetic acid for 60 minutes, or 1500 unitdm1 hyaluronidase in PBS prior to immunofluorescence staining. This was followed by two washes of PBS (3 minutes at room temperature). Negative controls were performed by replacing the primary antibody with PBS containing 1% BSA. The sections were examined under a Zeiss fluorescence microscope equipped with a 09 filter set (BP450-490, FT 510, LP 520). Photographs were taken on Kodak Ektachrome 400 films.
306
R. KITTELBERGER er al.
Hematoxylin and eosin (H&E) staining of serial sections was performed according to standard procedures.
Downloaded by [Monash University Library] at 04:09 17 April 2016
Electrophoresis and Immunoblotting Electrophoresis was performed on homogenous gels (7.5% T, 2.67% C) using a 3% stacking gel with the discontinuous Tris-glycine buffer system described by Laemmlilz in a MiniProtean II electrophoresis cell (Bio-Rad, Richmond CA, USA). Freeze-dried samples were dissolved in sample buffer (62mM Trisphosphate, pH 6.9,10% SDS) at a concentration of 1 mg/ml, denatured at 70°C for 30 minutes and applied to the gels (10 p1 per well). After a runin of 10 minutes at 50 V the gels were run for 45 minutes at 200 V.Gels were either stained with Coomassie blue or used for electrophoretic blotting. For immunoblotting the electrophoreticallyseparated proteins were transferred to Immobilon membranes (Millipore Corp., Bredford, MA, USA) according to Towbin,l3 in a Mini Trans-Blot cell (Bio-Rad, Richmond, CA, USA) in a buffer of 25 mM Tris, 192 mh4 glycine, 20% v/v methanol, pH 8.3, at a constant current of 0.2 A for 1 hour. After protein transfer, the membrane was washed 3 X 5 minutes in PBS-Tween (50 mM sodium phosphate, 0.15 M sodium chloride, 0.05% 'been 20, pH 7.4) and then incubated for 5 minutes in a solution of 10% Blotto (fat-free milk powder) in PBS. The membrane was then incubated for 30 minutes at room temperature in a solution of primary antibody (diluted 1:lOO) in PBS, followed by 3 washes of 3 minutes in PBS-lkreen and then incubated with donkey anti-rabbit antibody in PBS, 1% BSA (1500) for 30 minutes at room temperature, followed by another 3 washes (3 minutes each) of PBS-'been. The membrane was then incubated in a solution of pre-formed streptavidin-biotin-peroxidasecomplex (1500) in PBS, 1% BSA for 30 minutes at room temperature followed by three washing steps in PBS-meen of 3 minutes, each. Afterwards the membrane was incubated in the dark for 5 minutes at room temperature in a solution of 30 mg 4-chloro-l-naphthol in 10 ml of methanol, added to 50 ml PBS, and containing 50 p1 of 30% hydrogen peroxide solution. The reaction was stopped by rinsing the membrane in distilled water.
RESULTS Tissue Processing Tissues from adult sheep were selected for this study of the tissue distribution of type VIII collagen, in which we used a highly specific anti-type VIII collagen polyclonal antibody.8 Autofluorescenceof elastic fibers is a major problem in vascular tissues but counterstaining with eriochrome black T has been shown to be very useful in suppressing autofluorescence in our previous study in which antibodies to procollagen 111, type IV collagen, laminin, nidogen and fibronectin were used for immunofluorescence histochemistry.10 Using this method yellow green autofluorescence was converted to a dark red stain. In the present study, this counterstain was routinely applied. Figures la and lb (Color Plate I) are serial sections of sheep kidney cortex a tissue with few elastic fibers. The sections were reacted with antibodies to type VIII collagen and show the staining of glomerular arterioles. In a, no counterstain was applied: specific green staining is visible in the intimal and adventitial regions. The yellowish color of cell nuclei is caused by p-phenylenediamine, which was used
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
307
FIGURE 1 Serial sections of kidney specifically stained for type. VIII collagen showing two arteries beside a glomerulus. a and b: Without and with eriochrome black T counterstaining, respectively. Eriochrome black T does not obscure the specific staining. The two arterioles are even better visualized with the counterstain. GLOM indicates a glomerulus. Original magnification X 100, bar = 50 pm. See Color Plate I.
as an anti-fading agent in the mounting medium. In b, eriochrome black T was used resulting in much clearer specific staining with no adverse effect on the binding of the type VIII collagen-specific antibodies. We have recently shown that type VIII collagen in vascular tissues is heavily masked, probably by globular protein structures.8 Unmasking was only possible after prolonged proteolytic treatment at high concentrations of protease. Figure 2 (Color Plate II) demonstrates the effect of pepsin predigestion on a carotid artery. Figure 2a is a negative control: elastic fibers appear red in color. Without pepsin digestion only parts of the media are specifically stained for type VIII collagen (Fig. 2b), but after pepsin treatment, strong fluorescence is visible throughout the blood vessel wall, with the most intense staining being in the subendothelial region (Fig. 2c). Tissue specimens other than aortic or arterial samples completely disintegrated, when we applied this harsh pretreatment (2000 units pepsidml, 60 minutes, room temperature). Even the use of lower protease concentrations resulted in either tissue damage or else were ineffective in unmasking possible additional epitopes. Therefore, unless otherwise stated all tissue sections illustrated in this paper, were stained without proteolytic pretreatment.
Downloaded by [Monash University Library] at 04:09 17 April 2016
308
R. KITTELBERGER et al.
FIGURE 2 Cryosections of larger blood vessels: carotid artery a-c, aorta d-f. a, Negative control exhibitingred autofluorescence of elastic fibers caused by eriochrome black T-counterstaining. b, c: Antibody staining for type VIII collagen. The tissue in b is untreated, while the tissue in c was preincubated with pepsin (2000 units/ml in 0.1Macetic acid for 60 minutes at room temperature). The unmasking of additional epitopes is clearly visible in c. Type IV collagen staining of aorta in d is compared to type VIII collagen stainingin e on pepsin pretreated sections. Strong fluorescence of the subendothelium is visible with both antibodies. f: A transmittedlight image of e. Lumen in all figures is to the left. Original magnification X100. Bar = 50 pm. See Color Plate II.
Tissue Distribution of Q p e VIII Collagen
Larger Blood Vessels. The distribution of type VIII collagen in aorta and carotid artery has been previously reported.8 As shown in Figure 2, epitopes of this protein were heavily masked and were accessible to the antibodies only after extensive pepsin pretreatment of the tissue samples. The most intensive staining in the artery was in the subendothelial region
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
309
(Fig. 2c) although strong staining was found throughout the vascular wall. Similarly, in the aorta strong staining was visible beneath the endothelial cell layer (Fig. 2e) though staining in the media and the adventitia was sparse. This intimal distribution of type VIII collagen covered long stretches of the subintima as an uninterrupted layer. The media staining often seemed to be codistributed with elastic fibers (Fig. 2b, c). Immunolocalization of type VIII collagen on tissue sections of larger veins was not possible because of the tissue disintegration as a result of protease pretreatment. Without protease digestion no staining was visible. Therefore, we made pepsin digests of jugular vein, and separated them by gel electrophoresis and immunoblotting. AM, 50,000 Da band in the blot revealed a strong staining reaction for type VIII collagen fragment, identical to the staining reported previously.8No other bands were immunostained by the antibody (data not shown). Nuchal Ligament. Figure 3 (Color Plate 111) shows a cross section of nuchal ligament comparing H&E staining (Fig. 3a) with the immunostaining of type VIII collagen on a serial section (Fig. 3b). Virtually no specific fluorescence is visible. Strong autofluorescence (red color) of cross-cut elastic fibers dominates the image. Even after pepsin pretreatment of tissue sections, no specific staining could be revealed. The appearance was identical to the untreated section (Fig. 3b). Eye. As reported previously,3J Descemet’s membrane and corneal stroma stained for type VIII collagen. This collagen was also found in the sclera, but not in the Bowman’s
FIGURE 3 a, b: Serial cross sections of nuchal ligament stained with H&E and anti-type VIII collagen antibodies, respectively. No specific staining is visible, even after unmasking of possibly hidden epitopes by protease-treatment. The red color of elastic fibers is caused by eriochrome black T counterstaining. Original magnification X100. Bar = 50 km. See Color Plate III.
Downloaded by [Monash University Library] at 04:09 17 April 2016
310
R. KITTELBERGER er al.
membrane, in the rectus muscles and in the vortex veins. While the staining in the Descemet’s membrane appeared as a bilayer, fibrous staining was found in the corneal stroma and sclera and also in the dura of the optic nerve (Fig. 4). (Color Plate IV) Blood vessels in the eye exhibited no specific fluorescence. Muscle. Skeletal muscle tissue was devoid of antibody staining except for small blood vessels and the perineurium of nerve bundles. In a H&E-stained section (Fig. 5a) (Color Plate V) a nerve bundle as well as muscle fibers are shown. In a serial section (Fig. 5b) strong specific fluorescence is observed only around the nerve bundle. Cross-cut myelin sheath appear as a bright red color. In cardiac muscle, tissue staining was found in the walls of blood vessels only. Figures 5c and 5d compare H&E-stained, longitudinally cut myocardium with antibody staining. Specific fluorescence is visible in a small branch coronary artery (Fig. 5d). S p e VIII collagen-specific fluorescence was also found in coronary venules where strong fluorescence of fibrillar appearance was present throughout the blood vessel wall (Fig. 6). Liver. Hepatic tissue samples were virtually free of staining, though one tissue sample showed specific staining in a restricted area around portal venules and hepatic arterioles (Fig. 7). Similar structures in other parts of the sections did not stain. This suggests that type VIII collagen is present but must be masked in most areas of the liver tissue. Attempts to unmask additional epitopes with low pepsin concentrations in the preincubation revealed no additional staining. Kidney. Antibody reaction for type VIII collagen in kidney cortex revealed staining of all
FIGURE 4 a, b Serial section of the optic nerve stained with H&E and anti-type Vm collagen antibodies, respectively. A ring of fibrous appearance surrounds the optic nerve completely. Myelin appears red. Original magnification X100. Bar = 50 pm. See Color Plate IV
311
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
FIGURE 5 Serial cryosections of muscle tissues: a, b: skeletal muscle; c, d: cardiac muscle. a, c: stained with H&E; b, d: stained with type VIII collagen antibodies. Strong fibrous staining around a nerve bundle is visible in b while the endomysium stays free of staining. No staining was found in the myocardium except for small blood vessels, as shown in d. Yellow dots in b and d are cell nuclei, stained by p-phenylenediamine. Original magnification X100. Bar = 50 pm. See Color Plate V.
Downloaded by [Monash University Library] at 04:09 17 April 2016
312
R. KITTELBERGER et a[.
FIGURE 6 a: H&E staining of cardiac tissue, showing a cardiac venule, compared to b: staining with type Vm collagen antibodies on a serial section. Fibrillar appearance of type VIII collagen is visible throughout the blood vessel wall. Original magnification X 100. Bar = 50 pm.
arteries and arterioles (Fig. la, lb). Fluorescence was strong in the internal elastic lamina and weaker in the adventitia of these blood vessels. While the interstitum was virtually free of staining, very weak fluorescence was observed in a few glomeruli. Spleen. Arterioles of the spleen showed staining in the adventitia only, but not in the intimal region (Fig. 8b). Trabeculae and capsule exhibited type VIII collagen-staining which seemed to be codistributed with elastic fibers (Fig. 8d). Lung. 5 p e VIII collagen-specific fluorescence was found in pulmonary arterioles and venules (Fig. 9b). These vessels showed fibrillar staining throughout the wall. Staining was also found in the perichondrium of the cartilaginous rings around bronchi (Fig. 9d). Skin. In the skin, type VIII collagen was only localized in the connective tissue layer surrounding hair follicles. Strong fibrous staining of this layer was visible (Fig. 10). Cartilage. Ear cartilage, tracheal cartilage and bronchial cartilage exhibited strong fibrillar staining of the perichondium (Fig. 11). Hyaluronidase treatment revealed no additional staining of the tissue. Specific fluorescence was also observed inside some chondrocytes close to the perichondrium (not shown).
DISCUSSION We have used affinity-purified, highly specific polyclonal antibodies to type VIII collagens to identify sites of distribution of this protein in a variety of tissues from adult sheep. We
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
313
FIGURE 7 Cryosection of liver. a: Transmitted light image of b, showing a hepatic venule and arteriole. b: Stained with 5 p e VIII collagen antibodies. Original magnification X 100. Bar = 50 pm.
have shown previously that this collagen variant is a component of larger arteries and aorta, being predominantly localized in the subintima,g and so we expected to find it in venous tissue and smaller blood vessels as well. Unmasking of the epitopes, using high protease concentrations and prolonged incubation times, was necessary to visualize it in larger arteries. While this harsh treatment disintegrated most tissues other than elastin-rich tissues, reducing the protease concentrations did not result in additional staining for type VIII collagen. Thus, in most of the results shown untreated tissue samples were reacted with the antibody. Under such conditions, small blood vessels in many tissues exhibited specific fluorescence. This was especially obvious in kidney arterioles where there was strong staining of the internal elastic lamina and in the adventitia. Weaker staining of blood vessels was found in the skeletal muscle, heart muscle and lung. In heart and lung the veins also showed type VIII collagen-specificstaining. The presence of type VIII collagen was also demonstrated in a pepsin extract of jugular vein by immunoblotting. Q p e VIII collagen was detected in only a few blood vessels of the liver tissue. In the spleen the staining of arterioles was found to be restricted to the adventitia. Thus type VIII collagen is obviously a mural component of most blood vessels. The reasons why previous investigators have failed to demonstrate this variant in vascular tissue are probably: masking of epitopes, the application of a single monoclonal antibody and the use of fetal tissue.’ The differences in vascular distribution which we have found amongst the various tissues may also partially be explained by masking of epitopes, as in
R. KITTELBERGER et al.
Downloaded by [Monash University Library] at 04:09 17 April 2016
3 14
FIGURE 8 Serial sections of spleen tissue. a, c: H&E staining; b, d: stained with type Vm collagen-specific antibodies. An arteriole in b exhibits only weak type VIII collagen-specific staining in the adventitia. c, d Serial sections of a trabeculae. Specific staining seems to be codistributed with elastic fibers. Original magnification XIOO. Bar = 50 pm.
315
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYF'E VIII COLLAGEN
FIGURE 9 Serial sections of lung tissue. a. c: H&E staining; b, d: stained with type VIII collagen antibodies. a, b: Pulmonary venule with strong staining for type VIII collagen in b. Strong fibrous staining was also found in the perichondrium of cartilaginous rings around bronchi, as shown in c, d. Original magnification X100. Bar = SO Frn.
Downloaded by [Monash University Library] at 04:09 17 April 2016
316
R. KITTELBERGER et al.
FIGURE 10 Serial sections of skin tissue. a: H&E staining;b: stained with type VEI collagen antibodies. Specific staining of fibrous appearance was only found around hair follicles. Original magnification X 100. Bar = 50 pm.
liver sections, or it may reflect a tissue-specific distribution, e.g., the restriction of type VIII collagen to the adventitia of vascular walls in the spleen arterioles. Q p e VIII collagen has been shown previously to be a constituent of a basement membrane, namely Descemet’s membrane.7.8 It is clear from the data presented above that it is neither a component of epithelial basement membranes, such as those seen in alveolar membranes in the lung or at the dermal-epidermal junction in skin, nor was it found to be distributed around vascular smooth muscle cells. Weak staining in some glomeruli indicates its possible presence in the glomerular basement membrane. The predominant location of this protein in the subendothelium of many blood vessels suggests that it seems to be associated with basement membranes and may perhaps be a component of the subendothelial basement membrane. This assumption is strengthened by the very similar distribution of type VIII collagen and the basement membrane-specific type IV collagen in the intima of arteries and aorta (Fig. 2d, 2e). The predominant location of Qpe VIII collagen in the vascular subendothelium indicates a potential function of this collagen variant as an important contributor to the architectural integrity of the intima. By light microscopy type VIII collagen exhibits a linear pattern in vascular tissues, in the connective tissue surrounding hair follicles, in the stroma of the corneas, in the perichondrium of cartilagenous tissues and around nerve bundles in skeletal muscle and the dura of the optic nerve. Preliminary attempts to visualizing type VIII collagen distribution in corneal endothelial cell culture matrix on the ultrastructural level using antibody-directed
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
317
FIGURE 11 Serial sections of ear cartilage, showing the perichondrial area. a: H&E staining; b: stained with type VIII collagen antibodies. Strong fibrous fluorescence is visible throughout the periochondrium in b. Original magnification XIOO. Bar = 50 pn.
colloidal gold deposition showed staining along fine linear filaments (
ISSN: 0300-8207 (Print) 1607-8438 (Online) Journal homepage: http://www.tandfonline.com/loi/icts20
Distribution of Type Viii Collagen in Tissues: An Immunohistochemical Study Reinhold Kittelberger, Paul F. Davis, David W. Flynn & Nicholas S. Greenhill To cite this article: Reinhold Kittelberger, Paul F. Davis, David W. Flynn & Nicholas S. Greenhill (1990) Distribution of Type Viii Collagen in Tissues: An Immunohistochemical Study, Connective Tissue Research, 24:3-4, 303-318, DOI: 10.3109/03008209009152157 To link to this article: http://dx.doi.org/10.3109/03008209009152157
Published online: 07 Jul 2009.
Submit your article to this journal
Article views: 22
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=icts20 Download by: [Monash University Library]
Date: 17 April 2016, At: 04:09
Connective Tissue Research, 1990, Vol. 24, pp. 303-318 Reprints available directly from the publisher Photocopying permitted by license only 0 1990 Gordon and Breach, Science Publishers, Inc. Printed in the United States of America
DISTRIBUTION OF TYPE VIII COLL-GEN IN TISSUES: AN IMMUNOHISTOCHEMICAL STUDY
Downloaded by [Monash University Library] at 04:09 17 April 2016
REINHOLD KITTELBERGER,t PAUL E DAVIS,$ DAVID W. FLYNN,§ and NICHOLAS S . GREENHILLII fDepartment of Medicine, Wellington School of Medicine, Wellington South, New Zealand tMalaghan Institute of Medical Research, Wellington School of Medicine, Box 7060, Wellington South, New Zealand §Department of Pathology, Wellington School of Medicine, Private Box 7343, Wellington South, New Zealand IlDepartment of Chemistry, Victoria University of Wellington, Private Box 600, Wellington, New Zealand (Received June 23, 1989; in revised form October 25, 1989; in finalform November 2, 1989)
S p e VIU collagen was first detected as a secretion product of diverse endothelial cell cultures, including those derived from aorta, arteries and veins. Initial studies of its tissue distribution (using a monoclonal antibody) showed it to be present in a restricted number of tissues and failed to find it in the vasculature. Recently, type VIII collagen was shown (using monospecific polyclonal antibodies) to be a component of large blood vessels with predominant localization in the subendothelium. We applied an improved immunofluorescence technique using these antibodies to define the tissue distribution of type VIII collagen. We show that it is a component of arterioles and venules in muscle, heart, kidney, spleen, liver and lung and is also found in connective tissue layers around hair follicles, around nerve bundles in muscle, in the dura of the optic nerve, in cornea and sclera, and in the perichondrium of cartilaginous tissues. This collagen variant appears to have a wider distribution than originally assumed. It is a macromolecular component of the subendothelium, possibly a constituent of the vascular intimal basement membrane. KEYWORDS: blood vessels, collagen, eriochrome black T, immunofluorescence, tissue distribution, type VIII
INTRODUCTION Collagens are the major macromolecular constituents of the extracellular matrix, representing a family of genetically distinct molecules. An increasing number of new collagen variants have been discovered during the last two decades in addition to the most abundant representative: type I collagen. 1 At least twelve distinct collagen types have been characterized so far and it is probable that more will be found. According to a classification scheme introduced by Miller,1 type VIII collagen belongs to the group 2 molecules, with a-chains larger than M , 95,000 Da and triple-helical regions interrupted by globular domains. The so-called “cassette model” of type VIII collagen proposes a molecule composed of three, possibly identical, triple-helical cassettes (M,50,000 Da), two interconnecting noncollagenous sequences and one or two C- or N-terminal short segments of about M , 303
Downloaded by [Monash University Library] at 04:09 17 April 2016
304
R. KITTELBERGER er al.
10,000 Da.2 The whole molecule is composed of three a-chains of M, 180,000 Da (EC1). Although these a-chains, as well as fragments of them (M,125,000 Da (EC2) and 100,000 Da (EC3)) have been found in the media of certain cell cultures, and a pepsin-resistant fragment of M, 50,000 Da has been purified from corneal tissue,3 an intact tissue form of type VIII collagen has yet to be isolated. 5 p e VIII collagen was originally discovered by Sage et aZ.4 in the culture media of bovine aortic endothelial cell cultures and thus named endothelial collagen.4 Subsequently they found5 that most endothelial cells that had been established in culture, and several tumor cell lines secreted type VIII collagen. Its presence in the culture medium of corneal endothelial cells, derived from Descemet’s membrane, and in a number of vascular endothelial cell cultures of aortic, arterial and venous origin6 suggested it might be found in intact corneal and vascular tissues. It was immunolocalized in Descemet’s membrane and in the corneal stroma’ using an anti-serum, produced to a mixture of type VIII and type V collagen. However, a later study using a monoclonal anti-type VIII antibody, located the protein only in Descemet’s membrane and not in the corneal stroma.7 The same study demonstrated a restricted distribution of type VIII collagen in specialized extracellular matrices of fetal tissues, such as sclera, perichondria, periostia, cartilage growth plate and uncalcified calvarial bone. Surprisingly, no staining of the vascular tree was observed. Recently, however we have successfully used affinity-purified polyclonal antibodies to demonstrate type VIII collagen in vascular tissues,s where it is predominantly located in the subendothelial region. These discrepancies prompted us to determine the distribution of type VIII collagen in other tissues using the affinity-purified polyclonal antibodies. We report the distribution in a number of tissues of adult animals.
MATERIALS AND METHODS Materials Reagents were obtained as follows: Monospecific antiserum to collagen type IV from Bioscience Products AG, Emmembrucke, Switzerland; FITC-conjugated anti-rabbit antiserum from Miles Laboratories, Naperville IL, USA; O.C.T. compound from Lab-Tek Products, Naperville, IL, USA; 30% hydrogen-peroxide solution from Merck, Darmstadt, West Germany; Pepsin (porcine stomach mucosa, 1950 unitdmg), hyaluronidase (bovine testes, 1000 unitdmg), poly-L-lysine (M,> 300,000 Da), lbeen 20, eriochrome black T (C.I. 14645), p-phenylenediamine and 4-chloro-1-naphthol from Sigma, St. Louis MO, USA. Anti-rabbit (donkey) IgG and preformed streptavidin-biotin-peroxidasecomplex were purchased from Amersham, Little Chalfont, UK. Isolation of 5 p e VIII Collagen and Production of Antibodies A mixture of type VIII and type V collagen, as well as the pure M,50,000 Da fragment of type VIII collagen were isolated and purified from ovine Descemet’s membrane following the method of Kapoor et aZ.3 Affinity-purified monospecific antibodies to the pepsin-resistant fragment of type VIII collagen were produced as described previously.8 These antibodies were tested for their specificity to ELISA against the type VIII collagen fragment, bovine and human type I
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIIl COLLAGEN
305
collagen, ovine and human types 111, IV and V collagen, human fibronectin and murine laminin. The microtiter plates (Nunc, Roskilde, Denmark) were coated with 10p,g/well of the collagens or 1Fg/well of glycoproteins in 0.1M carbonate buffer, pH 8.3,O.5M NaCl at 4°C overnight. After washing with PBS, 0.05% Tween 20, 3 times for 3 minutes at room temperature, the plates have been incubated for 30 min at room temperature with horseradish peroxidase-conjugated goat anti rabbit Ig (Sigma, St Louis, MO; lOOFl/well, diluted 1500) and then washed again. Substrate incubation was performed for 15 min with lOOp,l/well of 41mM disodiumhydrogen phosphate, 24mM citric acid, pH 5.3, 2.2mM o-phenylene diamine, 3.5mM hydrogenperoxide at room temperature. The reaction was stopped by adding lOOp,l of 2.5M sulfuric acid into each well. Reading of the plates was done in a Titertek ELISA reader (Flow Laboratories, Irvine, Scotland) at 492 nm. Titers were defined as the intersection of the extrapolated linear part of the dilution curve to the x-axis on a semilogarithmic plot. The affinity-purified antibodies showed titers around 1:1000 against type VIII collagen but virtually no reactivity against other collagens and the glycoproteins. The protein concentrations of the antibody solutions were typically around 0.2 mg/ml, as determined by the absorbance at 280 nm. Further characterization of the antibody was performed by immunoblotting (as described below) of the gelelectrophoretically separated mixture of type V and type VIII collagen. The M, 5OkDa band (type VIII collagen fragment) reacted strongly with the antibody, but not with an antibody to type VI collagen.
Immunofluorescence Microscopy Tissue samples as fresh as possible were obtained from the local abbatoir. They were embedded in O.C.T. compound and quick frozen in liquid nitrogen within 30 minutes of acquisition. Cryosections of 5-8 p,m thickness were cut, mounted on poly-L-lysine-coated glass slides,g and air dried. For immunostaining the recently described immunofluorescence techniques10 was used. Briefly it involves: Incubation of the tissue sections with primary antibody (type VIII collagen antibodies are applied undiluted, type IV collagen antibodies 1:50) in phosphate-buffered saline (PBS = 50 mM sodiumlpotassium phosphate, 0.15 M sodium chloride, pH 7.4), containing 1% bovine serum albumin (BSA) at 4°C overnight. The sections are washed three times (each of three minutes) with PBS. Incubation with FITC-conjugated goat anti-rabbit serum (diluted 1:40 in PBS-1% BSA) at room temperature for 30 minutes. Then follow 3 X 3 minute washes with PBS. Staining for 5 minutes at room temperature with 0.3% eriochrome black T in PBS followed by washing twice (3 minutes each) with PBS. Mounting in glyceroYPBS (9:1), containing 0.1% p-phenylenediamine. 11 Some sections were preincubated with 2000, 200, 20 or 2 units/ml pepsin in 0.1M acetic acid for 60 minutes, or 1500 unitdm1 hyaluronidase in PBS prior to immunofluorescence staining. This was followed by two washes of PBS (3 minutes at room temperature). Negative controls were performed by replacing the primary antibody with PBS containing 1% BSA. The sections were examined under a Zeiss fluorescence microscope equipped with a 09 filter set (BP450-490, FT 510, LP 520). Photographs were taken on Kodak Ektachrome 400 films.
306
R. KITTELBERGER er al.
Hematoxylin and eosin (H&E) staining of serial sections was performed according to standard procedures.
Downloaded by [Monash University Library] at 04:09 17 April 2016
Electrophoresis and Immunoblotting Electrophoresis was performed on homogenous gels (7.5% T, 2.67% C) using a 3% stacking gel with the discontinuous Tris-glycine buffer system described by Laemmlilz in a MiniProtean II electrophoresis cell (Bio-Rad, Richmond CA, USA). Freeze-dried samples were dissolved in sample buffer (62mM Trisphosphate, pH 6.9,10% SDS) at a concentration of 1 mg/ml, denatured at 70°C for 30 minutes and applied to the gels (10 p1 per well). After a runin of 10 minutes at 50 V the gels were run for 45 minutes at 200 V.Gels were either stained with Coomassie blue or used for electrophoretic blotting. For immunoblotting the electrophoreticallyseparated proteins were transferred to Immobilon membranes (Millipore Corp., Bredford, MA, USA) according to Towbin,l3 in a Mini Trans-Blot cell (Bio-Rad, Richmond, CA, USA) in a buffer of 25 mM Tris, 192 mh4 glycine, 20% v/v methanol, pH 8.3, at a constant current of 0.2 A for 1 hour. After protein transfer, the membrane was washed 3 X 5 minutes in PBS-Tween (50 mM sodium phosphate, 0.15 M sodium chloride, 0.05% 'been 20, pH 7.4) and then incubated for 5 minutes in a solution of 10% Blotto (fat-free milk powder) in PBS. The membrane was then incubated for 30 minutes at room temperature in a solution of primary antibody (diluted 1:lOO) in PBS, followed by 3 washes of 3 minutes in PBS-lkreen and then incubated with donkey anti-rabbit antibody in PBS, 1% BSA (1500) for 30 minutes at room temperature, followed by another 3 washes (3 minutes each) of PBS-'been. The membrane was then incubated in a solution of pre-formed streptavidin-biotin-peroxidasecomplex (1500) in PBS, 1% BSA for 30 minutes at room temperature followed by three washing steps in PBS-meen of 3 minutes, each. Afterwards the membrane was incubated in the dark for 5 minutes at room temperature in a solution of 30 mg 4-chloro-l-naphthol in 10 ml of methanol, added to 50 ml PBS, and containing 50 p1 of 30% hydrogen peroxide solution. The reaction was stopped by rinsing the membrane in distilled water.
RESULTS Tissue Processing Tissues from adult sheep were selected for this study of the tissue distribution of type VIII collagen, in which we used a highly specific anti-type VIII collagen polyclonal antibody.8 Autofluorescenceof elastic fibers is a major problem in vascular tissues but counterstaining with eriochrome black T has been shown to be very useful in suppressing autofluorescence in our previous study in which antibodies to procollagen 111, type IV collagen, laminin, nidogen and fibronectin were used for immunofluorescence histochemistry.10 Using this method yellow green autofluorescence was converted to a dark red stain. In the present study, this counterstain was routinely applied. Figures la and lb (Color Plate I) are serial sections of sheep kidney cortex a tissue with few elastic fibers. The sections were reacted with antibodies to type VIII collagen and show the staining of glomerular arterioles. In a, no counterstain was applied: specific green staining is visible in the intimal and adventitial regions. The yellowish color of cell nuclei is caused by p-phenylenediamine, which was used
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
307
FIGURE 1 Serial sections of kidney specifically stained for type. VIII collagen showing two arteries beside a glomerulus. a and b: Without and with eriochrome black T counterstaining, respectively. Eriochrome black T does not obscure the specific staining. The two arterioles are even better visualized with the counterstain. GLOM indicates a glomerulus. Original magnification X 100, bar = 50 pm. See Color Plate I.
as an anti-fading agent in the mounting medium. In b, eriochrome black T was used resulting in much clearer specific staining with no adverse effect on the binding of the type VIII collagen-specific antibodies. We have recently shown that type VIII collagen in vascular tissues is heavily masked, probably by globular protein structures.8 Unmasking was only possible after prolonged proteolytic treatment at high concentrations of protease. Figure 2 (Color Plate II) demonstrates the effect of pepsin predigestion on a carotid artery. Figure 2a is a negative control: elastic fibers appear red in color. Without pepsin digestion only parts of the media are specifically stained for type VIII collagen (Fig. 2b), but after pepsin treatment, strong fluorescence is visible throughout the blood vessel wall, with the most intense staining being in the subendothelial region (Fig. 2c). Tissue specimens other than aortic or arterial samples completely disintegrated, when we applied this harsh pretreatment (2000 units pepsidml, 60 minutes, room temperature). Even the use of lower protease concentrations resulted in either tissue damage or else were ineffective in unmasking possible additional epitopes. Therefore, unless otherwise stated all tissue sections illustrated in this paper, were stained without proteolytic pretreatment.
Downloaded by [Monash University Library] at 04:09 17 April 2016
308
R. KITTELBERGER et al.
FIGURE 2 Cryosections of larger blood vessels: carotid artery a-c, aorta d-f. a, Negative control exhibitingred autofluorescence of elastic fibers caused by eriochrome black T-counterstaining. b, c: Antibody staining for type VIII collagen. The tissue in b is untreated, while the tissue in c was preincubated with pepsin (2000 units/ml in 0.1Macetic acid for 60 minutes at room temperature). The unmasking of additional epitopes is clearly visible in c. Type IV collagen staining of aorta in d is compared to type VIII collagen stainingin e on pepsin pretreated sections. Strong fluorescence of the subendothelium is visible with both antibodies. f: A transmittedlight image of e. Lumen in all figures is to the left. Original magnification X100. Bar = 50 pm. See Color Plate II.
Tissue Distribution of Q p e VIII Collagen
Larger Blood Vessels. The distribution of type VIII collagen in aorta and carotid artery has been previously reported.8 As shown in Figure 2, epitopes of this protein were heavily masked and were accessible to the antibodies only after extensive pepsin pretreatment of the tissue samples. The most intensive staining in the artery was in the subendothelial region
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
309
(Fig. 2c) although strong staining was found throughout the vascular wall. Similarly, in the aorta strong staining was visible beneath the endothelial cell layer (Fig. 2e) though staining in the media and the adventitia was sparse. This intimal distribution of type VIII collagen covered long stretches of the subintima as an uninterrupted layer. The media staining often seemed to be codistributed with elastic fibers (Fig. 2b, c). Immunolocalization of type VIII collagen on tissue sections of larger veins was not possible because of the tissue disintegration as a result of protease pretreatment. Without protease digestion no staining was visible. Therefore, we made pepsin digests of jugular vein, and separated them by gel electrophoresis and immunoblotting. AM, 50,000 Da band in the blot revealed a strong staining reaction for type VIII collagen fragment, identical to the staining reported previously.8No other bands were immunostained by the antibody (data not shown). Nuchal Ligament. Figure 3 (Color Plate 111) shows a cross section of nuchal ligament comparing H&E staining (Fig. 3a) with the immunostaining of type VIII collagen on a serial section (Fig. 3b). Virtually no specific fluorescence is visible. Strong autofluorescence (red color) of cross-cut elastic fibers dominates the image. Even after pepsin pretreatment of tissue sections, no specific staining could be revealed. The appearance was identical to the untreated section (Fig. 3b). Eye. As reported previously,3J Descemet’s membrane and corneal stroma stained for type VIII collagen. This collagen was also found in the sclera, but not in the Bowman’s
FIGURE 3 a, b: Serial cross sections of nuchal ligament stained with H&E and anti-type VIII collagen antibodies, respectively. No specific staining is visible, even after unmasking of possibly hidden epitopes by protease-treatment. The red color of elastic fibers is caused by eriochrome black T counterstaining. Original magnification X100. Bar = 50 km. See Color Plate III.
Downloaded by [Monash University Library] at 04:09 17 April 2016
310
R. KITTELBERGER er al.
membrane, in the rectus muscles and in the vortex veins. While the staining in the Descemet’s membrane appeared as a bilayer, fibrous staining was found in the corneal stroma and sclera and also in the dura of the optic nerve (Fig. 4). (Color Plate IV) Blood vessels in the eye exhibited no specific fluorescence. Muscle. Skeletal muscle tissue was devoid of antibody staining except for small blood vessels and the perineurium of nerve bundles. In a H&E-stained section (Fig. 5a) (Color Plate V) a nerve bundle as well as muscle fibers are shown. In a serial section (Fig. 5b) strong specific fluorescence is observed only around the nerve bundle. Cross-cut myelin sheath appear as a bright red color. In cardiac muscle, tissue staining was found in the walls of blood vessels only. Figures 5c and 5d compare H&E-stained, longitudinally cut myocardium with antibody staining. Specific fluorescence is visible in a small branch coronary artery (Fig. 5d). S p e VIII collagen-specific fluorescence was also found in coronary venules where strong fluorescence of fibrillar appearance was present throughout the blood vessel wall (Fig. 6). Liver. Hepatic tissue samples were virtually free of staining, though one tissue sample showed specific staining in a restricted area around portal venules and hepatic arterioles (Fig. 7). Similar structures in other parts of the sections did not stain. This suggests that type VIII collagen is present but must be masked in most areas of the liver tissue. Attempts to unmask additional epitopes with low pepsin concentrations in the preincubation revealed no additional staining. Kidney. Antibody reaction for type VIII collagen in kidney cortex revealed staining of all
FIGURE 4 a, b Serial section of the optic nerve stained with H&E and anti-type Vm collagen antibodies, respectively. A ring of fibrous appearance surrounds the optic nerve completely. Myelin appears red. Original magnification X100. Bar = 50 pm. See Color Plate IV
311
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
FIGURE 5 Serial cryosections of muscle tissues: a, b: skeletal muscle; c, d: cardiac muscle. a, c: stained with H&E; b, d: stained with type VIII collagen antibodies. Strong fibrous staining around a nerve bundle is visible in b while the endomysium stays free of staining. No staining was found in the myocardium except for small blood vessels, as shown in d. Yellow dots in b and d are cell nuclei, stained by p-phenylenediamine. Original magnification X100. Bar = 50 pm. See Color Plate V.
Downloaded by [Monash University Library] at 04:09 17 April 2016
312
R. KITTELBERGER et a[.
FIGURE 6 a: H&E staining of cardiac tissue, showing a cardiac venule, compared to b: staining with type Vm collagen antibodies on a serial section. Fibrillar appearance of type VIII collagen is visible throughout the blood vessel wall. Original magnification X 100. Bar = 50 pm.
arteries and arterioles (Fig. la, lb). Fluorescence was strong in the internal elastic lamina and weaker in the adventitia of these blood vessels. While the interstitum was virtually free of staining, very weak fluorescence was observed in a few glomeruli. Spleen. Arterioles of the spleen showed staining in the adventitia only, but not in the intimal region (Fig. 8b). Trabeculae and capsule exhibited type VIII collagen-staining which seemed to be codistributed with elastic fibers (Fig. 8d). Lung. 5 p e VIII collagen-specific fluorescence was found in pulmonary arterioles and venules (Fig. 9b). These vessels showed fibrillar staining throughout the wall. Staining was also found in the perichondrium of the cartilaginous rings around bronchi (Fig. 9d). Skin. In the skin, type VIII collagen was only localized in the connective tissue layer surrounding hair follicles. Strong fibrous staining of this layer was visible (Fig. 10). Cartilage. Ear cartilage, tracheal cartilage and bronchial cartilage exhibited strong fibrillar staining of the perichondium (Fig. 11). Hyaluronidase treatment revealed no additional staining of the tissue. Specific fluorescence was also observed inside some chondrocytes close to the perichondrium (not shown).
DISCUSSION We have used affinity-purified, highly specific polyclonal antibodies to type VIII collagens to identify sites of distribution of this protein in a variety of tissues from adult sheep. We
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
313
FIGURE 7 Cryosection of liver. a: Transmitted light image of b, showing a hepatic venule and arteriole. b: Stained with 5 p e VIII collagen antibodies. Original magnification X 100. Bar = 50 pm.
have shown previously that this collagen variant is a component of larger arteries and aorta, being predominantly localized in the subintima,g and so we expected to find it in venous tissue and smaller blood vessels as well. Unmasking of the epitopes, using high protease concentrations and prolonged incubation times, was necessary to visualize it in larger arteries. While this harsh treatment disintegrated most tissues other than elastin-rich tissues, reducing the protease concentrations did not result in additional staining for type VIII collagen. Thus, in most of the results shown untreated tissue samples were reacted with the antibody. Under such conditions, small blood vessels in many tissues exhibited specific fluorescence. This was especially obvious in kidney arterioles where there was strong staining of the internal elastic lamina and in the adventitia. Weaker staining of blood vessels was found in the skeletal muscle, heart muscle and lung. In heart and lung the veins also showed type VIII collagen-specificstaining. The presence of type VIII collagen was also demonstrated in a pepsin extract of jugular vein by immunoblotting. Q p e VIII collagen was detected in only a few blood vessels of the liver tissue. In the spleen the staining of arterioles was found to be restricted to the adventitia. Thus type VIII collagen is obviously a mural component of most blood vessels. The reasons why previous investigators have failed to demonstrate this variant in vascular tissue are probably: masking of epitopes, the application of a single monoclonal antibody and the use of fetal tissue.’ The differences in vascular distribution which we have found amongst the various tissues may also partially be explained by masking of epitopes, as in
R. KITTELBERGER et al.
Downloaded by [Monash University Library] at 04:09 17 April 2016
3 14
FIGURE 8 Serial sections of spleen tissue. a, c: H&E staining; b, d: stained with type Vm collagen-specific antibodies. An arteriole in b exhibits only weak type VIII collagen-specific staining in the adventitia. c, d Serial sections of a trabeculae. Specific staining seems to be codistributed with elastic fibers. Original magnification XIOO. Bar = 50 pm.
315
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYF'E VIII COLLAGEN
FIGURE 9 Serial sections of lung tissue. a. c: H&E staining; b, d: stained with type VIII collagen antibodies. a, b: Pulmonary venule with strong staining for type VIII collagen in b. Strong fibrous staining was also found in the perichondrium of cartilaginous rings around bronchi, as shown in c, d. Original magnification X100. Bar = SO Frn.
Downloaded by [Monash University Library] at 04:09 17 April 2016
316
R. KITTELBERGER et al.
FIGURE 10 Serial sections of skin tissue. a: H&E staining;b: stained with type VEI collagen antibodies. Specific staining of fibrous appearance was only found around hair follicles. Original magnification X 100. Bar = 50 pm.
liver sections, or it may reflect a tissue-specific distribution, e.g., the restriction of type VIII collagen to the adventitia of vascular walls in the spleen arterioles. Q p e VIII collagen has been shown previously to be a constituent of a basement membrane, namely Descemet’s membrane.7.8 It is clear from the data presented above that it is neither a component of epithelial basement membranes, such as those seen in alveolar membranes in the lung or at the dermal-epidermal junction in skin, nor was it found to be distributed around vascular smooth muscle cells. Weak staining in some glomeruli indicates its possible presence in the glomerular basement membrane. The predominant location of this protein in the subendothelium of many blood vessels suggests that it seems to be associated with basement membranes and may perhaps be a component of the subendothelial basement membrane. This assumption is strengthened by the very similar distribution of type VIII collagen and the basement membrane-specific type IV collagen in the intima of arteries and aorta (Fig. 2d, 2e). The predominant location of Qpe VIII collagen in the vascular subendothelium indicates a potential function of this collagen variant as an important contributor to the architectural integrity of the intima. By light microscopy type VIII collagen exhibits a linear pattern in vascular tissues, in the connective tissue surrounding hair follicles, in the stroma of the corneas, in the perichondrium of cartilagenous tissues and around nerve bundles in skeletal muscle and the dura of the optic nerve. Preliminary attempts to visualizing type VIII collagen distribution in corneal endothelial cell culture matrix on the ultrastructural level using antibody-directed
Downloaded by [Monash University Library] at 04:09 17 April 2016
DISTRIBUTION OF TYPE VIII COLLAGEN
317
FIGURE 11 Serial sections of ear cartilage, showing the perichondrial area. a: H&E staining; b: stained with type VIII collagen antibodies. Strong fibrous fluorescence is visible throughout the periochondrium in b. Original magnification XIOO. Bar = 50 pn.
colloidal gold deposition showed staining along fine linear filaments (
