Ultrastructural studies of bovine retinal microvascular basement membranes with the cationic dye ruthenium red.код для вставкиСкачать
THE ANATOMICAL RECORD 219:363-368 (1987) Ultrastructural Studies of Bovine Retinal Microvascular Basement Membranes With the Cationic Dye Ruthenium Red JAMES W. SCHMIDLEY Department of Neurology, School of Medicine, Case Western Reserve University, Cleveland, Ohio 44106-5000 ABSTRACT Our recent observation that the basement membranes of brain microvessels do not stain with the cationic dye ruthenium red has raised the question of whether the basement membranes of this and other vascular beds functioning as barriers between blood and neural tissues are deficient in the polyanionic macromolecules, such as glycosaminoglycans, which are responsible for the ruthenium red staining of other vascular basement membranes. We therefore attempted to produce staining in the only barrier-type microvascular basement membrane known to contain heparan sulfate. Bovine retinas were fixed by immersion in aldehyde fixatives containing ruthenium red, buffered with either 10 mM or 100 mM sodium cacodylate. We found discrete, electron-dense deposits of ruthenium red in vascular basement membranes, quite similar to those seen in vascular basement membranes of nonneural tissues after exposure to ruthenium red. These deposits were more distinct and more frequent in tissue exposed to ruthenium red-aldehyde solutions buffered with 10 mM cacodylate. They were not seen if ruthenium red was omitted from the fixative. The results demonstrate that anionic macromolecules in basement membranes of barrier-type microvessels can be stained with cationic dyes, and suggest that the failure of brain microvessels to stain with ruthenium red may be the result of a relative or total lack of polyanion in this basement membrane, or of other unique properties. After exposure to fixatives containing ruthenium red (RR), the basement membranes (BM) of capillaries exhibit regularly spaced, discrete, electron-dense deposits of this cationic dye, which binds electrostatically to anionic macromolecules (for reviews, see Charonis and Wissig, 1983; Schmidley and Wissig, 1986). Biochemical and immunohistochemical studies further suggest that in some vascular BM these sites of RR accumulation represent the glycosaminoglycan (GAG) chains of proteoglycans (PG) (Charonis et al., 1983; Farquhar, 1981; Kanwar and Farquar, 197913; Simionescu et al., 1984). The BM of the continuous capillaries forming the blood-brain barrier CBBB) is the only vascular BM studied thus far that does not exhibit such discrete electrondense deposits following fixation in RR, either in situ (following vascular perfusion) or in vitro (following isolation from cerebral cortex) (Schmidley and Wissig, 1986; Schmidley, in press). The restrictive properties of the BBB are probably not responsible for this lack of staining, because attempts to circumvent the BBB by including detergents or dimethylsulfoxide in the perfusion solutions (Schmidley and Wissig, 1986), and by immersing thin slices of cerebral cortex (Schmidley and Wissig, 1986) or isolated cerebral microvessels (Schmidley, in press) in RR-aldehyde solutions, did not produce discrete deposits of RR in the blood-brain barrier BM. The lack of discrete staining in the BM of isolated cerebral 0 1987 ALAN R. LISS, INC. microvessels cannot be attributed to loss of PGs during isolation, since anionic macromolecules in other BMs survive disruptive isolation procedures (Kanwar and Farquhar, 1979a,b). These findings contrast sharply with those in other vascular and nonvascular BM (including the endothelial and epithelia1 BM of the choroid plexus, and circumventricular organs, regions of the CNS with fenestrated capillaries, lacking a BBB), which are readily stained with RR (Charonis and Wissig, 1983; Thurauf et al., 1983; Schmidley and Wissig, 1986). Possible explanations for the failure to demonstrate discrete sites of RR deposition in blood-brain barrier BM might include 1)a relative or absolute absence of polyanionic macromolecules, such a s the GAG component of PGs, or 2) a difference in the structure of this BM that would render the GAG molecules inaccessible to RR or would prevent them from collapsing, after dehydration, to form aggregates of sufficient electron density to be detectable in thin sections (Hascall, 1980). Against explanation (1) is the immunohistochemical evidence for a HSPG core protein in blood-brain barrier BM of the rat (Laurie et al., 1983)and the mouse (Schmidley and Wissig, unpublished observations). Retina1 capillaries, like brain capillaries, are interposed between the circulation and a neuronal tissue Received February 19, 1987; accepted June 22, 1987 364 J.W. SCHMIDLEY requiring a precisely regulated extracellular environment for optimal function. They have structural and functional characteristics very similar to, if not identical with, those of BBB capillaries, which enable them to serve a similar barrier function (Raviola, 1977). The BM of microvessels in the bovine retina is the only “barrier type” microvessel BM that has been shown, by direct biochemical analysis, to contain GAG (heparan sulfate) (Kennedy et al., 1986; Cohen and Ciborowski, 1981). It thus might reasonably be expected to show, after exposure to fixatives containing RR, discrete electron-dense accumulations of this cationic dye similar to those found in the BM of nonneural capillaries. A failure to produce this discrete pattern of staining in a BM known to contain polyanionic macromolecules would support the idea that the architecture of the BM of barrier vessels differs from that of nonbarrier vessels, and that this different structure, rather than a n absence of polyanionic macromolecules, is responsible for the failure of staining with RR (i.e., explanation 2 above). On the other hand, demonstration of discrete RR deposits in BM of retina1 capillaries would suggest that our failure to produce similar staining in rat BBB capillary BM is the result of a relative or total deficiency of polyanionic macromolecules in this BM. MATERIALS AND METHODS Fig. 1. Deposits of RR in the interphotoreceptor matrix. In this and all subsequent figures, except Figure 6, the retinas were fixed and Cow eyes were obtained from a local slaughterhouse processed according to the high-ionic-strength protocol. 63,000 x . and transported on ice. Within 45 min, the retinas were removed, minced with a razor blade, and immersed in fixative overnight. Two fixatives were used: The first, referred to as high-ionic-strength fixative, consisted of 2% glutaraldehyde (GA), 1% paraformaldehyde (PF),in iments, because it is possible that the CEC for RR stain0.1 M Na-cacodylate buffer, pH 7.4, with 0.2% RR. The ing of anionic sites in bovine microvessel BM is low-ionic-strength fixative consisted of the same concen- approximately the same as that of buffers customarily trations of GA, PF, and RR, but was buffered with 0.01 used for electron microscopy (that is, approximately 0.1 M cacodylate, and contained 6.8% sucrose. The following M Na?. day, the tissues were washed in solutions containing 6.8% sucrose and 0.1% RR, buffered with either 0.15 M or 0.015 M cacodylate for 1 hr, then postfixed in 1% RESULTS Os04, 0.05% RR, buffered with either 0.1 M, or 0.01 M Because of the delay before fixation and the trauma cacodylate. The tissues were dehydrated in ethanol and embedded in Spurr. Additional fragments of retina were caused by mincing the tissue while it is still fresh, the immersed in high-ionic-strength fixative, but without ultrastructural appearance of the retinas was not as RR. These tissues were subsequently processed in solu- good as that of either freshly immersion-fixed retina or tions of high ionic strength that did not contain RR. perfused tissue. There was myelin figure formation, and Ultrathin sections were stained with lead citrate and/or some swelling of cellular processes. Preservation was ethanolic uranyl acetate and examined with a Zeiss EM more than adequate, however, for identification and ultrastructural study of capillaries and the components of 109 electron microscope. According to Scott and Dorling (1965), the staining of their walls. The internal limiting membranes ( E M ) and interphoanionic micromolecules by cationic dyes can be reversibly inhibited by increasing concentrations of inorganic toreceptor matrices (IPM) of retinas fixed and processed cations. The concentration of cation at which this occurs according to both the high-ionic-strength and the lowis termed the critical electrolyte concentration (CEC). ionic-strength protocols were diffusely electron-dense, Ultrastructural studies by Charonis and Wissig (1983) without any suggestion of discrete or punctate deposihave recently shown the CEC to be in the range of 0.5- tion of the RR (Fig. 1). These findings are in accord with 1.3 M for capillary BM in nonneural organs. Although I previous observations on the ILM (Matsusaka, 19711, have recently demonstrated (Schmidley, in press) that and with the presence of GAGS in the IPM (Hewitt, BM of isolated rat cerebral cortical microvessels do not 1986). exhibit RR-staining sites even after immersion in RRThe BMs of capillaries, small arterioles, and veins in aldehyde solutions containing only 0.01 M Na+, the low- retinas fixed and processed in solutions of low ionic ionic-strength fixative protocol was used in these exper- strength exhibited discrete, electron-dense sites along RUTHENIUM RED STAINING OF BASEMENT MEMBRANES Figs. 2, 3. Deposits of RR in the BM of retina1 capillaries. A few of these are indicated by arrows. In this and subsequent figures, N=nucleus, P=pericyte, E=endothelial cell, L=lumen. 61,000~. 365 366 J.W. SCHMIDLEY Fig. 4.Where the BM splits to enclose a pericyte, RR-staining sites (some of which are marked by arrows) are seen along the endothelial and retinal surfaces of the BM, as well as in the BM on both sides of the pericyte. 61,000 X . their inner and outer surfaces; as in other BMs, these sites were 10-20 nm in diameter and arranged in a more or less regular fashion at intervals of 20-70 nm (Figs. 2, 3) (Charonis and Wissig, 1983). Where the BM split to enclose a pericyte or smooth muscle cell, sites could be seen along both surfaces of the enclosed cell (Fig. 4). Where the BM was sectioned tangentially, the electrondense sites could be seen in a more or less regular array (Fig. 5). The sites were seen much less frequently, and when seen were less distinct, in BM of capillaries of retinas fixed and processed in solutions of high ionic strength (not illustrated). Electron-dense sites were never observed in BM of retinal microvessels fixed and processed in solutions not containing RR (Fig. 6). DISCUSSION The results suggest that anionic macromolecules in the BM of “barrier-type” microvessels can be stained in a discrete fashion with the cationic dye RR. The sites of RR deposition in BM are thought to represent accumulation of this cationic molecule bound to the polyanionic GAG chains of PG molecules (Farquhar, 1981).Heparan sulfate is a known constituent of bovine retinal microvessel BM (Cohen and Ciborowski, 1981; Kennedy et al., 1986), and it is very likely responsible for RR accumulation. However, in vitro RR staining is not specific for GAGs (Luft, 1971),and therefore it cannot be considered as proven that the RR deposits observed in these experiments represent individual PG molecules. The pattern of staining is very similar to that seen in the BM of nonneural capillaries when they are fixed in aldehydes containing RR (see, for example, Charonis and Wissig, 1983). The only difference between retinal and nonneural capillaries seems to be that in the former, staining is more frequent and more distinct after exposure to fixatives of low ionic strength, suggesting that the molecule(s) stained are relatively less anionic than those of other BM. If the stained molecules were GAGs, for example, they might be less sulfated than GAGs in other BMs. The latter hypothesis is testable by direct biochemical analysis (Robinson et al., 1984). With regard to the original observations that stimulated these experiments (Schmidley and Wissig, 1986), the findings suggest that the failure t o produce discrete RR-staining sites in the BBB of rat brain microvessels is not the result of a unique or universal ultrastructural characteristic of the BM of microvessels forming interfaces between blood and neural tissue, which would RUTHENIUM RED STAINING OF BASEMENT MEMBRANES 367 Fig. 6. Bovine retina fixed in aldehydes, without RR. No electrondense sites are seen in the capillary BM. 62,000X . Fig. 5. Tangential section of BM. RR staining sites are arranged in a quasi-regular array. 61,000 X . possible that, the immunohistochemical studies (Laurie et al., 1983) notwithstanding, PG molecules, or at least their GAG components, are absent from rat BBB capillary BM. Alternatively, the GAGS could be undersulprevent the formation of discrete, electron-dense depos- fated (Robinson et al., 1984). This possibility is also its of RR. Of course, it is still possible that the structures testable by direct biochemical analyses of BM, which of the two “barrier-type” BM are different, such that the are underway in our laboratory. structure of the BM of BBB capillaries in the rat proACKNOWLEDGMENTS hibits, while that of the BM of bovine retina1 capillaries permits, the formation of discrete RR deposits in associIt is a pleasure to thank Patricia Blue for excellent ation with GAG molecules or other polyanions. It is also technical assistance, and Sandy D’Amico for help in 368 J.W. SCHMIDLEY preparing the manuscript. This study was supported by grants NS-22524 and HL-35617 from the U S . Public Health Service, and by Clinician-Scientist Award 81428 from the American Heart Association. LITERATURE CITED Charonis, A.S., and S.L. Wissig (1983)Anionic sites in basement membranes, differences in their electrostatic properties in continuous and fenestrated capillaries. Microvasc. Res., 25:265-285. Charonis, A.S., P.C. Tsilibary, R.H. Kramer, and S.L. Wissig (1983) Localization of heparan sulfate proteoglycan in the basement of continuous capillaries. Microvasc. Res., 26t108-115. Cohen, M.P., and C.J. Ciborowski (1981) Presence of glycosaminoglycans in retinal capillary basement membrane. Biochim. Biophys. Acta, 674r400-406. Farquhar, M.G. (1981) The glomerular basement membrane. In: Cell Biology of Extracellular Matrix. E.D. Hay, ed. Plenum, New York, pp. 335-378. Hascall, G.K. (1980)Cartilage proteoglycans: Comparison of sectioned and spread whole molecules. J. Ultrastruct. Res., 70:369-375. Hewitt, A.T. (1986) Extracellular matrix molecules. In: The Retina: A Model for Cell Biology Studies. R. Adler and D. Farber, eds. Academic, San Diego, pp. 169-214. Kanwar, Y.S., and M.G. Farquhar (1979a) Presence of heparan sulfate in the glomerular basement membrane. Proc. Natl. Acad. Sci. U.S.A., 76:1303-1307. Kanwar, Y.S., and M.G. Farquhar (1979b) Anionic sites in the glomerular basement membrane. In vivo and in vitro localization to the laminae rarae by cationic probes. J. Cell Biol., 81t746-754. Kennedy, A., R.N. Frank, and M.A. Mancini (1986)In vitro production of glycosaminoglycans by retinal microvessel cells and lens epithelium. Invest. Ophthalmol. Vis. Sci., 27t746-754. Laurie, G.W., C.P. LeBlond, and G.R. Martin (1983) Light microscopic immunolocalization of Type IV collagen, laminin, heparan sulfate, proteoglycan and fibronectin in the basement membranes of a variety of rat organs. Am. J. Anat., 167r71-82. Luft, J.H. (1971) Ruthenium red and violet. I. Chemistry, purification, methods for use in electron microscopy and mechanism of action. Anat. Rec., 171t347-368. Matsusaka, T. (1971)The fine structure of the internal limiting membrane of the rat retina as revealed by ruthenium red staining. J. Ultrastruct. Res., 36:312-317. Raviola, G. (1977) The structural basis of the blood-ocular barriers. Exp. Eye Res., 25(Suppl.):27-63. Robinson, J., M. Viti, and M. Hook (1984) Structure and properties of a n under-sulfated heparan sulfate proteoglycan synthesized by a rat hepatoma cell line. J. Cell Biol., 98:946-953. Schmidley, J.W. Ultrastructural studies of basement membranes of microvessels isolated from rat brain: Lack of staining with ruthenium red. Microvasc. Res., in press. Schmidley, J.W., and S.L. Wissig (1986) The basement membrane of CNS capillaries lacks ruthenium red staining sites. Microvasc. Res., 32.'300-314. Scott. J.E., and J. Dorling (1965) Differential staining of acid glycosaminoglycans by Alciin Blue in salt solutions. Hiscochemie, 5:221233. Simionescu, M., N. Simionescu, and G.E. Palade (1984) Partial chemical characterization of the anionic sites in the basal lamina of fenestrated capillaries. Microvasc. Res., 28t352-367. Thurauf, N., R. Dermietzel, P. Kalweit (1983) Surface changes associated with fenestrated brain capillaries. I. In vitro labelling of anionic sites. J. Ultrastruct. Res., 84:103-110.