Contractile proteins in pericytes at the blood-brain and blood-retinal barriers.

Evidence from a variety of sources suggests that pericytes have contractile properties and may therefore function in the regulation of capillary blood flow. However, it has been suggested that contractility is not a ubiquitous function of pericytes, and that pericytes surrounding true capillaries apparently lack the machinery for contraction. The present study used a variety of techniques to investigate the expression of contractile proteins in the pericytes of the CNS. The results of immunocytochemistry on cryosections of brain and retina, retinal whole-mounts and immunoblotting of isolated brain capillaries indicate strong expression of the smooth muscle isoform of actin (alpha-SM actin) in a significant number of mid-capillary pericytes. Immunogold labelling at the ultrastructural level showed that alpha-SM actin expression in capillaries was exclusive to pericytes, and endothelial cells were negative. Compared to alpha-SM actin, non-muscle myosin was present in lower concentrations. By contrast, smooth muscle myosin isoforms, were absent. Pericytes were strongly positive for the intermediate filament protein vimentin, but lacked desmin which was consistently found in vascular smooth muscle cells. These results add support for a contractile role in pericytes of the CNS microvasculature, similar to that of vascular smooth muscle cells.

Antidepressants as analgesics: an overview of central and peripheral mechanisms of action.

Antidepressants, given systemically, are widely used for the treatment of various chronic and neuropathic pain conditions in humans. In animal studies, antidepressants exhibit analgesic properties in nociceptive, inflammatory and neuropathic test systems, with outcomes depending on the specific agent, the particular test, the route of administration and the treatment method used. Although early studies focused on central (i.e., supraspinal, spinal) actions, more recent studies have demonstrated a local peripheral analgesic effect of antidepressants. These peripheral actions raise the possibility that topical formulations of antidepressants may be a useful alternative drug delivery system for analgesia. Antidepressants exhibit a number of pharmacological actions: they block reuptake of noradrenaline and 5-hydroxytryptamine, have direct and indirect actions on opioid receptors, inhibit histamine, cholinergic, 5-hydroxytryptamine and N-methyl-D-aspartate receptors, inhibit ion channel activity, and block adenosine uptake. The involvement of these mechanisms in both central and peripheral analgesia produced by antidepressants is considered. Data illustrating the preclinical peripheral analgesic actions of antidepressants are presented, as are some aspects of the mechanisms by which these actions occur.

Peripheral antinociceptive actions of desipramine and fluoxetine in an inflammatory and neuropathic pain test in the rat.

Amitriptyline, a non-selective noradrenaline (NA) and 5-hydroxytryptamine (5-HT) reuptake inhibitor, has recently been demonstrated to produce a peripheral antinociceptive action in an inflammatory (formalin test) and a neuropathic pain model (spinal nerve ligation). In the present study, we determined whether desipramine, a selective NA reuptake inhibitor, and fluoxetine, a selective 5-HT reuptake inhibitor, could produce peripheral antinociceptive actions in these same tests. Effects on paw volume also were determined. In the 2.5% formalin test, desipramine and fluoxetine 10-300 nmol produced a dose-related reduction in phase 2 (16-60 min) flinching and biting/licking behaviours when coadministered with the formalin. Phase 1 flinch behaviours (0-12 min) were significantly reduced at the highest dose. These actions are peripherally mediated, as they were not seen when desipramine or fluoxetine (100, 300 nmol) were injected into the contralateral hindpaw. The peripheral action of desipramine and fluoxetine was not altered by coadministration of caffeine 1500 nmol. In the spinal nerve ligation model, desipramine 100 nmol, but not fluoxetine 100 nmol, produced a peripheral anti-hyperalgesic action in the hindpaw corresponding to the ligated side when thresholds were determined using a thermal paw stimulator. In paw volume experiments, desipramine, at doses which are maximally effective in behavioural tests, produced only a slight increase in paw volume, but fluoxetine (10-300 nmol) produced a robust and sustained dose-related increase in paw volume. Amitriptyline also produced minimal effects on paw volume. When coinjected with formalin, no agent significantly altered the degree of paw swelling produced by formalin. The increase in paw volume produced by fluoxetine was inhibited by ketanserin (5-HT2 receptor antagonist), mepyramine (histamine H1 receptor antagonist) and phentolamine (alpha-adrenergic receptor antagonist), but not by the other selective 5-HT receptor antagonists tested or caffeine. The pronounced peripheral pain alleviating actions in the absence of marked changes in paw volume produced by desipramine and amitriptyline, but not fluoxetine, in the formalin test and the spinal nerve ligation model suggest that these agents could be developed as cream or gel formulations to recruit a peripheral antinociceptive action in inflammatory and neuropathic pain states. Such a formulation might permit the attainment of higher and more efficacious concentrations in the region of the sensory nerve terminal, with limited systemic side effects.

A comparison of blood-brain barrier and blood-nerve barrier endothelial cell markers.

A number of major properties of endothelial cells (EC) at the blood-brain barrier (BBB) have been shown to be astrocyte-dependent. Whether analogous properties at the blood-nerve barrier (BNB) are induced and maintained by Schwann cells has not been investigated. As a preliminary investigation we have undertaken a comparative study of six EC membrane markers at the BBB and BNB and perineurium. Employing immunoblotting and immunocytochemistry the relative distribution between rat brain cortex and sciatic nerve was determined for the glucose transporter (GLUT-1), the transferin receptor (OX-26), the endothelial barrier antigen (EBA) and the OX-47 antigen. Using enzyme cytochemistry the same comparison was made for gamma-glutamyl transpeptidase (GGTP) and alkaline phosphatase. By immunocytochemistry GLUT-1 was uniformly strongly represented in brain EC, nerve EC and perineurium. OX-26 was strongly positive in brain EC but present only in trace quantities in nerve EC and perineurium. EBA similarly showed strong positivity in brain EC and trace amounts in nerve EC but was absent from perineurium. OX-47 was present moderately in brain EC and perineurium but absent from nerve EC. Quantitative immunoblotting of brain and sciatic nerve homogenates showed statistically significant differences in the level of expression of EBA and OX-26 between the two tissues. Enzyme cytochemistry showed that GGTP was strongly positive in brain EC but absent from nerve EC and perineurium. Alkaline phosphatase stained strongly in brain and nerve EC and was absent from perineurium. In summary the six membrane markers were heterogeneously represented in nerve compared with brain. This pattern of distribution in the nerve cannot simply be accounted for by the absence of astrocytes and their inductive influences. Any inductive influences of Schwann cells require investigation.

Peripheral antinociceptive action of amitriptyline in the rat formalin test: involvement of adenosine.

The present study determined (1) whether amitriptyline could produce a local antinociceptive action in the formalin test, (2) whether endogenous adenosine was involved in this action, and (3) which other systems might contribute to such an action. Coadministration of amitriptyline 10-100 nmol with 2.5% formalin produced a dose-related reduction in phase 1 (0-12 min) and phase 2 (16-60 min) flinching behaviours, as well as in phase 2 biting/licking time (no phase 1 expression). This action was not seen or only partially expressed at low concentrations of formalin (0.5%, 0.75%). Coadministration of caffeine with amitriptyline partially reversed the antinociceptive action of amitriptyline against both behaviours at 2.5% formalin. At 1.5% formalin, caffeine still produced only a partial reversal of effect; this appeared to be due to a block of adenosine A1 receptors, as it was also seen with the selective adenosine A1 receptor antagonist, 8-cyclopentyl-1,3-dimethylxanthine. Using antagonists for a number of other systems, no evidence for an involvement of alpha-adrenergic, histamine, excitatory amino acid or opioid receptors in the action of amitriptyline was observed or inferred. A local anaesthetic action for amitriptyline remains a possibility for the residual action. These results indicate that amitriptyline can produce a local peripheral antinociceptive action which is mediated, in part, by an interaction with endogenous adenosine, most likely an inhibition of the cellular uptake of adenosine with a consequent activation of adenosine A1 receptors on sensory nerve terminals. Local application of amitriptyline by cream or gel might prove to be a useful method of drug delivery in inflammatory pain states.

Cerebral and pial microvessels: differential expression of gamma-glutamyl transpeptidase and alkaline phosphatase.

Pial microvessels have several important blood-brain barrier (BBB) characteristics in common with cerebral microvessels, despite lacking their astrocytic ensheathment. We have therefore determined whether they have the same distribution of two enzymes, gamma-glutamyl transpeptidase (GGTP) and alkaline phosphatase, both of which are known to be astrocyte-dependent. GGTP was absent from all rat pial microvessels but strongly present in brain cortical capillaries. Alkaline phosphatase was heterogeneously expressed in pial microvessels, including capillaries, but strongly positive in brain cortical capillaries. Diffusible, inductive factors produced by astrocytes could account for these differences in enzyme distribution between the two vessel types. Furthermore, differences in expression between the two markers may reflect their differing sensitivities to the astrocytic factors. Caution is urged in the common usage of the pial microvessel as a model system in BBB studies.

Corneal glycoconjugates: an ultrastructural lectin-gold study.

The oligosaccharide chains of cell surface and extracellular matrix glycoconjugates are essential for the biological properties of these molecules. We have, therefore, investigated carbohydrate residues in the rat cornea using biotinylated lectin-gold probes. Fixed corneas were removed and embedded in Lowicryl HM20 or LR White. Ultrathin sections were incubated in one of the lectins: Triticum vulgare (WGA), Canavalia ensiformis (Con A), Griffonia simplicifolia (GS-1), Limax flavus (LFA) and Allomyrina dichotoma (Allo A), followed by streptavidin-gold, or the sections were incubated in cationic colloidal gold. Semi-quantification of gold labelling was determined for corneal endothelium, Descemet's membrane, stroma and epithelium from electron micrographs. WGA and Con A binding sites were expressed either moderately or strongly throughout the cornea, suggesting a preponderance of alpha-mannose and N-acetylglucosamine residues. A particular concentration of these sugars was found in Descemet's membrane. In contrast, GS-1 (specific for alpha-galactose) and Allo A (specific for beta-galactose) labelled all regions weakly. Sialic acid residues, as defined by LFA labelling and the expression of neuraminidase-sensitive cationic colloidal gold binding sites, were sparsely distributed throughout the stroma, Descemet's membrane and endothelium. In contrast, sialoglycoconjugates were found in significant concentrations in the epithelium. Electron microscopy proved useful in providing new information on the cellular and subcellular localization of these lectin binding sites.

Adenosine A3 receptor activation produces nociceptive behaviour and edema by release of histamine and 5-hydroxytryptamine.

This study evaluated the pain enhancing properties of the adenosine A3 receptor agonist N6-benzyl-5'-N-ethylcarboxamidoadenosine (N6-benzyl-NECA) by assessing behavioural effects following s.c. administration alone to the dorsal hindpaw of the rat, or in combination with a low concentration of formalin (0.5%). Edema formation was monitored by determining paw volume with plethysmometry. N6-benzyl-NECA (0.005-10 nmol) produced a dose-related increase in intrinsic flinching behaviours, as well as an increase in phase 2A flinch responses in the presence of formalin. Intrinsic effects were blocked by the histamine H1 receptor antagonist mepyramine and the 5-hydroxytryptamine2 (5-HT2) receptor antagonist ketanserin, but not by other 5-HT receptor antagonists or adenosine A1 or A2 receptor antagonists. N6-benzyl-NECA also produced an increase in paw volume, both alone and in the presence of formalin, with higher doses being required to produce this effect than for the flinch response. The increase in paw volume was also blocked by mepyramine and ketanserin but not by other antagonists. These results indicate both a nociceptive response and a proinflammatory response resulting in edema formation following activation of adenosine A3 receptors which is mediated by both 5-HT and histamine released most likely from mast cells.

Molecular characteristics of pial microvessels of the rat optic nerve. Can pial microvessels be used as a model for the blood-brain barrier?

Pial microvessels have commonly been used in studies of the blood-brain barrier because of their relative accessibility. To determine the validity of using the pial microvessel as a model system for the blood-brain barrier, we have extended the comparison of pial and cerebral microvessels at the molecular level by a partial characterization of the glycocalyx of pial endothelial cells, in view of the functional importance of anionic sites within the glycocalyx. Rat optic nerves were fixed by vascular perfusion. Anionic sites on the endothelium were labelled with cationic colloidal gold by means of post- and pre-embedding techniques. The effects of digestion of ultrathin sections on subsequent gold labelling was quantified following their treatment with a battery of enzymes. Biotinylated lectins, viz. wheat germ agglutinin and concanavalin A with streptavidin gold, were employed to identify specific saccharide residues. The results demonstrate that the luminal glycocalyx of pial microvessels is rich in sialic-acid-containing glycoproteins. Neuraminidase, which is specific for N-acetylneuraminic (sialic) acid, and papain (a protease with a wide specificity) significantly reduce cationic colloidal gold binding to the luminal endothelial cell plasma membrane. Wheat germ agglutinin (with an affinity for sialic acid) binds more to the luminal than abluminal plasma membrane, whereas concanavalin A, which binds mannose, binds more to the abluminal surface. Similar results have been obtained for cerebral cortical endothelial cells. With respect to these molecular characteristics, therefore, the pial and cortical microvessels appear to be the same. However, since the two vessel types differ in other respects, caution is urged regarding the use of pial microvessels to investigate the blood-brain barrier.

Dural microvessels: molecular properties of their luminal anionic sites.

(1) Neurogenic inflammation has been implicated in the pathogenesis of the vascular headaches of migraine and cluster headaches. (2) Dural blood vessels are both pain-sensitive and show neurogenic plasma extravasation. (3) Endothelial cell (EC) surface anionic sites appear to be a determinant of vascular permeability. We therefore examined the anionic sites of dural EC to determine whether they are different from those of pial and parenchymal vessels. Luminal anionic sites of rat optic nerve EC were labelled with cationic colloidal gold (CCG) and cationic ferritin (CF) and examined by electron microscopy. Employing a battery of enzymes, the effects of digestion of ultrathin sections on subsequent labelling with CCG was quantified using image analysis software. In addition, a gold-labelled lectin, wheat-germ agglutinin (WGA), was employed to locate specific saccharide residues. Of the enzymes with a narrow specificity, only neuraminidase substantially reduced CCG binding. Of the proteolytic enzymes, papain was most effective in reducing labelling. These results show that the luminal EC anionic sites are chiefly composed of sialoglycoproteins. The labelling with biotinylated WGA-streptavidin gold was similar to that with CCG without enzyme digestion. This suggests that WGA is binding to N-acetylneuraminic (sialic) acid residues and not to the neutral N-acetylglucosamine (since CCG would not label uncharged molecules). These results do not differ significantly from those for pial and parenchymal EC. It is therefore likely that factors other than anionic site molecular composition account for the susceptibility of dural vessels to neurogenic plasma extravasation. The relevance of these observations in an experimental animal model to the human clinical condition remains to be determined.

Optic nerve microvessels: a partial molecular definition of cell surface anionic sites.

Incorporated in the luminal glycocalyx of vascular endothelia (EC) are negatively charged microdomains (anionic sites). These sites are considered functionally important (a) in their interaction with circulating blood constituents, and (b) as a determinant of vascular permeability. The molecular composition of these EC sites, described for a number of tissues, has demonstrated a heterogeneity dependent on their anatomical location. Luminal anionic sites have not been characterized for EC of optic nerve. Optic nerves were removed from Sprague-Dawley rats previously fixed by vascular perfusion. EC anionic sites were labelled with the probes cationic colloidal gold (CCG) and cationic ferritin (CF), using the pre- and post-embedding techniques, and examined by electron microscopy. The effects of enzyme digestion of ultrathin sections on subsequent CCG labelling were determined using a battery of enzymes in association with the post-embedding technique. CCG labelling was quantified following each enzyme treatment using image analysis software. The biotinylated lectin wheat germ agglutinin (WGA) with streptavidin gold was also used to localize specific monosaccharide residues. The luminal front of intraneural EC showed a uniform labelling with CCG and CF which was greater than on the abluminal surface. Extracellular matrix components and basal laminae were moderately labelled. Digestion of tissue sections with heparitinase and trypsin had no significant effect on subsequent CCG labelling. Proteinase K was less effective than papain but both produced a significant reduction. Neuraminidase almost completely eliminated labelling. CCG binding to the luminal plasma membrane of optic nerve EC can be significantly reduced with proteolytic and glycolytic enzymes. The results demonstrate that sialoglycoproteins principally constitute these luminal EC anionic sites. Biotinylated WGA-streptavidin gold, which detects both N-acetylneuraminic (sialic) acid and N-acetylglucosamine, gave a similar pattern of labelling to CCG alone on the luminal versus abluminal EC fronts. These findings suggest that WGA is binding predominantly to N-acetylneuraminic acid residues since CCG would not label the neutral (uncharged) N-acetylglucosamine.

Distribution of a putative endothelial barrier antigen in the ocular and orbital tissues of the rat.

BACKGROUND: A rat endothelial barrier antigen (EBA) recognised by a monoclonal antibody has been shown to be expressed strongly by endothelial cells of brain capillaries possessing a blood-brain barrier and only weakly expressed by fenestrated brain vessels. METHODS: In this study immunocytochemical methods for light and electron microscopy were used to study EBA distribution in the eye and orbital tissues of the rat. RESULTS: Blood-ocular barrier vessels in the optic nerve, retina, iris, and some vessels in th choroid and ciliary body were immunopositive for EBA. By pre-embedding immunocytochemistry for electron microscopy the antigen was observed on the luminal endothelial cell surface. CONCLUSION: Surprisingly, some non-barrier vessels in the ciliary body and choroid expressed EBA suggesting that it may play a broader role in endothelial properties than previously recognised. The functional significance of EBA remains to be elucidated.

Differential expression of an endothelial barrier antigen between the CNS and the PNS.

A monoclonal antibody to an antigen (EBA) expressed by neural endothelial cells (EC) was used to investigate any difference in the distribution of EBA between the CNS and PNS. Pre-embedding ultrastructural cytochemistry of rat sciatic and optic nerves was undertaken using anti-EBA, detected with a silver-enhanced gold-conjugated secondary antibody. LM immunocytochemical localisation of EBA was also performed using an HRP-conjugated secondary antibody. EC of pial and parenchymal optic nerve vessels were strongly immunopositive for EBA. Vessels of the dura were negative. At the EM level EBA was observed on the EC luminal surface. In contrast, EC of sciatic nerve were either negative or only weakly immunopositive. The molecular characteristics and function of EBA are largely unknown. Therefore the functional significance of the present findings remains to be determined.

Molecular characterization of anionic sites on the luminal front of endoneurial capillaries in sciatic nerve.

The distribution of anionic microdomains has been described in cerebral vessels and more recently in capillaries of peripheral nerve. Evidence is accumulating that these sites play a role in the barrier function of vascular endothelia in the PNS and CNS. The chemical nature of anionic sites has been at least partly determined for cerebral vessels but not in peripheral nerve. This study reports our preliminary investigations to determine the nature of endothelial anionic sites in sciatic nerve. The effects of digestion of ultra-thin sections of nerve with a battery of proteolytic and glycolytic enzymes (papain, trypsin, proteinase K, hyaluronidase, heparinase, heparitinase and neuraminidase) on the distribution of anionic sites was determined using the label, cationic colloidal gold. Papain, a proteolytic enzyme of broad specificity, succeeded in removing the majority of cationic colloidal gold-binding sites on the luminal surface of vascular endothelia. In contrast trypsin and proteinase K were less effective, reflecting their narrower specificity. Hyaluronidase, heparinase and heparitinase did not significantly affect cationic colloidal gold-labelling. However, a considerable reduction in cationic colloidal gold-binding occurred following neuraminidase digestion. These results suggest that, as in cerebral vessels, sialic acid-containing glycoproteins are largely responsible for the negatively charged domains on the luminal membrane of endothelial cells in peripheral nerve.

ATP release from dorsal spinal cord synaptosomes: characterization and neuronal origin.

The present study determined the Ca(2+)-dependence of the release of adenosine 5'-triphosphate (ATP) from dorsal spinal cord synaptosomes evoked by depolarization with K+ and capsaicin, and the effect of intrathecal capsaicin pretreatment, dorsal rhizotomy and intrathecal pretreatment with 6-hydroxydopamine on such release. Release of ATP evoked by K+ was Ca(2+)-dependent, while that evoked by capsaicin was Ca(2+)-independent. Capsaicin pretreatment (60 micrograms, 7 days), which lesions small diameter afferents, did not alter release of ATP evoked by either K+ or capsaicin. Dorsal rhizotomy, which lesions all afferents, produced a significant reduction in the amount of ATP released from the rhizotomized side compared to the intact side. Pretreatment with 6-hydroxydopamine (100 micrograms, 7 days) to destroy adrenergic nerve terminals, markedly reduced spinal cord noradrenaline levels, but did not alter the K(+)-evoked release of ATP. These results suggest that some K(+)-evoked release of ATP could originate from large but not small diameter afferent nerve terminals in the spinal cord. ATP does not appear to originate from small diameter afferents as, although ATP is released by in vitro exposure to capsaicin, such release occurs only at high concentrations, release is Ca(2+)-independent and it is unaltered by pretreatment with capsaicin. The bulk of the ATP released from the spinal cord does not originate from descending noradrenergic nerve terminals.

Blood-nerve barrier: ultrastructural and endothelial surface charge alterations following nerve crush.

Nerve crush results in an enhanced vascular permeability of the endoneurial vessels distal to the lesion. Vascular permeability at the blood-nerve barrier (BNB) to serum proteins is influenced by many factors, including anionic surface charge, endothelial vesicular transcytosis and the presence or absence of fenestrated vessels. Using mice and rats, the present ultrastructural investigation examined the effect of nerve crush (axonotmesis) on: (1) the distribution of endothelial anionic sites and (2) the appearance of fenestrations in endoneurial vessels after 4 and 14 day intervals as demonstrated with cationic probes. Transient anionic fenestrations developed in a minority of mouse endoneurial vessels in 4-day crushed nerves, but were not found in 14-day crushed nerves of mice nor in crushed nerves of rats. The known increase in the permeability of endoneurial vessels in rats and mice was not associated with reduced luminal labelling with cationic ferritin at physiological pH. At pH 2.0 the labelling of glycocalyx moieties (such as sialic acid) with cationic colloidal gold was disrupted in some epi- and endoneurial vessels of 4-day rats, but in a greater proportion after 14 days. The enhanced permeability of the BNB during degeneration and regeneration is related to the formation of anionic fenestrations in endoneurial vessels of mice and to the reduced and uneven distribution of endothelial glycocalyx moieties that are anionic at pH 2.0 in rats.

Lanthanum tracer and freeze-fracture studies suggest that compartmentalisation of early bone matrix may be related to initial mineralisation.

In adult bone the calcified matrix and enclosed osteocytes are separated from the extracellular space by a continuous layer of bone lining cells. It thus appears that bone matrix is compartmentalised and, as such, may constitute a 'milieu intérieur' which is different from the general extracellular space. Since adult bone matrix is compartmentalised and matrix vesicles also form a microcompartment, it is conceivable that compartmentalisation, in early osteogenesis, may be a requirement for the initial events of the mineralisation process. We have therefore conducted an ultrastructural, tracer, and freeze-fracture study to determine the stage in which bone matrix becomes compartmentalised and also to find out whether there are tight junctions between osteoblasts. The results show that in early nonmineralised stages and in incipient mineralisation, lanthanum penetrates all intercellular spaces and the newly forming bone matrix which is rich in matrix vesicles and collagen. With the progression of mineralisation, when all matrix vesicles appear mineralised and calcification is 'spreading' to the surrounding matrix, lanthanum is restricted to intercellular spaces and conspicuous macular tight junctions are present between osteoblasts. We suggest that matrix vesicles act as microcompartments for calcification when the early bone matrix is in continuity with the surrounding extracellular space. In later stages, when lanthanum fails to penetrate the matrix, matrix vesicles may no longer be necessary because the bone matrix itself is compartmentalised, thus allowing for localised changes in composition that might favour mineral deposition.

Blood-nerve barrier: distribution of anionic sites on the endothelial plasma membrane and basal lamina of dorsal root ganglia.

Previous investigations of the blood-nerve barrier have correlated the greater permeability of ganglionic endoneurial vessels, compared to those of nerve trunks, with the presence of fenestrations and open intercellular junctions. Recent studies have demonstrated reduced endothelial cell surface charge in blood vessels showing greater permeability. To determine the distribution of anionic sites on the plasma membranes and basal laminae of endothelial cells in dorsal root ganglia, cationic colloidal gold and cationic ferritin were used. Electron microscopy revealed the existence of endothelial microdomains with differing labelling densities. Labelling indicated that caveolar and fenestral diaphragms and basal laminae are highly anionic at physiological pH, luminal plasma membranes and endothelial processes are moderately charged and abluminal plasma membranes are weakly anionic. Tracers did not occur in caveolae or cytoplasmic vesicles. In vitro tracer experiments at pH values of 7.3, 5.0, 3.5 and 2.0 indicated that the anionic charge on the various endothelial domains was contributed by chemical groups with differing pKa values. In summary, the labelling of ganglionic and sciatic nerve vessels was similar except for the heavy labelling of diaphragms in a minority of endoneurial vessels in ganglia. This difference is likely to account in part for the greater permeability of ganglionic endoneurial vessels. The results are discussed with regard to the blood-nerve and -brain barriers and vascular permeability in other tissues and a comparison made between the ultrastructure and anionic microdomains of epi-, peri- and endoneurial vessels of dorsal root ganglia and sciatic nerves.

Ultrastructural visualisation of proteoglycans in early unmineralised dentine of rat tooth germs stained with cuprolinic blue.

The ultrastructural distribution and localisation of proteoglycans (PGs) of early developing rat dentine were examined using cuprolinic blue in a critical electrolyte concentration procedure. Results show that the cuprolinic blue method produces images of higher morphological quality than other cationic dyes. PGs appeared as ribbon-like electron-opaque precipitates of various sizes, ranging between 1.4 and 0.2 microns in length, distributed throughout the matrix and in close association with well preserved matrix vesicles and collagen fibrils. Matrix vesicles revealed tightly packed PG filaments which appeared to be attached to their membrane. It is possible that the close association of PG filaments with matrix vesicles and collagen indicates that PGs are related to the process of mineralisation of dentine.