Ultrastructural and histochemical studies of the rat iris: identified neuronal inputs and supportive glia

Innervation of rat iris by trigeminal and ciliary neurons expressing pChAT, a novel splice variant of choline acetyltransferase

We have recently discovered a splice variant of choline acetyltransferase (ChAT) mRNA and designated the variant protein pChAT because of its preferential expression in peripheral neuronal structures. In this study, the presence of pChAT in rat iris was examined by immunohistochemistry and Western blot using a pChAT antiserum, in combination with RT-PCR analysis and ChAT enzyme assay. For comparison, the conventional ChAT (cChAT) was studied in parallel. By pChAT immunohistochemistry, intense labeling was found to occur in nerve fibers of the iris and in neurons of the ciliary and trigeminal ganglia. Denervation studies, analyzed by semiquantitative morphometry, indicated that these iridial pChAT fibers originated about half from the ciliary ganglion and the other half from the trigeminal ganglion. The presence of pChAT protein in the iris and trigeminal ganglion was confirmed by Western blot. The expression of pChAT mRNA in the ciliary and trigeminal ganglia was proved by RT-PCR. Although cChAT protein and mRNA were detected in the ciliary ganglion, neither was detectable in the trigeminal ganglion. The contributions of the ciliary and trigeminal ganglia to the iridial ChAT enzyme activity were verified by the present ChAT assay. Here, we provide evidence that iridial pChAT nerves are composed of postganglionic parasympathetic efferents from the ciliary ganglion and, more interestingly, somatic sensory afferents of the trigeminal ophthalmic nerve.

Specialised sympathetic neuroeffector associations in rat iris arterioles

Vascular sympathetic neuroeffector associations have been examined in rat iris arterioles using serial section electron microscopy and reconstruction techniques. Examination of random sections showed that, of all profiles of varicosities (199) seen to lie closer than 4 microns to vascular smooth muscle cells, only a small proportion (29/199) were found in close association with vascular smooth muscle cells, where adjacent membranes were separated by less than 100 nm. However, serial section examination, from intervaricose region to intervaricose region, of 79 varicosities similarly observed lying within 4 microns of vascular smooth muscle cells showed that 54 formed close associations with vascular smooth muscle cells. In serial sections, all these varicosities were also closely associated with melanocytes and of the 25 remaining varicosities, 22 formed close associations with melanocytes alone, whilst 3 did not come into close association with any effector cell. The increased observation of close associations with vascular smooth muscle cells in serial sections, compared with random sections, is consistent with the demonstration that the area of contact only occupies, on average, a small percentage (5%) of the total surface area of the varicosity as seen in the 3-dimensional reconstructions. In both random and serial sections, close associations were observed between varicosities and vascular smooth muscle cells or melanocytes irrespective of whether fibres were present singly or in small nerve bundles. Three-dimensional reconstruction of associations of varicosities and vascular smooth muscle cells demonstrated several common features, such as accumulations of synaptic vesicles and loss of Schwann cell covering at the region of membrane facing the effector cell. The similarity in the appearance of the neuroeffector association seen in this study and those described in previous studies provides evidence for the existence of a common sympathetic neuroeffector association, irrespective of the receptor subtype involved in neurotransmission.

Sympathetic innervation of the rat's eye and peripheral ganglia: an electron microscopic autoradiographic tracing study

The sympathetic innervation of the rat anterior eye segment and related peripheral ganglia was studied using light and electron microscopic autoradiography after anterograde tracing with 3H-leucine from the superior cervical ganglion. In the trigeminal and pterygopalatine ganglia, unmyelinated nerve fibers were labeled. Some fibers contained accumulations of small vesicles. Close apposition of these labeled sympathetic fibers with other unmyelinated fibers was common, and was also observed at sites where accumulations of vesicles were found. In the iris, ciliary body and trabeculum, numerous fibers and vesicle-containing varicosities were labeled, which all had a similar morphology. No labeling was found in the cornea. Sympathetic fibers traversing the trigeminal and pterygopalatine ganglion closely appose other unmyelinated fibers and contain clusters of vesicles without forming classical synaptic contacts. However, non-synaptic information transfer needs further investigation. The anterior eye segment, except for the cornea, is densely innervated by sympathetic nerve terminals.

The relationship of astrocyte-like cells to the vessels that contribute to the blood-ocular barriers

Brain capillaries form a selective interface, the blood-brain barrier (BBB), between the neural parenchyma and the blood. The factors which regulate this interface are poorly understood. Both the iris and retina possess vascular beds that express some BBB characteristics; therefore, they provide attractive models to further our understanding of how blood-tissue interfaces are regulated. We have determined whether three BBB markers: the transferrin receptor, P-glycoprotein, and gamma-glutamyl transpeptidase (gamma-GTP), can be localized in the capillaries of the rat retina and iris. We have also compared, in retina and iris, the relationship which GFAP-positive cells have with the blood vessels to the expression of the three BBB markers by the vessels. Immunocytochemistry revealed that capillaries throughout the retina express P-glycoprotein and the transferrin receptor. Retinal vessels do not show detectable gamma-GTP activity. GFAP-positive cells ensheath capillaries in the nerve fibre layer of the retina. Of the three BBB characteristics we examined, iridial vessels expressed only one of them: P-glycoprotein. In the iris, GFAP-positive cells do not ensheath capillaries. From our results we conclude that all BBB characteristics do not have to be expressed and regulated in capillaries as a unit. Our results, in combination with those of earlier studies, suggest that the expression of some BBB features does not require intimate contact between capillaries and astrocytes or astrocyte-like cells. Barrier maintenance appears to be a complex process which involves the integration of several factors.

Sympathetic transmission to the dilator muscle of the rat iris

The responses of the dilator layer of the rat iris to sympathetic nerve stimulation were examined using intracellular recording techniques. Three different cell types were detected. In two of these, which were assumed to reflect recordings from myoepithelial cells, sympathetic nerve stimulation initiated excitatory junction potentials. These started after a delay of several hundred milliseconds and lasted for several seconds. The excitatory junction potentials were abolished by low concentrations of prazosin and were relatively insensitive to yohimbine, indicating that neurally released noradrenaline activated an alpha 1-adrenoceptor. The adrenoceptor was further characterised as being of the alpha 1b subtype using chlorethylclonidine. The time course of excitatory junction potentials was slowed when the preparation was cooled, suggesting that a second messenger pathway was being activated. The contractions triggered by sympathetic nerve stimulation persisted after excitatory junction potentials had been abolished by reducing the external concentration of chloride ions and after adding the organic calcium antagonist, nifedipine. Thus it seems likely that contractions of the dilator are triggered by the release of calcium ions from internal stores. These observations are discussed in relation to the idea that alpha 1b-adrenoceptors are coupled to a messenger pathway which involves inositol triphosphate and the pulsatile release of calcium ions from internal stores. The second section of the paper deals with the structure of neuro-myoepithelial contacts in the dilator layer. The majority of sympathetic varicosities formed organized neuroeffector junctions with either myoepithelial cells or melanophores. At the junctions the effector cell membrane and varicosity membrane were separated by less than 80 nm, with synaptic vesicles concentrated towards the neuroeffector junction. The synaptic vesicles in varicosities that failed to form junctions did not aggregate towards their regions of exposed membrane. These observations are discussed in relation to the idea that transmission at autonomic varicosities occurs at organised neuroeffector junctions.

Cells Expressing mRNA for Neurotrophins and their Receptors During Embryonic Rat Development

In situ hybridization analysis of cells expressing messenger RNAs (mRNAs) for the neurotrophins nerve growth factor (NGF), brain-derived neurotrophic factor (BDNF) and neurotrophin-3 (NT-3) and their high-affinity receptors (trk, trkB and trkC) in the rat embryo revealed a complex but specific expression pattern for each of these mRNAs. For all mRNAs a developmentally regulated expression was seen in many different tissues. BDNF and NT-3 mRNAs were expressed in the sensory epithelia of the cochlea and vestibule macula of the sacculus and utricle, and both trkB and trkC mRNA were expressed in the spiral and vestibule ganglia innervating these sensory structures. NGF and NT-3 mRNA were found in the iris, innervated by the sympathetic neurons of the superior cervical ganglion and sensory neurons from the trigeminal ganglion, which expressed both trk and trkC mRNAs. Both NGF and NT-3 mRNAs were also expressed in other target fields of the trigeminal ganglion, the epithelium of the whisker follicles (NT-3 mRNA) and in the epithelium of the nose, tongue and jaw. NT-3 mRNA was found in the cerebellar external granule layer and trkC mRNA in the Purkinje layer of the cerebellar primordia. These sites of synthesis are consistent with a target-derived neurotrophic interaction for NGF, BDNF and NT-3. However, in some cases mRNAs for both the neurotrophins and their high-affinity receptors were detected in the same tissue, including the dorsal root, geniculate, superior, jugular, petrose and nodose ganglia, as well as in the hippocampus, frontal cortical plate and pineal recess, implying a local mode of action. Combined, these data suggest a broad function for the neurotrophins and their receptors in supporting neural innervation during embryonic development. The results also identify several novel neuronal systems that are likely to depend on the neurotrophins in vivo.

Nerve growth factor synthesis in cultured rat iris: modulation by endogenous transmitter substances

Organ cultures of rat iris show a characteristic change in the levels of both nerve growth factor (NGF) and its mRNA: a rapid but transient initial increase is followed by a smaller but persistently elevated NGF synthesis. This time course may be influenced by release of a factor(s) from degenerating nerve terminals and/or by the lack of some factor(s) repressing NGF synthesis in vivo. We therefore analyzed the influence of biogenic amine transmitter substances and putative neuropeptides on this elevation of NGF synthesis in cultured iris. The marked increase of NGF synthesis seen initially in culture was not completely mimicked by any of the substances tested. A specific increase in NGF production up to 150% of control was observed only with cGMP. We also obtained some evidence that reaction to trauma following the culture procedure could enhance NGF production: cutting of irides into small pieces increased NGF production in culture up to 250% of control and, vice versa, treatment with 1 microM dexamethasone decreased NGF production to about 60% of control. However, the sympathetic neurotransmitter norepinephrine (NE) decreased both NGF and its mRNA levels specifically in a dose-dependent manner (0.01-1 mM) to a minimum of about 25% of control. In situ hybridization with mRNA(NGF)-specific probes showed that in cultures of dissociated iris cells all cells were capable of expressing mRNA(NGF), but that 0.1 mM NE preferentially decreased expression of mRNA(NGF) in smooth muscle cells. Thus, our results indicate that the sympathetic transmitter NE is capable of downregulating NGF synthesis in the target cells of sympathetic neurons.

An increase of vasoactive intestinal polypeptide-, but not neuropeptide Y-, substance P- or catecholamine-containing nerves in the iris of the streptozotocin-induced diabetic rat

The distribution of adrenergic and vasoactive intestinal polypeptide-, neuropeptide Y- and substance P-immunoreactive nerves was studied histochemically and immunohistochemically in the irides of rats 8 weeks after the induction of diabetes with streptozotocin. In the control animals, catecholamine-containing, vasoactive intestinal polypeptide- and substance P-immunoreactive nerve fibres were found in the constrictor pupillae, dilator muscle and the ciliary processes. They also formed perivascular nerve plexuses of blood vessels in the dilator muscle. Neuropeptide Y-immunoreactive nerve fibres were only observed in the dilator muscle and ciliary processes. In the irides from diabetic animals, a considerable increase was observed in the fluorescence intensity and/or density of vasoactive intestinal polypeptide-immunoreactive nerves. Some varicosities of the vasoactive intestinal polypeptide-immunoreactive nerves appeared enlarged. In contrast, no apparent change in the density and/or fluorescence intensity of catecholamine-containing, neuropeptide Y- and substance P-immunoreactive nerve fibres was observed in the irides from diabetic animals when compared with controls. The results are discussed in relation to the symptoms of autonomic neuropathy of the irides in diabetes.

Co-existence of neuropeptides in sympathetic, cranial autonomic and sensory neurons innervating the iris of the guinea-pig

We have used double-labelling immunofluorescence to identify the peptide content of autonomic and sensory neurons innervating the iris of albino guinea-pigs. Four major classes of neurons were identified on the basis of their distributions, origins and immunohistochemical characteristics. A dense plexus of noradrenergic axons in the constrictor and dilator muscles of the iris originated from the superior cervical ganglion, and contained immunoreactivity (IR) to both neuropeptide Y (NPY) and dynorphin (DYN). The constrictor and dilator muscles were also supplied with a dense plexus of axons with IR to substance P (SP). These axons probably originated from SP-IR nerve cell bodies located along the ciliary nerves, and are almost certainly the same axons as those producing cholinergic pupilloconstriction. The iris was also innervated by unmyelinated, capsaicin-sensitive axons with IR to both SP and calcitonin gene-related peptide. Most of these axons also contained IR to DYN and some were also IR for cholecystokinin. These axons are almost certainly sensory. Axons containing IR to both NPY and vasoactive intestinal peptide (VIP) were common in the ciliary processes, and also formed a sparse plexus near the ciliary margin of the dilator muscle. Following surgical sympathetic denervation these axons showed IR for dopamine-beta-hydroxylase; they seemed to originate from the sphenopalatine ganglion. These results demonstrate that there are well-defined patterns of coexistence of neuropeptides in the autonomic and sensory neurons supplying the iris of guinea-pigs. To understand the physiological roles of these peptides, it will be necessary to consider the possibility of complex interactions between them.

Sympathetic nerve fibers in rat orofacial and cerebral tissues as revealed by the HRP-WGA tracing technique: a light and electron microscopic study

Horseradish peroxidase-wheat germ agglutinin injected into the superior cervical ganglion is taken up by the neuronal perikarya and transported anterogradely in large quantities into peripheral fibers and axonal terminals in a variety of orofacial and cerebral tissues, including the iris, cornea, pineal gland, facial skin and the adventia of cerebral and facial blood vessels. Subsequent histochemical processing according to the tetramethylbenzidine procedure produces a unique, highly visible intraaxonal marker that makes the identification and 3-dimensional tracing of the sympathetic nerve fibers clear and unambiguous at both the light and electron microscopic levels.

Immature and mature neurofilament-immunoreactive trigeminal fibers can innervate the iris as studied by intraocular grafting of iris and trigeminal ganglia

Using immunohistochemistry on stretch-prepared whole mounts of adult rat irides, a dense, well-organized plexus of neurofilament-positive nerves originating in the trigeminal ganglion can be visualized. Such a two-dimensional tissue preparation is well-suited for studies on sensory and autonomic nerve fiber growth. In the present study the growth capacity of such neurofilament-positive nerves has been studied immunohistochemically. In irides homologously transplanted to the anterior eye chamber of adult albino rats, the intrinsic neurofilament-positive network had almost completely disappeared 4 days postoperatively. In whole mounts of iris grafts after 15 days and 4 weeks in oculo a gradually increasing plexus of nerves was observed. After 3.5 months in oculo a dense, regular network of fluorescent fibers had formed in the iris grafts to the same magnitude as in situ. However, whereas large axon bundles constituted a prominent feature of the distribution of neurofilament-positive nerves in situ, only a few and relatively thin axon bundles were seen in the grafts. The growth capacity of the neurofilament-positive trigeminal nerves was also studied by grafting fetal trigeminal ganglia to the anterior eye chamber. As visualized in cryostat sections, trigeminal grafts contained a large number of strongly fluorescent perikarya and a high density of positive fibers after intraocular maturation. Such grafts readily innervated the host iris. In the area immediately adjacent to the grafts, thin, parallel, rather weakly fluorescent fibers radiated out from the ganglia. When mature trigeminal grafts with attached host iris were regrafted to the anterior eye chamber of adult animals for a few days, in order to remove the intrinsic host iris innervation, such irides showed outgrowing fibers, often organized in small axon bundles, at long distances from the ganglion graft. The present report shows that both mature and immature neurofilament-immunoreactive neurons are capable of innervating the iris. Furthermore, this ingrowth can occur both in the presence and absence of normal intrinsic neurofilament-positive nerve fibers.

Ultrastructural and histochemical evidence for differentiation of intraocular locus coeruleus grafts and invasion of the host iris by central neurites and glia

Intraocular grafts of dorso-lateral pons, including the noradrenaline-containing cell group locus coeruleus, have been studied with ultrastructural and histochemical techniques. Also, the invasion of neuronal and glial constituents from the grafts into the iris of the host animal is described. In mature brain grafts, aggregates of locus coeruleus neurons were easily discernible with monoamine histofluorescence. These cells had an ultrastructural appearance very similar to that in situ. Numerous somatic spines were frequently associated with synaptic specializations, and monoamine-containing vesicles could be found scattered in the cytoplasm of the locus coeruleus cells. Large neurons of the nucleus tractus mesencephalici nervi trigemini were also found. These cells were neurofilament-immunoreactive just as in situ, and were ultrastructurally characterized by size, distribution of the granular endoplasmic reticulum and abundant large terminals in synaptic contact with their somata and processes. All grafts showed a vigorous astroglial proliferation, evidenced both with immunohistochemistry of glial fibrillary acidic protein and electron microscopy. The astroglial cells were more numerous, larger and with more processes than in adult in situ counterparts. At the attachment site of the brain stem grafts, the iris dilator plate was entirely changed ultrastructurally by a vigorous invasion of neuronal and astrocytic processes. The normal, loose connective tissue stroma of the iris was replaced by layers of almost exclusively central nerve fibres and astrocytes respectively. Monoamine histofluorescence demonstrated an extreme adrenergic hyperinnervation of the iris at the attachment site of the graft, compared to the normal sympathetic ground plexus, whereas neurofilament immunohistochemistry did not visualize any substantial ingrowth of such positive central nerve fibres. Immunohistochemistry of glial fibrillary acidic protein strongly supported the ultrastructural evaluation, showing profound astroglial invasion deep into the iris stroma. Electron microscopic identification of central nerve fibres in the iris showed numerous adrenergic locus coeruleus fibres with small dense-core vesicles. Also, bundles of thin, central, unmyelinated axons were found deep in the iris as well as occasional dendrites. Both large dense-cored and small clear vesicles were encountered in the iris fibres of brain graft origin. Axo-dendritic synaptic specializations formed by locus coeruleus-derived adrenergic fibres were found in the iris.(ABSTRACT TRUNCATED AT 400 WORDS)