Abstract
To better understand the functional role of EphA5 in the adult human central nervous system (CNS), we performed an immunohistochemical mapping study. EphA5, like other members of the Elk/Eph family of receptor tyrosine kinases, was widely distributed in CNS neurons. However, the distribution of the neuronal staining was not uniform. The abundance of stained neurons appeared to increase from the forebrain to the hindbrain and spinal cord. Glial and endothelial tissue was unstained. These findings are consistent with the existence of receptor and ligand gradients in different brain regions. The localization of EphA5 to motor and sensory neurons is consistent with a role of EphA5 in neural plasticity, cell-cell recognition, and topographical orientation of neuronal systems.
T
In general, RTKs have different functions at various stages of development. There is considerable information on RTKs as receptors for peptide growth factors and their successive roles during embryonic development in driving cell proliferation, survival, and differentiation (Raff et al. 1985; Ullrich and Schlessinger 1990; Barres et al. 1992). EphA5 has been identified in various mammalian species as a member of the large subfamily of Eph-related RTKs (Maisonpierre et al. 1993; Sajjadi and Pasquale 1993; Siever and Verderame 1994; Zhou et al. 1994; Taylor et al. 1995; Miescher et al. 1997). The main features of Eph-like RTKs, such as an amino terminal immunoglobulin-like domain followed by a 19-20 cysteine residue-rich region and two fibronectin Type III repeats, give them an extracellular structure resembling that of various adhesion molecules (Tuzi and Gullick 1994). The catalytic cytoplasmic domain of all RTKs exhibits the greatest homology to the corresponding domain of intracellular tyrosine kinases, such as Src, rather than transmembrane RTKs and, by analogy, their intracellular signals may differ significantly from those of peptide growth factor receptors. The Eph-related RTKs bind to different membrane-anchored ligands and are believed to mediate cell-cell interactions.
Four related ligands are presently known to induce autophosphorylation of EphA5, i.e., ephrin-A1 (Davis et al. 1994), ephrin-A2 (Cheng et al. 1995), ephrin-A3 (Beckmann et al. 1994; Davis et al. 1994), and ephrin-A5 (Drescher et al. 1995; Winslow et al. 1995). Ephrin-A5 and ephrin-A2 in the rat and chicken play an important role in the guidance of retinal axons by inducing growth cone collapse and repulsion when receptor and ligand interact (Drescher et al. 1995). Ephrin-A2 and one of its receptors, EphA3, are expressed in matching gradients in the tectum and retina, respectively, suggesting a role in establishing topographic retinotectal connections (Cheng and Flanagan 1994; Shao et al. 1995). Similarly, mouse EphA5 and its ligands are expressed during development in the hippocampus and lateral septum, respectively, in matching gradients (Zhang et al. 1997), also suggesting a role in the development of hippocamposeptal topographical projections. Mouse EphA5 is widely distributed during neurogenesis, compared to its rather specialized localizations during embryogenesis. This suggests that EphA5 may play a role during neurogenesis because neuron target interactions would not yet have begun (Zhang et al. 1997).
Neural fasciculation is another important phenomenon during neural development in which EphA5 and its ligands have been implicated. In vitro studies using soluble ephrin-A5 in co-cultures of rat astrocytes expressing EphA5 ligands and of rat neurons resulted in the blocking of neurite fasciculation (Winslow et al. 1995).
The majority of research on the Eph family has focused on development, during which changes in receptor and ligand expression levels are very marked. The role/function of continued expression of EphA5 in adulthood remains unclear. However, other members of the RTK superfamily, such as epidermal growth factor and fibroblast growth factor, serve in the mature nervous system to maintain the long-term survival (Hughes et al. 1993) and synaptic integrity (Falls et al. 1993) of certain neuronal populations. It is therefore likely that EphA5 also has an important role in maintenance of the adult CNS.
This report describes the distribution of EphA5 in the adult human CNS, using a sensitive immunohistochemical procedure (Miescher et al. 1997) and a polyclonal affinity-purified antibody against the kinase domain of EphA5.
Materials and Methods
Tissue Specimens
Human CNS samples were obtained from autopsies without neurological disease 6 hr or less post mortem. All samples were from three women and one man between the ages of 60 and 75 years. The sample collection was approved by the Hospital Ethics Committee.
Biochemical Analysis
Autopsy samples were snap-frozen and stored at −70C. The Western blot analysis was performed as described elsewhere (Taylor et al. 1995). The affinity-purified polyclonal antibody used for this study was raised against the intracellular catalytic kinase domain of EphA5 and was tested for specificity in in vitro transcription/translation experiments (Taylor et al. 1995). Another affinity-purified polyclonal antibody raised to the second extracellular fibronectin domain of EphA5 was shown immunohistochemically to give the same results, further confirming the specificity of the kinase domain antibody (Miescher et al. 1997).
Light Microscopic Immunohistochemistry
Human brain autopsy specimens embedded in paraffin were sectioned at 4-μm thickness, mounted on gelatin-chromalum-coated glass slides, and deparaffinized. Endogenous peroxidase activity was blocked by bathing the sections in 80% methanol, 0.6% H2O2 for 20 min at room temperature. The sections were incubated successively with affinity-purified polyclonal antibody to the kinase domain of EphA5 and peroxidase-labeled goat anti-rabbit IgG (Vectastain ABC Kit; Vector Laboratories, Burlingame, CA). The staining reaction using the peroxidase substrate 3-amino-9-ethylcarbazole (AEC) was performed according to the manufacturer's instructions (Vector Laboratories). The samples were washed for 20 min between antibody incubations and were then counterstained with Mayer's hemalum.
Results
Biochemical Analysis
Western analysis of selected brain regions of autopsy material revealed a 120-kD protein in all brain regions tested, with the exception of white matter (Figure 1). Therefore, the white matter was used as a control for both Western analyses and immunostaining studies. The above results are consistent with previously published data showing EphA5 to be neuronally localized (Miescher et al. 1997).
Light Microscopic Immunocytochemistry
Western blot analysis of human brain regions with affinity-purified anti-EphA5 antibody. Note the human white matter control showing no reactivity to the antibody. Following a differential centrifugation protocol to enrich for EphA5 (Taylor et al. 1995), samples containing 25 μg of protein were loaded onto a 7.5% polyacrylamide gel. Molecular weight markers in Mr (1000).
The data were tabulated in semiquantitative form, listing the major centers that show distinct immunoreactivity on a scale of +/++++ (Table 1). This scale was used to illustrate the number of positively stained neurons, including processes, as a comparison among the brain areas studied. Individual nuclei are indicated according to the nomenclature in the human brain atlas by Nieuwenhuys et al. (1978).
Distribution of EphA5 immunoreactivity in human CNS
In the telencephalon, the olfactory bulb (Figure 2A and the olfactory nucleus stained poorly for EphA5. In these neurons, EphA5 appeared to aggregate in a dense perinuclear pattern. The hippocampal formation revealed particularly strong staining, of which the majority was found in the large pyramidal neurons and dendrites of the CA3 region (Figure 2B, the pyramidal cell layer, and the entorhinal cortex. EphA5 appeared to accumulate in a perinuclear pattern and, in some instances, along the plasmalemma of these neurons. In the dentate gyrus, only the plexiform layer contained EphA5 (Figure 2C. Here, again, staining was localized in the cytosol and along the dendrites. The neocortex (Figure 2D, rich in neurons, showed a relatively high staining pattern for EphA5. The dendritic arbors of the pyramidal neurons in the plexiform layer (Layer I) stained lightly for EphA5 compared to the strong perinuclear staining pattern in the cell bodies. Staining intensity increased proceeding into the deeper cortical layers, where it was found in all pyramidal neurons and their processes. Nonpyramidal neurons in the external pyramidal lamina (Layer II) and the multiform layer (Layer VI) also expressed EphA5. However, this staining was more dispersed (Figure 2D. The majority of neurons in the amygdaloid body showed strong immunoreactivity to EphA5 antibodies.
In the diencephalon, the hypothalamus (Figures 2E and 2F) contained a variety of small unidentified neurons that stained with anti-EphA5 antibodies. In contrast to the olfactory bulb and neocortex, which showed a perinuclear accumulation of EphA5, these neurons exhibited a distinct Nissl intracellular localization for the receptor. These small neurons were scattered throughout the hypothalamus and between positive nuclei, such as the anterior hypothalamic and periventricular nuclei. A large proportion of the nuclei of the thalamus showed anti-EphA5 staining, with the majority of staining restricted to the anterior thalamic, reticular thalamic, and anterior ventricular nuclei.
(
In the mesencephalon, both the superior and inferior colliculi contained EphA5. In these regions, EphA5 was found in large neuronal cells of the gray matter. Tegmental nuclei, such as the ruber nucleus (Figure 3A, mesencephalic reticular nucleus (of the trigeminal nerve), and oculomotor and interpeduncular nuclei, showed strong staining for the anti-EphA5 antibody. In the neurons of these regions, EphA5 was found in a perinuclear staining pattern, with a few large neurons exhibiting a terminal accumulation of the receptor in what appeared to be the Nissl region of the cell body (Figure 3A.
The substantia nigra, forming part of the so-called extrapyramidal motor system, was one of only a few brain regions that was distinctly negative for EphA5. On the other hand, EphA4 in the rat showed a strong staining pattern within the same area (Martone et al. 1997). The significance of the difference between human and rat remains unclear at present.
By far the highest levels of EphA5 were found in hindbrain structures. The strongest staining was found in the neurons of the pontine nucleus, locus ceruleus mesencephalic nucleus of the trigeminal nerve (Figure 3B, and many of the cranial motor nuclei. Both the inferior (Figure 3C and superior olivary complexes showed strong staining. In these regions, neurons exhibited both a perinuclear and Nissl localization of receptor. The medulla oblongata as a whole was particularly reactive and showed high levels of EphA5. Dorsal column nuclei of this region, such as the cuneate nucleus and nucleus gracilis, were strongly reactive. In the cerebellar cortex, only the Purkinje cells and their extensive arborization were positive, whereas the vast majority of cells in the granular layer and other neurons were negative (Figures 3E and 3G).
In the spinal cord, staining was restricted to the neurons of lamina Layers I and III to IX, with lamina Layers VII-IX showing the greatest intensity (Figure 3F. In some instances, staining was also observed along the plasmalemma of these neurons (Figure 3F. No staining was observed in the substantia gelatinosa (lamina Layer II), ependymal layer, and glial tissue of the spinal cord. This was also true for the glial tissue of the brain. This is in contrast to the results presented by Martone et al. (1997), who demonstrated intense staining of the substantia gelatinosa of the rat.
Discussion
This study documents the extensive distribution of EphA5, an Eph-related RTK, in the human CNS. The cellular localization of EphA5 indicates that it is expressed in neurons. In all instances, the staining appears to show a marked intracellular accumulation of EphA5 in a perinuclear pattern, with the exception of a few large neurons of the hippocampus, mesencephalon, and hindbrain, which show a distinct Nissl localization for the receptor. In neurons that show large arbors such as the pyramidal neurons of the cortex, neurons of the spinal cord, hippocampus, and all the Purkinje cells of the cerebellum, staining is found along the dendrites. Similar results, particularly in the cerebellum, were observed by Martone et al. (1997) in adult rat CNS. However, in mouse cerebellum EphA5 is expressed only in the posterior Purkinje cells (Zhang et al. 1997). Conversely, ligand expression appears to be restricted to the deep nuclei (the targets of the Purkinje cell) of the mouse cerebellum. The exact meaning of this distribution in terms of topographical interactions between receptor and ligand in the adult mouse remains unclear (Zhang et al. 1997).
At present, it cannot be certain what proportion of the staining is located at the membrane surface. The role of intracellular accumulation of EphA5 has been discussed previously (Miescher et al. 1997). However, Western blot analysis of brain membrane preparations (results not shown) and immunostaining (Figure 3F show a significant presence of the receptor at the neuronal cell membrane. Although the significance of this localization is at present unclear, EphA5, as in the case of EphA4, could be important for maintenance and/or proper functioning of multiple neuronal structures (Martone et al. 1997). The absence of biochemically and immunohistochemically detectable EphA5 in axon-rich tissue, such as white matter, may indicate an alternative role for Eph-A5 in adulthood compared to the guidance, fasciculation, and targeting cues during embryonic development. This idea is supported by the distribution of EphA4, a similar RTK, in the optic chiasm in the developing rat but not in the adult rat (Martone et al. 1997).
EphA5 parallels other RTKs with its neuronal expression. Furthermore, the expression of EphA5 is not restricted to one set of neurons but rather to a well-distributed population dispersed thoughout the CNS, as is true for other RTKs. An example includes the members of the Trk family of RTKs which, together, with their ligands, the neurotrophins, show distinct patterns of expression (Klein et al. 1990; Tessarollo et al. 1993) It is therefore possible that EphA5, like other members of the RTKs, may have complementary and overlapping roles in supporting the development and maintenance of various neuronal cell populations. Further support for this is found with mouse EphA5, the homologue to human EphA5. Mouse EphA5 is expressed at particularly high levels in the limbic system and, similar to the Trk family of RTKs, is believed to promote neuronal survival and maintenance (Zhou et al. 1994; Zhang et al. 1996, 1997). Human EphA5 appears to be more widespread than mouse EphA5 in the adult CNS and could therefore have a broader functional role owing to its more widespread distribution.
Coronal sections through (
Highest levels of EphA5 expression are found in distinct neuronal populations of the brainstem and spinal cord. These include some of the principal cholinergic nuclei (lateral nucleus of the amygdala and interpeduncular nucleus), monoaminergic nuclei (locus ceruleus, superior and inferior colliculi, entorhinal cortex, neocortex), motor and sensory nuclei (trigeminal motor and sensory nuclei, hypoglossal nucleus, vagal motor nucleus). These are regions known to have considerable neural plasticity.
In conclusion, our study has shown that EphA5 is not uniformly distributed in brain but is expressed most prominently in basal brain ganglia, various large pyramidal neurons, cerebellar Purkinje cells, and spinal motor neurons. The continued expression of EphA5 in these neural populations during adulthood may indicate a role for this RTK in direct cell interactions such as neural plasticity and topographical orientation.
Footnotes
Acknowledgements
Supported by grant no. 31-45′953.95 from the Swiss National Science Foundation and by the Swiss Cancer League of the Kantons Basel-Stadt and Basel-Land.
We thank A. Probst (University of Basel) for autopsy samples and paraffin-embedded tissue and B. Erne for invaluable immunohistochemical advice.