Raf Serine/Threonine Protein Kinases: Immunohistochemical Localization in the Mammalian Nervous System

the of the of cellular

Immunohistochemical studies have revealed that B-raf and C-Raf are widely distributed in central nervous system (CNS) areas including the hippocampus, neocortex and spinal cord (Mihály et al. , 1993Mihály & Rapp 1993). They are present not only in neurons but also in astrocytes (Mihály & Rapp 1994). Although most CNS neurons express Raf immunoreactivity, the possibility remains that Raf expression varies according to the type of growth factor receptor that a neuron expresses (Mihály & Endrész, 2000). Our ultrastructural immunohistochemical studies proved that Raf protein-like immunoreactivity was localized primarily in postsynaptic densities, dendritic spines, dendrites and soma of the neurons . On the basis of these observations the possibility arose, that Raf kinases participated in some receptor-mediated postsynaptic phenomena which might initiate long-term changes in postsynaptic neurons (Mihály et al., 1990). The present review collects the relevant localization data of Raf protein kinases in CNS structures. The localization of the Raf proteins will be discussed in light of some plasticity experiments (Mihály et al., 1990;1996).

Detection of Raf protein kinases in the brain with Western blotting
The first literature data about the presence of Raf protein kinase in the mammalian brain are those of . In these studies we used the polyclonal v-Raf antiserum which was produced in the laboratory of Dr. Ulf Rapp. The serum was raised in rabbit, against a large (30 kDa) C-terminal Raf protein, expressed in E. coli, by inserting the v-Raf oncogene into the bacterial DNA (Kolch et al., 1988). Adult Wistar rats were deeply anesthetized with halothane, decapitated and the brains were homogenized, lysed and immunprecipitated. The precipitates were subjected to gel electrophoresis, then proteins were transferred to nitrocellulose membranes and incubated with the polyclonal v-Raf antiserum . One conspicuous 95 kDa protein band was detected with this method . The same antibody was used on isolated subcellular fractions of the rat brain, where the cytosolic-and microsome fractions displayed positive signals, again at 95 kDa (Mihály et al., 1996). Efforts aiming the detection of C-Raf and B-Raf separately, were also successful (Morice et al., 1999): Western blotting revealed the presence of C-Raf and B-Raf proteins in different areas of the rat brain (Morice et al., 1999). Immunoblotting of guinea pig brain www.intechopen.com Raf Serine/Threonine Protein Kinases: Immunohistochemical Localization in the Mammalian Nervous System 425 homogenates also resulted in positive signals at 95 kDa with the aforementioned polyclonal anti-v-Raf serum (Mihály & Endrész, 2000). Several years later, using specific, polyclonal Braf antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA), we detected the 95 kDa protein again in rat brain homogenates (Mihály et al., 2007). Therefore, we conclude, that brain B-Raf kinase is ubiquitous and detectable through Western blotting in the rodent brain, as a 95 kDa protein (Fig. 1).  (Mihály et al., 1996). Fig. 1b: "Ctx" represents the signal obtained from brain cortex homogenate; "Bst" represents the signal obtained from brain stem homogentes (Mihály et al., 2007). The two antibodies detect one 95 kDa protein, indicating the presence of B-raf kinase in the rat brain.

Light microscopic localization of Raf proteins in the central nervous system (CNS)
Light microscopic immunohistochemistry of brain areas from laboratory rats, laboratory guinea pigs and domestic cats have been performed with help of the polyclonal v-Raf antibodies , polyclonal B-raf antibodies (Santa Cruz Biotechnology, Santa Cruz, CA, USA), monoclonal B-Raf antibodies (produced against the last 12 amino acids of the C-terminal region of B-Raf) and polyclonal C-Raf antibodies (anti-SP 63 serum). The polyclonal anti-v-Raf-, anti-SP 63-and monoclonal anti-B-Raf antibodies were prepared in the laboratories of Dr. Ulf Rapp (National Cancer Institute, Frederick Cancer Research and Development Center, Frederick, MD, USA; see descriptions of the antibodies in Mihály et al., 1993;1996). Two human hippocampi were investigated, too. The human brain samples were obtained from autopsy in the Department of Pathology, Szeged University. The autopsy was performed 16 h after death. The hippocampal tissue was kept in fixative solution (4% paraformaldehyde in 0.1 M phosphate buffer) for 12 days, then transverse plane frozen sections were made and stained with the monoclonal B-Raf antibody. Rats, guinea pigs and cats were deeply anesthetized and perfused through the heart with fixative solution (4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4). The brains were sectioned with freezing microtome, cryostat and vibratome. The tissue sections were immunostained as glass-mounted-or free-floating sections. Immunofluorescence was performed on cryostat sections, mounted on glass slides. The slides were incubated with primary antibodies (anti-Raf serum), then with secondary antibodies conjugated with fluorescein isothiocyanate (FITC). In peroxidase-based techniques, free-floating sections were used, incubated with the primary antibody, then treated according to the avidin-biotin method, using the Vectastain ABC kits (Vector, Burlingame, CA, USA). Independent from the antibody and method used, the staining pattern of the different brain areas was consistent, showing neuronal and glial localization of Raf kinases. The differences between B-Raf and C-Raf localizations are indicated in the text. No differences were detected in the staining pattern comparing the different species (rat, cat, guinea pig, human). The descriptions below refer to the results obtained with polyclonal v-Raf antibodies, if not indicated otherwise. Staining obtained with v-Raf antibodies will be referred to as B-Raf-like staining, staining with anti-SP 63 will be described as C-Raf-like immunoreactivity. Controls of the immunohistochemical procedure included absorption controls performed with the recombinant v-Raf protein and the SP 63 peptide. One mg of the protein and 10 mg of the peptide were reconstituted in 1 ml of deionized water, mixed with the undiluted antibody (0.01 ml), and kept at room temperature overnight with slow agitation. The mixture was diluted to working concentration (1:500), centrifuged at 70,000xg for 20 min, and used for the immunohistochemical procedure (Mihály et al., 1993). No immunostaining was seen after the procedure (Mihály et al., 1993). Other controls included incubations without the primary anti-Raf sera: in these experiments the primary antibody was omitted, and the sections were incubated in normal horse serum, diluted to 1:100, and secondary antibodies as usual. In these sections, no specific immunostaining was detected (Fig. 6e).

Localization of B-Raf in the spinal cord and brain stem
Neuronal and glial localization of Raf-protein-like immunoreactivity (RPI) was encountered in the spinal cord of the laboratory rat and guinea pig (Mihály & Rapp, 1993;. Neurons in laminae I, V, VI and IX displayed strong cytoplasmic staining. Glial cells in the white matter and grey substance were stained, too (Fig. 7). Axonal staining (e.g.: in the dorsal horn, where afferent axons terminate in large numbers) was not observed (Mihály et al., 1996). Different levels of the brain stem displayed similar staining pattern: cytoplasmic staining in large-and medium-sized nerve cells. The motor neurons of the cranial nerve nuclei were strongly stained. Scattered RPI was detected in sensory nuclei, such as the superior and medial vestibular nuclei and the ventral cochlear nucleus. The raphe nuclei of the pons and medulla also contained RPI (Mihály & Endrész, 2000;Mihály et al., 2007). Large neurons of the reticular formation and the inferior olive were containing strong RPI (Fig. 3). The neurons of the mesencephalic trigeminal nucleus were strongly stained (Mihály et al., 2007). On the other hand, the spinal trigeminal nucleus displayed only faint staining (except lamina I neurons, which were strongly stained; Mihály et al., 2007). The red nucleus displayed strong staining, whilst the substantia nigra cells contained less RPI (Fig. 3). The immunostaining was always localized in cell bodies, no axonal staining was observed in any of the brain stem areas. Detailed examination of the brain stem of the rat and the guinea pig did not reveal any species differences (Mihály & Endrész, 2000;Mihály et al., 2007).

Localization of B-Raf and C-Raf proteins in the cerebellum
The C-Raf and B-Raf proteins were detected in the vermis of the guinea pig cerebellum, using polyclonal anti-SP 63 (anti-C-Raf), and polyclonal anti-v-Raf (Mihály et al., 1993) antibodies. The B-Raf-like staining was very strong in neuronal cell bodies and glia-like cells. The staining was localized in neuronal cell bodies and dendrites. Dendritic staining was conspicuous in Purkinje cells, mainly in primary and secondary, large dendrites. Cell bodies of Purkinje cells were stained, too. Strong RPI was seen in the granular layer, where cell bodies of the granule cells and Golgi cells were immunoreactive. Small groups of glia cells stained in the white matter, and strong RPI was detected in the neurons of the fastigial nucleus (Fig. 4). Fig. 3. Representative cross sections of the rat brain stem (Swanson, 1992), showing the cellular distrubution of Raf-protein-like immunoreactivity. Large dots represent strongly stained neurons, small dots represent medium or faint staining. Triangles represent large nerve cells with outstanding RPI. The upper mesencephalon (a), the upper pons (b), the lower pons (c) and the lower medulla (d) are represented (Mihály et al., 2007) The C-Raf staining was pale, displaying Purkinje cells and their large (primary and secondary) dendrites. Other neuronal elements (e.g.: granule cells, basket cells, Golgi cells) did not stain. C-Raf-like staining was seen in the nerve cells of the fastigial nucleus, and faint labeling was detected in some glia-like cells in white matter (Fig. 5).

Localization of B-Raf and C-Raf proteins in the cerebral cortex
RPI has been detected with polyclonal anti-v-Raf antibody and polyclonal anti-SP 63 antiserum. The latter is specific for C-Raf (Schultz et al., 1985;Morice et al., 1999). The RPI was detected in the neocortex (Fig. 6), cingular, pyriform, perirhinal, entorhinal areas and in the hippocampus (Mihály et al., 1993). RPI was localized in neurons and glia-like cells. No attempt was made to identify the cell types with double immunostainings. However, on the basis of the shape of the RPI-containing cells, we state that pyramidal and non-pyramidal cells of the Ammon's horn, granule cells of the dentate fascia and multipolar cells of the hilum of the dentate fascia were strongly stained (Mihály et al., 1993). Similar localizations were detected with the two antibodies, although C-Raf-like staining was faint compared to the B-Raf-like immunoreactivity. The intensity differences of the staining obtained with serial dilutions of the antibodies indicated that the cytoplasmic concentration of B-Raf kinase was much higher than that of the C-Raf in every region of the cerebral cortex. Detailed analysis of C-Raf-like staining and localization was not performed (Mihály et al., 1993). The brains of three domestic cats were also studied, using vibratome sections of the motor neocortex (sigmoid gyrus). Similarly to rats and guinea pigs, RPI has been detected in neuronal and glial cells bodies. Layer V pyramidal cells were stained in their cell bodies and proximal, large dendrites (Fig. 6). Scattered, small, glia-like cells were labeled, too. The human hippocampus displayed faint immunostaining with the B-Raf antiserum. Neuronal cell bodies were stained, which were similar to pyramidal cells, located in the stratum pyramidale of CA1 (Fig. 6).

Localization of Raf proteins in glia cells
B-Raf immunostaining was observed in small, glia-like cells, measuring 8-12 μm, in the cerebral cortex (gray and white matter), cerebellum (mentioned above), brain stem and spinal cord. Glial cells in the grey substance and in the whitte matter were regularly seen ( Fig. 4; Fig. 7). Marginal astrocytes contained RPI in the spinal cord. Radial glia processes were stained in the cerebellar cortex (Fig. 4). No double immunolabelling was attempted in order to identify the cells. RPI was localized in the cytoplasm of the cell body and the processes (Fig. 7). No systematic study was made for the exploration of RPI in other glial cell types (oligodendroglia, microglia).

Electron microscopic localization of RPI in neurons
Ultrastructural immunohistochemistry was performed on vibratome sections of rat and cat cerebral cortex, using peroxidase-based preembedding methods Mihály & Rapp, 1994). These reports were the first in the literature, describing the electron microscopic localization of Raf kinases in the brain Mihály & Rapp, 1994). Comparison of the Raf kinase localization in a rodent and a carnivore, did not reveal differences: the localization pattern of RPI was consistent and similar in the two species.  Fig. 4a: Immunostaining depicts the layers of the cerebellar cortex: (1) white matter; (2) granular layer; (3) molecular layer; arrowheads: Purkinje cell layer; bar: 0.5 mm. Fig. 4b: the Purkinje cell RPI was detected not only in cell bodies, but also in dendrites (Fig. 4g). Arrowheads: radial glia processes; arrows: small neurons in the molecular layer (mol). Bar: 50 μm in Fig. 4b; 5 μm in Fig. 4g. The RPI in granule cells (Fig. 4c), Golgi cells (Fig. 4d), Purkinje cells (Fig. 4e) and outer stellate cells (Fig.  4f) are pointed by arrows. Bar as in Fig. 4g. Fig. 4h: the B-Raf staining of the neurons of the fastigial nucleus (arrow). Bar: 10 μm. Fig. 4i: white matter RPI apparently localized in glia cells (arrow). Arrowheads point to white matter axons containing RPI. Bar: 50 μm.  Fig. 5a shows the overall weak staining (arrowheads: Purkinje cell layer; 1,2,3: as in Fig. 4a; bar: 0.5 mm). Fig. 5b: Purkinje cells contain C-Raf, which is localized in cell bodies and dendrites (arrowheads). Arrow points to outer stellate cell, which displays weak staining. Bar: 50 μm. Fig. 5c: Purkinje cell (P) staining with high magnification. Arrows point to dendrites (n: unstained granule cells; bar: 50 μm). Fig. 5d: neurons of the fastigial nucleus (arrows) containing C-Raf. Fig. 5e: white matter containing weakly stained glia cell (arrow) and weak axonal staining (arrowheads). Bars: 50 μm. Neurons contain strong Raf kinase immunoreactivity in the cytoplasm of the cell body and dendrites (Figs 8-11). Not every neuron did stain: in the granule cell layer of the dentate fascia, cca. 20% of the granule cells remained unstained. This feature was visible already under the light microsocope, and was applicable to other cortical areas, as well (Mihály et al., 1993). As to the staining of the dendrites, very strong RPI was seen in dendritic spines (Figs 8-9; Fig. 11). The Raf-kinase-like staining was strong in postsynaptic densities, and in spine apparatuses ( Fig. 9; Fig. 11; . Not only the spine apparatuses, but also dendritic subsynaptic cisternae contained RPI (Fig. 10d). In these membrane cisternal organelles the immunprecipitate was localized between the membrane cisternae; the cavity of the cisternae did not contain RPI (Fig. 11). We did not observe immunostained axon terminals in the tissues studied so far. None of the synapsing axons terminating on dendrites, dendritic spines or neuron somas did contain Raf-kinase-like immunostaining (see . There are no comprehensive electron microscopic data about the ultrastructural localization of C-Raf, so this issue requires and deserves further, extensive experimentation.  Fig. 7a displays RPI in astroglia-like cells of the white matter (bar: 10 μm). Fig. 7b shows the ultrastructural localization of B-Raf immunoreactivity in the cytoplasm of astrocyte-like cell of the rat hippocampus (solid arrows point to the contours of the astrocyte; empty arrow point to a synapse in the vicinity of the astrocyte; bar: 5 μm).

Electron microscopic localization of RPI in astrocytes
We observed several immunostained astrocytes in the cerebral cortex of rats (Fig. 7) and cats (Mihály & Rapp, 1994). The cytoplasmic staining of the cell body was conspicuous: very often in the vicinity of bundles of intermediate GFAP filaments (Mihály & Rapp, 1994). RPI was also observed in perivascular glial processes, and glial processes in the neuropil, around synapses (Mihály & Rapp, 1994). The RPI containing astrocytes were detected in the white matter, too -similarly to light microscopy observations. No RPI containing oligodendroglia cells were observed -in some sections we have seen RPI containing astrocytes together with unstained oligodendroglial cells (Mihály & Rapp, 1994). However, no systematic studies were performed to prove or disprove the lack of Raf proteins in oligodendroglia cells. The endothelial cells and pericytes were not labelled with any of the antibodies.

Comparison of B-Raf and C-Raf brain immunolocalization patterns
The localization of the two isozymes did not show substantial differences. The main observation in this respect was, that C-Raf kinase was expressed only in a few neurons and glia cells -compared to the expression of B-Raf. The staining was also weaker in most of the experiments. The C-Raf staining was specific, because the absorption controls did not show immunostaining (Mihály et al., 1993). The C-Raf was detected in the cerebellum and the cerebral cortex: these structures are compared in Table 1. B-Raf antibodies stained several neurons, irrespective of the size and type. However, staining was stronger in large cells (Figs 2-4). The B-Raf was also localized in astrocytes, as strong cytoplasmic staining. Under the electron microscope, the neuronal immunostaining was cytoplasmic, and exceptionally strong in dendritic spines, spine apparatuses and postsynaptic densities. Axon terminals did not stain. The C-Raf was localized in large neurons and their proximal dendritic branches. A few glia-like cells also displayed faint C-Raf-like staining (Fig. 5). We did not perform systematic electron microscopic studies for the localization of C-Raf kinase: only one series of experiments was done, in which we investigated the ultrastructural localization of C-Raf www.intechopen.com Protein Kinases 434 in guinea pig hippocampal slices . In this electron microscopic study we observed neuronal and dendritic localization of C-Raf, thus corroborating the light microscopy findings .

Neuronal B-Raf kinases
The immunolocalization of Raf kinases shows consistent patterns in different mammalian species (rat, guinea pig, cat and human). The main Raf molecule of the brain is certainly the B-Raf isozyme. The 95 kDa protein was detected in rat and guinea pig 1996;2007;Mihály & Endrész, 2000). This observation conforms the literature data which give the same molecular weight in other experiments (Stephens et al., 1992;Dwivedi et al., 2006). Based on the specificity of the polyclonal v-Raf serum we refer to the immunostaining detected as to the B-Raf kinase immunostaining. This statement is strongly supported by the repeated Western blot experiments, which gave very consistent results: the polyclonal v-Raf serum and the B-Raf antibodies resulted in the same 95 kDa band in brain homogenates 2007;Mihály & Endrész, 2000). We did not perform immunoblotting with the C-Raf antibody because the polyclonal anti-SP 63 serum was thoroughly investigated (Schultz et al., 1985;Morice et al., 1999). The synthetic SP 63 peptide is the C-terminal part of the C-Raf protein (Schultz et al., 1985;. This 12 amino acids peptide was used for immunization in rabbits (Schultz et al., 1985;. The polyclonal serum was affinity purified, tested with immunoprecipitation and was found to be specific for the murine and human C-Raf terminal sequences in cell cultures and in tissue sections (Schultz et al., 1985;Mihály et al., 1993). This immunostaining is therefore referred to as the immunolocalization of C-Raf kinase. Neuronal B-Raf and C-Raf localization at the light microscopic level was found to be very similar in other experiments, years later (Morice et al., 1999). We think therefore, that our descriptions concerning the brain localization of Raf kinases are the first in the literature, and give a precise picture of neuronal B-Raf kinases 1993).
The main features of neuronal B-Raf kinase localization are as follows: 1. Cytoplasmic localization in the cell body was a general feature. This reflects the existence of a standard pool of the B-Raf kinase in the neuron. In case of growth factor signals, this cytoplasmic kinase pool can be mobilized quickly, in form of recruitment to the cell membranes or to the nucleus, if necessary. 2. Localization in dendrites was another characteristic feature of neuronal B-Raf. Dendritic staining was extremely strong in dendritic spines and in postsynaptic densities, meaning that a certain amount of B-Raf was already attached to the membrane, in proximity of the synaptic receptors. This localization suggested a very effective signal transduction process, mediated by the B-Raf.

Neuronal signal transduction with B-Raf: experimental proofs
The first observations pointing to the possible significance of Raf kinases in learning and memory were those of Mihály et al., (1990). In these experiments long term potentiation (LTP) was induced in rats, then the animals were subjected to immunohistochemical detection of Raf kinase with the polyclonal v-Raf antibodies (Mihály et al., 1990). The RPI was investigated in the granule cell layer of the dentate gyrus, by counting the immunoreactive neuronal cell bodies (Mihály et al., 1990). The number of RPI-containing cell bodies increased significantly in successfully potentiated animals (Mihály et al., 1990). This observation has been discussed either as the post-LTP increase of Raf kinase expression in dentate granule cells; or as the translocation of the Raf kinase from the distal dendrites to the cell body, following LTP (Mihály et al., 1990). We were not able to decide between these two possibilities in these in vivo experiments (Mihály et al., 1990). Years later, with help of in situ hybridization, it turned out that 24 h following LTP, the B-Raf expression in the dentate gyrus increased significantly (Thomas et al., 1994). Recently, experiments on B-Raf knockout mice proved, that the B-Raf is absolutely necessary for learning and memory consolidation (Chen et al., 2006;Valluet et al., 2010). However, the detailed mechanisms of the participation of B-Raf in LTP are unknown. www.intechopen.com

The possible intracellular translocation of B-Raf in neurons
Since the early immunohistochemical experiments on LTP (Mihály et al., 1990), the possibility of intracellular Raf translocation was an important issue, because the translocation could be one intracellular mechanism for signal transduction. The translocation was observed in vitro (Leevers et al., 1994) and in vivo  in different experimental conditions. The recruitment to the cell membrane is considered to be a necessary step of Raf activation (Matallanas et al., 2011). However, the translocation to the cell nucleus is difficult to explain, although the experimental facts seem to be firm . Another, translocation-like phenomenon was observed in sensory ganglion cells (Mihály et al., 1996). Primary sensory neurons contain B-Raf in their cell body, in form of homogeneous immunoreactivity (Mihály et al., 1996). Two-eight days after the transection of the peripheral nerve, this cytoplasmic staining pattern displays translocation towards the periphery, and at 8 days, most of the RPI was localized beneath the cell membrane of the sensory ganglion cells (Mihály et al., 1996). The significance of this phenomenon is not known, possibly it is connected to the process of chromatolysis (Mihály et al., 1996). The phenomenon of the translocation has been proved in cell cultures (see Mor & Philips, 2006), but not in vivo, in plasticity or pathological conditions. Further experiments are needed for the understanding of the possible in vivo translocation processes.

Possible functions of C-Raf in the neuron
Although C-Raf was found to be ubiquitous (Storm et al., 1990;Matallanas et al., 2011), not every neuron displayed immunoreactivity in our studies. These findings are in conform with other immunohistochemical studies (Morice et al., 1999). On the other hand, it seems, that C-Raf expression can be induced by some (yet unknown) conditions. Hippocampal slices maintained in vitro do not express C-Raf at the beginning of the in vitro incubation. Incubation of the slice for 2-4 h in oxygenized environment induced the appearance of the C-Raf staining in the nerve cells . Upregulation of C-Raf was observed in Alzheimer-brains, too (Mei et al., 2006). In vitro grown neurons were protected by C-Raf inhibitors against the toxicity of amyloid peptides (Echeverria et al., 2008). It seems therefore, that the upregulation of C-Raf in the cell causes cell damage and death: it can happen through the activation of apoptosis signals (Echeverria et al., 2008). No data exist about the causes of C-Raf upregulation: excess excitatory amino acids (excitotoxic effects), hypoxic conditions, and ageing may operate through unknown signaling pathways in brain cells.

Conclusions
Raf protein kinases are widespread in the mammalian brain and spinal cord. Raf protein kinases are activated by several cytokines and growth factors through their membrane receptors. We investigated the immunohistochemical localization of B-Raf kinase and C-Raf kinase in neurons and glial cells. The B-Raf kinase is localized in the cell body, in dendrites, dendritic spines, spine apparatuses, subsynaptic cisternae and postsynaptic densities. The localization does not depend on neuronal type, but the intensity of the immunostaining is greater in large neurons. Neuronal B-Raf was not found in presynaptic axons, suggesting, that in these cells B-Raf kinase is coupled to postsynaptic signaling pathways. The similarities of the presence and localization of B-Raf in three mammalian orders (rodents, carnivores and hominids) suggest that this protein kinase has an important, and phylogenetically conserved function in normal, adult neurons. Data are available that learning and memory consolidation are processes in which B-Raf participates importantly. The C-Raf kinase has a similar CNS localization, but the amount of C-Raf in the cells is generally lower, than that of B-Raf. On the other hand, results from different experiments suggest, that the neuron adapts and alters the C-Raf expression, and thus regulates the amount of C-Raf present in the cytoplasm and in the membrane compartments. Therefore, we can state, that the regulation of the two isozymes (B-Raf and C-Raf) are different (although the downstream cellular events are similar):