Neuron,

Vol. 6, 397-409, March, 1991, Copyright

0 1991 by Cell Press

Expression of Acidic Fibroblast G rowth Factor mRNA in the Developing and Adult Rat Brain Barbara J. Wilcox* and James R. Unnerstall*+*§ *Department of Neurology and the Alzheimer Center +Department of Pharmacology *Department of Neuroscience Case Western Reserve University Cleveland, Ohio 44106

Summary We have localized acidic fibroblast growth factor (aFCF) mRNA in the developing and adult rat brain using in situ hybridization histochemistry. Prenatally, hybridization to aFCF mRNA was observed throughout the brain, with the strongest signal associated with cells of the developing cortical plate. Postnatally, labeling was localized to specific neuronal populations. In the hippocampus, labeling of the pyramidal cell layer and dentate granule cells was observed and became progressively more intense with maturation. labeling was also observed in both the external and internal granule cell layers of the developing cerebellum. Pyramidal cells of the neocortex as well as neurons of the substantia nigra and locus ceruleus also express aFGF. This pattern persists into adulthood, although the intensity of the labeling is significantly reduced in the adult brain. These patterns of hybridization correlate with specific developmental events and suggest that aFCF plays a significant role in both central nervous system development and neuronal viability in the adult brain. introduction Fibroblast growth factor (FGF) is found in the brain in two closely related forms: acidic fibroblast growth factor (aFGF) and basic fibroblast growth factor (bFGF) (Gospodarowicz et al., 1986). These two forms have 55% amino acid sequence homology and have been shown to exert similar biological actions in culture. bFGF has been extracted from numerous tissues including brain, retina, pituitary gland, kidney, adrenal gland, and thymus. aFGF, however, has a more restricted distribution.Thisform of FGF isfound primarily in brain and retina, although recent reports have demonstrated low levels of aFGF mRNA in other tissues (Gospodarowicz et al., 1986; Winkles et al., 1987; Weiner and Swain, 1989). The apparent preferential localization to the brain suggests that aFGF may have a more selective role for neural tissue than does bFGF. Both FGFs exhibit potent neurotrophic activity. In culture, they induce proliferation and differentiation of embryonic cortical neurons as well as enhance sur3 Present address: Department (M/C 512), University of Illinois Chicago, Illinois 60680.

of Anatomy and Cell Biology at Chicago College of Medicine,

viva1 of hippocampal, cortical, and spinal cord neurons (Walicke, 1988; Morrison et al., 1986; Walicke et al., 1986; Gensburger et al., 1987; llnsicker et al., 1987; Whittemore et al., 1989, Sot. Neurosci., abstract). Furthermore, both FGFs induce neurite outgrowth in PC12 cells (Rydel and Greene, 1987; Schubert et al., 1987). These data suggest that FGF could act as a neurotrophic factor in mammalian brain and have a significant role in central nervous system development and maintenance. Low levels of aFGF mRNA have been extracted from brain tissue and identified by Northern blot analysis as a band of approximately 4.8 kb (Jaye et al., 1986; Alterio et al., 1988; Logan, 1988). An important question in understanding the function of this neurotrophic factor in the brain is to determine the regional and cellular sites of synthesis. To date, this information regarding the source of FCF in the central nervous system has not been clearly determined. Glial cells in culture are capable of synthesizing FGF (Hatten et al., 1988). However, this has not been determined in vivo. lmmunohistochemical studies have localized FGF-like immunoreactivity to neurons in the hippocampus (Pettmann et al., 1986). However, the characterization of the antiserum used for those studies was not complete, and it was not clear which form of FGF was recognized by the antiserum. In the present study, we have addressed the question of FGF localization in the rat central nervous system using in situ hybridization histochemistry, which provides the means for selective identification of specific mRNA species in individual cells. Because of the reported preferential localization of aFGF in neural tissue, we have chosen to focus our present studies on the pattern and developmental regulation of aFGF mRNA expression in the rat brain. Our results reveal a predominant localization of aFGF mRNA to neurons in many brain regions and suggest that the expression of this putative neurotrophic factor may have a significant role in brain development ancl neuronal maturation. Results Probe Characterization At the initiation of these studies, only the bovine nucleotide sequence for aFGF mRNA was available. Thus, much of ourwork has been done using a 36-mer oligonucleotide probe complementary to aFGF mRNA. To assure selectivity of the oligonucleotide probe that was used for in situ hybridization histochemistry, we have performed a series of specific control experiments. First, serial sections were hybridized with labeled sense-sequence probe that is exactly complimentary to the antisense probe. This revealed extremely low background labeling and no visible distinct localization (for example, see Figure 38 and Figure 48). In addition, further characterization of

Neuron 398

pattern observed is identical to that observed after hybridization with the original bovine sequence probe (Figure 2). Developmental

20s

18s

Figurel. Probe

Northern

Blot Analysis

Using a 36mer

Oligonucleotide

Forty micrograms of total RNA extracted from P9 rat brain was electrophoresed on an agarose gel and transferred to nitrocellulose by standard methods. The filter was hybridized with the 32P-labeled oligonucleotide probe by standard methods and washed at a final temperature of 47OC in 0.1 x SSPE containing 0.1% SDS. Under these conditions, a single band was labeled at approximately 4.6 kb, which is in agreement with the reported size for aFGF mRNA in brain.

probe specificity was carried out using Northern blot analysis. Autoradiograms revealed a single band labeled by our oligonucleotide probe corresponding to an mRNA size of approximately 4.8 kb (Figure 1). This size is consistent with results reported by others for aFGF mRNA in brain (Jaye et al., 1986; Alterio et al., 1988; Logan, 1988). Furthermore, hybridization of similar blots with the human cDNA probe resulted in labeling of an mRNA band of a corresponding size (data not shown). We have also conducted systematic empirical studies to determine the T, of our probe both in situ and by Northern analysis. The experimentally measured T, for the aFGF oligonucleotide probe (50°C) is in agreement with the calculated T, for the hybridization conditions used in these assays. These data indicate that the probe used for in situ hybridization studies was selective for a single mRNA species of the appropriate size for aFGF message. In addition, when the nucleotide sequence became available for rat aFGF, oligonucleotide probes were synthesized complementary to a sequence near the 3 end of the message. This is the opposite end of the molecule from the sequence used for the bovine probe and provides an additional control for probe specificity. Using these probes, we have confirmed that the

Expression

of aFGF mRNA

Nine time points in brain development from embryonic day 18 (E18) to postnatal day21 (P21) were selected for examination. In general, the pattern of specific hybridization evolved from a generalized distribution at early time points to more selective association with specific neuronal populations at later stages of development. Cortex Prenatally (El8 and E20), specific hybridization was observed throughout the developing brain (Figure 3A). The most intense labeling was associated with cortical plate cells and ventricular zone cells of the developing cerebral cortex. By PI, the levels of specific hybridization were reduced, although the more intense labeling of superficial laminaeof the maturing cortex was observed (Figure 3B). This pattern was more distinct by day 4. By P9, specific hybridization was observed as interpreted from film autoradiograms to be predominantly in lamina III and, to a lesser extent, lamina V of the cerebral cortex (Figure 3D; Figure 7). At P9, the degree of hybridization was generally higher in the frontal cortex as opposed to parietal or occipital areas (Figure 7). This pattern persisted into adulthood, although the amount of hybridization was significantly reduced (Figure 4A). The cellular localization of aFGF mRNA in the cerebral cortex was confirmed by examination at high magnification (Figure 5). Grains were observed preferentially localized to pyramidal neurons in laminae III and V, in confirmation of the laminar distribution observed in the film autoradiograms. Hippocampal

Formation

A similar progression in aFCF mRNA expression was observed in the developing hippocampus. Prenatally, hybridization associated with the developing hippocampal formation was low and uniform (Figure 3). By E20, higher levels of hybridization were seen associated with maturing CA1 and CA3 pyramidal neurons (Figure 3B). By PI, pyramidal neurons were clearly labeled, as were cells within the developing hilar region. At P4, the specific labeling in the hippocampus was clearly localized to the pyramidal cell layer in all subregions of the hippocampus (Figure 3C). In addition, distinct labeling was also observed in the developing dentate gyrus associated with the granule cell layer. Bythisage, labeling showed distinct localization to the developing lateral blade. High levels were also seen in the subiculum at this developmental time point. Between P6 and P12, distinct labeling in the granule cell layer of the medial blade of the dentate gyrus became visible (Figures 3B-3E). By P15, the entire extent of the pyramidal layer of the hippocampus was discretely labeled as well as the entire extent of the granule cell layer of the dentate gyrus. By P21, the specific labelinginthegranulecell layerofthedentate

Expression of Acidic Fibroblast 399

Figure 2. Comparison

Growth

of Labeling

Facto1

Patterns

between

the Bovine

and Rat Oligonucleotide

Probes

(A) Bovine. (B) Rat. This photo is a film image of hybridization of each probe to serial sagittal sections of rat brain at P6. The distribution of hybridization appears to be identical in both images with relatively high density found in the hippocampus (h) and cerebellum (cb). Differences in signal intensity can be accounted for by variations in specific activity of the probe. Exposure time, 5 weeks. Bar, 1 mm.

gyrus and in the pyramidal cell layer of the hippocampus was more intense and distinct (Figure 3F). Conversely, hybridization seen in the subiculum early in development decreased to near background by PI2 (Figure 3E). As described for the cerebral cortex, this pattern of labeling is also observed in the adult rat brain (Figure 4A). Examination of the autoradiograms produced by dipping labeled tissue sections in liquid emulsion confirmed that hybridization was associated preferentially with neurons (Figure 6A). Grain :ounts from these regions revealed a41 ratio of grains

over neurons versus those over neuropil. When compared with serial sections labeled with sense probe (Figure 6B), a low level of specific “background” hybridization was revealed. Thus, the possibility that aFGF mRNA may also be expressed at low levels in a subset of glial cells under certain conditions cannot be ruled out.

Other Areas Specific labeling appeared in the developing cerebellum in the few days after birth and was clearly localized to the external granule cell layer by P4. By P9,

Neuron 400

Figure 3. Developmental

Expression

of aFGF mRNA

in the Rat Hippocampus

Film images of3Wabeled oligonucleotide probe hybridized to parasagittal sections of developing rat brain taken through the hippocampus. Exposure time, 10 days. (A) At early time points, labeling is found throughout the brain with enrichments in the cortical plate (cp) and ventricular zone (v) of the developing cerebral cortex and the olfactory bulb (ob). (B) By PI, labeling of the hippocampal pyramidal cells of CA3 can be seen (h). (C) By P4, labeling in the developing cerebral cortex is reduced, but that of the hippocampal pyramidal cell layer (CAI, CA3) is increased, and the lateral blade of the granule cell layer is visible (dg). Note that CA1 and CA3 subregions and the subiculum (s) are clearly defined. (D) By P9, both blades of the dentate gyrus granule cell layer are visible; labeling of the CA1 and CA3 pyramidal regions is more defined, but the labeling of the subiculum is reduced. (E) By P12, labeling of the subiculum is reduced to near background levels, but discrete labeling of the pyramidal cell layer of the hippocampus and the granule cell layer of the dentate gyrus remains. (F) By P21, the adult pattern of hybridization in the hippocampal formation is observed. Bar, 1000 Km.

/ \

Figure 4. In Situ Hybridization

Labeling

of aFGF mRNA

in the Adult

Rat Brain

Coronal sections of adult male rat brain were processed as described with the antisense (A) and sense (B) oligonucleotide probes for aFCF. (A) Film image demonstrates the discrete pattern of labeling observed in the adult hippocampus (exposure time, IO weeks; longer exposure was necessary for the purpose of visualization and is not intended to be compared quantitatively with those in Figure 2). In the mature brain, distinct labeling is localized to the pyramidal cell layer of all subregions of the hippocampus (CAI, CA3) as well as the granule cell layer of the dentate gyrus (dg). In addition, labeling is also observed associated with the substantia nigra pars compacta (snc) and the cerebral cortex. (6) Film image of a serial section to that in (A) hybridized with the sense-sequence probe. Only low background labeling is observed. Bar, 1500 urn.

Nt?UVXl 402

Figure 5. Photomicrograph Demonstrating Localization of aFCF mRNAto Neurons in the Cerebral Cortex Photographic grains are shown over pyramidal neurons in lamina V of sensory motor cortex of the rat brain $t P21 after hybridization with the antisense probe (A). However, not all neurons appear to be labeled above background (asterisk). Background levels are demonstrated by hybridization with the sense-sequence probe to a serial section (B). Bar, 25 pm.

labeling associated with both external and internal granule cell layers was clearly distinguished (Figure 7; Figure 8A). By P21, the mature granule cell layer was distinctly labeled (Figure 8B). High magnification examination confirmed this and also revealed that Purkinje cells were not labeled (Figure 9). Significant levelsof labelingwerealsoobserved in theolfactory bulb throughout development. By P9, selective labeling of external and internal granule cell layers as well as the mitral cell layer was observed (Figure 7) In the P21 and adult brain, labeling was also detected over neurons in the substantia nigra pars compacta (Figure 10) and the locus ceruleus (Figure 11). Other brain regions that exhibited significant levels of specific hybridization during development into

adulthood include the nucleus accumbens, nucleus, and ventromedial hypothalamus Figure 4; Figure 7).

caudate (Figure 2;

Discussion These data demonstrate that aFGF mRNA is found in the developing rat brain as early as El8 and its expression continues into adulthood. Furthermore, these results clearly demonstrate that, although low levels may be expressed in some glial populations, aFGF mRNA is predominantly expressed in specific neuronal populations in the mammalian CNS. In development, the expression of aFGF mRNA in

Expression 403

of Acidic Fibroblast

Growth

Factor

Figure 6. High Resolution graph of In Situ Hybridization pocampuswith Antisenseand nucleotide Probes

Photomicro in Adult Hip SenseOhgo

(A) Antisense probe. (B) Sense probe. Labeled sections are dipped in liquid emulsion as described and counterstained with cresyl videtafter6week exposure. (A) Photographic grains are found concentrated over pyramidal neurons of CA3 of the adult rat hippocampus. (6) Very few grains are observed when hybridized with the labeled sense-sequence probe. Bar, 25 Wm.

the brain correlates with both the differentiation and migration of neurons to their adult positions as well as the establishment of synaptic contacts. Analysis of the probe hybridization to aFGF mRNA in the developing hippocampal formation provides a good example of this developmental correspondence. Discrete labeling of aFGF mRNA associated with the hippocampal pyramidal cell layer was visible during prenatal development and became progressively more distinct and intense with development. Studies using [3H]thymidine incorporation have demonstrated that, in rodent brain, these pyramidal neurons are born prenatally and migrate into place early in development with very few neurons differentiating after birth (Angevine, 1965). Granule neurons in thedentategyrus, however, undergo final mitosis and migration almost entirely after birth, with the highest rate of division occurring

during the second postnatal week (Schlessinger et al., 1975). It has been shown clearly that the neuroblasts from which the granule cells differentiate are located in a proliferative zone in the hippocampal hilar area and migrate to their adult positions with the lateral blade of the dentate gyrus forming first followed by the medial blade. As demonstrated by the present study, this temporal developmental pattern is also reflected in the expression of aFGF mRNA in these neurons. Although the autoradiograms show enrichments of signal in areas of proliferating neurons, such as the ventricular zone of the developing cortex and the proliferative zone of the dentate gyrus, the strongest aFGF mRNA hybridization signal is observed after the neurons have migrated into their adult position. This may simply reflect more dense packing of neurons in these cellular layers. However, when the auto-

Figure 7. In Situ Hybridization

of Oligonucleotide

Probe to aFGF mRNA

in P9 Rat Brain

Parasagittal section hybridized with the %-labeled probe as descibed. High levels of hybridization are observed in the olfactory bulb (ob), hippocampal formation (h), cerebellum (cb), nucleus accumbens (acb), as well as the hypothalamus (particularly the ventromedial nucleus, vmh). Exposure time was 8 weeks. Thus, as pointed out in Figure 4, long exposure times were necessary for demonstration of low-level hybridization and are not intended to be quantitatively compared with the images in Figure 2. Bar, 1600 urn.

radiograms were examined with high resolution microscopy, there appeared to be a higher density of autoradiographic grains over the outer granule neurons of the dentate gyrus than the inner neurons. Because this layer matures in an “outside-in” pattern, the outer granule cells are older than the inner cells, suggesting that there may indeed be an increase in aFGF mRNA expression after the neurons have reached their adult position. Further quantitative evaluation will be necessary to confirm this. We have consistently seen our strongest, most discretely localized hybridization to aFGF mRNA expression in the hippocampus emerge at approximately P9 (Figure 2 and Figure 5). This is a significant time point in development. At this time, the projections of the perforant path from the entorhinal cortex first arrive in the dentate gyrus (Singh, 1977). Importantly, it is also at this time that we have observed strong labeling both in the dentate gyrus and entorhinal cortex. Furthermore, at P9, the mossy fiber projections from the dentate gyrus to the hippocampal subfields CA3 and CA4 begin to make their connections, and synaptic spines begin to appear on dendrites of hippocampal pyramidal cells. Again, this correlates with increased signal intensity seen in both the dentate granule cell layer and the CA3 pyramidal cell layer, which is observed beginning at this time point. Correlation between these developmental events and the expression of aFGF mRNA suggest that this neurotrophic factor could have an important function in neurite

outgrowth and synaptogenesis for these populations of neurons. It is important to note that these same neuronal populations express aFGF mRNA in the adult brain. While the function of aFGF in the adult brain is unknown, the potential role for aFGF in neurite outgrowth and synaptogenesis that we have hypothesized for normal brain development suggests that aFGF could function in regulating synaptic remodeling and plasticity in the adult brain. Furthermore, a trophic role for aFGF in maintaining neuronal viability and connectivitycannot beoverlooked.This hypothesis is supported by studies using adult rats that have shown increased levels of aFGF mRNA in brain after injury (Logan, 1988). Moreover, FGF can promote the survival of basal forebrain cholinergic neurons after fimbria-fornix transections in adult rat brain (Anderson et al., 1988). In culture, FGFs promote the survival of primary hippocampal neurons challenged with excitotoxins (Mattson et al., 1989). While many of these studies used bFCF, both forms of FGF can act at the same receptors (Neufeld and Gospodarowicz, 1985, 1986). Since the relative potencies of bFGF or aFGF in producing these effects were not determined, it is possible that aFGF expressed in specific neuronal populations could be the primary effector molecule in mediating similar effects in vivo. In some areas of the brain such as the locus ceruleus, we noted that the photographic grains were distributed near the nucleus rather than over the cyto-

Expression 405

of Acidic Fibroblast

Growth

Factor

Figure mRNA

8. In Situ Hybridization of aFCF in the Developing Rat Cerebellum

Film images of parasagittal sections of rat cerebellum at P9 (A) and P21 (B) hybridized with the labeled oligonucleotide probe for aFCF mRNA. (A) At P9, high levels of labeling areobserved localized to both the internal and external granule cell layer of the developing cerebellum. (B) By P21, labeling of the mature cerebellar granule cell layer is apparent. Bars, 1000 urn.

plasm. This restricted localization in specific neuronal populations may be significant for understanding the biological role of the FGFs in these neuronal systems. For example, Murphy et al. (1990) have reported that FGF mRNA has a short half-life and that levels of the message are regulated by posttranscriptional processes. Based on these observations, we would hypothesize that, under normal conditions, aFGF is rapidly degraded in the cytoplasm, where message and

protein levels are below the level of detection. Only under specific stimulatory conditions would significant levels of mRNA and protein be found in these neurons. Confirmation of this hypothesis and identification of specific regulatory signals will be an important area of further research. The physiological actions of the FGFs are mediated through activation of a selective neuronal population(s). A recent report has demonstrated the distribu-

Neuron 406

Figure 9. Photomicrograph of Sagittal Section of Cerebellar Cortex from Rat at P21 Hybridized with aFCF Antisense Oligonucleotide Probe and Counterstained with Cresyl Violet Labeling is observed over thegranule ceils (G) but not Purkinje cells. M, molecular layer. Bar, 25 pm.

tion of an mRNA for a FGF receptor in rat brain using in situ hybridization (Wanaka et al., 1990). It is of interest to note that this study demonstrates expression of FGF receptor mRNA in many of the same neuronal populationsthatwe havedetected mRNAfortheaFGF hormone, including hippocampus, substantia nigra pars compacta, locus ceruleus, as well as the cerebellar granule cell layer. Furthermore, no signal peptide has been identified for aFGF, raising questions about the mode of secretion for this hormone (Barde, 1989). These data, coupled with the localization of aFGF mRNA in both efferent and target neuronal populations, such as the entorhinal cortex and dentate gyrus, suggest that aFGF could serve not only a trophic function, but may also have paracrine or autocrine actions in neuron survival, depending upon the location and pharmacological selectivity of the receptors. We have observed aFGF mRNA expression in many areas of the brain that are known to be significantly

involved in human neurodegenerative disease, including substantia nigra, locus ceruleus, hippocampus, and cerebral cortex, as well as caudate putamen and nucleus accumbens. In culture, FGF has been shown to promote survival and development of mesencephalicdopaminergic neurons(Ferrari et al., 1989). Recently, FGF infused into the caudate in vivo has been shown to promote the survival of substantia nigra neurons challenged with the neurotoxin l-methyl4 phenyl-1,2,3,6-tetrahydropyridine and promote the induction of tyrosine hydroxylase activity and dopamine turnover in the caudate (Otto and Unsicker, 1990). These data obtained from different neuronal systems that are involved prominently in neurodegenerative diseases such as Alzheimer’s or Parkinson’s disease suggest that enhancement of endogenous aFGF expression in these brain regions could have important effects in ameliorating the neurodegenerative process.

Figure 10. Photomicrograph Demonstrating Hybridization for aFGF mRNA in Substantia Nigra Pars Compacta of the Adult Rat Relatively high densities of grains are found over neurons in this region. Bar, 25 urn.

1 xpression ‘107

of Acidic Fibroblast

‘Ggure 11. Photomicrograph

Growth

Factor

Demonstrating

Localization

of aFGF mRNA to Neurons

in the Locus Ceruleus

.‘he two frames show serial parasagittal sections from a rat brain at P21 taken through the locus ceruleus. In both frames the middle r:erebellar peduncle is located in the upper right corner. Photographic grains appear to be preferentially located near the nucleus in these neurons (A). Background levels are demonstrated by hybridization of a serial section with the sense-sequence probe (B). Bar, 25 pm.

It cannot be overlooked that the effects of FGF paral!el in many instances theeffects of nerve growth factor !n these same systems. In the hippocampus, the pat:ern of aFGF expression is strikingly similar to that reported for nerve growth factor (Ayer-Le Lievre et al., 1988; Whittemore et al., 1988). While it is possible that 1hese different factors play independent roles in neul,al development and neuron survival, a significant body of evidence suggests that these as well as other growth factors and cytokines work in concert to modulate thecomplex events involved in establishing specific cell-cell contacts, synaptic remodeling and mainlenance, and neuron survival (Sporn and Roberts, .1988). These potential interactions in vivo remain an area for further research. In conclusion, these studies provide evidence that aFGF is expressed in specific neuronal populations. These data suggest that aFGF plays a significant role both in central nervous system development as well ,IS in the regulation of neuronal viability and plasticity n the adult brain.

Experimental

Procedures

In Situ Hybridization Histochemistry Tissue Preparation Timed pregnant rats (Zivic-Miller Sprague-Dawley, Indianapolis, IN) were anesthetized with Equithesin at selected gestational time points. Brains from individual fetuses and neonates (El& P21) were removed and flash frozen in liquid nitrogen. Tissue was stored at -80°C until use. Parasagittal sections 10 pm thick were cut on a cryostat and briefly thaw-mounted onto RNAasefree gelatin-coated glass slides. The sections were quickly air dried and stored at -80°C until use. Probe Preparatiofl The antisense and sense probes used for these studies were 36-mer oligonucleotides complementary to amino acids 26-37 of bovine aFGF, which is 90% homologous to that of rat aFGF, and 45-mer oligonucleotides complementary to amino acids 125-139 of rat aFGF (Goodrich et al., 1989; Abraham et al., 1986; Halley et al., 1988). The oligonucleotide sequence of bovine aFGF chosen was 87% homologous with the corresponding sequence of rat aFGF and 67% homologous with the corresponding sequenceof rat bFCF. C:C ratio of the bovine oligonucleotides was 61% and that of the rat sequence probes was 60%. These probes were chemically synthesized in the Department of Biochemistry on an Applied Biosystems Model 380A DNA synthesizer.

The oligonucleotide probes were labeled at the 3’ end in a standard reaction using terminal deoxynucleotidyl transferase (Tdt; BRL) and [?j]deoxyadenosine 5’(a-thio)triphosphate (Du Pant/New England Nuclear) (Siegel, 1989). The reaction mixture consisted of oligonucleotide probe at a final concentration of 0.5 PM, 5x potassium cacodylate tailing buffer, 70 PCi of [?j]dATP, and sterile distilled water in a final volume of 50 ~1. To initiate the reaction, 75 U of enzyme was added, and the mixture was incubated at 37OC for 10 min. The reaction was stopped by the addition of 175 ~1 of Tris-EDTA buffer and 1~1 of yeast tRNA. The sample was extracted with 225 1.11 of phenol, chloroform, isoamyl alcohol (50:49:1). The aqueous phase was transferred to a clean tube and extracted with 225 ~1 of chloroform, isoamyl alcohol (49:l). The sample was again transferred to a clean tube and precipitated by the addition of l/20 vol of 5 M NaCl and 2 l/2 vol of 100% ethanol and placement on dry ice for 30-60 min. The tubewas then centrifuged for 45 min at4V; the supernatant was discarded, and the resulting pellet was rinsed in ice-cold 100% ethanol. After drying under vacuum, the pellet was reconstituted in 50 PI of Tris-EDTA buffer and 1 PI of 2.25 M dithiothreitol (Sigma). The labeled probes were stored at 4OC until use. Hybridization The frozen sections were brought to room temperature, postfixed in phosphate-buffered 4% formaldehyde for 5 min, and rinsed in three changes of phosphate-buffered saline (pH 7.3). The sections were then incubated in 0.25% acetic anhydride in 0.1 M triethanolamine HCI (pH 8.0) for 10 min at room temperature, followed by two brief washes in 2x SSC (1 x SSC contains 0.15 M NaCI, 0.015 M sodium citrate [pH 7.01). Finally, the tissues are incubated in a series of ethanol (70%, 80%, 95%, 100%) for 5 min each and allowed to air dry. To reduce nonspecific probe binding, the sections were prehybridized at room temperature in hybridization buffer without probe. Small pieces of Parafilm are placed over each section to prevent evaporation. For oligonucleotide probes, the hybridization buffer consisted of 4x SSC, 50% formamide, 500 pglml sheared single-stranded salmon sperm DNA, 250 pglml yeast tRNA, 1 x Denhardt’s solution (0.02% Ficoll, 0.02% polyvinylpyrrolidone,0.02% bovineserum albumin), and 10% dextran sulfate. After 1 hr, the coverslips were removed by soaking in 2x SSC. Each section was then covered with 45 ~1 of hybridization buffer containing IO6 cpm of probe, covered as before with Parafilm, and incubated 18-24 hr in a humid chamber at room temperature. The Parafilm coverslips were then removed, and the sectionswere washed through a series of buffers of increasing stringency to remove free and nonselectively hybridized probe. For these studies we determined that 4 washes of 15 min each in 2x SSC, 50% formamide at 40°C, followed by 2 washes in 1 x SSC for 1 hr each at room temperature, and a final wash in 0.5x SSC for 1 hr at room temperature provided optimal hybridization with minimal crosshybridization. After the final wash, the sections are dipped briefly in dHzO and air dried (Siegel, 1989). For all experiments, serial sections from all developmental time points were processed simultaneously to allow more accurate comparisons. Autoradiography Labeled slides from each experiment were placed together against X-ray film (Kodak X-OMAT AR) in a cassette and exposed at 4’C. Sections from adult brain were routinely exposed 4-8 weeks. Sections from developing brain were routinely exposed 10 days to 2 weeks. For photographic purposes, some films were allowed to expose for longer periods of time. For higher resolution, the slides were then dipped in liquid emulsion (Kodak NTB3) and allowed to dry for at least 3 hr. The dipped slides were then stored in light-tight slide boxes with desiccant at 4oC for 6-8 weeks. Autoradiograms were interpreted with reference to Paxinos and Watson (1986) and Sherwood and Timiras (1970). Northern Blot Analysis Probe Preparation Oligonucleotide probes were prepared by end labeling at the 5 end with [r-32P]adenosine Striphosphate (Du Pant/New England Nuclear) using T4 polynucleotide kinase (BRL) and standard

methods (Sambrook et al., 1989). Oligonucleotide (50 pmol) was added to a tube containing 10 ~1 of 10x kinase buffer, 150-200 PCi of [$*P]ATP, IO-20 lJ of T4 polynucleotide kinase, and water in a final volume of 50 ~1. The reaction was carried out at 37OC for 1 hr and stopped by the addition of 450 ~1 of Tris-NaCCEDTA buffer and 1 PI of yeast tRNA (25 mg/ml). The sample was then extracted with 500 PI of phenol, chloroform, isoamyl alcohol (SO: 49:1), and the aqueous phase was removed to a clean tube. The sample was then extracted with chloroform alone. The aqueous phase was transferred to a clean tube and precipitated on dry ice after the addition of l/20 vol of 4 M NaCl and 2 l/2 vol of 100% ethanol. Aftercentrifugation, the resulting pellet was rinsed with cold ethanol and allowed to dry. The pellet was then reconstituted in 500 PI of Tris-EDTA buffer and 10 ~1 of ethanol. Northern Hybridization Total RNA was isolated from brain tissue taken from rats at P9 and P21 by the method of Chirgwin et al. (1979). Forty to sixty micrograms of total RNA was electrophoresed on a horizontal formaldehyde-agarose gel (1%) and transferred to a nitrocellulose membrane by standard methods (Sambrook et al., 1989). The filter was prehybridized for at least 3 hr at 42’C in a sealed plastic bag containing hybridization buffer without probe. Hybridization buffer consisted of 5x SSPE (1 x SSPE contains 10 m M sodium phosphate [pH 7.01, 0.8 M sodium chloride, and 1 m M EDTA) containing 50% formamide, 5x Denhardt’s solution, 0.1% SDS, 2.5 mg of sheared single-stranded salmon sperm DNA, and 10% dextran sulfate. Labeled probe (IO6 cpm/ml) was added to the bag and incubated 20-24 hr at 42OC. The filters were then washed according to standard methods with a final wash in 0.1 x SSPE, 0.1% SDS at 50°C and placed against X-ray film between two Du Pont Quanta III intensifying screens for varying amounts of time at -8OOC. Acknowledgments The authors wish to thank Ms. Hua Hun He for her excellent technical assistance; Drs. Gary Landreth, Jeffery Twiss, Ruth Siegel, Steven Younkin, Todd Golde, and Story Landis for scientific advice; and Drs. Landreth, Siegel, and Landis again for critical appraisal of the manuscript. This work was supported by a grant from the Alzheimer’s Disease and Related Disorders Association (PRG-88-105) and in part by Center grants from the National Institute on Aging (lP50A68012-Ol), the Ohio Department on Aging (ADR-2), and the Alzheimer Center, University Hospitals of Cleveland. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement”in accordance with 18 USC Section 1734 solely to indicate this fact. Received

July 25, 1990; revised

January

3, 1991.

References Abraham, J. A., Mergia, A., Whang, J. L., Tumolo, A., Friedman, J., Hjerrild, K. A., Gospodarowicz, D., and Fiddes, J. C. (1986). Nucleotide sequence of a bovine clone encoding the angiogenic protein, basic fibroblast growth factor. Science 233, 545-548. Alterio, J., Halley, C., Brou, C., Soussi, T., Courtois, Y., and Laurent, M. (1988). Characterization of a bovine acidic FGF cDNA clone and its expression in brain and retina. FEBS Lett. 242,41-46. Anderson, K. J., Dam, D., Lee, S., and Cotman, C. W. (1988). Basic fibroblast growth factor prevents death of lesioned cholinergic neurons in viva. Nature 332, 360-361. Angevine, 1. B. (1965). Time of neuron origin region. Exp. Neural. (Suppl.) 2, 1-41.

in the hippocampal

Ayer-Le Lievre, C., Olson, L., fbendal, T., Seiger, A., and Persson, H. (1988). Expression of the p nerve growth factor gene in hippocampal neurons. Science 240, 1339-1341. Barde, Y.-A. (1989). Trophic ron 2, 1525-1534.

factors

and neuronal

survival.

Neu-

Expression of Acidic Fibroblast

409

Growth

Factor

Chirgwin, I. M., Przybyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979). Isolation of biologically active ribonucleic acid from sources enriched in ribonuclease. Biochemistry 78, 5294-5299. Ferrari, G., Minozzi, M.-C., Toffano, G., Leon, A., and Skaper, S. D. (1989). Basic fibroblast growth factor promotes the survival and development of mesencephalic neurons in culture. Dev. Biol. 733, 140-147. Gensburger, C., Labourdette, G., and Sensenbrenner, M. (1987). Brain basic fibroblast growth factor stimulates the proliferation of rat neuronal precursor cells in vitro. FEBS Lett. 217, 1-5. Goodrich, S. P., Yan, C. C., Bahrenburg, K., and Mansson, P. E. (1989). The nucleotide sequence of rat heparin binding growth factor 1 (HBCF-I). Nucl. Acids Res. 17, 2867. Cospodarowicz, broblast growth

D., Neufeld, C., and Schweigerer, L. (1986). Fifactor. Mol. Cell. Endocrinol. 46, 187-204.

Halley, C., Courtois, Y., and Laurent, M. (1988). Nucleotide sequence of bovine acidic fibroblast growth factor. Nucl. Acids Res. 76, 10913. Hatten, M. E., Lynch, M., Rydel, R. E., Sanchez, J., Joseph-Silverstein, J., Moscatelli, D., and Rifkin, D. B. (1988). In vitro neurite extension by granule neurons is dependent upon astroglialderived fibroblast growth factor. Dev. Biol. 725, 280-289. Jaye, M., Howk, R., Burgess, W., Ricca, G. A., Chiu, I. M., Ravera, M. W., O’Brien, S. J., Modi, W. S., Maciag, T., and Drohan, W. (1986). Human endothelial cell growth factor: cloning, nucleotide sequence, and chromosome localization. Science 233, 541-545.

the Developing Press).

Rat Brain (Berkeley,

CA: IJniversity

of California

Siegel, R. E. (1989). In situ hybridization histochemistry. In Methods in Neurosciences, Vol. 1, P.M. Conn, ed. (Florida: Academic Press, Inc.), pp. 136-150. Singh, S. C. (1977). The development of olfactory and hippocampal pathways in the brain of the rat. Anat. Embryol. 751,183-199. Sporn, M. B., and Roberts, A. B. (1988). Peptide are multifunctional. Nature 332, 217-219.

growth

factors

Unsicker, K., Reichert-Preibsch, H., Schmidt, R., Pettmann, B., Labourdette, G., and Sensenbrenner, M. (1987). Astroglial and fibroblast growth factors have neurotrophic functions for cultured peripheral and central nervous system neurons. Proc. Natl. Acad. Sci. USA 84, 5459-5463. Walicke, P. (1988). Basic and acidicfibroblast trophic effects on neurons from multiple rosci. 8, 2618-2627.

growth factors have CNS regions, J. Neu-

Walicke, P., Cowan, W. M., Ueno, N., Baird, A., and Cuillemin, R. (1986). Fibroblast growth factor promotes the survival of dissociated hippocampal neurons and enhances neurite extension. Neurobiology 83, 3012-3016. Wanaka, A., Johnson, E. M., Jr., and Milbrandt, J. (1990). Localization of FCF receptor mRNA in the adult rat central nervous system by in situ hybridization. Neuron 5, 267-281.

Logan, A. (1988). Elevation of acidic fibroblast growth factor mRNA in lesioned rat brain. Mol. Cell. Endocrinol. 58, 275-278.

Weiner, H. L., and Swain, J. L. (1989). Acidic fibroblast growth factor mRNA is expressed by cardiac myocytes in cultureand the protein is localized to the extracellular matrix. Proc. Natl. Acad. Sci. USA 86, 2683-2687.

Mattson, M. P.,Murrain, M.,Cuthrie, P. B.,and Kater, S. 8. (1989). Fibroblast growth factor and glutamate: opposing roles in the generation and degeneration of hippocampal neuroarchitecture. J. Neurosci. 9, 3728-3740.

Whittemore, S. R., Friedman, P. L., Larhammar, D., Persson, H., Gonzalez-Carvajal, M., and Holets, V. R. (1988). Rat p-nerve growth factor sequence and site of synthesis in the adult hippocampus. J. Neurosci. Res. 20, 403-410.

Morrison, R. S., Sharma, A., DeVellis, J., and Bradshaw, R. A. (1986). Basic fibroblast growth factor supports the survival of cerebral cortical neurons in primary culture. Proc. Natl. Acad. Sci. USA 83, 7537-7541.

Winkles, J. A., Friesel, R., Burgess, W. H., Howk, R., Mehlman, T., Weinstein, R., and Maciag, T. (1987). Human vascular smooth muscle cells both express and respond to heparin-binding growth factor I (endothelial cell growth factor). Proc. Natl. Acad. Sci. USA 84, 7124-7128.

Murphy, P. R., Cuo, J. Z., and Friesen, H. G. (1990). Messenger RNA stabilization accounts for elevated basic fibroblast growth factor transcript levels in a human astrocytoma cell line. Mol. Endocrinol. 4, 196-200. Neufeld, C., and Cospodarowicz, D. (1985). The identification and partial characterization of the fibroblast growth factor receptor of baby hamster kidney cells. J. Biol. Chem. 25,13860-1X%8. Neufeld, C., and Cospodarowicz, D. (1986). Basic and acidic fibroblast growth factors interact with the same cell surface receptors. J. Biol. Chem. 267, 5631-5637. Otto, D., and Unsicker, K. (1990). Basic FGF reverses chemical and morphological deficits in the nigrostriatal system of MPTPtreated mice. J. Neurosci. 70, 1912-1921. Paxinos, C., and Watson, C. (1986). The Rat Brain in Stereotaxic Coordinates (New York: Academic Press). Pettmann, B., Labourdette, G., Weibel, M., and Sensenbrenner, M. (1986). The brain fibroblast growth factor (FGF) is localized in neurons. Neurosci. Lett. 68, 175-180. Rydel, R. E., and Greene, L. A. (1987). Acidic and basic fibroblast growth factors promote stable neurite outgrowth and neuronal differentiation in cultures of PC12 cells. J. Neurosci. 7,3639-3653. Sambrook, j., Fritsch, E. F., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual (Cold Spring Harbor, New York: Cold Spring Harbor Laboratory). Schubert, D., Ling, N., and Baird, A. (1987). Multiple influences of a heparin-binding growth factor on neuronal development. J. Cell Biol. 704, 635-643. Schlessinger, A. R., Cowan, W. M., and Gottlieb, D. I. (1975). An autoradiographic study of the time of origin and the pattern of granule cell migration in the dentate gyrus of the rat. J. Comp. Neural. 759, 149-176. Sherwood,

N. M., and Timiras,

P. S. (1970). A Stereotaxic

Atlas of