Expression of Basic Fibroblast Growth Factor and Its Receptor in the Rat Subfornical Organ

Abstract

Basic fibroblast growth factor (FGF), its mRNA and the mRNA that encodes for its receptor have all been localized in the rat subfornical organ (SFO). Basic FGF is widely distributed throughout the SFO; it is present in neurons, in the vascular basement membrane of lateral blood vessels (but not those within the SFO) and in ependymal cells surrounding the SFO. Results of in situ hybridization show that the expression of basic FGF mRNA is detected throughout the organ. Similarly, the expression of flg, the gene for the putative basic FGF receptor, can also be detected in the SFO. The results all support the possibility that this growth factor may modulate the known physiological functions of the SFO.

Although the brain is one of the richest sources of basic fibroblast growth factor (FGF), its physiological function in the CNS is unknown [1, 2]. It is thought to act as an autocrine as well as a paracrine factor whose activities are restricted to and regulated at its site of synthesis [3–5]. Although it is known to modulate cell growth, function and survival in vitro [6–12], it is presumed to be a neurotrophic factor for neurons of the central and peripheral nervous system [6–9]. Because the injection of basic FGF into the brain elicits a strong angiogenic response [12], it has also been implicated in injury repair in the CNS [13] and more recently in neurodegenerative diseases such as Alzheimer’s disease [14].

Virtually nothing is known about the physiological role of growth factors in the CNS. Accordingly, we have been examining the distribution of basic FGF [15], its mRNA [16] and more recently its receptor in order to assess its potential functions in the brain. Contrary to expectations [1, 2], the growth factor is not ubiquitous in the CNS. Indeed, specific loci of synthesis appear in such areas as the hippocampus and cingulate cortex [16], and can be altered by disease [14]. As an initial step towards identifying a potential neuroendocrine function for basic FGF, we investigated whether the subfornical organ (SFO) is a locus containing basic FGF, its mRNA and its receptor’s mRNA. Because the SFO regulates the sensory, motor, autonomic, endocrine and behavioral processes that influence water balance, we propose that growth factors like basic FGF may be local signals modulating the neurohumoral response [17, 18]. If so, this would be a new, heretofore unexpected role for basic FGF that requires further investigation.

Methods

Animals and Experimental Procedure

Under deep anesthesia (acepromazine 9.5 mg/kg, ketamine 185 mg/kg and xylazine 9.5 mg/kg, i.m.), 7 Sprague-Dawley rats were perfused transcardially using the pH shift method (pH 6.5–9.5) with a solution of phosphate-buffered saline containing 4% paraformaldehyde and 0.05% glutaraldehyde [19]. Brains were postfixed in 4% PFA in PBS containing 10% sucrose for 24 h and snap frozen in Tissue-Tek OCT (Miles Laboratories, Elkhard, Ind) and stored at −80 °C. Brain sections of 20 or 40 μm thickness were obtained with a Reichert-Jung Cryocut 1800 cryostat and collected in a 20% glycerol and 30% ethylene glycol cryoprotectant solution and stored at −20 °C for processing.

Immunohistochemistry

The procedure for the immunoperoxidase staining of basic FGF has been described extensively elsewhere [15]. Briefly, sections were first cleared of cryoprotectant with 4 × 10 min washings using 10 mM Tris buffer (containing 0.15 M NaCl and 0.2% Triton X-100) and then treated with 0.05% trypsin for 5 min. After extensive washing, the endogenous peroxidase was quenched by incubating the sections in 0.3% hydrogen peroxide in PBS for 30 min. The sections were washed and incubated with 3% normal goat serum for 1 h at room temperature to block nonspecific binding, and then with an antibody against basic FGF [15] for 2 h at room temperature. To further reduce nonspecific binding, 1.5% normal goat serum was included in the buffers of subsequent steps, and 1.5% normal rat serum was included with the antirabbit biotinylated antibody (Vector Labs, Burlingame, Calif.). In some instances, adjacent sections were stained with cresyl violet. Control sections were processed by either omitting the primary antibody, replacing it with rabbit IgG or using the flowthrough of a basic FGF affinity column [15].

In situ Hybridization

The procedure that we used for in situ hybridization was performed as described by Emoto et al. [16] using a variation of the procedure described by Simmons et al. [19]. Briefly, the XhoI-XhoI fragment derived from a rat basic FGF cloned cDNA [20] was subcloned into Bluescript (Stratagene, San Diego, Calif.), and the antisense RNA of the coding sequence was transcribed using T7 RNA polymerase and ³⁵S-UTP. The sense RNA was transcribed with T7 RNA polymerase and used on control sections. The in situ hybridization for the basic FGF receptor mRNA was performed as described by Buscaglia et al. [21]. DNA coding the extracellular domain of flg [22] was cut, subcloned into Bluescript (Stratagene, San Diego, Calif.) and the antisense RNA was transcribed using T7 RNA polymerase and ³⁵S-UTP. The sense RNA was transcribed with T3 RNA polymerase and ³⁵S-UTP and used on adjacent control sections. All tissue sections were exposed to Kodak NTB-2 liquid autoradiograph emulsion for 3 weeks, developed with Kodak D19 developer, fixed, and then counterstained with hematoxylin.

Results

Localization of Basic FGF in the Rat SFO

Basic FGF is localized in the rat SFO (fig. 1a). Very high levels of staining are observed in the basement membrane of the septal venules lateral to the SFO (dark arrows in fig. 1a). There is intense staining for basic FGF in cellular structures within the SFO. Higher magnification (fig. 1b) reveals that basic FGF is localized mostly in neuronal perikarya, although some glial cells appear immunopositive. The distribution of staining is widespread throughout the SFO, and staining of both nucleus and cytoplasm of the cells is readily detected. Basement membrane staining is not as striking in the microvasculature within the organ. The ependymal cells appear to be the darkest cell structure that stain for basic FGF. Analysis at high magnification (fig. 1c) shows the basic FGF immunoreactivity is localized in cells scattered throughout the ependymal layer and the staining is cytoplasmic. Adjacent sections stained with a preparation of antibody depleted of anti-FGF IgG [15] fail to stain the cells indicating that the staining is specific (fig. 1d). The staining of the ependymal cells, while not completely abolished, is significantly reduced.

Fig. 1.

Immunostaining for basic FGF in the rat SFO. Floating sections were stained with an anti-FGF antibody as described in the text. a, d 40-μm-thick sections. b, c 20-μm-thick sections. a Low magnification (× 40) of the immunoreactive basic FGF in the SFO. The dark arrows highlight the septal blood vessels of the SFO and the open arrow shows small vessels of the septum of the SFO. The staining is associated with the basement membrane. IIIV = 3rd ventricle. b Higher magnification (× 100) revealed that basic FGF was contained within neurons and ependymal cells lining the 3rd ventricle. Staining was relatively absent in cells surrounding the vessels in the SFO. c The flowthrough eluate of a basic FGF Affi-gel column failed to stain adjacent sections (× 100). d Basic FGF can be localized to tanycyte-like cells of the ependymal layer (arrows). × 500.

Localization of Basic FGF mRNA in the Rat SFO

In an effort to determine if the immunoreactive basic FGF detected in the SFO is locally synthesized in this tissue, we processed rat brains and examined them for the presence of basic FGF mRNA. As shown in figure 2a, an intense signal of basic FGF mRNA is observed throughout the organ. The hybridization is specific and the signal is absent in sections incubated with sense RNA (fig. 2b). High magnification analyses of bright field images reveals that the intense signal emanates from selected cells scattered throughout the organ (fig. 2c). The localization of the signal is compatible with the hypothesis that neuronal and ependymal cells located in the SFO are responsible for basic FGF expression. Incubation with the sense RNA fails to show the presence of significant hybridizing signal (fig. 2d).

Fig. 2.

In situ hybridization of basic FGF mRNA in the rat SFO. Tissue sections of the rat brain (20 μm) were processed as described in the text. a A dark field image revealed the intense labeling in the SFO when hybridized with the antisense RNA. b Sections labeled with an RNA probe encoding the sense RNA. c A high magnification bright field micrograph taken of the lateral SFO region shows that the signal is cell-associated. d Bright field corresponding to an adjacent section hybridized with the sense RNA. Bar = 100 μm.

Localization of flg mRNA in the Rat

In an effort to determine if the immunoreactive basic FGF and its mRNA that are detected in the SFO are locally utilized, we examined them for the presence of mRNA encoding flg, a basic FGF receptor [22]. As shown in figure 3a, a low magnification image shows that mRNA for flg is concentrated in the SFO and appears continuous throughout the organ. The hybridization is specific and there is absence of signal in sections incubated with the sense RNA (fig. 3b). At high magnification, bright field images demonstrate that the distribution pattern of flg mRNA is very similar to that of basic FGF mRNA (fig. 3c). The localization of the flg mRNA in the SFO suggests local utilization of basic FGF. Incubation with the sense RNA fails to show the presence of a significant hybridizing signal (fig. 3d).

Fig. 3.

In situ hybridization of flg mRNA in the rat SFO. Tissue sections of the rat brain (20 μm) were processed as described in the text. a A dark field micrograph showed intense labeling of the SFO when hybridized with the antisense RNA that encodes for the basic FGF receptor. b Sections hybridized with an RNA probe encoding the sense RNA. c Bright field micrograph taken of the lateral SFO region at higher magnification shows the signal is cell-associated. d Bright field corresponding to an adjacent section hybridized with the sense RNA. Bar = 50 μm.

Discussion

The SFO is an almost hemispherical protrusion of the rostrodorsal wall of the third ventricle that is involved in the control of oxytocin and vasopressin release, in the secretion of anterior pituitary hormones, in fluid uptake and in the regulation of blood pressure [23, 28–33]. Located at the level of the intraventricular foramina, it shares several morphological characteristics with other circumventricular organs. First, it lacks a blood brain barrier [23, 24]; second, it contains ependymal tanycytes that transport substances from the CSF and plasma [18, 25], and, finally, it is a rich source of neuroendocrine cells and terminals that link to the hypothalamoneurohypophysial system [26–28]. The observation that the genes for basic FGF and its receptor are specifically expressed in cells of the rat SFO strongly supports the hypothesis that basic FGF plays a biological function in this tissue. This function remains unknown.

Several studies have established the neurotrophic activity of basic FGF in vitro and in vivo [6–9]. However, recent evidence supports the notion that it also plays a neuroendocrine function. As an example, basic FGF can potentiate the effects of thyrotropin-releasing hormone on rat pituitry cells in vitro [11]. Furthermore, acidic FGF, a structural homologue to basis FGF [4], suppresses food intake when injected intracerebroventricularly [34]. lnterleukin-1, another member of the FGF family of growth factors [4], stimulates adrenocorticotropin hormone secretion [35–37], thyroid-stimulating hormone secretion [38, 39] and growth hormone [37, 38] secretion in vitro and elevates body temperature and enhances slow wave sleep in vivo [40]. Because neither acidic FGF nor interleukin-1 have been localized in the SFO, the expression of basic FGF and its receptor in this tissue emphasizes the importance of investigating its potential neuroendocrine function.

Based on the known effects of basic FGF on the microvascular endothelium, the possibility that basic FGF has a non-neuroendocrine function in maintaining the growth and function of the complex and highly permeable capillary network [17, 18, 41] that characterizes the SFO must also be investigated. The angiogenic effects of this growth factor are extensively documented [4, 10, 12], and it is now well recognized as one of the most potent angiogenic substances known. Remarkably, however, there is little basic FGF immunoreactivity associated with the vasculature of the SFO suggesting that any direct angiogenic function of an SFO-derived basic FGF is presumably mediated and regulated by cells within the SFO. The high density of neurons which innervate the SFO capillary system may thus be essential for supporting the extent to which the SFO is vascularized.

For many reasons, it is particularly interesting to observe basic FGF and its mRNA in ependymal cells of the SFO. Morphological studies have suggested that the ependymal tanycyte extend processes well into the rat SFO so as to contact capillaries, glial cells and neurons and transfer CSF-borne substances [18]. Because ependymal cells, some which appear to be tanycytes, contain basic FGF (see fig. 1d) and express its gene, the SFO-derived basic FGF could be involved in tanycyte-mediated functions such as regulating osmosensitivity. Furthermore, although basic FGF is not thought to be secreted in the classical sense [4], the protein is in an ideal location to be released into the third ventricle. In as much as injury is now well recognized to release basic FGF [42, 43, 44], its localization in the ependyma makes it ideal for release into the CSF after CNS trauma. Finally, it is important to emphasize that the SFO is a unique site within the CNS inasmuch as its cells express high levels of the genes for both basic FGF and its receptor. Accordingly, the present results support the need for further investigation into the physiological role of this potent growth and differentiation factor in the SFO.