INTRODUCTION

Top Abstract Introduction Materials and methods Results Discussion References

FIBROBLAST GROWTH FACTOR (FGF)-2, also known as basic FGF, is a mitogenic and angiogenic protein that has been found in all tissues examined thus far (3, 16). The effects of FGF-2 are exerted through cell-surface, high-affinity tyrosine kinase receptors (FGFR) and low-affinity sites consisting of heparan sulfate proteoglycans (13, 20). Receptors for FGF-2 (FGFR-1) are present in the embryonic and adult heart (15, 17, 22, 23) and were shown to be essential for normal heart development (26). Although FGF-2 is present in the heart into adulthood (15, 23), its role in the postnatal heart is less clear. FGF-2 is found intracellularly as well as outside the cell, where it is able to exert its effect on cell-surface receptors. However, the mechanism for FGF-2 export is unclear because it contains no signal peptide (3). There is evidence for unconventional release of FGF-2 via a pathway independent of the endoplasmic reticulum-Golgi complex (25). Also, factors involved in facilitating or inhibiting the export of FGF-2 have been identified (10, 36). With regard to the postnatal heart, studies have shown that FGF-2 is released on contraction and that this can be regulated by increasing the heart rate and force of contraction by electrical or adrenergic stimulation (8, 18). The mechanism of release is reported to involve transient, nonlethal disruptions of the plasma membrane, a phenomenon that occurs in many tissues exposed to high levels of mechanical stress (24). Factors influencing FGF-2 release may also be expected to influence FGF-2 synthesis. The increases in FGF-2 mRNA as well as protein after more damaging types of tissue injury (5, 9, 30) raise the possibility that regulation at the level of transcription is a component of FGF-2 release associated with the transient remodeling of the membrane in the contracting heart.

Detection of FGF-2 mRNA appears to be more difficult than detection of protein, reflecting, presumably, low activity of the FGF-2 promoter and/or relatively unstable transcripts. This has necessitated the use of large amounts of RNA for blotting studies, often in conjunction with tumor cells overexpressing FGF-2 (4, 28). An alternative approach for transcriptional studies is the reporter gene assay. We recently cloned a 1.4-kb fragment of the rat genome containing ~1 kb of FGF-2 upstream flanking DNA, including a promoter region, and reported the sequence for the 552/+252 domain (34). We have now examined an extended FGF-2 promoter region (1,058/+54) for a response to adrenergic stimulation both in vitro, using the luciferase reporter gene for transient transfection of neonatal rat cardiac myocytes, and in vivo, through the generation of transgenic mice expressing the hybrid FGF-2/luciferase gene. We have also completed the sequencing of the 1.4-kb clone and examined it for the presence of adrenergic responsive DNA elements. Our data indicate that FGF-2 promoter activity is stimulated both in vitro and in vivo after adrenergic stimulation, and these results correlate with increases in endogenous FGF-2 RNA accumulation. Characterization of this response in vitro through the use of an ₁-selective antagonist, prazosin, suggests that the majority of this effect is mediated through the ₁-adrenergic pathway, but it does not appear to be dependent on either Ca²⁺ influx or myocyte contraction. These results are discussed in relation to a role for FGF-2 in maintaining a healthy myocardium during postnatal development and in adulthood.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and methods Results Discussion References

Cell culture and tissue extracts. Neonatal rat cardiac myocyte cultures were prepared essentially as previously described (35). Briefly, ventricles from rat pups (at 1-36 h after birth) were dissected, and the cells were dissociated in a spinner flask using a combination of trypsin (GIBCO BRL, Burlington, ON, Canada) and DNase I (Sigma-Aldrich, Oakville, ON, Canada). Myocytes were separated from nonmuscle cells on a discontinuous Percoll gradient and plated on collagen-coated plates at a density of 1 × 10⁶ cells per 35-mm dish. Cells were initially plated in Ham's F-10 medium containing 10% fetal bovine serum (FBS), 10% horse serum, antibiotic (1,000 U/ml penicillin, 1 mg/ml streptomycin), and CaCl₂ supplemented to 1.05 mM. All cultureware was purchased from Corning (Fisher Scientific, Nepean, ON, Canada), and all media and culture reagents were from GIBCO BRL.

Sequencing. Isolation of the genomic clone -rFGF2-c4 from a Sprague-Dawley rat testis genomic library, which included a 1.4-kb BamH I fragment (B2) containing FGF-2 coding sequences and upstream flanking DNA, including a promoter region, was described previously (34). The 1.4-kb B2 fragment was subcloned into pUC119, and nucleotide sequence was determined by the dideoxy method using a Promega Femtomole Sequencing Kit (Fisher Scientific). The sequence was analyzed for consensus binding sites for known transcription factors using TFSEARCH (Y. Akiyama, Kyoto University, Kyoto, Japan).

Plasmids and constructs. The hybrid genes 1058FGFp.luc, 911FGFp.luc, and 313FGFp.luc, containing fragments of the rat FGF-2 gene fused upstream of a promoterless firefly luciferase gene (-p.luc, contained in the vector pXP1) (29) were described previously (34). A 250-bp fragment of the myosin light chain-2 (MLC-2) promoter cloned upstream of luciferase was described previously (14).

RNA analyses. Total RNA was isolated from transfected rat cardiomyocyte cultures or mouse hearts using guanidine isothiocyanate (7). Total RNA was denatured with formaldehyde and resolved by electrophoresis through a 1.0% agarose gel. The RNA was blotted to nitrocellulose, probed with a radiolabeled rat FGF-2 (Sma I/Xho I) cDNA fragment, and assessed by autoradiography. The rat FGF-2 fragment corresponds to a full-length FGF-2 cDNA from which those sequences peculiar to the high-molecular-weight isoforms have been deleted (33). Hybridization with the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA (Pst I/Bgl I) was used to assess loading.

Transient gene transfer. Cardiac myocytes were transfected by the calcium phosphate-DNA precipitation method. Sixty micrograms of plasmid were made up to a volume of 1.0 ml in 252 mM CaCl₂ and added gradually to an equal volume of aerated 2× HEPES-buffered saline (280 mM NaCl, 50 mM HEPES-KOH, pH 7.10, and 1.5 mM Na₂PO₄). Precipitate was allowed to form at room temperature for 30 min, and 325 µl were added to each of six culture dishes containing DMEM-10% FBS. After a 16-h transfection period, the cells were washed thoroughly with Ca²⁺- and Mg²⁺-free phosphate-buffered saline. For stimulation, the medium was changed to DMEM F-12 containing 1× insulin-transferrin-selenium-A (GIBCO-BRL), 0.02 mg/ml ascorbic acid, and antibiotics. These "identical" plates of transfected myocytes were then treated with adrenergic agonists or antagonists, as described in the text, to ensure a direct comparison of promoter activity in untreated and treated cells. Phenylephrine, 2,3butanedione monoxime, and nifedipine were purchased from Sigma-Aldrich; norepinephrine, prazosin, and atenolol were purchased from Research Biochemicals International (Natick, MA).

Reporter gene assays. After stimulation, transfected cardiac myocytes were harvested with trypsin-EDTA, pelleted, rinsed with Dulbecco's phosphate-buffered saline, and lysed in 100 mM Tris · HCl, pH 7.8, containing Triton X-100. Insoluble material was removed by centrifugation, and the luciferase activity in the supernatant was measured using the Promega Luciferase Assay System (Fisher Scientific) and a luminometer (ILA900 Luminometer, Tropix, Bedford, MA) according to the manufacturer's instructions. Luciferase activity was normalized against lysate protein content as determined by the Bradford Assay (Bio-Rad Laboratories, Mississauga, ON, Canada).

Transgenic mice. All animal experiments were done in accordance with the standards of the Canadian Council for Animal Care. The plasmid 1058FGFp.luc was linearized with Pvu I and Pst I and introduced into the male pronuclei of single-cell zygotes from CD1 mice. Injected embryos were subsequently transferred to the oviduct of surrogate mothers and brought to term. Tail tips of 3-wk-old progeny were removed, and genomic DNA was extracted using Proteinase K digestion followed by phenol-chloroform extraction and ethanol precipitation. The DNA was digested with Xba I, electrophoresed on a 1% agarose gel, and blotted to nitrocellulose membrane. The presence of the transgene was determined by probing with an EcoR I fragment of 1058FGFp.luc. The fragment was labeled with ³²P using the random priming method (Promega Prime-A-Gene Kit, Fisher Scientific). Animals giving positive DNA signals were bred and their neonatal progeny tested for luciferase activity in the heart and brain. This resulted in the establishment of two lines (P300 and P66), which were tested for luciferase activity in the adult heart and for response to phenylephrine. Adult animals (8-10 wk old) were divided into two groups and injected intraperitoneally with vehicle (saline) or phenylephrine (50 mg/kg in saline). After 6, 24, or 48 h, the animals were euthanized, hearts were dissected and fast-frozen on dry ice, and luciferase activity was determined from tissue homogenates as described in Reporter gene assays.

Statistical analysis. Data presented in text and Figs. 1-5 are means ± SE. Statistical analysis of the results was carried out using the Mann-Whitney (nonparametric) test. In all cases, a value was considered statistically significant if P was 0.05.

	RESULTS

Top Abstract Introduction Materials and methods Results Discussion References

FGF-2 RNA levels and promoter activity are stimulated by norepinephrine. Neonatal rat cardiac myocytes were isolated and treated without (control) or with 0.01 mM norepinephrine or norepinephrine with an -adrenergic antagonist (0.01 mM prazosin) for 6 h to assess any effect on FGF-2 RNA levels. RNA was separated by gel electrophoresis, blotted to nitrocellulose, probed with a radiolabeled fragment of the rat FGF-2 cDNA, and visualized by autoradiography (Fig. 1A). The 28S RNA band seen with ethidium bromide staining before transfer to nitrocellulose is also shown to allow a comparison of RNA levels (Fig. 1A). Although not as evident because of a discrepancy in loading, the 6.1-kb FGF-2 transcript level was increased after norepinephrine treatment for 6 h. Consistent with this observation, the 6.1-kb FGF-2 transcript was reduced and barely detectable after treatment with norepinephrine and the ₁-specific antagonist prazosin.

View larger version (35K): [in this window] [in a new window]   Fig. 1.   Effect of an -adrenergic antagonist, prazosin (Praz), on stimulation of fibroblast growth factor (FGF)-2 RNA levels and promoter activity by norepinephrine (NE). A: cultured neonatal rat cardiac myocytes were treated without (control) or with 0.01 mM NE or with NE and 0.01 mM Praz (NE + Praz) for 6 h. Isolated RNA was electrophoresed, blotted, probed with radiolabeled rat FGF-2 or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA fragments, as indicated, and visualized by autoradiography. The 6.1-kb FGF-2 and 1.4-kb GAPDH transcripts are indicated. The 28S RNA band for each sample, stained with ethidium bromide and photographed before blotting, is also shown. B: neonatal cardiac myocytes were transfected with 1058FGFp.luc and treated with NE, NE + Praz, or NE and 0.01 mM atenolol (NE + Atl) for 6 h. Cells were subsequently harvested, and luciferase activity and protein concentration were assessed. Promoter activity (luciferase/ng protein) for 1058FGFp.luc gene is shown as mean from multiple determinations (n = 9-15). Data are means ± SE.

A putative phenylephrine-responsive element is present in upstream FGF-2 flanking DNA. The complete sequence of a 1,389-bp genomic fragment containing rat FGF-2 5' flanking DNA is shown in Fig. 2. This corresponds to nucleotide positions 1,058 through +331 based on the primary transcription start site (+1) described for the brain (34). Analysis of these sequences revealed two copies, in tandem, of putative phenylephrine-responsive elements (PEREs). These sequences (5'-AGGGAGGG-3'), located at nucleotide positions 780 and 761, were identified on the basis of their high degree of similarity to sequences present in the human skeletal actin (5'-AGGGAGGG-3') and rat atrial natriuretic factor (ANF) promoters (5'-GGGGAGGG-3') that have been implicated in the response to ₁-specific adrenergic activation by phenylephrine (1). In the latter case, these sequences were shown to bind a specific protein complex and confer phenylephrine responsiveness (1). Consensus binding sites for known transcription factors identified in the FGF-2 sequences are also shown in Fig. 2.

View larger version (53K): [in this window] [in a new window]   Fig. 2.   Sequence of 1,389-bp genomic fragment of rat FGF-2 gene. Primary transcription start site in brain (34) is indicated as +1. Consensus binding sites for transcription factors, including promoter-specific factor Sp-1, are underlined and labeled accordingly, with exception of early growth receptor protein-1 (Egr-1) binding site, which is shaded. Putative phenylephrine (PE)-responsive elements (PEREs) are double underlined. An A/T-rich region conserved between rat and human FGF-2 promoters is boxed. The 2 leucine (leu) and single methionine (met) start codons are indicated. GenBank accession no. for these sequences is U78079. TRE, thyroid hormone-responsive element.

FGF-2 RNA levels and promoter activity are increased by phenylephrine, but this effect appears to be independent of putative PEREs. To initiate a character ization of the putative PEREs in the FGF-2 DNA, conditions were established for phenylephrine stimulation of endogenous FGF-2 RNA levels. Neonatal rat cardiac myocytes were isolated and treated with the ₁-adrenergic agonist phenylephrine for 48 h (1) and then assessed by RNA blotting. An increase in the 6.1-kb FGF-2 transcript was detected with the FGF-2 cDNA probe after phenylephrine treatment for 48 h (Fig. 3A). The GAPDH and 28S RNA transcripts were also assessed as controls for RNA loading, and the results are included for comparison.

View larger version (28K): [in this window] [in a new window]   Fig. 3.   Effect of PE treatment on endogenous FGF-2 RNA levels and transfected hybrid FGF-2/luciferase gene expression in neonatal rat cardiac myocytes. A: cultured neonatal rat cardiac myocytes were treated without (Cont) or with 0.1 mM PE for 48 h. Isolated RNA was electrophoresed, blotted, probed with radiolabeled rat FGF-2 or GAPDH cDNA fragments, as indicated, and visualized by autoradiography. The 6.1-kb FGF-2 and 1.4-kb GAPDH transcripts are indicated. The 28S RNA band for each sample, stained with ethidium bromide and photographed before blotting, is also shown. B, top: truncated regions of rat FGF-2 5'-flanking DNA containing (1058FGF, 911FGF) or not containing (313FGF) putative PEREs were inserted (hatched regions) upstream of luciferase coding sequence in promoterless plasmid (-p.luc). Calcium-phosphate/DNA precipitates made for each hybrid gene were divided between plates of cardiac myocytes to generate "identical" transfected cultures. B, bottom: cultures were subsequently treated with or without PE for 48 h before harvesting and assessment of luciferase activity and protein concentration to allow a direct assessment of effect of PE treatment on expression of each hybrid FGF-2 gene. Results are promoter activities (luciferase/ng protein) expressed as means ± SE from multiple determinations (n = 13-26). Basal levels for -p.luc in presence or absence of PE were 0.078 ± 0.007 and 0.030 ± 0.004 (n = 6), respectively.

Adrenergic stimulation of FGF-2 promoter activity was not affected by contraction arrest or Ca²⁺ influx. In an effort to investigate the role of contraction in the response of basal FGF-2 promoter activity to norepinephrine, cardiomyocytes transiently transfected with 1058FGFp.luc were treated with 0.01 mM norepinephrine in the presence or absence of 50 mM KCl or 1.0 mM 2,3-butanedione monoxime. Stimulation with 0.01 mM norepinephrine in the presence of either KCl or 2,3-butanedione monoxime caused a visible arrest of contraction of the cardiac myocytes during the 6-h incubation period but did not have a significant effect on norepinephrine-stimulated luciferase activity (Fig. 4A).

View larger version (31K): [in this window] [in a new window]   Fig. 4.   Effect of myocyte contraction or nifedipine (Nif) on stimulation of FGF-2 promoter activity by adrenergic stimulation. A: neonatal rat cardiac myocytes were transfected with 1058FGFp.luc and treated with 0.01 mM NE in absence or presence of 50 mM KCl (NE + KCl) or 1.0 mM 2,3-butanedione monoxime (NE + BDM) for 6 h. B: neonatal cardiac myocytes were transfected with 1058FGFp.luc and treated with 0.01 mM PE in absence or presence of Nif (PE + Nif) for 48 h. For experiments in both A and B, cells were subsequently harvested and luciferase activity and protein concentration were assessed. Promoter activities (luciferase/ng protein) for 1058FGFp.luc gene are shown as means ± SE from multiple determinations (n = 3-6).

Phenylephrine treatment increased 1058FGFp.luc transgene expression in the heart. Two independent transgenic mouse lines (P300 and P66) expressing 1058FGFp.luc were established. The relative levels of transgene expression in the brain and heart of nontransgenic and transgenic mice were determined and are expressed as luciferase activity per microgram of protein in the tissue homogenates (Fig. 5A). To assess whether the FGF-2 promoter responds to ₁-adrenergic stimulation in the heart in vivo, adult mice (8-10 wks old) from the P300 line were injected intraperitoneally with 50 mg/kg phenylephrine and then euthanized 6, 24, or 48 h after injection. The hearts were removed, and luciferase activity per microgram of protein was determined (Fig. 5B). A significant 3.7-fold increase in luciferase gene expression compared with that in mice injected with saline alone was observed 6 h after phenylephrine treatment (P < 0.0005, n = 6). At 24 h, the difference, although not quite significant (P = 0.057), remained at 3.6-fold but was lost at 48 h. No significant changes in luciferase activity were observed in saline-injected animals between 6 and 48 h after injection (not shown). The phenylephrine response observed was confirmed in the P66 line, in which the difference at 6 h after phenylephrine administration was 6.9-fold (P < 0.005, n = 4). A parallel assessment of endogenous mouse heart FGF-2 RNA levels at each time point after phenylephrine treatment was done by RNA blotting (Fig. 5C). The level of endogenous mouse 6.1-kb FGF-2 transcript was increased at 24 h but was decreased again by 48 h after administration of phenylephrine. A second FGF-2 transcript of 3.6 kb was also observed in mouse preparations and showed the same pattern of response as the 6.1-kb mRNA. This 3.6-kb transcript was not seen in rat RNA preparations (Figs. 1 and 3).

View larger version (50K): [in this window] [in a new window]   Fig. 5.   Detection of luciferase activity in 1058FGFp.luc transgenic mice and in vivo stimulation of FGF-2 gene expression by 1-adrenoceptor activity in heart. A: brain and heart from neonatal (2-day-old) P300 and P66 mice were dissected, homogenized, and assayed for luciferase activity. Nontransgenic (NT; n = 5) and transgenic (P300 and P66, n = 7) littermate values are shown as means ± SE. B: adult P300 mice expressing 1058FGFp.luc gene were injected intraperitoneally with 50 mg/kg PE (+PE) or saline vehicle (PE). Mice were euthanized 6, 24, or 48 h later (6 h for saline-injected animals) and their hearts assayed for luciferase activity. Values for luciferase per µg protein are shown as means ± SE (n = 3-7). C: RNA was isolated from hearts of mice injected intraperitoneally with PE and maintained for 6, 24, and 48 h. RNA was electrophoresed, blotted, probed with radiolabeled rat FGF-2 or GAPDH cDNA fragments, as indicated, and visualized by autoradiography. Positions of 6.1-kb FGF-2 transcript and a 3.6-kb transcript that was not observed in rat preparations, as well as a 1.4-kb GAPDH RNA, are indicated.

	DISCUSSION

Top Abstract Introduction Materials and methods Results Discussion References

Previously it was reported that FGF-2 is released from adult rat heart on contraction and that both release and contraction can be increased through adrenergic stimulation (8, 18). We have used RNA blotting and gene transfer to demonstrate that FGF-2 is under positive transcriptional control in the heart via the -adrenergic pathway (Figs. 1-5). Norepinephrine is a naturally occurring catecholamine that acts through both - and -adrenoceptors (41). Endogenous FGF-2 RNA levels and transfected rat FGF-2 promoter (1,058/+54) activity were both increased in neonatal rat cardiac myocytes after treatment for 6 h with norepinephrine (Fig. 1). This response was completely blocked with the ₁-selective antagonist prazosin. Also, endogenous FGF-2 RNA accumulation and transfected FGF-2 promoter activity were stimulated in response to treatment with the -specific agonist phenylephrine (Fig. 3). Although -adrenergic signaling was implicated in FGF-2 release from cardiac myocytes (8), we were unable to confirm a direct effect of -receptor activation on synthesis at the transcriptional level using the 1058FGFp.luc gene. The apparent decrease in FGF-2 promoter activity seen after atenolol addition to norepinephrine-treated cells was not quite significant (P = 0.07). Of course, this does not rule out the possibility that -stimulation exerts its effect elsewhere in the synthetic pathway or that the 1,058 bp of FGF-2 5'-flanking DNA used were insufficient to respond to a -adrenergic stimulus. Regardless, our data with phenylephrine as well as norepinephrine, in the presence or absence of prazosin, strongly implicate ₁-adrenoceptors in the regulation of FGF-2 synthesis in cardiac myocytes.

The rat genomic clone containing FGF-2 5'-flanking DNA includes promoter elements, which suggests that a wide variety of mechanisms may control FGF-2 transcription (Fig. 2). Like its human homologue, the rat FGF-2 gene promoter does not contain typical TATA or CAAT boxes (34, 38, 42). Instead, activation of transcription may involve binding of the promoter-specific factor Sp1 or early growth response protein-1 (Egr-1), because binding sites for these proteins are contained in a GC-rich region associated with the location of transcription initiation sites (2, 34). A search for known transcription factor consensus binding sites revealed putative binding sites (Fig. 2) for homeodomain- or basic helix-loop-helix domain-containing proteins, zinc finger proteins, and GATA proteins, and a 6-bp thyroid hormone-responsive element is present that has been described as a half-site for the thyroid hormone receptor (19). The ability of the 1058FGFp.luc gene to respond to phenylephrine in the same positive manner as observed with the endogenous FGF-2 gene suggested that the genetic information contained within the region 1,058/+54 of the FGF-2 gene is sufficient for this response. Two tandem 8-bp elements (5'-AGGGAGGG-3'), which appear closely related to the reported PERE in the ANF promoter (5'-GGGGAGGG-3') (1) were found within this region. Clearly, these elements were candidates for conferring ₁-adrenergic responsiveness on the FGF-2 promoter.

Although the activity of the FGF-2 promoter used in these studies was increased by both norepinephrine and phenylephrine, deletion of the putative PEREs had no effect on this response (Fig. 3B), suggesting that they do not play an essential role in ₁-adrenergic stimulation of the rat FGF-2 promoter. In the case of ANF, Sprenkle et al. (40) reported that the PERE (referred to as an Sp-1-like element by these authors) was by itself responsible for less than threefold phenylephrine induction but that inclusion of an upstream serum-response element (SRE) increased the response to a level of 5.3-7.4-fold. Responses of ANF and MLC-2 promoters to phenylephrine were attributed in part to elements with an A/T-rich core (12, 40, 44, 45). In this context, it is interesting to note that an A/T-rich sequence is contained at nucleotide position 113/108 (5'-TTTAAA-3') in the rat FGF-2 promoter (Fig. 2) and is identical to the core sequence of the SRE in the ANF promoter (40). These A/T-rich sequences are also conserved in the human FGF-2 gene (34, 38). Additional sequences conserved between the human and rat FGF-2 promoters are binding sites for Egr-1 (2, 38). Egr-1 binds to the consensus sequence 5'-GCGGGGGCG-3' (Fig. 2) and was shown to regulate FGF-2 transcription in astrocytes (2). Because protein kinase C (PKC) activation is known to upregulate Egr-1 in multiple cell types (6, 27, 37, 43), and because ₁-adrenergic-receptor effects are mediated through PKC (41), the Egr-1 elements might also be involved in the observed response of the FGF-2 promoter to phenylephrine treatment.

Although cardiac myocyte contraction can be stimulated through both ₁- and -adrenoceptors (8, 41), our data indicate that adrenergic stimulation of FGF-2 promoter activity is not dependent on stimulation of contraction. Arrest of contraction using high extracellular KCl or 2,3-butanedione monoxime did not interfere with the increase in FGF-2 promoter activity after norepinephrine treatment (Fig. 4A). Interestingly, when KCl was used alone, a significant 1.5-fold increase in FGF-2 promoter activity was observed (not shown), raising the possibility that changes in intracellular Ca²⁺ may be a component of the response to adrenergic stimulation. The slow, permanent depolarization of the sarcolemma by high extracellular KCl would cause a corresponding permanent increase in intracellular Ca²⁺. However, blocking of the major (L-type) Ca²⁺ channels in cardiac myocytes with nifedipine had no effect on phenylephrine-induced FGF-2 promoter activity (Fig. 4B). This does not rule out the possible contribution of Ca²⁺ via minor channel types and intracellular stores.

In addition, we used transgenic mice to show that FGF-2 synthesis can be regulated at the transcriptional level by ₁-adrenergic stimulation in vivo (Fig. 5). The rat FGF-2 promoter, like its human counterpart, possesses properties associated with a housekeeping gene, and its product is found in all tissues studied (3, 16, 28, 42). This was reflected in the detection of luciferase activity in both brain and heart of transgenic mice (Fig. 5A). The increase in endogenous mouse FGF-2 RNA levels observed 24 h after administration of phenylephrine is consistent with the stimulation of FGF-2 promoter activity observed at this time and preceding this event at 6 h (Fig. 5, B and C). The loss of increased FGF-2 promoter activity as well as the decrease in FGF-2 RNA levels at 48 h likely reflects metabolism and clearance of the phenylephrine and a corresponding reduction in adrenergic stimulation. Regardless, the accumulation of FGF-2 RNA and, more specifically, the stimulation of FGF-2 promoter activity via ₁-adrenoceptors in the transgenic mice indicate a role for this regulatory pathway in vivo. Furthermore, the transfection and transgenic mouse data (Figs. 3 and 5) suggest that the 1,058/+54 region of the rat FGF-2 gene contains sufficient information to allow adrenergic regulation of the FGF-2 promoter in vitro and in vivo.

FGF-2 has many properties that make it a candidate for maintaining a healthy myocardium as well as offering protection against injury. FGF-2 enhances de novo angiogenesis (21) and promotes cell survival (11). Also, FGF-2 was shown to be protective against free radical damage in isolated cardiac myocytes (31), and, as shown with the use of a model of ischemia-reperfusion, FGF-2 improved recovery of function of the intact myocardium (31, 32). Clearly, an increase in chronic levels of endogenous FGF-2 in the heart would potentially limit the extent of damage and improve recovery from an ischemic episode. Our data show that endogenous FGF-2 synthesis in the heart can be regulated at the transcriptional level by adrenergic stimulation.