Fibroblast growth factors (Fgfs) are polypeptide growth factors with diverse biological activities. Fgfs widely expressed in developing and adult tissues play important roles in development and metabolism. The human/mouse Fgf family comprises 22 members. Most Fgfs except for Fgf11–Fgf14 bind to and activate Fgf receptors on the cell surface, resulting in the activation of several cytoplasmic signal transduction pathways. They function in a paracrine or endocrine manner (Itoh and Ornitz,2004, 2008; Thisse and Thisse,2005).
Fgf16 was originally identified in the heart by a homology-based polymerase chain reaction (PCR; Miyake et al.,1998). Fgf16 was efficiently secreted by an uncleaved bipartite signal sequence (Miyakawa and Imamura,2003). Among postnatal tissues, Fgf16 was predominantly expressed in the heart (Miyake et al.,1998; Sontag and Cattini,2003). In addition, Fgf16 was abundantly expressed in the embryonic brown adipose tissue (Miyake et al.,1998; Konishi et al.,2000). However, the physiological roles of Fgf16 remain unclear.
To elucidate the physiological roles of Fgf16, we generated Fgf16 knockout mice by homologous recombination. The mice were apparently normal and fertile. However, a slightly decreased number of cardiomyocytes was observed in the postnatal heart. In addition, the proliferation of embryonic cardiomyocytes was significantly decreased. In this study, we report that Fgf16 is a growth factor playing a role in the proliferation of embryonic cardiomyocytes.
Expression of Fgf16 in Mice
We examined the expression of Fgf16 in mouse tissues at postnatal day (P) 56 by reverse transcription-PCR (RT-PCR; Fig. 1a). Fgf16 was predominantly expressed in the heart among the major tissues examined. In addition, we examined the expression of Fgf16 in the ventricles and atria of the heart by RT-PCR. Fgf16 was expressed in the ventricles but not in the atria (Fig. 1b). We also examined the expression of Fgf16 in mouse embryos at embryonic day (E) 14.5 and E18.5 by whole-mount in situ hybridization. However, clear signals for Fgf16 expression could not be detected in the embryos (data not shown). In addition, we also examined the expression of Fgf16 in cardiomyocytes and noncardiomyocytes, which were prepared from the hearts of neonatal mice, by RT-PCR (Fig. 1c). As α-MHC is preferentially expressed in cardiomyocytes (Lyons et al.,1990), we also examined the expression of α-MHC as a control. Fgf16 as well as α-MHC was predominantly expressed in cardiomyocytes.
Targeted Disruption of Fgf16
To address roles of Fgf16 in mice, we generated Fgf16 knockout mice. The mouse Fgf16 gene, the coding region of which is divided into three exons, exons 1–3, is located on chromosome X (Itoh and Ornitz,2008; Fig. 2a). We replaced exons 2 and 3 with an IRES-LacZ-polyA/PGK-neo cassette in ES cells. Homologous recombination was confirmed by Southern blot analysis using a 3′ probe. The wild-type and mutant alleles resulted in 5.4-kbp and 4.4-kbp EcoRI-digested DNA fragments, respectively (data not shown). PCR genotyping of mice using primers specific for the wild and mutant loci resulted in 509-bp and 273-bp DNA fragments, respectively (Fig. 2b). Fgf16 expression in the heart at P56 was examined by RT-PCR using primers specific for exons 1 and 3. The Fgf16 cDNA of expected size was detected in the wild-type heart not but the Fgf16 knockout heart (Fig. 2c). Mating of wild-type male mice (Fgf16X/Y) with heterozygous female mice (Fgf16X/−) resulted in offspring of four genotypes (Fgf16X/Y, Fgf16X/X, Fgf16X/−, and Fgf16−/Y) at normal Mendelian ratios (Fig. 2d). Fgf16 knockout mice were fertile.
Analysis of Fgf16 Knockout Mice
As Fgf16 was abundantly expressed in the heart, we examined the heart in Fgf16 knockout male mice at 6 months of age. The Fgf16 knockout heart was apparently normal (Fig. 3a). However, as it appeared to be slightly smaller than the wild-type heart, we quantitatively examined it at 6 months of age. Body weight and the weights of major tissues including the liver, kidney, and lung were essentially unchanged in Fgf16 knockout mice (Fig. 3b and data not shown). In contrast, heart weight in Fgf16 knockout mice was slightly but significantly decreased (Fig. 3b). We also examined the number of cardiomyocytes in the Fgf16 knockout heart at 6 months of age by morphometric analysis using isolated cardiomyocytes. The number of cardiomyocytes in Fgf16 knockout mice was also slightly but significantly decreased (Fig. 3b). We also examined the size of cardiomyocytes by morphometric analysis using isolated cardiomyocytes (Fig. 3b). The size of cardiomyocytes in the Fgf16 knockout heart was essentially unchanged. These results indicate that the decrease in heart size was caused mainly by the reduction in cardiomyocyte cell numbers. We also examined the heart by histological analysis with hematoxylin and eosin staining. Upon microscopic observation, cardiomyocytes were apparently normal in the Fgf16 knockout heart. Necrosis or cardiomyocyte disarray was not observed (Fig. 3c).
Proliferation and Survival of Cadiomyocytes in Fgf16 Knockout Mice
As cardiomyocytes mostly proliferate at embryonic stages, their number in the heart was mainly determined by cell proliferation during embryonic development (Pasumarthi and Field,2002). As Fgf16 knockout mice indicate a potential role for Fgf16 in the proliferation of cardiomyocytes at embryonic stages, we examined the proliferation at E14.5, E16.5, and E18.5 by determining the incorporation of bromodeoxyuridine (BrdU) into cardiomyocytes. The proliferation of cardiomyocytes of wild-type mice at E14.5 was highest (Fig. 4a). However, after E14.5, the proliferation gradually decreased as reported (Pasumarthi and Field,2002). Although the proliferation was slightly decreased in both the septum and the ventricle wall of Fgf16 knockout mice at E16.5 and E18.5, the proliferation was significantly decreased at E14.5 (Fig. 4a). These results indicate that Fgf16 is required for the proliferation of cardiomyocytes during embryonic development. However, the transient reduction of the proliferation at E14.5 might result in a small difference in cardiomyocyte cell numbers.
Fgf signaling is also involved in cell survival in several tissues during development (Orinitz and Itoh,2001; Thisse and Thisse,2005). Therefore, we also examined the survival of cardiomyocytes in the Fgf16 knockout heart by conducting a terminal deoxynucleotidyl transferase–mediated deoxyuridinetriphosphate nick end-labeling (TUNEL) assay. Few TUNEL-positive cells were detected at E18.5 or P56 (Fig. 4b). These results indicate that Fgf16 is not required for cell survival in the heart.
Fgf9, Fgf16, and Fgf20, members of the Fgf9/16/20 subfamily, potentially share similar biochemical properties (Ornitz and Itoh,2001; Zhang et al.,2006). Fgf9 is involved in the proliferation of cardiomyocytes at embryonic stages (Lavine et al.,2005). As the number of cardiomyocytes in Fgf16 knockout mice was significantly reduced, we also examined the expression of Fgf16 as well as Fgf9 and Fgf20 in the heart at different developmental stages by RT-PCR (Fig. 1d,e). Fgf9 expression was essentially unchanged in the heart at all embryonic stages examined, and thereafter slightly increased at postnatal stages. Fgf16 expression was weakly detected in the heart at E14.5, and thereafter gradually increased. However, Fgf20 expression was not detected in the heart at any developmental stage examined.
Cardiac Marker Gene Expression in Fgf16 Knockout Heart
As described above, no obvious abnormality was observed in the Fgf16 knockout heart by histological analysis. To further examine the role of Fgf16 in the heart, we determined the expression of cardiac marker genes in the Fgf16 knockout heart at E18.5 and 6 months of age by RT-PCR. The expression of α-myosin heavy chain (α-MHC), myosin light chain 2V (MLC2V), (Lyons et al.,1990), atrial natriuretic peptide (ANP; Argentin et al.,1994), and brain natriuretic peptide (BNP; Nakagawa et al.,1995) was essentially unchanged at E18.5 (Fig. 5a), indicating that Fgf16 did not affect the differentiation of embryonic cardiomyocytes. The expression of α-MHC, MLC2V, and ANP was also essentially unchanged at 6 months of age (Fig. 5b). In contrast, the expression of BNP was significantly decreased.
BNP is secreted from the cardiac ventricle (Ogawa et al.,1990). BNP has been extensively studied for its predictive value in patients with heart failure. Focal fibrotic lesions were observed in the heart of BNP knockout mice (Tamura et al.,2000). Therefore, we examined the Fgf16 knockout heart for fibrosis by conducting Masson trichrome staining. However, no cardiac fibrosis was found in the ventricles (Fig. 5c).
Fgf16 Knockout Heart Function
As described above, heart weight and the number of cardiomyocytes in Fgf16 knockout mice were slightly but significantly reduced. These results indicate heart function to be impaired in Fgf16 knockout mice. We examined tail-cuff systolic blood pressure at 6 months of age (Table 1). Blood pressure was essentially unchanged in Fgf16 knockout mice. We also examined cardiac performance at 6 months of age by echocardiography (Table 1). Heart rate, fractional shortening (FS), and the ejection fraction (EF) were essentially unchanged in the Fgf16 knockout heart. The diastolic interventricle septal wall thickness (IVSd) and diastolic left ventricle posterior wall thickness (LVPWd) were also essentially unchanged. However, the diastolic left ventricle internal dimension (LVDd) was slightly but significantly decreased. This is consistent with the finding that the Fgf16 knockout heart was slightly but significantly smaller than the wild-type heart.
Table 1. Blood Pressure and Echocardiographic Parameters in Fgf16 Knockout Mice
Results are expressed as means ± SEM for X/Y (n = 6) and −/Y (n = 7) mice at 6 months of age. EF, ejection fraction; FS, fractional shortening; IVSd, diastolic interventricle septal wall thickness; LVPWd, diastolic left ventricle posterior wall thickness, LVDd; diastolic left ventricle internal dimension. X/Y, wild-type male; −/Y; knockout male. An asterisk indicates statistical significance compared with the wild-type (*p < 0.07).
Systolic blood pressure (mmHg)
101.7 ± 5.0
97.3 ± 3.4
Diastolic blood pressure (mmHg)
52.5 ± 2.3
51.0 ± 2.2
Heart Rate (beats/min)
635 ± 19
625 ± 29
37.00 ± 0.86
40.57 ± 1.09
75.17 ± 0.91
79.00 ± 1.25
0.92 ± 0.06
0.87 ± 0.05
0.85 ± 0.04
0.87 ± 0.04
3.63 ± 0.13
3.27 ± 0.12*
Most Fgf genes have been disrupted in mice. The phenotypes of Fgf knockout mice indicate that Fgf signaling plays crucial roles in development and metabolism (Itoh and Ornitz,2008). Although the expression profile of Fgf16 and the activity of recombinant Fgf16 in vitro suggest potential roles for Fgf16 in the heart and brown adipose tissue (Miyake et al.,1998; Konishi et al.,2000; Sontag and Cattini,2003), the actual roles in vivo remain unclear. In this study, we examined roles of Fgf16 in vivo by generating Fgf16 knockout mice by homologous recombination. Although most tissues examined in these mice were essentially normal, heart weight and the number of cardiomyocytes were slightly but significantly decreased.
Cardiomyocytes mostly proliferate at embryonic stages (Pasumarthi and Field,2002). During the postnatal period, cardiomyocytes stop proliferating and increase in size. Postnatal cardiac growth is mediated by cardiomyocyte hypertrophy (Li et al.,1996). Several studies using culture models suggest that Fgf signaling is involved in these developmental events. For example, exogenous Fgf2 promoted the proliferation of cardiomyocytes in culture (Kardami,1990). However, heart development was not impaired in Fgf2 knockout mice (Zhou et al.,1998). In contrast, Fgf9, which is a member of the Fgf9/16/20 subfamily, plays a role in the proliferation of embryonic cardiomyocytes in vivo (Lavine et al.,2005). Fgf9 knockout mice died shortly after birth (Colvin et al.,2001). The embryonic heart of Fgf9 knockout mice was slightly smaller than that of wild-type mice. The proliferation of cardiomyocytes was significantly decreased in the Fgf9 knockout heart at embryonic stages. The phenotype is similar to that of the Fgf16 knockout heart. Furthermore, both recombinant Fgf9 and Fgf16 had a proliferative effect on heart explants in vitro (Lavine et al.,2005). In addition, Fgf9 and Fgf16 potentially share similar biochemical properties (Ornitz and Itoh,2001; Zhang et al.,2006) and their expression profiles in the embryonic heart were similar. These findings suggest Fgf9 and Fgf16 to synergistically promote the proliferation of embryonic cardiomyocytes. We also examined the expression of Fgf9 in the Fgf16 knockout heart at E18.5. However, the expression was essentially unchanged (data not shown).
Fgf signaling is also thought to be involved in the acquisition of cardiac fate. For example, Fibroblast growth factor receptor 1 (Fgfr1) is essential for the development of cardiomyocytes in vitro (Dell'Era et al.,2003). The disruption of both Fgfr1 and Fgfr2 in the embryonic mouse heart resulted in cellular hypertrophy (Lavine et al.,2005). Therefore, we examined the differentiation of cardiomyocytes in the embryonic Fgf16 knockout heart. However, the expression of cardiac markers, α-MHC, MLC2V, and ANP, was essentially unchanged, indicating that Fgf16 is not required for the differentiation of embryonic cardiomyocytes. In addition, no obvious abnormality was observed in the Fgf16 knockout heart by histological analysis.
Fgf16 was also expressed in the postnatal heart, indicating potential roles for Fgf16. The decreased heart weight and the decreased number of cardiomyocytes in Fgf16 knockout mice indicate heart function to be impaired. However, blood pressure and cardiac performance were essentially normal in Fgf16 knockout mice, indicating that heart function was not impaired.
We also examined the expression of cardiac marker genes in the postnatal Fgf16 knockout heart. Although the expression of α-MHC, MLC2V, and ANP was essentially unchanged, the expression of BNP was significantly decreased. Focal fibrotic lesions were observed in BNP knockout mice (Tamura et al.,2000). However, cardiac fibrosis was not found in the ventricles of any Fgf16 knockout mice examined. Therefore, Fgf16 is not involved in cardiac fibrosis under physiological conditions. On the other hand, BNP expression is increased in the heart in response to pressure overload (Nakagawa et al.,1995). In the heart of BNP knockout mice, multifocal fibrotic lesions were significantly increased in size and number in response to ventricular pressure overload (Tamura et al.,2000). In addition, although the development of the heart was almost normal, Fgf2 knockout mice developed significantly less hypertrophy than wild-type mice in response to pressure overload (Schultz et al.,1999). Therefore, Fgf16 might play roles in the heart under pathological conditions, including pressure overload and myocardial infarction.
In conclusion, Fgf16, which is expressed in cardiomyocytes, is a growth factor for embryonic cardiomyocytes, and may synergistically act with Fgf9.
Expression of Fgf9, Fgf16 and Fgf20 in Adult Tissues and Embryonic Heart Examined by RT-PCR
Total RNA was extracted from mouse tissues, cardiomyocytes, or noncardiomyocytes using an RNeasy mini kit (Qiagen). Cardiomyocytes and noncardiomyocytes were prepared from mouse neonatal hearts as described (Nakagawa et al.,1995). cDNA was synthesized from the RNA (1 μg) as a template in a reaction mixture containing moloney murine leukemia virus reverse transcriptase (Gibco BRL) and a random hexadeoxynucleotide primer (Takara, Japan). The cDNA was amplified by PCR with Taq DNA polymerase (Wako, Japan) and primers specific for Fgf9 (sense primer, 5′-GTC CTC TGA TGG CTC CCT TA-3′; antisense primer, 5′-AGA CAC TGT CTT TGT CAG CTT-3′), Fgf16 (sense primer, 5′-CCG CTT CGG AAT TCT GGA AT-3′; antisense primer, 5′-GGA CAT GGA GGG CAA CTT AGA A-3′) or Fgf20 (sense primer, 5′-CCA TGG CTC CCT TGA CCG AA-3′; antisense primer, 5′-GGC TCT AGA TTC ATC AAG TG-3′). As a control, the expression of mouse 18S rRNA was also examined by PCR using primers specific for 18S rRNA (sense primer, 5′-CTT AGA GGG ACA AGT GGC G-3′; antisense primer, 5′-ACG CTG AGC CAG TCA GTG TA-3′). The PCR product was separated by electrophoresis on a 2% agarose gel, visualized by ethidium bromide staining, and quantified with Image J software.
Mouse Fgf16 gene fragments, a 7.6-kbp fragment for the 5′ homology recombination arm and a 1.6-kbp fragment for the 3′ homology recombination arm, were amplified from the genomic DNA of 129 mouse embryonic stem (ES) cells as a template by PCR with KOD+ DNA polymerase (TOYOBO). A targeting vector was constructed by ligation of the fragments, the 5′ and 3′ homology recombination arms and a 6.3-kbp fragment for an IRES-LacZ-polyA/PGK-neo cassette. A diphtheria toxin A (DTA) expression cassette was inserted at the 5′ end of the targeting vector (Ohbayashi et al.,2002). The coding region of the mouse Fgf16 is divided into three exons, exons 1–3. Most of exon 2 and all of exon 3 of Fgf16 were replaced with the IRES-LacZ-polyA/PGK-neo cassette. The targeting vector was linearized with NotI and electroporated into C57BL/6 ES cells. The selection in G418 produced five homologous recombinant ES cell clones that were confirmed by Southern blot analysis using a 3′ probe. Mouse Fgf16 is located on chromosome X. Germ-line chimeras were produced by the simple aggregation method (Wood et al.,1993) with Fgf16 disrupted ES (−/Y) cells and morulae isolated from 129 Sv mice. Male chimeras were mated with C57BL/6 females. All mice were housed in a temperature-controlled environment with a 12-hr light/dark cycle.
Genotyping of Mice
Genotypes of mice were determined by PCR using the following primers: P1 (5′-GTC TTG CCT CAC AAT CTA CC-3′), P2 (5′-CCC GTG ATA TTG CTG AAG AG-3′), and P3 (5′-TGG CCA GCC TCT TCA TTC TA-3′). P1 and P3 produced a 509-bp fragment of the wild-type Fgf16 locus. P2 and P3 produced a 273-bp fragment of a mutant Fgf16 locus. The sex was determined by PCR using primers specific for chromosome Y (sense primer, 5′-CGC CCT TTA ATA TCG AAT CAC-3′; antisense primer, 5′-TCC AGT TCA TTT AGC CTC TGA-3′).
The heart at 6 months of age was fixed overnight in 4% paraformaldehyde, dehydrated, embedded in paraffin, and sectioned at 6 μm. Sections were stained with hematoxylin and eosin or with Masson trichrome and examined by light microscopy.
Morphometric Measurement of Isolated Cardiomyocytes
Paraformaldehyde-fixed hearts (the atria were removed) were digested with 12.5 M KOH for 20 hr as previously described (Gerdes et al.,1998). After careful washing with phosphate-buffered saline, rod-shaped cells, cardiomyocytes, were examined by light microscopy, and counted using a hemocytometer. Images of cardiomyocytes were captured and cell areas were measured for at least 50 cells per mouse using ImageJ software.
Time-mated female embryos were injected IP with BrdU (100 μg / g body weight) 1 hr before killing. BrdU immunohistochemistry on sections was performed as described (Naski et al.,1998). Sections were counterstained with hematoxylin. The number of BrdU-positive nuclei relative to the total number of nuclei was counted.
The heart at E18.5 and P56 was fixed. Sections of the heart were prepared as described (Naski et al.,1998). Apoptotic cells on the sections were detected using a DeadEnd Colorimetric TUNEL System kit (Promega).
Expression of Genes in Heart Examined by Quantitative RT-PCR
Total RNA was prepared from mouse heart at E18.5 and 6 months of age using an RNeasy mini Kit (Qiagen). The expression of genes was examined by RT-PCR with Taq DNA polymerase (Wako, Japan) and primers specific for mouse α-myosin heavy chain (MHC; sense primer, 5′-AGA TGG CTG ACT TCG GGG CAG-3′; antisense primer, 5′-CAT GGC CAT GTC CTC GAT CTT GT-3′), myosin light chain 2V (MLC2V; sense primer, 5′-TGT TCC TCA CGA TGT TTG GG-3′; antisense primer, 5′-CTC AGT CCT TCT CTT CTC CG-3′), atrial natriuretic peptide (ANP) (sense primer, 5′-CGG TGT CCA ACA CAG ATC TG-3′; antisense primer, 5′-AAG CTG TTG CAG CCT AGT CC-3′), brain natriuretic peptide (BNP; sense primer, 5′-GAT CTC CTG AAG GTG CTG TCC-3′; antisense primer, 5′-ATC CGG TCT ATC TTG TGC CCA-3′), and 18S rRNA. The PCR product was separated by electrophoresis on a 2% agarose gel, visualized by ethidium bromide staining, and quantified with Image J software. As a control, the expression of mouse 18S rRNA was also examined by RT-PCR using primers specific for 18S rRNA.
Blood Pressure Measurement
The blood pressure of mice at 6 months of age was measured by the tail-cuff method (Nakanishi et al.,2007). At least 20 readings were taken for each measurement, and each mouse was measured 5 times for 2 weeks.
Mice at 6 months of age were examined by conscious ecocardiography. During the echocardiography, the animals were restrained by grasping the skin on the back of the neck and wrapping the tail (Xu et al.,2007). Diastolic interventricle septal wall thickness (IVSd), diastolic left ventricle internal dimension (LVDd), diastolic left ventricle posterior wall thickness (LVPWd), fractional shortening (FS), ejection fraction (EF), and heart rate were calculated using the echocardiographic system (Toshiba Power Vision 8000) equipped with a 12-MHz imaging transducer (Nakanishi et al.,2007).
Results are expressed as the mean ± standard error of measurement (SEM). The statistical significance of differences in mean values was assessed with Student's t-test.
M.K and N.I. were funded by the Ministry of Education, Science, Culture, Sports, and Technology, Japan and N.I. was funded by the Takeda Science Foundation and the Mitsubishi Foundation.