Expression of Foxm1 Transcription Factor in Cardiomyocytes Is Required for Myocardial Development

Forkhead Box M1 (Foxm1) is a transcription factor essential for organ morphogenesis and development of various cancers. Although complete deletion of Foxm1 in Foxm1−/− mice caused embryonic lethality due to severe abnormalities in multiple organ systems, requirements for Foxm1 in cardiomyocytes remain to be determined. This study was designed to elucidate the cardiomyocyte-autonomous role of Foxm1 signaling in heart development. We generated a new mouse model in which Foxm1 was specifically deleted from cardiomyocytes (Nkx2.5-Cre/Foxm1fl/f mice). Deletion of Foxm1 from cardiomyocytes was sufficient to disrupt heart morphogenesis and induce embryonic lethality in late gestation. Nkx2.5-Cre/Foxm1fl/fl hearts were dilated with thinning of the ventricular walls and interventricular septum, as well as disorganization of the myocardium which culminated in cardiac fibrosis and decreased capillary density. Cardiomyocyte proliferation was diminished in Nkx2.5-Cre/Foxm1fl/fl hearts owing to altered expression of multiple cell cycle regulatory genes, such as Cdc25B, Cyclin B1, Plk-1, nMyc and p21cip1. In addition, Foxm1 deficient hearts displayed reduced expression of CaMKIIδ, Hey2 and myocardin, which are critical mediators of cardiac function and myocardial growth. Our results indicate that Foxm1 expression in cardiomyocytes is critical for proper heart development and required for cardiomyocyte proliferation and myocardial growth.


Introduction
The heart is the first organ to function during embryonic development, the beating heart can be detected as early as embryonic day 8 (E8) in the mouse [1,2]. Proper cardiac development requires strict adherence to a temporal and spatial pattern of gene expression. Embryonic development of the heart is mediated by proliferative growth, with cardiomyocytes rapidly progressing through the cell cycle and multiplying [3]. In the postnatal period, cardiomyocytes withdrawal from the cell cycle and cardiac growth becomes dependent on hypertrophy of individual cardiomyocytes [3]. Transcriptional regulation of cardiomyocyte proliferation during embryogenesis has been extensively studied, and several cardiac transcription factors were found to be critical for cardiomyocyte progression into the cell cycle. These include GATA family members 4 and 6 [4], myocardin [5], Twist family members 1 [6] and 2 [7], Hey2 [8,9], Sox4 [10] and Nkx2.5 [11].
Foxm1 (previously known as HFH-11B, Trident, Win, or MPP2) is a member of the Forkhead Box (Fox) family of transcription factors which share homology in the Winged Helix/Forkhead DNA binding domain. Foxm1 is expressed in proliferating cells of all embryonic tissues, including cardiac progenitor cells and the early myocardium [12,13]. However, expression wanes postnatally and Foxm1 can only be detected in a few adult tissues such as intestinal crypts, thymus and testis [14,15]. Foxm1 signaling has been shown to be a critical mediator of both G 1 -S and G 2 -M transitions of the cell cycle, and to be upregulated in various human cancers [16,17,18,19,20,21]. In addition, Foxm1 was determined to play a role in tissue repair following injury in the lungs and liver [15,22,23].
Foxm1-null (Foxm1 2/2 ) mouse lines have been previously generated and characterized by two separate labs [24,25]. Foxm1 2/2 mice in which the DNA binding and C-terminal transcriptional activation domains of the Foxm1 protein were deleted die in utero between E13.5 and E16.5 due to multiple abnormalities in various organ systems, including liver, lungs, blood vessels, brain and heart [13,25,26,27]. Although these studies showed that Foxm1 plays a cell autonomous role for organ development in multiple cell types, the role of Foxm1 in cardiac development and function remains unknown. Given widespread organ defects in Foxm1 2/2 mice, it remains unclear whether Foxm1 is critical for heart development or if cardiac abnormalities are secondary to defects in other organ systems which could alter embryonic growth. Therefore, a direct role of Foxm1 in cardiomyocyte growth and/or function awaits elucidation.
As Foxm1 is widely expressed during embryogenesis [12,28,29], the recent focus has been to elucidate the cell-specific roles of Foxm1 in different tissues using conditional knockout mouse models. Specific deletion of Foxm1 from hepatoblasts resulted in embryonic lethality around day E18.5 with disruption of hepatic cords and vasculature, as well as a lack of intrahepatic bile ducts [25]. Deletion of Foxm1 from precursors of cerebellar granule neurons interfered with Shh-induced signaling to delay brain development [30]. Foxm1 deletion from T lymphocyte lineage decreased proliferation of early thymocytes and activated mature T cells without affecting apoptosis or T cell differentiation [31]. However, mice with endothelial-or macrophage-specific Foxm1 deletions developed normally [32,33], indicating Foxm1 is dispensable in these cells lines during embryogenesis. Furthermore, while deletion of Foxm1 specifically from the pancreas did not affect pancreatic development [34], male mice developed islet dysfunction and diabetes resulting from impaired postnatal b-cell mass expansion [34] and females were prone to gestational diabetes [35], indicating Foxm1 requirements differ during pancreatic development. Deletion of Foxm1 specifically from smooth muscle cells did not affect differentiation, but mice died immediately after birth from severe pulmonary hemorrhage, structural defects in the arterial wall and esophageal abnormalities [28]. When Foxm1 was deleted conditionally in developing respiratory epithelium proliferation rates of respiratory epithelial cells were unaltered [12], suggesting that Foxm1 is not required for epithelial proliferation during lung development. However, deletion of Foxm1 from respiratory epithelium impaired lung maturation, decreased expression of surfactant-associated proteins SPA, SPB and SPC and delayed differentiation of type I cells from epithelial precursors causing respiratory failure after birth [12]. Thus, Foxm1 is essential for surfactant homeostasis and lung maturation during lung development. Studies in conditional Foxm1 knockout models have shown that Foxm1 plays unique roles in different tissues during embryonic development; the cardiomyocyte-specific role of Foxm1 in heart development remains unexplored.
In this study, we utilized the Cre-LoxP system to conditionally delete Foxm1 from cardiomyocytes to ascertain the cardiomyocyte-autonomous role of Foxm1 in heart development. Deletion of Foxm1 from cardiomyocytes resulted in chamber dilation and myocardial thinning, culminating in embryonic lethality in late gestation. Cardiac Foxm1 deletion caused decreased cardiomyocyte proliferation and altered expression of cell cycle regulators Cdc25B, Cyclin B 1 , nMyc, Plk-1 and p21 cip1 . We also identified CaMKIId, Hey2 and myocardin as new potential targets of Foxm1 signaling and mediators of myocardial thinning. This study shows that Foxm1 is critical for expression of cell cycle regulatory genes in developing cardiomyocytes and is required for proper heart development.

Foxm1 conditional knockout mice
We have previously described the generation of Foxm1 LoxP/ LoxP (Foxm1 fl/fl ) mice, in which LoxP sequences flank exons 4 through 7 of the Foxm1 gene encoding the DNA binding and transcriptional activation domains [25]. Foxm1 fl/fl mice were bred with Nkx2.5-Cre mice [36] to generate Nkx2.5-Cre/Foxm1 fl/fl double transgenic mice with deletion of Foxm1 from the myocardium. Using lineage tracing experiments previous studies demonstrated that Nkx2.5-Cre was expressed in the early myocardium as well as epithelium of the first pharyngeal arch [36]. However, no gross morphological abnormalities were observed in any non-cardiac tissues examined including thyroid and thymus (data not shown). Nkx2.5-Cre/Foxm1 fl/fl homozygous embryos exhibited an embryonic lethal phenotype with the exception of one mouse from an early breeding which survived to postnatal day 11 (P11). To generate Nkx2.5-Cre/Foxm1 fl/fl embryos, Nkx2.5-Cre/Foxm1 fl/+ heterozygous males were bred with Foxm1 fl/fl females. Foxm1 fl/fl or Nkx2.5-Cre/Foxm1 fl/+ embryos from the same litter were used as controls. Animal studies were reviewed and approved by the Animal Care and Use Committee of Cincinnati Children's Hospital Research Foundation.

Quantitative real-time RT-PCR (qRT-PCR)
Total cardiac RNA was prepared from individual Nkx2.5-Cre/ Foxm1 fl/fl and control hearts using RNA-STAT-60 (Tel-Test ''B'' Inc. Friendswood, TX). cDNA was generated using the Applied Biosystems High Capacity cDNA Reverse Transcription kit (Applied Biosystems, Foster City, CA). Evaluation of expression levels of specific genes was performed by qRT-PCR using Taqman probes (Table 1) and StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA) as previously described [12,20,28].

Western blot analysis
Hearts from E14.5 embryos were harvested and used to prepare protein extract. Three hearts from embryos with matching genotypes were pooled. Protein extracts were run on PAGE-SDS gels and transferred to PVDF membranes followed by incubation with primary antibodies specific for Foxm1 (1:1000; C20; Santa Cruz), Cyclin B 1 (1:500; BD Pharmingen), or p21 cip1 (1:300; BD Pharmingen). Secondary antibodies were conjugated with horseradish peroxidase. b-actin was used as a loading control.  [37] and counterstained with nuclear fast red (red nuclei). Foxm1 was detected in cardiomyocytes (arrows) and endocardial cells (arrowheads) from the ventricles (E-H, M-P), interventricular septum (inset E-H), and atrial walls (I-L) throughout cardiac development although expression progressively waned. Foxm1 was also detected in valves at E14.5 and E17.5 (inset I-J) but not postnatally (inset K-L). Foxm1-positive nuclei were detected in the embryonic pericardium (inset M-N) and in the coronary vasculature (inset O-P) until P7. The number of Foxm1-positive nuclei decreased during gestation in cardiomyocytes but was unaltered in endocardial cells (Q). Mean6SEM was determined from 5 random sections at E14.5 and P7 with 3 hearts in each group. qRT-PCR of total heart RNA demonstrated a decrease in Foxm1 mRNA from E14.5 to P20 and Foxm1 mRNA was undetectable in the adult heart (R). Foxm1 expression was normalized to b-actin mRNA. Significant differences (p,0.05) were indicated by asterisk. ''N'' values were represented by boxes inside or above bars. Scale bars represent 100 mm in A-D insets, 200 mm in A-D, 50 mm in E-L (main and insets) and 10 mm in M-P (main and insets). doi:10.1371/journal.pone.0022217.g001

Statistical analysis
Student's T-test was used to determine statistical significance. P values ,0.05 were considered significant. Values for all measurements were expressed as mean6standard error of mean (SEM).

Results
Foxm1 protein and mRNA expression declines during embryonic and postnatal development Hearts were collected from wild type mice at multiple timepoints during cardiac development. To identify cells express-ing Foxm1, microtome sections of paraffin-embedded hearts were immunohistochemically stained using anti-Foxm1 antibodies. Foxm1 protein was highly abundant in the embryonic period and easily detectable in nuclei of cardiomyocytes and endothelial cells up to postnatal day 7 (P7) ( Figure 1A-H). Foxm1 staining was observed throughout the developing heart, in the ventricles ( Figure 1E-H and M-P), interventricular septum (insets of Figure 1E-H), atria ( Figure 1I-L), heart valves (insets of Figure 1I-L), pericardium (insets of Figure 1M-N) and vasculature (insets of Figure 1O-P). The percentage of Foxm1-positive cardiomyocytes decreased from 29% at E14.5 to 20% at P7 ( Figure 1Q) and nuclear Foxm1 staining was seldom observed in the adult heart as only 461% of cardiomyocytes were Foxm1positive (data not shown). In addition, hearts were used to isolate total RNA and Foxm1 mRNA expression was examined by qRT-PCR. Similar to Foxm1 protein expression, Foxm1 mRNA was most highly expressed in the early embryonic period. Foxm1 expression declined by 30% from E14.5 to E17.5, but maintained a similar level of expression until P2. Foxm1 mRNA levels decreased rapidly in the postnatal period from P2 to P20, and Foxm1 mRNA was undetectable in the adult heart (11 weeks) ( Figure 1R).

Cardiomyocyte-specific deletion of Foxm1 causes embryonic lethality
Previous studies demonstrated that complete deletion of Foxm1 (Foxm1 2/2 mice) caused embryonic lethality between E13.5 and E16.5 and presented with malformation of multiple organ systems [25]. To ascertain the cardiomyocyte-autonomous role of Foxm1 signaling in cardiac development, Foxm1 fl/fl mice (in which LoxP sites flank exons 4-7) were crossed with Nkx2.5-Cre mice to generate a mouse line in which the DNA binding and transcriptional activation domains of the Foxm1 protein were excised in cardiomyocytes (Figure 2A). Breeding pairs between Foxm1 fl/fl and Nkx2.5-Cre/Foxm1 fl/+ heterozygous mice were used to generate embryos with the Nkx2.5-Cre transgene and homozygous for the Foxm1 fl/fl allele (Nkx2.5-Cre/Foxm1 fl/fl ) at an expected ratio of 1:4. The majority of Nkx2.5-Cre/Foxm1 fl/fl embryos did not survive to birth (Table 2). Although a near Mendelian ratio was observed for Nkx2.5-Cre/Foxm1 fl/fl embryos at E14.5 (Table 2), between E14.5 and E17.5 nearly 50% of Nkx2.5-Cre/Foxm1 fl/fl embryos were lost, with the remaining 50% lost between E17.5 and birth (Table 2). Only one Nkx2.5-Cre/Foxm1 fl/fl pup survived to postnatal day 11 but exhibited severe growth retardation prior to harvest (data not shown). Thus, Foxm1 deletion from cardiomyocytes is sufficient to induce embryonic lethality.
Embryonic deletion of Foxm1 from cardiomyocytes causes myocardial thinning, ventricular hypoplasia and disorganization of the myocardium Hearts from Nkx2.5-Cre/Foxm1 fl/fl embryos exhibited abnormal cardiac morphology in comparison to Foxm1 fl/fl and Nkx2.5-Cre/ Foxm1 fl/+ littermate controls. The ventricular lumen was dilated and ventricular walls were thinner in Foxm1 deficient embryos compared to control embryos ( Figure 3A-L). Ventricular wall thickness was significantly decreased at E14.5 and E17.5 ( Figure 3E-L, Table 3). Thickness of the interventricular septum (IVS) was decreased by 42% and 62% at E14.5 and E17.5, respectively, ( Figure 3M-P, Table 3) and cardiomyocyte organization within the IVS was in disarray (Figure 3M-P). There was, however, no significant change in the overall size of the heart at these two embryonic timepoints ( Figure 3A-D). Furthermore, although an increased network of extracellular matrix was observed in valves from Nkx2.5-Cre/Foxm1 fl/fl mice, there was no overall effect on valve size by deletion of Foxm1 from cardiomyocytes ( Figure 3Q-X). In addition, there were no gross morphological changes in other organs known to delineate from Nkx2.5 expressing cells such as the thymus (data not shown).
Decreased cardiomyocyte proliferation in Nkx2.5-Cre/ Foxm1 fl/fl embryos Cardiac proliferation was examined in Nkx2.5-Cre/Foxm1 fl/fl and control embryos by immunohistochemical staining with antibodies against either phospho-histone 3 (PH3) or Ki-67. PH3-positive cardiomyocytes undergoing mitosis were observed in both Nkx2.5-Cre/Foxm1 fl/fl and control Foxm1 fl/fl mice at all timepoints studied ( Figure 4A-H). However, cardiomyocyte proliferation was significantly diminished in Nkx2.5-Cre/Foxm1 fl/fl hearts compared to littermate controls as indicated by significant decreases in the percentage of PH3-positive cardiomyocytes of 39% at E14.5 and 43% at E17.5 ( Figure 4I). Yet, there was no change in cellular proliferation within the lung or in endocardial cells at these same timepoints ( Figure 4I), indicating that proliferation defects were restricted to cardiomyocytes in Nkx2.5-Cre/Foxm1 fl/fl embryos. Consistent with decreased numbers of PH3-positive cardiomyocytes, Nkx2.5-Cre/Foxm1 fl/fl hearts displayed a dramatic reduction in Ki-67, a proliferation-specific protein ( Figure 4C-D). Furthermore, decreased mRNA expression of cell cycle regulators Cdc25B, Cyclin B 1 , Polo-like kinase 1 (Plk-1) and nMyc was found in E14.5 Nkx2.5-Cre/Foxm1 fl/fl hearts by qRT-PCR ( Figure 4M). There was no change in the expression of Cyclin D 1 or cMyc ( Figure 4M). In addition, Nkx2.5-Cre/Foxm1 fl/fl hearts displayed increased mRNA levels of p21 cip1 ( Figure 4M), a known cell cycle inhibitor critical for activity of cyclin-dependent kinase 2 (cdk2). Western blot analysis confirmed decreased protein levels of Foxm1 and Cyclin B 1 as well as increased p21 cip1 expression in Nkx2.5-Cre/Foxm1 fl/fl hearts ( Figure 4L). These results demonstrated that Foxm1 deletion from cardiomyocytes altered expression of cell cycle regulatory genes, contributing to proliferation defects and structural abnormalities in the developing heart.
Decreased capillary density and cardiac fibrosis in postnatal Nkx2.5-Cre/Foxm1 fl/fl hearts Microtome sections of paraffin-embedded hearts from Nkx2.5-Cre/Foxm1 fl/fl and control mice were subjected to immunohisto- chemical staining with antibodies against PECAM-1 or a-smooth muscle actin (aSM). PECAM-1 staining indicated a clear paucity in capillary density in the Nkx2.5-Cre/Foxm1 fl/fl mouse heart at P11 ( Figure 5A-B). Subsequent quantification showed a greater than 50% decrease in capillary density in the Nkx2.5-Cre/Foxm1 fl/fl heart compared to Foxm1 fl/fl control at P11 ( Figure 5I). Despite the decreased capillary density, there was no difference in the number or morphology of coronary vessels in Nkx2.5-Cre/Foxm1 fl/fl mouse
Expression of calcium/calmodulin-dependent kinase IId (CaM-KIId) was significantly reduced in Nkx2.5-Cre/Foxm1 fl/fl hearts ( Figure 6A). CaMKIId is an intracellular signaling molecule that is known to play a critical role in beat-to-beat cardiac physiology and has been shown to be an essential mediator of pressure overload and pharmacologically induced hypertrophy [37]. In addition, CaMKIId mediates calcium signaling and growth mechanisms within the cardiomyocyte [38]. Our results suggest that CaMKIId is a downstream effector of Foxm1 signaling as well as a contributing factor in myocardial thinning and embryonic lethality associated with cardiomyocyte-specific Foxm1 deletion.
Expression of both Hey2 and myocardin was reduced in Nkx2.5-Cre/Foxm1 fl/fl hearts ( Figure 6B). Hey2 knockout mice exhibited a thin walled left ventricle and decreased cardiomyocyte proliferation [8], similar to the phenotype in Nkx2.5-Cre/Foxm1 fl/fl mice. Myocardin is a cofactor for the serum response factor (SRF) and is essential for cardiogenesis [39,40]. Furthermore, SRF knockout mice exhibited a thin myocardium with dilated chambers and disorganized IVS [41], a phenotype similar to Nkx2.5-Cre/Foxm1 fl/ fl hearts. Therefore, decreased expression of Hey2 and myocardin may contribute to the cardiac abnormalities and embryonic lethality of the Nkx2.5-Cre/Foxm1 fl/fl mouse. Altogether, our results demonstrate that Foxm1 deletion from cardiomyocytes alters expression of cardiac genes critical for heart morphogenesis and cardiomyocyte proliferation.

Discussion
We have previously generated and characterized a Foxm1-null (Foxm1 2/2 ) mouse line in which mice die between embryonic days E13.5 and E16.5 due to severe defects in multiple organ systems including lungs, liver, heart and blood vessels [25]. We further demonstrated that tissue-specific deletion of Foxm1 from hepatocytes [25], respiratory epithelium [12] or smooth muscle cells [28] was sufficient to cause lethality in utero or shortly after birth. Therefore, Foxm1 is essential for organ morphogenesis in multiple organ systems. However, it remained to be determined whether cardiac malformation in Foxm1 2/2 embryos was due to cardiomyocyte derived effects of Foxm1 signaling or if these Foxm1 2/2 defects were indirect, resulting from abnormalities in other organ systems and altered embryonic homeostasis. To elucidate the cardiomyocyte-autonomous role of Foxm1 in embryonic heart development, we used the Cre-LoxP system to generate a conditional Foxm1 knockout mouse line in which Foxm1 is selectively deleted from cardiomyocytes under control of the Nkx2.5 promoter, one of the earliest known cardiac markers.
Deletion of Foxm1 from cardiomyocytes caused thinning and disorganization of the muscular walls of the heart, including both ventricles and the interventricular septum. Myocardial thinning was due to decreased cardiomyocyte proliferation accompanied by altered expression of multiple cell cycle regulatory genes. Ultimately, the myocardial hypoplasia in Nkx2.5-Cre/Foxm1 fl/fl embryos caused lethality in late gestation. Although the myocardial phenotype exhibited similarities to that observed in hearts from Foxm1 2/2 mice, many unique features were observed in the conditional knockout model suggesting cardiomyocyte-autonomous roles for Foxm1 signaling. These include a delay in the onset of lethality, unaltered heart size, diminished cardiac capillary density and myocardial fibrosis. The decrease in cardiomyocyte proliferation in Nkx2.5-Cre/Foxm1 fl/fl embryos was significantly less than in Foxm1 2/2 embryos suggesting that abnormalities in other cell types or tissues contributed to cardiac malformation in mice with complete deletion of Foxm1. Although we observed increased deposition of extracellular matrix in the atrio-ventricular valves of Nkx2.5-Cre/Foxm1 fl/fl hearts the size of the valves was unaltered. These results are in contrast to valve thickening in Foxm1 2/2 mice [13] and suggest that although Foxm1 does mediate atrioventricular valve formation, this signaling is not cardiomyocytedependent. Alternatively, valve defects Foxm1 2/2 mice may result from altered blood pressure caused by structural abnormalities in blood vessels that were previously reported [26].  To date, embryonic lethality associated with ventricular hypoplasia and myocardial thinning has been linked to several signaling cascades including the transcription factor Hey2 [8], members of the NFAT family [42] or inactivation of serum response factor (SRF) [41]. In this study we described a model of myocardial thinning owing to multiple factors and resulting in embryonic lethality. In addition to altered expression of various cell cycle regulatory genes, this study identified Hey2, myocardin and CaMKIId as novel targets of Foxm1 signaling in vivo and as potential mediators of the thin ventricular phenotype.
We previously showed decreased expression of NFATc3 in Foxm1 2/2 hearts and in Foxm1-depleted cardiomyocytes in vitro [13]. This study confirmed that Foxm1 is a positive regulator of cardiac NFATc3 expression and further identified cardiomyocytes as the cell type responsible for Foxm1-regulated NFATc3 expression in vivo. It has been previously shown that dual deletion of NFATc3 and NFATc4 causes thin ventricles, decreased proliferation of ventricular myocytes and pericardial effusion culminating in embryonic lethality [42]. Therefore, decreased NFATc3 expression could be a contributing factor in myocardial thinning and embryonic lethality associated with Nkx2.5-Cre/ Foxm1 fl/fl mice.
Decreased expression of Hey2 can contribute to the cardiac phenotype and embryonic lethality of Nkx2.5-Cre/Foxm1 fl/fl mice as evidenced by embryonic lethality and myocardial thinning in Hey2 2/2 mice, a phenotype similar to Nkx2.5-Cre/Foxm1 fl/fl mice. Hey2 has also been shown to interact with the serum response factor (SRF) to inhibit activity of myocardin [43], which is essential for cardiogenesis, cardiomyocyte proliferation, migration and deposition of the extracellular matrix [39,44,45]. Furthermore, deletion of SRF from cardiomyocytes resulted in lethality between E10.5-13.5 with thin myocardium, dilated chambers and disorganized IVS [41], a phenotype similar to that observed in the Nkx2.5-Cre/Foxm1 fl/fl hearts. Therefore, Foxm1 may directly influence myocardial development by decreasing expression of Hey2 and myocardin and possibly interfering with SRF-mediated signaling in cardiomyocytes.
CaMKIId deficiency caused augmented cardiac function in multiple heart injury models which manifested as severe alterations in cardiac structure. The finding of decreased CaMKIId mRNA in Nkx2.5-Cre/Foxm1 fl/fl hearts suggests a role for Foxm1 in regulating cardiac CaMKIId expression. Interestingly, interleukin-1b (IL-1b) mRNA was increased in Nkx2.5-Cre/ Foxm1 fl/fl hearts. IL-1b has been reported to be a critical mediator of cardiac fibrosis [46]. Since significant fibrosis was observed in the postnatal Nkx2.5-Cre/Foxm1 fl/fl heart, increased expression of IL-1b may contribute to fibrotic deposition.
In summary, we demonstrated that Foxm1 plays a cellautonomous role in cardiomyocytes during cardiac development. Foxm1 deletion in developing cardiomyocytes caused embryonic lethality, decreased cardiomyocyte proliferation, diminished vascular density in the myocardium and induced cardiac fibrosis in the early postnatal period. This study further identified Hey2, myocardin, NFATc3, CaMKIId and various cell cycle regulatory genes as in vivo targets of Foxm1 signaling and potential mediators of the myocardial thinning and ventricular hypoplasia associated with the Nkx2.5-Cre/Foxm1 fl/fl phenotype.  A-B), a-smooth muscle actin (aSM) (C-D) or Masson's Trichrome (E-H). Capillary density was decreased in Nkx2.5-Cre/Foxm1 fl/fl hearts at P11 (A-B, I), while coronary vessel formation was not altered (C-D). Numbers of capillaries were counted in PECAM-stained hearts using 10 random sections and mean6SEM was determined to confirm decreased capillary density (I). Significant fibrosis was detected in the IVS and ventricular walls of Nkx2.5-Cre/Foxm1 fl/fl hearts (F, H) while none was detected in control hearts (E, G). Significant differences (p,0.05) were indicated by asterisk. Scale bars represent 100 mm in A-F and 50 mm in G-H. doi:10.1371/journal.pone.0022217.g005 b-actin mRNA. Decreased gene expression of Foxm1, Cyclin B 1 , and increased p21 cip1 expression translated to changes in protein levels as demonstrated by Western blot analysis (L). b-actin was used as a loading control. Significant differences (p,0.05) were indicated by asterisk. ''N'' values were represented by boxes inside bars. Scale bars represent 50 mm in A-H and J-K and 10 mm in insets E-H. doi:10.1371/journal.pone.0022217.g004 Figure 6. Reduced mRNA expression of CaMKIId, NFATc3, Hey2 and myocardin in Nkx2.5-Cre/Foxm1 fl/fl hearts. Whole-heart RNA was isolated from embryonic Nkx2.5-Cre/Foxm1 fl/fl and control Foxm1 fl/fl hearts and used for qRT-PCR analysis. Taqman primers specific for CaMKIId, b-catenin, NFATc3, GATA4, GATA 6, Hey2, myocardin, Twist1, Twist2, Sox4, Notch1, Notch4, FGF10, IL-1b, TGF-b 1 , BMP2, BMP4, Wnt5a and Wnt7b (Table 1) were used to evaluate cardiac mRNA expression. Gene expression was normalized to b-actin mRNA. Significantly diminished mRNA levels of the intracellular signaling molecule CaMKIId and transcription factors NFATc3, Hey2 and myocardin were detected in Nkx2.5-Cre/Foxm1 fl/fl hearts. Increased mRNA expression of IL-1b was also detected. Significant differences (p,0.05) were indicated by asterisk. ''N'' values were represented by boxes inside bars. doi:10.1371/journal.pone.0022217.g006