The retinoblastoma gene product (Rb) is developmentally regulated in the myocardium, as shown by Western blot analysis. Little or none is expressed in midgestation in myocardium and Rb is the predominant pocket protein expressed in the adult. p107 has a reciprocal pattern of expression. p130 is expressed at all three ages. This is noteworthy because it had been postulated that functional differences between Rb and p107 and the developmental regulation of these proteins was responsible for cell cycle exit and the irreversible loss of cell cycling in post-mitotic skeletal muscle. That model is somewhat paradoxical, since Rb is a protein whose function is totally reversible upon protein phosphorylation. Work in their laboratory has sought a model in which irreversible cell cycle exit might be constructed by the developmental regulation of Rb, versus p107, in concert with other inhibitors, such as p21 and p27, abundant in the post-mitotic heart but expressed at low or negligible levels in cardiac proliferation.

The signaling cascade for mitogen signal transduction has been elucidated through their work with the adenoviral E1A (Fig.1). Mitogens activate D-type cyclins whose targets are CDK4 and CDK6 whose essential substrates are pocket proteins. Pocket protein phosphorylation then disinhibits E2F-dependent gene transcription. This results in activation of cyclins E and A and other E2F-induced genes, and the increase in CDK2 activity necessary for entry into S phase and DNA synthesis.

In contrast, forced cell cycle re-entry in cardiac myocytes, triggered by E2F-1 to bypass pocket proteins, remain sensitive to blocks to CDK2 activity, including p21 or dominant negative CDK2. This inactivates pocket proteins using E1A, resulting in DNA synthesis that, surprisingly, proved to be independent of an increase in CDK2 activity. S phase entry driven by E1A was resistant to p21 and dominant negative CDK2 inhibitors, and occurred at levels of CDK2 activity no greater than those seen in serum-starved, growth-arrested post-mitotic cells. Four potential mechanisms for these observations are being studied: 1) E1A itself results in substrate activation or inactivation of a substrate through protein-protein interactions, 2) alternative E2F family members, 3) other pocket protein targets, and 4) other E1A targets, including p400, a poorly characterized protein known to bind to regions of E1A necessary for CDK2 independent effects.

However, this work does not show whether the endogenous genes play a role in cell cycle exit. The loss of p21 resulted in a significant delay in cell cycle exit in cardiac myocytes, as shown by flow cytometry in newborn mice (wild type, missing p21 or p27, or both). A 2-fold increase in the number of cells with S or G2M DNA was found. However, this is a delay in cell cycle exit, not a permanent ability to continue cycling, as animals even 14 days of age have negligible DNA synthesis in myocardium. In p27 null mice, p27 has a roughly 2-fold larger effect than loss of p21. Mice lacking even one copy of p27 are sufficient for effects as large as those seen in p21 null mice. Interestingly, with the combined deletion of p21 and p27, DNA synthesis is shut down and little or none is seen even at 14 days of age. Roughly 30% of these cardiac myocytes have DNA content greater than G0/G1 at this stage in the newborn.

The nominal gold standard gene deletion approach for determining protein function in vivo could not used with Rb. The conventional Rb knockout results in embryonic lethality in midgestation with mammoth defects in hematopoiesis and neurogenesis. Thus, technologies for conditional gene deletion using the Cre/lox system have been used: when a gene is tagged in innocuous regions, such as within its introns, with 34-base-pair motifs known as loxP sites and the Cre recombinations are introduced, DNA recombination is triggered between the loxP sites.

Schneider's lab has used a Cre-dependent reporter gene as a knock-in to the ubiquitously transcribed rosa26 locus. In the absence of Cre, little or no recombination as measured by activation of a Cre-dependent lacZ gene is seen. When these mice are mated to mice containing Cre-recombinase driven by the alpha-myosin heavy chain promoter, nearly uniform recombination is seen in atrial and ventricular myocytes by 10.5-11 days gestation. The Cre-dependent promoter is sometimes dismissed in the literature as being expressed predominantly in the embryonic atria and only comes up in the ventricle after birth. But, the levels of expression seen in the future ventricle at the linear heart tube stage are sufficient to trigger recombination in nearly all ventricular myocytes.

A recombination-specific gene product appears in the myocardium of the crossed alpha-myosin heavy-chain Cre mice and lox-p Rb mice. There is loss of the intact Rb allele, almost complete loss of the Rb protein from the myocardium, and spontaneous DNA synthesis even under basal conditions, as shown by immunoperoxidase staining and two-color immunofluourescence. The absolute frequency for this event is low, roughly one cell in a thousand, similar to that reported in mice myocardium overexpressing cyclin D1. Thus, other inhibitors are likely present either in series with Rb, such as p21 and p27, or in parallel in with Rb, notably p130, the other pocket protein that coexists with Rb in myocardium.

This work is the first demonstration of an essential function for pocket proteins in cardiac growth control in vivo, they believe. This is based on the observation that when the relatively innocuous cardiac-restricted Rb knockout is made into the conventional, totally innocuous p130 knockout, about a 2-fold increase in heart size occurs between 4 and 8 weeks of age. Promiscuous reactivation of a number of cell-cycle regulators, including cyclin B, usually transcriptionally suppressed in adult heart results. The CDK inhibitor knockout and Cre knockout work is ongoing.

A cardiac-specific knockout of the bone morphogenic protein (BMP) was developed. BMP is a member of the transforming growth factor beta (TGF-beta) superfamily that has been implicated, through circumstantial evidence and explant model systems, as potential regulators of cardiac development. Komuro reported recently that BMP signaling was important for cardiac myogenesis in p19 embryonic carcinoma cells in the mouse model. The role of BMPs or their type-1 receptor ALK3 can not be studied in myocardium by conventional deletion because of lethality at gastrulation or other early stages before heart formation in BMP2, BMP4, and ALK3.

In their lab, no survivors to late gestation were seen when mating alpha MHC-Cre mice to mice loxP tagged for the ALK3 gene. Marked resorption even by embryonic day 15 and internal hemorrhages by embryonic day 13 were seen. All mice had large atrial septal or ventricular defects. In some animals, the endocardial cushion was present but hypocellular and disorganized. In others, the endocardial cushion was absent and a single four-chambered heart observed. At least a subset of the animals was defective for expression of NKX2.5, the cardio-restricted homeobox gene; mechanistic studies are ongoing.

A marked increase in apoptosis in the septum was caused by the absence of BMP signaling through ALK3. Thus, they think a BMP signal, directly or indirectly, is responsible for cell survival in cardiac morphogenesis. A number of genome-wide technologies to identify the BMP-dependant genes both for survival in septum and for normal morphogenesis involving the atrioventricular cushion is being pursued by their lab.

BMP signaling is known to occur through at least two pathways. One involves Smad transcription factors. A different set of Smad transcription factors is implicated in signaling for TGF-beta itself. Both the BMP and TGF beta pathways have been reported to involve signaling through MAP kinases, including a TGF beta-activated kinase (TAK), which lies upstream of JNK and p38.

TAK1 activity in myocardium in vivo is upregulated as a delayed response to aortic banding, as shown in their lab by immune complex kinase assays in which endogenous TAK1 was immuno-precipitated and incubated with its substrates (Fig.2, upper panel). Upregulation of TAK levels is not seen. Activation of a kinase as a delayed response following pressure overload is seen, consistent with its involvement in time-dependant autocrine-paracrine mechanisms (Fig 2, bottom panel).

In tissue culture, the TAK1 signal was sufficient to mimic the effect of TGF beta on the skeletal alpha actin promoter, a marker of hypertrophy. Importantly, TAK1 kinase activity was necessary for TGF-beta signaling and was completely extinguished by dominant negative TAK1. This process involved the p38 pathway, not JNK, and the protein ATF6, recently implicated in adrenergic and endothelin-dependant signaling to other serum response factor-dependant cardiac promoters.

Schneider's lab created transgenic mice overexpressing TAK1 in the myocardium that have the same 3- to 4-fold increase in TAK-1 activity seen with pressure overload. A 40-50% increase in heart size by ten days of age, associated with fulminant mortality, resulted. No F1 animal survived beyond 15 days of age. The F0 founders survived for three to six months, ultimately dying of hypertrophy and heart failure. Mosaicism in the founder mice accounts for the difference between F0s and F1s. The TAK1 transgenic mice had 1) myocyte enlargement, 2) myocyte drop out and fibrosis, 3) reactivation of fetal gene expression, 4) a decrease in systolic and diastolic function, and 5) a marked increase in apoptosis, as shown by TUNEL staining. No apoptosis in wild-type animals and the expected 2-5% prevalence of apoptosis in mammary epithelium as a positive control was seen.

Thus, they believe TAK1 is activated in vivo as a response to mechanical load. It signals via p38 to ATF6 and other downstream proteins, and is sufficient to trigger the molecular, morphological, and functional hallmarks of cardiac hypertrophy. The early mortality makes it especially attractive as a potential test bed for counter measures aimed at cardiac apoptosis or other components of the heart failure phenotype.

A cardiac-specific Bcl-2 transgenic mouse model has been created in their lab. Bcl-2 is the prototype of a family of pro- and anti-apoptotic proteins that act by preventing the mitochondrial transitions that couple upstream apoptotic signals to downstream effector pathways, including downstream caspase activation as part of the apoptotic cascade. Human Bcl-2 was shown to be expressed in the 4-copy and 14-copy lines. The endogenous protein is not altered in expression in the presence of the transgene, and the levels of are increased about 3-fold in the 4-copy line and 7-fold in the 14-copy line.

A marked reduction in infarct size associated with rescue of systolic and diastolic function was shown after one hour of ischemia and 24 hours of reperfusion in the high-copy Bcl-2 line. Rescue was not seen in the lower copy line, suggesting a dose-dependant relationship. The Bcl-2 protein is markedly downregulated early with ischemia-reperfusion injury, and may potentially explain the all-or-nothing effect seen. In the early hours of ischemia-reperfusion there is an 80% loss of protein, which is partially recovered within 24 hours.

Two reported mechanisms for Bcl-2 degradation are likely active in ischemia-reperfusion. Bcl-2 is downregulated through inhibition of MAP-kinase-dependent phosphorylation, which can be triggered by TNF-alpha, among other agonists. Bcl-2 is also a reported target for degradation via proteolysis by caspases. Their second-generation Bcl-2 mice contain a phosphomimetic mutation at the sites for MAP-kinase phosphorylation; a reportedly constitutively stable mutation in tissue culture, even under TNF-alpha stimulation. Third-generation mice with caspase cleavage site mutations are underway.

They expect that mutations of Bcl-2, engineered forms of Bcl-2 resistant to the two degradation pathways, may be capable of providing cardioprotection at lower basal levels of expression. And, may be resistant to the degradation seen with ischemia-reperfusion in injury, and ultimately may be the forms of Bcl-2 selected for translational studies using viral gene delivery to the myocardium.