ABSTRACT
Class IA phosphoinositide 3-kinases (PI3Ks) are activated by growth factor receptors, and they regulate, among other processes, cell growth and organ size. Studies using transgenic mice overexpressing constitutively active and dominant negative forms of the p110α catalytic subunit of class IA PI3K have implicated the role of this enzyme in regulating heart size and physiological cardiac hypertrophy. To further understand the role of class IA PI3K in controlling heart growth and to circumvent potential complications from the overexpression of dominant negative and constitutively active proteins, we generated mice with muscle-specific deletion of the p85α regulatory subunit and germ line deletion of the p85β regulatory subunit of class IA PI3K. Here we show that mice with cardiac deletion of both p85 subunits exhibit attenuated Akt signaling in the heart, reduced heart size, and altered cardiac gene expression. Furthermore, exercise-induced cardiac hypertrophy is also attenuated in the p85 knockout hearts. Despite such defects in postnatal developmental growth and physiological hypertrophy, the p85 knockout hearts exhibit normal contractility and myocardial histology. Our results therefore provide strong genetic evidence that class IA PI3Ks are critical regulators for the developmental growth and physiological hypertrophy of the heart.
In response to increased workload, the adult heart grows due to an increase in the size of cardiac myocytes (26). Cardiac hypertrophy can result from either physiological or pathological stimuli. Physiological stimuli such as exercise lead to compensatory growth of the heart that is characterized by normal cardiac structure, preserved or improved cardiac function, and minimal alteration in cardiac gene expression pattern (9, 22). In diseased states, such as hypertension and valvular disorders, the chronic increase in workload leads to pathological hypertrophy of the heart. Though cardiac function is preserved initially (compensated growth), over time such pathological hypertrophy leads to tissue damage, left ventricular dilation, and deterioration in heart function (decompensated growth) and may eventually result in heart failure (19, 26). In addition, pathological hypertrophy is associated with an altered pattern of cardiac gene expression (21). As heart failure is almost invariably associated with cardiac hypertrophy, it is of clinical importance to understand the molecular mechanisms underlying various forms of cardiac hypertrophy (11).
Phosphoinositide 3-kinases (PI3Ks) are a family of lipid kinases that regulate a wide range of cellular processes including cell growth, survival, metabolism, and motility (6). Class I PI3Ks phosphorylate the 3′-hydroxyl on the inositol ring of phosphatidylinositol-4,5-bisphosphate to generate the lipid second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3). PIP3 in turn activates a number of downstream signaling proteins, including the protein serine-threonine kinase Akt (PKB), through binding to their pleckstrin homology domains. The lipid phosphatase PTEN dephosphorylates PIP3 at the 3′ position and therefore terminates PI3K signaling (47).
The PI3K/Akt pathway controls cell size, at least in part, through the regulation of the mammalian target of rapamycin (mTOR) pathway. Akt can phosphorylate and inactivate the tuberosclerosis complex 2 (TSC2) gene product, tuberin, and thereby reduce its ability to stimulate GTP hydrolysis on the Ras-like G protein, Rheb (16, 20, 24, 34, 49, 50). As a consequence, Rheb-dependent activation of the mTOR protein Ser-Thr kinase is enhanced, leading to increased protein synthesis and cell growth (10).
Both growth factor receptors (e.g., the insulin, insulin-like growth factor 1, and epidermal growth factor receptors) and G protein-coupled receptors can activate class I PI3Ks. Growth factor receptors engage class IA PI3Ks, which are heterodimers consisting of a p85 regulatory subunit and a p110 catalytic subunit. The p85 regulatory subunit is essential for the stabilization of the p110 catalytic subunit and for its recruitment to phosphotyrosines on activated receptors or their adaptor molecules (14). In both humans and mice, three genes encode five isoforms of the p85 regulatory subunit: the three major isoforms, p85α, p55α, and p50α, are encoded by a single gene, pik3r1, via alternative transcription initiation sites, and the two minor isoforms, p85β and p55γ, are encoded by the genes pik3r2 and pik3r3, respectively. While the p85α, p55α, p50α, and p85β proteins are ubiquitous, p55γ is enriched in the brain and the testis. Three genes encode the class IA p110 catalytic subunit isoforms p110α, p110β, and p110δ, with the first two being ubiquitous and p110δ being restricted to leukocytes (14). G protein-coupled receptors activate class IB PI3K, which is a dimer of the regulatory subunit p101 and the catalytic subunit p110γ, with the p101 regulatory subunit conferring sensitivity to activation by the Gβγ subunit of trimeric G proteins (44, 54).
The PI3K pathway has been implicated in the regulation of both heart size and cardiac function, with class IA and class IB PI3Ks both playing distinct roles (1, 35). Class IA PI3K appears to regulate heart size without affecting systolic function under resting conditions: transgenic mice overexpressing a constitutively active form of p110α in the heart have enlarged hearts, while those expressing a dominant negative form of p110α (dnPI3K) have reduced heart size, with cardiac contractility unaltered in both cases (40). In accordance with class IA PI3K being downstream of growth factor receptors, mice overexpressing the insulin-like growth factor 1 receptor in the heart also show increased heart size, which is blocked by the coexpression of dnPI3K (27). Furthermore, class IA PI3K seems to regulate physiological but not pathological cardiac hypertrophy, as the expression of dnPI3K in the heart blunts exercise-induced hypertrophy but not pressure overload-induced hypertrophy (29). The cardiac hypertrophy driven by class IA PI3K is likely to be mediated, at least in part, through Akt. Overexpression of a constitutively active form of Akt in the heart results in an increase in heart size, and coexpression of a kinase dead form of Akt reduces the cardiac hypertrophy driven by constitutively active p110α (25, 41). Class IB PI3K, on the other hand, regulates contractility and pathological hypertrophy of the heart (1, 36). Under normal conditions, mice with homozygous deletion of p110γ show hypercontractility of the heart without any changes in heart size (8). The p110γ−/− heart, however, does show attenuated hypertrophy in response to chronic pressure overload induced by transverse aortic constriction (33). Interestingly, while the p110γ-driven pathological hypertrophy requires its kinase activity and is likely to be mediated through Akt, the negative regulation of cardiac contractility by p110γ is mediated through its inhibition of phosphodiesterase-3B, which is independent of the kinase activity (33).
The goal of the current study is to investigate the role of class IA PI3K in regulating heart size and function via the targeted deletion of p85 genes, thereby circumventing potential complications that may arise from the overexpression of constitutively active or dominant negative forms of p110α. It has been difficult to assess the function of class IA PI3K in the adult animal by gene deletion because germ line knockout of the p85α/p55α/p50α regulatory subunits results in perinatal lethality (13), whereas germ line knockout of either the p110α or the p110β catalytic subunit causes embryonic lethality (2, 3). To overcome this problem, we used the Cre/lox approach to generate mice in which all three p85α isoforms are deleted in striated muscles. Previous studies with germ line knockout mice suggest that the p85α isoforms are likely to have the most impact on class IA PI3K signaling, as their deletion results in death of mice within days after birth (13), whereas mice lacking the p85β isoform are viable (52). The primary aim of the current study was to generate muscle-specific p85α/p55α/p50α knockout (KO) mice to investigate the role of the p85α regulatory subunits in regulating heart size and function during developmental growth and exercise-induced hypertrophy. In addition, we also examined the contribution of the p85β regulatory subunit by crossing the p85α/p55α/p50α muscle KO mice with p85β−/− mice.
In this report, we show that mice with the p85 regulatory subunits (p85α/p55α/p50α and p85β) of class IA PI3K deleted in the heart show reduced heart size, which is a consequence of a reduction in cardiac myocyte size. The p85 knockout heart shows greatly reduced class IA PI3K activity and diminished Akt activation in response to growth factor signaling. Furthermore, exercise-induced cardiac hypertrophy is attenuated in the p85 knockout heart. Despite such defects in growth under normal conditions and in physiological hypertrophy, the p85 knockout heart shows normal contractility and myocardial histology. These results provide strong genetic evidence that class IA PI3K regulates developmental heart growth and physiological hypertrophy without affecting cardiac function.
MATERIALS AND METHODS
Mice and genotyping.A 2.6-kb XbaI fragment of the mouse pik3r1 gene containing exon 7 was excised from a BAC clone and subcloned into cKO-II vector that carries a FRT-flanked PGK-neo cassette. Two loxP sites were inserted to flank exon 7. A 6-kb XhoI-XbaI fragment containing exons 1b and 1c was used as the long homology arm for targeted recombination. Correctly targeted embryonic stem cells were identified by Southern blotting with a probe outside the homology arms. Chimera mice were mated with rosa26-FLP mice (a gift from S. M. Dymecki) to remove the PGK-neo cassette and transmit the p85α loxP allele in the germ line. The final structure of the p85α loxP locus was confirmed with Southern blotting, PCR, and sequencing. The p85α loxP allele can be distinguished from the wild-type (WT) p85α allele by PCR using the primer pair CAC CGA GCA CTG GAG CAC TG and CCA GTT ACT TTC AAA TCA GCA CAG, which amplifies a 252-bp fragment from the WT allele and a 301-bp fragment from the loxP allele. The p85α loxP/loxP mice were mated with mck-Cre mice (5) and with p85β−/− mice (52). The Cre-mediated deletion of exon 7 of the pik3r1 gene was verified by Southern blotting and by PCR using the primer pair GGT TTC TTA CTT TAG ACG GAG CTG and CCA GTT ACT TTC AAA TCA GCA CAG, which amplifies a 1,275-bp fragment from the undeleted p85α loxP allele and a 298-bp fragment from the deleted allele.
All mice were maintained on a 129Sv-C57BL/6-FVB mixed background. Animal care and experimentations were approved by the Institutional Animal Care and Use Committees of Beth Israel Deaconess Medical Center and Harvard Medical School.
Western blot, kinase assay, and Northern blot.Heart lysates were prepared as described previously (52). Briefly, heart tissue was homogenized in buffer containing 25 mM Tris-HCl (pH 7.4), 10 mM EDTA, 10 mM EGTA, 100 mM NaF, 50 mM NaPPi, 1% NP-40, 10 mM Na3VO4, 1 mM dithiothreitol, 2 mM phenylmethylsulfonyl fluoride, 8 μg/ml leupeptin, 8 μg/ml aprotinin, and 8 μg/ml pepstatin. The lysate was clarified by centrifugation at 150,000 × g for 60 min. An in vitro PI3K assay was carried out as described previously (15). Briefly, pTyr immunoprecipitate from heart lysate was incubated with phosphatidylinositol (Avanti) and [γ-32P]ATP in kinase buffer containing 20 mM Tris-HCl (pH 7.4), 100 mM NaCl, and 0.5 mM EGTA. The lipid product was separated by thin-layer chromatography and quantified using a phosphorimager screen. The p85 pan antibodies, described before (15), reacted strongly with all p85α isoforms and weakly with p85β. Antibody against the insulin receptor β-subunit was from Santa Cruz Biotechnologies. Antibody against p110α was from BD Bioscience. Monoclonal antibody against phosphotyrosine (4G10) was a gift from T. Roberts. All other antibodies were from Cell Signaling Technologies. Northern blotting was carried out as previously described (29).
Analytical procedures and tissue histology.Analytical procedures for mouse heart weight have been described before (29). The measurement of myocyte cross-section area was carried out as described previously (42). Briefly, a section of the left ventricular epicardial free wall was stained with fluorescein isothiocyanate-conjugated wheat germ agglutinin to visualize membranes, and images were acquired on a confocal microscope (Bio-Rad). Masson's trichrome staining on paraformaldehyde-fixed tissue and muscle citrate synthase assay were carried out as described previously (29). Statistical analysis was carried out using PRISM 4 (Graphpad Software). One-way and two-way analyses of variance (ANOVA) were followed by Tukey's and Bonferroni's tests, respectively.
RESULTS
Generation of muscle-specific p85α knockout mice.The murine gene for the major regulatory subunit of class IA PI3K, pik3r1, encodes three proteins (p85α, p55α, and p50α) that arise from alternative transcription initiation sites (12). All three proteins share the same last nine exons (exons 7 to 15) that encode their C-terminal SH2 domains and p110-binding domain; therefore, all are able to mediate class IA PI3K signaling (14). The full-length p85α isoform has an SH3 domain and a BH domain at its N terminus that are encoded by exons 1a to 6. The two shorter isoforms, p50α and p55α, have alternative transcription initiation sites, and their unique first exons (exons 1b and 1c, respectively) splice directly into exon 7, the first common exon for all three isoforms (12). Mice deficient in either the full-length p85α protein or the two shorter p50α and p55α proteins are viable (7, 48, 51), while mice lacking all three isoforms exhibit perinatal lethality (13, 15). In order to better understand the role of class IA PI3K in the adult, we used the Cre/lox approach to generate mice in which all three p85α isoforms can be deleted in a tissue-specific fashion (39). We engineered a pik3r1 allele in which exon 7 is flanked by loxP sequences. Upon Cre-mediated deletion of exon 7, splicing of upstream exons (exon 6, 1b, or 1c) directly into the downstream exon (exon 8) results in a frameshift mutation that introduces an immediate stop codon. This deletion should prevent the translation of the SH2 and p110-binding domains, eliminating the ability to form a functional protein from any of the three transcription initiation sites. As a result, p85α becomes truncated and the expression of p55α and p50α is essentially abolished (Fig. 1A).
Generation of muscle-specific p85 knockout mice. (A) Exon 7 of the murine pik3r1 locus was replaced by homologous targeting with a gene construct that contains exon 7 flanked by a loxP sequence and a PGK-neo cassette flanked by a FRT sequence. The PGK-neo cassette was subsequently removed by FLP-mediated recombination, leaving only the floxed exon 7. b, BamHI; k, KpnI; x, XbaI; shaded boxes, FRT sites; black arrowheads, loxP sites. (B) Southern blot of gastrocnemius muscle DNA from WT, p85α loxP/loxP, and p85α mKO mice, confirming the deletion of the p85α gene by using a probe shown in panel A (filled bar at bottom). (C) Western blot of heart lysates from WT, p85α loxP/loxP, p85α mKO, and p85α mKO p85β−/− mice with anti-p85 pan and anti-p110α antibodies confirming the loss of p85α, p55α, p50α, and p110α in the heart.
We crossed the p85α loxP/loxP mice with the mck-Cre transgenic line that expresses Cre in striated muscles including skeletal and cardiac myocytes (5). Normally, all three p85α isoforms are expressed in the heart and in skeletal muscles, albeit p55α and p50α are at much lower levels than p85α (Fig. 1 and data not shown). In the p85α/p55α/p50α muscle KO mice (hereafter referred to as p85α mKO mice for clarity), we observed the expected genetic recombination at the p85α loxP locus (Fig. 1B). We also observed a significant loss of p85α, p55α, and p50α proteins both in the heart (Fig. 1C) and in skeletal muscles (data not shown). The residual p85α protein (10 to 20%) might be attributed to nonmuscle cells in the tissue and/or incomplete deletion of the gene in a few myocytes. The truncated p85α fragment that contains the N-terminal SH3 and BH domains appeared to be unstable, as it was not detectable with an antibody against the SH3 domain of p85α (data not shown and reference 15). In agreement with the notion that the p85 regulatory subunit is necessary not only for the activation of the p110 catalytic subunit but also for its stability (4, 55), we also observed a concomitant loss of the p110α protein in these tissues (Fig. 1C).
The p85β regulatory subunit of class IA PI3K is encoded by a separate gene, pik3r2, and mice lacking p85β are viable (52). We wished to investigate a possible contribution of the p85β regulatory subunit to regulating heart size and function in the absence of p85α. We therefore crossed p85α mKO mice with p85β−/− mice in order to achieve further reduction of class IA PI3K in the heart and the skeletal muscles.
p85α mKO p85β−/− mice have reduced heart size and attenuated class IA PI3K signaling in the heart.Both p85α mKO mice and p85α mKO p85β−/− mice were viable and had normal appearance and normal body weights (Table 1 and data not shown). The skeletal muscles of these mice appeared normal in morphology, and these mice had normal blood glucose levels (data not shown). This was not entirely unexpected, as mice deficient in either PTEN or PDK1 in muscle also lacked a striking skeletal muscle phenotype (8, 32).
p85α mKO 85β−/− mice have reduced heart weighta
We next examined the heart sizes of these mice. Despite p85α being considered the most abundant class IA regulatory subunit, p85α mKO mice had no apparent cardiac phenotype. The heart sizes of p85β−/− mice were also not different from those of WT mice (Fig. 2 and data not shown). Presumably, in the absence of one p85 isoform the other could compensate for its function. The hearts of p85α mKO p85β−/− mice, however, were significantly smaller for both females (Fig. 2A and 2B) and males (data not shown). When normalized against either body weight or tibia length (a measurement of body size), the heart size of these mice was approximately 20% smaller than that of control mice (Fig. 2C and D and Table 1). The small heart size of p85α mKO p85β−/− mice was associated with a reduction in myocyte size, as the cross-section area of left ventricular myocytes was smaller in these mice than in control mice (Fig. 3A and B). While the majority of the cardiac myocytes in the p85α mKO p85β−/− mice were smaller, we occasionally found very large fibers that were not seen in the control hearts (Fig. 3A). These fibers might represent myocytes that have escaped p85α gene deletion.
p85α mKO p85β−/− mice show reduced heart size. Representative images of whole hearts (A) and four-chamber sections (B) from 18- to 20-week-old mice of similar body weights are shown. Heart weight (HW)-to-body weight (BW) ratios (C) and HW-to-tibia length (TL) ratios (D) of 18- to 20-week-old mice show reduced heart size in p85α mKO p85β−/− mice (n = 11; *, P of <0.001 compared to all other genotypes, by one-way ANOVA). Bars represent standard errors of the means.
p85α mKO p85β−/− mice have reduced cardiac myocyte size and normal myocardium structure. (A) Wheat germ agglutinin-fluorescein isothiocyanate staining of transverse sections of the left ventricular epicardial free wall, illustrating the membrane boundary of individual myocytes (asterisk indicates a very large myocyte occasionally observed in the p85α mKO p85β−/− heart). (B) Quantification of myocyte cross-section area of the left periventricular wall as shown in panel A (n = 4, one-tailed t test). Bars represent standard errors of the means. (C) Masson's trichrome staining of left ventricle sections, showing no obvious fibrosis lesion in the p85α mKO p85β−/− heart.
As we expected, class IA PI3K signaling in the p85α mKO p85β−/− heart was greatly attenuated. At a physiological (submaximal) dose of insulin (0.5 U/kg of body weight), there was an approximately threefold activation of PI3K in the hearts of control animals, whereas there was no appreciable activation of PI3K in the p85α mKO p85β−/− hearts despite normal insulin receptor phosphorylation in these animals. A saturating dose of insulin (5 U/kg) elicited only a threefold increase of PI3K activity in the p85α mKO p85β−/− hearts, compared to a nearly 10-fold increase of PI3K activity in the WT hearts (Fig. 4A). Together, these results suggest that the p85α mKO p85β−/− hearts, despite their smaller sizes and altered gene expression profiles, are functionally normal under resting conditions.
Class IA PI3K signaling is greatly attenuated in the p85α mKO p85β−/− heart. (A) Insulin-stimulated PI3K activities in anti-pTyr immunoprecipitate were measured by in vitro kinase assay (n = 4 per genotype per treatment; *, P of <0.05 compared to other two genotypes of the same treatment group, by two-way ANOVA). Bars represent standard errors of the means. (B) Insulin receptor β-subunit (IRβ) tyrosine phosphorylation levels in the heart were measured by pTyr blotting and were comparable among the different genotypes. Insulin-stimulated Akt activation levels were assessed using phospho-specific antibodies against Thr-308 and Ser-473 on Akt. Insulin-stimulated phosphorylation levels of p44/42 Erk were comparable among the different genotypes. Representative blots of four experiments are shown. IP, immunoprecipitation.
p85α mKO p85β−/− mice have attenuated myocardial hypertrophy in response to exercise training.We next investigated whether the p85α mKO p85β−/− hearts became hypertrophic in response to increased workload. Previous studies suggest that class IA PI3K is required for physiological but not pathological hypertrophy of the heart (29). We therefore investigated the effect of exercise (in the form of swim training) on the heart size of p85α mKO p85β−/− mice. Mice were subjected to 4 weeks of swim training to induce myocardial hypertrophy. The loss of PI3K in skeletal muscle did not seem to significantly impact the exercise capacity of the muscle, and the p85α mKO p85β−/− mice proceeded through the 4-week swim training protocol normally. The analysis of muscle citrate synthase activity of these mice supported the notion that they exercised to a comparable extent as did the control mice (increase in muscle citrate synthase activity in p85α mKO p85β−/− mice following swim training was 49% ± 7%, compared to an increase of 45% ± 2% in WT mice; n = 6).
Analysis of heart size following swim training revealed that exercise-induced hypertrophy was attenuated in the p85α mKO p85β−/− heart. While the control mice had an average of a 23% increase of heart weight (normalized to body weight) in response to exercise, the p85α mKO p85β−/− mice had an average of only a 14% increase (P < 0.01, two-way ANOVA). This deficiency was even more pronounced when heart weight was normalized to tibia length. Thus, while the hearts of control mice had significantly enlarged following swim training, the heart sizes of p85α mKO p85β−/− mice following swim training were comparable only to those of control mice of the nonswim group (Fig. 6 and Table 3).
Echocardiography of p85α mKO 85β−/− mice reveals normal cardiac functiona
Swim training of p85α mKO 85β−/− mice reveals attenuated cardiac hypertrophya
We examined the phosphorylation state of Akt in the hearts at the end of the 4-week swim training and found it to be elevated in both WT and p85α mKO p85β−/− mice from that of the nonswim control group. Paradoxically, the p85α mKO p85β−/− mice exhibited higher basal Akt phosphorylation than did WT mice after swim training (Fig. 7A). It is possible that the enhanced Akt phosphorylation in p85α mKO p85β−/− mice was due to elevated adrenergic tone in response to the attenuated hypertrophy. Alternatively, there may be less negative feedback regulation of PI3K/Akt activation via the mTOR /S6-kinase (S6K) pathway in the p85α mKO p85β−/− mice (see Discussion). Northern blot analyses revealed that exercise partially alleviated the abnormal gene expression pattern in the p85α mKO p85β−/− hearts: the mRNA of α-skeletal actin returned to normal levels in response to swim training (Fig. 7B).
Despite the attenuated myocardial hypertrophy, the function of the p85α mKO p85β−/− heart was preserved. Echocardiography showed normal systolic function in the p85α mKO p85β−/− hearts following swim training (fractional shortening of p85α mKO p85β−/− hearts following swim training was 40% ± 6%, compared to 32% ± 5% of WT hearts following swim training; n = 5 to 6 mice per genotype). Both WT and p85α mKO p85β−/− mice showed comparable reductions in heart rate following 4 weeks of swim training (data not shown). Furthermore, no tissue damage or fibrosis was observed in the p85α mKO p85β−/− hearts following swim training (data not shown). Taken together, these results suggest that class IA PI3K is required for the physiological hypertrophy of the heart.
DISCUSSION
Previous studies with transgenic mice overexpressing various components of the class IA PI3K signaling pathway in the heart have suggested that class IA PI3K is a critical regulator of heart growth under physiological conditions (27, 29, 40). However, to our knowledge, this is the first study in which the role of class IA PI3K has been studied in the heart by using a knockout approach. In this study, we generated mice with muscle-specific deletion of the p85α gene of class IA PI3K. When crossed onto the p85β−/− background, these mice showed reduced heart size and attenuated physiological hypertrophy, thus demonstrating genetically that class IA PI3K is an important regulator of heart size.
To better understand the in vivo roles of class IA PI3K in various tissues, we generated mice carrying a floxed p85α allele that can be deleted via Cre-mediated genetic recombination, which results in the loss of p85α, p55α, and p50α in a tissue-specific manner (Fig. 1). When p85α loxP/loxP mice were crossed with the mck-Cre transgene (5), we found efficient deletion of p85α, p55α, and p50α proteins in the heart and in skeletal muscles. We also observed a concomitant loss of the p110α catalytic subunit, in accordance with previous reports that the p85 regulatory subunit is essential for the stability of the p110 catalytic subunit (4, 55). The residual amount of p85α in the heart (10 to 20%) might be due to nonmyocyte cells in the tissue or incomplete deletion of p85α in a small number of myocytes.
Despite the p85α isoforms being considered the major regulatory subunit of class IA PI3K and evidence from germ line KO studies suggesting that their deletion may impact heart size, the p85α mKO mice showed no apparent cardiac phenotype (Fig. 2 and data not shown). This suggests that p85β was able to compensate for the loss of the p85α isoforms. In contrast, with the additional elimination of p85β, heart size is significantly reduced. To our knowledge, the current study has uncovered for the first time a role of the p85β subunit in the regulation of tissue size. The p85α mKO p85β−/− mice were viable and had greatly diminished class IA PI3K activity in the heart (Fig. 4). Consequently, these mice had reduced heart size due to a reduction in cardiac myocyte size (Fig. 2 and 3). This is similar to the reduction observed both in mice with cardiac overexpression of dominant negative p110α (dnPI3K) (40) and in mice with cardiac deficiency of PDK1, the upstream activator of Akt (32). The p55γ isoform has been shown to be up-regulated in mouse embryonic fibroblasts deficient in p85α/p55α/p50α and p85β (4). We could not detect significant levels of the p55γ protein in either wild-type or p85α mKO p85β−/− hearts; thus, it is unlikely that p55γ was dramatically up-regulated in the p85α mKO p85β−/− hearts.
While the p85α mKO p85β−/− hearts exhibited attenuated Akt activation in response to insulin stimulation, the activation of p44/p42 Erk was enhanced (Fig. 4B). Previous studies have shown that in postdifferentiation myotubes Akt can inhibit Erk activation by interacting with Raf (37), and Akt has been demonstrated to directly phosphorylate Raf in vitro (56). Our observation extends these findings in vivo and indicates that in the heart the PI3K/Akt pathway acts to suppress Erk signaling. Such cross talk, however, is likely to be specific to cardiac myocytes, as we did not observe enhanced Erk activation in the skeletal muscles of p85α mKO p85β−/− mice (data not shown). In mouse embryonic fibroblasts deficient in p85α/p55α/p50α and p85β, platelet derived growth factor-stimulated Erk activation is attenuated (4). Thus, the PI3K/Akt pathway could influence Erk signaling in different ways specific to tissue and growth factor.
The small hearts of p85α mKO p85β−/− mice also exhibited an altered gene expression pattern, characterized by the up-regulation of fetal cardiac genes ANP and BNP (Fig. 5), that is typically associated with pathological hypertrophy (19, 29). It is possible that this is part of a compensatory response to the loss of class IA PI3K signaling. Despite the reduction in size and the transcriptional changes, cardiac structure and function remained normal in the p85α mKO p85β−/− mice (Fig. 3 and Table 2). Taken together, these results clearly demonstrate that class IA PI3K signaling is required for normal growth of the heart, but a severe attenuation of this pathway does not affect cardiac systolic function under these conditions. This is an important observation, as class IA PI3K inhibitors are currently being developed for cancer therapy.
p85α mKO p85β−/− hearts exhibit altered expression levels of ANP and BNP. (A) Northern blots of ANP, BNP, and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) mRNA levels in heart. (B) The relative levels of ANP and BNP mRNA (normalized to GAPDH mRNA levels) were quantified from three experiments (*, P of <0.05 compared to all other genotypes, by one-way ANOVA). Bars represent standard errors of the means. a.u., artibrary units.
Since class IA PI3K has been implicated in the induction of physiological but not pathological hypertrophy of the heart (29), we also tested the hypertrophic response to swim training in p85α mKO p85β−/− mice. These mice showed attenuated cardiac hypertrophy in response to exercise training (Fig. 6), demonstrating that class IA PI3K is indeed critical for the physiological hypertrophy of the heart. Interestingly, the hearts of p85α mKO p85β−/− mice did show limited growth in response to exercise, accompanied by enhanced basal Akt signaling (Fig. 7). It is possible that this is due to enhanced adrenergic signaling to the class IB PI3K p110γ in an attempt to compensate for the loss of class IA PI3K. Alternatively, it is possible that reduced signaling to the mTOR pathway in the p85α mKO p85β−/− hearts from that in wild-type hearts during the 4 weeks of swim exercise circumvents a negative feedback regulation of IRS-1 (53).
p85α mKO p85β−/− mice show attenuated cardiac hypertrophy in response to exercise training. (A) Representative whole-heart images of 5-month-old mice of similar body weights that were either nonswim controls or swim trained for 4 weeks. Heart weight (HW)-to-body weight (BW) ratios (B) and HW-to-tibia length (TL) ratios (C) of nonswim control and swim-trained mice show attenuated hypertrophy of the p85α mKO p85β−/− hearts (n = 5 to 9 per genotype per training group). *, P of <0.01 compared to the same training group of the other genotypes; **, P of <0.001 compared to the same training group of the other genotypes; #, P of <0.001 compared to nonswim control group of the same genotype. All results were determined by two-way ANOVA. Bars represent standard errors of the means.
p85α mKO p85β−/− mice show enhanced basal Akt activity and partial correction of gene expression in the heart. (A) Western blots of heart lysates from nonswim controls and swim-trained mice. (B to E) Levels of α-skeletal actin, β-myosin heavy chain (β-MHC), and ANP mRNA in the hearts of nonswim controls and swim-trained mice (n = 3 per genotype per training group). *, P of <0.05 compared to WT nonswim control group; #, P of <0.05 compared to nonswim control group of the same genotype. All results were determined by two-way ANOVA. Bars represent standard errors of the means. a.u., arbitrary units.
The mTOR signaling pathway controls protein translation and is a major regulator of cell size and cell growth in many cell types. Consistent with this idea, inhibition of mTOR attenuates cardiac hypertrophy (42). Muscle-specific deletion of PDK1 (a protein kinase required for the activation of Akt, p70S6K, and other AGC family protein kinases) results in severe heart failure at an early age (32). As discussed in the introduction, Akt phosphorylates tuberin to turn off its Rheb-GAP activity, thereby allowing Rheb-dependent activation of mTOR and p70S6K (10). Consistent with this, mice with ventricular deletion of the tuberosclerosis complex 1 gene product hamartin, the heterodimeric partner of tuberin that is required for its stability and activity, develop hypertrophic cardiomyopathy (30). Thus, there is considerable evidence that mTOR mediates cardiac hypertrophy. However, it does not appear that S6K is a critical mediator of mTOR-dependent cardiac hypertrophy. Mice doubly deficient in S6K1 and S6K2 genes have normal heart weight-to-body weight ratio and normal hypertrophic response to both exercise and aortic banding, and the loss of S6K1 and S6K2 was unable to attenuate the increase in heart size induced by cardiac overexpression of constitutively active p110α (28). We did not observe clear alternation in S6 phosphorylation in the hearts of p85α mKO p85β−/− mice (data not shown).
The PI3K/Akt pathway could also regulate the size of cardiac myocytes via the FOXO family of transcription factors. A recent report suggests that Akt controls cardiac myocyte size through phosphorylation and inactivation of the transcription factor FOXO3a, which regulates the transcription of the ubiquitin ligase atrogin-1 that directs muscle protein degradation (38, 43, 46).
Although we could not exclude the possibility that a reduction in myocyte number also contributes to small heart size in the p85α mKO p85β−/− mice, it is likely that PI3K regulates myocyte size rather than myocyte number. Previous analysis of transgenic mice with cardiac overexpression of constitutively active or dominant negative forms of p110α, as well as muscle-specific PDK1 knockout mice, showed that all exhibit altered myocyte size but no change in myocyte number (32, 40). In addition, the muscle creatine kinase promoter used to drive the expression of the Cre transgene in our study is activated only in differentiating myotubes but not in proliferating myoblasts (5, 45). Indeed, the reduction in heart size of the p85α mKO p85β−/− mice can be largely accounted for by a reduction in myocyte size (Fig. 2C and 3B).
Our current study, in conjunction with a number of previous reports, clearly establishes the differential regulation of cardiac growth and contractility by class IA and class IB PI3K. This study and others (29, 40) demonstrate that class IA PI3K regulates the physiological growth and hypertrophy of the heart without affecting its contractility. Class IB PI3K, on the contrary, does not affect normal heart growth but negatively regulates cardiac contractility through its kinase-independent regulation of phosphodiesterase 3B. Class IB PI3K also mediates pathological hypertrophy in response to adrenergic signaling (8, 33). Hence, cardiac deficiency of the PIP3 phosphatase PTEN leads to both enlarged heart and compromised contractility, with the hypertrophy corrected upon the introduction of dominant negative p110α and the contractility normalized upon the additional deletion of p110γ (8). A better understanding of the various signaling pathways regulating myocardial hypertrophy and heart function is critical for devising new therapeutic strategies in the treatment of heart failure (19, 36). The diverse roles PI3Ks play in cardiac physiology make them attractive targets in the treatment of myocardial hypertrophy due to chronic pressure overload: the activation of class IA PI3K would promote physiological growth of the myocardium that normalizes wall stress, and the inhibition of class IB PI3K would reduce pathological hypertrophy and improve contractility.
ACKNOWLEDGMENTS
We thank S. L. White and D. A. Brown for assistance with histology, S. E. Fostello for assistance with echocardiography, and J. D. Chang, J. Blenis, T. M. Roberts, and J. Yuan for discussion and advice.
This work is supported by a Howard Hughes Medical Institute predoctoral fellowship to J.L., by NIH grants GM41890 and CA089021 to L.C.C., and by NIH grant R01-HL-65742 to S.I.
FOOTNOTES
- Received 2 June 2005.
- Returned for modification 28 July 2005.
- Accepted 16 August 2005.
- Copyright © 2005 American Society for Microbiology
