Microtubule Actin Cross-Linking Factor 1 Regulates Cardiomyocyte Microtubule Distribution and Adaptation to Hemodynamic Overload

Aberrant cardiomyocyte microtubule growth is a feature of pressure overload induced cardiac hypertrophy believed to contribute to left ventricular (LV) dysfunction. Microtubule Actin Cross-linking Factor 1 (MACF1/Acf7) is a 600 kd spectraplakin that stabilizes and guides microtubule growth along actin filaments. MACF1 is expressed in the heart, but its impact on cardiac microtubules, and how this influences cardiac structure, function, and adaptation to hemodynamic overload is unknown. Here we used inducible cardiac-specific MACF1 knockout mice (MACF1 KO) to determine the impact of MACF1 on cardiac microtubules and adaptation to pressure overload (transverse aortic constriction (TAC).In adult mouse hearts, MACF1 expression was low under basal conditions, but increased significantly in response to TAC. While MACF1 KO had no observable effect on heart size or function under basal conditions, MACF1 KO exacerbated TAC induced LV hypertrophy, LV dilation and contractile dysfunction. Interestingly, subcellular fractionation of ventricular lysates revealed that MACF1 KO altered microtubule distribution in response to TAC, so that more tubulin was associated with the cell membrane fraction. Moreover, TAC induced microtubule redistribution into this cell membrane fraction in both WT and MACF1 KO mice correlated strikingly with the level of contractile dysfunction (r2 = 0.786, p<.001). MACF1 disruption also resulted in reduction of membrane caveolin 3 levels, and increased levels of membrane PKCα and β1 integrin after TAC, suggesting MACF1 function is important for spatial regulation of several physiologically relevant signaling proteins during hypertrophy. Together, these data identify for the first time, a role for MACF1 in cardiomyocyte microtubule distribution and in adaptation to hemodynamic overload.

MICROTUBULES (MTS) ARE DYNAMIC filamentous elements, which regulate cell size, morphology, and subcellular organization. Microtubules act as structural elements that resist actomyosin contractility (36) and also serve as tracks for transport of mRNA (25), proteins (16,32), and organelles (17,42). Although microtubules play important roles in many cell functions, stabilization and aberrant accumulation of microtubules in cardiomyocytes is a feature of pressure overload hypertrophy that may impair contractile function and contribute to heart failure (15,41). Aberrant stabilization and accumulation of MTs are believed to contribute to contractile dysfunction, in part, by increasing the viscous load on the myofilaments (36,41). Microtubules also play an important role in ion channel trafficking (29), mRNA transport (22,27), and intracellular organization (38,40,43), and disruption of these microtubule functions by aberrant stabilization in pressure overload may also contribute to adverse remodeling and heart failure. In support of a role for MT stabilization in contractile dysfunction, we found that chronic (every other day) treatment with colchicine to reduce MT densification attenuated hypertrophy and improved contractile function in Balb-c mice exposed to pressure overload produced by transverse aortic constriction (TAC) (7). Identification of endogenous pathways that regulate cardiac microtubule dynamics may thus provide pharmacological targets for attenuating microtubule-related ventricular dysfunction in the hypertrophied heart.
AMPK is a metabolic sensor that limits anabolic processes when ATP production is insufficient to meet the energetic demands of the cell. Activated by increased AMP/ATP ratio, AMPK reduces energy-consuming processes, such as protein synthesis by inhibiting the mTOR signaling pathway (10,13,28). In the heart, deletion of the AMPK ␣ 2 catalytic subunit exacerbates cardiac hypertrophy and heart failure in response to pressure overload, and this is associated with up-regulated mTOR signaling (46). While mTOR signaling is important for the hypertrophic response (31), recent evidence suggests mTOR also has antiapoptotic, anti-inflammatory, and metabolic functions that may be important for adaptive hypertrophy (30,33,45). Thus, while increased mTOR signaling may contribute to cardiac hypertrophy, it is not sufficient to explain the worsened cardiac function in AMPK ␣ 2 KO hearts after TAC.
In addition to regulating protein synthesis, AMPK can regulate the cytoskeleton. AMPK is a member of the MARK/PAR kinase subfamily (microtubule affinity-regulating kinase), which phosphorylates MAPs (microtubule-associated proteins) to influence MT dynamics and cell polarity (18). AMPK was recently shown to directly phosphorylate the microtubuleassociated protein tau and alter its binding to neuronal microtubules (39). In addition, AMPK phosphorylation of the plus end tracking protein clip-170 increased MT polymerization speed and was found necessary for directed cell migration, while reduced CLIP-170 phosphorylation by AMPK resulted in less dynamic MTs (20). Whether AMPK regulates MT dynamics in cardiomyocytes, however, and how this influences the hypertrophic response is not known.
Here, we investigated AMPK regulation of cardiac microtubules during hypertrophy of cultured rat neonatal cardiomyocytes and in pressure overload-induced cardiac hypertrophy using the TAC model in mice. Our data identify, for the first time, the cardiac microtubule cytoskeleton as a sensitive target of AMPK activity, and suggest a novel role for AMPK in limiting accumulation and densification of microtubules in response to pressure overload.

MATERIALS AND METHODS
Animals and TAC. Animals were housed in an air-conditioned room with a 12:12-h light-dark cycle, received standard rodent chow, and drank tap water. AMPK ␣2 KO and control wild-type mice were generated and used for sham or subjected to TAC, as previously described (11). This study was approved by the Institutional Animal Care and Use Committee of University of Minnesota.
Echocardiography. Mice were anesthetized with 1.5% isoflurane, and echocardiographic images were obtained with a Visualsonics high-resolution Veve 660 system, as previously described (46). Left ventricular (LV) diameter, shortening fraction, and wall thickness were measured from two-dimensional guided short-axis M-Mode views of the LV.
Neonatal rat ventricular cardiomyocyte culture and protein analysis. Neonatal rat ventricular cardiomyocytes (NRVMs) were isolated from 2-day-old Sprague-Dawley rats by enzymatic digestion and separated from nonmuscle cells on a discontinuous Percoll gradient. NRVMs were isolated from 2-day-old Sprague-Dawley rats by enzymatic digestion and separated from nonmuscle cells on a discontinuous Percoll gradient, as previously described (7). A total of 2-4 million viable myocytes were isolated per ventricle with very little fibroblast contamination (Ͻ2%) and plated on gelatin-coated plates in serum-containing DMEM (0.5-1 ϫ 10 5 cells/cm 2 ) and incubated for 48 h to allow attachment and spreading, after which, the medium was replaced with serum-free media for 24 h prior to treatment. Regarding adenovirus infection, one day after plating, cardiomyocytes were infected with the indicated multiplicity of infection of constitutively active AMPK (caAMPK), WT AMPK (8), or ␤-Gal-expressing adenovirus. After 8 h to allow infection, media were changed, and cells were allowed to grow for an additional 16 h in serum before replacement with serum-free media and treatments.
Knockdown of microtubule-associated protein. For RNAi treatment, cells were plated at 1 ϫ 10 5 cells/cm 2 . After 24 h, cells were transfected overnight using Lipofectamine 2000 (Invitrogen) with 10 pm/well rat microtubule-associated protein 4 (MAP4) siRNA or nontargeting control siRNA (Dharmacon, Lafayette, CO) in the continued presence of 10% FCS. The next day, the media were replaced with DMEM containing insulin/transferrin/selenium solution (Sigma) and incubated for an additional 24 -48 h before replacement with growth factor-free media (DMEM) for 24 h before treatment with phenylephrine (PE) or 5-aminoimidazole carboxamide ribonucleotide (AICAR).
Isolation of cytoskeletal fractions from NRVMs. Cell lysates were collected in Triton lysis buffer and separated into soluble (cytosol and membrane) and insoluble (cytoskeletal) fractions, as previously described (7) for Western blot analysis. Regarding immunofluorescence staining of microtubules, NRVMs were fixed in ice-cold methanol for 10 min, permeabilized with 0.2% Triton/PBS, blocked 30 min at RT with 2% FCS/PBS, and incubated with primary antibodies overnight. Primary antibodies were detected using Alexa Fluor 488 or Alexa Fluor 555 secondary antibodies (Molecular Probes, Carlsbad, CA). For cell area measurement, cells were plated at 0.5 ϫ 10 5 cells/cm 2 . After treatment, cells were fixed with 10% methanol as for immunofluorescence and then stained using 0.3% Coomassie blue for 5 min, followed by three rinses in PBS and photographed under phase contrast. Cell area was analyzed using Image J 1.34s (U.S. National Institutes of Health, Bethesda, MD). At least 100 individual cells were measured per experiment.
Leucine incorporation by cardiomyocytes. One microcurie of [3-H]leucine (Perkin Elmer, Waltham, MA) was added per well in 24-well plates. After 48 h of treatment, cells were washed once in PBS, fixed 10 min in 10% TCA, washed again in TCA, dissolved in 5% SDS, 0.1 M NaOH, and [ 3 H] incorporation was measured in a liquid scintillation counter.
Isolation of microtubules and other fractions from mouse cardiac tissue. Mouse hearts were processed for separation of free and polymerized tubulin, as previously described (41), with minor modifications. Mouse hearts were snap frozen at the time of collection, then pulverized under liquid nitrogen. Approximately one-half of the pulverized tissue was homogenized in 1 ml of MT stabilization buffer at 20,000 rpms for 20 s and processed for separation of free and polymerized tubulin, as previously described (41). The other half of frozen powdered tissue was homogenized in 1% Triton X-100, 1 mM EDTA, 100 mM NaCl, 1 mM vanadate, 1 mM PMSF, 1 ϫ protease inhibitor cocktail (Boehringer Mannheim, Mannheim, Germany), and 10% glycerol, in 10 mM Tris·HCl at pH 7.4. Lysates were centrifuged at 14,000 g, and Triton-soluble supernate was separated from the Triton-insoluble pellet. The pellet was washed again in lysis buffer, and then resuspended in 2 ϫ SDS loading buffer containing phosphatase and protease inhibitors and 2.5% ␤-mercaptoethanol and boiled. For protein concentrations, lysates were precipitated using 10% TCA, and measured using the BCA assay (Pierce Biotechnology, Rockford, IL). Data and statistical analyses. All values were expressed as means Ϯ SE. Statistical significance was defined as P Ͻ 0.05. One-way ANOVA was used to test each variable for differences among the treatment groups with StatView (SAS Institute, Cary, NC). If the ANOVA demonstrated a significant effect, post hoc pairwise comparisons were made with the Student's t-test.

RESULTS
AMPK regulation of cell size, protein synthesis, and microtubule dynamics. Cell enlargement requires increased protein synthesis, as well as coordinated expansion of the cytoskeletal networks that support the new cell dimensions. While investigating the role of AMPK in cardiomyocyte hypertrophy, we observed that AICAR treatment was able to limit cell area at a low dose (0.1 mM) that did not block protein synthesis ( Fig. 1D-F). This suggests an AICAR effect on cell spreading or cell shape that may occur independently of the well-known AMPK inhibitory effects on protein synthesis. Because cell spreading is mediated by changes in cytoskeletal dynamics, we examined expression and partitioning (cytosolic vs. cytoskeletal) of cardiomyocyte cytoskeleton proteins in response to PE stimulation and whether this is altered by AICAR treatment. Gentle lysis of cardiomyocytes into Triton-soluble (membrane and cytosol; sol) and insoluble (cytoskeleton and nucleus; csk) fractions revealed that activation of AMPK with AICAR had little effect on PE-induced assembly of myofilaments (csk-sarcomeric actin or ␤-actin) or intermediate filaments (csk-desmin), while it dosedependently reduced microtubule (csk-tubulin) levels ( Fig. 1, A-C). Examination of stabilized MTs using antibodies to detyrosinated tubulin (the c-terminus tyrosine is cleaved in longer-lived polymerized microtubules, leaving a c-terminal glutamine, Glu-tubulin) confirmed that microtubule longevity was much reduced in the presence of AICAR (Fig. 1, A-C). Interestingly, MT stability was more sensitive to AMPK activation than inhibition of mTOR, EEF2, or protein synthesis, as AICAR attenuated microtubule stability and cell area (Fig. 1,  A, B, and E and F) at doses (0.1 mM) that did not block mTORC1 signaling, EEF2 phosphorylation (Fig. 1, A and B), or protein synthesis (Fig. 1D). Similar to low-dose AICAR treatment, nocodazol inhibition of microtubule growth reduced cell area without inhibiting protein synthesis (Fig. 1, D and E), indicating that blocking microtubule proliferation can attenuate cell spreading, independent of overall changes in protein synthesis. One explanation for the decreased cell area despite equal levels of protein synthesis is that MT-dependent distribution of proteins or organelles to the cell periphery requires more persistent elongated microtubule tracks, which were diminished in the presence of AICAR. This possibility is supported by Coomassie staining of cellular protein distribution in fixed cells (Fig. 1F), which shows proteins more highly concentrated in perinuclear regions of cells treated with AICAR or nocodazole, while cells treated with PE alone have more evenly distributed proteins that reach the cell periphery.
To rule out the possibility that reduction in MT stability is an indirect effect of less hypertrophy, cardiomyocytes were treated with phenylephrine for 72 h to establish hypertrophy and microtubule densification, then treated briefly (3 h) with AICAR. Three hours of AICAR treatment reduced Glu-tubulin levels (Fig. 2, A-C), but it did not significantly reduce overall cytoskeletal tubulin, indicating that AMPK reduction of microtubule stability can occur independent of AMPK effects on hypertrophy, or on overall microtubule levels.
To determine whether the reduction of MTs is due to activated AMPK rather than an indirect effect of AICAR, cells were infected with adenovirus expressing constitutively active AMPK (CA-AMPK) for 24 h, then treated with phenylephrine for an additional 48 h. Similar to AICAR treatment, CA-AMPK reduced microtubule density and Glu-tubulin levels in phenylephrine-treated cells as demonstrated by immunofluorescence staining (Fig. 2D). Similar to low-dose AICAR treatment, CA-AMPK reduced cytoskeletal tubulin levels (Fig. 2E) and cell area (Fig. 2F) without inhibiting protein synthesis (Fig. 2G). These data indicate that AMPK regulates MT dynamics, which can influence cell spreading independent of protein synthesis.
To further examine the role of AMPK in microtubule regulation, we overexpressed wild-type AMPK-␣ 2 in neonatal cardiomyocytes and examined whether this influences phenylephrine-induced MT stabilization and accumulation. Adenoviral overexpression of WT AMPK resulted in spontaneously increased AMPK Thr172 phosphorylation, a reduction in stabilized microtubules (detyrosinated tubulin), and a reduction in tubulin association with the Triton-insoluble cytoskeleton, similar to the effects of activating AMPK using AICAR (Fig. 2, H  and I), These data are further evidence that AMPK regulates microtubule stability in cardiomyocytes.
The finding that MTs are a sensitive target of AMPK in neonatal cardiomyocytes led us to investigate whether AMPK may influence cardiac MT dynamics in vivo, where microtubule densification is believed to contribute to pressure overload-induced heart failure.
AMPK regulation of cardiac MT dynamics in pressure overload hypertrophy. We have previously demonstrated that TAC-induced hypertrophy, LV dysfunction, and heart failure are exacerbated by AMPK ␣ 2 KO (46). Because we observed that AMPK activation can reduce MT stabilization in neonatal cardiomyocytes exposed to hypertrophic signals, we examined the influence of AMPK ␣ 2 (AMPK ␣ 2 KO), the predominant AMPK catalytic subunit in the heart (34), on cardiac microtubule dynamics in mice under basal conditions or hypertrophic stress (transverse aortic constriction, TAC) (Fig. 3, A and B). Interestingly AMPK ␣ 2 KO mice had ϳ60% increase in polymerized tubulin levels compared with WT mice under basal conditions. Three weeks of TAC resulted in further increases in free and polymerized tubulin levels in both WT and AMPK ␣ 2 KO mice, but the increase in polymerized and free tubulin was significantly greater in the AMPK ␣ 2 KO mice. Detyrosinated tubulin (Glu-tubulin) levels were also significantly higher in AMPK ␣ 2 KO mice exposed to TAC, indicating an increase in microtubule stability. Interestingly, in AMPK KO mice, the levels of polymerized tubulin exhibited a striking correlation with increased lung to body weight ratio (indicative of pulmonary congestion) and reduced ejection fraction (indicative of contractile dysfunction) (Fig. 3, C and  D). Taken together, these results suggest AMPK ␣ 2 plays a role in limiting pressure overload-induced accumulation and stabilization of cardiac microtubules, and they identify an association between increased MT accumulation and left ventricular dysfunction.
AMPK regulation of microtubule-associated protein 4. To investigate mechanism(s) of AMPK regulation of cardiac MTs, we examined MAP4. MAP4 binds to and stabilizes microtubules in response to pressure overload, but its affinity for microtubules is reduced by phosphorylation at Ser-914 (ms) [equivalent to Ser-915 (rat), Ser-924 (cat), and Ser-941(hu) (4,6,12)]. To determine whether MAP4 expression or phosphorylation is regulated by AMPK ␣ 2 in vivo, we analyzed the Triton-soluble and insoluble ventricular lysates of WT and AMPK ␣ 2 KO mice exposed to sham or TAC conditions. In Triton X-100 tissue homogenates, mechanical homogenization in the presence of detergent, phosphatase inhibitors, and low temperature causes most of the tubulin to fractionate with the Tritonsoluble fraction. Interestingly, most of the MAP4 protein partitioned in the Triton-insoluble fraction (not shown). This suggests that MAP4 also associates with other more stable or Tritoninsoluble cytoskeletal elements of the cell in addition to microtubules, consistent with a recent report demonstrating MAP4 also binds to actin (19). Under basal conditions, insoluble MAP4 levels were significantly higher in AMPK ␣ 2 KO hearts than WT, while phosphorylation of insoluble MAP4 relative to total MAP4 was reduced in AMPK ␣ 2 KO samples (Fig. 4, A and  B). These differences in expression level or phosphorylation were not observed after TAC, however, whereas MAP4 expression increased in both WT and AMPK ␣ 2 KO (albeit with a higher relative increase in WT hearts compared with the low levels under sham conditions). In response to TAC, phosphorylation of Triton-insoluble MAP4 Ser914 was reduced as previously observed in cats (4), with no significant difference between WT and KO. In the Triton-soluble fraction, however, MAP4 Ser914 phosphorylation was increased in WT samples, while the increase in MAP4 Ser914 phosphorylation was significantly attenuated in AMPK ␣ 2 KO, consistent with a role for AMPK ␣ 2 in promoting MAP4 phosphorylation and reducing MT stability.
To determine whether activation of AMPK can increase MAP4 phosphorylation, we treated prehypertrophied (after 48 h of phenylephrine exposure) neonatal cardiomyocytes for an additional 24 h with phenylephrine in the absence or presence of 0.2 mM AICAR. AICAR exposure resulted in increased MAP4 phosphorylation and a trend toward reduced total MAP4 levels. At the same time, stabilized MTs, but not total MTs, in the same fraction were dramatically reduced, consistent with a role for MAP4 in MT stabilization. (Fig. 5, A and  B). Recently, AMPK was found to phosphorylate serine-262 within the KXGS motif in the microtubule-associated protein tau (39). Ser-262 in human tau is analogous to Ser-914 in the KXGS motif in mouse MAP4 (87% homology in the 22 amino acids surrounding Ser-914). To determine whether AMPK can directly phosphorylate Ser-914 of MAP4 in vitro, we incubated active WT, AMPK ␣ 2 /␤ 2 /␥ 2 with a peptide corresponding to amino acids 903 to 925 in mouse MAP4 and used LCMS to detect phosphorylation of the peptide. LCMS analysis showed that this peptide was monophosphorylated by AMPK on serine-914 within the KXGS motif (data not shown), suggesting AMPK can directly phosphorylate MAP4 at serine-914.
To verify the role of MAP4 in cardiomyocyte microtubule stability, we used RNAi to reduce MAP4 expression. MAP4 RNAi treatment of neonatal cardiomyocytes reduced stabilized MT levels (Glu-tubulin), compared with control nontargeting RNAi (Fig. 5, C and D). Interestingly, MAP4 depletion also altered MT organization, so that microtubules were sparsely distributed in some areas of the cell, while densely aligned along the cell periphery in others. Some cells treated with MAP4 RNAi also exhibited abnormal nuclear positioning, so that nuclei were observed along the cell periphery rather than the center (Fig. 5C). MAP4 also has been shown to bind to and regulate actin filaments (19), so MAP4 may influence cytoskeletal organization and nuclear positioning through effects on actin and/or microtubules. Interestingly, AICAR treatment still further reduced stabilized MTs in MAP4-deficient cells (Fig. 5D), suggesting AMPK activation may also reduce MT stabilization through mechanisms that are independent of MAP4.  Fig. 3. In response to pressure overload, AMPK ␣2 knockout (KO) mice show an exacerbated microtubule stabilization and accumulation, which correlates with contractile dysfunction. Free and polymerized MT fractions were isolated from WT and AMPK ␣2 KO mice after 3 wk of transverse aortic constriction (TAC) and analyzed by Western blot for ␣-tubulin or Glu-␣-tubulin (Fig. 5, A  and B). Lung weight to body weight ratio (LW/BW) (C) or % left ventricular ejection fraction (EF; D) of WT () and AMPK KO mice (OE) exposed to TAC were plotted vs. polymerized tubulin levels from the corresponding hearts. *P Ͻ 0.05 comparing WT to KO under same conditions. †P Ͻ 0.05 comparing sham to TAC.

DISCUSSION
AMPK has long been recognized as a metabolic sensor that reduces energy-consuming processes (such as protein synthesis), while increasing energy production (glucose transport, fatty acid oxidation, mitochondria biogenesis). Our data suggest a novel role for cardiac AMPK in attenuating the stabilization and densification of MTs that occurs in response to hypertrophic stress. AMPK prevention of microtubule proliferation can limit expansion of cell area in cultured cells, independent of previously identified AMPK inhibitory effects on protein synthesis. More importantly, our data suggest AMPK limits microtubule stabilization that occurs in response to pressure overload, where microtubule proliferation is believed to contribute to heart failure.
Microtubules and heart failure. There is evidence that microtubule densification contributes to contractile dysfunction through mechanical impairment of contractility (15,36,41). Reducing the viscous load upon myofilaments imposed by microtubules in the heart may allow AMPK to improve contractile function without additional energy expenditure. However, because microtubules play an important role in intracellular transport and subcellular organization, altered microtubule dynamics in pressure overload could prove maladaptive through other mechanisms in addition to increasing myofilament load. For instance, microtubule decoration by MAP4 (which plays a role in pressure overloadinduced MT stabilization) can inhibit transport of receptors (2) and mRNAs (25,26). Disruption of mRNA and protein transport may impair compensatory hypertrophy and pro-mote the transition to decompensation, which is associated with increased MT accumulation in pressure overload-induced heart failure (35). Microtubules also play an important role in organelle transport and distribution (1,9,38,40,44), and decoration of microtubules by MAPs may, thus, alter intracellular organization. Interestingly, microtubules were recently shown to mediate stretch-induced NADPH oxidase activation, oxidative stress, and calcium sparks (14,23), suggesting a potential role for increased MTs in oxidative stress and altered calcium dynamics observed in heart failure. In support of a role for MT accumulation in contractile dysfunction in AMPK KO mice, we observed that polymerized microtubule levels correlated strikingly with reduced left ventricular ejection fraction and pulmonary congestion in mice exposed to TAC. Furthermore, our previous study demonstrated that periodic colchicine treatment attenuated chamber dilation and improved contractile function in Balb-c mice exposed to pressure overload (7). The present data suggest that agents that activate AMPK may similarly preserve contractile function by preventing MT densification.
How does AMPK regulate microtubules? Recently, a mechanism for MT stabilization during pressure overload hypertrophy was identified. MAP4, which is upregulated in response to pressure overload (24), binds to and promotes MT stabilization (12,37). However, when phosphorylated at Ser-924 of the KXGS motif (equivalent to Ser-914 in mice) within the MT binding domain, this interaction is inhibited and MT stability is reduced (4). Cheng et al. (3) showed that increased phospha- tase activity during pressure overload results in dephosphorylation of MAP4 Ser924 , thereby promoting MT assembly and stabilization. We find that activation of AMPK increases MAP4 phosphorylation at the analogous site in rat cardiomyocytes. Conversely, MAP4 phosphorylation at this site is reduced in AMPK ␣ 2 KO mice. Thus, it appears that AMPK plays a role in counteracting the effects of PP2A on MT stability by reducing MAP4 affinity for MTs. AMPK is a member of the MARK/PAR kinase subfamily of Ser/Thr kinases, some of which have been shown to directly phosphorylate MAP4 on KXGS motifs to reduce affinity for microtubules (5,12). The MAP4 kinases in the heart have not been identified. It was recently demonstrated that AMPK can directly phosphorylate Ser-262 of the KXGS motif in the microtubule binding domain of tau, the neuronal MAP associated with Alzheimer's disease (39). The 22 amino acids flanking Ser-914 in MAP4 and Ser-262 in tau are ϳ87% homologous (19/22). In support of direct phosphorylation of MAP4 by AMPK, we find that purified active AMPK can phosphorylate this same site in vitro. Because MAP4 phosphorylation was not completely removed in AMPK KO mice, it is likely that other kinases also regulate MAP4 phosphorylation in the heart. This could be due to AMPK ␣ 1 activity, or the activity of an AMPK-related kinase, such as MARK2. MARK2 exhibits reduced activity in response to right ventricular pressure overload in cats at the same time that MAP4 phosphorylation is reduced (3). However, in this mouse model, we did not observe reduced MARK2 activation (as indicated by phosphorylation at Ser-215 in the activation loop; not shown) in response to TAC. Although our results suggest a role for AMPK in cardiac MAP4 phosphorylation, it will be important to determine the role of other MARKs and AMPK-related kinases in cardiac MAP4 phosphorylation, as well as the counterpoised MAP4 phosphatases (PP2A), to acquire a more complete understanding of the signaling pathways that regulate MAP4 and cardiac MT dynamics.
In addition to phosphorylation of MAP4, AMPK may influence MT dynamics through MAP4-independent mechanisms. This is suggested by the reduction in MT stability by AICAR, even when MAP4 was eliminated using RNAi. AMPK has also been shown to regulate microtubule dynamics through phosphorylation of clip 170 (21). However, RNAi depletion of clip 170 did not result in obvious effects on MT stability (not shown). We did observe a slight reduction in free tubulin in cells treated with AICAR, while total tubulin was higher in AMPK KO mice after TAC. Persistently high levels of tubulin synthesis are necessary for MT growth and stability, so it is possible that AMPK reduction of tubulin heterodimer synthesis may also contribute to reduced MT levels. However, we still observe rapid MT destabilization in prehypertrophied cells, even as free or polymerized tubulin levels change only slightly, so other AMPK-dependent mechanisms of MT destabilization are also likely to be involved. While our data identify MAP4 phosphorylation as one mechanism by which AMPK activity may promote MT destabilization, additional experiments will be necessary to identify the MAP4-independent mechanisms of cardiac MT regulation by AMPK.
In summary, we have identified a new role for AMPK in limiting stabilization and accumulation of microtubules during pressure overload-induced cardiac hypertrophy. Future studies will be designed to determine whether treatments that activate cardiac AMPK can prevent microtubule accumulation during pressure overload, and the extent to which AMPK regulation of microtubule dynamics contributes to the cardioprotective effects of AMPK.