Caveolin-3 Knock-out Mice Develop a Progressive Cardiomyopathy and Show Hyperactivation of the p42/44 MAPK Cascade*

A growing body of evidence suggests that muscle cell caveolae may function as specialized membrane microdomains in which the dystrophin-glycoprotein complex and cellular signaling molecules reside. Caveolin-3 (Cav-3) is the only caveolin family member expressed in striated muscle cell types (cardiac and skeletal). Inter-estingly, skeletal muscle fibers from Cav-3 ( (cid:1) / (cid:1) ) knock-out mice show a number of myopathic changes, consistent with a mild-to-moderate muscular dystrophy phenotype. However, it remains unknown whether a loss of Cav-3 affects the phenotypic behavior cardiac myocytes in vivo . Here, we present a detailed characterization of the hearts of Cav-3 knock-out mice. We show that these mice develop a progressive cardiomyopathic phenotype.

Interestingly, Cav-3 expression is necessary for caveolae formation in skeletal muscle fibers (7). Individual Cav-3 molecules homo-oligomerize to form high molecular mass multimers (ϳ14 -16 monomers per oligomer), both in vitro and in vivo (2). This self-assembly is thought to drive the invagination of the plasma membrane through the interaction of caveolin oligomers with cholesterol, sphingolipids, and other membrane protein components.
Cav-3 expression is detectable at embryonic day 10 in mouse heart (8), and Cav-3 has been shown to associate with the developing T-tubule system in skeletal myoblasts (9). In addition, Cav-3 (Ϫ/Ϫ) knock-out mice demonstrate dilated and longitudinally oriented T-tubules in their skeletal muscle fibers (7). Likewise, the skeletal muscles of patients with mutations in the human CAV-3 gene (LGMD-1C) 1 also show a disorganized T-tubule network (10).
Members of the dystrophin-glycoprotein (DG) complex have been shown to localize to muscle caveolae (6,7). Although not an integral member of the DG complex (11), (i) Cav-3 can directly interact with ␤-dystroglycan, (ii) Cav-3 is necessary for the localization of some DG complex members to lipid raft domains/caveolae in skeletal muscle fibers, and (iii) Cav-3 expression increases with the loss of dystrophin, as in mdx mice and Duchenne muscular dystrophy (7,(12)(13)(14). Thus, Cav-3 appears to dynamically interact with the DG complex. As a consequence, it is not surprising that mutations in many dystrophin-associated proteins, such as Cav-3, lead to similar forms of muscular dystrophy (15)(16)(17).
Caveolae have also been shown to function as "pre-assembled" signaling complexes through the compartmentalization of signaling molecules that interact with the caveolin proteins and/or "liquid-ordered" caveolar lipids (18). In the heart, a variety of signaling molecules co-fractionate with cardiac caveolae, and their residence in caveolae, or movement out of caveolae, is important for their function (19 -27). Furthermore, multiple studies have now shown that Cav-3 expression is dramatically decreased in different models of cardiac hypertrophy (25,26,28). This suggests that reduction of Cav-3 expression may be a pivotal event in the ensuing hypertrophic program, perhaps by allowing hypertrophy-inducing signaling cascades to remain constitutively activated.
Although a role for Cav-3 in multiple skeletal muscle processes has now been investigated, the functional role of Cav-3 in the heart remains unknown. Here, we present a thorough characterization of the hearts of Cav-3 (Ϫ/Ϫ) knock-out mice. Interestingly, we show that Cav-3 knock-out mice develop a progressive, mild-to-moderate, cardiomyopathic phenotypecharacterized by myocyte hypertrophy. Thus, loss of Cav-3 expression and cardiac myocyte caveolae is sufficient to induce the activation of a hypertrophic program in cardiac myocytes.

Animal Studies
Mice were housed and maintained in a barrier facility at the Institute for Animal Studies, Albert Einstein College of Medicine. The generation of Cav-3 KO mice was as we previously described (7).

Preparation of Caveolae-enriched Membrane Fractions
WT or Cav-3 KO hearts were harvested, minced with a razor blade, and homogenized for 30 s using a Polytron tissue grinder in 2 ml of MES-buffered saline with 1% (v/v) Triton X-100, at 4°C. Samples were centrifuged (1000 ϫ g for 5 min at 4°C), and the supernatant was adjusted to 40% sucrose by the addition of 2 ml of 80% sucrose in MES-buffered saline. A 5-30% linear sucrose gradient was formed above the homogenate and centrifuged at 39,000 rpm for 16 h in a SW41 rotor (Beckman Instruments). A light-scattering band in the ϳ15-20% sucrose region was observed. Twelve 1-ml fractions were collected, starting from the top of the gradient. For SDS-PAGE/Western blotting, an equal amount of total protein from each fraction (25 g) was analyzed.

Transmission Electron Microscopy
Heart tissue samples were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate buffer, post-fixed with OsO 4 , and stained with uranyl acetate and lead citrate. Microtome sections were examined under a Jeol 1200 EX transmission electron microscope and photographed at a magnification of 15,000ϫ. Caveolae were identified by their characteristic flask shape, size (50 -100 nm), and location at or near the plasma membrane.

Preparation of Heart Paraffin Sections
Mice were sacrificed, and their hearts were removed and placed in buffered formalin (10%). The tissue was fixed for ϳ24 h, washed in PBS for 20 min, dehydrated through a graded series of ethanol washes, treated with xylene for 40 min, and then embedded in paraffin for 1 h at 55°C. Paraffin-embedded 5-m-thick sections were then prepared using a Microm (Baxter Scientific) microtome and placed on super-frost plus slides (Fisher). Slides were then stained with hematoxylin and eosin (H & E) or Trichrome, according to standard laboratory protocols. Samples were examined by an experienced cardiac pathologist (Dr. Stephen M. Factor).

Non-invasive Cardiac Imaging
Gated Cardiac Magnetic Resonance Imaging-MRI experiments were performed using a General Electric Omega 9.4T vertical bore MR system equipped with a micro-imaging accessory and custom-built coils designed specifically for mice, as described previously (31). Just prior to each image acquisition, the heart rate was determined from the electrocardiogram, and the spectrometer gating delay was set to acquire data in diastole and systole. Multislice spin-echo imaging with an echo time of 18 ms and a repetition time of ϳ100 -200 ms was performed. A 35-mm field of view (with a 256-ϫ 256-pixel image matrix) was used. Short and long axis images of the heart were acquired, and MRI data were processed off-line with MATLAB-based custom-designed software.
Transthoracic Echocardiography-Transthoracic echocardiography was performed as described previously (32). Echocardiography was performed with mice in a supine position on a heating pad set at 38°C. Light anesthesia was achieved using isoflurane inhalation. Continuous, standard electrocardiograms were taken from electrodes placed on the extremities. Echocardiographic images were obtained using an annular array, broadband, 10/5-MHz transducer attached to an HDI 5000 CV ultrasound system (Advanced Technology Laboratories, Bothell, WA). A small gel standoff was placed between the probe and chest. Two-dimensional and M-mode images of the heart were obtained from the basal short axis view of the heart and stored on 3/4-inch SVHS videotapes for off-line measurements using the Nova-Microsonic (Kodak) Imagevue DCR workstation (Indianapolis, IN). All measurements were made in three to six consecutive cardiac cycles, and the averaged values were used for analysis. Left ventricular end-diastolic and end-systolic diameters as well as diastolic ventricular septal and posterior wall thicknesses were measured from M-mode tracings. Diastolic measurements were performed at the point of greatest cavity dimension, and systolic measurements were made at the point of minimal cavity dimension, using the leading edge method of the American Society of Echocardiography (33). Additionally, the following parameters were calculated using the above-mentioned measurements: left ventricular diastolic FIG. 1. Cav-3 KO mice do not express caveolin-3 in the heart and lack cardiac myocyte caveolae. A, Western blotting. Hearts were harvested from wild-type and Cav-3 KO mice. Tissue lysates were prepared (see "Experimental Procedures") and subjected to SDS-PAGE/transfer to nitrocellulose. Blots were probed with isoform-specific mAbs that selectively recognize either caveolin-1, caveolin-2, or caveolin-3. Note that there is a complete loss of caveolin-3 in Cav-3 KO mouse hearts, without any changes in the expression levels of caveolin-1 or caveolin-2 as compared with wild-type (WT) control mice. Immunoblotting with anti-actin IgG is shown as a control for equal protein loading. B, immunostaining (cross-sections). Frozen sections of the heart were prepared from wild-type and Cav-3 KO mice and immunostained with antibodies directed against either caveolin-1, caveolin-2, or caveolin-3. Note that caveolin-3 is localized to the plasma membrane (sarcolemma) of wild-type cardiac myocytes (panel a) but is completely absent in heart tissue derived from Cav-3 KO mice (panel d). In contrast, caveolin-1 and caveolin-2 expression is exclusively restricted to the endothelium and endocardium (see white arrowheads) as expected and remains unchanged in Cav-3 KO heart tissue (caveolin-1, panels b and e; caveolin-2, panels c and f). chamber, left ventricle chamber. C, immunostaining (longitudinal sections). Frozen sections of the heart were prepared from wild-type mice and immunostained with antibodies directed against caveolin-3. The fluorescence image and the corresponding phase image are shown; arrowheads indicate the Z-lines. Note that the Z-lines identified in the phase image clearly coincide with the immunostaining pattern observed for caveolin-3. D, cell fractionation. Hearts were harvested from wild-type and Cav-3 KO mice. Tissue lysates were prepared and subjected to sucrose density gradient analysis. In the wild-type heart, note that caveolin-1 and caveolin-3 are localized to the "light" buoyant density area of the gradient that contains lipid rafts/caveolae (fractions 5 and 6). In Cav-3 KO heart, caveolin-1 remains localized to the lipid raft/caveolae fractions as expected, because caveolin-1 is expressed in the endothelium and endocardium of the heart. E, transmission electron microscopy. Heart tissue samples were fixed and embedded as described under "Experimental Procedures." Caveolae were identified by their characteristic flask shape, size (50 -100 nm), and location at or near the plasma membrane. Arrowheads indicate detached caveolae, whereas arrows are used to indicate caveolae that remain attached to the plasma membrane. Note that in wild-type animals caveolae are present in both the cardiac myocyte and adjacent endothelial cell. In contrast, in Cav-3 KO animals there is a selective loss of muscle caveolae in the cardiac myocyte, whereas the adjacent endothelial cell retains its non-muscle caveolae. Myo, cardiac myocyte; Endo, endothelial cell; Lumen, blood vessel lumen that may contain red blood cells.

Cardiomyopathy in Caveolin-3 Null Mice 38991
wall thickness was calculated as the average of ventricular septal and left ventricular posterior wall thicknesses; left ventricular percent fractional shortening was measured using the equation, 100 ϫ [(end-diastolic diameter minus end-systolic diameter)/end-diastolic diameter]; and relative wall thickness was measured using the equation, (2 ϫ left ventricular diastolic wall thickness)/end-diastolic diameter. Note that differences between the "absolute" wall thicknesses measured using MRI and echocardiography are commonly observed and are likely due to technical factors such as differences in the time of gating; echocardiography may underestimate these values, whereas MRI may overestimate these values. Most importantly, however, the relative changes measured in left ventricular wall thickness using MRI and echocardiography are in agreement.

Immunoblotting with Phospho-specific Abs
For the analysis of phospho-proteins, WT and Cav-3 KO hearts were harvested, quickly rinsed in PBS (1ϫ), and immediately frozen in liquid nitrogen. Frozen hearts were then homogenized in 3 ml of boiling lysis buffer (10 mM Tris, pH 7.4, 1% SDS, 1.0 mM sodium orthovanadate), heated in a microwave for 15 s, and centrifuged for 5 min at 16,000 ϫ g to pellet any insoluble material. The supernatant was transferred to a new tube, and aliquots (1:10 dilution) were prepared for protein concentration analysis. Twenty micrograms of protein was run on SDS-PAGE gels and analyzed by Western blotting.

Blood Pressures
Blood pressure measurements were taken on both WT and Cav-3 KO mice by placing them in a mouse restrainer (RTBP007, Kent Scientific, Inc.) and applying a mouse occlusion cuff (RTBP050, Kent Scientific, Inc.) and mouse plethysmographic cuff (XBP051, Kent Scientific, Inc.) to the tail of the mouse. A heat lamp was used to warm the mice. Systolic and diastolic blood pressure measurements were taken using the XBP1000 apparatus (Kent Scientific, Inc.) connected to a data acquisition system. The occlusion pressure corresponding to the minimal and maximal plethysmographic signals were taken to be the systolic and diastolic pressures, respectively.

Cav-3 KO Mice Do Not Express Caveolin-3 in the Heart and
Lack Cardiac Myocyte Caveolae-Hearts were harvested from WT and Cav-3 KO mice and examined by Western blot analysis (Fig. 1A). Note that genetic ablation of the Cav-3 gene resulted in the complete loss of the Cav-3 protein from the heart. However, the expression levels of Cav-1 and Cav-2 were unaffected by a lack of Cav-3 expression. Immunoblotting with anti-actin IgG was also performed as a control for equal protein loading.
Immunofluorescent microscopic analysis showed that Cav-3 localized to the plasma membrane of WT cardiac myocytes, whereas Cav-1 and Cav-2 were found exclusively in the endocardium and the endothelium of the heart (Fig. 1B, panels a-c). As expected, Cav-3 expression was not detectable in sections of Cav-3 KO hearts; most importantly, Cav-1 and Cav-2 expression remained properly restricted to the endocardium and endothelium in Cav-3 KO hearts, indicating that there was no   compensatory up-regulation of Cav-1 and Cav-2 in Cav-3 null cardiac myocytes (Fig. 1B, panels d-f). Longitudinal sections demonstrate that Cav-3 co-localized with the Z-line patterning of the myocardium as well as the plasma membrane (Fig. 1C). This is consistent with an association between Cav-3 and the T-tubule system in cardiac myocytes, because the T-tubule system is in register with the Z-lines in the heart, as opposed to the A-I bands within skeletal muscle. Due to their "liquid-ordered" and "buoyant" properties, caveolae can be isolated by tissue solubilization in the detergent Triton X-100 at 4°C, followed by sucrose density-gradient centrifugation (34). In wild-type hearts, note that caveolin-1 and caveolin-3 were localized to the "light" buoyant density area of the gradient that contained lipid rafts/caveolae (Fig. 1D, fractions  5 and 6). In Cav-3 KO hearts, caveolin-1 remained localized to the lipid raft/caveolae fractions as expected, because caveolin-1 is expressed in the endothelium and endocardium of the heart.
Transmission electron microscopy revealed that caveolae are found in both the endothelium and myocardium of wild-type mouse hearts (Fig. 1E). However, disruption of the Cav-3 gene resulted in the complete loss of caveolae only in the cardiac myocytes, thus demonstrating at the structural level the necessity of Cav-3 expression for the formation of cardiac myocyte caveolae. However, the morphology and number of Cav-1-generated caveolae within the endothelium was unaffected.
Cav-3 KO Mice Show Left Ventricular Wall Thickening, as Assessed by Gated Cardiac MRI-Wild-type and Cav-3 KO hearts were next analyzed using magnetic resonance imaging (MRI) (31). Measurements of left ventricular wall thickness were obtained for hearts in both the systolic and diastolic phases of the cardiac cycle. Fig. 2A shows representative short axis (transverse) images at the mid-level of wild-type and Cav-3 KO mice during diastole. At 2 months of age, the Cav-3 KO hearts showed a moderate increase (ϳ10%) in left ventricular wall thickness, as compared with age-matched wild-type control mice ( Fig. 2B and Table I). However, by 4 months of age, the Cav-3 KO hearts showed an even more dramatic increase in left ventricular wall thickness (ϳ20%), indicating that this is a progressive cardiomyopathic phenotype. Interestingly, the increase in left ven- FIG. 3. Histological examination of Cav-3 KO heart tissue reveals cardiac myocyte hypertrophy, interstitial/peri-vascular fibrosis, and cellular infiltrates. Representative H & E staining of WT (a and b) and Cav-3 KO (d and e) heart paraffin-embedded sections is shown. Note the marked hypertrophic cardiac myocytes (arrows) and cellular infiltrates (arrowheads) in the Cav-3 KO sample. Interestingly, Trichrome staining of 11-month-old WT (c) and Cav-3 KO (f) hearts reveals increased interstitial/peri-vascular fibrosis at the junction of the right and left ventricles in Cav-3 KO hearts at this age.  (32). Importantly, the heart rates of all animals tested were not statistically different, thereby allowing for meaningful comparisons. Multiple measurements of chamber size and wall thickness were made during both diastole and systole. At 4 months of age, the Cav-3 KO hearts showed a significant increase (ϳ20%) in left ventricular chamber diameter during diastole, as compared with age-matched wild-type control mice. During systole, the increase in Cav-3 KO left ventricular chamber diameter was even more pronounced (ϳ50%) (Table II). Interestingly, the 4-month-old Cav-3 KO hearts showed a marked increase in left ventricular chamber diameter during both diastole and systole, as compared with Cav-3 KO hearts at 2 months of age (diastole: 3.08 Ϯ 0.07 3 3.33 Ϯ 0.12 mm; systole: 1. 63 Ϯ 0.04 3 1.91 Ϯ 0.15 mm). Thus,

FIG. 4. Analysis of the dystrophin-glycoprotein complex in Cav-3 KO mouse hearts: ␣-Sarcoglycan is no longer properly targeted to lipid rafts in the absence of Cav-3.
A, immunostaining (cross-sections). Frozen sections of the heart were prepared from wild-type and Cav-3 KO mice and immunostained with antibodies directed against components of the dystrophin-glycoprotein (DG) complex, including dystrophin, the sarcoglycans (␣, ␤, ␦, and ␥) and ␤-dystroglycan. Note that there is no change in the expression levels or the distribution of the DG complex in Cav-3 KO cardiac myocytes. B, Western blot analysis. Hearts were harvested from wild-type and Cav-3 KO mice. Tissue lysates were prepared and subjected to SDS-PAGE/transfer to nitrocellulose. Note that the total amount of ␣-sarcoglycan remains unchanged in Cav-3 KO heart tissue. Immunoblotting with anti-actin IgG is shown as a control for equal protein loading. C, cell fractionation. Hearts were harvested from wild-type and Cav-3 KO mice. Tissue lysates were prepared and subjected to sucrose density gradient analysis. In the wild-type heart, note that a significant portion of total ␣-sarcoglycan is localized to the light buoyant density area of the gradient that contains lipid rafts/caveolae (fractions 5 and 6). In contrast, in the Cav-3 KO heart the distribution of ␣-sarcoglycan is altered; ␣-sarcoglycan is now excluded from the lipid raft/caveolae fractions.

2-4 months is an important time frame in the development of this cardiomyopathic phenotype.
Echocardiography was also utilized to determine other parameters of cardiac structure and function such as inter-ventricular septum as well as the posterior and anterior wall thicknesses. For each of these areas, the measured wall thickness was ϳ20% greater in the Cav-3 KO when compared with wild-type hearts (Table II). Because the increase in thickness was uniform for all the wall thicknesses measured, these data are consistent with an eccentric hypertrophy profile.
Functionally, Cav-3 KO hearts resulted in a decrease of ϳ20% in fractional shortening, consistent with the observed chamber dilation and increases in wall thickness. In addition, our blood pressure measurements showed that Cav-3 KO mice have normal diastolic and systolic blood pressures, thus ruling out the possibility of pressure-overload-induced cardiac hypertrophy (Table III).
Histological Examination of Cav-3 KO Heart Tissue Reveals Cardiac Myocyte Hypertrophy, the Presence of Cellular Infiltrates, and Progressive Interstitial/Peri-vascular Fibrosis-H & E-stained sections of Cav-3 KO hearts were examined at low and high magnification. No abnormal histopathological features were evident in 2.5-week-old Cav-3 KO hearts (not shown). By 2 months of age, however, Cav-3 KO hearts clearly showed hypertrophic myocytes and an increase in the overall number of nuclei per field, as compared with age-matched wild-type control hearts (Fig. 3, panels a, b, d, and e). Similar findings were also present at 4 and 11 months of age (not shown).
The identification of fibrosis at the junction of the left and right ventricles, a recognized point of stress, is common in both aging mouse and human hearts. Trichrome staining showed greater interstitial/peri-vascular fibrosis at this junction in older Cav-3 KO hearts, as compared with age-matched wildtype control mice (Fig. 3, panels c and f). However, signs of ischemia were not observed in Cav-3 KO hearts.
␣-Sarcoglycan, a Marker of the DG Complex, Is Specifically Excluded from Lipid Rafts Domains in the Absence of Cav-3-A significant fraction of the DG complex expressed in myocytes is localized within lipid rafts/caveolae microdomains (7,22). In addition, mice with null mutations in many of the DG complex proteins demonstrate a cardiomyopathic phenotype. Thus, we next examined the expression and localization DG complex members in Cav-3 KO hearts. However, immunofluorescence analysis clearly demonstrated that the expression levels and membrane localization of each of the DG complex members examined (dystrophin; ␣-, ␤-, ␦-, and ␥-sarcoglycans; and ␤-dystroglycan) remained unchanged in Cav-3 KO heart tissue sections (Fig. 4A).
The DG complexes present within caveolae may function in cellular signaling, because caveolae have been shown to serve as platforms for organizing and integrating a variety of signal transduction processes. In this regard, ␣-sarcoglycan is the best-studied member of the DG complex that has been implicated in signaling (35)(36)(37)(38). Thus, we further examined the expression and membrane localization of ␣-sarcoglycan, as a biochemical marker for the DG complex. Fig. 4B shows that the expression levels of ␣-sarcoglycan remained unchanged in the Cav-3 KO hearts, as seen by Western blot analysis. However, sucrose density gradient fractionation revealed that ␣-sarcoglycan was specifically excluded from lipid rafts (fractions 5 and 6) in the Cav-3 KO hearts (Fig.  4C). These results suggest that Cav-3 expression is normally required for maintaining the localization of the DG complex within cardiac myocyte lipid rafts/caveolae. (39 -43) have previously demonstrated that both Cav-1 and Cav-3 can function as inhibitors of the Ras-p42/44 MAPK cascade (using a variety of in vitro approaches), probably through a direct interaction with MEK or ERK. Because the Ras-p42/44 MAPK cascade has been clearly implicated as a mediator of cardiac hypertrophy (44), we next assessed the activation state of ERK1/2 in Cav-3 KO hearts, using a phospho-specific antibody probe that selectively recognizes activated ERK1/2. Fig. 5 shows that ERK1/2 was hyperactivated in Cav-3 KO hearts, as compared with hearts derived from wild-type control mice. These results provide the first in vivo evidence that Cav-3 can function as a negative regulator of the p42/44 MAPK cascade. Importantly, immunoblot analysis with a phospho-independent antibody revealed that total levels of ERK1/2 remain unchanged in Cav-3 KO hearts. DISCUSSION Since their discovery in the 1950s, cardiac myocyte caveolae have been postulated to perform a variety of important functions. However, a molecular understanding of cardiac myocyte caveolae only began recently in the mid 1990s with the identification of Cav-3, a muscle-specific caveolin-related protein (2,4,6). Since the molecular identification and cloning of Cav-3, a wealth of in vitro data has accumulated demonstrating a role for Cav-3 in cardiac myocyte signaling (19 -27). Interestingly, several distinct in vivo animal models of induced cardiac hypertrophy have shown reductions in Cav-3 protein expression within the heart (25,26,28). Taken together, these data argue that Cav-3 may play an important functional role as a negative regulator of hypertrophic signaling in the heart. However, this hypothesis remains untested.

Cav-3 KO Hearts Show Hyperactivation of the p42/44 MAPK Cascade-We and others
Mutations in CAV-3, as well as mutations in any one of the sarcoglycans (␣, ␤, ␦, or ␥), result in a Limb-Girdle muscular dystrophy (LGMD) phenotype. In addition to skeletal muscle symptoms, many LGMD patients show cardiac involvement FIG. 5. Hyperactivation of the p42/44 MAPK cascade in Cav-3 KO heart tissue. Hearts were harvested from wild-type and Cav-3 KO mice. Lysates were prepared and subjected to immunoblot analysis with antibodies directed against phospho-ERK1/2. Immunoblotting with phospho-independent antibodies to ERK was also performed as a control for equal protein loading. Note that Cav-3 KO mouse hearts show hyperactivation of ERK1/2 (upper panel) without any changes in the total cellular levels of ERK1/2 (lower panel). (45,46). Analysis of skeletal muscle tissue biopsies from these patients reveals the loss or dramatic reductions in the expression of all the sarcoglycans when there is a disease-related mutation in a single sarcoglycan family member. The interdependence of sarcoglycan expression is also evident from studies employing sarcoglycan-null mouse models, although cardiac abnormalities are found in only ␤-, ␦-, and ␥sarcoglycan-null mice (47)(48)(49)(50). Unlike these sarcoglycan-null mouse models, Cav-3 KO mice show no changes in expression or overall membrane localization of the sarcoglycans. However, using ␣-sarcoglycan as a marker for the DG complex, we demonstrate that the DG complex is no longer correctly targeted to lipid rafts/ caveolae in the hearts of Cav-3 KO mice. Although distinct functional roles for DG complexes that localize to different membrane microdomains of the plasma membrane have not been elucidated, it is possible that specific signaling functions of the DG complex take place within caveolae (51).
Multiple lines of evidence implicate the DG complex in cellular signaling. It has been proposed that ␣-dystroglycan, as well the ␤-, ␦-, and ␥-sarcoglycans, may possess a receptor function, whereas ␣-sarcoglycan acts as a downstream effector (37,52). In addition, ␤-dystroglycan and the ␣and ␥-sarcoglycans have been shown to be tyrosine-phosphorylated upon stimulation with different ligands, implicating them as signaltransducing molecules (35,53). Specifically, ␣-sarcoglycan may function in bi-directional signaling in concert with the integrinadhesion system as well as possess ecto-ATPase activity (35,36).
Interestingly, the DG complex and caveolins share the feature of serving as scaffolds for signaling molecules, the perturbation of which may result in muscle pathology (51). The loss of ␣-sarcoglycan from detergent-insoluble domains may thus correspond with altered DG complex cell signaling. Animal models of ␤and ␦-sarcoglycanopathies demonstrate vascular constriction/focal narrowing that initiates ischemic events in the cardiac muscle (48,50,54,55). However, signs of ischemia were not observed in Cav-3 KO hearts, suggesting that the Cav-3 KO cardiomyopathy is not due to pathological constriction of the coronary arteries. Vascular constriction in humans with sarcoglycanopathies has not been demonstrated (56); however, a more detailed analysis is needed. Activated p42/44 MAPK (ERK1/2) has been shown to play an important role as an effector of the cardiac hypertrophic response (57-59). ERK1/2 and known upstream activators MEK1/2 and multiple membrane receptors have all been shown to co-localize to caveolae (60 -62). In vitro data also support a role for Cav-1 and Cav-3 as negative regulators of p42/44 MAPK signaling, because overexpression of Cav-1 and Cav-3 inhibits p42/44 MAPK activation, and targeted downregulation of Cav-1 using an antisense approach results in the hyperactivation of the p42/44 MAPK cascade in NIH 3T3 fibroblasts (42,43).
Upon examination of Cav-3 KO hearts, we observed hyperactivation of the p42/44 MAPK cascade, as predicted. These findings are consistent with the notion that loss of Cav-3 expression results in dys-inhibition of the p42/44 MAPK cascade, thereby contributing to the development of a hypertrophic cardiomyopathy. As such, this is the first in vivo demonstration that a loss of Cav-3 causes the activation of a hypertrophic signaling program related to p42/44 MAPK activation.
In summary, we have presented the first detailed characterization of the hearts of Cav-3 KO mice. We clearly demonstrate that there are no derangements in the expression or localization of the other caveolin family members within Cav-3 KO hearts. Using a combination of non-invasive techniques (cardiac-gated MRI; transthoracic echocardiography) and histologi-cal analysis, we showed that Cav-3 KO mice develop a progressive, mild-to-moderate cardiomyopathic phenotype. Because we show that loss of Cav-3 results in the mis-localization of the DG complex and hyperactivation of the p42/44 MAPK cascade, these alterations could mechanistically explain the observed cardiac pathology.