|Home | About | Journals | Submit | Contact Us | Français|
Germline ablation of the cytoskeletal protein nonmuscle myosin II-B (NMII-B) results in embryonic lethality with defects in both the brain and heart. Tissue specific ablation of NMII-B by a Cre-recombinase strategy should avoid embryonic lethality and permit study of the function of NMII-B in adult hearts.
To understand the function of NMII-B in adult mouse hearts and to see if the brain defects found in germline ablated mice influence cardiac development.
We used a loxP/Cre-recombinase strategy to specifically ablate NMII-B in the brains or hearts of mice. Mice ablated for NMII-B in neural tissues, die between postnatal day 12 and 22 without showing cardiac defects. Mice deficient in NMII-B only in cardiac myocytes (BαMHC/BαMHC mice) do not show brain defects. However BαMHC/BαMHC mice display novel cardiac defects not seen in NMII-B germline ablated mice. Most of the BαMHC/BαMHC mice are born with enlarged cardiac myocytes some of which are multinucleated, reflecting a defect in cytokinesis. Between 6–10 months they develop a cardiomyopathy which includes interstitial fibrosis and infiltration of the myocardium and pericardium with inflammatory cells. Four of five BαMHC/BαMHC hearts develop marked widening of intercalated discs.
By avoiding the embryonic lethality found in germline-ablated mice we were able to study the function of NMII-B in adult mice and show that absence of NMII-B in cardiac myocytes results in cardiomyopathy in the adult heart. We also define a role for NMII-B in maintaining the integrity of intercalated discs.
Nonmuscle myosin II (NMII) plays an important role in maintaining the integrity of the actin-myosin cytoskeleton, which in turn helps to determine cell shape, and functions in cell migration, cell polarity, and cytokinesis. Like conventional myosin IIs, such as skeletal, cardiac and smooth muscle myosins, NMII consists of a pair of heavy chains and two pairs of light chains. Three isoforms of the nonmuscle myosin heavy chain (NMHC), II-A, II-B and II-C which are encoded by three genes Myh9, 10 and 14 respectively, and are located on different chromosomes, have been identified in humans and mice.1–3 Although there is some overlap in the localization of these three isoforms, growing evidence suggests that they perform distinct functions during cell migration and embryonic development.4–6 Ablation of NMII-A in mice results in lethality at embryonic day (E)6.5 due to the lack of a normal functioning visceral endoderm which results in a markedly abnormal body pattern. These embryos fail to undergo gastrulation.7 In contrast, ablation of NMII-B in mice results in embryonic lethality between E14.5 and birth, with defects in the brain and heart,8;9 suggesting that NMII-B is critical for the development of both. Unfortunately, the embryonic lethality in NMII-B-null mice has impeded further efforts to understand the physiological roles of NMII-B in adult mice. Hypomorphic mice expressing low amounts of NMHC II-B can survive to adulthood and also display defects in both brains and hearts, however severe NMII-B hypomorphs also die before adulthood.10 Moreover, since the physiological activities of the heart are continuously regulated by the nervous system, questions are raised as to whether any of the heart defects in NMII-B ablated or hypomorphic mice are secondary to the brain defects.
In this report, we ablated NMHC II-B in mice, either in the nervous system alone or in the cardiac myocytes alone, using a loxP/Cre-recombinase strategy. We crossed the NMHC II-B floxed mice with a line of mice expressing Cre-recombinase regulated by the neural cell-specific nestin promoter to ablate NMHC II-B in the nervous system.11 In separate experiments we crossed the NMHC II-B floxed mice with a line of mice expressing Cre-recombinase under control of the α-myosin heavy chain (αMHC) promoter to ablate NMII-B in cardiac myocytes.12 Below we present results showing that NMII-B plays distinct physiological roles in the brain and heart and provide evidence that absence of NMII-B in the cardiac myocytes (and not in the non-myocytes) results in myocyte enlargement and cardiomyopathy. Moreover we demonstrate a role for NMII-B in the intercalated disc (ID) of adult mice.
Detailed methods are described in Online Supplemental Material. All experiments were conducted following animal protocols approved by ACUC, NHLBI. Nestin-Cre transgenic mice were from The Jackson Laboratory (#003771).
H&E and immunofluorescence staining, electron microscopy, and immunoblotting were performed as described previously.8
The size of cardiac myocytes was measured following wheat germ agglutinin staining using a Zeiss measuring tool.
Echocardiography was performed using an Acuson Sequoia 256c imaging system with the 15L8 multi-frequency transducer. Quantitation was performed using M-mode with Prosolv Software Version 3.0.
Three-leads electrocardiograms were recorded with a model MAC 1200, G.E. Medical Systems.
The data was expressed as mean ±SD. Student’s t-test was used to compare data between two groups.
To ablate NMHC II-B specifically in the brain or in the heart, we used a loxP/Cre recombinase strategy to delete exon2, the first coding exon of Myh10 (see Figure 1). Using the gene targeting method, we generated a mouse line designated Bflox/Bflox in which both the Neor expression cassette and exon2 of Myh10 are flanked by loxP sites (Figure 1c). Bflox/Bflox mice express normal amounts of NMHC II-B protein and are indistinguishable from wild-type. We crossed Bflox/Bflox mice with two different transgenic lines expressing Cre-recombinase under control of different promoters. To ablate NMHC II-B in neural tissue we used a mouse line expressing Cre-recombinase with a nestin promoter and a nervous system-specific enhancer11 (Figure 1e) to generate B+/Bnest and Bnest/Bnest mice. To ablate NMHC II-B in cardiac myocytes we used a mouse line expressing Cre-recombinase driven by the α-cardiac myosin heavy chain (α-MHC) promoter12 (Figure 1f) to generate B+/BαMHC and BαMHC/BαMHC mice. The transgenic mouse lines nestin-Cre and α-MHC-Cre are well characterized and demonstrate that nestin-Cre mice express functional Cre recombinase in the nervous system starting at E10.511 and that α-MHC-Cre mice express functional Cre recombinase specifically in cardiac myocytes starting at E9.0 and at high levels by E11.5.13
Immunoblot analysis of Bnest/Bnest mice on postnatal day (P)18 shows that NMHC II-B protein is markedly reduced in the cerebellum (Figure 2A, compare lanes 1 and 2), as well as throughout the entire brain (data not shown), but not in the hearts of Bnest/Bnest mice (Figure 2A, lanes 3 and 4). Immunofluorescence staining using an antibody for NMHC II-B confirms that NMHC II-B protein is ablated in the cerebellum of these mice at P18 (Figure 2B). As shown in the figure, Purkinje cells in Bflox/Bflox mice express high levels of NMHC II-B (red; Figure 2Ba). However, the NMHC II-B protein level is significantly reduced in the cerebellar Purkinje cells of Bnest/Bnest mice (Figure 2B, compare a with e). In contrast, staining for calbindin (Purkinje cell marker) is unaltered in these cells (Figure 2B, compare b with f). Panels d and h confirm the loss of NMHC II-B (yellow in d and green in h), in the cerebellar Purkinje cells. Both immunoblot analysis and immunofluorescence staining confirm that NMHC II-B protein is also significantly reduced in the cerebral cortex of Bnest/Bnest mice (data not shown).
All of the Bnest/Bnest mice die between P12 and P22 due to a severe hydrocephalus which results in enlargement of the lateral ventricles and a paper-thin cortex with absence of most brain cortical tissue (Figure 2C, compare b and d). Figure 2Cc also shows an underdeveloped cerebellum (ellipse) which correlates with defects in motor activity in these mice and which was also seen in hypomorphic mice that have a point mutation in NMHC II-B.14 The arrows in Figure 2Cc point to deformities following the decompression of the lateral ventricles and loss of cerebral-spinal fluid. Of note is our finding that, similar to hypomorphic mice with a point mutation in NMHC II-B15 the spinal canal of Bnest/Bnest mice is completely ablated at P7 (Figure 2D). This is consistent with a role for NMII-B in cell-cell adhesion in the spinal canal.
Sectioning of the hearts confirms that there are none of the cardiac abnormalities in the Bnest/Bnest mice that are found in B−/B− mice. Specifically, there is no evidence for a ventricular septal defect (VSD), double outlet of the right ventricle (DORV), myocyte hypertrophy, decreased myocyte number or increased cardiac myocyte binucleation, as seen in B−/B− mice.8;16 This demonstrates that ablation of NMHC II-B in the nervous system does not affect cardiac development. B+/Bnest mice appear normal in all respects.
To address the role of NMII-B in heart development and in the adult mouse we crossed the Bflox/Bflox mice with α-MHC Cre mice (Figure 1f), so that ablation would occur specifically in the cardiac myocytes starting at mid-gestation or after approximately E11.5 to avoid the early lethality found in germ line ablated B−/B− mice. Similar to other investigators who worked with this particular line of mice expressing Cre-recombinase, we found no adverse effects of the enzyme on the tissues in which it was expressed17 (See also Figures 4,,66 and Online Figure I). The NMHC II-B protein level is significantly reduced in the hearts of BαMHC/BαMHC mice compared to Bflox/Bflox mice at P0 as demonstrated in the immunoblot in Figure 3A. However, the NMHC II-B level is not affected in the brain of BαMHC/BαMHC mice (Figure 3A). The presence of residual NMHC II-B in the heart of BαMHC/BαMHC mice can be attributed to non-myocyte cells in the heart, which continue to express wild-type amounts of NMHC II-B. Immunofluorescence staining of E13.5 mouse heart sections using antibodies to NMHC II-A, II-B and II-C helps to clarify the ablation of NMHC II-B from the myocytes alone. It also demonstrates that ablation of NMHC II-B in cardiac myocytes has no effect on NMHC II-A and II-C expression. Figure 3Ba and d show that NMHC II-A (green) is only present in the non-myocyte cells in the heart. This is indicated by the lack of green signal co-incident with desmin (red) in the cardiac myocytes, since NMHC II-A is not present at this age in these cells. In contrast, in Figure 3Bb NMHC II-B (green) and desmin (red) co-stain the myocytes (yellow) and stain the non-myocytes green, showing that in the Bflox/Bflox heart NMHC II-B is expressed in both cell types. The arrows are pointing to the non-myocytes (green). However in the BαMHC/BαMHC heart the cardiac myocytes now appear red since NMHC II-B has been ablated but desmin remains. Again the arrows point to the non-myocytes, which stain green for II-B, indicating that NMII-B is not ablated in non-myocytes. Panels Bc and Bf show that there is very little or no NMHC II-C in these hearts and that this does not change after II-B is ablated.
H&E staining on the day of birth (P0; Figure 4) shows evidence for an increase in the size of the cardiac myocytes in BαMHC/BαMHC hearts (f) compared to B+/BαMHC hearts (c). Figure 4f also shows examples of bizarrely shaped nuclei (arrows) reflecting an abnormality in cytokinesis due to the loss of NMHC II-B.16 A VSD in a BαMHC/BαMHC heart (Figure 4d enlarged in 4e) is also shown. These abnormalities were not seen in B+/B+ or B+/BαMHC mice (Figure 4a,b). Unlike B−/B− mice, in which the heart phenotype of a membranous VSD and DORV is almost 100% penetrant,8 only 2 of 9 BαMHC/BαMHC mice examined were born with a VSD and neither displayed a DORV. We did not see any difference in the deletion of NMII-B in the BαMHC/BαMHC cardiac myocytes with or without the presence of a VSD. However 5 of 9 newborn BαMHC/BαMHC mice examined had an obvious increase in cardiac myocyte size along with nuclear changes. The absence of a DORV and the small percentage of mice with VSDs compared to B−/B− mice are likely due to the timing of the loss of NMHC II-B from the cardiac myocytes.
An advantage of these cardiac myocyte-specific NMHC II-B knockout mice is that most of them survive to adulthood permitting analysis of the role of NMII-B in the adult heart. H&E stained sections of BαMHC/BαMHC hearts examined at 6 months, similar to hearts examined at P0 (Figure 5a,e), show evidence for myocyte hypertrophy (Figure 5b,f). This hypertrophy is even more evident by 10 months (Figure 5c,g). Figure 6 shows wheat germ agglutinin (WGA) staining to more easily visualize and quantify the increase in cardiac myocytes size in BαMHC/BαMHC hearts at 4 and 6months of age. As shown in Figure 6f there is a progressive increase in the size of the cardiac myocytes.
In addition to cardiac myocyte hypertrophy BαMHC/BαMHC mice at 10 months also display additional pathological changes including interstitial fibrosis (Figure 5d,h). There is infiltration with inflammatory cells, such as lymphocytes, plasma cells and macrophages in the interstitium of the myocardium and the pericardium of BαMHC/BαMHC mouse hearts. Some of these changes can be seen as early as 6 months of age in BαMHC/BαMHC mice (Figure 5f), but are much more prominent at 10 months of age (5d,h; n=4). We also observed vacuolation in cardiac myocytes, suggesting that the myocytes have undergone marked degeneration (5h, arrow). The presence of vacuolated cells in the heart prompted us to examine them for evidence of an increase in apoptosis. Online Figure II shows the results of a TUNEL assay comparing BαMHC/BαMHC and Bflox/Bflox mouse hearts. There is a small increase in the number of cells undergoing apoptosis in the mutant heart as indicated by the arrows in the figure.
The presence of the inflammatory response in these hearts raises the possibility of a viral myocarditis. Of note however, we did not observe inflammation in the hearts of Bflox/Bflox littermates. Analyses of both control and BαMHC/BαMHC mice for the presence of viruses associated with myocarditis were negative for enterovirus (coxsackie viruses and echo virus, data not shown). In two out of eight BαMHC/BαMHC mice thrombi were seen in the H&E stained sections of the left atrium of the heart at 10 months of age, consistent with severe pathological changes in the heart and compromised cardiac function (see Online Figure III).
We also addressed the question of whether the fetal cardiac program was reactivated in the hearts of these mice by performing both immunoblot analyses and immunofluorescence microscopy on the NMII-B ablated and normal hearts. In contrast to our previous findings for mice that were rendered hypomorphic for NMII-B, where there is a 40 fold increase in the expression of β-cardiac myosin,10 we found only a 2–3 fold increase in the expression of β-myosin heavy chain in the 6 month old BαMHC/BαMHC mouse heart. Microscopy confirmed this and showed that the increase was only detected in relatively few cardiac myocytes and did not correlate with the extent of myocyte hypertrophy (data not shown)
To obtain information about the cardiac function of BαMHC/BαMHC mice, we carried out echocardiography at 4, 6 and 10 months of age. The Table shows that although there are no significant differences between Bflox/Bflox and BαMHC/BαMHC mice at 4 and 6 months, there are differences at 10 months of age. These include an increase in left ventricular internal diameter at the end of systole and a marked decrease in the percent of fractional shortening from 44±8% (n=7) for Bflox/Bflox to 29±9% (n=12) for BαMHC/BαMHC mice. These results confirm that cardiac function is significantly compromised in the BαMHC/BαMHC mice at 10 months and are consistent with a cardiomyopathy. We also performed electrocardiography (EKG; three standard leads) to determine whether abnormalities could be detected in BαMHC/BαMHC mice at this age. The electrical axis ranged from +30° to +90° for the control animals. However, 4 out of 5 of the BαMHC/BαMHC mice displayed an abnormal right axis deviation, ranging from +90° to >+210° and in three of them the severity of the deviation (>+120°) is consistent with an abnormality in cardiac conduction (Online Figure IV).
To understand a possible cause of the defect in conduction in BαMHC/BαMHC hearts we carried out an electron microscopy study. Previous work from this laboratory has demonstrated that in adult mice NMII-B was detected in the IDs in the heart.18 Moreover deletion and mutation of proteins associated with the IDs are often associated with defects in cardiac conduction.19;20 It was therefore of interest to see if the IDs of the BαMHC/BαMHC mice were normal in structure. Figure 7A is electron micrographs showing that the IDs of BαMHC/BαMHC mice are widened and distorted (4/5 mice examined) compared to Bflox/Bflox mice. Approximately 20% of the IDs of BαMHC/BαMHC mice show this abnormality which is not found in Bflox/Bflox mice. Careful inspection of the affected IDs shows that whereas the adhesion junctions are severely disrupted (large arrow, 7Ab,c), the desmosomes (arrowheads) and gap junctions (arrows) remain mostly intact and are less affected (panels b,c).
To gain insight into the cause of the disruption of the ID we carried out an immunoblot analysis of a number of proteins known to be present at the disc at 6 months. Figure 7B shows that of the proteins analyzed, including a number of adhesion molecules, only the actin binding protein mXinα is decreased (a decrease of 78.5±4.8% compared to the wild type, n=2 mice, performed in triplicate). In contrast, expression of connexin 43 is increased, most likely due to cardiac myocyte hypertrophy. Figure 7C shows the distribution of both mXinα and connexin 43 in wild type and BαMHC/BαMHC hearts at 10 months using confocal immunofluorescence microscopy. The staining confirms the decreased expression of mXinα at the ID. It also shows that unlike the wild type disc, mXinα is not uniformly associated with connexin 43 in many of the discs. We propose that the loss in NMII-B at the ID is the primary cause of the disruption of cell-cell adhesion in the NMII-B ablated heart. Moreover the loss (NMII-B) and decrease (mXinα) of two actin binding proteins at the ID, the latter of which also binds to β-catenin21 are expected to contribute to instability at the adhesion junction (see Discussion). Of note, no defects in the brain or other organs were found in BαMHC/BαMHC mice at any age.
We have previously reported that global ablation of NMHC II-B in mice resulted in hydrocephalus as early as E11.5 associated with defects in cell-cell adhesion of the neural epithelial cells lining the spinal canal and cerebral ventricles.9;15 In Bnest/Bnest mice, ablation of NMHC II-B was initiated at E10.5 controlled by the nestin promoter, which is consistent with the delayed onset of hydrocephalus. Of particular note despite the death of these mice between 12 to 22 days of age due most likely to severe hydrocephalus, there were no abnormalities found in the heart.
In addition to learning whether the defects we found in the B−/B− mouse hearts were related directly or indirectly to those found in the nervous system, we also wanted to study the role of NMII-B in the adult mouse heart. Most BαMHC/BαMHC mice manifested progressive cardiac abnormalities starting with myocyte hypertrophy, which was apparent as early as P0 and increased during postnatal development to 6 and 10 months of age. At 10 months there was also evidence for myocyte vacuolation and cell degeneration, interstitial fibrosis and an infiltration of the cardiac tissue with inflammatory cells. We hypothesize that the cardiac phenotype in the BαMHC/BαMHC mice is initiated by abnormalities specific to the cardiac myocytes, since NMII-B is ablated in these cells but not in the non-myocytes in these mice. This loss of NMII-B (and lack of compensation by NMII-A or II-C) results in a failure in cytokinesis as manifested by multinucleation and the bizarre nuclei found in these cells. It most likely contributes to abnormal enlargement of the cardiac myocytes as well as their decreased numbers at P0. We therefore reasoned that the interstitial fibrosis and infiltration of inflammatory cells are secondary to the primary abnormality in the cardiac myocytes, which is most likely myocyte degeneration.
The pathological changes in the hearts of cardiac-specific NMHC II-B knockout mice are in agreement with the echocardiographic and EKG studies. The marked decrease in the fractional shortening at 10 months is consistent with the compromised contractility of cardiac muscle. The abnormalities noted in EKGs (an abnormal electrical axis) could reflect the striking defects found in the IDs. Previous work has shown that in the adult heart NMII-B is localized to the Z-lines and IDs.18 The IDs are composed of adherans junctions, desmosomes and gap junctions that form cell-cell boundaries and connections between cardiac myocytes and allow the myocardium to function in synchrony. As noted above, work from a number of laboratories has shown that NMIIs play an important role in cell-cell adhesion7;15;20;22 and that abnormalities in a number of adhesion proteins result in either loss or structural changes in the cardiac IDs.20;23;24 Figure 7A provides evidence that loss of NMII-B primarily affects the adhesion junctions rather than the gap junctions or desmosomes. Moreover, BαMHC/BαMHC hearts at 6 months show a milder defect in the adhesion junction of the IDs and no defects in the desmosomes and gap junctions (data not shown).
We have analyzed the expression of a number of ID proteins and found a significant decrease in the expression of mXinα in BαMHC/BαMHC hearts compared to the wild type hearts. Mice ablated for mXinα also show abnormal IDs.19 Unlike the mXinα knockout hearts, NMII-B ablated hearts show no decrease in expression levels of N-cadherin orβ-catenin. Moreover there was no change in the distribution of β-catenin. We therefore attribute the primary cause of the disruption of the IDs to the loss of NMII-B. We speculate that the decrease of mXinα is secondary to the loss of NMII-B and the mechanism of this decrease is of ongoing interest. The decrease in both of these proteins, one of which (mXinα) has been demonstrated to also bind to β-catenin,21 would explain the marked disruption of the IDs.
The finding of a role for NMII-B in the cardiac ID is similar to the findings for NMII-B in the spinal canal. Our hypothesis is that NMII exerts tension and stabilizes actin filaments which in turn are required for maintenance of adhesion complexes between cells or in this case between the cardiac myocytes. The loss of NMII-B from the adhesion complex could therefore result in the gradual deterioration in the cardiac adhesion complex, including the loss of mXinα, over a period of time and this would account for our failure to observe abnormal discs in B−/B− mice which died before birth. Interestingly, generation of mice in which NMII-A replaced NMII-B did not produce defects in the IDs.25 This is consistent with the hypothesis that in cases where NMIIs are apparently playing a structural rather than a motor role, one isoform is more likely to substitute for the other in vivo as well as in cultured cells.6 When myosin is playing more of a motor role, for example in neural cell migration, because of significant differences in the kinetics of MgATPase hydrolysis and actin-binding properties between the myosin isoforms, successful substitution is much less likely, at least in vivo.15 These findings further support the idea that disruption of the IDs in BαMHC/BαMHC mice is due to the loss of NM II and is secondary to the development of the cardiomyopathy.
These conditionally ablated mice demonstrate that the defects we observed in the hearts and brains of the B−/B− mice are independent of each other. The availability of NMHC II-B floxed mice will allow conditional ablation of NMHC II-B in a variety of tissues and cells and thus help to further define its role both in situ and in vivo.
We are grateful to Charles W. Birdsall, NHLBI and to the NIH Mouse Imaging Facility for help with mouse echocardiograms and EKGs. We thank Douglas R. Rosing, NHLBI for his expert advice on EKGs. We would like to acknowledge the professional skills and advice of Christian A. Combs and Daniela Malide of the Light Microscopy Core Facility, NHLBI, regarding microscopy-related experiments performed in this paper. The authors wish to thank the members of the Laboratory of Molecular Cardiology for suggestions and criticism especially Mary Anne Conti and Sachiyo Kawamoto. We thank Stephanie Jackson for editorial assistance.
SOURCE OF FUNDING
This work was supported by the Division of Intramural Research, National Heart, Lung, and Blood Institute, National Institutes of Health.