土方 貴雄

廃用性筋萎縮のモデル動物を作成し筋萎縮の段階を筋タンパク質の免疫染色により明らかにした。これにより廃用性筋萎縮の臨床的なステージングが可能となった。<br /> （共同研究者として抗体を供与し貢献した）

骨格筋の筋形質膜には中間径フィラメントがアンカーして筋形質膜や筋線維構造を補強・安定化しているが、この中間径フィラメントと筋形質膜との間の分子構築を分子構築を明らかにした。<br /> （共同研究であるが主導的役割を果たした。）

Background: Skeletal muscles are composed of heterogeneous collections of muscle fiber types, the arrangement of which contributes to a variety of functional capabilities in many muscle types. Furthermore, skeletal muscles can adapt individual myofibers under various circumstances, such as disease and exercise, by changing fiber types. This study was performed to examine the influence of dystrophin deficiency on fiber type composition of skeletal muscles in canine X-linked muscular dystrophy in Japan (CXMD(J)), a large animal model for Duchenne muscular dystrophy. Methods: We used tibialis cranialis (TC) muscles and diaphragms of normal dogs and those with CXMDJ at various ages from 1 month to 3 years old. For classification of fiber types, muscle sections were immunostained with antibodies against fast, slow, or developmental myosin heavy chain (MHC), and the number and size of these fibers were analyzed. In addition, MHC isoforms were detected by gel electrophoresis. Results: In comparison with TC muscles of CXMDJ, the number of fibers expressing slow MHC increased markedly and the number of fibers expressing fast MHC decreased with growth in the affected diaphragm. In populations of muscle fibers expressing fast and/or slow MHC(s) but not developmental MHC of CXMDJ muscles, slow MHC fibers were predominant in number and showed selective enlargement. Especially, in CXMDJ diaphragms, the proportions of slow MHC fibers were significantly larger in populations of myofibers with non-expression of developmental MHC. Analyses of MHC isoforms also indicated a marked increase of type I and decrease of type IIA isoforms in the affected diaphragm at ages over 6 months. In addition, expression of developmental (embryonic and/or neonatal) MHC decreased in the CXMDJ diaphragm in adults, in contrast to continuous high-level expression in affected TC muscle. Conclusion: The CXMDJ diaphragm showed marked changes in fiber type composition unlike TC muscles, suggesting that the affected diaphragm may be effectively adapted toward dystrophic stress by switching to predominantly slow fibers. Furthermore, the MHC expression profile in the CXMDJ diaphragm was markedly different from that in mdx mice, indicating that the dystrophic dog is a more appropriate model than a murine one, to investigate the mechanisms of respiratory failure in DMD.

Using murine models, we have previously demonstrated that recombinant adeno-associated virus (rAAV)-mediated microdystrophin gene transfer is a promising approach to treatment of Duchenne muscular dystrophy (DMD). To examine further therapeutic effects and the safety issue of rAAV-mediated microdystrophin gene transfer using larger animal models, such as dystrophic dog models, we first investigated transduction efficiency of rAAV in wild-type canine muscle cells, and found that rAAV2 encoding beta-galactosidase effectively transduces canine primary myotubes in vitro. Subsequent rAAV2 transfer into skeletal muscles of normal dogs, however, resulted in low and transient expression of b-galactosidase together with intense cellular infiltrations in vivo, where cellular and humoral immune responses were remarkably activated. In contrast, rAAV2 expressing no transgene elicited no cellular infiltrations. Co-administration of immuno-suppressants, cyclosporine and mycophenolate mofetil could partially improve rAAV2 transduction. Collectively, these results suggest that immune responses against the transgene product caused cellular infiltration and eliminated transduced myofibers in dogs. Furthermore, in vitro interferon-g release assay showed that canine splenocytes respond to immunogens or mitogens more susceptibly than murine ones. Our results emphasize the importance to scrutinize the immune responses to AAV vectors in larger animal models before applying rAAV-mediated gene therapy to DMD patients.

Here, we present the first report of the molecular cloning of zebrafish protocadherin 10 (Pcdh10, OL-protocadherin) and describe its functional analyses in the development of segmental plate. Epitope-tagged Pcdh10 expressed in embryos was localized on cell peripheries of adjacent cells. In situ hybridization showed that pcdh10 was expressed in the paraxial mesoderm (PAM) and developing somites, and in the pineal body, the diencephalon, and the vicinity of otocysts. Expression in PAM increased in the last few presumptive somites, reached the maximum level in the latest segmenting somites, and decreased thereafter during somite maturation. These expression patterns suggested that Pcdh10 is involved in development of PAM and somites. This was confirmed by morpholino knockdown and dominant-negative inhibition of Pcdh10 in embryos, which disturbed movements of PAM cells and somite segmentation. Comparative studies showed that pcdh10 expression lasted up to approximately three times longer in maturing somites than that of paraxial protocadherin (pcdh8). They also indicated that the adaxial cells expressed pcdh8 but not pcdh10. We propose that Pcdh10 is involved in the morphogenic movements of PAM cells and somite segmentation and that differential adhesion of Pcdh8 and Pcdh10 plays a role in the morphogenic machinery of somites and adaxial cells.

Plectin is a versatile cytoskeletal linker protein that preferentially localizes at interfaces between intermediate filaments and the plasma membrane in muscle, epithelia] cells, and other tissues. Its deficiency causes muscular dystrophy with epidermolysis bullosa simplex. To better understand the functional roles of plectin beneath the sarcolemma of skeletal muscles and to gain some insights into the underlying mechanism of plectin-deficient muscular dystrophy, we studied in vivo structural and molecular relationships of plectin to subsarcolemmal cytoskeletal components, such as desmin, dystrophin, and vinculin, in rat skeletal muscles. Immunogold electron microscopy revealed that plectin fine threads tethered desmin intermediate filaments onto subsarcolemmal dense plaques overlying Z-lines and I-bands. These dense plaques were found to contain dystrophin and vinculin, and thus may be the structural basis of costameres. The in vivo association of plectin with desmin, (meta-)vinculin, dystrophin, and actin was demonstrated by immunoprecipitation experiments. Treatment of plectin immunoprecipitates with gelsolin reduced actin, dystrophin, and (meta-)vinculin but not desmin, implicating that subsarcolemmal actin could partly mediate the interaction between plectin and dystrophin or (meta-)vinculin. Altogether, our data suggest that plectin, along with desmin intermediate filaments, might serve a vital structural role in the stabilization of the subsarcolemmal cytoskeleton.

Plectin is a versatile linker protein which is associated with various types of cytoskeletal components and/or filaments including intermediate filaments. To better understand the functional roles of plectin in smooth muscle cells, we examined the distribution of plectin and other related proteins in rat colon smooth muscles by confocal laser and electron microscopy. The sarcolemma of smooth muscle cells exhibits two ultrastructurally distinct domains, domains associated with dense plaques and caveola-rich domains. Staining with anti-plectin and anti-desmin antibodies showed that plectin was localized along the sarcolemma in an intermittent manner and desmin was distributed in the sarcoplasm and intermittently at the cell periphery where it was codistributed with desmin. Plectin exhibited complementary and non-overlapping distribution to caveolin-1 and dystrophin, components of caveola domains, whereas plectin was codistributed with vinculin, talin and integrin beta1, components of dense plaques. Plectin was also codistributed with beta2-chain laminin but not with beta1-chain laminin, Electron microscopic observations on the sarcolemma revealed close association of intermediate filaments with dense plaques. Correlated confocal and electron microscopy clearly demonstrated that anti-plectin fluorescence corresponded to dense plaques but not to caveola domains in electron microscopic images. These findings indicate that plectin is confined to dense plaques to which desmin intermediate filaments may be anchored in rat colon smooth muscle cells.

While over a dozen I-Z-I proteins are expressed in postmitotic myoblasts and myotubes it is unclear how, when, or where these first assemble into transitory I-Z-I bodies (thin filament/Z-band precursors) and, a short time later, into definitive I-Z-I bands. By double-staining the growth tips of transfected myotubes expressing (a) MYC-tagged s-alpha-actinins (MYC/s-alpha-actinins) or (b) green fluorescent protein-tagged titin cap (GFP/T-cap) with antibodies against MYC and I-Z-I band proteins, we found that the de novo assembly of I-Z-I bodies and their maturation into I-Z-I bands involved relatively concurrent, cooperative binding and reconfiguration of, at a minimum, 5 integral Z-band molecules. These included s-alpha-actinin, nebulin, titin, T-cap and alpha-actin, Resolution of the similar to 1.0 mu m polarized alpha-actin/nebulin/tropomyosin/troponin thin filament complexes occurred subsequent to the maturation of Z-bands into a dense tetragonal configuration. Of particular interest is finding that mutant MYC/s-alpha-actinin peptides (a) lacking spectrin-like repeats 1-4, or consisting of spectrin-like repeats 1-4 only, as well as (b) mutants/fragments lacking titin or alpha-actin binding sites, were promptly and exclusively incorporated into de novo assembling I-Z-I bodies and definitive I-Z-I bands as was exogenous full length MYC/s-alpha-actinin or GFP/T-cap.

Plectin is a versatile linker protein which is associated with various types of cytoskeletal components and/or filaments including intermediate filaments, and its deficiency causes the disruption of myofibrils, or muscular dystrophy. To better understand the functional role of plectin in skeletal muscle fibers, we have examined the topological and structural relationships of plectin to intermediate filaments and Z-discs in rat diaphragm muscles by confocal and immunoelectron microscopy. Immunofluorescence analysis revealed that plectin was colocalized with desmin at the periphery of Z-discs. This plectin localization around Z-discs was constantly maintained irrespective of the contracted or extended state of the muscle fibers, suggesting either direct or indirect association of plectin with Z-discs. Immunogold labeling in skinned muscle fibers clearly demonstrated that plectin-labeled fine threads linked desmin intermediate filaments to Z-discs and connected intermediate filaments to each other. These results indicate that through plectin threads desmin intermediate filaments form lateral linkages among adjacent Z-discs, preventing individual myofibrils from disruptive contraction and ensuring effective force generation.

It has been biochemically shown that dystrophin and alpha- and beta-dystroglycan form an oligomeric complex which links laminin, a component of the basement membrane, to components of the subsarcolemmal cytoskeleton in skeletal muscle fibers. In the present study the dystrophin-glycoprotein complex and its structural relationships to laminin and subsarcolemmal cytoskeleton were ultrastructurally examined in crude surface membranes prepared from rat skeletal muscles. Sarcolemmal vesicles within crude surface membranes were identified and characterized by fine protrusions on their outer surface and electron-dense materials or patches associated with the inner surface. These two components were seen to be in register with each other across the sarcolemma. The fine protrusions were immunolabeled by anti-alpha-dystroglycan and reassociated with exogenous laminin. Immunolabeling in combination with laminin reassociation demonstrated that the electron-dense materials contained dystrophin at laminin-binding domains of the membrane. In addition, they were often associated with very fine filaments. These results provide morphological evidence for the biochemically proposed model of molecular array of dystrophin complex from the basement membrane to the subsarcolemmal cytoskeleton. ((C) Elsevier, Paris).

To explore the roles of specific domains of sarcomeric a-actinin (s-alpha-aetinin) in the assembly and maintenance of striated myofibrils, myogenic cultures were transfected with four MYC-tagged s-alpha-actinin peptides. They were: (1) full-length sarcomeric alpha-actinin, (2) an N-terminal deletion that removed the actin-binding site only (MYC/A(-)), (3) a peptide that consisted of the actin-binding site only (MYC/A(+)), and (4) an N-terminal deletion that removed the EF-hands and titin-binding domains (MYC/EFT-). While cytotoxic in replicating myogenic cells, as they were in PtK2 cells, the four MYC peptides were not cytotoxic in postmitotic myotubes. In myotubes each of the four different MYC peptides were promptly and selectively incorporated into normal Z bands. The incorporation of MYC/A(-), MYC/A(+), and MYC/EFT- into Z bands suggests that (a) the actin-binding site, (b) the spectrin-repeats believed to be responsible for anti-parallel dimerization, and (c) the C-terminal EF-hands and titin-binding domains are each dispensable for targeting s-alpha-actinin/MYC peptides into Z bands. These findings could not have been predicted from the behavior of alpha-actinin (a) in binding assays in cell-free systems or (b) when expressed in transfected nonmuscle cells. (C) 1998 Academic Press.

In the skeletal muscle fiber organization of many vertebrate muscles, serial arrangements or linkages of muscle fibers along the muscle or fascicle are commonly found. These serially linked muscle fibers employ distinct junctional morphologies from muscle to muscle. Notable are the end-to-end linkages of muscle fibers through tendinous intersections (TIs), where many fibers end onto a continuous connective tissue plate with folded terminations similar to myotendinous junctions. Besides this end-to-end linkage, overlapping linkages or arrangements occur among nonspanning fibers terminating intrafascicularly. These nonspanning fibers bear tapering terminations with direct cell-cell (myomuscular) junctions or without any specialized junctions. Despite their overlapping linkages or tapering profiles, nonspanning fibers maintain a uniform sarcomere length along the linked fibers, suggesting that the overlapping-linked nonspanning fibers are equivalent to the end-to-end linked fibers in their mechanical capacity. However, the junctional compliance could differ in their extracellular elastic components and their organization at junctional sites, e.g., direct mechanical (myomuscular) junctions vs, indirect linkages through connective tissue. Increasing evidence suggests that the elastic components, including muscle fibers as well as connective tissues, are more critical than previously thought for the mode and/or the efficiency of tension transmission among serially arranged fibers and thus for the mechanical properties of the muscle.

To understand the multiple roles of alpha-actinin in the assembly of (1) Z bands in muscle, and (2) a variety of cytoskeletal structures in non-muscle cells, 4 sarcomeric alpha-actinin derived cDNAs tagged with a MYC epitope were constructed. The constructs were: (1) full-length (FL/MYC); (2) minus EF-hands (-EF/MYC); (3) actin-binding site ((+)A/MYC); and (4) minus actin-binding site ((-)A/MYC). These four cDNAs were individually transfected into PtK2 cells. The exogenous sarcomeric alpha-actinin (s-alpha-actinin/MYC) was followed with labeled anti-MYC, the endogenous non-sarcomeric alpha-actinin (non-s-alpha-actinin) with labeled anti-non-s-alpha-actinin. The salient findings were: (1) the selective intracellular localizations of each expressed MYC-tagged peptide differed one from the other; (2) their respective localizations in the 10-24-h post-transfection (p.t.) period differed from their localizations in the 48-72-h p.t. period; (3) each MYC-positive peptide was cytotoxic, but each in a distinctive way; and (4) while the selective targeting of FL/MYC to dense bodies, adhesion plaques, adherens junctions, and ruffled membranes was consistent with binding studies in cell-free systems, the incorporation of the mutated peptides, particularly (+)A/MYC and -A/MYC was not. Changes in localization over time and the distinctive cytopathologies probably reflect domain-specific targeting. They also suggest unexpected cooperative involvement of multiple domains of alpha-actinin with specific receptors in distal cytoskeletal structures. To date, such qualitative in vivo interactions have not been described either in in vitro binding studies, or in short-term experiments involving localization and/or fate of microinjected labeled molecules into living cells. (C) 1997 Wiley-Liss, Inc.

The earliest expression of truncated desmin in transfected PtK2 cells results in the formation of dispersed microprecipitates containing not only the truncated desmin, but also endogenous vimentin and cytokeratin proteins. Desmin microprecipitates without vimentin or vimentin microprecipitates without desmin are not observed. The microprecipitates involving cytokeratin invariably are also positive for desmin and vimentin. Over time, the precipitates enlarge into 1- to 2-mum spheroids and then fuse into amorphous chimeric juxtanuclear masses that can occupy &gt;30% of the cell volume. Concurrently, first the vimentin and then the cytokeratin networks are resorbed. The chimeric precipitates are not recognized or marked for degradation by the lysosomal system. Ultimately the cell nucleus fragments and the cell dies. Similar protein complexes appear in many human and animal pathologies, suggesting that a similar protein-precipitation sequence initiated by the introduction of a mutationally or environmentally altered protein molecule is at work.

A procedure for classifying histochemically stained skeletal muscle fibers using a personal computer was established. Frozen sections of rat rectus abdominis muscle were stained for myosin ATPase with preincubation at low and high pH, and for a mitochondrial enzyme, NADH tetrazolium reductase. Light microscopic images of the sections were converted through a video camera into computerized images which were then analyzed with digital image-processing software in a personal computer to recognize individual muscle fibers and to categorize them into slow-twitch, oxidative type; fast-twitch, oxidative-glycolytic type; and fast-twitch, glycolytic type. This procedure yielded not only number of fibers but also cross-sectional area of fibers for each fiber type. Video copies of the identical fields were used for the visual assessment of fiber types to compare the results with those obtained by the computer. The discrepancy between the computer-aided procedure and the conventional visual assessment on video copies was less than 10 percent, except for 30 percent in 'intermediate' fibers in sections stained for the mitochondrial enzyme. The present procedure, employing a relatively small-sized multi-purpose personal computer system, proved to be effective and reliable in the analysis of local fiber-type composition.

Using digital image analysis and several anatomical methods, morphometric analysis of nonspanning fibers which had tapering profiles at their intrafascicular termination sites and represented overlapping arrangements within the fiber fascicles was performed in the rat rectus abdominis. Special emphasis was focused on dimensional relationships occurring between overlapping portions and tapering segments and sarcomere lengths in non- and overlapping portions. Nonspanning fibers were found to overlap each other for more than 40% of their length. In length, their overlapping portions generally corresponded to their tapering segments, which were also greater than 40% of the fiber length. In addition, despite the presence of overlapping linkages, nonspanning fibers maintained a fairly uniform length irrespective of their overlapping and non-overlapping portions. Overlapping linkages in fibers without tapering profiles have a larger cross-sectional area in the overlapping portion than in the non-overlapping one, resulting in a phenomenon which will cause different sarcomere lengths between the two portions during fiber stretching. The present results suggest that tapering profiles in the overlapping portion ensure uniform sarcomere lengths within nonspanning fibers, thereby providing mechanical stability in each fiber. (C) 1993 Wiley-Liss, Inc.

The rectus abdominis muscle is architecturally compartmentalized by tendinous intersections and is supplied by multiple thoracic nerves. In this study, the rectus abdominis of the rat has been qualitatively and quantitatively examined with regard to muscle dimensions, fiber organization, fiber-type composition, and innervation. The muscle exhibits architectural heterogeneity and different patterns of innervation among its thoracic, epigastric, and hypogastric parts. The epigastric part, adherent to the rectus sheath via tendinous intersections, represents relatively simple design. It is formed by serially arranged compartments with shorter fibers, compared with the other parts. These compartments are segmentally supplied by thoracic nerves. The hypogastric part is more complex, forms an interdigitation of muscular slips, and has segmental distribution of thoracic nerves in mediolateral direction. The thoracic part much differs from the other parts. It has smaller cross-sectional areas, compartments composed of abundant nonspanning fibers with intrafascicular termination, and non-segmental distribution of thoracic nerves. In addition to these craniocaudal specializations among the three parts, the muscle exhibits mediolateral differences in fiber-type composition. Slow-twitch oxidative fibers are more densely distributed in the medial half region than the lateral, whereas fast-twitch glycolytic fibers follow an inverse pattern. The mediolateral differences in fiber-type composition as well as the craniocaudal specializations in architectural design and innervation imply regionally differentiated recruitments of the muscle in various behaviors.

The ultrastructure of the basement membrane of the rat proximal tubule was observed by transmission electron microscopy after the use of a cold dehydration technique. The basement membrane of the P1 segment is thick and possesses several structural specializations that are rare in other basement membranes; these include intraepithelial ridges, dense bars, and basement membrane vesicles. The intraepithelial ridges are found in the intercellular spaces between interdigitating processes of the proximal tubule cells. The ridges and the interdigitating processes run circumferentially around the tubule. The dense bars are frequently found in the intraepithelial ridges. They are especially prominent on the concave side of the tubular bends and to a lesser extent near sites where intracellular actin filaments anchor onto the basal cell membranes. The basement membrane vesicles are bounded by unit membranes; they are variable in both their electron density and their size. They are usually found in association with dense bars, and the grade of their accumulation is positively correlated with the development of the dense bars. These three specializations have no topographical relationship with the interstitial structures, such as fibroblasts and collagen fibrils. The specializations are best developed on the concave side of tubular bends where the circumferential stresses caused by the intraluminal hydraulic pressure are presumably the largest; we therefore propose that they are an adaptation to, or a manifestation of, the increased wall stress in the proximal tubule.