Abnormal Distribution of Calcium-Handling Proteins: A Novel Distinctive Marker in Core Myopathies

Transcription

1 J Neuropathol Exp Neurol Copyright Ó 2007 by the American Association of Neuropathologists, Inc. Vol. 66, No. 1 January 2007 pp. 57Y65 ORIGINAL ARTICLE Abnormal Distribution of Calcium-Handling Proteins: A Novel Distinctive Marker in Core Myopathies Muriel Herasse, PhD, Karine Parain, MSc, Isabelle Marty, PhD, Nicole Monnier, PhD, Angela M. Kaindl, MD, Jean-Paul Leroy, MD, Pascale Richard, PhD, Jöel Lunardi, PhD, Norma B. Romero, MD, PhD, and Ana Ferreiro, MD, PhD Abstract Central core disease (CCD) and multi-minicore disease (MmD) are muscle disorders characterized by foci of mitochondria depletion and sarcomere disorganization ( cores ) in muscle fibers. Although core myopathies are the most frequent congenital myopathies, their pathogenesis remains elusive and specific diagnostic markers are lacking. Core myopathies are mostly caused by mutations in 2 sarcoplasmic reticulum proteins: the massive Ca 2+ -release channel RyR1 or the selenoprotein N (SelN) of unknown function. To search for distinctive markers and to obtain further pathophysiological insight, we identified the molecular defects in 12 core myopathy patients and analyzed the immunolocalization of 6 proteins of the Ca 2+ -release complex in their muscle biopsies. In 7 cases with RYR1 mutations (6 CCD, one MmD), RyR1 was depleted from the cores; in contrast, the other proteins of the sarcoplasmic reticulum (calsequestrin, SERCA1/2, and triadin) and the T-tubule (dihydropyridine receptor-> 1 subunit) accumulated within or around the lesions, suggesting an original modification of the Ca 2+ -release complex protein arrangement. Conversely, all Ca 2+ -related proteins were distributed normally in 5 MmD cases with SelN mutations. Our results provide an appropriate tool to orientate the differential and molecular diagnosis of core myopathies and suggest that different pathophysiological mechanisms lead to core formation in SelN- and in RyR1-related core myopathies. Key Words: Calcium-release proteins, Core myopathies, RyR1, SelN. From Inserm U582 (MH, KP, NBR, AF), Institut de Myologie, Paris, France; Université Pierre et Marie Curie-Paris 6 (MH, KP, NBR, AF), Paris, France; Inserm U607 (IM, NM, JL), CIS-DRDC, CEA-Grenoble, France; Laboratoire de Biochimie et Génétique Moléculaire (NM, JL), CHU Grenoble, France; Department of Pediatric Neurology (AMK), Charité, University Medicine, Berlin, Germany; Service d Anatomopathologie (J-PL), CHU Brest, France; and AP-HP, Groupe Hospitalier Pitié- Salpêtrière (PR), UF de Cardiogénétique et Myogénétique, Paris, France. Send correspondence and reprint requests to: Ana Ferreiro, MD, PhD, INSERM U582, Institut de Myologie, Groupe Hospitalier Pitié-Salpêtrière, 47 Bd. de l Hôpital, Paris, France; a.ferreiro@myologie. chups.jussieu.fr Drs. Herasse and Parain contributed equally to this work. This work was financially supported by the INSERM, the AFM/Institut de Myologie, the Clinical Research Protocol (PHRC) for study of Malignant Hyperthermia and Central Core Disease, and the GIS/Institut des maladies rares Bioclinical network for the study of pathologies of the calcium release complex in skeletal muscle. INTRODUCTION Core myopathies (CM) are the most prevalent congenital myopathies (1). They manifest clinically with early-onset muscle weakness and are defined by characteristic histopathologic changes of the muscle fiber structure, termed cores. Core lesions are localized areas of mitochondria depletion and sarcomere disorganization. Their presence in muscle fibers characterizes 2 congenital myopathies: central core disease (CCD) (2) and multi-minicore disease (MmD) (3). Classically, CCD and MmD were considered independent entities. CCD was defined as an autosomaldominant disorder characterized by central, well-delimited cores spanning the entire fiber length. Most CCD cases carry heterozygous mutations of the ryanodine receptor type 1 gene (RYR1) (4Y6). In contrast, MmD is an autosomalrecessive disorder whose hallmarks are multiple, poorly defined cores that spread only a few sarcomeres in the longitudinal fiber axis. Four different clinical phenotypes have been distinguished in MmD (7, 8). The most prevalent one, termed classical MmD, is associated in most but not all cases with mutations in the selenoprotein N gene (SEPN1) (9Y11). Despite these classic differences between CCD and MmD, recent progress in the molecular characterization of CM has disclosed an unexpected overlap between both disorders. In RYR1-related CCD, the spectrum of morphologic manifestations can be highly pleomorphic, including multiple cores (5) or presence of only minicores (12, 13). Besides, homozygous or compound heterozygous RYR1 defects are responsible for 2 MmD forms (moderate MmD and MmD with ophthalmoplegia) (12, 14Y16). In some of these cases, an age-related evolution of the morphologic lesions, from minicores to cores, has been observed (14, 17). Additionally, other known or unknown genes can be more rarely involved in CM (18); RYR1 and/or SEPN1 implication has been excluded in a number of patients, and mutations of the ACTA1 gene have been identified in 2 families with a dominant CM (19). Therefore, establishing a precise diagnosis in CM is frequently challenging, particularly in young and/or sporadic cases. These difficulties are exacerbated by the lack of specific morphologic markers that could help selecting the most appropriate genetic test (20). The massive size of RYR1 J Neuropathol Exp Neurol Volume 66, Number 1, January

2 Herasse et al J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 (160 kb, 106 exons) (21) precludes systematic whole-gene screen. Consequently, many CM cases remain molecularly undefined, which prevents genetic counseling and accurate assessment of their risks regarding respiratory failure and orthopedic complications. This lack of diagnostic markers relates to our incomplete understanding of the pathophysiology of CM. RYR1 encodes the skeletal muscle isoform of a sarcoplasmic reticulum (SR) channel (RyR1), which mediates Ca 2+ release during excitationycontraction coupling (E-C coupling) and interacts with other proteins to form the triadic junction (Fig. 1). Yet, the pathophysiological consequences of RYR1 mutations remain controversial (22). Moreover, selenoprotein N (SelN) function is unknown (23). SelN contains a potential Ca 2+ binding domain (EF-hand) and, like RyR1, is localized in the SR membrane and associated with core lesions. These common features allow hypothesizing a potential functional link between RyR1 and SelN. To our knowledge, the distribution of the major proteins implicated in Ca 2+ homeostasis and E-C coupling has not been ascertained in CM (except for sarco/endoplasmic reticulum Ca 2+ ATPase SERCA [13]). This prompted us to study the immunolocalization of 6 of these proteins (RyR1, the dihydropyridine receptor DHPR->1 subunit, SERCA isoforms 1 and 2, triadin, and calsequestrin) in FIGURE 1. Simplified representation of the main components of the sarcoplasmic reticulum (SR) and triadic junction in skeletal muscle. The longitudinal SR contains the Ca 2+ - ATPase pump SERCA, which uptakes Ca 2+ from the cytoplasm to the SR lumen. The terminal cisternae are the SR portion involved in releasing Ca 2+ for muscle contraction. They are rich in calsequestrin (which binds and stores Ca 2+ in the SR lumen) and have a highly specialized zone, the junctional SR, which contains the homotetrameric RyR1 together with other proteins such as triadin. RyR1 serves as a Ca 2+ release channel of the SR and also as a bridging structure connecting the SR and the T-tubule. Two terminal cisternae interact with one T-tubule to form a triadic junction. The T-tubule membrane contains the voltage-sensitive Ca 2+ channel DHPR (dihydropyridine receptor) composed of 4 subunits (>, A, F, and >2/C). Arrival of the membrane potential to the T-tubule activates the DHPR, which, in turn, triggers RyR1 opening; this produces a massive Ca 2+ release from the SR stores to the cytoplasm where it binds to contractile filaments and initiates muscle contraction (E-C coupling). SRm, sarcoplasmic reticulum membrane; Ttm, T-tubule membrane. 58 skeletal muscles from genetically characterized CCD and MmD patients carrying RYR1 or SEPN1 mutations. MATERIALS AND METHODS Patients Open diagnostic muscle biopsies from 12 ambulant patients with a clinical, histologic, and molecular diagnosis of SEPN1- orryr1-related MmD or CCD were included in this study. Histologic criteria in the 6 MmD cases included type I fiber predominance and relative hypotrophy as well as the presence of short cores in most type I and type II fibers. Five of these cases had a classical MmD phenotype (9): congenital weakness predominantly of axial muscles, scoliosis, and respiratory failure requiring nighttime ventilation. The sixth patient presented in infancy with the moderate form of MmD (pelvic girdle weakness, no ophthalmoplegia). The 6 CCD patients showed long, well-delimited cores in 90% to 100% of type I fibers. All showed early-onset muscle weakness, mainly involving the pelvic girdle. One patient (CCD6) belongs to a previously reported family with recessive CCD (14) in which the affected siblings showed minicores in the muscle biopsies taken in childhood (no longer available) but long cores with rods in those taken in adulthood. To validate our results, the 12 genetically characterized patients were compared with 6 CM patients (MmD7YMmD9, CCD7YCCD9) without SEPN1 or RYR1 mutations selected using the same clinical and morphologic criteria. SEPN1 mutations were excluded in the 3 MmD cases; RYR1 mutations were excluded in the 6 patients after screening of the whole RYR1 cdna (in 5 cases) or the C-terminal encoding exons (in MmD7). Also, we excluded ACTA1 mutations in MmD8, CCD7, and CCD8 (the latter belonging to the same dominant family as CCD9). Linkage Analysis and Mutation Screening SEPN1 and RYR1 linkage studies were performed in the genetically informative families as described (9, 12). SEPN1 mutation screening was performed by direct sequencing as described (9). Because of its large size, the whole RYR1 coding sequence was screened when possible using complementary DNA (cdna) obtained from a skeletal muscle biopsy (12). When muscle samples were unavailable, we screened RYR1 using genomic DNA and focused on the 15 3`-terminal exons containing most of the known CCD mutations (5). All samples were obtained after informed consent according to local ethics committees. Immunohistochemical and Ultrastructural Procedures Muscle samples were frozen immediately in isopentane cooled in liquid N 2. Histologic and histochemical stainings were performed according to standard procedures using 10-Km cryosections (24). For immunohistochemistry, we used the following primary antibodies: polyclonal antibodies developed by one of the authors (IM) and directed Ó 2007 American Association of Neuropathologists, Inc.

3 J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 Ca 2+ Handling Proteins in Core Myopathies against purified pig skeletal muscle RyR1 (anti-ryr, 1/500) (12) and the N-terminal domain of the rabbit triadin (anti- TNter, 1/20) (25, 26); commercial antibodies directed against purified alpha1 subunit of rabbit skeletal muscle DHPR (monoclonal anti-dhpr> 1 s; ABCAM, Cambridge, UK; 1:300), rabbit calsequestrin (polyclonal anti-calsq; Affinity BioReagents, Golden, CO; 1:200), and mouse SERCA 1 and 2 isoforms (monoclonal anti-serca1/2; Novocastra, Newcastle, UK;1:100). Serial 7-Km cryosections were fixed with paraformaldehyde 4% in phosphatebuffered saline (PBS). After washing and incubation in PBS with 3% bovine serum albumin, labeling with primary antibodies was performed overnight at 4-C. Sections were then washed in PBS with Tween 0.1% and incubated 1 hour at room temperature with the following secondary antibodies: fluorescein isothiocyanate-conjugated (FITC) swine anti-rabbit immunoglobulins (DAKO, Carpinteria, CA; 1:300); FITC mouse anti-rabbit (DAKO; 1:300); or CY3- conjugated AffiniPure F(ab`)2 fragment goat anti-mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA; 1:1000). Immunofluorescence was visualized with an Axiophot2 fluorescence microscope (Zeiss, Oberkochen, Germany) and recorded with a Photometrics Cool Snap fx camera (Roper Scientific, Tucson, AZ). For electron microscopy (EM) studies, muscle samples were fixed in glutaraldehyde, processed using standard procedures and observed using a Philips EM120 microscope. RESULTS Mutational Analysis The mutations identified are shown in Table 1. The five patients with classical MmD carried SEPN1 mutations. Three of them were compound heterozygous for a missense mutation associated with either a splicing or a nonsense mutation (out-of-frame deletion or insertion). The 2 remaining patients were homozygous; MmD2 carried a common missense mutation, whereas in MmD3, an out-of-frame nucleotide insertion in exon 5 led to an absence of SelN protein verified by Western blot on fibroblast cultures (data not shown). The 6 CCD patients and the moderate MmD case (MmD6) had RYR1 mutations. Five of them carried missense heterozygous changes in the C-terminal coding region of the gene (exons 100 and 102), a hot spot where most of the known CCD mutations cluster. The remaining 2 patients showed more uncommon genetic defects. Patient CCD5 had an in-frame heterozygous deletion of 3 bases in exon 91, predicting loss of one amino-acid residue. Patient CCD6 belonged to a consanguineous family with recessive CCD; she carried a homozygous missense change in exon 71, resulting in the replacement of a proline by serine (p.p3527s). Histologic and Immunohistochemical Results In all cases and with all the antibodies used, the preserved areas of the muscle fibers showed a labeling of the intermyofibrillar network indistinguishable from that observed in control muscle. Multi-Minicore Disease Patients With SEPN1 Mutations In these cases, both fiber types were well represented; type 1 fibers were slightly more numerous. Classic histochemical techniques (Fig. 2) showed minicores in a variable percentage of fibers, which was lowest (around 50%) in MmD3, maximum (100%) in MmD1 and MmD2, and intermediate in MmD4 and MmD5. However, these lesions were not visible by immunohistochemistry, because there was no focal modification of immunolabeling with any of the antibodies analyzed. The expression patterns of all the Ca 2+ -related proteins studied were normal and identical to those observed in normal controls (Fig. 3; Table 2). TABLE 1. Results of the Mutational Studies in Multi-Minicore Disease (MmD) and Central Core Disease (CCD) Patients, Age at the Time of Muscle Biopsy, and Muscle Analyzed Cases Genes Exons Mutations Pattern of Inheritance Muscle Analyzed Age at Muscle Biopsy MmD with SEPN1 mutations MmD1 SEPN1 6/12 p.h293r/g.ivs11-2:a>g AR Deltoid 9 years MmD2 SEPN1 5 p.g273r Ho AR Deltoid 11 years MmD3 SEPN1 4 c.713_714insa Ho AR Deltoid 12 years MmD4 SEPN1 5/10 c deltg/p.w490l AR Deltoid 12 years MmD5 SEPN1 4/6 c.713_714insa/p.g315s AR Deltoid 35 years MmD with RYR1 mutation MmD6 RYR1 100 p.h4803y Sporadic Deltoid 25 years CCD with RYR1 mutations CCD1 RYR1 102 p.h4887y AD Deltoid 33 years CCD2 RYR1 100 p.r4825c AD Quadriceps 37 years CCD3 RYR1 100 p.y4796c AD Deltoid 31 years CCD4 RYR1 100 p.l4793p Sporadic Deltoid 37 years CCD5 RYR1 91 c.del AD Quadriceps 30 years CCD6 RYR1 71 p.p3527s Ho AR Quadriceps 21 years Mutations are designed according to the published reference sequences. Ho, homozygous; AD, autosomal-dominant; AR, autosomal-recessive. Ó 2007 American Association of Neuropathologists, Inc. 59

4 Herasse et al J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 FIGURE 2. Minicore lesions in multi-minicore disease (MmD) (AYC) and core lesions in central core disease (CCC) (DYH). Cores are visible with oxidative stainings as clear zones devoid of enzymatic activity (arrowheads). In transverse muscle fiber sections, typical MmD minicores are multiple, poorly defined small areas (A), but they can occasionally be larger in diameter and unique (B). Although the archetypal cores in CCD are central, sharply delimited, unique lesions, only one CCD biopsy (CCD4) showed this pattern (D); all other CCD samples disclosed multiple cores that were often subsarcolemmal and less well-defined (E). Thus, core length as evidenced by longitudinal muscle fiber sections is the most reliable criterion to differentiate minicores (C) from cores (F). The dark rim surrounding some cores ([G, H], arrows) has often been considered a sign of mitochondria accumulation. However, this rim is more intense with NADH-TR (G), reflecting both mitochondrial and sarcoplasmic reticulum (SR) oxidative activity, than with SDH (H), which is exclusively mitochondrial. This suggests a perilesional accumulation of SR (and possibly mitochondria). NADH-TR-stained transverse muscle sections from MmD5 ([A], 40), MmD1 ([B], 40), CCD4 ([D], 20; [G, H], 40), and CCD1 ([E], 25); longitudinal electron microscopy sections from MmD2 (C), and CCD1 (F). Scale bar = 2.5 Km. Central Core Disease Patients With RYR1 Mutations In all CCD samples, myosin ATPase reactions disclosed an extreme predominance of type I fibers (approximately 90Y100%), all expressing SERCA2. Oxidative enzyme stainings revealed long cores, which were multiple and peripheral in all cases but one (Fig. 2DYH). Some cores were bordered by a dark rim, suggesting focal accumulation of SR (Fig. 2GYH). In 3 cases (CCD1, CCD3, and CCD6), cores were associated with rods. EM studies demonstrated different degrees of sarcomere disorganization, including structured cores (CCD3), unstructured cores (CCD2, CCD5), or coexistence of both (CCD1). Independent of the core distribution and associated lesions, all samples displayed focal immunostaining abnormalities with all the antibodies. These abnormalities were consistent and restricted to the core lesions (Fig. 4). Strikingly, RyR1 immunolabeling was virtually absent from the cores, suggesting a focal depletion of this protein. In contrast, most cores were strongly immunolabeled for DHPR> 1 s, triadin, SERCA1/2 and, particularly, calsequestrin, indicating a focal accumulation of these proteins within and/or around the cores (Table 2). Some cores were diffusely and homogeneously labeled with the 5 latter antibodies ( homogeneous pattern ), suggesting a focal accumulation of SR and triads FIGURE 3. Normal immunolabeling in multi-minicore disease (MmD) with SEPN1 mutations. Serial transversal sections of muscle samples from MmD2 ([AYC], 40) and MmD5 ([DYF], 25) labeled with anti-ryr1 (A), anti-triadin (B), anti-serca2 (D), and anticalsequestrin (E). The minicore lesions observed with NADH-TR staining ([C, F], arrowheads) were not revealed by any of the antibodies. 60 Ó 2007 American Association of Neuropathologists, Inc.

5 J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 Ca 2+ Handling Proteins in Core Myopathies TABLE 2. Summary of the Immunohistochemical Changes of Ca 2+ -Related Proteins in Genetically Characterized Core Myopathies MmD With SEPN1 Mutations MmD With RYR1 Mutations CCD With RYR1 Mutations Cores RyR1 N,,,,,, Calsequestrin N jjj jjj Triadin N jj jj SERCA1/2 N j j DHPR>1s N jj jj Oxidative-positive dark areas No Yes, labeled for triadin T DHPR>1s and RyR1 No Reduction and increase in immunolabeling (arrows) refer to the core lesions and not to the normal fiber areas. MmD, multi-minicore disease; CCD, central core disease; N, normal. devoid of RyR1. Conversely, other lesions showed a peripheral rim of immunocytochemical protein accumulation, which contrasted with a variable degree of labeling reduction in their center ( perilesional/halo pattern ) (Figs. 4, 5A). Both patterns of immunocytochemical protein accumulation could coexist in the same patient (Fig. 5B) or even in the same fiber, but never in the same core (Fig. 4). However, one of these patterns tended to predominate in a single patient. There was no correlation between the immunocytochemical pattern and the type of RYR1 mutation or the degree of sarcomere disorganization observed by EM (structured vs. unstructured cores, presence or absence of rods). No alteration in the distribution of SR or T-tubules within or around the cores was apparent on ultrastructural sections. Interestingly, in 3 CCD patients (CCD1, CCD3, and CCD4), NADH-TR staining showed sparse fibers with only small, poorly defined cores, virtually identical to the minicores in MmD. All the labeling abnormalities described previously were retrieved in these RYR1-related minicores (Fig. 6A). Multi-Minicore Disease Patient With RYR1 Mutation Muscle from MmD6 showed short cores associated with centrally located nuclei and multiple foci of increased oxidative activity, staining dark blue with NADH-TR (Fig. 6B). The minicores showed a labeling pattern identical to that observed in CCD with RYR1 mutations (Fig. 6BaYc, f, g, arrows). Noticeably, the dark oxidative-positive areas were often labeled for triadin and, occasionally, for DHPR> 1 s and RyR1, suggesting focal accumulation of these proteins outside the cores (Fig. 6BfYh, asterisk). FIGURE 4. Abnormal immunolabeling in central core disease with RYR1 mutations. Serial (HYK) and nonserial (AYG) longitudinal (AYC) and transverse (DYK) muscle sections from patients CCD4 (AYC, F, HYK), CCD6 (D), CCD5 (E), and CCD3 (G). Core lesions devoid of RyR1 (A, D, E, H); strong labeling of DHPR> 1 s (B, J), calsequestrin (C, F, G), triadin (I), and SERCA2 (K) within or around the cores. Each individual core displays either a perilesional (C, F, IYK) or a homogeneous (G) protein accumulation pattern. In (A) and (C), the muscle fiber borders are delimited by dotted lines and the core lesions are indicated by arrowheads. Original magnifications: (A) 16; (B) 100; (CYE) 25; (FYK) 40. Ó 2007 American Association of Neuropathologists, Inc. 61

6 Herasse et al J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 FIGURE 5. Perilesional versus homogeneous pattern in RYR1-mutated cases. (A) Serial transversal muscle sections from CCD1 (100). NADH-TR (a); RyR1 (b), triadin (c), and DHPR> 1 s (d) immunolabeling. In the perilesional/halo lesions, triadin tended to be reduced in the core center (c), whereas DHPR> 1 s was often increased (d). This suggests that, at least in some cores, the junctional sarcoplasmic reticulum (SR) component of the triad (RyR1 and triadin) is preferentially depleted from the core center, whereas the tubular component (DHPR> 1 s) is present or accumulated. (B) Serial transverse muscle sections from patient CCD3 (40); NADH-TR (e), myosin-atpase ph 4.35 (f), SERCA2 (g), and SERCA1 (h). Those cores showing some degree of SR oxidative activity in their center (i.e. staining blue with NADH-TR; [e], arrow) tended to be homogeneously labeled ([g], arrow). In contrast, the cores whose center lacked NADH-TR staining tended to show a perilesional pattern ([e, g], arrowheads). Some fibers, which stained as type 1 with myosin-atpase (f), expressed high levels of both SERCA2 and SERCA1 ([fye], asterisk). Multi-Minicore Disease and Central Core Disease Patients Without SEPN1 or RYR1 Mutations Labeling in these nonmutated cases was relatively variable, but none of them showed a pattern identical to either of those in SEPN1- or in RYR1-mutated patients (Fig. 7). DISCUSSION Recent progress in the molecular characterization of CM has shown that the classic diagnostic criteria based on core morphology alone are not sufficient for an exact diagnosis, stressing the necessity of novel markers. Previous studies demonstrated that all types of core lesions show increased immunoreactivity for several proteins such as filamin C, >B-crystallin, actin, dystrophin, gelsolin, desmin, >1-antichymotrypsin, A-amyloid precursor protein, A2- microglobulin, NCAM, and myotillin (10, 13, 20, 27Y30). Filamin C and >B-crystallin have been considered the strongest immunohistochemical markers of core lesions (13, 20). However, the described immunocytochemical alterations are nonspecific; they are present in minicores, cores, and in other structural muscle defects such as target lesions (20, 28). Thus, they probably represent a secondary, stereotyped cellular response to sarcomere disorganization, nonindicative of the primary molecular defect. In contrast, our study revealed a particular generelated immunohistochemical pattern in CM. The distribution of Ca 2+ -handling proteins was normal in SEPN1-related 62 MmD but was consistently altered in CCD and MmD with RYR1 mutations. Within or around the cores, we observed a virtual absence of RyR1 and a focal accumulation of calsequestrin, SERCA1/2, triadin, and DHPR> 1 s. This pattern was present in RYR1-related minicores as well as in cores; therefore, the immunohistochemical differences between RYR1- and SEPN1-related myopathies cannot be fully explained by the distinct core sizes but inform on the different molecular defects. These distinctive immunohistochemical patterns represent a useful marker for differential diagnosis and may help choosing the appropriate genetic test for disease confirmation, especially in young CCD patients whose biopsies do not show the classic core structures. Cores showing the described immunolabeling abnormalities are not indicative of SEPN1 mutations, but advocate RYR1 screening. In our experience, RyR1 depletion and calsequestrin accumulation were the most conspicuous markers in RYR1- mutated cases. Therefore, we recommend the use of these 2 antibodies as a first immunohistochemical screening in CM. Calsequestrin accumulation within the cores is particularly noticeable as a result of the low level of labeling in normal fiber areas. In contrast, SERCA focal abnormalities are frequently inconspicuous as a result of the strong basal labeling for this protein in muscle fibers; actually, no modification in SERCA immunolabeling was observed in a previous study of 3 families with RYR1 mutations (13). Noticeably, in 5 CM cases without RYR1 or SEPN1 mutations, we observed dark oxidative-positive areas that were immunolabeled for calsequestrin, triadin, and DHPR>1s, whereas the cores showed Ó 2007 American Association of Neuropathologists, Inc.

7 J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 Ca 2+ Handling Proteins in Core Myopathies FIGURE 6. Abnormal immunolabeling of minicores in central core disease (CCD) and multi-minicore disease (MmD) patients with RYR1 mutations. (A) Serial transverse muscle sections from patient CCD3 (40); NADH-TR (Aa) and immunolabeling for RyR1 (Ab). RyR1 reduction is observed both in the typical CCD cores and in the adjacent minicores or areas of uneven oxidative staining (arrows). (B) Serial longitudinal (ayc) and transverse muscle sections from MmD6. NADH-TR (Ba, d); DHPR> 1 s (Bb), RyR1 (Bc, g, h), calsequestrin (Be, i), and triadin (Bf) immunolabeling. Minicores ([Bd], arrows) were short in length ([BaYc], arrows) and showed RyR1 reduction ([Bc, g], arrows) and DHPR> 1 s (Bb), triadin ([Bf], arrows), and calsequestrin accumulation. Calsequestrin accumulation was restricted to the cores ([Bd, e, i], arrowheads); in contrast, triadin and RyR1 accumulation was occasionally observed in areas staining dark-blue with NADH-TR or in oxidatively normal areas adjacent to the cores ([Bd, fyh], asterisk). The muscle fiber boxed is in (BdYf) and enlarged in (Bh, i). Original magnifications: (ayf) 25; (gyi) 40. no labeling abnormality. Thus, identification of immunolabeled spots on a CM biopsy is not reliable in itself; comparison with serial NADH-TR or SDH-stained sections is necessary to verify whether they correspond to the cores. This could partly explain the RyR1 accumulation described in a previous RYR1-mutated case in which correlation with cores or serial sections are not illustrated (31). Aside from its diagnostic applications, our results can provide some prospective indications about the pathophysiology of CM. It has been hypothesized that SelN could be implicated in skeletal muscle Ca 2+ handling and directly related with the SR Ca 2+ -release structures (9). Our results do not support this hypothesis; on the contrary, they suggest that SelN is not directly involved in maintaining the structural integrity of the Ca 2+ -release complex. Thus, different mechanisms could lead to core formation in muscles lacking normal SelN and in those carrying RyR1 defects. Two models have been proposed to explain the functional consequences of RYR1 mutations (22) based on data obtained using nonhuman and/or nonmuscular cells. According to the first model, RyR1 mutations would lead to a sustained calcium leak. The second model suggests that the mutated RyR1 would lose its functional connection with DHPR, leading to E-C uncoupling. Alternatively, protein accumulation in muscle fibers can be a nonspecific phenomenon. However, neither this possibility nor any of the 2 former pathophysiological models can explain the focal RyR1 immunocytochemical depletion that we observed in human Ó 2007 American Association of Neuropathologists, Inc. 63

8 Herasse et al J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 FIGURE 7. Immunolabeling patterns in multi-minicore disease (MmD) and central core disease (CCD) patients without SEPN1 or RYR1 mutations. Serial transverse muscle sections from MmD9 (AYC), CCD8 (DYF), CCD7 (GYI), MmD 7 (J, K), and CCD9 (L, M). NADH-TR (A, D, G, J, L); immunolabeling for DHPR> 1 s (B), RyR1 (EYH), and triadine (C, F, I, K, M). In 5 of the 6 nonmutated patients (MmD7YMmD9, CCD8, and CCD9), most cores showed no labeling abnormality ([AYF], arrowheads). Yet, these cases differed from those with SEPN1 mutations by the presence of rare, small areas positively labeled for calsequestrin, triadin, DHPR> 1 s and occasionally for RyR1 ([AYF], arrows). Some of these areas colocalized partially with the cores (DYF), but most corresponded to oxidative-positive dark areas adjacent to the cores ([JYM], arrows). An exception to this pattern was CCD7 (GYI), whose cores were labeled for DHPR> 1 s, triadin, SERCA1/2, and calsequestrin but tended to have a halo of perilesional RyR1 accumulation (H) instead of the RyR1 depletion observed in RYR1-mutated cases. Original magnifications: (AYC) 25; (DYM) 40. muscle expressing RYR1 missense mutations. Our results point to a focal accumulation of SR virtually depleted of RyR1 but not of the other Ca 2+ handling proteins. This can be best understood according with a novel, hypothetical pathophysiological mechanism that we tentatively propose for core generation: some RYR1 mutations would lead to a relative and focal RyR1 depletion, thereby modifying the stoichiometry of proteins in the Ca 2+ -release complex. This could cause initially a focal compensatory increase in SR and T-tubules; the absence of RyR1, and thus of a correct functional bridging between SR and T-tubules, would eventually provoke a loss of the normal SR distribution and its accumulation around the lesions, whereas T-tubules remain present within the cores. This evolving character of the cores is difficult to document and cannot be proven by 64 this study, but would explain why 2 different patterns of protein accumulation ( homogeneous and perilesional ) may coexist in the same biopsy. So far, definition of cores has been based on focal mitochondria and sarcomere abnormalities. Our results suggest that, in cases with RYR1 mutations, SR and T-tubule abnormalities are a more distinctive change than the conventional EM defects in mitochondria and contractile apparatus. Based on the central role of SR and T-tubules in muscle cells, it is tempting to speculate that their pathologic modifications would lead to E-C coupling dysfunction, thereby causing muscle weakness. Along these lines, the mitochondrial and sarcomere modifications would represent secondary morphologic changes whose functional relevance remains to be proven. Ó 2007 American Association of Neuropathologists, Inc.

9 J Neuropathol Exp Neurol Volume 66, Number 1, January 2007 Ca 2+ Handling Proteins in Core Myopathies In conclusion, our findings provide an appropriate tool to discriminate between RYR1- and SEPN1-related CM and suggest some novel pathophysiological indications. Further studies, including additional mutations and Western blot analysis, are needed to extend our results and gain a deeper insight into the generation of core lesions and muscle weakness in CM. ACKNOWLEDGMENTS The authors thank Emmanuelle Lacène, Melanie Pucelle, and Céline Ledeuil for technical assistance. The authors also thank Prof. Michel Fardeau and Dr. Frédéric Chevessier for useful advice and discussions. AMK is a member of the German Network on Muscular Dystrophies (MD-NET). AF is the recipient of an Avenir Career Development Award. REFERENCES 1. Fardeau M, Tomé F. Congenital myopathies. In: Engel AG, Franzini- Armstrong C, eds. Myology. New York: McGraw-Hill; 1994:1487Y Shy GM, Magee KR. A new congenital non-progressive myopathy. Brain 1956;79:610Y21 3. Engel AG, Gomez MR, Groover RV. Multicore disease. A recently recognized congenital myopathy associated with multifocal degeneration of muscle fibers. Mayo Clin Proc 1971;46:666Y81 4. Zhang Y, Chen HS, Khanna VK, et al. A mutation in the human ryanodine receptor gene associated with central core disease. Nat Genet 1993;5:46Y50 5. Monnier N, Romero NB, Lerale J, et al. Familial and sporadic forms of central core disease are associated with mutations in the C-terminal domain of the skeletal muscle ryanodine receptor. Hum Mol Genet 2001;10:2581Y92 6. Wu S, Ibarra MC, Malicdan MC, et al. Central core disease is due to RYR1 mutations in more than 90% of patients. Brain 2006;129: 1470Y80 7. Ferreiro A, Fardeau M. 80th ENMC International Workshop on Multi-minicore Disease: 1st International MmD Workshop. 12Y13th May, 2000, Soestduinen, The Netherlands. Neuromuscul Disord 2002; 12:60Y68 8. Jungbluth H, Beggs A, Bonnemann C, et al. 111th ENMC International Workshop on Multi-minicore Disease. 2nd International MmD Workshop, 9Y11 November 2002, Naarden, The Netherlands. Neuromuscul Disord 2004;14:754Y66 9. Ferreiro A, Quijano-Roy S, Pichereau C, et al. Mutations of the selenoprotein N gene, which is implicated in rigid spine muscular dystrophy, cause the classical phenotype of multi-minicore disease: Reassessing the nosology of early-onset myopathies. Am J Hum Genet 2002;71:739Y Ferreiro A, Ceuterick-de Groote C, Marks JJ, et al. Desmin-related myopathy with Mallory body-like inclusions is caused by mutations of the selenoprotein N gene. Ann Neurol 2004;55:676Y Clarke NF, Kidson W, Quijano-Roy S, et al. SEPN1: Associated with congenital fiber-type disproportion and insulin resistance. Ann Neurol 2006;59:546Y Monnier N, Ferreiro A, Marty I, et al. A homozygous splicing mutation causing a depletion of skeletal muscle RYR1 is associated with multi-minicore disease congenital myopathy with ophthalmoplegia. Hum Mol Genet 2003;12:1171Y Sewry CA, Muller C, Davis M, et al. The spectrum of pathology in central core disease. Neuromuscul Disord 2002;12:930Y Ferreiro A, Monnier N, Romero NB, et al. A recessive form of central core disease, transiently presenting as multi-minicore disease, is associated with a homozygous mutation in the ryanodine receptor type 1 gene. Ann Neurol 2002;51:750Y Jungbluth H, Müller C, Brockington M, et al. Autosomal-recessive inheritance of RYR1 mutations in a congenital myopathy with cores. Neurology 2002;59:284Y Jungbluth H, Zhou H, Hartley L, et al. Minicore myopathy with ophthalmoplegia caused by mutations in the ryanodine receptor type 1 gene. Neurology 2005;65:1930Y Quinlivan RM, Muller CR, Davis M, et al. Central core disease: Clinical, pathological, and genetic features. Arch Dis Child 2003;88: 1051Y Fananapazir L, Dalakas MC, Cyran F, et al. Missense mutations in the beta-myosin heavy-chain gene cause central core disease in hypertrophic cardiomyopathy. Proc Natl Acad Sci U S A 1993;90: 3993Y Kaindl A, Ruschendorf F, Krause S, et al. Missense mutations of ACTA1 cause dominant congenital myopathy with cores. J Med Genet 2004;41:842Y Bonnemann CG, Thompson TG, van der Ven PF, et al. Filamin C accumulation is a strong but nonspecific immunohistochemical marker of core formation in muscle. J Neurol Sci 2003;206:71Y Phillips MS, Fujii J, Khanna VK, et al. The structural organization of the human skeletal muscle ryanodine receptor (RYR1) gene. Genomics 1996;34:24Y Dirksen RT, Avila G. Altered ryanodine receptor function in central core disease: Leaky or uncoupled Ca(2+) release channels? Trends Cardiovasc Med 2002;12:189Y Petit N, Lescure A, Rederstorff M, et al. Selenoprotein N: An endoplasmic reticulum glycoprotein with an early developmental expression pattern. Hum Mol Genet 2003;12:1045Y Romero NB, Nivoche Y, Lunardi J, et al. Malignant hyperthermia and central core disease: Analysis of two families with heterogeneous clinical expression. Neuromuscul Disord 1993;3:547Y Marty I, Thevenon D, Scotto C, et al. Cloning and characterization of a new isoform of skeletal muscle triadin. J Biol Chem 2000;275: 8206Y Thevenon D, Smida-Rezgui S, Chevessier F, et al. Human skeletal muscle triadin: Gene organization and cloning of the major isoform, Trisk 51. Biochem Biophys Res Commun 2003;303:669Y Vita G, Migliorato A, Baradello A, et al. Expression of cytoskeleton proteins in central core disease. J Neurol Sci 1994;124:71Y De Bleecker JL, Ertl BB, Engel AG. Patterns of abnormal protein expression in target formations and unstructured cores. Neuromuscul Disord 1996;6:339Y Ferreiro A, Estournet B, Chateau D, et al. Multi-minicore diseasev searching for boundaries: Phenotype analysis of 38 cases. Ann Neurol 2000;48:745Y Schroder R, Reimann J, Salmikangas P, et al. Beyond LGMD1A: Myotilin is a component of central core lesions and nemaline rods. Neuromuscul Disord 2003;13:451Y Scacheri PC, Hoffman EP, Fratkin JD, et al. A novel ryanodine receptor gene mutation causing both cores and rods in congenital myopathy. Neurology 2000;55:1689Y96 Ó 2007 American Association of Neuropathologists, Inc. 65