Fourteen patients from 13 unrelated families from France, the French Antilles, and Argentina were included in the present study. Patients were classified into three groups according to their clinical disease severity. P4 and P5 are brothers. P1 to P5 (Group 1) presented a severe/lethal congenital myopathy leading to death in the first days of life. Their muscle biopsy was performed between 2 days and 15 days of life. P6 to P9 presented an intermediate congenital myopathy and a biopsy effectuated between 2 and 10 months (Group 2); P10 to P14 presented typical or mild (P13, and P14) nemaline myopathy and a muscle biopsy performed during childhood or adolescence/early adult life (6 months-21 years; Group 3). The clinical data of these patients were systematically retrieved and retrospectively analysed. Patients were personally examined by one of 6 of the authors. Clinical and genetic characterization of P1, P2, P4, and P5 has been previously reported 4 15 26 .

Patients or parents gave informed consent for the genetic analysis according to French legislation (Comité de Protection des Personnes Est IV DC-2012-1693). Genomic DNA was extracted from blood by standard methods. As nemaline myopathy (NM) is genetically heterogeneous and as the immense size of the nebulin gene significantly impedes classical sequences approaches 12 , we performed exome sequencing on 5 μg of genomic DNA from the patients and their parents as in Böhm et al. 15 .

Exome sequencing was performed at the BGI (Shenzhen) on a Hiseq 2000 (Illumina) by using the Agilent 44 M v2 SureSelect Exon enrichment kit. Variant calling was done with the SOAP software. Variants filtering and prioritization were performed by comparison with SNP databases and with the VaRank program 27 . We discarded polymorphisms with a minor allele frequency (MAF) of more than 0.5% and excluded all variants with a frequency <20% of the total reads for a specific position. Additionally dHPLC and Sanger sequencing was performed in 7 patients as reported in Lehtokari et al. 17 . The mutations are reported according to the coding sequence of the nebulin cDNA reference sequence NM_001164508.1, and its translation.

An open muscle biopsy was performed in all patients after informed consent. Age at biopsy varied from 29 weeks of adjusted gestational age to 21 years. The biopsied muscle is reported in Table 1 and was deltoid in 9 patients (P1, P2, P3, P5, P6, P9, P10, P13, and P14) and vastus lateralis in 5 (P4, P7, P8, P11, and P12). In order to make a precise and comparative study of muscle biopsy findings in Group 1 we standardized the age of new-borns calculating their ‘gestational adjusted age’ as described in Shichiji et al. 28 (Table 1). Samples were analysed in our research laboratory at the Myology Institute in Paris or in the Neuropathology laboratory of FLENI Institute and Garrahan Hospital in Buenos Aires, Argentina. For conventional histochemical techniques 10 μm thick cryostat sections were stained with haematoxylin and eosin (H&E), modified Gomori trichrome (mGT), Periodic acid Schiff technique (PAS), Oil red O, reduced nicotinamide adenine dinucleotide dehydrogenase-tetrazolium reductase (NADH-TR), succinic dehydrogenase (SDH), cytochrome c oxidase (COX), and adenosine triphosphatase (ATPase) preincubated at pH 9.4, 4.63, 4.35. Digital photographs of each biopsy were obtained with a Zeiss AxioCam HRc linked to a Zeiss Axioplan Bright Field Microscope and processed with the Axio Vision 4.4 software (Zeiss, Germany). The fibre type pattern was determined by counting 1000 fibres from each patient in ATPase 9.4 and 4.35 reactions, and by calculating the percentage of type 1 and type 2 fibres.

Based on our experience, and the fibre type proportion reported in the literature regarding the muscle analysed, we considered type 1 fibres predominance to be present when there were more than of 60% type 1 fibre in deltoid muscles, and more than 40% in vastus lateralis muscle 29 . Fibre type distribution in G1 patients was analysed comparing the data on muscle fibre patterns during the main phases of skeletal muscle development obtained from individual with no neumuscular disorder 30 31 . Moreover, where possible, we analysed fibre type proportion in age-matched control biopsies corresponding to G2, and G3 patients.

For the analysis of the proportion of fibres with rods, 800–1000 fibres of the muscle sections of each patient were analysed, and the percentage of fibres appearing with and without rods on the total number of fibres of a muscle sections was calculated; four consecutive, non-overlapping fields were counted. In addition, a classification of the rods and their pattern was effectuated. We defined the rods as being cytoplasmic when localised mainly inside the fibres sparing the subsarcolemmal areas, scattered when they were randomly distributed in the muscle fibre, diffuse, when several small rods were distributed across the whole fibres homogenously occupying the majority of their area, central when distributed mainly in the centre of the cytoplasm, and subsarcolemmal when they were localized in a compact manner close to the fibre membrane (as clusters). We also evaluated the shape of nemaline bodies being mainly globular/ovoid, squared, or elongated.

Frozen muscle samples for immunohistochemical and immunofluorescence analyses were available for 7 patients (P1, P6, P7, P9, P10, P11, and P14). Myosin heavy chain fast (NCL-MHCf, Novocastra Laboratories, Newcastle Upon Tyne, United Kingdom), myosin heavy chain slow (NCL-MHCs, Novocastra Laboratories, Newcastle Upon Tyne, United Kingdom), myosin heavy chain developmental (NCL-MHCd, Novocastra Laboratories, Newcastle Upon Tyne, United Kingdom), myosin heavy chain neonatal (NCL-MHCn, Novocastra Laboratories, Newcastle Upon Tyne, United Kingdom). Antibodies were visualized using immunoperoxidase techniques 28 . Myosin alpha and beta-slow heavy chain, fast 2A heavy chain, and 2X myosin heavy chain (BA-D5, SC-71, and 6H1, Developmental Studies Hybridoma Bank, University of Iowa, Iowa City, USA) immunofluorescence were assessed on 10-μm-thick cryosections over night at 4°C. Subsequently, sections were incubated with appropriate conjugated secondary antibodies for one hour (Alexa Fluor-488 green goat anti-rabbit antibody, and Alexa Fluor-594 red goat anti-mouse antibody, Molecular Probes, Cergy Pontoise, France). A set of control slides was prepared with omission of the primary antibodies.

Detailed electron microscopy analysis was prospectively performed in thirteen patients. Small muscle specimens were fixed with glutaraldehyde (2.5%, pH 7.4), post fixed with osmium tetroxide (2%), dehydrated and embedded in resin (EMBed-812, Electron Microscopy Sciences, USA). Ultra-thin sections from at least three small blocks from each patient were stained with uranyl acetate and lead citrate. The grids were observed using a Philips CM120 electron microscope (80 kV; Philips Electronics NV, Eindhoven, The Netherlands) and were photo documented using a Morada camera (Soft Imaging System, France).

To investigate whether the contractile performance is affected in muscle biopsies from patients with mutations in the nebulin gene, we performed skinned muscle fibre contractility experiments. Small strips were dissected from muscle biopsies of patients P1 (Group 1), P7, P8 and P9 (Group 2), and P10, P11 and P14 (Group 3) and were skinned overnight as described previously 32 . The skinning procedure renders the membranous structures in the muscle fibres permeable, which enables activation of the myofilaments with exogenous Ca²⁺. Preparations were washed thoroughly with relaxing solution and stored in 50% glycerol/relaxing solution at -20°C. Small muscle preparations (cross-sectional area ~0.002 mm²) were dissected from the skinned strips, and were mounted using aluminum T-clips between a length motor (ASI 403A, Aurora Scientific Inc., Ontario, Canada) and a force transducer element (ASI 315C-I, Aurora Scientific Inc., Ontario, Canada) in a single fibre apparatus (ASI 802D, Aurora Scientific Inc., Ontario, Canada) that was mounted on an inverted microscope (Zeiss Axio Observer A1). Sarcomere length was set using a high speed VSL camera and ASI 900B software (Aurora Scientific Inc., Ontario, Canada). Mechanical experiments were performed at a sarcomere length of ~2.2 μm, a length selected to minimize force differences due to shorter thin filaments in fibres from nemaline myopathy patients with nebulin mutations 32 . Fibre width and diameter were measured at three points along the fibre and the cross-sectional area was determined assuming an elliptical cross-section. Two different types of bathing solutions were used during the experimental protocols: a relaxing solution (40 mM BES; 10 mM EGTA; 6.86 mM MgCl₂; 5.96 mM Na-ATP; 3.28 mM K-propionate; 33 mM creatine phosphate; 1 mM DTT; 0.5 mM PMSF; 0.2 mM Leupeptin; 0.05 mM E64) and an activating solution (40 mM BES; 10 mM CaCO₃-EGTA; 6.64 mM MgCl₂; 6.23 mM Na-ATP; 2.1 mM K-propionate; 15 mM creatine phosphate; 1 mM DTT; 0.5 mM PMSF; 0.2 mM Leupeptin; 0.05 mM E64). The temperature of the bathing solutions was controlled by a TEC controller (ASI 825A, Aurora Scientific Inc. Ontario, Canada). Muscle preparations were mounted in a relaxation solution at 1°C. Subsequently, the muscle preparations were pre-activated by switching to an activation solution at 1°C. In that way the myofibres are loaded with calcium, but no force is generated. By rapid switching to an activation solution at 20°C, the fibres were activated and force was generated. When the force trace reached a plateau, the muscle fibres were slacked to 70% of their original length followed by a rapid restretch to the original length after 30 milliseconds. This procedure allows the force to redevelop from zero 33 . The rate of tension redevelopment was calculated by fitting a bi-exponentional through the force redevelopment curve. The first-order rate constant k ₁ reflects crossbridge cycling kinetics and was therefore used in the analyses 34 .

For determination of the myosin heavy chain isoform composition of the muscle fibre preparations we used specialized SDS-PAGE 32 . In brief, muscles fibres were denatured by boiling for 2 minutes in SDS sample buffer. The stacking gel contained a 4% acrylamide concentration (pH 6.7), and the separating gel contained 7% acrylamide (pH 8.7) with 30% glycerol (v/v). The gels were run for 24 h at 15°C and a constant voltage of 275 V. Lastly, the gels were silver-stained, scanned, and analysed with One-D scan EX software (Scanalytics Inc., Rockville, MD, USA).

The ethical committee of La Pitié-Salpêtrière Hospital (CCPPRB) approved this study.

Five patients were female and 9 were male. Clinical summary, laboratory features, a complete list of morphological methods, functional studies applied to muscle biopsies, and genetic characterization of all patients is provided in Table 1. Patient 6 and 7 were classified to be a part of intermediate congenital nemaline myopathy group because they were breathing and moving at birth. However, these patients developed soon after birth a severe clinical picture and they were never able to achieve respiratory independence and/or ambulation. They eventually deceased at 5 months, and 32 months respectively. They therefore present a phenotype in between a severe and an intermediate congenital nemaline myopathy.

To identify the genetic cause in this cohort of patients, we performed exome enrichment and sequencing on genomic DNA from the patients and their parents. Exome sequencing allows a rapid and parallel screening of most human genes, and is suitable and efficient for the diagnosis of neuromuscular diseases and the analysis of large genes such as NEB, frequently mutated in NM 15 . This approach also covers any newly discovered gene for the disorder. For all patients presented here we found known or novel variations in the NEB gene. These changes were confirmed by Sanger sequencing, and their familial segregation validated when parent DNA available. In all the patients, at least one of the two compound heterozygous mutations was a truncating mutation (frameshift or nonsense mutation) leading to protein truncation or degradation, or both (Table 1). The second mutations were, either a frameshift mutation, deletion of several amino acids in frame (splice site mutations), a non-conservative missense change close to an actin binding site (P13), or synonymous variants in P3, P8, and P11 that all impacted on splicing. Most of the mutations (18) affected ubiquitously expressed exons, while six were in the alternatively spliced exons and one in an exon of the triplicated region of eight exons. Overall, most mutations were predicted to lead to a truncated or absent protein. There was no obvious correlation between the type of mutation or its location on the protein and the clinical severity.

Overall, severe NM was associated with a fibre size variability, presence of rods in about 1/3 of fibres, and a high percentage of undifferentiated fibres. By contrast a higher percentage of fibres with rods, and a type 1 fibre predominance/uniformity was noted in the intermediate and typical NM patients. The amount of nemaline bodies seems to be inversely related to clinical severity.

Taken together, our results suggest that myofibrillar dissociation correlated with clinical severity.

The maximal force generation capacity of the muscle fibre preparations was normalized to their cross-sectional area (i.e. maximal active tension) (Figure 5). An overview of the maximal active tension and the rate of tension redevelopment of the muscle fibres are summarized in Table 2. Data are presented as mean ± SEM.

The myosin heavy chain (MHC) gel electrophoresis experiments revealed that muscle preparations from biopsies from patient P1 contained both neonatal isoforms as well as type 1 and type 2A isoforms: (9.2 ± 2.8% MHC neonatal; 40.0 ± 3.2% MHC type 1; 50.8 ± 1.3% MHC type 2A). Muscle preparations from biopsies of P7 and P8 (group 2) contained both MHC type 1 and MHC type 2A isoforms: P7 (66.6 ± 6.6% MHC type 1 and 33.4 ± 6.6% MHC type 2A) and P8 (40.8 ± 12.0% MHC type 1 and 59.2 ± 12.0% MHC type 2A), respectively. All other patient biopsies (P9, P10, P11 and P14) showed exclusively myosin heavy chain type 1 isoforms.