Abstract

Hereditary inclusion body myopathy (HIBM) is an adult onset, slowly progressive distal and proximal myopathy. Although the causing gene, GNE, encodes for a key enzyme in the biosynthesis of sialic acid, its primary function in HIBM remains unknown. To elucidate the pathological mechanisms leading from the mutated GNE to the HIBM phenotype, we attempted to identify and characterize early occurring downstream events by analyzing the genomic expression patterns of muscle specimens from 10 HIBM patients carrying the M712T Persian Jewish founder mutation and presenting mild histological changes, compared with 10 healthy matched control individuals, using GeneChip expression microarrays. When analyzing the expression profile data sets by the intersection of three statistic methods (Student’s t-test, TNoM and Info score), we found that the HIBM-specific transcriptome consists of 374 differentially expressed genes. The specificity of the HIBM transcriptome was assessed by the minimal transcript overlap found between HIBM and the transcriptome of nine additional muscle disorders including adult onset limb girdle myopathies, inflammatory myopathies and early onset conditions. A strikingly high proportion (18.6%) of the overall differentially expressed mRNAs of known function were found to encode for proteins implicated in various mitochondrial processes, revealing mitochondria pathways dysregulation. Mitochondrial morphological analysis by video-rate confocal microscopy showed a high degree of mitochondrial branching in cells of HIBM patients. The subtle involvement of mitochondrial processes identified in HIBM reveals an unexpected facet of HIBM pathophysiology which could at least partially explain the slow evolution of this disorder and give new insights in the disease mechanism.

INTRODUCTION

Hereditary inclusion body myopathy (HIBM) is a neuromuscular disorder characterized by adult onset, slowly progressive distal and proximal muscle weakness and a typical muscle pathology including cytoplasmatic rimmed vacuoles and cytoplasmatic or nuclear filamentous inclusions composed of tubular filaments. This disease is common in the Jewish Persian community, with a prevalence of 1:1500 (1,2), and in other Jewish Middle Eastern population clusters, but with an unusual clinical feature: the sparing of the quadriceps muscle, even in advanced stages of the disease. The same disease, termed distal myopathy with rimmed vacuoles (DMRV), has been described in a cluster of Japanese patients (3,4). The responsible gene, GNE, presents a single homozygous missense mutation (M712T) in all Persian and other Middle Eastern Jewish and non-Jewish HIBM patients (5), confirming the founder effect hypothesis of this disorder in Middle Eastern Jews. More than 50 different mutations in this same gene have now been identified worldwide in quadriceps sparing HIBM non-Jewish patients (5), including Japanese DMRV patients (6,7), establishing HIBM and DMRV as the same entity.

GNE encodes the enzyme UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase (UDP-GlcNAc 2-epimerase/ManNAc kinase; GNE), which is the key enzyme in the biosynthetic pathway of sialic acid (8). Sialic acid, the most abundant terminal monosaccharide on glycoproteins and glycolipids of eukaryotic cells, is involved in multiple biological pathways, playing an important role in many cellular functions, such as adhesion processes, cell migration, inflammation, wound healing and also in metastasis (9,10). GNE catalyzes the first two steps of the sialic acid biosynthesis, the formation of N-acetylmannosamine from UDP-N-acetylglucosamine and the consecutive phosphorylation of the sugar at C6. As a bifunctional enzyme, it consists of two functional domains, an epimerase domain and a kinase domain (11). Most Middle Eastern patients are homozygous for a single mutation, whereas most of the others are found as compound heterozygotes, carrying mutations either both at the epimerase domain, both at the kinase domain, or one in each domain of the enzyme. The broad distribution of mutations along the gene tends to exclude the impairment of a solitary aspect of the enzyme’s function as the underlying cause for HIBM. It suggests instead a more non-specific disruption that results in a general loss of function. Although the role of GNE has been thoroughly recognized as a key enzyme in the biosynthetic pathway of sialic acid, the process by which the mutations in the enzyme lead to muscle disease is not understood. In spite of a slight reduction in the enzymatic activity of most mutated GNE proteins, and some reports of slightly reduced sialylation in histological sections of some patients, there is no clear effect of the mutations upon the overall production of sialic acid in HIBM patients and therefore the data do not support a priori an explanation for the pathology of the disease through the conventional sialic acid pathway (12–17). Although a mutated GNE transgenic mouse model generated on a GNE−/− cell basis (using a D176V GNE missense mutation occurring in the epimerase domain of the enzyme, a prevalent mutation in Japan) showed a strong hyposialylation of most organs in vivo (18), the primary function of GNE in HIBM, impaired by the different mutations, remains to be elucidated.

In recent years, microarray studies and genome-wide expression profiling of normal and diseased skeletal muscle in human and in mouse models have generated a detailed insight into the molecular processes accompanying different conditions (19–24) such as nemaline myopathy (NM) or Duchenne muscular dystrophy (DMD), facilitating our understanding of the underlying disease mechanism. In an effort to elucidate the pathological mechanisms of the mutated GNE gene leading to the HIBM phenotype, we have attempted to identify and characterize the early occurring downstream events of this pathway by analyzing the genome scale expression patterns in the disease target tissue. We have compared gene expression patterns of muscle specimens from HIBM patients carrying the M712T mutation, presenting with mild histological changes, against healthy matched control individuals, using HG-U133A GeneChip expression microarrays. Clusters of genes functionally related to the mutated GNE, and possibly involved in the phenotype of HIBM, were defined and analyzed. In particular functional assessment of mitochondrial involvement was performed by refined mitochondrial morphology analysis.

RESULTS

A distinct expression signature for HIBM-affected muscle

To define gene expression patterns and pathways specifically dysregulated in HIBM and involved early in the disease process, the global mRNA expression profile from 10 HIBM patients carrying the founder Persian Jewish mutation in GNE along with 10 muscle samples from matched healthy individuals (Table 1; Supplementary Material, Fig. S1) was analyzed using Affymetrix HG-U133A GeneChip arrays. Following array signal quantitation and normalization, a set of 7530 active transcripts which represent 34% of the 22 000 different genes present on the array were identified as expressed in muscle tissue, and showed variance among subjects. The differential expression pattern of HIBM was generated from these data by statistical analysis. A signature of 374 transcripts, constituting 5% of the 7530 valid active genes represented on the microarray, were defined as significantly differentially expressed in HIBM compared with control muscles with a minimum geometric fold change of 1.2 (P < 0.05) (Supplementary Material, Table S1). Of these, 194 transcripts were found as overexpressed and the remaining 180 were underexpressed in HIBM versus unaffected muscle tissue. The highest fold change of more than 2.7 was observed for transferrin receptor (TFRC) (P = 0.00034) and with the mitochondrial proton carrier uncoupling protein 3 (UCP3) transcript with a downregulation of 2.3-fold (P = 0.0021). Interestingly, the fold change of the remaining 372 disease signature transcripts was only of a small magnitude, with most of the transcripts being dysregulated in the range of 1.5 to −1.6-fold change; only six genes had a fold higher than 2, and 34 genes higher than 1.5-fold. Table 2 details the 35 genes with the highest fold difference in expression between healthy and HIBM muscle tissue.

Table 1.

Clinical information

Sample id Pathology stage Tissue type Clinical status Age Gender Anesthesia 
B150 Deltoid Affected 26 Local 
B143 Deltoid Affected 55 Local 
B146 Deltoid Affected 45 Local 
B153 Deltoid Affected 29 Local 
B118 Biceps Affected 29 Local 
B155 Quadriceps Affected 59 Local 
B149 Deltoid Affected 40 Local 
B144 Deltoid Affected 46 Local 
B81 Biceps Affected 25 Local 
B77 Tibialis anterior Affected 26 Local 
B100 Quadriceps Normal 18 Local 
B101 Quadriceps Normal 52 Local 
B162 Gluteus Normal 56 General 
B127 Deltoid Normal 26 Local 
B119 Deltoid Normal 19 Local 
B158 Paraspinal Normal 73 General 
B163 Triceps Normal 31 General 
B165 Deltoid Normal 74 General 
B129 Quadriceps Normal 21 Local 
B170 Triceps Normal 48 General 
Sample id Pathology stage Tissue type Clinical status Age Gender Anesthesia 
B150 Deltoid Affected 26 Local 
B143 Deltoid Affected 55 Local 
B146 Deltoid Affected 45 Local 
B153 Deltoid Affected 29 Local 
B118 Biceps Affected 29 Local 
B155 Quadriceps Affected 59 Local 
B149 Deltoid Affected 40 Local 
B144 Deltoid Affected 46 Local 
B81 Biceps Affected 25 Local 
B77 Tibialis anterior Affected 26 Local 
B100 Quadriceps Normal 18 Local 
B101 Quadriceps Normal 52 Local 
B162 Gluteus Normal 56 General 
B127 Deltoid Normal 26 Local 
B119 Deltoid Normal 19 Local 
B158 Paraspinal Normal 73 General 
B163 Triceps Normal 31 General 
B165 Deltoid Normal 74 General 
B129 Quadriceps Normal 21 Local 
B170 Triceps Normal 48 General 

1, no pathological changes; 2, mild changes; 3, moderate–severe changes.

Table 2.

Top differentially expressed transcripts in HIBM

Probe ID Symbol Gene description Biological functiona Fold change P-valueb 
208737_at ATP6V1G1 ATPase, H+ transporting, lysosomal 13 kDa, V1 subunit G isoform 1 ATP biosynthesis 1.491 0.00206 
213379_at CL640 Hypothetical protein CL640 Biosynthesis 1.514 0.01234 
201193_at IDH1 Isocitrate dehydrogenase 1 (NADP+), soluble Glyoxylate cycle 1.524 0.00206 
212328_at KIAA1102 KIAA1102 protein Muscle contraction 1.528 0.00206 
201849_at BNIP3 BCL2/adenovirus E1B 19 kDa interacting protein 3 Apoptosis 1.536 0.00253 
212224_at ALDH1A1 Aldehyde dehydrogenase 1 family, member A1  1.537 0.01234 
218428_s_at REV1L REV1-like (yeast) Mutagenesis 1.548 0.01234 
201273_s_at SRP9 Signal recognition particle 9 kDa  1.548 0.01234 
206765_at KCNJ2 Potassium inwardly rectifying channel, subfamily J, member 2 Ion transport 1.560 0.01234 
204041_at MAOB Monoamine odxidase B Electron transport 1.568 0.00030 
218263_s_at LOC58486 Transposon-derived Buster1 transposase-like protein  1.592 0.00206 
209656_s_at TM4SF10 Transmembrane 4 superfamily member 10  1.605 0.01234 
213019_at RANBP6 RAN-binding protein 6  1.638 0.00206 
203811_s_at DNAJB4 DnaJ (Hsp40) homolog, subfamily B, member 4 Protein folding 1.649 0.01234 
217801_at ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex ATP biosynthesis 1.662 0.01234 
209732_at CLECSF2 C-type (calcium-dependent, carbohydrate recognition domain) lectin Cell adhesion 1.694 0.04923 
208652_at PPP2CA Protein phosphatase 2 (formerly 2A), catalytic subunit, alpha isoform Cell cycle regulation 1.697 0.01234 
218139_s_at C14orf108 Chromosome 14 open-reading frame 108  1.698 0.00206 
201597_at COX7A2 Cytochrome c oxidase subunit VIIa polypeptide 2 (liver) Electron transport 1.702 0.00206 
200777_s_at BZW1 Basic leucine zipper and W2 domains 1 Translation regulation 1.702 0.00206 
217869_at HSD17B12 Hydroxysteroid (17-beta) dehydrogenase 12 Metabolism 1.706 0.01234 
216814_at ACTR3 ARP3 actin-related protein 3 homolog (yeast) Cell motility 1.711 0.01234 
201005_at CD9 CD9 antigen (p24) Cell motility 1.817 0.01234 
203789_s_at SEMA3C Sema domain, immunoglobulin domain (Ig), short basic domain Immune response 1.899 0.00992 
220556_at ATP1B4 ATPase, (Na+)/K+ transporting, beta 4 polypeptide Transport 1.903 0.00206 
200884_at CKB creatine kinase, brain  1.911 0.01234 
203186_s_at S100A4 S100 calcium-binding protein A4  2.007 0.02918 
201744_s_at LUM Lumican Cartilage condensation 2.011 0.03133 
215000_s_at FEZ2 Fasciculation and elongation protein zeta 2 (zygin II) Signal transduction 2.143 0.01362 
213765_at MFAP5 Microfibrillar associated protein 5  2.182 0.01234 
213564_x_at LDHB Lactate dehydrogenase B  2.557 0.01234 
208691_at TFRC Transferrin receptor (p90, CD71) Ion transport 2.746 0.0034 
219827_at UCP3 Uncoupling protein 3 (mitochondrial, proton carrier) Energy pathways −2.272 0.00206 
201010_s_at TXNIP Thioredoxin-interacting protein Signal transduction −1.785 0.02575 
203068_at KIAA0469 KIAA0469 gene product  −1.562 0.01234 
Probe ID Symbol Gene description Biological functiona Fold change P-valueb 
208737_at ATP6V1G1 ATPase, H+ transporting, lysosomal 13 kDa, V1 subunit G isoform 1 ATP biosynthesis 1.491 0.00206 
213379_at CL640 Hypothetical protein CL640 Biosynthesis 1.514 0.01234 
201193_at IDH1 Isocitrate dehydrogenase 1 (NADP+), soluble Glyoxylate cycle 1.524 0.00206 
212328_at KIAA1102 KIAA1102 protein Muscle contraction 1.528 0.00206 
201849_at BNIP3 BCL2/adenovirus E1B 19 kDa interacting protein 3 Apoptosis 1.536 0.00253 
212224_at ALDH1A1 Aldehyde dehydrogenase 1 family, member A1  1.537 0.01234 
218428_s_at REV1L REV1-like (yeast) Mutagenesis 1.548 0.01234 
201273_s_at SRP9 Signal recognition particle 9 kDa  1.548 0.01234 
206765_at KCNJ2 Potassium inwardly rectifying channel, subfamily J, member 2 Ion transport 1.560 0.01234 
204041_at MAOB Monoamine odxidase B Electron transport 1.568 0.00030 
218263_s_at LOC58486 Transposon-derived Buster1 transposase-like protein  1.592 0.00206 
209656_s_at TM4SF10 Transmembrane 4 superfamily member 10  1.605 0.01234 
213019_at RANBP6 RAN-binding protein 6  1.638 0.00206 
203811_s_at DNAJB4 DnaJ (Hsp40) homolog, subfamily B, member 4 Protein folding 1.649 0.01234 
217801_at ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex ATP biosynthesis 1.662 0.01234 
209732_at CLECSF2 C-type (calcium-dependent, carbohydrate recognition domain) lectin Cell adhesion 1.694 0.04923 
208652_at PPP2CA Protein phosphatase 2 (formerly 2A), catalytic subunit, alpha isoform Cell cycle regulation 1.697 0.01234 
218139_s_at C14orf108 Chromosome 14 open-reading frame 108  1.698 0.00206 
201597_at COX7A2 Cytochrome c oxidase subunit VIIa polypeptide 2 (liver) Electron transport 1.702 0.00206 
200777_s_at BZW1 Basic leucine zipper and W2 domains 1 Translation regulation 1.702 0.00206 
217869_at HSD17B12 Hydroxysteroid (17-beta) dehydrogenase 12 Metabolism 1.706 0.01234 
216814_at ACTR3 ARP3 actin-related protein 3 homolog (yeast) Cell motility 1.711 0.01234 
201005_at CD9 CD9 antigen (p24) Cell motility 1.817 0.01234 
203789_s_at SEMA3C Sema domain, immunoglobulin domain (Ig), short basic domain Immune response 1.899 0.00992 
220556_at ATP1B4 ATPase, (Na+)/K+ transporting, beta 4 polypeptide Transport 1.903 0.00206 
200884_at CKB creatine kinase, brain  1.911 0.01234 
203186_s_at S100A4 S100 calcium-binding protein A4  2.007 0.02918 
201744_s_at LUM Lumican Cartilage condensation 2.011 0.03133 
215000_s_at FEZ2 Fasciculation and elongation protein zeta 2 (zygin II) Signal transduction 2.143 0.01362 
213765_at MFAP5 Microfibrillar associated protein 5  2.182 0.01234 
213564_x_at LDHB Lactate dehydrogenase B  2.557 0.01234 
208691_at TFRC Transferrin receptor (p90, CD71) Ion transport 2.746 0.0034 
219827_at UCP3 Uncoupling protein 3 (mitochondrial, proton carrier) Energy pathways −2.272 0.00206 
201010_s_at TXNIP Thioredoxin-interacting protein Signal transduction −1.785 0.02575 
203068_at KIAA0469 KIAA0469 gene product  −1.562 0.01234 

aGO annotation.

bIntersection (TNoMl t-test; Info) P-value.

Notably, a strikingly high number of 56 transcripts, representing 18.6% of the overall differentially expressed mRNAs of known function, were found to encode for proteins implicated in various mitochondrial processes (Table 3).

Table 3.

Mitochondria-related transcripts differently expressed in HIBM

Probe id Symbol Gene description Fold change P-valuea 
208737_at ATP6V1G1 ATPase, H+ transporting, lysosomal 13 kDa, V1 subunit G isoform 1 1.491 0.002057 
210149_s_at ATP5H ATP synthase, H+ transporting, mitochondrial F0 complex, subunit d 1.190 0.002057 
217801_at ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex, ϵ subunit 1.662 0.012341 
216591_s_at SDHC Succinate dehydrogenase complex, subunit C 1.336 0.000080 
204041_at MAOB Monoamine oxidase B 1.568 0.000296 
201931_at ETFA Electron-transfer flavoprotein, alpha polypeptide 1.475 0.002057 
218561_s_at C6orf149 Chromosome 6 open-reading frame 149 1.143 0.012341 
220495_s_at C5orf14 Chromosome 5 open-reading frame 14 1.321 0.012341 
201634_s_at CYB5-M Cytochrome b5 outer mitochondrial membrane precursor 1.109 0.012341 
217329_x_at COX7B Cytochrome c oxidase subunit VIIb 1.277 0.019201 
203609_s_at ALDH5A1 Aldehyde dehydrogenase 5 family, member A1 1.256 0.020496 
201175_at TMX2 Thioredoxin-related transmembrane protein 2 1.329 0.021177 
207328_at ALOX15 Arachidonate 15-lipoxygenase 0.897 0.012341 
217995_at SQRDL Sulfide quinone reductase-like (yeast) 0.753 0.012341 
221028_at MGC11335 Hypothetical protein MGC11335 0.858 0.012341 
203067_at PDHX Pyruvate dehydrogenase complex, component X 1.364 0.012341 
201754_at COX6C Cytochrome c oxidase subunit Vic 1.291 0.000217 
201597_at COX7A2 Cytochrome c oxidase subunit VIIa polypeptide 2 1.702 0.002057 
203613_at NDUFB6 NADH dehydrogenase 1 beta subcomplex, 6, 17 kDa 1.289 0.003729 
217188_s_at C14orf1 Chromosome 14 open-reading frame 1 1.151 0.004065 
201568_at QP-C Low molecular mass ubiquinone-binding protein (9.5 kDa) 1.266 0.002057 
213758_at COX4I1 Cytochrome c oxidase subunit IV isoform 1 1.150 0.012341 
218201_at NDUFB2 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 2, 8 kDa 1.273 0.012341 
214241_at NDUFB8 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 8, 19 kDa 1.262 0.012341 
202026_at SDHD Succinate dehydrogenase complex, subunit D 1.320 0.000652 
204067_at SUOX Sulfite oxidase 1.255 0.012341 
203028_s_at CYBA Cytochrome b-245, alpha polypeptide 1.284 0.016893 
212224_at ALDH1A1 Aldehyde dehydrogenase 1 family, member A1 1.537 0.012341 
218671_s_at ATPIF1 ATPase inhibitory factor 1 1.367 0.012341 
219827_at UCP3 Uncoupling protein 3 0.449 0.002057 
214214_s_at C1QBP Complement component 1, q subcomponent-binding protein 1.285 0.012341 
200807_s_at HSPD1 Heat shock 60 kDa protein 1 1.214 0.033861 
208846_s_at VDAC3 Voltage-dependent anion channel 3 1.329 0.002057 
212604_at MRPS31 Mitochondrial ribosomal protein S31 1.326 0.002057 
217408_at MRPS18B Mitochondrial ribosomal protein S18B 1.154 0.012341 
203517_at MTX2 Metaxin 2 1.388 0.002057 
203095_at MTIF2 Mitochondrial translational initiation factor 2 1.294 0.012639 
202298_at NDUFA1 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 1, 7.5 kDa 1.268 0.000975 
216409_at DKFZp547D104 Hypothetical protein DKFZp547D104 0.894 0.017367 
201816_s_at GBAS Glioblastoma-amplified sequence 1.213 0.040420 
201849_at BNIP3 BCL2/adenovirus E1B 19 kDa interacting protein 3 1.536 0.002527 
214126_at MCART1 Mitochondrial carrier triple repeat 1 1.077 0.013712 
200657_at SLC25A5 Solute carrier family 25, member 5 1.355 0.023624 
213149_at DLAT Dihydrolipoamide S-acetyltransferase 1.365 0.002432 
211569_s_at HADHSC l-3-hydroxyacyl-coenzyme A dehydrogenase, short chain 1.416 0.012341 
218654_s_at MRPS33 Mitochondrial ribosomal protein S33 1.316 0.002095 
217386_at MRPS11 Mitochondrial ribosomal protein S11 1.096 0.012341 
215171_s_at TIMM17A Translocase of inner mitochondrial membrane 17 homolog A (yeast) 1.291 0.002057 
200692_s_at HSPA9B Heat shock 70 kDa protein 9B (mortalin-2) 1.309 0.012341 
219133_at FLJ20604 Hypothetical protein FLJ20604 1.323 0.002057 
217869_at HSD17B12 Hydroxysteroid (17-beta) dehydrogenase 12 1.706 0.012341 
208652_at PPP2CA Protein phosphatase 2 (formerly 2A), catalytic subunit, alpha isoform 1.697 0.012341 
200884_at CKB Creatine kinase, brain 1.911 0.012341 
218049_s_at MRPL13 Mitochondrial ribosomal protein L13 1.427 0.012341 
201193_at IDH1 Isocitrate dehydrogenase 1 (NADP+), soluble 1.524 0.002057 
218027_at MRPL15 Mitochondrial ribosomal protein L15 1.341 0.021037 
Probe id Symbol Gene description Fold change P-valuea 
208737_at ATP6V1G1 ATPase, H+ transporting, lysosomal 13 kDa, V1 subunit G isoform 1 1.491 0.002057 
210149_s_at ATP5H ATP synthase, H+ transporting, mitochondrial F0 complex, subunit d 1.190 0.002057 
217801_at ATP5E ATP synthase, H+ transporting, mitochondrial F1 complex, ϵ subunit 1.662 0.012341 
216591_s_at SDHC Succinate dehydrogenase complex, subunit C 1.336 0.000080 
204041_at MAOB Monoamine oxidase B 1.568 0.000296 
201931_at ETFA Electron-transfer flavoprotein, alpha polypeptide 1.475 0.002057 
218561_s_at C6orf149 Chromosome 6 open-reading frame 149 1.143 0.012341 
220495_s_at C5orf14 Chromosome 5 open-reading frame 14 1.321 0.012341 
201634_s_at CYB5-M Cytochrome b5 outer mitochondrial membrane precursor 1.109 0.012341 
217329_x_at COX7B Cytochrome c oxidase subunit VIIb 1.277 0.019201 
203609_s_at ALDH5A1 Aldehyde dehydrogenase 5 family, member A1 1.256 0.020496 
201175_at TMX2 Thioredoxin-related transmembrane protein 2 1.329 0.021177 
207328_at ALOX15 Arachidonate 15-lipoxygenase 0.897 0.012341 
217995_at SQRDL Sulfide quinone reductase-like (yeast) 0.753 0.012341 
221028_at MGC11335 Hypothetical protein MGC11335 0.858 0.012341 
203067_at PDHX Pyruvate dehydrogenase complex, component X 1.364 0.012341 
201754_at COX6C Cytochrome c oxidase subunit Vic 1.291 0.000217 
201597_at COX7A2 Cytochrome c oxidase subunit VIIa polypeptide 2 1.702 0.002057 
203613_at NDUFB6 NADH dehydrogenase 1 beta subcomplex, 6, 17 kDa 1.289 0.003729 
217188_s_at C14orf1 Chromosome 14 open-reading frame 1 1.151 0.004065 
201568_at QP-C Low molecular mass ubiquinone-binding protein (9.5 kDa) 1.266 0.002057 
213758_at COX4I1 Cytochrome c oxidase subunit IV isoform 1 1.150 0.012341 
218201_at NDUFB2 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 2, 8 kDa 1.273 0.012341 
214241_at NDUFB8 NADH dehydrogenase (ubiquinone) 1 beta subcomplex, 8, 19 kDa 1.262 0.012341 
202026_at SDHD Succinate dehydrogenase complex, subunit D 1.320 0.000652 
204067_at SUOX Sulfite oxidase 1.255 0.012341 
203028_s_at CYBA Cytochrome b-245, alpha polypeptide 1.284 0.016893 
212224_at ALDH1A1 Aldehyde dehydrogenase 1 family, member A1 1.537 0.012341 
218671_s_at ATPIF1 ATPase inhibitory factor 1 1.367 0.012341 
219827_at UCP3 Uncoupling protein 3 0.449 0.002057 
214214_s_at C1QBP Complement component 1, q subcomponent-binding protein 1.285 0.012341 
200807_s_at HSPD1 Heat shock 60 kDa protein 1 1.214 0.033861 
208846_s_at VDAC3 Voltage-dependent anion channel 3 1.329 0.002057 
212604_at MRPS31 Mitochondrial ribosomal protein S31 1.326 0.002057 
217408_at MRPS18B Mitochondrial ribosomal protein S18B 1.154 0.012341 
203517_at MTX2 Metaxin 2 1.388 0.002057 
203095_at MTIF2 Mitochondrial translational initiation factor 2 1.294 0.012639 
202298_at NDUFA1 NADH dehydrogenase (ubiquinone) 1 alpha subcomplex, 1, 7.5 kDa 1.268 0.000975 
216409_at DKFZp547D104 Hypothetical protein DKFZp547D104 0.894 0.017367 
201816_s_at GBAS Glioblastoma-amplified sequence 1.213 0.040420 
201849_at BNIP3 BCL2/adenovirus E1B 19 kDa interacting protein 3 1.536 0.002527 
214126_at MCART1 Mitochondrial carrier triple repeat 1 1.077 0.013712 
200657_at SLC25A5 Solute carrier family 25, member 5 1.355 0.023624 
213149_at DLAT Dihydrolipoamide S-acetyltransferase 1.365 0.002432 
211569_s_at HADHSC l-3-hydroxyacyl-coenzyme A dehydrogenase, short chain 1.416 0.012341 
218654_s_at MRPS33 Mitochondrial ribosomal protein S33 1.316 0.002095 
217386_at MRPS11 Mitochondrial ribosomal protein S11 1.096 0.012341 
215171_s_at TIMM17A Translocase of inner mitochondrial membrane 17 homolog A (yeast) 1.291 0.002057 
200692_s_at HSPA9B Heat shock 70 kDa protein 9B (mortalin-2) 1.309 0.012341 
219133_at FLJ20604 Hypothetical protein FLJ20604 1.323 0.002057 
217869_at HSD17B12 Hydroxysteroid (17-beta) dehydrogenase 12 1.706 0.012341 
208652_at PPP2CA Protein phosphatase 2 (formerly 2A), catalytic subunit, alpha isoform 1.697 0.012341 
200884_at CKB Creatine kinase, brain 1.911 0.012341 
218049_s_at MRPL13 Mitochondrial ribosomal protein L13 1.427 0.012341 
201193_at IDH1 Isocitrate dehydrogenase 1 (NADP+), soluble 1.524 0.002057 
218027_at MRPL15 Mitochondrial ribosomal protein L15 1.341 0.021037 

aIntersection (TNoM; t-test; Info) P-value.

Differentially expressed transcripts in HIBM

To detect a finer structure of transcriptional variation within each group, hierarchical clustering (Fig. 1) performed on the 374 mRNAs differentially expressed genes selected by the intersection analysis, identified separate clusters for HIBM muscles and normal controls and demonstrated clear segregation of muscle biopsies taken from HIBM patients from the control unaffected samples. The two normal outliers classified with the affected individuals were both from relatively adult individuals (48 and 56 years old) but no direct explanation could be given for their outlier clustering, together with the oldest affected individuals. On the normal branch of the cluster, samples distribution did not correlate with either age or muscle type and no obvious correlation between the disease state and muscle type was observed within the group of affected individuals (Fig. 1, top panel). Conversely, the disease expression distribution as displayed in the cluster does correlate with the severity of muscle pathology, from specimens with moderate changes on the left side of the cluster (B77 and B143), then mild changes and finally with very mild or no pathological changes (Fig. 1). This expression pattern suggests that disease state has a stronger impact on gene expression clustering compared with other clinical and demographic factors examined.

Figure 1.

Hierarchical clustering differentiates HIBM from control individuals. Hierarchical clustering dendogram of the 374 transcripts (rows) showing gene expression patterns that most reflect differences between muscle samples (columns) from 10 HIBM-affected individuals and 10 unaffected controls (from Supplementary Material, Table S1). Genes were clustered using Cluster software (53). All muscle specimens from affected individuals (blue) fell into one branch of the dendogram, whereas, with two exceptions, samples from unaffected controls (orange) fell into a separate branch. Muscle type for each sample is shown. Heatmap tiles show relative intensity signal levels represented by color, with red denoting high expression and green indicating low expression levels.

Figure 1.

Hierarchical clustering differentiates HIBM from control individuals. Hierarchical clustering dendogram of the 374 transcripts (rows) showing gene expression patterns that most reflect differences between muscle samples (columns) from 10 HIBM-affected individuals and 10 unaffected controls (from Supplementary Material, Table S1). Genes were clustered using Cluster software (53). All muscle specimens from affected individuals (blue) fell into one branch of the dendogram, whereas, with two exceptions, samples from unaffected controls (orange) fell into a separate branch. Muscle type for each sample is shown. Heatmap tiles show relative intensity signal levels represented by color, with red denoting high expression and green indicating low expression levels.

Notably, when applying an unsupervised Principal Component Analysis (PCA) approach to the data set (data not shown), the 20 samples did not cluster into affected versus control groups, but rather most of the HIBM-affected samples clustered together, whereas the healthy controls seemed to cluster by age, thus suggesting a common process in HIBM muscles, independent of the variable degrees of pathology. In order to address the possibility that differential expression could be associated with the age of the donor, as particularly in the normal control group age was not entirely uniform, we estimated the effect of age on gene expression. Among the 374 differentially expressed genes, Pearson correlation test pointed to correlation between age and expression of 12 genes (LYPLA1, MRPS33, BNIP3, CL40, JMID3, MGC4504, UROS, SDBCAG84, CDON, GBAS, TMX2, MRPL13). However, covariance analysis of the expression of these 12 genes for age in all examined muscle specimens remained significantly different between the affected and the control groups (P = 0.004; P = 0.001; P ≤ 0.001; P = 0.005; P = 0.007; P = 0.002; P = 0.001; P ≤ 0.001; P = 0.034; P = 0.046; P = 0.023; P = 0.005, respectively).

Quantitative real-time PCR confirmed the differential expression identified by the microarray analysis

To validate the differential expression identified by microarray analysis, using an independent method, we performed quantitative real-time PCR analysis for a subset of eight genes with the highest fold change (LDHB, S100A4, UCP3, BNIP3, COX7A2, ATP5E, TFRC, TXNIP). Four of the samples previously analyzed on the arrays (two HIBM patients and two control muscle samples) were assayed in this experiment. As illustrated in Fig. 2, the expression level of six of the transcripts, as assessed by real-time PCR, correlated with the microarray results and verified the directional change mode and the magnitude of the change. COX7A2, ATP5E, LDHB, BNIP3 and TFRC were overexpressed in the range from 2 to 6-fold in the HIBM muscle biopsies, whereas UCP3 was downregulated ∼3-fold in muscle biopsies taken from HIBM patients. In contrast to the microarray results, S100A4 and TXNIP expression evaluated by real-time PCR did not clearly distinguish between normal and HIBM muscle in those samples.

Figure 2.

Validation of differential gene expression by quantitative real-time PCR. Eight genes identified as dysregulated in HIBM patients were analyzed using quantitative real-time PCR: COX7A2, cytochrome c oxidase subunit VIIa polypeptide 2; BNIP3, BCL2/adenovirus E1B 19 kDa interacting protein 3; ATP5PE, ATP synthase, H+ transporting, mitochondrial F1 complex, epsilon subunit; TFRC, transferrin receptor; LDHB, lactate dehydrogenase B; UCP3, uncoupling protein 3; TXNIP, thioredoxin-interacting protein; S100A4, S100 calcium-binding protein A4. Relative expression levels of genes are indicated by the fold changes in expression levels for HIBM patients and control-unaffected individuals. The expression level of a specific gene in the different samples is represented as the fold value of its expression relative to sample B144, which has been assigned a value of 1 (RQ, relative quantity). PCR reactions were performed in duplicate using RNA samples from two HIBM-affected and two control individuals previously analyzed on the microarray platform.

Figure 2.

Validation of differential gene expression by quantitative real-time PCR. Eight genes identified as dysregulated in HIBM patients were analyzed using quantitative real-time PCR: COX7A2, cytochrome c oxidase subunit VIIa polypeptide 2; BNIP3, BCL2/adenovirus E1B 19 kDa interacting protein 3; ATP5PE, ATP synthase, H+ transporting, mitochondrial F1 complex, epsilon subunit; TFRC, transferrin receptor; LDHB, lactate dehydrogenase B; UCP3, uncoupling protein 3; TXNIP, thioredoxin-interacting protein; S100A4, S100 calcium-binding protein A4. Relative expression levels of genes are indicated by the fold changes in expression levels for HIBM patients and control-unaffected individuals. The expression level of a specific gene in the different samples is represented as the fold value of its expression relative to sample B144, which has been assigned a value of 1 (RQ, relative quantity). PCR reactions were performed in duplicate using RNA samples from two HIBM-affected and two control individuals previously analyzed on the microarray platform.

Functional annotation revealed mitochondria pathways dysregulation

Our analyses identified a signature of 374 differentially expressed transcripts whose expression levels correlate with HIBM muscle phenotypes. To gain a better understanding of these transcripts and to uncover common functions among the differentially expressed genes, we examined their functional classifications and analyzed for enrichment in the gene ontology (GO) (25) categories for cellular processes using the DAVID bioinformatics resource (26). Although the largest resulting category regrouped 73 genes of unknown function, several functional clusters were identified among the additional genes representing the disease signature. Most notably, genes related to mitochondrial processes and structure (P = 6.6E−14) and transcripts encoding for proteins involved in transport activity (P = 6.8E−5) were most significantly enriched among the genes dysregulated in HIBM patients. Genes related to oxidative phosphorylation, including ATP5E, ATP5H, ATP6V1G1, COX4I1, COX6C, COX7A2, COX7B, NDUFA1, NDUFB2, NDUFB6, NDUFB8 and SDHC (P = 7.5E−5) and intracellular transport (P = 7.0E−4) as well as transcripts encoding for signal recognition particle (P = 1.8E−4) and cytoplasmic vesicles (P = 1.9E−3) were also enriched in the set of differentially expressed genes.

To identify functional modules of gene expression and interacting partners that are relevant to the HIBM gene expression signature, we applied annotation enrichment analysis through the Ingenuity Pathways Knowledge Base (IPA) (27). Fisher’s exact test was performed to determine the likelihood of the 374 genes to participate in a given function or pathway, relative to the total number of occurrences of these genes in all functional annotation in the IPA. Principal functions associated with the signature were cell cycle (P = 2.6E−4), molecular transport (P = 4.81E−4), cell morphology (P = 7.6E−4) and cellular growth and proliferation (P = 5.21E−4) in which four out of the 51 molecules in the group, ALOX15, GNAQ, HBEGF and NFATC4, had been previously related to hypertrophy of muscle cells (28–31). Metabolic signaling pathways consisting of molecules such as ALDH1A1, MAPK3, MAOB, PPP2CA, GSTP1 and others were also predicted to be involved in the disease process (27).

HIBM transcriptome specificity

Several of the individual genes found to be dysregulated in this study have been implicated in previous studies on other muscular disorders. To examine whether the transcriptome signature of HIBM muscle and particularly mitochondria hyperactivation is a common phenomena also in the expression profile of other human muscular disorders and to exclude the possibility that the relatively mild fold change observed in HIBM represents in fact non-specific secondary events occurring similarly in many other muscular diseases as common downstream pathways, we have analyzed the signature set of the 374 genes for specificity. In order to evaluate the specificity of the transcriptional profile to HIBM, the complete Gene Expression Omnibus (GEO) data sets from nine different muscular disorders were normalized and reanalyzed with the same statistical tools as HIBM (intersection of TNoM, t-test and Info score), and subsequently the resulting status of the 374 transcripts were looked at specifically. We applied this new analysis to the following conditions: DMD, Becker muscular dystrophy (BMD), NM, limb girdle muscular dystrophy 2I (LGMD2I), all of which are congenital or early onset muscular disorders; adult onset myopathies such as dysferlinopathy (LGMD2B, limb girdle muscular dystrophy 2B), limb girdle muscular dystrophy 2A (LGMD2A), facioscapulohumeral dystrophy (FSHD) and inflammatory myopathies including dermatomyositis (DM) and inclusion body myositis (IBM). Interestingly, among the genes commonly dysregulated in almost all these conditions were the mitochondrial UCP3 and S100A4, a calcium-binding protein known to be involved in the regulation of a number of cellular processes including cell cycle progression and differentiation. Although the fold-change expression of S100A4 found in the microarray for HIBM could not be confirmed by real-time PCR, it maybe worth mentioning that S100A4 was also recently shown to be increased in hypertrophic rat hearts and to have a pro-cardiomyogenic effects in embryonic stem cell (32). Figure 3 summarizes the overall results and divides the analysis into three subclasses, showing only minor overlap between HIBM and these other muscular diseases. Interestingly, despite no known phenotypic or pathological similarity, 52 (14%) and 45 (12%) different transcripts were found to be commonly dysregulated in HIBM and LGMD2A or LGMD2B, respectively. In the case of IBM, a sporadic inflammatory muscle disorder sharing many pathological features with HIBM and sometimes considered as the sporadic form of the hereditary disorder, only a subset of 27 dysregulated genes were found to overlap with HIBM. For the additional muscle disorders analyzed, only a small overlap was found and usually at a much higher fold change (Fig. 3; Supplementary Material, Tables S2a–i). These findings emphasize the ability of the present expression signature to distinguish between HIBM and other muscle disorders and lend further support to the hypothesis that mitochondria hyperactivation in HIBM is a unique phenomenon.

Figure 3.

HIBM transcriptome specificity. Expression data sets from previous studies on nine different muscular disorders were collected through GEO database and analyzed using the same statistical approach (Intersection of TNoM, t-test and Info) as the current HIBM data set. Number of genes dysregulated in HIBM which are also dysregulated in the nine other muscular diseases are presented in three separated subgroups. (A) Overlapping dysregulated transcripts in HIBM and inflammatory myopathies (inclusion body myopathy, IBM and dermatomyositis, DM). (B) Overlapping dysregulated transcripts in HIBM and adult onset dystrophies LGMD2B, LGMD2A and FSHD. (C) Overlapping dysregulated transcripts in HIBM and early onset muscular disorders including DMD/BMD, NM and LGMD2I. Numbers in white represent the number of overlapping transcripts differentially expressed in two or more conditions, as labeled, among the 374 transcripts representing the HIBM transcriptome signature. Numbers in yellow represent the number of differentially expressed transcripts unique to HIBM. Circle areas are not to scale. The identity of the genes overlapping with the HIBM transcriptome for each condition is depicted in Supplementary Material, Tables S2a–i.

Figure 3.

HIBM transcriptome specificity. Expression data sets from previous studies on nine different muscular disorders were collected through GEO database and analyzed using the same statistical approach (Intersection of TNoM, t-test and Info) as the current HIBM data set. Number of genes dysregulated in HIBM which are also dysregulated in the nine other muscular diseases are presented in three separated subgroups. (A) Overlapping dysregulated transcripts in HIBM and inflammatory myopathies (inclusion body myopathy, IBM and dermatomyositis, DM). (B) Overlapping dysregulated transcripts in HIBM and adult onset dystrophies LGMD2B, LGMD2A and FSHD. (C) Overlapping dysregulated transcripts in HIBM and early onset muscular disorders including DMD/BMD, NM and LGMD2I. Numbers in white represent the number of overlapping transcripts differentially expressed in two or more conditions, as labeled, among the 374 transcripts representing the HIBM transcriptome signature. Numbers in yellow represent the number of differentially expressed transcripts unique to HIBM. Circle areas are not to scale. The identity of the genes overlapping with the HIBM transcriptome for each condition is depicted in Supplementary Material, Tables S2a–i.

Mitochondrial morphology is affected in HIBM muscle cells

Alterations in mitochondrial metabolism and protein expression are tightly linked to mitochondrial structure (33,34), Therefore, since the expression of mitochondria- related transcripts was remarkably altered, we quantitatively analyzed mitochondrial shape. To this end, primary cell cultures were established from three HIBM and three control individuals' biopsies. Two muscle cultures of affected individuals were derived from two of the biopsies analyzed in the microarray experiment (ms17 from B150 and ms15 from B149) and were matched to two novel controls by age, gender and muscle type (ms28, 20-year-old male, deltoid, and ms35, 46-year-old male, deltoid). The third pair was derived from the tibialis anterior muscle of a 30-year-old female (ms6; HIBM affected) and from the quadriceps of a 28-year-old control female (ms7).

For mitochondrial morphology analysis, cells were stained with the mitochondria-specific fluorescent cation rhodamine 123 (R123) and analyzed in three independent experiments (Fig. 4). We previously validated this approach and demonstrated its suitability for morphology analysis in living cells (34–37). Parameters investigated were the number of mitochondria (Nc), their aspect ratio (AR) as well as their form-factor (F), reflecting the degree of mitochondrial branching. No significant differences were observed for Nc between the individual control and HIBM cell cultures (Supplementary Material, Figs S2 and 4B; gray bars). This demonstrates that the mitochondrial network is not fragmented in HIBM cells. Although AR was higher in HIBM cells in one pair (ms6 versus ms7), lower for HIBM cells in another pair (ms17 versus ms28) and similar for the third pair (ms15 versus ms35) (Supplementary Material, Fig. S2), the differences in average AR were not statistically significant (Fig. 4B; open bars). In contrast, the quantification of F revealed a statistically significantly higher value in patient cells, indicating that mitochondria are more branched in cells of HIBM patients than in controls (Supplementary Material, Fig. S2; Fig. 4A and B; black bars).

Figure 4.

Quantification of mitochondrial shape in HIBM myoblasts. (A) Typical example of a control (ms35; top left panel; CT) and HIBM patient myoblast (ms17; lower left panel) stained with the mitochondria-specific cation rhodamine 123 (RAW). Following image processing, a binary image (BIN) was obtained from which mitochondrial shape was quantified. (B) Average values of mitochondrial aspect ratio (AR), degree of branching (F) and number of mitochondria per cell (Nc) in a cohort of three control (CT) and three patient cells.

Figure 4.

Quantification of mitochondrial shape in HIBM myoblasts. (A) Typical example of a control (ms35; top left panel; CT) and HIBM patient myoblast (ms17; lower left panel) stained with the mitochondria-specific cation rhodamine 123 (RAW). Following image processing, a binary image (BIN) was obtained from which mitochondrial shape was quantified. (B) Average values of mitochondrial aspect ratio (AR), degree of branching (F) and number of mitochondria per cell (Nc) in a cohort of three control (CT) and three patient cells.

DISCUSSION

We have identified a transcriptional signature in HIBM consisting of 374 genes, most of them displaying a low fold change differential expression. This finding could reflect the mild myopathic changes examined or the adult onset and slowly progressive course of HIBM. Alternatively, the summation of different cell types including connective and endothelial tissue cells, as well as affected and spared muscle cells, may result in averaging down the expression differences specific to affected muscle cells.

The major group of differentially expressed transcripts in HIBM muscle compared with healthy tissue represents processes taking place in mitochondria. Mitochondria are the essential organelles for energy supply of cells and it is now well established that mitochondria are not only the major suppliers of ATP, but also that they can respond to multiple physiological stress via signaling processes, cell growth and differentiation events. In particular, they play a major role in cell death as sensors of apoptotic signals, by releasing various pro-apoptotic molecules into the cell cytoplasm. Mitochondrial intrinsic apoptosis pathway signaling cascade can be induced by several stimuli causing cell damage, oxidative stress, ER stress and other insults. Interestingly, the transcriptome signature of HIBM muscle includes transcripts involved in all these pathways, electron transport, signal transduction, cytoskeleton and cell organization, as well as in cell death. The precise events upstream to mitochondria apoptotic involvement remain to be fully characterized, however it is well accepted that Akt is a key factor in this apoptotic/cell survival process. Notably, our previous findings of a primary impairment of apoptotic events and survival defects in HIBM cells, as shown by the lack of activated Akt response to apoptotic stimuli (38), are in line with the major mitochondrial involvement reported in these studies. Also consistent with our present observations, our previous studies pointed to a slow, but accumulating process of apoptosis of HIBM myotubes; indeed this process is in line with the mild mitochondrial alterations described here, which very likely stimulate a cascade of events leading eventually to increased apoptosis, as illustrated by alterations in the cytochrome c and various cell death protein genes such as DEDD and DAXX. Myotubes and myofibers are long-lasting structures which demand a very orchestrated combination of activities to ensure proper maintenance for survival. Mild defects in even a small number of targets in this cascade, such as traffic, ubiquitination, cytoskeleton organization or cell adhesion functions, could impair the subtle functional balance in the myofiber, causing maintenance difficulties and eventual apoptosis. Mitochondria can form a more or less complex functional network within the cytosol, which is thought to be linked to mitochondrial and cellular metabolism both during healthy and pathological conditions (33,34). Here we show that in HIBM muscle cell cultures, mitochondria are significantly more branched than in control cells, whereas their length and number are the same. Interestingly, fibroblasts from complex I-deficient patients containing highly branched mitochondria displayed a much less severe biochemical deficiency than patient cells with fragmented mitochondria (35). This suggests that in HIBM cells, the increased mitochondrial branching (compatible with mitochondrial biogenesis) could be part of a compensatory response. These most likely very slow and accumulating events could explain why no major changes in mitochondria had been reported in HIBM biopsies in any previous study when analyzed by more conventional techniques which detect more severe changes such as important deformations, ragged fibers as a parameter of mitochondrial proliferation or Cox deficiency as mitochondrial functional impairment, which are well recognized in sporadic IBM (39), but not in HIBM (40) (Supplementary Material, Fig. S3). Only when examined by more sensitive tools, very subtle morphological changes as the branching structure of mitochondria could be traced. Changes in mitochondrial morphology are observed in a variety of conditions, including apoptosis (41) and oxidative stress (35), and have been shown to affect signal propagation and cell response to apoptotic stimuli (42).

The vast majority of the 374 differential transcripts detected in this study represents an HIBM-specific, unique mRNA signature. Only few of these transcripts were also identified in other muscular disorders, either early onset dystrophies or myopathies such as DMD, BMD, NM, adult onset myopathies such as LGMD2B, LGMD2A, or the inflammatory myopathies DM and even the pathologically close entity IBM, where the same pathological changes have been described as in HIBM, such as rimmed vacuoles and inclusion bodies. These findings point to the fact that although the final muscle pathology of several myopathies is characterized by loss of muscle mass, the HIBM-specific transcriptome profile reported reflects a combination of events occurring most probably upstream of common pathological pathways and therefore represents either relatively early events in the mechanism initiated by the GNE mutation or specific reactions to such events.

Recently, accumulating evidence points to the role of glycosphingolipids, and especially gangliosides, in mitochondrial- mediated regulation of apoptosis events and the development of pathological processes (43,44). Interestingly, in human embryonic kidney cells, GNE has been reported to regulate the levels of GM3 and GD3 synthases, the enzymes responsible for the incorporation of the first and the second sialic acid structure, respectively, on glycospingolipids, and subsequently the amounts of their respective biosynthetic products, the gangliosides GM3 and GD3 (45). In particular, it has been shown that GD3 can directly interact with mitochondria, recruiting this organelle to the apoptosis program (46). GD3 ganglioside is thought to reach mitochondria by the physical continuity occurring between the endoplasmic reticulum and early Golgi with mitochondrial membranes; this mechanism of recruitment is slower than the diffusion process of soluble products within the cytosol and therefore is expected to generate slower kinetics of apoptosis (45). Furthermore, the specificity of GD3 effects seems to be due to specific structural features of the two sialic acid residues because other gangliosides, such as GM3 and GD1a, do not have any effects on mitochondrial changes (43).

Although GNE transcript and protein levels were found to be similar in HIBM and control in this study and in an earlier one (47), respectively, the GNE mutation, by affecting either sialylation or still unrecognized mechanisms, could initiate a slow process of signal transduction, mostly engaging mitochondrial pathways and mildly affecting a variety of physiological processes, including remodeling of the cytoskeleton, which would lead eventually to apoptosis fate. Although mitochondria play a major role in muscle and indeed various muscular conditions eventually show mitochondria defects and cell injury, those are robust events recognized relatively easily (39,48). The rather subtle involvement of mitochondrial processes identified in HIBM reveals an unexpected facet of HIBM pathophysiology which could at least partially explain the slow evolution of this disorder and give new insights into the disease mechanism.

MATERIALS AND METHODS

Patients and muscle specimens

A total of 20 muscle specimens were available for this study based on clinical, histological and molecular diagnosis (Table 1), all in compliance with an approved protocol by the Institutional Review Board of Wolfson Hospital, Holon, Israel. All muscle biopsies were collected and immediately frozen either in liquid nitrogen until used for RNA extraction or in isopentane for the preparation of frozen sections. An additional muscle specimen was placed in PBS and used for the establishment of primary myoblast culture for some individuals. The 10 normal muscle specimens were collected from consenting individuals who had undergone either muscle biopsy which eventually was diagnosed as normal, or orthopedic surgery. All patients were of Jewish Persian descent carrying the homozygous M712T mutation in GNE. In order to detect changes as primary as possible, we selected only mildly HIBM-affected muscles biopsies when possible.

The pathological changes in the patients’ biopsies were evaluated by hematoxilin–eosin staining of frozen sections (Supplementary Material, Fig. S1) and ranged from non-pathological changes in two cases, mild changes in six cases (including myopathic changes and central nuclei), to moderate changes presenting atrophic fibers, central nuclei, cytoplasmic rimmed vacuoles and some extent of fibrosis, in two cases. Notably, the quadriceps muscle biopsied from a severely affected patient (B155), already in a wheelchair, was also mildly affected.

Muscle cell cultures

Primary muscle cultures from HIBM patients carrying the M712T mutation in GNE and from non-affected control individuals were established as described previously (38). Myoblasts were grown in growth medium (GM) until ∼70% confluence. To initiate differentiation, GM was replaced by differentiation medium (DM) containing 2% horse serum (HS). All studies were carried out on cultures at passage 4–8.

RNA preparation and microarray hybridization

Following tissue disruption in dry ice and homogeneization, total RNA was isolated using TriReagent (Molecular Research Center, Cincinnati, OH, USA) according to the manufacturer’s recommended protocol. The integrity of the isolated RNA was analyzed using a standard 1% agarose gel electrophoresis, and purified total RNA was quantified using Nanodrop ND1000. Subsequently, RNA was labeled, fragmented and individually hybridized to Affymetrix HG-U133A oligonucleotide arrays (Affymetrix Incorporated, Santa Clara, CA, USA). The complete recommended protocol for preparation and microarray processing as well as detailed annotations of HG-U133A arrays are available at the Affymetrix URL (http://www.affymetrix.com). Briefly, 5 µg of mRNA was used to generate first-strand cDNA by using a T7-linked oligo (dT) primer. After second-strand synthesis, in vitro transcription was performed with biotinylated UTP and CTP (Enzo Diagnostics, Farmingdale, NY, USA), resulting in ∼300-fold amplification of RNA. The target cDNA generated from each sample was processed as per the manufacturer’s recommendation using an Affymetrix Gene Chip Instrument System. Briefly, spike controls were added to 15 µg fragmented cRNA before hybridization. Arrays were then washed and stained with streptavidin–phycoerythrin, before being scanned on an Affymetrix GeneChip scanner, and the expression value for each gene was calculated using Affymetrix Microarray Software Suite 5.0 (MAS5.0).

Microarray analysis and normalization

For data normalization, all array scans were pre-processed with RMAExpress software (49), which employs RMA normalization (background adjustment and quantile normalization), to summarize the final values from the scanner calls. We considered genes to be valid for analysis if the gene probe was defined as ‘Present’ (P) according to MAS 5.0 for at least three out of the 20 samples analyzed. This resulted in 7530 valid genes. For each gene g and each sample i, the expression value eg,i was divided by the geometric mean of the gene across all samples n√(Πi =1neg,i). Expression ratios were then transformed to log (base 2). The complete microarray raw and processed data have been deposited in the GEO database (series number GSE12648).

Statistical analysis and clustering

The general approach to analysis has been previously outlined (50) and performed using the ScoreGenes package (http://compbio.cs.huji.ac.il/scoregenes/). To identify probe sets which are significantly differentially expressed in HIBM versus normal muscle tissue and to best distinguish between the groups, three powerful statistical tests were applied to the database: threshold number of misclassifications, TNoM (a non-parametric test that measures the number of classification errors committed when using the best simple threshold to distinguish between two classes, based on the expression levels of the given gene) (51), Student’s t-test (two-tailed t-test to measure whether the mean expression of the gene in the two classes are significantly different) and Info (a non-parametric test that estimates the uncertainty remaining about the class of a sample after observing the expression of the individual gene). A lower Info score indicates a higher predictive value for a given gene (51). In all three methods, we standardize the score by using P-values that report the probability of getting this score under the null hypothesis. For the t-test, the hypothesis is that the gene has the same mean expression in both classes; in the TNoM and Info, the null hypothesis is that the sample labels are independent of the expression value. We compute t-test P-values using Student’s t distribution. The computation of P-value for TNoM and Info is based on exact computation (52). Eventually, differentially expressed genes between the two tagged groups of samples were evaluated by the intersection of the three methods: to select statistically significant genes, we chose all overlapping genes that displayed a P-value of at most 0.05 in all three scoring methods.

Biological pathway analysis

Functional annotation clustering was performed by DAVID (http://david.abcc.ncifcrf.gov) (26) with medium classification stringency. The co-regulated functional categories among the differentially expressed genes were ranked according to a modified Fisher’s exact test and referred to as ‘enrichment score’. Changes in gene expression between the experimental groups were tested for enrichment in functional annotations. Ingenuity Pathway Analysis (IPA) was further used to find significant pathways related to the genes dysregulated in HIBM patients. This significance is expressed as P-value calculated using the right-tailed Fisher’s exact test to measure the likeliness of genes from the data set file to participate in that biological function or pathway, as defined in IPA.

Quantitative real-time RT–PCR

To validate the microarray results, we quantified the expression of several representing transcripts: COX7A2, ATP5E, LDHB, UCP3, BNIP3, TRFC, S100A4 and TXNIP transcripts in two affected patients and two control individuals by quantitative real-time PCR (Applied Biosystems, Foster City, CA, USA). One microgram of total RNA, from the same RNA preparation that was analyzed in the microarray, was reverse-transcribed using the Reverse Transcription Reagents Kit (Invitrogen, Carlsbad, CA, USA), according to the manufacturer’s instructions, using random primers in a 20 µl reaction. Subsequent PCR reactions were performed in duplicates, and 2 µl of the cDNA reaction was used for each assay. The quantitative real-time PCR analysis was performed with the ABI Prism 7500 Sequence Detection System (Applied Biosystems) by using TaqMan Universal PCR Master Mix and Assays-on- Demand Gene Expression probes (Applied Biosystems) (COX7A2, assay ID: Hs01652418_m1; ATP5E assay ID: Hs00953807_g1; LDHB, assay ID: Hs00929956_m1; UCP3, assay ID: Hs01106050_g1; BNIP, assay ID: Hs00969291_m1; TRFC, assay ID: Hs00174609_m1; S100A4, assay ID: Hs00243202_m1; TXNIP, assay ID: Hs01006899_g1). Human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Hs99999905_m1) was used as an endogenous control. Transcripts expression levels and fold changes between HIBM-affected and control samples were normalized to GAPDH in each RNA sample.

GEO data sets analysis

In order to identify commonly differentially expressed genes in various muscle diseases and to assess HIBM transcriptome specificity, mRNA expression profile data sets of affected and normal muscle for each of the following diseases were compiled and analyzed in the same manner as the current experimental samples (intersection between TNoM, Student’s t-test and Info). The following data sets were analyzed in this process: GSE1007, Duchenne muscular dystrophy (DMD); GSE3307, Becker muscular dystrophy (BMD), limb girdle muscular dystrophy type2A (LGMD2A), type 2B (LGMD2B), type 2I (LGMD2I) and Fascioscapulohumeral dystrophy (FSHD); GSE1551, dermatomyositis (DM) and inclusion body myopathy (IBM), and nemaline myopathy (NM) (22). Among the significantly differentially expressed genes obtained by this analysis for each condition, transcripts overlapping with the HIBM signature were compiled.

Mitochondrial morphology analysis

Primary myoblasts cultures, derived from affected and control muscle specimens, were loaded with the mitochondria-specific fluorescent cation rhodamine 123 and visualized by video-rate confocal microscopy, as described in detail previously (36). The parameters examined were the mitochondrial aspect ratio (AR, reflecting the mitochondrial length/width ratio), the mitochondrial form-factor (F, which is a measure of the degree of mitochondrial branching) and the number of mitochondria per cell (Nc).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG Online.

FUNDING

This work was supported by a grant from the German–Israeli Foundation for Research and Development (GIF), Jerusalem, Israel, and in part by grants from the Association Francaise contre les Myopathies (AFM) and from the Neuromuscular Disease Foundation (NDF).

ACKNOWLEDGEMENTS

We are grateful to all patients who made this study possible. We thank Drs Peter Kang and Alvin Kho for PCA analysis and for useful comments and suggestions. We also thank Tanya Goltser and Dr Ronnen Segman for great technical assistance and helpful discussion, and Dr Laura Canetti for statistical analysis. We are grateful to Jasmine Jacob-Hirsch, Dr Ninette Amariglio, Dr Gidi Rechavi and Dr Naftali Kaminski from the Functional Genomics Unit at the Sheba Medical Center for excellent assistance with array samples and for array analysis. Special thanks to Professor Zohar Argov and Dr Ron Dabby for biospy material; to Zippora Shlomai for the establishment of the myoblast cultures used in this study and to Sharita Timal for assistance with the cell mitochondrial assays.

Conflict of Interest statement. None declared.

REFERENCES

1
Argov
Z.
Eisenberg
I.
Mitrani-Rosenbaum
S.
Genetics of inclusion body myopathies
Curr. Opin. Rheumatol.
 , 
1998
, vol. 
10
 (pg. 
543
-
547
)
2
Argov
Z.
Yarom
R.
‘Rimmed vacuole myopathy’ sparing the quadriceps. A unique disorder in Iranian Jews
J. Neurol. Sci.
 , 
1984
, vol. 
64
 (pg. 
33
-
43
)
3
Nonaka
I.
Noguchi
S.
Nishino
I.
Distal myopathy with rimmed vacuoles and hereditary inclusion body myopathy
Curr. Neurol. Neurosci. Rep.
 , 
2005
, vol. 
5
 (pg. 
61
-
65
)
4
Nonaka
I.
Sunohara
N.
Ishiura
S.
Satoyoshi
E.
Familial distal myopathy with rimmed vacuole and lamellar (myeloid) body formation
J. Neurol. Sci.
 , 
1981
, vol. 
51
 (pg. 
141
-
155
)
5
Eisenberg
I.
Avidan
N.
Potikha
T.
Hochner
H.
Chen
M.
Olender
T.
Barash
M.
Shemesh
M.
Sadeh
M.
Grabov-Nardini
G.
, et al.  . 
The UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase gene is mutated in recessive hereditary inclusion body myopathy
Nat. Genet.
 , 
2001
, vol. 
29
 (pg. 
83
-
87
)
6
Kayashima
T.
Matsuo
H.
Satoh
A.
Ohta
T.
Yoshiura
K.
Matsumoto
N.
Nakane
Y.
Niikawa
N.
Kishino
T.
Nonaka myopathy is caused by mutations in the UDP-N-acetylglucosamine-2-epimerase/N-acetylmannosamine kinase gene (GNE)
J. Hum. Genet.
 , 
2002
, vol. 
47
 (pg. 
7
-
9
)
7
Nishino
I.
Noguchi
S.
Murayama
K.
Driss
A.
Sugie
K.
Oya
Y.
Nagata
T.
Chida
K.
Takahashi
T.
Takusa
Y.
, et al.  . 
Distal myopathy with rimmed vacuoles is allelic to hereditary inclusion body myopathy
Neurology
 , 
2002
, vol. 
59
 (pg. 
1689
-
1693
)
8
Keppler
O.T.
Hinderlich
S.
Langner
J.
Schwartz-Albiez
R.
Reutter
W.
Pawlita
M.
UDP-GlcNAc 2-epimerase: a regulator of cell surface sialylation
Science
 , 
1999
, vol. 
284
 (pg. 
1372
-
1376
)
9
Lowe
J.B.
Glycan-dependent leukocyte adhesion and recruitment in inflammation
Curr. Opin. Cell. Biol.
 , 
2003
, vol. 
15
 (pg. 
531
-
538
)
10
Varki
A.
Angata
T.
Siglecs—the major subfamily of I-type lectins
Glycobiology
 , 
2006
, vol. 
16
 (pg. 
1R
-
27R
)
11
Effertz
K.
Hinderlich
S.
Reutter
W.
Selective loss of either the epimerase or kinase activity of UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase due to site-directed mutagenesis based on sequence alignments
J. Biol. Chem.
 , 
1999
, vol. 
274
 (pg. 
28771
-
28778
)
12
Broccolini
A.
Gliubizzi
C.
Pavoni
E.
Gidaro
T.
Morosetti
R.
Sciandra
F.
Giardina
B.
Tonali
P.
Ricci
E.
Brancaccio
A.
Mirabella
M.
Alpha-dystroglycan does not play a major pathogenic role in autosomal recessive hereditary inclusion-body myopathy
Neuromuscul. Disord.
 , 
2005
, vol. 
15
 (pg. 
177
-
184
)
13
Hinderlich
S.
Salama
I.
Eisenberg
I.
Potikha
T.
Mantey
L.R.
Yarema
K.J.
Horstkorte
R.
Argov
Z.
Sadeh
M.
Reutter
W.
Mitrani-Rosenbaum
S.
The homozygous M712T mutation of UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase results in reduced enzyme activities but not in altered overall cellular sialylation in hereditary inclusion body myopathy
FEBS Lett.
 , 
2004
, vol. 
566
 (pg. 
105
-
109
)
14
Noguchi
S.
Keira
Y.
Murayama
K.
Ogawa
M.
Fujita
M.
Kawahara
G.
Oya
Y.
Imazawa
M.
Goto
Y.
Hayashi
Y.K.
Reduction of UDP-N-acetylglucosamine 2-epimerase/N-acetylmannosamine kinase activity and sialylation in distal myopathy with rimmed vacuoles
J. Biol. Chem.
 , 
2004
, vol. 
279
 (pg. 
11402
-
11407
)
15
Penner
J.
Mantey
L.R.
Elgavish
S.
Ghaderi
D.
Cirak
S.
Berger
M.
Krause
S.
Lucka
L.
Voit
T.
Mitrani-Rosenbaum
S.
Hinderlich
S.
Influence of UDP-GlcNAc 2-epimerase/ManNAc kinase mutant proteins on hereditary inclusion body myopathy
Biochemistry
 , 
2006
, vol. 
45
 (pg. 
2968
-
2977
)
16
Salama
I.
Hinderlich
S.
Shlomai
Z.
Eisenberg
I.
Krause
S.
Yarema
K.
Argov
Z.
Lochmuller
H.
Reutter
W.
Dabby
R.
, et al.  . 
No overall hyposialylation in hereditary inclusion body myopathy myoblasts carrying the homozygous M712T GNE mutation
Biochem. Biophys. Res. Commun.
 , 
2005
, vol. 
328
 (pg. 
221
-
226
)
17
Sparks
S.E.
Ciccone
C.
Lalor
M.
Orvisky
E.
Klootwijk
R.
Savelkoul
P.J.
Dalakas
M.C.
Krasnewich
D.M.
Gahl
W.A.
Huizing
M.
Use of a cell-free system to determine UDP-N-acetylglucosamine 2-epimerase and N-acetylmannosamine kinase activities in human hereditary inclusion body myopathy
Glycobiology
 , 
2005
, vol. 
15
 (pg. 
1102
-
1110
)
18
Malicdan
M.C.
Noguchi
S.
Nonaka
I.
Hayashi
W.K.
Nishino
I.
A Gne knockout mouse expressing human GNE D176V mutation develops features similar to distal myopathy with rimmed vacuoles or hereditary inclusion body myopathy
Hum. Mol. Genet.
 , 
2007
, vol. 
16
 (pg. 
2669
-
2682
)
19
Greenberg
S.A.
Sanoudou
D.
Haslett
J.N.
Kohane
I.S.
Kunkel
L.M.
Beggs
A.H.
Amato
A.A.
Molecular profiles of inflammatory myopathies
Neurology
 , 
2002
, vol. 
59
 (pg. 
1170
-
1182
)
20
Haslett
J.N.
Sanoudou
D.
Kho
A.T.
Bennett
R.R.
Greenberg
S.A.
Kohane
I.S.
Beggs
A.H.
Kunkel
L.M.
Gene expression comparison of biopsies from Duchenne muscular dystrophy (DMD) and normal skeletal muscle
Proc. Natl Acad. Sci. USA
 , 
2002
, vol. 
99
 (pg. 
15000
-
15005
)
21
Lennon
N.J.
Kho
A.
Bacskai
B.J.
Perlmutter
S.L.
Hyman
B.T.
Brown
R.H.
Jr
Dysferlin interacts with annexins A1 and A2 and mediates sarcolemmal wound-healing
J. Biol. Chem.
 , 
2003
, vol. 
278
 (pg. 
50466
-
50473
)
22
Sanoudou
D.
Haslett
J.N.
Kho
A.T.
Guo
S.
Gazda
H.T.
Greenberg
S.A.
Lidov
H.G.
Kohane
I.S.
Kunkel
L.M.
Beggs
A.H.
Expression profiling reveals altered satellite cell numbers and glycolytic enzyme transcription in nemaline myopathy muscle
Proc. Natl Acad. Sci. USA
 , 
2003
, vol. 
100
 (pg. 
4666
-
4671
)
23
Tsukahara
T.
Tsujino
S.
Arahata
K.
CDNA microarray analysis of gene expression in fibroblasts of patients with X-linked Emery–Dreifuss muscular dystrophy
Muscle Nerve
 , 
2002
, vol. 
25
 (pg. 
898
-
901
)
24
Winokur
S.T.
Chen
Y.W.
Masny
P.S.
Martin
J.H.
Ehmsen
J.T.
Tapscott
S.J.
van der Maarel
S.M.
Hayashi
Y.
Flanigan
K.M.
Expression profiling of FSHD muscle supports a defect in specific stages of myogenic differentiation
Hum. Mol. Genet.
 , 
2003
, vol. 
12
 (pg. 
2895
-
2907
)
25
Ashburner
M.
Ball
C.A.
Blake
J.A.
Botstein
D.
Butler
H.
Cherry
J.M.
Davis
A.P.
Dolinski
K.
Dwight
S.S.
Eppig
J.T.
, et al.  . 
Gene ontology: tool for the unification of biology. The Gene Ontology Consortium
Nat. Genet.
 , 
2000
, vol. 
25
 (pg. 
25
-
29
)
26
Dennis
G.
Jr
Sherman
B.T.
Hosack
D.A.
Yang
J.
Gao
W.
Lane
H.C.
Lempicki
R.A.
DAVID: Database for Annotation, Visualization, and Integrated Discovery
Genome Biol.
 , 
2003
, vol. 
4
 pg. 
P3
 
27
Calvano
S.E.
Xiao
X.
Richards
D.R.
Felciano
R.M.
Baker
H.V.
Cho
R.J.
Chen
R.O.
Brownstein
B.H.
Cobb
J.P.
Tschoeke
S.K.
, et al.  . 
A network-based analysis of systemic inflammation in humans
Nature
 , 
2005
, vol. 
437
 (pg. 
1032
-
1037
)
28
Reddy
M.A.
Thimmalapura
P.R.
Lanting
L.
Nadler
J.L.
Fatima
S.
Natarajan
R.
The oxidized lipid and lipoxygenase product 12(S)-hydroxyeicosatetraenoic acid induces hypertrophy and fibronectin transcription in vascular smooth muscle cells via p38 MAPK and cAMP response element-binding protein activation. Mediation of angiotensin II effects
J. Biol. Chem.
 , 
2002
, vol. 
277
 (pg. 
9920
-
9928
)
29
Adams
J.W.
Sakata
Y.
Davis
M.G.
Sah
V.P.
Wang
Y.
Liggett
S.B.
Chien
K.R.
Brown
J.H.
Dorn
G.W.
Enhanced Galphaq signaling: a common pathway mediates cardiac hypertrophy and apoptotic heart failure
Proc. Natl Acad. Sci. USA
 , 
1998
, vol. 
95
 (pg. 
10140
-
10145
)
30
Yoshioka
J.
Prince
R.N.
Huang
H.
Perkins
S.B.
Cruz
F.U.
MacGillivray
C.
Lauffenburger
D.A.
Lee
R.T.
Cardiomyocyte hypertrophy and degradation of connexin43 through spatially restricted autocrine/paracrine heparin-binding EGF
Proc. Natl Acad. Sci. USA
 , 
2005
, vol. 
102
 (pg. 
10622
-
10627
)
31
Kyriakis
J.M.
Avruch
J.
Mammalian mitogen-activated protein kinase signal transduction pathways activated by stress and inflammation
Physiol. Rev.
 , 
2001
, vol. 
81
 (pg. 
807
-
869
)
32
Schneider
M.
Kostin
S.
Strom
C.C.
Aplin
M.
Lyngbaek
S.
Theilade
J.
Grigorian
M.
Andersen
C.B.
Lukanidin
E.
Lerche Hansen
J.
Sheikh
S.P.
S100A4 is upregulated in injured myocardium and promotes growth and survival of cardiac myocytes
Cardiovasc. Res.
 , 
2007
, vol. 
75
 (pg. 
40
-
50
)
33
Benard
G.
Rossignol
R.
Ultrastructure of the mitochondrion and its bearing on function and bioenergetics
Antioxid. Redox Signal.
 , 
2008
, vol. 
10
 (pg. 
1313
-
1342
)
34
Koopman
W.J.
Verkaart
S.
Visch
H.J.
van Emst-de Vries
S.
Nijtmans
L.G.
Smeitink
J.A.
Willems
P.H.
Human NADH:ubiquinone oxidoreductase deficiency: radical changes in mitochondrial morphology?
Am. J. Physiol. Cell. Physiol.
 , 
2007
, vol. 
293
 (pg. 
C22
-
C29
)
35
Koopman
W.J.H.
Verkaart
S.
Visch
H.J.
van der Westhuizen
F.H.
Murphy
M.P.
van den Heuvel
L.W.
Smeitink
J.A.
Willems
P.H.G.M.
Inhibition of complex I of the electron transport chain causes oxygen radical-mediated mitochondrial outgrowth
Am. J. Physiol. Cell. Physiol.
 , 
2005
, vol. 
288
 (pg. 
C1440
-
C1450
)
36
Koopman
W.J.H.
Visch
H.J.
Smeitink
J.A.M.
Willems
P.H.G.M.
Simultaneous, quantitative measurement and automated analysis of mitochondrial morphology, mass, potential and motility in living human skin fibroblasts
Cytometry A
 , 
2006
, vol. 
69A
 (pg. 
1
-
12
)
37
Koopman
W.J.H.
Visch
H.J.
Verkaart
S.
van den Heuvel
L.W.
Smeitink
J.A.
Willems
P.H.G.M.
Mitochondrial network complexity and pathological decrease in complex I activity are tightly correlated in isolated human complex I deficiency
Am. J. Physiol. Cell. Physiol.
 , 
2005
, vol. 
289
 (pg. 
C881
-
C890
)
38
Amsili
S.
Shlomai
Z.
Levitzki
R.
Krause
S.
Lochmuller
H.
Ben-Bassat
H.
Mitrani-Rosenbaum
S.
Characterization of hereditary inclusion body myopathy myoblasts: possible primary impairment of apoptotic events
Cell Death Differ.
 , 
2007
, vol. 
14
 (pg. 
1916
-
1924
)
39
Oldfors
A.
Moslemi
A.R.
Jonasson
L.
Ohlsson
M.
Kollberg
G.
Lindberg
C.
Mitochondrial abnormalities in inclusion-body myositis
Neurology
 , 
2006
, vol. 
66
 (pg. 
S49
-
S55
)
40
Argov
Z.
Soffer
D.
Karpati
G.
Hereditary inclusion body myopathies
Structural and Molecular Basis of Skeletal Muscle Diseases
 , 
2002
Basel
ISN Neuropathology Press
(pg. 
274
-
276
)
41
Pacher
P.
Hajnoczky
G.
Propagation of the apoptotic signal by mitochondrial waves
EMBO J.
 , 
2001
, vol. 
20
 (pg. 
4107
-
4121
)
42
Szabadkai
G.
Simoni
A.M.
Chami
M.
Wieckowski
M.R.
Youle
R.J.
Rizzuto
R.
Drp-1-dependent division of the mitochondrial network blocks intraorganellar Ca2+ waves and protects against Ca2+-mediated apoptosis
Mol. Cell.
 , 
2004
, vol. 
16
 (pg. 
59
-
68
)
43
Malisan
F.
Testi
R.
GD3 ganglioside and apoptosis
Biochim. Biophys. Acta
 , 
2002
, vol. 
1585
 (pg. 
179
-
187
)
44
Morales
A.
Colell
A.
Mari
M.
Garcia-Ruiz
C.
Fernandez-Checa
J.C.
Glycosphingolipids and mitochondria: role in apoptosis and disease
Glycoconj. J.
 , 
2004
, vol. 
20
 (pg. 
579
-
588
)
45
Wang
Z.
Sun
Z.
Li
A.V.
Yarema
K.J.
Roles for UDP-GlcNAc 2-epimerase/ManNAc 6-kinase outside of sialic acid biosynthesis: modulation of sialyltransferase and BiP expression, GM3 and GD3 biosynthesis, proliferation, and apoptosis, and ERK1/2 phosphorylation
J. Biol. Chem.
 , 
2006
, vol. 
281
 (pg. 
27016
-
27028
)
46
Rippo
M.R.
Malisan
F.
Ravagnan
L.
Tomassini
B.
Condo
I.
Costantini
P.
Susin
S.A.
Rufini
A.
Todaro
M.
Kroemer
G.
Testi
R.
GD3 ganglioside directly targets mitochondria in a bcl-2-controlled fashion
FASEB J.
 , 
2000
, vol. 
14
 (pg. 
2047
-
2054
)
47
Krause
S.
Aleo
A.
Hinderlich
S.
Merlini
L.
Tournev
I.
Walter
M.C.
Argov
Z.
Mitrani-Rosenbaum
S.
Lochmuller
H.
GNE protein expression and subcellular distribution are unaltered in HIBM
Neurology
 , 
2007
, vol. 
69
 (pg. 
655
-
659
)
48
Smeitink
J.
van den Heuvel
L.
DiMauro
S.
The genetics and pathology of oxidative phosphorylation
Nat. Rev. Genet.
 , 
2001
, vol. 
2
 (pg. 
342
-
352
)
49
Bolstad
B.M.
Irizarry
R.A.
Astrand
M.
Speed
T.P.
A comparison of normalization methods for high density oligonucleotide array data based on variance and bias
Bioinformatics
 , 
2003
, vol. 
19
 (pg. 
185
-
193
)
50
Kaminski
N.
Friedman
N.
Practical approaches to analyzing results of microarray experiments
Am. J. Respir. Cell. Mol. Biol.
 , 
2002
, vol. 
27
 (pg. 
125
-
132
)
51
Ben-Dor
A.
Bruhn
L.
Friedman
N.
Nachman
I.
Schummer
M.
Yakhini
Z.
Tissue classification with gene expression profiles
J. Comput. Biol.
 , 
2000
, vol. 
7
 (pg. 
559
-
583
)
52
Ben-Dor
A.
Friedman
N.
Yakhini
Z.
Overabundance Analysis and Class Discovery in Gene Expression Data 2002–2005
 , 
2002
School of Computer Science & Engineering, The Hebrew University of Jerusalem
53
Eisen
M.B.
Spellman
P.T.
Brown
P.O.
Botstein
D.
Cluster analysis and display of genome-wide expression patterns
Proc. Natl Acad. Sci. USA
 , 
1998
, vol. 
95
 (pg. 
14863
-
14868
)

Author notes

Present address: Howard Hughes Medical Institute, Program in Genomics, Division of Genetics, Children’s Hospital, Harvard Medical School, Boston, MA, USA.

Supplementary data