Abstract
Objective: To define the spectrum of clinical and biochemical features in 51 children with isolated complex I deficiency.
Background: Mitochondrial respiratory chain defects are one of the most commonly diagnosed inborn errors of metabolism. Until recently there have been technical problems with the diagnosis of respiratory chain complex I defects, and there is a lack of information about this underreported cause of respiratory chain dysfunction.
Methods: A retrospective review of clinical features and laboratory findings was undertaken in all diagnosed patients who had samples referred over a 22-year period.
Results: Presentations were heterogeneous, ranging from severe multisystem disease with neonatal death to isolated myopathy. Classic indicators of respiratory chain disease were not present in 16 of 42 patients in whom blood lactate levels were normal on at least one occasion, and in 23 of 37 patients in whom muscle morphology was normal or nonspecific. Ragged red fibers were present in only five patients. Tissue specificity was observed in 19 of 41 patients in whom multiple tissues were examined, thus the diagnosis may be missed if the affected tissue is not analyzed. Nine patients had only skin fibroblasts available, the diagnosis being based on enzyme assay and functional tests. Modes of inheritance include autosomal recessive (suggested in five consanguineous families), maternal (mitochondrial DNA point mutations in eight patients), and possibly X-linked (slight male predominance of 30:21). Recurrence risk was estimated as 20 to 25%.
Conclusion: Heterogeneous clinical features, tissue specificity, and absence of lactic acidosis or abnormal mitochondrial morphology in many patients have resulted in underdiagnosis of respiratory chain complex I deficiency.
Complex I (reduced nicotinamide-adenine dinucleotide [NADH]-coenzyme Q reductase, EC 1.6.5.3) is the most complicated of the five respiratory chain enzyme complexes. It consists of more than 40 polypeptides, seven of which are encoded by mitochondrial DNA (mtDNA). It is the first enzyme in the electron transport chain, accepting electrons from NADH and transferring them via a series of redox centers to ubiquinone.1
Although first recognized in 1979,2 there are relatively few published reports of respiratory chain complex I deficiency compared with the most widely reported defects of energy production—namely, respiratory chain complex IV (cytochrome c oxidase) and the pyruvate dehydrogenase complex. This is partly due to technical difficulties in assaying complex I, including the presence of considerable rotenone-insensitive “background” activity in many tissues.3 More recently, we and others4,5 have come to regard complex I deficiency as the most common energy generation disorder.
Complex I deficiency may present in a variety of ways: as a multisystem disorder, or as an isolated myopathy or liver disease. Robinson and colleagues5,6 have reported two series of complex I cases totaling 44 patients in all, with presentations including fatal infantile lactic acidosis, moderate lactic acidosis with spastic quadriplegia, Leigh disease, cardiomyopathy and cataracts, hepatopathy and tubulopathy, and mild symptoms with lactic acidemia. Isolated complex I deficiency is one of the most common causes of Leigh disease, affecting 12 of 35 Australian patients,7 and 7 of 25 patients in a British study.8
Many of the reported cases of complex I deficiency are associated with syndromes known to involve mtDNA mutations, such as mitochondrial myopathy, encephalopathy, lactic acidosis, and strokelike episodes (MELAS)9 and Leber’s hereditary optic neuropathy (LHON).10 It has also been reported in Alpers disease and Kearns–Sayre syndrome,11 neurodegenerative diseases such as PD and dystonia,10 Huntington’s disease,12 and has been implicated in aging.4 In addition, many of the reported cases show deficiency of one or more of the other respiratory chain complexes.11
The ability to measure complex I activity reliably in frozen tissues and cultured cells has enabled us to diagnose 51 cases of isolated complex I deficiency in pediatric patients. Metabolic indicators, tissue morphology, mtDNA mutations, and residual enzyme activity in different tissues varied between and within the clinical groups, and are described in this report.
Methods.
Patients.
The patients reported in this study represent all the patients we have investigated and regard as having a definite isolated complex I defect. Samples were referred to our laboratory by clinicians throughout Australia and New Zealand between 1975 and 1997. One patient was referred from the United States and one from Hong Kong. The patients were from 47 pedigrees, five of which were consanguineous. Eight families had affected siblings, and there were three two-generation pedigrees, although tissues or cultured cells were not available for study in all of these individuals. The male-to-female ratio was 30:21. Age at onset, age at death when applicable, and (when available) blood and CSF lactate levels, and morphology of skeletal and cardiac muscle, liver, and kidney are presented in the table. Skeletal muscle morphologic studies were performed at multiple centers; 13 samples were examined at the Victorian State Neuropathology Service.
Summary of features of 51 patients with isolated complex I deficiency
Lactate was raised (>2.5 mmol/L) in blood in 33 of 42 patients and in CSF in 27 of 35 patients in whom it was measured. Nine patients had only normal blood lactate levels recorded, and seven had episodes of increased lactate but normal lactate measured at other times. Three patients had normal blood lactate, but CSF lactate was elevated.
The patients were categorized by main presenting symptoms as described in the following paragraphs.
Leigh disease.
The 19 patients in this group complied with criteria presented previously,7 with two in the “Leighlike” category (Patients 16 and 17). Thirteen of these were mentioned in that report, as indicated in figure 1 . Typical findings of Leigh disease were present (i.e., progressive neurologic disease with motor and intellectual developmental delay, symptoms or signs of brainstem or basal ganglia disease, and typical neuroimaging or postmortem findings). Blood lactate level was elevated in 9 of the 13 patients in whom it was recorded. In those patients in whom lactate was measured on more than one occasion, it was variably elevated. In three patients it was normal on some days but was elevated on others (Patients 1, 2, and 16). CSF lactate was elevated in 13 of the 17 patients in whom measurements were recorded. Siblings were affected in four families (Patients 7, 13a and b, 14a and b, and 16), and presentations were similar within each family.
Figure 1. Residual complex I activity in patient tissues and cell lines expressed as percentages of control means of ratio of complex I activity to citrate synthase (CS) activity. The numbers in brackets were assigned to patients in a previous report.7 
The complex I-to-complex II ratio was less than 25%. Control means and observed ranges for complex I-to-CS ratios are the following: skeletal muscle homogenates, 281 (100 to 470), n = 16; isolated skeletal muscle mitochondria, 524 (270 to 890) n = 16; liver homogenates, 284 (200 to 340) n = 6; liver homogenates with CS corrected for nonspecific thiolase activity, 416 (270 to 570), n = 6; heart homogenates, 449 (248 to 572) n = 7; lymphoblast mitochondria (LB), 353 (275 to 410) n = 6; and fibroblast mitochondria (FB), 358 (180 to 519), n = 35. LIMD = lethal infantile mitochondrial disease; MELAS = mitochondrial myopathy, encephalopathy, lactic acidosis, and strokelike episodes; FTT = failure to thrive; LA = lactic acidosis.
Developmental regression.
Five patients had a history of developmental regression and neurologic abnormalities such as spasticity and ataxia without optic atrophy, episodes of hyperventilation, or brain imaging suggestive of Leigh disease. Three had both blood and CSF lactate levels measured: one (Patient 18) had elevated levels in both fluids, one (Patient 21) had normal blood but slightly elevated CSF lactate levels, and one (Patient 19) had normal CSF and blood lactate levels. A fourth patient (Patient 20) had intermittently raised blood lactate, and the fifth (Patient 22) had no lactate measurements.
Lethal infantile mitochondrial disease (LIMD).
In the 11 infants in this group, death occurred in the first year after birth. All had a profound illness, with lactic acidemia documented in nine patients. Lethargy, hypotonia, and seizures were the main features. Lactate was not measured in one patient (Patient 25), whereas blood and CSF lactate levels were normal in one in whom liver involvement was prominent (Patient 30). Death usually occurred shortly after presentation. Two of the early-presenting children had hypoglycemia in the neonatal period, attributed to hyperinsulinism in one (Patient 24b) and to growth retardation in a 31-week gestation neonate (Patient 26b). Twin boys were affected in one family (Patients 26a and 26b), and sisters in another (Patients 24a and 24b). A detailed clinical description of Patient 28 has been reported elsewhere.13
Cardiomyopathy.
Six patients presented with cardiomyopathy. This number excludes those who subsequently developed cardiomyopathy as part of multisystem involvement. The two patients with dilated cardiomyopathy presented with symptoms at 6 months (Patient 32) and 9.5 months (Patient 37), and died at 7 months and 14 months respectively. Neither was dysmorphic; cardiomyopathy was the only clinical feature of their disease. One patient (Patient 37) had a similarly affected sibling, although no biochemical investigations were performed on that child. Two of the four with hypertrophic cardiomyopathy (Patients 33 and 35) presented within the first few days of life with cardiac failure. Both infants were dysmorphic, and corneal opacities were noted in one (Patient 35). Both had consistently raised blood lactate levels, however it was not possible to know the degree of cardiovascular compromise, and therefore perfusion, when the lactate levels were recorded. One infant (Patient 33) had EKG changes consistent with Wolff–Parkinson–White syndrome. These two babies died within the first week after birth. In a third child (Patient 34), hypertrophic cardiomyopathy was noted at 14 months, but failure to thrive and delayed motor development were present from 8 months. This child also had EKG changes of Wolff–Parkinson–White syndrome. The fourth child (Patient 36) presented with cardiac failure, generalized hypotonia, and muscle wasting.
Other.
The remaining 10 patients presented with various symptoms. Two presented with isolated myopathy and lactic acidosis; one at 13 months with neck extensor weakness that developed into generalized myopathy. This patient (Patient 39) showed a response to riboflavin therapy.14 The other patient (Patient 38) presented with easy tiring on exertion at 3 years, with striking weakness of the flexor muscles of the neck and milder proximal weakness of shoulder and lower limb girdle muscles. His condition is slowly progressive. Three patients presented with MELAS syndrome, all with elevated lactate levels in CSF and blood, and one intermittently in blood (Patient 41). One patient (Patient 43) had significant renal tubular dysfunction, with elevated blood urea, a reduction in glomerular filtration rate to 50%, and elevated lactate in blood and CSF. Another patient (Patient 47) had normal development but collapsed suddenly at 9 years and died several days later with no sign of cardiomyopathy or lactic acidosis. Another patient (Patient 46) presented with severe lactic acidosis in infancy then failed to thrive, with persistent lactic acidosis, and died at 2 years 4 months. Two patients had symptoms and signs of Alpers disease with liver and neurologic involvement: one (Patient 44) with elevated blood lactate, and the other (Patient 45) with intermittently raised blood lactate and normal CSF lactate.
Recurrence risk.
Information about siblings was available in 40 of the 47 families in the study. This enabled us to calculate an empirical recurrence risk for complex I deficiency. There were 67 siblings, excluding seven half-siblings (six maternal, one paternal). There were 12 affected individuals, ascertained by proven complex I deficiency or near-identical clinical presentation and course as the affected sibling, and four patients who were possibly affected. This included two unexplained deaths in childhood, one child with learning difficulties, and one child with bilateral sensorineural deafness. This gives an empirical recurrence risk of 18% for the definitely affected siblings, and 24% when the four possibly affected siblings are included. Excluding known consanguineous (presumably autosomal recessive inheritance) and maternally inherited pedigrees (those with known mtDNA mutations), the recurrence risk is 17% for definitely affected siblings and 21% when the possibly affected siblings are included.
Using all siblings in calculations of recurrence risk can lead to overestimation of incidence due to unascertained siblingships, and it is often recommended that only siblings born after the proband be included.15 There were 22 such siblings: five definitely affected and one possibly affected. If only these data are used, the recurrence risk is 23% for the definitely affected siblings and 27% when the possibly affected sibling is included. Given the limitations of sample size, lack of complete information in some families, and the known genetic heterogeneity, we regard it as reasonable to quote a value of approximately 20 to 25% as an empirical recurrence risk figure for complex I deficiency.
Materials.
Coenzyme Q1 (CoQ1) was a gift of Eisai Chemical Company (Tokyo, Japan). Decylbenzylquinone, horse heart cytochrome c (III), rotenone, and antimycin A were provided by Sigma Chemical Corporation (St. Louis, MO), and 5,5′-dithio-bis (2-nitrobenzoic acid) was provided by Calbiochem (San Diego, CA). Fatty acid free bovine serum albumin and all other substrates were from Boehringer-Mannheim (Mannheim, Germany). All other chemicals were of analytical reagent (AR) grade. The ThermoSequenase cycle sequencing kit was from Amersham Life Science (Little Chalfont, Buckinghamshire, UK); AmpliTaq polymerase and GeneAmp buffer 1 from Perkin-Elmer (Foster City, CA); and deoxynucleotide triphosphates (dNTPs), restriction enzymes, and appropriate buffers were from Boehringer-Mannheim.
Tissues.
Skeletal muscle and liver samples were obtained either from open or needle biopsies, or taken at various times postmortem. Heart muscle was obtained at surgery or postmortem. Tissues were stored at −70 °C. Fibroblast cell lines were grown from forearm skin biopsies. Lymphoblasts were transformed using Epstein-Barr virus.
Methods.
Enzyme and functional studies.
Respiratory chain complexes I, II, II+III, III, IV, and the mitochondrial marker enzyme citrate synthase (CS) were assayed in skeletal muscle, liver, and heart homogenates (post-600-g supernatants) and in isolated mitochondria from fibroblasts, transformed lymphoblasts, and skeletal muscle as described,7 except that in more recent assays, CS in liver was assayed with and without 0.1 mM oxaloacetate to correct for nonspecific thiolase activity, which is significant in liver homogenates (approximately 35% of total activity) but negligible in other samples. Patient and control tissue samples were stored at −70 °C before analysis. The same human control skeletal muscle and liver samples were thawed, homogenized, and assayed with each batch of two or three patient samples. Repeated measurements in these control tissues over a 4 to 5-year period show that there is no significant decline in activity of complex I in tissues stored in this way (data not shown). Normal cultured fibroblast and transformed lymphoblast cell lines were analyzed with each batch of patient cell lines. Rates of adenosine triphosphate (ATP) synthesis in digitonin-permeabilized fibroblasts using complex I- and complex II-linked substrates were determined,16 and growth of some cell lines was tested in medium containing 25 μM NaN3 and in which glucose was replaced by galactose (galactose/azide medium).17
mtDNA mutation analysis.
mtDNA analysis was performed as described previously7 for the detection of rearrangements by Southern blotting, and the following point mutations: A3243G (MELAS), A8344G (myoclonic epilepsy with ragged red fibers [MERRF]), T8993G (neurogenic muscle weakness, ataxia and retinitis pigmentosa [NARP]), and T8993C. Three mutations associated with decreased complex I activity and LHON, or dystonia or other neurologic features,10,18 were also tested for: G3460A (as described),19 T4160C, and G14459A. The common G11778A LHON mutation was not studied because it does not cause a substantial complex I enzyme defect.10 For the T4160C mutation, the forward primer was nucleotide (nt)4137 to nt4159 (5′ to 3′) with a mismatched G at nt4159, creating an AluI restriction site in the normal sequence, but not when the mutation is present.20 The reverse primer was nt4301 to nt4281 (5′ to 3′). For the G14459A mutation, the forward primer was nt14430 to nt14458 (5′ to 3′) with a mismatched G at nt14456, creating a MaeIII site when the transition is present, as described.21 The reverse primer was nt14594 to nt14563 (5′ to 3′) with a mismatched C at nt14587, creating a MaeIII site in all patients, enabling the completeness of MaeIII digestion to be monitored. The PCR product is 164 base pair (bp), and digestion of this results in products of 139 bp and 25 bp in individuals with the normal sequence, and in 110 bp, 29 bp, and 25 bp in patients with the mutation.
Sequencing of transfer RNA (tRNA) leucineUUR and parts of the ND1 and ND6 genes was performed using the ThermoSequenase cycle sequencing kit according to the manufacturer’s instructions.
Results.
Diagnostic criteria.
The criteria used to establish the diagnosis of complex I deficiency were as follows:
-
1. %Complex I activity relative to CS or complex II was <25% of the control mean in one or more tissues.
-
2. %Complex I ratios (as described earlier) were 25 to 40% plus one of:
-
a. Similar activity in multiple tissues, cell lines, or siblings
b. Supportive histology or electron microscopy (EM; i.e., presence of abnormal mitochondria in at least one tissue)
c. Supportive functional tests in fibroblasts; in other words, low ratio of ATP synthesis with glutamate and malate (complex I-linked substrate) compared with the rate with succinate and rotenone (complex II-linked substrate), or inability to grow in galactose/azide medium
Complex I deficiency was considered isolated if other respiratory chain complexes had CS ratios that were not clearly deficient (i.e., >25%) and had residual activities at least twofold higher than complex I.
Complex I activities.
Assay of complex I has become more reliable with the use of CoQ analogs as substrates, and with improved methods of sample preparation, particularly the use of hypotonic lysis in the preparation of isolated mitochondria from cultured cells.22 The hypotonic lysis step minimizes the rotenone-insensitive background activity. In our hands the rotenone-sensitive activity is 54 ± 8% (mean ± SD of 50 estimations on nine different cell lines) for cultured fibroblast mitochondria, and 69 ± 9% (n = 7) for those from cultured transformed lymphoblasts. For the muscle, liver, and heart homogenates (post-600-g supernatants), sonication is used to lyse the samples. The rotenone-sensitive complex I activity is 49 ± 16% (n = 8) for skeletal muscle, 48 ± 7% (n = 6) for liver, and 83 ± 5% (n = 7) for heart homogenates. For isolated mitochondria from skeletal muscle the rotenone-sensitive activity is 78 ± 11% (n = 17) after sonication.
Respiratory chain complexes and CS were assayed in isolated skeletal muscle mitochondria in 6 patients, skeletal muscle homogenates in 32 patients (i.e., 38 skeletal muscle samples in all), liver homogenates in 22, heart homogenates in 4, isolated mitochondria from fibroblasts in 45, and from transformed lymphoblasts in 10 patients. The results of complex I enzymology are presented in figure 1. Complex I activity is expressed as a percentage of the mean of control values relative to the mitochondrial matrix enzyme CS (as a CS ratio) to account for different tissue mitochondrial contents and day-to-day variations in sample preparation.
All but five patients fulfilled diagnostic criterion 1 (i.e., complex I activity relative to CS or complex II was <25% of the control mean in one or more tissues). Patient 1 had 31 to 50% complex I relative to CS in four tissues, and abnormal mitochondria in liver. Patient 3 fibroblasts had an activity of 30% relative to both CS and complex II, and normal ATP synthesis, but were unable to grow on galactose/azide medium. Patient 8 fibroblasts had 32% complex I activity and an ATP synthesis rate of more than 1 SD below the mean (i.e., moderate support for the diagnosis), and was galactose sensitive. Patients 15 and 34 had similarly reduced complex I activity in two cell lines (27% in fibroblasts and transformed lymphoblasts for Patient 15, and 34% in fibroblasts and 37% from transformed lymphoblasts for Patient 34), and ATP synthesis more than 1 SD below the mean in fibroblasts. Patient 34 also had abnormal mitochondria in skeletal muscle. Thus Patients 3 and 8 were the only patients who did not fulfill criterion 1 or at least two parts of criterion 2 in addition to the enzyme activity.
Tissue specificity.
Tissue specificity was defined as residual complex I activity complying with the diagnostic criteria described in one or more tissues or cell lines, and >50% in one or more other tissues or cell lines, and was apparent in 19 of 41 patients in whom multiple tissues were examined (see figure 1). Seven patients expressed the defect in liver or heart but not in skeletal muscle or cultured cells, and 11 expressed the defect in one or more tissues but not in cultured cells. One patient (Patient 43) was complex I deficient in skeletal muscle and fibroblasts but not in transformed lymphoblasts.
Fibroblast ATP synthesis.
Digitonin-permeabilized fibroblasts from 45 patients were used to estimate ATP synthesis rates with glutamate and malate (a complex I-linked substrate) and succinate and rotenone (a complex II-linked substrate). The results are shown in figure 2 , in which ATP synthesis is expressed as a ratio of the rate with glutamate and malate to the rate with succinate and rotenone (ATP synthesis ratio), and is related to residual complex I activity expressed as a ratio to CS. The ATP synthesis ratio for 41 estimations on 13 normal fibroblast lines was 1.70 ± 0.38 (mean ± SD). Of the 28 patients with residual complex I activity <50% in fibroblasts, five had ATP synthesis ratios more than 3 SDs below the mean, six between 3 and 2 SDs below the mean, 11 between 2 and 1 SDs below the mean, and six between 1 SD below the mean and the mean. The one patient (Patient 26b) with activity between 41 and 50% was included in this group because he was the sibling of another patient with residual activity of 25%. Of the 17 patients with >50% residual activity in fibroblasts, only two had ATP synthesis ratios more then 1 SD below the mean (Patients 37 and 41). Linear regression analysis using SPSS for Windows (SPSS; Chicago, IL) showed a significant association between complex I activity and ATP synthesis ratios in the group as a whole (p < 0.0005, r = 0.67) and for the 28 patients with <50% residual complex I activity (p < 0.0005, r = 0.69), but not for the 17 with activities >50% (p = 0.377).
Figure 2. Ability of permeabilized fibroblasts from patients to generate adenosine triphosphate (ATP) via complex I (glutamate [Glut]/succinate [Succ]) relative to residual complex I activity (complex I/citrate synthase [CS]). (A) Data for 13 different normal fibroblast lines showing mean and observed range for both ATP synthesis and complex I activity, each determined at least three times for each cell line. (B) Data for the 45 patient fibroblast lines, with the area bounded by the normal fibroblast data superimposed. ○ = Leigh disease; ▪ = regression; ▴ = lethal infantile mitochondrial disease; ⧫ = cardiomyopathy; • = other.
Fibroblast galactose sensitivity.
The ability of fibroblasts to grow in galactose/azide medium was tested in five fibroblast cell lines in which residual complex I activity was between 22 and 32%, and in which no other tissue from the patients was available for assay of enzyme activity, nor was there supporting evidence from sibling data. ATP synthesis ratios were between 2 and 1 SDs below the mean in three patients (Patients 7, 8, and 9), and between 1 SD below the mean and the mean in the other two (Patients 3 and 36). All five cell lines were unable to maintain growth in galactose/azide medium, dying after 2 to 12 days in culture. The same cell lines grown in glucose medium were still growing healthily after 20 days in culture. Three normal cell lines tested in the same way were still healthy after 20 days in culture in both types of medium. Five of the most severely affected fibroblast cell lines, with residual complex I activities between 3 and 13% relative to CS, were also tested (in Patients 2, 23, 24b, 25, and 46). All died in galactose/azide medium between 4 and 6 days in culture.
Patients with only fibroblasts studied.
Nine patients had only cultured fibroblasts available for analysis. Two were siblings of other patients in whom the complex I defect had been demonstrated in muscle or liver (Patients 14a and 24b). Two others (Patients 13a and 13b) were a sibling pair, two of four siblings with Leigh disease. Both showed a similar moderate decrease in fibroblast complex I activity relative to CS of 28% and 31% of control means, with complex II ratios of 20% for both. One patient (Patient 36) who presented predominantly with cardiomyopathy expressed a complex I defect in fibroblasts, with 29% complex I activity relative to CS, but only 20% relative to complex II. The ATP synthesis ratio was not decreased significantly, but the cell line was unable to maintain growth in galactose/azide medium. Skeletal muscle and liver, which were not available for enzyme analysis, showed increased numbers of enlarged mitochondria with disarray of cristae on EM, thus supporting the diagnosis. For the other four patients, ATP synthesis ratios were between 1 and 2 SDs below the mean in three (Patients 7, 8, and 9; i.e., moderate support for the diagnosis from ATP synthesis rates), and were normal in the other (Patient 3). All four patients were unable to maintain growth in galactose/azide medium.
Morphology.
Histology, histochemistry, and EM can contribute to the diagnosis of complex I deficiency, particularly when the residual enzyme activity is >25%. Skeletal muscle was examined in 37 patients. Only 14 patients supported a mitochondrial abnormality, with five (Patients 2, 32, 38, 40, and 41) showing ragged red fibers (RRF) and nine others (Patients 14a, 19, 23, 24a, 34, 35, 36, 39, and 43) exhibiting abnormal mitochondria. The NADH-tetrazolium reductase (NADH-TR) stain was performed on samples from 29 patients, and was elevated in five (Patients 19, 23, 32, 38, and 43) and within normal limits in 23 patients. Only one patient (Patient 14a) showed decreased NADH-TR staining, but all other mitochondrial stains were also decreased. Patient 23 had two quadriceps samples examined—one taken perimortem, the other postmortem. The perimortem sample had subsarcolemmal accumulations of mitochondria, whereas the postmortem sample had a paucity of subsarcolemmal staining. Patient 32 also had two muscle samples examined, both taken postmortem. The changes in quadriceps were mild and nonspecific, but diaphragm showed increased lipid and RRF, much greater than the mitochondrial proliferation that can be found in diaphragm of patients with respiratory abnormalities. These two patients highlight the variability that may be present in skeletal muscle morphology in the same patient. For liver, 18 patients were examined histologically, of whom six had abnormal mitochondrial morphology (Patients 1, 12, 22, 33, 36, and 45) and 10 had nonspecific changes such as increased lipid, degeneration, atrophy, necrosis, or increased glycogen. Five hearts were examined, four in the cardiomyopathy group, three of which (Patients 32, 33, and 37) showed mild to marked endocardial fibroelastosis, one with increased mitochondrial staining and lipid, and one with lipid droplets. Another cardiomyopathy patient (Patient 35) showed subendocardial fibrosis. The patient who collapsed suddenly (Patient 47) showed mildly abnormal mitochondrial cristae on EM of heart. In Patient 43, kidney showed increased numbers of abnormal mitochondria in the renal tubules.
mtDNA mutations.
Southern blots to search for mtDNA deletions and insertions were performed on 50 patients, with no alteration detected. mtDNA from all patients was analyzed for the common point mutations causing MELAS (A3243G), MERRF (G8344A), NARP, and Leigh disease (T8993G or T8993C), and for three point mutations causing LHON with or without other neurologic symptoms, and associated with complex I deficiency (G3460A, T4160C, and G14459A). The common MELAS A3243G mutation was found in two of the patients who presented with MELAS syndrome (Patients 40 and 41). The G14459A mutation, which has been described previously in patients with LHON and dystonia,23 was found in two brothers who presented with Leigh disease (Patients 14a and 14b), in whom dystonia was not present. Part of the ND6 gene was sequenced, confirming the presence of the mutation.
Sequences of the mtDNA tRNA leucineUUR gene were examined in all patients. Nucleotide changes were detected in six patients. The presence of the A3243G mutation in two of the MELAS patients was confirmed (Patients 40 and 41), and the T3271C mutation24 was detected in the other (Patient 42). A C3303T mutation described previously as associated with cardiomyopathy and myopathy25 was detected near homoplasmically in one of the patients presenting with isolated myopathy (Patient 38), and was confirmed on restriction digestion with HpaI. In the other patient (Patient 39) with isolated myopathy, the T3250C mutation was identified, as we described previously.14 The patient presenting with renal disease (Patient 43) has a novel pathogenic mutation—G3242A (manuscript in preparation). The activities of complexes I, II, II+III, III, and IV, and the mitochondrial marker enzyme CS in skeletal muscle from these six patients are shown in figure 3 . There is a suggestion of complex IV being reduced in most, but in all patients complex I is by far the most decreased. Fibroblasts were unavailable from the patient with the T3271C mutation, but were available in the other five. Only the patient with the G3242A mutation expressed the enzyme defect in fibroblasts.
Figure 3. (A–F) Respiratory chain enzyme profiles in skeletal muscle of patients with mutations in the transfer RNA leucineUUR gene of mitochondrial DNA. Enzyme activities are expressed as percentages of control means with observed ranges represented by the vertical lines. (A, D) Data are from assays of isolated mitochondria. (B, C, E, F) Data are from assays of post-600-g supernatants. NT = not tested; CS = citrate synthase.
Discussion.
The 51 patients with isolated complex I deficiency presented here reflect a broad spectrum of clinical presentations. The main features of the group are summarized in the table. Although not all patients had skeletal muscle examined histologically, most showed no specific signs of mitochondrial abnormalities (see the table). Definite RRF, the original diagnostic hallmark of adult mitochondrial dysfunction, were found in only five patients. Subsarcolemmal mitochondrial aggregates or morphologically abnormal mitochondria were found in another nine patients. These results contrast markedly with those reported previously,9 when RRF were present in 14 of 15 patients with muscle complex I deficiency. Two factors probably contribute to the higher incidence of histologic abnormalities in that study. First, their patients had a narrower range of clinical features (13 of 15 had MELAS) and most are likely to have had tRNA mutations. Five of our six patients with tRNA mutations identified had either RRF or abnormal mitochondria (all in the “Other” group of the table) compared with only 9 of 31 without a tRNA mutation identified. Second, the median age at muscle biopsy in the earlier study was 13 years (range, 3 to 27 years) compared with 2 years (range, 0 to 19 years) in our study. In our experience it is very unusual to find RRF in children less than the age of 4 years and it is likely that other symptoms prompted investigation of most of our patients before such changes could develop.
NADH-TR staining was normal or increased in 28 patients, and decreased in only one (in that patient, other mitochondrial stains were also reduced) and is not a reliable marker of complex I deficiency. The NADH-TR stain is largely insensitive to rotenone, presumably because the bulk of the staining is due to the rotenone-insensitive enzyme NADH-cytochrome b5 reductase.3 Elevations of blood and CSF lactate were not present in some patients (see the table), and in several, blood lactate was only raised intermittently. When measurements were available in both fluids, elevated CSF, but not blood, lactate was observed in three patients. The presence of lactic acidosis and abnormal muscle morphology are criteria often used to initiate further investigation of respiratory chain function, particularly enzymology, in patients. Eleven patients (22%) had normal blood lactate on at least one occasion and unremarkable muscle morphology, or no morphologic studies performed on muscle. In two patients, muscle morphology was markedly different in two separate skeletal muscle samples. It is therefore important to remember that when considering respiratory chain disease, normal lactate and muscle morphology do not rule out the diagnosis of a complex I defect. Similar observations have been reported by others.5,26
The tissue samples analyzed had been stored at −70 °C for periods ranging from 1 day to 9 years. There can be considerable loss of activity postmortem, particularly in liver, but we have observed that muscle collected and frozen at −70 °C within 6 hours of death, and liver within 2 hours, remain suitable for respiratory chain enzyme analysis. In our experience, complex II loses activity most rapidly after death and on inappropriate storage. Storage at −20 °C results in a dramatic decrease of activity of respiratory chain complexes, with relative preservation of CS activity. The complex II activity can therefore be considered a marker of the integrity of the tissue and the reliability of respiratory chain enzymology, particularly if the tissues have been taken at longer or unknown times postmortem. Tissues from two patients (Patients 16 and 47) reported here were taken approximately 12 hours after death, but complex II activity was still >85% of normal mean activity in liver and skeletal muscle in one (Patient 16) and in heart in the other (Patient 47). In this second patient, all skeletal muscle enzymes were decreased, but the complex II-to-CS ratio was 77% of normal mean. We therefore consider that the results are valid for both these patients. With the difficulty of obtaining fresh tissue for analysis in our widespread pediatric population, these observations provide support for the validity of using frozen tissue for patient studies. Other workers27,28 have reported difficulty with the complex I assay in cultured cells, but our experience is that it can be extremely useful both in diagnosis and in confirmation of enzyme defects.
A disadvantage of frozen tissues is that they cannot be used for functional studies, however such studies can be very useful with skin fibroblasts, particularly if they are the only sample available. Overall, ATP synthesis rates strongly supported the diagnosis in 11 of the 28 fibroblast lines with complex I activity <50%, and provided moderate support in an additional 11 cell lines. In a few patients (Patients 3, 10, 18, 21, and 26b) ATP synthesis was higher than expected from the residual complex I activity, suggesting that deficient ATP synthesis may not be the primary defect in these patients and that the symptoms may be due to another mechanism, such as free radical damage.29 Our patient data suggest that fibroblast complex I activity must be decreased by approximately 50% before it has a significant impact on ATP synthesis (see figure 2). The effect on ATP synthesis of inhibiting complex I activity (i.e., its “control strength”) appears to vary between, and even within, different cell types. Only 25% inhibition of complex I activity in synaptic mitochondria was required before ATP synthesis was compromised, but in nonsynaptic mitochondria the corresponding threshold value was 72%.30
A prominent feature of the group of patients reported here is tissue specificity of complex I deficiency. Overall, 46% of patients (19 of 41) in whom multiple tissues were studied exhibited some degree of tissue specificity. This included seven patients with expression of the defect in skeletal muscle but not fibroblasts (Patients 4, 17, 19, 38, 39, 40, and 41), four of whom had mutations in the mtDNA tRNA leucineUUR gene, and seven patients with normal skeletal muscle complex I activity. Possible causes of tissue specificity include tissue-specific subunits of complex I,31 threshold effects with mtDNA heteroplasmy,32 and tissue differences in RNA processing.33
There are several possible modes of inheritance in this group of patients. In five patients, the parents of the patients were first cousins, implying autosomal recessive inheritance and therefore a defect of a nuclear encoded gene (see the table). Recently, the first nuclear gene mutation in complex I deficiency was identified in the 18-kDa (AQDQ) subunit gene. One of 20 patients studied was homozygous for a frame shift mutation inherited in an autosomal recessive manner.1 Cybrid studies are underway in our patients with fibroblast expression to determine whether the defect is nuclear or encoded mitochondrially. We consider that the number of consanguineous pedigrees in this study is low compared with our group of definite complex IV-deficient patients, in which there are nine consanguineous pedigrees among 25 families. This comparison suggests that autosomal recessive inheritance may be operative in only a minority of complex I-deficient patients, unlike complex IV deficiency, with which most patients are likely to have autosomal recessive defects.34
The possibility of X-linkage in our families is suggested by the slight male predominance. One complex I gene, the NDUFA1 gene, has been mapped to the X chromosome, but to date no mutations have been reported in this gene in patients.1 There is evidence for at least two X-linked complex I complementation groups from cell hybrid studies using mutagenized Chinese hamster fibroblasts.35
Maternal inheritance is present in the seven families with mtDNA mutations. Two children in one family had a mutation in the mtDNA ND6 gene encoding a subunit of complex I. The other six families had a mutation in the tRNA leucineUUR gene. Mutations in this gene affect complex I preferentially32 (see figure 3), perhaps by interfering with processing of the adjacent ND1 gene from the polycistronic mtDNA transcript.36 Other mutations in mtDNA complex I genes or tRNA genes could also cause complex I deficiency, but were not tested in this study.
Morphologic studies are an important component of investigating patients with suspected respiratory chain disease, but provide diagnostic results in only a minority of affected children. Given the complexities of tissue specificity, and the possible absence of lactic acidosis, it is advisable to consider assaying respiratory chain enzymes either in a clinically involved tissue or in multiple tissues or cell lines from any patient with a high suspicion of respiratory chain dysfunction. In our experience, tissues collected within 2 hours of death for liver, and within 6 hours for skeletal muscle, and stored at −70 °C are suitable for respiratory chain enzyme analysis. Expression of the defect in cultured fibroblasts means that confirmatory functional studies can be performed using ATP synthesis rates and sensitivity to galactose as a carbon source. mtDNA mutations can lead to complex I defects, but autosomal recessive and probably X-linked inheritance are other likely modes of inheritance in affected families. We estimate that these three modes of inheritance together give an empirical recurrence risk of approximately 20 to 25%.
Acknowledgments
Supported in part by an institute grant from the National Health and Medical Research Council of Australia.
Acknowledgment
The authors acknowledge the following people for referral of patients, clinical details, and morphologic studies: Drs D. Danks, I. Hopkins, L. Shield, K. Collins, D. Reddihough, G. Thompson, and C.W. Chow (Royal Children’s Hospital, Melbourne, Victoria); Drs. P. Procopis, J. Christodoulou, R. Ouvrier, K. North, B. Wilcken, M. Wilson, E. Fagan, S. Arbuckle, and A. Kan (New Children’s Hospital, Sydney, NSW); Drs. F. Collins, V. Tobias, G. Wise, G. Morgan, M. Freckman, and H. Johnston (Sydney Children’s Hospital, NSW); Dr. I. Wilkinson (John Hunter Hospital, Newcastle, NSW); Drs. D. Thomas, J. Fletcher, J. Manson, E. Robertson, E. Haan, B. Lewis, A. Bourne, and N. Poplawski (Adelaide Women’s and Children’s Hospital, SA); Dr. J. McGill (Royal Children’s Hospital, Brisbane, Qld); Drs. D. Cowley, F. Bowling, and A. Tannenberg (Mater Misericordiae Hospital, Brisbane, Qld); Drs. J. Dixon, T. Stanley, and D. Kenwright (Wellington Hospital, New Zealand); Drs. S. Sacks and F. Mayall (Waikato Hospital, Hamilton, New Zealand); Drs. D. Jamison and J. Allen (Starship Children’s Hospital, Auckland, New Zealand); Dr. R. Matalon (Miami Children’s Hospital, FL); Dr. M. Tang (Tsan Yuk Hospital, Hong Kong); and Dr. Kin Shing Lun (Queen Elizabeth Hospital, Hong Kong). The authors also thank Stefania Toombs, Tiffany Symes, Tamara Gough, Alison Blake, Ivan Biros, Tom Milovac, Wendy Hutchison, and Sarah White for technical assistance; Veronica Collins for the linear regression analysis; and Dr. Stephen Kahler for helpful discussion.
- Received August 26, 1998.
- Accepted December 24, 1998.
References
Disputes & Debates: Rapid online correspondence
REQUIREMENTS
If you are uploading a letter concerning an article:
You must have updated your disclosures within six months: http://submit.neurology.org
Your co-authors must send a completed Publishing Agreement Form to Neurology Staff (not necessary for the lead/corresponding author as the form below will suffice) before you upload your comment.
If you are responding to a comment that was written about an article you originally authored:
You (and co-authors) do not need to fill out forms or check disclosures as author forms are still valid
and apply to letter.
Submission specifications:
- Submissions must be < 200 words with < 5 references. Reference 1 must be the article on which you are commenting.
- Submissions should not have more than 5 authors. (Exception: original author replies can include all original authors of the article)
- Submit only on articles published within 6 months of issue date.
- Do not be redundant. Read any comments already posted on the article prior to submission.
- Submitted comments are subject to editing and editor review prior to posting.