Article Text
Abstract
Background The natural history of all known patients with French-Canadian Leigh disease (Saguenay-Lac-St-Jean cytochrome c oxidase deficiency, MIM220111, SLSJ-COX), the largest known cohort of patients with a genetically homogeneous, nuclear encoded congenital lactic acidosis, was studied.
Results 55 of 56 patients were homozygous for the A354V mutation in LRPPRC. One was a genetic compound (A354V/C1277Xdel8). Clinical features included developmental delay, failure to thrive, characteristic facial appearance and, in 90% of patients, acute crises that have not previously been detailed, either metabolic (fulminant lactic acidosis) and/or neurological (Leigh syndrome and/or stroke-like episodes). Survival ranged from 5 days to >30 years. 46/56 patients (82%) died, at a median age of 1.6 years. Of 73 crises, 38 (52%) were fatal. The immediate causes of death were multiple organ failure and/or Leigh disease. Major predictors of mortality during crises (p<0.005) were hyperglycaemia, hepatic cytolysis, and altered consciousness at admission. Compared to a group of SURF1-deficient Leigh syndrome patients assembled from the literature, SLSJ-COX is distinct by the occurrence of metabolic crises, leading to earlier and higher mortality (p=0.001).
Conclusion SLSJ-COX is clinically distinct, with acute fatal acidotic crises on a backdrop of chronic moderate developmental delay and hyperlactataemia. Leigh syndrome is common. Stroke-like episodes can occur. The Leigh syndrome of SLSJ-COX differs from that of SURF1-related COX deficiency. SLSJ-COX has a different spectrum of associated abnormalities, acidotic crises being particularly suggestive of LRPPRC related Leigh syndrome. Even among A354V homozygotes, pronounced differences in survival and severity occur, showing that other genetic and/or environmental factors can influence outcome.
- Metabolic diseases
- mitochondrial disorders
- COX deficiency
- lactic acidosis
- Leigh syndrome
- metabolic disorders
- clinical genetics
- neurology
Statistics from Altmetric.com
- Metabolic diseases
- mitochondrial disorders
- COX deficiency
- lactic acidosis
- Leigh syndrome
- metabolic disorders
- clinical genetics
- neurology
Introduction
Leigh syndrome (LS) is a classic sign of mitochondrial disease, characterised by recurrent and often fatal episodes of neurological deterioration involving brain stem and basal ganglia. Typical lesions include necrotic regions and cystic changes, pronounced capillary proliferation in the adjacent grey matter, and relative preservation of neurons.1 One of the best known causes is deficiency of cytochrome c oxidase (COX, respiratory chain complex IV). COX deficiency due to inactivating mutations in SURF1 are the best documented autosomal cause of LS worldwide.2 3 Mutations in other genes such as COX15 or COX10 can also cause COX deficiency and LS, and deficiencies of respiratory chain complexes I or V and of pyruvate dehydrogenase can also cause LS.4
LRPPCR encodes leucine-rich pentatricopeptide repeat-containing protein that may regulate the stability and handling of mature mitochondrial mRNA.5 6 To date, LRPPRC mutations have only been reported in patients from a genetic isolate in northern Québec, Canada. They cause Saguenay-Lac Saint-Jean (SLSJ) COX deficiency, also called LS, French-Canadian type (MIM 220111). In SLSJ-COX, COX activity is notably deficient in liver and brain, and to a lesser extent in muscle, fibroblast, and kidney.7 This distribution of COX deficiency is similar to the normal expression pattern of wild-type LRPPRC, which is low in liver and brain. Due to a founder effect, the carrier rate of SLSJ-COX is elevated in SLSJ (1/23) and ∼1/2000 births are affected.8 9 Two LRPPRC mutations are reported to cause SLSJ-COX: A354V, which accounted for all but one mutant allele in the original series, and C1277Xdel8. Heterozygote screening for A354V is now routinely offered to couples with Saguenay-Lac Saint-Jean ancestry.
This retrospective study reports all known SLSJ-COX deficient patients. Because most are homozygous for a single mutation, A354V, it is possible to study the variability in clinical course that is unrelated to the primary mutation. By comparing these data with that from reported SURF1 deficient LS patients, we identify clinical features that may be useful in suspecting LRPPRC deficiency in patients from Québec and elsewhere.
Patients and methods
Subjects
We included all patients known to the Québec physicians who follow paediatric inborn errors. We also included patients who died before identification of the LRPPRC gene if (1) stored biological samples revealed two mutant LRPPRC alleles, and/or if (2) they died in infancy of no other proven cause, if the ages of birth and death were known, and if both parents were heterozygous for an LRPPRC mutation. Clinical histories were assembled from medical records at the three genetics centres in Québec in which SLSJ-COX patients had been identified before 2007. All available clinical, laboratory, imaging and pathological records were reviewed. The completeness of data varied greatly among patients.
Metabolic crises were defined as acute clinical deterioration and metabolic acidosis with plasma bicarbonate <12 mmol/l. Neurological crises were defined as acute neurological distress without evident cause. This included LS and Leigh-like syndromes fulfilling Rahman's criteria4 and stroke-like episodes with cerebral lesions with a non-vascular distribution.
Molecular studies
DNA was purified from peripheral blood leucocytes. LRPPRC mutations were identified using PCR amplification of two LRPPRC segments, each including one of the known mutations (c.C1061T [A354V] in exon 9 and c.3830_3837del + T3839G [C1277Xdel8] in exon 35), followed by allele specific oligonucleotide hybridisation. PCR primers for exon 9 were 5′TTTGCAATTCTTATTTTGTCTTCA3′ (LRPPRC-9S) and 5′CAACTGGTCTAATCTTATAAATGTTCC3′ (LRPPRC-9AS) (337 bp amplicon). PCR primers for exon 35 were 5′TTAATTGGAAATATTCAGTTTGATCC3′ (LRPPRC-35S) and 5′TAAATAAGGTCCTGAAACAACAGG3′ (LRPPRC-35AS) (250 bp amplicon). Oligonucleotide probes were: for exon 9, wild type, 5′GGAAGATGTAGCGTTGCA3′ (ASO-COX-9N) and mutant, 5′GGAAGATGTAGTGTTGCA3′ (ASO-COX-9M), and for exon 35, wild type, 5′GATGTGGTGCAATTGCTG3′ (ASO-COX-35N) and mutant, 5′GATAGTGCTGAACAAACC3′ (ASO-COX-35M).
Patient consent
Molecular studies were performed with informed consent of patients or their legal guardians.
Statistical analysis
Median values for clinical and biochemical parameters were compared using a Mann–Whitney test. Proportions were compared with the χ2 test. We used Kaplan–Meier survival analysis and the Cox's proportional hazard ratio model to compare survival rates for the SLSJ-COX cohort and reported SURF1 COX deficient patients. Mortality rates were compared using the log rank test. Multiple logistic regression was used to identify prognostic indicators of crisis mortality. To take into account clustering of data (many patients experienced several crises), regression was performed using generalised estimating equations methods.10 Models were adjusted for sex and age at crisis. Statistical analyses were performed using the SAS statistical software (version 9.1, SAS Institute, Inc).
Literature search
We performed a Medline search using ‘SURF1’ and ‘Leigh’ as key words. We reviewed all clinical reports identified. In addition to the articles of which the authors were previously aware, reports of four other SURF1 deficient patients were added.
Results
Patient identification and genotypes
The study group of 56 patients (26 males and 30 females) represents all documented SLSJ-COX patients born before 2007 (1957 and 2006). Genotype was established from patient DNA in 29 cases and was deduced for 27 others by clinical suspicion plus heterozygosity for A354V in both parents. Six patients were identified presymptomatically because of a positive family history. Fifty-five patients were A354V homozygotes. One was a genetic compound (A354V/C1227Xdel8). Spectrophotometric COX assays were available from at least one tissue of 26 patients and were consistent with the tissue specific pattern previously described.7 Clinical information other than dates of birth or death was available for 51 patients.
Chronic clinical course
Reason for initial consultation
The median age for the first detected manifestation of the disease was 5 months (range 0–24 months). Patients were referred for neonatal distress (10), psychomotor delay (23), failure to thrive (5), ataxia (2), and acute metabolic acidosis (11).
Neonatal period
Information was available for 36 patients. Pregnancy histories were unremarkable. Of 34 children born at term, birth weight was 3140±652 g. Hypotonia was noted in 21 patients (58%); transient tachypnoea of the newborn in 17 (47%); poor suckling, 16 (44%); tremulations, 10 (28%); and hypoglycaemia, six (17%). Blood lactic acid concentration measured in six asymptomatic affected neonates born to at-risk families ranged from 3.5–5.5 mmol/l. Six patients experienced a neonatal metabolic crisis, with profound metabolic acidosis, initial HCO3 ranging from 3.5–11.8 mmol/l and blood lactic acid concentrations between 12–26 mmol/l (normal <2.2 mmol/l). One patient died from fulminant lactic acidosis and multiple organ failure at age 5 days.
Morphotype
Mild but distinct craniofacial features were present in many patients. Typical findings included prominent forehead, midfacial hypoplasia, broad midline, mild hirsutism, and a characteristic arched form of the eyebrows (figure 1).
Developmental and physical examination
Global developmental delay was documented in all patients older than 12 months (supplementary figure 1). Language delay was particularly pronounced. Median age at walking was 19 months. Neurological examination revealed hypotonia, weakness, and paucity of facial expression. Older ambulatory patients had truncal ataxia with wide based gait and mild intention tremor. The prevalence of clinical features in SLSJ-COX patients is presented in table 1. One adolescent girl had hypergonadotrophic hypogonadism. None of the following were documented: peripheral neuropathy, cardiomyopathy, other endocrine abnormalities, renal tubulopathy or chronic liver or kidney disease.
Biological assessment and acid–base balance in stable patients
Patients who were not in acute crisis typically had respiratory alkalosis and mildly elevated blood lactic acid. Means and standard deviations of individual median values were: pH 7.42±0.03; Pco2 27.7±3.1 mm Hg; bicarbonate 18.9±2.5 mmol/l; and lactate 3.8±2.2 mmol/l. Blood lactate concentration was sometimes normal: 13/30 (43%) patients had at least one sample ≤2.2 mmol/l. Seven patients had lumbar puncture revealing increased lactate concentration in the cerebrospinal fluid (CSF) in all cases, ranging from 3.6–6.2 mmol/l. The activities of alanine and aspartate aminotransferases were normal, as were plasma ammonia concentrations. No characteristic urinary organic acid profile was observed except mild lactaturia.
Acute crises
Forty-five patients (45/50, 90%) experienced one or more acute decompensations requiring intensive care. Five patients (10%), all of whom have survived, never had a crisis (median age 8.3 years). For eight patients who died in infancy, we have insufficient information about the circumstances of death. Based on the rest of the cohort, acute metabolic or neurological crises are the most probable cause of death. Table 2 shows the outcomes of the first and subsequent crises. In the first crisis, 22/45 patients died (49%). Of clinical note, most survivors had a second crisis in the following weeks, which is a particularly vulnerable period for patients. Overall, the minimal estimate of mortality from crises is 38/56 (68%). If the eight unexplained infant deaths are attributed to crises, the overall death rate from crises is 46/56 (82%).
Of 73 crises, 53 (73%) were of the acute metabolic type, and 12 (16%) were neurological crises (without severe acidosis) including LS (n=8) and stroke-like episodes (n=4). Eight (11%) were mixed metabolic–neurologic crises. Crises were frequently associated with infectious illness: among 73 crises, prodromal symptoms included upper respiratory tract infection (21/73), gastrointestinal infection (9/73), lower respiratory tract infection (4/73) or febrile illness without apparent cause (7/73). In 6/73, crises were abrupt, without prodroma. Typically, metabolic crises had an acute onset, with signs of acidosis developing over 1 h or less. Patients presented with Kussmaul respiration, poor skin perfusion and sometimes dehydration, profound lactic acidosis and frequently, associated hyperglycaemia. Five patients presented seizures. Many had severe hypotonia and stupor, rapidly progressing to coma. At this point, patients either stabilised, or rapidly developed shock, acute respiratory distress syndrome with parenchymal lung opacities, peripheral oedema, multiple organ failure and death, despite aggressive support including mechanical ventilation, continuous intravenous bicarbonate infusion, and inotropic support. The course of the acid–base status of patients during the first 3 days of fatal versus non-fatal metabolic crises is shown in supplementary figure 2. Fatal crises were characterised by lower pH, lower HCO3 and higher Pco2. Urinary organic acids profile showed lactaturia and, in some instances, non-specific organic aciduria including ketone bodies and Krebs cycle intermediates. In the terminal phase of multiple organ failure, liver cytolysis occurred with an increase of creatine kinase (rarely amylase and lipase), and rarely with mild hyperammonaemia (<200 μmol/l). In fatal acute metabolic crises, the median time from initial presentation to death was 1 day (range 0–7 days). Some patients showing an initial metabolic improvement had a biphasic course, with secondary deterioration. Most commonly, these patients progressively developed a mixed metabolic–neurologic crisis. Plasma lactate values were usually near normal during this phase. Peripheral oedema was common in patients surviving beyond the first days of the crisis.
Neurological crises manifested as acute or sub-acute deterioration in the absence of severe metabolic or lactic acidosis. In LS, patients were severely hypotonic with ataxia, stupor or coma, presented periods of tachypnoea alternating with hypoventilation and apnoea, and other signs of brainstem dysfunction (ophthalmoplegia, swallowing disturbances, and labile arterial hypertension). Nine patients presented seizures. Electroencephalograms showed multiple diffuse high voltage spike discharges, followed by generalised slow waves. This progressive neurological deterioration led to death with a median time between initial presentation and death of 34 days (range 6–66 days). Most patients died in a vegetative state, usually with major loss of respiratory control and total dependency on mechanical ventilation. Brain MRI examination showed T2 weighted hyperintensities in the brainstem, most commonly in the mesencephalic periaqueductal grey matter and the pontine and medullary tegmenta (figure 2). Occasionally, lesions were also observed in cerebellar white matter, thalami, lenticular, caudate and subthalamic nuclei. Proton MR spectroscopy showed lactate accumulation (figure 2). Four patients aged 4, 10, 13, and 13 years experienced neurological crises with cerebral imaging suggestive of stroke-like episodes (wide hemispheric lesions in a non-vascular distribution).11 Clinically, they presented with lethargy (3/4), weakness (3/4), myoclonus (1/4), hemiparesis (1/4), seizures (3/4), and ophthalmoplegia (1/4). The three oldest patients recovered over a few weeks, but the 4-year-old progressively deteriorated and died after 63 days.
Outcome
Global mortality rate in SLSJ-COX deficiency was 46/56 patients (82%), with a median age at death of 1.6 years. A Kaplan–Meier curve illustrates survival of the cohort (figure 3). Looking for prognostic indicators at the start of the crises, using the first biological assessment of the patients at admission and before any parenteral infusion, we identified hyperglycaemia as a major predictor of mortality: patients with blood glucose above 10 mmol/l had a 30-fold increase in mortality risk (95% CI 7.5 to 126.8) compared to non-hyperglycaemic patients. Impairment of consciousness and high plasma liver aminotransferase values were also associated with an increased risk of mortality (table 3).
Ten patients were living at the time of the study, with a median age of 14.5 years (range 0.5–25 years). After puberty, patients have had a stable clinical course, with mild mental retardation (IQ ranging between 64–78) and seizures (3/6 patients). No metabolic crisis has been recorded after the age of 7 years and no neurological crisis after 13 years. Available neuropathological samples from patients who died with clinical evidence of LS were often too limited to diagnose LS formally. However, lesions were consistent with LS. In contrast, patients who died in fulminant lactic acidosis showed non-specific ischaemic encephalopathy with brain oedema but without characteristic lesions of LS.
Comparison of SLSJ-COX with SURF1 deficient LS
Clinical information was extracted from reports of 39 patients with SURF1 deficiency published between 1999 and 2007. The main clinical features of these patients, their outcome and age at last follow-up are presented in supplementary table 1. Median age at onset was 10 months (range 0–48 months) for SURF1 patients versus 5 months (0–24 months) for SLSJ-COX patients (p=0.01). Precise information about psychomotor development was unavailable for most SURF1 patients. When reported, case reports suggest that most if not all SURF1 patients have psychomotor delay. Seven of 39 SURF1 patients (18%) had a distinctive facial appearance, including frontal bossing, brachycephaly, hypertrichosis, maxillary hypoplasia, and telecanthus,12 similar to those in SLSJ-COX (table 1). Sixteen of 39 (41%) had hirsutism or hypertrichosis (table e-1) versus 21/39 (54%) in SLSJ-COX patients (p=0.4). Clinical features of SURF1 patients not reported in SLSJ-COX deficiency include: peripheral neuropathy, 11 patients (28%); cardiomyopathy, two patients (5%); and renal tubulopathy in four patients (10%). Conversely, no acute severe episode of lactic acidosis was reported in SURF1 disease and only one patient was described as having ‘several episodes of metabolic acidosis with mottling and cold extremities’.13
The median age at death was 72 months for SURF1 deficient patients and 20 months in SLSJ-COX patients (figure 3). The hazard ratio for death in SLSJ-COX compared to SURF1 related COX deficiency is 2.29 (95% CI 1.34 to 3.92, p<0.005).
Discussion
SLSJ-COX patients form the largest known cohort of genetically homogeneous autosomal recessive LS patients. Analysis of the variability of this group provides a unique opportunity to assess the extent to which environmental factors plus genetic factors other than the causal gene can influence outcome. Comparison of the typical presentation of SLSJ-COX patients with that of other genetically defined LS or COX deficiencies may reveal genotype–phenotype correlations useful for clinical diagnosis.
In clinically stable SLSJ-COX patients there is much similarity among the morphological, developmental, neurological, and biochemical findings (table 1, supplementary figure 1). Conversely, there are pronounced differences of severity among patients.
The most striking features of SLSJ-COX are the acute metabolic or neurologic crises. Acute neurological changes are characteristic of LS but acute metabolic crises are not typically described in the LS literature. The liver is a key organ in lactate homeostasis, responsible for recycling lactate to glucose via the Cori cycle. In SLSJ-COX deficiency, the COX activity is more severely reduced in liver and brain than in other tissues.7 The severe hyperlactacidaemia in metabolic crises may be related to metabolic liver dysfunction. We speculate that, in such circumstances, the liver changes from the main consumer of lactate to a major site of lactate production.
In SLSJ-COX, the occurrence of crises is the main determinant of mortality. All five SLSJ-COX patients who did not develop a crisis have survived. Conversely, of the 45 patients who developed one or more crises, only five survived (11%). Two anecdotal but intriguing observations deserve comment. First, in the neonatal period, only one of five metabolic crises was fatal. Although the numbers of patients are small, we speculate that tolerance to crises may be greater in the neonatal period. Furthermore, no metabolic crises occurred after the age of 7 years, showing a clear vulnerability period in infancy and early childhood. Second, the only patient to survive more than four crises was the genetic compound who has a C1277Xdel8 mutation on one allele. C1277Xdel8 is predicted to be a null allele since it creates a frame shift. Clearly, mutations other than A354V can permit the development of metabolic crises. At present we cannot draw any conclusions about the significance, if any, of the observation that this patient survived multiple crises. It cannot be excluded that the A354V mutation may confer special characteristics to LRPPRC that are particularly deleterious to survival.
Most episodes of acute and subacute neurological deterioration were similar to LS episodes reported in other infantile mitochondrial diseases such as SURF1 deficiency. Some older patients developed crises resembling stroke-like episodes, with focal neurological deficits, seizures, and neuroradiological evidence of abnormal cortical metabolism. We have previously speculated that the neurological episodes of LS and stroke-like episodes may result from similar mechanisms.11
There were many clinical similarities between SLSJ-COX patients and SURF1 deficient LS patients (table 1 and supplementary table 1). The survival curves of these two groups of LS patients are distinct (figure 3). However, because pronounced differences in lifespan are possible among patients with the same disease, differences in mean survival are of little use clinically in discriminating between the two conditions. Furthermore, the group data do not provide precise prognostic information for individual patients.
In contrast, severe metabolic crises were not described in SURF1 deficient LS. Perhaps the reports of these patients focused on other aspects of the clinical presentation and neglected to mention these episodes. However, it seems highly unlikely that the frequent, dramatic, fatal episodes of metabolic acidosis of SLSJ-COX would have escaped mention in at least some of the 39 case reports of SURF1 deficient LS, had they occurred at a similar rate. Thus it is likely that metabolic crises are a true difference between SLSJ-COX and SURF1 related LS. We suggest that if a child with LS develops acute, severe metabolic acidosis, LRPPRC might be considered as a candidate causal gene, even in non-French-Canadian patients.
The occasional occurrence of cardiomyopathy and renal tubular dysfunction in SURF1-deficient LS also differs from their absence in SLSJ-COX, despite full evaluations in recent years that would have detected such problems. Interestingly, in SLSJ-COX, COX deficiency is most severe in liver and brain, and kidney and heart have considerable residual COX activity.7 This correlates well with the metabolic and neurologic symptoms of SLSJ-COX.
The clinical spectra of other types of COX deficiency are not well defined. Three patients with COX15 mutations have been reported with cardioencephalopathy or LS.14 Patients with COX deficiency caused by SCO2 mutation have developed encephalomyopathy and fatal hypertrophic cardiomyopathy15 or neurogenic muscle atrophy.16 Metabolic crises and LS were not described in these patients. Only two families with SCO1 mutations have been reported, presenting with neonatal encephalopathy and liver failure,17 and with progressive hypertrophic cardiomyopathy, encephalopathy, and lactic acidosis.18
In conclusion, SLSJ-COX is a clinically and biochemically distinct form of LS. The occurrence of acute acidotic crises in a child with suspected mitochondrial disease may be suggestive of LRPPRC related COX deficiency, even in non-French-Canadian patients. Pronounced differences in survival and severity of clinical signs occur among homozygotes for the A354V mutation, suggesting that naturally occurring variants in other genes and/or environmental factors have a major influence on phenotype. By extension, this permits guarded optimism for the development of effective treatment in this and possibly other forms of LS.
Acknowledgments
We thank the patients and their families for their participation, the clinicians involved in the care of the patients, the Brandon J Teresi Foundation and the Consortium of Lactic Acidosis of Saguenay-Lac Saint-Jean. The financial support of the Consortium of Lactic Acidosis of Saguenay-Lac Saint-Jean is gratefully acknowledged.
References
Supplementary materials
Web Only Data
Files in this Data Supplement:
Footnotes
Funding F-GD was financed in part by L'Association de l'Acidose Lactique du Saguenay-Lac Saint-Jean, Québec, Canada for his contribution to the design and conductance of the study.
Competing interests None to declare.
Patient consent Obtained.
Provenance and peer review Not commissioned; externally peer reviewed.