Glutaric acidemia type I (GA1, MIM 231670) is characterized clinically by dystonia and dyskinesia in childhood; pathologically by striatal degeneration, in particular of the caudate and putamen; and biochemically by tissue deficiency of glutaryl-CoA dehydrogenase. Large amounts of glutaric and 3-hydroxyglutaric acids are usually present in urine. The disorder is transmitted as an autosomal recessive trait. Treatment with L-carnitine and prompt treatment of catabolic events with intravenous fluids, glucose, and insulin may prevent neurologic symptoms in patients without striatal damage at diagnosis, but the effect of treatment with riboflavin and dietary restriction of glutarigenic amino acids (lysine and tryptophan) in such patients is much less clear.
Macrocephaly at or appearing shortly after birth is common. A period of apparently normal development averaging 1 year is typical; although soft neurologic signs such as irritability and jitteriness are often present, symptoms rarely if ever have their onset after 5 years of age. The disease usually presents as an acute encephalopathic episode during a respiratory or gastrointestinal illness, with the sudden onset of hypotonia, loss of head control, and seizures, followed by opisthotonus, grimacing, fisting, tongue thrusting, rigidity, and dystonia (Fig. 95-5); recovery is slow and incomplete. Neurologic symptoms may then remain static, in which case the condition resembles extrapyramidal cerebral palsy, or may progress slowly, punctuated by episodes of ketosis, vomiting, hepatomegaly, and encephalopathy (coma and convulsions) associated with infection. In some patients motor delay, hypotonia, dystonia, and dyskinesia develop gradually during the first few years of life. Whatever the type of onset, there is relative preservation of intellect. Episodes of high fever in patients with severe dystonia may be due to constant and unabated activity of large skeletal muscles. Patients with severe neurologic impairment may survive well into adult life or may succumb during the first decade with intercurrent illnesses or Reye syndrome-like episodes; the latter become less frequent with age. Some enzyme-deficient individuals, including the occasional sibling of a severely affected child, do not develop neurologic manifestations.40,42
A 13-month-old boy with glutaric acidemia, showing dystonia of face, tongue, neck, back, arms, and hands. This child did not have abnormal urine organic acids (Courtesy of Dr. I. Bergman.)
Most routine laboratory studies, including serum electrolytes and pH, are normal except during acute episodes, when acidosis, hypoglycemia, ketonemia and ketonuria, hyperammonemia, and mild parenchymal liver disease may appear.34,49,50 Routine studies on CSF may show slightly increased protein or may be normal. Amino acids in blood and urine are usually normal, but during acute episodes serum 2-aminoadipic acid may be greatly elevated and the urine may show generalized aminoaciduria, with special prominence of glutamine, glutamic acid, 2-aminoadipic acid, and saccharopine.34,49 Similar changes have been described in patients with typical Reye syndrome and are not diagnostic. Serum L-carnitine is almost always very low at diagnosis, with an increased esterified fraction, and increased glutarylcarnitine can be demonstrated in plasma and urine.38,39,41,42,45
Urine organic acids are usually abnormal (Fig. 95-6), showing large quantities of glutaric acid (up to 22 mg/mg creatinine; normal less than 15 μg/mg creatinine) and lesser amounts of 3-hydroxyglutaric acid (up to 0.7 mg/mg creatinine) and (occasionally) glutaconic acid.51 Excretion of glutaconic acid may become prominent, exceeding that of 3-hydroxyglutaric acid, during episodes of ketosis.34,49,50
Urine organic acids in glutaric acidemia. A = dimethylmalonic acid (internal standard), B = succinic acid, C = glutaric acid, D = 3-hydroxyglutaric acid, and E = hippuric acid. (Goodman, unpublished.)
Glutaric acid concentrations are also increased in blood and CSF30,31 but may be demonstrable only by stable isotope dilution GC/MS. Glutaric acid in urine, blood, and CSF is elevated in some patients only when they are acutely ill39,41,43,44 but is normal in other patients even during catabolic episodes. A few patients are known in whom glutaric acid was elevated only in CSF at diagnosis.48,52
Neuroradiologic Phenotype (Fig. 95-7)
CT and MRI scans of brain in glutaric acidemia. Left: CT scan of 10-month-old boy, 3 months after the onset of movement disorder, showing loss of cerebral volume, most notable in the temporal lobes, and widening of Sylvian fissures. (Courtesy of Dr. I. Bergman.) Middle: CT scan of a 5-month-old boy without movement disorder but with abnormal urine organic acids and glutaryl-CoA dehydrogenase deficiency, showing the same changes. (Courtesy of Dr. S. Seshia.) Right: MRI scan of a 3-year-old girl showing cortical atrophy, dilated lateral ventricles, and shrinkage and increased intensity of the caudate and putamen, suggesting fibrosis.
The neuroradiologic findings in GA1 have been reviewed by Brismar and Ozand.53 CT and MRI scans often show collections of fluid over the frontal lobes and at the base of the brain, with atrophy of the frontal and temporal lobes and failure of the temporal lobes to cover the insular region. In some instances the extracerebral fluid collections in the Sylvian region are in arachnoid (or subdural) cysts.54,55 These changes may be present at birth and have been observed in otherwise clinically normal infants in whom the diagnosis was made during evaluation for macrocephaly.46,56,57 White matter changes may be observed in association with these abnormalities. Subdural hematomas have been noted in several patients,58-61 possibly because the fluid collections between the brain and cranium stretch the bridging vessels, making them more vulnerable to minor trauma. Changes in the basal ganglia, in particular atrophy of the caudate and putamen, are rarely present at birth but can develop within a few days or weeks of an encephalopathic episode.62 The globus pallidus and the thalamus are usually spared.
The striatum bears the brunt of neurotoxicity in GA1. Moderate fatty changes in neurons in the caudate and increased numbers of astrocytes in the putamen were observed in two severely affected patients who died at about 1 year of age.37,39 Changes in the basal ganglia were much more marked in older patients,49,63-65 all of whom showed severe neuronal loss and extensive fibrous gliosis in the caudate and putamen (Fig. 95-8). Some also showed degeneration of the globus pallidus and spongy degeneration of cortical white matter; in one case this was due to splitting of myelin sheaths.65
Histologic section of the caudate nucleus from a 10-year-old boy with glutaric acidemia, showing almost total neuronal loss and replacement by gliotic tissue. H & E; original magnification × 300. (Courtesy of Dr. M. D. Norenberg.)
Most autopsied patients have also shown microvesicular fatty infiltration of liver parenchymal cells, cells of the proximal renal tubule, and myocardial cells.37,39,49 These changes are not pathognomonic and are seen in several other disorders, including Reye syndrome, other organic acidemias, and disorders of the urea cycle.
Molecular Basis of Disease
GCD cDNA from pig, human, and mouse has been cloned and characterized, as have the human (GenBank accession number U69141) and murine (GenBank accession number U18992) GCD genes.7,66,67 The human gene has been mapped to chromosome 19p13.2,68 is about 7 kb long, and contains 11 exons.69 More than 60 pathogenic GCD mutations have been found in GA1 patients (Fig. 95-9),70 but there is no single common mutation except in inbred populations like Island Lake Indians in Canada (IVS1+5G>T)71 and the Old Order Amish in Lancaster County in Pennsylvania (A421V).69 The most common mutation outside of these populations is R402W,69 which accounts for fewer than 20 percent of mutant alleles; most patients are compound heterozygotes for mutations found in comparatively few individuals.
Pathogenic mutations in glutaryl-CoA dehydrogenase. (Reprinted by permission of John Wiley & Sons, Inc. from Goodman SI, Stein DE, Schlesinger S, Christensen E, Schwartz M, Greenberg CR, Elpeleg O: Glutaryl-CoA dehydrogenase mutations in glutaric acidemia (type 1): Review and report of thirty novel mutations. (Hum Mutat 12: 141, 1998.))
There is no apparent correlation between genotype and clinical severity. Patients in populations in which single mutations predominate show a variety of clinical phenotypes,43,44 and in other populations sibships contain GCD-deficient individuals who are clinically unaffected.40,42 This is true even when unaffected and affected sibs are homozygous for a mutation that terminates translation upstream of the substrate binding site.72 It appears that, whatever the responsible mutation(s), GCD deficiency confers a high but not absolute risk for developing striatal damage in infancy or childhood.
Specific mutations do, however, correlate with the presence and/or severity of organic aciduria. The R227P allele has been associated with little or no excretion of glutaric acid,73 even in heterozygous patients, and there may be a similar association with A293T, G178R, and R88C.70 The reason for this is not clear, as GCD containing any of these substitutions produces stable subunits with very low specific activity when expressed in E. coli. The possibility that intragenic complementation may occur between particular mutations has not been examined.
The excretion of glutaric acid is probably due to accumulation of glutaryl-CoA behind deficient glutaryl-CoA dehydrogenase, whose activity varies from 0 to 10 percent of normal in most GA1 patient tissues, and subsequent hydrolysis to the free acid by intracellular thioesterases. Glutaconyl-CoA, the presumed source of 3-hydroxyglutaric acid, is not produced by mutant GCD 16,49 but may be formed from accumulated glutaryl-CoA by MCAD in mitochondria or by glutaryl-CoA oxidase in peroxisomes.18 Because 3-hydroxyglutaric acid is not excreted in glutaric acidemia type II, in which MCAD and other mitochondrial dehydrogenases are deficient and in which peroxisomal glutaryl-CoA oxidase is presumably normal, its origin in GA1 is more likely to be MCAD or a related mitochondrial enzyme.
Glutaric acid concentrations have been increased in all tissues that have been examined at death, including brain. The concentration of glutaric acid in brain may be different in different areas; in one patient it was 0.67 μmol/g wet weight in frontal cortex and 1.25 μmol/g wet weight in basal ganglia.63 In another patient, the glutaric acid concentration in frontal cortex was 0.83 μmol/g wet weight (normal = undetected).49 Glutaconic acid has not been detected in brain, even when large amounts of the compound were being excreted in urine just before death.49
The concentrations of γ-aminobutyric acid (GABA) were extremely low in the caudate (0.60 μmol/g wet weight; normal = 3.03 ± 0.83) and putamen (0.87 μmol/g wet weight; normal = 5.69 ± 0.75) of one patient and correlated with severe deficiency of neuronal glutamate decarboxylase in these areas. Enzyme activities, and the concentrations of GABA, were much nearer to the normal range in frontal, occipital, and cerebellar cortex.63
Any theory to explain the pathogenesis of GA1 must explain the particular vulnerability of the striatum, and in particular the caudate and putamen, to neuronal degeneration; why striatal damage rarely if ever occurs after 5 years of age; and why it occurs in patients with only minimal elevations of abnormal metabolites in blood and CSF. No theory has been put forth that addresses all these points adequately. Nonetheless, several studies have examined the possibility that striatal degeneration in some way relates to accumulation of glutaric acid or one of its derivatives.
Different effects of accumulated metabolites on striatal cultures have been noted. One study showed that glutaric acid caused neurodegenerative changes in cortical-striatal cultures and that, unlike kainic acid and quinolinic acid, glutaric acid toxicity did not require prior synapse formation but occurred even before synapses were well established.74,75 Another study on the effects of glutaric, 3-hydroxyglutaric, and glutaconic acids on cultures from rat hippocampus, cortex, and striatum showed only a reversible cytotoxic effect from 5 mM glutaric and glutaconic but also showed that 3-hydroxyglutaric acid produced concentration-dependent toxicity starting at 1.5 mM.76 This effect could be reduced by glutamate receptor antagonists, suggesting that an excitotoxic mechanism is involved.
Glutaric, glutaconic, and 3-hydroxyglutaric acids competitively inhibit neuronal glutamic acid decarboxylase (GAD), the enzyme responsible for GABA biosynthesis, and the Ki of glutarate (1.3 × 10−3 molar)77 approximates the concentration in which it is found in brain in glutaric acidemia. The low concentrations of GABA found in the caudate and putamen in one patient63 are consistent with such an effect, but it is not known if inhibition of GAD alone could produce neuronal damage and loss. Further, decreased GAD activity and GABA concentration in the striatum may be secondary to cell death from some other cause, as it is in Huntington chorea.78
Evidence is accumulating that excitotoxicity plays a major role in the pathogenesis of disorders such as Huntington disease, and glutaric and/or 3-hydroxyglutaric acid might cause cell death in the striatum if it excited excessively one or all classes of glutamate receptors. High affinity uptake of glutamic acid by rat synaptosomes is inhibited 80 percent by 10 mM glutaric acid79 but not at all by 0.1 mM,80 but it is not known if the 1.0 to 1.5 mM concentrations found in brain in glutaric acidemia can inhibit glutamate uptake and thus cause high and possibly stimulatory concentrations of the neurotransmitter to be retained in the synaptic cleft. The effects of 3-hydroxyglutaric and glutaconic acid have not been studied.
Another possible pathogenetic mechanism involves quinolinic acid (QA), an intermediate in tryptophan metabolism in brain (Fig. 95-10). QA is a potent neurotoxin when injected into the CNS of experimental animals,81 and neurones are so sensitive to its effects that even endogenous concentrations may be toxic.82 A block in glutaryl-CoA dehydrogenase could shunt tryptophan down the kynurenine pathway, particularly when viral infections provoke the production of the α- and γ-interferons that induce indolamine-2,3-dioxygenase.83
The two pathways of tryptophan metabolism in brain, one through glutaryl-CoA and one through quinolinic acid.
Whatever the toxic compound(s), the limited time during which the striatum can apparently be damaged suggests particular vulnerability of a receptor that is spatially and temporally regulated, i.e., that is present in the striatum only for a limited time during development. Glutamate receptors that fit this model have been described in the rat.84
The cause of the fatty changes seen at autopsy in the liver, kidneys, and heart is also unknown, but they may merely be due to nonspecific mitochondrial toxicity. The observation of severe ketosis during Reye syndrome-like episodes in this condition suggests that carnitine deficiency, if present, does not seriously limit the capacity of the liver to oxidize long-chain fatty acids. It is also possible that glutaconic acid, which is excreted in large amounts during ketotic episodes, is a mitochondrial toxin; excretion may rise during ketotic episodes because the enoyl-CoA intermediates of fatty acid oxidation competitively inhibit hydration of glutaconyl-CoA to 3-hydroxyglutaryl-CoA. Whatever its cause, the accumulation of 2-aminoadipic acid and saccharopine observed during acute episodes appears to relate more to general mitochondrial dysfunction than to back-up of metabolites proximal to glutaryl-CoA, because the same abnormalities have been described in Reye syndrome.85,86
Pedigree analysis, which shows that males and females are affected with approximately equal frequency, and enzyme analysis, showing partial GCD deficiency in cells from obligate heterozygotes, establish inheritance of GA1 as an autosomal recessive trait.87,88 Heterozygous carriers are clinically normal, and two carriers did not excrete detectable glutaric acid even after oral loads of L-lysine.30 There has not been enough experience with carrier detection based on GCD activity in leukocytes or fibroblasts to be certain about its reliability in carrier detection, but the elucidation of pathogenic mutations in the GCD gene and the development of rapid and accurate methods for mutation screening, e.g., allele specific hybridization, now permit carriers to be identified with certainty in families and populations in which disease-causing mutations are known.
Well over 200 patients have been detected since GA1 was first described in 1975, and the condition is certainly one of the more common organic acidemias. There are no accurate figures about incidence, however, because the disorder is very often misdiagnosed as “dystonic cerebral palsy” and because it is not usually screened for in newborns. Screening for glutarylcarnitine in dried blood spots from newborns by tandem MS has detected nine confirmed patients in about 450,000 livebirths,89 but most of them were from the Old Order Amish community in Lancaster County, a population in which the condition is especially frequent.44 The incidence of disease in the Island Lake Indians in Canada is on the order of 1:225 livebirths,71 and the incidence in Sweden may be about the same as that of phenylketonuria there, i.e., about 1 in 30,000 livebirths.90
In most cases the diagnosis is made on the basis of increased glutaric and 3-hydroxyglutaric acids in urine and confirmed by finding deficiency of glutaryl-CoA dehydrogenase in cultured fibroblasts. The diagnosis is almost certain if 3-hydroxyglutaric acid is identified, as it has not been found in any other condition. Increased urine glutaric acid may also be found in the setting of glutaric acidemia type II (Chap. 103), glutaryl-CoA oxidase deficiency,91 α-aminoadipic acidemia,25 short-gut syndrome,92 unusual intestinal flora,93 and use of infant formulas that contain medium-chain triglycerides. Because glutaric acid concentrations in urine, blood, and cerebrospinal fluid may be normal, GCD assay and/or mutation analysis should be carried out whenever the condition is strongly suspected on clinical or neuroradiologic grounds.
GCD activity in fibroblasts (or leukocytes) is usually less than 5 to 10 percent of normal in affected patients, whether measured by decarboxylation of [1,5-14C]glutaryl-CoA30 or by tritium release from [2,3,4-3H]glutaryl-CoA,16 but can be as high as 30 percent of normal in patients with the R227P/V400M genotype.73,94 This amount of enzyme activity is not unusual in heterozygous carriers, and positive diagnosis of such patients may be possible only by mutation analysis.
A rationale for newborn screening comes from the observation that treatment can prevent development of striatal damage if begun before the onset of symptoms61 (see below). Other than in populations with one prevalent mutation, in which screening for GA1 might be done by mutation analysis, the only test available that appears to have the required sensitivity and specificity is to measure glutarylcarnitine in dried blood spots from newborns by tandem MS. A program based on this methodology has been initiated in a US population comprised in part of the Old Order Amish, in whom the disease's incidence is particularly high. While false positive results with this technique are expected to be rare, and have not been detected, there are still too little data available to know the incidence of the disease in the population at large and the frequency of false negative results. It is, however, expected that GA1 will increasingly be the focus of newborn screening programs.
Prenatal diagnosis of an affected fetus has been achieved on several occasions, in some cases by demonstrating increased glutaric acid levels in amniotic fluid and in others by showing enzyme deficiency in cultured cells or directly in chorionic villus samples. One affected fetus showed increased glutaric acid in amniotic fluid (4.35 to 13.3 μg/ml; normal <0.39) and decreased GCD activity in cultured amniotic cells; the diagnosis was confirmed by the presence of large amounts of glutaric acid in fetal liver (17 μg/g wet weight), brain (112 μg/g wet weight), and kidney (70.2 μg/g wet weight) and by decreased GCD activity in liver and kidney. Minimal and possibly insignificant changes were observed in the striatum.95 Four other affected fetuses have been detected by assaying GCD activity in chorionic villus samples directly and/or in cultured chorionic cells.96 Prenatal diagnosis by mutation analysis, while possible, has not yet been reported.
It may be that measuring glutaric acid in amniotic fluid is predictive only when the proband in the family consistently excretes large amounts of glutaric acid and that enzyme assay and/or mutation analysis is necessary in other instances.
There is increasing evidence that striatal degeneration can be prevented in many patients by treatment with L-carnitine (100 mg/kg per day) and by prompt and vigorous treatment of intercurrent illnesses with fluids, insulin, and glucose.61,97 Without treatment, almost all patients with GA1 develop an incapacitating dystonic-dyskinetic disorder, but more than 80 percent of patients treated in this way have developed normally.61,97 The role of protein, or lysine and tryptophan, restriction in treatment is not clear, but few patients have been helped by diet alone. Results of treatment when striatal damage is present at the time of diagnosis are less favorable, although it is possible that further damage is prevented.61
Observations that glutaric, glutaconic, and 3-hydroxyglutaric acid inhibit neuronal GAD 77 and that GAD activity and GABA concentrations are low in the basal ganglia of glutaric acidemia patients63 have prompted treatment with pharmacologic agents that increase GABA concentrations in brain. Those that have been tried are baclofen, valproic acid, and vigabatrin.
It was originally thought that baclofen, i.e., β-(4-chlorophenyl)GABA, activated GABA receptors, but it has since been learned that it activates neurons that are normally inhibited by this neurotransmitter.98,99 Whatever effect it has in this disorder may thus not relate to a GABA-like function. Baclofen therapy has been tried in several patients, with variable effect. In some patients there has been no effect whatever,37 but Brandt et al. have reported significant improvement in two of three Danish patients given 2 mg/kg per day, even when administration was controlled in a double blind fashion.100 Effects varying from none to questionable have been observed in several additional patients, but we know of no other case in which the effects were as dramatic as those noted in the Danish patients.
Valproic acid therapy has also been tried in several patients, with the rationale that the drug causes selective increase of GABA in synaptic areas by inhibiting GABA transaminase or succinic semialdehyde dehydrogenase or by inhibiting GABA uptake by glial cells and nerve endings. Some improvement on valproic acid has been noted in one patient in the literature,36 but many researchers feel that, because it competes with glutaryl-CoA for esterification with carnitine, valproic acid should be avoided whenever possible.45
There is one brief report of significant clinical improvement, and increased concentrations of GABA and glutaric acid in CSF, following the use of vigabatrin (γ-vinyl-GABA) at an oral dose of 35 to 50 mg/kg per day.101