Human Molecular Genetics

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (39) Request Permissions Google Scholar Articles by Koeller, D. M. Articles by Goodman, S. I. Search for Related Content PubMed PubMed Citation Articles by Koeller, D. M. Articles by Goodman, S. I. Social Bookmarking          What's this?

Human Molecular Genetics, 2002, Vol. 11, No. 4 347-357
© 2002 Oxford University Press

Biochemical, pathologic and behavioral analysis of a mouse model of glutaric acidemia type I

David M. Koeller^1,+, Michael Woontner², Linda S. Crnic^2,3, Bette Kleinschmidt-DeMasters^4,5, Janet Stephens⁵, Edgar L. Hunt⁵ and Stephen I. Goodman²

¹Departments of Pediatrics and Molecular and Medical Genetics, Oregon Health and Science University, Portland, OR 97201, USA, and ²Department of Pediatrics, ³Department of Psychiatry, ⁴Department of Neurology and ⁵Department of Pathology, University of Colorado Health Sciences Center, Denver, CO 80262, USA

Received September 6, 2001; Revised and Accepted November 20, 2001.

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Glutaric acidemia type I (GA-I) is an autosomal recessive disorder of amino acid metabolism resulting from a deficiency of glutaryl-CoA dehydrogenase (GCDH). Patients accumulate glutaric acid (GA) and 3-OH glutaric acid (3-OHGA) in their blood, urine and CSF. Clinically, GA-I is characterized by macrocephaly, progressive dystonia and dyskinesia. Degeneration of the caudate and putamen of the basal ganglia, widening of the Sylvian fissures, fronto-temporal atrophy and severe spongiform change in the white matter are also commonly observed. In this report we describe the phenotype of a mouse model of GA-I generated via targeted deletion of the Gcdh gene in embryonic stem cells. The Gcdh^–/– mice have a biochemical phenotype very similar to human GA-I patients, including elevations of GA and 3-OHGA at levels similar to those seen in GA-I patients. The affected mice have a mild motor deficit but do not develop the progressive dystonia seen in human patients. Pathologically, the Gcdh^–/– mice have a diffuse spongiform myelinopathy similar to that seen in GA-I patients. However, unlike in human patients, there is no evidence of neuron loss or astrogliosis in the striatum. Subjecting the Gcdh^–/– mice to a metabolic stress, which often precipitates an encephalopathic crisis and the development of dystonia in GA-I patients, failed to have any neurologic effect on the mice. We hypothesize that the lack of similarity in regards to the neurologic phenotype and striatal pathology of GA-I patients, as compared with the Gcdh^–/– mice, is due to intrinsic differences between the striata of mice and men.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Glutaric acidemia type I (GA-I) is an autosomal recessive disorder of amino acid metabolism due to mutations in the gene for glutaryl-CoA dehydrogenase (GCDH) (EC 1.3.99.7). GCDH catalyzes the oxidative decarboxylation of glutaryl-CoA (Fig. 1), an intermediate in the catabolism of tryptophan, lysine and hydroxylysine (1). The lack of GCDH activity in patients with GA-I leads to the accumulation of glutaric acid (GA) and 3-OH glutaric acid (3-OHGA) in the blood, urine and CSF.

View larger version (10K): [in this window] [in a new window]   Figure 1. Metabolism of glutaryl-CoA.

 
Clinically, GA-I is characterized by macrocephaly, progressive dystonia and dyskinesia, which usually is apparent within the first 3 years of life. Symptoms may have a gradual rate of onset and progression, or occur suddenly after an acute metabolic crisis. Degeneration of the caudate and putamen of the basal ganglia, widening of the Sylvian fissures, and fronto-temporal atrophy are commonly demonstrable by radiographic imaging studies of the brain. Pathologically the most characteristic feature of GA-I is a loss of neurons in the caudate and putamen (2). Severe spongiform change in the white matter is also commonly observed (2,3).

In rare instances, patients with an inherited deficiency of GCDH remain asymptomatic (4), leading to the hypothesis that environmental factors are important in the pathophysiology of GA-I (5). The onset of symptoms often occurs following an acute viral illness, suggesting that an acute rise in serum and/or CSF levels of GA and 3-OHGA associated with catabolism may be the trigger that eventually results in neuronal cell death. Consistent with this hypothesis, patients who are diagnosed with GA-I pre-symptomatically can escape the typical striatal damage via aggressive treatment with IV glucose to prevent the catabolism associated with acute illness (6,7). The cellular mechanisms for the neuronal damage that occurs in GA-I are not known. Several reports have demonstrated that 3-OH glutarate causes excitotoxic damage to cultured neurons in vitro (8–10). The ability to block 3-OH glutarate toxicity by specific glutamate receptor antagonists has led to the hypothesis that these receptors have a role in the pathophysiology of GA-I. An alternative hypothesis is that elevations of interferon associated with viral infection may increase production of quinolinic acid, an intermediate in tryptophan oxidation in brain and a potent neurotoxin (11). An additional question about the pathophysiology of GAI is the basis for the localization of the primary neuropathology to the caudate and putamen nuclei of the striatum. We have generated GCDH deficient mice via homologous recombination in embryonic stem (ES) cells. In this report we describe the initial phenotypic analysis of this murine model of GA-I.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Biochemical phenotype of Gcdh^–/– mice
A line of Gcdh^–/– mice [Gcdh^{tm1Dmk (–/–)}] was generated via homologous insertion of a gene targeting vector which resulted in a deletion of the first 7 exons of the Gcdh gene, and the insertion of a ß-galactosidase reporter gene (nlacF) controlled by Gcdh chromosomal regulatory elements (Fig. 2) (12). Homologous insertion of the targeting vector was identified by PCR analysis of both the 5' and 3' ends of the locus. Enzymatic assay of glutaryl-CoA dehydrogenase activity from samples of liver confirmed a complete loss of activity in Gcdh^–/– animals (not shown). Genotype analysis of the progeny of heterozygote-by-heterozygote matings (Gcdh^+/– x Gcdh^+/–) showed the expected Mendelian segregation ratio, indicating that Gcdh^–/– animals have normal fetal and post-natal viability. There was no effect of genotype on birth weight, neonatal growth or final adult weight.

View larger version (8K): [in this window] [in a new window]   Figure 2. Gene targeting vector for disruption of the Gcdh gene. Approximately 8.5 kb of 5' homologous DNA ending at the translation start site in exon 1 of the Gcdh gene was fused in frame with the ATG start codon of the nlacF gene. The 3' homologous DNA includes all of exon 8 and its flanking intron sequences. Homologous insertion of the targeting vector results in deletion of the 3' end of exon 1 and all of exons 2–7. Homologous recombination at the 3' end was confirmed by PCR analysis using a forward primer in the NEO gene and a reverse primer in intron 8 (arrows). Analysis of the 5' end was via a PCR reaction using a reverse primer in the nlacF gene and a forward primer from genomic sequence upstream of the 5' homologous DNA (arrows).

 
Patients with GA-I accumulate large amounts of both GA and 3-OHGA and the Gcdh^–/– mice excreted levels of GA and 3-OHGA similar to those in patients with GA-I (Table 1). Measurement of urinary metabolite excretion from a cross-section of animals of different ages did not demonstrate any significant age-dependent variation (Fig. 3A and B). The level of urine 3-OHGA did tend to increase as the level of GA increased, however, the correlation was not statistically significant. Elevation of glutaconic acid is also occasionally seen in patients with GA-I, particularly during periods of ketosis. We did not detect any elevation of glutaconic acid in the Gcdh^–/– mice, which may reflect the lack of ketosis in these animals. Aside from the elevations of GA and 3-OHGA, the urine organic acids in the Gcdh^–/– mice were the same as wild-type animals. Brain levels of GA and 3-OHGA were also elevated in Gcdh^–/– mice (Table 1). The levels of GA and 3-OHGA were not determined in the CSF, but given the elevations in the brain parenchyma and blood are likely to be elevated. As in human GA-I patients, the Gcdh^–/– mice have elevations of both serum glutaryl-carnitine and urine glutaryl-glycine (not shown). The urinary loss of large amounts of carnitine in patients with metabolic disorders can result in serum and whole body carnitine depletion. To look for evidence of carnitine depletion we measured both serum and tissue levels and found that the serum total, free and acyl-carnitine levels, as well as the acyl:free carnitine ratios were normal in Gcdh^–/– mice (data not shown). Tissue carnitine levels in liver, skeletal muscle and heart were also unaffected by genotype.

View this table: [in this window] [in a new window]   Table 1. GA and 3-OHGA levels in Gcdh–/– mice

 

View larger version (11K): [in this window] [in a new window]   Figure 3. Cross-sectional analysis of urine glutaric and 3-OHGA excretion in Gcdh–/– mice. Each point represents the level of urine glutarate (A) or 3-OH glutarate level (B) from a single animal.

 
Expression of Gcdh in the mouse brain
The gene targeting vector used to disrupt the mouse Gcdh gene inserted the nlacF gene in frame with the GCDH translation start site in exon 1 (Fig. 2). Expression of the nuclear localized ß-galactosidase encoded by the nlacF gene is therefore controlled by the chromosomal elements that regulate the expression of Gcdh, allowing this reporter to be used to evaluate the expression of the Gcdh gene in the brain. ß-Galactosidase expression was highest in the granule cells of the hippocampus and cerebellum (Fig. 4). Expression was also detected in some layers of the cerebral cortex, with very little staining in the striatum. The expression of the lacZ reporter corresponds to the regions of the brain previously demonstrated to have the highest level of GCDH by western blotting (12).

View larger version (82K): [in this window] [in a new window]   Figure 4. ß-Galactosidase staining of coronal sections of brains from control mice (A and B) and mice expressing the nlacF gene under the control of the Gcdh promoter (C and D). Higher magnification demonstrates that ß-galactosidase activity is highest in the granule cells of the hippocampus (E) and cerebellum (F).

 
Pathologic analysis of Gcdh^–/– mice
There were no external physical differences between wild-type, Gcdh^+/– and Gcdh^–/– mice. Total body weights were similar for animals of all three genotypes, but there was a significant difference in kidney weights. Both male and female Gcdh^–/– mice had larger kidneys than either wild-type or heterozygous animals (Table 2). For all genotypes, hepatic lipid increased with age. The increase in hepatic lipid level was also apparent by measurement of total liver fatty acid levels (data not shown). Heart and kidney histology was no different in wild-type and knockout animals.

View this table: [in this window] [in a new window]   Table 2. Comparison of kidney size between control and Gcdh–/– mice

 
Light microscopic study of the brain demonstrated a focal, spongiotic change, maximal in the frontal cortex, where indentation of neurons was apparent (Fig. 5). Similar pathology was also present in the striatum, where it was almost exclusively limited to the striate white matter pencil fibers. Older animals also demonstrated spongiotic changes in the cerebellar white matter and hippocampus. Two pathologists blinded to the genotype of the animal detected spongy changes in 9/9 Gcdh^–/– mice, whereas no change was seen in nine Gcdh^+/+ and 10 Gcdh^+/– mice examined. Spongiotic change was never associated with cellular necrosis, microgliosis, astroctyosis or an inflammatory response. No myelin loss was evident in the affected striate white matter pencil fibers. In all cases the spongiotic changes were patchy and never uniformly distributed throughout the affected structure. The spongiotic changes were more severe in older animals, particularly in the striatum and cerebellar white matter (Fig. 6). Spongiotic changes in the thalamus and brain stem became more pronounced in some aged animals. The changes were similar on the two genetic backgrounds. Necrotic changes were apparent in the livers of many of the older mice, but did not correlate with genotype.

View larger version (115K): [in this window] [in a new window]   Figure 5. Low (top) and high (bottom) power photomicrographs of the vacuolization that characterizes brains of the Gcdh–/– mice; vacuoles indent and displace neurons in the cerebral cortex. Note the absence of inflammation, microgliosis, astrocytosis or neuronal necrosis (at low power) and the delicate strands of cell membranes that traverse the vacuoles (high power). Hematoxylin and eosin, 330x, 800x.

 

View larger version (118K): [in this window] [in a new window]   Figure 6. Photomicrographs taken from the longest surviving Gcdh–/– mice demonstrate striking vacuolization in the bundles of white matter fibers in the corpus striatum (top) and in the white matter of the cerebellum (bottom). Hematoxylin and eosin, 500x, 200x.

 
Electron microscopic examination of the striatum from an 18-month-old Gcdh^–/– animal showed numerous large vacuoles not associated with necrosis or inflammation. Vacuole size was variable, but some were estimated to be at least 20 times the cross-sectional diameter of nearby intact myelinated axons. Some vacuoles contained an amorphous fluid of moderate electron density (Fig. 7). The nature of the accumulated fluid could not be determined, the fixation used for electron microscopy would have preserved lipid had it been present. However, in most areas the vacuoles appeared devoid of fluid and were sometimes traversed by delicate membranous strands. No residual axons or sub-cellular components were identified within the large vacuoles. In a few instances, electron-dense myelin lamellae bordered both sides of the vacuoles, suggesting the vacuoles to have formed within and between the myelin layers due to myelin splitting. However, in other areas the vacuolization occurred adjacent to, rather than within, individual myelin sheaths (Fig. 7). These changes in the white matter are similar to those described in human patients with GA-1 (2,3), as illustrated in a photomicrograph taken from the brain of a young girl with the disorder (Fig. 8).

View larger version (198K): [in this window] [in a new window]   Figure 7. Electron micrograph prepared from the striatum of an 18-month-old Gcdh–/– animal illustrates a vacuole containing amorphous fluid of moderate electron density. The vacuole contains neither axons nor sub-cellular components and is eccentrically displacing myelin sheaths (black ovoid structures, upper left of photomicrograph), rather than occurring between myelin lamellae. Uranyl acetate/lead citrate stain, 2700x.

 

View larger version (185K): [in this window] [in a new window]   Figure 8. Photomicrograph of the frontal white matter from a young girl with GA-1 shows the widespread white matter vacuolization (‘spongy myelinopathy’) seen in actual patients with this disorder; note the similarity to the vacuolization seen in the Gcdh–/– mice in Figures 5 and 6. Hematoxylin and eosin, approximately 400x.

 
Behavioral analysis of Gcdh^–/– mice
Motor measures. Gcdh^–/– mice had difficulty righting themselves when placed on their backs during the first days of life, but no other overt signs of a movement disorder or decreased motor coordination were apparent in older animals. Similarly, the Gcdh^–/– mice could not be distinguished from normal littermates in home cage behavior or social interaction. To formally test motor coordination and balance, mice were evaluated in several ways. Male and female mice did not differ on these tests, nor did Gcdh^+/– and Gcdh^+/+ animals and their data for genders and control groups was combined. The Gcdh^–/– mice had a mild but statistically significant impairment on the rotorod task with significantly shorter times spent on the rod at asymptote (day 4) compared to control mice (F = 5.26, df = 1, 25, P = 0.03) (Fig. 9A). When the same animals were re-tested at 60 days of age, the difference was similar to that at 20 days, but did not reach significance (P = 0.10). To evaluate balance and fine motor coordination we measured the rate at which 80-day-old mice, which had previously been tested on the rotarod, could traverse a 5 mm square wooden beam. The Gcdh^–/– mice were slower than either wild-type or heterozygous animals at crossing the 5 mm wide beam (Fig. 9B). This was evident in both the time to traverse the rod (F = 4.56, df = 1, 23, P = 0.043) and the number of slips and falls made during the traverse (F = 4.60, df = 1, 23, P = 0.013) (Fig. 9C). We also evaluated the response to being suspended by their tails. Neurologically abnormal mice often demonstrate a clasping response in which they press their paws against their body (13). In contrast, normal mice struggle and flail with their limbs. The Gcdh^–/– mice showed a normal response to tail suspension

View larger version (17K): [in this window] [in a new window]   Figure 9. Evaluation of motor function in Gcdh–/– mice. (A) Performance of control (wild-type, Gcdh+/– and Gcdh+/+) and Gcdh–/– mice at asymptote (day 4) on rotating rod across the three trials given on that day. (B) Latency to cross a 5 mm square rod for wild-type and Gcdh–/– mice. (C) Number of slips of the foot from the rod combined with the number of times mice lost their footing and fell under the 5 mm rod, combining both trials. Gcdh–/– mice reached a significantly lower asymptote on the rotorod, were slower to traverse the narrow rod and made more slips and falls while crossing the rod than did the control group (see text for statistics).

 
Pre-pulse inhibition of the auditory startle response. Pre-pulse inhibition of the auditory startle response refers to the ability of a brief, low intensity sound pulse presented in advance of the startle stimulus to inhibit the startle response. This task is sensitive to striatal damage produced by treatment with the mitochondrial toxin 3-nitropropionate (14) and is impaired in both humans with Huntington’s disease (HD) (15) and in the mouse model for HD (16). There were no significant differences between Gcdh^+/+ and Gcdh^+/– animals, so their data were combined into one comparison group. All mice began to startle significantly to the 80 dB white noise pulse in the startle sensitivity test (see below) and as the pre-pulse should be a non-startling stimulus, the 81 dB pre-pulse was eliminated from the data analysis, leaving 67, 69 and 73 dB pre-pulses in the analysis. The data for percent inhibition of startle at the three pre-pulse intensities were first analyzed for gender and age effects. As is usual with this task (17), there were no significant effects of gender, so genders were combined in further analyses. Age was analyzed by splitting the mice into groups younger and older than 60 days of age. There was a significant interaction between age grouping and pre-pulse level (F = 3.09, df = 2, 140, P = 0.049) as well as a significant increase in inhibition with increasing pre-pulse level (F = 49.83, df = 3, 204, P = < 0.00001). The age interaction effect consisted of an increase in effects of age at higher pre-pulse levels. To further explore the effects of genotype, the youngest group of animals was analyzed separate from the older group and this analysis revealed a significant interaction between genotype and pre-pulse level (F = 4.09, df = 2, 50, P = 0.023) with the Gcdh^–/– mice who were less than 60 days of age at testing having increased inhibition of startle at all three pre-pulse levels (Fig. 10).

View larger version (16K): [in this window] [in a new window]   Figure 10. Prepulse inhibition of the auditory startle response across three prepulse intensities for the control (C, Gcdh+/– and Gcdh+/+) and Gcdh–/– (KO) mice that were less than 60 days old. The Gcdh–/– mice at this age have increased prepulse inhibition compared to control mice (see text for statistics).

 
Habituation of the auditory startle response. Trials were grouped into 10 blocks of 10 trials each for analysis. Data from Gcdh^+/+, Gcdh^+/–, and male and female animals did not differ significantly and were pooled into one comparison group. There was a significant genotype by block interaction (F = 2.16, df = 9, 234, P = 0.025), and a significant effect of block (F = 5.95, df = 9, 234, P < 0.000001), such that the control animals showed a significant decrease in startle amplitude, whereas the Gcdh^–/– mice showed a low, flat curve as demonstrated by a significant planned linear contrasts for the control group (P = 0.003) but not the Gcdh^–/– group (P = 0.54) (Fig. 11). The latency to reach maximum amplitude of the startle response was equal in the two genotypes at the start of testing, indicating that the motor mechanisms for the expression of the startle response are intact, although the trend toward reduced startle amplitude may indicate motor impairment.

View larger version (17K): [in this window] [in a new window]   Figure 11. Habituation of the auditory startle response. Responses were averaged over 10 blocks of 10 trials each. Values given are means and SEs of the mean of the maximum amplitude of the startle response measures in arbitrary units for control (C, Gcdh+/– and Gcdh+/+) and Gcdh–/– (KO) mice. The Gcdh–/– mice fail to habituate their startle response (see text for statistics) and tend to have a lower amplitude response.

 
Startle sensitivity curve. Sensitivity to white noise of various intensities was examined to ensure that differences between genotypes in pre-pulse inhibition and habituation of the startle response are not due to hearing deficits nor to deficits in ability to perform the startle response. There was no difference between Gcdh^+/+ and Gcdh^+/– mice or between males and females on maximum amplitude of startle adjusted for baseline amplitude, nor on startle latency, so the control groups and genders were combined for analysis. There was no significant effect of genotype on the maximum amplitude of startle and both groups showed a significant linear trend with startle amplitude rising with increasing startle stimulus intensity (overall test F = 18.35, df = 5, 130, P < 0.000001; planned comparison for control, F = 17.7, df = 1, 26, P = 0.00027, for Gcdh^–/–, F = 4.91, df = 1, 26, P = 0.036). The control mice showed a significant quadratic trend (F = 10.80, df = 1, 26, P = 0.003), whereas the Gcdh^–/– mice did not (P = 0.36). As in the habituation, there is a non-significant trend for the Gcdh^–/– mice to be lower in their startle response to the 120 dB stimulus. All groups of animals showed a significant increase in the startle response to 80 as compared to 70 dB (F = 21.6, df = 1, 25, P = 0.00009), thus the 81 dB pre-pulse data was eliminated from analysis as noted above.

Response of Gcdh^–/– mice to metabolic stress
Patients with GA-1 frequently have their initial onset of symptoms during or after an acute viral illness. To determine if an acute metabolic stress would precipitate symptoms similar to those seen in GA-I patients, we subjected 5- and 9-day-old mice to several different types of metabolic stress which in GA-I patients are associated with increased excretion of GA and 3-OHGA. These included injections of the interferon inducer poly-inosine/cytosine (poly-IC) or -interferon to mimic the cytokine response to acute viral infection, and lipopolysaccharide to mimic a gram-negative bacterial infection. All three treatments produced transient slowing of growth in all of the animals, but none had any effect on neurologic function. Catabolic stress produced by placing mice without food at 4°C also produced transient growth failure, but there was no difference between the different genotypes in this effect, and no neurologic impairment. In human GA-I patients it is known that there is a window of vulnerability to acute striatal degeneration during the first several years of life. To see whether in utero exposure to elevated GA and 3-OHGA would have any effect on neurologic outcome we evaluated the progeny of matings between homozygous Gcdh^–/– animals. The animals derived from these matings, which are all homozygous Gcdh^–/–, showed no evidence of increased severity of their neurologic phenotype as compared to the progeny of heterozygote x heterozygote matings.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Mice with a targeted deletion of the Gcdh gene had a biochemical phenotype similar to that seen in patients with GA-I; they had high levels of GA and 3-OHGA, and of the glycine and carnitine conjugates of GA. Urine levels of GA and 3-OHGA in the Gcdh^–/– mice showed little variation between animals, unlike patients with GA-I who can be divided into two classes (high and low), based upon their levels of GA excretion. In humans, the level of GA excretion correlates with genotype; however, neither the genotype nor the level of GA excretion correlates with clinical outcome (18,19). The lack of correlation between GA level, genotype and clinical outcome in GA-I suggests that environmental and genetic background differences are important in the development of the severe neurologic impairment in affected patients.

Macrocephaly and sub-dural hemorrhages are common in patients with GA-I (6). Neuropathologic studies of the brains of patients with GA-I demonstrate several consistent features, including fronto-temporal atrophy, widening of the Sylvian fissures and flattening of the anterior portion of the temporal lobes (2,3). A slight enlargement of the lateral ventricles with marked loss of volume of the caudate and putamen nuclei is also characteristic. Microscopically, there is loss of neurons and astrogliosis in the caudate and putamen and a diffuse spongiform myelinopathy.

Many of the neurologic features of GA-I are present in patients before any encephalopathic crisis or striatal damage (6). Such changes include macrocephaly, fronto-temporal atrophy, white matter changes and sub-dural hemorrhages. The universal presence of these changes, even in patients who never have an acute encephalopathic crisis, suggests that their pathogenesis is different from that of the acute striatal degeneration seen in GA-I. There were no gross size or structural changes in the brains of Gcdh^–/– mice, nor was there any neuron loss or gliosis in the striatum. However, the Gcdh^–/– mice did have the diffuse spongiform myelinopathy seen in human GA-I. The distribution of this neuropathology includes structures with the highest levels of Gcdh expression (cerebellum and hippocampus), as well as regions with lower levels of Gcdh expression (striatum and cortex). Spongiotic change within the brain is not unique to patients with GA-1. Spongiform change of gray matter is classically seen in prion disorders. Spongy myelinopathy is a feature of vitamin B12 deficiency, vacuolar myelopathy associated with human immunodeficiency virus, hexachlorophane toxicity, some mitochondrial cytopathies and Canavan’s disease (20). Spongy myelinopathy is also characteristic of other disorders of amino acid metabolism, including phenylketonuria, non-ketotic hyperglycinemia and maple syrup urine disease (20). Within mice, spongiotic change in brain may be seen in several types of mutants, with cerebral spongiform change most recently reported in mice with mutations in the mahogany gene Atrn (21). Nevertheless, the striking similarity between the spongiotic changes demonstrated in this mouse model and the human patients with GA-1 cannot be overlooked, and probably corresponds to the shared biochemical defects. The pathogenesis of the spongiotic change seen in these disorders is unknown.

The behavioral phenotype of the mice is consistent with the neuropathologic phenotype. The motor effects were very subtle, in keeping with the lack of cell death in the caudate and putamen. There was no defect in pre-pulse inhibition of the auditory startle response, a measure that is sensitive to striatal damage (14,15), again in keeping with the neuropathology. The youngest mice did show significant accentuation of pre-pulse inhibition, which may be evidence for abnormal function or development of the striatum. Deficits on this task develop in the mouse model of HD (16). In addition, while normal startle latencies demonstrate that motor functions required for this reflex are intact, the lower amplitude of startle in these mice at the 120 dB sound pressure level (SPL) in both the habituation and loudness function tasks does indicate some defect in processing or responding to this startle stimulus. Failure to habituate their responding to repeated startle stimuli may indicate some defect in this elementary form of learning, or may be a consequence of the lower startle response to the 120 dB stimulus.

Studies of HD mouse models clearly demonstrate that human genetic disorders of the striatum can be modeled in the mouse. In this experiment we have shown that another mouse model of striatal disease, GA-1, can also show pathological features in common with the corresponding human disorder. However, like the HD mouse models, the pathological features in our GA-1 mouse model appear to be milder than is seen in most patients at autopsy and may represent the earliest phases of the disorder. As in GA-I, patients with HD have striatal atrophy, neuron loss and astrogliosis. Mouse models of HD with progressive neurologic symptoms similar to those of human HD patients have been reported (13,16,22). The brains of the HD mice are smaller than control animals, but there is no specific decrease in striatal volume, or neuron loss in the caudate putamen (22). Striatal gliosis was noted in one report (13). The authors suggest that the pathology in these mice corresponds to that seen in the early stages of the human disease, and that the rate of progression of the phenotype in the mice may be too rapid for the striatal atrophy to occur before death. Alternatively, this may reflect an intrinsic difference in striatal metabolism between mice and men. Regardless, studies of HD mouse models clearly demonstrate that human genetic disorders of the striatum can be modeled in the mouse. The Gcdh^–/– mice have levels of GA and 3-OHGA similar to those of GA-I patients during a catabolic episode, and yet they lack overt behavioral or neuroanatomic evidence of significant striatal damage. We also failed to induce acute striatal damage in response to a variety of metabolic stresses such as cytokines and fasting. The lack of similarity in regards to the neurologic phenotype and striatal pathology of GA-I patients, as compared with the Gcdh^–/– mice, suggests that a better understanding of the differences between the striata of mice and men (such as their sensitivity to the excitotoxic effects of 3-OHGA) may be the key to unraveling the pathophysiology of GA-I.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
Gene targeting
A gene targeting vector was created by replacing exons 1–7 of the Gcdh gene with the nlacF and NEO genes (Fig. 2). nlacF (provided by Dr Jacques Peschon) is a derivative of the bacterial lacZ gene; it includes a nuclear localization signal from the SV40 large T antigen, which targets the ß-galactosidase activity to the nucleus (23). The nlacF gene was cloned in frame in exon 1 of the Gcdh gene at the AflIII site overlapping the ATG initiation codon. Expression of the nlacF gene, which lacks its own promoter, is controlled by the regulatory elements located 5' of the Gcdh gene (12). The NEO cassette (provided by Dr Philippe Soriano) containing the neomycin phosphotransferase gene was cloned immediately 3' of the nlacF gene (24). The final vector contained 8.5 kb of homologous DNA 5' of the nlacF gene and 700 bp of homologous DNA 3' of the NEO cassette.

Transfected J1 ES cells derived from 129X1/SvJ (revised nomenclature) mice were selected for resistance to G-418, and screened for homologous insertion by PCR analysis using two different pairs of primers. Homologous insertion at the 3' end of the vector was confirmed with a forward primer derived from the NEO sequence and a reverse primer complementary to chromosomal DNA in intron 8, 3' of the Gcdh sequence in the targeting vector, yielding an amplification product of 600 bp. At the 5' end a reverse primer corresponding to the lacZ gene and a forward primer derived from genomic DNA outside of the targeting vector were used to amplify an 8.7 kb fragment specific for the correctly targeted allele. For this reaction PCR was done using the Boehringer Mannheim Expand^TM Long Template PCR System according to the manufacturer’s directions. The 5' PCR product was confirmed to be correct by restriction analysis. Correctly targeted ES cells were injected into C57Bl/6J blastocysts, which were transferred to the uteri of pseudopregnant females for gestation. Chimeric male animals were crossed to C57Bl/6J females and the progeny were screened for the presence of the targeted Gcdh mutation by PCR. The genotypes of mice were checked by a multiplex PCR procedure using the primers for the 3' fusion as described above, with the addition of a forward primer from exon 3 and a reverse primer from exon 5. The amplification product from the exon 3/exon 5 primers is unique to wild-type chromosomes and is 580 bp in length. Transfection of ES cells, blastocyst injection and generation of the chimeric mice were done at Genome Systems (St Louis, MO). Mice used in the experiments described below were from the F2 to F4 129X1/SvJ x C57Bl/6J intercross unless otherwise noted.

Biochemical testing
Quantitative measurement of GA and 3-OHGA in serum and urine was done by stable isotope dilution GC/MS using pentanyldioic-2,2,4,4-d₄ acid (MDS Isotopes, Montreal, Canada) and 3-OH-glutaric-2,2,4,4-d₄ acid (Dr H.Buchel, Marburg, Germany) as internal standards.

ß-Galactosidase staining of fresh tissue
Five chimeric mice that failed to produce carrier offspring and five control mice were killed at 21 months of age by anesthesia with pentobarbital (65 mg/kg), followed by perfusion with normal saline and 2% formaldehyde and 0.2% glutaraldehyde. After post-fixation for 30 min with the same fixative, brains were sliced with a precision brain slicer (Zivic Miller, Zelienople, PA) into 1–2 mm sections and fixed for a further 90 min at room temperature. After a further post-fixation in 0.2% glutaraldehyde for 10 min, sections were rinsed three times for a total of 90 min in a solution containing 2 mM MgCl₂, 0.1% deoxycholate and 0.2% NP-40 and incubated overnight in a solution of 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide, 40 mg/ml X-gal and Tris buffer at pH 7.5 in the dark at 37°C. Blue stain indicative of expression of the Gcdh gene was visualized using a dissecting microscope.

Pathologic analysis
Twenty-nine mice ranging in age from 2 to 18 months were anesthetized with pentobarbital (65 mg/kg) and killed by perfusion with normal saline followed by 10% neutral buffered formalin. Ten of the mice were on the inbred 129 background and 20 were F2 hybrid mice of the 129X1SvJ x C56Bl/6J background. After gross necropsy, tissues were post-fixed in 10% neutral buffered formalin for 2–5 days, embedded in paraffin and sectioned at 4 µm. All sections were stained with Harris hematoxylin and eosin, with liver sections additionally stained using the periodic acid Schiff method, with and without diastase (24). Ultrastructural studies were also conducted on selected liver and brain sections. For these studies, fresh tissues were fixed in 3% glutaraldehyde in phosphate buffer, post-fixed in 1% osmium tetroxide and embedded in epoxy resin. Thin sections were cut and stained with 2% uranyl acetate, followed by lead citrate and examined with an electron microscope. All liver and brain sections were examined for morphologic abnormalities; pathologists were blinded to genotype prior to assessment of sections.

Behavioral testing
General methods. All tests were performed between 12:00 and 16:00 h. All testing equipment was thoroughly cleaned with Odormute (R.C.Steele Company, Brockport, NY), a detergent containing an enzyme that removes odors, and dried between subjects. Mice were kept on a 12 h light–dark cycle with lights on at 07:00 h and were supplied with acidulated water (pH 3) and autoclaved chow (NIH 5K67; Richmond, St Louis, MO) ad libitum. For the rotorod testing at 20 and 60 days of age, all subjects were F2 hybrids produced by heterozygous males and females who were F1 hybrids of C57BL/6J and 129X1/SvJ. All 38 animals in the study were the offspring of one pair of mice and thus were tested in replications approximately 20 days apart. The 74 animals used in the pre-pulse inhibition studies included these 38 F2 hybrid mice, as well as F3 and F4 hybrid mice from subsequent generations, for a total of 27 Gcdh^+/+, 29 Gcdh^–/– and 18 Gcdh^+/– mice ranging in age from 48 to 124 days. Both male and female mice were examined.

Rotorod. Using an accelerating rotorod apparatus (Ugo Basile, Varese, Italy), mice were required to ambulate on a rod (diameter = 6 cm; length = 8.5 cm; grooved surface) to avoid falling 24 cm to a plastic surface below. The rod rotation speed automatically increased steadily from 0.8 to 30 r.p.m. during a 5 min trial, so that the longer the mouse remained on the rod, the faster it rotated. Each mouse was run each day for three trials with 10–15 min between them. For each trial, the mouse was placed on the rod facing the opposite direction of the rotation and was allowed to ambulate for 5 min or until it fell off. The times for each trial were averaged to yield a daily ‘time on rod’ score. Daily training was carried out for 4 days for the mice at 20 days of age, with a 1 day test at 60 days of age.

Rodwalk. Mice that had been tested on the rotorod were evaluated with the rodwalk test at 80 days of age. The procedure used was a modification of the method described by Carter et al. (16). Mice were first trained to run down a 1 m rod that was 28 mm square by placing the mouse part of the way down the rod on the first three trials, and subsequent trials required the mice to run the full 1 m to escape to their holding container. Four trials were given the first day, and by the end of the four trials on the second day, all the mice reached the end of the rod in well under 20 s. On the third day, mice were run for two consecutive trials each on square rods of 28, 12 and 5 mm in width. The time taken to traverse the rod, and the number of times the mice slipped or fell under the rod, were recorded.

Pre-pulse inhibition of the auditory startle response. For all three startle tests, the peak amplitude of the startle response as well as the latency (time to peak amplitude) of the startle response were measured using the San Diego Instruments Startle Response Monitor (SDI, San Diego, CA). Mice were placed in a small chamber to position them for the presentation of the startle stimulus. This chamber was placed on a movement-sensitive platform that was located in a foam-lined acoustic cubicle to isolate sound. The background white noise level in the chamber was 65 dB SPL. The session began with a 5 min acclimation period, after which the mice were exposed to 36 presentations of 120 dB SPL white noise stimuli 50 ms in duration, with a 1 ms rise/fall time and a 10–20 s variable inter-stimulus interval. There were six blocks of six stimuli each: startle stimulus only, and startle stimulus preceded by a 67, 69, 73 or 81 dB pre-pulse stimulus (2, 4, 8, 16 dB over background), presented in a balanced, semi-random order. The pre-pulse stimuli were 30 ms in duration and delivered 100 ms prior to the startle stimuli. The percent inhibition of the startle reflex was calculated for each pre-pulse level as follows: [1 – (startle amplitude after pre-pulse-pulse pair/startle amplitude after pulse only)] x 100.

Short-term habituation of the acoustic startle response. One week after pre-pulse inhibition testing, the last 28 male and female mice tested on pre-pulse inhibition were tested on short-term habituation of the auditory startle response. They ranged in age from 48 to 74 days. After the 5 min acclimation period, mice were exposed to 100 acoustic stimuli (120 dB – day 10 or 100 dB – day 24, white noise, 50 ms duration with a 1 ms rise/fall time) with a fixed inter-stimulus interval (10 s). For each trial, the baseline amplitude, which is the amplitude registered at the onset of the auditory stimulus, was subtracted from the maximum amplitude to control for any movements the animals were making at stimulus onset. The baseline values were typically extremely small compared to the maximum amplitude.

Acoustic startle response sensitivity. One week later, the same 28 mice were tested on startle response sensitivity. After the 5 min acclimation period, 60 white noise startle stimuli, 50 ms in duration, were presented. Against a background of 60 dB white noise, the mice were presented with 10 stimuli at each of six intensities (70, 80, 90, 100, 110, 120 dB) in a balanced, semi-random order with a variable (20–40 s, mean 30 s) inter-stimulus interval to avoid habituation.

Statistical analyses. Data were analyzed by analysis of variance (ANOVA) with genotype as the independent variable and with repeated measures for trial block or stimulus intensity. Significant main effects were analyzed with Newman–Keuls post hoc tests for comparisons among treatment means or least squares difference planned comparison for pairwise comparisons. Statistical analysis was performed using Statistica software.

	ACKNOWLEDGEMENTS

 
We wish to thank Cheryl Peck and Drs Piero Rinaldo and K.Michael Gibson for metabolite analysis, and Dr Gary Mierau for the electron microscropic analyses. This work was supported by National Institutes of Health grants HD 04024, HD 08315 and NS 32841.

	FOOTNOTES

 
⁺ To whom correspondence should be addressed. Tel: +1 503 494 2783; Fax: +1 503 494 2781; Email: koellerd@ohsu.edu

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS REFERENCES

 
1 Besrat,A., Polan,C.E. and Henderson,L.M. (1969) Mammalian metabolism of glutaric acid. J. Biol. Chem., 244, 1461–1467.[Abstract/Free Full Text]

2 Chow,C.W., Haan,E.A., Goodman,S.I., Anderson,R.M., Evans,W.A., Kleinschmidt-DeMasters,B.K., Wise,G., McGill,J.J. and Danks,D.M. (1988) Neuropathology in glutaric acidaemia type 1. Acta Neuropathol. (Berl.), 76, 590–594.[Medline]

3 Soffer,D., Amir,N., Elpeleg,O.N., Gomori,J.M., Shalev,R.S. and Gottschalk-Sabag,S. (1992) Striatal degeneration and spongy myelinopathy in glutaric acidemia. J. Neurol. Sci., 107, 199–204.[Web of Science][Medline]

4 Bjugstad,K.B., Goodman,S.I. and Freed,C.R. (2000) Age at symptom onset predicts severity of motor impairment and clinical outcome of glutaric acidemia type 1. J. Pediatr., 137, 681–686.[Web of Science][Medline]

5 Goodman,S.I. and Frerman,F.E. (1995) Organic acidemias due to defects in lysine oxidation: 2-ketoadipic acidemia and glutaric acidemia. In Scriver,C.R., Beaudet,A.L., Sly,W.S. and Valle,D. (eds), The Metabolic and Molecular Bases of Inherited Disease, 7th edn. McGraw-Hill, New York, Vol. 1, pp. 1451–1460.

6 Baric,I., Zschocke,J., Christensen,E., Duran,M., Goodman,S.I., Leonard,J.V., Muller,E., Morton,D.H., Superti-Furga,A. and Hoffmann,G.F. (1998) Diagnosis and management of glutaric aciduria type I. J. Inherit. Metab. Dis., 21, 326–340.[Web of Science][Medline]

7 Hoffmann,G.F., Athanassopoulos,S., Burlina,A.B., Duran,M., de Klerk,J.B., Lehnert,W., Leonard,J.V., Monavari,A.A., Muller,E., Muntau,A.C. et al. (1996) Clinical course, early diagnosis, treatment and prevention of disease in glutaryl-CoA dehydrogenase deficiency. Neuropediatrics, 27, 115–123.[Web of Science][Medline]

8 Flott-Rahmel,B., Falter,C., Schluff,P., Fingerhut,R., Christensen,E., Jakobs,C., Musshoff,U., Fautek,J.D., Deufel,T., Ludolph,A. et al. (1997) Nerve cell lesions caused by 3-hydroxyglutaric acid: a possible mechanism for neurodegeneration in glutaric acidaemia I. J. Inherit. Metab. Dis., 20, 387–390.[Web of Science][Medline]

9 Kolker,S., Ahlemeyer,B., Krieglstein,J. and Hoffmann,G.F. (1999) 3-Hydroxyglutaric and glutaric acids are neurotoxic through NMDA receptors in vitro. J. Inherit. Metab. Dis., 22, 259–262.[Web of Science][Medline]

10 Ullrich,K., Flott-Rahmel,B., Schluff,P., Musshoff,U., Das,A., Lucke,T., Steinfeld,R., Christensen,E., Jakobs,C., Ludolph,A. et al. (1999) Glutaric aciduria type I: pathomechanisms of neurodegeneration. J. Inherit. Metab. Dis., 22, 392–403.[Web of Science][Medline]

11 Heyes,M.P. (1987) Hypothesis: a role for quinolinic acid in the neuropathology of glutaric aciduria type I. Can. J. Neurol. Sci., 14, 441–443.[Web of Science][Medline]

12 Woontner,M., Crnic,L.S. and Koeller,D.M. (2000) Analysis of the expression of murine glutaryl-CoA dehydrogenase: in vitro and in vivo studies. Mol. Genet. Metab., 69, 116–122.[Web of Science][Medline]

13 Lin,C.H., Tallaksen-Greene,S., Chien,W.M., Cearley,J.A., Jackson,W.S., Crouse,A.B., Ren,S., Li,X.J., Albin,R.L. and Detloff,P.J. (2001) Neurological abnormalities in a knock-in mouse model of Huntington’s disease. Hum. Mol. Genet., 10, 137–144.[Abstract/Free Full Text]

14 Kodsi,M.H. and Swerdlow,N.R. (1997) Mitochondrial toxin 3-nitropropionic acid produces startle reflex abnormalities and striatal damage in rats that model some features of Huntington’s disease. Neurosci. Lett., 231, 103–107.[Web of Science][Medline]

15 Swerdlow,N.R., Paulsen,J., Braff,D.L., Butters,N., Geyer,M.A. and Swenson,M.R. (1995) Impaired prepulse inhibition of acoustic and tactile startle response in patients with Huntington’s disease. J. Neurol. Neurosurg. Psychiatry, 58, 192–200.[Abstract/Free Full Text]

16 Carter,R.J., Lione,L.A., Humby,T., Mangiarini,L., Mahal,A., Bates,G.P., Dunnett,S.B. and Morton,A.J. (1999) Characterization of progressive motor deficits in mice transgenic for the human Huntington’s disease mutation. J. Neurosci., 19, 3248–3257.[Abstract/Free Full Text]

17 Willott,J.F., Carlson,S. and Chen,H. (1994) Prepulse inhibition of the startle response in mice: relationship to hearing loss and auditory system plasticity. Behav. Neurosci., 108, 703–713.[Web of Science][Medline]

18 Busquets,C., Merinero,B., Christensen,E., Gelpi,J.L., Campistol,J., Pineda,M., Fernandez-Alvarez,E., Prats,J.M., Sans,A., Arteaga,R. et al. (2000) Glutaryl-CoA dehydrogenase deficiency in Spain: evidence of two groups of patients, genetically and biochemically distinct. Pediatr. Res., 48, 315–322.[Web of Science][Medline]

19 Goodman,S.I., Stein,D.E., Schlesinger,S., Christensen,E., Schwartz,M., Greenberg,C.R. and Elpeleg,O.N. (1998) Glutaryl-CoA dehydrogenase mutations in glutaric acidemia (type I): review and report of thirty novel mutations. Hum. Mutat., 12, 141–144.[Web of Science][Medline]

20 Graham,D.I. and Lantos,P.L. (1997) Greenfield’s Neuropathology, 6th edn. Arnold, London, UK.

21 Bronson,R.T., Donahue,L.R., Samples,R., Kim,J.H. and Naggert,J.K. (2001) Mice with mutations in the mahogany gene Atrn have cerebral spongiform changes. J. Neuropathol. Exp. Neurol., 60, 724–730.[Web of Science][Medline]

22 Mangiarini,L., Sathasivam,K., Seller,M., Cozens,B., Harper,A., Hetherington,C., Lawton,M., Trottier,Y., Lehrach,H., Davies,S.W. et al. (1996) Exon 1 of the HD gene with an expanded CAG repeat is sufficient to cause a progressive neurological phenotype in transgenic mice. Cell, 87, 493–506.[Web of Science][Medline]

23 Bonnerot,C., Schneider,K.A., Briand,P. and Grimber,G. (1987) A ß-galactosidase hybrid protein targeted to nuclei as a marker for developmental studies. Proc. Natl Acad. Sci. USA, 84, 6795–6799.[Abstract/Free Full Text]

24 Soriano,P., Montgomery,C., Geske,R. and Bradley,A. (1991) Targeted disruption of the c-src proto-oncogene leads to osteopetrosis in mice. Cell, 64, 693–702.[Web of Science][Medline]

25 Prophet,E.B. (ed.) (1994) Laboratory Methods in Histotechnology. American Registry of Pathology, Washington, DC.

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				I. Harting, E. Neumaier-Probst, A. Seitz, E. M. Maier, B. Assmann, I. Baric, M. Troncoso, C. Muhlhausen, J. Zschocke, N. P. S. Boy, et al. Dynamic changes of striatal and extrastriatal abnormalities in glutaric aciduria type I Brain, July 1, 2009; 132(7): 1764 - 1782. [Abstract] [Full Text] [PDF]

				S. Kolker, S. W. Sauer, J. G. Okun, G. F. Hoffmann, and D. M. Koeller Lysine intake and neurotoxicity in glutaric aciduria type I: towards a rationale for therapy? Brain, August 1, 2006; 129(Pt 8): e54 - e54. [Full Text] [PDF]

				W. J. Zinnanti, J. Lazovic, E. B. Wolpert, D. A. Antonetti, M. B. Smith, J. R. Connor, M. Woontner, S. I. Goodman, and K. C. Cheng New insights for glutaric aciduria type I. Brain, August 1, 2006; 129(Pt 8): e55 - e55. [Full Text] [PDF]

				W. J. Zinnanti, J. Lazovic, E. B. Wolpert, D. A. Antonetti, M. B. Smith, J. R. Connor, M. Woontner, S. I. Goodman, and K. C. Cheng A diet-induced mouse model for glutaric aciduria type I Brain, April 1, 2006; 129(4): 899 - 910. [Abstract] [Full Text] [PDF]

				S. W. Sauer, J. G. Okun, M. A. Schwab, L. R. Crnic, G. F. Hoffmann, S. I. Goodman, D. M. Koeller, and S. Kolker Bioenergetics in Glutaryl-Coenzyme A Dehydrogenase Deficiency: A ROLE FOR GLUTARYL-COENZYME A J. Biol. Chem., June 10, 2005; 280(23): 21830 - 21836. [Abstract] [Full Text] [PDF]

				K. A. Strauss Glutaric aciduria type 1: a clinician's view of progress Brain, April 1, 2005; 128(4): 697 - 699. [Full Text] [PDF]

				H. S. Toogood, A. van Thiel, J. Basran, M. J. Sutcliffe, N. S. Scrutton, and D. Leys Extensive Domain Motion and Electron Transfer in the Human Electron Transferring Flavoprotein{middle dot}Medium Chain Acyl-CoA Dehydrogenase Complex J. Biol. Chem., July 30, 2004; 279(31): 32904 - 32912. [Abstract] [Full Text] [PDF]

				S. Kolker, G.F. Hoffmann, O. Bahr, and J.B. Schulz Adult onset glutaric aciduria type I presenting with a leukoencephalopathy Neurology, April 22, 2003; 60(8): 1399 - 1399. [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (39) Request Permissions Google Scholar Articles by Koeller, D. M. Articles by Goodman, S. I. Search for Related Content PubMed PubMed Citation Articles by Koeller, D. M. Articles by Goodman, S. I. Social Bookmarking          What's this?

Online ISSN 1460-2083 - Print ISSN 0964-6906

Copyright © 2009 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics