J Autism Dev Disord (2012) 42:105–115 DOI 10.1007/s10803-011-1216-y
ORIGINAL PAPER
Proton Magnetic Resonance Spectroscopy and MRI Reveal No Evidence for Brain Mitochondrial Dysfunction in Children with Autism Spectrum Disorder Neva M. Corrigan • Dennis. W. W. Shaw • Todd L. Richards • Annette M. Estes • Seth D. Friedman Helen Petropoulos • Alan A. Artru • Stephen R. Dager
•
Published online: 15 March 2011 Ó Springer Science+Business Media, LLC 2011
Abstract Brain mitochondrial dysfunction has been proposed as an etiologic factor in autism spectrum disorder (ASD). Proton magnetic resonance spectroscopic imaging (1HMRS) and MRI were used to assess for evidence of brain mitochondrial dysfunction in longitudinal samples of children with ASD or developmental delay (DD), and cross-sectionally in typically developing (TD) children at 3–4, 6–7 and 9–10 years-of-age. A total of 239 studies from 130 unique participants (54ASD, 22DD, 54TD) were acquired. 1HMRS and MRI revealed no evidence for brain mitochondrial dysfunction in the children with ASD.
These data were presented in preliminary form at a trans-NIH conference, Mitochondrial Encephalopathies: Potential Relationships to Autism? Indianapolis, IN, 29 June 2008. N. M. Corrigan Dennis. W. W. Shaw T. L. Richards H. Petropoulos S. R. Dager (&) Department of Radiology, University of Washington, Seattle, WA, USA e-mail:
[email protected] Dennis. W. W. Shaw S. D. Friedman Seattle Children’s Hospital, Seattle, WA, USA A. A. Artru Department of Anesthesiology, University of Washington, Seattle, WA, USA A. M. Estes S. R. Dager University of Washington Autism Center, Seattle, WA, USA A. M. Estes Department of Speech and Hearing Sciences, University of Washington, Seattle, WA, USA S. R. Dager Department of Bioengineering, University of Washington, Seattle, WA, USA
Findings do not support a substantive role for brain mitochondrial abnormalities in the etiology or symptom expression of ASD, nor the widespread use of hyperbaric oxygen treatment that has been advocated on the basis of this proposed relationship. Keywords Autism Developmental disorders MRS MRI Mitochondrial disorders Brain metabolism Lactate
Introduction Brain mitochondrial dysfunction has been implicated in a number of neurodegenerative disorders (Calabrese et al. 2001; Mancuso et al. 2006) and has been proposed to be a contributing factor to the pathogenesis of autism spectrum disorder (ASD) (Lombard 1998). Evidence in support of this association has largely been based on peripheral markers of mitochondrial dysfunction or abnormal mitochondrial DNA (mtDNA) sampled from body compartments outside the brain, such as blood or skeletal muscle. The association was first suggested by a report of four patients with ASD exhibiting blood lactic acidosis (Coleman and Blass 1985). A small number of subsequent case studies and case series investigations reported the occurrence of autistic behavior in children with mitochondrial abnormalities (Filiano et al. 2002; Filipek et al. 2003; Graf et al. 2000; Poling et al. 2006; Pons et al. 2004). From a study of blood lactate levels obtained from 69 children with ASD, and skeletal muscle biopsies from a subsample of 11 of these children who exhibited higher lactate, mitochondrial dysfunction was suggested to be one of the most common medical conditions associated with autism, with an estimated prevalence rate of 7% (Oliveira et al. 2005). A recent study that tested lymphocytes from blood samples of
123
106
ten children with autistic disorder reported a high occurrence of mitochondrial dysfunction and increased mtDNA abnormalities compared to controls, although as only one child fulfilled stringent diagnostic criteria for a mitochondrial respiratory chain disorder, the authors suggested a prevalence rate of 10% (Giulivi et al. 2010). Additionally, a recent literature review and meta-analysis of reports that assessed for mitochondrial dysfunction across a wide range of possible markers concluded that whereas the prevalence rate for severe mitochondrial disorder was approximately 5% within the general population of ASD, a much higher proportion likely exhibited a spectrum of mitochondrial dysfunction of varying severity and without discernable mtDNA abnormalities (Rossignol and Frye 2011). One limitation, however, to assessing links between ASD and mitochondrial dysfunction through evaluation of peripheral markers in body compartments outside the brain is that phenotypic expression is tightly linked to the location of impaired mitochondrial function within the body. Further, unlike nuclear DNA, abnormal mtDNA can coexist with normal mtDNA within the same cell, and there may be substantially different loading of abnormal mtDNA between tissue compartments. Since the highest levels of abnormal mtDNA and altered bioenergetics occur in low mitotic tissue with high energy demands (Cortopassi et al. 1992), the brain is one of the most commonly affected organs in mitochondrial disorder. Since ASD is generally considered a neurodevelopmental disorder, if mitochondrial dysfunction contributes to the etiology or primary symptom expression of ASD, evidence for mitochondrial dysfunction would be expected specifically in brain tissue. Neuroimaging techniques, primarily single-voxel and multi-voxel proton magnetic resonance spectroscopy (1H MRS) and MRI, are used clinically to non-invasively assess chemical and structural abnormalities associated with mitochondrial dysfunction in the brain (‘‘Mitochondrial Encephalopathies: Potential Relationships to Autism?’’ 2008). Primary evidence for brain bioenergetic disturbances associated with impaired mitochondrial function, most notably elevation of brain lactate, can be detected with 1H MRS (Castillo et al. 1995; Dager et al. 2004; Kapeller et al. 1996; Kuwabara et al. 1994; Mathews et al. 1993). Elevations of brain lactate occur when mitochondrial dysfunction results in a shift in cellular ATP production from oxidative phosphorylation towards anaerobic glycolysis. On MRI, a variety of characteristic brain structural abnormalities are observed with mitochondrial disorder (Finsterer 2009; Saneto et al. 2008). The widespread availability of these noninvasive imaging tools provides a valuable means of screening for brain-specific abnormalities that may reflect mitochondrial dysfunction in ASD. Several of the previously cited case series that assessed muscle biopsies and/or blood samples also collected MRI
123
J Autism Dev Disord (2012) 42:105–115
(Filiano et al. 2002; Filipek et al. 2003; Graf et al. 2000; Poling et al. 2006; Pons et al. 2004), with two reporting MRI abnormalities. MRI abnormalities were found in two of twelve patients with mitochondrial abnormalities and autistic behavior (Filiano et al. 2002), with cerebral atrophy, as evidenced by widened sulci and ventricles and thinned centrum semiovale, observed in a 2-year-old boy and extensive multifocal patches of white matter signal intensity observed in a 20-year-old male. In a study of 5 children with ASD or autistic features and a family history of mtDNA diseases (Pons et al. 2004), 4 of the children exhibited abnormal mtDNA or reduced respiratory chain complex activity. Of these children, a 5-year old girl exhibited abnormal signal intensities in the cerebellum, brainstem and thalamus, as well as agenesis of the corpus callosum, and a 5-year old boy exhibited abnormal MRI signal intensities bilaterally in the caudate and pallidum, as well as elevated lactate by 1H MRS. To date, two published studies have used 1H MRS to systematically measure brain lactate in children with ASD. One study of nine children with ASD ranging in age from 3 to 12 years, selected for study on the basis of elevated blood lactate levels, reported elevated brain lactate in the frontal lobe of one child (Chugani et al. 1999). A study by our group of forty-five 3–4 year-old children with ASD found no evidence of elevated lactate (Friedman et al. 2003). To our knowledge, there have been no systematic investigations of 1H MRS and structural MRI alterations associated with mitochondrial dysfunction in children with ASD. In this study, multi-voxel 1H MRS and MRI were used to evaluate for evidence of brain mitochondrial dysfunction in children with ASD compared to children with idiopathic developmental delay (DD) and typical development (TD) at 3–4, 6–7, and 9–10 years-of-age. Volumetric MRI and 1 H MRS data from the 3–4 year-old age-point were previously reported (Friedman et al. 2003, 2006a; Sparks et al. 2002). The primary aims were to systematically investigate mitochondrial dysfunction in relation to age in a sample of children with ASD compared to children with DD or TD by: (1) employing a recently developed multi-voxel 1H MRS analytic method (Corrigan et al. 2010) to measure lactate concentrations both diffusely and within specific brain compartments from 1H MRS data collected at each age-point and (2) examination, blinded to diagnosis, of T1 and T2-weighted MRI images collected at each time point to assess for structural abnormalities associated with mitochondrial disorders. A secondary aim was to identify individual children at each age-point exhibiting brain lactate levels at the upper range of the sample distribution which might reflect occult metabolic imbalance, and to heuristically assess whether any of the identified children in the ASD group demonstrated atypical behavioral or clinical features.
J Autism Dev Disord (2012) 42:105–115
107
Table 1 Demographic information for participants with usable data, by age and diagnostic group Group
3 years
6 years
9 years
N
Gender
Mean age in years (SD)
N
Gender
ASD
45
38 m/7f
4.0 (0.35)
35
28 m/7f
6.6 (0.39)
29
24 m/5f
9.7 (0.48)
DD
14
6 m/8f
4.0 (0.47)
14
9 m/5f
6.4 (0.35)
15
10 m/5f
9.5 (0.29)
TD
14 N = 73
11 m/3f
3.7 (0.49)
20 N = 69
12 m/8f
6.6 (0.43)
33 N = 77
29 m/4f
9.7 (0.39)
Methods Participants Seventy-three children (45 ASD, 14 DD, 14 TD) at 3–4 years-of-age, 69 children (35 ASD, 14 DD, 20 TD) at 6–7 years-of-age, and 77 children (29 ASD, 15 DD, 33 TD) at 9–10 years-of-age were scanned. Demographic information for children with usable data is shown in Table 1. At the 3–4 year age-point, TD participants compared to participants with ASD had significant age differences (F = 3.5, p \ 0.04); there were no significant group differences in age at the other age-points. Group differences in gender were demonstrated at age 3 using Pearson chisquared (chi-squared = 10.0, p \ .01), a result of more girls in the DD group (57%) relative to ASD (16%) and TD groups (21%); there were no significant group gender differences at other age-points. A subset of children at each age did not have usable 1H MRS data for lactate measurement (3–4 years: 4 TD, 4 ASD; 6–7 years: 3 TD, 1 ASD, 4 DD; 9–10 years: 10 TD). Four additional TD children studied did not have usable MRI or 1H MRS data. Data from a total of 130 unique children across all 3 age-points (54 ASD, 22 DD, 54 TD) were included in these analyses. Children in the ASD and DD groups were primarily evaluated longitudinally at 3–4, 6–7 and 9–10 years-of-age. Participant dropout and technical difficulties resulted in the assessment of some children in these diagnostic groups at only 1 or 2 age-points. Additional children from a larger subject pool followed behaviorally from age 3 were added to each diagnostic group at the later age-points to partially compensate for participant attrition. The TD group was primarily assessed cross-sectionally, with a smaller subset evaluated longitudinally. Usable data for more than one age-point were available from 65% of the children in the ASD group, 55% of the children in the DD group and 20% of the children in the TD group. Children were recruited from local parent advocacy groups, the University of Washington Infant and Child Subject Pool, public schools, clinics, hospitals and the Department of Developmental Disabilities. Exclusion
Mean age in years (SD)
N
Gender
Mean age in years (SD)
criteria included: identifiable genetic abnormalities (including Fragile X syndrome, Norrie syndrome, neurofibromatosisis, tuberous sclerosis, phenylketonuria), cerebral vascular disease, severe sensory or motor impairments, significant pulmonary disease, unstable cardiovascular status, major physical abnormalities, medically significant perinatal difficulties that required hospitalization, such as prematurity or intrauterine growth retardation, documented pre- or post-natal head trauma, and implanted medical prostheses or other devices incompatible with MR scanning. Written parental/guardian informed consent to participate in the study, approved by the University of Washington Internal Review Board, was obtained from each participant. Diagnosis of children with ASD utilized the Autism Diagnostic Interview-Revised (ADI-R, (Lord et al. 1994) and all participants were assessed with the Autism Diagnostic Observation Schedule-Generic (ADOS-G, (Lord et al. 2000) and by an experienced clinician using DSM-IV diagnostic criteria (American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders 1994), as previously described (Friedman et al. 2003; Petropoulos et al. 2006; Sparks et al. 2002). The ASD sample included children with a range of intellectual abilities. Diagnoses was determined at the time of the first MRI scan for each child and confirmed at subsequent age-points. Study Procedures Children in the ASD and DD groups were imaged using continuous IV infusion of propofol, as previously described (Amundsen et al. 2005). Children in the TD group were studied late at night during natural sleep. All imaging data were acquired on the same 1.5 Tesla GE Signa Horizon whole body MR scanner (v.5.8 Genesis software) using a custom-built receive-only linear birdcage radiofrequency head coil. High-resolution coronal T1weighted images were acquired for all participants using a three-dimensional spoiled gradient-recalled (SPGR) imaging sequence (TR = 33.3 ms, TE = 30 ms, flip angle = 30°, FOV = 22 cm, 256 9 256 matrix, 1.5 mm slice thickness-zero-filled from 3 to 1.5 mm during acquisition at the 3 year age-point). Axial proton density (PD) and
123
108
J Autism Dev Disord (2012) 42:105–115
filtering (Guillemaud 1998). The corrected images were then classified using a k-means algorithm to produce binary gray matter, white matter, and CSF images and averaged together. These images were filtered to produce final images with a point-spread function approximating that of the spectroscopic images (Friedman et al. 2006b). Spectroscopic Data Analysis
Fig. 1 Example MRS slab locations for the 3–4 and 6–7 year agepoints. For the 9–10 year age-point, the single slab was placed at a position that coincided with the middle of the volume of acquisition for the other two age-points
Spectroscopic data were reconstructed and pre-processed with the LCModel software package (v.6.2, (Provencher 1993) and lactate concentrations calculated for the entire acquired PEPSI volume, as well as specific brain compartments, as described previously (Corrigan et al. 2010). An example spectrum averaged from the acquired PEPSI volume for one child with ASD studied at each age-point, with superimposition of the LCModel fit, is shown in Fig. 2. The gray and white matter compartments included voxels comprised of 60% or more of the respective tissue type, based on the segmented structural images. As
T2-weighted images covering the whole brain were also acquired (TE = 13/91 ms, TR = 2,000 ms, FOV = 22 cm, matrix 256 9 160, 2.5 mm slice thickness). Proton echo-planar spectroscopic imaging (PEPSI, (Posse et al. 1997) was used for rapid acquisition of multivoxel MRS data. Unsuppressed water (TE = 20 ms) and water-suppressed (TE = 272/20 ms) PEPSI scans (TR = 2,000 ms, 32 9 32 spatial matrix, nominal voxel size = 1cm3, FOV = 22 cm) were acquired. For the 3–4 and 6–7 year age-points, PEPSI data were collected from 2 contiguous axial sections, each 20 mm thick, centered on the anterior commissure and through the basal ganglia, as shown in Fig. 1. Acquisition time for each PEPSI slab was approximately 20 min. For the 9–10 year age-point, in order to increase compliance and maximize usable data acquired from children in the TD group, the scan duration was decreased (by approximately 20 min) by acquiring PEPSI data from a single axial section through the basal ganglia, at a level falling midway between slab locations for the earlier age-points. Slabs were placed by visual inspection and, when available, guided by images acquired at previous age-points. The long-echo (TE = 272 ms) metabolite data and unsuppressed water scans were used for lactate quantification (Corrigan et al. 2010). Structural Image Segmentation PD and T2-weighted images were segmented to produce gray matter, white matter and cerebral spinal fluid (CSF) images. This was performed by first correcting the PD and T2-weighted images that corresponded to the location of the PEPSI slab for RF inhomogeneity using homomorphic
123
Fig. 2 Spectra for the whole slab region from a single participant with ASD followed longitudinally. The spectrum for each age-point is shown (black line) with the LCModel fit superimposed (red line)
J Autism Dev Disord (2012) 42:105–115
109
another at age 6 who had lactate levels [3SD above the mean for all children in the same age group (all p values [ 0.091). For heuristical purposes, children with lactate levels[2SD above the mean of all children in the same age group were identified. The clinical and behavioral histories of ASD children with these higher lactate values were reviewed to identify any atypical clinical or behavioral features. Structural MRI Assessment
Fig. 3 Example gray matter (green), white matter (blue) and ventricular (yellow) regions for a single participant superimposed on the participant’s structural image
elevated lactate associated with mitochondrial disorder may have the highest yield in CSF (Bianchi et al. 2007; Lin et al. 2003), voxels in or surrounding the ventricles were included in a CSF compartment. Example PEPSI voxel locations for the gray matter, white matter, and CSF compartments for an individual participant are shown in Fig. 3, superimposed on the participant’s T2-weighted structural image. Statistical Analyses Statistical analyses were performed with PASW Statistics 18.0 (SPSS Inc, Chicago, IL), using a two-tailed significance level of 0.05. At individual age-points, mean and standard deviation of lactate concentrations were calculated across all participants for each brain compartment and diagnostic group. One-way ANOVA was used to test for significant differences in lactate concentration between the three diagnostic groups for each brain compartment and age-point, as well as all age-points combined by averaging lactate concentrations across multiple time-points for individual subjects. Two-way repeated measures ANOVA with GeisserGreenhouse correction was performed on the data from children in the ASD and DD groups for which MRS data were available at all 3 age-points (ASD = 17; DD = 9) to assess for lactate changes over time, or differences in changes over time by diagnostic group. For each brain compartment and age group, the Shapiro–Wilk test confirmed that lactate concentrations followed a normal distribution, with the exclusion of one TD outlier at age 3 and
T1 and T2-weighted structural MR images acquired at each age-point were systematically assessed by a board-certified pediatric neuroradiologist who was blinded to diagnosis (D.W.W.S.). T2 images at one age-point were not available for 7 TD participants. Images were evaluated for any evidence of cerebral injury that could be related to mitochondrial disorder. Specifically, cerebral sulcal prominence, ventriculomegaly and prominence of cerebellar folia were evaluated for evidence of volume loss. Evidence for focal injury was also evaluated, including the presence of parenchymal T2 hyperintensity in the basal ganglia, cerebral white matter, brainstem and cerebellum.
Results Mean lactate concentrations for diagnostic groups at each of the 3 age-points, and for the combined age group, for each brain compartment are shown graphically in Fig. 4. These values are detailed in Table 2, along with the corresponding one-way ANOVA F-values and post-hoc p-values Bonferroni–corrected for multiple comparisons. No significant differences in lactate concentration were found between diagnostic groups for any brain compartment at any age-point, or when combined across all agepoints. Repeated measures assessment of age effects and ageby-diagnosis interactions for the participants with ASD and DD for whom data were collected at 3 age-points revealed no significant age or age-by-diagnosis interactions for any of the brain regions (all F values \ 0.04 for time; F \ 0.02 for diagnosis by time). Lactate concentrations [2SD above the mean of all participants in the corresponding age-group for at least one brain compartment were observed for 2 ASD and 3 TD participants at age 3–4, 2 ASD and 2 TD participant at age 6–7, and 1 ASD, 1 DD and 2 TD participants at age 9–10. No participant had lactate levels [2SD above the mean at more than one age-point. Behavioral characteristics of children with ASD having brain lactate values [2SD are detailed in Table 3. An extensive review of behavioral
123
110
J Autism Dev Disord (2012) 42:105–115
Fig. 4 Mean lactate concentrations for the whole slab, gray matter, white matter, and ventricular regions for the 3, 6 and 9 years-old age-points, as well as for all ages combined, by diagnostic group. Black error bars indicate one standard deviation from the mean
Table 2 Mean (SD) lactate concentrations for all participants by age, diagnostic group and brain region, with corresponding one-way ANOVA F value and significance Age
3
6
9
Combined
Lactate concentration (mM) ROI
ASD
DD
TD
F value
F sig
Whole slab
0.725 (0.156)
0.640 (0.209)
0.737 (0.302)
1.144
0.325
Gray matter
0.734 (0.205)
0.645 (0.270)
0.741 (0.337)
0.788
0.459
White matter Ventricular
0.661 (0.182) 0.625 (0.228)
0 562 (0.217) 0.588 (0.165)
0.742 (0.266) 0.700 (0.371)
2.408 0.548
0.098 0.581
Whole slab
0.678 (0.210)
0.688 (0.169)
0.704 (0.199)
0.099
0.906
Gray matter
0.664 (0.247)
0.658 (0.224)
0.792 (0.277)
1.658
0.199
White matter
0.614 (0.200)
0.665 (0.208)
0.663 (0.208)
0.465
0.630
Ventricular
0.610 (0.320)
0.515 (0.262)
0.566 (0.282)
0.507
0.605
Whole slab
0.662 (0.208)
0.646 (0.293)
0.672 (0.263)
0.042
0.959
Gray matter
0.578 (0.283)
0.561 (0.274)
0.715 (0.340)
1.576
0.215
White matter
0.621 (0.240)
0.586 (0.275)
0.580 (0.303)
0.162
0.851
Ventricular
0.627 (0.397)
0.703 (0.375)
0.730 (0.347)
0.499
0.610
Whole slab
0.689 (0.127)
0.658 (0.184)
0.698 (0.246)
0.310
0.734
Gray matter
0.675 (0.179)
0.620 (0.237)
0.747 (0.306)
2.054
0.133
White matter
0.627 (0.140)
0.588 (0.177)
0.646 (0.263)
0.563
0.571
Ventricular
0.628 (0.262)
0.615 (0.227)
0.671 (0.323)
0.382
0.683
observations, clinical course and medical history revealed no consistent clinical characteristics or atypical medical or behavioral features for these children with ASD. Review of MRI images revealed no global or focal atrophy, abnormal basal ganglia signal, or evidence of gliosis to suggest residue of parenchymal injury commonly associated with mitochondrial disorder in the ASD or TD groups. One child in the TD group had a single punctate
123
non-specific focus of T2 hyperintensity in the cerebral white matter. Two DD children were noted to have structural abnormalities including one with periatrial and frontal white matter T2 hyperintensities and a small cerebellar vermis, stable between 3 and 9 years-of-age. A second DD subject imaged only at 6 years had periatrial white matter hyperintensity and volume loss involving cerebellar vermis and hemispheres.
J Autism Dev Disord (2012) 42:105–115
111
Table 3 Clinical measures for participants with ASD who exhibited higher brain lactate values at any age-point ASD participant ASD subgroup Age of interest Regression (0–10) Mullen scales (age 3) Std sum
V
DAS (age 6)
NV
GCA V
DAS (age 9) NV
GCA V
NV
SP
1
AD
3
1 (social)
97
98
97
105
92 100 115
81 151 107
2
PDD-NOS
3
0
96
105
87
116
120 109 108
100 116 106
3
PDD-NOS
6
0
81
73
89
99
105
116 112 104
4
AD
6
0
52
49
56
55
57
59
62
51
76
73
5
PDD-NOS
9
0
51
52
49
73
69
88
78
74
77
93
97 113
ASD abbreviations—AD autistic disorder, PDD-NOS pervasive developmental disorder not otherwise specified Age of interest: age at which child had a brain lactate value exceeding twice the standard deviation above the mean of all children studied at that age Regression: sum of regression scores (each on a scale of 0–2) across 5 domains as reported on ADI at 3–4 years. Mullen/DAS Abbreviations—DAS differential ability scales, Std Sum standard summary score, V verbal, NV non-verbal, GCA general conceptual ability, SP spatial performance
Discussion This longitudinal investigation of children with well-characterized ASD and DD, in comparison to children with TD, revealed no evidence for abnormal brain lactate elevations or structural MRI abnormalities indicative of mitochondrial dysfunction in any child with ASD at any age-point. There also were no diagnostic group differences in brain lactate levels, nor a shift in the range of individual lactate levels within the ASD group, that might reflect mitochondrial compromise. The structural abnormalities observed in 2 DD subjects likely reflect residua of prenatal and/or perinatal insult, although white matter abnormalities, usually extensive or confluent, and cerebellar volume loss, are seen in some cases of respiratory chain defects. These findings extend previous work using 1H MRS that evaluated these children at age 3–4 (Friedman et al. 2003) utilizing a less sensitive analytic approach for detecting brain lactate (Corrigan et al. 2010). Additionally, brain lactate levels were observed to remain stable between 3–4 and 9–10 years-of-age. The children with ASD who were included in this multiyear study represent the largest sample evaluated to date using MRI and MRS to systematically assess for evidence of mitochondrial dysfunction. The longitudinal design and narrowly defined age-groupings allowed for repeated assessments over a critical time course of development. Sampling at multiple time-points increased the likelihood of detecting an intermittent abnormality, since mitochondrial disorders can exhibit periods of quiescence during which metabolic evidence of dysfunction may be difficult to detect. Analyses of longitudinal data from the ASD and DD groups also allowed us to determine that brain lactate levels did not increase over time. Additionally, systematic assessment of longitudinal MRI by an expert observer blinded to diagnosis did not reveal either cross-sectional or
progressive structural evidence of injury from mitochondrial compromise. If recent population prevalence estimates (Oliveira et al. 2005; Giulivi et al. 2010; Rossignol and Frye 2011) held true for the brain, approximately 3–6 children in the ASD group would have been expected to exhibit brain metabolic and structural evidence for a brain mitochondrial disorder, and a much greater proportion would be expected to exhibit more subtle brain metabolic or structural alterations in the setting of compromised mitochondrial function, as a forme fruste of the disorder. The children with ASD who exhibited [2SD higher brain lactate concentrations (two children at age 3–4, two children at age 6–7 and one child at age 9–10) did not exhibit any clinical or behavioral features that distinguished them from the other children in the ASD group. Moreover, a similar proportion of children with DD and TD also exhibited [2SD higher lactate levels across all age-points (ASD = 4.8%; DD = 2.6%; TD = 10.0%), suggesting that these children with ASD do not represent specific cases of altered oxidative metabolism unique to ASD. The most parsimonious explanation for lactate elevations [3SD above the mean of all children observed in two TD children is subtle hyperventilation during the scan. This consideration does highlight that although 1H MRS provides sensitive measurement of lactate levels localized to the brain, lactate elevations are not specific to mitochondrial dysfunction and can result from acute physiological alterations, such as hyperventilation (Dager et al. 1995; Friedman et al. 2007; Posse et al. 1997) or following caffeine ingestion (Dager et al. 1999). In this study, the use of propofol sedation allowed us to image the children with ASD and DD at rest, which minimized the effects of emotional or stress-driven physiological changes on lactate levels in these groups. It is possible that 1H MRS failed to detect subtle alterations of brain energy metabolism
123
112
between diagnostic groups, but the sensitivity of 1H MRS for detecting subtle lactate elevations due to oxidative metabolic abnormalities in bipolar disorder has been demonstrated using data acquisition procedures similar to those used in the present study (Dager et al. 2004). Although, the use of propofol could potentially produce physiological alterations in the respiratory chain, this would be expected to accentuate or unmask occult mitochondrial dysfunction, if present, in the children with ASD or DD. In consideration of this, it is important to note that the range of measured lactate values were relatively consistent across all three diagnostic groups. Compared to more commonly measured metabolites with strong 1H MRS signals, brain lactate is challenging to detect by 1H MRS, and basal lactate levels are generally considered to be at or below the detection limit when using conventional 1H MRS acquisition and analytic approaches (Dager et al. 2008). For this reason, normative brain lactate values for adults have been highly variable across reported studies. In many investigations, detection of a 7 Hz doublet centered at 1.3 ppm that is clearly distinguishable from the background noise has been considered to represent an abnormal elevation of brain lactate, possibly indicative of a pathological condition (Bonavita et al. 1999; Burtscher and Holta˚s 2001; Castillo et al. 1995; Dinopoulos et al. 2005; Lin et al. 2003). In studies where lactate has been quantified in healthy adults, reported values have ranged from approximately 0.2–1.0 mM (de Graaf 2007), reflecting both individual subject variance as well as variability caused by differences in spectral acquisition (such as single-voxel versus multi-voxel acquisition) and processing techniques across studies. We have previously reported mean resting brain lactate concentration of approximately 0.86 mM in healthy adults (Corrigan et al. 2010; Dager et al. 2004). In comparison, lactate concentrations in normal-appearing gray matter of adults with mitochondrial disorder (MELAS) have been reported to be in the range of 3.0 mM (Mo¨ller et al. 2005). In healthy infants, brain lactate concentrations have been reported to be in the range of 0.8 mmol/kg (Kreis et al. 2002). A cross-sectional study of children and adults ranging from less than 1 year to 39 years of age reported gray matter lactate concentrations between 0.4 and 0.5 mM, with little variation in lactate concentrations with age (Pouwels et al. 1999). In the current report, steps were taken to improve the detection sensitivity and measurement accuracy of brain lactate by using a novel analytic technique that maximized spectral SNR for lactate measurement within regions of interest (Corrigan et al. 2010). Additionally, an echo time of 272 ms was selected for data acquisition which, due to the j-coupling properties of lactate, resulted in a positive and in-phase lactate doublet (Dager and Steen 1992). This echo time also reduced the contribution of macromolecules
123
J Autism Dev Disord (2012) 42:105–115
and lipids to the spectroscopic signal in the chemical shift frequency range of the lactate peak. Most studies implicating mitochondrial dysfunction in children with ASD have been based on examination of peripheral blood or tissue samples that do not directly reflect metabolic processes in the brain. Moreover, in pediatric populations elevated blood lactate, for example, can result from spurious measurements due to use of a tourniquet or a struggling child (Haas et al. 2007). Biochemical testing of peripheral tissues is also prone to false positives due to problems with proper storage and handling of tissue samples (‘‘Mitochondrial Encephalopathies: Potential Relationships to Autism?’’ 2008). There is also a high degree of variability in findings for the sample tissue sample across laboratories and there is not a recognized ‘gold standard’ for laboratory-based diagnosis of mitochondrial dysfunction (Gellerich et al. 2004; Thorburn and Smeitink 2001). Since data from this study were collected over approximately 10 years, subject attrition was inevitable and, as a result, these analyses include both longitudinal as well as cross-sectional data. Additionally, the TD children were the most likely to be studied at only one age-point due to difficulties with retaining these children in the study or their ability to remain sleep during data acquisition. Although the total numbers of unique children scanned between ASD and TD diagnostic groups were equivalent, the ASD group had substantially more studies performed overall (109 ASD scans as compared to 67 TD scans). The large number of scans for the ASD group improved the likelihood of detecting mitochondrial-associated abnormalities for at least one age-point, if intermittently expressed, as compared to the TD and DD diagnostic groups. A limitation of including both cross-sectional and longitudinal data in analyzing each age-point is the potential of auto-correlation across age-points for individuals with multiple scans. However, this auto-correlation would tend to accentuate the likelihood of detecting mitochondrial-associated abnormalities across age-points if there were affected children in the ASD or DD groups, which we did not find. Although 1H MRS brain lactate elevation is a sensitive, non-invasive biomarker for brain mitochondrial dysfunction, other 1H MRS chemical measures, such as glutamate (Dager et al. 2004) or glutathione (Meister 1995), may additionally be altered in association with mitochondrial dysfunction and provide complementary indices of brain metabolic alterations. However, these measurements require different pulse sequences and analytic strategies (Dager et al. in press), as well as an increase in data acquisition time. Other multinuclear MRS techniques may also prove useful for investigating bioenergetic alterations associated with mitochondrial dysfunction. For example, in
J Autism Dev Disord (2012) 42:105–115
keeping with some earlier PET studies of ASD that found evidence for wide-spread brain hypermetabolism (De Volder et al. 1987; Horwitz et al. 1987; Rumsey et al. 1985), a preliminary report using phosphorus (31P) MRS to measure high energy phosphates and membrane phospholipids in 11 adolescents and young adults with ASD found decreased PCr, as well as reduced esterified ends (e.g., the alpha phosphates of ATP and ADP) in the frontal lobe, also interpreted as being consistent with brain hypermetabolism (Minshew et al. 1993). Although brain hypermetabolism is opposite in direction to what would be expected with mitochondrial dysfunction, 31P MRS findings of reduced PCr could, conversely, reflect decreased high-energy substrate production secondary to mitochondrial dysfunction. To date, there have been no additional published 31P MRS studies of ASD to follow-up on that initial report. More recent work using 31P MRS to measure brain pH shifts associated with metabolic alterations (Friedman et al. 2006b) and rapidly interleaved 31P-1H MRS techniques at high-field to characterize the effects of hypocapneainduced lactate changes on high-energy phosphates (Friedman et al. 2007) suggest that 31P MRS has the potential to broaden clinical investigations of brain bioenergetic status in ASD. However, the specificity and sensitivity of 31P MRS for detecting high-energy phosphate alterations associated with mitochondrial dysfunction, as well the range of normal values in the absence of mitochondrial dysfunction, requires further assessment. In summary, no 1H MRS or MRI evidence for brain mitochondrial dysfunction, or shifts in brain oxidative metabolism, were observed in this sample of children with ASD at 3–4, 6–7 or 9–10 years-of-age. Our findings do not support the widespread and increasingly common use of hyperbaric oxygen to treat ASD, advocated on the basis of this presumed relationship. Since ASD is a heterogeneous disorder, and the true prevalence of mitochondrial disorders in the general population is unknown (Kirby et al. 1999), it is probable that brain mitochondrial disorder or less severe mitochondrial dysfunction occurs in some individuals with ASD. However, our findings suggest that the occurrence rate for brain involvement in ASD must be relatively rare, and much lower than recent literature estimates of 5–10% or higher from peripheral markers, as no child in this group of 54 children with ASD exhibited 1H MRSI or MRI evidence for brain mitochondrial disorder, or more subtle mitochondrial-related alterations in brain energy metabolism. Re-examination of individual children within the ASD group at multiple age-points also did not reveal any evidence for intermittent nor progressive mitochondrial dysfunction. Not addressed in this study is the possibility that a time-limited expression of mitochondrial dysfunction, not discernable at later ages in the developmental
113
progression, may precede or be concurrent with initial symptom expression, which can occur as early as 12–18 months-of-age in children subsequently diagnosed with ASD (Zwaigenbaum 2010). Acknowledgments This study was supported by NIH grants 2P01 HD 35465, 1P50 HD 55782 and 1R01 HD 065283. The authors thank Denise Echelard and Marie-Anne Domsalla for their technical and administrative support.
References American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4 edn.). Washington, DC: American Psychiatric Association. Amundsen, L. B., Artru, A. A., Dager, S. R., Shaw, D. W. W., Friedman, S., Sparks, B., et al. (2005). Propofol sedation for longitudinal pediatric neuroimaging research. Journal of Neurosurgical Anesthesiology, 17(4), 180–192. Bianchi, M. C., Sgandurra, G., Tosetti, M., Battini, R., & Cioni, G. (2007). Brain magnetic resonance in the diagnostic evaluation of mitochondrial encephalopathies. Bioscience Reports, 27(1–3), 69–85. Bonavita, S., Di Salle, F., & Tedeschi, G. (1999). Proton MRS in neurological disorders. European Journal of Radiology, 30(2), 125–131. Burtscher, I. M., & Holta˚s, S. (2001). Proton MR spectroscopy in clinical routine. Journal of Magnetic Resonance Imaging, 13(4), 560–567. Calabrese, V., Scapagnini, G., Giuffrida Stella, A. M., Bates, T. E., & Clark, J. B. (2001). Mitochondrial involvement in brain function and dysfunction: Relevance to aging, neurodegenerative disorders and longevity. Neurochemical Research, 26(6), 739–764. Castillo, M., Kwock, L., & Green, C. (1995). MELAS syndrome: Imaging and proton MR spectroscopic findings. AJNR. American Journal of Neuroradiology, 16(2), 233–239. Chugani, D. C., Sundram, B. S., Behen, M., Lee, M. L., & Moore, G. J. (1999). Evidence of altered energy metabolism in autistic children. Progress in Neuro-Psychopharmacology and Biological Psychiatry, 23(4), 635–641. Coleman, M., & Blass, J. P. (1985). Autism and lactic acidosis. Journal of Autism and Developmental Disorders, 15(1), 1–8. Corrigan, N. M., Richards, T. L., Friedman, S. D., Petropoulos, H., & Dager, S. R. (2010). Improving 1H MRSI measurement of cerebral lactate for clinical applications. Psychiatry Res: Neuroimaging, 182(1), 40–47. Cortopassi, G. A., Shibata, D., Soong, N. W., & Arnheim, N. (1992). A pattern of accumulation of a somatic deletion of mitochondrial DNA in aging human tissues. Proceedings of the National Academy of Sciences of the United States of America, 89(16), 7370–7374. Dager, S. R., Corrigan, N. M., Richards, T. L., & Posse, S. (2008). Research applications of magnetic resonance spectroscopy to investigate psychiatric disorders. Topics in Magnetic Resonance Imaging, 19(2), 81–96. Dager, S. R., Corrigan, N., Richards, T., & Shaw, D. (in press). Brain chemistry: Magnetic resonance spectroscopy. In D. Amaral, G. Dawson, & D. Geshwind (Eds.), Autism spectrum disorder. New York: Oxford University Press. Dager, S. R., Friedman, S. D., Parow, A., Demopulos, C., Stoll, A. L., Lyoo, I. K., et al. (2004). Brain metabolic alterations in medication-free patients with bipolar disorder. Archives of General Psychiatry, 61(5), 450–458.
123
114 Dager, S. R., Layton, M. E., Strauss, W., Richards, T. L., Heide, A., Friedman, S. D., et al. (1999). Human brain metabolic response to caffeine and the effects of tolerance. American Journal of Psychiatry, 156(2), 229–237. Dager, S. R., Strauss, W. L., Marro, K. I., Richards, T. L., Metzger, G. D., & Artru, A. A. (1995). Proton magnetic resonance spectroscopy investigation of hyperventilation in subjects with panic disorder and comparison subjects. American Journal of Psychiatry, 152(5), 666–672. Dager, S. R., & Steen, R. G. (1992). Applications of magnetic resonance spectroscopy to the investigation of neuropsychiatric disorders. Neuropsychopharmacology, 6(4), 249–266. de Graaf, R. A. (2007). In vivo NMR spectroscopy: Principles and techniques (2nd ed.). Chichester, West Sussex, England: John Wiley and Sons, Ltd. De Volder, A., Bol, A., Michel, C., Congneau, M., & Goffinet, A. M. (1987). Brain glucose metabolism in children with the autistic syndrome: Positron tomography analysis. Brain and Development, 9(6), 581–587. Dinopoulos, A., Cecil, K. M., Schapiro, M. B., Papadimitriou, A., Hadjigeorgiou, G. M., Wong, B., et al. (2005). Brain MRI and proton MRS findings in infants and children with respiratory chain defects. Neuropediatrics, 36(5), 290–301. Filiano, J., Goldenthal, M. J., Rhodes, C. H., & Marı´n-Garcı´a, J. (2002). Mitochondrial dysfunction in patients with hypotonia, epilepsy, autism, and developmental delay: HEADD syndrome. Journal of Child Neurology, 17(6), 435–439. Filipek, P. A., Juranek, J., Smith, M., Mays, L. Z., Ramos, E. R., Bocian, M., et al. (2003). Mitochondrial dysfunction in autistic patients with 15q inverted duplication. Annals of Neurology, 53(6), 801–804. Finsterer, J. (2009). Central nervous system imaging in mitochondrial disorders. Canadian Journal of Neurological Sciences, 36(2), 143–153. Friedman, S. D., Jensen, J. E., Frederick, B. B., Artru, A. A., Renshaw, P. F., & Dager, S. R. (2007). Brain changes to hypocapnia using rapidly interleaved phosphorus-proton magnetic resonance spectroscopy at 4 T. Journal of Cerebral Blood Flow and Metabolism, 27(3), 646–653. Friedman, S. D., Shaw, D. W. W., Artru, A. A., Dawson, G., Petropoulos, H., & Dager, S. R. (2006a). Gray and white matter brain chemistry in young children with autism. Archives of General Psychiatry, 63(7), 786–794. Friedman, S. D., Mathis, C. M., Hayes, C., Renshaw, P., & Dager, S. R. (2006b). Brain pH response to hyperventilation in panic disorder: Preliminary evidence for altered acid-base regulation. American Journal of Psychiatry, 163(4), 710–715. Friedman, S. D., Shaw, D. W., Artru, A. A., Richards, T. L., Gardner, J., Dawson, G., et al. (2003). Regional brain chemical alterations in young children with autism spectrum disorder. Neurology, 60(1), 100–107. Gellerich, F. N., Mayr, J. A., Reuter, S., Sperl, W., & Zierz, S. (2004). The problem of interlab variation in methods for mitochondrial disease diagnosis: enzymatic measurement of respiratory chain complexes. Mitochondrion, 4(5–6), 427–439. Giulivi, C., Zhang, Y.-F., Omanska-Klusek, A., Ross-Inta, C., Wong, S., Hertz-Picciotto, I., et al. (2010). Mitochondrial dysfunction in autism. JAMA, 304(21), 2389–2396. Graf, W. D., Marin-Garcia, J., Gao, H. G., Pizzo, S., Naviaux, R. K., Markusic, D., et al. (2000). Autism associated with the mitochondrial DNA G8363A transfer RNA(Lys) mutation. Journal of Child Neurology, 15(6), 357–361. Guillemaud, R. (1998). Uniformity correction with homomorphic filtering on region of interest. IEEE ICIP, 2, 872–875.
123
J Autism Dev Disord (2012) 42:105–115 Haas, R. H., Parikh, S., Falk, M. J., Saneto, R. P., Wolf, N. I., Darin, N., et al. (2007). Mitochondrial disease: A practical approach for primary care physicians. Pediatrics, 120(6), 1326–1333. Horwitz, B., Rumsey, J., Grady, C., & Rapoport, S. (1987). Interregional correlations of glucose utilization among brain regions in autistic adults. Annals of Neurology, 22(1), 118. Kapeller, P., Fazekas, F., Offenbacher, H., Stollberger, R., Schmidt, R., Berglo¨ff, J., et al. (1996). Magnetic resonance imaging and spectroscopy of progressive cerebral involvement in Kearns Sayre Syndrome. Journal of the Neurological Sciences, 135(2), 126–130. Kirby, D. M., Crawford, M., Cleary, M. A., Dahl, H. H., Dennett, X., & Thorburn, D. R. (1999). Respiratory chain complex I deficiency: An underdiagnosed energy generation disorder. Neurology, 52(6), 1255–1264. Kreis, R., Hofmann, L., Kuhlmann, B., Boesch, C., Bossi, E., & Hu¨ppi, P. S. (2002). Brain metabolite composition during early human brain development as measured by quantitative in vivo 1H magnetic resonance spectroscopy. Magnetic Resonance in Medicine, 48(6), 949–958. Kuwabara, T., Watanabe, H., Tanaka, K., Tsuji, S., Ohkubo, M., Ito, T., et al. (1994). Mitochondrial encephalomyopathy: Elevated visual cortex lactate unresponsive to photic stimulation—A localized 1H-MRS study. Neurology, 44(3 Pt 1), 557–559. Lin, D. D. M., Crawford, T. O., & Barker, P. B. (2003). Proton MR spectroscopy in the diagnostic evaluation of suspected mitochondrial disease. AJNR. American Journal of Neuroradiology, 24(1), 33–41. Lombard, J. (1998). Autism: A mitochondrial disorder? Medical Hypotheses, 50(6), 497–500. Lord, C., Risi, S., Lambrecht, L., Cook, E. H., Leventhal, B. L., DiLavore, P. C., et al. (2000). The autism diagnostic observation schedule-generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30(3), 205–223. Lord, C., Rutter, M., & Le Couteur, A. (1994). Autism diagnostic interview-revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders, 24(5), 659–685. Mancuso, M., Coppede, F., Migliore, L., Siciliano, G., & Murri, L. (2006). Mitochondrial dysfunction, oxidative stress and neurodegeneration. Journal of Alzheimers Disease, 10(1), 59–73. Mathews, P. M., Andermann, F., Silver, K., Karpati, G., & Arnold, D. L. (1993). Proton MR spectroscopic characterization of differences in regional brain metabolic abnormalities in mitochondrial encephalomyopathies. Neurology, 43(12), 2484–2490. Meister, A. (1995). Mitochondrial changes associated with glutathione deficiency. Biochimica et Biophysica Acta, 1271(1), 35–42. Minshew, N. J., Goldstein, G., Dombrowski, S. M., Panchalingam, K., & Pettegrew, J. W. (1993). A preliminary 31P MRS study of autism: Evidence for undersynthesis and increased degradation of brain membranes. Biological Psychiatry, 33(11–12), 762–773. Mitochondrial Encephalopathies: Potential Relationships to Autism? (2008). Retrieved September 1, 2010 from http://www.ninds.nih. gov/news_and_events/proceedings/20090629_mitochondrial.htm. Mo¨ller, H. E., Kurlemann, G., Pu¨tzler, M., Wiedermann, D., Hilbich, T., & Fiedler, B. (2005). Magnetic resonance spectroscopy in patients with MELAS. Journal of the Neurological Sciences, 229–230, 131–139. Oliveira, G., Diogo, L., Grazina, M., Garcia, P., Ataı´de, A., Marques, C., et al. (2005). Mitochondrial dysfunction in autism spectrum disorders: A population-based study. Developmental Medicine and Child Neurology, 47(3), 185–189.
J Autism Dev Disord (2012) 42:105–115 Petropoulos, H., Friedman, S. D., Shaw, D. W. W., Artru, A. A., Dawson, G., & Dager, S. R. (2006). Gray matter abnormalities in autism spectrum disorder revealed by T2 relaxation. Neurology, 67(4), 632–636. Poling, J. S., Frye, R. E., Shoffner, J., & Zimmerman, A. W. (2006). Developmental regression and mitochondrial dysfunction in a child with autism. Journal of Child Neurology, 21(2), 170–172. Pons, R., Andreu, A. L., Checcarelli, N., Vila`, M. R., Engelstad, K., Sue, C. M., et al. (2004). Mitochondrial DNA abnormalities and autistic spectrum disorders. Journal of Pediatrics, 144(1), 81–85. Posse, S., Dager, S. R., Richards, T. L., Yuan, C., Ogg, R., Artru, A. A., et al. (1997). In vivo measurement of regional brain metabolic response to hyperventilation using magnetic resonance: Proton echo planar spectroscopic imaging (PEPSI). Magnetic Resonance in Medicine, 37(6), 858–865. Pouwels, P. J., Brockmann, K., Kruse, B., Wilken, B., Wick, M., Hanefeld, F., et al. (1999). Regional age dependence of human brain metabolites from infancy to adulthood as detected by quantitative localized proton MRS. Pediatric Research, 46(4), 474–485. Provencher, S. W. (1993). Estimation of metabolite concentrations from localized in vivo proton NMR spectra. Magnetic Resonance in Medicine, 30(6), 672–679.
115 Rossignol, D. A., Frye, R. E. (2011). Mitochondrial dysfunction in autism spectrum disorders: a systematic review and metaanalysis. Molecular Psychiatry, 25. [Epub ahead of print]. Rumsey, J. M., Duara, R., Grady, C., Rapoport, J. L., Margolin, R. A., Rapoport, S. I., et al. (1985). Brain metabolism in autism. Resting cerebral glucose utilization rates as measured with positron emission tomography. Archives of General Psychiatry, 42(5), 448–455. Saneto, R. P., Friedman, S. D., & Shaw, D. W. W. (2008). Neuroimaging of mitochondrial disease. Mitochondrion, 8(5-6), 396–413. Sparks, B. F., Friedman, S. D., Shaw, D. W., Aylward, E. H., Echelard, D., Artru, A. A., et al. (2002). Brain structural abnormalities in young children with autism spectrum disorder. Neurology, 59(2), 184–192. Thorburn, D. R., & Smeitink, J. (2001). Diagnosis of mitochondrial disorders: Clinical and biochemical approach. Journal of Inherited Metabolic Disease, 24(2), 312–316. Zwaigenbaum, L. (2010). Advances in the early detection of autism. Current Opinion in Neurology, 23(2), 97–102.
123