|Home | About | Journals | Submit | Contact Us | Français|
To review progress in understanding pediatric obsessive-compulsive disorder (OCD). The focus is on the frontal-striatal-thalamic model of OCD, neurobiological and genetic studies of the disorder and their influence on recent advances in treatment.
Computerized literature searches were conducted with the keywords “obsessive-compulsive disorder” in conjunction with ‘pediatric”, “genetics” and “imaging”.
Neuroimaging studies find evidence to support the frontal-striatal-thalamic model. Genetic and neurochemical studies also implicate glutamate in the pathology of OCD. This has led to application of glutamate modulating agents to treat OCD.
Studies of pediatric OCD have led to a refined frontal-striatal-thalamic model of pathogenesis and are having an evidence-based impact on treatment. Despite this progress, fully explanatory models are still needed that would allow for accurate prognosis and the development of targeted and efficacious treatments.
Newer, non-invasive brain imaging approaches offer promise in enhancing understanding not only of brain development but also of the neurobiologic underpinnings of childhood-onset neuropsychiatric disorders. These techniques permit unprecedented in vivo ‘biopsies’ of brain structure, chemistry and function. Here, we present research aimed at generating a mechanistic understanding of the pathogenesis and treatment response of pediatric obsessive-compulsive disorder (OCD). With as many as 80% of all cases beginning during childhood and adolescence 1, pediatric studies are especially critical in advancing our understanding of the disorder. In this review we discuss the neurobiology of pediatric OCD, recent genetic findings and the novel application of glutamate-modulating agents for OCD. Special attention is focused on the glutamate hypothesis of OCD, first proffered by Rosenberg and Keshavan 2.
Neurocognitive testing of frontal cortical functions is under explored in pediatric OCD. Spatial-perceptual deficits similar to those of frontal lobe lesion patients were reported in OCD adolescents7 but not replicated 4, 5. Recently, deficits in visual attention and executive functioning were found in children with OCD 5. There is more evidence for prefrontal oculomotor abnormalities in pediatric OCD 6, 7 including in: ability to suppress responses, volitional execution of delayed responses, and anticipation of predictable events. Patients with OCD had more response-suppression failures than controls 6, 7. No significant differences between patients with OCD and controls were observed on other prefrontal cortical functions, such as the delayed-response task. Severity of OCD symptoms was related to response-suppression deficits 6. Woolley et al 8, using functional magnetic resonance imaging (fMRI) found that during the ‘stop’ task, OCD patients showed reduced activation in right orbitofrontal cortex, thalamus and basal ganglia compared to controls. These disturbances of inhibition in OCD may underlie the repetitive behavior that characterizes the illness and indicate abnormalities in orbital prefrontal ventral striatal circuits 6, 7. Indeed, pediatric OCD patients had significantly greater gray matter density in the orbital frontal cortex than healthy controls 9, confirmed by manual region of-interest (ROI) measurements. Furthermore, among patients, gray matter density in right lateral orbital frontal cortex correlated significantly with OCD symptom severity, but not with anxiety or depression 9.
Gray matter volume of the anterior cingulate is greater in pediatric OCD patients compared to age- and sex-matched controls 2, 10. Anterior cingulate volume was positively correlated with age in controls but not in OCD patients 2. There was no difference between groups for anterior cingulate white matter 10. In an independent sample, greater anterior cingulate gray matter volume was noted in patients than in controls using volumetric MRI, akin to previous findings 2, 10, but not with VBM 9. Converse to those reports 2, 9, 10, a VBM analysis by Carmona et al 11 found significantly lower gray matter density in OCD patients compared to healthy controls in the anterior cingulate bilaterally. This may indicate sensitivity issues with both techniques and fundamental differences in what is being measured 12. Single-voxel proton magnetic resonance spectroscopy (1H-MRS) of the anterior cingulate found lower glutamatergic concentrations (Glx) in OCD patients than in healthy controls 13. Below-normal anterior cingulate Glx was also noted in adult females with OCD 14. The lower anterior cingulate glutamate correlated with symptom severity in these patients. While no volumetric effects in dorsolateral prefrontal cortex (DLPFC) were noted 2, proton magnetic resonance spectroscopic imaging (1H-MRSI) did reveal above-normal concentration of the putative neuronal marker, N-Acetyl-Aspartate (NAA) in left but not right DLPFC in pediatric OCD patients 15. Higher NAA in left DLPFC may indicate abnormal cortical pruning in OCD.
OCD patients had significantly smaller striatal volumes than age- and sex-matched healthy controls 16. In the OCD patients, striatal volumes correlated inversely with symptom severity but not with illness duration 16. In a second sample, a smaller globus pallidus was noted in pediatric OCD patients than in healthy volunteers 10. Interestingly, a VBM study showed above-normal gray matter density in the bilateral putamen in pediatric OCD patients 9. It should be noted that VBM and manual tracing methods for evaluating brain volume have not been well validated against each other 12 and may not reflect identical aspects of brain morphology (see issues that require further study below). Above-normal striatal Glx concentrations, which normalized after successful treatment with an SSRI, were noted in pediatric OCD 17, 18. This reduction in striatal Glx may persist after SSRI discontinuation 19. Interestingly, cognitive behavioral therapy (CBT) did not change caudate Glx concentrations in pediatric OCD patients despite a reduction in symptoms 20. Finally, subcortical hyperintensities occur more frequently in children with OCD than in controls 21.
Thalamic volume is larger in pediatric OCD patients than in age- and sex-matched controls 22. After 12 weeks of treatment with paroxetine, thalamic volume normalized in OCD patients concurrent with a reduction in OCD symptoms. This reduction in thalamic volume in OCD appears specific to medication as no changes in thalamic volume were noted with CBT 23. Interestingly, below-normal ratios of bilateral medial-thalamic NAA/Cr and NAA/Cho were noted in OCD 24. Lower ratios could mean lower NAA. That might be inconsistent with the aforementioned above-normal overall thalamic volume in OCD 22, in so far as lower NAA could imply lower neuron mass and thus smaller thalamic volume. However, these results need not be associated with death or reduced size of neurons, A more advanced quantification technique, if fact, using validated phantom-replacement methodology that allowed for absolute quantification indicated greater medial-thalamic Cho 25, 26 and Cr 27 but not lower NAA 25–27 in OCD patients than in healthy controls. This discrepancy highlights the risk inherent in using metabolite ratios to describe MRS data. The finding of altered Cho in pediatric OCD is specific to the medial-thalamus as no difference was noted in lateral thalamus 25. OCD patients differed not only from controls with regard to medial-thalamic Cho and Cr, but also from pediatric MDD patients 26, 27.
Aside from the frontal-striatal-thalamic circuit, other brain regions have been implicated in pediatric OCD. In the corpus callosum, all subregions except for the isthmus were larger in OCD patients than in controls 28. Callosal area correlated significantly with OCD symptom severity but not with illness duration. Secondly, the age-related increase in callosal size observed in normal subjects was not present in OCD patients. MRI signal intensity, related to myelination in the corpus callosum, was lower in genu of the corpus callosum in OCD patients than in healthy controls 29. Lower signal intensity on T1-weighted images may indicate greater myelination of the region of the corpus callosum, leading to increased volume, as noted in the earlier study 28. The genu connects ventral prefrontal cortex with striatum, regions noted above to be critical in pediatric OCD. Dysregulation of the limbic-hypothalamic-pituitary-adrenal (LHPA) axis has been reported in patients with OCD 30–33. The pituitary gland is significantly smaller in treatment-naive pediatric OCD patients than in healthy controls, with a more prominent difference in males 34. Interestingly, the smaller gland volume found in OCD contrasts with the enlarged pituitary seen in MDD 35, 36 and bipolar disorder 37. It is not known if SSRI treatment changes pituitary volume in OCD as it does in MDD 38 or as antipsychotic medications do in schizophrenia 39.
Estimates of the heritability of obsessive-compulsive symptoms in children and adolescents range from 45% to 65% 40 indicating a strong genetic component to the illness. To date, two glutamate-related genes (transporter and receptor) have shown promise in explaining the above-described neurobiology of the illness.
Three independent groups found that the 3′ region of SCL1A1 may contain a susceptibility allele for OCD, primarily in male offspring 41–43. The protein product of this gene is the high-affinity neuronal and epithelial transporter (EAAT3, EAAC1) for L-glutamate, L- and D-aspartate, and cysteine 44, 45. EAAT3/EAAC1 is present in cortex, basal ganglia, and hippocampus, and is detected in all parts of the neuron 46. EAAT3/EAAC1 binds and transports cysteine more effectively than astrocyte glutamate transporters 47. Furthermore, EAAT3 is localized to some GABAergic neurons, where it may play a role in regulating GABA synthesis 48. In the adult brain, glutamate transport keeps extracellular glutamate below neurotoxic concentrations 49. However, in adults EAAT3/EAAC1 exhibits rather low expression and is thought to make a minor contribution to the removal of synaptic glutamate compared to EAAT1 and EAAT2 50. It is expressed during early brain development, before astrocytes are functional. This suggests that EAAT3/EAAC1 is involved in the developmental role of glutamate and, possibly, GABA, which is also excitatory in certain brain regions during early brain development 50. This role of EAAT3/EAAC1 in brain development is consistent with the linkage and association findings supporting SLC1A1 as a primary candidate gene in pediatric OCD 41–43 and in autistic spectrum disorders (Autism Genome Project Consortium, 2007). Expression of EAAT3/EAAC1 is regulated by testosterone and prolactin 45. Increased expression of EAAT3/EAAC1 by testosterone is consistent with association of OCD with SLC1A1 being strongest in males 41, 42. Mice deficient in EAAC1 develop dicarboxylic aminoaciduria 51, reduced neuronal glutathione and, with aging, brain atrophy, increased aggressiveness, and impaired self-grooming 44. These results in EAAT3/EAAC1 knockouts suggest that pediatric OCD may be associated with increased rather than with decreased EAAT3 expression. Pharmacogenetically, increased SLC1A1 expression might be a compensatory response of the brain that tends to suppress OCD symptoms and that could be supported by glutamate receptor antagonists or EAAT3 agonists if these can be identified. Under-expression of SLC1A1, glutamate receptor agonists and EAAT3 antagonists, in contrast, could all aggravate OCD symptoms. Under-expression of SLC1A2 and SLC1A3 could produce OCD symptoms that would be aggravated by EAAT1 and EAAT2 antagonists, while enhanced expression of these genes or EAAT1 and EAAT2 agonism could have therapeutic effects. In adult OCD, cognitive-behavioral therapy (CBT) has multiple effects on MRS metabolites 52–54, including reduction of above-normal baseline glutamate in anterior cingulate 55. Although different physiologic conditions may prevail in adult vs. pediatric OCD, the relative expression of neuronal and astrocytic glutamate transporters may again be instrumental in the production of symptoms in OCD and their remediation with CBT.
The 5072T/G variant of GRIN2B is significantly associated with OCD in pediatric patients 56. Additionally, the 5072G 5988T haplotype was associated with OCD. The NMDA subunit 2B gene [GRIN2B, (MIM 138252)] on chromosome 12p encodes for the NR2B subunit of the ionotropic glutamate receptor. GRIN2B is expressed in the striatum and the prefrontal cortex 57 consistent with regions demonstrating glutamatergic abnormalities in pediatric OCD patients 13, 18. GRIN2B has also been linked to schizophrenia 58, attention deficit hyperactivity disorder 59 and bipolar disorder 60. GRIN2B is thought to play a role in plasticity during cortical development 61. Additionally, neurotoxic levels of glutamate during the neonatal period increase the expression of NMDA NR2B in the striatum and cortex 62. The increased expression of GRIN2B in response to excess glutamate 63 suggests that pediatric OCD is associated with increased GRIN2B expression in the striatum.
Selective serotonin reuptake inhibitors (SSRI’s) are the only FDA-approved medications for OCD. However, SSRI’s are typically only effective in 40 to 60% of patients, leaving a substantial number still ill 64. Indeed, as treatment response is defined by a 20 to 40% reduction in symptoms, many “responders” are still markedly symptomatic 64. Given the persistence of symptoms and levels of treatment response, it is clear that the serotonin paradigm of understanding OCD does not fully account for the neurobiology of the disorder.
As discussed in the previous sections, evidence of glutamate abnormalities in OCD is mounting 13, 14, 18, 41, 42, 56, 65–68. Indeed, all of the 1H-MRS and CSF measures of glutamate concentration in OCD demonstrated very large effect sizes (d > 1.00) indicating robust differences in regional glutamate concentrations in OCD patients as compared to controls (see figure 2). This neurobiological evidence has led to the search for agents that modulate glutamate 69. Indeed, the glutamate-modulating agent riluzole (1-amino-6-trifluoromethoxybenzothiazole) has shown promise in neuropsychiatric disorders 70–75. Riluzole is well tolerated and is FDA approved for treatment of amyotrophic lateral sclerosis (ALS; 76–78). Riluzole is primarily an inhibitor of glutamate release but also inactivates voltage-dependant sodium channels in cortical neurons and blocks GABA reuptake 79–81. In a case report 70 and an open-label trial in adults 71, riluzole has shown promise for ameliorating the symptoms of OCD. More recently, an open-label trial in children (8 to 16 years) with OCD found riluzole to be beneficial and well tolerated 82. A larger placebo-controlled trial at NIMH is underway.
In 1998, Rosenberg and Keshavan 2 first hypothesized a role for glutamate in pediatric OCD. The first reports of in vivo differences in Glx between pediatric OCD patients and healthy were published shortly thereafter 17, 18. Since these initial reports, in OCD patients a lower concentration of glutamate has been noted in the anterior cingulate 13, 14 while greater Glx/Cr levels have been seen in orbital frontal white matter 65. These neuroimaging reports found further support in genetic studies that noted an increased susceptibility to OCD in those expressing alterations in the neuronal glutamate transporter gene 41, 42 and certain glutamate receptor genes 56, 67. Furthermore, peripheral markers 66 and animal models 68 have provided additional support for glutamate dysfunction in OCD. Clinically, glutamate-modulating agents are showing promise for OCD 70, 71, 82. Hence, 1H-MRS, CSF, genetic, animal and clinical studies have all implicated glutamate in OCD.
The traditional strategy of going from pharmacology to pathophysiology has failed to demonstrate substantive progress in our ability to understand psychiatric illness 83. It is becoming clear that investigators need to combine strategies (genetic, neuroimaging, pharmacological, animal models, etc) to allow for the most advancement 83, 84. Research into diabetes, heart disease and oncology is focused on cure and prevention. In psychiatry, the bar is typically set lower, with an eye only on incremental advances 83. A roadmap of how to achieve these advances is only now coming into focus. The progress described in pediatric OCD in this review is a rare occurrence in psychiatry, an example where neurobiological studies of a disorder have directly informed its treatment. Indeed, if one compares what is known regarding diabetes and what we are starting to see in pediatric OCD, one can see how the disorder is coming into greater focus using multiple methods (see table 2). Brain imaging has demonstrated great potential for aiding in the diagnosis, treatment, prevention, and cure of neuropsychiatric disorders 85. When coupled with advances in assessment, genetics, pharmacology and animal models, the potential to have meaningful clinical impact becomes profound.
To date, 1H-MRS studies of OCD have reported the combined Glx measure (glutamate and glutamine) 13, 14, 17, 18, 65. Given the differing physiological roles of glutamate and glutamine, techniques that allow for the separation of the two similar resonances need to be applied in OCD. Techniques include improved spectral editing 86–88 or the use of higher field MRI scanners 89.
The combination of genetics and imaging methods offers tremendous potential for advancing our understanding of psychiatric illness 90. First-order studies combining genetic and imaging findings in pediatric OCD are needed. The initial studies linking genetic markers with neuroimaging findings in pediatric OCD are currently underway by our group. Second-order studies are also needed to look at what cellular mechanisms linked to gene polymorphisms may be responsible for the changes noted in the imaging studies.
Additional studies of glutamate related genes are required. These two candidate genes (SLC1A1 and GRIN2B) mentioned in this review represent only the start of tying genetic studies into glutamate-related findings with pediatric OCD, as there are many glutamate related genes that have not yet been explored. Indeed, there are at least 25 genes for glutamate receptors and 5 genes for neuronal and glial glutamate transporters 91. We are unaware of investigations of potential associations between OCD and expression of the SLC1A2 gene (encoding for EAAT2 at 11p13-p12) or the SLC1A3 gene (encoding for EAAT1 at 5p13), though unpublished negative findings may exist. These two astrocyte glutamate transporters may influence the pathophysiology of OCD and its pharmacologic regulation by regulating regional glutamate levels. Glutamate enters astrocytes through these transporters 50 and is rapidly converted to non-toxic glutamine 92 which thence is exported to neurons (through monocarboxylate transporters; 93, 94) for re-conversion to glutamate 95. Enhanced expression of SLC1A2 and SLC1A3 would therefore increase the local residence time of glutamate and glutamine through this neuron-astrocyte cycling, resulting in higher Glx, as observed in the caudate in pediatric OCD 17, 18. Under-expression of SLC1A2 and SLC1A3, in contrast, would result in diversion of incoming synaptic glutamate to neurons, causing faster neuron firing, greater remote synaptic export of glutamate and/or consumption in the Krebs Cycle 96 to sustain the higher metabolic rate, and lower overall levels of tissue Glx, as observed in the anterior cingulate in pediatric OCD 13. Increased expression of SLC1A1 and the glutamate receptor gene GRIN2B (see above) might therefore be secondary responses, i.e., the generation of more EAAT3 transporters and glutamate receptors to handle higher extracellular glutamate concentrations, to a primary astrocyte defect in OCD. SLC1A1 accounts for 59% of cases of OCD 41. It might be possible to account for additional OCD cases by considering the combinations of relative expression of SLC1A1, SLC1A2, and SLC1A3 that lead to suboptimal synaptic glutamate distribution between astrocytes and neurons.
It is not known why some VBM studies conflict with volumetric MRI studies using manual tracing methods 12. For example, in the anterior cingulate in OCD patients Carmona et al 11 noted lower gray matter density bilaterally while Rosenberg and Keshavan 2 found greater gray matter volume. Interestingly, Szeszko et al 9 found greater anterior cingulate gray matter volume in patients as compared to healthy volunteers using volumetric MRI, but not with VBM. As the operational definition for gray matter density has not been resolved, it may be that the methods measure two very different things (i.e. volume vs. the statistical probability of a voxel being gray matter). Further work is needed to validate VBM and to resolve the conflict noted with traditional manual tracing techniques.
Given that the clinical phenomenology and nosology of OCD is, for the most part, well worked out, applying techniques developed in the emerging field of imaging genetics can further explicate the underlying developmental neurobiology of pediatric OCD. Such studies may provide further support for the glutamate hypothesis of OCD. The combined study of biological, genetic and behavioral/symptom variables also responds to the call for translational approaches to mental illness made by the National Institutes for Mental Health. Such approaches may lead to better understanding of pediatric OCD and, in turn, to new diagnostic and treatment approaches.
Financial Support: This work was supported in part by the State of Michigan Joe F. Young Sr. Psychiatric Research and Training Program, the Miriam L. Hamburger Endowed Chair of Child Psychiatry at Children’s Hospital of Michigan and Wayne State University, Detroit, MI, and grants from the National Institute of Mental Health (R01MH59299, R01MH65122, R01MH081864, K24MH02037) and the Mental Illness Research Association (MIRA).
Statistical Expert: None