The role of anti-glutamic acid decarboxylase autoantibodies in mood disorders

Introduction

Mood disorders (MDs) are a relatively heterogeneous spectrum of psychiatric conditions. Differences in clinical course (single or recurrent episodes), severity and frequency of mood episodes, and population prevalence may characterize each syndrome [major depressive disorder (MDD), bipolar disorder (BD), cyclothymic disorder, dysthymia] within this broad nosological definition. These disorders generally have a substantial burden on the life of patients as well as on the public health systems.^[1,2] In fact, they have been increasingly recognized as leading causes of the worldwide burden of disease and disability.^[3]

Despite the recent substantial progress in unraveling the complex biological underpinnings of MDs,^[4] in which several biological pathways have been implicated,^[5-7] the pathophysiological mechanisms underlying these conditions are still unclear. Among these, it has been hypothesized that the gamma-aminobutyric acid (GABA) pathway takes the major role.^[8] Specifically, a low GABAergic function might be associated with the biological disruption leading to clinical symptomatology. Furthermore, specific alterations of the GABAergic molecular pathway might be present in patients manifesting distinct symptoms. One of these possible alterations may involve the role of autoantibodies for the L-glutamic acid decarboxylase (GAD), a key enzyme responsible of the synthesis of GABA.

We reviewed the limited research on the mechanisms responsible for the decreased GABAergic function in MDs using the following key words: "gamma-amino butyric acid" OR "GABA" OR "L-glutamic-acid decarboxylase" OR "GAD" OR "autoantibodies" OR "GADA" AND "mood disorder" OR "major depressive disorder" OR "bipolar disorder" OR "depression". We found more than 200 publications and we selected all those with pertaining information on the association of any MD and abnormalities in the GABA pathway.

The aim of this review is to focus on: (1) reviewing briefly on the role of GABAergic pathway in MDs; (2) describing the molecular functions of GAD; (3) discussing the specific role of GAD as an important factor in the pathophysiology of MDs; and (4) providing future directions for the implementation of GAD autoantibodies (GADA) screening in clinical practice.

GABA in MDs

GABA is an inhibitory neurotransmitter present exclusively in the central nervous system (CNS). In a pivotal case series of 4 patients reported by Emrich et al.,^[9] these authors demonstrated a marked mood-stabilizing effect of valproic acid (VPA) in the management of acute manic episodes as well as in the maintenance treatment of other 7 patients with recurrent episodes of manic or manic-schizoaffective psychosis, irresponsive to lithium prophylaxis. The beneficial effect of VPA suggested a role of the GABAergic pathway in MDD since this anticonvulsant leads to increased cerebral concentrations of GABA by inhibiting GABA-transaminase which degrades GABA and by facilitating the reuptake of released GABA into cells. VPA also stimulates synthesis of GABA by increasing activity of glutamic acid decarboxylase.^[10] Subsequently, Emrich et al.^[9] discovered a lack of GABA in the CNS of mood disorder patients, which was restored by VPA, hypothesizing that modifications of this neurobiological pathway may be associated with MDs. A more recent study has suggested that, in general, GABAergic anticonvulsants possess antimanic properties and that the specific antimanic effect of lithium is associated with an increased action of GABA.^[11] Furthermore, the same authors^[11] suggested that the increased inhibitory neurotransmission induced by long-term lithium treatment counteracts the increased excitatory neurotransmission resulting from elevated levels of glutamate (GLU) which was detected in post-mortem brain tissue of BD patients. Following the same line of investigation, GABA plasma level may represent a biological trait marker for MDD. Indeed, Petty and Schlesser^[12] found significantly decreased GABA plasma levels, compared to healthy controls, in 40% of depressed patients, but higher levels of plasma GABA were found in manic BD individuals. Moreover, they observed that patients with different types of depressions, particularly those with familial loading, had plasma GABA levels significantly lower than control groups. Instead, GABA plasma levels in patients with reactive or bipolar depression did not differ from those of controls.^[12] As a result, it has been proposed that plasma levels of GABA might be a useful marker to predict the susceptibility to a depressive disorder in people with a family history of MDs. Furthermore, plasma levels of GABA may be specific and predictive of response to treatment,^[13] although GABA plasma level appears to show a low sensitivity as a test for depression.

Several other studies showed a decreased concentration of GABA in cerebrospinal fluid (CSF) of patients with severe depressive disorder.^[14-17] In particular, MDD patients over 40 years of age had significantly lower CSF levels of GABA than younger subjects.^[17] In addition, GABA levels in CSF of patients with depression and schizoaffective disorder are lower than those with schizophrenia or neurological conditions,^[16] Parkinson’s disease, Huntington’s disease, and dementia, all conditions which present at times depressive features.^[18] Of note, free GABA levels in CSF were lower in depressive disorders than in BD manic patients or healthy subjects.^[16,17] In addition to the association of GABA levels with depressive disorders, low levels of GABA have been also found in anxiety disorders^[19,20] and chronic migraine,^[21] which is often comorbid with MDD. In a recent study, Mann et al.^[20] found an inverse correlation between psychic anxiety severity and free GABA levels in CSF, independently of age. Interestingly, benzodiazepines, the most used anti-anxiety agents, increased GABA synthesis in the CNS.^[22]

Other proofs of the association between plasma GABA levels and depressive disorders may derive from the effect of electro-convulsive therapy (ECT) on severe refractory depression, since this treatment has been associated with a down-regulation of the GLU/GABA ratio (i.e. an increase in GABA and a decrease of GLU levels) in the hippocampus of rats.^[23,24] In fact, this measure of the GABAergic tone appears to be more informative than single neurotransmitter levels, given that GLU (a precursor of GABA) and GABA exert their effects in a neuromodulatory conjunction.^[25] Similar findings were observed in humans. GABA concentrations, measured with proton magnetic resonance spectroscopy, were significantly elevated in the occipital cortex of depressed patients following ECT.^[26] The increased levels of GABA in association with ECT may explain its antidepressant actions. In addition, increased GABA concentrations in the occipital cortex were also found during treatment of MDD with serotonergic antidepressants.^[13]

Significantly lower levels of GABA were observed in the anterior cingulate cortex (ACC) of adolescents with MDD compared with healthy subjects.^[27] Levels were measured by the means of proton magnetic resonance spectroscopy and expressed as ratios to unsuppressed voxel tissue water (VTW). In this age group, significant differences were found for the ratio GABA/VTW in the ACC of adolescents with and without anhedonia but not in those nonanhedonic compared to healthy controls.^[27]

GABA pathway: focus on GAD

GABA is synthesized from GLU via decarboxylation by GAD, a pyridoxal phosphate-dependent enzyme.^[28] Studies of GAD began in the early 50s with the work of Roberts and Frankel,^[29,30] and Awapara^[31] who independently discovered that GABA is synthesized in GABAergic neurons in CNS. However, GAD and GABA were also detected in pancreas where the latter is stored in synaptic-like vesicles in islet beta cells.^[32-34] GAD exists in two isoforms: one of a molecular size of 65kDa is termed GAD₆₅, whilst the other, of 67kDa, is termed GAD₆₇. Each of them plays a distinct role: GAD₆₅ is the product of a gene located on chromosome 10 whereas GAD₆₇ gene is on chromosome 2.^[35-37] These two proteins appear to be essential for maintaining homeostasis and viability of complex organisms. Studies carried out in GAD₆₇ knockout mice showed a reduction in the levels of GABA and died at birth of a severe cleft palate.^[38] Instead GAD₆₅ knockout mice presented normal basal levels of GABA and appear normal at birth but developed fatal seizures and anxiety-like phenotypes.^[39] The two isoforms were localized in different neuronal compartments. GAD₆₅ lies primarily in axon terminals and synthesizes GABA for neuronal transmission. GAD₆₇ is widely distributed throughout the neuron for the synthesis of GABA for general metabolic activity.^[35,36,40]

GAD in MDs

A substantial amount of evidence has suggested a role of GAD in MDs. Karolewicz et al.^[41] measured the levels of GAD₆₅ and GAD₆₇ in post-mortem brain samples from the gray matter of left dorsolateral prefrontal cortex (DLPFC, Brodmann's Area 9) of MDD patients, treated and untreated with antidepressants. GAD₆₅ and GAD₆₇ were reduced in antidepressant-free MDD subjects compared to matched controls. This reduction was not present in MDD patients medicated with antidepressants at the time of death, suggesting that GAD₆₅ and GAD₆₇ might play a role in depressive syndromes.^[41] Fatemi et al.^[42] investigated the cerebellar levels of Reelin 410, 330 and 180 kDa, GAD₆₅, and GAD₆₇ in subjects with BD, schizophrenia, MDD, and controls using the well-characterized Stanley Brain Consortium Collection. They found a reduction in levels of GAD₆₅ and GAD₆₇ proteins in all psychiatric subjects. These findings could explain increased blood and CSF glutamate and glutamine levels in schizophrenic,^[43,44] and depressed subjects,^[45] probably due to an accumulation of these two precursor compounds of GABA. Indeed, normal production of GAD₆₅ and GAD₆₇, as well as of Reelin 410, 330 and 180 kDa, reflects normal GABAergic cell function in several parts of the brain including cerebellum.^[46] Of note, Reelin, a protein responsible for correct lamination of the CNS during the embryonic period, may be involved in the etiology of schizophrenia, BD and autism.^[47-54] Remarkably, deficits in CNS levels of Reelin can affect memory processing, learning, synaptic organization and cognition in adult brain.^[55]

Interestingly, GAD₆₇ appeared to be a promising biomarker for BD and schizophrenia, since it discriminated these illnesses among a number of psychiatric conditions with relatively high specificity and sensitivity.^[56] Moreover, decreased levels of GAD were found in the postmortem left DLPFC of BD patients compared with MDD individuals and controls, giving rise to the possibility of differentiating the brain areas involved in unipolar and bipolar depression. Heckers and colleagues first made the description of the distribution of GAD₆₅ and GAD₆₇ mRNA-positive neurons in the human hippocampus.^[57] They found that abnormalities of hippocampal GAD expression are more prominent in BD than in schizophrenia, whereas another study showed a larger reduction of GAD₆₇ mRNA-containing neurons in BD patients than in those with schizophrenia.^[58] Of interest is the evidence that decreased GAD₆₇ expression leads to reduction of levels of GABA, with a net effect of a reduced GABAergic inhibitory control over glutamatergic cells. Therefore, it has been hypothesized that GAD₆₇ levels could be a surrogate marker for psychosis liability^[59] and pharmacological agents that raise GAD₆₇ expression levels could represent novel targets for antipsychotic therapy.^[59] Taken together, these findings demonstrate that GAD might have a role in modulating the psychopathological presentation in distinct subset of patients with MDs, possibly those with more prominent mood congruent or incongruent psychotic features.

GADA in MDs

As we have previously described, MDs may be associated with low levels of GABA following a decreased activity of GAD. It would therefore be reasonable to assume that its antibody, GADA, can somehow be involved in the pathogenesis of MD inhibiting GABAergic function. Support to this hypothesis came from research on the Stiff Person Syndrome (SPS). In 1956, Moersh and Woltman^[60] observed this syndrome in 14 patients over the age of 35 years, characterized by fluctuating rigidity and spasms without pyramidal tract dysfunction, or any other known neurologic disorders that could explain the stiffness. Moreover, Levy et al.^[61] observed muscular rigidity and episodic spasms superimposed on the rigidity in 20 consecutive patients. They reported as the hallmark sign the continuous contraction of the agonist and antagonist muscles in the trunk that caused hyperlordosis and respiratory problems. Interestingly, several patients initially received a tentative diagnosed of psychogenic process because their presentation was dominated by task-specific phobias and their stiffness was precipitated by unexpected noises or mental anticipation. In addition, seizures were observed in 10% of cases. Anxious and depressive symptoms in SPS can be explained by alterations in GABAergic neurotransmission. It was demonstrated that stiffness was caused by a reduction of GABA or glycine, the two main inhibitory neurotransmitters, and was improved by drugs increasing brain levels of GABA, such as diazepam or VPA.^[62] It was found later that up to 65% of patients may have antibodies (GADA) against both GAD₆₅ and GAD₆₇: GAD₆₅ and GAD₆₇^[62,63] In contrast, patients with type-1 diabetes had anti-GAD antibody titers 50 times lower than those of patients with SPS. The epitope of the GAD antigen may also differ between patients with type-1 diabetes and those with the SPS.^[64] These autoantibodies cause a functional impairment in the synthesis of GABA in persons with SPS; therefore, GADA should be considered to play a pathogenic role in this disease. In this context, it is noteworthy that patients with type-1 diabetes mellitus, particularly those with poor glycemic control, often undergo CNS related changes with low cognitive performance and depression. Short-term treatment of depression in patients with diabetes improves their dysphoria and other signs and symptoms of depression.^[65] In Batten disease, a rare genetic neurodegenerative disorder characterized by severe mental impairment, Chattopadhyay et al.^[66] studied a mouse model reporting the presence of an autoantibody to GAD₆₅. These authors hypothesized that an autoimmune response to GAD₆₅ may contribute to a preferential loss of GABAergic neurons associated with Batten disease.^[66] Several groups observed an increased prevalence of autoimmune disease and/or autoantibodies in patients with BD, including autoimmune thyroiditis and autoimmune atrophic gastritis. More specifically, Padmos et al.^[67] studied 239 patients with DSM-IV BD, 74 patients with DSM-IV schizophrenia, and 220 healthy control subjects for detection of GAD₆₅, GAD₆₇ and thyroperoxidase antibodies (TPOA), formerly reported to have an increased prevalence in patients with BD. The presence of GAD₆₅ (and not that of TPOA and H⁺/K⁺ adenosine triphosphatase antibody) tended to be associated with BD.^[67] Psychiatric symptoms, such as depression and anxiety, may be prominent, resulting in an incorrect diagnosis. Culav-Sumić et al.^[68] described a case of a woman who initially presented with anxious depression and remained resistant to treatment with different classes of antidepressants and additional therapy with lithium and atypical antipsychotics until the detection of GADA supported the diagnosis of SPS. Even the benefit obtained with immunosuppressive treatment with methylprednisolone might support the findings of anxious and depressive symptoms in SPS following the abnormal GABAergic neurotransmission. Finally, Yarlagadda et al.^[69] found elevated, but not statistically significant, levels of GADA in 12 patients with chronic psychotic disorders (schizophrenia and schizoaffective disorder) compared to healthy controls suggesting a link between antibodies to GAD₆₅ and chronic psychotic disorders as well as an autoimmune mechanism in the pathogenesis of these disorders.

Future directions

Existing evidence supports the role of GADA in the pathophysiology of a set of heterogeneous disorders that share clinical manifestations of severe motor, behavioral and mood symptoms. It is of interest that BD patients might present increasing titers of GADA compared to healthy controls. It is conceivable (and remains to be tested) that GADA levels might be one of the causative phenotypic manifestations of BD. Indeed, findings of a decreased GABAergic tone during mania might be explained by the diminished synthesis of this inhibitory neurotransmitter due to the action of GADA. Carefully designed studies targeting subsets of BD patients could clarify this hypothesis. The implications for diagnosis and treatment are significant. Detecting GADA in peripheral tissues is a feasible procedure that may assist the diagnostic assessment and, depending on the specificity and sensitivity, could be considered a screening test for BD patients. Moreover, GADA could be tested as a marker of response to treatment, particularly to GABAergic agents, such as VPA. Further works are needed to identify the exact pathophysiological mechanism through which GADA develop and to clarify whether a genetic liability may play a role. Regarding the latter, it is of interest that a recent pharmacogenomic analysis^[70] in Han Chinese BD patients found a strong association between glutamate decarboxylase-like protein 1 (GADL1) gene and response to lithium. The physiological functions of GADL1 gene are not clear, which may be, however, similar to those of GAD.

Conclusion

Several studies found that a decrease of brain GABA levels in mood disorder cases can be associated with manic or depressive states. This apparent incongruity may indicate more a mood-stabilizing role of GABA rather than an action on different mood phases. It remains to be established why GABA levels, both in CNS and peripherally, might also present elevation in specific clinical cases. In an attempt to clarify the mechanisms behind these abnormal levels of GABA in the brain, it has been hypothesized an abnormal function of the enzyme GAD that catalyzes the conversion from GLU to GABA. A weak action of this enzyme would justify decreased levels of the neurotransmitter. The aforementioned lines of evidence suggest that the autoantibody to GAD may be a possible causative factor. If this is confirmed, a relatively simple test to assess the level of GADA may provide a better diagnosis of a mood disorder and to improve treatment. If an abnormal level of this antibody is present as a trait rather than being associated with illness episodes, it would allow an early diagnosis of such prevalent and disabling disorder.