Depression in Epilepsy: A Neurobiologic Perspective

Abstract

Depression is the most frequent psychiatric comorbidity in patients with epilepsy. By the same token, patients with depression are at higher risk of developing epilepsy than are controls. Such bidirectional relations raise the question of whether both disorders share common pathogenic mechanisms, presenting with common neurotransmitter abnormalities and involvement of the same neuroanatomic structures. In this article, some of the available data in support of this hypothesis are reviewed.

Around 400 B.C., Hippocrates wrote, “Melancholics ordinarily become epileptics, and epileptics, melancholics: what determines the preference is the direction the malady takes; if it bears upon the body, epilepsy, if upon the intelligence, melancholy” (1). Hippocrates' supposition that “epileptics become melancholics” reflects current thinking of a unidirectional relation between depression and epilepsy, because depression is the most frequent psychiatric comorbidity in epilepsy (2–6). In contrast, Hippocrates' suggestion that patients with depression are at increased risk of developing epilepsy comes as a surprise to most clinicians, investigators, and patients alike. However, the premise is supported by two studies published in the last decade.

In a population-based, case control study carried out in patients with newly diagnosed epilepsy, Fosgren and Nystrom (7) found that a history of depression (preceding the onset of epilepsy) was seven times more frequent among patients than among age- and sex-matched controls. Similarly, in a population-based, case–control study of the incidence of new-onset epilepsy among adults aged 55 and older, Hesdorffer et al. (8) found that patients were 3.7 more likely to have a history of depression preceding their initial seizure than were controls. In this study, the authors also controlled for medical therapies of depression. Whereas data of these two studies suggest a bidirectional relation between depression and epilepsy, they cannot be interpreted as an indication of a causal relation. However, the high comorbidity prevalence rates of these two disorders suggest that depression and epilepsy may share pathogenic mechanisms. The purpose of this review is to examine some of the available data on their common pathogenic mechanisms.

Neurotransmitter Abnormalities in Epilepsy and Depression

It is reasonable to assume that neurotransmitter abnormalities in epilepsy and depression account for the antiepileptic and psychotropic properties of several antiepileptic drugs (AEDs), such as carbamazepine (CBZ), oxcarbazepine, valproic acid (VPA), and lamotrigine (LTG). Ample evidence exists that serotonin (i.e., 5-hydroxytryptamine [5-HT]), norepinephrine (NE), dopamine, γ-aminobutyric acid, and glutamate are operant in the pathogenesis of both disorders (9–14). This review focuses on common abnormalities of serotonergic and noradrenergic transmission—both of which are pivotal pathogenic mechanisms of mood disorders and the bases for development of antidepressant pharmacologic treatment (9). Likewise, decreased serotonergic and noradrenergic activity facilitates the kindling process of seizure foci, exacerbates seizure severity, and intensifies seizure predisposition in some animal models of epilepsy (10).

Are Abnormal Serotonergic and Noradrenergic Transmission Common Pathogenic Mechanisms to Epilepsy and Depression?

Experimental Data

In animal models of epilepsy, compelling experimental data on the pathogenic role played by 5-HT and NE in seizure predisposition are illustrated in studies of two strains of genetic epilepsy-prone rats (GEPR), GEPR-3 and GEPR-9, which are characterized by predisposition to sound-induced generalized tonic–clonic seizures (15–17) and, particularly in GEPR-9s, a marked acceleration of kindling (10). Both strains of rats have innate serotonergic and noradrenergic pre- and postsynaptic transmission deficits. Noradrenergic deficiencies in GEPRs appear to result from deficient arborization of neurons arising from the locus coeruleus (18,19), coupled with excessive presynaptic suppression of NE release in the terminal fields and lack of postsynaptic compensatory upregulation (10,20). GEPR-9 rats have a more pronounced NE transmission deficit and, in turn, exhibit more severe seizures than do GEPR-3 rats (21). Evidence also exists of deficits in serotonergic arborization in the GEPR brain as well as deficient postsynaptic serotonin_1A-receptor density in the hippocampus (22). Of note, patients with major depressive disorder (MDD) display endocrine abnormalities similar to those identified in GEPRs, including increased corticosterone serum levels, deficient secretion of growth hormone, and hypothyroidism (23).

Increments of either NE or 5-HT transmission can prevent seizure occurrence, whereas reduction will have the opposite effect (10,24). For example, drugs that interfere with the release or synthesis of NE or 5-HT exacerbate seizures in the GEPRs, including NE storage vesicle inactivators; reserpine or tetrabenazine; the NE false transmitter, α-methyl-m-tryosine; the NE synthesis inhibitor, α-methyl-p-tyrosine; and the 5-HT synthesis inhibitor, p-chlorophenylalanine. Conversely, drugs that enhance serotonergic transmission, such as the selective serotonin reuptake inhibitor (SSRI) sertraline, resulted in a dose-dependent seizure frequency reduction in the GEPR that correlates to the extracellular thalamic serotonergic concentration (25). The 5-HT precursor 5-hydroxy-l-tryptophan (5-HTP) has anticonvulsant effects in GEPRs when combined with the SSRI, fluoxetine (26). SSRIs and monoamine oxidase inhibitors (MAOIs) can exert anticonvulsant effects in experimental animals, such as mice and baboons, which are genetically prone to epilepsy (24,27), as well as nongenetically prone cats (28), rabbits (29), and rhesus monkeys (30). In addition, an antiepileptic effect of 5-HT_1A receptors has been correlated to a membrane-hyperpolarizing response, which is associated with increased potassium conductance in hippocampal-kindled seizures in cats and in intrahippocampal kainic-acid–induced seizures in freely moving rats (31,32).

As mentioned, AEDs with established psychotropic effects (CBZ, VPA, and LTG) can cause an increase in 5-HT (33–38). In GEPRs, the anticonvulsant protection of CBZ can be blocked with 5-HT–depleting drugs (34). In addition, the anticonvulsant effect of the vagal nerve stimulator (VNS) in the rat may be mediated by activation of the locus coeruleus (39). Deletion of noradrenergic and serotonergic neurons in the rat prevents or significantly reduces the anticonvulsant effect of VNS against electroshock or pentylenetetrazol-induced seizures (40). Furthermore, the effect of VNS on the locus coeruleus may be responsible for its antidepressant effects identified in humans.

Clinical Data

Deficits in 5-HT transmission in human depression may, in part, be related to a paucity of serotonergic innervation in terminal areas, which is suggested by a scarcity of 5-HT levels in brain tissue, plasma, and platelets, as well as by a deficit in serotonin transporter–binding sites in postmortem human brain (41–55). A deficit in the density or affinity of postsynaptic 5-HT_1A receptors has been identified in the hippocampus and amygdala of untreated depressed patients who committed suicide (56). Furthermore, in suicide victims with MDD, impaired serotonergic transmission is associated with defects in the dorsal raphe nuclei that result from suppression of 5-HT_1A autoreceptors caused by excessively dense serotonergic somatodendritic impulses (57).

In a positron emission tomography (PET) study, using the 5-HT_1A receptor antagonist [¹⁸F]trans-4-fluoro-N-2-[4-(2-methoxyphenyl)piperazin-1-yl]ethyl-N-(2-pyridyl) cyclohexanecarboxamide ([¹⁸F]FCWAY), reduced 5-HT_1A binding was found in mesial temporal structures ipsilateral to the seizure focus in patients with temporal lobe epilepsy (TLE), with and without hippocampal atrophy (58). In addition, a 20% binding reduction was found in the raphe and a 34% lower binding in the thalamic region ipsilateral to the seizure focus (these differences yielded a statistical trend). In a separate PET study aimed at quantifying 5-HT_1A–receptor binding in 14 patients with TLE, a binding reduction was identified in the raphe nuclei; in the epileptogenic hippocampus, amygdala, anterior cingulate, and lateral temporal neocortex ipsilateral to the seizure focus; and in the contralateral hippocampi, but to a lesser degree (59).

In contrast to animal studies, research on the impact of pharmacologic augmentation or reduction in 5-HT and NE transmission on seizures in humans has been rather sparse and based mostly on uncontrolled data. For example, depletion of monoamines from use of reserpine is associated with an increase in frequency and severity of seizures in patients with epilepsy (60,61), whereas the use of reserpine at doses of 2–10 mg/day was found to reduce the electroshock seizure threshold and the severity of the resulting seizures in patients with schizophrenia (62–64). The tricyclic antidepressant imipramine, with reuptake inhibitory effects of NE and 5-HT, was reported to suppress absence and myoclonic seizures in the only double-blind, placebo-controlled studies carried out so far (65–67). Open trials with the tricyclic antidepressant, doxepin, and the SSRIs, fluoxetine and citalopram, yielded an improvement in seizure frequency, but no controlled studies with SSRIs have been performed (68–71).

Are Common Neuroanatomic Structures Involved in Depression and Epilepsy?

A review of the literature reveals structural and functional abnormalities of the same neuroanatomic regions in primary depression and in epileptic seizure disorders that are frequently associated with comorbid depression (72,73). In epilepsy, relevant areas include mesial and orbitofrontal regions as well as mesial temporal and subcortical structures, such as thalamic nuclei. In primary MDD, Sheline (73) described the existence of morphologic and volumetric changes in various neuroanatomic structures that form a “limbic–cortical–striatal–pallidal–thalamic tract.” The tract consists of two branches: (1) a limbic–thalamic–cortical branch that includes the amygdala, hippocampus, and medial-dorsal nucleus of the thalamus as well as the mesial and ventrolateral prefrontal cortex; and (2) a branch running in parallel and linking the caudate, putamen, and globus pallidus with limbic and cortical regions. It is not surprising to find prevalence rates of depression ranging from 19% to 65% among patients with epilepsy of mesial temporal or frontal lobe origin (72). Evidence of common structural involvement will next be reviewed in greater detail.

Temporal Lobe Abnormalities

Hippocampal atrophy is among the most frequently identified abnormality in patients with epilepsy and primary depression. Furthermore, neuroimaging studies performed in patients with epilepsy and comorbid depression have identified a correlation between the severity of depression and severity of mesial temporal structural abnormalities, as identified in studies using magnetic resonance imaging (MRI), single-photon emission computed tomography (SPECT) scans with the tracer [^99mTc]-hexamethylene propylene amine (^99mTc-HMPAO), and proton magnetic resonance spectroscopy (¹H-MRS) (74–76).

In two separate studies of patients with a history of primary MDD in remission, Sheline et al. (77,78) reported bilateral, smaller hippocampal volumes than those of age, sex, and height-matched normal controls. They also identified large hippocampal low-signal foci (≥4.5 mm in diameter), and their number correlated with the total number of days depressed. A significant inverse correlation between the duration of depression and left hippocampal volume also was demonstrated. More recently, in a study of 38 female patients with a history of MDD, Sheline and colleagues (79) established that hippocampal atrophy was prevented with antidepressant drug therapy. They found a significant correlation between reduction in hippocampal volume and the duration of untreated depression, whereas no correlation was found between hippocampal volume loss and length of depression for patients taking antidepressant medication.

Both the neuropathologic findings and the magnitude of hippocampal volume reductions differ significantly between the two disorders, with reductions in TLE being significantly greater than those in MDD. In mesial temporal sclerosis, neuropathologic findings consist of neuronal cell loss and astrocytosis in hippocampal formation, amygdala, entorhinal cortex, and occasionally in parahippocampal gyrus. In hippocampus, neuronal cell loss is prominent in areas CA1 and CA4, the dentate gyrus, and the subiculum (82). Unfortunately, few neuropathologic studies of the human hippocampal formation in patients with primary MDD are available. However, Lucassen et al. (83) compared 15 hippocampi of patients with a history of MDD with 16 matched controls and nine steroid-treated patients (high steroids are associated with hippocampal atrophy). In 11 of 15 depressed patients, three steroid-treated patients, and one control, rare but convincing apoptosis was identified in entorhinal cortex, subiculum, dentate gyrus, CA1 and CA4.

Hippocampal atrophy in primary MDD has been attributed to two potential pathogenic mechanisms: (a) an alteration in brain-derived neurotrophic factors (BDNF), resulting from the mood disorder; and (b) high glucocorticoid exposure. It has been suggested that a decrease in BDNF levels in the dentate gyrus and pyramidal cell layer of hippocampus, amygdala, and neocortex is mediated by glucocorticoids and can be reversed with antidepressant therapy (84,85). Antidepressant drugs increased hippocampal BDNF levels in humans (86). High glucocorticoid exposure stems from excessive activation of the hypothalamic–pituitary–adrenal axis, with almost half of all individuals with depression having impaired dexamethasone suppression of adrenocorticotropic hormone (ACTH) and cortisol. These changes are also reversible to treatment with antidepressants (87). In animal studies with rats and monkeys, deleterious effects of prolonged glucocorticoid exposure were associated with damage to hippocampal neurons; impeded granule cell development in the adult hippocampal dentate gyrus; transient and reversible atrophy of the CA3 dendritic tree; and finally, results in cell death in extreme and prolonged conditions (88–91). In a neuropathologic study of amygdala and entorhinal cortex of seven patients with MDD, 10 with bipolar disorder (BPD), and 12 controls the specimens of MDD patients and those of patients with BPD had a significant reduction of glial cells and of the glial/neurons ratio in left amygdala and to a lesser degree in left entorhinal cortex (81).

Therefore are the neuropathologic changes of TLE magnified in the presence of a chronic, untreated depressive disorder? Whereas no answer to this question is available, some data suggest a negative impact of a psychiatric history on seizure outcome after pharmacologic (92) and surgical treatment (93,94). In a study of 90 patients who underwent an anterotemporal lobectomy for the management of a refractory TLE, a lifetime history of depression (identified at the time of the presurgical evaluation) was a predictor of a worse seizure outcome (94). Thus could these data suggest that depression may be a biologic marker for more severe epilepsy?

Frontal Lobe Abnormalities

Functional disturbances of frontal lobe structures have been recognized in TLE, particularly among patients with comorbid depression, and correlate to bilateral reduction in inferofrontal metabolism (95–98). Additionally, neuropsychological testing with the Wisconsin Card Sorting Test, which is highly sensitive to frontal-lobe–mediated executive dysfunction, has revealed poor performance in patients with TLE and comorbid depression (99).

Involvement of frontal lobes in primary depression has been demonstrated with functional neuroimaging (e.g., PET, SPECT) and neuropsychological studies (100,101). Executive abnormalities consistently are found in studies on depressive disorders, with stronger results apparent with more severe pathology. These neuropsychological disturbances correlated to reduced blood flow in mesial prefrontal cortex (102,103). Furthermore, in tests demanding executive function, cingulate cortex and striatum could not be activated in patients with MDD (104).

Likewise, structural changes have been identified in the cingulate gyrus and white matter of the orbitofrontal and prefrontal cortex, including smaller orbitofrontal cortex volumes in young adults (105,106) and in geriatric patients with MDD (107,108). Of note, the magnitude of prefrontal volume changes was related to the severity of the depression, as elderly patients with minor depression had lesser changes than did those with MDD (109).

Neuropathologic studies have documented structural cortical changes in frontal lobes of depressed patients. Rajkowska et al. (110) found decreases in cortical thickness, neuronal sizes, and neuronal densities in layers II, III, and IV of the rostral orbitofrontal region in the brains of depressed patients. In the caudal orbitofrontal cortex, significant reductions in glial densities in cortical layers V and VI associated with decreases in neuronal sizes were identified. Finally, in all cortical layers of the dorsolateral prefrontal cortex, the authors demonstrated a decrease in density and size of neuronal and glial cells.

Conclusions

Clearly, these data appear to suggest the involvement of common neuroanatomic structures and neurotransmitters in depression and epilepsy, which may explain the bidirectional relation between the two disorders and their frequent comorbid occurrence. This review only begins to examine the very complex interplay between neurobiologic aspects of mood disorders and epilepsy. Importantly, the review illustrates that depression in epilepsy is much more than a “psychosocial” complication!