FCD is a developmental malformation of the cerebral cortex that is associated with the development of pharmacoresistant epilepsy. FCD patients usually present with seizures in early childhood, and some also exhibit other clinical manifestations, including cognitive impairment, developmental delay, behavioral disorders, autism spectrum disorders, and focal neurological deficits[1]. Seizure origin and the occurrence of additional clinical symptoms are determined by the anatomical location of FCD, which can be detected by MRI or histopathological examination of resected tissue specimens. Although the pathogenesis of FCD remains to be defined, the association between the adenosine system and FCD-associated epilepsy was suggested by our previous study[13]. Adenosine is an endogenous anticonvulsant and acts as a seizure terminator in the brain[19]. Our previous data indicated that overexpression of ADK and consequent adenosine deficiency in FCD lesions are strongly associated with the development of pharmacoresistant epilepsy[13]. In this study, through immunohistochemistry and Western blot analysis, we first evaluated maladaptive changes in the adenosine system within FCD type I and type II lesions and found increased expression of the two major adenosine-metabolizing enzymes ADK and ADA, as well as the major adenosine-producing enzyme CD73, suggesting increased turnover of adenosine. We also found an increase in A2AR expression in conjunction with alterations in two downstream pathways of A2AR, as shown by a decrease in the expression of the astroglial glutamate transporter GLT-1 and increased expression of mTOR. These results suggest that dysregulation of the adenosine system and its downstream pathways is a common pathological feature of FCD. Adenosine-metabolizing enzymes or adenosine receptors might therefore be considered potential targets for the treatment of epilepsy associated with FCD.
Adenosine is a ubiquitous inhibitory modulator released during seizures and has been shown to mediate seizure arrest, postictal refractoriness, and suppression of epileptogenesis[19]. Neuronal excitability in the brain is modulated by activation of G protein-coupled adenosine receptors, specifically inhibitory A1Rs and facilitatory A2ARs[18]. The equilibrium of adenosine receptor expression levels, as well as the availability of endogenous adenosine to activate the receptors, plays a crucial role in the development and progression of epilepsy[18]. In both epilepsy patients and animal models, adenosine dysfunction, including upregulation of the major adenosine-metabolizing enzyme ADK, subsequent reduction in adenosine tone[13–15], and increased expression of A2ARs[27], has been demonstrated to promote epileptogenesis and seizures[44].
Astrocytic ADK is the major adenosine-metabolizing enzyme in the brain. Extracellular levels of adenosine are regulated mainly by an astrocyte-based adenosine cycle, with astrocytes serving as the key metabolic sink for adenosine through its clearance via ADK[45]. In line with this role, genetic disruption of ADK expression increases extracellular adenosine levels[46]. Upregulation of ADK and the consequent reduction in extracellular adenosine concentrations are sufficient to induce seizures through insufficient activation of A1Rs[14, 47]. Thus, overexpression of ADK in conjunction with astrogliosis has been demonstrated to be a pathological hallmark of experimental epilepsy as well as temporal lobe epilepsy, RE, and tumor-associated epilepsy[31, 33, 48]. In the present study, we found robust overexpression of ADK in reactive astrocytes and neurons within FCD type I lesions, as well as in dysmorphic enlarged neurons and BCs in FCD type II lesions. These results indicate that overexpression of astrocytic ADK might play a role in epileptogenesis in FCD patients.
There are two alternatively spliced forms of ADK: ADK-long (ADK-L), which is expressed in the nucleus, and ADK-short (ADK-S), which is expressed in the cytoplasm[49]. Cytoplasmic ADK-S determines intra- and extracellular levels of adenosine, which are balanced through equilibrative nucleoside transporters. Therefore, intracellular ADK-S determines the activation of adenosine receptors and thereby contributes to the regulation of neuronal excitability[36, 50]. ADK-L is localized in the nucleus and is needed to maintain methylation reactions, including DNA and histone methylation, which play a role in the regulation of epileptogenesis and cell proliferation[19]. Overexpression of both subtypes of ADK was detected within FCD lesions in the present study. These results suggest that dysregulation of ADK expression may impact FCD on different levels: the increase in ADK-L levels may drive epileptogenesis through an epigenetic mechanism and may affect cell proliferation during cortical development, whereas the increase in ADK-S levels may directly decrease seizure thresholds through insufficient activation of A1Rs and thereby contribute to seizures in FCD.
In addition to the overexpression of ADK in astrocytes, we found ectopic expression of both isoforms of ADK in neurons and BCs within FCD type I and type II lesions. Developmental studies performed in mice suggest that there is a switch in the ADK expression pattern in the brain from neuronal expression during the perinatal period to nearly exclusive astrocytic expression in adulthood, suggesting an association of neuronal ADK with developmental processes in the immature brain[51]. Consistently, ADK was found to be strongly expressed in total cortical homogenates of the human fetal brain (gestational week 13; temporal cortex), including in deep compartments of the cortical wall (ventricular/subventricular zone) in early stages of corticogenesis[50]. Analysis of human specimens collected at autopsy and surgical controls demonstrated that ADK is weakly expressed in glial cells while not expressed in most neurons of the cortex[13, 15, 33, 48]. These results suggest that ADK expression during brain development may affect important neural progenitor cell processes, including proliferation, survival, and neuroplasticity. Thus, our current findings of ectopic neuronal expression of ADK in FCD suggest an earlier, more immature expression profile of components of the adenosine system, which may indicate impaired brain maturation in FCD[50, 51]. The immature pattern of ADK expression in neurons suggests that these cells may retain a more embryonic phenotype and failed to mature fully, therefore continuing to express embryonic genes and proteins. In FCD patients, insults during the prenatal period might affect neural stem cells in the ventricular zone, leading to abnormalities in their proliferation, maturation, migration and terminal differentiation[13]. We also found changes in the expression of a second adenosine-metabolizing enzyme, ADA, that paralleled changes in ADK expression suggesting a marked change in the extracellular adenosine pool, which is largely responsible for the activation of inhibitory A1Rs.
While the extracellular adenosine pool is regulated by ADK and ADA, there is a distinct synaptic adenosine pool that is largely under the control of the adenosine-producing enzyme CD73 and allows the control of A2AR activation at the single-synapse level. In line with this association, in our present study, we found a concomitant increase in CD73 and A2AR immunoreactivity similar to what we found previously in specimens from patients with epilepsy associated with RE[27]. Overexpression of CD73 within FCD lesions suggests that increased CD73-mediated adenosine production might elevate the activation of A2ARs in FCD. In support of increased A2AR activation, we also detected increased A2AR immunoreactivity in FCD, which might represent a compensatory response to the reduction in extracellular adenosine tone. A2ARs can promote synaptic neurotoxicity by mediating the release of adenosine or its precursor ATP from neurons, thus aggravating the progression of epilepsy[52, 53]. A2ARs are mainly activated through CD73-dependent adenosine production via the frequency-dependent release of ATP[23]. While the role of A1Rs in epilepsy has been established, there are only a few reports on the role of A2ARs in epilepsy, and the findings are controversial[24, 34]. Pharmacological evidence from in vivo and in vitro studies suggests that A2ARs have proconvulsive and proepileptic functions[54–57]. Neuronal hyperexcitability, such as epileptic seizures or seizure-like activity, can further induce A2AR activation, thereby exacerbating damage to neural circuits and increasing the severity of epilepsy[19]. In the present study, we observed overexpression of A2ARs in reactive astrocytes, microglial cells, and neurons in FCD, which in general suggests the promotion of epileptogenesis, which may lead to further seizures. Interestingly, we found A2AR immunoreactivity in the cytoplasm of neurons and astrocytes, as well as in the cytoplasm of dysmorphic enlarged neurons and BCs in FCD specimens. These unexpected findings are consistent with previous results reporting that A2ARs is localized in the cytoplasm in models of Alzheimer’s and Parkinson’s disease, as well as in specimens from RE patients[25, 27, 58]. The function and pathological involvement of cytoplasmic A2AR are currently unknown. We propose that once the membrane is saturated with receptors, excess receptors accumulate in the cytoplasm[27].
According to previous studies, an increase in the transmembrane expression of A2AR is expected to have a net proepileptic effect[59, 60]. In an in vitro study, administration of the A2AR antagonist ZM241385 shortened the duration of epileptiform activity, suggesting that A2AR activation may impair the anticonvulsant activity of A1Rs[59]. In A2AR knockout mice, the severity of epileptic seizures induced by pentylenetetrazol or pilocarpine and the proportion of mice that experienced convulsions were found to be significantly reduced, indicating that activation of A2ARs can modulate excitatory neurotransmission and exacerbate limbic seizures[60]. A2AR blockade can also prevent cisplatin-induced impairment of neurogenesis and cognitive function[61]. Consistent with these results, the increase in A2AR immunoreactivity in neurons and reactive astrocytes in FCD specimens observed in the present study may indicate profound changes in A2AR-related synaptic plasticity, neuronal excitability, and inflammation in FCD patients[62, 63].
Glutamate is the major excitatory neurotransmitter in the mammalian central nervous system, and excess extracellular glutamate plays a critical role in excitotoxicity and neuronal cell death. The astroglial glutamate transporter GLT-1, a key regulator of extracellular glutamate concentrations, mediates the uptake of glutamate by astrocytes to prevent excitotoxicity and cell death[64]. GLT-1 is under the control of A2ARs, which regulate astrocytic glutamate clearance through their physical associations with GLT-1[25]. In primary cultured astrocytes as well as in an ex vivo preparations enriched with glial plasmalemmal vesicles, prolonged activation of A2ARs led to decreased activity and expression of GLT-1, thus inducing a sustained reduction in astrocytic glutamate uptake[25]. Abnormalities in astrocytic A2AR expression and activity can therefore affect the communication between astrocytes and neurons, thereby altering glutamate homeostasis via regulation of GLT-1 activity[14]. This mechanism leads to excessive extracellular accumulation of glutamate and is thought to be involved in the pathology of neurodegenerative diseases[64]. In our previous study of specimens from RE patients, astrocytes in the brains of RE patients were found to exhibit a reduction in GLT-1 expression concomitant with an increase in A2AR expression, which might explain the increases in excitotoxicity and cell death[27]. Consistent results were also observed in the present study. An increase in A2AR expression accompanied by a decrease in GLT-1 expression in reactive astrocytes was found within lesions in the brains of FCD patients, indicating the contribution of A2ARs to the regulation of glutamate uptake in FCD. Thus, interactions between A2ARs and GLT-1 may influence cell death and cognitive dysfunction associated with FCD via modulation of glutamate signaling, and this combined effect may be one of the key pathological bases of glutamate synaptic dysfunction and increased glutamate excitotoxicity in FCD.
FCD constitutes a heterogeneous group of developmental disorders caused by germline or somatic mutations in genes modulating the PI3K/Akt/mTOR pathway[65]. mTOR is a widely expressed kinase that controls protein synthesis, cell size, and cell proliferation according to the availability of amino acids, growth factors, and nutrients and cell energy status[66]. Overactivation of the mTOR pathway has been found to promote epileptic seizures[67, 68]. The mTOR inhibitor rapamycin can effectively reduce the frequency and severity of epileptic seizures in patients with tuberous sclerosis[67]. AMP-activated protein kinase (AMPK), a cellular energy sensor and signal transducer regulated by a wide array of metabolic stressors, acts as an important upstream inhibitory of the mTOR pathway[69]. An increased AMP/ATP ratio in vivo usually indicates increased AMPK activity[70]. As an important substrate for AMP generation, adenosine may inhibit the downstream mTOR pathway through activation of AMPK, thus exerting an anticonvulsant effect[68]. In addition, the nuclear isoform of ADK may play a crucial role in the regulation of cell proliferation through a combination of epigenetic and other adenosine receptor-independent mechanisms, such as interactions with the mTOR pathway[36]. Over the past decade, increasing evidence has shown that FCD type II is associated with hyperactivation of the mTOR pathway[71]. Numerous studies have confirmed that the mTOR pathway is activated in dysmorphic enlarged neuron BCs in FCD type II patients, as well as in malformed neurons, microglia and immature cells in a variety of epilepsy-related pathologies[72, 73]. In the present study, we detected overexpression of mTOR in reactive astrocytes in white matter but not the gray matter in FCD type I lesions. However, we found upregulation of mTOR in dysmorphic enlarged neurons in both the gray and white matter in FCD type IIA lesions and BCs in both the gray and white matter in FCD type IIB lesions, as well as in reactive astrocytes in both types of lesions. In addition, within FCD type IIB lesions, we found strong colocalization of mTOR with the adenosine-metabolizing enzymes ADK and ADA and A2ARs. Based on the association of adenosine with the mTOR pathway, reduced adenosine concentrations in combination with attenuated inhibition of the mTOR pathway may be involved in the pathological mechanisms of epilepsy associated with FCD. In line with this notion, in a study focused on gastric cancer, the accumulation of adenosine in the hypoxic tumor microenvironment was found to activate the PI3K/Akt/mTOR signaling pathway by increasing the activation of A2ARs, which may enhance cancer cell stemness and promote cancer progression[74].
In conclusion, within FCD type I and type II lesions, the adenosine-metabolizing enzymes ADK and ADA and the synaptic adenosine-producing enzyme CD73 were upregulated, A2AR immunoreactivity was increased, and the expression of putative downstream mediators of adenosine were altered, with GLT-1 expression being decreased and mTOR expression being increased. Together, these maladaptive changes may contribute to epileptogenesis in FCD. Our results suggest that dysregulation of the adenosine system and its downstream mediators is a common pathological feature of both FCD type I and FCD type II. Enzymes involved in adenosine metabolism as well as adenosine receptors might be potential targets for the treatment of FCD-associated epilepsy. Our descriptive findings from brain samples from FCD patients warrant the design of mechanistic studies to unravel the pathology underlying FCD-associated epilepsy.