Address correspondence and reprint requests to Dr. S.M. Sisodiya at UCL Institute of Neurology, Queen Square, London WC1N 3BG, U.K. E-mail: email@example.com
Summary: Purpose: Accumulating evidence implicates drug-transporter proteins ABCB1 and ABCC1 in resistance to multiple antiepileptic drugs (AEDs) in refractory epilepsy. These proteins are upregulated in surgically resected human brain tissue containing epileptogenic pathologies, including cortical dysplasia. In surgically resected cases, no upregulation is seen in normal adjacent tissue, suggesting that neither seizures nor prolonged exposure to AEDs need induce ABCB1 or ABCC1 expression. We wished to determine if status epilepticus might cause upregulation of these proteins.
Methods: Immunohistochemistry was performed for ABCB1 and ABCC1 in postmortem human brain tissue from a patient who died in status epilepticus and was found to have unihemispheric cortical dysplasia.
Results: There was upregulation of both proteins, as expected, in the hemisphere containing dysplasia. There also was widespread upregulation of both proteins in glia from the normal hemisphere. Previous work shows that drug treatment does not cause such upregulation.
Conclusions: Upregulation of these proteins in histologically normal brain tissue is most likely the result of seizures in status, as seen in animal models. The findings provide a possible explanation for the appearance of AED resistance in prolonged status and emphasise the importance of prompt treatment of status epilepticus.
Resistance to antiepileptic drug (AED) treatment is an important clinical problem in one third of patients with epilepsy: such refractory epilepsy is associated with increased mortality and morbidity. A growing body of evidence suggests innate drug transporters known to cause drug resistance in cancer cells, particularly members of the adenosine triphosphate (ATP)-binding cassette (ABC) protein superfamily, also may contribute to drug resistance in epilepsy. Thus members of the ABC superfamily ABCB1 (also known as MDR1 P-glycoprotein) and ABCC1 (also known as multidrug resistance–associated protein 1) are upregulated in glia and neurons in resected and postmortem human brain tissue containing pathologies causing refractory epilepsy (1–6). In animal studies, both ABCB1 and ABCC1 have been shown to transport a variety of AEDs, including phenytoin (PHT) (1,7–9), phenobarbitone (PB) (10), carbamazepine (CBZ) (8,9), lamotrigine (LTG), topiramate (TPM), and felbamate (FBM) (8,11). Other AEDs are likely to be ABCB1 substrates, as most AEDs share a planar lipophilic chemical structure.
Recently Rizzi et al. (9) showed that in rodent limbic epilepsy, seizures induce expression of ABCB1-encoding mRNA in hippocampus, and that such increased expression is associated with reduced hippocampal PHT levels. They also showed that repetitive treatment for 7 days with CBZ or PHT did not induce ABCB1 mRNA overexpression. We have shown in human brain that ABCB1 and ABCC1 protein upregulation in pathologies causing chronic epilepsy, including cortical dysplasia, is limited within a resection specimen to the region of the histologic abnormality and is not seen in the adjacent nonlesional brain tissue. We argued that this implies that chronic epilepsy and prolonged exposure to a variety of AEDs need not lead to upregulation of ABCB1 and ABCC1 in normal human brain tissue (6).
Status epilepticus is a potentially fatal manifestation of epilepsy. The outcome is worse in patients with a longer duration of status (12). Delay in the initiation of therapy may reduce the response to AED treatment in humans (13) and in animal studies (14). There are likely to be a number of reasons for poorer response to treatment with longer duration of status, including dynamic alterations in γ-aminobutyric acid subtype A (GABAA)-receptor function and systemic organ failure (12). We studied postmortem brain tissue from a patient with drug-resistant generalized tonic–clonic status epilepticus caused by cortical dysplasia to determine whether widespread upregulation of ABCB1 and ABCC1 occurred that could have contributed to drug resistance in status.
The study was approved by the Joint Research Ethics Committee of the Institute of Neurology, University College London. and the National Hospital for Neurology and Neurosurgery. We used archival brain tissue surplus to diagnostic needs.
Archival tissue was selected from an individual who had died in status epilepticus and was found to have cortical dysplasia. Seizure onset was at 7 years, with rare nocturnal generalized tonic–clonic seizures, treated with PHT and PB. A year before status, new complex partial seizures (facial grimacing, salivation, right-arm jerking, lasting 5–30 s) developed and became frequent, occurring six to eight per day, with nightly nocturnal generalized tonic–clonic seizures. Scalp EEG showed frequent spikes over the left hemisphere. Treatment latterly included valproate (VPA), CBZ, and clonazepam (CZP). At age 21 years, tonic–clonic status epilepticus occurred spontaneously 4 days before admission, leading to admission to another hospital and treatment with intravenous diazepam (DZP) and chlormethiazole infusion. Seizure frequency decreased temporarily to one per hour, with recovery of consciousness between. However, control worsened, and the patient was transferred to our centre. Further witnessed seizures occurred, characterised by forced eye deviation to right, and nystagmoid eye movements for 20–30 s. EEG showed diffuse slow activity with spike discharges, left more than right, with gradual build-up and sudden offset with nystagmoid eye movements. On computed tomography (CT) brain scanning, the left hemisphere was diffusely enlarged. Although seizures ceased for a short interval, they recurred, and a thiopentone infusion was given. EEG showed further subclinical activity, and the patient required intubation and ventilation with increased thiopentone infusion rate (8 mg/min). CSF analysis showed an opening pressure of 140 mm CSF with normal protein, normal CSF/blood glucose ratio, and no CSF pleocytosis. Seizures continued, and intravenous PHT was given. Further EEG showed continuing seizures; thiopentone was discontinued, and chlormethiazole was reintroduced, but with only transient abolition of seizures. Seizures continued for 11 days after admission despite continuing therapy with PHT, CBZ, and chlormethiazole. Monitoring showed no evidence of significant metabolic derangement, such as hypoxia or acidosis: the patient remained ventilated and comatose. Agonally on ventilation, hypotension and asystole occurred, followed by ventricular tachycardia, from which the patient could not be resuscitated. At postmortem, the left cerebral hemisphere was enlarged. On histologic examination, there was widespread cortical dysplasia (Fig. 1A) in the left frontal, temporal, and parietal lobes, with no abnormality detected in the right cerebral hemisphere. Brain tissue was archived and subsequently used in this study.
Antibodies and Immunohistochemistry
The methods used have previously been described in detail (6). For detection of ABCC1, monoclonal antibody MRPr1 (1:100 dilution; Alexis Corporation, Nottingham, U.K.) was used. The antibody has been well characterized and is believed to be specific for ABCC1. Paraffin-embedded sections of human kidney and choroid plexus epithelium, known to express ABCC1, were used as positive controls. ABCB1 was detected by using the well-characterized monoclonal antibody C494 (1:250 dilution; Alexis Corporation). C494 is known to cross-react with pyruvate carboxylase, but we previously showed similar patterns of brain immunoreactivity with C494 and another ABCB1 antibody, C219, directed against a separate ABCB1 epitope, suggesting that C494 brain immunoreactivity patterns are very likely to reflect ABCB1 presence (6). Normal liver was used as positive control for ABCB1 antibodies, whereas within experimental sections, labeling of capillary endothelium acted as a positive internal control. Immunohistochemistry was undertaken exactly as described previously, including positive and negative controls (6).
As expected, extensive immunoreactivity with both antibodies was seen throughout the areas of cortical dysplasia in the left cerebral hemisphere, which contained extensive cortical dysplasia (Fig. 1B and C). In the right cerebral hemisphere, which was normal on routine pathologic examination, extensive white matter glial immunoreactivity for both ABCB1 (Fig. 1D) and ABCC1 (Fig. 1E) was seen. Not all glia in either hemisphere were immunolabeled. Neurons in the right cerebral hemisphere appeared morphologically normal, and none was immunoreactive with either antibody. Endothelium in all sections was immunoreactive for ABCB1.
In this case report, we have shown widespread upregulation of drug-transporter molecules ABCB1 and ABCC1 in glia throughout both cerebral hemispheres, even in histologically normal-appearing brain that did not contain microscopic epileptogenic pathology. The availability of the normal right hemisphere allowed examination of histologically normal brain tissue, as with extensive cortical dysplasia, it can be difficult to be certain that tissue from the affected hemisphere that is not clearly dysplastic is entirely normal. The observed immunopositivity in the normal hemisphere contrasts with findings from patients in whom surgical resection specimens containing focal cortical dysplasia were studied: in those cases, normal cortex and white matter beyond the histologic abnormality did not show upregulation of either ABCB1 or ABCC1, even though such tissue was adjacent to the epileptogenic pathology and had been exposed to a variety of AEDs (6). Therefore either status epilepticus or its drug treatment, in this case consisting of thiopentone, PHT, and chlormethiazole, can cause widespread upregulation of ABCB1 and ABCC1 in normal-appearing brain tissue. Although PB may upregulate ABCB1 in cell lines (10), prolonged treatment with PHT or PB in humans (1,6) or repetitive short-term treatment in animals with PHT or CBZ (9) does not appear to cause upregulation of ABCB1 in normal tissue. No information is available for chlormethiazole, although this drug is not known to induce other enzymes. In parallel with animal studies (9), therefore, it would appear that status epilepticus can induce widespread ABCB1 and ABCC1 upregulation, even in histologically normal tissue.
The time course of upregulation of ABCB1 in this case cannot be determined, and it is unlikely that such determination will prove possible in cerebral parenchyma in humans. In mice, increased expression of ABCB1 mRNA was noted as early as 3 h after ≥90 min of recurrent seizures, and when tested at 6 h, increased ABCB1 levels were associated with significantly lower hippocampal PHT levels (9). In rats, increased ABCB1 levels in hippocampal astrocytes were noted by 24 h after kainate injection (15). In both studies, upregulation persisted for ≥10 weeks after the initial insult.
Our findings support those of other workers that seizures can induce ABCB1 expression. With accumulating evidence that upregulation of ABCB1 and ABCC1 in cerebral parenchyma contributes to drug resistance in epilepsy, it is possible that acute upregulation of drug-transport proteins contributes to poorer response to AED treatment with increased duration of seizures in status epilepticus. We propose that the widespread upregulation of ABCB1 and ABCC1 leads to a reduction in brain parenchymal penetrance of AEDs, in keeping with the evidence that these proteins transport AEDs (7–11). The findings emphasise the need to treat status epilepticus urgently, and also raise the possibility of additional treatment options in refractory status epilepticus.
Acknowledgment: This work was supported by grants from the Institute of Neurology and the National Hospital for Neurology and Neurosurgery, the Epilepsy Research Foundation, U.K., the Patrick Berthoud Trust, and University of London Central Research Fund. We thank W-R. Lin and Steve Durr for excellent technical assistance.