This study demonstrates persistent HHV-6B infection in most patients with MTS-MTLE, but no detectable infection in patients with other pathology and constitutes the first report of primary isolation and maintenance of virus-infected astrocytes from the human brain. In addition, we provide direct evidence for an etiological link of a ubiquitous human herpesvirus in a subset of patients with intractable MTLE with MTS, a syndrome of unknown origin. Active HHV-6 infection was confirmed by the detection of viral DNA, mRNA viral transcripts, and viral protein expression in human primary astrocytes isolated and cultured from resected brain tissue from patients with MTLE. HHV-6B in MTS-MTLE is unlikely to be a consequence of nonspecific inflammation or seizures, since none of the patients without MTLE patients, all of whom had intractable epilepsy, were positive for virus. Moreover, there was no evidence for inflammatory changes in any of the HHV-6B–positive patients.
Studies associating HHV-6 with neurologic disorders, including epilepsy, are based on detection of viral DNA, RNA, and antigen, suggesting pathogenic reactivation of latent HHV-6. Establishment of HHV-6 latency in the CNS can follow primary childhood infection [37
]. Although HHV-6 reactivation in the brain appears to occur predominately in immunosuppressed adults and is associated with neurological complications following bone marrow and stem cell transplantation, viral reactivation may also play a role in disorders such as multiple sclerosis and epilepsy. Several studies have demonstrated the presence of low levels HHV-6 DNA (typically nested PCR) in normal human brains that can range between a frequency of 0%–75% [15
]. These discrepancies can be attributed to different patient populations, methods of detection, and sensitivities and specificities of reagents. Studies demonstrating HHV-6 DNA in the normal human brain have typically used nested PCR, indicating that levels of viral DNA are low, supporting the observations by Blumberg et al. that HHV-6 may be present as a latent commensal pathogen that is able to infect the human brain without causing any apparent neurological disease [41
]. The result in this study demonstrating high levels of viral DNA detected in MTLE specimens by non-nested TaqMan PCR, together with the presence of viral antigen and RNA, suggests association with disease. Detection of HHV-6 DNA by PCR in 25%–50% of patients with temporal lobe epilepsy indicates an association of viral reactivation with a neurological disorder that may be independent of immunosuppression [15
]. The results in this study demonstrated higher HHV-6 levels in hippocampal tissue than in the surrounding temporal neocortex. High viral loads in the hippocampus from patients with MTLE could reflect viral reactivation in a specific brain region associated with latency resulting from early childhood infection.
Alternatively, persistent noninflammatory encephalitis may also lead to the development of epilepsy. Several lines of evidence, including clinical, imaging, and neuropsychological data, suggest that MTS/MTLE is a progressive disorder [45
]. Imaging studies have shown that febrile seizure history and epilepsy duration are associated with increasing hippocampal atrophy, independent of seizure frequency [24
]. A history of febrile seizures may be difficult to establish retrospectively, many years after the event. In our overall cohort of 38 patients, which represents our collective experience at NIH and the CNMC (), five of nine patients with a definite history of febrile seizures (all in the MTS-MTLE group), compared with nine of 28 reporting no history of febrile seizures, were positive for HHV-6B (a clear history could not be obtained for one HHV-6B–positive patient with MTS-MTLE), suggesting a trend towards HHV-6B–positive patients with MTLE having febrile seizure history. The long latency between childhood febrile seizures and the appearance of persistent unprovoked seizures suggests these patients may have chronic HHV-6 infection rather than reactivated virus. The presence of chronic viral infection in these patients would be supported by the progression of hippocampal atrophy. Collectively, these data suggest an ongoing process; the latency between occurrence of an early risk factor such as febrile seizures and onset of chronic epilepsy is consistent with either persistent or reactivated infection [25
We also had the unique opportunity to follow longitudinally a patient who experienced recurrent seizures after three focal resections and a hemispherectomy for a period of 11 mo. Consistent detection of high levels of HHV-6B DNA in each of the three resections and in frontal and temporal/parietal lobes following hemispherectomy suggests this patient had a persistent and widespread infection with HHV-6B. Following hemispherectomy, this patient has been seizure-free without anticonvulsant treatment for more than 1 y. It is intriguing to speculate that these data demonstrate a widespread and persistent HHV-6 infection that was associated with epilepsy, and that once the viral infection was removed, the patient became seizure-free.
An association between chronic epilepsy and persistent or reactivated HHV-6 infection of astrocytes suggests the possibility that viral infection of astrocytes are associated with changes in cell function that may contribute to disease. Astrocytes are known to interact closely with neurons and are critical in modulating synaptic transmission [49
]. Astrocytes can modulate neurotransmission by maintaining low concentrations of extracellular glutamate by the glial glutamate transporters EAAT-1 and EAAT-2 [50
]. Elevated extracellular glutamate, the main excitatory neurotransmitter, may be involved in epilepsy by triggering excitotoxicity through loss of glutamine synthetase [51
], an enzyme that metabolizes glutamate in astrocytes, and/or by malfunctioning astrocytic glutamate transporters [50
]. The CA1 and CA3 neurons lost in MTS/MTLE are particularly susceptible to glutamatergic-mediated cell death. Sclerotic hippocampi from temporal lobe epilepsy demonstrate reduced EAAT-2 immunoreactivity [52
], and are prone to alternative EAAT-2 mRNA splicing [53
]. A unique finding in this study is the isolation ex vivo of cultured astrocytes from patients with MTLE who are infected with HHV-6. These primary HHV-6–infected astrocytes demonstrated low levels of EAAT-2 mRNA. In support of our ex vivo findings, astrocytes infected with HHV-6 in vitro also demonstrated a remarkable decrease in EAAT-2 mRNA. Detection of high levels of HHV-6 DNA in MTLE brain tissue, isolation of HHV-6 from primary astrocytes isolated from MTLE brain tissue, and decreased expression of EAAT-2 mRNA demonstrates an association between HHV-6 infection and astrocytic dysfunction. Functional changes in virus infected glia or in glia harboring reactivated virus may lead to secondary injury of the exquisitely sensitive hippocampal neuron, and ultimately to development of MTLE and epilepsy. The potential relationship between HHV-6 astrocytic infection and MTLE deserves further investigation.
Overall, our cumulative experience at the NIH Clinical Center and the CNMC consists of a large cohort of patients with epilepsy, from which 60% with clinically defined MTLE had detectable HHV-6B sequences in surgical brain resections. The highest HHV-6 viral loads were demonstrated in the hippocampus. This appeared to be specific for MTLE, since non-MTLE epilepsy material had no detectable levels of HHV-6 DNA by TaqMan or nested PCR. The statistically significant correlation of HHV-6 and MTLE () in a large cohort of patients with epilepsy and dysregulation of glutamate transporter expression by HHV-6 suggests a novel pathophysiological mechanism of disease.