Skip to Main Content

Ognen Petroff

Associate Professor Emeritus of Neurology

Contact Information

Ognen Petroff, MD, FAAN (Neurology), FACNS, FAES, FANA

Office Location

Mailing Address

  • Epilepsy & Seizures

    Yale SCHOOL OF MEDICINE, Department of Neurology, Division of Epilepsy and EEG, 15 York Street, P.O. Box 208018

    New Haven, Connecticut 06520

    United States

Research Summary

Epilepsy is a common neurological disorder affecting two to three million Americans. We have focused our research on brain glutamate, glutamine and GABA metabolism in epilepsy. Glutamate is the primary excitatory and GABA the primary inhibitory neurotransmitters. Alterations in both glutamatergic and GABAergic systems have been strongly implicated as the final common pathway in many epilepsy syndromes.

Audio-video-electroencephalography (AV-EEG) monitoring remains the primary tool in the evaluation and management of provoked seizures, first seizure, epilepsy, status epilepticus and seizure mimics. Advances in computerized EEG recording and analysis have revolutionized the ability to monitor for electrographic seizures and EEG disturbabces that often herald seizures.

However, the accurate prediction of when a seizure is likely to happen requires further study. My research interest is focused on the metabolic and EEG changes that occur prior to the onset of a seizure.

Specialized Terms: Epilepsy & seizure disorders; Cerebral metabolism; Brain glutamate, GABA & glutamine metabolism; Antiepileptic drug actions & brain metabolism; CNS physiology & neurochemistry; Neuroprotection; periodic discharges, non-convulsive seizures

Extensive Research Description

The significance of understanding
the relationship between GABA levels and cortical excitability.

cortical excitability is a hallmark of several human epileptic syndromes,
facilitating the spread of the seizure in the localization related epilepsies
and as a final common pathway for some of the primary generalized and myoclonic
epilepsies. Photosensitive epilepsy is the most common of the
stimulus-triggered (reflex) epilepsies, and there is considerable evidence,
developed primarily in animal models, that it is associated with impairments in
GABAergic inhibition, possibly secondary to alterations in GABA metabolism.
GABA is the major inhibitory neurotransmitter and has a crucial role in
regulating cortical excitability. The concentration of GABA in both the
vesicular and cytosolic pools may have an important influence on GABAergic
inhibition. Our development of non-invasive magnetic resonance spectroscopy
(MRS) based measurements of cellular GABA in human cerebral cortex allowed us
to study GABA metabolism in human epilepsy. Our studies found profound
decreases in the level of cortical GABA in many patients with complex partial
seizures (CPS) and juvenile myoclonic epilepsy (JME). Our preliminary studies
suggest that GABA synthesis rates are decreased in patients with low GABA in
functional occipital cortex remote from the ictal onset zone. We found that, light-deprivation,
a condition that enhances excitability in the visual cortex, decreases GABA
levels in non-epileptic subjects in parallel with increased excitability measured
with paired-pulse stimulation using transcranial magnetic stimulation. These
results support an important role for cortical GABA level on cortical
excitability thereby facilitating the spread of seizure discharges from the
ictal onset zone to involve more normal functional brain. The primary goal is
to better understand the mechanisms by which GABA levels influence cortical
excitability, and the role of altered GABA levels in the etiology and
pathogenesis of epilepsy. Our primary general hypothesis is that reduced
cellular GABA levels result in an increase in cortical excitability through
decreased GABA release.

The significance of the relationship
between GABA metabolism and cortical excitability.

There is a
renewed awareness of the interaction between metabolism and excitability as the
relationships between different aspects of neuronal-glial neurotransmitter
cycling are established. In order to fully understand the physiologic
consequences on GABAergic inhibition of altering neurotransmitter cycling, our approach
was to start at the final step and work backwards. We used a variety of
physiologic and analytic methods to examine the consequences of glutamic acid
decarboxylase (GAD) inhibition on network excitability in the rat hippocampus. We
used field and intracellular recordings from the CA1 region of rat hippocampal
slices to determine the physiological effects of blocking GABA synthesis with
the convulsant, 3-mercaptoproprionic acid (MPA). We measured the rate of
synthesis of GABA and glutamate in slices using 2-13C-glucose as a label source
and liquid chromatography-tandem mass spectrometry. The primary goal was to
establish whether GAD inhibition alone was sufficient to induce network
hyperexcitability in an isolated preparation comparable to that seen with GABA
receptor inhibition. The key findings were that inhibition of GABA synthesis
via GAD only produces hyperexcitability following repetitive stimulation,
demonstrating a pronounced use dependence when studied both intra and
extracellularly. This is consistent with the data showing that the regulation
of IPSC amplitudes and of GABA release can be regulated by glutamine
availability in use-dependent fashion. In addition, our data with NO-711, which
blocks neuronal GABA uptake, indicated that under baseline conditions, neuronal
GABA uptake does not provide a significant degree of GABA that is available for
release under resting conditions, but does once frank hyperexcitability has
been produced by repetitive stimulation. Despite this clear disinhibition, we
were unable to observe seizure-like activity comparable to that seen with GABAA
receptor blockade until basal excitability was increased with elevated extracellular
potassium concentrations. Finally, our data demonstrated that there is no
significant change in the size of the GABA pool, even under conditions where
synchronized activity is present. However, as expected, GAD inhibition with MPA
significantly decreased the amount of newly formed GABA. These biochemical data
demonstrated that full blown network excitability is seen only with a
combination of >70% decrease in newly synthesized GABA with a concomitant
moderate elevation in neuronal excitability. Thus, these data are comparable to
those in circuit modeling studies showing that GABAergic inhibition is highly

Understanding the role of
homocarnosine in modulating human cerebral excitability.

One aspect
of GABA biochemistry that has not received significant attention is the role of
GABA-containing compounds. One such compound is homocarnosine, a dipeptide
formed from histidine and GABA. The role
of homocarnosine in the human brain is poorly understood. Homocarnosine has
been proposed as an inhibitory neuromodulator, which is hydrolyzed into GABA
and histidine in the extracellular fluid (ECF), thereby increasing GABAergic
activity. Consistent with this proposal, studies of refractory human epilepsy
using vigabatrin show that CSF homocarnosine concentrations are significantly
higher in patients whose seizure control improved than in those who failed to
benefit. Free GABA concentrations were the same. Our own observational studies
of human epilepsy suggest that increased cortical levels of intracellular
homocarnosine appear to be associated with decreased cortical excitability. Low
intracellular homocarnosine and GABA levels, measured in the occipital lobe
using magnetic resonance spectroscopy (MRS), are associated with frequent
complex partial seizures in patients treated with valproate or lamotrigine. Patients
with juvenile myoclonic epilepsy with excellent seizure control and treated
with the same drugs usually have high homocarnosine levels, but very low
intracellular GABA levels. What is unknown is whether higher levels of intracellular
homocarnosine are a characteristic of patients with primary epilepsies, perhaps
compensating for the low intracellular GABA levels and contributing to better
seizure control by decreasing cortical excitability.

Topiramate, gabapentin and
levetiracetam are three new drugs that increase cortical homocarnosine
. The
mechanisms through which these drugs increase human homocarnosine levels are
unknown. Unlike vigabatrin, none of these three drugs alter intracellular GABA
concentrations in rodent models Topiramate and gabapentin increase human
cortical GABA levels within two hours of the first dose and homocarnosine
levels rise after one day (with topiramate) to one week (with gabapentin) of
daily use. Levetiracetam was studied only after two weeks of treatment. Our
studies using three antiepileptic drugs show that patients with refractory
complex partial seizures with better seizure control had higher homocarnosine
levels than those with poor seizure control. Cortical intracellular GABA levels
were the same in patients, who responded, compared with those, who failed to
benefit. The findings suggest, but do not prove, that homocarnosine may
decrease cortical excitability. The alternative explanation would be frequent
seizures either decreases the synthesis of homocarnosine or enhances its
catabolism. Taken as a whole these findings suggest, but do not prove, that
homocarnosine may decrease cortical excitability.

Homocarnosine appears to directly
decrease neuronal excitability of the human hippocampus.

studies have been carried out in whole animal and cell culture models, which
may not apply to humans because of the very low endogenous levels of
homocarnosine present in non-primate models. However, under conditions when
GABA is elevated, abundant homocarnosine is found in the rodent brain and CSF,
demonstrating that it can be synthesized de novo in the rodent. Moreover, biochemical studies showed that
homocarnosine blocks GABA uptake into synaptosomes, suggesting that homocarnosine
has the potential to be an endogenous modulator of GABAergic function. Our preliminary data using hippocampal slices
incubated in ACSF indicate that homocarnosine decreases neuronal excitability
in both rat and human hippocampus. We studied the effects of bath-applied
homocarnosine using electrophysiological recording techniques in the CA1 region
of rat hippocampal slices. We found that both homocarnosine and GABA shifted
the input-output relationship for evoked synaptic responses rightward and thus
was inhibitory. GABA and homocarnosine did not have additive effects,
suggesting a common mechanism of action. However, in intracellular recordings,
homocarnosine had no significant effects on the membrane potential and slightly
decreased the input resistance of the cells. Therefore, the dipeptide is not a
direct GABAA agonist.

An excess of extracellular glutamate
in the sclerotic hippocampus may be one of the key molecular causes of seizures
and brain damage in human mesial temporal lobe epilepsy
. Research by the Yale Epilepsy Clinical
Research group revealed that several important characteristics of the
epileptogenic human hippocampus include: above normal interictal extracellular
glutamate levels and enhanced glutamate release during spontaneous seizures
with abnormally slow post-ictal glutamate clearance. Paradoxically, interictal
extracellular glutamate concentrations were considerably higher in patients
with hippocampal sclerosis (MTLE) than in patients without this pathology (non-MTLE),
despite the 60–80% neuronal loss and doubling of glial density in the sclerotic
hippocampus. Surprisingly, a considerable (35–40%) loss of glutamine
synthetase protein and activity was demonstrated in astrocytes of the epileptogenic
hippocampus resected from
patients with MTLE. Isotopic tracer (13C-glucose) studies during epilepsy
surgery suggested that the accumulation and impaired clearance of glutamate in
MTLE is due to a slowing of the glutamate–glutamine cycle metabolism in the
sclerotic hippocampus compared with the non-gliotic epileptogenic hippocampus. Based
on these data we hypothesize that elevated extracellular glutamate is a
consequence of impaired glial function due to both decreased rates of glial
uptake and/or metabolism. The isotopic enrichment of microdialysis (extracellular)
glutamine was higher for probes with nearly normal glutamate after infusion of
labeled substrates, which suggests the rate of glutamine synthesis, thus
glutamate detoxification, is lower in those areas of the brain with above
normal extracellular glutamate. The data obtained with an infusion of
13C-glucose or 13C-acetate on different days in the same subjects were the
same, which increased our confidence in the findings. The variation among
multiple probes in the same patient reflected regional variation in glutamine
synthesis and extracellular glutamate. Glutamine synthesis is limited to glia;
therefore, our data suggested glial dysfunction in regions with above normal
extracellular glutamate concentrations.


Research Interests

Brain; Cerebellum; Epilepsy; Glutamates; Metabolism; Neurochemistry; Neurology; Physiology; Seizures; Health Care

Selected Publications