The Baboon in Epilepsy Research: Revelations and Challenges

Abstract

The baboon offers a natural model for genetic generalized epilepsy with photosensitivity. In this review, we will summarize some of the more important clinical, neuroimaging and elctrophysiological findings form recent work performed at the Southwest National Primate Research Center (SNPRC, Texas Biomedical Research Institute, San Antonio, Texas), which houses the world’s largest captive baboon pedigree. Due to the phylogenetic proximity of the baboon to humans, many of the findings are readily translatable, but there may be some important differences, such as the mutlifocality of the ictal and interictal epileptic discharges on intracranial EEG and greater parieto-occipital connectivity of baboon brain networks compared to juvenile myoclonic epilepsy in humans. Furthermore, there is still limited knowledge of the natural history of the epilepsy, which could be transformative for research into epileptogenesis in GGE and sudden unexpected death in epilepsy (SUDEP).

1. Photosensitivity, or the predisposition to seizures provoked by visual stimuli, was accidentally encountered in a baboon subspecies, Papio hamadryas papio (or P. h. papio) in Senegal.

Robert Naquet, with his collaborators Eva and Keith Killam, first discovered and characterized photosensitivity in the P. h. papio, as well as in other baboon subspecies housed at the Southwest National Primate Reserch Center (San Antonio) [1,2]. Over the course of more than three decades, Naquet and his associates evaluated the electrophysiological pathways and neurochemical mechanisms involved in the photoepileptic responses, with the goal to develop a platform for drug development and testing [3,4,5,6,7]. Much of his pioneering work seemed to corroborate the role of frontothalamic networks in the generation of generalized ictal and interictal epileptic discharges as evoked by intermittent light stimulation (ILS). Nonetheless, despite elaborate neuroanatomical studies and thorough electrophysiological approaches, the connections between visual cortices and motor cortices were elusive. One challenge for studying photosensitivity in this model may have been the limited reliability for activation of the photoconvulsive responses, which led to the premedication of baboons with allyl-glycine, both photosensitive and non-photosensitive; by reducing the production of gamma-aminobutyric acid, or GABA, ally-glycine lowers the seizure threshold and facilitatesphotoconvulsive responses [4,5,7]. However, ally-glycine can also trigger seizures from alternative regions or obscure the spread of the epileptic discharge to the frontorolandic cortices. Naquet did foresee the potential importance of neuroimaging to elucidate mechanisms underlying epilepsy in the baboon and embraced positron emission tomography as a research tool [8]; however, the tools for evaluating the structural and dynamic markers of generalized epilepsies were still limited in his time.

The baboon’s stature as a model for photosensitivity, and its suitability for the development and testing of antiseizure medications, declined over time, in part because of the limited availability of baboons, their higher cost and more expensive housing and maintenance, as well as general ethical concerns related to using nonhuman primates in research. More importantly, technological development allowed adaptation of less expensive rodent and murine models for epilepsy research and drug testing. Several questions remained unanswered as intertest waned; Naquet’s group had not established visuomotor connections underlying the photoepileptic response and sought genetic markers of photosensitivity [9]. And, as a model of photosensitivity, the epileptic baboon nevert gained recognition as a model for GGE.

The purpose review is to shed light on some misconceptions related to the baboon as a model for photosensitivity and GGE, and to present the advantages and challenges of the baboon in epilepsy research. The areas of research which offer the greatest promise lie in the improved understanding of the natural history of GGEs, in particular with regards epileptogenesis and sudden unexpected death in epilepsy (SUDEP), and as a model to evaluate and test devices developed for seizure detection or neurostimulation (see companion report in this issue).

2. Photosensitivity is Occurs in Most Baboon Subspecies.

While photosensitivity was always mentioned in relationship to “Papio papio”, it affects other baboon subspecies as well. Indeed, P. h. papio was the first baboon subspecies identified with photosensitivity, with a large prevalence in Senegal [1,5]. The prevalence of photosensitivity showed regional variations in P. h. papio, as high as 60% from the Casamance or Tiaffene regions from where Naquet acquired his study animals [5,10]. The regional differences in photosensitivity was the first indication of a potential genetic influence [10]. As most of his studies utilized P. h. papio, the historical association of photosensitivity with this subspecies persisted. The SNPRC houses the largest captive pedigree in the world, consists mainly of P. h. anubis, P. h. cynocephalus, and their hybrids, as well as P. h. hamadryas, mainly from East Africa [11]. The first scalp EEG studies at the SNPRC were in unsedated animals, comparing the rate of photosensitivity among different subspecies, including for P. h. anubis and P. h. cynocephalus, but confirming the higher susceptibility for photosensitivity in P. h. papio [2]. In more recent scalp EEG studies at the SNPRC, albeit using low-dose ketamine for sedation, which is known to activate interictal epileptic discharges (IEDs) [12,13], the overall prevalence of photosensitivity was found to be 40% in the pedigree consisting mainly of Papio h. anubis and anubis/cynocephalus hybrid subspecies, but also noted in several Papio h. hamadryas [13].

3. The Epileptic Baboon Models Genetic Generalized Epilepsy in Humans.

Over 700 baboons were electroclinically characterized in the captive pedigree housed at the SNPRC, and 26% were witnessed to have generalized tonic-clonic seizures (GTCS) in the colony [14]. All of these seizures were witnessed in the daytime, and 60% in the morning hours. However, in a recent study of seven baboons being monitored for a median of 2 (range 1-20) weeks by video only, we recorded 46 GTCS, 89% of which occurred nocturnally or upon awakening, indicating that the prevalence of seizures could be underestimated in the colony [15]. Over 90% of the epileptic baboons presented with GTCS by age 6 years old, the last year of their adolescence [16]. As mentioned above, the use of ketamine as a sedative for scalp electrode placement may have reduced seizure thresholds during the electroencephalography (EEG) studies, myoclonic seizures were noted in almost all of 217 baboons with seizures recorded [13]. All of the myoclonic seizures affected the eyelids and face, but in one third of the baboons, they also involved the trunk and extremities. Myoclonic, absence and GTCS were also recorded in baboons implanted with intracranial electrodes, combining depth, as well as subdural grid and strip electrodes [17]. Surprisingly, in addition to the generalized seizures, subclinical seizures were also recorded with multifocal onsets, originating from the visual cortices as well as the frontal and parietal association cortices. In a recent characterization of spontaneous GTCS in baboons, focal features were apparent in 41%, mainly characterized by gyratory seizures to one side, or sequentially to both sides, prior to evolution of the tonic-clonic semiology [15]. Focal motor, including gyratory seizures, as well as IEDs, also occur in juvenile myoclonic epilepsy [18,19,20]; nonetheless, there are limited intracranial recordings in human generalized epilepsies, and the generalized ictal and interictal discharges are usually not adequately sampled or characterized [21,22]. Finally, a preliminary analysis of 1400 pedigreed animals at the SNPRC demonstrated heritability (h²) of spontaneous seizures was 0.33 (p<1×10⁻⁷) and of IEDs to be 0.19 (p<0.002), which suggests a substantial genetic effect in seizure liability [16], and, at least one gene, RBFOX1, has been identified to be significantly linked to the diagnosis of epilepsy (Kos and Szabó, personal communication), establishing the epilepsy syndrome as genetic in etiology. RBFOX1 is a candidate gene in human GGE as well [23], which potentially regulates cerebral cortex development [24], crosstalk with microRNA miR-129–5p that impacts homeostatic downscaling of excitatory synapses [25], and affects transcriptomic expression and splicing patterns of neuronal genes, some which code for channels or receptors implicated in genetic epilepsies [26,27].

4. Cortical Brain Regions Implicated in the Generation of Ictal and Interictal Epileptic Discharges Extend Beyond the Frontothalamic Networks.

Naquet’s research added further credibility toward the theory that the frontothalamic network (motor cortices, medial frontal cortices) was the generator of generalized seizures associated with GGE [3,5]. This theory persisted in the literature, despite the electroclinical differences between seizure types, and a proposed genetic etiology, which would likely affect brain regions outside of the frontothalamic networks. The intracranial EEG and functional neuroimaging studies performed by our group demonstrated a more diffuse pathophysiology, implicated multiple brain region as and networks [17,28,29]. While generalized IEDs involving the motor cortices predominated, focal discharges also arose from the orbitofrontal, parietal and primary and secondary visual cortices, and less frequently in the opercular regions [17]. Intracranial recordings not only demonstrated multifocal subclinical seizures, but also focal parietal or frontal discharges that triggered some clinical seizures. Correlation of IED rate with regional cerebral blood flow on H₂¹⁵O-PET demonstrated an epileptic network that included the visual cortices, parietal association cortices, insula, as well as the motor and premotor cortices [28]. BOLD-fMRI demonstrated increased connectivity of most brain networks to the parieto-occipital cortices [29]. More diffuse cortical involvement was also demonstrated by evidence of neuron loss in the frontal, parietal and temporal association cortices on flow cytometry studies [30]. It is unclear why the cortical changes encountered in the pedigree housed at SNPRC were not evident in the Papio h. papio studied by Naquet’s group, begging the question: were these changes driven by interspecies and genetic differences between the baboons, or by differences between research hypotheses and invasive EEG recording technologies? The approach of combining of depth, and subdural grid and strip electrodes, may have provided better cortical EEG sampling but also led to irritation of underlying cortices to produce subclinical ictal and interictal epileptic discharges. Nonetheless, this approach enabled the documentation cortico-cortical propagation of photoparoxysmal responses from the occipital to frontorolandic cortices [17], which was not evident with by stereotactic EEG [9]. Functional neuroimaging data acquired in epileptic baboons further corroborated these findings [28,29].

Functional neuroimaging, including blood oxygen level dependent (BOLD) functional MRI (fMRI) and magnetoencephalography (MEG), is the only feasible approach to studying GGE networks in humans [31,32,33,34]. Phase clustering in the high-frequency gamma range is enhanced by ILS prior to photoepileptic responses; this is most dramatic in parieto-occipital cortices during photoparoxysmal responses, and in the frontocentral cortices and parietal lobes before myoclonic and and parietal lobes before absence seizures [31]. BOLD activations were predominant in the parieto-occipital regions in most photoparoxysmal responses, resulting in the activation generalized spike-and-wave discharges [32]. Before the onset of generalized spike-and-waves discharges and absence seizures, BOLD fMRI demonstrated activations and MEG increased fast frequency activity in the precuneus and occipital lobes, respectively [33,24]. Evidence from praxis-induced generalized ictal and interictal dsicharges [35] as well as the aforementioned MEG studies demonstrating increases in posterior gamma activity in the parietal and occipital cortices [31,34], suggested that increasing synchronization of high-frequency activity across brain regions can prime, or recruit, the frontothalamic network [36]; analogous to a tea kettle, the subsequent discharge of the frontothalamic network may be the most efficient pathway for the brain to relieve the excessive cortical excitation. Evidence of focal sharp waves on scalp EEG recordings in human GGE [18,19,20] and focal ictal semiologies [19,20,37,38] also support a more diffuse cortical pathophysiology underlying GGE in humans.

5. Future Research.

One of the potential advantages of the large baboon pedigree housed at SNPRC is that it provides an excellent resource to study the natural history of GGE. The epileptic baboon demonstrates seizure onset in childhood and adolescence, which is consistent with human GGE, only that in baboons, these developmental phases are compressed within a six-year period instead of over two decades [13,14]. Using pedigree data and/or better genetic markers, subpopulations at greatest risk for epilepsy could be prospectively followed to identify potential epigenetic, behavioral, molecular and neuroimaging markers predicting the eventual onset of clinical epilepsy. In other words, studying epileptogenesis in baboon model of GGE can be achieved within the framework of a five-year grant mechanism. And once baboons develop epilepsy, these biomarkers, along with genetic or acquired autonomic markers, could be re-examined prospectively to identify predictors for SUDEP [39].

The epileptic baboon can also be utilized for device studies, namely for seizure detection and neurostimulation (see companion study). Non-invasive neuromodulatory epilepsy treatments—such as transcranial magnetic stimulation (TMS)—may also be explored in this model [40,41,42]. This model also provided preclinical confirmation of the efficacy and safety of high-frequency microburst VNS Therapy in two baboons monitored for 24 weeks [43]. Diurnal, even catamenial, seizure patterns in the epileptic baboons are similar to people with GGE, and JME in particular [15,43]. Furthermore, knowledge of the natural history of epilepsy can provide the appropriate context for long-term investigations of seizure detections and therapeutic interventions. Social, cognitive and behavioral outcomes can also be evaluated in this model with direct translatability to humans.

6. Conclusions:

Recent electrophysiological, neuroimaging and histopathological studies in the epileptic baboon suggest that it is not only an excellent model for photosensitivity but also for GGE. This model also demonstrates a more diffuse pathophysiology underlying GGE, which is not restricted to the frontothalamic networks. Insights into epileptogenesis and SUDEP will require a better understanding of the natural history of GGE in the baboon. Finally, disease severity, such as GTCS frequency, can be correlated with electrophysiological and neuroimaging findings, even neuropathological changes, to explore the mechanisms underlying the biological and neurobehavioral effects of chronic GGE.

Highlights:

• The epileptic baboon is a model for genetic generalized epilepsy, with a 40% prevalence of photosensitivity

• Intracranial EEG demonstrates generalized and multifocal ictal and interictal epileptic discharges in frontal, parietal and occipital cortices.

• Functional neuromaging also confirms a generalized epileptic network associated with parieto-occipital connectivity

• The baboon provides a model for electrophysiological, neuroimaging, autonomic and neurobehavioral effects of chronic GGE