Jump to ContentJump to Main Navigation
The Neuroethics of BiomarkersWhat the Development of Bioprediction Means for Moral Responsibility, Justice, and the Nature of Mental Disorder$

Matthew L. Baum

Print publication date: 2016

Print ISBN-13: 9780190236267

Published to Oxford Scholarship Online: March 2016

DOI: 10.1093/acprof:oso/9780190236267.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2018. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 25 September 2018

(p.171) Appendix II Seizure Prediction

(p.171) Appendix II Seizure Prediction

Source:
The Neuroethics of Biomarkers
Author(s):

Matthew L. Baum

Publisher:
Oxford University Press

Approximately 1–2 of 20 people will experience a seizure at some point in life (Bragin et al. 2010; Nestler et al. 2009). Particularly pertinent to our investigation here is the presentation of first seizures in those persons of driving age; this is the category that might feasibly fall into Probability’s shoes. Of those presenting with first seizures in Australia, for example, 60 percent are above 16 years of age, the age at which persons are eligible to enter with supervision the driver’s seat of a car; new onset epilepsy is the most common cause of first seizure with 50 percent of those with a first unprovoked seizure experiencing another in the next 8 years (Adams & Knowles 2007).

Genetics

Although we are concerned with adolescent/adult onset events, it is worth starting off by mentioning that about 10 percent of early-onset forms of absence epilepsy are attributable to mutations in the gene encoding glucose transporter type 1, SLC2A1 (Pal et al. 2010). The genetic components of other epilepsies are less well categorized, and it is likely that most alter seizure thresholds.1

What is reasonably clear, however, is that interfering with the proper expression of a range of neuronal-excitability-related genes can lead to seizures and epilepsy. More than 20 strains of knockout mice, for example, develop spontaneous seizures2 (Nestler et al. 2009). Most of these mutant strains exhibit seizure onset before mid-adolescence (and are thus less interesting to us here), but several mutants exhibit late-adolescent to adult onset: Weaver (g-protein gated inward rectifier girk2) has adult onset; Centromere bp-b (brain specific DNA binding protein) has onset at 3–4 months; Synapsin 1,2 has onset after 2 months. In these mice raised in standard laboratory conditions, the genotype can predict seizure onset. No human knockout of this type has been described. There is, however, a possibility that future research will describe risk variants in humans.

(p.172) Along these lines, and perhaps more relevant in our case, there exist a large number of mouse mutants that exhibit decreased seizure threshold and susceptibility: these are in the signaling proteins, Tyn tyrosine kinase receptor, tPA, and BDNF. It is thought that the absence of these proteins upsets the balance of excitation and inhibition in key areas of the brain, leading to seizure susceptibility (Nestler et al. 2009). A decreased seizure threshold becomes important when we remember that we bathe our brains in insults that at any given moment may tip the scales of excitation and inhibition; if the scales tip too far by this chaotic process, a seizure might ensue. In other words, differences in seizure susceptibilities at population levels will translate to differences in seizure incidence.

To highlight even more directly the combinatory requirement of biology and environment, we can consider the following cases. One mutation called γ‎2R43Q in the GABAa receptor, a major inhibitory receptor in the brain, decreases the seizure threshold in adult mice; because even temporary changes in GABAa receptors early in life can have lasting impact on seizure threshold in adulthood (Chiu et al. 2008) and because temperature increase impairs this particular mutant receptor’s activity (Galanopoulou 2010), it raises the possibility that childhood fever might permanently alter the seizure threshold in those individuals who possess this variant. Mice with a mutation in the nicotinic acetylcholine receptor (L247T) display a higher sensitivity specifically to nicotine-induced seizures than mice without the mutation (Steinlein & Bertrand 2010). Finally, head injury in rodents leads to differential seizure-threshold lowering depending on genetic background (Pitkänen et al. 2011), which suggests that the rodents’ specific genetic makeup can provide resilience against or susceptibility to environmental insults.

Because the electrical activity of a seizure seems to depend on the balance of excitation and inhibition across key circuits, much attention has recently been paid to the proteins most closely tied to these electrical properties, namely, ion channels. Aside from the GABA channels noted earlier, seizures have been associated with mutations in the major glutamate channels (NMDA and AMPA) and with potassium channels (Endele et al. 2010; Klassen et al. 2011; Olson & Terzic 2010). Further corroborating the channel hypothesis, Rasmussen encephalitis, a disorder in which auto-antibodies bind to and interfere with glur3 subunit of the AMPA receptor, can cause seizure onset usually at 2–10 years of age in humans (Nestler et al. 2009).

One group of researchers recently endeavored to construct a “channotype,” a profile of variation in all known ion channel genes (Klassen et al. 2011). It was hoped that the channotype would be a barometer that would enable individualized epilepsy onset predictions. The group sequenced 237 channel genes in 139 healthy humans and 152 with idiopathic epilepsy (i.e., epilepsy with no (p.173) known cause). Rather than a signature of epilepsy, the group found a variety of polymorphisms that did not significantly predict individual risk; even healthy people could have enrichment of deleterious mutations. The authors conclude that “personalized ion channel disease prediction will require the integration of genomic profiles with single-cell proteomics and computational modeling of dynamic circuit behavior to translate unique channel variant portfolios into more informative predictors of personal risk for excitability disorders.” In other words, this goal might be difficult. One critique is that this study is cross-sectional (done at one point in time), so it is unclear whether any of these healthy controls with the deleterious mutations might develop seizures in the future. The other critique is the environmental one outlined earlier: that perhaps a channotype combined with an outline of key environmental exposures could identify subpopulations with identifiable predictive risk.

Event Related

Immature rodents exposed to viral RNA or LPS (a component of the bacterial capsule) show lower seizure threshold as adults, an effect that seems mediated by pro-inflammatory immune response that triggers long-term glutamate receptor subunit changes in the hippocampus (Ravizza et al. 2011)—all effects that could, in theory, be monitored and entered into an actuarial risk calculator.

After stroke, 2–4 percent of patients develop recurring seizures (compared to 1 percent of the general population) (Kwan 2010). After head trauma, up to 25 percent develop seizure; even after mild head trauma (i.e., concussion), some patients exhibit residual sub-threshold epileptiform activity 12–14 months after the incident (Pitkänen et al. 2011).

Interestingly, 5-year risk for first seizure was 23 percent with arteriovenous malformations with intracranial hemorrhage or focal neurological deficit (n = 119; CI 9–37) (those that did not present with seizure; all those presenting with seizures had further seizures, as pointed out in Hauser & Mohr [2011]; Josephson et al. [2011]).

Transcriptomics

As all of the previously noted event mediators—viral infection, stroke, head trauma, and hemorrhage—have inherent immune system components, and as emerging research in the field of neuroimmunology is beginning to show that the neural and immune systems are much more intimately associated than once thought, one group recently endeavored to assay the transcriptome (p.174) of peripheral blood cells in animals for association with later epilepsy. The authours summarize their results:

The obtained microarray data point to distinct groups of genes associated with animals that later develop seizures. This supports the hypothesis that the molecular signature preceding the development of epilepsy is present in the peripheral blood transcriptome, and this in turn may allow the development of a prognostic test that can be used to both screen and diagnose potential epilepsy patients, and to prospectively evaluate the effectiveness of ant epileptogenic therapy. However, more work is required in order to fully demonstrate the usefulness of this approach.

(Karsten et al. 2011, p. 215)

Electrophysiological

Pathological high-frequency oscillations (pHFOs), which from animal studies are thought to reflect fast and synchronized activity across populations of neurons, might reflect the neuronal abnormalities responsible for epilepsy. Normal HFOs are 100 Hz, whereas pathological ones are 250–600 Hz and are only present in the dentate gyrus3 of rats that later exhibit spontaneous seizures (Bragin et al. 2010). Though HFOs have been detected in humans, their successful differentiation from pHFOs is an ongoing project. One difficulty is that the pHFOs in rats are detected with implanted electrodes in the brain, whereas we might seek to avoid this option in humans.

Another strategy is to train an algorithm over time to pick out patterns in neural activity that tend to precede seizure in a particular individual. In 2013, Cook and colleagues reported the results of a proof of principle study in which algorithms were trained with electrophysiological data recorded from electrodes implanted in the brains of people with epilepsy in order to predict short time periods of low, medium, and high risk of seizure (Cook et al. 2013). Though this study showed variability in the quality of predictions person to person, it holds the potential for a personal and temporally restricted risk prediction of just the sort that I argued in Chapter 6 could provide a “solution from technology” to the problem of obligations to others becoming more burdensome as the number of forseeable biological risks we pose to others increases.

Sum and Combinations

There may be even greater potential in these technologies through their combination. If we combine an already high risk category, like the patients presenting (p.175) with arteriovenous malformations, with genomics, transcriptomics, and electrophysiological assays, the possibility of identifying very high-risk groups may soon exit from science fiction.

Notes

(1.) A seizure threshold, approximately speaking, might be considered to be set at the intensity level of an insult (audio, visual, chemical, thermal) necessary to trigger a seizure.

(2.) A knockout mouse is a strain in which a specific gene has been genetically manipulated by an experimenter to be non-functional (i.e., it is as if the gene has been “knocked out of the genome”).

(3.) The dentate gyrus is a brain area thought to be implicated in epilepsy because of its integral inhibitory function to the hyper-excitable hippocampus.

(p.176)

Notes:

(1.) A seizure threshold, approximately speaking, might be considered to be set at the intensity level of an insult (audio, visual, chemical, thermal) necessary to trigger a seizure.

(2.) A knockout mouse is a strain in which a specific gene has been genetically manipulated by an experimenter to be non-functional (i.e., it is as if the gene has been “knocked out of the genome”).

(3.) The dentate gyrus is a brain area thought to be implicated in epilepsy because of its integral inhibitory function to the hyper-excitable hippocampus.