Click for next page ( 12


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 11
3 Glutamate Biomarkers Biomarkers can be categorized in a variety of ways. The most com- mon categorization is by use, such as biomarkers of etiology, pathogenesis, diagnosis, diagnostic subtype, treatment, susceptibility, and progression of disease (FDA, 2010). Biomarkers of use also can extend to the regulatory and public health arena, where they are categorized as surrogate endpoints for Food and Drug Administration approval, biomarkers for clinical prac- tice, clinical practice guidelines, and public health practice (IOM, 2008). Because no glutamate biomarkers are currently validated, this workshop fo- cused primarily on research in the early stages of development. The research discussed ranged from molecular to behavioral. Their applications were primarily aimed at pathophysiology, diagnosis, and treatment. Yet the cat- egorization of biomarkers, posed to participants in the earlier presentations, was more fundamental: What is the best way to conceptualize biomarkers? Two speakers presented overlapping conceptualizations. The first, by Jeffrey Conn, professor of pharmacology at Vanderbilt University, divided glutamate biomarkers into three general types: (1) biomarkers of structural engagement with a molecular target; (2) biomarkers of functional engage- ment with a molecular target; and (3) biomarkers of efficacy. The concep- tualization by Kalpana Merchant, chief scientific officer at Eli Lilly and Company, divided biomarkers into two types: (1) “proof of mechanism,” which includes both target engagement and target modulation based on pharmacodynamic markers; and (2) “proof of concept,” which include biomarkers that allow prediction of efficacy or safety. 11

OCR for page 11
12 GLUTAMATE-RELATED BIOMARKERS IN DRUG DEVELOPMENT BIOMARKERS OF ENGAGEMENT AND EFFICACY A biomarker of structural engagement seeks to answer the seminal question: Does the biomarker bind to, or in any way directly engage with, the molecular target of interest, such as a receptor on a glutamate neuron, a nearby cell, or a transporter? Structural engagement is the rate-limiting step for developing and validating any potential glutamate biomarker. Without demonstrating target engagement, Merchant suggested, it is difficult to de- termine whether any apparent biomarker is associated with or attributable to the treatment or the underlying disease. Finding a structural biomarker is a formidable task. It must directly measure factors such as penetration into the brain, degree of receptor oc- cupancy, or another type of direct interaction with the intended molecular target in the central nervous system (CNS). One of the biggest hurdles is penetration into the CNS. The blood–brain barrier (BBB) can prevent penetration or actively extrude certain molecules, often large molecules, by brain reflux transporters. The CNS also can metabolize certain molecules after successfully penetrating the BBB, but before reaching their target (Pike, 2009). Structural and/or functional biomarkers are best studied by imaging with positron emission tomography (PET), a technology used to detect functional activity in regions of the brain in real-time based on radiotracer ligands binding to targets, in this case at the glutamate synapse. PET affords anatomical and quantitative measurement of displacement of a high-affinity endogenous ligand with a labeled one known as a probe. PET provides information about kinetics with high sensitivity and can map whether the probe can fully occupy a receptor once a sufficient dose reaches the CNS. Dosing information is important because an otherwise excellent biomarker can fail to be identified inside the brain if the dose is insufficient. PET displacement studies of potential biomarkers binding with endogenous ligands, however, do not specify the functional aspects of the probe’s inter- action with the target. For that purpose, PET can be combined with other imaging techniques, electrophysiology, or another biomarker. Combining techniques, many participants said, may yield more progress than any single technique alone. A functional biomarker provides a direct measure of target engagement or an indirect measure of downstream actions following target engagement. The biomarker may come from functional magnetic resonance imaging (fMRI), electrophysiological, or electroencephalogram (EEG) response. It could also arise from pharmacodynamic studies using biochemical, physio- logical, and multimodal imaging techniques. But these and other functional methods must be understood as secondary because they do not directly as- sess structural target engagement, the primary goal. If a signal appears from

OCR for page 11
13 GLUTAMATE BIOMARKERS one of these functional assays, there is greater confidence that the target is engaged at some level. However, target engagement is not directly assessed except, for example, through a dose–response curve and detailed pharma- cokinetic and pharmacodynamic studies. Conn indicated that biomarkers of functional engagement have the strong advantage of providing insight into whether or not a compound may demonstrate or predict efficacy. Functional biomarkers are the most common type of glutamate biomarker currently being studied. Other types of biomarkers that are important to develop are pharmacodynamic-based measures to stratify diseased individuals based on their response to a given drug. Mark Bear, Picower Professor of Neurosci- ence at the Massachusetts Institute of Technology, indicated that biomark- ers for diagnostic stratification are urgently needed. As with any biomarker for any disease, the biomarker should be minimally invasive. The range of biomarkers presented below is grouped by modality, that is, the methodological tools used to identify a purported biomarker of structural or functional engagement with a glutamate-related target. Biomarker measurement tools are used for many purposes. They can en- compass electrophysiology, genomics, pharmacological response, receptor expression patterns, radiological or other imaging, and behavioral or neu- ropsychological testing, among others. The methods can be used as direct or indirect measures of glutamate transmission. Any single method or group of methods can be used to shed light on structural or functional engage- ment, whether at the level of genes, proteins, cells, neurocircuits, cognition, or complex behavior. The methods can be used alone or in combination. PHYSIOLOGICAL BIOMARKERS Several functional glutamate biomarkers have gained currency from years of electrophysiological research providing new avenues of research. For example, until recently dopamine transmission was believed to be primarily responsible for the pathophysiology of schizophrenia. However, electrophysiological studies of schizophrenia and the N-methyl D-aspartate (NMDA) receptor have demonstrated that glutamate dysfunction partici- pates in its pathogenesis (Javitt et al., 1996; Umbricht et al., 2000;). In particular, multiple workshop presentations highlighted advancements in electrophysiology techniques as potential methods by which biomarkers for diseases and disorders with glutamate pathology might be developed. Event-Related Potentials The brain’s processing of sensory information has been studied with electrophysiological techniques for several decades. The human environ-

OCR for page 11
14 GLUTAMATE-RELATED BIOMARKERS IN DRUG DEVELOPMENT ment is rich with sensory information, much of which is filtered out for its irrelevancy. By contrast, the human brain must be attentive to novel or salient sensory stimuli because that information is evolutionarily crucial to trigger the flight-or-fight response needed for survival. An “event-related potential” (ERP), studied with EEG electrodes on the surface of the scalp, is an umbrella term covering several electrophysiological methods of mea- suring the CNS response to sensory signals. Several speakers described elec- trophysiological biomarkers of glutamate dysfunction that rely on specific types of auditory or visual ERPs. ERP biomarkers have been studied in relation to cognitive impairment and negative symptoms in schizophrenia. ERP enables study of schizophrenia impairments not improved by marketed antipsychotic medications. Participants noted that this method helps to un- derstand schizophrenia’s cognitive symptoms (e.g., disorganized thinking, inability to plan ahead, impaired memory) or negative symptoms that affect emotion and behavior (e.g., blunted affect, avolition, alogia). ERP assesses functional areas such as visual and auditory information processing, sensory gating, and slow-wave activity, among others. Some ERP abnormalities manifest early in life, especially in vision and hearing, before onset of schizophrenia. Biomarkers in these sensory systems may help identify those at risk for schizophrenia, as well as monitoring the pro- gression of schizophrenia and measuring drug efficacy for any new drugs targeting negative and cognitive symptoms. For improvements in CNS lo- calization of the source of the response, electrophysiological techniques also can be combined with other techniques, such as magnetoencephalography (MEG), to gain higher spatial resolution of the CNS location generating the signal, as well as with imaging techniques. Auditory ERPs Several biomarkers of information processing have links to glutamate dysfunction, including mismatch negativity (MMN) and P3a (Javitt et al., 2008). They are two sequential components of an EEG-recorded waveform, measured by field potentials, which represent the summed synchronous activity of up to millions of neurons. MMN is an ERP measure known as mismatch negativity, a measure of auditory discrimination. It specifically measures the decline in EEG amplitude (in µvolts) in response to an audi- tory stimulus that is distinct from a stream of otherwise repetitive auditory stimuli (Figure 3-1). The singularly dissimilar stimulus is referred to as the “oddball stimulus.” Detected by a scalp electrode above the auditory cortex, the MMN is recorded about 200 milliseconds after the oddball stimulus is introduced. It is shown in the figure below as the nadir in the waveform. It is evoked even when subjects are not told to attend to it. Several hundred milliseconds after introduction of the oddball stimu-

OCR for page 11
15 GLUTAMATE BIOMARKERS Auditory S S S S Oddball S S S S S Stimuli Healthy Volunteers ERP Responses to nses to Schizophrenia Pts P3a 5.0 “Standard” (90%) (90%) S mulii 2.5 Difference 0.0 Waveforms µV -2.5 RON ERP Responses to “Oddball” (10%) -5.0 S muli MMN -7.5 -50 0 50 100 150 200 250 300 350 400 -50 0 50 100 150 200 250 300 350 400 ms ms FIGURE 3-1 Auditory biomarkers 2 Broadside P3a from electrophysiology. Figure MMN and SOURCE: Light et al., 2010. Bitmapped graphs lus, the waveform amplitude surges upward, marking the “P300” peak amplitude, with P signifying positive amplitude at 300 milliseconds. One component of the P300 peak, known as P3a, is measured by electrode placement above the frontocentral lobe. The P3a component is detected by EEG.1 Thus, the CNS’s passive registration of the oddball stimulus is measured first by an amplitude decrement (MMN) prior to an amplitude increase (P3a) in healthy people. In individuals with schizophrenia, the waveform pattern is noticeably different. The general waveform is similar to that of healthy people, but significantly blunted in negative and then positive amplitude (Figure 3-2). Gregory Light, associate professor at the University of California–San Di- ego, explained that changes in MMN and P3a are linked to a broad array of other features of schizophrenia, including decrements in higher order cognitive processes, measures of drug efficacy, and patients’ daily function- ing, among other measures of global assessment. MMN abnormalities have been found to be heritable in people with schizophrenia (Hall et al., 2006a, 2006b). NMDA antagonists reproduce the neurophysiological waveform 1 When the lead is placed at another site on the skull, the so-called P3b component of the ERP is associated with cognitive processing because the subject reports having detected the oddball stimulus at that latency in time.

OCR for page 11
16 P3a 5.0 2.5 µV 0.0 RON -2.5 MMN -5.0 135 ms - 210 ms 240ms - 315ms 350ms - 475ms -100 0.0 100 200 300 400 500 -4.50 µV 4.50 µV -4.50 µV 4.50 µV -1.00 µV 1.00 µV FIGURE 3-2 Automatic sensory information processing abnormalities across illness course of schizophrenia. Figure 3 Broadside SOURCE: Jahshan et al., in press. Bitmapped

OCR for page 11
17 GLUTAMATE BIOMARKERS of schizophrenia in both animal and human models suggesting that MMN has the potential for use as a technique for glutamate biomarker identifica- tion (Javitt et al., 2008). The role of serotonin, dopamine, nicotinic, and other receptors in the generation of MMN is less clear; however recent studies have found that MMN amplitude and latency are altered following antagonist treatments, such as haloperidol and psilocybin (dopamine and serotonin receptor antagonists respectively) and following nicotinic recep- tor stimulation (Garrido et al., 2009). Changes in MMN and P3a are consistently replicated abnormalities in schizophrenia, noted Gregory Light, associate adjunct professor at the University of California–San Diego. In a meta-analysis, the effect size of the MMN waveform differential between healthy subjects and those with schizophrenia is approximately one standard deviation (Umbricht and Krljes, 2005). When examining the relationship between the waveform and the course and severity of schizophrenia, the findings are striking: The group “at risk” for schizophrenia tracks the more normal waveform (in P3a pattern), yet as disease progression occurs, the waveform becomes less and less pronounced (Figure 3-2). In healthy subjects, the waveform pattern shows no changes over the course of time. Finally, when given to healthy people, the NMDA antagonist ketamine induces MMN (Umbricht et al., 2000) and P3a (Watson et al., 2009) attenuation similar to that seen in schizophrenia. While clinical application of ERPs to schizophrenia is promising it has also proven useful for investigating other diseases including dyslexia and learning disabilities (Garrido et al., 2009). Interestingly, the changes to the P3a component are not unique to schizophrenia and have been found in Alzheimer’s disease, bipolar disease, and attention-deficit hyperactivity disorder (Javitt et al., 2008). Steady-State, Visual-Evoked Potentials The visual system is rich in glutamate neurotransmission from the ret- ina through nuclei en route to the visual cortex. Visual defects are manifest in schizophrenia, affecting about a third of patients (Butler et al., 2005). Brian O’Donnell, professor of psychology at Indiana University, noted that people with schizophrenia display deficits in early-stage processing of visual information by the magnocellular pathway of the visual system, as assessed by steady-state, visual-evoked potentials. The magnocellular pathway transmits visual information of low resolution from the retina through the thalamus to the visual cortex, as opposed to high-resolution information transmitted by the parvocellular pathway. As measured by psychophysical tests, the deficits include dot-motion trajectory discrimination, grating velocity discrimination, and contrast sen-

OCR for page 11
18 GLUTAMATE-RELATED BIOMARKERS IN DRUG DEVELOPMENT sitivity, among others. Although the underlying mechanism is unknown, an impact is seen on the perception of motion/spatial processing as op- posed to object processing. Steady-state, visual-evoked potentials consist of presenting a stimulus that is periodically varying, such as a flickering patch or grading on a screen. People with schizophrenia display a selective reduction in these steady-state evoked potentials, with deficits greatest at frequency bands of 17 Hz and higher (Krishnan et al., 2005). There is also a visual amplitude drop in the P300 test, described earlier, but the effect is not as strong as that seen in the auditory system, O’Donnell stated. Vi- sual deficits are correlated with problems in independent living scales, one component of measurement on the Global Assessment of Functioning scale (Butler et al., 2005). Although only tested in normal animals, not humans, the NMDA antagonist ketamine impairs discrimination of horizontal and vertical lines formed by spatial proximity of dots (Kurylo and Gazes, 2008). Although the animals discriminated solid patterns normally, they performed abnormally on the type of visual deficits found in people with schizophre- nia. O’Donnell suggested that because the visual system offers noninvasive access to the CNS and perceptual defects are associated with glutamate dysfunction, biomarker development in this area is worth pursuing. Sensory Gating Sensory gating is another type of biomarker obtained by EEG. It refers to an automatic process that enables the brain to adjust and habituate to a series of repeating sensory stimuli. After an initial stimulus, the brain sup- presses its response to a repeated presentation of the same stimulus. Two types of auditory gating measures―P50 and prepulse inhibition (PPI)―were described by Mark Geyer, professor at the University of California–San Diego, and Bruce Turetsky, associate professor at the University of Penn- sylvania, as potential glutamate biomarkers. These two measures have been studied in animals and humans with schizophrenia, among other disorders. Both measures are abnormal in schizophrenia, suggesting poten- tial biomarkers of pathophysiology. Although both are widely recognized biomarkers, neither is specific to schizophrenia; P50 sensory gating is ab- normal in Alzheimer’s disease while both are abnormal in bipolar disorder. P50 is also found abnormal in yet another disorder, cocaine abuse (Javitt et al., 2008). The two measures are neither strongly correlated with each other, nor with cognitive abnormalities in schizophrenia (Greenwood et al., 2007). P50 does not measure glutamate function if the latter is defined by response to ketamine.2 Turetsky indicated that ketamine has no effect on 2 Ketamine is a non-competitive antagonist that blocks the NMDA receptor channel. Other competitive agonists and antagonists for the NMDA receptor might be different ways to assess glutamate function or dysfunction. See the final section of this workshop report.

OCR for page 11
19 GLUTAMATE BIOMARKERS the P50 gating response in both rodents and human volunteers. In contrast, the pharmacological evidence linking PPI and glutamate is stronger. Geyer reported that NMDA antagonists disrupt PPI in rodents, whereas clozapine, a widely used drug for schizophrenia, prevents the disruption. The animal data suggest that PPI could be used as a biomarker in animal models for pharmacological studies of glutamate-related medications, but the applica- bility to humans is not completely clear. EEG/MEG Combination A combination of EEG and MEG is being tested as a potential bio- marker of depression and its early response to treatment (Tononi and Cirelli, 2006). Depression is one of the leading causes of worldwide dis- ability (WHO, 2001). Most antidepressants fall under the umbrella of serotonin- and norepinephrine-targeted drugs and, for bipolar depression, anticonvulsants or antipsychotics. The drugs usually take several weeks to achieve full effects. Given these drawbacks, glutamate has been studied as another target for modulating depression, possibly with more rapid effects. Glutamate pathways appear to contribute or modulate depression in animal models and humans, as shown in studies using MRI and post-mortem tissue analysis. α-Amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor potentiators of synaptic plasticity and an NMDA antagonist both displayed rapid antidepressant effects in clinical trials (Brennan et al., 2010; Sanacora et al., 2008; Zarate et al., 2006, 2010). To analyze the role of glutamate pathways and antidepressant effects, Carlos Zarate, chief of the mood and anxiety disorders research unit at the National Institute of Mental Health, reported that studies are being conducting in depressed patients that have failed most treatments, such as lithium and quetiapine, with a combination of high-density EEG and MEG activity to analyze slow-wave activity above the anterior cingulate region. High-density EEG uses multichannel recordings to localize neurocircuits arising from extracellular currents, and MEG3 is used to identify intracel- lular currents. The union of techniques has allowed Zarate and colleagues to predict which patients will show the greatest improvement in symptoms following administration of a low-dose of ketamine. In these patients, ket- amine was found to have a significant rapid antidepressive effect, within 110 minutes of administration (Salvadore et al., 2010; Zarate et al., 2006). The union of techniques can help to localize, in this particular context, 3 MEG records the magnetic fields engendered by electrical currents within the brain. It uses arrays of superconducting quantum interference devices. The main application is to localize pathological reasons prior to surgical removal, neural feedback, and research (with the last goal to determine the function of various parts of the brain).

OCR for page 11
20 GLUTAMATE-RELATED BIOMARKERS IN DRUG DEVELOPMENT changes in synaptic potentiation as evidence by increases in slow-wave ac- tivity. Data from this preclinical study suggests that the effects of ketamine are immediate through directly targeting the NMDA receptor complex which indirectly enhances the AMPA throughput, leading to rapid antide- pressant effects. In summary, these newly combined electrophysiological measures may serve as a biomarker to predict drug efficacy. Animal Models The basis of many electrophysiological findings, which have been devel- oped in humans, may be correlated with positive outcomes in disease treat- ment, but they may not be fully understood. Animal models offer insight about molecular, cellular, and neurocircuits that underlie electrophysiologi- cal findings. They can help to clarify the mechanisms of signal transduc- tion, the cell populations most responsible for generating signaling, and the circuitry responsible for neural oscillations, and they can enable testing of new drugs. One presentation addressed the value of primate models in investigating ERPs, such as MMN and other noteworthy electrophysiologi- cal biomarkers. Schroeder and his team of investigators have developed the use of multiarray electrodes recording at different lamina within the auditory and visual cortices of non-human primates. Studying selective attention, they have sought to localize the source of the oscillations responsible for selec- tive attention in visual and auditory tasks, and to interpret its causation. The best method they have found to understand the ERPs is by examining a particular parameter known as current-source density (Lakatos et al., 2008). Applying that technique, Charles Schroeder at the Nathan Kline Institute has reported cell type; cell population (e.g., pyramidal cells); the pattern of circuit activation (e.g., feedforward, feedback); the physiological identity of transmembrane currents; and net local excitation versus inhibi- tion. These and other studies signify the importance of animal testing in biomarker development to find applications to humans, many participants noted. COGNITIVE BIOMARKERS Cognition represents a high-functioning capability of the brain, cov- ering areas such as memory, language, planning, abstract reasoning, and reasoning speed. Cognition is measured by neuropsychological tests, many of which have been validated for purposes of studying cognitive impair- ment (e.g., impairment in psychiatric disorders, substance dependence, and memory impairment in Alzheimer’s disease). The tests are delivered by questionnaires and/or verbal examination administered by professionals.

OCR for page 11
21 GLUTAMATE BIOMARKERS One cardinal form of cognition is a psychological construct known as “working memory.” It is carried out in the dorsolateral region of the prefrontal cortex of the brain in conjunction with neurons from the pa- rietal, temporal, and cingulate cortices (Friedman and Goldman-Rakic, 1994). Working memory refers to the ability to maintain and manipulate information over a short period of time, usually 30 seconds, in order to plan, solve problems, and reason―capacities collectively described as goal- oriented behavior. Glutamate likely mediates these functions (Lewis et al., 2003). Serious problems in working memory are responsible for many symptoms of schizophrenia (Goldman-Rakic, 1994; Verrall et al., 2010). These symptoms are difficult to treat and responsible for the most disabling characteristics of schizophrenia (Buchanan et al., 2005). One major psychological test of working memory function is the AX- Continuous Performance Task (CPT) (Barch et al., 2009). The task requires subjects to press a button when they pick out the letter “X” preceded by the letter “A,” amid many other letters preceding the “X.” Performance on this task, which measures goal maintenance, is abnormal in people with schizophrenia. In healthy people, the NMDA antagonist ketamine produces schizophrenia-like deficits on the AX-CPT (Umbricht et al., 2000). While AX-CPT may be sensitive to other neurotransmitter antagonists, such as muscarinic receptor antagonists, the ketamine challenge finding suggests that NMDA receptors are dysfunctional in schizophrenia. Another working memory task is known as “n-back.” In this task, the volunteer is required to follow a series of stimuli and is instructed to respond whenever a stimulus is presented that is the same as the one pre- viously introduced n trials ago, wherein n is a pre-specified integer, most typically 1, 2, or 3. Conducting a large study of more than 1,000 healthy individuals, Angus MacDonald and his colleagues at the University of Minnesota found that those screening positive for D-amino oxidase (DAO) single nucleotide polymorphisms (SNPs) showed significantly more errors on the AX-CPT and the n-back tasks. DAO metabolizes the amino acid D-serine, which is a coagonist of the NMDA receptor (Verrall et al., 2010). This finding suggests a potential link between DAO SNPs and schizophrenia, as the disease has been linked to poor working memory. Still, more work needs to be done to establish AX-CPT and n-back neuropsychological tests as potential bio- markers for schizophrenia and/or glutamate dysfunction. One problem is that poor performance on these cognitive tests is not specific to schizophre- nia (Javitt et al., 2008). MacDonald explained that these psychological tests have not been studied sufficiently in large samples and by other methods of psychometric validation.

OCR for page 11
22 GLUTAMATE-RELATED BIOMARKERS IN DRUG DEVELOPMENT IMAGING BIOMARKERS Imaging biomarkers have flourished over recent decades as a result of striking advances in technology. In fact, MRI and computed tomography (CT) scans are frequently named in physician surveys as among the fore- most advances in medicine during the 20th century. Both are now standard diagnostic tools in medical practice. But many newer imaging techniques are used in research and have not yet been adopted in clinical practice, largely because of a combination of high costs and the high level of so- phistication required for proper use. Imaging can also be used with other methods to corroborate or refute findings. Knowledge gained through these methods can streamline biomarker development and drug development, said most workshop presenters. Many presenters described a variety of imaging techniques used in the pursuit of glutamate biomarkers. Roughly, the techniques described below deal with structure, function, and/or quantification of functional activity. Function can be objectively measured, often by signal density, intensity, duration, and location. Imaging can be used alone or in combination with electrophysiology. As a general rule, electrophysiology is better suited to detect functional changes within a time frame of milliseconds. This is especially vital considering that action potentials take up to 130 ms to propagate. Imaging is superior for spatial resolution, but its time resolution typically ranges from seconds to minutes. Both have the capacity to measure functional activity. So imaging is generally superior for spatial resolution, whereas electrophysiology is superior for temporal resolution. MRI MRI is ideally poised to study in vivo brain structure and function. The use of non-ionizing radio frequency signals for image acquisition is tailored to soft tissue and safer than CT, which exposes the patient to high levels of radiation. fMRI and pharmacological MRI (phMRI) are broadening the understanding of glutamate sites of action, mechanisms of action and treat- ment localization, dose, and effects. fMRI and phMRI depend on changes in the amount of blood oxygenation level–dependent (BOLD) signal. Lack- ing reserves for glucose and oxygen, active neurons require more immediate glucose and oxygen delivery than do inactive neurons. A BOLD perturba- tion in hemodynamic activity is associated with higher activity in nervous system pathways and regions under study. fMRI is frequently undertaken during execution of specific tasks assigned to subjects, such as neuropsy- chological tests measuring cognition, emotion, or substance dependence. A union of pharmacological techniques and fMRI, phMRI is used to detect drug-induced effects on activity levels in the brain at the site of

OCR for page 11
23 GLUTAMATE BIOMARKERS action. Drug administration is given to examine signaling changes from baseline. The technique seeks many answers, such as the location of the drug’s physiological target, the response to treatment, appropriate dose levels of a drug to achieve a desired effect, and the pharmacokinetics and pharmacodynamics at its site of action (Paulus et al., 2005). Activity- dependent changes with phMRI also may identify the consequences of drug action over time, its distribution to other areas of the brain, and aspects of pathophysiology, disease onset, and course of disease. phMRI is especially useful for studying glutamate because glutamate’s action is highly energy dependent and thus readily detectable. Another advantage is that phMRI may identify sites of adverse effects that could be targeted for protection. Finally, an advantage of phMRI over PET is that no radioligands are neces- sary to find and produce, which is the most difficult and sophisticated part of the process. Magnetic Resonance Spectroscopy One variant of magnetic resonance techniques is proton magnetic reso- nance spectroscopy (1H MRS). It can assess in vivo function of brain chemistry by exploiting the fact that hydrogen atoms in distinct chemical environments, depending on the molecules in which they are bonded, pos- sess distinct resonant properties. Glutamate and glutamine are among the few molecules detectable by MRS, but at extremely low concentrations (6.0–12.5 mmol/kgww). The concentrations can be inferred from the spec- tra generated by the technique. MRS is spatially and temporally averaged to detect molecules of interest. In his presentation, Robert Mather, principal scientist at Pfizer Pharmaceuticals, described his application of MRS to the study of retigabine, an anticonvulsant drug that binds to voltage-dependent potassium channels (Kv7 or M-channels). MRS enabled him to determine that retigabine reduced concentrations of glutamate/glutamine in the hip- pocampus. His study demonstrated the feasibility of using MRS as a mecha- nistic biomarker of changes in glutamate/glutamine ratios associated with the action of a particular medication at its target location. Recent studies using MRS have found abnormalities in the glutamine/glutamate ratio in bipolar disorder and major depressive disorder with reductions during de- pressive episodes and increases during mania, similar to those seen during the first episode of schizophrenia (Yüksel and Öngür, 2010). Dr. Mather remarked that MRS can be a valuable biomarker of glutamate levels in many psychiatric and neurological diseases.

OCR for page 11
24 GLUTAMATE-RELATED BIOMARKERS IN DRUG DEVELOPMENT PET and SPECT PET is a non-invasive, in vivo nuclear medicine imaging technique ad- ept at studying CNS function. PET radioligands, which must be short lived to reduce radiation exposure, are injected intravenously and synthesized to interact with particular molecular targets. PET’s foremost strength is its level of sensitivity (10–12M) and capacity for kinetic analysis. Drawbacks are its slower temporal resolution (minutes) and low spatial resolution (2–6 mm). To improve both, PET can be combined with a number of other imag- ing methods performed simultaneously and built into the same machinery, such as single photon emission computed tomography (SPECT), which furnishes 3-D images that can be manipulated, among other distinctions. The combined techniques provide better understanding of molecular inter- actions, molecular environment, and understanding of receptor occupancy. During the first few years of PET’s introduction, starting in the 1960s, PET radioligands were limited; the foremost ligands, oxygen-15 and fluo- rine-18, among others, were too nonspecific for studying glutamate syn- apses. What also hindered PET’s application to glutamate was its limited spatial resolution. Today, however, PET has dramatically improved, as has the introduction of more PET ligands and new techniques to reconstruct PET images for better spatial resolution. The foremost barrier holding back PET applications for glutamate neurotransmission is the scarcity of PET radioligands that expressly bind to particular molecules at the glutamate synapse. PET remains a challenge because of the complex interplay of 30 or more molecular targets vying for glutamate modulation, observed Schoepp. Other problems specifically with PET ligands have been lipophilicity, which leads to nonspecific reten- tion in the high levels of CNS lipids. Radioligands have also been beset by high binding to plasma proteins, which limits entry of the radioligand to the CNS. One of the few examples of PET radioligands for target engagement in glutamate neurotransmission is a specific radioligand for the mGluR5 receptor. Speaker Robert Innis, of the National Institute of Mental Health, and his team developed an 18F-labeled ligand for the compound SP203, which antagonizes the mGluR5 receptor. He reported finding that PET scanning could be used to visualize and quantify the labeled antagonist in the healthy human brain. PET provides a level of detail not available with other techniques. MRI could not be used for reconstructing images because its sensitivity was insufficient to detect the potential biomarker. He also found that animal models could not have predicted target engagement in humans because rats and monkeys defluorinated the radioligand before its uptake into the CNS. The glutamate antagonist is highly important to study in vivo because its antagonism of the metabotropic receptor blunts

OCR for page 11
25 GLUTAMATE BIOMARKERS drug-seeking behavior, which is of major significance to the study of drug dependence. One other radioligand is an 18F-labeled inverse agonist for the canna- binoid receptor CB1 (Pacher et al., 2006; Terry et al., 2010). The receptor mediates marijuana’s psychotropic effects. Located pre-synaptically, CB1 receptors inhibit the release of glutamate. CB1 receptors are located widely throughout the brain, including the cortex, hippocampus, and cerebellum. In glutamate-related diseases and injuries, these receptors function patho- logically, likely by releasing excess glutamate and producing excitotoxicity and oxidative stress on the post-synaptic neuron (Pacher et al., 2006; Terry et al., 2010). SPECT and PET are similar techniques except that SPECT directly emits gamma radiation, whereas PET emits two gamma photons in op- posite directions. A PET scanner has the distinct advantage of generating significantly higher resolution images, about two to three orders of magni- tude greater than SPECT. But SPECT images can be manipulated in three dimensions. To protect human health, the gamma-emitting radiotracers used in PET must be short lived, whereas SPECT uses radiotracers that are longer lived isotopes.

OCR for page 11