Biomarkers in schizophrenia
This May, the American Psychiatric Association (APA) published the latest edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM), which is commonly referred to as the bible of psychiatry . Now in its fifth edition, DSM enables clinicians to diagnose psychiatric problems based on a list of symptoms that are considered to be the essential components of each disorder. This approach enables consistency in diagnosis across clinicians. However, DSM V, like its predecessors, is not based on pathophysiology and does little to incorporate biology into its diagnostic criteria.
While there have been significant advances in our understanding of basic neuroscience over the last 60 years, this has not translated into psychiatric clinical practice. Indeed, psychiatry remains the only branch of medicine that does not routinely use diagnostic laboratory tests . Much of this is due to the following factors: the underlying biology of psychiatric disorders is poorly understood; the preclinical models for psychiatric disorders are not very good; and the organ of interest is hard to access [3,4].
This leads to a situation where, as a direct result of these issues, there is a lack of translational models and therefore the application of basic neuroscience to the clinical arena is more challenging than in other medical specialties.
One way forward is to develop biomarkers to aid in prognosis, patient stratification and diagnosis. Once validated, these biomarkers can be used for the development of diagnostics, which can be used for improved clinical decision-making. The papers in this special thematic issue on biomarkers in schizophrenia discuss different approaches to the same common goal: the identification of biomarkers to enable better patient care.
Sabine Bahn’s group has a long track record in the identification of proteomic biomarkers to aid in the diagnosis of schizophrenia [5,6]. Their approach rests on the hypothesis that mental illness is a systemic disorder that will have perturbations across the whole body, including the composition of the blood. They review an extensive literature that has provided evidence of disturbances in inflammatory, hormonal and metabolic pathways. The data generated through these proteomic approaches give credence to the idea that the periphery can be used to detect perturbation in mental illness and demonstrating this has been of immense importance.
Identification of useful biomarkers for psychiatric disorders is very challenging without a concerted effort. Colin Dourish and Gerard Dawson describe some of the work being generated by an innovative collaboration orchestrated by the clinical research organization P1vital [8,9]. They have developed a framework in which a precompetitive consortium between industry and academia enables the development and validation of new biomarkers. This is accomplished through implementation of experimental medicine studies whereby potential biomarkers useful for psychiatric research are validated. This approach is exemplified by some of the imaging studies undertaken by this consortium that have used individuals categorized as high schizotypes, which is a personality disorder in the milder part of the schizophrenia spectrum. Research assessing spatial working memory in these individuals, who are in essence healthy but with some of the genetic load of schizophrenia, reveals that these high schizotypes have significant differences in their brain activation patterns, suggesting they have an inefficient encoding strategy. These findings suggest that these subjects may be useful surrogates for schizophrenic patients. This is important in that use of this subject population may make it easier to perform clinical studies in the schizophrenia field as these individuals are typically not on medications and are potentially easier to recruit for studies.
Larry Ereshefsky and his colleagues outline some of the essential work that is needed in Phase I studies to detect early signs of efficacy in drug development . Their manuscript covers a variety of biomarker platforms that aim to address what has been termed the three pillars of drug development demonstration of: exposure at the site of action, target binding and pharmacodynamic activity . These methods include cognitive testing, imaging, EEG and specialized pharmacodynamic biomarkers. The selection of the appropriate approach depends on the particular attributes of the drug being tested and which system it will alter.
The use of EEG biomarkers such as mismatch negativity (MMN), a preattentive auditory event-related potential, have been explored for their utility in schizophrenia. MMN is a translational biomarker that offers a unique window to some of the sensory processing deficits that are present in schizophrenia. As Perez and her colleagues write, this biomarker has shown a consistent and robust reduction in schizophrenia patients . Interestingly, initial studies suggest that it may also be helpful in the prediction of conversion to psychosis in those subjects who are at high risk. These initial observations will need confirmation in longitudinal studies.
Looking to other neurological conditions, current drug development strategies for Alzheimer’s disease are moving to earlier stages of the disease with the idea that the most impact will be had prior to the development of full-blown disorder . The Alzheimer’s Disease Neuroimaging Initiative has been enormously successful in identifying biomarkers that can be used to predict conversion to disease from the prodromal stage . Investigators studying prodromal schizophrenia have a similar goal in mind. Current research in this area is identifying subjects in the prodromal phase of schizophrenia to discover biomarkers that can be used to predict conversion . Kristin Cadenhead and colleagues describe some of the efforts in identifying prognostic biomarkers . These prognostic biomarkers may enable early intervention to improve outcomes, or even more excitingly, prevent the onset of schizophrenia. These biomarker platforms include neuroimaging, electrophysiology, and most robustly and consistently found in the prodrome, neurocognitive deficits. Interestingly, it appears that targeting specific cognitive deficits may be helpful in delaying or even preventing psychosis.
All these efforts will ultimately move the field forward by identifying biomarkers that can then be developed for patient stratification, early diagnosis and prognosis, or to identify those subjects who are responders or nonresponders. These biomarkers can then form the core from which diagnostic products can be made to enable the goal of personalized medicine; that is, giving the right drug to the right patient at the right time. Improvement of psychiatric nosology will only come from a better understanding of the pathophysiology of these disorders. This can be achieved only through studying relatively homogenous patient populations that have some common biology. There is a catch-22 here, that to identify the biology one needs patient groups that are defined somewhat biologically. An alternative is to use particular phenotypic characteristics to define more homogenous groups and then explore the biology of this clustering. One example that used this method is the clinical data on the identification of CFHR1 protein as a response predictor biomarker for bitopertin . These data suggest that baseline levels of CFHR1 may predict clinical response. One of the potential reasons a peripheral biomarker was identified in this example is that the patient population was comprised of negative symptom schizophrenic patients and was therefore more phenotypically homogenous than a take all comers schizophrenia trial.
Ultimately, the use of phenotyping combined with biomarker discovery may enable the identification of subpopulations that will inform our diagnostic nosology such that we can expect future editions of the DSM to include the biological characteristics of these mental disorders.