Role of biomarkers and endophenotypes in prevention and treatment of psychopathological disorders

The term biomarker is used in numerous scientific disciplines. Accordingly, several different types of biomarkers exist, ranging from biosignatures in geology to surrogate end points in medicine. Although any of these types of biomarkers could provide for an interesting discussion, this article focuses on biomarkers as endogenous, measurable characteristics that mark either risk for or manifestation of psychiatric illness. For example, recent proton magnetic resonance spectroscopy studies indicate that choline concentrations in the cingulate cortex correlate positively with depression ratings among patients with bipolar disorder when compared with controls [1]. Thus, choline levels in the cingulate appear to be a state-dependent biomarker of the depressive phase of bipolar illness. This state dependence suggests that choline concentrations in the cingulate cortex may serve as an objective biomarker of treatment response.

Biomarkers versus endophenotypes

An important subtype of biomarkers are endophenotypes, which have a more restrictive definition, as outlined by Gould and Gottesman [2]:

• ▪ An endophenotype must segregate with illness in the general population;

• ▪ An endophenotype must be heritable;

• ▪ An endophenotype must be state independent, manifesting whether illness is present or in remission;

• ▪ An endophenotype must co-segregate with the disorder within families;

• ▪ An endophenotype must be present at a higher rate within affected families than in the general population;

• ▪ An endophenotype should be a characteristic that can be measured reliably, and be specific to the illness of interest [3].

When these six criteria are met, the endophenotype is assumed to mark genetic risk of illness, whether or not illness is expressed phenotypically. Note that by this definition, choline concentrations in the cingulate cortex of bipolar patients do not qualify as an endophenotype because they are dependent on clinical state. Therefore, this marker cannot be used to identify those who are at genetic risk for bipolar disorder, yet present no symptoms.

A good example of an endophenotype in psychiatric research can be found in the literature on schizophrenia. Patients with schizophrenia exhibit irregularities in smooth-pursuit eye tracking. These irregularities are measured by eye tracking devices while patients follow a moving stimulus. Although the pathophysiology of eye tracking dysfunction is not fully understood, as many as 80% of schizophrenia patients exhibit this trait, as do up to 45% of their first-degree relatives, compared with only 10% of healthy controls [4]. Interestingly, this 10% figure is very close to the estimated prevalence rate of schizophrenia predisposition in the general population [5]. However, because schizophrenia predisposition is not fully penetrant, the majority of these individuals never develop the disorder.

Following from the above discussion, eye tracking dysfunction qualifies as an endophenotype of schizophrenia predisposition because it:

• ▪ Appears to segregate with risk of illness in the population;

• ▪ Is heritable;

• ▪ Is state independent;

• ▪ Cosegregates with illness within families [6];

• ▪ Is present at a higher rate in affected families than in the general population, and;

• ▪ Can be measured reliably and is specific to schizophrenia liability.

Endophenotypes & prevention

Given their state independence, endophenotypes hold much promise in prevention research because they can be used to identify those who are at a high risk of developing psychiatric disorders [7]. Once again, a good example comes from research on schizophrenia. Using a statistical technique known as taxometrics, Erlenmeyer-Kimling et al. selected three putative endophenotypes (sustained visual attention, neuoromotor performance and selected measures of intelligence) to identify a high-risk group of 7–12-year-old children with a parent with schizophrenia [8]. Their analysis identified 47% of children as high risk, compared with 4% of controls. Importantly, by age 22–29 years, 43% of the high-risk group had been hospitalized or received significant treatment for a schizophrenia spectrum disorder. Using similar methods, these findings were later replicated by an independent research group [9].

As we have noted elsewhere [10], the implications of these findings are difficult to overstate. Enrolling all children of a parent with schizophrenia in a prevention program is woefully inefficient because only 10–15% will develop a schizophrenia spectrum disorder. However, analysis of putative endophenotypes can identify a group at nearly 50% risk, making prevention much more efficient. Importantly, prevention programs can be quite effective for those at high risk for schizophrenia. For example, by administering a low dose of risperidone, McGorry et al. substantially reduced the emergence of first episode psychosis in a group of high-risk patients [11]. Those who took risperidone and participated in a cognitive behavioral intervention were 95% psychosis-free 3 years later, compared with 30% who received typical needs-based care.

Extending the use of endophenotypes to other psychiatric disorders

In psychiatric research, biomarkers have been identified for many disorders. Yet, despite their usefulness for prevention, very few endophenotypes have been discovered. One reason for this is that psychiatric disorders are often etiologically heterogeneous. For example, recent research suggests that there are distinct subtypes of major depression [12]. However, these subtypes are very difficult to distinguish from one another solely based on symptoms. Yet symptoms are usually the only source of information used for making a diagnosis. As a result, etiologically heterogeneous samples are recruited for most research studies, which obfuscates identification of endophenotypes. Thus, until researchers subdivide their samples based on best guesses about alternative etiologies, identification of endophenotypes will remain elusive.

In medicine, it is universally accepted that etiological diagnosis is superior to syndromal (symptom-based) diagnosis. When etiology is understood, prevention and treatment strategies usually improve because mechanisms of illness can be targeted directly. Thus, the search for endophenotypes is an important development in psychiatric research insofar as it brings us closer to etiological mechanisms of psychopathology. Hopefully, future efforts will identify endophenotypes for a host of psychiatric disorders, leading to improved prevention and treatment programs, and legitimizing the current popularity of the term.