Biomarkers for schizophrenia and autism

Currently, mental disorders are diagnosed by analysis of behavior and reports from the patient on their own thoughts. This way of diagnosing disease can be difficult, subjective and messy, particularly with regard to symptom overlaps between disorders. It also necessitates waiting until the disorder is fully manifest before the diagnosis can be made. Moreover, these disorders are heterogeneous, with wide variation in the abnormal behaviors within a given disorder, and very likely wide variation in genetic and environmental causes of the disorder. These are the reasons why we need biomarkers – objective, quantifiable, molecular signatures of each disease, and each type of disorder within a given disease category. It would be ideal to have a test that could be done on serum as is done for high cholesterol, which raises the concern for cardiovascular disease. Other possible sources of biomarkers could be urine or even fecal samples, all of which are non-invasive tests. MRI is another non-invasiv test of interest, but is very expensive. Taking samples of cerebral spinal fluid is both invasive and expensive, but is of course more relevant for the brain than are serum and urine.

A new paper from a multi-national collaboration provides what may be a first step towards “Identification of a biological signature for schizophrenia in serum”. While this title of the paper overstates the case, some progress towards this goal is reported. The authors measured the concentrations of 181 proteins and small molecules in serum taken from newly diagnosed schizophrenia (Sz) patients as well as serum from patients with major depressive disorder (MDD), bipolar disorder (BPD) and Asperger syndrome (AS). In one large group of Sz samples, they found that the levels of 34 molecules could be used to define or separate Sz from control serum – with high sensitivity (detect most of the Sz subjects) and high specificity (identify Sz subjects without mistakenly identifying controls as Sz). However, when they used this 34 molecular “signature” on 4 other groups of Sz subjects, they were able to separate only 60-75% of Sz subjects from controls. The sensitivity and specificity in comparing Sz against BPD was 86 and 78% (respectively). Comparing Sz against MDD was 87 and 94%. Comparing Sz against AS was 96 and 96%. Thus, the 34 molecular signature was extremely useful in deciding between Sz and AS, very good between Sz and MDD, and quite good between Sz and BPD, but not as good comparing Sz and controls. Since it can be difficult, at least at the outset, to make a differential diagonsis between Sz and BPD based on behavior, this molecular approach may be helpful in that case.

So what are these 34 molecules? Can they tell us something useful about Sz? Many of these molecules have been implicated in acute and chronic inflammatory conditions, and as I emphasize in my book, there are many other connections between inflammation and  Sz, autism and MDD .

A puzzling aspect of this work is why the investigators started with only 181 proteins and small molecules when it is technically possible to assay several thousand molecules in serum. However, a strength in this work is that they were able to obtain samples from newly diagnosed Sz subjects, before they were taking medications. This is not the case in prior publications of this type. This is the case for a Chinese paper. However, the latter paper used a global metabolic method (far broader and unbiased compared to the new paper) to look for changes in Sz serum. They reported evidence suggesting upregulated fatty acid metabolism in Sz. The authors of both of these papers suggest Sz has a metabolic component.

What about biomarkers for autism? There have been several molecular analyses of urine from ASD children that found differences from controls. One from Tony Persico and colleagues in Rome particularly caught my eye. They reported that a molecule called p-cresol is significantly higher in young autistic children (particularly females) compared to controls. This is intriguing because this molecule is produced by particular species of gut bacteria, and the composition of gastrointestinal bacterial species differs between ASD and control children. This is discussed in more detail in some of my prior posts.

This entry was posted in General and tagged , , , , , , , , , , . Bookmark the permalink.

5 Responses to Biomarkers for schizophrenia and autism

  1. Pingback: Biomarkers for Schizophrenia and Autism | asdresearchinitiative

  2. Earlier this year an interesting study came out with regards to biomarkers for schizophrenia. (full-text). The paper received very little media attention despite the fact that the words ‘AUC=1’ were used and data were confirmed on both a test and training set – five serum metabolites and one urinary marker.
    I blogged about it here:
    OK it needs further validation and there may be some issues with ethnicity, but as far as technology and methodology goes, they were pretty comprehensive.

    • phpatterson says:

      Excellent point Paul. The paper you cite has the advantage over the one I cited of being non-biased – looking at virtually all small molecules in serum rather than a selected subset. The former paper’s downside is that a large fraction of the subjects were hospitalized when their samples were taken. Thus, they had more of a chronic disease than the subjects in the study I cited, which were sampled just at the first time they suffered a psychotic event. Chronic disease could cause a multitude of aberrant metabolic changes that are not directly related to the pathology that caused the disease in the first place. Ideally, one would like to test those at high risk for schizophrenia before an after the psychotic event that led to their clinical presentation. It is annoying indeed to see so little overlap between the 3 studies of serum molecules differentially found in sz vs controls. Many possible reasons for this, of course.

  3. Anonymous says:

    There are some seriously funky things afoot. Tyrosine is a precursor to phenol and p-cresol. SNPs in the PTPN22 gene (coding for a tyrosine phosphatase) are linked to autoimmune disorders and a hyper-immune response to infectious disease. “Old Friend” bacteria (adapted to upregulate Tregs in humans) break down excess tyrosine. A few pathogens such as C.Diff, Pseudomonas, and Desulfovibrio can also metabolize p-cresol and hydroxybenzoate, which are elevated in autistic kids. (Desulfovibrio is usually found in oil fields. Phenol and p-cresol can mimic petroleum contamination.) So, in the absence of Old Friends and against the genetic background of PTPN22 SNPs, you’d expect tyrosine to build up, fostering the growth of neurotoxin-producing pathogens seen in autism/schizophrenia. Right? I know it’s reaching-for-the-moon thinking based on very very scant data. But it’s still interesting to think about. (scroll a bit to MacFabe article)

Leave a Comment

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s