Showing posts with label pathway analysis. Show all posts
Showing posts with label pathway analysis. Show all posts

Monday, September 7, 2015

Patentable Biomarkers of Suicide

From: Understanding and predicting suicidality using a combined genomic and clinical risk assessment approach (reference 1 with permission).





One of the most interesting aspects of biological psychiatry is the attempt to characterize complex biological systems.  It may not have been apparent but complex biological systems factored in a recent post about bronchitis.  Lungs are certainly complex with two different blood supplies and complicated immunology, but the lungs are not thinking organs.  They don't come up with any secondary concepts that need to be analyzed as possible derivatives of the biological substrate.  And even then, basic syndromes that we all learned about in medical school and in clinical rotations, defy more useful classifications.  I have previously posted on endophenotypes and their usefulness in the treatment of asthma and only recently noted that they have proliferated to include an obese endophenotype and how that affects response to therapy.  Diagnostic and treatment approaches to asthma and bronchitis are necessarily crude, largely because the biological complexity in these processes is not fully appreciated and addressed.

The brain is certainly the most complex organ in the body.  Cellular arrays in the brain produce a stream of consciousness, robust unconscious processing, unique conscious states, and all forms of emotional, social and intellectual constructs that can be observed, monitored, and changed.  That brings me to a paper from Molecular Psychiatry on possible biomarkers for suicide.  Not just any paper - at this point it is the most downloaded paper from the top-ranked psychiatry journal (1/140) in the world.  Molecular Psychiatry has an impact factor of 14.496 and that is the highest impact factor of all psychiatry journals.  In part that is probably driven by how absurdly expensive that similar journals like Biological Psychiatry are or other barriers to purchase like needing to be a member of the sponsoring society.  This is a public access journal that uses Creative Commons Licenses for their content.  The authors in this case have provided a 20 pages article and 124 pages in Supplemental Information.

The idea of a biomarker for suicide is very attractive to psychiatrists, because assessing suicide risk is a big part of what we do.  Current clinical guidelines suggest that we need to make that assessment at every patient visit.  The actual prediction of suicide is difficult due to the fact that mental states change over time and people may not be able to communicate their true level of risk.  I have had people tell me in retrospect that they lied about their degree of suicidal thinking and level of control when I asked them about it.  I have had acute care colleagues tell me that they were weary of having to guess about whether a person was going to try to kill themselves or not - many times a day.  The assessment is further complicated by a lack of acceptable acute care options that may further hinder complete self disclosure.  A biomarker would potentially be beneficial.  I qualify that by the fact that the dexamethasone suppression test was once considered a biomarker of suicide (1), but these days it is rarely done and certainly not as part of a suicide assessment.  A study by Coryell, et al (9) notes that the DST was not able to differentiate patients who died from suicide or cardiovascular disease when long term mortality was determined by the National Death Index.  Those authors suggest it may be useful as a predictor of suicide only in patients with depression.

In this article the authors take a look at possible biomarkers in blood that could predict both suicide and some associated markers like risk of hospitalization.  There is a lot going on in this paper.  All the research participants were men.  They studied four different patient cohorts including 217 patients followed longitudinally.  This group was called the Discovery Cohort because markers were discovered based on 37 patients who had a switch from a no suicide state to a high suicide state defined as a score of 2 - 4 on the HAMD question about suicide.  26 deceased patients who committed suicide were used to validate the initial markers.  Two psychiatric cohorts of 108 and 157 to look at prediction of suicidal ideation and hospitalization with the chosen tests.  The flow of these experiments in depicted in the graphic at the top of this post from the original paper.  In the diagram, the designations AP (absent-present) and DE (differential expression) are techniques to capturing genes that are turning of and turning on and off and gradual  changes in gene expression.  The respective genes in this analysis are color-coded based on those properties.  The Convergent Functional Genomic (CFG) Approach is depicted in the box.  Candidate genes are ranked in the triangles according to CFG score.  The CFG score was the sum of various weighted factors including evidence of human brain expression, evidence of human peripheral presence, human genetic evidence and linkage with weighted scores in the CFG box.  Using their discovery and validation sequence the authors were able to pare down the total number of genes down from 412 to 208 to 143 and ultimately to 76 genes.  The supplementary information provides the validation of biomarkers and a table that looks at each gene and prior human genetic evidence, prior evidence of brain expression and prior human evidence of peripheral expression.

The authors discussion of the biological relevance of their findings was interesting.  They did pathway analysis looking at Ingenuity, KEGG, and GeneGO databases.  Of these only the Kyoto Encyclopedia of Genes and Genomes (KEGG) is publicly available without a subscription fee.  It is very useful to know about KEGG because of the relevance of pathway analysis in the psychiatric literature.  As an example, I have been teaching about the mTOR pathway discussed in this article in my neurobiology of addiction lectures for the past 4 years.    

This article is very interesting and can be read at  several levels.  It is premature to consider it definitive at this point and based on this paper and the work of the associated lab these authors are working on additional validation strategies.  If they are  correct,  suicidality may be captured in time as a polygenic event based on a combination of genes that are turned off and on and others that gradually change.  I titled this post as "patentable genes" because the only conflict of interest cited is the lead author is listed as an inventor on a patent application being filed by Indiana University.  For trainees and early career psychiatrists a familiarity with this technology and its potential uses and limitations would be one of the reading goals and including Molecular Psychiatry and its sister journal from the same group Translational Psychiatry (8) is probably a good idea.  Both are potentially good sources of neuroscientific information in psychiatry and if popularity is any indication - fill a niche in the field.  Some of the tools that they developed along the way are useful to think about from a clinical perspective (4, 5).  The thought that the CFI-S Scale was particularly interesting because it is a 22 point binary scale that looks at factors (excluding suicidal ideation) that they determined to be important.  The factors are also classified as to whether they represent increased reasons (IR) or decreased barriers (DB) to suicide.  The emphasis on suicide as a discrete syndrome independent of diagnosis is a research strategy that has been called for recently based on the need to come up with better ways to diagnose and treat the problem.  In a clinical setting I think that clinicians are still frequently surprised by suicide attempts and suicides being able to determine if a patient is in a high risk state based on a blood test independent of their clinical presentation and statements would be useful both in terms of the test but also the associated dialogue.

What I really like about this paper is that it is an attempt to deal with a common psychiatric problem at the appropriate level of complexity.  Clinical trials do exactly the opposite.  As an example, clinical trials in psychiatry will look at heterogeneous groups of patients pulled together under a vague diagnostic category.  There may be rating scales or global ratings just because the rating scales don't seem to have much discriminatory power.  In the end, the entire study is generally collapsed for a very simple statistical analysis.  Getting to those final variables and what has been ignored in the process is always the critical question.  I think it is trendy these days to commiserate about the fact that there are inconclusive, weak and non-reproducible results from the standard clinical trials technology.  I don't know why anyone would expect a different result.  If anything this paper illustrates that a lot of biological information can be considered and analyzed.  The popularity of this paper leaves me hopeful that this is a positive trend for the future.            


George Dawson, MD, DFAPA


References:

1:  Niculescu AB, Levey DF, Phalen PL, Le-Niculescu H, Dainton HD, Jain N,Belanger E, James A, George S, Weber H, Graham DL, Schweitzer R, Ladd TB, Learman R, Niculescu EM, Vanipenta NP, Khan FN, Mullen J, Shankar G, Cook S, Humbert C, Ballew A, Yard M, Gelbart T, Shekhar A, Schork NJ, Kurian SM, Sandusky GE, Salomon DR. Understanding and predicting suicidality using a combined genomic and clinical risk assessment approach. Mol Psychiatry. 2015 Aug 18. doi: 10.1038/mp.2015.112. [Epub ahead of print] PubMed PMID: 26283638.

2:   Lee BH, Kim YK. Potential peripheral biological predictors of suicidal behavior in major depressive disorder. Prog Neuropsychopharmacol Biol Psychiatry. 2011 Jun 1;35(4):842-7. doi: 10.1016/j.pnpbp.2010.08.001. Epub 2010 Aug 11. Review. PubMed PMID: 20708058.

3:   Collection of references for biomarkers in suicide.

4:  Simplified Affective State Scale (SASS).

5:  Convergent Functional Information for Suicide (CFI-S) Scale.

6:  Laboratory of Neurophenomics Web Site.

7.  Niculescu AB Medline Collection on additional convergent functional genomics references.

8.  Translational Psychiatry Web Site.

9.  Coryell W, Young E, Carroll B.  Hyperactivity of thehypothalamic-pituitary-adrenal axis and mortality in major depressive disorder.  Psychiatry Res. 2006 May 30;142(1):99-104. Epub 2006 Apr 21. PubMed PMID: 16631257.

Attribution:

The figure at the top of the post is from the original article listed completely in reference 1 under a Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License.  To view the condition of that license view it here.

Supplementary:

1.  There is a Mayo Clinic Conference coming up this fall for anyone interested in translational approaches to psychiatric disorders and addictions.  Further information is available at this web site.

2.  There is also the 3rd Annual Update and Advances in Psychiatry conference at the US Madison and one of presentations is by Daniel Weinberger, MD on the neuroscience of schizophrenia and psychotic disorders.   Information on that conference and the conference brochure is available at this web site.