Wednesday, August 24, 2016






The Signal, the History, and the Context

      What a signal does/means depends on the context.  A hormone or biomarker level in the blood is not as informative as knowing the longitudinal history of levels and  their contexts. What is the   % change of the current level- is the marker increasing or decreasing, and how fast is that occurring? Has the marker in the past been this high or higher, and then it went down?  What made it go down? What was the maximal level attained in the past? What was the maximal % change? A specific driver is having  a personal and/or family history of having illnesses related to that marker. A first layer of context is if they have other signs of illness.  A second layer of context is  provided by information about living in an environment or having been subjected to events  relevant to the illness. Context, while non-specific, may give rise to or enable the illness manifesting, if the specific driver is in place to channel it into disease.  

      For exemplification, think PSA levels and prostate cancer. While absolute PSA levels provide some information, the slope of increasing PSA is more informative.  Have they had in the past high levels like this, which then went down?  What made them go down? What were the maximal levels attained in the past? What was the maximum slope?  A driver is a personal history or a family history of prostate illness.  A first layer of context is a physical exam or imaging test revealing an enlarged prostate. A second layer of context is provided by age, ethnicity, and environmental factors such as high fat diet. Context, while non-specific,  may  give rise to or  enable disease, if the specific driver/mutations are in place to channel it into disease .

      Another example, this time from psychiatry, is that of suicidal ideation.  While absolute levels of suicidal ideation intensity provide some information, the slope of increasing suicidal ideation may be more informative regarding risk of committing suicide.  Have they had in the past high suicidal ideation comparable to this or higher, that then resolved?  What made it resolve?  What was the maximum suicidality (ideation, attempts) they had in the past? What was the maximum slope? A specific driver of behavior is a personal history or a family history of suicidal behavior, or knowing somebody who attempted or committed suicide.  A first  layer of context is a mental status exam or test revealing increased anxiety, low mood, or distorted thinking.   A second layer of context is provided by gender, age, ethnicity, and environmental factors such as stressful life events and addictions. Context, while non-specific, may  give rise to or  enable suicide, if the specific driver is in place channel it into to suicidal ideation and action. 


Alexander B. Niculescu, III, MD, PhD

Monday, March 28, 2016

A brief proposal for improving drug discovery

       We need to move from the primacy of therapeutics to the primacy of diagnostics, from “companion diagnostics” to “companion therapeutics”. Drug discovery greatly benefits from being based on correcting biological markers of disease.  Once you have such markers, you can identify drugs that modulate them individually, or modulate the biological pathways they are in, or modulate in opposite direction the gene/phene expression signature of a panel of top markers.
         The later approach is probably the best, as most diseases are complex, and by modulating a biosignature you are more likely to be more efficacious as well as more specific for the disease. For example, using a gene expression biomarkers signature, “dry lab” approaches such as connectivity map, or “wet lab” approaches such as de novo screening of compounds in cells or model organisms, may yield useful compounds. Using several low dose combinations of such compounds in subsequent clinical trials may synergize their abilities and minimize their liabilities, resulting in a broad-spectrum therapeutic with favorable side effect profile.  The biomarker signature can be monitored as a primary measure of response to treatment, providing real-time feedback.
      After the clinical trials are successfully completed, in subsequent actual clinical practice, tailoring the compounds in the combinations and their dosages based on the profile of biomarkers in a particular patient may yield additional benefit. Again, the biomarker signature can be used to monitor not only the efficacy and make any needed adjustment in dosages, but also as a way of gaining additional insights about stratification in a wider context, taking into account gender, possible subtype of disease, age, ethnicity, and, importantly, environmental factors.


Alexander B. Niculescu, III, MD, PhD

Saturday, April 11, 2015

Suicidality: from phenomenology to treatment

      Suicidality can be viewed as whole-organism apoptosis (self-poptosis). First, suicide might have evolved to occur adaptively due to life factors, to minimize detrimental impact and burden on the extended kin when the present is perceived as bleak and unsuccessful for the individual, and there is little hope for future improvement and successes, with the individual perceiving it is irreversibly damaged, not needed by family, extended kin, peer group, or society at large. Suicidality is a decision and choice leading to self-harming actions and behaviors driven by feelings and thoughts of pain or fear, hopelessness or anger, despondency or perceived uselessness. Conversely, a life that is perceived to be successful and meaningful is protective. Second, suicide often occurs maladaptively, due to mind abnormalities, with the individual who commits suicide being vulnerable due to a psychiatric illness, misperceiving circumstances and/or overreacting in an impulsive fashion. Related to that, suicide can be an attempt to assuage perceived guilt, or an attempt to harm (through social opprobrium or guilt) the individual(s) perceived to be the source of the lack of success of the suicidal person. Conversely, a well-balanced and functioning mind is protective. Third, body health abnormalities. Having severe health problems or pain may make an individual more prone to suicide. Conversely, a healthy, strong and resilient body is protective. Fourth, environmental factors. Environmental stressors, such as a hostile environment and loss of social connections and status can make individuals more prone to suicide. Conversely, a pleasant environment and good social standing are protective. Fifth, addictions. Addictions may destroy an individual’s life, mind and body, making them more prone to suicide. Conversely, being free from and immune to addictions is protective. Sixth, cultural factors. Cultural enablers (a personal or family history of suicidality, or being part of an environment or culture where suicide is an option) also come into play. Conversely, cultural and religious beliefs that make suicide not be an option are protective.

      Suicidality is a combination of increased risk factors/drivers (increased reasons why to do it) and decreased protective factors/brakes (decreased reasons not to do it). Younger people, older people, and males are more at risk for it. The evolutionary rationale in the young is primarily to avoid transmitting damaging combinations of genes. In the old, it is primarily to avoid being a burden. In males, it may be a price paid for increased testosterone-related drive and impulsivity.

      Discrepancies between where you are in your life vs. your standards and goals (where you would have liked to be in life) can lead to suicidality. Treatment and prevention need to involve reducing this perceived gap, in addition to improving hope for the future, and improving the mind (feelings and thoughts) that can color how life is perceived, improving body health, improving environment, improving any addictions, and improving cultural factors.


 Alexander B. Niculescu, III, MD, PhD

Sunday, March 15, 2015

Neuroscience and Heisenberg Uncertainty

The current sources of data for neuroscience research are naturalistic data (where there is no intervention and the subjects do not know they are being observed), and experimental data (where there is an intervention, and/or the subjects know they are being observed). Any naturalistic data may suffer from the fact that it was not specifically designed apriori to answer a particular question. Any experimental data may suffer from the Heisenberg uncertainty principle, where the intervention/observer modifies to some extent what is being observed. That holds true from induced pluripotent stem cells to testing of patients in clinical trials. What is the solution to circumvent these limitations and transform them into strengths? We believe that a convergent combination of naturalistic data with experimental data has the best yield, and should be programmatically pursued at all levels of neuroscience. For example, in developing tools to predict psychiatric disease outcomes such as suicide, a combination of naturalistic medical records and other life records mining with experimental neuropsychological and laboratory tests will yield the best outcome. Similar arguments can be made for drug development, and so on.

Monday, January 13, 2014

Less is more: the secrets of success in human (psychiatric) genetic and biomarker studies


1. Phenotype
- Study discrete quantitative phenotypes ("phenes"), not broad diagnostic categories.
- Study these phenes (for example,  mood, hallucinations, suicidality ) in high –risk populations ( bipolar schizophrenia), which provide an enriched pool.
- Study these phenes longitudinally, over time. Wireless devices and mobile health applications are crucial.

2. Cohorts
- When enrolling subjects, for reliability of phene measures use an longitudinal within-subject design whenever possible.
- Validate phene measures by the convergence of internal feelings and thoughts ( as measured by self-report scales) and of external actions and behaviors ( as measured by external raters).
- Separate cohorts by gender and by ethnicity, as this homogeneity leads to a reduction in noise.

3. Gene expression (biomarker discovery) is much more powerful than genetics (mutations discovery)
- One expressed gene may integrate the effects of up to  ~ 103 SNPs, epigenetic changes, as well as the current effects of the environment.
- Focus on discovering  first state biomarkers, correlated with phenes measured at the time of biomarker testing, not trait biomarkers, unless you have good longitudinal phene data. State over time is trait.

4. Study design 
- A within -subject design is the best, as it factors out genetic variability. You can do aggregates of n of 1 studies. You may need n~101 for gene expression studies, and n~ 104 for family based genetic studies (the closest you can come to “within-subject” in genetics).
- A case-case design is second best, as it factors out some disease related variability. You may need n~102 for gene expression studies, and n~ 105 for genetic studies.
- A case-control design is the least powerful, due to heterogeneity and noise that is not factored out, i.e. many of the differences do not have something to do directly with the phenotype you are studying. You may need n ~103 for gene expression studies, and n~ 106 for genetic studies.

5. Convergent Functional Genomics (CFG) at a gene level
- ~ 102 more reproducibility at a gene level than at a SNP level.
- Reproducibility in independent cohorts  is more important than strength of signal in the discovery cohort, as that could be a fit-to-cohort effect.
- CFG is like a magnet that finds the biomarker “needle” in the genetic or blood gene expression “haystack”. It  uses in a Bayesian way the whole prior body of work in the field to identify,  prioritize and give credence to disease-related genes from the long lists of differentially tagged genes in genetic association studies ( GWAS), and from differentially expressed genes in the brain or blood.
-Moreover, the genes and biomarkers prioritized by CFG are fit-to-disease, not fit to cohort. Because of that, they travel well, reproduce and are predictive in independent cohorts, which is the ultimate litmus test for any genetic or biomarker finding.  

Monday, August 12, 2013


How to prevent suicide     
 
      Suicidal behavior can occur due to a combination of existential reasons, biological vulnerabilities, and environmental factors. The existential reasons include lack of satisfaction with current life (health, finances, social importance), lack of hope for the future, and not being/feeling needed (especially by progeny and family). The biological vulnerabilities include mental health issues and addictions, including a genetic vulnerability to suicide as reflected in a family history of suicides among biological relatives. The environmental factors include increased stress and pain (physical and psychological), as well as cues that enable this behavior (previous attempts, knowing examples of other people who have done it, living in an environment and (sub-) culture where suicide is an option, seems attractive, and the means are available). 
      The opposites of each of the above items are protective. Individuals can have a mixture of risk factors and protective factors.
      A simple checklist of all these factors, tabulating risk factors and protective factors, along with improved objective biomarkers, should lead to very high levels of identification of individuals at risk, combining specificity with sensitivity of detection. This would permit preemptive intervention- changing and saving lives.
 
      We are working on that.

Friday, January 11, 2013

Trend of the Year 2013:

Genomics out, biochemistry in, for curing diseases?

Genomic variation and complexity are such that the genetic basis of most  human diseases is  going to require another decade or more to unravel completely, and even then it will be only partially explanatory, due to the profound role of the environment and of developmental history. Even when driver mutations are found and are targeted by a drug, the disease process may not be completely blocked, and alternative biological pathways may be recruited by the disease, as is the case in cancers.

However, diseases have key "vulnerabilities" at a metabolic and biochemical level that can be therapeutically attacked much sooner than that. Such metabolic targets are the result of the combinatoric integration of myriad mutations and environmental effects. For example, in psychiatry, the use of DHA (an omega-3 fatty acid), may correct membrane, signalling and inflammatory abnormalities that are responsible for vulnerability to stress, anxiety, mood and cognitive symptoms.

Genomics will still be important for scientific understanding in the long run, and for risk stratification and diagnostics in the short run.


Alexander B. Niculescu, III, MD, PhD