How are genes involved in tailoring psychiatric treatment?

Psychiatrists need an understanding of genome-wide association studies (GWAS), pharmacogenomics, and polygenic risk scores (PRS) so they can tailor treatment for their patients. At present, the utilities of these techniques are evolving. GWAS are used in research, pharmacogenomic testing is only occasionally indicated in the clinic to improve treatment outcomes, and PRS testing is expected to play a substantial role in the assessment of risk for disease, subtypes of disease, and treatment response in the future, explained Professor John Nurnberger, Indiana University School of Medicine, Indiana, to a large audience of clinicians at Psych Congress 2019.

Not one gene, but multiple genetic factors

GWAS screen single nucleotide polymorphisms (SNP) of the entire genome at intervals of 10–50 Kb. Thousands of samples are required, said Professor Nurnberger. Findings relevant to psychiatry so far include:

  • the identification of 128 independent associations spanning 108 conservatively defined loci in schizophrenia1
  • significant loci containing genes encoding ion channels, neurotransmitter transporters and synaptic components in bipolar disorder (BPD)2

Pharmacogenomics provides the basis for precision medicine and should improve patient care

Pharmacogenomics testing is currently only occasionally indicated in the clinic

Responses to and the adverse event profiles of pharmacologic treatments are modulated by genetic factors. For instance, the cytochrome P450 (CYP) enzyme system is involved in the metabolism of antidepressants and antipsychotics. Two of these enzymes — CYP2C19 and CYP2D6 — are featured in FDA warnings, which highlight a maximum dose of a given medication for slow metabolizers of these enzymes.

Pharmacogenomic tests therefore provide the basis for precision medicine in psychiatry and, with careful development through clinical trials, should improve patient care, said Professor Nurnberger.3

Patients receiving pharmacogenetic-guided therapy were 1.71 times more likely to achieve symptom remission

A systematic review and meta-analysis of 1737 eligible subjects in five randomized controlled trials (RCTs) that examined pharmacogenetic-guided decision support tools (DSTs) in major depressive disorder (MDD) has shown that individuals receiving pharmacogenetic-guided DST therapy were 1.71 times more likely to achieve symptom remission compared to individuals who received treatment as usual.4

The Genomics Used to Improve Depression Decisions (GUIDED) Trial was an RCT of 1167 outpatients. Pharmacogenomic testing was effective in improving response and remission rates among patients with treatment resistance, particularly for patients who are treated with medications incongruent with their genetic profile. However, the primary outcome — symptom improvement at week 8 — did not reach statistical significance.5

Nevertheless, the International Society of Psychiatric Genetics Statement on Testing for Choice of Treatment, 2019, comments “evidence to support widespread use of other pharmacogenetic tests at this time is inconclusive, but when pharmacogenetic testing results are already available, providers are encouraged to integrate this information into their medication selection and dosing decisions. Genetic information for CYP2C19 and CYP2D6 would likely be most beneficial for individuals who have experienced an inadequate response or adverse reaction to a previous antidepressant or antipsychotic trial.”

The quality and utility of pharmacogenomic tests should increase over time

It is reasonable to consider a commercial pharmacogenomic test for treatment-resistant patients, said Professor Nurnberger, but skepticism is warranted because:

  • they contain a combination of well-supported assessments (e.g., CYP2D6) and poorly-supported tests (e.g., serotonin transporter assessment)
  • recommendations are generally based on proprietary algorithms that combine information from multiple gene variants using rules not revealed to practitioners or consumers
  • most testing companies have not conducted controlled trials of their products

PRS is expected to play a substantial role in assessment and treatment response in the future

The PRS reflects the genomic burden of genetic risk variants for a particular disease in an individual, said Professor Nurnberger, and methods for calculating it have been developed in the last 10 years. Each risk variant is given a value based on its effect size, and each individual in a target sample receives a total score based on the number and value of the risk variants they carry.

The optimal predictive power of a PRS is obtained by using hundreds or thousands of variants.6 The current predictive power of PRS for discrimination between patients and controls is about 82% for schizophrenia, 65% for BPD, 58% for MDD, and 54% for anxiety,7 said Professor Nurnberger. The predictive power can be increased:

  • using statistical methods
  • by weighting by gene expression or other biological processes
  • by increasing sample sizes for discovery sample8,9

PRS can indicate liability for developing MDD

A recently published study has demonstrated that PRS liability for MDD is associated with first depression in the general population.10

Professor Nurnberger concluded that although a PRS is generally not informative of disease status for psychiatric diagnoses at an individual level at present, with further development the PRS is expected to result in substantial clinical utility in the assessment of risk for disease, subtypes of disease, and treatment response in the future.

Our correspondent’s highlights from the symposium are meant as a fair representation of the scientific content presented. The views and opinions expressed on this page do not necessarily reflect those of Lundbeck.

References

  1. Ripke S, et al. Nature. 2014;511:421–7.
  2. Stahl EA, et al. Nat Genet. 2019;51:793–803.
  3. Nurnberger J, et al. J Clin Psychiatry. 2019;80(1).
  4. Bousman C, et al. Pharmacogenomics. 2019;20:37-47.
  5. Greden JF, et al. J Psychiatr Res. 2019;111 :59-67.
  6. Fullerton J, Nurnberger J. F1000Res. 2019; 8: F1000 Faculty Rev-1293.
  7. So H, et al. Bioinformatics 2017;33:886–92.
  8. Khera AV, et al. Nat Genet. 2018;50:1219–24.
  9. Wang D, et al. Science. 2018;362(6420).
  10. Musliner K, et al. JAMA Psychiatry. 2019;76:516–25.