Common polymorphisms for the time of living and death?

Abstract

EDITORIAL Common Polymorphisms for the Time of Living and Death? A ll organisms on Earth live with the daily light–dark cycle and adapt their physiology and behavior to environmental circadian rhythm factors. Many cellular mechanisms and physiological functions are known to exhibit circadian rhythmicity, from regulation of the cell cycle to metabolism. 1 Elegant genetic studies in model organisms revealed a molecular clock conserved from invertebrates to humans for circadian regulation. At the core of the molecular clock that drives the daily rhythm is an interlocking, positive and negative, transcriptional– translational feedback loop comprised of CLOCK, BMAL1, PER1-3, CRY1-2, and other genes. 2,3 For humans living under culturally enforced time schedules and artificial illumination, there are expected interactions between the body clock (endogenous rhythm) and the social clock. 4 Many measurable human-specific biorhyth- mic behaviors may exist due to these interactions. Fur- thermore, as the human population has expanded explo- sively in recent history, a large number of genetic variations deposited in human genomes on a population level may exert subtle influences on human biorhythmic behaviors. Defining more measurable biorhythmic behav- iors and identifying the genetic basis for interpersonal differences may represent the next exciting frontier in human circadian research. The report by Lim et al in this issue is an example of research in this direction. 5 They carried out a candi- date gene association study with 135 tagging single nu- cleotide polymorphisms (SNPs) covering 18 candidate genes within circadian pathways for normal participants from an unrelated aging study. They determined a com- mon SNP, rs7221412, in the adjacent region downstream of PER1, to be associated with 2 traits, the activity acrop- hase from actigraph (midpoint of active hours) and the time of death. Subjects with the rs7221412 GG genotype on average were found to have an activity acrophase 67 minutes later than subjects with the rs7221412 AA geno- type. Subjects with rs7221412 GG were also found on av- erage to have a time of death 7 hours later than subjects with rs7221412 AG or rs7221412 AA in participants from a nonoverlapping cohort mostly from the same study. These findings were replicated in a small independent cohort with participants of similar ethnic background and of a much younger age. In addition, they validated the activity acrophase measures using cosinor analysis, the time of dim light melatonin onset, and body temper- ature nadir in the replicate cohort, further strengthening the value of using activity acrophase as a real world marker of circadian rhythm. Interestingly, the rs7221412 genotype did not show association with the period length of intrinsic biological rhythm, although it had a clear influence on entrained phase, supporting separate regula- tory mechanisms for period length and phase of entrainment. In many ways, this study is innovative and reflects many difficulties in identifying common polymorphisms for circadian rhythm in human populations. First of all, it is always more difficult to work with human subjects for studying circadian rhythm compared with other model organisms. In classical studies, 6 subjects were put under controlled lighting conditions to block environ- mental time cues. However, using actigraph, as Lim et al did for their discovery cohort, has the benefit of being convenient for recording activity for days with minimum interruption in the normal lives of a large number of subjects. This method also gives an objective measure of activity instead of relying on self-reporting. However, this approach inevitably increases the variance of measure- ments, as many unexpected and unrecorded factors may influence results. For example, differences in active hours between weekdays and weekends are expected, although they may be smaller for retired people or inpatient sub- jects. This may have contributed to the effect size observed in the Lim et al paper. Sophisticated statistical treatments are needed for repeated measure data from circadian studies to correct for these factors. Validation of novel measures of biorhythmic activities in additional cohorts is also necessary. One intrinsic difficulty in association studies with common SNP markers is demonstrating biological rele- vance. The Lim et al paper shows suggestive differences in daytime PER1 expression between subjects with differ- ent genotypes at rs7221412. For circadian genes with oscillating expression patterns, comparison of expression can be skewed by the time of sampling. Expression profiling at precise timing or multiple standardized time points is necessary to compare circadian gene expression, which is often not feasible when using human subjects. C 2012 American Neurological Association V