No other aspect of behavioral biology is as well understood at the molecular level as the mechanism that generates and sustains circadian rhythms. If you are following science in general, or this blog in particular, you are probably familiar with the names of circadian clock genes like per, tim, clk, frq, wc, cry, Bmal, kai, toc, doubletime, rev-erb etc.
The deep and detailed knowledge of the genes involved in circadian clock function has one unintended side-effect, especially for people outside the field. If one does not stop and think for a second, it is easy to fall under the impression that various aspects of the circadian oscillations, e.g., period, phase and amplitude, are determined by the clock genes. After all, most of these genes were discovered by the study of serendipitously occuring mutations, usually period-mutations.
If the circadian properties are really deteremined by clock genes, then the predictions from this hypothesis are that: 1) every cell in the body shows the same period (phase, amplitude, etc.), 2) every cell in the body has the same period throughout its life, 3) every cell removed from the body and placed in the dish continues to oscillate with the same period as it had inside the body, and 4) the properties of the circadian rhythms are not alterable by environmental influences.
Stated this bluntly, one has to recoil in horror: of course it is not determined by genes! But without such an exercise in thinking, much work and writing (especially by the press) tacitly assumes the strict genetic determination. However, the experimental data show this not to be true. Period (and other properties) of the whole organism’s rhythms are readily modified by environmental influences, e.g., light intensity (Aschoff Rule), heavy water, lithium, sex steroids and melatonin. They change with age and reproductive state. There is individual variation even in clonal species (or highly-inbred laboratory strains). The period of the rhythms measured in cells or cell-types in a dish is not always the same as exhibited by the same cells inside the organism. Finally, the occurrence of splitting (of one unified circadian output into two semi-independent components differing in period)suggests that two or more groups of circadian pacemakers simultaneously exhibit quite different periods within the same animal.
Several years ago, Dr. Eric Herzog (disclosure – a good friend) has shown that even the individual pacemaker cells within the same SCN (suprachiasmatic nucleus – the site of the main mammalian circadian clock in the brain) exhibit different periods. When dispersed in culture, pacemaker neurons (originally taken from a single animal) tend to show a broad variation in periods (amplitudes, etc…).
As they grow cellular processes, two neurons in a dish may touch and form connections. As soon as they do, their periods change and from then on the two cells show the SAME period, i.e., they are synchronized. As more and more cells connect, they build an entire network of neurons, all cycling in sync with each other – same period, same phase, same amplitude. This is assumed to happen inside the whole animal as well – the unconnected SCN neurons of the fetus start making connections just before (or immediately after, depending on the species) birth and as a result, an overt, whole-organism rhythm emerges out of arrhytmic background.
But, what molecules are involved in cell-cell communication that allows the pacemaker cells to synchronize their rhythms? For several years, the most likely candidate was thought to be GABA, an inhibitory neurotransmitter produced by all SCN cells. Sara Aton, a graduate student in Eric Herzog’s lab (now postdoc at UPenn), set out to test this hypothesis. Over a few years and several papers, a different story emerged, culminating in this months paper in PNAS:
GABA and Gi/o differentially control circadian rhythms and synchrony in clock neurons. (by Aton SJ, Huettner JE, Straume M and Herzog ED).
What the paper shows – and there is a lot of detail there, so you can read the paper for yourself if interested, or at least the media coverage (here, here and here) – is that various perturbations of the GABA system, either at the synthesis end or the reception end, have, at best, some mild effects on amplitude and phase. There was no effect of GABA on period of individual pacemaker neurons. Yet, effect on period is neccessary for mutual synchronization of cells into a network. Instead, VIP (vasoactive intestinal polypeptide) was shown to be the agent that, by modulating period, allowed spatially coupled cells to also temporally couple – to synchronize their circadian oscillations.
This is a much more important finding than you may think at the first glance. The naive idea of a single clock driving all the overt rhythms has been abandoned for more than half a century. Every important problem in chronobiology – coherence of rhythms, temperature compensation, communication between pacemakers and peripheral clocks, entrainment to environmental cycles, etc. – hinges on the properties of the multi-clock networks. Understanding the biochemical mechanism by which pacemaker cells syncronize with each other is thus a key finding that will allow us to study those phenomena at a cellular and molecular level. Right now, and due to Sara Aton’s work, VIP is the “handle” we will use to revisit those old problems and test our pet hypotheses about the coupling of circadian systems in various animals (the reasonable assumption being that mouse is not unique in using VIP and that this molecule is probably used for the same function in all vertebrates). We can now study exactly how two cells communicate by VIP to synchronize their clocks – is the pattern of VIP release, for instance, used as a kind of temporal “code”?
Probably the most important such phenomenon to study is splitting. Different kinds of spliting have been observed in lizards, starlings, tree-shrews (Tupaia), mice, rats, hamsters, marmosets and many other animals under various experimental conditions, e.g., constant light, constant darkness, removal of the pineal, infusion with testosterone, or exposure to skeleton photocycles. Splitting can be induced by highly artificial experimental protocols, e.g., alternate eye-patching, or they may appear spontaneusly in animals out in the wild, something that can be replicated in the laboratory.
In some cases, it has been shown that the splitting is lateral, i.e., left SCN drives one component and the right SCN drives the other. In the case of alternate eye-patching, it is reasonable to hypothesize that the same thing is happening. But in other cases, it is more likely that each SCN splits into two subsets of neurons, synchronized within but not between the two groups. Is VIP the synchronizer in both groups? In all cases? In all animals?
If, for instance, rhythms split into two componenets under the influence of testosterone, is VIP used for coupling within each of those two semi-independent “clocks”? Does one group of neurons, insensitive to testosterone, use VIP to synchronize its output, while the other group of neurons, under the influence of testosterone, changes its period and also uses VIP to synchronize its output?
Now that we know to use VIP and not GABA as an entry into the system, all of these questions will be much more amenable to future research – an exciting prospect for me and many others in the field.