Full Conversation between Dr. Gibbons and Dr. Susan Harbison 

Dr. Gary H. Gibbons, Director, National Heart, Lung, and Blood Institute:  Today my guest is Dr. Susan Harbison, one of our Earl Stadtman tenure track researchers in the NHLBI Intramural Program.

Susan, thank you very much for taking the time to share with me some of your thoughts today, and describing your program and your interest in science. Why don’t you just start by telling me a little bit about your program? What is exciting you right now about what you’re working on?

Dr. Susan Harbison, NHLBI:  Well, one of the things we did recently was a genome-wide association study of sleep in the fruit fly, drosophila melanogaster. So we’re using flies as a model system. There is a community resource, the Drosophila Genetic Reference Panel. This is a group of lines that was created by obtaining flies from—they were wild-caught in Raleigh, North Carolina. They were inbred for 20 generations to produce identical sets of lines. So, [flies in] each line have the same genotype. I measured sleep on these lines. I measured things like sleep duration, the number of sleep-bouts or naps, the average sleep-bout length that the flies take and waking activity which is a measure of how hyperactive the flies are. 

Dr. Gibbons:  So, how did you measure how a fruit fly naps?

Dr. Harbison:  Oh okay, so we used the drosophila activity monitoring system. This is an infrared based system. The flies are placed; each one is singly housed in a glass tube that’s about three inches long. The fly walks back and forth across the tube. He has got food there and a little plug to keep him from flying away. We place those in the drosophila activity monitoring system, and that monitor has an infrared beam that bisects the tube. So, when the fly walks back and forth, he just breaks the infrared beam, and that tells us whether or not he’s active. So, from that data we get a series of text files that just have numbers of counts per minute of recorded time. We can decipher from that how long the fly has been awake, how long he has been sleeping, and we can calculate these other parameters that I was mentioning, how many sleep bouts or naps that they had, and what the average sleep-bout length is. 

Dr. Gibbons:  That’s fascinating. Another naïve question, so in humans, when you’re asleep, your metabolic rate changes, other physiologic changes. The brain activity changes, does the fruit fly simulate those elements of sleep in mammals? 

Dr. Harbison:  Well research in sleep in fruit flies is relatively new. They were first shown to have all the same behavioral characteristics of mammalian sleep in 2001 in two different papers. The characteristics are that they have an increased arousal threshold, so that you or I, if you’re asleep, it may take a little bit longer to get your attention. So that is one thing that the flies have. They also have a homeostatic response. So, if you deprive them of sleep, they want to make up that lost sleep. They have a defined circadian period of quiescence. They are diurnal as we are, so they tend to sleep at night. 

They also have a defined resting place, and that is close to the food. So they like to be near their food source when they’re asleep. So, all of those characteristics are very similar to mammalian behavior except perhaps the fact that we don’t sleep next to a refrigerator. 

Dr. Gibbons:  Some people do. 

Dr. Harbison:  But mammals out in the wild will tend to sleep in a den close to their food if they store it. 

Dr. Gibbons:  That’s fascinating, so some of the physiology sounds similar. What have you found so far in terms of some of the molecular pathways? Do they have BMAL and PER and CLOCK, and all those sorts of networks in the same way? 

Dr. Harbison:  Yes absolutely. In fact, circadian rhythms, the canonical clock genes for circadian rhythms, were first discovered in fruit flies. So, the genes period and timeless, clock and cycle were first discovered in flies in the late 1960s and then also in the ’80s. So there is a definite relationship between, or conservation between flies and mammalian circadian systems. We think that that conservation extends to sleep behavior as well. We are finding a lot of genes that have mammalian and human homologs in, for example, my genome wide association study of sleep.   

Dr. Gibbons:  Okay so tell us a little bit more about what you have unlocked in doing this GWAS related to this phenotype. What insights has it given us to understanding sleep better?

Dr. Harbison:  So, there are several things that we found that were really intriguing. When we did the genome-wide association study, we first associated the sleep phenotypes with just over 2.4 million SNPs that were available in the Drosophila Genetic Reference Panel. So, that in and of itself is a little different from human genome-wide association studies that would normally take advantage of SNP chips or something like that has on the order of 500,000 different SNPs. This is taking advantage of all the polymorphisms that are available. 

So, one thing that we found was that there were many different polymorphisms that are found in numerous developmental genes, including genes involved in central nervous system development. So that is really intriguing. It begs the question, are these genes a reflection of the neural circuitry of the fly? So, is it something that sort of comes about as a result of the neural development, or is the case that these genes are operant in the adult stage? We don’t know which yet, but that was one of the things that came out of the study. 

The second thing that is really interesting is, as I mentioned before, each of these lines has an identical genotype, or each of the flies in the line has an identical genotype. So, what we observed was a variation in sleep in flies of identical genotypes. That is unusual. It’s not the sort of thing that you could really see in a human study unless you are talking about looking at monozygotic twins. But with the flies, it’s a very powerful tool in that we can do multiple measurements on identical genotypes, and I did that for 64 flies per line. So, we got quite a range, quite a number of flies per line. 

We were looking at the data and I decided to just plot out the differences, rather the sleep phenotypes, for each fly. What I found was the short sleeping lines were far more variable than the long sleeping lines, and that this variability we could actually quantify as the coefficient of environmental variation. So, this is a reflection of the environmental sensitivity of the flies to changes or random perturbations in the environment. That could have a potential impact on human genome-wide association studies, because, as I’m sure you’re aware, there are many association studies that have, or find a disease allele and that allele is present in the population that does not have the disease. So, if you are environmentally sensitive, then perhaps your sleep phenotype could be quite variable. At least in the flies, this was true of the short sleeper lines and not of the long sleepers. 

Dr. Gibbons:  That’s fascinating. Is there any data suggesting the role of epigenetic changes in explaining any of that at all? What do you think is driving that influence? 

Dr. Harbison:  That has not been well characterized yet. So, I don’t want to say at this point. 

Dr. Gibbons:  That’s interesting. 

Dr. Harbison:  I think it is a very interesting question for further study. 

Dr. Gibbons:  Where do you see your work going next? Is there a particular finding that you are intrigued by that you would like to forge further in?

Dr. Harbison:  So, one of the things that we’re doing is verifying the polymorphisms that we found in the genome-wide association study. We are using an artificial selection scheme to do that. So, we started with the five most extreme short sleeping lines of the DGRP, and the five longest sleepers, and we crossed them together to produce an outbred population. Now we’re selecting these flies for short sleep and for long sleep, and what we’re finding is that if you sort of stack the deck in this way, then it doesn’t take very many generations to get an extreme response in the sleep. The short sleepers now sleep about six and a half hours less during the night than they did at the beginning of the experiment after eleven generations. The long sleepers sleep about three and a half hours more. So, it’s very interesting and because we know where the SNPs are, we can use high-throughput genotyping methods and also genome DNA sequencing to follow up on that and find out which SNPs were most important across generations in producing these phenotypes. 

Dr. Gibbons:  That is fascinating. Recognizing that the parallels of this model back to mammals and humans, any particular implications of your work that might relate to, again, human sleep biology or things that are related to the disease phenotypes in humans? 

Dr. Harbison:  Well, of course, we did look at how many genes overlap with known human homologs first of all, and human homologs that had been previously discovered in sleep studies. There were ten of them that had been seen before in human sleep studies. I want to say several of them were involved in day-time sleepiness in humans, usual bedtime, and this was your normal sleep characteristic. This was a GWAS study previously done by Dan Gottlieb and colleagues. There were also some studies done on obstructive sleep apnea, narcolepsy, and restless leg syndrome, and we found overlap from the human genes to our fly genes as well. So, there is definitely a suggestion of conservation there, and we are following up on that, pursuing it a little bit further. 

Dr. Gibbons:  Interesting. I am aware of some studies in humans that again, if you perturb sleep, it affects appetite, diet so forth. Are these same sort of perturbations mimicked again in the fly model that you’re studying? 

Dr. Harbison:  I think people are very interested in measuring that. There hasn’t been a whole lot of work done on it yet. There have been some studies looking at the altering circadian rhythms in flies, and seeing how that affects feeding behavior. That was done in Amita Segal’s lab. Also when I was a post-doc with Amita, we did sort of a preliminary experiment where we looked at sleep deprivation and how that affected three different measures of energy stores in the fly. It was triglycerides, protein and glycogen content. It was interesting that sleep deprivation affected glycogen content in males and it tended to affect the triglycerides in females. 

Dr. Gibbons:  Any thought about the sex differences? What is driving that?

Dr. Harbison:  It’s very interesting. There are huge sex differences in male and female flies. I sort of speculate that that might be related to the males’ courtship behavior. They have a burst of activity at dawn, and one at dusk. The females tend to be sort of active at a lower level than the males, but throughout the entire day. All of the flies take a little siesta during the day, but for the females, it’s much shorter than the males. The males, in some cases, sleep all day long, and just wake up for dawn and dusk. 

Dr. Gibbons:  Interesting. 

Dr. Harbison:  That’s when most of the mating takes place. 

Dr. Gibbons:  I see. 

Dr. Harbison:  So, you can see what they woke up for. 

Dr. Gibbons:  Fair enough. I guess another simulation to the human model. I also understand you have an interesting background in that you started off as an engineer before moving into genetics. Tell me a little bit about that.

Dr. Harbison:  I did. I was an aerospace engineer, and I worked for McDonnell Aircraft Company as well as for the Naval Aviation Depot. That was for the Department of Defense. I wrote maintenance and structural repair instructions on a variety of different naval aircraft, the A-4, the C-130, and the H-53 helicopter. I did some work too on the Tomahawk Cruise Missile. So, one thing I realized while I was doing that work was that I was really interested in those projects where I had to gain a lot of data and then try to synthesize or figure out what was going on with it. Like, looking at maintenance records to try to figure out how long a structural component would last, and therefore what the fatigue life or the flight critical part would be. So, I knew I was really interested in research. 

I was in sort of a production environment at the time. And along came the human genome project, and I really got fascinated with genetics. I thought it was really intriguing that there was the possibility that we could quantify or define risk to disease and understand, perhaps, our own evolution by comparing different organisms and different species, to also look at, or understand physiological processes like lifespan. So, I got really interested in it. I was also, I had a couple of basenjis, those are dogs, and I was showing them. So, I got very interested in animal breeding. So, I thought perhaps a venture into quantitative genetics would be well-suited to my computational and mathematical background.  So, I started working as a PhD student in Trudy Mackay’s laboratory at NC State. 

Dr. Gibbons:  That’s fascinating. As part of our audience, we often have young investigators, perhaps related to your unique journey, any words of advice or wisdom that you might impart to young investigators?

Dr. Harbison:  Okay. So, one thing to definitely have in mind is a research question that you are very passionately interested in, and the willpower to follow it up with a lot of perseverance.  This is because as anyone who has spent a lot of time in the lab realizes that most experiments don’t work out. So, you have to be able to continually try, try again, in the face of repeated failure. Then the other thing to recognize too is that even when your experiment does work out, that you must realize that now you have to publish that work, and it’s not always easy to do that. Sometimes it will take a submission to several different journals before it’s accepted, especially if the work is really novel or controversial. That process usually strengthens the paper, and it’s something that you end up being proud of at the end. But for the very young investigators, there is this tendency and it was certainly true for me as well to think that well I’m going to go in and do this experiment and everything is going to be great, and a lot of times it doesn’t work out that way. 

Dr. Gibbons:  Good, good. Well thank you very much.