AUSTRALIAN FRONTIERS OF SCIENCE, 2003
Canberra, 31 July to 1 August 2003
Evolutionary
developmental physiology – how does individual developmental plasticity
lead to heterochrony?
by Dr Sandra Orgeig
 |
Sandra Orgeig is an ARC Research Fellow in Environmental Biology at the University of Adelaide. After completing her honours degree at the University of Cape Town, she gained an Overseas Postgraduate Research Scholarship in 1990 to undertake a PhD in Physiology at Flinders University. Following postdoctoral work in physiology at the University of Adelaide, she gained an ARC Postdoctoral Fellowship in 1996 and an ARC Research Fellowship in Environmental Biology in 1999. She is the recipient of the Fenner Medal and Young Tall Poppy Award in 2002 for her work on the evolution of the pulmonary surfactant system, and the role of cholesterol in regulating the biophysical function of this lipid-protein mixture at the air-liquid interface of the lung during thermal fluctuations. Sandra has published over 50 scientific papers. Her current interests are in evolutionary developmental physiology, where she uses the surfactant system as a model to examine the effect of environmental factors on lung development. |
In my view, just about any biological system is inherently complex, but the way in which we analyse or examine a system can mean that we can increase the level of complexity of that system. For example, if we are looking at the development of a physiological process within an embryo, we are adding other factors such as time and the environment, and that increases the level of complexity. If we then look at the evolution of the development of that physiological process across a whole range of species, we are adding another level of complexity.
So what we are particularly interested in is to determine how variability and changes in a developmental process within an individual can lead to the incredible variability that we see in that same developmental process across a whole range of different species.
When we are looking at the evolution of development, there are a couple of terms that are particularly important and that I will be concentrating on today. The first of these is 'heterokairy', which is a brand-new term coined by Burggren and Spicer in a publication appearing this year. It is defined as the changes in the time and/or rate of appearance of a particular system within a species. 'Heterochrony' is a much older term, which is defined as the evolution of different developmental rates or times between species or groups of animals.
What we want to know is: can heterokairy (variability within an individual species) lead to heterochrony (changes in evolution across species) via classical Darwinian selection?
In order to answer this sort of question we need to not simply describe a developmental process but we need to understand how it occurs, how it is controlled and, in particular, how it evolves. So what are the specific selective criteria that are important in shaping the evolution of that developmental process?
For example, across the vertebrate species there are a range of systems that consist of fairly similar cell types and tissue types, and that have a similar overall anatomical structure, yet they differ dramatically in function. In essence, we have a subset of relatively conserved genes that can give rise to the spectacular phenotypic diversity in function. What we are interested in is how this phenotypic diversity evolves. Presumably, it requires some sort of process of heterokairy leading to heterochrony, but what are the selective pressures that these processes can act upon?
Once we have identified those evolutionary or environmental forces, we need to know how they drive the developmental process. That is, how do they exert their control? This means that we have to identify extrinsic and intrinsic controllers, and determine how they interact. So, once we have defined the environmental force, how is its effect mediated within the organism? That is, what intrinsic controllers are affected by it?
This leads us to one of the most basic, fundamental questions in developmental physiology: when do environmental conditions and when do pre-programmed events control the development of vertebrate embryos? Is the initial development purely controlled by the genetic program, and do environmental factors become important later on in modifying the functional development? What is the interplay between genetic and environmental factors?
The system that we have been concentrating on as a model is the pulmonary surfactant system. This is ideal for these sorts of studies, because it is highly conserved across vertebrate species and is also a complex system – it consists of numerous components – but it is essentially located within a particular cell.
Click on image for a larger version of figure 1
We all know that our lungs consist of ~25 generations of airways that end with these blind-ended sacs called alveoli. This is where gas exchange takes place. Figure 1 shows a cross-section through an alveolus, it is lined by a thin layer of fluid, called the hypophase. If this hypophase consisted purely of water; it would have a high surface tension, which would lead to high collapse pressures and therefore lead to the sticking together of the epithelial surfaces. In order to overcome these physical problems, the type II cell of the alveolar epithelial lining produces a substance called pulmonary surfactant. This is a lipid-protein mixture that lines the air-liquid interface and modifies the surface tension with changing lung volume.
Click on image for a larger version of figure 2
Surfactant is synthesised within the rough endoplasmic reticulum and golgi apparatus of the type II cell. It is stored in these onion-like appearing organelles called lamellar bodies, and upon an appropriate signal these lamellar bodies are exocytosed into the hypophase (figure 2). There they unravel to form this cross-hatched structure called tubular myelin. And it is from the tubular myelin that the individual lipids bud off to form the functional surfactant film.
Click on image for a larger version of figure 3
This film functions to vary the surface tension as the lung inflates and deflates. It does this through a series of lipids. We classify them into the disaturated phospholipids (DSP) – the main surface tension controlling agents, which at low lung volumes are able to virtually eliminate surface tension – and the unsaturated phospholipids (USP) and cholesterol, which tend to act as fluidisers (figure 3). They enable the surface film to respread across the alveolar surface area as the lung expands. They are also particularly important in animals with a low body temperature, because they enable the surfactant film to remain in a spreadable and fluid state at those low body temperatures.
Also important are four specific surfactant-associated proteins, SP-A, SP-B, SP-C and SP-D. SP-B and SP-C are hydrophobic proteins; they interact directly with the lipids and enhance the biophysical activity of this film. SP-A and SP-D are hydrophilic surfactant proteins, which are involved in the innate immune functions of the surfactant system.
Click on image for a larger version of figure 4
What about the evolution of the surfactant system? We know that all vertebrate groups contain members with lungs, and lungs evolved at least twice – first within the ancient placoderms, about 500 million years ago, and at least once but possibly more often within the Osteicthyes, which are the fish. The Osteicthyes include the primitive ray-finned airbreathing fishes, the modern teleosts – which include, for example, the goldfish – and the lobe-finned lungfish. The lungfish are closely related to the stem ancestor that gave rise to all the terrestrial vertebrates, or the tetrapods (figure 4).
The lungs differ dramatically in structure, function and also embryological origin, particularly amongst the fish – some have swim bladders and some have lungs, and these have different embryological origins. Yet despite these differences they all have a surfactant system. We have shown that the surfactant system is highly conserved, in terms of morphology and biochemistry and also in the way its synthesis and secretion are controlled.
Click on image for a larger version of figure 5
We have also shown that the surfactant system has had a single evolutionary origin, that pre-dates the evolution of lungs. We have shown this through analysis of the conservation of the surfactant proteins. Figure 5 is a western blot showing lung-washings (called lavage) from a whole range of species that have been probed with an anti-human SP-A antibody. We can see that the lung-washings of all these species cross-react with the anti-human SP-A antibody, so it is highly conserved. In particular, we have shown the presence of the protein here in two lungfish, two airbreathing fish and in the goldfish swim bladder. In particular, the 66 kD dimer is the prevalent form.
So the fact that these different lungs and swim bladders, with their different embryological origins, all express this system means that it actually evolved before the evolution of the lungs. We have shown this for the other proteins and for the messenger RNA as well.
Click on image for a larger version of figure 6
If the surfactant system is so highly conserved, is its development also conserved? That is, does it develop at the same rate in the different species? And if not, why not, and what can this tell us about the evolution of the process?
We can look at development of the process morphologically. These are isolated type II cells from crocodile embryos at day 60 of incubation through to hatching, which is at about day 80. We can see that there is an increase in the number of the lamellar bodies, which are the storage organelles for surfactant, through the developmental process (figure 6).
Click on image for a larger version of figure 7
We can also look at the development biochemically. Figure 7 shows one indicator: the amount of phospholipid within the lavage fluid (lung-washings). We show that for a range of species across the final 30 per cent of incubation. We can see that in general there is an exponential increase in the amount of phospholipid that accumulates within the lungs over this developmental period. Note the break in the axis [between 7 and 20 on the Y axis] - so the sea turtle here is really very large.
There are some differences, though, in that the chicken, for example, peaks at 80 to 85 per cent. This correlates with the time when the chicken actually begins airbreathing, because it breathes internally in an air cell. Another difference is here at 70 to 75 per cent, where [compared with the chicken and sea turtle] there is virtually no phospholipid in the crocodile and nothing in the bearded dragon. So there are these differences in the timing, and this is an example of heterochrony.
Click on image for a larger version of figure 8
We have looked at a few of the underlying factors that might be controlling this heterochrony, or leading to it. We have looked at basal secretion of phosphatidylcholine, which is the major phospholipid in surfactant, by isolating the type II cells and simply measuring the amount of phosphatidylcholine secreted (figure 8). We can see in general that there is not a huge difference – they are relatively constant. There is a kick up here at 100 per cent, but it is nowhere as dramatic as in the previous slide. So basal activity of the cells themselves cannot account for that large increase. There are, obviously, other factors involved.
Click on image for a larger version of figure 9
When it comes to type II cells in surfactant synthesis and secretion, a range of factors are important. Adenergic and cholinergic agonists interact directly with receptors on the type II cell membrane. Glucocorticoids interact indirectly via fibroblasts to increase surfactant synthesis and secretion. Thyroid hormones such as T3 interact with nuclear receptors to increase surfactant synthesis and secretion (figure 9).
If we are trying to identify any of these factors that could account for the observed heterochrony, we need to determine whether there is a difference in the time at which these factors stimulate surfactant. And, if there is a differential effect, then turning on or off any of these neurohormonal factors during the developmental process can have a dramatic effect.
Click on image for a larger version of figure 10
First we looked at adrenaline, by stimulating isolated type II cells with adrenaline and observing the secretory response (figure 10). We are looking here at levels greater than 100 per cent, and we see that adrenaline basically functions at all times, in all species, before, during and after hatching. So it is not a good candidate for explaining the observed heterochrony.
Click on image for a larger version of figure 11
What about glucocorticoids? We used dexamethasone as a synthetic glucocorticoid, and again we measured PC secretion from type II cells (figure 11). We see that in this case there are discrete windows of opportunity where glucocorticoids are acting within the different species. And these discrete windows are different for the different species. Here is the sea turtle; it is responsive here [at 70-75%] but not again thereafter. The crocodile responds here [at 90-95%] but not before and not after. So glucocorticoids are potentially a good candidate for explaining the observed heterochrony.
Click on image for a larger version of figure 12
What about heterokairy? Figure 12 illustrates the process of heterokairy in more detail. Here, in panel A, is a developing egg [left side of time line], which develops along this time frame here [horizontal baseline]. Hatching occurs at this time point here [vertical line on the right]. This [hatched area] illustrates a single regulatory system that consists of three different components [3 horizontal lines]. Development begins here at this time point [point of deviation from horizontal timeline] and is completed here [plateau] for this particular regulatory system, once all three components are established.
In panel B we have three different regulatory systems, each consisting of three different components. There are different aspects here to development. There is a different onset of timing for when the developmental process begins in the different systems, there are different rates at which each of them develops [slopes], and there is a specific sequence in which the development of the different processes is completed.
In panel C we can see what happens if there is a disturbance to the developmental process, be it genetic, environmental or phylogenetic. We can see that the onset of development has been altered – it has been brought forward for this one [white] and this one [grey] – the rates have altered, and the sequence in which these developmental processes are completed has been altered. All these alterations in the timing of the development fall under the banner of heterokairy. So how can we demonstrate whether these alterations through heterokairy can lead to the observed heterochrony?
In order to do this we need to be able to first demonstrate whether environmental factors can alter the developmental process. We do this by experimentally manipulating embryos. This can help us, because by experimentally manipulating the environment of an embryo, we can evaluate those particular stimuli as potential forces for shaping the evolution of that developmental process. We know, for example, that there are particular factors that can accelerate or retard the development of an organism. These environmental changes can therefore induce an extrinsic rate of change on the intrinsic rate of development, which is presumably set by the genes.
We know from humans that the environmental conditions in utero can have a dramatic effect on the final phenotype of the newborn. This has been termed fetal origins of adult disease or adult physiology. Whether this occurs in any other species, we don't know. Does the developmental milieu of a vertebrate embryo affect the outcome of the adult? What are the long-term effects, for example, of developing in very harsh or unpredictable environments?
Finally, if we alter endogenous controllers such as neurohormonal factors, we can induce an uncoupling of one developing system from another, or at least from the inherent time clock which is used to synchronise all the developing processes. This can help us to determine which particular factors are important in which processes.
Click on image for a larger version of figure 13
So we come back to the surfactant system. Figure 13 shows a particularly sensitive indicator of surfactant maturation: the cholesterol to phospholipid ratio in the lavage fluid. We show it for a range of species, again over the final 30 per cent of gestation or incubation, but we also show the cholesterol to phospholipid ratio for a number of adult species here.
We can see this is a good factor to analyse, because in mature animals – that is, those at 100 per cent – and adults, the ratio is kept within fairly tight limits. But each different species gets there in a very different manner. So if we look at this more closely, and look first at the chicken [slope] here and the oviparous bearded dragon [slope] here, we can see that they have very similar slopes. Again we need to be comparing this slope here [between 75 and 85%], because the chicken begins airbreathing at this point here [85%]. So these slopes are very similar.
If we then look at the oviparous bearded dragon – that is an egg-laying lizard – compared with a viviparous sleepy lizard, that is a live-bearer, we can see that they have a dramatically different pattern. These are the two most closely related species, yet they have the most different pattern. In fact, the sleepy lizards have a very similar pattern to humans. It is also similar to the green sea turtle.
Obviously, phylogeny is not determining the developmental process here, because the most closely related species have the most different pattern, and vice versa. So one of the factors that we thought we might investigate as a potential candidate for explaining these differences in developmental process is hypoxia, because live-bearing animals, in particular, experience quite significant hypoxia late during gestation as the demand for oxygen by the foetus outweighs the supply by the placenta. Similarly, green sea turtles lay large clutches of eggs deep in wet sand, and they experience quite significant hypoxia, as low as 13 per cent oxygen. Yet when we actually did the experiments, we did this all at 21 per cent.
Click on image for a larger version of figure 14
So we have looked at the effect of hypoxia – relatively mild hypoxia of 17 per cent – on developing chicken embryos (figure 14). Here we look at a different factor. This [y-axis] is the disaturated phospholipid to total phospholipid ratio, which is the saturation index of the surfactant – which is a sensitive indicator of maturation. We can see that, under normoxic conditions, day 16 chicken embryos have virtually no disaturated phospholipid. This then increases dramatically to about 40 per cent at day 21, when the chickens hatch.
However, under hypoxia, day 16 embryos have a dramatically elevated disaturated phospholipid ratio. This then increases gradually to the same levels at day 21. So this exogenous factor, hypoxia, has greatly accelerated the surfactant maturation rate. Some of these organisms even hatched a day earlier.
Click on image for a larger version of figure 15
What could be the intrinsic controllers that are mediating this effect? We injected various drugs into chicken embryos at 48 and 24 hours before a sampling time point. So if we sampled at day 16 we injected at days 14 and 15. Control animals were injected with saline; we injected dexamethasone, which is a synthetic glucocorticoid, T3, which is a thyroid hormone, or a combination of the two. We see that in our day 16 control animals there is virtually no disaturated phospholipid, but once the animals have been treated, before day 16, they have a dramatically increased amount of disaturated phospholipid, almost identical to that normally seen only in day 20 animals (figure 15). But, importantly, when they are injected before day 18 or before day 20, we see no effect; no difference between the control and the treated animals. So dexamethasone is operating only at a very discrete window of opportunity during the ontogeny of surfactant development. It can only induce developmental plasticity at this time point.
We have looked at the effect of hypoxia on thyroid hormones and we know that hypoxia does not stimulate plasma T3. In mammals, hypoxia stimulates glucocorticoids, and in chickens glucocorticoids peak immediately before hatching. So we strongly hypothesise that in chickens the effect of hypoxia is mediated via the glucocorticoids. Hypoxia is the extrinsic controller, glucocorticoids the intrinsic controller, and they interact to uncouple surfactant development from the remainder of the development of the lung.
What the underlying molecular mechanisms of this interaction are, we don't know. So this is where future questions would be coming from. Looking at the molecular mechanisms: that is using the information flow, from the genotype through to the phenotype. So we would be particularly interested in whether the expression of any genes is induced by these endogenous and exogenous cues; if so, which genes are involved; and, in particular, how the temporal inducibility is regulated – that is, how these windows of opportunity are regulated.
In conclusion: we have shown that the surfactant system is a good model. It is highly conserved, in terms of morphology and biochemistry and also development, in that they all have to overcome the same developmental hurdles, but the timing of the development is very different between the species. And this is heterochrony.
We have also shown that extrinsic forces such as hypoxia can alter the timing within a particular species – this is heterokairy – and this is presumably regulated through very specific intrinsic controllers.
Click on image for a larger version of figure 16
This leads us to our working hypothesis, which is that the developmental plasticity, which is induced by environmental factors, can be exploited or acted upon by classical Darwinian selection, or natural selection, in order to select a particular phenotype (figure 16). So you can imagine that if an animal is growing up in chronically hypoxic environments such as at altitude, the early hatching or the early maturation of the surfactant system may be of specific selective advantage. That phenotype is selected, eventually leading to an alteration in allele or gene frequency. That is a shifting in the window of opportunity, and this will lead to heterochrony.
Finally, I would like to acknowledge my colleagues Chris Daniels and also Warren Burggren, and our students Lucy Sullivan, Sonya Johnston and Helen Blacker, and the ARC for funding.
