Essays on Science and SocietyGE PRIZE ESSAY

The Molecular Basis of Size Differences

See allHide authors and affiliations

Science  04 Dec 2009:
Vol. 326, Issue 5958, pp. 1360-1361
DOI: 10.1126/science.1184444

Size differences account for a great deal of the diversity found in the animal kingdom, but we still have much to learn about how sizes are programmed. Generally, the cells of different animals are comparable in size, and all animals begin as a single cell. This leaves the number of cells accumulated as the main determinant of animal size. We can reasonably expect the genes controlling cell number to be conserved among animals. So it seems that size-determining genes must be deployed in the elephant in such a way that it amasses several hundred thousand times more cells than the mouse. Which are these genes and how do they control size? I asked this question in a more experimentally tractable context: How do body parts of a single animal become different sizes? Fingers, toes, and ribs are sets of structures whose members are similar in form but differ in size. Although we know that Hox transcription factors specify the identity of individual fingers, toes, and ribs, little is known about how their individual sizes are programmed.

Fortunately, a similar situation exists in an experimental model system, the fruit fly, Drosophila melanogaster. Drosophila has two true wings that develop from imaginal discs of 50,000 cells and two smaller balancing organs called halteres that form from imaginal discs of 10,000 cells. We know that all differences between these appendages are specified by the expression of the Hox gene Ultrabithoax (Ubx) in the haltere and its absence from the wing (1). Thus I was able to ask how the expression of this single gene limits the size of the haltere.

My first insight came from removing Ubx function from random clusters of haltere cells (2). As expected, the resulting Ubx mosaic halteres are larger than wild-type halteres. Surprisingly, however, the overgrowth is not restricted to the Ubx mutant tissue but spreads to the wild-type tissue as well. This means that Ubx does not control haltere size cell-autonomously by, for example, acting directly on cell cycle or apoptotic checkpoints. Rather, I hypothesized that, because of their fundamental and non-autonomous role in the control of tissue growth and patterning, alterations in morphogen signaling might underlie changes in organ size. Indeed, transcription of decapentaplegic (dpp), a growth-promoting morphogen of the BMP family, is reduced in the haltere in comparison with the wing (2). Furthermore, the pattern of Dpp pathway activation is altered between the haltere and the wing, not just quantitatively but also qualitatively (see the figure). In the wing, Dpp travels far from its source to form a broad activity gradient. In the haltere, I found Dpp signaling to be largely restricted to the cells in which it is produced (2). Because receptor binding impedes morphogen mobility, I examined the expression pattern of the Dpp receptor thickveins (tkv) and found it to be strongly up-regulated in the haltere compared with the wing. I was able to show that the Dpp mobility restriction in the haltere is due in large part to transcriptional up-regulation of tkv. The alterations in transcription of dpp and tkv account for much of the reduced size of the haltere relative to the wing (2).

GE Healthcare and Science are pleased to present the prize-winning essay by Michael A. Crickmore, a regional winner from North America, who is the Grand Prize winner of the GE & Science Prize for Young Life Scientists.

Morphogen distribution in the haltere and wing-size regulation.

Ubx reduces the size of the haltere imaginal disc relative to that of the wing by decreasing the production and mobility of growth-promoting morphogens (for example, Dpp). Ubx impedes morphogen mobility in the haltere by up-regulating the receptor throughout the haltere and repressing an HSPG in the posterior half of the haltere. Reduced morphogen availability causes a reduction in haltere size relative to the wing.


The mechanism by which Ubx orchestrates these changes in the haltere is a telling example of selector gene function. Ubx converts one tkv repressor into a repressor of a second tkv repressor, thereby upregulating tkv levels in all haltere cells. The resulting elevated receptor density traps secreted Dpp at or near its site of production, leading to high Dpp signaling in the cells that transcribe dpp. This triggers a negative autoregulatory loop through which Dpp signaling represses dpp transcription, leading to a global reduction in the amount of Dpp in the tissue (2). This demonstrates how a fairly subtle modification of a regulatory network, effected by a selector gene, can set off a chain of events that has powerful ramifications for morphogen signaling and organ size.

Hox modulation of components of morphogen signaling pathways seems likely to define a general principle of size regulation (3). For example, the Wnt and Hedgehog morphogen pathways are also altered in the haltere relative to the wing (4, 5). The mobilities of BMPs, Wnts, Hedgehogs, and other morphogens are promoted by heparan sulfate proteoglycans (HSPGs) (6). In the haltere, I found that Ubx works with the posterior transcription factor engrailed to selectively repress the HSPG dally in the posterior compartment. As a consequence, morphogen mobility is severely impaired in the posterior side of the tissue, causing the posterior of the haltere to be smaller than the anterior (5) (see the figure).

As is clear from the examples above, it is critical that the production levels of size-regulating genes such as dpp, tkv, and dally be precisely controlled. In the haltere, the same is true of the regulator of the size regulators: Heterozygous Ubx mutant flies have enlarged halteres (1), and flies with extra copies of the Ubx locus have shrunken halteres (7). While studying how Ubx levels are controlled in wild-type halteres, I found that when Ubx protein levels rise, subsets of overlapping transcriptional input into the Ubx promoter become silenced (8). This silencing likely buffers against inappropriately high cellular Ubx concentrations and may stabilize Ubx levels against the unpredictability of genetic variation. I found that when Ubx enhancer-reporting lab strains are outcrossed to various wild-collected D. melanogaster strains, dramatic enhancer silencing is seen in a large fraction of the resulting progeny (8). Despite (or perhaps because of) the silencing of subsets of Ubx enhancers, Ubx protein levels remain normal in the outcrossed progeny, and their halteres develop perfectly. Only through the use of the sensitive transcriptional reporter lines available in the Ubx locus is the chaos beneath this calm surface revealed. It may be that if we look close enough, we will find similar systems operating to generate stereotyped expression levels of many concentration-dependent transcription factors.

In conclusion, my thesis work has shown that alterations in morphogen signaling landscapes underlie differences in tissue sizes. Transcription factors manipulate the extent and intensity of morphogen signaling through the control of morphogen production and mobility. At least some size-controlling genes autoregulate their production levels to ensure reliable size outcomes in the face of natural genetic variation. These findings help explain how size differences are generated and how the genes that orchestrate size changes are themselves regulated. But is the regulation of morphogen signaling capable of explaining size differences as great as those between elephants and mice? Or are the growth effectors downstream of morphogen signaling also differentially tuned to create specific sizes for specific contexts? And how exactly do changes in patterns of morphogen activity translate into changes in size? Sizeable questions, all.


Navigate This Article