By showing that knockout of the Drosophila homologue of parkin produces a substantial phenotype, Greene et al. provide an important in-vivo model for this genetic form of PD. The phenotype is clearly related to mitochondrial triggering of apoptosis, echoing results in vitro (Tanaka et al., 2001 and Darios et al., 2003). This reinforces the longstanding idea that there is a link between mitochondrial function and sporadic PD. The consistency between such different models supports a contention that Greene et al. eloquently make, that parkin’s function is conserved between divergent species, and some of the mechanisms that lead to cell death in PD may also be well enough conserved to be amenable to analysis in invertebrate systems such as Drosophila or C. elegans.

Although this is a landmark paper and a useful tool, the Drosophila parkin knockout is an imperfect model for PD, as the tissue types affected in the fly...  Read more

By showing that knockout of the Drosophila homologue of parkin produces a substantial phenotype, Greene et al. provide an important in-vivo model for this genetic form of PD. The phenotype is clearly related to mitochondrial triggering of apoptosis, echoing results in vitro (Tanaka et al., 2001 and Darios et al., 2003). This reinforces the longstanding idea that there is a link between mitochondrial function and sporadic PD. The consistency between such different models supports a contention that Greene et al. eloquently make, that parkin’s function is conserved between divergent species, and some of the mechanisms that lead to cell death in PD may also be well enough conserved to be amenable to analysis in invertebrate systems such as Drosophila or C. elegans.

Although this is a landmark paper and a useful tool, the Drosophila parkin knockout is an imperfect model for PD, as the tissue types affected in the fly are muscle and testes rather than the neurons that degenerate in PD. Why is this? It is too early to make definitive statements, but a problem that occupies much of my time is how close a homologue does a homologue have to be before we can make functional predictions? Evolution has co-opted proteins for different functions that retain detectable sequence identity. And the converse is also true, that proteins with different sequences may have overlapping functions. I think this problem gets worse when we consider protein families. The different families of E3 protein-ubiquitin ligases (reviewed by Jackson et al., 2000) have the same function at one level, namely, to add ubiquitin to proteins, but at another level they are functionally distinct. Different E3 and E3-like proteins vary in their domain architecture, they require different cofactors and, crucially, they may target different substrates, possibly with redundancy between different E3 ligases. Therefore, perhaps the knockout experiments reported here may mimic PD only partially because the nature of the substrates regulated varies between humans and flies. At least two putative substrates of parkin, α-synuclein and synphilin, do not have readily identifiable Drosophila homologues. Ubiquitination plays a role in other processes at synapses, and we do not yet know what role parkin might play in this.

If the function of parkin-mediated ubiquitination in either signaling events or protein turnover varies among species, then the attempt to model the human condition in lower species by this approach may only ever be partially successful.

View all comments by Mark Cookson

While overt efforts to create models of human disease in lower systems represent a relatively new area of investigation, in fact, this work has been going on in another guise for a very long time. Much of this work can be described as genetic analysis, or the use of mutations in model organisms to investigate basic biological processes. In the course of these studies, a number of genes were characterized which later proved to exhibit significant sequence similarity to genes responsible for heritable human disease. What has this work told us? Perhaps the most important lessons that have been learned from this work, and indeed the most important biological insights that have been gleaned over the past 30 years, are that protein sequence conservation is strongly correlated with protein function, and that protein function and biochemical pathways are highly conserved across evolution. Indeed, much of our current knowledge of vertebrate molecular biology owes its origin to work carried out in simple model organisms.

Does this mean that we should expect models of disease in lower...  Read more

While overt efforts to create models of human disease in lower systems represent a relatively new area of investigation, in fact, this work has been going on in another guise for a very long time. Much of this work can be described as genetic analysis, or the use of mutations in model organisms to investigate basic biological processes. In the course of these studies, a number of genes were characterized which later proved to exhibit significant sequence similarity to genes responsible for heritable human disease. What has this work told us? Perhaps the most important lessons that have been learned from this work, and indeed the most important biological insights that have been gleaned over the past 30 years, are that protein sequence conservation is strongly correlated with protein function, and that protein function and biochemical pathways are highly conserved across evolution. Indeed, much of our current knowledge of vertebrate molecular biology owes its origin to work carried out in simple model organisms.

Does this mean that we should expect models of disease in lower systems to recapitulate human disease characteristics closely? In many cases, probably not. For example, genetic analysis has revealed many examples of developmental signaling pathways that are highly conserved in vertebrates and invertebrates. However, these signaling pathways are often used in different contexts in different species, such that the phenotypes resulting from mutations in orthologous genes in different organisms can differ substantially from one another, despite the underlying molecular conservation. There is also an increasing number of examples of human disease models involving lower species (e.g., the Drosophila models of long QT and Fragile X syndromes) that appear at first glance to share little in common with the corresponding human diseases, but have contributed much to our understanding of these diseases. These and many other lessons tell us that cellular phenotypes and molecular mechanisms tend to be conserved, whereas tissue specificities, and morphological and behavioral phenotypes are less conserved.

Only time will tell how much Drosophila parkin mutants will contribute to our understanding of dopaminergic neuron loss in humans. Meanwhile, it is imperative that we apply realistic expectations to studies of disease mechanisms in simple model organisms. Failure to do so will encourage efforts to make these models into something they are not at the expense of extracting what they can indeed provide, and it will likely lead to premature termination of many promising avenues of research.

View all comments by Leo Pallanck