ComBio special: Zebrafish and the presenilin genes

Associate Professor Michael Lardelli is senior lecturer in genetics at University of Adelaide's School of Molecular and Biomedical Science, whose research group is trying to unravel the molecular pathology of Alzheimer's disease using zebrafish as a genetic model, focusing on the central role of the presenilin genes.

Zebrafish embryos are increasingly popular as a model system for studying gene function as they are vertebrates, small and easy to breed and handle, and are numerous and transparent. In addition, their organ systems are very similar to those of humans and zebrafish mutants provide excellent models of human development and disease.

Using these stripy aquaria critters, Lardelli and colleagues have uncovered a genetic mechanism that may provide new insights into the most common, sporadic form of Alzheimer's disease, and potentially, other disease such as cancer.

Lardelli started his molecular studies in Drosophila and then moved to mice, before shifting his focus to zebrafish as a vertebrate model about 10 years ago. Growing up in Canberra, Lardelli did his undergraduate degree in science at Sydney University before embarking on a research career in the laboratory of David Ish-Horowicz in what used to be the Imperial Cancer Research Fund at Oxford. There he completed a PhD in Drosophila development before moving on to a postdoc at the Karolinska Institute in Stockholm with Urban Lendahl.

Together, Landelli and Lendahl discovered some of the vertebrate Notch genes and are best known for their work on Notch 3. After 3.5 years, Lardelli moved to Uppsala University, where he first used the zebrafish model.

"I set up my own little system there in 1994 with the aim of cloning and analysing zebrafish Notch genes," Lardelli says. "On moving back to Australia a few years later, I set up the zebrafish facility at Adelaide Uni, which was only the second in Australia at the time (after one at the Ludwig Institute for Cancer Research in Melbourne). Now, this area is just exploding."

In 2000, Lardelli secured NHMRC funding to look at presenilin genes in zebrafish in the context of his continuing interest in Notch signalling, after a link was established between the two in C. elegans. Presenilin genes came to prominence in the mid 90s in association with familial cases of Alzheimer's, (FAD), the early-onset, inherited form.

"The PS gene encodes a protein that sits in the lipid bilayer and cuts the transmembrane domains of other proteins, such as Notch 1, to release a signal-active cytoplasmic domain," Lardelli says. "We know now that presenilins cleave a whole bunch of proteins in cell membranes, and that this is very important for many cell signalling pathways."

This list includes the amyloid precursor protein (APP), the cleavage of which yields amyloid peptides that deposits in the brain as plaques in Alzheimer's.

"Presenilins also seem to be multifunctional - for example, they are involved in the phosphorylation of signalling proteins like beta-catenin and are also found in the nuclear membrane. We now know that these are very important proteins in the cell performing many tasks."

Familial and sporadic

When Lardelli started this research project, presenilin genes were also the major locus for mutations involved in FAD. In fact, more than 50 per cent of the characterised mutations associated with FAD are in the presenilin genes (latest count at 160 in presenilin 1 alone).

"We know a lot about the mechanisms underlying FAD, because we can do genetics on it, and we know many of the genetic players involved. However, little is known about the non-familial or sporadic form of AD (SAD).

This is the late-onset form and accounts for about 95 per cent of all cases. All we really know about SAD in a general sense is that the disease 'looks' the same as FAD - we get the amyloid plaques, and we get the neurofibrilliary tangles (same pathology and cell biology), but why and how is unknown.

"We don't have a real lead to the molecular mechanisms underlying the sporadic disease."

With this background, recent data from a number of laboratories pointed to decreased presenilin activity playing a central, causative (rather than facilitating) role in FAD.

One of the mysteries about the many presenilin mutations is that they are all dominant, so it only needs one mutated copy of the gene to get the disease. How this myriad of all-dominant mutations actually have their effect in AD remains a subject for debate.

"Most of the mutations in presenilin that are associated with FAD are point or missense mutations, resulting in an amino acid change," he says.

"There are very few mutations leading to a truncated protein, which led to the belief that truncations in presenilin generate a dead protein or null protein, while missense mutations create forms of presenilin that dominantly interfere with or inhibit normal presenilin function."

The cleavage of APP by presenilin produces amyloid beta-peptides of different lengths, and it appears that the ratio of the longer to the shorter forms (usually called the 42:40 ratio) is critical to normal function. If the ratio increases past a threshold, amyloid might start to precipitate in the brain.

"It is now thought that the plaques may not actually be the damaging agents. Instead, it could be the oligomers that form before the plaques - so the amyloid molecules start to aggregate and form these soluble oligomers that turn out to be quite toxic. Thus, the plaques actually may be protective by bringing the toxic oligomers out of solution."

Near the beginning of this year, a paper came out that generated new ideas to explain how all of the presenilin mutations can have a dominant effect. Presenilin mutations can shift the cleavage site slightly, thus slowing the rate of cleavage and giving the longer forms more time to leave the membrane.

This leads to an increase in the 42:40 ratio. The longer forms are more aggregation-prone and therefore oligomerise faster. "According to the amyloid hypothesis, this is going to promote AD," he says.

Aberrant splicing

So, where does Lardelli's work come into this story? A group in Sydney had identified presenilin mutations that resulted in loss of the entire exon 8 sequence, but still produced FAD.

"Rather naively we tried to model this human mutation in zebrafish by injecting morpholino antisense oligonucleotides (morpholinos), which are extremely effective tools to block protein translation or interfere with splicing."

The results were unexpected and in fact quite cryptic, with the injected fish showing a very strong phenotype consistent with complete loss of presenilin activity.

It turned out that the genetic changes being introduced produced low levels of aberrant splicing, potentially forming truncated presenilin proteins and a consequent dominant-negative effect on activity.

Many zebrafish and human hairs later, Lardelli confirmed that truncating the presenilin within a certain region produced a potent dominant-negative protein with no obvious effect on the levels of normal presenilin protein (which is very tightly regulated), yet with almost complete loss of presenilin function.

In addition, mutations causing such truncations would be dominant lethal in embryo development and therefore not detected previously in mouse or other models.

Lardelli's results are highly significant for two reasons. Firstly they show that truncations in presenilin genes do not necessarily produce null protein; instead, very strong dominant-negative phenotypes are a possible outcome. Secondly, the data highlighted the potentially very small amounts of this truncated protein needed to decrease presenilin activity.

"Everyone has assumed truncated presenilin to be inactive, and what we are saying is wait a minute, these are incredibly dominant negative if they are truncated in the right region.

Thus, any mechanism in a brain cell such as ageing or oxidative stress that could interfere with splicing, or in some other way truncate presenilin protein, would potentially reduce presenilin activity and lead to AD, in effect acting like a familial mutation. Therefore could a molecular common link between FAD and SAD be presenilin?"

Presenilins and SAD

Lardelli says it would be very satisfying if presenilins were also important for sporadic AD: "it would all make sense," he says. "We didn't really set out to explain sporadic Alzheimer's; we really just wanted to find out about presenilins in our zebrafish models."

The researchers still have much to do to prove their idea, such as identifying in vivo mechanisms. The immediate next step for the group is to demonstrate the dominant-negative effect in human cells, and Lardelli is pursuing this currently in collaboration with Ralph Martin and Giuseppe Verdile in Western Australia.

The other exciting thing about this mechanism is that decreased presenilin activity potentially affects other disease due to the myriad of roles played by this ubiquitously expressed and multi-talented protein.

"For example, aberrant splicing can occur in cancer so here is a mechanism that could wipe out presenilin activity in cells during cancer, and presenilin was recently identified as a tumour suppressor in skin."