Building a continuous and dynamic but neglected cell compartment: axonal endoplasmic reticulum

Animal and human movement depends on the ability of nerve cells to carry signals along narrow projections known as axons, which in humans can extend as far as a metre from the centre of the cell, the cell body. Maintaining long axons to ensure good communication requires a lot of engineering. This is reflected in problems that occur when it goes wrong - conditions such as axon degeneration, paralysis, or lack of sensation. Some of these, like Hereditary Spastic Paraplegia (HSP), which causes selective paralysis of the lower body, preferentially affect the axons furthest from the cell body, and therefore may affect processes like communication or transport that long axons are most vulnerable to.

One structure that may be vulnerable to diseases that affect axons are tubules known as smooth endoplasmic reticulum (ER), which run lengthwise through the axon, fusing and splitting from each other to form a network. Due to their length and continuity, and their potential to carry signals for long distances, they have been termed a "neuron within a neuron". In support of an important role for this network, HSP is often caused by mutations in proteins that help model ER, by inserting in one face of the ER membrane, and curving it.

Axonal ER is an underexplored compartment that barely featured in the scientific literature for 2-3 decades. It is only recently that we have developed tools to visualize it and see defects in it; in this way, we have found that removing some HSP membrane-curving proteins, using fruitfly mutants, causes moderate disruption of the ER network in motor axons.

Even when we remove the best known curvature-inducing proteins, the network is still substantially intact, with most axons still possessing at least one tubule. We hypothesise the existence of multiple mechanisms to achieve a continuous network of at least one tubule, and avoid too many tubules. Tubules must be shaped; tubule growth must require new membrane synthesis at the right location; since unattached ER tubules move up and down axons, there must be mechanisms to transport them, and possibly sense where they are needed.

To detect mechanisms that are important to make the axonal ER network and help it respond to the needs of the cell, we will use fruitfly genetics. Fruitfly neurons function very similarly to our own. Their short life cycle, and ease of handling in large numbers, makes them amenable to genetics, by removing or altering genes and studying the consequences, and thus learning how the affected processes work.

FIRST, since there are additional HSP genes that we hypothesise are important for shaping axonal ER, we will make flies that lack these genes, and test whether axonal ER is affected, and how.

However, removing only known genes will not identify new mechanisms that must exist. Large-scale random mutagenesis and screening for defects is a proven approach for this, having led to several Nobel prizes for biological processes such as embryo development, or biological clocks. Therefore SECOND, we will generate enough random mutations to disrupt most genes in one part of the genome. Visualising ER with a fluorescent marker, in axons that are visible through the insect cuticle, is rapid enough to allow screening of new mutant flies for defects in axonal ER, and thus find most of the genes involved in its organisation.

Once we have established stable mutant lines and confirmed their phenotypes, we will use whole-genome sequencing to identify the affected genes in several new mutants. This information will tell us what protein is affected in each mutant, and help us to predict its role. Finally, we will test some of these predictions by in depth analysis of how the affected proteins behave in normal axons, and how ER behaves in the mutants that appear most promising.

By studying a few genes with new phenotypes in depth, we will gradually build up a picture of how axonal ER is formed, regulated and responds to needs.

Endoplasmic reticulum (ER) forms a dynamic network of sheets and tubules. It is central to Ca2+ and lipid homeostasis, and forms close contacts with other organelles. In neurons, it appears continuous through dendrites, cell body and axon, and is therefore termed a "neuron within a neuron" - potentially carrying local, regional, or long-range signals, independent of action potentials or physical transport, up to 1m in humans.

The ubiquity and continuity of axon ER implies important roles for it, supported by the finding that several axon degeneration genes encode ER modelling proteins. In Drosophila we have developed tools to visualise axon ER, and shown that some ER-modelling proteins can affect the amount, continuity, or tubule dimensions of axon ER. Therefore our tools can help reveal how this underexplored compartment is assembled and regulated.

We still understand little of how the axon ER network is assembled and regulated and responds to cell needs. While we see clear mutant phenotypes affecting axon ER, all phenotypes so far retain ER tubules in most parts of axons, implying that we have not yet fully disrupted the processes that localise it and maintain its continuity and local density.

We will therefore screen for additional mutant phenotypes that affect axon ER, to understand their cellular and molecular basis, and the transport, synthesis, modelling, or regulatory processes that they reveal. We will first test fly homologs of axon degeneration genes that may model ER. We will next initiate a forward genetic screen for new mutations that cause ER defects, using MARCM to generate axons with labelled ER that are visible without dissection, and homozygous for new mutations on one chromosome arm. We will also screen for mutations in genes with some functional redundancy. We will identify affected genes in several new mutants by whole-genome sequencing, and select a small number for phenotypic study, to build a picture of the affected processes.

While this work is basic in nature and is not primarily aimed at understanding specific disease mechanisms, basic biology of both ER and axons fits squarely within the BBSRC strategic priority of "Bioscience for Health". Axon dysfunction is central to some of the most common conditions of ageing. In Alzheimer's disease, axon pathology is an early symptom. In diabetes, it is a major component of morbidity - approximately 60%-70% of patients have neuropathy that can promote injuries that require amputation, and >50% of lower limb amputations are in diabetics (www.cdc.gov). Basic understanding of the axon biology that underpins degeneration is an area of great need and promise, for impacts on understanding and amelioration of disease mechanisms.

Beneficiaries outside the immediate academic field include those with an interest in human axon degeneration, where basic Drosophila biology can generate hypotheses for human disease mechanisms. For example, we were the first to show (Wang et al 2007, highlighted in the BBSRC 2007 Annual Report) that a HSP gene had a role in BMP signaling, and susbequent work (e.g. Tsang et al 2009) showed roles for other HSP genes in BMP signaling. The ubiquity of BMP signaling means that great care is needed before using it as a therapeutic target, and good mouse models of most HSP diseases are still lacking, but as these models emerge, BMP signaling may deserve exploration as a therapeutic target, e.g. via localised delivery of agonists or antagonists. A better understanding of local calcium signaling mechanisms in axons, or other consequences of ER defects, could also lead to plausible therapies to ameliorate dysregulation. Development of plausible therapies would ultimately benefit both patients and pharmaceutical businesses - although this would require more clinically based research for several years beyond the immediate work proposed here.

Basic research on Drosophila axon maintenance replaces the need for much experimentation on protected animal species. Early characterisation of Wallerian degeneration was performed in mice (much of it in a BBSRC Institute, by Dr. Michael Coleman in Babraham), but Drosophila screens have enabled identification of several new genes that protect against degeneration, at least one a druggable NADase, without large numbers of mice. If the equivalent of our unbiased forward screen of up to 2000 flies were to be undertaken in mice, it would require establishment and screening of 2000 mouse families, many of them producing pups with homozygous mutant defects, an enormous impact on animal welfare, as well as prohibitively expensive.

The University has an active public engagement programme (www.cam.ac.uk/public-engagement) that includes publication of a research magazine, public events, open lectures, and the annual Cambridge Science Festival. CJO'K contributes through regular schools talks, occasional radio interviews, public panel discussions, and talks to encourage minority participation; a recent talk to a lay audience on the challenges faced by long axons is available online (www.youtube.com/watch?v=Kno1GvpjmSk). To foster public awareness of basic disease research with whole animals, we will contribute a demo with live flies and their use for "Bioscience in Health" at the annual Cambridge Science Festival. When appropriate, we will communicate via the University press office, using traditional and new media. Research findings of high public impact will be disseminated via written and broadcast media.

While our work is not specifically focused on disease mechanisms, an important audience is patients suffering from neurodegenerative diseases, and their families. I have contacts with UK HSP and ALS charities through funding applications and reviewing, have hosted visits by their donors, and given a talk to the AGM of the UK HSP support group to emphasise the importance of basic science; I intend to continue this work.