Re-examining the nature of fragile X

In the wake of negative results from several high-profile clinical trials in fragile X, we find ourselves questioning all of our previous assumptions about the nature of this disorder. After all, understanding the basic pathology of the disease is critical to development of new treatments---this is true across the board, in all branches of medicine. In the early days, shortly after the FMR1 gene was discovered and the normal protein product of the gene (FMRP) was identified, it was noted that FMRP was an RNA binding protein. So, whatever the normal function of the missing protein was, it seemed to have something to do with RNA metabolism. Since RNA is the template that is used to make new proteins, this meant that the fragile X protein was involved in regulating protein synthesis. A bit more work over the next few years led to the “mGluR Theory” of fragile X, the idea the FMRP normally regulates protein synthesis in dendrites in response to synaptic activity, and that in the absence of FMRP (in fragile X) there was abnormal protein synthesis, leading to excessive activity in some signaling pathways.

While this theory was extensively validated, pointing to the potential therapeutic value of mGluR5 antagonists as treatments for fragile X, there was always the possibility that this represented only a piece of the puzzle. After all, most biomedical research is done in fairly narrowly-focused model systems; the whole point of using models (animals, cells, etc.) is to simplify the incredibly and incomprehensibly complex human brain to the point where experimentation is feasible. Thousands of experiments by hundreds of scientists around the world confirmed that absence of FMRP led to consistent signaling abnormalities, and that these could be normalized by decreasing the function of metabotropic glutamate receptors, either genetically or pharmacologically. So, the mGluR Theory isn’t necessarily wrong, but clearly there’s more to the story.

Keep in mind that all of the above takes place in the postsynaptic compartment, otherwise known as the dendrite, or the dendritic spine to be more specific. However, it has been known all along that FMRP is normally present in the presynaptic compartment, otherwise known as the axon, as well as in surrounding glial cells like astrocytes (these are the non-neurons in the brain that perform many of the household chores.) So, the question has always been, what’s FMRP doing in all those other places? There’s no protein synthesis in axons, and there’s no synaptic activity in glia, so FMRP must be doing more than just regulating activity-dependent protein synthesis.

In the years since the original conception of the mGluR Theory, some valuable clues have come along. For example, FRAXA scientist Dr. Len Kaczmarek first found that FMRP interacts directly with certain ion channels to regulate neuronal excitability; specifically, he reported that one section of FMRP (not the RNA binding part) interacts with a large potassium channel called Slack to fine-tune the fidelity of auditory circuits. More recently, another FRAXA researcher, Dr. Vitaly Klyachko found that FMRP interacts directly with the presynaptic BK channel---literally the Big K (potassium) channel---to regulate neuronal excitability, and that absence of FMRP results directly in hyperexcitability of this system. In collaboration with Dr. Steve Warren, they examined samples from a patient with a rare mutation of FMR1. This patient has a point mutation in the gene, which results in a single amino acid change in the protein, which is produced in normal amounts. The patient has what is described as a “partial fragile X phenotype” of intellectual disability and seizure disorder, but without the physical features or typical behaviors of fragile X syndrome. Upon examining the protein produced from this mutation, Dr. Klyachko found that it had entirely normal RNA binding properties, resulting in normal activity-dependent protein synthesis and normal mGluR-LTD. The only problem caused by this single amino acid change was that this mutated FMRP couldn’t interact with BK channels, leading to increased excitability of the axon and greater neurotransmitter release. In other words, there is an entirely separate and very different presynaptic function for FMRP, not related to its RNA binding function in dendrites, and this represents another facet of fragile X, a distinct presynaptic phenotype unrelated to mGluR5.

It has long been known that FMRP is highly expressed in the non-neuronal cells of the brain (glia). Additionally, early experiments showed that normal neurons growing with fragile X glia demonstrated many of the same abnormalities we would expect to see in fragile X neurons. Conversely, growing fragile X neurons with normal glia reversed many of the usual structural abnormalities. Clearly, something is going on in fragile X glia, and that represents a potentially significant part of the disease mechanism. More recently, FRAXA researcher Dr. Yongjie Yang has shown that fragile X astrocytes (one main type of glial cell) have low levels of a key glutamate transporter, and this deficiency contributes to excessive glutamate levels in the synapse, resulting in hyperexcitability. This abnormality does not respond to mGluR5 antagonists; in fact, they may aggravate the problem. Thus, we have identified key glial phenotypes which appear to be independent of the other phenotypes described in neurons.

As we move forward, we may need to consider the contributions of these distinct presynaptic, postsynaptic, and glial phenotypes to the clinical presentation of fragile X. We may need to treat them separately for best overall effect, and the knowledge that different things are going on in different places in the brain may help us to develop combination drug strategies which can significantly alter the course of fragile X.