WASHINGTON – "There is nothing more beautiful than 20/20 hindsight," Huda Zoghbi told the audience at the 2017 Annual Meeting of the Society for Neuroscience.
In the case of Zoghbi, who is a professor in the Departments of Pediatrics, Molecular and Human Genetics, Neurology and Neuroscience at Baylor College of Medicine, an investigator at the Howard Hughes Medical Institute, and the director of the Jan and Dan Duncan Neurological Research Institute at Texas Children's Hospital, that hindsight spans a career that has included research on neurodevelopment and neurodegeneration.
Her research has led her to conclude that on both ends of that spectrum, small changes in protein levels can have big effects.
In her talk, titled "On balance: fine-tuning protein levels for neurological health," Zoghbi told the audience that insight "crystallized as the years went by."
The initial crystal was deposited during Zoghbi's work on Rett syndrome, a neurodevelopmental disorder that is due to mutations in the protein MeCP2. Since the protein is on the X chromosome, the disorder is usually fatal in boys. Affected girls have one normal and one dysfunctional copy of MeCP2 and express the normal MeCP2 on average in half their cells. They survive, but can have a wide range of symptoms from motor problems to autism-like behaviors.
Some girls have comparatively mild disease because by roll of the genetic inactivation dice, they express their normal MeCP2 in a high proportion of cells.
But there are also both girls and boys who have mild symptoms because their mutations are "hypomorphic," that is, they express lower levels of MeCP2, but the protein is still present.
The relationship between protein levels and symptoms is complex. But overall, there is a reproducible constellation of symptoms that tends to go with different protein levels, and relatively small changes in protein levels can affect which constellation a patient will have.
The idea was reinforced when Zoghbi's lab began work on spinocerebellar ataxia 1 (SCA-1), a dominantly inherited balance disorder that ultimately kills patients due to loss of brainstem neurons.
SCA-1 is a polyglutamine disorder, caused by an expanded series of glutamine repeats in the protein ataxin-1. It does not, however, lead to any fundamental changes in the protein's function.
"We never found an interaction that is exclusive just to the mutant or just to the wild-type; what we found is that some interactions are enhanced," she said.
"In SCA-1, lowering protein [levels] by 20 percent is enough to rescue many phenotypes."
From thesis to therapy
Zoghbi is now testing the idea that subtle increases in some proteins can cause late-onset neurodegeneration in more common neurodegenerative disorders. If the idea turns out to be correct, targeting regulatory proteins might be a way to head neurodegeneration off at the pass.
"That modest changes make a big difference provides therapeutic opportunity," Zoghbi said.
Modest changes are easier to achieve in the first place, and may also be more sustainable.
A basic biological organizing principle from cells to organisms is that they are homeostatic – when their status quo is perturbed, it oftentimes sets off compensatory mechanisms to get back to the way things were in the good old days.
As a result, losing weight is perhaps not easy, but not nearly as hard as keeping it off. Blocking tumor cell proliferation by shutting down a specific protein is possible for a time, but not for a long time.
Modifying protein levels slightly, rather than radically, may help avoid setting off such compensatory mechanisms.
Slight modifications are also preferable from a safety standpoint, which will be critical if the current attempts to treat neurodegenerative disorders earlier are to be successful.
"For neurodegenerative diseases, you need to think about treating people before they are symptomatic, and that means maybe 40 or 50 years," Zoghbi said.
That time frame means that long-term treatment needs to be feasible from a toxicity standpoint, and that treating patients who are clinically asymptomatic needs to be safe enough to justify the risk-benefit profile.
But large effects stemming from small changes can also work against therapeutics development in the form of a narrow therapeutic window.
The MeCP2 gene, where both overexpression and loss of function cause overlapping neurodevelopmental syndromes, is an excellent example of that issue. Effective therapies will need to be carefully titrated.
Zoghbi's work with a protein called Atonal also showed the pitfalls of trying to find subtle changes.
Atonal is expressed in the inner ear early during development. To study basic mechanisms of development, Zoghbi and her team generated mice lacking one copy of the Atonal gene, Atoh1.
When they ran the animals through a standard battery of tests, they discovered no gross behavioral abnormalities. "The mice looked good," she said.
It was only during later studies that the team unexpectedly found, in the course of other experiments, that mice with only one copy of Atoh1 suffered age-related hearing loss.
Zoghbi stressed that low levels of Atoh1 causes cochlear hair loss well after the protein is no longer expressed at all.
"Atonal is expressed in cochlear hair cells only early on. At the time you see this degeneration, Atonal has been turned off for weeks," she said.
Her team's accidental finding underscores how easy it is to miss the effects of changes in protein levels. "For some genes that we think about as developmental . . . we may not think about haploinsufficiency as being important," she said. But her lab's work on Atonal demonstrates that "we need to more closely evaluate animal models [that we have] signed off as normal."