Why Doesn’t Everyone Get Alzheimer’s?

by Guy McKhann, M.D.

July 30, 2012

Hardly a week goes by without an article about Alzheimer’s disease. Many don’t make sense or are rehashes of what is already common knowledge. In the last few weeks, however, there have been worthwhile reports on two novel studies.

Why don’t some people get the disease?

The field of medicine, almost by definition, is focused on disease. For a given disease process, physicians ask about prevention, treatment, and the disease’s natural history. Rarely do we ask, “Why doesn’t everybody get the disease?”

There is no question that certain diseases run in families. This is true for diabetes, heart disease, multiple sclerosis, and Alzheimer’s disease. There are no clear-cut genetic mutations for these conditions, but having family members with the disease is clearly a risk factor. Is the reverse possible? Are there genetic factors that protect you from the disease?

That is exactly the question that Kari Stefansson and his colleagues in Iceland asked (see The New York Times story or find the study, published in Nature). This group works through DeCODE Genetics, a company which scanned the entire DNA of 1,795 Icelanders. This population is a relatively homogeneous group, with good family records, and is geographically stable. In other words, from a genetic and epidemiologic point of view, it is ideal for study.

Dr. Stefansson and his colleagues focused on the elderly, those 75 and older. They divided the Icelanders into two groups: those with a family history or symptoms of Alzheimer’s disease and those without either. They performed a genetic screening, looking for genes that occurred in the normal group. They found a mutation in a particular gene significantly over-represented in this group, suggesting that this mutation might be involved in protection from Alzheimer’s. They went on to do two other clinical studies.

In the first they looked at 85-year-olds who were cognitively intact, and found the mutation over-represented in this group. In a second study, they did a series of cognitive tests in those 80 to 100 with and without the protective mutation. Those with the mutation performed better on cognitive tests far longer than the others. Finally the researchers carried out studies with cells in culture, comparing those with and without the mutation. Those with the mutation produced lower levels of amyloid beta, a component of the plaques found in Alzheimer’s disease.

This report suggests that there are genetic mutations which protect against Alzheimer’s disease, and that in this case they work by diminishing the production of amyloid beta. As I have written about in previous columns, the “amyloid hypothesis” proposes an overproduction of amyloid, particularly of a fragment of the amyloid protein, Abeta42, which is thought to be toxic to nerve cells. Thus if you could either slow down the production of Abeta42 or remove it from the brain there would be a treatment for the disease. Drug companies have spent millions of dollars attempting to do just that, so far with no success.

The DeCODE study suggests that the basic idea of slowing production of amyloid may be correct. However, those with the protective mutation may have had lower amyloid levels all their lives. A treatment program of only a few months may not be sufficient. So when should one start such a program?

How early can the disease be detected?

In another study (published in the New England Journal of Medicine) mentioned in the same New York Times article, this question is asked of another unique population. Most of the people with Alzheimer’s disease may have a family history, but no specifically identified genetic abnormality. But a small population (about 1 percent) has a dominantly inherited form. In this situation, if you have mutations in one of three identified genes, you will get the disease. Clinically, the dominant form of the disease is somewhat different than the much more common sporadic form: It starts at a younger age and is likely to have a more rapid course.

An international consortium to study these dominantly inherited subjects, known as the Dominantly Inherited Alzheimer Network (DIAN), has provided interesting data on early detection. Based on the assumption that the disease will breed true within a family—that is, you will get the disease at about the same age as your parent—one can ask how soon before your expected disease recognition are you starting to show either laboratory or clinical evidence of the disease?

In 128 subjects, DIAN looked at known biomarkers for Alzheimer’s disease, involving cerebrospinal fluid (CSF), imaging of brain volume, and accumulation of amyloid in brain, as well as changes in cognitive function. Their basic strategy was to determine the subject’s age when a marker of disease first appeared, and then to subtract that age from the expected age of disease onset in their parent.

There is a progression of the appearance of these biomarkers: (1) abnormalities of Abeta in CSF appear first, up to 25 years before expected age of disease; (2) amyloid deposition, as measured by PIB imaging, appears later, about 15 years before, as do abnormalities of tau (the other protein involved in the mechanism) and changes in brain volume; (3) impairment of memory starts 10 years before disease recognition and more global cognitive impairment starts five years prior.

This study reinforces the idea that to treat this disease we must start earlier, and probably treat for a long time. We do this now for those with higher levels of cholesterol to prevent subsequent heart disease. Hopefully we can develop the same strategy for Alzheimer’s disease.