Spotlight on Research: A Specific Cognitive Weakness Is Attributed to a Single Gene Deletion

Behavioral neuroscience has passed another milestone; for the first time, a specific weakness in cognitive ability has been linked to deletion of a single gene. APS Fellow Carolyn B. Mervis (Emory University) is part of a collaborative research group that made this breakthrough via recent studies of Williams syndrome.

The findings were based on National Institutes of Health supported research published in a paper in the July 12 issue of the journal Cell. The other principal investigators on the project include a molecular geneticist at the University of Utah, Mark Keating, and a medical geneticist at the University of Nevada School of Medicine, Colleen Morris.

Williams syndrome (WS) is a multifaceted genetic condition of varying severity occurring about once in 20,000 live births. The majority of these cases are the result of spontaneous genetic mutations, and are not inherited. (Few individuals with WS have children of their own, so even if there were a genetic transmission of the defect, the disorder would remain relatively rare.)

People with classic WS generally have mental retardation (the average IQ is 55 to 60, with a range of 35 to 105); an unusually outgoing, loquacious and friendly manner (sometimes too friendly to strangers); overly sensitive to other’s emotions; a great deal of anticipatory anxiety; a prominent attractive smile; characteristic facial features; small teeth; joint problems; and heart and circulatory symptoms. Associated supravalvar aortic stenosis (SV AS), a narrowing of the ascending aorta, may be mild, moderate, or life-threatening.

Until recently, the diagnosis of WS was usually made between 4 and 6 years of age, with SV AS or other heart problems often being one of the most important pieces of evidence. Now, a blood test for WS is available, and most newborns with SV AS are tested immediately. Earlier detection is also facilitated by the fact that most children with any developmental delay are tested by age 2.

Cognitive Profile

Individuals with classic WS have a unique pattern of cognitive strengths and weaknesses. They perform very poorly, relative to their overall IQ level, at copying patterns or constructing shapes in response to pictures (visuo-spatial construction, VSC), while their auditory short-term memory is typically above that predicted by IQ, and their language skills are in line with or slightly above IQ.

Mervis and her colleagues, Jacquelyn Bertrand, Byron Robinson, and Bonnie Klein, have studied the characteristic WS cognitive profile (WSCP), and described it in terms of subscores on the Differential Ability Scales (DAS), a standard test that assesses a wide range of intellectual abilities. The criteria for WSCP are a low absolute score on pattern construction, with pattern construction also low relative to auditory sbort-tenn memory, and low relative to the overall score. The sensitivity of this operational definition of the WSCP in classic WS is 0.94; the specificity is 0.92. That is, among people with the classic WS genetic condition, 94 percent of them test positive by these criteria; and in a group of people without WS but who have borderline normal intelligence, mental retardation stemming from any of various syndromes or from uncertain etiology, 92 percent do not exhibit the cognitive profile.

Homing in on It

Classic WS has recently been shown to be caused by hemizygosity for at least 500 kilobases of DNA on chromosome 7 (i.e., a deletion of genes on one of the two copies of chromosome 7). “Geneticists hypothesize that there are about ten genes missing,” reports Mervis. To elucidate the mechanisms of WS, the researchers have sought individuals with SV AS, but without all the other characteristics of WS, hoping to identify people with smaller deletions in this region of chromosome 7.

Indeed, researchers have identified individuals with hemizygosity limited to the gene for elastin, a structural protein found in large arteries and other elastic connective tissue, and have established that SV AS results from this deletion. Elastin is expressed at negligible levels in the brain, and these people have no cognitive or personality abnormalities. More recently, they identified two kindreds with SV AS combined with a history of academic problems, but with normal IQs, and without the personality characteristics of WS.

When Mervis and colleagues tested the affected members of these families, they fit the WS cognitive profile, visuo-spatial construction being their greatest cognitive weakness. “They do not show a dramatic loss in IQ if they are otherwise normal, because they have strategies to compensate. They lose a few points on IQ tests, but are not knocked out of the normal range,” explains Mervis. “The observed cognitive deficiency so far is quite specific for Visuo-Spatial construction. Even in classic WS with moderate retardation, the VS problem primarily involves construction.

Retarded people with WS can match patterns as well as anyone of their equivalent IQ, they just can’t construct them.” In contrast, high functioning (normal intelligence) people with classic WS, like people with partial WS, compensate for their deficiency in VSC through other cognitive skills. “It is not that they can’t do these VSC problems, but that they are very slow at them,” says Mervis.

The Fine Points

The genome of the affected individuals of one partial WS family has a deletion of just 85 kilobases of DNA on chromosome 7, and is hemizygous for just two genes. One is elastin, accounting for the SVAS. The other is a newly discovered gene of unknown function called LIM-kinase1 (LIMKI). The LIMK 1 gene product has bits of amino acid sequence that resemble sequences in three known classes of proteins. “LIM” domains are types of “zinc finger” domains that interact with DNA; other proteins of this group are known to be involved in gene regulation, particularly in the context of differentiation and the determination of cell fate during development.

However, unlike other LIM family zinc finger proteins, this one is also a kinase. That means it phosphorylates something, which means it probably regulates some cellular function in the short term. The presence of a “PEST” (proline, glutamate, serine, threonine) amino acid domain in LIMK1 also suggests that the level of LIMK1 protein itself is regulated on a short time scale.

The main hypothesis for the function of LIMK1 is that it is important for the formation of certain networks of neurons that carry out parallel processing involved in visuospatial construction. “However,” says Mervis, “it cannot be ruled out that LIMKI might also be involved in [neural] signal transmission.” LIMK1 messenger RNA appears at very high levels in both fetal and adult human brain. Thus, it wonld appear that hemizygosity for LIMKI is implicated in the extreme weakness for VSC shown by individuals with classic or partial WS.

Asked whether LIMK I is the gene for VSC? Mervis replied quickly, “No, we don’t want to say that. Really, there is a cascade, involving lots of genes. If you’re hemizygous for LlMKI, you have trouble with visuospatial construction, but this is not the only way to have trouble with visuospatial construction.” Is the damage limited to VSC? “LIMK1 does not affect the other cognitive traits and abilities we measured in the two SV AS kindreds, but more could be measured that may tum out to be affected,” explained Mervis. “For instance, there may be an effect on the executive function of planning. People with classic WS have a great deal of difficulty on tasks such as the Tower of London. There is a generally strong correlation between spatial planning and ability in VSc.” But as of yet, the executive function abilities of the members of the two SVAS kindreds have not been studied.

Implications for Intelligence

And what does this all tell us about the nature of intelligence, and the sometimes controversial issue of Ug”? Mervis explains, “This is the way we think about it: Ability on pattern construction within the WS population is highly correlated with g. Those who have good backward digit span (i.e., working memory) are better at both language and pattern construction than those who don’t. This suggests that there is some role for basic processes or central processes that many people equate with g. People with WS who are better at analytical tasks (e.g., the Raven matrices, generally considered the best single measure of g) are better at both language and pattern construction than those who do not perform well on analytical tasks. Visuospatial construction is definitely not decoupled from these people’s other abilities, but it is an extreme weakness.”

The study of LIMK1 is far from complete. Keating’s lab is working to develop mice which have the LIMK1 gene deleted in one copy of chromosome 7 (hemizygous “knockout mice”). Asked whether there are any cases were known in which a person was missing both copies of LIMKI , Mervis replied, “We are not aware of any homozygotes for the deletion. It remains to be seen whether such individuals [either mice or humans] are viable.” When the knockout mice are ready, it may be difficult to test them for VSc.

“Wild mice construct nests, but it isn’t clear that laboratory mice can construct anything. Remember, the problem is quite specific for construction-perceptual matching is at the level expected for IQ.”

No structural lesions are detectable by Magnetic Resonance Imaging (MRI) in the brains of people with WS who have been examined, and no brain imaging has yet been done on the partial WS individuals. Nor has anyone done functional brain imaging on people with WS. “One problem,” explains Mervis, ” is that people with WS are very sensitive to sound and have a lot of anticipatory anxiety,” making them especially uncomfortable in an MRJ device. This research on LIMK1 provides a striking ex.ample of the potential of interdisciplinary research. Through the study of a rare genetic syndrome, a new window has been opened onto the field of human cognition and its development.