Levy and colleagues previously demonstrated that cystatin C binds amyloid-β in vitro (Sastre et al., 2004). “I was initially interested in cystatin C because a variant of this inhibitor forms amyloid deposited in the cerebral vasculature of patients with hereditary cerebral hemorrhage with amyloidosis, Icelandic type (HCHWA-I), leading to cerebral hemorrhages early in life,” said Levy (see ARF related conference story). Immunohistochemical studies in the early 1990s had indicated colocalization of cystatin C with Aβ in the brain of AD patients. To assess whether these two proteins interact in the brain, Levy crossed mice ubiquitously overexpressing the cystatin C gene CST3 with Tg2576 mice, which overexpress human APP with the Swedish FAD mutation.

Levy presented some of her initial findings of this study at last year’s Society for Neuroscience meeting in Atlanta, where she speculated that cystatin C may act as a neuroprotective agent (see ARF related news story). In the present animal study, first author Weiqian Mi and colleagues report that cystatin C binds soluble amyloid-β in vivo, acting to sequester it and to reduce amyloid plaque load. Further, the New York group found that cystatin C significantly reduced cystatin C dimerization, which commonly occurs in HCHWA-I. From these findings, Levy posits that cystatin C, a secreted protein that originates from within the cell, acts as a carrier of soluble amyloid-β in blood, cerebrospinal fluid, and the brain, where it inhibits Aβ aggregation into plaques. Increasing CST3 expression might affect amyloid deposition and potentially alter the evolution of AD pathology, the authors write, speculating that even small increases could affect pathogenesis.

Jucker’s group started out trying to generate a mouse model for HCHWA-I. Surprisingly, none of the various lines the scientists generated to express either wild-type or mutant cystatin C proved to have either cystatin C amyloid or cerebral hemorrhages. Suspecting that Aβ might promote cystatin C aggregation, the scientists next crossed their mutant (HCHWA-I) cystatin C mice to APP23 transgenic mice. This did not generate cystatin C amyloid, either. On the contrary, it halved the number of parenchymal amyloid plaques compared to what APP23 mice usually have. It was not just the mutant protein: a different line, of wild-type cystatin overexpressors bred to APP/PS1 mice, also had fewer amyloid plaques than did the APP/PS1 mice alone. Closer in-vitro analysis of the interaction between the two proteins showed that the proteins form Aβ-cystatin C heterodimers, that Aβ likely inhibits the dimerization of mutant cystatin C, as Levy’s team had found also, and that cystatin C inhibits Aβ fibril formation. This means that, instead of the expected coaggregation between two amyloidogenic proteins, the scientists saw that one amyloidogenic protein interfered with the aggregation of the other. Amyloid interference of a similar sort has been described for Aβ40, which inhibits Aβ42 fibril formation (Kim et al., 2007) and for β-synuclein and α-synuclein, respectively (Tsigelny et al., 2007). This group could not find evidence for a neuroprotective effect of cystatin C.

To assess whether the two proteins interact inside or outside of the neurons, first author Stephan Kaeser and colleagues generated transgenic mice that overexpress wild-type cystatin C only in astroglia and crossed them with APP/PS1 mice. These mice had a similar phenotype as the previous strains that express cystatin C in neurons driven by the Thy1 promoter. The German scientists take this to mean that the interaction between cystatin C and Aβ occurs mostly outside neurons, where it might be more accessible to any future therapeutic intervention. In short, a half-dozen new mouse lines after the beginning of their study, the researchers still don’t have a model for HCHWA-I, but they know more about why cystatin C variants can be a risk factor for AD. The cystatin C variant associated with increased risk for AD appears to reduce secretion of the protein; more generally, people with AD tend to have low cystatin C serum levels (Benussi et al., 2003; Chuo et al., 2007). Cystatin C levels vary among people by some 10 to 20 percent—over one’s adult lifetime, this can make for a significant difference in AD risk, Jucker speculated by e-mail.

Taken together, the data available to date would suggest that the level of cystatin C in the brain affects a person’s risk for AD because the cystatin C counteracts aggregation of Aβ (see also Selenica et al., 2007). In this way, the findings broadly support John Hardy’s hypothesis of mass action (Singleton et al., 2004; ARF related conference story), whereby genetic changes that lead to a rise, over time, in the presence of amyloidogenic proteins increase the risk for disease.

Ástrídur Palsdottir, University of Iceland, Reykjavik, who studies cystatin C gene mutations, e-mailed ARF that “together, these findings underline and provide a mechanism for the reported genetic association between cystatin C and Alzheimer disease, and have implications regarding therapeutic interventions.“ To this end, Levy is currently studying the mechanism by which cystatin C protects neuronal cells from a variety of insults in tissue culture and in vivo. Understanding the function of this inhibitor may allow researchers to identify a way to keep amyloid-β in its soluble form and to facilitate its degradation.—Gabrielle Strobel and Rachel Ahmed.

Rachel Ahmed is a Ph.D. student at the University of Kentucky.

References:
Kaeser SA, Herzig MC, Coomaraswamy J, Kilger E, Selenica ML, Winkler DT, Staufenbiel M, Levy E, Grubb A, Jucker M. Cystatin C modulates cerebral beta-amyloidosis. Nat Genet. 2007 Nov 18; Abstract

Mi W, Pawlik M, Sastre M, Jung SS, Radvinsky DS, Klein AM, Sommer J, Schmidt SD, Nixon RA, Mathews PM, Levy E. Cystatin C inhibits amyloid-beta deposition in Alzheimer's disease mouse models. Nat Genet. 2007 Nov 18; Abstract