Cystatin C inhibits amyloid-b deposition in Alzheimer’s disease mouse models Weiqian Mi 1 , Monika Pawlik 1,2 , Magdalena Sastre 2,5 , Sonia S Jung 1,5 , David S Radvinsky 1 , Andrew M Klein 1 , John Sommer 1 , Stephen D Schmidt 1 , Ralph A Nixon 1,3,4 , Paul M Mathews 1,3 & Efrat Levy 1–3 Using transgenic mice expressing human cystatin C (encoded by CST3), we show that cystatin C binds soluble amyloid-b peptide and inhibits cerebral amyloid deposition in amyloid-b precursor protein (APP) transgenic mice. Cystatin C expression twice that of the endogenous mouse cystatin C was sufficient to substantially diminish amyloid-b deposition. Thus, cystatin C has a protective role in Alzheimer’s disease pathogenesis, and modulation of cystatin C concentrations may have therapeutic implications for the disease. Our findings are consistent with evidence 1 for a role of cystatin C in Alzheimer’s disease, a neurodegenerative disorder characterized by the deposition of amyloid-b in the brain. Cystatin C, a ubiquitous cysteine protease inhibitor 1 , co-localizes with amyloid-b in amyloid plaque cores, amyloid-laden vascular walls 2–5 and pyramidal neurons within brain regions most prone to amyloid deposition 5 . In humans, the B/B polymorphism in CST3, which results in less efficient cleavage of the signal peptide and reduced cystatin C secretion 6,7 , is associated with increased risk of Alzheimer’s disease 1 . This genetic linkage is note- worthy given that cystatin C binds amyloid-b1–40 and amyloid-b1–42 with high affinity in vitro, inhibiting amyloid-b fibril formation 8 . To investigate the in vivo effects of cystatin C on amyloid-b deposition, we crossed transgenic mice overexpressing human CST3 (ref. 9) with mice overexpressing human APP (Tg2576 (ref. 10) or APP23 (ref. 11) mice). Human wild-type cystatin C (CysC-W) or the L68Q variant (CysC-V) were expressed under transcriptional control of the CST3 promoter 9,12 . The L68Q cystatin C variant in individuals with hereditary cerebral hemorrhage with amyloidosis–Icelandic type (HCHWA-I) 12 causes cystatin C amyloid deposition in the central nervous system vasculature, resulting in cerebral hemorrhages 13 . a b c d APP Dimer 28 17 14 Glycosylated Monomer 0.6 0.3 0.25 0.2 0.15 0.1 0.05 0 Tg2576/CysC-V APP23/CysC-V 0.5 0.4 0.3 0.2 Dimer/monomer Dimer/monomer 0.1 0 APP – /CysC + APP + /CysC + APP – /CysC + APP + /CysC + CysC CysC-V CysC Aβ Aβ CysC Aβ CysC APP CysC 1 2 3 4 5 6 CysC-V CysC-W – – – – + + + + + – + fIAPP CTER Aβ 7 8 1 2 3 4 5 6 9 7 8 10 11 12 98 IP Western APP 64 22 17 14 6 3 16 6 – – – – – – + + + + + – – – – + + + – + + + + + + – – ** CysC Aβ Figure 1 APP overexpression diminishes cystatin C (CysC) dimerization. (a) Protein blot analysis with anti–cystatin C of brain homogenates of mice expressing the variant human CysC-V(M11) or wild-type human CysC-W(F6) and of those crossbred with Tg2576. Cystatin C monomeric, glycosylated and dimeric forms are indicated, as are molecular size markers (kDa). (b) The protein bands were scanned and quantified, and the results are expressed as mean ± s.e.m. for the dimer relative to the monomer (for all methods, P values and n, see Supplementary Methods). (c) Cystatin C overexpression does not affect either APP expression levels or proteolytic processing. Protein blot analysis with anti–APP 22C11 (lanes 4–6) or 6E10 (lanes 1–3, 7–12) of mouse brain homogenates of 2-month-old CysC-W(F6) mice crossbred with APP23 (lanes 1–3), 3-month-old CysC-V(M11) mice crossbred with Tg2576 (lanes 4–9) or 12-month-old CysC-V(F6) mice crossbred with APP23 (lanes 10–12). Full-length APP (flAPP), C-terminal fragments of APP (CTER) and amyloid-b are marked. (d) In vivo binding of cystatin C to amyloid-b. Protein blot with anti–amyloid-b of proteins immunoprecipitated (IP) with anti–cystatin C from brain homogenates of 18-month-old Tg2576 or Tg2576/CysC-V(M11) mice (lanes 1–4), and with anti–cystatin C of proteins immunoprecipitated with anti– amyloid-b from brain homogenates (lanes 5 and 6) or plasma (lane 7) of 6-month-old mice and from brain homogenate of a neuropathologically normal individual (lane 8). Cystatin C and amyloid-b are marked. No amyloid-b or cystatin C were immunoprecipitated when the immunoprecipitating antibody was omitted or replaced with an irrelevant mouse antibody (Supplementary Fig. 1 online). Received 20 July; accepted 7 September; published online 18 November 2007; doi:10.1038/ng.2007.29 1 Nathan S. Kline Institute, Orangeburg, New York 10962, USA. 2 Departments of Pharmacology, 3 Psychiatry and 4 Cell Biology, New York University School of Medicine, New York, New York 10016, USA. 5 Present addresses: Department of Cellular and Molecular Neuroscience, Imperial College London, London W12 0NN, UK (M.S.) and Centocor Research and Development Inc., Radnor, Pennsylvania 19087, USA (S.S.J.). Correspondence should be addressed to E.L. (elevy@nki.rfmh.org). 1440 VOLUME 39 [ NUMBER 12 [ DECEMBER 2007 NATURE GENETICS BRIEF COMMUNICATIONS © 2007 Nature Publishing Group http://www.nature.com/naturegenetics