Fibrinogen levels were pharmacologically (A) or genetically (B) reduced in TgCRND8 mice and Tg6799 (C) AD mice. The distribution of amyloidosis in the brain was analyzed after treatment. The decrease in fibrinogen levels provoked a significant decrease in the total amount of CAA in both transgenic AD mouse lines. CAA was determined by Thioflavin S staining in 2-4 sections from 4-7 mice/group. Bars represent mean ± SEM. *p<0.05.

(A) Clots formed after topical application of 10% FeCl₃ in TgCRND8 and WT mice were treated with tPA, and clot size was followed and determined over time. Representative images of the time series of clot dissolution after tPA treatment of a pre-formed clot in WT (B) and TgCRND8 mice (C). Blood flow is interrupted with dark zones representing clot formation (arrows). Scale bar, 50 μm. *p<0.05, **p<0.01, ***p<0.001; TgCRND8 vs WT mice.

TgCRND8 and their non-transgenic littermates (WT) were stereotaxically injected with fluorescently-labelled fibrinogen (FBG-488). Representative confocal tile scan images of FBG-488 (green) at the site of the injection in 6-month-old WT (A) and TgCRND8 (B) mice one day after injection. Nuclei were counterstained with DAPI (blue). Scale bar, 200 μm. Quantification of FBG-488 area in 9-week and 6-month-old injected mice (C) showed the amyloid burden and degree of pathology affect fibrin(ogen) clearance. A control molecule, tetramethylrhodamine-BSA (BSA-Rhod), was also injected and quantified in another set of 6-month-old mice (D). Values represent the mean ± SEM from 3-4 mice/group. *p<0.05; TgCRND8 vs WT.

(A) TgCRND8 mice were implanted with pumps delivering ancrod or saline for 4 weeks and tested in the Morris water maze. TgCRND8 mice with reduced fibrinogen levels spent significantly more time in the target quadrant than TgCRND8 controls, indicating that the pharmacological reduction of fibrinogen improves spatial memory retention. (B) TgCRND8 mice heterozygous for fibrinogen (TgCRND8-fbg+/-) and their littermates were tested in the Y-maze. TgCRND8-fbg+/- mice spent significantly more time exploring the novel arm than TgCRND8 control mice, indicating that the genetic reduction in fibrinogen improves working memory. (C) In contrast, mice deficient in plasminogen (plg-/-), which are predisposed to severe thrombosis and present fibrinogen deposits in different organs, showed memory impairment compared to WT mice in the Morris water maze. Bars represent the mean ± SEM. *p<0.05, **p<0.01.

(A) Combined fibrin formation/degradation assays. Clotting of purified fibrinogen was initiated with thrombin in the presence of tPA and plasminogen and either 10 μM of Aβ42 or vehicle (control) (n=4/group). (B) Total formation/degradation time for pure fibrin clots in (A) determined from initiation to complete dissolution; **p <0.01. (C) Dose response of the effect of Aβ42 on clot formation by adding human thrombin to fibrinogen at time 0. Control (vehicle); Aβ42 100 nM; Aβ42 500 nM; Aβ42 5 μM. (D) The addition of other β-pleated sheet amyloids (calcitonin and amylin; 5 μM) does not affect the turbidity of the clot like Aβ42 (5 μM). The lower general turbidity obtained in (D) is due to DMSO added in each curve, as explained in Experimental Procedures. Curves in C & D are representative of 4 experiments.

Perfused TgCRND8 mice at 6 months-of-age have fibrin(ogen) deposits at sites of CAA, detected by Congo Red fluorescence, in the cortex (A-C) and hippocampus (D-F). Fibrin(ogen) immunohistochemistry was performed on human post-mortem sections from the frontal cortex of 5 control and 9 AD patients. The number of parenchymal vessels >20 μm that contained fibrin(ogen) was quantified (G), and patients diagnosed with AD presented nearly twice the number of fibrin(ogen)-positive vessels compared to control subjects; **p<0.01. Representative 20× images showing fibrin(ogen)-positive vessels (arrows) in controls (H) and AD patients (I). High power magnification showed that fibrin(ogen) in AD patients co-localized with CAA (green, via Thioflavin S staining; J-L), lined the interior of the vessel wall (M-O), and filled the lumen (P-R) in addition to being deposited in the tunica media of CAA-positive vessels (S-U). Scale bars, 20 μm.

Confocal images (inverted gray levels) at low magnification of a control fibrin clot (A) or an Aβ42-influenced clot (B) using platelet-deficient human plasma and thrombin in the presence of fluorescent-conjugated fibrinogen. Scale bar in A & B, 36.5 μm. (C) Fluorescent fibrin forms a network with aggregates in the presence of Aβ42. (D) Congo Red fluorescence in the same field as in (C). Scale bar in C & D, 36.5 μm. Confocal image of fibrin clot showing aggregates acquired 15 (E) and 90 (F) min after thrombin addition. Scale bar in E & F, 8.75 μm. SEM image obtained from control fibrin clot formed from pure fibrinogen and thrombin (G) or in the presence of Aβ42 (H). Scale bar in G & H, 1.25 μm. Inset, 1 μm × 1 μm. (I) Red and green show the edge of a clot formed from plasma in the presence of Aβ42 before and 5 min after the addition of tPA, respectively. Scale bar, 36.5 μm. (J) The lysis front retreat rate (μm/min) was determined from 5 min time-lapse confocal acquisitions (n=4); ***p <0.001. Aggregates remaining after fibrinolysis contain fibrin(ogen) (K) and Aβ42 (L).

A cranial window was opened in 6-month-old TgCRND8 and WT littermates, and in vivo imaging and clot formation were carried out. Representative images of time series of clot formation before (A) and after the addition of 2.5% (B), 10% (C) and 20% (D) FeCl₃. Fluorescent-labeled blood flow (gray) is interrupted with dark zones representing clot formation (black). (E) The number of occluded vessels after FeCl₃ treatment at incremental doses was recorded in TgCRND8 mice and WT littermates over time. TgCRND8 mice show occlusion earlier and with lower doses of FeCl₃. Clot formation was also induced using the near-infrared laser of a two-photon laser scanning fluorescence microscope. Representative images of maximum projections (Z-stack) from the same area taken before (F-H) and after (I, J) the laser-induced injury in a TgCRND8 mouse. Methoxy-X04 was injected to label Aβ deposits and identify CAA-positive vessels (G, pseudocolored red). The white boxes show the region of interest where the laser was focused to form a clot (arrow in J). Scale bars, 40 μm. (K) Percent vessel occlusion after injury was quantified. TgCRND8 vessels showed a significant increase in the degree of occlusion compared to WT vessels. Note that under the same conditions as TgCRND8 mice, some of the vessels in WT mice did not occlude. (L) The same vessels quantified in (K) were classified and plotted as CAA-positive and CAA-negative based on the Methoxy-X04 staining. *p<0.05, **p<0.01, ***p<0.001; TgCRND8 vs WT mice.