Animals and Tissue Plasminogen Activator Treatment
Male 23-month-old APP23 mice were anesthetized, and either saline (n = 12), 1mg/kg tPA (n = 12), or 10mg/kg tPA (n = 12) (Alteplase, Actilyse, Boehringer Ingelheim, Germany) was administered by the tail vein. The low dose corresponds to the dose used in humans for thrombolysis. The high dose was chosen because the clot-lysis system in rodents has been reported to be less responsive to tPA than that of humans. An additional group of nontransgenic littermate control mice (n = 12) was treated with 10mg/kg tPA. Ten percent of the dose was given within the 1st minute, and the rest was given over 30 minutes with a micropump. Five mice died within 10 days after treatment (saline, 1; low-dose tPA, 3; high-dose tPA, 1; nontransgenic controls, 2).
Histology and Immunohistochemistry
Ten days after treatment, mice were sacrificed, and serial 25m paraffin sections were cut throughout the brain. Cresyl violet, hematoxylin and eosin, Congo red staining, the Berlin blue method of Perls, and immunohistochemistry with polyclonal antibody NT12 to A were performed according to previously published protocols.
Quantification of Cerebral Amyloid Angiopathy and Hemorrhage
CAA was quantified with a previously described rating system. In brief, the CAA frequency was obtained from the total number of affected vessels in a systematically sampled set of every 20th A-immunostained section. For estimation of CAA severity, A-positive vessels were classified into three severity grades. A CAA score was obtained by multiplication of the CAA frequency with the CAA severity. One mouse was eliminated from the analysis because its CAA score exceeded the group mean by more than four times the standard deviation of the group.
Cerebral hemorrhage is followed by a delayed appearance of hemosiderin-positive microglia cells 5 to 7 days after the bleeding. For the quantification number, Berlin blue-stained clusters of hemosiderin were assessed on sets of every 10th systematically sampled section throughout the entire neocortex. An additional set of every 10th section was stained for hematoxylin and eosin and screened for remnants of acute intraparenchymal or subarachnoidal hematomas. All quantification was performed by two raters (D.T.W. and L.B.) with interobserver correlations of r = 0.80 to 0.93.
Magnetic Resonance Imaging
Cerebral magnetic resonance imaging scans of randomly chosen mice (saline, 5; high-dose tPA, 7; nontransgenic controls, 5) were performed 4 to 7 days after tPA administration. Magnetic resonance imaging scans were taken on a 4.7T DBX spectrometer (Bruker, Karlsruhe, Germany). In brief, the mice were anesthetized and subjected to one imaging cycle, in which T2-weighted 1mm-thick coronal brain slices were scanned with a RARE sequence (repetition time, 3,000msec; effective echo time, 88msec; spatial resolution in plane, 86m2). The total measuring time was 2.3 minutes.