Matrix metalloproteinase inhibition reduces oxidative stress associated with cerebral amyloid angiopathy in vivo in transgenic mice

Abstract

Cerebral amyloid angiopathy (CAA), characterized by extracellular beta-amyloid peptide (Abeta) deposits in vessel walls, is present in the majority of cases of Alzheimer's disease and is a major cause of hemorrhagic stroke. Although the molecular pathways activated by vascular Abeta are poorly understood, extracellular matrix metalloproteinases (MMP) and Abeta-induced oxidative stress appear to play important roles. We adapted fluorogenic assays for MMP activity and reactive oxygen species generation for use in vivo. Using multiphoton microscopy in APPswe/PS1dE9 and Tg-2576 transgenic mice, we observed strong associations between MMP activation, oxidative stress, and CAA deposition in leptomeningeal vessels. Antioxidant treatment with alpha-phenyl-N-tert-butyl-nitrone reduced oxidative stress associated with CAA (approximately 50% reduction) without affecting MMP activation. Conversely, a selection of agents that inhibit MMP by different mechanisms of action, including minocycline, simvastatin, and GM6001, reduced not only CAA-associated MMP activation (approximately 30-40% reduction) but also oxidative stress (approximately 40% reduction). The inhibitors of MMP did not have direct antioxidant effects. Treatment of animals with alpha-phenyl-N-tert-butyl-nitrone or minocycline did not have a significant effect on CAA progression rates. These data suggest a close association between Abeta-related MMP activation and oxidative stress in vivo and raise the possibility that treatment with MMP inhibitors may have beneficial effects by indirectly reducing the oxidative stress associated with CAA.

Figure 1

Representative example of cerebrovascular amyloid angiopathy, oxidative stress and MMP activation in the same vessel segment from a living APPswe/PS1dE9 mouse. The signal from amyloid angiopathy signal assessed with thioflavin S (red, A), Amplex Red (blue, B) and the fluorescein conjugate DQ™ gelatin (green, C) spatially coincide as shown after thresholding, segmenting, and aligning the image from each marker (D). Scale bar=100 μm.

Figure 2

Characterization of the in situ and in vivo imaging of MMP activation with DQ Gelatin substrate. Panels A-D show representative examples of in situ MMP zymography in fresh frozen tissue from Tg2576 mice showing MMP activation associated with CAA in leptomeningeal vessels. MMP activation was observed in the vessels with CAA (A) whereas preincubation with 1-10-phenanthroline monohydrate (B) or precleavage of DQ™ gelatin with collagenase IV from Clostridium histolyticum (C) inhibit MMP activation associated with CAA. Similarly, no MMP activation was observed in sections from age-matched wild type mice (D). Panels E-L show a representative example of the time-course of activation of MMP associated with CAA after DQ™ gelatin application in vivo in APPswe/PS1dE9 mice. The same vessel segment was imaged before DQ™ gelatin was topically applied (A) and after 10 minutes (B), 15 minutes (C), 20 minutes (D), 30 minutes (E), 40 minutes (F) and 50 minutes (G). By 20 minutes after the application of the substrate, the signal observed was saturated, and similar to that observed at subsequent time points. The same area was also imaged after thioflavin S incubation for histochemical assessment. Scale bar: 50 μm.

Figure 3

Quantification of MMP activity and oxidative stress associated with amyloid angiopathy in two different mouse models. The number of fluorescent pixels resulting from proteolytic digestion of DQ™ gelatin by MMP (A and B) or Amplex red (C and D) are expressed as a percentage of thioflavin S pixels on the same vessel segments. A) In APPswe/PS1dE9 mice, the percentage of MMP positive pixels is not significantly affected after antioxidant treatment with PBN (100 mg/Kg, 2 days). However, the MMP inhibiting agents minocycline (100mg/Kg, 14 days), simvastatin (20mg/Kg, 14 days) and GM6001 (100mg/Kg 3 days) significantly reduced MMP activity when compared to untreated control values. Data are representative of 3-7 animals and 21-32 vessel segments. One-way ANOVA followed by Tukey-b test {F_(5,160)=4.896, *P<0.001 vs. Untreated and PBN}. B) In Tg2576 mice, the percentage of MMP-positive pixels was significantly reduced by the MMP inhibitor minocycline (50 and 100mg/Kg, 14 days), whereas it was not affected by antioxidant treatment with PBN (15mg/Kg, 14 days) when compared to untreated control values. Data are representative of 3-7 animals and 18-75 vessel segments. One-way ANOVA followed by Tukey-b test {F_(3,187)=18.344, *P<0.001 vs. Untreated}. C) In APPswe/PS1dE9 mice the percentage of AR pixels is significantly reduced after antioxidant treatment with PBN (100mg/Kg 2 days) and also by the MMP inhibiting agents minocycline (100mg/Kg, 14 days), simvastatin (20mg/Kg, 14 days) and GM6001 (100mg/Kg 3 days) when compared to untreated control values. Data are representative of 3-7 animals and 20-36 vessel segments. One-way ANOVA followed by Tukey t test {F_5,203=7.894; *P<0.001 vs. Untreated and minocycline 7 days}. D) In Tg2576 mice, the percentage of AR-positive pixels was significantly reduced after PBN treatment (15 mg/Kg, 14 days) and also by minocycline treatment (100mg/Kg and 50 mg/Kg, 14 days). Data are representative of 3-7 animals and 25-99 vessel segments. Oneway ANOVA followed by Tukey-b test {F_3,226=5.728, *P<0.001 vs. Untreated.

Figure 4

Representative example of CAA, oxidative stress, and MMP activation in individual segments of leptomeningeal vessels from APPswe/PS1dE9 mice using in vivo multiphoton microscopy. CAA deposits were identified by fluorescence from thioflavin S staining. The oxidative stress signal was derived from Amplex Red oxidation and MMP activation was derived from the green fluorescent substrate (DQ™ gelatin). In untreated mice, ~70% of thioflavin S-positive pixels were Amplex Red and MMP positive. In PBN treated mice, only ~30% of thioflavin S-positive pixels were Amplex Red-positive. Minocycline administered for 7 days (Min7), showed no effect on the oxidative stress signal, although MMP-positive pixels were reduced to ~50% of the thioflavin S-positive pixels. Minocycline administered for 14 days (Min 14) reduced both the Amplex Red signal (~40%) and MMP signal (~50%). A similar profile was observed after simvastatin (Sim) and GM6001 treatment. Scale bar=100 μm.

Figure 5

Inhibitors of MMP do not reduce oxidative stress from senile plaques ex vivo. The quantitative oxidative stress index using the ratio of Amplex Red signal to Thioflavin S fluorescence from individual senile plaques in sections from Tg2576 mouse brain was significantly reduced with PBN treatment (100 μM), whereas no effect was observed after minocycline, sinvastatin or GM6001 treatments. Data are representative from 14-41 plaques. Significant differences were determined by one-way ANOVA followed by Tukey-b test {F_(7,202)=2.442; *P<0.05 vs. Untreated and minocycline}.

Figure 6

Representative examples of cerebrovascular amyloid angiopathy progression in the same vessel segments from Tg2576 mice. A) Longitudinal imaging of identified vessels labeled with thioflavin S at day 0 and with systemic methoxy-XO4 administration at 7 and 14 days in untreated and treated animals. An ALZET minipump with PBN (15mg/Kg/day) was implanted s.c. for 14 days before cranial window surgery was performed, and replaced with a new one for an additional 14 days during the surgery. Minocycline (50mg/Kg) was administered daily, ip, for 14 days prior to cranial window implantation, and continued for the next 2 weeks. Scale bar=100 μm. B) Quantitative measurements of CAA progression in Tg2576 mice after PBN and minocycline treatment . No statistical differences were observed in CAA progression from 3-4 mice and 14-55 vessel segments (P-values: Untreated vs. PBN=0.94, Untreated vs. minocycline=0.75 and PBN vs. Minocycline 0.73). .