The role of flotillins in regulating aβ production, investigated using flotillin 1-/-, flotillin 2-/- double knockout mice

Abstract

Flotillin 1 and flotillin 2 associate in the plasma membrane to form microdomains that have roles in cell signaling, regulation of cell-cell contacts, membrane-cytoskeletal interactions, and endocytosis. They are thought to be involved in the trafficking and hence processing of the Amyloid Precursor Protein, APP. In this study we set out to obtain in vivo confirmation of a link between flotillins and cleavage of APP to release amyloidogenic Aβ peptide, and to generate tools that would allow us to ask whether flotillins are functionally redundant. We used a mouse model for Aβ-dependent cerebral amyloidosis, APPPS1 mice, combined with deletion of either flotillin 1 singly, or flotillin 1 and flotillin 2 together. There was a small but significant reduction in Aβ levels, and the abundance of congo-red stained plaques, in brains of 12 week old mice lacking flotillin 1. A similar reduction in Aβ levels was observed in the flotillin 1-/-, flotillin 2-/- double knockouts. We did not observe large effects on the clustering or endocytosis of APP in flotillin 1-/- mouse embryonic fibroblasts. We conclude that flotillins are likely to play some role in APP trafficking or processing, but the relevant cellular mechanisms require more investigation. The availability of flotillin 1-/-, flotillin 2-/- mice, which have no overt phenotypes, will facilitate research into flotillin function in vivo.

Figure 1. Deletion of flotillin 1 causes a small reduction in APP endocytosis in mouse embryonic fibroblasts.

All cells are primary MEFs A. Cells expressing GFP-APPswe were labelled on ice, or at 37°C with Alexa555 conjugated anti-GFP antibodies and subsequently analysed by flow cytometry. The right hand panel overlays signals from 4°C and 37°C. The difference between the signals is due to endocytosis of the antibody at 37°C. B. Colocalisation of GFP-APPswe with internalised anti-GFP(Alexa555) antibodies following 30 minute incubation at 37°C C. Primary MEFs from flotillin 1-/- mice or congenic controls, expressing GFP-APPswe, were labelled for 30 min at 37°C with Alexa555 anti-GFP antibodies as in A. In order account for variable expression levels, APP uptake was calculated as the ratio of median Alexa555 fluorescence (anti-GFP) over median GFP fluorescence, after gating for GFP-positive cells as shown by the black line overlaid in A. This ratio was normalised so that the mean of the control values equalled 1, to allow comparison of different experiments. Data from 5 separate flow cytometry experiments are shown. Bars are SEM.

Figure 2. Deletion of flotillin 1 does not alter clustering of APP in mouse embryonic fibroblasts.

A. Immunofluorescence staining for endogenous flotillin-1, which resides in flotillin microdomains at the plasma membrane, does not overlap with GFP-APPswe in images acquired with TIR illumination. Pearson's correlation coefficient calculated from 10 similar images is shown. B. GFP-APPswe colocalised extensively with antibody against clathrin heavy chain, detected by indirect immunofluorescence and TIR illumination. Pearson's correlation coefficient calculated from 10 similar images is shown. C. Consistent with previous reports, no significant overlap was observed between flotillins and clathrin, detected as in B. Pearson's correlation coefficient calculated from 10 similar images is shown. D. PALM during TIR illumination was used to determine the size of mEos2-APPswe clusters at the plasma membrane of MEFs. A representative image after particle detection and reconstruction with QuickPALM is shown. Fitted centroids from the PALM analysis are represented as a single pixel. The intensity of that pixel is proportional to the accuracy of the fitted centroid, and the intensities of centroids fitted to the same pixel are summed. E. Frequency distribution of cluster size for mEos2-APPswe at the plasma membrane, comparing control and flotillin 1-/- MEFs. At least 10 images were analysed for each genotype.

Figure 3. Deletion of flotillin 1 reduces the accumulation of both soluble Aβ, and Aβ in formic-acid extractable plaques, in brains of APPPS1 mice.

A. APP from RIPA buffer solubilised lysates of mouse brain was immunoprecipitated with the monoclonal antibody 6E10, and the precipitates analysed by Western blotting with the same antibody. The bands corresponding to full length APP (APP), β C-terminal fragment of APP (βCTF), and Aβ are indicated. Bands with an * are present in mice not expressing human APP, and are most likely antibody heavy and light chains. 3 mice of each genotype were analysed. Approximate positions of protein molecular weight markers are indicated. B. Brain tissues from 12 week old APPPS1 or APPPS1, flotillin 1-/- mice were harvested and Aβ levels were measured quantitatively using ELISA. Soluble Aβ40 and Aβ42 were present in the supernatant after tissue homogenisation and centrifugation at 20,000 rcf. Each data point represents assay from the brain of one mouse. Bars are SEM. C. Brain tissues were harvested as in B above, but Aβ40 and Aβ42 were extracted with 70% formic acid from the pellet, after tissue homogenisation and centrifugation, and the levels assayed using ELISA. Each data point represents assay from the brain of one mouse. P values were calculated using Student's t-test. Bars are SEM.

Figure 4. Deletion of flotillin 1 reduces the accumulation of Congo Red stained plaques in brains of APPPS1 mice.

A. Brain hemispheres from flotillin 1-/- and congenic control mice were sectioned on a cryotome, stained with Congo Red, imaged on a stereo microscope and subsequently analysed with ImageJ. B. Quantification of the proportion of the area of each brain slice stained with Congo Red. Each data point represents one slice, data are pooled from 10 flotillin 1-/- and 10 control mice. Bars are SEM.

Figure 5. Deletion of flotillin 1 and flotillin 2 causes the same reduction in Aβ levels as is observed when flotillin 1 alone is deleted.

A. The insertion of a gene trap in the flotillin 2 locus causes a band shift from 500 bp to 350 bp in PCR-based genotyping. The left hand panel indicates position of DNA size markers. B. In flotillin1 -/- MEFs the amount of flotillin 2 is slightly reduced, while in flotillin 2-/- MEFs there is a dramatic reduction in the amount of flotillin 1 present. As expected, no flotillin proteins can be detected in double knockout cells. C. Brain tissues were harvested and Aβ levels were measured quantitatively using ELISA. Soluble Aβ40 and Aβ42 were present in the supernatant after tissue homogenisation and centrifugation at 20,000 rcf. Each data point represents assay from the brain of one mouse with the genotype shown. Bars are SEM. D. Brain tissues were harvested as in C above, but Aβ40 and Aβ42 were extracted with 70% formic acid from the pellet, after tissue homogenisation and centrifugation, and the levels assayed using ELISA. Each data point represents assay from the brain of one mouse. P values were calculated using Student's t-test. Bars are SEM.