Supplementary Figure 1 PICALM expression in brain capillary endothelium in human brain and in mouse brain. a, Double immunostaining for PICALM (red, left) and lectin positive endothelial profiles (blue, right) in the prefrontal cortex (Brodmann 9/10; pink merged, right) of an age matched control Braak I and AD individual Braak VI. Bar=20 µm. b, Immunoblotting for PICALM relative to tubulin in brain microvessels and microvessel depleted brains in 2 3 month old control mice. c d, Immunoblotting for PICALM relative to β actin in brain microvessels (c) and microvessel depleted brains (d) in 15 month old APPsw/0 and age matched littermate control mice. Mean + s.e.m., n=6 mice per group. e, Immunoblotting for PICALM relative to β actin in primary mouse brain endothelial cells (BEC) cultured with or without Aβ40 (1 µm) for 72 hr. Mean ± s.e.m. from 3 independent primary isolates in triplicate. P<0.05 by Student s t test; NS, non significant.
Supplementary Figure 2 Generation of Picalm +/ mouse, biochemistry and behavioral characterization. Generation of Picalm +/ mouse: a, A diagram showing the targeting strategy for generating the Picalm knockout mouse. See additional Methods for more detailed information. Biochemistry: b m, Picalm +/ mice show no changes in blood glucose, liver and renal tests and serum electrolytes compared to wild type littermate controls. The tests include: (b) Glucose; (c) Serum alkaline phosphatase; (d) Serum alanine aminotransferase (ALT); (e) Serum aspartate aminotransferase (AST); (f) Creatine phosphokinase (CPK); (g) Albumin; (h) Total protein; (i) Total bilirubin; (j) Blood urea nitrogen; (k) Creatinine level; (l) Calcium level; (m) Phosphorous level. Mean + s.e.m., n=5 mice per group. Behavioral tests: n q, Nest construction (n), burrowing (o), novel object recognition (NOR) (p) and location (NOL) (q) tests showing no cognitive difference in 9 month old Picalm +/ mice compared to age matched Picalm +/+ littermates. Means + s.e.m. n=12 mice per group. Statistical significance by Student s t test.
Supplementary Figure 3 Additional characterization of APP sw/0 Picalm +/ mice. a b, Representative western blots and densitometry analysis of PICALM in brain microvessels (a) and microvessel depleted brain homogenate (b) in 3 month old APP sw/0 Picalm +/ mice and age matched APP sw/0 Picalm +/+ littermate controls. Mean + s.e.m., n=3 4 mice per group. Statistical significance by Student s t test. β actin was used as a loading control. c, Representative images of brain tissue sections stained with human Aβ specific antibodies show accelerated Aβ deposition in the hippocampus and cortex and development of early cerebral amyloid angiopathy () in 6 month old APP sw/0 ; Picalm +/ mice compared to age matched APP sw/0 ; Picalm +/+ littermate controls. d, Increased Aβ load in the hippocampus and cortex in 6 month old APP sw/0 ; Picalm +/ mice compared to age matched APP sw/0 ; Picalm +/+ littermate controls. Mean + s.e.m., n=5 6 mice per group. Statistical significance by Student s t test.
Supplementary Figure 4 Endothelial-specific expression of Tdtomato after injection of the AAV-Flex-Tdtomato into the hippocampus of APP sw/0 ; Picalm +/ ; Tie2-Cre mice, and Aβ clearance from the hippocampus in APP sw/0 ; Picalm +/ ; Tie2-Cre mice after injection of the AAV-Flex-Picalm into the hippocampus and endothelial-specific rescue of PICALM. a, Representative confocal images in 5 month old APP sw/0 ; Picalm +/ ; Tie2 Cre mice show Tie2 Cre dependent expression of Tdtomato (red) in >50% of lectin positive endothelial vascular profiles after injection of the AAV Flex Tdtomato into the hippocampus. Co injection of AAV Synapsin GFP shows that ~3% of hippocampal neurons express Tdtomato indicating minimal leakage. b, Coronal brain section of a 6 month old APP sw/0 ; Picalm +/ ; Tie2 Cre mouse injected into the hippocampus with AAV Flex (control, left) or AAV Flex Picalm (right) show clearance of Aβ from the hippocampus after endothelial specific re expression of PICALM (right; see also main Fig. 4b). Aβ immunoststaing green; Dapi blue. Data are representative from 3 5 independent experiments.
Supplementary Figure 5 Characterization of Picalm +/ mice and APP sw/0 Picalm +/ mice. a d, APP abundance relative to β actin (a), β secretase activity (b) sappβ levels (c) and γ secretase activity determined by the production of Notch intracellular domain (NICD) fragment from Notch protein (d) were studied in the forebrain lysates from 3 month old Picalm +/ and age matched littermate controls. e j, Picalm +/ mice show no changes in the expression of major Aβ transporters in the brain microvessels, including Pgp (e), LRP1 by immunoblottting of microvessels (f) and double staining for LRP1 and endothelial specific lectin (g) and RAGE (h), and no change in the levels of Aβ degrading enzymes in the brain including neprilysin (i) and insulin degrading enzyme (IDE) (j). k n, APP abundance relative to β actin (k), β secretase activity (l) sappβ levels (m) and γ secretase activity (n) were studied in the forebrain lysates from 3 month old APP sw/0 Picalm +/ and age matched littermate APP sw/0 Picalm +/+ controls. o s, APP sw/0 Picalm +/ mice show no changes in the expression of major Aβ transporters in the brain microvessels, including Pgp (o), LRP1 (p) and RAGE (q), and no change in expression of Aβ degrading enzymes in the brain including neprilysin (r) and IDE (s). Mean + s.e.m., from 3 4 mice per group. Statistical significance was determined by Student s t test. NS, non significant.
Supplementary Figure 6 Binding of Aβ to LRP1 and interactions of Aβ-LRP1 complexes with AP-2 and PICALM. a, Proximity ligation assay (PLA) showing binding of Aβ40 (1 nm) to LRP1 on the cell surface of brain endothelial cells (BEC) at 4 o C for 30 min. b, LRP1 Aβ40 complexes colocalize with AP 2 30 s after treatment of BEC with FAM Aβ40 (250 nm). c, LRP1 Aβ42 complexes colocalize with PICALM 30 s after treatment of BEC with FAM Aβ42 (250 nm) for 30 s. Dapi, nuclear staining (blue). Insets: high magnification depicting colocalizations. Bar=10 µm. d, Minimal internalization and lack of colocalization of scramble FAM Aβ42 (green) with LRP1 (red). Bar=10 µm. Representative findings are from 3 independent primary isolates determined in triplicates (i.e., 3 different cell cultures per isolate; >20 cells in each replicate). e, Control assay for Fig. 5g shows no binding between GST and recombinant human PICALM.
Supplementary Figure 7 Apolipoprotein E and activated α2-macroglobulin do not initiate PICALM binding to LRP1. a, PICALM does not bind to LRP1 in BEC treated with lipidated apoe3 (40 nm), apoe4 (40 nm) or activated α2 Macroglobulin (α2 M*, activated with methylamine, 0.25 nm) (left). Binding of PICALM to LRP1 in BEC treated with Aβ40 (1 nm, positive control), and loss of PICALM binding to LRP1 in BEC treated with pre formed complexes between Aβ40 and apoe3, apoe4 or α2 M* (right). In these experiments co immunoprecipitation of PICALM was performed by LRP1 specific antibody (IP: LRP1) within 1 min of BEC exposure to different LRP1 ligands using assay conditions as in main Fig. 5e and described in Supplementary Methods b, Control LRP1 internalization assay with Aβ. BEC were incubated with Aβ40 (230 nm) at 4 o C for 15 min and then transferred to 37 o C for 1 min for LRP1 internalization assay. The cell surface LRP1 was immunodetected with the N terminus specific LRP1 antibody (N 20). c, LRP1 internalization is not triggered by apoe3, apoe4 or α2m*. BEC were incubated with Aβ40 (230 nm, control) or pre formed complexes between Aβ40 and apoe3, apoe4 or α2m* at 4 o C for 15 min and then transferred to 37 o C for 1 min for LRP1 internalization assay. Values are means + s.d. from 3 independent BEC isolates from 3 different donors using 3 replicate cultures. The number of cells counted for each culture was 20. The total number of cells in each group was 180. Statistical significance by ANOVA followed by Tukey s posthoc test.
Supplementary Figure 8 Endothelial cell polarity of LRP1 and RAGE in brain capillaries in human brain in situ and in an in vitro blood-brain barrier model. a, A representative confocal scanning analysis of lectin positive endothelium (blue) and LRP1 immunodetection in capillary endothelium (red) in a control human brain. Merged: purple. Bar represents 5 μm. Chart: LRP1 relative signal intensity (red) plotted over the endothelial specific lectin signal intensity (blue). b, A representative confocal scanning analysis of lectin positive endothelium (blue) and RAGE immunodetection in capillary endothelium (red) in a control human brain. Merged: purple. Bar represents 5 μm. Chart: RAGE relative signal intensity (red) plotted over the endothelial specific lectin signal intensity (blue). c, Expression of RAGE and LRP1 in the endothelial monolayer (left) and polarized expression of RAGE (red) and LRP1 (green) to the apical and basolateral side of the monolayer (right), respectively. Representative findings are from 3 independent primary isolates determined triplicates (3 different cultures per isolate; >20 cells in each replicate). d, Inhibition of synthetic human Aβ40 basolateral to apical transendothelial transport across an in vitro BBB monolayer by RAP (the receptor associated protein) and anti LRP1 specific antibody, but not by anti LRP2, anti VLDLR and anti LDLR antibodies, and in monolayers after BEC transfection with si.lrp1 but not control si.scarmble. Aβ40 transcellular transport was determined by ELISA within 30 min and was corrected for the paracellular diffusion of simultaneously measured 14 C inulin basolateral to apical transport as described in detail in the Supplementary Methods. Aβ40 transport in the absence of potential inhibitors or competitors was arbitrarily taken as 100%. Mean + s.e.m., from 3 independent isolates determined in triplicate.
Supplementary Figure 9 PICALM regulates Aβ transcytosis in brain endothelial monolayers. a, A representative proximity ligation assay (PLA) showing colocalization of Aβ LRP1 complexes (green) near the apical membrane of the endothelial monolayer within 5 min of Aβ40 (1 nm) application to the basolateral side. Bar=20 µm. b, Aβ40 (1 nm) transendothelial transport across an in vitro BBB monolayer expressed as the percentage of dose (%) of Aβ transported from the basoalateral to apical chamber. Primary human brain endothelial cells have been transfected with scramble si.rna (si.scramble), or si.rna targeting RAB11a or RAB11b, respectively. Mean + s.e.m., from 3 independent primary isolates determined in triplicate (3 different cultures per isolate; >20 cells in each replicate). c, Immunoblotting of different Rab proteins in Madin Darby canine kidney (MDCK) epithelial cells and human brain endothelial cells (BEC), showing that Rab11b is mainly expressed in BEC, while Rab11a is mainly expressed in MDCK cells. Immunoblot is representative of 3 independent experiments. For each experiment we used at least 3 different samples for each condition. d, Left panel: HPLC elution profile of 125 I Aβ40 in the apical chamber after 30 min transport from the basolateral to apical chamber shows a single peak corresponding to the elution profile of Aβ40 standard indicating no degradation. Right panel: SDS PAGE analysis showing intact 125 I Aβ40 in the apical chamber after 30 min of transendothelial transport across an in vitro BBB monolayer.
Supplementary Figure 10 PICALM and LRP1-Aβ complexes associate with EEA1, Rab5 and Rab11, but not Rab7 or LAMP1 (a lysosomal marker), in human brain endothelial cells cultured with Aβ. a, Colocalization between PICALM (red) and EEA1 (green) in BEC cultured with and without FAM Aβ40 (250 nm) for 2 min. Dapi, nuclear staining (blue). Insets: high magnification depicting colocalization. Bar=10 µm. Graph (on the left) quantification of colocalized PICALM and EEA1 puncta. b, Colocalization between Aβ LRP1 complexes visualized by the proximity ligation assay (PLA) (green) and Rab5 (red) in cultures incubated with FAM Aβ40 (250 nm) for 2 min. c, Lack of association between PICALM (red) and a lysosomal marker LAMP1 (green) in BEC cultured without and with FAM Aβ40 (250 nm) for 5 min. Dapi, nuclear staining (blue). Insets: high magnification. Bar=10 µm. Graph (on the left) quantification of colocalized PICALM and LAMP1 puncta. Mean + s.d., from 3 independent primary isolates determined in triplicates (3 different cultures per each isolate; >20 cells in each replicate). Statistical significance by Student s t test. NS, non significant.
Supplementary Figure 11 Association of Aβ42 with Rab5 and Rab11 and low colocalization between PICALM and Rab GTPases in the absence of Aβ in primary human endothelial cell cultures. a b, Association of Aβ42 with Rab 5 (a) and Rab 11 (b) in brain endothelial cells cultured with FAM Aβ42 (250 nm) for 90 s and 5 min, respectively. Dapi, nuclear staining (blue). Insets: high magnification depicting colocalization. c e, Control experiment showing barely detectable colocalizations between PICALM (red) and Rab5 (green) (c), Rab7 (d) and Rab11 (e) in primary human brain endothelial cells cultured with vehicle ( Aβ) for 2 and 5 min, respectively. Dapi, nuclear staining (blue). Insets: high magnification depicting colocalization. Bar: 10 µm. Mean + s.d., from 3 independent primary isolates determined in triplicate (3 different cultures per isolate; >20 cells analyzed in each replicate).
Supplementary Figure 12 Reduced PICALM expression in brain endothelial monolayers derived from Alzheimer s disease patients. a, A primary AD brain endothelial monolayer showing ZO 1 tight junction protein (green) and reduced expression of PICALM (red). A representative image was taken from 6 independent primary isolates determined in triplicate (3 different cultures for each isolate; >20 cells in each replicate). b, Transendothelial electrical resistance (TEER) showing normal TEER values in in vitro BBB monolayers from AD patients compared to controls. Mean + s.e.m. for 8 isolates per group studied in triplicate. c, Permeability of an in vitro BBB model derived from AD cells compared to controls to 40 kda and 2,000 kda FITC Dextran. Mean + s.e.m. from 3 isolates per group in triplicate. d, A representative western blot analysis showing the expression of Flag PICLAM, LRP1 minigene (mlrp1), and Flag PICLAM+mLRP1 in the BBB AD monolayers after adenoviral expression of PICALM and mlrp1. Tubulin is used as loading control. Immunoblot is representative of 3 independent experiments. For each experiment we used at least 3 different samples for each condition.
Supplementary Figure 13 Diagram The diagram showing PICALM regulates PICALM/clathrin dependent internalization of Aβ LRP1 complex and guides Aβ trafficking to Rab5 and Rab11 leading to Aβ endothelial transcytosis and clearance across the BBB.
Supplementary Figure 14 Original western blots.
Supplementary Table 1. Clinical and neuropathological characteristics of controls and Alzheimer s disease (AD) patients participating in this study. 1a. Patients participating in histological analysis. Patient Number Diagnosis Age Gender PMI (hr) Vascular Risk Factors Braak CERAD CDR MMSE Disease Duration Source A load (%) SD (%) PICALM/ lectin (%) SD (%) 845 Control 90 F 5 Hypertension; Atherosclerosis I Negative 0 30 N/A USC ADRC 0.50 0.08 46.44 7.89 870 Control 95 F 9 Hypertension II Negative 1 Unknown N/A USC ADRC 0.98 0.15 57.78 6.93 824 Control 87 F 7 Hypertension 0 Negative 0 30 N/A USC ADRC 0.71 0.11 64.02 10.24 861 Control 74 F 10 None I Negative 0 18 N/A USC ADRC 0.23 0.04 57.01 9.69 907 Control 84 M 3 None I II Negative 0 28 N/A USC ADRC 0.91 0.10 50.88 9.16 804 Control 82 M 9 None 0 Negative 0 28 N/A USC ADRC 0.34 0.04 58.58 5.86 877 Control 93 M 3.75 None I Sparse Unknown 18 N/A USC ADRC 0.71 0.13 78.84 10.25 559 Control 70 M 10.5 Atherosclerosis I Negative 0 Unknown N/A USC ADRC 0.32 0.04 48.36 5.80 720 Control 63 M 3.5 Hypertension 0 Negative 0 Unknown N/A USC ADRC 0.37 0.06 70.34 11.25 10473384 Control 84 M 5 Hypertension III Negative Unknown 26 N/A Rush ADRC 1.67 0.17 57.24 9.73 20683921 Control 85 F 6 Hypertension III Sparse Unknown 26 N/A Rush ADRC 0.54 0.05 44.94 4.49 20993308 Control 83 F 1 None I Negative Unknown 26 N/A Rush ADRC 0.61 0.07 59.17 5.33 11072071 Control 88 M 8.5 Hypertension II Negative Unknown 27 N/A Rush ADRC 0.17 0.02 45.08 8.11 15844425 Control 77 M 8 Hypertension III Negative Unknown 28 N/A Rush ADRC 2.29 0.34 64.32 7.08 11342432 Control 90 M 2.5 None II Negative Unknown 29 N/A Rush ADRC 0.20 0.03 54.88 7.13 11190734 Control 77 M 4.5 None II Negative Unknown 27 N/A Rush ADRC 0.25 0.05 68.57 5.49 10383017 Control 88 F 5 Hypertension I Negative Unknown 27 N/A Rush ADRC 0.18 0.03 64.40 10.30 20998065 Control 82 M 8.5 None I Negative Unknown 28 N/A Rush ADRC 0.50 0.04 69.30 6.24 11697592 Control 78 M 9.75 None I Sparse Unknown 29 N/A Rush ADRC 0.61 0.05 56.05 5.61 20800682 Control 85 F 5.75 None II Negative Unknown 30 N/A Rush ADRC 0.71 0.06 51.48 5.66 785 AD 99 F 12.3 Atherosclerosis III Moderate 2 Unknown 7 years USC ADRC 6.80 0.82 21.12 1.69 896 AD 83 F 13.25 800 AD 96 F 5 V VI Moderate 2 0 5 years USC ADRC 14.38 1.58 21.59 3.89 V Frequent 3 Unknown 9 years USC ADRC 7.62 1.22 22.88 2.06 689 AD 78 F 7 None V Moderate 2 Unknown 8 years USC ADRC 10.47 1.47 16.66 2.33 719 AD 92 F 4.25 V VI Frequent 2 Unknown 10 years USC ADRC 12.72 2.16 16.22 2.27
Patient Number Diagnosis Age Gender PMI (hr) 728 AD 86 M 10 Vascular Risk Factors Braak CERAD CDR MMSE Disease Duration Source A load (%) SD (%) PICALM/ lectin (%) IV Moderate 1 Unknown Unknown USC ADRC 8.32 1.42 25.07 3.51 905 AD 81 M 8 Atherosclerosis III Moderate 2 11 5 years USC ADRC 5.24 0.47 28.08 2.81 890 AD 71 M 5.5 None V VI Moderate 2 8 4 years USC ADRC 17.20 2.75 10.57 0.95 796 AD 76 M 5 V Moderate 3 0 2 years USC ADRC 11.65 1.28 14.20 1.28 825 AD 68 F 9.75 Hypertension III Moderate 1 15 4 years USC ADRC 4.99 0.60 29.14 5.25 902 AD 91 F 3.25 Atherosclerosis V VI Frequent 3 6 3 years USC ADRC 18.57 1.86 6.25 1.00 900 AD 87 M 5.5 V VI Frequent 2 19 Unknown USC ADRC 9.45 1.70 17.01 3.06 799 AD 95 F 6 Atherosclerosis V Frequent 2 Unknown 17 years USC ADRC 11.20 1.68 22.56 3.16 854 AD 87 F 5.5 V Moderate 2 0 17 years USC ADRC 5.00 0.75 31.31 3.76 888 AD 51 F 6 None V VI Frequent Unknown Unknown 1 month USC ADRC 7.26 0.94 38.61 3.09 427 AD 74 M 9 V Frequent 3 Unknown 15 years USC ADRC 13.53 2.30 21.42 2.14 851 AD 87 F 4.75 Atherosclerosis V Frequent 3 16 6 months USC ADRC 12.70 2.29 28.00 4.20 808 AD 98 M 6 Hypertension, V VI Frequent 1 28 2 years USC ADRC 14.11 1.13 36.18 3.62 910 AD 100 F 4.75 Atherosclerosis IV Moderate 1 11 Unknown USC ADRC 7.55 1.06 28.13 3.66 911 AD 88 M 7 Atherosclerosis III Moderate 1 19 Unknown USC ADRC 3.40 0.61 26.97 2.70 915 AD 86 F 7.5 V Frequent 3 Unknown Unknown USC ADRC 9.91 1.19 23.52 3.53 20195344 AD 94 F 5 Hypertension IV Moderate Unknown 11 2 years RUSH ADRC 5.17 0.93 19.74 2.96 10248033 AD 88 M 18 Hypertension III Moderate Unknown 12 2 years RUSH ADRC 0.66 0.11 30.53 3.66 11606935 AD 98 M 4 Hypertension III Moderate Unknown 18 1 years RUSH ADRC 4.31 0.78 32.30 3.88 21403995 AD 86 F 21 Hypertension III Frequent Unknown 15 3 years RUSH ADRC 2.51 0.33 30.08 3.61 20865035 AD 101 F 8.75 Hypertension IV Moderate Unknown 19 2 years RUSH ADRC 11.44 1.94 16.35 1.64 20254902 AD 94 M 6 Hypertension IV Moderate Unknown 24 2 years RUSH ADRC 9.92 1.79 23.66 1.89 20187042 AD 84 F 13 Hypertension III Moderate Unknown 22 6 months RUSH ADRC 4.94 0.74 22.75 3.41 20506587 AD 92 F 4.25 None III Moderate Unknown Unknown 2 years RUSH ADRC 2.16 0.28 40.85 6.13 11326252 AD 95 M 3.25 Hypertension IV Moderate Unknown 20 3 years RUSH ADRC 9.92 1.79 30.24 4.84 SD (%)
1b. Patients participating in a rapid brain autopsy study to establish brain endothelial cell (BEC) cultures. Patient Number Diagnosis Age Gender PMI (hr) Vascular Risk Factors Braak CERAD CDR Source 17 Control 59 F 4.5 None 0 Negative 0 Rochester ADRC 29 Control 96 F 6 None 0 Negative 0 Rochester ADRC 38 Control 58 F 5.5 None 0 Negative 0 Rochester ADRC 39 Control 72 M 4.3 Atherosclerosis Myocardial Infarction 0 I Sparse 0 Rochester ADRC 40 Control 73 M 4.7 Atherosclerosis I II Negative 0 Rochester ADRC 804 Control 82 M 9 None 0 None 0 USC ADRC 845 Control 90 F 5 Hypertension Atherosclerosis I Sparse 0 USC ADRC 877 Control 93 M 3.75 None I Sparse Unknown USC ADRC 37 AD 80 F 3.0 Atherosclerosis V VI Frequent 3 Rochester ADRC 41 AD 80 F 3.1 None V VI Frequent 2 Rochester ADRC 42 AD 80 F 5.2 None III V Moderate 2 Rochester ADRC 43 AD 77 M 2.8 None V VI Frequent 2 Rochester ADRC 49 AD 78 M 5.0 Atherosclerosis V VI Frequent 2 Rochester ADRC 799 AD 95 F 6 Atherosclerosis V Frequent 2 USC ADRC 854 AD 87 F 5.5 Atherosclerosis V Moderate 2 USC ADRC 888 AD 51 F 6 None V VI Frequent Unknown USC ADRC