SUPPLEMENTARY INFORMATION from reactive astrocytes impairs memory in mouse models of Alzheimer disease Seonmi Jo *, Oleg Yarishkin *, Yu Jin Hwang, Ye Eun Chun, Mijeong Park, Dong Ho Woo, Jin Young Bae, Taekeun Kim, Jaekwang Lee, Heejung Chun, Hyun Jung Park, Da Yong Lee, Jinpyo Hong, Hye Yun Kim, Soo-Jin Oh, Seung Ju Park, Hyo Lee, Bo-Eun Yoon, YoungSoo Kim, Yong Jeong, Insop Shim, Yong Chul Bae, Jeiwon Cho, Neil W. Kowall, Hoon Ryu, Eunmi Hwang, Daesoo Kim and C. Justin Lee Nature Medicine doi:1.138/nm.3639
Kainic acid Control Number of plaques a Plaque 2 months 4 months 6 months b 2 SFig 1 15 1 8 months 11 months 23 months 5 All 2 4 6 8 11 23 3 mm WT Age (month) c Plaque / / Iba1 / DAPI Plaque Iba1 Merge + DAPI d Plaque Iba1 Merge + DAPI 2 mm 5 mm e DG CA1 CA3 CA3 PI DAPI 5 mm 1 mm Nature Medicine doi:1.138/nm.3639
Supplementary Figure 1. Amyloid plaques, reactive astrocytes, and a lack of neuronal death. (a) Thioflavin-S staining of amyloid plaques in the hippocampus. mice start to develop amyloid plaques at the age of 4 months and develop more with aging. (b) Average number of plaques as a function of age (a). The number sigmoidally increases with age while wild-type mice show no plaque at any ages tested. WT: wild-type. 6 slices were analyzed for each group. Data are means SEM (error bars). (c) Confocal images show numerous reactive astrocytes and activated microglia clustered around amyloid plaques in mice at 9 months of age. : glial fibrillary acidic protein, Iba1: ionized calcium binding adaptor molecule-1. DAPI: 4', 6-diamidino-2- phenylindole. (d) High-magnification images of the molecular layer of DG from white box in (c). (e) Confocal images of propidium iodide (PI) and DAPI staining show no apparent cell death in the hippocampus of both wild-type and mice (11 12 months of age). For positive control, wild-type mice were injected with kainic acid (25 mg kg -1, i.p.) 48 h before PI injection. CA: Ammon s horn of hippocampus (Cornu Ammonis), DG: dentate gyrus. Nature Medicine doi:1.138/nm.3639
/ Iba1 / GFAP / Plaque / / / DAPI a b SFig 2 2 months 23 months 2 months 4 months 6 months 8 months 11 months 23 months Plaque Plaque Merge + DAPI Merge + DAPI 2 mm 2 mm c d 5XFAD Iba1 MoDG GrDG Merge Iba1 2 mm 3 mm Supplementary Figure 2. in the reactive astrocytes of and 5XFAD mice. (a) Representative confocal images of thioflavin-s staining, and immunostaining in the molecular layer of DG. mice didn t show amyloid plaques Nature Medicine doi:1.138/nm.3639
and -containing reactive astrocytes even at the age of 23 months. -negative and -positive staining indicates ergic interneurons. (b) Representative confocal images of mice at various ages. mice start to develop amyloid plaques at the age of 4 months. Although the exact level of and varies among amyloid plaques, -positive reactive astrocytes are consistently observed at all ages with amyloid plaques. (c) Representative confocal images of, Iba1, and immunostaining in the molecular layer of DG. (12 months of age) show that is not colocalized with Iba1 but co-localized with. (d) Representative confocal images of 14-month-old 5XFAD mice. is accumulated in reactive astrocytes of 5XFAD mice. Nature Medicine doi:1.138/nm.3639
Intensity of Gat3 in + pixels (AU) Intensity of Gat1 in + pixels (AU) Intensity of Gad67 in + pixels (AU) Intensity of Gad65 in + pixels (AU) Intensity of Maoa in + pixels (AU) Intensity of Abat in + pixels (AU) a / Gad67 / Gad67 Merge b / Gad65 / Gad65 Merge 1 mm 1 mm c / Maoa / d / Abat / Maoa Merge Abat Merge 1 mm 1 mm e f g h Gad67 in astrocytes Gad65 in astrocytes Maoa in astrocytes Abat in astrocytes 15 15 12 24 1 1 8 16 5 7 8 5 7 1 4 5 5 8 18 13 i / Gat1 / / DAPI Gat1 Merge + DAPI 6 Gat1 in astrocytes 4 2 2 mm 1 8 WT j / Gat3 / / DAPI Gat3 Merge + DAPI 6 Gat3 in astrocytes 4 2 2 mm 1 9 WT Nature Medicine doi:1.138/nm.3639
Supplementary Figure 3. Gad67, Gad65, Maoa, Abat, Gat1, and Gat3 expression in reactive astrocytes. (a d) Expression of Gad67 (a), Gad65 (b), Maoa (c), and Abat (d) in the molecular layer of DG was tested by immunostaining. Each enzyme is minimally expressed in astrocytes and not changed in -containing reactive astrocytes near the amyloid plaque (9 months of age). (e h) Quantification of confocal images. Mean intensity of Gad67 (e), Gad65 (f), Maoa (g), and Abat (h) was measured in -positive pixels. P >.5 (t-test). (i j) Expression of Gat1 (i) and Gat3 (j) in the molecular layer of DG was tested by immunostaining. Mean intensity of Gat1 (i) or Gat3 (j) was measured in -positive pixels. Expression of both transporters is not changed in -containing reactive astrocytes near the amyloid plaque (9 months of age). P >.5 (t-test). Gad: glutamate decarboxylase. Maoa: monoamine oxidase A, Abat: transaminase, Gat: transporter, WT: wild-type. AU: arbitrary unit. Number on each bar refers to the number of cells analyzed. Data are means ± SEM (error bars). Nature Medicine doi:1.138/nm.3639
Maob activity (mol g -1 hr -1 ) Maoa activity (mol g -1 hr -1 ) / Maob / Maob / Intensity of putrescine in + pixels (AU) Pearson s correlation Iba1 (-GFP) GFP a / Putrescine / / DAPI Putrescine Merge + DAPI b Putrescine SFig 4 Merge MoDG -GFP Putrescine Merge 2 mm GrDG Putrescine Merge 5 mm Iba1 Putrescine Merge e c 2 15 1 5 SAT Putrescine ** 1 1 WT Mono-acetyl putrescine MAOB d.4.3.2.1 N-acetyl-γaminobutyrate N-acetyl-γaminobutyraldehyde 6 6 -GFP WT ALDH2 6 Iba1 Unknown Acetyl-CoA Coenzyme A, H + O 2, H 2 O H 2 O 2, NH 4 + NAD +, H 2 O 2H +, NADH H 2 O Acetate f g Maob Maob Merge Merge 5 mm 1 mm h i.8 *** j.3 Maob Maoa Water Selegiline.6.4.2 4 4 Water Sele.2.1 4 4 Water Sele Nature Medicine doi:1.138/nm.3639
Supplementary Figure 4. Putrescine, Maob expression, and Maob enzyme activity. (a) Presence of putrescine, a precursor, was detected by immunostaining. Representative confocal images show global existence of putrescine in the DG molecular layer including -positive areas. Putrescine is increased in mice, and most intense staining is observed around amyloid plaques (arrowheads) and near the containing reactive astrocytes. In both genotypes, putrescine is absent in neurons which are clustered at the granule cell layer (1 months of age). (b) Confocal images show that intense staining of putrescine is well colocalized with Iba1- positive microglia in mice (1 months of age). (c) Quantification of confocal images. Mean intensity of putrescine was measured in positive pixels. ** P <.1 (t-test). (d) Pearson s correlation coefficients show that putrescine is colocalized with Iba1 better than or -GFP. (e) Putrescine degradation III pathway for production. MAOB mediates the second reaction step. SAT1: putrescine acetyltransferase, ALDH2: mitochondrial aldehyde dehydrogenase. (f g) Confocal images representing labeling with two commercially available antibodies, which are directed against different epitopes, show increased expression of Maob in containing reactive astrocytes of mice (1 months of age). Antibodies were purchased from Santa Cruz Inc.(f) and Sigma Inc.(g). (h) Colorimetric enzymatic activity assay of Maob. Maob activity in the hippocampus is selectively inhibited by drinking water administration of selegiline (5 1 mg kg -1 day -1 for 3 days) in wild-type mice (13 months of age). (i) Maob activity after selegiline administration. Sele: selegiline. *** P <.1 (t-test). (j) Maoa activity after selegiline administration. n.s. P >.5 (t-test). AU: arbitrary unit. Number on each bar refers to the number of hippocampi analyzed. Data are means ± SEM (error bars). Nature Medicine doi:1.138/nm.3639
Amplitude (pa).6.4.2 1 pa 5 ms b-actin.2 + TTX + TTX 2 * 1 11 11 2, 4, 6, sepsc interevent interval (ms) * 6 3 1 1 1..8.6.4.2 4 2 1 2 3 4 Stimulus intensity (ma) + TTX + TTX 1.5 1..5 1 1 2, 4, 6, mepsc interevent interval (ms) i + control + selegiline + control + selegiline 6 8 ** 6 4 2 9 9 1 1 Cont Sele Cont Sele 5 Selegiline 1.5 12 12 1.5 1.6 1.2.8.4 3 Water.8 3 Sele Ab oligomer normalized to b-actin Water Tetramer Trimer Dimer Monomer 8 Control Selegiline happ.4 5 1 15 2 25 3 mepsc peak amplitude (pa) h 12.6 eepsc amplitude (pa) Stim 12.8 5 1 15 2 25 3 sepsc peak amplitude (pa).8 PP Age (month) 1. Patch j 11 Cumulative probability 5 pa 5 ms Control Selegiline 11 1. f Cumulative probability + TTX 3 e + TTX +.2 mepsc g.4 ** 6 Ab monomer normalized to b-actin d.6 Frequency (Hz) 5 pa 5 ms.8 Cumulative probability 1. Amplitude (pa) Cumulative probability SFig 5 c Frequency (Hz) b sepsc eepsc amplitude (pa) a.6.4.2 3 3 Water Sele Supplementary Figure 5. Spontaneous EPSCs (sepsc), miniature EPSCs (mepsc), input-output relationship, and Ab species in DG of wild-type and mice. (a c) sepscs recorded from granule cells of DG in wild-type and mice (N = 2 for each group; both sexes at 12 12.5 months of age). (a) Representative traces of sepscs. (b) Cumulative probability of sepsc peak amplitude. Inset bar graph: Average amplitude of sepscs. **P <.1 (t-test). (c) Cumulative probability of sepsc interevent interval. Inset bar graph: Average frequency of sepscs. * P <.5 (t-test). Nature Medicine doi:1.138/nm.3639
(d f) mepscs recorded from granule cells of DG in wild-type and mice (N = 2 for each group; both sexes at 12 12.5 months of age). (d) Representative traces of mepscs. (e) Cumulative probability of mepsc peak amplitude. Inset bar graph: Average amplitude of mepscs. * P <.5 (t-test). (f) Cumulative probability of mepsc interevent interval. Inset bar graph: Average frequency of mepscs. P >.5 (t-test). (g i) Input-output relationship of granule cells of DG in wild-type and mice with or without pretreatment of selegiline (1 mm) (N = 2 for each group; both sexes at 12 12.5 months of age). (g) Schematic diagram for the input-output relationship experiments and representative traces of eepscs evoked by stimuli of 3 ma intensity. (h) Relationships between stimulus intensity and eepsc amplitude. ** P <.1 for genotype, P >.5 for drug (Two-way repeated measures ANOVA). *** P <.1 for + selegiline at 5 ms (One-way ANOVA and bonferroni test). (i) Mean amplitudes of eepscs evoked by stimuli of 3 ma intensity. ** P <.1 for genotype, P >.5 for drug (Two-way ANOVA). (j) Effect of selegiline on the level of Ab oligomer and monomer in mice. Western blots of Ab in DG using 6E1 antibody (N = 3 for each group; male at 1.5 12 months of age) revealed that the level of Ab monomer and oligomer are not significantly altered by oral administration of selegiline for 1 week. P >.5 (t-test). EPSC: excitatory post-synaptic current. Number on each bar refers to the number of cells (b, c, e, f, i) and mice (j) analyzed. N refers to the number of animals studied. Data are means (c, f, h, i, j) or medians (b, e) ± SEM (error bars). Nature Medicine doi:1.138/nm.3639
Spike probability Spike probability Spike probability a b SFig 6 5 mv 5 mv 5 ms 5 ms Before + BIC, CGP Before + BIC, CGP c + water d + selegiline 1. 1. Probability max.8.8.6.6.4.4.2 Before (n = 12) + BIC, CGP (n = 12).2 Before (n = 15) + BIC, CGP (n = 15) 2 4 6 8 1, Stimulus intensity (ma) ES 5 2 4 6 8 1, Stimulus intensity (ma) e f 1..8 5 mv.6 5 ms.4 Control Selegiline.2 Control (n = 5) Selegiline (n = 3) 2 4 6 8 1, Stimulus intensity (ma) Supplementary Figure 6. Effect of selegiline on spike probability of dentate granule cells in wild-type and mice. (a d) Selegiline has no apparent effect on spike probability in wild-type littermates (N = 2 for WT + water; N = 3 for WT + selegiline; both sexes at 12 13 months of age). (a) Example traces of evoked EPSP and action potential in wild-type mice at 3 ma stimulation, before and after bath application of receptor antagonists (BIC: 1 mm bicuculline. CGP: 5 mm CGP55845). (b) Example traces in selegiline-administrated wild-type mice (5 1 mg kg -1 day -1, oral administration for 7 days) at 3 ma stimulation, before and after bath application of receptor antagonists. (c) Summary graph of spike probability in wild-type mice versus stimulus intensity. (d) Summary graph of spike probability in selegiline-administered wild-type mice versus stimulus intensity. Nature Medicine doi:1.138/nm.3639
(e f) Incubation of slices with selegiline (1 mm, 2 h) has a similar effect with its oral administration in mice (13 months of age). (e) Example traces in mice at 3 ma stimulation, incubation of slices with or without selegiline. (f) Summary graph of spike probability in mice versus stimulus intensity. EPSP: excitatory post-synaptic potential, Probability max : maximum probability, ES 5 : 5% effective stimulus. n refers to the number of cells analyzed. N refers to the number of animals studied. Data are means ± SEM (error bars). Nature Medicine doi:1.138/nm.3639
Time spent in quadrants (%) Target crossing number Swim speed (cm s -1 ) Time spent in quadrants (%) Target crossing number Swim speed (cm s -1 ) Escape latency (s) Swim speed (cm s -1 ) a Day 1 Handling Day 15 Day 23 Day 29 b SFig 7 Spatial cues Hidden platform Releasing points c 6 5 Acquisition d 3 Acquisition 4 2 3 2 1 + water + selegiline + water + selegiline 1 2 3 4 5 6 7 8 9 1 11 12 13 14 Training day * *** * 1 + water + selegiline + water + selegiline 1 2 3 4 5 6 7 8 9 1 11 12 13 14 Training day e 5 4 3 f 1 st probe test + water + selegiline 6 1 st probe test 3 1 st probe test + water + selegiline 4 g 2 2 2 1 1 8 4 7 4 Target Right Left Opposite 8 4 7 4 Water Sele Water Sele 8 4 7 4 Water Sele Water Sele h 5 4 3 i 2 nd probe test + water + selegiline 6 2 nd probe test 3 2 nd probe test + water + selegiline 4 j 2 2 2 1 1 8 4 7 4 Target Right Left Opposite 8 4 7 4 Water Sele Water Sele 8 4 7 4 Water Sele Water Sele Supplementary Figure 7. Morris water maze experiment for testing learning and memory in wild-type and mice treated or non-treated with selegiline. (a b) Experimental protocol for Morris water maze test for wild-type and mice with or without oral administration of selegiline (both sexes at 1 12 months of age). After selegiline administration (1 mg kg -1 day -1 for 14 days), mice were released in a Morris water maze with a hidden platform and spatial cues. Acquisition is tested by 4 trials day -1 for 14 days. Probe test was done on the 23 th and 29 th day in the absence of the platform. (c) Escape latency during acquisition test. Nature Medicine doi:1.138/nm.3639
(d) Swimming speed during acquisition test. (e) Time spent in each quadrant during 1 st probe test. (f) Target crossing number during 1 st probe test. (g) Swimming speed during 1 st probe test. (h) Time spent in each quadrant during 2 nd probe test. (i) Target crossing number during 2 nd probe test. (j) Swimming speed during 2 nd probe test. For escape latency, two-way repeated measures ANOVA followed by Bonferroni s post hoc analysis revealed significant effects for the genotype [F(1,88) = 35.977, P =.], no drug effect, and significant interaction effect between genotype and drug [F(1,88) = 9.411, P =.3]. ** P <.1, *** P <.1, P >.5. Difference in swim speeds and probe test results among groups are not significant. Number on each bar refers to the number of mice (e j) analyzed. Data are means ± SEM (error bars). Nature Medicine doi:1.138/nm.3639
a (a) Ab plaque Pre SFig 8 release A R Cl (b) Putrescine B R (c) MAOB (d) Synapse (e) BEST1 (f) R Reactive astrocyte Post b from reactive astrocytes Decrease in presynaptic release (PP-DG synapse) Impairment in spike probability and synaptic plasticity Learning and memory impairment Supplementary Figure 8. Model diagram of memory impairment in AD. In AD, astrocytes near Ab plaques (a) have more putrescine (b). Putrescine is degraded by MAOB (c) to produce the inhibitory neurotransmitter (d). is then abnormally released via BEST1 (e) which is redistributed away from microdomains. Then the released binds to extrasynaptic A and B receptors (f) and strongly inhibits presynaptic release and spike probability. Consequently, granule cells of the dentate gyrus receive less glutamatergic inputs at perforant path synapses and show reduced synaptic plasticity. This finally leads to memory impairment in Alzheimer disease. Pre: presynaptic terminal, Post: postsynapse, NMDAR: N-methyl-D-aspartate receptor, AMPAR: -amino-3-hydroxy-5-methyl-4-isoxazole propionate receptor. Nature Medicine doi:1.138/nm.3639
Intensity of neuronal (AU) Intensity of In + pixels (AU) Intensity of (AU) a SFig 9 b Contralateral Stab wound injury Contralateral Stab wound injury CA1 Py c 4 3 * SR 2 d 1 3, 2, 6 5 Cont Injury *** NeuN 1, 6 5 Cont Injury e 15 Merge 1 5 4 25 Cont Injury 3 mm Supplementary Figure 9. in the reactive astrocytes which is induced by stab wound injury in the CA1 of hippocampus. (a) Target site for stab wound injury. Stab wounds were stereotaxically introduced into the hippocampus (blue arrow). CA1 region adjacent to the injured site (red box) and contralateral non-injured site was observed (N = 3 for each group; males at 1 weeks of age) (b) Representative confocal images for stab wound injury and the contralateral site. (c) Mean intensity of. * P <.5. (d) Mean intensity of in -positive area. *** P <.1. (e) Mean intensity of interneuronal. P >.5. AU: arbitrary unit. Number on each bar refers to the number of slices (c, d) or cells (e) analyzed. N refers to the number of animals studied. Data are means ± SEM (error bars). Nature Medicine doi:1.138/nm.3639
Supplementary Table 1. 5% effective stimulus (ES 5 ), maximum spike probability (Probability max ), resting membrane potential (RMP), and input resistance (R input ) of granule cells (n refers to the number of granule cells recorded). Genotype Oral administration Bath application n ES 5 (ma) ± SEM Probability max ± SEM Water 18 229.482 ± 1.47.989 ±.8 Selegiline 15 197.288 ±.752.993 ±.7 Water 21 415.956 ± 6.651.638 ±.98 Selegiline 23 227.438 ± 4.795.974 ±.18 Water before 8 22.853 ± 7.375.983 ±.11 +BIC,CGP 8 159.374 ± 1.625 1. ±. Selegiline before 15 231.155 ± 4.544.993 ±.7 +BIC,CGP 15 187.384 ± 7.39 1. ±. Water before 15 415.93 ± 4.57.713 ±.112 Genotype +BIC,CGP 15 269.59 ± 4.171 1. ±. Selegiline before 11 26.27 ±.874.973 ±.27 Oral administration +BIC,CGP 11 177.49 ±.8 1. ±. Bath application RMP (mv) ± SEM R Input (MW) ± SEM Water 29-79.379 ± 1.13 284.724 ± 19.25 Selegiline 18-81.5 ±.85 295.444 ± 32.218 Water 21-79.238 ± 1.494 315.381 ± 24.593 Selegiline 23-8.652 ±.732 323.522 ± 18.745 Nature Medicine doi:1.138/nm.3639
Supplementary Table 2. Detailed information for statistical analysis of behavioral test results. Behavior tests Passive avoidance test Morris water maze test Result from statistical analysis Two-way repeated measures ANOVA (at the.5 level) Genotype effect: F(1,42) = 18.63, P =. Drug effect: F(1,42) = 12.494, P =.1 Genotype Drug interaction: not significant One-way repeated measures ANOVA F(3,42) = 14.6, P =. Scheffe's post hoc analysis: WT + water vs. + water, P =. WT + selegiline vs. + water, P =. +water vs. + selegiline, P =.23 One-way ANOVA with Scheffe's post hoc analysis (for each day) Day 1 : WT + water vs. WT + selegiline, P =.4 WT + selegiline vs. + water, P =.42 WT + selegiline vs. + selegiline, P =.43 Day 2 : WT + water vs. + water, P =. WT + selegiline vs. + water, P =. + water vs. + selegiline, P =.13 Two-way repeated measures ANOVA Genotype effect: F(1,88) = 35.977, P =. Drug effect: not significant Genotype Drug interaction: F(1,88) = 9.411, P =.3 One-way repeated measures ANOVA F(3, 88) = 21.468, P =. Scheffe's post hoc analysis: WT + water vs. + water, P =. WT + water vs. + selegiline, P =.34 WT + selegiline vs. + water, P =. + water vs. + selegiline, P =.12 One-way ANOVA with Scheffe's post hoc analysis (for each day) Day 1 6: not significant Day 7: WT + water vs. + water, P =.48 Day 8: WT + water vs. + water, P =.33 Day 9: WT + water vs. + water, P =.15 Day 1: WT + water vs. + water, P =.13 Day 11: WT + water vs. + water, P =.42 Day 12: WT + water vs. + water, P =.4 Day 13: WT + water vs. + water, P =.4 Day 14: not significant Nature Medicine doi:1.138/nm.3639
Supplementary Table 3. Human tissue information for control subjects and subjects with Alzheimer disease. Case Age Sex Braak stage Cause of death Control 1 87 F I Unknown Control 2 88 M I Adenocarcinoma Control 3 86 M II Unknown Control 4 87 F II Unknown Control 5 67 M I Sudden accidental death Control 6 82 M I Colon cancer Control 7 61 M I Unknown Control 8 11 F I Myocardial infarction Control 9 89 M III Renal failure Control 1 68 M I Congestive heart failure Control 11 78 F I Unknown AD 1 82 M V AD AD 2 79 F VI AD AD 3 7 M VI AD AD 4 59 M VI AD AD 5 8 F V AD AD 6 92 M V AD AD 7 9 F V AD AD 8 1 M V AD AD 9 75 M V AD AD 1 83 M VI AD AD 11 79 F VI AD Nature Medicine doi:1.138/nm.3639