Acid sphingomyelinase deactivation post-ischemia promotes brain angiogenesis and remodeling by small extracellular vesicles

Abstract

Antidepressants have been reported to enhance stroke recovery independent of the presence of depressive symptoms. They have recently been proposed to exert their mood-stabilizing actions by inhibition of acid sphingomyelinase (ASM), which catalyzes the hydrolysis of sphingomyelin to ceramide. Their restorative action post-ischemia/reperfusion (I/R) still had to be defined. Mice subjected to middle cerebral artery occlusion or cerebral microvascular endothelial cells exposed to oxygen-glucose deprivation were treated with vehicle or with the chemically and pharmacologically distinct antidepressants amitriptyline, fluoxetine or desipramine. Brain ASM activity significantly increased post-I/R, in line with elevated ceramide levels in microvessels. ASM inhibition by amitriptyline reduced ceramide levels, and increased microvascular length and branching point density in wildtype, but not sphingomyelinase phosphodiesterase-1 ([Smpd1]^-/-) (i.e., ASM-deficient) mice, as assessed by 3D light sheet microscopy. In cell culture, amitriptyline, fluoxetine, and desipramine increased endothelial tube formation, migration, VEGFR2 abundance and VEGF release. This effect was abolished by Smpd1 knockdown. Mechanistically, the promotion of angiogenesis by ASM inhibitors was mediated by small extracellular vesicles (sEVs) released from endothelial cells, which exhibited enhanced uptake in target cells. Proteomic analysis of sEVs revealed that ASM deactivation differentially regulated proteins implicated in protein export, focal adhesion, and extracellular matrix interaction. In vivo, the increased angiogenesis was accompanied by a profound brain remodeling response with increased blood-brain barrier integrity, reduced leukocyte infiltrates and increased neuronal survival. Antidepressive drugs potently boost angiogenesis in an ASM-dependent way. The release of sEVs by ASM inhibitors disclosed an elegant target, via which brain remodeling post-I/R can be amplified.

Keywords: Antidepressants; Ceramide; Exosome; Focal cerebral ischemia; Sphingomyelin; Stroke recovery.

Fig. 1

Amitriptyline inhibits ASM activity in vivo and promotes angiogenesis after I/R in an Asm dependent way. A Asm activity in the reperfused ischemic striatum (labeled I/R) and contralateral non-ischemic striatum (labeled C), measured using BODIPY-labeled sphingomyelin in wildtype mice exposed to transient MCAO, which were intraperitoneally treated with vehicle or amitriptyline (2 or 12 mg/kg b.w., b.i.d.) immediately after MCAO or with 24 h delay, followed by animal sacrifice after 24 h or after 14 days. B Ceramide content, measured by LC–MS/MS in I/R and C of wildtype MCAO mice treated with vehicle or amitriptyline for 14 days as above. C Total microvascular length, D branching point density and E mean microvascular branch length, evaluated by LSM in I/R and C of wildtype MCAO mice treated with vehicle or amitriptyline for 14 days. F Microvascular length, G branching point density and (H) mean branch length in C and I/R of Smpd1^+/+ (wildtype) and Smpd1^−/− (that is, ASM-deficient) MCAO mice treated with vehicle or amitriptyline for 14 days. Note that amitriptyline increases angiogenesis in wildtype but not Smpd1^−/− mice. Representative 3D stacks post-I/R are shown in (I), ROIs for the evaluation of microvascular networks in (J), and maximum projection images inside these ROIs in (K). Data are means ± SD values. *p ≤ 0.05/**p ≤ 0.01/***p ≤ 0.001 compared with non-ischemic C; ^†p ≤ 0.05/^††p ≤ 0.01 compared with corresponding vehicle; ^‡p ≤ 0.05/^‡‡p ≤ 0.01 compared with corresponding Smpd1^+/+ (n = 4–7 animals/group [in (A)]; n = 7–9 animals/group [in (B)]; n = 7–8 animals/group [in (C–E)]; n = 5–7 animals/group [in (F–H)]; analyzed by one-way ANOVA followed by LSD tests). Scale bars in 3D reconstructions in (I), 500 µm; in horizontal sections in (I), 1000 µm

Fig. 2

Asm inhibitor amitriptyline promotes brain remodeling after I/R in vivo when administered with 24 h delay. A CD31⁺ microvessels, B serum IgG extravasation into brain parenchyma, C infiltrating CD45⁺ leukocytes and D surviving NeuN⁺ neurons evaluated by immunohistochemistry in the reperfused ischemic striatum of C57BL/6j mice exposed to transient MCAO, which were intraperitoneally treated with vehicle or amitriptyline (2 or 12 mg/kg b.w., b.i.d.) starting 24 h after MCAO, followed by animal sacrifice after 14 days. Representative photographs are also shown. Data are means ± SD values. *p ≤ 0.05/**p ≤ 0.01/***p ≤ 0.001 compared with corresponding vehicle (n = 6–8 animals/group; analyzed by one-way ANOVA followed by LSD tests). Scale bars in (A, C), 100 µm; in (B), 1000 µm; in (D), 200 µm

Fig. 3

Amitriptyline inhibits ASM activity in vitro and reduces the intracellular accumulation of ceramide-rich vesicles after I/R. ASM activity, evaluated using BODIPY-labeled sphingomyelin in hCMEC/D3 exposed to A non-ischemic control condition (C), 3 h OGD (that is, ischemia; I), or 24 h ischemia followed by 3 h reoxygenation/glucose re-supplementation (I/R), or to B non-ischemic C, 24 h ischemia (I) or 24 h ischemia followed by 24 h reoxygenation/glucose re-supplementation (I/R), which were treated with vehicle or amitriptyline (50 µM) during 3 h [in (A)] or 24 h [in (B)]. Note that ASM activity is reduced by I and I/R. C Immunohistochemistry for ceramide in hCMEC/D3 exposed to non-ischemic C, 24 h I or 24 h/3 h I/R treated with vehicle or amitriptyline (50 µM). Note the intracellular accumulation of ceramide-rich vesicles (in green) upon I/R, which is reduced by amitriptyline (selected vesicles labeled with arrow; nuclei were counterstained in blue with Hoechst 33342). The number of vesicles evaluated by Cell Profiler is shown in (D). **p ≤ 0.01/***p ≤ 0.001 compared with corresponding vehicle; ^††p ≤ 0.01/^†††p ≤ 0.001 compared with corresponding C; ^‡‡‡p ≤ 0.001 compared with corresponding I (n = 4–5 independent samples/group; analyzed by two-way ANOVA followed by LSD tests). Scale bar in (C), 10 µm

Fig. 4

Amitriptyline, fluoxetine and desipramine promote cerebral angiogenesis in vitro in an ASM dependent way. A–C Matrigel-based tube formation evaluated for the number of closed tubes, microvascular length and branching point density, D transwell migration, E, F VEGFR2 abundance examined by Western blot and G, H VEGF concentration in supernatants measured by enzyme-linked immunosorbent assay (ELISA) of hCMEC/D3 exposed to vehicle or amitriptyline (0–50 µM). In (F, H), analyses were made after 4 and 24 h amitriptyline exposure, respectively. I Tube formation and J transwell migration of hCMEC/D3 exposed to vehicle or fluoxetine (0–20 µM). K Tube formation and L transwell migration of hCMEC/D3 exposed to vehicle or desipramine (0–50 µM). Note that all three ASM inhibitors increase angiogenesis. M Tube formation, N transwell migration, O VEGFR2 abundance and P VEGF concentration in supernatants of hCMEC/D3 transfected with scrambled siRNA (used as control) or SMPD1 siRNA which were exposed to vehicle or amitriptyline (50 µM). In (O, P), measurements were made after 4 and 24 h amitriptyline exposure, respectively. Data are means ± SD values. *p ≤ 0.05/**p ≤ 0.01/***p ≤ 0.001 compared with corresponding vehicle; ^‡p ≤ 0.05/^‡‡p ≤ 0.01/^‡‡‡p ≤ 0.001 compared with corresponding scrambled siRNA (n = 3–7 independent samples/group [in (A–L)]; n = 5–8 independent samples/group [in (M,N,O)]; n = 3 independent samples/group [in (P)]; analyzed by one-way ANOVA [in (A–D, G–L)] or two-way ANOVA [in (M, N, P)] followed by LSD tests [in (A–D, G–N, P)] or paired two-tailed t tests [in (E, F, O)])

Fig. 5

Intracellular ceramide-rich vesicles express markers of late endosomes and multivesicular bodies. Immunocytochemistry for A ceramide (in green) and the late endosome marker Rab7 (in magenta) and B ceramide (in green) and the multivesicular body marker CD63 (in magenta) of hCMEC/D3 exposed to 24 h ischemia followed by 3 h reoxygenation/glucose re-supplementation (I/R). In the merged photographs, double labeled cells are shown in white (selected cells labeled with arrow; nuclei were counterstained in blue with Hoechst 33342). Scale bar in overview photograph, 10 µm; in magnification, 5 µm. Data are representative for 3 independent studies

Fig. 6

Amitriptyline promotes the extracellular release of vesicles with immunofluorescence characteristics of exosomes. Concentration of A, C CD9⁺ and B, D CD63⁺ sEVs in the supernatant of hCMEC/D3 exposed to 3 h C, 3 h I or 24 h/3 h I/R [in (A, B)] or 24 h C, 24 h I or 24 h/24 h I/R [in (C, D)] treated with vehicle or amitriptyline (50 µM). sEV concentration was evaluated by AMNIS flow cytometry. Note that the number of CD9⁺ and CD63⁺ sEVs, which is elevated upon I/R, further increases by amitriptyline. Data are means ± SD values. *p ≤ 0.05/**p ≤ 0.01 compared with corresponding vehicle; ^†p ≤ 0.05/^††p ≤ 0.01/^†††p ≤ 0.001 compared with corresponding C; ^‡p ≤ 0.05/^‡‡p ≤ 0.01 compared with corresponding I (n = 4–5 independent samples/group [in (A, B)]; n = 6–9 independent samples/group [in (C, D)]; analyzed by two-way ANOVA followed by paired two-tailed t tests)

Fig. 7

sEVs released by amitriptyline have angiogenic activity and exhibit enhanced uptake by endothelial cells. A–C Matrigel-based tube formation, D transwell migration and E VEGF release of hCMEC/D3, which were treated with vehicle, amitriptyline (Ami, 50 µM) or sEV preparations (25 µg protein/ml) isolated from supernatants of hCMEC/D3 that had been cultured in non-ischemic control condition (C), ischemia (I) or ischemia followed by reoxygenation/glucose re-supplementation (I/R) and had been treated with vehicle or amitriptyline (50 µM) during cell cultivation. VEGF release was determined by ELISA. Representative tube formation and migration assays are shown in Suppl. Fig. 27. F hCMEC/D3 uptake of sEV preparations obtained from hCMEC/D3 treated with vehicle or amitriptyline (50 µM) evaluated by PKH67 dye. Representative photographs of hCMEC/D3 exhibiting sEV uptake (in green) are depicted. Nuclei were counterstained in blue with Hoechst 33342. Data are means ± SD values. *p ≤ 0.05/**p ≤ 0.01 compared with corresponding vehicle (n = 4 independent samples/group [in (A–C)]; n = 3 independent samples/group [in (D)]; n = 3–4 independent samples/group [in (E)]); n = 5 independent samples/group [in (F)]; analyzed by one-way ANOVA followed by LSD tests [in (A–E)], or two-tailed t tests [in (F)]). Scale bar in (F), 10 µm

Fig. 8

sEVs taken up by endothelial cells accumulate in late endosomal compartment. hCMEC/D3 exhibiting accumulation of sEV particles labeled with PKH67 (in green) in intracellular vesicles expressing A the late endosome marker Rab7, B the multivesicular body marker CD63 and C the lysosome marker LAMP1 (in magenta). sEV preparations had been obtained from supernatants of hCMEC/D3 treated with amitriptyline (50 µM). Nuclei were counterstained in blue with Hoechst 33342. Scale bar in overview photograph, 10 µm; in magnification, 2 µm. Data are representative of 5 independent studies