Circulating endothelial cell-derived extracellular vesicles mediate the acute phase response and sickness behaviour associated with CNS inflammation

Abstract

Brain injury elicits a systemic acute-phase response (APR), which is responsible for co-ordinating the peripheral immunological response to injury. To date, the mechanisms responsible for signalling the presence of injury or disease to selectively activate responses in distant organs were unclear. Circulating endogenous extracellular vesicles (EVs) are increased after brain injury and have the potential to carry targeted injury signals around the body. Here, we examined the potential of EVs, isolated from rats after focal inflammatory brain lesions using IL-1β, to activate a systemic APR in recipient naïve rats, as well as the behavioural consequences of EV transfer. Focal brain lesions increased EV release, and, following isolation and transfer, the EVs were sequestered by the liver where they initiated an APR. Transfer of blood-borne EVs from brain-injured animals was also enough to suppress exploratory behaviours in recipient naïve animals. EVs derived from brain endothelial cell cultures treated with IL-1β also activated an APR and altered behaviour in recipient animals. These experiments reveal that inflammation-induced circulating EVs derived from endothelial cells are able to initiate the APR to brain injury and are sufficient to generate the associated sickness behaviours, and are the first demonstration that EVs are capable of modifying behavioural responses.

Figure 1

Extracellular vesicle (EV) release in vivo and in vitro after an IL-1β challenge. Animals received a single injection of IL-1β (10ng in 1 µl into CNS parenchyma within the striatum) or vehicle (saline) and blood was taken at 2 hours and analysed for EV number and size using Nanosight Tracking Analysis. Distribution in vehicle and IL-1β treated animals demonstrated a mixed population of vesicles (a - graph; EM inset). This was quantified as EVs × 10¹⁰/ml by NTA for vesicles <200 nm (b) as well as for total vesicles (c). Average size of the EV population in nm was also determined (d). Brain vascular endothelial cells (GP-8) were treated with IL-1β (10ng/ml) for 2 hours and supernatant was harvested and analysed for EV size and number using NTA. Distribution in vehicle and IL-1β treated cells demonstrated a mixed population of vesicles (e - graph; EM inset). This was quantified as EVs × 10⁶/ml by NTA for vesicles <200 nm (f), as well as for total vesicles (g). Average size of the EV population in nm was also determined (h). Data are mean ± SEM, n = 6, *p < 0.05; **p < 0.01 and ***p < 0.001. Scale bars on micrographs show 0.2 µm and 50 nm (inset).

Figure 2

Schematic overview of in vivo experiments (a). Animals received a single intracranial injection of IL-1β or vehicle (saline) and blood was collected and processed for EV isloation. Platelet free plasma was separated using centrifugation and EVs further separated by ultracentrifugation. Samples from 3 ‘donor’ animals were pooled and administered to each ‘recipient’ animal. The hepatic response to EV transfer of in vivo and in vitro-derived EVs (b–e): EVs or EV-free supernatant were transferred from donor animals challenged with IL-1β or saline into naive recipient animals and the hepatic response was studied after 4 hours. Specifically, expression of TNF mRNA in naïve animals receiving EV-free supernatant (b); EVs isolated from whole blood (c) or EVs from GP-8 cell supernatant (d). CXCL-1 mRNA in animals receiving EV-free supernatant (e); EVs isolated from whole blood (f) or EVs from GP-8 cell supernatant (g). All qPCR data are normalized to the expression of the housekeeping gene GAPDH and then further normalized to control animals. Data are mean ± SEM, n = 6, *p < 0.05 and ***p < 0.001.

Figure 3

Hepatic neutrophil numbers were counted in animals receiving EV-free supernatant (a); EVs isolated from whole blood (b) or EVs from GP-8 cell supernatant (c). Representative microscopy of PKH-67-labelled EVs (green) introduced into naïve animals tracked to the liver (d) where they were found largely in CD11b-positive (red) Kupffer cell populations. An administration of PKH-67 intravenously does not result in any staining (e). Scale bars on micrographs represent 40 µm.

Figure 4

The hepatic response to EV transfer of in vitro-derived EVs in clodronate-treated animals. (a) TNF mRNA expression in the liver of animals receiving EVs isolated from IL-1β treated GP-8 cell supernatant in the presence or absence of clodronate-filled liposomes (lipo) or denatured EVs (EV den). (b) CXCL-1 mRNA expression in the liver of animals receiving EVs isolated from IL-1β treated GP-8 cell supernatant in the presence or absence of clodronate-filled liposomes. All qPCR data are normalized to the expression of the housekeeping gene GAPDH and then further normalized to control animals. Data are mean ± SEM, n = 4; ^‡compared to saline/liposome treated animals; *compared to IL-1β treated animals.

Figure 5

The behavioural response to IL-1β-stimulated, in vivo-derived circulating EVs and in vitro-derived brain endothelial EVs. Number of rears and number of squares crossed in the open field are shown for rats given in vivo-derived EVs or EV-free plamsa (a–d), and also for EVs from IL-1β and saline-treated endothelial cells (e,f). Data are mean ± SEM, n = 6, *p < 0.05 and **p < 0.01.