Liver sinusoidal endothelial cells (LSECs) line the specialized liver vasculature. They have no basement membrane and present open fenestrae without diaphragm, creating a permeable and dynamic barrier between the blood and other liver cell types [1]. Situated at the interface between portal blood containing molecular signals from the pancreas, adipose tissue and gut, and hepatocytes, hepatic stellate cells (HSCs) and Kupffer cells, LSECs form a platform integrating these signals and transmitting them to neighboring cells [1]. Specifically, in physiological conditions, LSECs have anti-fibrotic properties, with a specific ability to inhibit the differentiation of hepatic stellate cells into extracellular matrix-depositing myofibroblasts, in part thanks to their release of nitric oxide [2].

The ability of LSECs to communicate to the environing cells is thus of paramount importance to their role. Classical means of cell communication are represented by cell junctions, adhesion contacts, and soluble factors, such as cytokines. Another mechanism of cell communication has recently emerged, namely communication by extracellular vesicles (EVs). EVs are vesicles enclosed by a double membrane emitted by all cells. In the liver, following exposure to detrimental stimuli, including alcohol, virus or metabolic stresses, various liver cells release EVs, with a changed cargo, and these EVs contribute to the pathophysiology of liver diseases [3]. Yet, the release and role of EV in physiological conditions has been mostly overlooked. So far, the role of EVs emitted by LSECs has been barely studied, either in physiological or pathological conditions.

In the current issue of Hepatology International, Chen and colleagues investigated the role of LSECs in liver fibrosis [4]. Several of their findings deserve comments.

First, Chen and colleagues unraveled a new way of communication between healthy LSECs and HSCs [4]. Using a rat model of aldosterone-induced liver fibrosis and in vitro studies culturing either primary LSECs or immortalized cell lines, as models of LSECs, with HSCs, they demonstrated that LSECs release EVs inhibiting activation and migration of HSCs in physiological conditions. These results are a step forward in our understanding of the mechanisms by which healthy LSECs prevent HSCs activation and fibrosis.

Second, authors showed an over-activation of autophagy in LSECs following aldosterone treatment both in rats with fibrosis induced by aldosterone and in vitro. Ruart and colleagues also observed an increase of autophagic flux in primary LSECs isolated from rats treated for 4 to 6 weeks with carbon tetrachloride injections [5]. They concluded that upregulation of endothelial autophagy may play a role in the adaptive response at early stages of liver injury. Conversely, our group observed a lower autophagy level in LSECs from 12 patients with NASH, half of them with liver fibrosis, as compared with 11 individuals with a normal liver at histological examination or with simple steatosis [6]. We showed that low concentrations of TNFα and IL-6, similar to those observed in the portal vein of patients with metabolic syndrome, could reproduce this effect in vitro. Altogether, these results suggest that autophagic response in LSECs varies according to the stimulus used. In this regard, the relevance to the human situation of high doses of aldosterone is questionable. Indeed, primary hyperaldosteronism is not a recognized cause of liver fibrosis and cirrhosis in patients. The impact of various stimuli leading to LSEC capillarization, such as metabolic stresses, alcohol or virus on LSEC-derived EVs numbers, contents and capacity to regulate HSCs should be investigated. An analysis of autophagy level in a large number of liver biopsies from patients, encompassing various causes of liver disease and several severity stages, would help reconcile those discrepancies.

Third, Chen and colleagues observed a protective effect of autophagy knock-down in LSECs against liver fibrosis provoked by continuous infusion of aldosterone via osmotic micro-pumps into rats [4]. Autophagy was inhibited by injecting rats with a si-ATG5 adeno-associated virus (AAV). These results are at variance of previously published results. Indeed, Ruart and colleagues, as well as our team, found opposite effects, namely more liver fibrosis in mice deficient in endothelial autophagy. Ruart and colleagues used transgenic Atg7lox/lox;VE-cadherin-Cre mice treated with carbon tetrachloride while we used Atg5lox/lox;VE-cadherin-Cre mice fed a high fat diet [5, 6]. Altogether, these observations suggest that autophagy level is finely regulated in LSECs. Whether the detrimental effect of LSEC autophagy is restricted to the model of aldosterone infusion or is more common will require future studies. More generally, the level of autophagic flux in the endothelium in patients with liver diseases warrants further research, taking into account the severity, etiology, nutritional status and other clinical parameters.

Fourth, Chen and colleagues delineated in LSECs the link between EV biogenesis and autophagy previously demonstrated in other cell types [7]. They also assessed EV protein and RNA composition. This first step where should be followed by more detailed studies looking at the influence of autophagy on EV function, both in physiological conditions and in the context of various liver diseases. These studies should also investigate the link between autophagy and EV release in LSECs cultured under shear stress conditions, to mimic as closely as possible the in vivo conditions. Indeed, our group and others demonstrated that shear stress strongly influences both the level of autophagy [6, 8] and the release of EVs—amount and content—[9, 10] by endothelial cells both in the liver and other vascular beds. For instance, physiological shear stress triples the level of autophagy in LSECs [6]. More generally, maintaining cells in an environment that is as close as possible to in vivo conditions is crucial, particularly in the case of primary LSECs that capillarize in a few days in vitro. In that regard, parameters such as shear stress, but also culture surface stiffness, co-culture with other hepatic cell types should be taken into account when looking at autophagy in LSECs and EV emissions. Future analysis of LSECs EVs in vivo should also be facilitated by the emergence of new technologies, such as intravital imaging and reporter mouse models [3]. These approaches allow to overcome the difficulties related to EV small size placing them under the diffraction limit of standard microscopes [11], and the relative paucity of EVs necessitating important amounts of starting material.

A previous study showed the uptake of LSEC-derived EVs by HSCs. This study was also conducted mostly in vitro, using immortalized cell lines. Authors showed that overexpression of SK1 in a LSEC line, as observed in LSECs isolated from mice with liver fibrosis provoked by carbon tetrachloride treatment, induces the release of EVs promoting AKT activation and migration in HSCs [12]. The conclusions of that study are thus opposite to the ones obtained by Chen and colleagues. Explanations for this discrepancy could be: (i) the different stimuli emitter cells were exposed to; (ii) the differential uptake and consequently effect of EVs. Indeed, EVs composition strongly influences the uptake of EVs. In vivo, some EV target certain organs with remarkable specificity [11, 13] while others are taken up following poorly selective interaction mechanisms, for instance via the phosphatidylserine certain EVs expose at the surface [11]. All this raises the question whether LSEC-derived EVs are taken up by HSCs simply because of their proximity inside the sinusoids or whether specific subtypes of LSECs-derived EVs are preferentially taken up by HSCs. This question is not only crucial to better understand the interactions between liver cells in cirrhosis, but also to be able in the long term to harness LSECs-derived EVs as vector of specific targeting HSCs for therapeutic approaches in cirrhosis.

In conclusion, Chen and colleagues advanced the knowledge of the role of LSECs, HSCs, autophagy and EVs in liver homeostasis and disease. In order to take full advantage of this understanding to consider targeting EVs for therapeutic purposes or using them as delivery means for drugs, a better grasp of the heterogeneity of LSEC-derived EVs, their interaction with other liver cell types and relevance of the observed mechanisms in other clinical contexts is needed.