5.1. Extracellular Vesicles in Viral InfectionExtracellular vesicles (EVs) are small membrane-enveloped vesicles that are actively released by a wide variety of cell types and contain information about the state of the cell at the time of its biogenesis [29]. This information is characterized by a distinct EV cargo (e.g., proteins, microRNAs, or other nucleic acids) that is transported from a cell of origin to a target cell [103,104]. Therefore, EVs are considered as important signalling vehicles and modulators of cellular functions such as the immune response. According to their mode of biogenesis, EVs are classified into different subgroups. The most prominent subgroups are exosomes, which are generated by inward budding into the multivesicular bodies (MVBs); microvesicles, which are released by outward budding of the plasmamembrane; and apoptotic bodies, whose formation is associated with cell death. In the following, the generic term “extracellular vesicles” will be used, since a separation and classification of already-released EVs into the individual subspecies is not possible with conventional purification methods [105]. These methods purify EVs according to their density (gradient centrifugation), diameter (size-exclusion chromatography) or on the basis of specific surface antigens (affinity purification), resulting in EVs of relatively high purity. Moreover, EVs can be purified based on their sedimentation speed (ultracentrifugation) or precipitated using reagents such as polyethylene glycol; these two methods isolate EVs of lower purity. However, the removal of cellular contamination during EV purification is not the only difficulty in immunological EV research. Due to the marked similarities between enveloped viruses and EVs in their size, density, and membrane composition, it is also not straightforward to separate them [106,107,108]. Despite all these controversies and difficulties, encapsidation of infectious viral genomes of whole enveloped and non-enveloped viruses has been reported by several groups to occur and contribute to disease progression [109,110,111,112,113,114]. This route of dissemination is of particular immunological interest and especially advantageous for the virus, as it is protected in this way from both enzymatic degradation and neutralization by antibodies [111,115,116,117]. Still, the role of EVs in infection biology goes far beyond EV-induced viral spread: even if they do not contain viral genomes, virus-induced EVs can act directly on the innate immune response, both activating and inhibiting its immune cells and signalling pathways [116,118,119]. Since this also occurs in the case of HBV infection, the present findings will be described in more detail in the following chapters. 5.2. Immunomodulatory EVs in HBV InfectionSince HBV, as a virus optimally adapted to the human host, skillfully manipulates the immune response, it is hardly surprising that it also uses EVs for this purpose (Figure 3). In recent years, several valuable studies have been published describing an impact of EVs on HBV infection. However, the authors of these publications likewise faced the difficulty of adequately separating EVs and virions, as these are highly similar in size, density, and envelope composition [107,108]. Consequently, the extent to which EV-induced immunomodulation can be addressed by the respective experimental setups must be critically considered.A particular mechanism of HBV-induced immunomodulation is associated with subviral particles (SVPs) such as filaments and spheres, which can be considered as a specific subset of EVs untypically released from the endoplasmic reticulum via the Golgi network [30,31,32]. However, since these SVPs also contain viral components such as HbsAg [120], they can be considered as an intermediate between EVs and virions [108]. This is reinforced by reports from Jiang et al. according to which SVPs, like HBV virions and exosomes, originate from multivesicular bodies, further emphasizing the relatedness between these different extracellular entities [121]. Although the data regarding the immunomodulatory role of SVPs are sparse, influences on both the adaptive and innate immune systems are becoming apparent. Not only have SVPs been shown to sequester neutralizing antibodies in patient sera [120], but an inhibitory effect of SVPs on interferon-α (IFN-α) release by human plasmacytoid dendritic cells (pDCs) has also been reported [75]. As the authors also observed inhibition of IFN-α release by treatment with recombinant HbsAg derived from yeast, the question arises as to whether immunosuppression by SVPs is due to their HBsAg content only, or also due to other EV cargo such as virus-induced miRNAs or other viral proteins [74,75]. These interesting observations call for further studies on the immunomodulatory role of SVPs.In addition, immunosuppressive effects of SVP-free EVs (non-SVP EVs) have been reported at the functional level. It was observed that EVs which were purified from the supernatants of HBV-producing cells, as well as from patient sera accumulated in liver, spleen, and intestine [122]. This was concluded to induce immune suppression and enhanced numbers of HBV core antigen-producing cells in a murine in vivo system. However, EVs were purified by ultracentrifugation, resulting in rather impure EVs and no separation between EVs and HBV virions. In another study, EVs were separated from SVPs and HBV virions by density gradient centrifugation, and a separation between fractions containing EV-marker CD9 on the one hand and fractions containing HbcAg and HBV DNA on the other hand was shown [123]. These EVs were then used to tread human peripheral blood mononuclear cells (PBMCs), and upregulation of programmed death ligand 1 (PD-L1) in macrophages and monocytes was measured. The authors suggest that this may be responsible for the frequent occurrence of T-cell exhaustion in CHB patients [123,124]. In line with this, a study by Shi et al. concludes that HBV-induced EVs are also partly responsible for the decreased response of some CHB patients to IFN-α treatment. The authors claim that interferon-induced transmembrane protein 2 (IFITM2), which is upregulated in CHB, will be shuttled to pDCs by exosomes and thus inhibit IFN-α synthesis [125]. However, EVs in this study were purified by precipitation, resulting in a decrease in the volume of an EV sample rather than removal of non-EV components. Thus, no distinction can be made at this point between effects due to exosomes, other EVs, and other components in the conditioned medium because all are precipitated.After all this evidence for an immunomodulatory role of EVs in HBV infection, the question arises as to which viral or virus-induced factors are responsible for these effects. The HBx protein, as the master regulator of HBV infection, also seems to play an important role in this context. Thus, HBx induces the release of EVs that transfer both HBx protein and mRNA to hepatic recipient cells, which stimulates them to proliferate, and thus may be partly responsible for the formation of HCC [126]. Moreover, HBx has been reported to enhance the EV-dependent secretion of the HBV replication-inhibiting apolipoprotein B mRNA-editing catalytic polypeptide 3G (APOBEC3G), lowering intracellular APOBEC3G levels and thus promoting HBV replication [127]. However, HBx is not the only immunoregulatory factor in HBV infection. For instance, cellular and viral microRNAs (miRNAs) can be transported by EVs, thus affecting the immune response to HBV infection. Two reports by the group of Hua Tang published in 2017 and 2020 address the HBV-encoded miRNA miR3, which promotes HBV persistence and controls viral replication [128,129]. They observed that miR3 is found in EVs as well as in the HBV virion itself, enhances IL-6 secretion via SOCS5 in M1 macrophages, and thus controls the innate immune response. In addition to these reports of immune regulation by HBV-encoded miRNAs, there are further reports that HBV also affects the expression of cellular miRNAs. Kouwaki et al. demonstrated that HBV infection leads to increased levels of cell-derived immunoregulatory miR21 and miR29a in EVs, thereby inhibiting the IL-12 response in THP-1 macrophages [103]. In this way, HBV can not only target EV pathways but also influences cellular miRNA biogenesis, thus linking two completely different regulatory pathways and using them for its own advantage.Like all areas of EV research, the focus of HBV-induced EVs is in an evolving process, and benefits greatly from new methods. Previously mentioned studies did not yet have access to an established method allowing for complete removal of HBV virions from EV samples, which is essential for a final differentiation between virus- and EV-mediated effects. The use of EVs in functional studies also requires that the samples are not contaminated with antibodies. However, EV samples obtained by affinity-based methods still contain EV-specific antibodies, which cannot be separated from the vesicles if they have been purified by positive selection. Therefore, purification of pure EV samples and removal of HBV contamination by negative selection is the method of choice. In 2020, we published a method showing a clear removal of HBV virions from EV-containing plasma samples [130]. This method is a combination of size exclusion chromatography and removal of HbsAg-positive particles by negative selection. The resulting EV samples are free of HBV-sized particles, HbsAg, HbcAg, and infectious potential. As the samples are also free of contaminating HbsAg-targeting antibody, they are suitable for functional studies. These new methodologies enable a clear distinction between virion and EV-mediated effects.Regarding the many observations on proviral and immunosuppressive EVs in HBV infection, it should not be forgotten that HBV-induced EVs may also play an immune-activating role. However, the nature of PRR-activating ligands could not be conclusively identified in this context. Indeed, EVs released from HBV-infected hepatocytes were reported to induce upregulation of NKG2D ligand on THP-1 and Hepatic F8/80+ macrophages, which resulted in increased IFN-γ secretion from NK cells when cocultured [103]. However, immune-activating EVs were suspected to contain ligands for both RLRs and TLRs, but the immune-activating nucleic acids were not identified.In another study, Dansako et al. primarily suspected immune activation through EV-dependent transfer of mitochondrial DNA [131]. However, it should be noted that this specific observation may be due to apoptotic bodies as a particular subset of EVs. One reason why this conclusion should be cautiously considered is that EV purification was performed via ultracentrifugation, which does not yield EVs of the highest purity, and no filtering step to remove apoptotic bodies was applied. The other reason is that although HBV infection did not lead to less cell viability, it did lead to a reduction in cell number.Compared to the publications of EVs in HBV infection, the number of publications on immunomodulatory EVs in HDV infection, which is a satellite virus dependent on HBV, is much lower. In the first publication ever to establish a link between EVs and HDV, we recently showed that HDV monoinfection, which does not induce virus release in the absence of HBV, mediates the release of immune-activating EVs [132]. These EVs induced a proinflammatory cytokine response in noninfected primary human immune cells. Another publication reports that EVs artificially loaded with HDAg can elicit a cellular immune response against HDV, but the role of HDV-induced EVs in HDV infection was not examined in this study [133]. Apart from these two publications mentioned above, we are not aware of any other original publications on the role of EVs in HDV infections. The various indications of immunomodulatory effects of EVs in HBV and HDV infections call for further research in this area. 5.3. Transfer of HBV Genomes in HBV-EVs

EVs not only influence disease courses by immune manipulation, but also via direct transfer of viral genomes, and this has also been reported for HBV-induced EVs. This way, EV-dependent transfer of replication competent HBV genomes could lead to antibody-independent viral spread, enhancing the severity of hepatitis virus disease. Given the challenging nature of separating EVs from HBV virions and the lack of an established method for EV-HBV separation at the time the experiments were conducted, these individual studies require detailed examination and in-depth discussion:

Yang et al. reported that EVs purified from the sera of chronic hepatitis B virus carriers contain viral nucleic acids, which mediate viral transmission [134]. What suggests a high purity of these EV samples is that they were obtained by CD63-specific affinity isolation, which is considered a method for purifying low-contaminant EVs. However, postulated EV samples still contained HbsAg, and an absence of hepatitis B virions was not shown. Additionally, it cannot be excluded that CD63 is also associated with infectious HBV particles, as it is necessary for the assembly of HBV in the multivesicular body [121,135]. In line with the previously mentioned publication, Kouwaki et al. also detected HBV RNA in EVs released from NTCP-expressing and HBV-infected HepG2 or Huh7 cells [103]. A CD81-specific positive selection was applied to purify the EVs, via which very pure EVs can also be obtained, though it is also not clear if CD81 is excluded from the HBV virion.In contrast, Kakizaki et al. suggest that HBV DNA copies are more likely to be present in the HbsAg- and HBcAg-containing virion fractions of a density gradient than in the EV-containing fractions [123]. In this context, it is important to note that EVs occur in a variety of subspecies that also differ in density, so the putative HBV viral fraction may also contain EVs [136,137].In a recent study, Wu et al. reported enclosure of full HBV virions inside EVs and demonstrated it by electron microscopy [138]. So far, this effect has been predominantly reported for naked viruses such as hepatitis A and E virus, and future work is required to show whether these EV-cloaked virions are specifically translocated into EVs or end up in apoptotic bodies through cell death [110,111,112,113,114,139].According to further publications, HBV DNA was detected in EVs from patient plasma as well as in EVs released from HBV-infected primary hepatocytes or cell lines [131,140,141]. However, these studies employed ultracentrifugation for EV purification, which does not separate EVs from virions. Consequently, they might indicate EV-mediated spread in HBV infection, but may not be sufficient to prove it, as effects could also be due to residual infectious potential of contaminating virions.In agreement with the results of other research groups, we also detected HBV genomes in EVs after complete HBV virion removal [130]. Since these EV samples were free of HBsAg and HBcAg, the HBV genomes could only be present in EVs or in naked capsids not enveloped by HBsAg [142]. Regarding HDV as the satellite virus to HBV, HDV genome-containing EVs were also present in the conditioned medium of HDV-monoinfected hepatoma cells releasing only EVs and no HDV virions due to the lack of HBsAg [132].

Taken together, these combined results shown by various groups demonstrate that encapsidation of HBV genomes occurs in EVs and may have an impact on infection progression. If the postulated EV-dependent HBV transmission is shown to be a fact in further studies, it would have a striking impact on the course of infection and the efficacy of therapies. Given the high number of chronically infected patients, further research in this area is absolutely essential.