The HDI HBV mouse model can be used to establish transient or persistent HBV replication with HBsAg secretion into the peripheral circulation from 1 week to more than 6 months. HBV persistence in this model is determined by different factors, including the genetic background of mouse strains, age, gender, plasmid backbone, and the dosage of plasmid ( Table 1 ). The genetic background plays a key role in HBV exposure outcome. It has been reported that 40–60% C3H/HeH and C57BL/6 mice remain HBV-positive up to 6 months after HDI of 10 μg pAAV/HBV1.2 (13, 18–20). In contrast, BALB/cJ, FVB/NJ, NOD/ShiLtJ, and 129×1/SvJ mice that received the same plasmid injection rapidly clear HBV within 3-5 weeks post injection (wpi). Even mice with the same background but different substrains display different outcomes after HDI of HBV plasmid. In C57BL/6N mice, HBsAg and HBV DNA rapidly decline below the limit of detection within 8 weeks, while HBsAg remains detectable in C57BL/6J mice at 26 wpi (21).

HBV persistence differs in mice of different ages or gender after HDI. 12-week-old adult C3H/HeN mice have been shown to clear HBV within 6 wpi; however, 50% of their younger counterparts (6-week-old) remain HBsAg-positive at 26 wpi (18). The same study suggested that adult mice possess mature gut microbiota that stimulates liver immunity, resulting in rapid HBV clearance, while younger mice are immune-tolerant through the Toll-like receptor (TLR)-4 dependent pathway (18). Another study using AAVS1 mice demonstrated that male AAVS1 mice receiving 10 μg pAAV/HBV1.2 express higher HBsAg and HBeAg levels than female mice (20). Similarly, male C57BL/6 mice that received 10 μg pSM2 plasmid expressed higher levels of HBV proteins than their female counterparts (15). Mechanistically, it is due to high frequency of regulatory T cells and low response of HBV-specific T cells in male mice (15).

The plasmid backbone also affects HBV persistence in mice after HDI. In contrast to pAAV/HBV1.2, injection of pGEM4Z with the same 1.2-fold HBV genotype A genome into C57BL/6 mice produces only transient antigenemia, and HBsAg rapidly disappears within 3 weeks (13). In line with this, mice injected with 15 μg pAAV/HBV1.3 containing 1.3-fold HBV genotype C genome clear the HBsAg within 8 weeks while 60% mice remain HBsAg-positive at week 20 post HDI of pcDNA3.1(+)-HBV1.3 (14).

Finally, the doses of plasmid backbone or HBV genome also greatly impact HBV persistence in HDI model. Approximately 40-80% C57BL/6 mice that receive 5 or 6 μg pAAV/HBV1.2 develop HBsAg persistence up to 6 months, while a high dose of 100 μg pAAV/HBV1.2 results in HBsAg clearance within 4-5 weeks (16, 17). Similarly, 20 μg pAAV/HBV1.2 or 6 μg pAAV/HBV1.2 plus 14 μg pAAV/control plasmid also results in HBV clearance within 4 weeks (17).

These data highlight the need for researchers to carefully choose mouse strains and plasmid doses when using this model in HBV research.

There is a consensus that adaptive immune responses, especially those of virus-specific T cells, play an essential role in mediating intrahepatic HBV clearance (22). In acute HBV infection in humans, viral clearance is associated with the development of a vigorous multispecific CD8+ T cell response and an acute necroinflammatory liver disease; it was therefore assumed that HBV clearance is principally mediated by virus-specific major histocompatibility complex class I-restricted cytotoxic T lymphocytes (CTLs) (23). The immune effectors required for HBV clearance were screened in an HBV HDI mouse model, which also demonstrated that CD8+ T cells are the key cellular effectors mediating hepatic HBV clearance (24). Consistent with these findings, chronic HBV infection is characterized by weak or undetectable HBV-specific CD8+ T cell responses and the presence of functionally exhausted HBV-specific CD8+ T cells that are unable to clear the virus (25–27).

The use of mouse models allows detailed analysis of immune responses to HBV in vivo. Several studies employed HDI mouse models to investigate immune responses to HBV and their relationship with HBV clearance. An acute HBV replication mouse model was established by HDI of pT-MCS-HBV1.3 and pCMV-SB. A variety of immunodeficient mouse strains have been used to determine the roles of different types of immune cells in HBV clearance in this model; the results showed that natural killer (NK) cell, type I IFN, CD4, and CD8, but not B cells, contribute to HBV clearance (24). Similarly, Tseng et al. showed that the molecules involved in innate immunity, including IFN-α/β receptor (IFNAR), RIG-I, MDA5, Myd88, NLRP3, ASC, and interleukin-1 receptor (IL-1R), are not essential for HBV clearance in the mouse model (28). Another study using a panel of mouse strains lacking specific innate immunity components demonstrated that HBV-specific T cell responses are functionally impaired in TLR signaling-deficient mice, thus allowing enhanced and prolonged HBV replication and expression (29). In line with this, when blocking the type I IFN receptor or TLR7 signaling pathway, CD8⁺ T cell activation and HBV clearance are significantly impaired (30).

The immune factors that suppress or contribute to HBV persistence could be also found by comparing HBV HDI mouse models that mimic acute-resolving and chronic HBV infection. By comparing C57BL/6N and C57BL/6J mice, Wang et al. found that regulatory T (Treg) cells suppress the follicular helper T (Tfh) cells response to HBsAg in C57BL/6J mice and thus lead to persistent HBV infection (21). Importantly, this is consistent with the impaired Tfh-cell response to HBsAg in patients with chronic HBV infection (21). Consistently, depletion of Tregs or inhibition of their function with a blocking antibody against cytotoxic T-lymphocyte-associated protein 4 (CTLA4) restored the Tfh-cell response to HBsAg both in the HBV-persistent mice and patients with chronic HBV infection (21). In line with this, our group demonstrated that intrahepatic HBV replication resulted in increased intrahepatic Treg infiltration and thus limited intrahepatic anti-HBV CD8+ T cell responses (31). We further demonstrated that the suppression of anti-HBV CTL function by Tregs is gender related, which nicely explains the gender-related differences in HBV infection outcomes in humans (15). It has also been reported that the chronic HBV replication mouse model established by HDI of pAAV/HBV1.2 is unable to respond to HBsAg vaccination, which is attributed to IL-10 secretion by Kupffer cells (KCs) (32). Moreover, extracellular vesicles secreted by HBV-infected cells from patients with HBV exert immunosuppressive functions and lead to HBV persistence in the HDI model (33). Recently, we demonstrated that during acute-resolving HBV infection in HDI model using pSM2 plasmid, the liver produces increased amounts of matrix metalloprotease (MMP)2, which, together with MMP9, mediate membrane CD100 shedding from the surface of T cells and NK cells in the secondary lymphoid organs and increase serum soluble CD100 levels (34). By interacting with CD72, sCD100 induces the activation of antigen-presenting cells in the spleen and liver in the HDI mice, thus promoting the intrahepatic anti-HBV CD8 T cell response and virus clearance (34).

Immune therapeutic strategies that aim to restore the antiviral function of HBV-specific CD8⁺ T cells have been intensively examined in HBV HDI mouse models that mimic chronic HBV infection. To mimic patients with chronic hepatitis B, the persistent HBV replication mouse models are usually established by HDI of 6-10 μg pAAV/HBV1.2 plasmid into 5-to-6-week-old male C57BL/6 or C3H mice (35, 36). Our group previously demonstrated that intravenous administration of TLR3 ligand Polyinosinic:polycytidylic acid [poly(I:C)] led to HBV clearance, which was associated with increased CD8⁺ T cell infiltration in the liver in an IFN-γ- and CXCR-3-dependent manner (37). Furthermore, we also showed that the NOD1 ligand, D-glutamyl-meso-diaminopimelic acid, enhanced the antigen-presenting function of liver sinusoidal endothelial cells to promote HBV-specific CD8⁺ T cells and HBV clearance in an HBV-persistent HDI mouse model (35). Therapeutic strategies targeting immune checkpoint programmed death-1/programmed death ligand-1 to restore the antiviral function of HBV-specific CD8⁺ T cells were also examined in an HBV persistent HDI mouse model (38).

HDI mouse models are also widely used to test antiviral agents such as lamivudine (14), RNA interference targeting HBV replicative intermediates (39), and the HBV genome-specific guide RNA-mediated clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 system (40). Anti-HBV vaccines such as cytomegalovirus-based vaccines (41) were also tested in this model.

Another advantage of the model is that plasmids of different HBV genotypes and variants or mutants can be constructed and injected into the mice. Therefore, they can be used for testing antiviral effects against mutated HBV (42) and exploring the influence of HBV genome strain on HBV persistence, as well as its underlying mechanisms (14, 43).

Although HDI mouse models are widely used to investigate immune responses and evaluate antiviral compounds, the limitations should be carefully considered. Firstly, the HDI method is stressful for animals as it involves injecting a large volume of liquid, and it also causes significant liver damage in the first few days after injection (10). Secondly, the non-HBV immune response elicited by the plasmid backbone should also be considered. Thirdly, mouse hepatocytes are extremely inefficient in forming cccDNA from rcDNA, resulting in a general absence of cccDNA (44). Fourthly, intrahepatic transfection efficiency is only ~6% in this model (12), much lower than in HBV-infected patients in the real world. Finally, due to the lack of the entry receptor in mice, this is not a natural infection model, so it cannot be used to study the entire infection process. Considering the huge differences between HDI mouse model and patients with HBV infection, researchers should be careful when evaluating the findings about the therapeutic value of immunological and pharmacological interventions against acute and chronic HBV infection in hydrodynamic transfection systems.

There are other aspects to consider when interpreting results from HDI mouse models. When modeling acute HBV infection, the levels of HBsAg and HBV DNA decline quickly in the initial phase. This also occurs under treatment with various of immunosuppressive agents (45), suggesting that this process is not immune-related and that the results obtained in such a situation should be interpreted with care. Although HBV can persist in the host and maintain its replication over a prolonged period, the immunological mechanisms leading to persistence are primarily related to lack of recruitment and intrahepatic accumulation of activated T cells to HBV. Intrahepatic immune activation by TLR ligands result in immune cell infiltration and HBV clearance (37). A major issue is the lack of inflammation in mice with persistent HBV replication, resulting in insufficient recruitment of immune cells into the liver. Thus, HDI mouse models may be useful for investigating host immune responses and viral clearance or persistence; however, a direct comparison with acute and chronic HBV infection in humans is not always possible.

Both Ad-HBV and AAV-HBV transfection mouse models can establish persistent HBV replication for more than 3 months or even more than 1 year (46, 52, 54). Notably, only low doses of Ad-HBV have been reported to establish such persistent HBV replication (46). The high doses of Ad-HBV can induce acute-resolving hepatitis, which is partially due to strong non-HBV immune responses induced by the vector itself (55, 56). Nevertheless, this permits the investigation of mechanisms of immune-mediated viral clearance, including the role of NK cells (57) and intrahepatic CTLs (58).

Compared with Ad-HBV, AAV-HBV vectors have minimal AAV genomes with only the essential AAV inverted terminal repeat sequence that is responsible for viral packaging and does not encode any AAV viral proteins. Thus, this vector has a clean background and does not induce an obvious non-HBV immune response (59), making it a more suitable tool to establish persistent HBV infection (10). More importantly, the AAV-HBV mouse model allows more efficient and homogeneous HBV transduction than the HDI model with ~60% of hepatocytes expressing HBcAg 12 weeks after injection (49).Serum HBsAg in AAV-HBV mice remains at relative high levels even in 3 or 4 month ( Table 2 ). Therefore, the AAV-HBV mouse model is preferred to evaluate therapeutic vaccines for chronic infection (61) and the antiviral effects of TLR agonists like the TLR9 ligand CpG (50). It is worth mentioning that Prof. Xiaobing Wu’s group in Shanghai developed a persistent HBV replication mouse model by transduction of recombinant AAV8 (rAAV8)-HBV 1.3 into C57BL/6 mice (62). This model is also used to evaluate antiviral effects of NAs such as entecavir and lamivudine (63). Moreover, two studies reported fibrosis in HBV mice transfected with AAV8-HBV1.2, which is one of the few replicative models that can be used to study the mechanism of liver fibrosis in the course of chronic hepatitis B (64, 65). Lucifora et al. reported that cccDNA can be detected by Southern blot in an AAV-HBV model (66). However, it remains uncertain whether AAV-HBV generates cccDNA in murine hepatocytes through rcDNA, the natural precursor of cccDNA.

As adenoviral vectors have large genomes and encode numerous non-HBV viral products, it is technically challenging to interpret the HBV-related immune response and pathogenesis in the Ad-HBV transfection model (55, 56). Moreover, except for the observed liver fibrosis in AAV8-HBV1.2-transfected mice mentioned above, Ad-HBV and AAV-HBV mice do not usually develop pathological changes in the liver. Although rAAV8-HBV1.3 can to some extent achieve persistent HBV replication after transfection, it cannot stimulate HBV-specific humoral immunity (67). This may be due to immune tolerance induced by the AAV8 vector (67). Moreover, similar to the HDI approach, this model does not represent real infection. For these reasons, it is controversial whether this model can be used to study HBV immunology, which limits its application.

The HBV protein-transgenic mice express HBV proteins such as HBsAg (77, 78), HBcAg (79), HBeAg (80), or HBx (81). These models can be used to study the virology and oncogenic potential of these HBV proteins.

In 1985, Chisari et al. developed the first HBV transgenic mice that express HBsAg (77). It was initially suggested that this model did not exhibit signs of pathology. Later, the same group described that increased expression of large HBsAg, rather than the small and middle HBsAg proteins, inhibits the secretion of HBsAg-containing particles (78). Furthermore, HBsAg accumulation in the endoplasmic reticulum of hepatocytes leads to continuous inflammation and liver injury, followed by HCC development (82). The same HBs-transgenic mice were crossed to C57BL/6 and BALB/c, which revealed that HBsAg protein-induced liver injury depends on host genetic background (83). It was recently shown that methylation of specific CpG sites controls gene expression in these HBs-transgenic mice, resulting in decreased cell stress and improved liver integrity (84). Although pathological aspects of HBsAg could be revealed with this model, the intracellular accumulation does not reflect the natural function of HBsAg in HBV-infected humans.

To address HBV biology and immunological aspects, HBeAg-transgenic mice were generated to investigate the role of immunological tolerance in chronic infection of newborns (80). Both HBeAg and HBcAg are expressed in this model. It has been suggested that HBeAg secretion may be one of the viral strategies for HBV persistence after perinatal infection (80). Another HBcAg-transgenic mouse model was used to examine factors that influence the intracellular localization of nucleocapsid particles in hepatocytes; it demonstrated that nucleocapsids can form de novo within the nucleus, and the reformed nucleocapsid cannot be transported across the intact nuclear membrane (79). Knowledge about the functional and structural significance of HBcAg and HBeAg could be gained with these models.

Mice transgenic for HBx protein have yielded controversial results regarding its oncogenic potential. Early reports showed that HBx-expressing mice develop HCC (81, 85, 86). However, others reported contradicting findings that HBx alone is not sufficient for inducing HCC development (87–89). Hepatocarcinogenesis has been directly compared in HBsAg and HBx knock-in transgenic mice. HBsAg and HBx genes were integrated into the mouse p21 locus using homologous recombination. Although p21-HBx transgenic mice developed HCC at the age of 18 months, p21-HBsAg transgenic mice have developed HCCs 3 months earlier. Herein, the expression of genes related to metabolism and genomic instability largely resembled the molecular changes during HCC development in humans (90). However, hepatocarcinogenesis in HBV-transgenic mice occurs in a cirrhosis-free condition and does not reflect the common situation in HCC-developing patients with chronic hepatitis B infection.

Single-HBV protein transgenic mice facilitated investigation of the assembly, secretion, immune responses, and functionality of these proteins in vivo. However, HBV-full genome transgenic mice support the whole HBV life cycle in vivo. The integrated HBV genome includes terminally redundant DNA representing a 1.3 overlength HBV genome leading to efficient viral replication, including rcDNA synthesis from pregenomic RNA, viral assembly, and production of infectious HBV particles that are morphologically indistinguishable from human-derived virions (76). Due to self-tolerance, full-genome transgenic mice do not suffer from T cell-related liver injury. The lack of hepatitis indicates that HBV viral proteins or HBV itself is not cytopathic in transgenic mice. Adoptive transfer of HBV-specific cytotoxic T cells caused acute liver disease and thus resulted in temporal clearance of HBsAg (91, 92). These HBV replication-competent transgenic mice have been extensively used to evaluate antiviral drugs that hinder HBV replication, including small interfering RNAs (93, 94); cytokines (95, 96); and various polymerase inhibitors such as adefovir dipivoxil (97), lamivudine (98), and entecavir (99). Thus, cellular, molecular pathogenic, and antiviral mechanisms occurring in acute viral hepatitis have been recapitulated in full-genome transgenic mice.

All lineages of HBV transgenic mice generated to date are immunologically tolerant to viral antigens, particularly at the T-cell level, and they do not develop acute or chronic hepatitis. Therefore, these models do not develop HBV-related liver diseases including liver injury and progressive fibrosis and cirrhosis. Due to the transgenic origin of HBV replication, these mice do not have the potential of viral clearance (100). Another restriction is the lack of cccDNA in these full-genome transgenic mice, suggesting that species-specific differences may play a role in determining the host range of this virus (74). Still, the full-genome transgenic mice are absent of the HBV entry factor NTCP. Although infectious particles are produced, hepatocyte infection does not occur. Furthermore, in human NTCP-expressing transgenic mice, HBV infection in mouse hepatocytes is also limited due to the lack of a host cell dependency factor (100). For these reasons, studies examining HBV entry, mechanisms underlying infection, compounds interfering with the infection process, and cccDNA synthesis are not suitable in full-genome transgenic mice.