SARS-CoV-2 infection can disrupt innate antiviral immune defences, leading to a massive release of cytokines (21–24). Severe cases show excessive elevation of blood concentrations of pro-inflammatory factors and chemokines such as IL-6, MCP1, TNF-α and IFN-γ. Over-activation of immunity leads to systemic cascade reactions such as acute respiratory distress syndrome, septic shock, multi-organ failure and even death (25–27).Using mathematical modelling, S.Q.Du (28) et al. demonstrated that delaying the onset of adaptive immunity in the early stages of the infection may be a promising therapeutic tool for patients at high risk of COVID-19 ( Figure 1 ).

Viruses can induce changes in the clinical features and progression of deterioration in a variety of autoimmune diseases. For example, herpesvirus-induced systemic lupus erythematosus (SLE), multiple sclerosis (MS), rheumatoid arthritis (RA), and hepatitis C virus-associated cryoglobulinaemic vasculitis (29, 30). This implies that viruses have some autoimmunomodulatory role, such as coxsackievirus B, rotavirus (31), influenza A virus as well as measles, rubella, and EBV (32). Similarly, SARS-CoV-2, a novel enveloped single-stranded RNA β-coronavirus, has been reported to induce diseases such as Guillain-Barré syndrome (GBS) (33), autoimmune haemolytic anaemia, anti-MDA-5 positive dermatomyositis (34), arthritis, psoriasis (35), and immune thrombocytopenic purpura (36).

COVID-19-mediated autoimmunity may be plausible. According to type II hypersensitivity, the production of autoantibodies after viral infection may lead to tissue damage (cross-reactivity). This is a possible mechanism of virus-induced autoimmunity. But it is not certain whether COVID-19-associated slave autoantibodies may act as a catalyst for disease exacerbation, or a mediator for triggering immune-disordered inflammation, or simply a concomitant phenomenon of intense inflammation. Studies have shown (37) that 10%-15% of COVID-19 patients have circulating autoantibodies, such as anti-type I interferon autoantibodies, antinuclear antibodies, lupus anticoagulant, anti-SSA/Ro antibodies (38), anticardiolipin antibodies (39), anti-IFN antibodies (40), dermatomyositis-associated anti-MDA5 antibodies, and anti-angiotensin converting enzyme 2 (ACE2) autoantibodies. As the disease worsens an imbalance occurs which in turn damages systemic tissues (41, 42). This phenomenon suggests that antibody activation of COVID-19 may be one of the factors triggering AIH. It is important to note that autoimmune dysregulation is more likely to occur in the elderly or in people with comorbid underlying diseases. Defective T-cell activation due to aging of the organism’s microenvironment is hardly implicated (43).

Interestingly, there is no direct evidence that SARS-CoV-2 can directly cause AIH. Bartoli A et al. (41) published a report that middle-aged women with Hashimoto’s thyroiditis developed primary sclerosing cholangitis and Guillain-Barré syndrome after infection with SARS-CoV-2. Specific manifestations were elevated liver enzymes and a positive AMA (1:640) as well as a liver biopsy showing Ludwig stage I, which is consistent with early manifestations of PBC. Certain patients may develop sero-immunological features of AIH during the course of COVID-19. Exceptionally, IgG elevation in AIH after SARS-CoV-2 infection was not significant. Whereas, given that Hartl et al. (44) reported a 10% probability of the presence of a subgroup of IgG-normal AIH, normal IgG is not a major basis for ruling out AIH.

What has been identified, but not yet definitively reported, is the presence of hypoalbuminaemia, elevated gamma-glutamyltransferase and alkaline phosphatase during the course of COVID-19 patients. Of these, 2-11% of COVID-19 patients had hepatic comorbidities, and 14-53% had abnormal levels of liver injury markers detected during the course of the disease (45). EASL-ESCMID (46) suggested that patients with severe pneumonia were more likely to have abnormal transaminases. Despite all this, transaminase abnormalities have not been shown to be a causative factor for COVID-19 exacerbations. Possible mechanisms include (45) immune-mediated cytokine storm, pneumonia-associated ischaemic-hypoxic injury, microthrombosis, and drug effects such as lopinavir (47). Interestingly, the direct cytotoxic effects of the virus are negligible because the viral titre in liver tissue is relatively low and viral inclusion bodies have not been detected. In addition, the genome sequence homology of the viruses seems to validate this phenomenon. For instance, 60% of SARS (85% homozygosity) can lead to hepatic impairment (48) and MERS-CoV (50% homozygosity) infection carries a risk of causing lobular hepatitis (49).

Gamma-glutamyltransferase is known to be a diagnostic biomarker of bile duct cell injury. Coincidentally, cholangiocytes show high expression of ACE2 receptor (50) and the spike (S) protein of SARS-CoV-2 binds to angiotensin converting enzyme 2 (ACE2) receptor (51). Subsequently, transmembrane serine protease 2 catabolism of the S protein occurs after the binding body enters the target cell ( Figure 2 ). Following this, the viral genome is released from the endosomes and later replicates and proliferates ( Figure 3 ). Notably, viral RNA has only been detected in faecal and blood samples, which is not sufficient to fully justify this mechanism. Therefore, further monitoring of the effect on the degree of cholestasis and biliary tract damage is still required.

Homologous self-proteins of virally shared peptide sequences can lead to pathological autoimmune responses. Based on this principle, Kanduc D et al. explored the possibility of molecular mimicry induced autoimmune diseases by quantitative analysis of peptide sharing (52). Coincidentally, Vorjani et al. (53) pointed out through an in vitro study that molecular mimicry is the cross-reactivity of antibodies against SARS-CoV-2 spike glycoprotein S1 with structurally similar host peptide protein sequences. The peptide sequences included the high-affinity human histone transglutaminase, nuclear antigen, anti-extraction nuclear antigen and myelin basic protein. This suggests that molecular mimicry is involved in the SARS-CoV-2-induced autoimmune response.

In addition, the same immune response may be triggered between COVID-19 and AIH by breaking immune tolerance (54). It is now known that SARS-CoV-2 can upregulate autoimmunity through molecular mimicry, epitope spreading, bystander activation (32), immortalisation of infected B cells, cryptic antigens, spiny synapse-mediated super antigenic activity, and inflammatory signalling stimulation ( Figure 4 ). Note that the efficacy of these effects depends on a variety of factors such as genetic background such as congenital B-cell autoimmune defects, viral strain type, viral load, host immune status and duration of infection.

Microbiomic analysis of COVID-19 suggested the presence of a gut-lung axis as well as autoimmune-associated microbial alterations. Nagata N et al. (55) found the presence of a gut-lung axis in COVID-19 by multi-omics analysis, demonstrating the existence of multiple gut microbial-metabolite-cytokine interrelationships. Katz-Agranov N et al. (56) found a loss of biodiversity, an increase in the number of pathogenic bacteria (e.g., Aspergillus), and a relative decrease in the number of commensal bacteria (anti-inflammatory, protective microorganisms, e.g., Lactobacillus) in both by comparing the microbiomes of COVID-19 and lupus. Analyses of the microbiomes of COVID-19 and AIH are still lacking. It’s worth noting that due to similar pathogenic mechanisms, the findings of the classic autoimmune disease lupus are informative about the triggers of AIH.

Genetic association studies can help identify genes and biological pathways behind specific physical outcomes or traits. AIH is known to be rich in genetic susceptibility. According to typical biomolecular mechanisms, the immune mechanisms such as risk of infection, disease progression, or cytokine triggering in SARS-CoV-2 must involve regulation and variation at the level of deep genetic pathways. Regrettably, there is a lack of cross-targeting results at this stage. In the future, improved genetic studies may further advance the exploration of the causal association between COVID-19 and autoimmune diseases such as AIH.

It has long been shown that hepatitis A, human papilloma and influenza virus vaccines (59) can induce AIH (60). The vaccine carries a greater risk of causing liver injury in people with allergies, a history of post-vaccination immune disorders, autoimmune disorders and a genetic family history (61). Vaccine development departments safeguard the safety of vaccines by conducting rigorous clinical drug trials. Nonetheless, as the vaccine became widely available, a series of confirmed cases of AIH were reported around the globe ( Table 1 ).

Fimiano and Tan CK et al. (63, 78) reported the types of COVID-19 vaccines that can induce AIH, including the second dose of AstraZeneca’s vaccine (AZD1222), Pfizer-BioNTech’s mRNA vaccine, Moderna’s COVID-19 vaccine (mRNA-1273), and CoronaVac’s vaccine followed by the third dose of Pfizer Biotech’s Technologies’ booster vaccine third dose. The main types of vaccines that can induce AIH in the cases reported so far include mRNA vaccines (most), viral vector vaccines and inactivated vaccines (least) (58, 65, 76, 77) ( Figure 5 ).

Bril et al. (65) reported the first case of new onset of AIH after vaccination in a 35-year-old postpartum woman who developed generalised pruritus, hepatomegaly, bilirubinuria, jaundice, and antinuclear antibody positivity (1:1,280) one week after vaccination with the COVID-19 vaccine (Pfizer Biotech, first dose). Exceptionally, the patient did not have elevated IgG levels and a liver biopsy suggestive of eosinophilia (44). Therefore, we cannot exclude hepatotoxic effects of the drug.

Clayton-Chubb et al. reported the case of a 36-year-old male physician of Iraqi origin (77), who had hypertension but no liver disease. Twenty-six days after the patient received his first dose of ChAdOx1 nCoV-19 vaccine (Oxford-AstraZeneca), he developed abnormal liver function (ALT, AST, γ-glutamyltransferase, etc.). He was positive for antinuclear antibodies (titre 1:160). The biopsy showed interfacial hepatitis with mixed inflammatory cell infiltration, which was consistent with the presentation of AIH. Of note, the patient’s immunoglobulin G was also not elevated. Obvious confounding factors such as pregnancy and other underlying medical history were ruled out in this case. This supports the idea that the COVID-19 vaccine induces autoimmune phenomena and is not related to the mechanism of action of the vaccine. Undeniably, despite the patient’s good health literacy, we cannot exclude the role of confounding factors such as drugs or toxicants.

Recently, Ghielmetti M et al. (75) reported a case of a 63-year-old male with no history of autoimmunity or natural SARS-CoV-2 infection. He developed acute severe autoimmune liver disease 7 days after the first dose of mRNA-1273. The patient’s immunoglobulin G was mildly elevated. Exceptionally, the patient’s AMA differed from the conventional PBC pattern and also had a unique ANA. he carried the PBC protective HLA DRB1*11:01 allele (82), whereas the determinants of AIH susceptibility are mainly the HLA alleles DRB1*03 and DRB1*04. The clinical significance of these new specific results needs to be further evaluated.

Cases of liver injury due to Pfizer/BioNTech BNT162b2 mRNA COVID-19 vaccine have been reported in the UK. Not coincidentally, a case of drug-induced liver injury due to COVID-19 vaccine was reported by Mann R (83) et al. A 61-year-old woman developed jaundice and malaise 9 days after vaccination (Pfizer, second dose). Her laboratory parameters showed elevated alkaline phosphatase and bilirubin. Inconsistently, the patient’s antibodies were all negative, and a liver biopsy showed only mild oedema and scattered infiltration of inflammatory cells, which is not consistent with the inference that the vaccine causes autoimmune liver disease. Of course, this study does not rule out a correlation with the patients’ history of irritable bowel, cholecystectomy, and fatty liver.

There are many similar reports ( Table 2 ). For example, Lee SK et al. reported a 57-year-old woman who was diagnosed with AIH-PBC overlap syndrome 2 weeks after vaccination (Pfizer/BioNTech BNT162b2 mRNA vaccine, first dose) (81, 84). In addition, Mekritthikrai et al. (85) reported a case of new-onset AIH in a 52-year-old woman after 2 doses of COVID-19 inactivated vaccine (CoronaVac). Although an increasing number of cases suggest an association between COVID-19 vaccine and the development of AIH, the mechanism remains unclear.

Vaccines, as biological agents that render organisms immune to disease, are known to trigger regular or irregular autoimmune responses. The prevailing mechanisms (86) include molecular mimicry, bystander activation, central tolerance defects, adjuvants, epitope spreading and sustained antigenic stimulation.

Talotta et al. (110) suggested that women with pre-existing autoimmune diseases may be at risk of immune dysregulation when vaccinated with COVID-19. Akinosoglou K (111) et al. further elaborated on this issue, stating that COVID-19 vaccination does not trigger new immune-mediated adverse events, but that immune responses have the potential to trigger a genetically predisposed individual’s pre-existing existing underlying dysregulated pathways (112).

Nevertheless, most of the clinical evidence does not allow for a full spectrum diagnosis of AIH. Consequently, a hypothesis was born that AIH after COVID vaccination could be an acute drug-induced hepatitis with autoimmune properties or an AIH-like syndrome due to molecular mimicry of the vaccine, which is supported by comparisons with statin-induced liver injury (113). Recently Mann R (83) et al. reported a case of drug induced liver injury due to COVID-19 vaccine. A 61 year old female developed conjunctival jaundice, malaise, abdominal tenderness and fever 9 days after the second dose of Pfizer COVID-19 vaccine. Laboratory indices showed elevation of alkaline phosphatase and bilirubin, cholestatic phenotype and mild elevation of aminotransferases. Exceptionally, the patient was negative for all antibodies and liver biopsy showed only mild oedema and scattered infiltration of inflammatory cells. This is not consistent with the inference that the vaccine causes AIH. And the follow-up found that the patient’s symptoms and liver function indices gradually resolved. Of course, this study does not exclude the possibility of medical history as a risk factor.

Of course, the case reports are different and cannot be generalised. We can have multiple explanations for the post-vaccination liver lesions. For example, the association of viral spiking proteins with vaccine-induced spiking-specific receptors may lead to autoimmune dysregulation (1), vaccine-associated pharmacological liver injuries such as furotoxin or minocycline may also be characteristic of autoimmune hepatitis (114), and pregnancy, among others, is one of the risk factors for autoimmune disorders such as rheumatoid arthritis and thyroiditis (115–117).

Troublingly, although the mechanism of induction is not clear, the development of liver injury (VILI) after COVID-19 vaccination is similar to autoimmune hepatitis (AIH) in its clinical and histomorphological manifestations. However, surprisingly, recently Uzun S (118) et al. compared the two groups of samples by histomorphological assessment, whole and spatial transcriptome sequencing, multiplex immunofluorescence and immune complex sequencing. They finally found that TRBV6-1, TRBV5-1, TRBV7-6 and IgHV1-24 genes have different roles in the two. VILI had more pronounced centrilobular necrosis, increased activation of metabolic pathways, greater predominance of T- and B-cell clones, and more pronounced CD8+ T-cell infiltration compared with AIH, which had a predominance of CD4+ effector T-cells and CD79a+ B-cells and plasma cells. This study differs from previous conceptions by proposing that VILI may be a separate entity that is distinct from AIH and more closely related to drug-induced autoimmune hepatitis with a better prognosis.

So only vaccines based on minimal immune determinants specific to the pathogen and not present in the human proteome can ensure the safety and efficacy of the vaccine (52). Vaccine-induced AIH or transient drug-induced liver injury remains to be determined, but what is clear is that the association of SARS-CoV-2 vaccination with AIH is by no means coincidental. Certainly, the number and conditions of exposure events in the current study are not sufficient to form sufficient evidence to elaborate a causal relationship between COVID-19 vaccine and autoimmune hepatitis. It also suggests that we should be alert to and pay attention to the disease background, pathogenetic mechanisms, and coping strategies for the occurrence of AIH as a concomitant adverse event.

Vaccination against SARS-CoV-2 significantly reduces the risk of COVID-19 death in AIH patients (119). Early reports have shown that immunosuppression, advanced chronic liver disease and liver transplantation lead to an increased risk of viral infection and impaired immune response after vaccination. A small prospective randomised controlled study (120) monitored antibody titres and spike-specific T-cell responses after SARS-CoV-2 vaccination in patients with AIH. They were found to have lower median antibody levels and 48% of antibody titres were below the 10% percentile of controls (9194 AU/mL, p<0.001). Also AIH patients demonstrated a higher risk of spike-specific T cell response failure and a lower frequency of spike-specific CD4+ T cells. The application of steroids (N=27, 7326 AU/mL, p=0.020) and MMF (N=14, 4542 AU/mL, p=0.004) can be considered as risk factors for them. Interestingly, all of the above phenomena were improved to varying degrees after booster vaccination. This suggests that our SARS -CoV-2 variant is more resistant to the neutralising effect of low titre antibodies and emphasises the need for routine assessment of antibody levels in AIH patients after vaccination and provision of booster vaccination according to individualised needs. On this basis, the appropriate adjustment of immunosuppressive agents may also be of significance for the improvement of vaccine response rates.

Barda N et al. (121), using a database from the largest healthcare provider in Israel, found that the third dose was more effective in preventing adverse outcomes of COVID-19 compared to patients who received only two doses of BNT162b2 mRNA vaccine at least 5 months earlier. In addition, vaccination with Pfizer or AstraZeneca vaccines can show 95% neutralisation of the Delta variant (122). Anti-Omicron antibody titres tended to rise significantly with the third dose of vaccine (123). Various such studies have demonstrated that the third dose of vaccine is effective in enhancing the post-vaccination immune response in immunosuppressed populations and may improve prognosis.

It is important to note that, despite the variety of vaccines, the antibody titres or activity required for effective resistance to infection are still not fully defined. Furthermore, immunosuppressive drugs such as steroids, mycophenolate mofetil or azathioprine suppress the strength of humoral or cellular immunity in the body to varying degrees. A similar phenomenon has been observed in patients with autoimmune rheumatic diseases (121), but the exact degree of immunity, such as the trend of cellular immunity, is not clear. As a result, it remains to be explored whether the adverse outcomes in autoimmune hepatitis patients infected with COVID-19 are due to the disease itself or to the immunological drugs. The standardised duration and efficacy of vaccine booster injections for the treatment of disease also remains to be analysed more comprehensively in larger prospective cohorts.

It is now known that comorbidities of liver transplantation also reduce the effectiveness of vaccination (124), and therefore revisions of their vaccination regimens are mostly advocated. Yet, to date, limited to sample size or deficiencies in control variables, there have been few breakthroughs in this direction as to the extent to which complications of AIH such as oesophageal varices, ascites and hepatic failure impact on neocoronary vaccination.

Research has indicated that the genotypes IL-6 rs1800795 CG+GG, rs1554606 (125), IL-8 rs2227306 CT+CC, and the C allele of IL-10 (rs1800896) (126) may serve as independent risk factors for severe progression and adverse outcomes in COVID-19 (127). Furthermore, the IL-6 rs1800795 GG genotype (128) and the IL-6R (rs12083537) GG genotype (129) can predict COVID-19 mortality. Conversely, compared to G/G-C/C, the IL-6 -174 G/C genotype may offer some protection against adverse COVID-19 progression (p = 0.03; OR = 0.41, 95% CI 0.18-0.96) (130). Likewise, IL-6 rs2069840 and rs2069837 (125) exhibit a similar protective effect (OR = 0.21, 95% CI = 0.07-0.69, P= 0.006) (131). Intriguingly, reports have shown that patients with autoimmune hepatitis (AIH) exhibit the IL-6 -174 GG genotype (132) and have significantly higher IL-6 polymorphism levels (133). Genetic testing has also demonstrated that patients carrying the IL-6 G (rs1800795), C (rs1800796), or G (rs1800797) alleles may be more susceptible to liver diseases (134). These findings suggest that SNPs in cytokine genes play a significant role in the pathological mechanisms underlying COVID-19-triggered AIH.

Additionally, the IL-1β(rs16944) CC genotype is significantly associated with severity and mortality in COVID-19, while IL-1βgene variation (rs1143634) may affect patient prognosis (135). In contrast, the IL1RN haplotype CTA and C/C variant of rs419598 can influence endogenous anti-inflammatory mechanisms, leading to attenuated cytokine release syndrome and reduced mortality in acutely infected males (136). Similarly, the rs1143634 SNP T allele exhibits a protective effect (130). Therefore, IL-1β rs16944 and IL-6R rs12083537 SNPs may serve as potential predictors of severity in SARS-CoV-2 infections.

Interestingly, a significant association exists between type 1 AIH and the rare 2308A allele (TNF*2), which is linked to increased TNF-α transcriptional activity (137). Other studies have shown that TNF-α -308G/A and -238G/A polymorphisms may contribute to increased susceptibility to autoimmune liver diseases (AILD) in Caucasians (128). Moreover, the A/A genotype of TNF-α (-308 G > A) positively correlates with TNF-α levels and susceptibility in AIH patients (138). Surprisingly, TNF-α (GG) significantly influences anti-SARS-CoV-2 spike protein IgG antibody levels (P= 0.005) (139). Individuals carrying the TNF-α G-308 A polymorphism A allele are more susceptible to COVID-19, and TNF-α (AA) is associated with a strong viral invasiveness (140). These findings align with SNPs in TNF-α in AIH. Furthermore, IFNs (125), particularly IFN-λ3 (IL-28B), are involved in antiviral functions. For instance, IFNλ4 rs73555604C can serve as a marker for COVID-19 severity (141). In SARS-CoV-2 infection, rs12979860 and rs1298275 were unfavourable genotypes in the context of hepatocellular carcinoma pathology, but further genetic information is needed.

Crucially, ACE2 polymorphisms can regulate immune responses and inflammatory processes. Hou Y et al. (142) examined DNA polymorphisms in ACE2 and TMPRSS2 (two key host factors for SARS-CoV-2) across approximately 81,000 human genomes and found that ACE2 polymorphisms are associated with cardiovascular and pulmonary diseases by altering angiotensinogen-ACE2 interactions. Meanwhile, p.Val160Met (rs12329760) in TMPRSS2 provides potential support for genetic susceptibility and risk factors in different individuals with COVID-19.

In summary, IL-6 rs1800795 and TNF-α G-308 A, as common SNPs in both diseases, may explain the genetic factors underlying COVID-19-induced AIH at the genomic level. However, no direct evidence currently links a specific ACE2 polymorphism to both AIH and COVID-19. Therefore, large-scale genomic sequencing can be actively conducted to determine if shared ACE2 polymorphisms exist between the two conditions. This may provide deeper insights into the causal associations between AIH and COVID-19 at the pathogenesis level, guiding the application of human genetic measures in combating virus-associated autoimmune diseases.

Specifically, there is a strong linkage disequilibrium between TNF*2 and HLA A1-B8-DR3 haplotypes, and an interdependence between TNF*2 and HLA DRB1*0301 (137). This reminds us that there are more genetic susceptibility mechanisms than just one SNP of cytokines. Studies on the nature of HLA gene polymorphisms may help to uncover the nature of disease associations. HLA polymorphisms, as a central component of the specific immune response, are involved in the recognition of foreign antigens, which triggers a precise immune response and plays a crucial role in warding off pathogen invasion and maintaining the health of the organism.

For instance, studies have demonstrated an association between HLA-A01 and both autoimmune hepatitis (AIH) and COVID-19 (143, 144), implying a notable immunogenetic similarity between these conditions. Nevertheless, disparities exist in the HLA polymorphic loci of AIH and COVID-19, with varying implications. Specifically, susceptibility to AIH type 1 (AIH-1) correlates with the presence of alleles such as DRB113, DRB10301, DRB30101, and DRB10401, among which DQB10301 exhibits protective effects (145). Conversely, AIH type 2 (AIH-2) is linked to HLA class II genes, particularly DRB107 and DRB103 (146), with the DQ locus emerging as a pivotal determinant in AIH-2 pathogenesis, notably DQB10201 as a primary susceptibility locus. Additionally, the HLA-DR (DRB107/DRB1*03) loci are implicated in diverse autoantibody expression profiles (131, 132).Utilizing Han-MHC reference panel analysis, Li Y et al. (147) revealed associations between AIH and HLA-B∗35:01, HLA-B∗08:01, and rs7765379 across the entire MHC region in the Han population, highlighting the significance of ethnic genetic specificity.

Notably, an HLA-B∗46:01 SNP has been implicated in COVID-19 susceptibility (125), while HLA-DQ rs7453920A and HLA-DPrs3077G have been suggested to influence COVID-19 severity (141).Furthermore, Bernasconi E et al. (148) conducted HLA testing on five patients who developed AIH following COVID-19 vaccination, all of whom harboured HLA DRB1 alleles (HLA DRB1*03, *11, *07, *13, *14) strongly associated with AIH susceptibility. These findings underscore the need to rigorously consider the hypothesis that COVID-19 infection or vaccination may precipitate AIH by modulating HLA polymorphisms, thereby presenting a variable and atypical clinical spectrum.