Polynucleotide cytidine deaminases, innate immunity and hepatitis B infection

More than 350 millions individuals worldwide suffer from chronic hepatitis B virus (HBV) infection, a condition that evolves towards liver insufficiency and hepatocellular carcinoma in approximately 15 to 40 percent of cases. Yet the majority of de novo HBV infections, except for those occurring during the perinatal period, are cleared during the acute phase. This largely appears to result from non-cytopathic host defense mechanisms involving cytokines and notably interferons (IFNs). A key role for IFNs in the biology of HBV infection has been further illustrated by studies in animal models of the disease. Besides, IFN-α is the standard treatment for chronic hepatitis B. Although type I and II IFNs are the most important immune-regulators of HBV replication, the downstream effectors mediating the inhibitory action of IFNs on HBV replication have not yet been identified. In a first place we showed that HBV replication is blocked by human APOBEC3G and APOBEC3F, two closely related editing enzymes belonging to the family of polynucleotide cytidine deaminases, whose catalytic activity results in the conversion of cytidine residues to uridine residues. Both proteins were initially recognized for their ability to inhibit retroviruses, a property which mainly depends on their editing activity, and as the targets of the virion infectivity factor (Vif) protein of human immunodeficiency virus (HIV), which induces their degradation. The family of polynucleotide cytidine deaminases also comprises Activation-Induced cytidine Deaminase (AID), APOBEC1, APOBEC2, and the products of the APOBEC3A through APOBEC3H gene cluster. AID is essential for class switch recombination and somatic hypermutation of the immunoglobulin locus in pre-B lymphocytes, thus for the generation of antibody diversity, and APOBEC1 edits the apolipoprotein B mRNA in the guts, thereby regulating cholesterol metabolism. By comparison, little is known about the physiological functions of the other family members. However, apart from APOBEC2 and APOBEC3A, sequence analysis of the genes encoding these enzymes suggested that they possibly evolved as a host defense against pathogens. This hypothesis was confirmed for most of APOBEC3 proteins, which were independently found to restrict the replication of various exogenous and endogenous retroelements, as well as the adeno-associated virus. For this reason, we next explored the range of cytidine deaminases active on HBV and found that besides APOBEC3G and APOBEC3F, APOBEC3B, AID and APOBEC1 are also potent inhibitors of HBV. In parallel, while examining in details the mechanisms underlying the antiviral activity of APOBEC3G against HBV, we found that this cytidine deaminase blocks HBV infection not by lethal editing of nascent reverse transcripts, as for retroviruses, but rather by impairing the accumulation of pregenomic RNA (pgRNA)-containing viral capsids, the subviral structures in which HBV reverse transcription takes place before virions containing a partially double-stranded DNA genome are released. Remarkably, this is similar to the decreased steady-state levels of pgRNA-containing capsids which are observed upon exposure of HBV infected cells to IFNs. It suggested that IFNs and cytidine deaminases might affect HBV replication through the same pathway. The first step to unveil the in vivo relevance of antiviral cytidine deaminases for the control of HBV replication consisted in our demonstration that type I and II IFN treatment of human hepatocytes induces the production of APOBEC3G, and to a smaller extent, that of APOBEC3F and APOBEC3B, but not that of AID and APOBEC1. These observations led us to ask whether the induction of antiviral cytidine deaminases by cytokines contributes to the IFN-induced clearance of HBV, as IFN-α and IFN-γ are both found in high amounts in the liver of acutely infected individuals. Importantly, we revealed that blocking APOBEC3G, APOBEC3F and APOBEC3B in hepatic cell lines does not abrogate the inhibitory effect of IFNs on HBV, indicating that antiviral cytidine deaminases are not the main mediators of the action of IFNs on this pathogen. Nevertheless, while HBV replication was not impaired in hepatic cells when these enzymes were expressed at their baseline levels, they could edit the HBV genome in these conditions, suggesting that cytidine deaminases might participate in the emergence of new HBV strains, notably those that have been recognized as being of clinical importance. Finally, the results of this work indicate that a therapeutic strategy based on the expression of antiviral cytidine deaminases in the liver could be envisioned to cure HBV chronic infections.