Shedding new light on the chicken-and-egg dilemma of hepatitis delta virus (HDV) replication initiation

HDV is a small circular RNA adopting a self-complementary rod-like secondary structure and replicating within the hepatocyte nucleus via a double rolling circle process. Circular genomic RNA (HDV-G) serves as a template for producing antigenomic RNAs (HDV-AG) and 0.8 kb viral mRNAs. The small and large forms of HDV antigens (S-HDAg and L-HDAg), arising from viral mRNAs, form the so-called HDV ribonucleoprotein (RNP) when associated with HDV-G. For decades, the HDV research has grappled with a chicken-and-egg dilemma following the discovery that S-HDAg is required for HDV replication.1 Mechanistic insights into S-HDAg function include histone mimicry, recruitment of the host transcriptional machinery, comprising the RNA polymerase II (Pol II), as well as the promotion of RNA elongation via the removal of inhibitory elongation factors.2,3 It has been suggested that S-HDAg brought by the incoming RNP would be sufficient to initiate transcription from the HDV-G, since only low amounts of S-HDAg were found to be required to initiate HDV RNA transcription.4 However, the study by Prawira et al.5 published in this issue of Hepatology sheds new light on this dilemma by proposing an alternative model for HDV replication initiation in the absence of S-HDAg. The primary post-translational modification regulating Pol II activity on DNA template is the phosphorylation of the C-terminal domain (CTD) of its RPB1 subunit, which contains multiple repeats of the heptapeptide sequence Y1S2P3T4S5P6S7 and whose phosphorylation pattern (“CTD code”) dynamically changes according to the stage of transcription (initiation, pause, or elongation). A series of kinases are involved in CTD phosphorylation, including some from the cyclin-dependent kinases (CDKs) family. In humans, the current accepted model involves CDK7 in CTD-S5 phosphorylation (initiation of transcription), and CDK9, CDK12, and CDK13 in CTD-S2 phosphorylation (release of Pol II from the proximal pause and elongation of transcription).6 CDK8 and CDK19 are paralogous kinases, included in the mediator kinase module together with the cyclin C, that may positively or negatively regulate Pol II transcriptional activity depending on the target gene, cell type, transcription-cycle stage, etc.6 Whether CTD phosphorylation patterns of Pol II were similar in the case of an RNA template, such as HDV was not explored so far. To investigate if CDK proteins were involved in the HDV replication cycle, Prawira et al.5 screened for small molecules targeting several CDKs. They found no effect of CDK7 and CDK12/13 inhibitors on HDV infection, whereas the CDK9 inhibitor was toxic. However, they report a dose-dependent decrease of HDV infection in various relevant in vitro hepatocyte models with CDK8/19 inhibitors. Using different treatment duration strategies and VSVG pseudo-type HDV particles, they demonstrated that CDK8/19 inhibitors target HDV replication at an early post-entry step of the viral life cycle. CDK8 and CDK19 double knockout (KO) cells (hepatocytes or NTCP-expressing cells unrelated to the liver) were not permissive to HDV infection, contrary to their wild-type (WT) counterparts. Similarly, cyclin C KO cells were resistant to HDV infection, suggesting that the entire mediator kinase module is necessary for productive HDV infection. Rescue assays using WT or dead kinase mutants of CDK8 and CDK19 in CDK8/19 KO cells demonstrated that CDK8 and CDK19 kinase activities are required for productive HDV infection. HDV RNAs quantification demonstrated that MSC2530818, one CDK8/19 inhibitor, reduced the levels of HDV-G, HDV-AG, and HDV mRNAs. Due to the replication strategy used by HDV, levels of each HDV RNA are tightly linked to each other. Consequently, it remains to be confirmed whether CDK8/19 is indeed required for the synthesis of all HDV RNAs using methods to monitor levels of nascent HDV-G, HDV-AG, and HDV mRNAs very early after treatment with CDK8/19 inhibitors. Interestingly, correlating with reduced HDV RNA levels, Prawira et al.5 observed reduced CTD-S2 and CTD-S5 phosphorylation of HDV RNA-associated Pol II upon treatment with MSC2530818. Whether CDK8/19 is directly involved in CTD-S5 and CTD-S2 phosphorylation or indirectly via the phosphorylation of other unexplored kinases also remains to be determined. This data strongly suggests, for the first time, that some CTD phosphorylation patterns of Pol II are conserved during HDV RNA synthesis (compared with canonical RNA synthesis from DNA templates), but related kinases might differ. Indeed, inhibitors blocking the canonical CDK7 and CDK12/13 kinases involved in Pol II CTD-S5 and CTD-S2 phosphorylation, respectively (upon RNA transcription from a DNA template), showed no effect on HDV infection. It was demonstrated that Pol II undergoes subunit remodeling and does not require canonical general transcription factors or elongation factors for successful RNA-templated transcription of plant viroids,7 also highlighting a distinct mechanism from canonical DNA transcription. Considering that viroids share structural features and replication strategies with HDV, it would be interesting to determine whether such Pol II remodeling also applies when HDV RNAs serve as templates and how CDK8 is generally involved in mechanisms requiring a specialized, smaller Pol II complex mediating RNA-templated transcription. In addition, Prawira et al.5 showed that HDV replication can be restored in CDK8/19 KO cells when S-HDAg was expressed before infection, and MSC2530818 treatment barely affected HDV replication in WT cells expressing S-HDAg before HDV infection, supporting the idea that S-HDAg can functionally substitute for CDK8/19 in HDV RNA synthesis. Using different mutants of S-HDAg, they demonstrated that this functional substitution requires S-HDAg–Pol II binding, RNA binding, histone mimicry functions, as well as S-HDAg methylation, but not S-HDAg phosphorylation. Kinetics analyses reported by Prawira et al.5 strongly support the hypothesis that S-HDAg accelerates HDV RNAs replication initiated via CDK8/19 in its absence. Although not as efficient as early treatments, Prawira et al.5 also reported decreased HDV RNA levels of infectious HDV particles production upon late treatments with MSC2530818. Notably, contrary to HDV RNA levels, the percentage of HDAg+ cells was not affected by late treatments with MSC2530818, suggesting a long half-life of HDV proteins. Since the production of L-HDAg has been reported to inhibit HDV replication1 and consequently S-HDAg production, it can be hypothesized that at late times post-infection, when S-HDAg levels decline, low levels of HDV replication are maintained through CDK8/19. Such a mechanism could at least partially explain how HDV persists for weeks in HDV-monoinfected liver chimeric humanized mice, despite the absence of hepatitis B virus infection and thus any possibility of viral propagation.8 This low-level replicative state may thus represent an effective viral strategy to evade innate immune surveillance by the monoinfected cells9 while remaining poised for new virion production upon HBV super-infection. Importantly, transcriptomic analyses performed by Prawira et al.5 did not reveal any impact of S-HDAg expression on the global cellular transcription or on the expression of genes regulated by CDK8/19, suggesting a specific interplay of S-HDAg with the cellular transcriptomic machinery restricted to HDV RNA synthesis. Moreover, recent work has revealed that a broad spectrum of cell signaling kinases (encompassing more than 100 enzymes from diverse families) can target previously unassigned residues within the Pol II CTD.10 It was further suggested that phosphorylation of noncanonical CTD residues may locally modulate polymerase activity, thereby enabling gene-specific regulation without widespread disruption of global transcription.10 In this context, it is tempting to speculate that the specificity of S-HDAg regarding Pol II activation for HDV RNAs only may be the result of S-HDAg promoting the modification of additional, noncanonical residues within the Pol II CTD during HDV transcription. Identifying such CTD modifications and the kinases responsible for them will be an important direction for future studies. Finally, it remains to be determined if CDK8/19/Cyclin C-mediated initiation of HDV replication, as evidenced by Prawira et al., is fully independent of S-HDAg or if S-HDAg molecules brought by the incoming RNP are still required. The authors5 further suggest that their findings may provide a basis for the development of new therapeutic strategies for chronic hepatitis D. In support of this view, they report that (i) pharmacological inhibition of CDK8/19, in combination with the entry inhibitor bulevirtide, exerted a strong antiviral effect on HDV with limited global transcriptional changes in cell culture systems, and that (ii) MSC2530818 monotherapy abrogated HDV replication in HBV/HDV–coinfected uPA-SCID mice (Figure 1). Notably, in both settings, the antiviral effect was assessed after short-term drug exposure (≤7 d). The potential therapeutic application of CDK8/19 inhibition warrants a cautious interpretation. Given the central role of CDK8 and CDK19 in context-dependent host transcriptional regulation, long-term in vivo tolerability, particularly under chronic administration, remains an open question. Subtle functional effects may escape detection in short-term assays but become clinically relevant over time. Notably, CDK8/19 inhibition has previously been explored in oncology precisely to impair cell proliferation and stress resilience, and in viral infection settings, short-term kinase inhibition has been shown to suppress replication by disrupting host metabolic reserve. Thus, although CDK8/19 inhibition represents an intriguing antiviral strategy, its clinical development is likely to be non-trivial and may require an unusually favorable therapeutic window.FIGURE 1: Shedding new light on the chicken-and-egg dilemma of HDV replication initiation and potential clinical application. Abbreviation: HDV, hepatitis delta virus.Overall, the study conducted by Prawira et al.5 elegantly contributes to our understanding of the early phases of HDV replication by providing compelling insights into how HDV replication is initiated (and possibly maintained) when S-HDAg levels are low, thereby solving the chicken-and-egg dilemma (Figure 1).