Exploring Predictive Factors for Bulevirtide Treatment Response in Hepatitis Delta-Positive Patients

Abstract

Background: Hepatitis delta virus (HDV) infection represents the most severe form of viral hepatitis and is a significant global health challenge. Bulevirtide (BLV) is a novel therapeutic treatment that has resulted in variable response rates in HBV/HDV-coinfected patients. We evaluated clinical, virological, and polymorphic factors for the purpose of predicting BLV treatment success. Methods: Thirty HBV/HDV-coinfected patients received BLV monotherapy (2 mg/day) for 24 to 48 weeks. Baseline (BL) serum samples were collected to assess clinical parameters and virological markers (HDV RNA, HBV DNA, HBsAg, HBcrAg, anti-HBc IgG) at treatment weeks 24 (TW24) and 48 (TW48). Additionally, full-genome HDV sequencing and a phylogenetic analysis were performed. Finally, analyses of the HDAg protein sequence and HDV RNA secondary structure were conducted to evaluate potential associations with treatment response. Results: A significant reduction in HDV RNA levels was observed at TW48, with a virological response (HDV RNA undetectable or ≥2 Log decline from BL) achieved by 58% of patients. Median BL levels of anti-HBc IgG were significantly different between virological responders (39.3 COI; interquartile range [IQR] 31.6-47.1) and virological non-responders (244.7 COI; IQR 127.0-299.4) (p = 0.0001). HDV genotype 1e was predominant across the cohort, and no specific HDAg polymorphisms predicted the response. However, secondary structure analysis of HDV RNA revealed that a specific pattern of internal loops in the region 63-100 nucleotides downstream of the editing site may influence treatment response by impacting editing efficacy. Conclusions: This study revealed key factors influencing BLV efficacy in HBV/HDV coinfection. Lower baseline anti-HBc IgG levels strongly correlated with a positive virological response, suggesting that the liver's inflammatory state affects treatment success. Additionally, the analysis of HDV RNA secondary structure in patients receiving BLV treatment revealed a higher editing efficiency in virological responders, highlighting areas for further research.

Keywords: Anti-HBc IgG; HDV; antiviral treatment; bulevirtide; genetic variability.

Figure 1

Virological kinetics. Median HDV RNA levels are shown with the interquartile range at different study time points for the entire patient cohort. The whiskers represent the interquartile range. Baseline, BL.

Figure 2

Box plot of immunoglobulin G antibody to hepatitis B core antigen (Anti-HBc IgG) values in v-responders (individuals who achieved a virological response) and v-non-responders (individuals who did not achieve a virological response). The box plot depicts anti-HBc IgG values for patients at baseline (BL), 24 weeks (TW24), and 48 (TW48) weeks of BLV treatment. V-responders are represented in blue. V-non-responders are represented in red. The whiskers indicate minimum and maximum values. v-R-BL, v-responders at baseline; v-NR-BL, v-non-responders at BL; v-R-TW24, v-responders at TW24; v-NR-TW24, v-non-responders at TW24; v-R-TW48, v-responders at TW48; v-NR-TW48, v-non-responders at TW48; ***, p = 0.0001; **, p = 0.0012.

Figure 3

Bayesian phylogenetic tree. The tree was constructed using fragment A, which was available for all 30 patients (see Materials and Methods, Section 2.5). Patient-derived BL sequences are colored blue (blue circles represent virological responders and red circles represent virological non-responders), while reference sequences are colored red (subtype 1b), orange (subtype 1c), and green (subtype 1e). As there were no patient sequences clustering with lineages for genotypes 2 through 8, as well as genotype 1 subtypes 1d and subtype 1a (partially), as indicated by a high posterior probability, they were collapsed for convenience. Sequences were aligned using MAFFT v7.511, the phylogenetic tree was inferred using BEAST v2.7.3 and Tree Annotator v2.7.3, and tree visualization was implemented with Figtree v1.4.4 and Adobe Illustrator 2025.

Figure 4

(a) Representation of baseline hepatitis delta antigen (HDAg) protein sequences derived from 14 HBV/HDV-positive patients who achieved virological response (v-responders) and (b) from those (n = 10) who did not achieve virological response (v-non-responders). Post-transcriptional modification (PTM) sites are indicated by circles, while functional domains are represented by boxes. Sequence logos were generated using WebLogo 3. (c) The line plot illustrates the Shannon entropy profiles for two groups of sequences derived from the HDAg protein: v-responders (indicated in blue) and v-non-responders (indicated in red). The x-axis corresponds to the amino acid positions along the HDAg sequence, while the y-axis represents the corresponding Shannon entropy values. P, phosphorylation; M, methylation; A, acetylation; I, isoprenylation; RBD, RNA-binding domains; CCD, coiled-coil domain; NLS, nuclear localization sequence; VAS, virus-assembly signal.

Figure 4

(a) Representation of baseline hepatitis delta antigen (HDAg) protein sequences derived from 14 HBV/HDV-positive patients who achieved virological response (v-responders) and (b) from those (n = 10) who did not achieve virological response (v-non-responders). Post-transcriptional modification (PTM) sites are indicated by circles, while functional domains are represented by boxes. Sequence logos were generated using WebLogo 3. (c) The line plot illustrates the Shannon entropy profiles for two groups of sequences derived from the HDAg protein: v-responders (indicated in blue) and v-non-responders (indicated in red). The x-axis corresponds to the amino acid positions along the HDAg sequence, while the y-axis represents the corresponding Shannon entropy values. P, phosphorylation; M, methylation; A, acetylation; I, isoprenylation; RBD, RNA-binding domains; CCD, coiled-coil domain; NLS, nuclear localization sequence; VAS, virus-assembly signal.