A Hidden Hepatitis B Problem May Depend on Combinations of Mutations, Not Just Single Genetic Changes
A new review is challenging a widely used way of thinking about occult hepatitis B infection, a form of infection that can remain invisible to routine screening even while the virus persists in the body. The focus is hepatitis B virus genotype D, the dominant HBV lineage across much of the Middle East, where it provides the genetic background for diagnostic testing, blood-screening strategies and vaccination programmes. The review argues that individual mutations in the virus’s surface protein cannot fully explain why some infected people test negative for hepatitis B surface antigen, or HBsAg. Instead, combinations of mutations may interact in unexpected ways, altering the virus’s ability to produce, secrete or display the antigen. Such interactions, known as epistasis, could help explain how replication-competent HBV DNA survives in people whose standard blood tests appear negative. However, the researchers emphasise that this is a plausible hypothesis, not an established mechanism.
Occult hepatitis B infection, or OBI, is defined by the presence of replication-competent HBV DNA in a person who tests negative for HBsAg, the viral marker most commonly used to identify active infection. Viral levels in OBI are often very low, which makes detection difficult and can allow cases to evade routine screening. The condition matters because apparently uninfected blood or organs may still carry transmissible virus, while immunosuppressive treatment can sometimes reactivate a quiet infection. HBV is a partially double-stranded DNA virus that replicates through an RNA intermediate, a process that permits genetic variation despite its DNA genome. Changes in the gene encoding the surface proteins can affect how antigens fold, assemble into particles or interact with antibodies. Some mutations are known to produce “immune escape” or “diagnostic escape,” enabling viral particles to avoid antibody recognition or reducing the amount of detectable HBsAg. Yet single substitutions explain only part of the clinical and laboratory variation observed in occult infection.
The review centres on the possibility that mutations may work as a network rather than as isolated alterations. In genetics, epistasis occurs when the effect of one mutation depends on the presence of another. Two changes can reinforce each other, cancel each other out or generate a new phenotype that neither produces alone. For HBV, a mutation in one portion of the surface protein could alter the local structure in a way that changes the effect of a second mutation elsewhere. The combined result might influence the stability of the protein, its movement through the endoplasmic reticulum, its incorporation into viral or subviral particles, or its exposure to antibodies. A cluster might therefore reduce HBsAg secretion without eliminating viral replication. That distinction is crucial: a virus can remain biologically active while becoming harder to identify through an antigen-based assay. The authors propose that this type of non-additive interaction could be especially important in genotype D, but they find that direct evidence remains sparse.
To separate observation from speculation, the researchers establish a four-level evidence hierarchy. The first level consists simply of mutations appearing together in sequence data, a pattern called co-occurrence. Co-occurrence can be suggestive, but it does not prove that the mutations interact biologically. The second level requires phylogenetically corrected covariation, which asks whether mutations continue to be associated after accounting for the fact that closely related viruses inherited the same changes from a common ancestor. Without this correction, researchers can mistake shared ancestry for functional cooperation. The third level requires an experiment comparing a candidate mutation cluster with each of its constituent single mutations on the same defined genotype D genetic background. The fourth and strongest level demands mechanistic validation, combining functional data with structural or biochemical evidence to explain how the mutations produce their effect. According to the review, existing studies reach only the first or second level.
That evidence gap is significant because viral sequence databases can create an illusion of certainty. If several substitutions repeatedly appear in the same HBV genomes, they may seem to form a functional unit. But the pattern could instead reflect founder effects, regional transmission chains, sampling bias or the evolutionary history of a particular genotype. Genotype D-specific datasets large enough for reliable cluster-level analysis are limited, and Middle Eastern datasets are even more restricted. The problem is compounded by the fact that many studies sequence only a portion of the viral genome, examine small numbers of patients or lack detailed clinical information. A mutation observed in an HBsAg-negative patient may be relevant, but it may also be incidental or linked to another unmeasured change. The review therefore calls for larger, geographically diverse datasets, full-length or carefully targeted sequencing, and statistical models that correct for phylogeny, recombination and population structure before any mutation cluster is labelled epistatic.
The authors also examine whether HIV or hepatitis C virus coinfection could shape the emergence of mutation combinations. Coinfections can alter immune pressure, treatment exposure and the cellular environment in which HBV replicates. In theory, these conditions might favour viral variants that would be rare during HBV infection alone. Patients receiving antiviral drugs may also experience selection pressures that change the composition of the viral population over time. However, the review finds no direct evidence linking HIV or HCV coinfection to the selection of particular HBV surface-protein mutation clusters associated with OBI. This absence does not rule out an effect; it indicates that the relevant studies have not yet been performed or have not included enough well-characterised patients. Demonstrating such a relationship would require longitudinal sampling, reliable coinfection histories, treatment data and deep sequencing capable of detecting low-frequency viral variants rather than only the dominant sequence in a blood sample.
The decisive experiment proposed by the researchers is deliberately straightforward in principle, although technically demanding in practice. Scientists would reconstruct candidate mutation clusters by site-directed mutagenesis on a defined genotype D backbone, creating viral or expression constructs containing each individual mutation as well as the full combination. These matched constructs could then be tested for HBsAg production and secretion, intracellular retention, antibody binding and viral replication. Assays would need to distinguish between reduced antigen synthesis and impaired release of the surface protein. They would also need to measure whether a mutation cluster changes the ratio of viral particles to non-infectious subviral particles, which are abundant HBV surface-protein structures released from infected cells. A genuine epistatic effect would be demonstrated if the combined mutations produced a phenotype that could not be predicted from the single mutants, such as a disproportionately large loss of HBsAg detectability while replication remained measurable. Repeating the comparison across independent genotype D backbones would help establish that the result is not an artefact of one laboratory strain.
Structural biology could reveal why a mutation cluster behaves differently from its individual components, but the review warns that current tools have limitations. Recent cryo-electron microscopy structures of HBsAg subviral particles have improved understanding of the surface-protein architecture, yet the flexible loops forming the so-called “a” determinant are not resolved at the detail required to interpret many cluster-level conformational changes. The a determinant is the principal immunological region recognised by antibodies generated through infection or vaccination. Small structural shifts there could affect assay binding without necessarily changing the virus’s capacity to replicate. But modelling such shifts from unresolved or low-resolution regions remains an inference, not a direct observation. The researchers suggest combining improved structural methods with biochemical measurements, antibody panels and computational modelling. A credible mechanism would need to connect the sequence changes to a measurable change in folding, oligomerisation, membrane insertion, secretion or antibody accessibility.
The implications extend beyond a single viral genotype or one region of the world. If combinations of mutations contribute to diagnostic escape, routine assays designed around well-characterised individual variants may fail to recognise some infections, particularly when viral loads are low. That could complicate blood and organ safety, clinical decisions before immunosuppression and surveillance of chronic HBV transmission. It could also influence how laboratories interpret apparently discordant results, such as detectable HBV DNA alongside a negative HBsAg test. Still, the review does not claim that current screening systems are broadly ineffective, nor does it establish that epistatic clusters are enriched among Middle Eastern populations. Its central message is methodological: a mutation pattern should not be promoted to a causal explanation until it survives phylogenetic analysis and controlled functional testing. By defining the experiments needed to move from sequence association to biological proof, the authors turn an intriguing explanation for occult infection into a testable research programme. The outcome could sharpen diagnostic design while clarifying how HBV evolution operates beneath the threshold of conventional detection.
