The study protocol was approved by the four study sites research ethics committees, namely: Jos University Teaching Hospital, Faith Alive Foundation, Bingham University Teaching Hospital and Plateau State Specialist Hospital (study approval numbers: JUTH/ADM/11/0517, FA/0112F, BUTH/EC/01-25-18B and PLASHREC/12018). After explaining the study objectives and procedures to the participants, a written informed consent was obtained from each patient before data and sample collection. The patients’ data were coded and reported without divulging personal health information. All other aspects of the study were conducted in accordance with the ethical standards of the Helsinki Declaration.

We conducted a cross-sectional study between February 2018 and May 2019 at Jos University Teaching Hospital, Plateau State Specialist Hospital, Faith Alive Foundation and Bingham University Teaching Hospital, all in Jos metropolis, Nigeria. The 101 consecutively enrolled adult patients, including 51 patients co-infected with HIV, were confirmed to have chronic HBV infection after testing positive for hepatitus B surface antigen antibodies. For HBV sero-diagnosis, 5 mL of blood sample was collected, centrifuged at 2500 revolutions per minute for 1 min, and 1 mL of plasma was used for testing HBsAg by third-generation enzyme-linked immunosorbent assay (Monolisa, Bio-Rad, Paris, France) according to manufacturer’s instruction; a positive HBsAg result indicates current infection. All participants were on antiviral therapy (HBV/HIV: lamivudine/tenofovir combination, HBV-mono: tenofovir alone) ≥ 12 months.

The 5 mL of blood samples collected were centrifuged, and the sera separated, aliquoted, and stored at –80 °C until use. The HBV DNA detection and quantification were performed using automated COBAS^® AmpliPrep TaqMan HBV Test version 2.0 (Roche Diagnostics International AG, Rotkreuz, Switzerland). A total of 30 samples with detectable viral load were selected for DNA sequencing reactions.

DNA extraction, amplification and direct sequencing of the full-length HBsAg and the reverse transcriptase genes were performed at Center for Human Virology and Genomics, Nigeria Institute of Medical Research, Lagos, Nigeria. Viral DNA was extracted from 200 µL of plasma samples using the ZR Viral DNA kit^TM (Zymo Research, Irvine, California, United States), according to the manufacturer’s protocol.

Universal primers from previous study were used for detection of HBV genotypes,²³ which were synthesised by Inqaba Biotech (Pretoria, South Africa). Two outer primer sets were used. Primer set 251F (5ʹ-GGA TGT GTC TGC GGC GTT T-3ʹ) and 1797R (5ʹ-GAC CCA CAA TTC KTT GAC ATA CTT TCC-3ʹ) were used for the initial amplification, and 251F (5ʹ-CGA ACC ACT GAA CAA ATG GC-3ʹ) and 1190R (5ʹ-TCA CCA TAT TCT TGG GAA CAA GA-3ʹ) used for the second amplification, as previously described.²³ These primer sets were designed to generate overlapping fragments, and polymerase chain reaction (PCR) was performed using Platinum™ SuperFi™ PCR Master Mix (Thermofisher, Waltham, Massachusetts, United States). Each 25.0 µL reaction mixture contained 12.5 µL of 2× premix, 1.25 µL of each 10.0 µM primer, 5.0 µL of PCR water and 5.0 µL DNA template for first PCR while for the second PCR, 3.0 µL of first PCR product was used. Both PCR were carried out in Applied Biosystems MiniAmp Plus Thermal Cycler (ThermoFisher, Waltham, Massachusetts, United States) with the following cycling conditions: 95 °C for 3 min followed by 35 cycles of 97 °C for 30 s, 54 °C for 60 s and 72 °C for 60 s for the first round of PCR while 31 cycles were done for the second round of PCR. Each round of PCR was followed by a final extension of 72 °C for 10 min. Amplified products were verified using 1% agarose gel stained with GelRed^® (Biotium, Fremont, California, United States) on a Cyfox Gel Documentation System (Sysmex-Partec, Münster, Germany). The amplified products were purified using ExoSAP-IT (Thermofisher, Waltham, Massachusetts, United States) and quantified using dsDNA high-sensitivity (Thermofisher, Waltham, Massachusetts, United States) reagent on the Qubit-4^TM (Thermofisher, Waltham, Massachusetts, United States). Samples with DNA concentration ≥ 5 ng/µL and gel bands of appropriate size were selected for further processing.

Verified nested PCR products were purified using ExoSAP-IT^TM PCR Product Cleanup reagent (Thermofisher, Waltham, Massachusetts, United States) at 37 °C for 15 min (to digest excess primers and deoxynucleoside triphosphates), and 80 °C for 15 min (to inactivate the enzymes) according to manufacturer’s protocol. The sequencing reaction mixture contained 1.0 µL of BigDye® Terminator version 3.1 Ready Reaction Mix (Applied Biosystems, Foster City, California, United States), 4.0 µL of 5× sequencing buffer, 2.0 µL of template, 11.0 µL of double distilled water, and 2.0 µL of 10.0 µM 251F and 1190R primers. Cycle sequencing was performed using 251F and 1190R primers in a MiniAmp Plus Thermal Cycler with the following program: 25 cycles of 96 °C for 10 s, 50 °C for 5 s and 60 °C for 4 min. The cycle sequencing product was purified using the BigDye Xterminator® Solution (Applied Biosystems, Waltham, Massachusetts, United States) followed by the capillary electrophoresis on the 3130XL Genetic Analyzer using a 50 cm capillary array, and POP-7 polymer (Applied Biosystems, Waltham, Massachusetts, United States). All the generated sequences were analysed and assembled using CLC Genomic Workbench version 8.0.3 (CLC Bio, Aarhus, Denmark) and then subjected to NCBI nucleotide BLAST (https://blast.ncbi.nlm.nih.gov/Blast.cgi) for quality check. In this study, a chromatogram was considered of good quality if (1) at least 80% of total amplicon length was sequenced and (2) the noise-to-signal ratio was estimated to be < 5%.

To infer the evolutionary relationship between the six HBV nucleotide sequences in our study and HBV sequences from different parts of the world, we used Molecular Evolutionary Genetics Analysis (MEGA) X version 10.1.7 (Mega Software, Dortmund, Germany).²⁴ First, a total 102 sequences (including a sequence from a chimpanzee) representing the HBV genotypes from different parts of the world were mined from the NCBI GenBank nucleotide database (https://www.ncbi.nlm.nih.gov/nucleotide/). The sequence from the chimpanzee was used as an out-group. The generated nucleotide sequence data were deposited in the GenBank database under the accession numbers ON236588–ON236593.

These sequences, together with the six from our cohort, were aligned using MUltiple Sequence Comparison by Log-Expectation (MUSCLE) software,²⁵ which is a sequence alignment option in MEGA. The multiple sequence alignment was trimmed to 846-base pairs, followed by determination of the DNA model that best describes the nucleotide substitution pattern. This was performed using the maximum likelihood (ML) statistical method. The General Time Reversible (GTR) model with discrete Gamma distribution (+G = 0.51, with 5 rate categories) and some invariant evolutionary sites (+I = 35.51% sites) was chosen as the best model, based on its low Bayesian information criterion score (18 980). For the phylogeny reconstruction the ML statistical method was used with a bootstrap of 1000 replicates and the sub tree-Pruning-Regrafting (Extensive; SPR level 5) as the ML heuristic method. Moreover, we used the option of ‘moderate branch length filter’ to enable a semi-stringent exhaustive optimisation of the branch lengths of the phylogenetic trees. Initial trees for the heuristic search were generated using Neighbor-Joining and BioNJ algorithms and the phylogenetic tree with superior log likelihood value selected. Phylogenetic trees were produced in Newick format employing interactive Tree of Life interface (https://itol.embl.de/tree/).

To determine the genetic diversity of the six HBV sequences, we performed an alignment analysis using MEGA.²⁴ The overall mean (average) and maximum distance were used as proxies for genetic diversity.

The obtained sequences of overlapping surface (S) and pol gene were translated to the protein sequences and aligned with the references in BioEdit version 7.1.3.0.²⁶ Escape mutant analysis and drug resistance analysis were performed using HBV Drug Resistance Interpretation and Geno2pheno (HBV) version 2.0 (https://hbv.geno2pheno.org/index.php). It is a program that searches for homology between the input sequence and other DNA sequences in the existing stored database for drug resistance and surface gene mutations. Genotype assignments were confirmed using the GenBank BLAST and HBVseq tools from the HIV Drug Resistance Database.^26,27

The obtained data were analysed using Statistical Package for Social Science software (version 19.0; IBM Corp., Armonk, New York, United States) and expressed as medians (with interquartile ranges [IQR]) for continuous variables, and as counts and percentages for discrete variables. Chi-squared test was used for discrete data.

All the generated sequences were from the HBV-mono infected patients. The genotypes of the six HBV nucleotide sequences were inferred from the phylogenetic tree (Figure 1), which showed that all the sequences were HBV genotype E. All the six sequences clustered with HBV genotype E, most of which were from the West African region.

According to the Geno2pheno (HBV) 2.0 report, all the six HBV genotype E sequences were susceptible to lamivudine, adefovir, entecavir, tenofovir disoproxil fumarate, and telbivudine. Two (33.3%) of the six genotypes, all from HBV-mono infected individuals, had escape mutations, the SH2 domain-containing adapter protein B protein of the surface-gene. One sample had L127R escape mutation while the other had G145A at a position first described as vaccine escape mutation. These mutations are amino acid substitutions within the major hydrophilic region called the ‘a’ determinant (124–147).

There are limitations to this study that need to be highlighted. It is a cross-sectional study with a small sample size, and low amplification and sequencing rate, yielding only 20.0% of the intended samples. Many factors could be responsible for this. For example, several freeze-thawing cycles and poor storage due to frequent power outage could have affected the sample integrity, resulting in decreased HBV DNA quantity. Other causes include intrinsic reagent-related issues and impact of HBV/E on the sensitivity of HBsAg assays due to surface antigen gene mutations located outside the sequenced region.⁴⁶ Also, there is a growing concern that the accumulation of escape mutations may lead to detection failure of HBsAg using the available commercial diagnostic assays. This in turn may lead to a false-negative HBV result. Moreover, this has an impact on risk of transmission, including rendering the vaccination ineffective due to undiagnosed mutant carriers.

Our study confirmed the dominance of genotype E, and it has great public health significance, given that the protective vaccine currently in use by the national vaccination programme is from the HBV-A2 strain. Additionally, no patients had HBV drug resistant mutations, suggesting that the use of tenofovir is an effective treatment in the management of HBV infection in both chronic HBV/HIV co-infected and HBV-mono infected patients. The study also showed the presence of mutation associated with immune escape mutant in chronic HBV-infected patients on long-term antiviral therapy. Therefore, further larger study is required to understand the dynamics of immune escape mutant on HBV diagnosis, and impact of genotypes on treatment outcomes.