Genotyping and subgenotyping of circulating HBV isolates in the samples analyzed (Table ) showed no significant differences in genotype distributions among acute and chronic infections under LAM selective pressure. In acute infection, 16 of 20 HBV isolates (80%) under study belonged to genotype A, three (15%) were from genotype D, and the remaining one (5%) belonged to genotype F. In samples from chronic cases, the following genotype distribution was found: 25/44 (56.8%) genotype A, 13/44 (29.5%) genotype D, 5/44 (11.4%) genotype F, and one (2.3%) genotype B. Among isolates from genotype A, subgenotypes A1 and A2 were found. The ratio of subgenotypes A1/A2 in acute cases (8/8, 50% each) was significantly different from that in chronic cases (22/3, 88% A1 and 12% A2; P
0.012). If the equal distribution of subgenotypes A1 and A2 among newly infected individuals (acute infection) reflects an increase in subgenotype A2 in Brazil, this suggests that the profile of circulating subgenotypes in Brazil could be changing. Alternatively, differences between the two subgenotypes could be related to disease progression (resolution of acute infection or progression to chroni-city). These possibilities warrant further investigation.
Comparisons of YMDD variants in serum of patients with acute and chronic HBV infection detected by direct sequencing and pyrosequencing
Surprisingly, acute HBV patients had relatively low HBV titers compared to what would have been expected for an acute HBV infection, ranging from 6.2 x 102 to 1.4 x 106 copies/mL (mean viral load, 2.0 x 105 copies/mL; median viral load, 2.0 x 104 copies/mL). Chronic patients had HBV titers ranging from 9.4 x 102 to 2.4 x 109 copies/mL. The mean viral load was 1.4 x 105 copies/mL, and the median was 5.6 x 104 copies/mL.
The direct PCR Sanger sequencing method (population-based sequencing approach) detected only the major population in our assays. Literature reports indicate that only minor populations present as more than 20% of the total quasispecies pool can be detected by the Sanger method [26
]. To test the ability of pyrosequencing to detect minor sequence variants of the YMDD population, we evaluated different proportions of plasmids containing WT (rtM204) and MUT (rtV204) sequences. Allelic quantification based on pyrograms indicated accurate detection when minor variants represented at least 5% of the total circula-ting population (Figure ). A value of 4% was subsequently used as the lower limit of detection of minor populations by pyrosequencing under our experimental conditions. Other methods capable of detecting (but not capable of quantifying) viral mutants that constitute as little as 4-5% of the total population, such as RFLP analyses and line-probe assays, are either labor intensive and thus not suitable for high-throughput screening, or are subject to a greater number of false-positive and false-negative results than sequencing [7
]. Our result confirms previous reports that pyrosequencing is the most sensitive method available for detecting small subpopulations of resistant virus and, as such, is likely to become the method of choice in the near future [7
Figure 1 Pyrosequencing analysis with allelic quantification of A/G for the first position of codon M/ATG and V/GTG in different mixtures of WT (YMDD) and MUT (YVDD) plasmids. (A) 100% WT-0% MUT; (B) 50% WT-50% MUT; (C) 66% WT33% MUT; (D) 90% WT-10% MUT; (E) 95% (more ...)
Comparisons of YMDD variants in serum of patients with acute and chronic HBV infection detected by direct sequencing and pyrosequencing are shown in Table . As expected, none of the individuals with acute hepatitis B had LAM-resistant isolates as a dominant virus population, whether detected by direct sequencing or pyrosequencing. However, because of its greater ability to detect viral subpopulations, pyrosequencing revealed that 11/20 (55%) of the individuals with acute hepatitis B had only WT isolates, whereas 9/20 (45%) had minor subpopulations of LAM-resistant isolates varying from 4% to 17%. The detection of pre-existing resistant variants in acute phase provides information helpful in choosing an appropriate antiviral regimen whether individuals have become chronic carriers, and thus need to start an antiviral regimen.
Thirty-eight patients (86.4%) with chronic hepatitis B were undergoing a LAM monotherapy regimen, whereas the other six (13.6%) were receiving combination therapy of LAM plus adefovir dipivoxil (ADV) or tenofovir disoproxil fumarate (TDF). There was no significant association between the treatment duration and the occurrence of LAM-resistant isolates. Direct sequencing methods determined that WT isolates were present in 19 of 44 patients (43.2%) and LAM-resistant isolates were present in 25 of 44 patients (56.8%), with a predominance of the YVDD variant (17/25, 68%) compared to the YIDD variant (8/25, 32%). Pyrosequencing confirmed the presence of exclusively WT isolates in 10 of 19 samples (52.6%) characterized as WT by direct sequencing. In the other nine samples (47.4%), pyrosequencing was able to detect the presence of minor subpopulations of LAM-resistant isolates. Of 25 samples characterized as LAM-resistant by direct sequencing, pyrosequencing confirmed the presence of only one population of resistant mutants (either YVDD or YIDD) in 14 (56%). The remaining 11 samples (44%) contained a mixture of mutant variant populations or a minor subpopulation composed of WT isolates. Pyrosequencing proved to be a powerful tool for detecting co-circulating strains in a complex population. This allowed resistant HBV to be detected before any evidence of virological or biochemical breakthrough, thus increasing the possibility of a correct choice of rescue therapy and increasing the likelihood of successful treatment.
Interestingly, all but two individuals whose major virus population was composed of WT isolates and a small percentage of resistant variants detected by pyrosequencing had a YIDD variant as a minor subpopulation, suggesting that the rtM204I mutation may naturally occur more often and replicate more efficiently than YVDD variants in environments with little or no selection pressures. The only disagreement between the results of direct sequencing and pyrosequencing was for sample NN124. The direct sequencing method detected nucleotides (GTG) coding for rt204V, although the electropherogram indicated mixtures with small quantities of nucleotides A and T corresponding to the first and third position, respectively, of codon rt204I (Figure ). In contrast, pyrosequencing indicated a majority (~60%) of rt204I variant and about 40% rt204V variant (Figure ). The same discrepant results were also obtained when the segment used as template for the direct sequencing method was amplified using pyrosequencing primers. This disagreement may be attributable to the similar amounts of YIDD and YVDD variants (60% vs. 40%) reported by pyrosequencing.
Figure 2 Discrepancy between direct sequencing and pyrosequencing in sample NN124. The direct sequencing method (A) detected the nucleotides (GTG) coding for the rtM204V variant, although the electropherogram indicated mixtures with small quantities of nucleotides (more ...)