Hepatitis B virus (HBV) infections are considered an important public health issue, particularly due to their endemicity in many countries in the world. Between 350 and 400 million people worldwide are infected with HBV and show signs of chronic HBV infection. The occurrence of coinfections between human immunodeficiency virus (HIV) and HBV is very common, mainly because of the similar infection routes that both viruses share, and according to the Global Hepatitis Report 2017 (World Health Organization), the prevalence of HBV infection in HIV-infected persons is 7.4% [1].

The main serological marker for the detection of HBV is the surface antigen (HBsAg) and the persistence of this antigen in the blood of infected patients for more than six months, defines the chronic phase of HBV infection the disease. In contrast, occult hepatitis B virus infection (OBI) is defined by undetectable levels of the HBsAg or anti-HBsAg antibodies in the blood of infected individuals accompanied by the presence of either, episomal HBV covalently closed circular DNA (cccDNA) in the liver, or detectable levels of HBV DNA in the blood [2].

There is a clear correlation with the prevalence of OBI and the level of endemicity of HBV in different geographic regions. Interestingly, highly variable levels of OBI prevalence, ranging between 0 and 90% have been reported, depending on the HBV genotype and the method used for the detection [3–5]. It is worth mentioning that HIV positive patients show a higher risk of developing OBI [4,6].

In Mexico, despite the low prevalence of chronic HBV infection (0.21%) and the high level of vaccine coverage (92%) [7–9], mandatory in the national vaccination scheme since 1999, two studies regarding the presence of OBI in HIV positive patients have been published. The first study by Torres-Baranda et al. reported an OBI prevalence of 18.4% in HIV-positive patients using only one molecular marker (S gene) and a nested PCR assay designed to detect the DNA-HBV in serum samples [10]. In the second study, we reported an OBI frequency of 49% using two out of three positive molecular markers as a criteria to detect OBI and a nested PCR to amplify the Core region and two real time PCR assays to amplify the X and S genes in plasma samples [11].

The molecular mechanisms associated to the development of OBI are still unknown, nevertheless, several hypotheses have been proposed to explain this phenomenon. Previous studies have suggested that mutations in the PreS/S ORF that codes for the HBsAg might be related to the lack of detection of the HBsAg in the blood of infected patients, possibly due to antigenic changes that reduce the affinity of diagnostic antibodies toward this protein or to the reduced secretion of the HBsAg into the bloodstream [12,13]. To date, few studies have reported the molecular features of OBI in Mexico. In this study, we describe the frequency and the mutations in the S gene of occult HBV in patients with HIV.

In this study, we report a frequency of occult hepatitis B virus infections of 36% in a cohort of 50 Mexican patients infected with HIV. Previous reports worldwide have demonstrated that prevalence data is variable, ranging from 1 to 89% [4]. In Mexico, only two studies have reported the frequency of OBI in HIV positive patients. Torres R et al. reported a prevalence of 18.4% (7/38) in HIV patients from Jalisco, using as a criteria for OBI infection the positivity with only one molecular marker, the S-gene [10]. The second report, from our research group, reported a prevalence of 49% in HIV patients from Mexico City, using as a criteria the positivity of at least two of three molecular markers [11]. The differences reported in these two studies might be associated with several factors which include the level of endemicity, the inclusion criteria used in the studies, the size and type of sample, the sensitivity and specificity of the assays, as well as the sets of primers used for the detection of HBV genome [15–19].

The golden standard for the confirmation of OBI is the detection of HBV-DNA in serum or liver tissue samples, and requires assays that demonstrate to be highly specific and sensitive, due to the fact that HBV DNA in OBI patients is found in very low concentrations (<200IU/mL), moreover, previous studies have demonstrated that intermittent detection of HBV-DNA in OBI patients is not a rare event [2,20].

In this study, based on the Taormina criteria [2], we determined the frequency of occult hepatitis B virus infections in a group of HIV positive patients from a tertiary care hospital in Mexico City, using three standardized nested PCR assays to amplify the Core, PreS and S genomic regions of HBV genotype H. These assays showed a limit of detection of around 100copies/mL (approximately 20 UI/mL), which is consistent with previous reports that demonstrated that the viral load in OBI patient is about 20IU/mL [17]. In spite of there are several commercial kits to quantify the HBV viral load by quantitative PCR, such as the COBAS AmpliPrep-COBAS TaqMan 48, which has a lower detection limit of 10–20 UI/mL [21], to date, they have not been validated for the detection of OBI.

OBI infections are multifactorial entities and although several mechanisms for the development of OBI have been suggested, none of them have been fully demonstrated. Among them, we can mention co-infections with other viruses such as HCV and HIV, factors related to altered immune responses, genome integration of some regions of the HBV that could be associated with the absence of HBsAg and the low level of virus replication, and mutations in all the HBV genome regions [22,23].

Most of the mutational analysis of the HBsAg that have been carried out to search for possible associations between them and the development of OBI, nevertheless, they have only focused on the MHR region of the S gene, particularly in the “a” determinant, a highly conserved antibody-neutralizing epitope cluster [24–26]. Some of these mutations have been associated to the lack of detection of the HBsAg due to structural and antigenic changes in the protein, causing the common tests to throw back false negative results. Moreover, mutations in other sited of the HBsAg might be causing alterations in the maturation, secretion and expression of the HBsAg, highlighting the importance of analyzing the whole protein in search for mutations [27–29].

We aimed to determine if mutations in the HBsAg, in this cohort of Mexican patients also infected with HIV, might be associated with the development of OBI. Considering that during the HBV replication, genetic variants or viral quasispecies are generated [30], the S gene sequences from ten OBI patients were amplified and cloned. We did not search for specific mutations associated to OBI; our approach was to compare the S gene sequences obtained from OBI patients with the set of three reference sequences genotype H from the NCBI genotyping tool. Then, the nonsynonymous mutations were listed and the frequency among clones in one sample, as well among all the samples was determined.

The cloning strategy resulted to be useful to confirm that several quasispecies of HBV can be found in one single patient. These quasispecies are not identical but are related between them (see Supplementary Fig. 1, Suppl. Fig. 2, Table 3). Moreover, we identified several mutations in all the domains that constitute de ORF-S. A total of four non-synonymous mutations were identified: one in the TM1 domain (C19Y), two in the MHR (Q129H, E164D) and one in the CL-2 (I195M).

The C19Y mutation was found with a frequency of 44%, and it is located in a region that is necessary for the insertion and correct folding of the S protein during its translation in the endoplasmic reticulum [31]. If any of these mutations is present in the TM1 domain, it could have an effect on the secretion of the HBsAg and the HBV virions during replication cycle [32,33]. The C19F mutation was also reported in a Mexican patient from the Nahua genotype H with OBI by Roman et al.[34]. Even though this study was descriptive, it is interesting that this mutation was found in an HIV negative patient, suggesting that his mutations might also be present in other groups of patients with OBI. The substitution is F19V was reported in a group of OBI donors with HBV genotype C, nevertheless, since it was found in a different genotype, the amino acid substitution is different. The authors did not report any function or association with OBI [35].

The Q129H substitution located within the “a” determinant of HBsAg, was found in 36% of the OBI patients. This mutation, described as an escape mutation [36], has also been reported in other genotypes but mainly in blood donors [27,37–39]. This mutation was reported in 5/21 blood donors recently diagnosed with HBV in Mexico [40]. It is interesting that this escape mutant has not been reported in OBI patients with genotype H in Mexico. More studies are needed to understand the role of these mutations in the OBI mechanisms.

Other frequent mutation found in the MHR was E164D (39%). This mutation is often found as a double mutation along with I195M. These mutations alter the antigenicity of HBsAg and change its affinity for anti-HBsAg specific antibodies. Some of these mutations, for example, those corresponding to the I195M, W196S, and the M198I mutants, affect a conformational region of the HBsAg protein consisting of two short alpha helixes separated by a disulphide bridge forming a turn [41]. Besides, this substitution can produce simultaneous mutations in both, the overlapping surface antigen, and the polymerase ORF's. When is present as a double mutation (sE164D/sI195M), it has been associated with the triple lamivudine resistance polymerase mutation pattern (rtV173L, rtL180M, rtM204V).

We detected the double mutation, sE164D/sI195M, in six OBI patients. Of these, two (OBI-41, OBI-62) included lamivudine (LMV), and four patients (OBI-50, 54, 75,106) tenofovir in their ARV treatment for HIV infection.

One patient (OBI-112) had only the I195M mutation, which simultaneously produces the rtM204V, that induces resistance to LMV. Because this patient was naive to treatment, this drug resistance mutation could be a naturally induced genetic variant of HBV. This phenomena has been previously reported in an untreated Mexican patient with chronic HBV infection genotype H and coinfected with HIV, probably due to immune pressure as suggested by the author [42].

We found two patients (OBI-43 y 44) that were under treatment with LMV and did not show resistance-associated mutations. In this case, some studies have suggested that these mutations might be found in a minority of the HBV quasispecies from a patient (<20%), thus to detect these mutations, highly sensitive cloning methods are required [43]. Nevertheless, the cloning and direct sequencing method used in this study has a low sensitivity for the detection of underrepresented variants (<20%), thus, we cannot rule out some genetic variants of HBV that might be present in a single sample could have not been detected.

This study has some limitations: the number of patients included (50), it would be necessary to analyze more HIV positive patients to associate the presence of these mutations with the OBI infection, and the association with clinical and demographic variables. We were not able to clone eight OBI samples, probably due to some technical issues and in consequence, these OBI samples were not included in the mutational analysis. The other limitation is the number of clones analyzed, ten per sample. In other studies it has been suggested to analyze up 100 clones per sample, or to use of more sensitive methods such as high throughput sequencing and ultradeep pyrosequencing as an alternative to study the genetic variability of HBV in a single sample [43,44].

In the last years, research carried out to detect occult hepatitis B virus infections have demonstrated the consequences of this type of infections on public health systems worldwide. Detection of OBI is expensive and technically demanding, thus, many patients with OBI might remain undiagnosed. Besides, OBI can be transmitted by blood transfusion and organ transplants and is considered a risk factor for future development of chronic hepatic diseases such as hepatocellular carcinoma and cirrhosis.

In this study, we report a frequency of occult hepatitis B virus infections of 36% in a cohort of Mexican HIV-infected patients highlighting the importance of using sensitive HBV-DNA detection techniques as a prophylactic measure in this type of highly susceptible population.

Several mutations, C19Y, Q129H, E164D, I195M, that were present with a high frequency were also identified and might be associated to the development of OBI. Nevertheless, there are still several questions to solve, and further studies are needed to evaluate the role and function of these mutations in the development of OBI.

HBV

hepatitis B virus