Mem Inst Oswaldo Cruz, Rio de Janeiro, 113(1) January 2018
Detection of occult hepatitis B in serum and oral fluid samples
1Fundação Oswaldo Cruz-Fiocruz, Instituto Oswaldo Cruz, Laboratório de Hepatites Virais, Rio de Janeiro, RJ, Brasil
2Universidade Federal do Rio de Janeiro, Hospital Universitário Clementino Fraga Filho, Rio de Janeiro, RJ, Brasil
3Universidade Federal do Estado do Rio de Janeiro, Hospital Universitário Gaffrée e Guinle, Rio de Janeiro, RJ, Brasil
4Hospital Geral de Nova Iguaçu, Nova Iguaçu, RJ, Brasil
5Fundação Oswaldo Cruz-Fiocruz, Instituto Oswaldo Cruz, Laboratório de AIDS e Imunologia Molecular, Rio de Janeiro, RJ, Brasil
In occult hepatitis B infection (OBI), hepatitis B virus DNA (HBV DNA) can be detected in serum samples; however, oral fluid collection for detection of HBV DNA has not yet been explored, despite the availability of collection devices. Serum and oral fluid samples from 45 hepatitis B core antibody (anti-HBc)-positive patients were collected for the amplification of the HBV polymerase gene. HBV DNA was detected in five serum and four oral fluid samples (the detection limit for oral fluid was 1.656 log IU/mL in paired serum). In conclusion, simple methodologies of sample collection and in-house polymerase chain reaction (PCR) allowed detection of HBV DNA, and these could be used to improve the diagnosis of OBI, especially in locations with limited resources.
Occult hepatitisB (OBI) is identified by the presence of hepatitis B virus DNA (HBV DNA) inthe liver (with detectable or undetectable HBV DNA in the serum) of individualsdeemed HBV surface antigen (HBsAg)-negative, using currently available assays.When detectable, the amount of HBV DNA in the serum is usually very low (<200 IU/mL). Patients can be divided into two groups: seropositive-OBI [hepatitisB core antibody (anti-HBc)-and/or hepatitis B surface antibody (anti-HBs)-positive]and seronegative-OBI (anti-HBc-and anti-HBs-negative) (Raimondo et al. 2008 13 ).
The persistenceof HBV DNA in HBsAg-negative individuals could be related to the poor laboratorydetection of HBsAg owing to the low levels of HBs antigenemia or HBsAg mutants,underlying coinfections, or other host-related factors (Bréchot et al.2001). It has been demonstrated that patients who are intravenous drug usersor have other diseases that are also transmitted parenterally, such as hepatitisC and HIV, are at a greater risk of developing occult hepatitis B (Morsica etal. 2009), and are more likely to develop complications such as cirrhosis (Squadritoet al. 2013).
Although HBV DNAis always present in liver tissue, its clinical assessment is difficult (Larrubia2011); therefore, the use of sensitive methods for HBV DNA detection in serumsamples is paramount. In contrast, oral fluid samples have been used in molecularstudies to detect hepatitis B (Van der Ejik et al. 2005 16 , Kidd-Ljunggren et al.2006, Heiberg et al. 2010 4 ) owing to the ease by which these samples can be obtainedin subjects with difficult venous access, more particularly children, the elderly,drug addicts, and patients on haemodialysis. Furthermore, oral fluid samplecollection is cheaper, less invasive, and painless compared to blood collection,and can easily be performed in remote laboratory environments. Therefore, weinvestigated HBV DNA detection in oral fluid samples from individuals with occulthepatitis B.
This study aimedto detect OBI in a cohort of anti-HBc- and/or anti-HBs-positive patients usingserum and oral fluid samples.
A total of 45 individualswere included in this study; however, all of them did not present HBsAg in serum.These individuals were recruited from the following ambulatories in the regionof Rio de Janeiro in the period from October 2010 to July 2014: Viral HepatitisAmbulatory (IOC/Fiocruz), Clementino Fraga Filho Hospital (Federal Universityof Rio de Janeiro/UFRJ), Gaffree and Guinle University Hospital (Federal Universityof Rio de Janeiro State), and Nova Iguaçu General Hospital. Individualsprovided informed consents prior to blood and oral fluid sample collection,and the study was approved by the Fiocruz ethics committee (CAAE 18281313.4.0000.5248).Individuals could be of any gender, race, or ethnicity, and had to be over 18years of age.
Blood samples werecollected by venipuncture, and the obtained serum was stored at -20ºC untilfurther use. Oral fluid samples were obtained using a Salivette collector (Sarstedt,Nümbrecht, Germany). This collection device consists of a polypropylenetube that contains an absorbent pad made from cotton. The swab was placed insideeach volunteer's mouth for 2 min to absorb oral fluid. Salivettes were checkedvisually for blood contamination and excluded if such event had occurred. Aftercollection, 1 mL of phosphate buffer saline (PBS, pH 7.2) was added to minimisethe effects of oral fluid degradation and facilitate pipetting. The vials werecentrifuged at 2000 x g for 10 min and stored at -20ºC until use.
The serum sampleswere subjected to commercial enzyme immunoassays (ELISA) for the detection ofHBsAg, anti-HBc, anti-HBc IgM, anti-HBs (Bioelisa anti-HBs, Biokit, Barcelona,Catalonia, Spain), hepatitis B e antigen (HBeAg), and anti-HBe (e411 Cobas,Roche Diagnostics, Manheim, Germany) according to the instructions of each manufacturer.Serum samples were also analysed to determine anti-HCV (Murex anti-HCV 4.0,DiaSorin, Kyalami, Republic of South Africa) and anti-HIV (DS-EIA-HIV-AGAB-SCREEN,RPC, Diagnostic System, Nizhny Novgorod, Russia) levels, as well as the biochemicaldosages of liver enzymes, such as aspartate aminotransferase (AST), alanineaminotransferase (ALT), alkaline phosphatase, total, direct, and indirect bilirubin,and gamma-glutamyl transferase (GGT), using a commercial kit (LabMax 560, LabTest,Lagoa Santa, Brazil).
HBV DNA was extractedfrom serum samples using a commercial kit (High Pure Viral Nucleic Acid Kit,Roche Diagnostics, Mannhein, Germany) by following the manufacturer's instructions.To extract HBV DNA from oral fluid samples, a RTP® DNA/RNA Virus Mini Kit(Stratec Biomedical Ag, Berlin, Germany) was used. The oral fluid volume wasincreased two-fold (400 µL), as determined previously (Portilho et al.2012), and the manufacturer's recommendations were followed.
Oligonucleotideswere used to amplify the polymerase gene region of HBV (Mallory et al. 2011 8 )via a single round of amplification, generating a product containingapproximately 940 base pairs. For serum samples, PCR analysis was performedusing a tube filled with 25 μLof reaction buffer containing the following components: 0.5 µM of eacholigonucleotide, 0.2 mM of a mixture of four deoxynucleotides, 10× PCRbuffer and 1.5 mM MgCl2, Platinum Taq polymerase (Invitrogen, SanDiego, CA, United States) (5U) at 1.5 U, and target DNA (5 µL). A target-freecontrol reaction tube contained 25 µL of reaction mixture only. Negativeand Positive HBV controls were included for each target tested. The thermocycler(T3 Thermocycler, Biometra, Göttingen, Germany) program incubated the samplesfor 3 min at 95ºC, followed by 45 cycles consisting of 95ºC for 10s, 58ºC for 20 s, and 72ºC for 40 s, followed by an additional extensionstep at 72ºC for 5 min. To increase the PCR sensitivity in oral fluid samples,the protocol was modified as follows: 0.5 µL (2.5 U) of 5 U/µL PlatinumTaq DNA polymerase, (Invitrogen) and 10 µL DNA.
Serum samples inwhich HBV DNA was detected using in-house PCR were also submitted for the quantificationof HBV DNA via Abbott Real Time HBV (Abbott Laboratories), and for viralsequencing employing the same oligonucleotides as the ones used for PCR amplification(Mallory et al. 2011 8 ), to determine HBV genotypes. Sequences were analysed usingthe Mega v6.0 software (Tamura et al. 2013 15 ), and HBV genotypes were identifiedusing blast alignment.
All individualscompleted a questionnaire, and a descriptive statistical analysis was performed,with the means, medians, and standard deviations being calculated. Statisticalanalysis was performed using the GraphPad InStat software (La Jolla, CA, UnitedStates).
Most of the patientswere men (29/45), and the mean age was 36.36 ± 20.74 years. All individualswere HBsAg-negative, anti-HBc-positive, and anti-HBs-negative. We could notaccess HBeAg or anti-HBe results for all patients. Among them, 30 were HBeAg-negative,and 12 out of 33 samples were anti-HBe-positive. Among the 45 individuals, 16had detectable levels of anti-HCV, and 24 were anti-HIV-positive. Eleven patientswere receiving treatment for hepatitis C and/or HIV infections during the study.Regarding the biochemical tests, the mean ALT value was 14.21 ± 13.18U/L, and the mean AST value was 21.33 ± 26.49 U/L. The mean total bilirubinvalue was 0.15 ± 0.15 U/L, the mean alkaline phosphatase value was 65.17± 37.91 U/L, and the mean GGT value was 94.27 ± 120.95 U/L.
All serum sampleswere subjected to the in-house PCR protocol for HBV polymerase gene determination.Among them, five (11.11%) showed the presence of HBV DNA, displaying a meanviral load of 2.246 ± 0.635 log IU/mL. Genotypes were determined viasequence analysis in three of the five samples, in which two were classifiedas genotype F and one as genotype A. In two samples, the quality of the sequencedata was very poor, which prevented their correct classification into genotypes.Among the patients with HBV DNA detected through the qualitative method in bothtypes of sample, one was anti-HCV positive, and three were anti-HIV positive(Table).
Oral fluid samples,for which HBV DNA had been detected in paired serum samples, were also subjectedto in-house PCR, and four of these contained HBV DNA. The oral fluid samplethat was HBV-DNA-negative presented a viral load of 1.623 log IU/mL in its pairedserum sample. The agarose gel electrophoresis results obtained for oral fluidsamples are represented in Figure.
The presence ofOBI was evaluated via detection of HBV DNA in serum and oral fluid samplesin a cohort of patients with no HBsAg but showing anti-HBc and/or anti-HBs positiveresults. In the studied population (n = 45), five cases of OBI were confirmedwhere patients with no HBsAg presented HBV DNA in serum.
We found that oneOBI patient was anti-HCV positive, and three were anti-HIV positive. Some studiesindicate that OBI infection is more common in patients who are coinfected withhepatitis C or HIV, varying from 1-62% in HIV patients (Piroth et al. 2008 11 ),and occurring in approximately one-third of subjects from the MediterraneanBasin and in more than 50% of East Asian populations (Coppola et al. 2015 3 ).The presence of occult HBV in coinfected HCV patients may indicate more severeliver damage, cirrhosis, and increased rates of hepatocellular carcinoma (Chenet al. 2016). In HIV patients, the identification of OBI cases may be due toHBV immune-escape, which reduces the humoral immune response and anti-HBs titres,recurrence of HBV replication, recovery of immune responses after HIV treatment,or the development of resistance to lamivudine therapy (Maldonado-Rodríguez et al. 2015 7 ).
One of the difficultiesin identifying OBI is the low level of HBV DNA in serum samples. Therefore,it is extremely important to use a sensitive PCR protocol. We evaluated theapplicability of an in-house PCR method for amplification of the polymerasegene of HBV, which was able to detect a positive result in a sample containing1.623 log IU/mL HBV DNA following the commercial method, even allowing the detectionof different HBV genotypes, and presenting 100% of concordance between them.
To the best ofour knowledge, oral fluid samples have not yet been evaluated for use in occulthepatitis B diagnosis. It is well-known that the viral load in oral fluid isproportional to that in serum (Zhang et al. 2008 17 ), and is higher in HBsAg- andHBeAg-positive individuals (Noppornpanth et al. 2000 10 ). Therefore, it is consideredextremely difficult to detect HBV DNA in oral fluid samples from OBI patients.In this study, it was possible to detect HBV DNA in oral fluid of four of thefive positive serum samples (80%).
HBV DNA could notbe detected in only one oral fluid sample. This sample presented a low viralload (1.623 log IU/mL) in the paired serum sample. Non-detection may have beenthe result of viral load fluctuation, the DNA extraction method, or the collectiondevice used. The extraction method (RTP® DNA/RNA Virus Mini kit, Stratec)and Salivette collector were used in a previous study in our laboratory (Portilhoet al. 2012) to detect serial dilutions of HBV DNA in artificially infectedsamples. In this case, it was possible to detect 2 x 101 copies ofHBV DNA/mL. However, another study showed that collectors based on the principleof mechanical friction may be more efficient (Heiberg et al. 2010 4 ) for HBV DNAdetection.
However, this studypresents some limitations. First, we did not include OBI seronegative patientsbecause of the large number of individuals already included in our laboratorycohort. Moreover, we did not have sufficient volumes of oral fluid samples toperform HBV DNA sequencing, and compare the detected genotypes using this typeof sample. We intend to collect samples from patients with the serological profileof OBI to identify more cases. Furthermore, we could not conduct any additionalsample collections from these individuals during the course of the infectionbecause of logistical problems, and therefore, we could not determine the potentialoccurrence of viral load fluctuation in the oral fluid samples showing negativeresults.
In conclusion,it was possible to detect HBV DNA in serum and oral fluid samples from individualswith occult hepatitis B infection, suggesting the usefulness of this methodand its ability to improve the identification of these cases in patients withdifficult venous access.
To Vanessa AlvesMarques, Juliana Custódio Miguel and Ana Carolina da Fonseca Mendonçafor their technical assistance.
LMV conceived thestudy; LMV, MMP and EL designed the study protocol; LCN, CAVN, CEBM, JHP andLLLX carried out the clinical assessment, subject selection and recruitment;GLF performed the immunoassays, analysis and interpretation of these data; MMPperformed all molecular assays, analysis and interpretation of these data. MMPand LMV drafted the manuscript; LCN, CAVN, CEBM, JHP, LLLX and EL criticallyrevised the manuscript for intellectual content. All authors read and approvedthe final manuscript. The authors disclose no actual or potential conflictsof interest, including any financial, personal, or other relationships withpeople or organisations within two years of the beginning of this study thatcould inappropriately influence the study.