ABSTRACT
The search for hepatitis C virus (HCV) in body fluids other than blood is important when assessing possible nonparenteral routes of viral transmission. However, the role of oral fluids in HCV transmission remains controversial. Here we quantitatively determined HCV RNA in saliva and gingival crevicular fluid (GCF) of anti-HCV-positive patients. Most patients (14 of 18; 78%) whose saliva specimens were negative had HCV RNA in their GCF. Most patients (20 of 26; 77%) had higher HCV RNA levels in their GCF than in their saliva. Although there was not a statistically significant correlation between the serum viral load and HCV level in saliva or GCF, patients with low serum HCV loads were less likely to have detectable HCV in their saliva. These findings have important implications for medical personnel and suggest that epidemiological studies designed to understand the significance of the oral route of transmission of HCV are warranted.
Hepatitis C virus (HCV) infection represents a major public health problem in the world today. The infection primarily causes liver disease; however, HCV infection has also been associated with extrahepatic abnormalities, including mixed cryoglobulinemia, malignant lymphoma, Sjögren's syndrome, and oral lichen planus (2, 12, 18, 19, 34, 39). Lymphotropism of HCV has been observed, and several laboratories have detected the virus in blood mononuclear cells (BMC) (16, 22, 26, 28, 35, 38). Common risk factors for HCV infection include blood transfusion from unscreened donors as well as injection drug use. Although sexual and vertical transmissions have also been reported, there remain a large number of HCV carriers in whom no route of infection has been identified.
Epidemiological surveys demonstrate that body fluids other than blood, including saliva, might be potential sources of HCV infection. Experimental inoculation of saliva obtained from chronic HCV carrier chimpanzees has been reported to transmit hepatitis to recipient animals (1). Several studies have demonstrated HCV RNA in the saliva of hepatitis C patients by reverse transcription (RT)-nested PCR. However, the detection rates of viral RNA within saliva have varied widely, and some groups have failed to demonstrate HCV RNA within saliva (6-11, 14, 17, 23, 25, 27, 29-33, 36-38). A potential source of HCV RNA within saliva includes gingival crevicular fluid (GCF), which might contain HCV-infected BMC in the setting of periodontal inflammation. To our knowledge, only one study has qualitatively identified HCV in GCF; HCV RNA was detected in 59% of GCF specimens from hepatitis C patients in the study (20). Since the efficiency of HCV transmission is likely related to its viral load, it is important to quantitate viral RNA levels within body fluids in order to properly evaluate possible nonparenteral routes of HCV infection.
Thus, we examined the presence of HCV RNA in the saliva and GCF of anti-HCV antibody-positive patients using real-time quantitative RT-PCR.
MATERIALS AND METHODS
Sample collection.Twenty-six dental patients attending the hospital of Nippon Dental University at Tokyo were studied. All of the patients were anti-HCV antibody seropositive on the basis of screening using a second-generation enzyme immunoassay (Abbott HCV PHA, Abbott Diagnostics, Abbott Park, IL). This study protocol was approved by the Ethics Committee of the hospital and was conducted according to Ethic Guideline for the Studies on Human Genome and Gene Analysis. Written informed consent was obtained from each patient participating in the study.
Blood samples were collected and centrifuged for 20 min at 5,000 rpm to separate the serum. Patients spit into a cup to obtain saliva samples. Whole saliva samples (approximately 2 ml) were then transferred into sterile containers. None of the samples were macroscopically observed to contain blood. GCF specimens were collected by first drying the gingival surface with sterile cotton, after which the area was isolated in order to prevent contamination with saliva. A paper strip (2 by 5 mm) was then subgingivally inserted for 30 s to collect specimens (approximately 50 μl). If there was visible contamination of the sample with blood, another sample without macroscopic blood contamination was taken from another site. The depth at gingival crevices was then measured by a periodontal probe, and the presence of bleeding on probing was examined. Serum, saliva, and GCF samples were collected simultaneously and were stored at −80°C before use.
RNA extraction.Total RNA was extracted from 100 μl of serum or saliva specimens and from paper strips with collected GCF using a QIAamp viral RNA kit (QIAGEN, Valencia, CA). In preliminary experiments using various amounts of serum, saliva, and GCF samples in the presence or absence of paper strips, we confirmed that (i) sample volumes of >40 μl yielded the same efficiencies of RNA extraction from each specimen and (ii) inclusion of a paper strip described above in the lysis buffer did not influence the efficiency of RNA extraction.
Quantitation of HCV RNA.To determine the quantity of HCV RNA, real-time RT-PCR involving single-tube reactions was performed using TaqMan EZ RT-PCR Core reagents (PE Applied Biosystems, Foster City, CA), as previously described (3). Briefly, the reaction mixture contained 1× TaqMan EZ buffer, 500 nM concentrations of each primer from the HCV 5′ noncoding region (5′-GAG TGT CGT GCA GCC TCC A-3′ and 5′-CAC TCG CAA GCA CCC TAT CA-3′), a 200 nM concentration of fluorogenic probe [5′-(6-carboxyfluorescein) CCC GCA AGA CTG CTA GCC GAG TAG TGT TGG (6-carboxytetramethylrhodamine)-3′], 200 μM concentrations of each deoxynucleoside triphosphate, 3 mM Mn(OAc)2, 5 U of Thermus thermophilus DNA polymerase, 0.5 U of AmpErase uracil N-glycosylase, and template RNA. The primers and probe were designed on the basis of the conserved sequences among HCV genotypes. The RT step was started with a 1-min incubation at 50°C, followed by 50 min at 65°C. Thermal cycling conditions were as follows: a precycling period of 5 min at 95°Cfollowed by 50 cycles of denaturation at 94°C for 15 s and annealing at 55°C for 10 s and extension at 69°C for 1 min. All reactions and analyses of the amplification plots were performed on an Applied Biosystems PRISM 7700 sequence detector (PE Applied Biosystems). Standard curves of the assays were obtained by plotting 10-fold serial dilutions of known concentrations of a synthetic HCV genotype 1b transcript. HCV RNA copy numbers of the synthetic transcript were calculated from the quantity and its molecular weight. Using a standard curve, the Sequence Detector software calculated automatically the concentration of RNA copies in the experimental samples. We found that results obtained from our in-house real-time RT-PCR method were well correlated with those from the COBAS AMPLICOR HCV MONITOR Test, version 2.0 (Roche Diagnostics, Tokyo, Japan) (15), and that 1 HCV RNA copy/ml in our method corresponded to approximately 1 international unit/ml by the above-mentioned commercial assay (data not shown).
HCV genotyping.HCV genotype was determined by RT-PCR of the core region sequence with genotype-specific primers for determination of HCV genotypes 1a, 1b, 2a, 2b, 3a, 3b, 4, 5a, and 6a, as described previously (24).
PCR amplification of β-globin DNA.Total DNA was extracted from saliva samples using a QIAamp DNA Mini kit (QIAGEN) according to the manufacturer's instructions. To characterize the degree of cell contamination in saliva, isolated DNA was subsequently used as a template to amplify the human β-globin gene fragment of 268 bp with the following primers: 5′-GAA GAG CCA AGG ACA GGT AC-3′ and 5′-CAA CTT CAT CCA CGT TCA CC-3′ (21).
Statistical analysis.The Spearman rank test was used for evaluating the correlation between variables: anti-HCV antibody levels and viral loads in serum, saliva, and GCF.
RESULTS
The clinical and virological characteristics of 26 patients are presented in Table 1. The study group consisted of 10 males (38%) and 16 females (62%) with a mean age of 69 years (range, 56 to 79 years). Their mean liver enzyme values were as follows: 30 IU/liter for alanine aminotransferase (ALT) and 33 IU/liter for aspartic aminotransferase (AST). HCV RNA levels in the serum of 20 patients (77%) were determined by real-time RT-PCR assay, which showed a detection limit of 102 copies/ml and a linear range over 5 logs. Four of six serum samples whose HCV RNA levels were below the detection limit in this measurement were found to have detectable HCV RNA by the qualitative nested RT-PCR (4). We found no difference in efficiency and specificity of HCV cDNA amplification among genotypes 1b, 2a, and 2b in the real-time RT-PCR assay (data not shown).
Clinical and virological characteristics of 26 patients examined in this studya
Figure 1 summarizes viral loads in the serum, saliva, and GCF specimens of the patients. A mean serum HCV RNA level of 5.1 × 105 copies/ml was observed among samples with viral loads greater than 102 copies/ml. As expected, serum viral RNA levels were significantly correlated with anti-HCV antibody levels (r = 0.80, P < 0.0001) (Fig. 2A). In a number of cases (20 of 26; 77%), the viral load of the GCF was greater than that of the saliva. HCV RNA was detected in 31% of the saliva samples and 85% of the GCF specimens using real-time RT-PCR. Mean viral RNA levels were 1.9 × 104 (saliva) and 3.1 × 104 (GCF) copies/ml in these samples. It should be noted that most (seven out of eight) of the saliva samples contained 1.4 × 102 to 8.2 × 103 copies/ml of HCV RNA, with a mean value of 2.0 × 103 copies/ml among these seven samples (Fig. 1).
HCV viral load in the serum, saliva, and GCF of anti-HCV-positive patients. Numbers of patients within each range of the viral load are indicated.
(A) Correlation between anti-HCV antibody levels and HCV RNA levels in serum. The Spearman rank test was used for testing the correlation between variables. There is a significant positive correlation (r = 0.80, P < 0.0001) between the serum levels of HCV antibody detected by the passive hemagglutination assay and those of HCV RNA determined by real-time RT-PCR. (B) Correlation between viral loads in the serum and those in saliva specimens. Results for patients whose HCV RNA levels in saliva were ≥102 copies/ml are plotted. No significant correlation was observed. (C) Correlation between viral loads in serum and those in GCF specimens. Results for patients whose HCV RNA levels in the GCF were ≥102 copies/ml are plotted. No significant correlation was observed.
Among the 18 patients with HCV RNA-negative saliva, 102 to 103 copies/ml of viral RNA were detected in the GCF of 3 patients, 103 to 104 copies/ml of viral RNA were detected in the GCF of 2 patients, and >104 copies/ml were detected in the GCF of 9 patients. No significant association was observed between viral RNA levels in the serum and viral RNA levels in the saliva (Fig. 2B) or GCF (Fig. 2C). However, relatively high serum viral loads (>105 copies/ml) were observed in five out of eight patients with HCV RNA-positive saliva, while serum viral loads were 1.5 × 103 copies/ml or less in most of the patients whose saliva specimens were negative (13 out of 18). Four patients with HCV RNA-positive saliva and/or GCF had no detectable serum HCV RNA by real-time RT-PCR (Fig. 2B and C); however, viral RNA was detectable in their sera by qualitative nested RT-PCR. Although no visible contamination of the saliva and GCF with blood was observed, there may be a small amount of cells or lysed cells in the fluids. To determine the degree of cell content in samples, total DNA was extracted from three saliva specimens, which contained >103 copies/ml of HCV RNA (Fig. 2B), and tested for the presence of cellular DNA by amplifying a human β-globin gene. A certain amount of cellular DNA was detectable in the saliva specimens (data not shown), suggesting some salivary HCV RNA may be derived from HCV-infected cells, such as BMC and mucosal epithelial cells, as discussed below. Various amounts of HCV-infected cells in the saliva and GCF may, in part, account for differences in the viral loads.
HCV RNA was detectable in most GCF and/or saliva specimens obtained from patients with clinical evidence of oral diseases: HCV RNA was detected in all 14 (100%) patients with periodontitis, 6 of 7 (85%) patients with squamous cell carcinoma, and 3 of 4 (75%) patients with lichen planus. Three out of four patients with HCV RNA-negative GCF, however, also had some oral epithelial lesions. On the other hand, among seven patients without oral diseases, HCV RNA was detected in the GCF and saliva of six and three patients, respectively. There was a trend toward increased viral loads in the oral fluids, especially GCF, among patients with bleeding on probing compared to those without the bleeding. The viral RNA levels in the GCF and saliva had no correlation with age, gender, or serum levels of ALT or AST. It also seems that their viral RNA levels were not correlated with HCV genotype, although the viral genotypes in 12 of 26 patients were not determined.
DISCUSSION
Identification of HCV in body fluids other than blood is important in order to evaluate possible nonparenteral routes of transmission. The role of oral fluids in HCV transmission remains controversial. Although the presence of HCV RNA in saliva has been reported by several research groups (6-11, 14, 17, 23, 25, 27, 29-33, 36-38), only one study has attempted to quantify HCV RNA in saliva, in which patients coinfected with HCV and human immunodeficiency virus were examined using a branched DNA assay (27). Moreover, limited information exists regarding the prevalence of HCV in the GCF of patients with hepatitis C, apart from one study in which a qualitative RT-PCR method was used to detect HCV in 59% of GCF and 35% of saliva specimens from patients with HCV viremia (20).
To the best of our knowledge, this study is the first to quantitate HCV loads within the saliva and GCF of anti-HCV antibody-positive patients using real-time RT-PCR. To search for a possible oral route of HCV transmission, whole saliva and GCF containing cell fractions were used to determine the viral loads in this study. Although any saliva and GCF samples tested were not macroscopically observed to contain blood, we cannot rule out the possible effect of a small amount of bleeding as a source of HCV RNA. Here we observed HCV more commonly in the GCF than the saliva of HCV-seropositive patients. We further found viral loads of 102 to 104 copies/ml and 103 to 105 copies/ml in saliva and GCF, respectively. This result may be partially due to the presence of PCR inhibitors in saliva. An internal control to measure the possible effect of PCR inhibitors was not included in our real-time RT-PCR. Although the mean viral load within the GCF was approximately 10-fold lower than that in the serum, GCF samples from 12 of 26 patients (46%) had viral titers similar to or greater than those observed in the sera. No significant correlation was observed between the serum viremia levels and viral levels in the saliva or GCF. However, there was a trend that patients with HCV RNA-positive saliva showed higher viral loads in sera than patients with HCV RNA-negative saliva. These findings suggest that GCF might be one of the sources of HCV RNA within the saliva.
Although HCV is a hepatotropic virus, convincing evidence of HCV lymphotropism has been demonstrated in tissue culture (13). HCV has been widely detected in BMC in patients with chronic HCV infection, and differences in quasispecies identification within serum and BMC suggest that viral replication occurs within BMC (16, 22, 26, 28, 35, 38). HCV-infected BMC might allow HCV to infiltrate the GCF and saliva, since BMC migrate from dentogingival vessels into gingival crevices. There also might be transudation of HCV-containing serum into the mouth. Generally, periodontal inflammation increases the excretion of BMC-rich GCF. There is also a possibility that HCV exists within mucosal epithelial cells. HCV has been identified in the mucosal tissue, as well as salivary glands, of anti-HCV-positive patients with oral lichen planus using various techniques, including in situ hybridization, strand-specific RT-PCR, and immunohistochemistry (5, 32). Thus, it is likely that several possible sources discussed above are involved in HCV penetration into the saliva and GCF. Whatever the sources or mechanisms are, the findings obtained provide important implications for medical personnel regarding HCV transmission in health care settings as well as for HCV epidemiology, as the origin of the viral infection remains unclear in up to 40% of cases.
In this study, although the numbers of specimens were limited, we quantitatively determined HCV RNA in oral fluids from dental patients, including some patients with oral diseases, and demonstrated frequent detection of HCV in the saliva and GCF. Further large-scale epidemiological studies employing real-time RT-PCR assays are required to clarify the clinical significance of HCV in the saliva and GCF, including the potential for viral transmission through exposure to these fluids.
ACKNOWLEDGMENTS
We thank Yasushi Inoue and Ryosuke Suzuki for technical advice and helpful discussion on data analysis. We also thank Makiko Yahata for technical assistance and Tomoko Mizoguchi for manuscript preparation.
This work was partly supported by grants-in-aid from the Ministry of Health, Labor, and Welfare of Japan to T.S. and T.S.
FOOTNOTES
- Received 8 February 2005.
- Returned for modification 28 May 2005.
- Accepted 2 June 2005.
- Copyright © 2005 American Society for Microbiology
