| |
Hepatitis Delta Virus Levels May Predict Disease Severity in HBV
|
| |
| |
"Quantitation of the Level of Hepatitis Delta Virus RNA in Serum, by Real-Time Polymerase Chain Reactionand Its Possible Correlation with the Clinical Stage of Liver Disease"
The Journal of Infectious Diseases April 1 2004;189:1151-1157
Tsuyoshi Yamashiro,1 Kazuyoshi Nagayama,2 Nobuyuki Enomoto,2 Hideki Watanabe,2 Tsuyoshi Miyagi,1 Hiroki Nakasone,1 Hiroshi Sakugawa,1 and Mamoru Watanabe2
1First Department of Internal Medicine, School of Medicine, University of the Ryukyus, Okinawa, and 2Department of Gastroenterology and Hepatology, Tokyo Medical and Dental University, Tokyo, Japan
ABSTRACT
Some hepatitis B virus (HBV) carriers with chronic hepatitis delta virus (HDV) superinfection show progressive chronic hepatitis, whereas others show no apparent signs of liver disease. In the present study, we established a sensitive method for the quantitation of the level of HDV RNA in serum on the basis of real-time reverse-transcription polymerase chain reaction (RT-PCR), to clarify the role that the level of HDV RNA in serum plays in the diverse natural course of clinical manifestation. In 48 subjects who were positive for hepatitis B surface antigen and for antihepatitis delta antibody, the levels of HDV RNA in serum were quantitated by RT-PCR. The levels of HBV DNA in serum were determined by a transcription-mediated amplification assay. The levels of HDV RNA in serum of subjects with chronic hepatitis and of subjects with liver cirrhosis were significantly higher than those in asymptomatic carrier subjects. The levels of HBV DNA in serum did not differ significantly among these 3 groups. In conclusion, HDV RNA quantification by real-time RT-PCR is possibly a useful tool for understanding the pathophysiology of HDV infection.
BACKGROUND
Hepatitis delta virus (HDV) is a hepatotropic, circular, single-stranded RNA viroid that codes hepatitis delta antigen (HDAg) as the sole HDV protein [1]. HDV is defective and is dependent on coinfection with hepatitis B virus (HBV) to provide hepatitis B surface antigen (HBsAg) for virion assembly. Infection with HDV is widespread, and some endemic areas have been reported, such as southern Italy and parts of Africa and South Asia [1]. In adults acutely infected with HBV, coinfection with HDV increases the risk of fulminant hepatitis, but only 2% of these adults become persistently positive for antihepatitis delta antibody (anti-HD), which indicates chronic HDV infection [1]. On the other hand, among chronic HBV carriers with HDV superinfection, 70% become persistently positive for anti-HD. Chronic HBV and HDV infections develop diverse clinical courses, including an asymptomatic phenotype, chronic hepatitis (CH), liver cirrhosis (LC), and hepatocellular carcinoma [1]. However, the mechanism underlying such diversity is not clear.
Previously, in the Miyako Islands (Okinawa, Japan), where HDV infection is endemic [25], we studied subjects positive for both HBsAg and anti-HD. The levels of HBV DNA in serum were generally low in these subjects, regardless of the severity of liver disease, raising the possibility that the level of HDV replication plays a major role in liver damage. Recently, Gudima et al. used the Northern-blot method to quantitate HDV RNA and revealed that the level of HDV RNA correlates with the number of HDV RNA particles [6]. Therefore, the use of a simple method to measure the level of HDV RNA in clinical samples was important in the analysis of the pathophysiology of HDV infection, because many laboratory tests for the detection of HDV RNA are rather complicated [7, 8]. By introducing the use of reverse-transcription polymerase chain reaction (RT-PCR) [9] for the measurement of the level of HDV RNA in serum, we showed that half of the patients in the Miyako Islands who were positive for anti-HD were negative for HDV [2]; most were clinically asymptomatic carrier (ASC) patients, which suggests that a low level of HDV replication is associated with an indolent clinical course, despite the presence of chronic HBV infection.
On the basis of these observations, we established in the present study a sensitive and simple quantification system for HDV RNA, using real-time RT-PCR. Subsequently, we applied the method to a cross-sectional study to clarify the relationship between the clinical presentation of chronic hepatitis delta and the level of HDV RNA in serum.
Subjects.
A total of 48 subjects were enrolled in the study: 16 men and 32 women who were 31-83 years old. All of them lived in the Miyako Islands (Okinawa, Japan), where both HBV and HDV infection are known to be endemic [4]. All subjects were positive for both anti-HD and HBsAg in serum. Of these 48 subjects, 14 were ASC subjects who showed consistently normal levels of alanine aminotransferase (ALT) at least bimonthly for >2 years; 20 had CH that was diagnosed on the basis of sustained abnormal levels of ALT; and 14 had LC, which was diagnosed on the basis of standard abdominal-ultrasound findings consistent with LC, including liver-surface irregularities, hypertrophy of the caudate lobe, and splenomegaly [10, 11]. The abdominal-ultrasound findings for all ASC subjects were normal, without any evidence of liver disease. All subjects were negative for hepatitis B e antigen (HBeAg). To control for other factors that contribute to an elevated level of ALT, the following subjects were excluded from the present study: those who were positive for either antibody to hepatitis C virus or antinuclear antibodies; those whose ultrasound examinations revealed fatty liver; and those with a history of either excessive alcohol intake or exposure to hepatotoxic drugs. The clinical backgrounds of the subjects are shown in table 1. Subjects with LC had significantly low platelet counts, compared with ASC subjects and subjects with CH, which indicated the presence of hypersplenism resulting from the cirrhotic change in the liver. The results of the zinc-turbidity tests were significantly higher in subjects with CH and in subjects with LC than in ASC subjects, which suggested hypergammaglobulinemia caused by chronic inflammation. The level of HBV DNA in the serum was quantitated by use of a commercial kit (DNA Probe Chugai-HBV; Chugai Diagnostics) and a transcription-mediated amplification assay [12] with a detection range of 0.5 X 101-0.5 X 106 kilocopies/mL (kC/mL).
DISCUSSION
In the present study, we used real-time PCR to develop a new, sensitive method for the quantitation of the level of HDV RNA in serum and determined these levels in 48 subjects positive for anti-HD who were living in the Miyako Islands (Okinawa, Japan) [25]; subsequently, we assessed the relationship between these levels and the extent of liver damage. The levels of HDV RNA in serum of subjects with either CH or LC were significantly higher than those in ASC subjects, but the levels of HBV DNA in serum did not differ among the 3 groups. In addition, in subjects with LC, levels of ALT in serum weakly correlated with levels of HDV RNA in serum. These findings suggest that the level of circulating HDV RNA in serum is possibly related to the clinical manifestation of chronic HDV infection and that the relationship is stronger than that for the level of HBV DNA in serum.
We used real-time RT-PCR to develop a sensitive quantitation method for the detection of HDV RNA in serum at levels as low as a single copy. In the past, semiquantitative methods (dot-blot hybridization [8] and semiquantitative RT-PCR [9]) have been available mainly for the measurement of the level of HDV RNA. The real-time RT-PCR system enables a more sensitive measurement of PCR products and has been successfully adopted for the quantitation, with high sensitivity and accuracy, of hepatitis C virus and HBV [15]. In addition, to avoid false-negative results, it is especially important to make PCR primers for highly conserved sequence regions, because RNA viruses, including HDV, are prone to mutations that can cause primer mismatch [1]. Before the present study, we determined, in our subjects, the complete sequence of the HDV genome and found a highly conserved sequence within the HDAg-coding region. In fact, the primers used in the present study enabled us to evaluate the level of HDV RNA in the serum of most subjects with chronic liver diseases, particularly those diseases caused by HDV genotype II.
Using real-time RT-PCR, we showed that liver damage is associated with the level of HDV RNA. We found that the levels of HDV RNA in the serum of subjects with CH and of subjects with LC were significantly higher than those in the serum of ASC subjects, thereby extending our previous finding that, for RT-PCR, the rate of detection of HDV RNA in ASC subjects positive for anti-HD is low, compared with that in subjects with either CH or LC [2]. The levels of HDV RNA in serum were below the threshold of detection (<1 kC/mL) in 9 of the 14 ASC subjects and also seemed to be low (<102 kC/mL) in the remaining 5 subjects. In contrast, 90% (31/34) of subjects with chronic liver disease consisting of either CH or LC had levels of HDV RNA in serum that were as high as 106 kC/mL. Thus, active HDV replication was associated with liver damage in subjects positive for anti-HD. Because, in 30% of acute HDV superinfection cases HDV is cleared and anti-HD alone persists [1], the subjects in this study who were negative for HDV RNA could be examples of such cases. However, our observation that 75% (9/12) of cases negative for HDV RNA, irrespective of HBV DNA levels, were ASC subjects indicates that the presence of active HDV replication is associated with liver disease. Furthermore, a previous study, which analyzed the relationship between intrahepatic HDAg and hepatitis, also demonstrated a positive correlation [16], and Kuo et al. [17] reported that HDAg is essential for HDV-genome replication. These findings are concordant with our own observations.
In contrast to the levels of HDV RNA in serum, the levels of HBV DNA in serum were not dispersed and were at relatively low levels in ASC subjects and in subjects with either CH or LC. In HBsAg carriers, superinfection with HDV frequently induces severe or progressive liver disease [2, 18, 19]. It has been unclear whether HBV or HDV is primarily responsible for liver damage in patients with chronic HDV. Some studies have suggested that liver-disease activity is related to HDV [2022], whereas others have indicated that it is related to HBV [2325]. In our previous study, we did not find in subjects positive for anti-HD any correlation between the level of HBV DNA (assayed by use of real-time RT-PCR) and the level of ALT; this suggests that HBV plays a minor role in the liver damage seen in patients positive for anti-HD [5]. Although the present study suggests that the level of circulating HDV RNA plays a major role in the pathogenesis of liver injury, further studies of large cohorts over long periods of time will be needed.
The pathogenesis of HDV hepatitis is not well understood. A direct cytopathic effect of the small form of the delta antigen, in the absence of HDV-genome replication, has been reported [26], and the mortality of HDV-infected mice is consistent with virus-mediated cell death [27]. On the other hand, as with the other hepatitis viruses, peak HDV replication precedes the peak histopathologic changesand then markedly diminishes after the immune response is initiated during peak pathology; this adds credence to the notion that humoral or cellular immune mechanisms are involved in the pathogenesis of liver damage and in HDV replication. The reason why the levels of HDV RNA and ALT correlate only in patients with LC is unclear. Investigating one possibility, we assumed that immune response against HDV in subjects with CH is different than that in subjects with LC. In some subjects with CH, active HDV replication did not elicit rigorous immune response in the presence of relatively low levels of ALT. On the other hand, in subjects with LC, active HDV replication closely correlated with high levels of ALT, presumably because of the rigorous immune response, and, consequently, resulted in severe liver damage. Therefore, subjects with rigorous immune response to HDV replication subsequently progressed to LC and thereby demonstrated a correlation between the level of ALT and the level of HDV RNA. In any case, further investigation of the pathogenesis of HDV hepatitis is needed.
The severity of liver disease is also influenced by HBV genotype and HDV genotype. In the Miyako Islands, HBV genotype B and HDV genotype II are prevalent [28, 29]. These genotypes are reported to be linked to liver disease that is milder than that associated with other genotypes [30]. Although we did not determine the genotypes of the subjects studied, the observed diversities in the level of HDV RNA in liver disease do not appear to be due to genotype differences alone, because HBV genotype and HDV genotype in most patients in the Miyako Islands are reported to be homogeneous [29].
In the present study, the levels of HDV RNA in serum were distributed in a relatively wide range (1106 kC/mL), as was HBV. It is known that host-derived double-strandedRNA adenosine deaminase modifies HDV's mRNA [31] and changes the balance of large HDAg and small HDAg, causing a change in the level of HDV RNA. However, in the present study, liver tissue was not accessible, and so we could not address this point. Thus, the mechanism that determines the level of circulating HDV RNA needs further study.
In conclusion, we have established a sensitive method for quantitation of the level of HDV RNA in serum and have determined these levels in subjects who were positive for anti-HD. As a result, we could successfully determine the level of HDV RNA in serum. In our cross-sectional study, the level of HDV RNA in serum seemed to correlate with the severity of liver disease, whereas the level of HBV DNA did not. This new strategy for quantification of HDV RNA should be useful in clinical tests and/or in further basic research on the biology of HDV.
RESULTS
The sensitivity and linearity of real-time RT-PCR when the Light Cycler System was employed were examined by use of synthetic HDV RNA (figure 1A). After RT-PCR was completed, logarithmic values of fluorescence (y-axis) for each dilution were plotted against cycle numbers (x-axis). A baseline was set just above the fluorescence background, and a Ct was determined on the basis of amplification curves obtained during the initial exponential phase of amplification. There was a direct relationship between the cycle number corresponding to the Ct and the log concentration of RNA molecules initially present in the RT-PCR reaction, and this linearity was conserved in serial dilutions of synthetic HDV RNA used as a standard (ranging from 1-106 copies of RNA/uL [figure 1B], corresponding to 1-106 kC/mL); the coefficient of correlation was 0.99. The final PCR products were resolved by gel electrophoresis, and the specific amplification of HDV cDNA of expected size (134 bp) was confirmed, even for a single molecule of HDV RNA standard (figure 1C). Melting curves were recorded by plotting fluorescence signal intensity against temperature (figure 1D).
As determined by the real-time RT-PCR method, the levels of HDV RNA in the serum of the 48 subjects ranged from below the threshold of detection (<1 kC/mL) to 9.6 X 105 kC/mL (median, 1.8 X 102 kC/mL) (figure 2A). (In the determination of median RNA values, levels of HDV RNA in serum that were below the threshold of detection were assumed to be 100 copies/mL; because we used the Mann-Whitney U test, the results of our statistical analyses were unaffected by this assumption). In 12 subjects, the levels of HDV RNA in serum were below the threshold of detection; among them, 9 were ASC subjects, 3 had CH, and 0 had LC. On the other hand, with the exception of 1 ASC patient, the subjects who had levels of HDV RNA that were >102 kC/mL also had either CH or LC. The levels of HDV RNA in serum of subjects with CH were <1-3.4 X 102 kC/mL (median, 1.1 X 104 kC/mL); those in subjects with LC were 6-9.6 X 105 kC/mL (median, 4.6 X 102 kC/mL); and those in ASC subjects were <1-3.8 X 103 kC/mL (median, <1 kC/mL). Therefore, the levels of HDV RNA in serum of subjects with CH and of subjects with LC were significantly higher than those of ASC subjects (P = .0012 and P = .0003, respectively; Mann-Whitney U test).
Among the 48 subjects, 47 had levels of HBV DNA in serum that were <1 X 103 kC/mL, and, in 33 of these 47, they were below the threshold of detection (5 kC/mL), as determined by a transcription-mediated amplification assay (figure 2B). The levels of HBV DNA in serum were not significantly different among the 3 groups (ASC subjects, <5-1.6 X 103 kC/mL; subjects with CH, <5-1.3 X 102 kC/mL; subjects with LC, <5-3.2 X 102 kC/mL [P = .5; Kruskal-Wallis test]). Thus, in the subjects enrolled in the present study, HBV replication was suppressed at low levels, and the HBV-replication levels were unrelated to the severity of liver disease. These findings indicate that the pathogenesis of liver damage in these subjects was unrelated to the level of circulating HBV.
Figure 2C shows the relationship between the level of HDV RNA and the level of HBV DNA. Among 10 subjects whose HDV level and HBV level were both below the threshold of detection, 7 were ASC subjects, and none showed signs of LC; in contrast, among 23 subjects with detectable levels of HDV and no detectable HBV, 19 had either CH or LC. Therefore, in most of the ASC subjects, HBV and HDV were not detectable (P = .003; Fisher's exact test), whereas most of the subjects with active liver disease (i.e., those with either CH or LC) showed active HDV replication (P = .0001; Fisher's exact test). Furthermore, among 13 subjects with both detectable HBV replication and detectable HDV replication, approximately half (7/13) had LC, which suggests progressive liver damage that is due to the active replication of both viruses (P = .03, Fisher's exact test).
Last, we analyzed the relationship between HDV RNA levels and the clinical parameters shown in table 1. The levels of HDV RNA in serum were found to significantly correlate only with the levels of ALT in serum of subjects with LC (figure 3). For 1 patient with LC, when we quantitated serial levels of HDV RNA in serum, we found that the changes in these levels correlated with those in the changes in the levels of ALT in serum; the ALT level increased from 34 to 234 U/mL over 6 years, and the HDV RNA level increased from 5.8 to 39.3 kC/mL during the same interval.
RNA extraction.
Total RNA was extracted by use of the acid guanidium phenol-chloroform method. In brief, 150 uL of serum was mixed with 700 uL of ISOGEN (Wako Pure Chemical Industries) and, in the aqueous phase, was extracted once, with 140 uL of chloroform. RNA was precipitated with isopropanol, with 20 mg of glycogen (Boehringer Manheim) used as a carrier, and then was washed once with ethanol, dissolved in 10 uL of water, and stored at -80 C until used.
Synthesis of cDNA for HDV RNA.
A 10-uL portion of the RT mixture was adjusted to contain 1 uL of RNA solution (which corresponded to 15 uL of serum); 50 U of Moloney murine leukemia virus reverse transcriptase (Gibco BRL) buffer, prepared in accordance with the manufacturer's instructions; 10 U of an RNase inhibitor (Promega); and 50 pg of random hexamers (Takara Bio). The mixture was incubated at 37 C for 40 min. To inactivate the reverse transcriptase, the mixture was treated with 2.5 uL of 1.0 N sodium hydroxide at 37 C for 30 min, followed by neutralization by the addition of 2.5 uL of 1.0 N hydrochloric acid. Thus, 15 L of final cDNA solution was obtained from 15 uL of serum.
Plotting of a standard curve by use of artificially synthesized cDNA. To synthesize standard RNA, the partial HDV genome coding HDAg was amplified by nested PCR with primer 888 [13] (5-GATGCCCAGGTCGGACCGCG[G/A]GGAG-3; nt 888-912), primer 914 (5-GGAGATGCCATGCCGACCCGAAGAG-3; nt 914-938), primer 1391 (5-GGCGAAGAGGCCCCGGACGGATCAG-3; nt 1391-1367), and primer 1421 (5-AAAAGGGAAGGACGGGGAGGG[G/A]GCT-3; nt 1421-1397); the nucleotides were numbered according to the HDV sequence reported by Wang et al. [14]. These primers were designed on the basis of HDV sequences in GenBank and newly isolated HDV sequences from patients in the Miyako Islands. The PCR product was inserted into the pGEM-T vector (Promega) by use of TA cloning. Synthetic RNA was transcribed in vitro by use of T7 RNA Polymerase and Ribomax In Vitro Transcription kits (Promega). For removal of vector DNA and enzymes, an RNeasy Kit (Qiagen) was used. The purified RNA was quantitated by measurement of optical density at a wavelength of 260 nm.
Quantitation of HDV RNA by use of real-time RT-PCR. Quantitative PCR was performed in 10 uL of Light Cycler DNA Master SYBR Green I mix (Boehringer Mannheim) containing 3.5 mM of MgCl2 and 1 uL of cDNA solution, with primer 1164 (5-CCGGCTACTCTTCTTTCCCTTCTCTCGTC-3; nt 1164-1192) and primer 1297 (5-CACCGAAGAAGGAAGGCCCTGGAGAACAA-3; nt 1297-1268), amplifying 134 bp of HDV cDNA. PCR was performed in 50 cycles, each comprising denaturation at 95 C for 0 s, annealing at 68 C for 5 s, and extension at 72 C for 20 s; after each cycle, fluorescence was detected at 72 C, by use of the Light Cycler System (Boehringer Mannheim). The primers were designed on the basis of regions highly conserved among different HDV genotypes, resulting in nucleotide homologies of 98% for genotype I, 98% for genotype IIa, and 100% for genotype IIb. After each cycle of amplification, real-time data acquisition was performed. After the final cycle, melting-point analysis of all the samples and controls was performed at 57 C-95 C. Once the threshold cycle was chosen, the time at which the amplification plot crossed the threshold was defined as the crossing time (Ct). The calculated Ct value was predictive of the quantity of target RNA copies present in the sample. The standard curve for this assay was calculated by use of a series of 10-fold dilutions of (1-106 copies) previously titrated synthetic HDV RNA.
|
|
| |
| |
|
|
|
