|Home | About | Journals | Submit | Contact Us | Français|
We have analyzed three cases of hepatitis C virus (HCV)-infected recipients who received blood from HCV-infected donors. Two recipients were exposed to two different HCV RNA-positive donors, and one was exposed to a single donor. All parental genomes from the actual infecting units of blood and the recipients were defined, and their presence in the follow-up serum samples was determined using sensitive strain-specific assays. The strain from one of the donors was found to predominate in all recipients' serum samples collected throughout the follow-up period of 10 to 30 months. In two recipients exposed to two infected donors, the strain from the second donor was occasionally found at very low level. However, the original recipients' strains were not detected. Our observations show that HCV-infected individuals can be superinfected with different strains, and this event may lead to eradication or suppression of the original infecting strain. Furthermore, our findings demonstrate that simultaneous exposure to multiple HCV strains may result in concomitant infection by more than one strain, although a single strain could rapidly establish its dominance. The results of the present study suggest the existence of competition among infecting HCV strains which determines the ultimate outcome of multiple HCV exposure.
Hepatitis C virus (HCV) is recognized as an important cause of chronic hepatitis, cirrhosis, and hepatocellular carcinoma (2, 15, 28). The overall prevalence of anti-HCV in the United States is 1.8%, and it is estimated that 2.7 million Americans carry the virus (1).
One of the unresolved questions associated with HCV infection is the sequence of events following superinfection with a new HCV strain. It was found in chimpanzees that persistent infection does not provide protection against subsequent infection with heterologous and even homologous strains (9, 22, 23). Occasionally, the superinfecting strain overtook the original strain (overtake phenomenon) or vice versa; however, the techniques used were not sensitive enough to allow for the detection of strains constituting a small minority of circulating sequences. Furthermore, the baseline HCV replication in chimpanzees is typically very low, and it is unclear whether these findings are relevant for clinical situations.
The evidence that the overtake phenomenon occurs in humans is limited—in a single case report, the superinfecting strain was only transiently detected (14). However, in a recently published study, Eyster et al. (8) reported on a common change of infecting HCV genotypes over time in high-risk patients with hemophilia. Although the genotype change was confirmed by direct sequence analysis of infecting strains in only three cases, these findings suggest that superinfection could lead to subsequent exclusion of the original infecting strain. However, the dynamics of this process are unknown, and it is also unclear whether such an overtake is complete, as the minor strain could replicate at a significantly lower level than the major strain and thus remain undetectable by routinely used techniques.
Whether superinfection could result in subsequent concomitant infection with dual or multiple HCV strains is also unclear. In support of such a possibility, there are a number of studies reporting the relatively common detection of various HCV genotypes in the same individual (21, 25, 26). The parental viral sequences, however, were not available for analysis in these studies. The evidence, therefore, is questionable, particularly in light of the reported unreliability of commonly used genotyping techniques for the detection of mixed infections (10, 31). It is also likely that some of these reports described transient dual infections at the time of superinfection, after which a dominant single infection was established. Importantly, in several large studies, mixed infections with various HCV genotypes were found to be extremely rare or nonexistent (3, 12, 33).
The major obstacle to the study of multiple HCV infections is the lack of clinical material allowing for unambiguous identification of genomes to which the persons were exposed. A unique opportunity to study HCV superinfection in vivo is provided by the Transfusion-Transmitted Viruses Study/National Heart, Lung, and Blood Institute Repository (TTVS/NHLBI). This prospective study of posttransfusion hepatitis was conducted in the years 1974 to 1981, when HCV was much more prevalent in the blood donor population. Here we analyze three patients from the TTVS/NHLBI study who were HCV positive at the time they received a transfusion of blood from an HCV-infected donor(s). These cases are unique, as parental genomes from the actual infecting units of blood and pretransfusion recipients' sera were defined and their presence in the follow-up recipients' sera was determined by sensitive strain-specific assays.
Some clinical, serological, and virological data on three blood recipients and their respective blood donors are presented in Table Table1.1. The female subjects 1 and 2 were enrolled in the TTVS in 1977 and 1978, respectively, when they were hospitalized for genitourinary surgery, while the male subject 3 was enrolled in 1974, when he was hospitalized for a cardiosurgical procedure. Prior to transfusion, all subjects were strongly anti-HCV positive by second-generation enzyme immunoassay (Ortho Laboratories, Raritan, N.J.) and were HCV RNA positive in serum by reverse transcriptase (RT)-PCR. Subjects 1 and 2 received blood from two different HCV RNA-positive donors, while recipient 3 received blood from a single HCV-infected donor. One of the HCV RNA-positive donors implicated in transfusion in subject 2 was anti-HCV negative (Table (Table1).1). The recipients received 1 U of whole blood from each of their respective donors. For recipients 1 and 2, transfusion from both donors were administered within hours on the same day. However, the sequence of transfusion was not recorded in the TTVS database. After transfusion, recipient 1 became symptomatic by day 35 and continued to have alanine aminotransferase (ALT) elevations throughout her observed course (Table (Table2)2) and recipient 2 remained asymptomatic with normal ALT activities, while subject 3 had intermittent elevations of ALT throughout the observation period (not shown).
Donors' serum samples were obtained on the day of donation, while recipients' pretransfusion samples were collected 1 day before and/or on the day of transfusion. For patients 1 and 2, the follow-up samples were collected at 1- to 2-week intervals for the first 10 months, and for patient 1, additional samples were collected every 3 months for the next 20 months. However, for patient 3, only sera collected 32, 102, and 327 days after transfusion were available for virological analysis. All serum samples were kept at −70°C until the present analysis; to avoid repeat freeze-thawing, they had originally been divided into 250-μl aliquots.
Samples from donors and recipients were tested as part of a coded panel that included other positive and negative samples as controls. The code was broken only after all donors' and recipients' samples had their sequence analyses reported.
RNA was extracted from 100 μl of serum by means of a modified guanidinium thiocyanate–phenol-chloroform technique using a commercially available kit (Ultraspec 3; Biotecx Laboratories, Houston, Tex.) and was dissolved in 20 μl of water. Ten microliters of this RNA solution was reverse transcribed with Moloney murine leukemia virus reverse transcriptase and PCR amplified. For the identification of virus strains, we have chosen the NS5b region for the following reasons. (i) It has been reported to be a stable region (30). Similarly, in our previous study, we observed a minimal number of nucleotide substitutions over time (19). Importantly, because of this stability, a minimal number of quasispecies are expected. (ii) The region is sufficiently divergent to allow for a reliable identification of individual virus strains and is the basis for genotype classification (27).
For amplification of the NS5 region, we used the following primers: 5′ GGCGGAATTCCTGGTCATAGCCTCCGTGAA 3′ (nucleotides [nt] 8645 to 8616) and 5′ TGGGGATCCCGTATGATACCCGCTGC/TTTC/TGA 3′ (nt 8245 to 8275) for the first round and primers 5′ CGCTTTCACAGATAACGACTAAGT 3′ (nt 8580 to 8557) and 5′ CTCCACAGTCACTGAGAGCGACAT 3′ (nt 8276 to 8299) for the nested second round. Subsequently, all PCR products were sequenced directly in both directions by the Sanger dideoxy chain termination method with a modified T7 DNA polymerase (Sequenase version 2.0 kit; United States Biochemical).
Extensive measures, described previously (19), were employed to prevent and detect carryover contamination. All RT-PCR runs included positive controls consisting of end point dilutions of synthetic RNA strands and negative controls included uninfected sera. The HCV genotypes were determined by direct sequencing of the NS5 region as described elsewhere (19). Quantification of donor sera for HCV RNA was carried out with the Bayer HCV Quantiplex 2.0 assay.
The E2 region, including the hypervariable region 1, was amplified as described previously (17). HCV quasispecies were compared by the single-strand conformation polymorphism (SSCP) assay as described elsewhere (19), with minor modifications. This assay is highly sensitive, as we were routinely able to detect any minor variant representing ≥3% of the whole population. In brief, PCR products were purified with a DNA binding resin system (Wizard PCR; Promega, Madison, Wis.) and resuspended in 50 μl of water. Next, 2 to 4 μl of the purified product was diluted in 15 μl of low-ionic-strength solution (10% saccharose, 0.5% bromophenol blue, 0.5% xylene cyanol), denatured by heating it at 97°C for 3 min, immediately cooled on ice, and subjected to nondenaturing 8% polyacrylamide gel electrophoresis in 1× Tris-borate-EDTA buffer with 400 V applied for 4 to 5 h at a constant temperature of 25°C. The bands were visualized by silver staining (Silver Stain; Promega). To lower the risk of artifactual polymorphism, each analysis was duplicated in an independent experiment using a new RNA template.
Prior to transfusion (Table (Table11 and and2),2), two different serum samples from recipient 1 were HCV RNA positive, and the infecting strain was determined to be genotype 2b. Both her donors were found to be HCV RNA positive; sequence analysis of the NS5 region revealed that donor A was infected by genotype 2b, whereas in donor B the infecting viral strain was type 1a. The earliest posttransfusion serum sample, at day 10, was HCV RNA positive, but the amplified strain was classified as type 1a and was closely related to that found in donor B (99.5% similarity). All subsequent recipient sera were found to have the same type 1a strain throughout the period recipient 1 was observed. At 913 days after transfusion, the analyzed NS5 region sequence was identical to the sequence amplified from serum on day 10 (Fig. (Fig.1).1).
Recipient 2 was initially infected by a genotype 2a strain, as was one of her donors (Table (Table1);1); the second donor was infected with genotype 1b. The infecting genotype could not be determined in the earliest posttransfusion sample, drawn on day 8; however, all subsequent sera contained a genotype 2a strain whose sequence was virtually identical to that found in donor B but differed by several nucleotide substitutions from the recipient's pretransfusion strain.
Recipient 3 and his only donor were both infected with a genotype 1b strain (Table (Table1);1); all three follow-up samples contained a genotype 1b strain which was 100% identical to that found in the donor but differed by four nucleotide substitutions from the initial recipient strain.
In each of the analyzed cases, the follow-up sera contained HCV strains from one of the donors; however, the pretransfusion recipients' strains and the other donors' strains could have been present below the sensitivity level of direct sequencing (20 to 25%). A commonly used alternative, sequencing of cloned PCR products, has its own pitfalls, as it may be vulnerable to the introduction of artifactual polymorphism (29). Also, to achieve high sensitivity, a large number of clones have to be processed and sequenced, making it laborious and thus impractical.
We decided to use sequence-specific primers that would allow specific amplification of one sequence from the background of other sequences. This strategy takes advantage of the observation that mismatches localized at the 3′ terminus of the primer can dramatically decrease amplification efficiency (16, 20).
The technique employed is shown in detail for patient 1. First, strain-specific primers were designed to match type 2b strains from donor A and the recipient prior to transfusion but to provide a 3′-end mismatch with respect to the 1a strain found in donor B and in the recipient posttransfusion (Fig. (Fig.1,1, primers A1 and A2). Thus, these primers should preferentially amplify both type 2b strains but not the type 1a strain. To provide the template necessary to optimize the reactions, viral sequences were amplified with two rounds of universal primers (see Materials and Methods), and the RT-PCR products were column purified (Wizard PCR). The approximate number of template copies was calculated from optical density readings.
Figure Figure22 illustrates the set-up of PCR for the specific detection of donor A's and the recipient's pretransfusion viral sequences (type 2b). As seen in row A, standard three-step PCR with annealing at 56°C and an MgCl2 concentration of 2.0 mmol allowed for a sensitive and specific detection of the proper template while the negative control (in this instance, RT-PCR products from donor B) was not amplified. Next, to test the sensitivity for the detection of the minor strain from the background of the major strain, PCR products representing the donor A-specific sequence were serially diluted in the RT-PCR product of the donor B strain diluted 1:10 (containing approximately 1010 template copies/μl). As the sensitivity of the assay was now lowered by approximately 1 log unit, it was theoretically capable of detecting the minor sequence in the RT-PCR product when the sequence was present at a concentration of ≥1:10−8 with respect to the major sequence. When applied to the detection of the recipient's pretransfusion strain from the background of donor B's sequence, the sensitivity of the assay was found to be similar.
Once its sensitivity and specificity were determined, the strain-specific assay was applied to recipient follow-up samples using both round I and II RT-PCR products (amplified with universal primers) as templates. Positive reactions were detectable in only one sample collected 56 days after transfusion (Fig. (Fig.3).3). For confirmation, this positive reaction was cloned into the TA cloning vector (Invitrogen), and 10 individual clones were sequenced. Sequence analysis confirmed that the strain-specific PCR detected the type 2b sequence. Cloned sequences were more closely related to that found in donor A than to that found in the recipient prior to transfusion. As seen in Fig. Fig.4,4, all unique substitutions differentiating donor A's strain from the recipient's pretransfusion strain matched the follow-up samples.
To decrease the possibility that these results were influenced by sampling errors, the assays were repeated starting from extracted RNA. Next, because the lack of detection of donor A's strain and the recipient's pretransplant strain in the follow-up samples hypothetically could have been caused by changes in the viral sequence primer annealing regions, analogous analysis was repeated using different sets of primers (both universal and strain specific). Again, the type 2b sequence was found only in the follow-up sample collected 56 days after transfusion. Analysis of serial dilutions of serum and RT-PCR products using primers matching the major (donor B) and minor (donor A) strains revealed that the approximate proportions of the strains when simultaneously detected in the circulation was 1:10−4.
Similar strain-specific assays were developed and employed for the analysis of infecting viral strains in the remaining two cases. In subject 2, the type 1b strain from the second donor (donor A) was repeatedly detectable in a single follow-up sample collected 300 days after transfusion. However, it constituted a small minority of the circulating sequences (≤0.1%). A strain-specific assay for the detection of the recipient 2 pretransfusion strain from the background of the dominant donor strain could not be successfully developed. To partially remedy this shortcoming, RT-PCR amplification products from three different follow-up samples (days 29, 63, and 300) were cloned into the TA cloning vector (Invitrogen), and 20 individual clones from each of the reactions were subsequently sequenced. Sequence analysis confirmed the presence of a type 2a strain which closely matched that found in donor B, while sequences matching the recipient's pretransfusion strain were not present. For subject 3, the pretransfusion recipient strain could not be detected in any of the analyzed follow-up serum samples.
Figure Figure55 shows SSCP analysis of the E2 region in patients 1 and 3 (in patient 2, the E2 region could not be amplified). As can be seen, the evolution of viral quasispecies was clearly different in the two cases. In patient 1, the number of major variants within quasispecies decreased after transfusion compared to the pretransfusion sample, and the quasispecies composition remained unchanged for the observation period. In contrast, in patient 3, the number of major variants increased, and the quasispecies showed constant evolution during the follow up.
In the present study, we have shown that after exposure of HCV-positive recipients to one or two blood donors, the original recipients' strains became undetectable while all donor strains established infection in the recipient. However, one donor strain was clearly predominant, while the other was detected only occasionally. These observations documenting the predominance of a single strain after superinfection are compatible with the results of our recent study on the outcome of liver transplantation in HCV-infected patients who received HCV-infected grafts (19, 32). In that study, analysis of sequential follow-up sera revealed that half of the patients retained their original HCV infecting strain while in the other half the donor strain took over. However, as the study was conducted with severely immunosuppressed liver transplant recipients and the viral load transplanted with the graft was exceedingly large, it is unclear whether this observation would be pertinent to other clinical situations.
The observed predominance of one strain could perhaps be explained from the evolutionary point of view. The major hypothesis of classical population biology, the competitive exclusion principle, states that in the absence of niche differentiation, one competing species will eventually eliminate or exclude the other (11). HCV is a rapidly replicating virus (37), and even small fitness differences, understood as the overall replication and survival ability, could result in overgrowth of one strain by another. The applicability of the competitive exclusion principle to RNA virus populations in cell culture is highly likely, as it has been demonstrated that even virus populations of approximately equal fitness are inherently unstable, since stochastic changes in the balance do eventually occur due to the sudden evolution of higher-fitness variants (5). Presumably, such a displacement of one virus population by another could also occur in infected animals and humans; however, in the case of an infected host, the equilibrium is infinitely more complex, and any shift in the virus population(s) might be mitigated by changes in the adaptive environment.
One theoretical way to prevent competitive exclusion is the development of a parasitic relationship, where one virus population adapts to use the resources of another virus population and in this way compensates for fitness disadvantages; the end result is the development of an equilibrium. Well-known example of such a relationship are “defective interfering particles,” described for many human and animal viruses (13) and recently also for HCV (24). An intriguing possibility is that the minor strain was more adept at infecting certain cells within the liver or even in extrahepatic sites. For example, it has been demonstrated for lymphocytic choriomeningitis virus that strains differing by a single amino acid substitution, when inoculated together into a mouse, are competitively selected either by the liver and spleen or by neurons (6). In light of recent findings (17, 18), the presence of extrahepatic HCV replication is highly likely. Nevertheless, regardless of the mechanisms that allowed the minor strain to survive, their effectiveness was probably limited, as it was detectable only transiently and constituted a small fraction of the circulating virus population.
Whether the immune system has played any role in the selection of infecting strains is unclear. It can be argued that the superinfecting strain is immunologically favored, as it represents, to a certain extent, an “escape mutant.” However, neither in our previous study of liver transplant recipients (19, 32) nor in published chimpanzee studies (9, 22, 23) did the donor strains seem to be privileged in any way, which suggests the likely role of viral factors, such as the replication fitness of individual strains, in establishing dominance in superinfection and coinfection.
Patients 1 and 2 received blood from two infected donors. While both donor strains established infection in the recipient, one was clearly predominant. What determined this strain predominance is unclear, although one possibility is the size of the inoculum. Interestingly, in case 1, the predominant strain was the one with the higher viral load in the donor. However, in both donors in case 2, the levels of viremia were below the sensitivity limit of the quantification assay. Whether the amount of virus introduced during transfusion determines the outcome of superinfection is unclear, as in all three analyzed cases the donor strain overtook the recipient strain. In any case, the overall amount of virus in the recipient, taking into account the amount of virus in body fluids and the liver, must have been much larger than the amount transfused in 1 U of blood.
The E2 region quasispecies analysis provided the opportunity to study the evolution of quasispecies of two different HCV strains in the same host. Interestingly, two different pictures emerged: in patient 1, the number of viral variants within quasispecies was reduced and the quasispecies composition remained largely unchanged during the follow up, while in patient 3 the number of variants increased after the superinfection and quasispecies showed constant evolution thereafter. These observations suggest that the interplay between different HCV strains and hosts may result in different, perhaps unique, quasispecies compositions. This is compatible with the assumption that at any given moment during the natural history of the infection, the quasispecies distribution represents the best-fitting population that has established a status of equilibrium with a particular host (7).
Studies of superinfection and coinfection phenomena have important practical implications, as repeated exposure to HCV is common in high-risk groups, such as drug addicts, but is also likely in many other epidemiological settings. Furthermore, detailed analysis of virological outcome in situations where more than one viral strain is involved, and particularly the evidence for eradication of one strain by another, could have implications for novel treatment options or for the development of a live, attenuated vaccine. The feasibility of exploiting viral interference for antiviral therapy has been demonstrated for influenza A virus infections. An attenuated cold-adapted influenza virus was shown to block the growth of wild-type virus in vitro (34) and, when administered to infected animals or humans, would prevent or reduce the disease symptoms (35, 36).
In summary, we have described three cases of blood transfusion from HCV-positive donors into HCV-positive recipients. In each case, the recipient's pretransfusion strain was apparently displaced by a predominant strain from the transfused blood. In both patients exposed to two HCV RNA-positive donors, the second donor's strain also produced infection. These observations are compatible with the presence of direct competition between infecting strains, which could result in the dominance of a single strain and the competitive exclusion or suppression of other strains.
The formation of the TTVS/NHLBI Repository was supported by contract no. NO1-HB-42972 of the National Institutes of Health. The study was supported by a grant from The Sigismunda Palumbo Foundation.
Sera were provided by James W. Mosley, University of Southern California, Luiz H. Barbosa, and George J. Nemo, Division of Blood Diseases and Resources.