点击显示 收起
1Viral Epidemiology Branch, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Bethesda, and 2Viral Epidemiology Laboratory, SAIC Frederick, National Cancer Institute, Frederick, Maryland; 3Departments of Microbiology and Internal Medicine, University of Virginia Health Systems, Charlottesville
Received 2 May 2002; revised 6 August 2002; electronically published 13 December 2002.
Tests for human herpesvirus8 (HHV-8) often disagree on which individuals are infected. To evaluate HHV-8 antibody changes over many years, 245 homosexual men enrolled in a prospective study between 1982 and 1999 were studied. Samples collected annually were evaluated by immunofluorescence to HHV-8 latent nuclear antigen and by an ELISA for antibody against the HHV-8 lytic antigen K8.1. When positive, samples from most subjects were persistently reactive in both assays. Titers to both lytic and latent antibodies increased for many years, usually in parallel, and were higher in human immunodeficiency virusinfected men. The increasing titers and expanding epitope recognition suggest ongoing, low-grade replication of HHV-8. Use of multiple assays is important in determining prevalence in asymptomatic populations. Because persons with more advanced infections have higher antibody titers, they are easier to detect, which may skew our understanding of the epidemiology of HHV-8.
Informed consent was obtained from participants, and guidelines for sample use from the US Department of Health and Human Services and National Cancer Institute (NCI) were followed.
Financial support: NCI, National Institutes of Health (contract N01-CO-12400).
Reprints or correspondence: Dr. Robert J. Biggar, National Cancer Institute, 6120 Executive Blvd., Rm. 8014, Bethesda, MD 20852 .
Since the discovery of human herpesvirus8 (HHV-8, also called the Kaposi sarcomaassociated herpesvirus) in Kaposi sarcoma tissue [1], this virus has been the subject of intense investigation. Despite remarkable advances in molecular studies, much remains unclear about its epidemiology. Soon after discovery, it was found that almost all patients with Kaposi sarcoma have HHV-8 antibodies [2 5], usually with relatively high titers [6 8]. Detection of asymptomatic HHV-8 infection has been much more problematic. In interlaboratory comparisons, using standard panels of samples from general populations in America and Europe [7, 9], all laboratories generally detected antibodies in the patients with Kaposi sarcoma, but there was much less agreement about results in asymptomatic persons. Similarly, serial samples from some studies [10 12], but not all [13, 14], have yielded inconsistent serologic results in individuals thought to be positive.
In the absence of a reference standard, establishing the true infection status in asymptomatic persons is problematic. Culturing HHV-8 has proven to be difficult. By use of polymerase chain reaction (PCR), most investigators find <10% of seropositive, asymptomatic Europeans and Americans to be viremic [15 17], and, when detected, the virus is usually present in minute amounts. Furthermore, in serially followed subjects, PCR-detected viremia may be found only intermittently [18]. Even among persons with Kaposi sarcoma who have PCR-positive tumors, virus levels in blood and saliva are modest, and only one-half or fewer are found to be viremic [19 21].
We examined consistency within and between latent and lytic HHV-8 antibody assays and how antibody levels changed over time. To this end, we studied samples collected annually for almost 2 decades from a cohort of homosexual men, a population in which the prevalence of HHV-8 antibodies is high [2 5, 8 11].
METHODS
In 1982, we began a longitudinal study of homosexual men in New York and Washington, DC, in response to the emerging AIDS epidemic. Details of enrollment, follow-up, and human immunodeficiency virus (HIV) testing have been presented elsewhere [22, 23]. Blood samples were sought at annual visits, except that Washington participants were not seen in 1983. Surviving HIV-infected subjects continued in follow-up study throughout the 1990s. Follow-up was discontinued on most HIV-negative subjects in the early 1990s, although some were retained as study control subjects. We have previously reported a study of this cohort in which we described HHV-8 prevalence and incidence, using an immunoflourescence antibody (IFA) assay for latent antigens [11]. In that study, we noted a peak in incidence in the early 1980s.
The main objective of the current study was to examine the longitudinal dynamics of the antibody response. Serial samples were screened for antibody to lytic and latent HHV-8 antigens. The methods used in our laboratories have been described in detail elsewhere [4, 11, 17]. An IFA assay was done using HHV-8 infected BCBL-1 [11] or BCP-1 [4, 17] cells with samples diluted 1 : 100. This assay detected reactivity to a latency-associated nuclear antigen encoded by orf73. The K8.1 assay was an ELISA using K8.1, a lytic phase surface glycoprotein, done on serum samples initially diluted 1 : 20. Because of sample depletion or equivocal reactions, results for only IFA or K8.1 were available for a few samples. The term "persistent" is used to describe sequential observations that were all or almost all of the same type, allowing for an occasional divergent result considered to be a sample or laboratory error.
Subsets of antibody-positive subjects were then titered. To avoid bias, we titered all HIV-negative subjects for whom we had at least 6 HHV-8positive samples and a similar number of HIV-infected HHV-8positive subjects. Because the objective of the study was to determine changes in asymptomatic men, none of those titered was reported to have developed Kaposi sarcoma. Titers were done at 2-fold dilutions in the IFA assay with BCP-1 cells and 4-fold dilutions in the K8.1 assay. Slopes were determined from log10 transformed titers on positive samples, starting from the first sample positive by that assay; the significance of the titer change in individuals was calculated by linear regression (SAS software version 8.2; SAS Institute).
RESULTS
Tests were done on 1529 samples from 245 subjects. The mean time span was 8.1 years, with a mean of 6.9 samples per person. Of these subjects, 134 (55%) were prevalently HIV-infected or became HIV-infected during the study. Thirty-one (23%) of the HIV-infected men were reported to have developed Kaposi sarcoma during the study, but none of 111 HIV-negative men developed the disease.
We classified 130 men (53%) as being persistently HHV-8 seronegative. This group included 2 men who were never seropositive while participating but who were reported to have developed Kaposi sarcoma after they discontinued active participation. Among the 128 other seronegative subjects, there were 8 positive IFA results (1.0%) among 789 tests and 19 positive K8.1 results (2.3%) among 821 tests. These positive results were both preceded and followed by a series of negative results, and only 1 sample was positive by both tests. Presumably, they represent either false-positive reactions or mislabeled samples.
Sixty-seven (27%) men were prevalently HHV-8 seropositive, defined as positive by either assay on their first tested sample . Sixteen (24%) were later reported to have Kaposi sarcoma. The first sample was both IFA- and K8.1-positive for 45 (67%) subjects, only K8.1-positive for 15 (22%) subjects, and only IFA-positive for 7 (10%) subjects, with the last group including 2 subjects not tested by K8.1 tests because no samples remained. Fifty-one (76%) of these men were persistently seropositive. In this expected-positive group, 9 samples had false-negative results by IFA and by K8.1 tests (2.4% each; generally not the same samples). Of 18 men with initially discordant results and multiple samples, 6 became positive by both tests in the second sample, 2 remained persistently positive only by K8.1 tests but did develop IFA reactivity 5 and 7 years later, 9 were persistently K8.1 positive but IFA negative for at least several years, and 1 was persistently IFA positive but remained K8.1 negative for at least 7 years.
fig.ommitted
Figure 1. Comparison of qualitative assessment by immunoflourescence antibody (IFA) and K8.1 tests for antibodies to human herpesvirus8 among persons considered to be prevalently positive on the basis of reactivity by either test in the earliest sample tested. Subjects (SUBJ) are arranged by duration of follow-up. Qualitatively, patterns appeared to be consistent by assay, but some subjects recognized only 1 antigen for many years. HIV, human immunodeficiency virus; KS, Kaposi sarcoma; +, positive; -, negative.
Forty-eight (20%) subjects were classified as HHV-8 seroconverters, defined as being antibody-negative in 1982 but developing persistent antibodies in 1 or both assays thereafter . Thirty-three (73%) subjects seroconverted by 1984. Thirteen seroconverters later developed Kaposi sarcoma. Of 41 subjects tested by both tests, 23 (56%) seroconverted to both assays at the same time, 11 (27%) seroconverted in the K8.1 assay first, and 7 (17%) seroconverted in the IFA assay first. Nineteen (40%) seroconverters had discordant antibody patterns in which one assay became persistently positive only years after the other, if ever, or else reverted to being persistently negative in 1 assay. Eleven of 19 were predominately K8.1 positive, and 8 subjects were predominately IFA reactive. At some point, reactivity by both assays usually occurred, but 5 subjects were never IFA positive while in the study.
fig.ommitted
Figure 2. Comparison of qualitative assessment by immunoflourescence antibody (IFA) and K8.1 tests for an antibody to human herpesvirus 8 (HHV-8) among persons considered to have seroconverted to HHV-8 reactivity while under evaluation. Data are arranged by year of first positive sample and then duration of follow-up. Qualitatively, patterns showed initial reactivity to both antigens was common but some subjects reacted to one antigen years before the second antigen, and a few, all human immunodeficiency virus (HIV) positive, lost reactivity to 1 antigen. KS, Kaposi sarcoma; SUBJ, subject; +, positive; -, negative.
Impact of HIV infection. Discordant serologic test results were more common in HHV-8 seroconverters than those who were prevalent positives, with reactivity being somewhat more common for the K8.1 test than the IFA assay. Both HIV-infected and -uninfected men showed patterns of delayed recognition to one or the other HHV-8 antigen. However, except for instances when only one sample reverted, the patterns of loss of epitope recognition were seen only in HIV-infected men. The consistency of most individual profiles in and indicated that the changes were not laboratory errors.
HHV-8 seroconversion was more common in subjects who were already HIV infected at study onset or who became HIV infected. Eighty-four men were already HIV positive when first HHV-8 tested. Of these men, 39 (46%) were also HHV-8 seroprevalent, and 18 (21%) later seroconverted to HHV-8. Among 50 subjects who became HIV infected while in the study, 11 (22%) were HHV-8 infected before HIV seroconversion, and 21 (42%) seroconverted to HHV-8 after HIV infection. Another 111 subjects never became HIV infected, of whom 17 (15%) were HHV-8 seroprevalent and 9 (8%) seroconverted to HHV-8.
Level of seroreactivity was assessed by examining K8.1 optical density for all subjects and by K8.1 or IFA titers for the subset with titer data. To avoid bias from using multiple samples from the same subject, we examined reactivity in the first and last positive sample in each subject. HIV status was assigned by the reactivity at the time of sampling. HIV-infected men had higher reactivity in the last positive sample, both in optical density values and, among those titered, K8.1 and IFA antibodies than HIV uninfected men. In both HIV-infected and -uninfected men, reactivity increased between the first and last sample, whether measured by optical density or titer (data not shown). Profiles of HHV-8 antibody titers varied in both 16 HIV-negative and 13 HIV-positive subjects (representative examples in ), but, in general, titers increased steadily over many years. shows the slopes of the titers over time for 27 men who had both IFA and K8.1 reactivity. Most (63%) had rising antibody titers in both IFA and K8.1 assays, and the increase was usually in parallel.
fig.ommitted
Figure 3. Titer changes in representative subjects over time. Identifiers link data to qualitative data presented in (PP; prevalent positive) or (SC; seroconverter). When titer trend was P < .15, the value is provided. Panels show in human immunodeficiency virus (HIV)uninfected subjects: A, Increasing immunoflourescence antibody (IFA) and K8.1 titers; B, delayed K8.1 response; C, delayed IFA response; and in HIV-infected subjects: D, increasing IFA and K8.1 titers; E, delayed K8.1 response; and F, loss of IFA reactivity.
fig.ommitted
Figure 4. Slope values for all individuals studied by immunoflourescence antibody (IFA) or K8.1 titer (log10 transformed), in which values in the upper right quadrant are increasing for both antibodies. Some values were off scale. Identifiers link data to qualitative data presented in (PP; prevalent positive) or (SC; seroconverter). Three subjects were not presented, 2 because the IFA was never positive and 1 because the K8.1 was never positive. Values generally fell along the diagonal, suggesting that antibody titers increased in parallel. HIV, human immunodeficiency virus.
Subjects diagnosed with Kaposi sarcoma. The relationship between HHV-8 and Kaposi sarcoma was not the primary purpose of the present analysis. In brief, all subjects who developed Kaposi sarcoma were HIV-infected, and, of 134 HIV-infected men, 31 (23%) reported with Kaposi sarcoma. Accepting a positive result in either test as evidence of HHV-8 infection, Kaposi sarcoma was reported in 2 (2%) of 130 persistent negative subjects, 13 (27%) of 48 seroconverters, and 16 (24%) of 67 prevalent positive subjects. The 2 men who were negative for HHV-8 were both tested before developing Kaposi sarcoma but not after. Presumably, both seroconverted. Compared with asymptomatic subjects, men who developed Kaposi sarcoma appeared to have generally similar profiles of qualitative results, with some being non- or late-reactors (n = 7) or seroreverters (n = 2), by either the IFA or K8.1 assay. Among 111 subjects who were never HIV-positive, none developed Kaposi sarcoma, although 17 were prevalently HHV-8 positive at study onset and 9 seroconverted to HHV-8. The risk of Kaposi sarcoma did not appear to be related to the relative timing of infection with HIV and HHV-8. Kaposi sarcoma was diagnosed in 6 (40%) of 15 subjects who were HHV-8 infected before HIV infection, 2 (15%) of 13 who seroconverted to both viruses at the same time, and 11 (50%) of 22 who seroconverted to HHV-8 after HIV infection (including 2 presumed to have become HHV-8 infected).
DISCUSSION
The use of serial samples over nearly 2 decades revealed the changing profiles of antibody reactivity in asymptomatic persons. Other studies have presented data on serial assay reactivity and performed antibody titers on selected cases in the years immediately preceding Kaposi sarcoma [12, 13, 24]. Although those investigators showed that titers increased immediately preceding or at the onset of Kaposi sarcoma, our study showed that antibody titers increased over time, even in asymptomatic subjects. In many studies, antibodies were also inconsistently detected within individuals followed over time [1012], with the profiles of reactivity typically being more erratic than the profiles we observed. Transient changes in serostatus and titer could be due to label or test error, but they could also be biological. In our data, we saw examples of apparent errors, but the consistency of the patterns argues strongly that long term patterns are due to the nature of the immune response to this virus. Two findings help to explain some of the difficulty with HHV-8 serologic analysis.
First, antibody titers were low, often near the threshold of detection, when first determined to be positive. However, both IFA and K8.1 antibody titers increased steadily in most subjects. The increases of both latent and lytic antibodies continued for nearly 2 decades of observation, demonstrating that the changes were not related to recent seroconversions. We propose that the increasing antibody titer resulted from active viral replication, probably at a low level, despite the presence of antibodies.
Second, there was a gradual broadening of the epitopes recognized. Although the majority of subjects were prevalently positive or seroconverted simultaneously in both assays, numerous subjects, both HIV-infected and -uninfected, were reactive to one antigen for years before becoming reactive to the other antigen. These findings may be explained by a gradual expansion of epitopes to which antibodies are made and are thus consistent with the increasing antibody titers. In addition, a few subjects lost reactivity to one antigen. Notably, these subjects were all HIV infected, suggesting that the loss might be related to an immune perturbation. Changes in CD4 count did not appear to be associated with titer changes (data not presented). Some subjects did not recognize all epitopes. Hence, disagreement as to which asymptomatic subjects are reactive is not surprising. These observations indicated the importance of using assays that detect more than one antigen.
The biology of immune responses could distort our current understanding of the epidemiology of HHV-8. We have shown that titers increase and epitope recognition broadens over time from infection. Thus, it is easier to identify persons with longer durations of infection than those with new infections, and the higher prevalences at older ages could be related to easier recognition of infection with advancing age. In addition, our subjects were predominantly white men. Infected females have higher antibody titers to another human gamma herpesvirus, Epstein-Barr virus, than males [25], and there could be differences by race as well. In the current study, we examined samples from homosexual men, but this group may have higher antibody titers than would be found in the general population. Unlike some studies [13, 14], we did not find that the relative timing of HIV versus HHV-8 infection affected the risk of Kaposi sarcoma. HIV-infected persons were more likely to be detected as HHV-8 infected, but they have higher antibody titers. Gambus et al. [8] reported similar observations.
In healthy subjects, low-antibody titers and limited epitope recognition may underlie the low level of agreement noted in interlaboratory assay comparisons [7, 9]. They also raise concern about our understanding of the epidemiology of HHV-8 infection. Tests that are marginally sensitive to the low-antibody levels seen with HHV-8 are most problematic in low prevalence populations, such as asymptomatic American and European populations. In such populations, the frequency of false positivity from nonspecific reactions may dominate true positive reactions. Furthermore, even with very specific tests, the epidemiologic profiles may be affected by factors that increase reactivity, such as the duration of infection and HIV status, rather than true infection.
Acknowledgments
We are grateful to Christine Gamache and Rachel Bagni (SAIC Frederick) for excellent technical assistance.
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Boshoff C, Weiss RA. Kaposi's sarcomaassociated herpesvirus. Adv Cancer Res 1998; 75:5786.
21.
22.
23.
24.
25.