Frequent and Asymptomatic Oropharyngeal Shedding of Human Herpesvirus 8 among Immunocompetent Men

Abstract

Background.

Little is known about the clinical and virologic manifestations of human herpesvirus (HHV)–8 infection in immunocompetent persons in the absence of malignancy.

Methods.

A total of 46 human immunodeficiency virus–negative, HHV-8–seropositive men collected saliva daily, and 25 recorded 15 common symptoms daily (gastrointestinal, constitutional, and oropharyngeal) and absences from work or school. Quantitative polymerase chain reaction measured HHV-8 DNA in saliva.

Results.

Some 44 (96%) of 46 men reported having sex with men (MSM). Of the 44 MSM, 27 (61%) had HHV-8 detected in saliva on ≥1 day; heterosexual men also shed HHV-8. In analyses restricted to MSM, HHV-8 DNA was detected on 636 (22%) of 2897 days. Among MSM with HHV-8 detected in saliva, the median rate was 20% (range, 1%–100%), with 30% shedding on >50% of days, and the median quantity was 4.5 log₁₀ copies/mL (range, 2.0−7.3 log₁₀ copies/mL). The quantity of HHV-8 shed was lower in nonwhites (P < .001) and younger participants (P = .03). The frequency of HHV-8 detection and quantity were correlated (r = 0.62; p < .001). Symptoms were reported on 10 (9%) of 114 days when HHV-8 was present, compared with 78 (9%) of 830 days without (odds ratio, 0.93 [95% confidence interval, 0.30−2.88]; P = .9).

Conclusions.

HHV-8 is detected frequently and intermittently in the saliva of chronically infected immunocompetent MSM, but this infection is asymptomatic.

Human herpesvirus (HHV)–8 (also known as Kaposi sarcoma [KS]–associated herpesvirus) was first identified in a KS lesion from a patient with AIDS in 1994 [1]. HHV-8 causes all forms of KS—including epidemic (HIV-associated), iatrogenic (transplant-associated), classic (Mediterranean or Middle Eastern), and endemic (African)–and primary effusion lymphoma and multicentric Castleman disease [2]. A decade of careful basic science, epidemiologic, and clinical research has begun to elucidate the natural history of infection with HHV-8, but many questions remain outstanding. We sought to characterize the clinical and virologic manifestations of chronic HHV-8 infection in persons from areas where KS is not endemic.

Characterizing HHV replication patterns at mucosal sites has led to important observations about the transmission and clinical manifestations of infection with herpes simplex virus (HSV) I and II [3–8], Epstein-Barr virus (EBV) [9, 10], cytomegalovirus (CMV) [11, 12], and HHV-6 [13]. In prior work, we demonstrated that the oropharynx is the most common mucosal site for the replication and transmission of HHV-8. HHV-8 is shed from the oropharynx of chronically infected men who have sex with men (MSM) in the United States [14] and in African adults [15]. Replication of HHV-8 occurs in salivary epithelial cells [14], virus isolated from saliva is infectious to cells in vitro [14, 16], and behaviors associated with exposure to saliva have been correlated with the acquisition of HHV-8 [14, 17]. Comprehensive characterization of HHV-8 replication in the oropharynx may improve the understanding of both the epidemiologic and the clinical course of HHV-8 infection. We therefore initiated a prospective study of men in Seattle, Washington, with documented HHV-8 infection, to characterize the virologic and clinical natural history of chronic HHV-8 infection in immunocompetent persons.

SUBJECTS AND METHODS

Study participants.

Men in Seattle, Washington, were recruited from outpatient clinics and advertisements in the community for participation in studies of the epidemiologic features of HHVs. Written, informed consent was obtained in accordance with a protocol approved by the University of Washington Human Subjects Division. Inclusion in the study was determined solely by the presence of antibodies to HHV-8 and the absence of antibodies to HIV. We identified 46 men with serologic evidence of HHV-8 infection and invited them to collect saliva daily at home, with some also completing a daily diary recording constitutional and oropharyngeal symptoms, medications ingested, visits to health care providers, and days of work missed.

Specimen collection and laboratory testing.

Whole-virus EIA or immunofluorescence assay was used as described elsewhere [18] for the detection of serum antibodies to HHV-8. Oropharyngeal secretions were collected by swabbing the buccal, lingual, palatine, and tonsillar mucosa with a Dacron swab, which was then placed in 1 mL of buffer [19]. Swabs were collected for periods of 25−135 consecutive days (“sessions”), with some men collecting for >1 session. HHV-8 DNA was measured quantitatively using a fluorescent-probe–based polymerase chain reaction (PCR) assay (TaqMan assay; Applied Biosystems), as described elsewhere [14, 19–22]. DNA was extracted from 200 μL of sample using the QIAmap96 DNA blood kit (Qiagen) and was eluted with 100 μL of AE buffer (10 mmol/L [pH 8.0]; Qiagen). Ten microliters of DNA was then used for each PCR using primers to the orf73 gene (ORF73 forward, 5^′-CCA GGA AGT CCC ACA GTG TTC-3^′; ORF73 reverse, 5^′-GCC ACC GGT AAA GTA GGA CTA GAC-3^′) [19, 23]. To quantify HHV-8 DNA in each sample, a standard curve was constructed with 10⁵, 10⁴, 10³, 10², and 10 copies of cloned plasmid DNA included in each run, with each sample run in duplicate. The quantity of HHV-8 DNA in each sample was calculated by comparing the threshold cycle with the standard curve. The run was accepted only if the 10-copy sample of plasmid DNA was detected. Multiple negative controls and 1 positive control were built into each PCR run. An internal control was spiked into the PCR master mix, to monitor for false-negative results. A negative result was accepted only if the internal control was detected. Limiting dilution assays revealed that as few as 100 copies of HHV-8 DNA/mL could be detected in 50 of 50 replicate runs. Therefore, any oropharyngeal sample with ≥100 copies of HHV-8 DNA/mL was considered to be positive.

Statistical analysis.

Participant characteristics, HHV-8 oropharyngeal detection patterns, and clinical symptoms were reviewed using graphical summaries and descriptive statistics. The distribution of copies of HHV-8 DNA was highly skewed and was thus log-transformed before analyses. Correlates of HHV-8 shedding frequency were examined using zero-inflated binomial models with person-level random effects to account for overdispersion [24]. Generalized estimating equations compared the mean quantity of HHV-8 in the oropharynx among selected characteristics such as age or ethnicity and examined the correlation between the detection of HHV-8 and the reporting of clinical symptoms.

RESULTS

Study participants.

Of the 46 men enrolled in the study, 44 reported being MSM. Analysis was restricted to the 44 MSM, and the 2 heterosexual men are described separately. The median age of the participants was 45 years (range, 26−72 years); 38 (86%) reported their race as “white,” 3 (7%) “black,” and 3 (7%) “other.” Thirty-four of the participants (77%) noted >100 lifetime sex partners. The 44 participants collected a total of 2897 days of oral swabs, for a median of 32 days per person (range, 25−248 days). A total of 32 (73%) men collected swabs during only 1 session, whereas the others collected during an additional 1−3 sessions separated by a median of 293 days between sessions (range, 108−1371 days).

Oropharyngeal shedding patterns.

Of the 44 MSM, 27 (61%) had HHV-8 detected in the saliva on ≥1 day. Overall, HHV-8 was detected on 636 (22%) of 2897 days sampled. Among men with HHV-8 detected in the oropharynx, the median proportion of days on which HHV-8 was detected was 20%, with a large range (1%–100%). Both heterosexual participants shed HHV-8 from the oropharynx. The first heterosexual participant shed HHV-8 DNA on 1 (1%) of 97 days sampled, and the second shed HHV-8 on 3 (8%) of 40 days sampled.

HHV-8 shedding often appeared to be sporadic. We characterized a total of 87 “episodes” of HHV-8 shedding, defined as ≥1 day with HHV-8 DNA detected in saliva followed and preceded by a day without HHV-8 DNA detected. Of these 87 episodes, 52 (60%) lasted only 1 day, with the remaining episodes having a duration of 2−36 days. On days when HHV-8 was detected, the median number of copies was 4.5 log₁₀ copies/mL (range, 2.0−7.3 log₁₀ copies/mL). Men observed to have HHV-8 detected frequently in the mouth tended to have high numbers of virus copies detected (r = 0.62; P ≤ .001) (figure 1).

Figure 1.

Correlation between human herpesvirus (HHV)–8 shedding frequency and quantity among 27 participants with HHV-8 detected on at least 1 day. Solid line, fitted regression line.

Figure 2 displays HHV-8 oropharyngeal shedding patterns in the 44 participants. It is notable that HHV-8–infected men seem to fall into 1 of 2 patterns: those with frequent, high-quantity HHV-8 shedding (figure 2,top) or infrequent (or absent) HHV-8 detection at low quantities (figure 2, bottom).

Figure 2.

Human herpesvirus (HHV)–8 oropharyngeal shedding frequency and quantity among immunocompetent men who have sex with men (MSM). Data from the first sampling session for each man are shown. Each row represents an individual subject. Within each row, each vertical bar depicts the quantity of HHV-8 DNA detected in 1 mL of saliva on a given day. Gray bars indicate days on which samples were either not collected or quantifiable, and colors are used to depict the quantity of HHV-8 DNA detected.

Correlates of HHV-8 shedding.

We also examined the correlates of HHV-8 detection in the oropharynx among the MSM. Age, race, HSV coinfection, and number of lifetime sex partners were not significantly associated with the frequency of HHV-8 detection in saliva. The mean quantity of HHV-8 detected in the oropharynx, however, differed by age and race in univariate models. On average, 0.71 log₁₀ fewer copies of HHV-8 were detected in nonwhites than in whites (95% confidence interval [CI], 0.35−1.06 log₁₀ copies lower; P ≤ .001). Similarly, younger men also tended to have HHV-8 detected at lower quantities (0.54 log₁₀ copies lower for those <45 years old vs. those ≥45 years old [95% CI, 0.06−1.02 log₁₀ copies lower]; P = .03). Small sample sizes in racial categories precluded multivariate models. Table 1 depicts the frequency and quantity of HHV-8 oropharyngeal shedding, stratified by race, in greater detail.

Table 1.

Human herpesvirus (HHV)—8 oropharyngeal shedding frequency and quantity, stratified by race/ethnicity.

	Race
Shedding summary	White (n = 38)	Black (n = 3)	Othera (n = 3)
Men with HHV-8 detected in the oropharynx on at least 1 day, no. (%)	23 (61)	3 (100)	1 (33)
HHV-8 oropharyngeal shedders
Days with HHV-8 detected, median percentage (range)b	20 (1.1-100)	94.7, 36.7, 12.9c	1.6c
Days with HHV-8 detected/days sampled, no. (%)d	548/2579 (21)	86/136 (63)	2/182 (1)
Maximum no. of copies per person, median log10 copies (range)b	5.5 (2.6-7.3)	3.0, 5.1, 6.3c	3.5c
Days with HHV-8 detected, median log10 copies (range)d	4.6 (2.0-7.3)	3.9 (2.1-6.3)	3.5d

^aIncludes 2 Latino men and 1 Asian man.

^bPerson-level summary.

^cIndividual data values are shown instead of medians and ranges because of small sample sizes.

^dDay-level summary.

Clinical manifestations of HHV-8 shedding.

A subset of 25 (57%) of 44 participants were asked to complete detailed daily diaries cataloging the presence or absence of common oropharyngeal or systemic clinical symptoms, medications ingested, and days missed from work or school. A total of 944 days were analyzed when both diaries were complete and samples were available for HHV-8 quantification. Participants reported the presence of symptoms on 88 of 944 days, including 10 (8.8%) of 114 when HHV-8 was detected and 78 (9.4%) of 830 when HHV-8 was not detected in the oropharynx (P = .9) (table 2).

Table 2.

Relationship between self-reported illness and human herpesvirus (HHV)—8 oropharyngeal shedding.

	Days when symptom was reported
Symptom	Oral HHV-8 PCR negative (n = 830)	Oral HHV-8 PCR positive (n = 114)	Total (n = 944)
Nausea	10	0	10 (1)
Vomiting	1	0	1 (0.1)
Diarrhea	16	2	18 (2)
Temperature >37.8°C	0	0	0 (0)
Subjective fever	2	1	3 (0.3)
Chills	0	0	0 (0)
Swollen glands	6	0	6 (0.6)
Sore throat	21	2	23 (2)
Rash	0	0	0 (0)
Abdominal cramps	1	1	2 (0.2)
Headache	17	1	18 (2)
Cold	23	0	23 (2)
Unable to work	1	0	1 (0.1)
Stayed in bed	8	4	12 (1)
Sought medical attention	0	0	0 (0)
At least 1 of the above symptoms, no. (%)a	78 (9.4)	10 (8.8)	88 (9.3)

NOTE. PCR, polymerase chain reaction.

^aDoes not equal summation of days with each symptom reported because of days with overlapping symptoms.

HHV-8 shedding over time.

A total of 12 participants returned to the clinic to provide additional sessions of oropharyngeal swabs (table 3). The median time between collection sessions was 293 days (range, 108−1371 days). Four (33%) of 12 participants were not observed to shed HHV-8 in the initial session but shed HHV-8 on ≥1 day during subsequent sessions. No clear trends could be discerned in shedding frequencies, with participants both increasing and decreasing the observed shedding frequency over time.

Table 3.

Human herpesvirus–8 oropharyngeal shedding over time.

Participant	Session 1 shedding rate, %	Time between sessions	Session 2 shedding rate, %	Time between sessions	Session 3 shedding rate, %	Time between sessions	Session 4 shedding rate, %
1	75.9	10.7 months	90.9
2	93.1	5.6 months	100	4.6 months	45.1
3	0	1.1 years	32.2	10.5 months	28.5
4	1.5	1.7 years	13.9
5	42.5	3.7 years	25.8	9.8 months	3.5
6	15.6	5.1 months	3.3	5.2 months	42.8
7	3.1	2.2 months	1.7	3.8 months	0
8	0	3.5 years	9.6	1 year	1.9	10 months	0
9	0	3.8 years	6.4	8.7 months	6.4	11 months	0
10	0	6.2 months	0	3.6 months	3.1
11	0	2 years	0
12	0	5.2 months	0	5.1 months	0	5.3 months	0

DISCUSSION

To our knowledge, our study is the first to extensively characterize HHV-8 replication patterns in the oropharynx, a site that appears to be important for both the acquisition and transmission of HHV-8 infection. We observed a dichotomous pattern of HHV-8 oropharyngeal shedding in HHV-8–seropositive immunocompetent MSM. Some men frequently had HHV-8 detected in high quantities in saliva, whereas others either did not shed, despite many days of observation, or shed infrequently and at low quantities. HHV-8 shedding was not detected in saliva on clusters of consecutive days. The large number of participants with HHV-8 detectable on only 1 or 2 days of a sampling session in our study underscores the importance of frequent sampling in the assessment of mucosal replication patterns of HHV-8 or correctly identifying the presence of HHV-8 replication in the oropharynx of a given individual.

A small number of previous studies have examined HHV-8 oropharyngeal shedding in immunocompetent persons. Our group initially reported that 4 of 9 HIV-negative/HHV-8–seropositive MSM who collected saliva daily at home for extended periods of time had HHV-8 detected on at least 1 day [14]. In areas where KS is not endemic, HHV-8 DNA could be detected in oropharyngeal secretions sampled either once [25] or quarterly [26] from very small groups of healthy individuals. In southern Italy, where classic KS is common, salivary shedding was not detected in a single sampling in one study of persons with serum antibodies to HHV-8 [27], and it was infrequently detected in single salivary specimens from HHV-8–seropositive healthy blood donors [28]. Finally, in areas where KS is endemic, HHV-8 DNA is frequently detected in single samples from the oral cavity of HIV-negative women [15] and from children with sickle-cell disease [29].

Very little has been reported about the clinical manifestations of HHV-8 infection before the development of malignancy. Evidence of a primary infection syndrome is supported by work in both immunocompromised and immunocompetent hosts. The symptoms have ranged from fever and maculopapular rashes in Egyptian children [30] to acute KS [31, 32], fever, hematologic abnormalities, and splenomegaly [31] in renal-transplant recipients and cervical and submental lymphadenopathy with rash in HIV-positive men [33]. No studies to date have defined the symptoms associated with chronic HHV-8 infection. It is possible that, with a more-detailed survey instrument, symptoms could have been identified to coincide with HHV-8 oropharyngeal shedding. It is more likely, however, that HHV-8 has similarities with other human HHVs, such as EBV, CMV, and HHV-6. In each of these infections, symptoms similar to those reported with HHV-8 are seen during primary infection, and salivary shedding in chronically infected persons is asymptomatic.

An important limitation of the present study is that the time of primary infection with HHV-8 could not be established. Previous studies have found that oropharyngeal shedding with both EBV [10] and HHV-6 [13] is more frequent and at higher numbers during primary infection. One plausible explanation for the apparent dichotomy in HHV-8 shedding frequencies seen among participants in our study is that some had been recently infected with HHV-8. This explanation, however, seems to be unlikely for a number of reasons. First, all participants in the study already had antibodies to HHV-8 detectable in serum at study entry. Second, despite the suggestion that primary HHV-8 infection may be associated with a primary infection syndrome, participants with more-frequent HHV-8 shedding were not more likely to report symptoms. Finally, in the small subset of participants for whom multiple observations of shedding frequency over time were available, no pattern could be easily discerned. For example, participant 1 was one of the highest shedders in the study and shed at high rates for a period of nearly 1 year. Conversely, participant 3 was initially not observed to shed HHV-8 from the oropharynx but, 2 years later, HHV-8 was frequently detected in his salivary samples. These results highlight the need for further description of the virologic and clinical characteristics of primary HHV-8 infection.

Two additional limitations of the present study are important to mention. Our cohort was relatively homogeneous in their demographics—they were predominantly white, middle-aged MSM. Conclusions about differences in shedding patterns by demographic characteristics were, as a consequence, based on a small number of participants, and larger studies are required for validation. Additionally, serodiagnostic testing for HHV-8 infection has proved challenging [34], and it is possible that persons who were not found to shed HHV-8 in saliva were erroneously considered to be HHV-8 infected on the basis of false-positive antibody testing. We have previously shown, however, that although our serodiagnostic testing strategy may fail to detect serum antibodies to HHV-8 in persons known to be infected with HHV-8, <3% of persons with detectable HHV-8 antibodies were unlikely to harbor HHV-8 infection [18]. A more plausible explanation for the failure to detect HHV-8 shedding in HHV-8–seropositive participants might be an insufficient duration of observation. Of the 44 MSM participants, the 27 observed to shed HHV-8 had a median of 75 days (range, 27−248 days) of samples, whereas the 17 not observed to shed HHV-8 only had a median of 31 days (range, 25−128 days) of samples. These results argue strongly for the importance of extended periods of observation to accurately characterize mucosal shedding patterns of HHV-8.

Asymptomatic mucosal shedding has been well described for all HHVs except for varicella-zoster virus. Work with EBV, the only other human γ-herpesvirus and the virus most closely related to HHV-8, has shown that asymptomatic shedding in saliva is both frequent and likely the predominant mode of transmission [35, 36]. If asymptomatic salivary shedding is also the predominant mode of HHV-8 transmission, it remains unclear why the virus is restricted to distinct geographic and demographic populations [37]; all other human herpesviruses transmitted by saliva (HSV-1, EBV, CMV, and HHV-6) are ubiquitous worldwide. One hypothesis is that the frequency and quantity of HHV-8 shedding may differ among the groups at high risk for HHV-8. No comparative studies of HHV-8 mucosal shedding between persons from high-risk groups and those in the general population have been performed to date, and the small number of heterosexual men in the present cohort precluded such a comparison. It is interesting to note, however, that the frequency and quantity of HHV-8 shed from the oropharynx in these 2 individuals was lower than the median rates observed in MSM. The frequent and occasionally high-titer shedding of HHV-8 in this group of immunocompetent Seattle MSM, as well as of others that we have previously described [38], may help explain why nearly one-quarter of HIV-negative MSM in the United States and western Europe are infected with HHV-8 [39] and why HHV-8 appears to be an uncommon infection in the general population of these regions.

The striking difference between those with frequent and infrequent oropharyngeal HHV-8 reactivation suggests that a combination of host and/or viral factors may govern the control of HHV-8 replication. Aside from immunodeficiency, the presence of replicating HHV-8 in the peripheral blood has been the only identified risk factor for the development of KS from chronic HHV-8 infection [40–44]. The degree to which the lack of control of HHV-8 replication in the oropharynx is associated with viremia or disease progression is not known, but it should be examined in future studies.

Despite impressive decreases in the incidence of KS in the United States and western Europe with the widespread implementation of highly active antiretroviral therapy (HAART) [45], HHV-8 infection remains a problem worldwide. KS is at present the most common malignancy in many areas hard hit by the intersection of HIV and HHV-8 [46], and treatment for persons with KS without access to HAART or with endemic and classic KS is lacking. Comprehensive studies of the mucosal replication of HHV-8 may allow for effective strategies to prevent acquisition and predict those at highest risk for complications from chronic infection.