Varicella-Zoster Virus DNA in Blood After Administration of Herpes Zoster Vaccine

Varicella-zoster virus (VZV) DNA is common in the blood after administration of live-attenuated herpes zoster vaccine. The extent that this reflects viral replication is determined by the nature of the VZV-specific T-cell immunity after vaccination.

Abstract

We studied the relationship between varicella-zoster virus (VZV) DNAemia and development of VZV-specific immunity after administration of live-attenuated zoster vaccine. VZV-DNAemia, detected by polymerase chain reaction (PCR), and VZV-specific effector (Teff) and memory (Tmem) T cells, was measured in 67 vaccinees. PCR was positive in 56% (9 direct, 28 nested) on day 1 and in 16% (1 direct, 10 nested) on day 14. Teff progressively increased in direct-PCR-positive vaccinees up to day 30, but Tmem did not. Conversely, Tmem, but not Teff, increased in direct-PCR-negative vaccinees on day 7. The kinetics of these immune responses and VZV DNAemia suggested that direct-PCR sample positive represented viremia.

Varicella, which results from the primary infection with varicella-zoster virus (VZV), is defined by the clinical manifestations of blood-borne spread of VZV. This viremia was documented more than 50 years ago by virologic techniques, and its kinetics in blood have been described in detail using molecular techniques to identify VZV DNA (reviewed in [1]). VZV viremia and DNAemia are also commonly detected when reactivation of latent VZV causes herpes zoster (HZ) [1].

Vaccine strain VZV administered as varicella vaccine is also likely to result in viremia. This is inferred from the presence of skin lesions distant from the site of inoculation in 3%–5% of varicella vaccine recipients [2]; by the occurrence of HZ caused by vaccine strain VZV in dermatomes distant from the site of vaccine administration [3]; and by the presence of VZV DNA in the oropharynx shortly after administration of HZ vaccine live (ZVL), presumably as a result of viremia [4]. This study describes the characteristics and kinetics of VZV DNAemia after administration of ZVL, and analyzes VZV DNAemia in the context of VZV-specific immune responses.

METHODS

Clinical

Heparinized blood (5 mL) was obtained from participants in a study of immune responses to the licensed ZVL (NCT01331161) [5]. The trial was approved by IRBs at the 2 study sites. Samples to determine VZV-specific adaptive immune responses were obtained at days 0, prior to vaccination and at days 1, 3, 7, 14, and 30. Samples for VZV DNAemia were obtained simultaneously, stored at −80°C, and the Emory site samples shipped on dry ice to the Colorado site for analysis.

Laboratory

DNA Extraction

This utilized the QuickGene DNA Whole Blood kit (Autogen). Samples were thawed and 200 μL of whole blood mixed with 250 μL of lysis buffer and 30 μL of protease, incubated at 56°C for 2 minutes, and 250 μL of ethanol added. The treated samples were transferred onto DNA-binding cartridges and extracted according to the manufacturer’s directions.

Polymerase Chain Reaction

Each polymerase chain reaction (PCR) tube contained 200 ng of extracted DNA in 5 μL of water and 15 μL Master Mix. VZV genome open reading frame 63 (ORF63) was detected with the first set of primers listed in Supplementary Table 1. In order to increase the sensitivity to detect VZV DNA ORF63, nested PCR analysis was carried out on 1/40 volume of any PCR product that did not initially contain detectable VZV DNA, using a second pair of VZV ORF63 primers. The downstream nested primer (V63-315d) was 5′-GTGCTGGGAGGAATTGTTACAG-3′ and the upstream primer (V63-500u) was 5′-TCGTCGCT ATCGTCTTCACCAC-3’. Amplification was performed with a RotorGene (RG-3000) real-time PCR System. Complete details on the PCR methods are in the Supplementary Methods.

VZV-Specific Responder Cell Frequency

VZV-specific CD4⁺ memory T cells were enumerated by adding a limiting dilution step to a conventional lymphocyte proliferation assay as previously described [6]. The details of this assay are provided in Supplementary Methods.

Flow cytometric enumeration of VZV-specific T-cell subsets were determined by flow cytometry as previously described [5]. Details are provided in Supplementary Methods.

Statistics

The database from the primary experiment linked demographic and immunologic data from each subject to the PCR data. This was done by correlation analysis using Prism 6.0 software (Graphpad). Significance was defined by P < .05.

RESULTS

Prior to measuring VZV DNA in the blood of ZV recipients, we evaluated 5 published primer sets for the PCR assay of VZV DNA in blood and saliva. The current study used a primer set that varied by 1 nucleotide from a prior publication [7]. The limit of detection for each of these is shown in Supplementary Table 1. The primer for ORF63 providing the greatest relative sensitivity was chosen for the DNAemia experiments reported here.

Detection of VZV DNA in human blood is influenced by common steps involved in preparing and storing DNA for PCR analysis. Supplementary Table 2 summarizes evaluation of 3 such steps. Brief vortexing and a small number of freeze-thaw cycles decreased detection of VZV DNA by PCR, while the presence of host DNA in clinical samples from whole blood was a greater impediment to detection of VZV DNA. The content of host DNA extracted was 4 to 8 µg/200 µL of whole blood; 200 ng was routinely present in PCR tubes. The sensitivity of the PCR assay calculated from reconstruction experiments while optimizing these factors was 200 copies (cp)/mL in water and 500 cp/mL in whole blood (Supplementary Figure 2). Because we measured VZV DNA in whole blood, we determined the effect of delay in processing. Survival of VZV DNA in whole blood was measured in a reconstruction experiment (Supplementary Figure 3). This demonstrated that the half-life of the input DNA of either source was 4–5 hours at 37°C. At room temperature only 2% of cell-free DNA was lost per hour, while the loss was approximately 3-fold greater at 37°C.

VZV DNAemia after each postvaccine interval is summarized in Table 1. Each clinical sample was analyzed in triplicate. Samples negative in the direct PCR assay (prior to being retested in a nested PCR format) were re-evaluated in triplicate in a nested PCR format. The negative results for 67 blood samples obtained prior to vaccination, including nested PCR assays, confirm the absence of contamination within the laboratory. The decline in positive PCR results with increasing interval after vaccination is an additional indication that direct and nested PCR results were specific and not randomly contaminated. The number of positive samples declined with successive intervals (37 to 11); the number of direct PCR-positive samples declined during these intervals (9 to 1); and the proportion of positive direct to nested PCR results fell (24% to 9%). The number of copies in positive specimens was variable, but the mean and upper limit of the range of direct PCR-positive samples increased with time after vaccination. The temporal clustering of sequential samples in subjects with any positive test for VZV DNA (Supplementary Table 3) did not reveal a consistent pattern of relationship to prior tests. Five samples were positive on day 3 and not positive on day 1, although most were positive on both days. Five samples were positive on day 7 only, but were usually positive on day 1 and 3 when positive on day 7. Two samples were positive on day 14 only. Four samples were positive at 11 to 13 days after a prior positive. There was no relationship of DNAemia to the age of the vaccinee. The results for each sample are shown in Supplementary Table 4.

Table 1.

Varicella-Zoster Virus (VZV) DNA Detected in Blood After Zoster Vaccination^a

Day	PCR Positive/Total testedb	Initial PCR Positive	Nested PCR positivec	Copies/mL whole blood (mean)d	Copies//mL whole blood (range)d
0	0/67	n/a	0	n/a	n/a
1	37/66	9	28	1680	140–2800
3	31/65	5	26	2135	35–4550
7	21/66	5	16	4200	1260–10150
14	11/65	1	10	4130	n/a

^aBased on Supplementary Table 1.

^b Polymerase chain reaction (PCR) results are based on 0.2 mL of whole blood.

^cNested PCR performed on 1/40 volume taken from each initial PCR tube that was negative for VZV DNA.

^dApplies only to samples positive in initial PCR.

Abbreviation: n/a, not applicable.

VZV cell-mediated immunity (CMI) was analyzed in conjunction with VZV DNAemia to test the hypothesis that the profile of VZV-CMI responses would correlate with the magnitude/duration of the VZV DNAemia as a reflection of viral replication. To test this relationship we compared VZV-specific memory T cells (Tmem) measured by the responder cell frequency (RCF) assay, and Tmem and effector T-cell (Teff) responses measured by flow cytometry, between participants who had VZV DNAemia at any visit by direct PCR and those who did not have a positive direct PCR result (ie, either DNAemia detected only by nested PCR or no DNAemia). Before vaccination, there were no significant differences in VZV-CMI between the direct DNAemia and no DNAemia groups. However, Tmem and Teff responses after immunization distinguished the 2 groups. In the first 30 days after immunization, participants without DNAemia detected by direct PCR had significant increases in the proportion of VZV-specific CD4⁺ and CD8⁺ Tmem measured by flow cytometry, whereas participants with DNAemia detected by direct PCR did not (P ≤ .049) (Figure 1A). This was confirmed by RCF assay (P = .01). Different kinetics was observed for the proportion of CD4⁺ and CD8⁺ Teff (Figure 1B). Participants with and without DNAemia detected by direct PCR had similar Teff responses at baseline and day 7 after immunization. However, only participants with DNAemia by direct PCR showed a continued increase in VZV CD4⁺ Teff by flow cytometry up to 30 days after immunization (P = .001); CD8⁺ Teff were also marginally higher in the DNAemia group compared with those who did not have DNAemia at day 30 (P = .11). A sensitivity analysis comparing participants that were positive for VZV DNA detected by direct PCR with those without any DNAemia (either direct or nested PCR) was confirmatory (Supplementary Figure 3).

Figure 1.

Relationship of DNAemia to varicella-zoster virus (VZV)-specific memory T-cell (A) and effector T-cell (B) responses after zoster vaccine. Data were derived from participants who had immunological test results. This included 21 participants who did not have DNAemia detected by direct polymerase chain reaction (PCR) and 13 who had DNAemia by direct PCR. Bars represent means and standard error of the means (SEM). P values were calculated by unpaired t test after verifying that the data were normally distributed. Abbreviation: RCF, responder cell frequency.

DISCUSSION

Viral nucleic acid in blood and body secretions is frequently used to: (1) measure the extent of virus replication during an infection; (2) evaluate response to therapy; or (3) measure dissemination of vaccine virus after vaccination [1, 4]. Detection of viral DNA or RNA in these settings is often equated with infectious virus. However, this concordance is rarely established, and there is typically a large discrepancy in viral genome number and infectious virus recovered in culture. For example, VZV DNA is readily detected in the oropharynx after varicella, when virus cannot be isolated [8]. VZV DNA can be detected in the oropharynx during herpes zoster, whereas isolation of infectious VZV has not been reported [9].

After optimizing PCR methods, we determined that VZV DNAemia was frequently detected by PCR within 3 days after administration of ZVL, and continued in some vaccinees at 7 and 14 days. DNAemia was detected by direct PCR only in 20% to 30% of the positive samples and the remainder by nested PCR. This may be an underestimate because in optimization experiments we demonstrated that blood sample processing influences DNA detection, and that large amounts of host DNA decrease PCR sensitivity. This caution is applicable to numerous previous reports of VZV DNAemia, especially for whole blood samples. Our finding of DNAemia after ZVL is consistent with the reported appearance of VZV DNA in the oropharynx within 1–3 days after vaccination [4].

VZV DNAemia could represent either virus or noninfectious VZV DNA that escaped from the vaccination site into the blood, possibly protected by cells or protein complexes. ZVL administration results in inoculation of 10¹¹ copies of VZV DNA, thus making passive presence in the blood a strong contributor. We also showed that the half-life of VZV DNA in blood in reconstruction experiments is surprisingly long. Infectious VZV is also likely to enter the blood stream after replication at the inoculation site. Several parallel observations suggest that VZV DNAemia detected by direct PCR is at least partially the consequence of VZV replication. These observations include the increase over the first week postvaccination in mean and upper limit of VZV copy numbers, and persistence beyond 3 days, both consistent with virus replication. Furthermore, our findings validate reports of rash or HZ at a site distant from that of ZVL inoculation, suggesting viremic spread of vaccine strain VZV [3].

In a previous report on VZV-CMI during HZ we showed that an early increase in VZV-specific Tmem was associated with limited VZV replication, as reflected by the absence of postherpetic neuralgia [10]. This is in accordance with a switch to memory responses after control of viral replication. In the current report, an early increase in VZV-specific Tmem was also associated with absence of DNAemia detected by direct PCR. Based on the Tmem kinetics during HZ and post-ZVL, we postulated that the absence of DNAemia detected by direct PCR reflected limited vaccine virus replication. Consistent with this is the observation of robust VZV-specific Teff responses as late as 30 days after ZVL when DNAemia was detected by direct PCR, suggesting that a prolonged effector immune response was required to limit viral replication.

Distinguishing VZV viremia from DNAemia would inform several issues in clinical virology: (1) it would indicate the potential for vaccine strain VZV to reach ganglia distant from the vaccination site and potentially cause HZ [3]; (2) it would indicate how long to delay immune suppression in patients receiving ZVL prior to such therapy [11]; (3) it would clarify how viremia, rather than noninfectious VZV components in the blood, influences the nature of the postvaccine immune response [12]; and (4) it would indicate how pre-existing immunity of recipients of ZVL might alter replication of vaccine strain virus and affect the efficacy of the vaccine.

Supplementary Data

Supplementary materials are available at The Journal of Infectious Diseases online. Consisting of data provided by the authors to benefit the reader, the posted materials are not copyedited and are the sole responsibility of the authors, so questions or comments should be addressed to the corresponding author.

Notes

Acknowledgments. We thank Mary Bower, Jennifer Whitaker, Ildefonzo Tellez, Srilatha Edugupanti, Tianwei Yu, David Rimland, and Bruce Ribner at Emory.

Financial support. This work was supported by NIH grant number U19AI090023.

Potential conflicts of interest. M. J. L. has research funding, intellectual property, and advisory board participation for Merck, Sharpe & Dohme; research funding and advisory board participation for GlaxoSmithKline. A. W. has research funding and spousal intellectual property with Merck, Sharpe & Dohme; research funding from GlaxoSmithKline. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.