Sensitive Method for Detection of Human Herpesviruses 6 and 7 in Saliva Collected in Field Studies

Abstract

To facilitate studies of the epidemiology and natural history of human herpesviruses 6 and 7 in infants, a practical method for collecting and quantifying the DNA of these viruses was developed. Saliva was collected using small strips of filter paper, and virus was detected using a real-time quantitative fluorescent-probe PCR assay. The sensitivity and specificity of this method even after prolonged drying of the specimens compared favorably to those of our traditional method of collecting and assaying saliva.

Human herpesvirus 6 (HHV-6) and human herpesvirus 7 (HHV-7) are almost universally acquired by 2 to 3 years of age (7, 8, 14, 16). Salivary viral shedding appears to begin within days of primary infection (2, 19) and persists in most individuals (6, 7). Thus, the detection of virus in the saliva provides a noninvasive means of recognizing HHV-6 and HHV-7 infections, and a new onset of salivary shedding can be used to detect recent acquisition of these viruses. We sought to develop a simple, reliable method for detecting HHV-6 and HHV-7 in the saliva of infants.

Small commercially available filter paper strips have been utilized as specimen collection devices for the detection of both viral antigens and antibodies in local secretions (3, 4, 20; P. Reighelderfer, R. Coombs, D. Wright, D. Burns, and A. Kovacs, Abstr. 38th Intersci. Conf. Antimicrob. Agents Chemother., abstr. I-251, 1998). We evaluated the use of these filter paper strips for the collection of saliva and the detection of HHV-6 and HHV-7 DNAs.

Experimental design and collection of saliva.

Sno strips (Chauvin Pharmaceuticals Ltd., Essex, England) are strips (60 by 6 mm) of sterile filter paper designed to quantify tear flow. Approximately 12.5 μl of fluid has been shown to saturate the filter paper (data not shown). Four different experiments were completed to evaluate both the sensitivity of Sno strips as a means of collecting saliva for HHV-6 and HHV-7 PCR detection and the effects of environmental conditions on this assay.

First, the detection of HHV-6 by PCR was used to compare saliva collected by Sno strips and saliva collected by expectoration into a cup (7). Saliva was collected from 33 healthy adults both by placement of five Sno strips in the mouth until the distal ends were saturated and by expectoration of approximately 0.4 ml of saliva into a sterile cup. Specimens were stored at −20°C until the DNA was extracted with phenol-chloroform, amplified in a Perkin-Elmer 9600 thermocycler, and assayed by liquid hybridization (method 1) (7). Second, using method 1, PCR to detect HHV-7 was performed on the remaining extracted DNA from the saliva collected by Sno strips. Third, duplicate sets of five Sno strips were collected from 10 individuals to compare method 1 with a more automated and quantitative method that uses a Perkin-Elmer 7700 automatic thermocycler (method 2). For method 2, DNA was extracted using Qiagen columns (QIAmp Blood Kits; Qiagen Inc., Valencia, Calif.), and PCR was performed using a real-time quantitative fluorescent-probe PCR assay (Perkin-Elmer Applied Biosystems, Foster City, Calif.). Finally, four additional samples of five Sno strips each were obtained from seven individuals. One sample from each individual was frozen immediately, while the rest were allowed to air dry for 1 day, 1 week, and 2 weeks before storage at −20°C to compare what effect prolonged drying might have on our ability to detect HHV-6 and HHV-7 viral DNAs. These samples were analyzed using method 2.

DNA extraction and PCR methods. (ii) Method 1.

Method 1, utilizing phenol-chloroform DNA extraction and liquid hybridization, was performed as previously reported (7). Expectorated saliva (0.4 ml) or the distal portions of five Sno strips cut at the shoulder (60 μl) were treated overnight with proteinase K. DNA was extracted using phenol-chloroform (7). The HHV-6 and HHV-7 DNAs from 12.5 μl of saliva from Sno strip specimens and 20 μl of saliva from expectorated specimens were then amplified by PCR (Perkin-Elmer 9600 thermocycler). The primer pair 5R and the probe 5R-P were used for strain common detection of HHV-6 (7). The HHV-7 primer sequences HHV7-1 (5′ CGG CGT TTT ACT CGG AAC TCC T 3′) and HHV7-2 (5′ TCC CCA TAA CAA ATG TGC CAT AAG A 3′) amplified a 116-bp portion of the major capsid protein. The probe HHV7-p1 (CAG ATT TTG TCC AAC GCC CTA TC), end labeled with ³²P, was used to detect the amplicon by liquid hybridization and autoradiography (7). These primers and probes have been found to reliably detect 10 copies of purified HHV-6 or HHV-7 DNA and are specific when tested with herpes simplex virus types 1 and 2, cytomegalovirus, Epstein-Barr virus, varicella-zoster virus, and HHV-6 (for HHV-7 primers) (data not shown), HHV-7 (for HHV-6 primers) (7), and HHV-8 (data not shown) target DNAs.

Negative controls included HSB-2 cultured T cells coprocessed with every five study specimens, and at least one sample with all reaction components, except input DNA, was included in every experiment. Positive controls were also included in every experiment. Fifty copies of the internal control, the HHV-6 or HHV-7 amplicon with 21 bp of the probe sequence replaced by unrelated Drosophila DNA (fly control) (5), were added to each PCR mixture prior to amplification. In addition, every experiment included a dilution series of cloned amplicon DNA that consisted of 10¹, 10², 10³, and 10⁴ copies.

(ii) Method 2.

Method 2 utilized Qiagen columns to isolate DNA from Sno strips. The saturated ends from five Sno strips in 400 μl of ATL tissue lysis buffer (Qiagen) with 800 μg of proteinase K were incubated overnight at 55°C. On the following day, 400 μl of AL lysis buffer (Qiagen) was added. After incubation at 70°C for 10 min, 420 μl of 100% ethanol was added. The mixture, including the strips, was then loaded into the columns and centrifuged at 6,000 × g for 2 min. The DNA was eluted with 100 μl of 10 mM Tris.

Real-time PCR, a sensitive and reproducible method for the detection of viral DNA (17), was used to detect HHV-6 and HHV-7 DNAs. Real-time PCR uses the 5′→3′ exonuclease activity of Taq polymerase to digest an internal probe labeled with two fluorescent dyes, the reporter and the quencher (10). When the probe is intact, the reporter and quencher dyes undergo fluorescent resonance energy transfer, thereby suppressing the fluorescence of the reporter dye (9). When target DNA is present, upon primer elongation, the probe is cleaved by the 5′→3′ exonuclease activity of Taq polymerase. The reporter dye is then no longer physically attached to the quencher dye on the probe, and fluorescent resonance energy transfer is interrupted. This results in an increase of reporter dye fluorescence that is proportional to the amount of PCR product accumulated. DNA copy number is determined by calculation of the number of PCR cycles necessary for a standard curve of known amounts of DNA (10⁴, 10³, 10², and 10¹) to cross a fluorescent threshold and then interpolation of the unknowns. The threshold is set at the beginning of the exponential phase for the amplification being run.

Each 50 μl of PCR mixture contained 20 μl of purified solution containing the specimen DNA, 830 nM primers, 100 nM probe, 8% glycerol, 60 nM internal passive control (6-carboxy-x-rhodamine conjugated to the 5′ end of 5′-GATTAG-3′), 5 mM MgCl₂, 200 μM each deoxynucleoside triphosphate (dNTP) (except for 400 μM dUTP), 2.5 U of AmpliTaq polymerase (Perkin-Elmer), 1.25 U of Taqstart Taq polymerase antibody (Clontech Inc., Palo Alto, Calif.), 0.05 U of uracil-N-glycosylase (UNG), and 50 copies of internal control DNA. The 5R probe for the HHV-6 amplicon was labeled with the reporter dye, 6-carboxyfluorescein (FAM), on the 5′ end and the quencher dye, 6-carboxytetramethylrhodamine (TAMRA), on the 3′ end. The forward primer for HHV-7 real-time PCR was 5′-TTT CCT GTG ACA AAA GAA GCA GTT A, and the reverse primer was 5′-ATC CCA CAC GCT TTA CGG G. The sequence and labels of the HHV-7 probe were 5′-FAM-TTC CTG CGC AAT AAA GTG AAA ACT GTT AGC ATT-3′-TAMRA. The internal control fly probe was labeled with the reporter 6-carboxytetramethylrhodamine on the 5′ end and TAMRA on the 14th base (A). A minor group binding protein (Epoch Inc., Bothell, Wash.) was added to the 3′ end of the fly probe to increase the melting temperature and optimize its performance for real-time PCR. With a Perkin-Elmer 7700 automatic thermocycler, the PCR cycling temperatures were as follows: after 2 min of incubation at 50°C followed by 2 min at 95°C, the samples were subjected to 45 cycles of 95°C for 20 s followed by 60°C for 1 min. A true FAM-negative PCR must have a positive 6-carboxytetramethylrhodamine signal; otherwise, the reaction is interpreted as inhibited.

Saliva from 31 of 33 adults (94%; 95% confidence interval [CI], 80 to 98%) had detectable HHV-6 DNA. The two individuals who did not have HHV-6 detected in their saliva did not have HHV-6-specific antibody found by Western blot testing and were considered non-HHV-6 infected. A comparison of HHV-6 detection in specimens collected from the infected individuals by expectoration and by Sno strips revealed six specimens that were discordant: three that were negative by Sno strip and positive by expectoration and three that were positive by Sno strip but not by expectoration (the PCR was inhibited). Thus, the Sno strip and expectoration methods were 90% sensitive (28 positive of the 31 positive by either method) (95% CI, 75 to 96%).

Of the available Sno strip specimens, 28 of 29 (97%; 95% CI, 83 to 100%) had amplified HHV-7 DNA. HHV-7 serologic testing of the individual with no HHV-7 in his saliva was not done, and the expectorated specimens were not tested.

A comparison of phenol-chloroform DNA extraction followed by conventional PCR and liquid hybridization (method 1) to DNA extraction by Qiagen columns followed by real-time PCR (method 2) revealed that the latter method had a sensitivity similar to that of the former in detecting HHV-6 DNA from specimens collected by Sno strips. Overall, 9 of 10 specimens (90%) analyzed by conventional PCR and liquid hybridization were positive for HHV-6 DNA compared to 10 of 10 specimens (100%) analyzed by real-time PCR. The DNA copy number, as determined by real-time PCR, ranged from 463 to 29,683 copies/ml. The sample with the lowest copy number, 463 copies/ml, was the sample that was found negative by method 1.

There did not appear to be a consistent difference in the detection of HHV-6 or HHV-7 DNA from samples that were frozen immediately or allowed to dry for 1, 7, or 14 days at room temperature prior to freezing and then were tested. The mean copy numbers between sample groups dried for different amounts of time were not significantly different (data not shown). Interperson variability was high; however, viral copy number did not vary much within an individual's samples (Fig. 1).

FIG. 1.

Quantities of salivary HHV-6 and HHV-7 DNAs detected in specimens collected by Sno strips and allowed to dry for different lengths of time prior to assay. (A) Saliva was collected from six different individuals and allowed to dry for different time periods prior to PCR. The graph depicts the log-transformed HHV-6 DNA copies per milliliter of saliva at the different time points for the six subjects. (B) Saliva was collected from six different individuals and allowed to dry for different time periods prior to PCR. The graph depicts the log transformed HHV-7 DNA copies per milliliter of saliva at the different time points for the six subjects.

Sno strips appear to provide a simple, noninvasive, yet sensitive means of collecting saliva in the field for the detection of HHV-6 and HHV-7 DNAs by PCR. Overall, 90% of HHV-6-infected adults had viral DNA detected in their expectorated saliva, and 90% had DNA detected in saliva collected by Sno strips. Previous studies that have assayed a larger volume of adult saliva (20 to 30 μl) by PCR have reported frequent (90 to 100%) detection of HHV-6 and HHV-7 DNAs (1, 7, 11, 12, 15). Other studies, in which saliva was collected using devices such as cotton swabs and throat swabs, have reported a lower and variable frequency of detection of HHV-6 (3 to 67%) (13, 18). Although these studies did not report the actual volume of saliva assayed, it may be that smaller volumes of saliva were used. Relatively less specimen was assayed when collected by Sno strips (10 to 15 μl) than by expectoration (20 μl). Autoradiographic images produced from the Sno strip specimens were less intense than those produced from the expectorated specimens (Fig. 2), indicating that relatively less HHV-6 DNA was amplified from the specimens collected by Sno strips. Furthermore, compared to the results for other subjects, there was relatively little HHV-6 DNA in the samples of the three individuals whose expectorated saliva was HHV-6 DNA positive and Sno strip saliva negative. This result suggests that the limit of detection had been reached for the individuals with relatively lower viral DNA levels in their saliva.

FIG. 2.

Autoradiogram of patient specimens (lanes 1 to 3 and 5 to 7) and a negative control (lane 4). Lanes: A, amplicons from specimens collected by Sno strips; B, those collected by expectoration. Lanes A have a relatively weaker signal or no signal (patient 6) compared to their paired lanes B. Patient 6 appears to have less viral DNA in her saliva than patients 1, 2, 5, and 7. HHV-6 DNA collected in 10 to 15 μl of saliva by Sno strips appears to be the limit of detection by method 1.

Increasing the volume of saliva above the 10 to 15 μl that can be wicked onto Sno strips would not necessarily increase the overall sensitivity of the test. Inhibitors of PCR became evident in the saliva of three individuals when more saliva (20 μl) was assayed from expectorated specimens. It is not evident whether the lack of PCR inhibition for the specimens collected by Sno strips was due to the smaller specimen volume or to the removal of inhibitors by the filter paper.

Method 2 may offer a slight increase in sensitivity over method 1. The 1 of 10 Sno strip specimens that tested negative for HHV-6 when phenol-chloroform was used to extract the DNA, followed by routine PCR (method 1), had a low copy number of HHV-6 when DNA was isolated with Qiagen columns and quantified by real-time PCR (method 2). DNA extraction with Qiagen columns followed by real-time PCR offered several clear advantages, including rapidity (it required approximately half the time to process specimens) and objective quantification of the DNA copy number, potentially allowing for a better understanding of viral dynamics. In addition, pertinent to field studies, drying saliva collected with Sno strips for up to 2 weeks did not appear to significantly affect the isolation of viral DNA.

In summary, Sno strips for saliva collection, Qiagen column isolation of DNA, and fluorescent, real-time PCR provide a convenient and sensitive method for the study of HHV-6 and HHV-7 DNA shedding.