Amplification of the Six Major Human Herpesviruses from Cerebrospinal Fluid by a Single PCR

Abstract

We used a novel type of primer system, a system that uses stair primers, in which the primer sequences are based on consensus sequences in the DNA polymerase gene of herpesvirus to detect herpesviruses by PCR. A single PCR in a single tube detected the six major herpesviruses that infect the central nervous system: herpes simplex virus type 1 (HSV-1), and type 2 (HSV-2), cytomegalovirus (CMV), Epstein-Barr virus (EBV), varicella-zoster virus (VZV), and human herpesvirus 6 (HHV-6). We used the technique to analyze 142 cerebrospinal fluid (CSF) samples that had been stored at −80°C and compared the results with those obtained previously for the same samples by standard, targeted PCR. Four hundred one targeted PCR tests had been run with the 142 samples to detect HSV-1, HSV-2, CMV, and VZV; screening for EBV and HHV-6 was not prescribed when the samples were initially taken. Eighteen CSF samples tested positive by classic targeted PCR. The herpesvirus consensus PCR detected herpesviruses in 37 samples, including 3 samples with coinfections and 17 viral isolates which were not targeted. Two samples identified as infected by the targeted PCR tested negative by the consensus PCR, and eight samples that tested positive by the consensus PCR were negative by the targeted PCR. One hundred three samples scored negative by both the targeted and the consensus PCRs. This preliminary study demonstrates the value of testing for six different herpesviruses simultaneously by a sensitive and straightforward technique rather than screening only for those viruses that are causing infections as suggested by clinical signs.

Human herpesviruses, particularly herpes simplex virus type 1 (HSV-1), and type 2 (HSV-2), cytomegalovirus (CMV), Epstein-Barr virus (EBV), varicella-zoster virus (VZV), and human herpesvirus 6 (HHV-6), are major agents of central nervous system infections. The clinical signs produced by the six viruses are not viral taxon specific. Accurate etiological diagnosis is essential now that effective and virus-specific therapy (for example, acyclovir, gancyclovir, and foscarnet) is available. Isolation of these viruses by culture of cerebrospinal fluid (CSF) gives poor results, and PCR is now the method of choice for virus detection (3, 10). Generally, a series of independent PCRs in which each PCR detects a single virus is performed. However, human herpesvirus DNA polymerase genes contain highly conserved spans of nucleotides that encode the same amino acids (7, 17). Since 1990 (16), several PCR methods for the amplification of these regions have been developed. In 1991, an amplification system became available for HSV-1, HSV-2, CMV, and EBV. This system uses a single pair of consensus primers whose sequences correspond to a conserved region of the DNA polymerase gene (11). Multiplex PCR was described in 1993 for the detection of HSV-1, HSV-2, CMV, EBV, and VZV (13). Consensus primers and virus-specific probes were designed for HSV-1, HSV-2, and VZV in 1994 (1). Degenerate primers have been described for 22 species of herpesviruses (human and animal viruses), but they have not been used for clinical diagnosis (14).

We have described a new type of primer, the stair primer, which is designed to overcome the problems encountered with standard primers if there are mutations in the sequence of the priming region (4). We used this approach to amplify a well-conserved region of the DNA polymerase gene and simultaneously detected the six major human herpesviruses in a single reaction (9). Here we report the results of a preliminary study that was performed to evaluate the feasibility of the technique for CSF samples.

MATERIALS AND METHODS

Biological samples.

We studied all available CSF samples sent to the laboratory in 1995 for herpesvirus testing, irrespective of the clinical reasons for sample collection. The samples had thus already been tested by targeted PCR at the request of the clinicians involved. All samples were stored frozen at −80°C.

Reference viral strains HSV-1 ATCC VR-733, HSV-2 ATCC VR-734, CMV AD169 (ATCC VR-538), EBV ATCC line VRL-1612, an HHV-6 strain (kindly provided by Hélène Colandre, Institut Pasteur, Paris, France), and a strain of VZV (VZRPX, which was isolated in our laboratory) were used as positive controls for amplification with herpesvirus consensus primers and specific primers. For HSV-1, HSV-2, CMV, and VZV, MRC-5 cells were infected in 75-cm² flasks at a dose that gave a 50% cytopathic effect after 48 to 72 h. The culture medium was removed and the cells were harvested by scraping. HHV-6 was obtained by coculture with donor lymphocytes, and EBV was obtained from B95-8 cells maintained in suspension; the cells were harvested by centrifugation.

Sample preparation.

DNA was extracted from 90-μl aliquots of CSF. Ten microliters of 10× lysis buffer (1 M Tris HCl [pH 7.4], 0.1 M EDTA, 5% sodium dodecyl sulfate, 2 mg of proteinase K per ml) was added. The samples were incubated for 1 h at 37°C and were then extracted twice with phenol-chloroform-isoamyl alcohol (25/24/1; vol/vol). The DNA was precipitated with absolute alcohol in the presence of 0.3 M sodium acetate. The DNA pellets were washed with 70% alcohol, dried, and suspended in 30 μl of sterile distilled water. The resulting DNA preparations were stored at −80°C. For reference viruses, the infected cells were lysed in 2× lysis buffer, and DNA was prepared as described above. DNA preparations were stored at −80°C. Negative controls were prepared by repeating the DNA preparation protocol, but CSF was replaced with distilled water and infected cells were replaced with uninfected cells.

Amplification with herpesvirus consensus primers.

The pair of stair primers (PolyHer1 and PolyHer2) were produced by Argène Biosoft, Varilhes, France (Hybridowell Herpes Consensus kit). The consensus sequences were selected from within the DNA polymerase gene and were 76 to 86% identical to the genomic sequences of the six herpesviruses. Each stair primer used comprised an equimolar mixture of 11 oligonucleotides corresponding to a consensus sequence: all primers had the same 5′ end but extended for 20 to 30 nucleotides in the 3′ direction. The amplified fragment corresponded to positions 2163 to 2481 of the CMV DNA polymerase gene (GenBank accession no. X17403). The primers were used at a final concentration of 2 μM per reaction mixture, with each oligonucleotide contributing 1/11 to the final molarity.

The reaction was carried out in a volume of 50 μl. The PCR mixture contained 15 μl of DNA preparation, 60 mM Tris HCl (pH 9), 17 mM (NH₄)₂SO₄, 0.017% bovine serum albumin, 2 mM MgCl₂, and 200 μM (each) deoxynucleotide triphosphate. The reaction mixture was overlaid with 100 μl of paraffin oil and was heated to 94°C for 4 min. The temperature was maintained at 70°C while 2 U of Taq polymerase (Pharmacia, Orsay, France) was added (hot start). The reaction mixture was placed in a thermal cycler (Omnigene-Hybaid Thermocycler supplied by Life Science International, Cergy Pontoise, France). The sample was then subjected to 5 cycles of 30 s at 94°C for denaturation, 50 s at 56°C for annealing (2 s of ramping/°C), and 50 s for elongation at 72°C; 15 cycles of 30 s of denaturation at 94°C, 50 s of annealing at 46°C (2 s of ramping/°C), and 50 s of elongation at 72°C; and finally, 20 cycles of 30 s of denaturation at 94°C, 50 s of annealing at 54°C (2 s of ramping/°C), and 50 s of elongation at 72°C (2 s of ramping/°C).

The amplified products were analyzed by hybridization in microtiter plates according to the manufacturer’s recommendations. Each PCR product was aliquoted and distributed into six wells. In each well, the amplicons were hybridized with one of the six biotinylated oligonucleotide probes. Each probe was 100% specific for the virus concerned (HSV-1, HSV-2, CMV, EBV, VZV, and HHV-6). Briefly, 15 μl of the amplification product for each specimen was incubated in a microcentrifuge tube in 60 μl of denaturing solution for 5 min. Fixative solution (600 μl) was added, and the mixture was transferred to 96-well microtiter plates (100 μl in each of six wells per sample). The plates were incubated at 37°C for 2 h or at 4°C overnight to allow chemical fixation of the amplicons to the plastic. Probe (100 μl, one probe per well) was added, and the plates were incubated at 37°C for 30 min to allow hybridization. The plates were washed, and streptavidin-conjugated peroxidase and a chromogenic substrate were used to detect hybridization. The optical density at 492 nm was read in a spectrophotometer (Dinex Technology, Issy les Moulineaux, France). In every series, we included one negative control for extraction and the six positive controls obtained from reference strains (described above in Sample Preparation). A negative control consisting of amplified products from the human proto-oncogene ETS₂ was supplied with the kit and was used to check the specificity of detection. As indicated by the manufacturer, the threshold value for a positive result was set at a value 10% higher than the mean of six values for the negative controls (described in the paragraph on sample preparation above).

Amplification by specific primers.

The samples sent to the laboratory had been tested for one or several herpesviruses by targeted PCR, as requested by the clinician. The primers used were those described by Griffais et al. (5). The sequences of these primers correspond to those of different genes in the six viral genomes. DNA was extracted as described above, and the DNA pellets were suspended in 30 to 75 μl of water, according to the number of specific viral infections requested to be diagnosed for a patient’s sample. The samples were amplified in a 50-μl reaction mixture containing 15 μl of the DNA preparation, 10 mM Tris HCl (pH 8.4), 50 mM KCl, 0.01% gelatin, 2 mM MgCl₂, 200 μM (each) deoxynucleotide triphosphate (Pharmacia, Orsay, France), and 2 U of Taq polymerase (Pharmacia). This reaction mixture was overlaid with 100 μl of paraffin oil. The amplification protocol (Thermocycler, as described above) consisted of 3 min of denaturation at 92°C and then 35 cycles of 30 s of denaturation at 92°C, 1 min and 15 s of annealing at 55°C, and 1 min and 15 s of elongation at 72°C (increments of 1 s/cycle). The amplification products were analyzed by Southern blotting and hybridization with the specific probes described by Griffais et al. (6) and were obtained by amplification with internal primers labeled by incorporation of digdUTP (Boehringer Mannheim, Meylan, France).

DNA was prepared and amplified, and the amplification products were analyzed in three separate, nonadjacent laboratories.

Statistical methods.

Fisher’s exact test and the χ² test with Yates’ correction were used, where appropriate, for comparison of the results obtained by the two PCR methods.

RESULTS

In 1995, our laboratory received 164 CSF samples for herpesvirus screening. One hundred forty-two samples were available for retesting by the herpesvirus consensus PCR. Four hundred one PCRs had been performed with these 142 samples to test for HSV-1, HSV-2, CMV, and VZV; tests for EBV and HHV-6 were not prescribed by the clinicians for any sample. Eighteen samples (12.7%) tested positive by the targeted PCR tests and 37 (26%) tested positive by the consensus PCR tests, including three samples that had coinfections (CMV, VZV, and HSV-2, VZV and HSV-2, and CMV and HHV-6), such that the number of individual virus detections was 41. The numbers of positive amplifications obtained by the two approaches were significantly different (P < 0.006). Of the 142 CSF samples, 103 were classified as negative by both the targeted and the consensus PCRs.

We present the results in more detail in Tables 1 and 2. The consensus PCR test detected a virus not tested for by the targeted PCR in 16 samples: 13 samples classified as negative and 3 coinfected samples (1 classified as false negative and 2 classified as positive). The targeted PCR detected HSV-1 in 3 samples (3 of 125; 2.4%), HSV-2 in 1 sample (1 of 120; 0.8%), CMV in 9 samples (9 of 90; 10%), and VZV in 5 samples (5 of 66; 7.5%). The consensus PCR detected HSV-1 in 4 of the 142 samples (2.8%) HSV-2 in 5 samples (3.5%), CMV in 11 samples (7.7%), EBV in 8 samples (5.6%), VZV in 11 samples (7.7%), and HHV-6 in 2 samples (1.4%). The four viruses for which tests were prescribed were detected in 18 samples by the targeted PCR and in 24 samples by the consensus PCR, giving sensitivities of 69 and 92%, respectively. Two samples identified as positive by the targeted PCR were classified as negative by the consensus PCR, and eight samples identified as negative by the targeted PCR were positive by the consensus PCR (Table 2, group B).

TABLE 1.

Overall comparison of the results of targeted PCR and herpesvirus consensus PCR

Targeted PCR result	No. of samples with the indicated herpesvirus consensus PCR result for the following virus:
	HSV-1		HSV-2		CMV		EBV		VZV		HHV-6
	Positive	Negative	Positive	Negative	Positive	Negative	Positive	Negative	Positive	Negative	Positive	Negative
Positive	2	1	1	0	9	0	0	0	4	1	0	0
Negative	2	120	3	116	1	80	0	0	2	59	0	0
Not determined	0	17	1	21	1	51	8	134	5	71	2	140

TABLE 2.

CSF samples testing positive for herpesvirus by targeted PCR or by herpesvirus consensus PCR

Groupa	CSF sample no.b	Result of targeted PCRc						Virus detected by consensus PCRc
		HSV-1	HSV-2	CMV	EBV	VZV	HHV-6
A	1	N	N			P		VZV
	29	N	N	P		N		CMV
	50	P	N					HSV-1
	51	N	N	P				CMV
	63	N				P		VZV + HSV-2
	72	P	N					HSV-1
	79	N	N	P		N		CMV
	85			P				CMV
	89	N	N	P		N		CMV + HHV-6
	93			P				CMV
	111	N	N	N		P		VZV
	115	N	N			P		VZV
	116			P				CMV
	119	N	P	N		N		HSV-2
	126			P				CMV
	127			P				CMV
B	4	N	Nd					HSV-2
	13	N	N	N		Nd		VZV
	17	N	Nd					CMV + VZV + HSV-2
	36	Nd	N	N		N		HSV-1
	42	Nd	N					HSV-1
	61	N	N			P		Nd
	94	N	N	N		Nd		VZV
	104	N	Nd	N		N		HSV-2
	106	N	N	Nd		N		CMV
	128	P	N					Nd
C	2	N	N	N		N		EBV
	12	N	N					VZV
	14	N	N	N		N		EBV
	19	N	N					VZV
	20	N	N					VZV
	27	N	N	N		N		EBV
	32	N	N	N		N		EBV
	39			N				HHV-6
	54	N	N	N		N		EBV
	90	N						EBV
	107			N				EBV
	122	N	N					EBV
	131	N	N	N				VZV

^aA, results consistent for both methods; B, different results for the two methods; C, positive results by consensus PCR for viruses not tested for by targeted PCR. 

^bSamples received in chronological order in 1995. 

^cN, PCR negative; P, positive by targeted PCR. 

^dClassified as a false-negative result. 

The specificity of amplification was verified for all samples: the positive controls (with reference strains) gave no hybridization signal with any probe other than that expected. The negative controls gave no hybridization signal with any of the six probes. This specificity was strongly suggested by the two EBV-positive samples (samples 2 and 14), both of which were shown to be from the same patient. Testing for EBV was not initially prescribed, and the two samples were handled in independent series. The same applied to the two samples (samples 12 and 20) positive for VZV.

DISCUSSION

The storage of CSF samples at −80°C did not appear to impair their suitability for the consensus PCR. The discordant results in this study involving a retrospective comparison could not be checked by retesting because many of the samples were completely used up. Nevertheless, it is clear that the consensus PCR was specific. The sensitivity of the consensus PCR was higher than that of the standard, targeted PCR, although the difference was not significant in this small series. The precautions taken (the use of testing in three geographically different locations) and the checking of negative and positive controls for each series made it possible to rule out false-positive results due to contamination of the samples. There are several possible explanations for this higher sensitivity: more amplification cycles were used for the consensus PCR than for the targeted PCR, more DNA was used in the PCR because a single reaction was performed, so the subdivision of the sample into several aliquots was not necessary, and the enzyme-linked immunosorbent assay may be a more sensitive detection method than Southern blotting with a nonradioactive probe (15). Industrially produced optimized products appear to perform better than our homemade reagents. The problems of standardization have been highlighted by the European Union Concerted Action on Virus Meningitis and Encephalitis (8) and must be resolved if PCR is to become the reference method for virological diagnosis of neurological herpesvirus infections.

The PCR technique described by Rozenberg and Lebon (11) does not detect VZV or HHV-6. The approach described here is much simpler and quicker than analysis involving gel electrophoresis and determination of the sizes of restriction fragments. Multiplex PCR may be less sensitive and, as described by Tenorio et al. (13), requires a second, nested amplification for virus identification. The diagnostic assay developed by Aono et al. (1) is useful for the detection of only three herpesviruses.

We detected viruses for which tests were not requested by the clinician in several cases, showing that the same clinical syndrome may be produced by different herpesviruses and that specific detection of these viruses in CSF on the basis of clinical signs is unsatisfactory. Our findings of mixed viral infection are consistent with those of Tang et al. (12), who reported coinfections with two herpesviruses in the central nervous system. In one sample, we detected three different herpesviruses. The proportion of samples testing positive for EBV by the consensus PCR was consistent with recent reports (2).

The herpesvirus consensus PCR described here appears to be reliable for use with clinical samples. It is the first system to detect six different herpesviruses in a single amplification by a technique suitable for routine use. There are clear advantages to a method that involves only one reaction: less sample material, reagents, and time are required. The results of this preliminary study should prompt a more exhaustive analysis of the clinical value of simultaneous detection. A sensitivity study that includes a rigorous parallel comparison of these PolyHer stair primers with the standard tests for individual viruses is also necessary.