Multiplex Detection of Human Herpesviruses from Archival Specimens by Using Matrix-Assisted Laser Desorption Ionization-Time of Flight Mass Spectrometry

Abstract

The human herpesviruses are involved in a variety of diseases. Large-scale evaluation of the clinical and epidemiological importance of different herpesviruses requires high-throughput methods. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a method that has a multiplex capacity enabling simultaneous detection of several viruses in a single sample. PCR-based methods for the multiplex detection of all known human herpesviruses were developed on the MALDI-TOF MS system. A variety of 882 archival samples, including bronchoalveolar lavage, conjunctival fluid, sore secretion, blister material, plasma, serum, and urine, analyzed for herpesviruses using PCR-based reference methods, were used to evaluate the MALDI-TOF MS method. The overall concordance rate between the MALDI-TOF MS method and the reference methods was 95.6% (κ = 0.90). In summary, the MALDI-TOF MS method is well suited for large-scale detection of all known human herpesviruses in a wide variety of archival biological specimens.

The herpesviruses infecting humans include herpes simplex virus (HSV) types 1 and 2, varicella-zoster virus (VZV), Epstein-Barr virus (EBV) types A and B, cytomegalovirus (CMV), human herpesvirus 6 (HHV6) types A and B, HHV7, and HHV8 (Kaposi's sarcoma-associated herpesvirus). Herpesvirus infections in healthy subjects are often asymptomatic but can cause oral and genital lesions (HSV), chicken pox or shingles (VZV), infectious mononucleosis (EBV and CMV) (13), or roseola (HHV6) (19). In immunologically compromised hosts, herpes infections are more severe, causing pneumonia (CMV and HHV6) (1, 19), lymphoproliferative diseases (EBV) (19), Kaposi's sarcoma (HHV8) (19), or encephalitis (HSV-1) (6). Most herpesviruses have been reported to traverse the placenta (21), and intrauterine infections may cause birth defects (CMV) (19), premature delivery (VZV) (4), and fetal mortality (HSV) (21). Herpesvirus infections during pregnancy have also been associated with childhood acute lymphoblastic leukemia (EBV) (14).

Common laboratory techniques in herpesvirus detection include antibody detection such as enzyme-linked immunosorbent assay (26), PCR (15, 17, 18, 25, 27), and dot blot hybridization (27). These techniques usually require separate methods for the detection of each herpesvirus and clinical information for selection of virus assay. Since most microbiological laboratories maintain large biobanks with clinical samples, large-scale evaluation of the clinical and epidemiological importance of different herpesviruses is today restricted only by cost and throughput considerations. There is therefore an increasing interest in rapid and large-scale herpesvirus detection, and methods for the multiplex detection of several herpesviruses have recently been developed using PCR and microarray techniques (8, 10, 23). Virus detection using the matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) system is a high-throughput technology with multiplex capacity that has been used for genotyping and detection of RNA viruses (3, 7, 9, 11, 12, 16) and human papillomavirus (HPV) (28).

The aims of the present study were to develop an efficient screening method for qualitative multiplex detection of all HHVs by MALDI-TOF MS and to investigate the usefulness of a wide variety of archival sample types for HHV detection.

MATERIALS AND METHODS

Sample selection.

A consecutive series of 860 archival aliquots of samples diagnostically tested for HSV-1, HSV-2, VZV, CMV, or EBV (EBV Virus Capsid Antigen-IgM analysis) at the Department of Medical Microbiology, Malmö University Hospital, Malmö, Sweden, was retrieved. All archived aliquots had been stored at −80°C or −20°C (the EBV samples) for at least 18 months. For each positive sample, the negative sample, with the same type of biological material, submitted closest in calendar time was retrieved. Sample materials included bronchoalveolar lavage, conjunctival fluid, sore secretion, blister material, plasma, serum, and urine (Table 1). Thirty-two of the selected samples had been analyzed for three herpesviruses (HSV-1, HSV-2, and CMV [n = 3] or HSV-1, HSV-2, and VZV [n = 29]), 403 samples had been analyzed for two herpesviruses (HSV-1 and HSV-2), and 425 samples had been analyzed for one herpesvirus by clinical diagnostic testing. An additional 22 archived aliquots of serum samples that had tested positive for HHV8-specific antibodies, collected between 1996 and 2005 and stored at −80°C, were obtained from the Department of Microbiology, Oncological Center, Aviano, Italy. Thus, a total of 1,349 herpesvirus analyses had been performed on a total of 882 samples (Table 1).

TABLE 1.

Samples used in the study and results of reference methods

Virus analysis	No. of analyses (no. positive)
	BALa	Blister material	Conjunctival fluid	Sore secretion	Plasma	Serum	Urine	Total
HSV-1	28 (14)	370 (80)	18 (8)	19 (6)				435 (108)
HSV-2	28 (0)	370 (106)	18 (1)	19 (3)				435 (110)
CMV	10 (5)				97 (48)	2 (1)	20 (10)	129 (64)
VZV		156 (78)	2 (1)	6 (3)				164 (82)
EBV						164 (70)		164 (70)
HHV8						22 (12)		22 (12)
Total	66 (19)	896 (264)	38 (10)	44 (12)	97 (48)	188 (83)	20 (10)	1,349 (446)

^aBAL, bronchoalveolar lavage.

This study was approved by the Institutional Review Board of Lund University.

Reference method.

DNA extracted from an aliquot of each sample, upon arrival at the Department of Medical Microbiology, was analyzed for the selected herpesvirus, and the remainder of the sample was stored as described above.

(i) PCR analysis of HSV-1, HSV-2, and VZV.

DNA was extracted from a 200-μl sample by using a MagNA Pure LC instrument and a total nucleic acid isolation kit (Roche Diagnostics, Penzberg, Germany) and eluted in 100 μl of buffer, except for suspensions of VZV blister material, for which DNA was released by heating the sample to 95°C for 10 min. The PCR methods were adapted from those of Aurelius et al. (5) and Puchhammer-Stockl et al. (20). Briefly, 10 μl of MagNA extract, corresponding to 10 μl (VZV blister material) or 20 μl of the original sample, was mixed with 1× Taq buffer, 2 mM MgCl₂, 0.2 mM deoxynucleoside triphosphate (dNTP), 1 U of Taq polymerase, and 0.4 μM concentrations of each HSV outer primer (HSV1 [see reference 5], HSV2 forward [5′-TCA GCC CAT CCT CCT TCG GCA GTA-3′], and HSV2 reverse [5′-GAT CTG GTA CTC GAA TGT CTC CG-3′]) or 28 nM concentrations of each VZV primer (forward [5′-TCC ACG TAT GAC TCT CTC AC-3′] and reverse [5′-GAT CAG ACA CAT GAC GAA TC-3′]) in a 50-μl reaction, and a layer of mineral oil was added. The PCR program consisted of 93°C for 5 min; followed by 20 cycles of 93°C for 50 s, 55°C for 50 s, and 72°C for 60 s; followed by 72°C for 5 min. A total of 5 μl of the primary PCR product was mixed with 1× Taq buffer, 2 mM MgCl₂, 0.2 mM dNTP, and 0.4 μM concentrations of each HSV inner primer (HSV1 [5], HSV2 forward [5′-AGA CGT GCG GGT CGT ACA CG 3′], and HSV2 reverse [5′-CGC GCG GTC CCA GAT CGG CA-3′]) or 115 nM concentrations of each VZV primer (20) and 1 U of Taq polymerase, in a 50-μl reaction, and a layer of mineral oil was added. The PCR program was as described above but extended by 10 cycles. The products were analyzed by electrophoresis on a 2% agarose gel. All PCR reagents were from Applied Biosystems (Foster City, CA), and the primers were from DNA-Technology A/S (Aarhus C, Denmark).

(ii) CMV analysis.

DNA from a 200-μl sample was extracted and analyzed by using the PCR- and hybridization-based COBAS Amplicor CMV monitor test and a COBAS Amplicor analyzer (Roche Diagnostics, Stockholm, Sweden). The template volume used in the PCR corresponds to 25 μl of the original sample.

(iii) Qualitative EBV PCR.

Qualitative EBV PCR was performed on all samples serologically tested for EBV for clinical diagnosis after DNA extraction of the archived aliquots as described below. Each 12-μl reaction mixture contained 4 μl of MagNA extract, corresponding to 16 μl of original sample, 5 μl of TaqMan PCR Master Mix (Applied Biosystems), a 1 mM concentration of additional MgCl₂ (Applied Biosystems), 900 nM concentrations of each primer (forward [5′-AAG GTC AAA GAA CAA GGC CAA G-3′] and reverse [5′-GCA TCG GAG TCG GTG GG-3′]; both from CyberGene AB, Huddinge, Sweden). The PCR program included 95°C for 10 min, followed by 50 cycles of 64°C for 1 min and 95°C for 15 s. The PCR products were visualized on a 2% agarose gel as previously described (22) and compared to the size of a molecular weight marker (GeneRuler 50-bp DNA ladder; MBI Fermentas) and EBV-positive controls. The detection limit of the assay was five copies of EBV plasmid control (prepared as described below).

(iv) Quantitative HHV8 analysis.

Quantitative HHV8 analysis was performed on all samples serologically tested for HHV8, after DNA extraction as described below, by TaqMan real-time PCR as described by Tedeschi et al. (24), using 2 μl of MagNA extract, corresponding to 4 μl of the original sample, in a 20-μl reaction supplemented with 0.1% bovine serum albumin (A2153-50G; Sigma-Aldrich, Stockholm, Sweden) and 3.5 mM MgCl₂ (Applied Biosystems) on a 7900 HT sequence detection system (Applied Biosystems). Samples with copy numbers below the lowest standard point detected were considered negative.

(v) HHV6 and HHV7.

No reference methods were available for analysis of HHV6 and HHV7.

DNA extraction of archived sample aliquots.

DNA from 200 μl of each sample (100 μl of HHV8 samples) was extracted by using a MagNA Pure LC instrument and a total nucleic acid isolation kit (MagNA) and eluted in 50 μl of buffer. Three controls composed of 200 μl each of NaCl, pooled HHV negative plasma (determined by the MALDI-TOF MS method), and pooled HHV negative plasma spiked with 25,000 copies of HPV type 51 (HPV-51) control/ml were included in each batch of 32 samples. In the batch with HHV8 samples, 100 μl of each control was used.

Virus controls.

Viral particles were obtained from the saliva from volunteers (HHV7) and the Swedish Institute for Infectious Disease Control in Solna, Sweden (HHV6A GS strain and HHV6B Z29 strain), and DNA was released by incubation with 2.5 mg of proteinase K (AM2542; Applied Biosystems)/ml at 56°C overnight. Viral DNA and HPV-51 plasmid control were obtained from the Department of Medical Microbiology, Malmö University Hospital, Malmö, Sweden (HSV-1, HSV-2, VZV, CMV, and EBV A), and the Department of Microbiology, Oncological Center, Aviano, Italy (HHV8). Primary PCR amplicons, using appropriate primers (Table 2), were cloned into TOPO TA cloning vectors (Invitrogen, Carlsbad, CA). Specific sequences were confirmed to correspond to those reported in the National Center for Biotechnology Information (NCBI) database by DNA sequencing of the plasmids. No positive control was available for EBV B.

TABLE 2.

Primers used in the MALDI-TOF MS analyses

Primer (NCBI reference sequence accession no., target gene)	Sequence (5′-3′)a	Primer 5′-position	UEPc size (Da)	Extension base(s)	EPd size (Da)
HSV1 (NC_001806, UL5)	F: CGTGGAGCTTGGCCATGTAC	13904
	R: AACCTCATGAAGGTGCTGGAGTAC	14075
	ME: GGAGCACATGCAGTT	14034	4,617	C	4,906
HSV2 (NC_001798, UL5)	F: CGTGGAGCTTGGCCATGTAC	14025
	R: AACCTCATGAAGGTGCTGGAGTAC	14196
	ME: GGAGCACATGCAGTT	14155	4,617	T	4,921
CMVb (X92746, UL122)	F: GCCAAGCGGCCTCTGATAAC	484
	R: TCATCCACACTAGGAGAGCAGACT	620
	ME: ATGCAGATCTCCTCAA	554	4,825	T	5,129
VZV (AY548171, ORF1)	F: CAACGTCGCCATCTGGAGTAC	652
	R: GCCACTCGTCGTTCAAACAAG	805
	ME: AGGTTATGCAACGTACT	696	5,209	GGGT	6,501
EBV A (K03333, EBNA2)	F: ACCTGATTCCCCCTGCTCAT	716
	R: CATTGGAAGGTGGCTGGTGAT	868
	ME: GTAGTGGCTTGGTTGT	761	4,974	A	5,287
EBV B (K03332, EBNA2)	F: ACCTGATTCCCCCTGCTCAT	713
	R: CATTGGAAGGTGGCTGGTGAT	865
	ME: GTAGTGGCTTGGTTGT	758	4,974	GA	5,616
HHV8 (U93872, ORF06)	F: AAAAAGAATAATGTCAGCTCTATGCTAAGA	6074
	R: TTAACCGCTTCTGCCTCGAA	6195
	ME: GATTATGACCCCCCT	6115	4,488	GGT	5,451
HHV6A (X83413, U66)	F: ATGAACCACGTTTTTCCATGTCTTCT	101834
	R: GAACCATTTCAGAAGAACATTCTCATAC	102033
	ME: CTAACGTATTCCTAGTCCC	101879	5,699	A	6,012
HHV6B (AF157706, U66)	F: ATGAACCACGTTTTTCCATGTCTTCT	103132
	R: GAACCATTTCAGAAGAACATTCTCATAC	103331
	ME: CTAACGTATTCCTAGTCCC	103177	5,699	GA	6,341
HHV7 (AF037218, U60)	F: ATGAACCATGTTTTACCGTGTCTTCT	99572
	R: GAGCCATTTCAAAAGAGCATTCTTATAC	99771
	ME: AGCTTTCTGTTTAAGTATG	99616	5,824	T	6,128

^aThe 5′ extension ACGTTGGATG was added to all forward and reverse primers. F, forward; R, reverse; ME, MassExtend.

^bThe forward and reverse primers were adapted from those of Mengoli et al. (18).

^cUEP, unextended primer.

^dEP, extended primer.

MALDI-TOF MS analyses.

All procedures were performed according to SEQUENOM standard protocols unless otherwise specified. Two qualitative, PCR-based, multiplex assays were designed to detect the HHVs by MALDI-TOF MS (SEQUENOM; MassARRAY, San Diego, CA); one assay was used to detect HSV-1, HSV-2, CMV, EBV A, EBV B, VZV, and HHV8, the other was used to detect HHV6A, HHV6B, and HHV7. The latter assay does not discriminate between HHV6A and HHV6B for the HST strain. Primer sequences, 5′-primer positions, target genes, NCBI reference sequence numbers, the masses of unextended primers, and the masses and extension base of extended primers are presented in Table 2. The 10-base 5′ extension ACGTTGGATC, recommended by SEQUENOM to increase the robustness in multiplex reactions, was added to each forward and reverse primer. After optimization of the annealing temperatures and the enzyme, MgCl₂, and primer concentrations, the SEQUENOM hME protocol was performed with the following adjustments: 2 μl of MagNA extract, corresponding to 4 (HHV8) or 8 μl of original sample, was used in a 6-μl primary PCR containing 1× PCR buffer, 200 μM dNTP, 0.15 U of TaqGold, 3.5 mM MgCl₂ (Applied Biosystems), and 0.5 μM concentrations of each primer (Metabion, Martinsried, Germany). The PCR program included 95°C for 10 min; 5 cycles of 64°C for 30 s, 72°C for 60 s, and 95°C for 30 s; 40 cycles of 72°C for 60 s and 95°C for 30 s; followed by 72°C extension for 10 min. The PCR product was treated with shrimp alkaline phosphatase (SEQUENOM), and 2 μl of secondary PCR mix (optimized for annealing temperature and primer concentration) containing 0.229 μl of ATC Terminator mix, 0.04 μl of MassExtend enzyme (SEQUENOM), and 1 μM concentrations of each MassExtend primer (Metabion) was added; a secondary PCR was then performed at 94°C for 2 min, followed by 99 cycles of 94°C for 5 s, 42°C for 5 s, and 72°C for 6 s. After desalting the mixture by the addition of 6 mg of clean resin to each sample, 15 nl of product from each sample was dispensed onto a 384-spot SpectroCHIP by using a MassARRAY Nanodispenser (all from SEQUENOM). The MALDI-TOF MS analysis was performed in a Bruker Autoflex (SEQUENOM). Each sample was exposed to nine individual laser pulses, the laser power was set to 35%, the linear detection voltage was 1,588 V, and the time delay for ion extraction was 300 ns. The spectra were acquired by using SpectroACQUIRE. The MassARRAY Typer (SEQUENOM), used for interpretation of the results, is programmed primarily for single nucleotide polymorphism genotyping. In contrast to single nucleotide polymorphism genotyping, in which abnormal peak height ratios generate “moderate” or “aggressive” calls, the clear identification of multiple peaks of appropriate mass, regardless of their height ratios, is interpreted as the presence of all corresponding viruses. Thus, “conservative,” “moderate,” and “aggressive” calls were interpreted as positive. “Low probability” and “no alleles” results were judged negative.

TaqMan real-time PCR.

All real-time PCR analyses were performed on a 7900 HT sequence detection system. HPV-51 (forward primer [5′-GCG CAC TAA TGA CAG CAA GGT-3′], reverse primer [5′ CGG TGC GTG TGA TAT ATT CTT CTG 3′], and probe [5′-FAM-TGC ACC TGT GTC TCG A-MGB-3′]) and CMV (primers and probe as described by Watzinger et al. [25]) analyses were performed in 6-μl reactions using TaqMan Assay-by-Design reagents (Applied Biosystems) and 2 μl of MagNA extract, supplemented with 0.1% bovine serum albumin and 3.5 mM MgCl₂. All MagNA extracts were from the same DNA extraction batch that was used in the MALDI-TOF MS analyses.

Statistical methods.

The kappa (κ) value (2) was calculated for the concordance between reference and MALDI-TOF MS methods (Table 3).

TABLE 3.

Comparison of the MALDI-TOF MS results with reference PCR-based methods

Virus analysis	Total	No. of samplesa				Concordance (%)	κ
		−Ref/−MS	+Ref/+MS	−Ref/+MS	+Ref/−MS
HSV-1	435	325	98	2	10	97.2	0.92
HSV-2	435	322	104	3	6	97.9	0.94
CMV	129	63	54	2	10	90.7	0.81
VZV	164	79	74	3	8	93.3	0.87
EBV	164	94	58	0	12	92.7	0.85
HHV8	22	10	9	0	3	86.4	0.73
Total	1,349	893	397	10	49	95.6	0.90

^aRef, reference results; MS, MALDI-TOF MS results.

RESULTS

DNA extraction controls.

All control, NaCl, HHV-negative plasma, and HHV-negative plasma samples spiked with HPV-51 were analyzed by HPV-51 real-time PCR. The mean viral yield of the plasma samples spiked with HPV-51 was 24.3% (standard deviation = 16.8, n = 32). All other controls were negative for HPV-51.

Detection limit of the MALDI-TOF MS method.

To determine the detection limit of the MALDI-TOF MS method, a 2-μl aliquot from a dilution series of each HHV control was used as a template in the MALDI-TOF MS method. Five copies of EBV, HHV6A, and HHV6B control DNA were always detected (100% sensitivity). The sensitivity for VZV and HHV7 was 100% at 10 copies, and the sensitivity for HSV-1, HSV-2, CMV, and HHV8 was 100% at 100 copies. To determine the detection limit for multiple infections, all nine HHV controls were combined in a single mixture. The sensitivity for detecting all controls in a combined mixture containing 100 copies of each control was 78%. A typical example of the mass spectra resulting from a sample, in which 100 copies each of all controls were combined and successfully detected, is presented in Fig. 1. The specificities of the MALDI-TOF MS method, calculated using 100 copies of each HHV control as a template, were 100% for HSV-2, VZV, HHV8, HHV6B, and HHV7; 98% for HSV-1 and CMV; 95% for HHV6A; and 92% for EBV. Analyses of dilution series were repeated a minimum of seven times.

FIG. 1.

MALDI-TOF MS spectra from a single sample containing 100 copies each of all nine HHV controls. UEP, unextended primer.

The detection limit of the entire analysis, including the DNA extraction step, was evaluated by the addition of serially diluted viral controls to HHV-negative plasma prior to DNA extraction. At 2,500 copies/ml of plasma, all HHV controls could be detected. If we assume a 100% yield in the extraction process, 2,500 copies/ml corresponds to 500 copies in the initial serum sample and 20 copies in the primary PCR template.

MALDI-TOF MS analyses.

A variety of 882 archival samples that had been analyzed for herpesviruses by reference PCR methods, 1,349 analyses in all, were used to evaluate the MALDI-TOF MS method. The concordance rate between MALDI-TOF MS analyses, obtained using template volumes corresponding to 4 μl (HHV8) or 8 μl of the original sample, and reference methods using template volumes corresponding to 4 μl (HHV8), 10 μl (VZV blister material), 16 μl (EBV), 20 μl (HSV and VZV), or 25 μl (CMV) of the original sample was 95.6% (κ = 0.90). A comparison of reference methods and MALDI-TOF MS results for each herpesvirus is presented in Table 3. Some analyses (n = 338) were repeated up to four times. The results for 53 (15.7%) of the analyses fluctuated between runs; 39 (73.6%) of these were positive in the reference analysis. An analysis was considered positive if the same virus was identified in all or three of four analyses. Six samples that were negative for HSV-1 and HSV-2 by the reference method were positive for VZV, and 25 samples with negative VZV reference test results were positive for HSV-1 or HSV-2 by the MALDI-TOF MS method. All of these samples consisted of blister material (n = 29) or sore secretions (n = 2). MALDI-TOF MS results in relation to the original clinical diagnostic request and multiple infections detected using the MALDI-TOF MS method are presented in Table 4. Since the method does not discriminate between HHV6A and the HHV6B HST strain and all HHV6-positive samples were positive for HHV6A, no discrimination between HHV6 types A and B is made in the table. All negative controls, including 2 to 24 nontemplate controls per analysis, were negative for all herpesviruses.

TABLE 4.

MALDI-TOF MS results in relation to the original clinical diagnostic request

Virus analysis requested (no. of samples)	No. of samples determined to be positive by MALDI-TOF MS analysis for:
	HSV-1	HSV-2	CMV	VZV	EBVA	EBVB	HHV8	HHV6	HHV7	Two HHVsa	Three HHVsa	Four HHVsa
HSV (435)	100	107	0	20b	11	1	0	5	21	16c	1d	1e
CMV (129)	0	0	56	0	4	1	0	2	3	6f
VZV (164)	14	16g	0	77	0	0	0	4	0	5h
EBV (164)	0	0	0	0	56	2	0	2	0	1i
HHV8j (22)	0	0	1	0	1	0	9	0	0	2k

^aEither two, three, or four herpesviruses, as indicated in the respective footnotes, were detected in the same samples; these values were also included in the column for respective virus.

^bFourteen samples were also submitted for VZV analysis.

^cThree EBVA and HHV7, one VZV and HHV6, two EBVA and HSV-1, one EBVA and HSV-2, one HSV-1 and HSV-2, two HHV6 and HSV-1, and six HHV7 and HSV-1.

^dHSV-1, HHV6, and HHV7.

^eEBVA, EBVB, HSV-1, and HHV7.

^fTwo EBVA and CMV, one EBVB and CMV, one HHV6 and CMV, one HSV-1 and CMV, and one HHV7 and CMV.

^gThree samples also submitted for HSV analysis.

^hThree HSV-1 and HHV6, one HHV6 and VZV, one HSV-1 and VZV, and one EBV and HHV7.

ⁱEBV and HHV6.

^jNot clinically diagnosed

^kOne HHV8 and CMV and one HHV8 and EBVA.

Real-time PCR.

The viral loads of CMV and HHV8 samples were analyzed by real-time PCR for the respective viruses. All samples with negative or inconsistent CMV and HHV8 results by the MALDI-TOF MS method had viral loads below the detection limit of the MALDI-TOF MS method. Both samples that had been negative for CMV in reference tests but positive using the MALDI-TOF MS method were also positive for CMV in the real-time PCR.

DISCUSSION

We have developed a multiplex MALDI-TOF MS method that successfully detects HHVs in a wide variety of archival biological specimens. The concordance rate between the MALDI-TOF MS method, using template volumes corresponding to 4 or 8 μl of the original sample, and reference methods, using template volumes corresponding to 4 to 25 μl of the original sample, was 95.6% (κ = 0.90). Consistently negative results for nontemplate controls and an overall 98% specificity for 100 copy plasmid controls validated this new method. By comparison, a concordance rate of 94% was found in a previous report comparing PCR and oligonucleotide microarray methods for multiplex herpesvirus detection (10).

Multiple infections, HHV6, and HHV7 were consistently demonstrated in repeated MALDI-TOF MS analyses in a substantial proportion of the study samples. Although no reference method was available for HHV6 and HHV7, DNA sequence validation of the plasmid controls, the unique combination of primer sequences according to NCBI BLAST analysis, and consistently positive MALDI-TOF MS results for the plasmid controls demonstrate the validity of the method. Since no EBV B control virus was available, the accuracy of discrimination between EBV A and EBV B must await further tests.

The detection limits of the MALDI-TOF MS method on all HHV controls were comparable to previous reports of multiplex herpesvirus detection using an oligonucleotide microarray and multiplex PCR techniques (8, 10). Since all negative control results were negative and detection limits could be documented with positive controls, we judge the fluctuations in results of some samples to reflect a low viral load near the detection limit of the method rather than contamination, as shown by real-time PCR quantitation for CMV and HHV8 samples. Low viral loads, or weak cross-reactivity in either the reference methods or the MALDI-TOF MS method, may also contribute to the few discrepant analyses. Routine analysis of duplicate samples and clear indications of sensitivity thresholds can be recommended for clinical use. Thus, we believe that the multiplex approach described here could be very useful for large-scale epidemiological research studies in which broad multiplex HHV detection is of interest.

In some cases, the requested diagnostic testing for VZV was negative, but the MALDI-TOF MS analysis detected HSV and vice versa. In other cases, unsuspected viruses were demonstrated alone or in combination with those whose detection had been evaluated. This observation indicates the difficulty in selecting the correct test based on clinical symptoms and suggests that a broad multiplex HHV analysis could improve both the economy and the diagnostic accuracy of clinical viral testing. MALDI-TOF MS analyses have been reported for the detection of hepatitis virus and HPVs (7, 9, 11, 12, 28). Our hospital, serving the southern region of Sweden, receives a sufficient number of samples per day to motivate 96 multiplex analyses for each of these three viral groups. Similar multiplex panels are being developed for human genetic disorders. Thus, several diagnostic questions could be addressed on separate 96-well systems using 96- or 384-format spectroCHIPs. All MALDI-TOF MS reactions are performed in the same reaction plate, with robotic pipetting, with the added quality advantages of traceability of samples and reduced human error throughout the entire process.

In summary, the MALDI-TOF MS methods for multiplex HHV detection will allow large-scale research studies on archival samples of various biological materials. The MALDI-TOF MS methods may also become highly useful for multiplex clinical diagnostic testing, following validation by parallel analyses of patient samples with established standard methods.