Diagnostic Consequences of Cytomegalovirus Glycoprotein B Polymorphisms

Abstract

Failure of a cytomegalovirus (CMV) real-time PCR assay targeting glycoprotein B (gB) was investigated. A multiplex assay targeting gB and immediate-early 2 (IE2) genes showed discordant results (gB negative and IE positive or a >10-fold-higher viral load with IE primers) in saliva from 14.6% of CMV-infected newborns. Sequencing revealed 3 patterns of gB variations.

TEXT

Cytomegalovirus (CMV) is a leading cause of congenital infection and hearing loss in children and an important pathogen in immunocompromised patients. PCR-based methods are used widely in diagnostic laboratories to detect and monitor CMV infections. Since the glycoprotein B (gB) gene of CMV is essential for the infectivity of the virus and has a largely conserved nucleotide sequence, various regions of the gB gene are commonly used as PCR assay targets (3, 5, 7, 9, 18). A real-time PCR assay used in our laboratory targets the conserved region that lies upstream of antigenic domain 1 (3) of the gB gene, between base pairs 1541 and 1612.

As part of the National Institute on Deafness and Other Communication Disorders (NIDCD) CMV and Hearing Multicenter Screening (CHIMES) study, newborns at seven medical centers were screened for congenital CMV infection by a multiplex real-time PCR with concurrent use of primers and probes targeting conserved regions of the gB gene and exon 5 of the immediate-early 2 (IE2) gene (2). During the course of this study, discordant results in the performance of primers/probes targeting the gB and IE2 genes were observed. The failure of the PCR using gB primers/probes was investigated in this study.

Newborn CMV screening was carried out using rapid culture and/or real-time PCR of saliva specimens. Between March 2007 and July 2010, 386 infants tested positive for congenital CMV infection. Of these, 246 infants (63.7%) had real-time PCR of saliva specimens completed. The real-time PCR assay protocol was previously described (2). Briefly, 5 μl of saliva samples was used as a template without DNA extraction by utilizing the ABI TaqMan 7500 system (Life Technologies Corporation, Carlsbad, CA). Primers and probes include gB fw (AGGTCTTCAAGGAACTCAGCAAGA), gB re (CGGCAATCGGTTTGTTGTAAA), gB FAM/TAMRA probe (6FAM-AACCCGTCAGCCATTCTCTCGGC-TAMRA) (where 6FAM is 6-carboxyfluorescein and TAMRA is 6-carboxytetramethylrhodamine), based on the AD169 sequence, and IE2 fw (GAGCCCGACTTTACCATCCA), IE2 re (CAGCCGGCGGTATCGA), and VIC/MGBNFQ probe (VIC-ACCGCAACAAGATT-MGBNFQ), based on the consensus CMV sequence. Quantified plasmid DNA containing target sequences of both primer sets served as calibration standards.

Samples were considered to be discordant if the IE2 gene was PCR positive and the gB gene was PCR negative or if the IE2 gene copy number was at least >10-fold higher than the gB gene copy number. Discordant samples were subjected to PCR amplification of the gB region containing the target sequences using gB forward (CACAGGTTGGTGGCTTTTCT) and reverse (GTCGTGAGTAGCAGCGTCCT) primers. PCR products were sequenced and aligned with CMV sequences retrieved from NCBI GenBank (sequences for AD169, Towne, Toledo, HAN20, and S3 [accession numbers BK000394, FJ616285, GU180092, GQ396663, and GU937742, respectively]) by using BioEdit software (Ibis Therapeutics, Carlsbad, CA). Ten random CMV-positive newborn saliva samples with concordant results between gB and IE genes were also amplified and sequenced. Informed consent was obtained from study participants, and the study was conducted in accordance with the guidelines of the Institutional Review Board of the University of Alabama at Birmingham.

Among the 246 CMV-positive newborns, discordant results were observed for 36 (14.6%) infants. Of the 36 specimens with discordant PCR results, 26 were PCR negative using the gB primers, and the remaining 10 samples showed >10-fold-higher CMV copy numbers per reaction with IE2 primers than that observed with gB primers. The viral load levels in the 36 specimens are shown in Table 1. Amplification and nucleotide sequence analysis for the gB target region were completed for 28/36 discordant samples due to sample availability. This analysis showed 3 different patterns of polymorphisms within the target region for the gB primers/probe compared to the reference AD169 sequence. All of the nucleotide substitutions were synonymous mutations (Fig. 1). Of the 10 samples with concordant results, 2 contained a silent single-nucleotide substitution in the reverse primer target region (Fig. 1, pattern E). The remaining 8 samples matched the reference AD169 sequence (Fig. 1, pattern D).

Table 1.

Quantitative PCR results for the gB and IE2 target genes of discordant samples and corresponding mutation patterns from 28 sequenced samples

Identifier	No. of copies/reaction		Mutation patterna
	gB	IE2
1	0.0	25.0	B
2	0.0	31.0	A
3	0.0	31.5	ND
4	0.0	85.0	ND
5	0.0	96.0	ND
6	0.0	110.0	A
7	0.0	115.0	A
8	0.0	125.0	A
9	0.0	199.0	ND
10	0.0	350.0	A
11	0.0	430.0	A
12	0.0	560.0	ND
13	0.0	690.0	A
14	0.0	900.0	A
15	0.0	1,100.0	A
16	0.0	1,500.0	ND
17	0.0	2,000.0	A
18	0.0	2,880.0	A
19	0.0	11,700.0	A
20	0.0	15,000.0	A
21	0.0	19,000.0	A
22	0.0	28,000.0	A
23	0.0	64,000.0	A
24	0.0	98,000.0	A
25	0.0	150,000.0	ND
26	0.0	310,000.0	C
27	0.6	24,000.0	A
28	1.2	43,000.0	A
29	2.5	8,500.0	A
30	2.6	200.0	A
31	3.0	9,700.0	A
32	3.5	7,900.0	A
33	5.3	190.0	ND
34	10.0	125,000.0	A
35	20.0	65,000.0	A
36	43.5	53,000.0	A

^aND, sequencing not performed.

Fig. 1.

Alignment of known sequences and the mutation patterns found in study samples. A, B, and C represent sequencing result of discordant samples. D and E are sequences obtained from positive concordant samples.

In this study, we demonstrated polymorphisms within the region of the gB gene that were previously thought to be highly conserved. Discordant results, in which IE2 primers/probes detected CMV DNA but gB primers/probes failed, were observed with real-time PCR analysis of newborn saliva specimens in 14.6% of congenitally infected infants. The gB sequence variability of the CMV present in the saliva specimens most likely resulted in a mismatch between our primers and probes, leading to significant lowering of the sensitivity of the real-time PCR assay. Similar findings were also reported by Lengerova et al., who experienced about a 5% false-negative rate with an in-house assay due to mismatches in the target region of primers/probes located within the major immediate-early exon 4 region (12).

Studies investigating the diversity within genes of CMV (4, 6, 15, 16) have demonstrated extensive variability and that multiple variants can coexist in an individual (14, 17). Falsely lower copy number readings might be obtained in cases where multiple strains are present in a sample, because minor variants may be detected while the major virus populations might not be detected due to primer/probe mismatch. In addition, infection with multiple virus strains can also lead to the generation of new CMV variants by recombination, and the newly formed variants may not be detected (8). These findings, in addition to the results of our study, emphasize the importance of selecting primers and probes from highly conserved regions of CMV among different strains in order to avoid false-negative PCR results which could significantly impact clinical care. A significant number of CMV-infected newborns would have been missed if our real-time PCR assay included primers/probes targeting gB alone. Since new polymorphisms are being described on a regular basis, it is important to maintain constant vigilance so that the molecular diagnostic assays remain highly sensitive for the detection of CMV in clinical specimens (11, 13). Although it may be difficult to completely avoid false-negative PCR results due to the high diversity of CMV strains, it may be possible to overcome imperfections in the primer/probe design by the use of degenerate primers for both qualitative and quantitative PCRs (1, 10). Alternatively, the use of multiplex assays targeting two or more independent target regions, the strategy used in our study, could be another approach to reduce the chances for false-negative results.