Whole blood samples from 336 pregnant women, who were registered at Yantai Yuhuangding Hospital of Qingdao University (Yantai, China), were used in the present study. Amniotic fluid samples (11) were obtained from pregnant women exhibiting CMV-positivity, which was identified using reverse transcription qPCR (RT-qPCR) and LAMP assays. Informed consent was obtained from the patients, and the study was permitted by the Human Research Ethics Committee of Yantai Yuhuangding Hospital of Qingdao University. Samples were initially detected using RT-qPCR and then evaluated using a LAMP assay. Total DNA from whole blood samples or amniotic fluid samples was extracted using a QIAamp DNA Mini kit (Qiagen GmbH, Hilden, Germany), according to the manufacturer’s instructions. Total extracted DNA was quantified by measurements at 260 nm optical density (OD) using a spectrophotometer (NanoDrop 2000; Thermo Fisher Scientific, Wilmington, DE, USA). Extracted DNA samples were stored at −20°C prior to use. Viral DNA isolation was performed from stock viruses of herpes simplex virus type 1 (HSV-1), HSV-2, varicella zoster virus (VZV), HSV-6, HSV-7 and CMV (Sinobio, Beijing, China) using the QIAamp DNA Mini kit.

The primers for LAMP amplification of the CMV glycoprotein B (gB) gene were designed based on CMV sequence data obtained from Genbank (accession number: M60931). Oligonucleotide primers that were used in the present study were designed using Primer Explorer V4 software (Eiken Chemical Co. Ltd., Tokyo, Japan). Designed primer sequences were subjected to BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) in order to exclude the possibility of cross-reactivity with HSV-1, HSV-2, VZV, HSV-6 and HSV-7. CMV specific primers consisted of two outer (F3 and B3) and two inner primers: Forward inner primer (FIP) and backward inner primer (BIP). Inner primers that recognized both forward and reverse strands of the target DNA were connected by a ‘TTTT’ linker. And additional loop primers [forward loop primer (LF) and backward loop primer (LB)] were used to promote both the amplification efficiency and acceleration of the reaction. Details of the sequence and location of each nucleotide primer in the target DNA sequences are provided in Fig. 1.

In order to determine the sensitivity of the CMV LAMP method, part of the gB gene containing the target DNA sequence was amplified using the following primers: Forward TGCCCGACGTCACGGTGGTC and reverse: ACCGACTTCAGGGTACTGG, which was cloned into pGEM-T-Easy plasmid (Promega Corporation, Madison, WI, USA). Optimization of LAMP conditions for CMV and sensitivity determination was determined by amplifying 10⁰–10⁷ copies of CMV gB-containing plasmids. The specificity of the LAMP assay was determined using HSV-1, HSV-2, VZV, HSV-6 and HSV-7 DNA samples as negative controls.

Amplifications were optimized using different conditions: Using 20, 25, 30 or 35 µl reaction volumes, including 2, 5 or 10 µl DNA template, 1 or 2 µM inner primers (FIP and BIP), 0.1, 0.3 or 0.5 µM outer primers (F3 and B3) and 0.5 or 1 µM loop primers (LF and LB), 0.5,1 or 2 µl Bst DNA polymerase (Large Fragment; New England Biolabs, Inc., Ipswich, MA, USA), 2 × reaction mix (0.5 of the total volume), and supplemented distilled and deionized water (ddH₂O). Reaction temperatures were screened at 59, 62 or 65°C and at the following reaction times: 5, 10, 15, 20, 25, 30 and 35 min. A LAMP turbidimeter TERAMECS (LA200; Teramecs, Co. Ltd., Kyoto, Japan) was used to incubate the mixtures and to measure the turbidity following the LAMP reaction. The turbidity cut-off value was set at >0.1 mean ± 3 standard deviation of the turbidity, from the turbidity values of three negative samples. The LAMP products were also subjected to 1.5% agarose gel electrophoresis in order to validate the experiments. Gels were visualized under an ultraviolet light following ethidium bromide staining.

Primers for the CMV RT-qPCR assay were designed according to previously reported sequences (23) and were synthesized by Sangon Biotech (Shanghai, China). Primers were dissolved in ddH₂O, to 100 µM and stored at −20°C. The RT-qPCR assay was performed using a One-Step PrimeScript RT-PCR kit (Takara Bio, Inc., Otsu, Japan), using LightCycle 2.0 (Roche Diagnostics GmbH, Mannheim, Germany).

In order to evaluate the LAMP assay in whole blood or amniotic fluid specimens from pregnant women, 336 whole blood samples were tested for CMV using RT-qPCR. Whole blood samples (336) and 11 amniotic fluid samples from RT-qPCR-confirmed CMV-positive pregnant women were then subjected to a LAMP assay using the optimized conditions (25 µl reaction volume, including 3 µL DNA template, 1 µM inner primers, 0.5 µM outer primers and 0.5 µM loop primers, 1 µl Bst DNA polymerase and ddH₂O. The reaction was performed at 62°C for 35 min). Sensitivity, specificity, positive predictive value and negative predictive value from the LAMP assays were then calculated using standard formulas and the results of RT-qPCR were used as standards.

In order to optimize the conditions for CMV LAMP detection, LAMP was conducted under different conditions, including different Mg²⁺ concentrations, different concentrations of loop primers, and different temperatures and durations. The results suggested that LAMP conditions were optimized at a 25-µl reaction volume. Final reaction mixtures consisted of 1.6 µM inner primers (FIP and BIP), 0.2 µM outer primers (F3 and B3), 0.8 µM loop primers (LF and LR), 10 mM MgSO₄, 1 µl Bst DNA polymerase and 5 µl DNA template. Amplifications were performed at 64°C for 30 min and reactions were terminated at 85°C for 5 min.

In the present study, the capability of the CMV LAMP assay for the discrimination of CMV from other members of herpesviridae, such as HSV-1, HSV-2, VZV, HSV-6 and HSV-7 was assessed. The results of the present study suggested that CMV exhibits bright turbidity following a LAMP assay, whereas other members of herpesviridae were not detected using LAMP (Fig. 2A). LAMP products were then analyzed using gel electrophoresis and the results suggested that CMV was successfully amplified, while other viral DNA was not amplified (Fig. 2B). Therefore, primer sets developed in the present study exhibit specificities for the target CMV sequences.

LAMP assay sensitivity was analyzed using the serially-diluted CMV gB-containing plasmid. Serial dilutions of recombinant pGEM-T Easy plasmid ranging from 10⁷–10⁰ copies per tube were used in order to determine the detection limits of CMV LAMP. Results demonstrated that the sensitivity of the CMV LAMP assay was 10 copies per tube, according to a real-time turbidimeter at 650 nm OD (Fig. 3A and B). Reactions were repeated three times. Serially-diluted plasmids were examined, ranging from 10⁶–10⁻¹ copies per tube, using RT-qPCR (Fig. 3C). The threshold of RT-qPCR was 10⁻¹ copies per tube, which was 10 times more sensitive than that of the CMV LAMP assay.

In order to further evaluate the performance of the CMV LAMP assay for CMV infection detection in pregnant women, CMV LAMP assays were conducted in 336 whole blood samples and 11 amniotic fluid samples from pregnant women. Samples were tested using RT-qPCR. Positive samples (10, 13, 20 or 21), with a threshold of 10 copies, were confirmed using RT-qPCR. RT-qPCR-positive samples, 10 and 13, were confirmed as positive according to the CMV LAMP, with a threshold of 10³ or 10⁴ copies per tube, respectively and 100% sensitivity. However, samples 2 and 3, which were negative according to RT-qPCR, were shown to be positive according to the CMV LAMP assay, with a specificity of 98.45 and 97.24%, PPV values of 86.67 and 76.92%, respectively; NPV for both thresholds were 100% (Table I). The following values were observed at 10¹ and 10² copies per tube compared with those at 10³ and 10⁴: Sensitivity decreased to 86.96 and 91.30%, specificity increased to 100 and 99.68%, PPV increased to 100 and 94.45%, and NPV decreased to 99.05 or 99.36%, at 10¹ and 10² copies per tube respectively. In order to reconfirm the sensitivity and specificity of the CMV LAMP assay, 11 amniotic fluid samples were examined from pregnant women with whole blood CMV-positive and 15 control amniotic fluid samples were examined, from pregnant women without active CMV infection. The results indicated that 11 samples were positive, with a threshold of 10², 10³ or 10⁴, while the 15 control samples were negative, with a threshold of 10¹, 10² or 10³ (Table II). Overall, the CMV LAMP method performed well in the detection of CMV infection.