Performance of the Trioplex real-time RT-PCR assay for detection of Zika, dengue, and chikungunya viruses

Abstract

The emergence and spread of Zika virus (ZIKV) presented a challenge to the diagnosis of ZIKV infections in areas with transmission of dengue (DENV) and chikungunya (CHIKV) viruses. To facilitate detection of ZIKV infections, and differentiate these infections from DENV and CHIKV, we developed the Trioplex real-time RT-PCR assay (Trioplex assay). Here, we describe the optimization of multiplex and singleplex formats of the assay for a variety of chemistries and instruments to facilitate global standardization and implementation. We evaluated the analytical performance of all Trioplex modalities for detection of these three pathogens in serum and whole blood, and for ZIKV in urine. The limit of detection for the three viruses and in different RNA-extraction modalities is near 10³ genome copy equivalents per milliliter (GCE/mL). Simultaneous testing of more than one specimen type from each patient provides a 6.4% additional diagnostic sensitivity. Overall, the high sensitivity of the Trioplex assay demonstrates the utility of this assay ascertaining Zika cases.

Fig. 1

Analytical performance of the Trioplex assay in multiplex format for all target viruses. Normal human serum was spiked with Zika virus, dengue virus type 1 (DENV-1), dengue virus type 2 (DENV-2), dengue virus type 3 (DENV-3) or dengue virus type 4 (DENV-4). Six serial dilutions (1:10) were tested to compare ZIKV detection performance of the Trioplex assay with the CDC ZIKV reference assay (a) or DENV detection with the CDC DENV reference assay (b). Data for DENV-4 is shown. Straight line represents linear regression of CDC reference test. Dashed line represents linear regression for Trioplex assay. Trioplex assay limit of detection evaluated for each DENV serotype and chikungunya virus testing 20 replicates per dilutions: 1:100 before LoD (BLoD), 1:10 before LoD (BLoD), LoD and 1:10 after LoD (ALoD) (c). Mean genome copy equivalents per milliliter (GCE/mL) of viral RNA are displayed at each dilution. Error bars represent GCE/mL standard deviation

Fig. 2

Analytical performance comparison between MP96 and LC 2.0 RNA extraction platforms in serum and urine. a Normal human serum or b urine pooled from healthy donors was contrived with ZIKV at a dilution of 1:10 before the limit of detection (1:10 BLoD), at the limit of detection (LoD), and at 1:10 after the limit of detection (1:10 ALoD). Twenty replicates of each dilution were extracted using the small volume protocol (0.2 mL) and were tested with the Trioplex assay on the ABI7500 Fast Dx instrument. The same serum (c) and urine (d) dilutions were tested on the QuantStudio Dx instrument. Mean genome copy equivalents per milliliter (GCE/mL) of viral RNA detected are displayed at each dilution. Error bars represent GCE/mL standard deviation. Straight line represents linear regression of MP96 platform (Roche). Dashed line represents linear regression of LC 2.0 platform (Roche)

Fig. 3

Analytical performance comparison between MP96 and LC 2.0 RNA extraction platforms in whole blood (EDTA). A pool of whole blood donated by healthy donors was contrived with ZIKV at a dilution of 1:10 before the limit of detection (1:10 BLoD), at the limit of detection (LoD) and at 1:10 after the limit of detection (1:10 ALoD). Twenty replicas of every dilution were extracted using small volume protocol (0.2 mL) and tested with Trioplex assay on the ABI7500 Fast Dx instrument (a) or in the QuantStudio Dx instrument (b). Mean genome copy equivalents per milliliter (GCE/mL) of viral RNA detected are displayed at each dilution. Error bars represent GCE/mL standard deviation. Straight line represents linear regression of MP96 platform (Roche). Dashed line represents linear regression of LC 2.0 platform (Roche)

Fig. 4

Analytical performance comparison between Trioplex assay multiplex and the ZIKV singleplex format assay using small volume and large volume RNA extraction. Normal human serum or urine was contrived with ZIKV at a dilution of 1:10 before the limit of detection (1:10 BLoD), at the limit of detection (LoD), and at 1:10 after the limit of detection (1:10 ALoD). Twenty replicates of every dilution were extracted using the MagNA Pure 96 instrument (Roche) and tested by Trioplex assay multiplex or ZIKV singleplex format assay on the ABI 7500 Fast Dx or the QuantStudio Dx instruments. a Compares the mean genome copy equivalents per PCR reaction (GCE/rxn) of viral RNA extracted from serum at each dilution on the ABI 7500 Fast Dx instrument. A linear regression was plotted for multiplex with small volume protocol (Sv) (0.2 mL) (black straight line), singleplex assay with small volume protocol (gray straight line), multiplex with large volume protocol (Lv) (1 mL) (black dashed line). and singleplex assay with large volume protocol (gray dashed line). b Compares the mean genome copy equivalents per PCR reaction (GCE/rxn) of viral RNA extracted from serum or urine at each dilution on the QuantStudio Dx instrument. A linear regression was plotted for multiplex with small volume protocol serum (Sv) (black straight line), multiplex with large volume protocol serum (Lv) (gray straight line), multiplex with small volume protocol urine (Lv) (black dashed line), and multiplex with large volume protocol urine (gray dashed line). Error bars represent GCE/mL standard deviation. The CT values for every dilution replicate in serum tested was plotted for c small volume and d large volume extractions

Fig. 5

Clinical performance of the Trioplex assay across specimen types. Clinical specimens collected concurrently from 373 cases with previous Zika determination in the acute stage were tested. RNA was extracted with the MagNA Pure 96 small volume external lysis protocol from 373 case-paired serum, 373 urine, and 345 whole blood-EDTA specimens and tested with the Trioplex assay in multiplex format in the ABI 7500 Fast Dx instrument. a Correlation of CT values between case-matching serum and urine specimens; R² = 0.36 p < 0.0001. b Correlation of CT values between case-matching serum and whole blood-EDTA specimens; R² = 0.33 p < 0.0001. c Dataset separated by DPO 1–5 and percent cases detected by the Trioplex assay in every specimen type. The white bar represents the percent of cases with ZIKV detected in serum specimens. The gray bar represents the percent of ZIKV-positive urine specimens that resulted ZIKV-negative in the matching serum specimen, and the checkered box represents the percent of ZIKV-positive whole blood specimens resulted ZIKV-negative in the matching serum specimen. Both gray and checkered boxes represent the added value of testing more than one specimen type with the Trioplex assay