A pilot study: VereMTB detection kit for rapid detection of multidrug-resistant mycobcterium tuberculosis in clinical sputum samples

The performance of VereMTBTM Detection Kit for detection of multidrug-resistant tuberculosis (MDR-TB) was validated by using 124 sputum samples. Comparing with MGIT culture, the sensitivity and specificity of VereMTB Detection Kit for MTBC detection were 97.0% and 98.3%, respectively. Compared with MGIT DST, the sensitivity and specificity of VereMTB Detection Kit for RIF resistance detection were 85.7% and 93.9%, respectively, and the sensitivity and specificity of VereMTB Detection Kit for INH resistance detection were 75.0% and 95.7%, respectively. 6 NTM samples were also detected and identified correctly. The VereMTB Detection Kit can detect MDR-TB rapidly and accurately in sputum samples from TB suspects.


Introduction
Tuberculosis (TB) is a respiratory infectious disease that seriously endangers human health. According to the World Health Organization (WHO), 10.0 million new pulmonary TB cases and 1.3 million deaths due to the disease were reported worldwide in 2017 [1]. There are large gaps in the detection and treatment of multidrug-resistant TB (MDR-TB)/rifampicin-resistant TB (RR-TB). In 2017, appropriate treatment was administered for only 25% of the estimated 558 000 MDR/RR-TB cases [1].
Conventional drug susceptibility testing is commonly used for detecting drug-resistant TB, however, this method is time-consuming that requires 8-12 weeks to generate results. Leveraging on PCR technology, relatively faster turnaround time of just a few hours can be achieved in the form of nucleic acid amplification test (NAAT). NAAT can also identify specific molecular markers related to drug resistance thus providing another layer of information for successful patient management [2]. Therefore, rapid detection and retrieving of molecular information of MDR-TB is essential for appropriate treatment and care. Thus, reducing morbidity, mortality, transmission of infection and the emergence of extensively drug-resistant TB (XDR-TB).

PLOS ONE
PLOS ONE | https://doi.org/10.1371/journal.pone.0228312 March 9, 2020 1 / 10 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 VereMTB TM Detection Kit is a NAAT Lab-on-Chip (LoC) assay used for detection and identification of MTBC and nontuberculous Mycobacterium (NTM) as well as the diagnosis of MDR-TB [3,4]. Drug susceptibility to RIF and INH is determined together with MTBC detection in a single sourced sample. The workflow comprises three main steps: PCR, Microarray Hybridization and Microarray Analysis. The initial version of this detection kit was previously evaluated and has shown to be user-friendly and fast with good diagnostic performance [4]. In this current version some primers and probes are redesigned to improve the sensitivity for drug susceptibility detection and NTM identification. The current VereMTB Detection Kit is able to identify and differentiate Mycobacterium species fortuitum ( � unable to differentiate the species in parentheses due to sequence similarity). IS6110 and 16srRNA sequences were used for detection and identification of MTBC and NTM respectively. Probes for detecting DST (RIF and INH resistance) remain unchanged-Wild-type rpoB hot-spot (codons 510-513, 515-518, 523-526, 530-533), inhA (-21/-7) and katG (codon 313-317); as well as mutations L511P (ctg/ccg), D516V (gac/gtc), H526D (cac/ gac), H526Y (cac/tac), S531L (tcg/ttg) for rpoB, t-8a, t-8c, c-15t for inhA, and S315T (agc/acc, agc/aca) for katG. In this study, we evaluated the diagnostic performance and the feasibility of this current version of VereMTB Detection Kit using sputum samples for MTBC/ NTM and MDR-TB detection.

Clinical specimens
A total of 124 sputum samples were prospectively collected from random pulmonary tuberculosis suspects in Beijing Chest Hospital, these pulmonary tuberculosis suspects were continuously enrolled. After smear test, all samples were processed for BACTEC MGIT 960 culture and VereMTB Detection Kit test.

DNA extraction
Sputum sediment from 1 ml aliquot was resuspended in 100 μL sterile water after centrifuge. The sediment was then incubated at 95˚C-100˚C for 20 minutes, followed by sonication in water bath sonicator for 15 minutes. Supernatant containing DNA was transferred into a new microcentrifuge tube.

VereMTB Detection Kit
VereMTB Detection Kit includes a single use disposable VereChip™ (Fig 1) which integrates a miniaturized PCR reactor, and a DNA microarray, as well as reagents and consumables necessary for nucleic acid amplification and DNA hybridization. The main steps for VereMTB test are shown in Fig 2. Each DNA sample was added to two different PCR Master Mixes in two different tubes respectively. The prepared sample mix was loaded into two separate chambers on the Vere-Chip through two inlets. The chip was sealed with clamps, placed into the Temperature Control Module (TCM) of the Temperature Control System (TCS). The PCR reaction was performed within the Chip chambers using the amplification protocol previously mentioned [4]. Upon completion of the PCR protocol, the hybridization buffer mix was loaded through the same inlets for PCR Master Mix loading. This pushed the PCR products out of the outlet to the microarray area for DNA hybridization. Chips were clamped and placed back into the TCM for the hybridization reaction. At completion of hybridization, the Chip was transferred into a 50 mL conical tube filled with wash buffer and centrifuged at 1500 x g for 2 minutes and dried in an empty tube using the same settings. After drying, a chip image was captured with the Optical Reader (Veredus Laboratories Pte Ltd) and the VerePLEX TM software automatically performed the analysis generating a report based on the diagnostic rules fixed for target detection.

Sequencing
All the culture positive strains were collected for DNA sequencing to identify Mycobacterium species related gene (16S rRNA) and drug resistance related gene mutation for RIF (rpoB) and INH (katG and inhA) at National TB Reference Laboratory ( Table 1). The sequencing results were entered into the Basic Local Alignment Search Tool (BLAST, http://www.ncbi.nlm.nih. gov/BLAST), for the alignment with reference Mycobacterium tuberculosis strain H37Rv to identify the mutations in rpoB, katG and inhA gene.

Ethics statement
This study was reviewed and approved by the ethics review board of the Chinese Center for Disease Control and Prevention. All the specimens we used were useless sample from clinic laboratory after routine examination. The requirement to obtain individual informed consent was waived by the review board.

The performance of VereMTB Detection Kit for MTBC and NTM detection
A total of 124 sputum samples were tested for the validation, comprising of 60 Acid-Fast Bacilli (AFB) smear-negative samples and 64 AFB smear-positive samples ( Table 2). Fifty-one out of the 60 smear-negative samples were culture negative and the remaining 9 smear-negative samples were culture positive. These 9 samples were confirmed to have MTBC according to sequencing result of 16S rRNA. All of the smear-positive samples were culture positive; 58 smear-positive samples were culture-positive for MTB and the remaining 6 were deemed as NTM based on 16S rRNA sequencing. Using the VereMTB Detection Kit, 57 out of these 58 smear-positive MTBC samples were detected. 5 out of the 6 NTM samples were identified as M. intracellulare and one sample was determined to be M. kansasii, and this result was in accordance with 16S rRNA sequencing ( Table 2).
Refer to Table 3 for data comparison of MGIT cultured method and VereMTB Detection Kit for AFB smear negative samples, smear positive samples, and all samples. 9 out of 60 AFB smear negative samples were MTBC culture positive, and 8 out of these 9 samples were Ver-eMTB MTBC positive. 1 AFB smear negative and cultured negative sample showed MTBC positive result by VereMTB detection kit. 58 out of 64 AFB smear positive samples were MTBC culture positive, and 57 out of these 58 samples were VereMTB MTBC positive. The remaining 6 were deemed as NTM based on 16S rRNA sequencing (refer to Table 2). In total, out of the 124 samples, 67 were MTBC culture positive, and 65 out of these 67 samples were VereMTB MTBC positive.
Summary of the performance of VereMTB Detection Kit for detection of MTBC is shown in Table 4. The overall sensitivity of the assay was 97.0% (65/67). The sensitivity in smear- Twelve of the samples determined as RIF resistant from MGIT DST was consistently determined as RIF resistant using VereMTB Detection Kit. Two specimens were, however, inconsistent with MGIT DST where these samples were identified as RIF susceptible by VereMTB Detection Kit. These were by sequencing results indicating genotypic mutation (H526L and H526Y) within the rpoB gene of both samples. Three MGIT DST RIF susceptible specimens were identified as RIF resistance by Ver-eMTB Detection Kit, further DNA sequencing analysis revealed that 2 of which were confirmed to be genotypically resistant due to mutation (H526Y and D516V) in rpoB gene (  In terms of diagnostic accuracy for RIF and INH resistance when comparing to the MGIT DST, VereMTB Detection Kit showed good sensitivity (85.7% and 75.00% respectively) and excellent specificity (93.9% and 95.7% respectively) ( Table 7).

Discussion
Early and accurate diagnosis is critical to interrupt transmission of TB and MDR TB, especially in the absence of an effective vaccine [6]. This is the first evaluation of sputum specimens from pulmonary TB suspects in China using the VereMTB Detection Kit. Our results showed that theVereMTB Detection Kit can effectively diagnose MTBC, NTM and MDR-TB from sputum samples. Relative to the MGIT DST culture method, the turnaround time of VereMTB Detection Kit was sharply shortened from 15 days to 3 hours.
MTB Detection Kit showed good sensitivity (97.0%) and specificity (98.3%) for MTBC detection. However, 1 false positive result was observed which might be due to the carry over contamination from previous tested samples since it is a PCR based detection kit.
For drug resistance detection, VereMTB Detection Kit was able to give valid result for all smear positive and MTBC detected samples, some of the smear negative and MTBC detected samples (5/8), and showed good sensitivity (85.7% for RIF and 75% for INH) and excellent specificity (93.9% and 95.7% respectively). This kit provided "indeterminate" drug resistance result when MTBC was detected but the kit sensitivity was not enough to conclude drug susceptibility information.
Recent studies have shown that rapid liquid culture systems such as MGIT 960 and the proportion method with shorter incubation time (4 weeks) often fail to detect strains that are resistance to RIF, particularly strains contain specific mutations within rpoB gene codon 511 (such as L511P), codon 516 (such as D516Y), codon 526 (such as H526N, H526L and H526S), and codon 533 (such as L533P) [7,8]. Patients infected with MTBC with the above disputed rpoB mutations often fail treatment or relapse due to low-level resistance to RIF (minimum inhibitory concentration, MIC of 0.5-2.0 μg/ml) [9,10]. VereMTB Detection Kit is able to detect L511P mutation in rpoB gene, and is indicative of mutation within the rpoB hot-spot region (rpoB wild type probe missing), which provide low-level RIF resistance information to doctors. In this study, two samples carrying rpoB H526L mutation were RIF resistant according to MGIT DST result. One was detected by VereMTB chips (RIF resistant and rpoB wild type probe missing) and the other was not detected. Mutations in katG gene and inhA gene are responsible for about 75% isoniazid-resistant isolates [11]. Drug resistance determination on the VereMTB Detection Kit interrogates mutations in these genes. The sensitivity of VereMTB Detection Kit for detection of INH resistance was 75.0%. The potential cause of this low performance could be due to mutations outside of the katG gene and inhA gene, not analyzed on the VereMTB Detection Kit. In this study, four INH phenotypically resistant samples showed INH susceptibility using VereMTB Detection Kit. Sequencing results revealed that all 4 samples carried no mutations in katG and inhA gene suggesting a potential mutation out of the VereMTB detection scope. According to recent research, the oxyR-ahpC gene region has been implicated as an indicator of the isoniazid resistance [12,13]. Thus, this lower sensitivity for INH resistance was likely associated with the high prevalence of oxyR-ahpC region mutations within China, which is not detected by the VereMTB Detection Kit.
One unique feature of VereMTB Detection Kit is its 'two in one' capability in combining MTB drug resistance detection and common NTM detection. The assay successfully differentiated and identified both MTB and NTM, with the identified NTMs confirmed using sequencing. The performance of NTM detection has, however, not been rigorously tested due to the small test size of NTM in this study. The VereMTB Detection Kit covers a 14 NTMs including some of the common NTMs in China-M. intracellulare, M. avium, M. abscessus/chelonae, M. kansasii and M. fortuitum [14,15]. With rising NTM infections in China [16] and cases of NTM pulmonary infection being misdiagnosed as pulmonary TB due to the similar clinical symptoms and current diagnostic criteria [17,18], the VereMTB Detection Kit may be a useful tool for diagnostic and clinical use.
However, due to the limitation of number of probes the current Verechip microarray is able to fix (maximum 126 probes include hybridization control probes, PCR control probes and all detection probes, spotted in 3 replicates to avoid "ambitious" situation), VereMTB Detection Kit is only capable to detect MTBC, main NTM, and stain susceptibility to rifampicin and isoniazid. This assay has potential of detecting more resistance-conferring mutations, and strain susceptibility to more antibiotics if higher density VereChip platform is in operation.
We also realized the limitation of this study. It was just a preliminary validation study on new molecular assay and we just collected relatively small number sample specimens. The performance looks good and large scale sample evaluation studies across multiple communicable disease centers in order to further examine the robustness of this assay will be needed in the future.

Conclusions
In conclusion, the VereMTB Detection Kit provides a much shorter turnaround time compared to conventional DST. According to our evaluation result, the VereMTB Detection Kit showed good sensitivity in detecting MTB and RIF and/or NIH resistance TB from sputum specimens, and it was also able to detect and identify NTM in the same test.