Validation of a Multiplex Real-Time PCR Assay for Detection of Mycobacterium spp., Mycobacterium tuberculosis Complex, and Mycobacterium avium Complex Directly from Clinical Samples by Use of the BD Max Open System

Abstract

A multiplex real-time PCR was validated on the BD Max open system to detect different Mycobacterium tuberculosis complex, Mycobacterium avium complex, and Mycobacterium spp. directly from clinical samples. The PCR results were compared to those with traditional cultures. The multiplex PCR assay was found to be a specific and sensitive method for the rapid detection of mycobacteria directly from clinical specimens.

TEXT

Mycobacterium tuberculosis complex (MTC) and Mycobacterium avium complex (MAC) are the most common slow-growing mycobacteria isolated from respiratory infections worldwide (1, 2). Tuberculosis is still a major global public health problem and one of the leading infectious causes of death, especially in developing countries (1, 3, 4). In contrast, the prevalence of nontuberculous mycobacterial infection has been increasing in developed countries. MAC isolates represent the organisms most frequently associated with nontuberculous mycobacterial lung diseases in most of the world (1, 2, 5). The identification of mycobacteria responsible for diseases has important ramifications for infection control and selection of antimicrobial therapy. Identification, however, is hampered by the slow growth of most mycobacteria, which may take as long as 2 months using traditional culture methods (6, 7).

Molecular methods represent a reliable and rapid alternative for laboratory diagnostics of mycobacteria in clinical samples (1, 3, 4). Several PCR assays have already been described for the detection of mycobacteria; however, some of them are conventional PCR requiring post-PCR processing, others use melting curve analysis and need additional interpretation, and some have not been used directly from clinical samples (8–11). The Cepheid Xpert MTB/RIF assay (Cepheid, Sunnyvale, CA) is an FDA-cleared assay that provides direct detection of MTC and rifampin (RIF) resistance in clinical samples, and it displays good sensitivity relative to that of culture (12). However, it does not detect MAC or of other Mycobacterium species. The BD Max system (BD Diagnostics, Sparks, MD) is an open fully integrated automated molecular platform that combines specimen processing and real-time PCR. In addition to a number of FDA-cleared assays, the BD Max also offers generic extraction kits and PCR reagents to be used on an open platform that allows users to create their own assay using their own set of primers and probes (13, 14). The aim of this study was to validate a multiplex PCR test to detect Mycobacterium spp. (pan-Mycobacterium [PAN]), MTC, and MAC directly from clinical respiratory samples using a user-developed protocol (UDP) on the BD Max open-mode system. This test is for primary diagnosis only and was not designed to detect RIF resistance. The PAN target usually encompasses broad-based characterizations of gene content in a given group of organisms. The pan-Mycobacterium primers used in this study were based on the amplification of the 16S rRNA gene from Mycobacterium species.

A total of 120 frozen clinical specimens previously identified by culture at Tampa General Hospital (TGH) Clinical Laboratory, Tampa, FL, were included in this study. Out of them, 111 were respiratory samples, seven were biopsy samples, and two were peritoneal fluid samples. All samples were first processed by the microbiology laboratory and at the molecular laboratory using the standard-of-care mycobacterial culture and singleplex multistep manual PCR protocols (using the same primer-probe sets described in this paper). Samples were then stored at −80°C. For the new PCR, the samples were thawed, and an aliquot of 500 μl was first treated with proteinase K for 30 min at 60°C, followed by 5 min at 95°C. After this pretreatment, 250 μl of the sample was inoculated into the BD Max sample preparation reagent tube. Extraction and multiplex PCR were performed by use of the BD Max system, using the BD Max ExK DNA-3 extraction kit and the BD Max DNA MMK master mix (BD Diagnostics, Québec, Canada), along with specific in-house-designed primers and probes for PAN, MTC, MAC, and β-globin (BG; internal control) detection (Table 1).

TABLE 1.

Primers and probes used for the BD max multiplex real-time PCR

Target	Primer/probe	Sequence (5′ to 3′)a	Geneb
PAN	Forward	CGAACGGAAAGGYCYCTTCG	16S rRNA
	Reverse	CCGTCGTCGCCTTGGTAG
	Probe	JOE-TTTWGCGGTGTGGGATGRGCCCG-BHQ1
MTC	Forward	CTGTGGGTAGCAGACCTCACCTA	IS6110
	Reverse	CGGTGACAAAGGCCACGTA
	Probe	FAM-TGTCGACCTGGGCAGGGTTCG-BHQ1
MAC	Forward	TTGGGCCCTGAGACAACACT	ITS
	Reverse	GCAACCACTATCCAATACTCAAACAC
	Probe	ROX-CCGTGTGGAGTCCCTCCATCTTGG-BHQ1
BG	Forward	GCAAGGTGAACGTGGATGAA	β-Globin
	Reverse	AACCTGTCTTGTAACCTTGATACCAA
	Probe	Quasar 705-TTGGTGGTGAGGCCCTGGGC-BHQ3

^aBHQ1, black hole quencher 1; FAM, 6-carboxyfluorescein; BHQ3, black hole quencher 3.

^bIS, insertion element; ITS, internal transcribed spacer.

The PCR master mix was distributed in two different tubes snapped into the DNA-3 extraction strip. The first tube, BD Max DNA MMK, was a lyophilized PCR reagent mix containing dinucleoside triphosphates (dNTPs), MgCl₂, Hot Start DNA polymerase, and buffers. The second master mix tube was prepared in house and contained a combination of four sets of primers and probes (concentrations of 1.8 μM of each primer and 0.4 μM of the probe), 2 μl of primer diluent (from BD Max MMK), and water to complete a 12.5-μl final volume. The mycobacterial primers and probes used in this study were modified from three different references (6, 7, 15) and validated to be used as a multiplex real-time PCR for the first time in this study. The PCR cycling conditions were 95°C for 10 min and 40 cycles of 95°C for 15 s and 60°C for 60 s. The PCR detector gain and threshold for each channel were set at 50 and 100 fluorescence detection, respectively.

The limit of detection (LOD) of the new multiplex PCR test for MTC and MAC was determined using the strains M. tuberculosis ATCC 27294 and M. avium ATCC 25291, respectively. For PAN, the LOD was tested three times using three different Mycobacterium isolates: M. abscessus ATCC 19977, M. tuberculosis ATCC 27294, and M. avium ATCC 25291. A 0.5 McFarland (1.5 × 10⁸ CFU/ml) suspension of each strain was prepared in ultrapure water (Sigma-Aldrich Co. Ltd., St. Louis, MO, USA), followed by seven 10-fold dilutions also prepared in ultrapure water. The highest dilution (i.e., lowest concentration) that was positive for PCR was tested in triplicate to determinate the LOD cycle threshold (C_T), mean C_T, and standard deviation (SD). The efficiency (R²) of the PCR was calculated from values generated by the standard curve. The specificity of the test was carried out by testing eight different Mycobacterium strains: M. tuberculosis ATCC 27294, M. avium ATCC 25291, M. intracellulare ATCC 13950, M. abscessus ATCC 19977, M. chelonae ATCC 35752, M. fortuitum ATCC 6841, M. porcinum ATCC 33776, and M. immunogenum ATCC 700505. In addition, extensive in silico analysis of the MTC primers indicated reactivity with the members of this complex. Similarly, in silico analysis of the MAC and PAN primer-probe set demonstrated specific coverage for the respective targets. To assess the ability of the multiplex PCR to correctly identify more than one species of Mycobacterium in a sample, an extra set of eight clinical samples was spiked with more than one Mycobacterium species. Samples were spiked with 125 μl of each McFarland suspension control. Two samples were spiked with M. tuberculosis and M. avium, two samples were spiked with M. tuberculosis and M. intracellulare, two samples were spiked with M. tuberculosis and M. abscessus, and two samples were spiked with M. avium and M. abscessus.

Out of 120 clinical samples included in this study, 78 samples were positive for mycobacteria by culture, and 42 samples were negative. Thirty-seven samples were identified by culture as MAC (17 M. avium, 15 M. intracellulare, 4 M. chimera, and 1 M. bouchedurhonense), 16 as M. tuberculosis complex, and 25 as different mycobacterial species (10 M. abscessus group, 6 M. gordonae, 4 M. fortuitum, 3 M. szulgai, 1 M. marinum, and 1 M. lentiflavum). Mycobacterial species were identified by hsp65 sequencing.

The new multiplex PCR was able to detect mycobacteria in 72 out of the 78 culture-positive samples and to identify 31 samples as MAC and 16 samples as MTC (Fig. 1A to C). Discrepant results between culture and PCR were observed in nine samples. Three of them were positive for MAC by culture and negative for PAN and MAC by PCR. Another three samples were identified as MAC by culture and as MTC and PAN by PCR. Finally, three samples positive for another mycobacterial species by culture (M. gordonae, n = 2; M. fortuitum, n = 1) were negative for PAN by PCR. All six samples negative for PAN PCR were acid-fast bacillus (AFB) smear negative, indicating a very low concentration of the organisms in the sample. Therefore, the discrepant results between PCR and culture in these 6 cases could be explained by the low concentration of the target, making DNA extraction more challenging, especially in sputum samples. The other three samples identified as MAC by culture and as MTC by PCR were from patients previously positive for M. tuberculosis complex infection. We believe that these patients might be infected by both MAC and MTC, and, as MAC grows faster than MTC, culture ended up not identifying MTC. On the other hand, MTC amplification is more efficient than MAC amplification, as indicated by a 2-log-lower LOD, which might explain why PCR identified MTC only.

FIG 1.

(A) BD Max PAN PCR and culture results for mycobacterial detection. (B) BD Max MAC PCR and culture results for M. avium complex detection. (C) BD Max MTC PCR and culture results for M. tuberculosis complex detection.

The sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for each target were: PAN, 89.7%, 100%, 100%, and 84%; MAC, 83.8%, 100%, 100%, and 93.3%; and MTC, 100%, 97.1%, 84.2%, and 100%. The LOD, mean C_T, SD, R², and efficiency for PAN, MAC, and MTC targets are described in Table 2. The efficiency of the test was very good, as indicated by R² values obtained for all 3 genomics probes. All six Mycobacterium strains tested for specificity, as well as the eight clinical samples spiked with more than one mycobacterial species, were correctly identified by the new PCR multiplex assay.

TABLE 2.

LOD dilution, C_T, Mean C_T, SD, R², and efficiency results for each of the multiplex targets^a

Target	Mycobacterial control	LOD dilution	CT	Mean (SD) CT	R2	Efficiency (%)
PAN	M. abscessus ATCC 19977	105	36.1, 38.6, 34.3	36.33 (2.16)	0.99	114.35
	M. avium ATCC 25291	104	41.8, 40.4, 42.7	41.63 (1.6)	0.99	103.98
	M. tuberculosis ATCC 27294	103	37.3, 34.6, 35.4	35.6 (0.87)	0.99	107.23
MAC	M. avium ATCC 25291	103	37.8, 35.9, 37.5	35.77 (1.39)	0.99	94.8
MTC	M. tuberculosis ATCC 27294	101	37.8, 35.9, 37.5	37.1 (1.02)	0.999	102.7

^aLOD, limit of detection; C_T, cycle threshold; R², coefficient of determination.

Culture is the most time-consuming test for the detection and identification of mycobacteria in clinical samples. Molecular techniques have been used to detect and identify mycobacterial species faster than culture, reducing the turnaround time for these results to be reported and accelerating treatment decisions (16, 17). Real-time PCR, in particular, has been used as an important tool for rapid differentiation among different mycobacterial species and for antibiotic therapy decisions, according to each organism's likely susceptibility profile (18). Although molecular assays are rapid and highly specific, negative PCR results do not guarantee that the sample is free of mycobacteria (19). Some limitations can be found when a new test for mycobacterial detection is being validated. This include choosing a sensitive and effective extraction method for mycobacterial detection in sputum samples that retain PCR inhibitors and, for this reason, present high rates of false-negative results (around 12%) (20, 21). Also, it is important to validate a highly sensitive test able to detect lower concentrations of cells, since the sensitivities of real-time PCR assays are usually lower in AFB smear-negative samples than those in AFB smear-positive samples (22, 23). Finally, a multiplex test should be able to detect coinfection with different mycobacterial species, which can be a common situation in human hosts (24). This can be a difficult challenge, since competition for common reagents used for different targets can occur with multiplex PCR tests (1).

The Xpert MTB/RIF assay performed on the GeneXpert platform (Cepheid, Sunnyvale, CA) is an automated molecular test for the simultaneous detection of tuberculosis and rifampin resistance (12). The advantages of this test include the simultaneous identification of M. tuberculosis complex and genetic mutations associated with resistance to rifampin from clinical samples and the potential to provide rapid access to patient results. However, this test cannot identify MAC, which occurs more frequently in our patient population. Moreover, there is a low incidence of multidrug-resistant MTC in our population and in the United States as a whole. Finally, the cost of the Xpert MTB/RIF assay is greater than that of the multiplex assay described here. A fully automated walk-away system, such as the BD Max, which combines extraction and amplification steps in an open-mode platform, has the advantage of reduced hands-on time and low risk of reaction contamination. Moreover, different from other fully automated systems that are capable of running only one sample at a time, the BD Max can obtain results in <4 h for up to 24 samples. There are several studies that demonstrated the use of a laboratory-developed test (LDT) or commercial kits to detect and identify M. tuberculosis and M. avium complex by real-time PCR directly from clinical samples (1, 6, 7, 15). The utility of the BD Max open system has been described for other targets (25, 26); however, this is the first report of a Mycobacterium multiplex PCR developed using this system.

In summary, the new multiplex PCR used on the BD Max open system platform described in this study proved to be a sensitive and specific method to detect mycobacterial species, as well as to identify M. tuberculosis complex and M. avium complex. The introduction of this new multiplex PCR test for the detection of mycobacteria in clinical samples does not exclude culture procedures, which continue to be the most sensitive and gold standard method and are required to obtain isolates for susceptibility testing. However, multiplex molecular tests might be useful for coinfection detection and to accelerate diagnosis in cases that are smear negative. The test can be performed in approximately 4 h and, by providing significantly faster results than those with culture, it can certainly accelerate the initiation of isolation protocols and targeted therapy while awaiting comprehensive broader susceptibility testing results. It could be used in conjunction with the Xpert MTB/RIF assay to obtain RIF resistance data in a more cost-effective approach or in combination with other tests, such as the Hain test (27), to obtain broader resistance testing on samples documented to be positive for MTC or MAC.