Evaluation of a Single-Tube Multiplex Real-Time PCR for Differentiation of Members of the Mycobacterium tuberculosis Complex in Clinical Specimens

Transcription

1 JOUNAL O CLINICAL MICOBIOLOGY, July 2011, p Vol. 49, No /11/$12.00 doi: /jcm Copyright 2011, American Society for Microbiology. All ights eserved. Evaluation of a Single-Tube Multiplex eal-time C for Differentiation of Members of the Mycobacterium tuberculosis Complex in Clinical Specimens Tanya A. Halse, Vincent E. Escuyer, and Kimberlee A. Musser* Wadsworth Center, New York State Department of Health, Albany, New York eceived 7 March 2011/eturned for modification 22 April 2011/Accepted 5 May 2011 Members of the Mycobacterium tuberculosis complex (MTBC) differ in virulence attributes, drug resistance patterns, and host preferences. The rapid differentiation of these species to determine zoonotic or human sources of tuberculosis disease or to direct treatment can benefit both public health and patient management. Commercially available assays cannot differentiate these species, and published assays have not been evaluated directly on clinical specimens. A real-time C assay for the differentiation of M. tuberculosis, M. bovis, M. bovis BCG, M. africanum, M. microti, and M. canettii was developed. The presence or absence of regions of difference (D) between the genomes of members of the MTBC allowed for the design of a single-tube five-plex real-time C assay to differentiate these species. This assay assesses the presence of D1, D4, D9, D12, and a region exterior to D9 which is present in all MTBC members. To evaluate the performance of this assay, 192 clinical specimens positive for MTBC by real-time C were tested, resulting in a 94% correlation of the real-time C with the identification results obtained with cultured material. Additionally, 727 Bactec MGIT 960-positive cultures were tested, resulting in a 97% concordance between the methods. This real-time C is an inexpensive and rapid (2.5-h) method performed in a closed-format system and requiring minimal hands-on time that can be implemented in a clinical laboratory and used directly on clinical specimens. The Mycobacterium tuberculosis complex (MTBC) comprises the closely related organisms M. tuberculosis, M. africanum, M. bovis, M. bovis BCG, M. microti, M. canettii, M. caprae, M. pinnipedii, and M. mungi (1). Differentiation between the species within this complex is important for public health surveillance and reference testing, as most species within the MTBC have been reported to infect humans (6). Additionally, it may be important information for physicians to rapidly identify severe side effects of M. bovis BCG in patients with bladder cancer or in vaccinated individuals or to assess transmission of M. bovis from animals or animal products. It is also helpful to direct patient treatment, as M. bovis is intrinsically resistant to pyrazinamide (ZA) (14). A recent report on the incidence of M. bovis in San Diego, CA, highlighted the importance of routine species-level identification in U.S. tuberculosis (TB) surveillance. That investigation reported the growing incidence of TB caused by M. bovis (45% of all culture-positive TB cases in children between 1994 and 2005), which may be representative of other communities in the United States (17). In New York, 3 to 7% of the MTBC specimens that our laboratory receives each year are identified as species other than M. tuberculosis. Currently there are few methods available to rapidly differentiate species within the MTBC. The existing methods are not * Corresponding author. Mailing address: Wadsworth Center, New York State Department of Health, 120 New Scotland Avenue, Albany, NY hone: (518) ax: (518) ublished ahead of print on 18 May always suitable; as conventional (biochemical) methods are highly subjective, high-pressure liquid chromatography (HLC) can identify only M. bovis BCG (3), and DNA probe and amplification assays based on 16S rna gene sequences can identify specimens only as MTBC. ublished molecular assays to differentiate members of the MTBC are not in realtime C formats (6, 14, 20), do not identify members other than M. tuberculosis, M. bovis, and M. bovis BCG (10, 15), and are not validated for use on clinical specimens (10, 15, 16). To date no single assay to perform this diagnostic test exists for probe-based differentiation of members of the MTBC directly from clinical specimens. Our laboratory has historically relied on a well-validated conventional C based on comparative genomics (14), specifically, regions of difference (D), that is utilized on cultured material to differentiate the MTBC species present. This assay is generally reliable but has limitations because culture may take 2 to 8 weeks and the conventional C assay involves two to five separate reactions requiring postamplification processing, thereby increasing the risk of contamination. or these reasons, a single-tube, multiplex, real-time C assay (MTBC-D real-time C) was developed for rapid differentiation of members of the MTBC from both clinical specimens and cultured material, based on the same comparative genomics. This new assay was evaluated on 919 patient samples received by the Wadsworth Center (WC), New York State Department of Health, over a 16-month period. (art of the content of this paper was presented at the 110th General Meeting of American Society for Microbiology, May 2010.) 2562

2 VOL. 49, 2011 EAL-TIME C TO DIEENTIATE MEMBES O THE MTBC 2563 TABLE 1. rimers, probes, and linearity data for assay targets Target and primer or probe a Sequence (5 33 ) luorophore r 2 value Calculated efficiency (%) D1 VIC-MGB CCC TTT CTC GTG TTT ATA CGT TTG A GCC ATA TCG TCC GGA GCT T CAC TCT GAG AGG TTG TCA D4 AM-MGB CCA CGA CTA TGA CTA GGA CAG CAA AAG AAC TAT CAA TCG GGC AAG ATC ACC AGT GAG GAA ACC D9 TEX615-BHQ TGC GGG CGG ACA ACT C CAC TGC GGT CGG CAT TG AGG TTT CA C CTT CGA C CC b D12 NED-MGB CGT TGG AAC GCG AAA TAC G CCA GGA TAT GGG CGC AAA T TGC GCT GAC CCC AC ext-d9 CY5-BHQ GCC ACC ACC GAC TCA TAC CGA GGA GGT CAT CCT GCT CTA G TT CTT CAG CTG GT C C b a, forward primer;, reverse primer;, probe. b Locked nucleic acid probe; indicates the insertion of an LNA base. MATEIALS AND METHODS Clinical specimens. To evaluate the MTBC-D real-time C assay directly on clinical specimens, 192 specimens (94.3% respiratory specimens and 5.7% other types, including abscess, aspirate, lymph node, body fluid, gastric fluid, urine, and wound specimens) initially positive for MTBC by real-time C targeting IS6110 (IS6110 real-time C) were included. A second study evaluated the MTBC-D real-time C on 727 Bactec MGIT 960-positive cultures. Specimens were all received in the Mycobacteriology Laboratory at the WC between 10 ebruary 2009 and 17 June All clinical specimens and Bactec MGIT 960 cultures were processed as previously described (5). Identification by conventional C, Accurobe, and sequencing. The different members of the MTBC were identified by conventional C as described previously (14). Briefly, each organism was tested for the presence or absence of genomic D, namely, D1, D4, D9, D10, and D12, to determine a distinct signature pattern, allowing the assay to differentiate among members of the MTBC. Members of the M. avium complex (MAC) were identified by using the Gen-robe Accurobe MAC test. Other nontuberculous mycobacteria (NTM) were identified by sequence analysis of three target genes, i.e., the 16S rna gene, hsp65, and rpob (8, 9, 12). MTBC-D real-time C primers and probes. The MTBC-D real-time C targets D (D1, D4, D9, and D12) (2, 4, 6, 18) with primers and probes designed using rimer Express v. 3.0 (Applied Biosystems, Inc. [ABI], Carlsbad, CA). An additional primer-probe set (ext-d9) was designed to target a conserved region that is found external to D9 and is present in all members of the MTBC. All primer and probe sequences are shown in Table 1 with the respective labeled reporter dye. The D1, D4, and D12 assays utilize MGB probes (ABI), and the D9 and ext-d9 assays utilize locked nucleic acid (LNA) probes (Integrated DNA Technologies [IDT], Coralville, IA). All primers were obtained from IDT. MTBC-D real-time C. ollowing initial decontamination (5), the clinical specimens were heat treated at 80 C for 1 h, processed using a GeneOhm lysis kit (BD Diagnostics, San Diego, CA), and tested by the MTBC-D real-time C at the original concentration and at a 1:5 dilution. The cultured material determined to be positive by Bactec MGIT was diluted 1:5, heat treated at 80 C for 1 h, and tested by the MTBC-D real-time C in duplicate. This assay was performed in a 25- l volume using the oche LightCycler-astStart DNA Master Hybridization robes kit (oche Diagnostics Corporation, Indianapolis, IN). Each reaction mixture consisted of 1 LightCycler-astStart DNA Master Hybridization robes mix, 4 mm MgCl 2, 450 nm forward and reverse primers, 125 nm probes (Table 1), sterile water, and 5 l of DNA template. Thermocycling conditions were as follows: 1 cycle at 95 C for 10 min, followed by 45 cycles at 95 C for 15 s and 60 C for 1 min. Thermocycling, fluorescence data collection, and data analysis were performed with a 7500 fast real-time C system (v ) (ABI) according to the manufacturer s instructions, with the passive reference dye, OX, turned off. Clinical specimens were assessed for inhibition in the initial IS6110 real-time C assay (5). A 1:5 dilution of material in culture-positive MGIT tubes was tested, which alleviated inhibition potentially found in the cultured material. Interpretation of MTBC-D real-time C. A positive target result (cycle threshold [C T ] value of 45) indicated that the D was present, and a negative target result (C T value undetermined) indicated that the D was absent. The resulting target patterns for each specimen were compared to the signature patterns in Table 2 to determine a match to the signature pattern for each species of MTBC. igure 1 illustrates a typical amplification plot for an M. tuberculosis-positive sample, depicting all five fluorophores shown together (ig. 1A) and separately (ig. 1B to ). To meet the assay criteria, the C T value of the ext-d9 target was expected to be on average 3 C T values higher than the result with the initial test by the IS6110 real-time C but could potentially be higher, depending on the number of copies of IS6110 present. C T values for each positive target were TABLE 2. Signature patterns of positive and negative real-time C results used to determine MTBC species Organism C result for target: D1 D4 D9 D12 ext-d9 M. tuberculosis M. bovis M. bovis BCG M. africanum M. microti M. canettii NTM a a NTM, nontuberculous mycobacteria.

3 2564 HALSE ET AL. J. CLIN. MICOBIOL. Downloaded from IG. 1. Amplification plots depicting five fluorophores used in MTBC-D real-time C. (A) A positive M. tuberculosis sample showing amplification with all five fluorophores is shown along with negative samples. (B to E) Each fluorophore is shown individually: VIC-MGB (B), AM-MGB (C), TEX615-BHQ (D), NED-MGB (E), and CY5-BHQ (). expected to be within 4 C T values of each other. When a signature pattern was identified, meeting the assay criteria, a positive identification result for the MTBC species was reported. Tests were repeated once if discrepant results between replicates were encountered or if the C T values for each positive target were not within 4 C T values of each other. The specimen result was considered inconclusive if the discrepancies between replicates or targets remained upon repeat testing. The identification of specimens with negative or inconclusive results was not reported until culture was complete. igure 2 illustrates the testing algorithm utilizing MTBC-D real-time C on clinical specimens. Specificity. The specificity of the assay was assessed by testing approximately 10 6 genome copies of DNAs purified from MTBC members, NTM, and other respiratory pathogens with similar clinical symptoms. This testing included the following fungal and bacterial species (the number of strains is specified in parentheses if more than one was tested): Coccidioides immitis, Histoplasma capsulatum, Blastomyces dermatitidis, Bordetella pertussis, Chlamydophila pneumoniae, Haemophilus influenzae type b, Legionella pneumophila, Mycoplasma pneumoniae, Streptococcus pneumoniae, methicillin-resistant Staphylococcus aureus (MSA), methicillin-susceptible Staphylococcus aureus (MSSA), Klebsiella pneumoniae, seudomonas aeruginosa, Escherichia coli, Moraxella catarrhalis, Streptococcus sanguinis, Neisseria mucosa, Neisseria sicca, M. abscessus, M. africanum (2), M. asiaticum, M. avium, M. bovis (5), M. bovis BCG (10), M. canettii, M. chelonae, M. flavescens, M. florentinum, M. fortuitum, M. gastri, M. gordonae, M. interjectum, M. intracellulare, M. kansasii, M. lentiflavum, M. malmoense, M. massiliense, M. microti (3), M. mucogenicum, M. nebraskense, M. neoaurum, M. paraffinicum, M. peregrinum, M. phlei, M. phocaicum, M. porcinum, M. scrofulaceum, M. simiae, M. szulgai, M. terrae, M. tuberculosis (6), and M. xenopi. No M. caprae, M. pinnipedii, orm. mungi strains were available for testing. Additionally, no ZA-susceptible M. bovis strains were available for testing (13). Strains were from either ATCC or the culture collection of the WC. All mycobacterial strains from the WC used in the specificity panel underwent 16S rna, hsp65, and rpob DNA sequence analysis for confirmation of identity. Sensitivity. The analytical sensitivity of the assay was determined by duplicate testing of a standard curve of 10-fold dilutions of DNA isolated from a strain of M. tuberculosis (containing one copy of IS6110), with known concentrations in CU. The amplification efficiency of the C for each target was calculated from the slope (m) of the line generated by the standard curve using the formula E (10 1/m 1) 100. The limit of detection (LOD) of the assay for sputum specimens was determined by seeding the dilutions described above into known negative sputum specimens. The sputum was processed as described previously, and each dilution was tested in triplicate in the MTBC-D real-time C (5). ESULTS Specificity and sensitivity. The results of specificity testing for the MTBC-D real-time C on all 26 samples representing organisms of the MTBC correlated 100% with the expected results. All 43 strains of NTM and other bacteria found in respiratory specimens tested negative, resulting in 100% specificity for the cohort (data not shown). The analytical sensitivity of the MTBC-D real-time C on December 13, 2018 by guest

4 VOL. 49, 2011 EAL-TIME C TO DIEENTIATE MEMBES O THE MTBC 2565 was determined to be 2.5 CU for cultures, and the LOD for sputum specimens was determined to be 5 CU. The efficiency of each target of the assay is shown in Table 1. MTBC-C-positive clinical specimen study. A total of 192 clinical specimens initially positive for MTBC by IS6110 realtime C were tested with the MTBC-D real-time C assay (Table 3). Of these, 165 had paired Bactec MGIT 960- positive cultures with conventional C results to determine the MTBC species for comparison. A total of 155 (94%) correlated and were identified as follows (n): M. tuberculosis (145), M. africanum (4), M. bovis (2), M. bovis BCG (2), and negative for MTBC (2). A small subset, 10 specimens (6%), were determined to be negative (5) or inconclusive (5) by MTBC-D real-time C but were later identified as having M. tuberculosis by conventional C and MTBC-D real-time C once they became culture positive. These 10 specimens produced high C T values in the initial IS6110 real-time C assay, indicating a low level of DNA present. Of note, 27 of the 192 (14%) were culture negative, and 18 (9%) of these were determined to contain M. tuberculosis or M. africanum by the MTBC-D real-time C. Clinical specimen testing results IG. 2. MTBC-D testing algorithm for clinical specimens. indicated that no other material present in the specimens interfered with or was inhibitory to the assay. Bactec MGIT 960-positive culture study. In a separate study, 727 Bactec MGIT 960-positive cultures were tested with the MTBC-D real-time C, and the results were compared to those of conventional C testing (Table 4). We found 708 specimens (97%) concordant by the two methods, and these were identified as follows (n): M. tuberculosis (373), M. africanum (11), M. bovis (15), M. bovis BCG (6), and negative for MTBC (303). Of the remaining 19 specimens, 15 (2.1%) had inconclusive results by MTBC-D real-time C and were negative for MTBC by conventional C. Two (0.3%) were incorrectly identified by MTBC-D real-time C as having M. africanum and were later determined to have M. tuberculosis. Two specimens (0.3%) were inconclusive by MTBC-D real-time C and were later identified as containing M. tuberculosis and M. africanum. DISCUSSION This study reports the development and evaluation of a single-tube, five-target, real-time C assay which can differ- TABLE 3. Comparison of MTBC-D real-time C on clinical specimens to conventional C on Bactec MGIT 960-positive cultures esult with MTBC-D real-time C on clinical specimens No. of specimens with result with conventional C on Bactec MGIT 960-positive cultures: M. tuberculosis M. africanum M. bovis M. bovis BCG NTM Culture negative M. tuberculosis M. africanum 4 1 M. bovis 2 M. bovis BCG 2 Negative Inconclusive 5 6 Total (n 192)

5 2566 HALSE ET AL. J. CLIN. MICOBIOL. esult with MTBC-D real-time C TABLE 4. Comparison of MTBC-D real-time C to conventional C on Bactec MGIT 960-positive cultures No. of specimens with result with conventional C: M. tuberculosis M. africanum M. bovis M. bovis BCG NTM M. tuberculosis 373 M. africanum 2 11 M. bovis 15 M. bovis BCG 6 Negative 303 Inconclusive Total (n 727) entiate the species of the MTBC in approximately 2.5 h directly on clinical specimens. The ease of use, decrease in hands-on time, and decreased potential for amplicon contamination found with this assay compared to separate Cs (14, 15) are invaluable. We found a substantial savings in both technician time and consumable costs when comparing this new assay to the existing assay, which requires multiple conventional Cs and postamplification analysis. Turnaround time was also significantly reduced by the run time of the assay combined with its use on clinical specimens positive for MTBC DNA. To date there is no single-probe-based assay for differentiation of members of the MTBC directly from clinical specimens. insky and Banaei (15) reported a multiplex real-time C used to differentiate the members of the MTBC, but that assay relies on SYB green I detection and melting curve analysis, which do not provide the added specificity of a probebased assay (7). The assay also requires two separate steps and was evaluated on culture isolates only. All other published assays have utilized conventional C (6, 14) or have been evaluated only on cultured material (10, 16). We found MTBC-D real-time C testing of IS6110 realtime C-positive clinical specimens to be concordant with conventional C testing of Bactec MGIT 960-positive cultures for 94% (155/165) of the specimens tested. The utilization of the MTBC-D real-time C allowed for rapid identification, providing diagnostic information within 2 days of receipt of the clinical specimen. The remaining 6% (n 10) of the specimens tested were negative with this assay, and we believe that they contained MTBC organisms at below the LOD of the assay as predicted by the IS6110 real-time C C T values. The LOD of this assay is 5 CU for sputum specimens, and the LOD for the IS6110 real-time C used to detect MTBC initially is 0.5 CU (5). or these specimens, a positive culture was obtained and tested to identify the species present, and an initial MTBC-D real-time C result differentiating the MTBC species in the clinical specimen was not reported (ig. 2). Interestingly, we determined the MTBC species for 9% (18/ 192) of the specimens that were ultimately found to be culture negative, which were not included in the comparative study. This is likely due to the presence of dead organism from a patient undergoing treatment (5, 11, 19) or to mixed cultures containing other microorganisms that have overgrown MTBC. These 18 specimens were reported, and the final result indicated the presence of DNA with no viable MTBC organisms isolated. In our analysis of 727 Bactec MGIT 960-positive cultures, 97% (708/727) of the samples were identified as containing the same species with the MTBC-D real-time C and conventional C. The MTBC-D real-time C provided a more rapid result without the potential for contamination within 2.5 h, compared to 6 to 72 h with the conventional C test. The remaining 3% (19) of the samples included 17 (2.3%) that were inconclusive by the MTBC-D real-time C, and the results were not reported. There were 2 culture-positive M. tuberculosis specimens that were misidentified by this assay as M. africanum due to a negative result for the D9 target. They were further tested with the D9 primers and probe as a singleplex assay, and one resulted in a positive result and one resulted in a negative result. The D9 target is the least efficient of all five targets, as shown in Table 1; therefore, it is likely that the complexity of the multiplex reaction combined with the content of this particular culture was the cause of this false-negative result. urther work to investigate this finding is ongoing. Nevertheless, the D9 target was successfully amplified in 372 other specimens identified as containing M. tuberculosis. In summary, this paper describes the evaluation of a singletube five-plex real-time C assay that can accurately differentiate members of the MTBC in 2.5 h and can be introduced in the molecular diagnosis of mycobacteria. This assay provides added value over the multiple single-target conventional C assays and other published assays because of its high level of specificity and sensitivity, short turnaround time, cost-effectiveness, and performance on clinical specimens determined to be positive for MTBC. The use of the MTBC-D real-time C assay may save up to 8 weeks of time to differentiate species of the MTBC and may provide information to achieve proper drug therapy and early insight into TB transmission. ACKNOWLEDGMENTS We acknowledge Jeremy ivenburg for screening clinical specimens with the IS6110 real-time C and Michele Isabelle, Andrea Doney, Susan Wolfe, and hyllis Cunningham for performing the conventional C assay used as a comparison in this study. EEENCES 1. Alexander, K. A., et al Novel Mycobacterium tuberculosis complex pathogen, M. mungi. Emerg. Infect. Dis. 16: Brosch,., et al Use of a Mycobacterium tuberculosis H37v bacterial artificial chromosome library for genome mapping, sequencing, and comparative genomics. Infect. Immun. 66: loyd, M. M., V. A. Silcox, W. D. Jones, Jr., W.. Butler, and J. O. Kilburn Separation of Mycobacterium bovis BCG from Mycobacterium tuber-

6 VOL. 49, 2011 EAL-TIME C TO DIEENTIATE MEMBES O THE MTBC 2567 culosis and Mycobacterium bovis by using high-performance liquid chromatography of mycolic acids. J. Clin. Microbiol. 30: Gordon, S. V., et al Identification of variable regions in the genomes of tubercle bacilli using bacterial artificial chromosome arrays. Mol. Microbiol. 32: Halse, T. A., et al Combined real-time C and rpob gene pyrosequencing for rapid identification of Mycobacterium tuberculosis and determination of rifampin resistance directly in clinical specimens. J. Clin. Microbiol. 48: Huard,. C., L. C. de Oliveira Lazzarini, W.. Butler, D. van Soolingen, and J. L. Ho C-based method to differentiate the subspecies of the Mycobacterium tuberculosis complex on the basis of genomic deletions. J. Clin. Microbiol. 41: Jamnikar Ciglenecki, U., J. Grom, I. Toplak, L. Jemersic, and D. Barlic- Maganja eal-time T-C assay for rapid and specific detection of classical swine fever virus: comparison of SYB Green and TaqMan MGB detection methods using novel MGB probes. J. Virol. Methods 147: Kim, B.-J., et al Identification of mycobacterial species by comparative sequence analysis of the NA polymerase gene (rpob). J. Clin. Microbiol. 37: Kirschner,., et al Genotypic identification of mycobacteria by nucleic acid sequence determination: report of a 2-year experience in a clinical laboratory. J. Clin. Microbiol. 31: Lee, H.., et al Novel multiplex C using dual-priming oligonucleotides for detection and discrimination of the Mycobacterium tuberculosis complex and M. bovis BCG. J. Clin. Microbiol. 48: Lin, J., and H. J. Harn Application of the polymerase chain reaction to monitor Mycobacterium tuberculosis DNA in the CS of patients with tuberculosis meningitis after antibiotic treatment. J. Neurol. Neurosurg sychiatry 59: McNabb, A., et al Assessment of partial sequencing of the 65-kilodalton heat shock protein gene (hsp65) for routine identification of Mycobacterium species isolated from clinical sources. J. Clin. Microbiol. 42: Niemann, S., E. ichter, and S. usch-gerdes Differentiation among members of the Mycobacterium tuberculosis complex by molecular and biochemical features: evidence for two pyrazinamide-susceptible subtypes of M. bovis. J. Clin. Microbiol. 38: arsons, L. M., et al apid and simple approach for identification of Mycobacterium tuberculosis complex isolates by C-based genomic deletion analysis. J. Clin. Microbiol. 40: insky, B. A., and N. Banaei Multiplex real-time C assay for rapid identification of Mycobacterium tuberculosis complex members to the species level. J. Clin. Microbiol. 46: ounder, J. I., et al Mycobacterium tuberculosis complex differentiation by genomic deletion patterns with multiplex polymerase chain reaction and melting analysis. Diagn. Microbiol. Infect. Dis. 67: odwell, T. C., M. Moore, K. S. Moser, S. K. Brodine, and S. A. Strathdee Tuberculosis for Mycobacterium bovis in binational communities, United States. Emerg. Infect. Dis. 14: Talbot, E. A., D. L. Williams, and. rothingham C identification of Mycobacterium bovis BCG. J. Clin. Microbiol. 35: Velayati, A. A., V. V. Bakayev, and A.. Bahrmand Use of C and culture for detection of Mycobacterium tuberculosis in specimens from patients with normal and slow responses to chemotherapy. Scand. J. Infect. Dis. 34: Warren,. M., et al Differentiation of Mycobacterium tuberculosis complex by C amplification of genomic regions of difference. Int. J. Tuberc. Lung Dis. 10: Downloaded from on December 13, 2018 by guest