| Background and objectivesThe majority of mycobacteria are saprophytic bacteria and the minority could cause human or animal diseases. For example, mycobacterium tuberculosis complex (MTC) is the causative agent of tuberculosis (TB) and Mycobacterium leprae is the pathogen of leprosy, and some non-tuberculous mycobacteria (NTM) could cause non-tuberculosis mycobacterial diseases. There are some differences between different mycobacterium in biological characteristics and pathogenesis and drug susceptibility. Thus, the differentiation and identification of mycobacteria is critical for diagnosis and treatment and prognosis of TB and non-tuberculosis mycobacterial diseases. It is needed to develop a rapid, simple, accurate and cost-effective method which could be used for the differentiation and identification of mycobacteria in all levels of laboratories.The traditional method for differentiation and identification of mycobacteria, which recognize phenotypic and biochemical characteristics, is slow, laborious, cumbersome and relatively subjective. Gas chromatography, high-performance liquid chromatography and mass spectroscopy are cumbersome, labour-consuming and technically demanding. Many molecular biology methods, including DNA hybridization, conventional Polymerase chain reaction (PCR), real-time PCR, DNA fingerprinting and DNA sequencing, have been used to differentiate and identify mycobacteria. However, these genotypic methods all have some disadvantages. DNA hybridization is costly since many probes are needed and conventional PCR has increased risk of cross contamination because subsequent product analysis is needed. Real-time PCR is expensive as the use of labeled probe and DNA fingerprinting is cumbersome and costly because of the restriction endonuclease ananlysis. Although DNA sequencing is the gold standard for molecular identification, it is not suitable to be applied in clinical laboratory since it needs expensive DNA sequencer.At the end of last century, a colloidal gold-based immunochromatographic assay using anti-MPT64monoclonal antibody labeled with colloidal gold particles was developed for rapid differentiation of mycobacterial culture. Recently, a rpoB-based duplex real-time PCR, which is rapid and cost-effective method for the differentiation between MTC and NTM, has shown promise for routine use in clinical laboratory. Then, which of the two methods is better for rapid differentiation of mycobacteria? Whether the two methods could be used under different conditions? How to choose and according to what principles to choose and what will happen? Part1of our study aimed to answer these questions.Different double-strand DNAs could generate different melting profiles due to distinct sequence length and sequence composition, so double-strand DNAs could be distinguished by melting curve analysis. High-resolution melting curve analysis (HRMA), which could detect single base substitution, is a highly sensitive method for DNA analysis, and it is developed on the basis of conventional melting curve analysis. If the above-mentioned duplex real-time PCR could rapidly differentiate MTC from NTM by conventional melting curve analysis, it is possible to identify NTM to species by HRMA. Moreover, rpoB is a single-copy gene and it is variable in mycobacterium species. Thus, rpoB is better than multi-copy genes for rapid identification of NTM theoretically. Part2of our study aimed to adopt simplex real-time PCR coupled with HRMA to identify NTM and evaluate the accuracy of this method.Clinical isolates not clinical specimens are used for differentiation and identification of mycobacteria in most cases, but mycobacterial cultivation usually takes several weeks to form enough colonies. The period of diagnosis and treatment would be shortened if clinical specimens are directly used for mycobacterial detection. The prevalence and incidence of TB is much higher than those of non-tuberculosis mycobacterial diseases, so the detection of MTC in clinical specimens is more important. However, currently available methods which detect MTC in clinical specimens are not accurate enough for quantitative analysis. For example, real-time PCR could not accurately quantitate MTC because multi-copy gene is used as target gene. gyrB is a single-copy gene which evolutes at a moderate speed and it is suitable for quantitative analysis. Part3of our study aimed to develop a gyrB-basod Taqman real-time PCR for rapid and accurate detection of MTC in clinical specimens and evaluate the performance of this method.MethodsPart1:First of all,16reference strains of16mycobacterium species and95clinical mycobacterial isolates were subcultured, which were identified by biochemical method. Then these strains were differentiated by colloidal gold-based immunochromatographic assay. These strains were also differentiated by duplex real-time PCR after DNA extraction. For those MTC isolates which were falsely differentiated, mpt64gene was amplified and sequenced. Finally, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) of the two methods were evaluated respectively by using biochemical method as the gold standard. The difference and the agreement between the two methods were also evaluated. Reference strains included Mycobacterium abscessus ATCC19977, Mycobacterium aurum ATCC23366, Mycobacterium avium ATCC25921, Mycobacterium chelonae ATCC35752, Mycobacterium fortuitum ATCC6841, Mycobacterium gordonae ATCC14470, Mycobacterium intracellulare ATCC13950, Mycobacterium kansasii ATCC12478, Mycobacterium neoaurum ATCC25795, Mycobacterium nonchromogenicum ATCC19530, Mycobacterium parafortuitum ATCC19686, Mycobacterium scrofulaceum ATCC19981, Mycobacterium smegmatis ATCC19420, Mycobacterium triviale ATCC23292, Mycobacterium bovis (M.bovis) ATCC19210, Mycobacterium tuberculosis (MTB) H37Rv。 Clinical mycobacterial isolates included40NTM clinical isolates and55MTC clinical isolates.Part2:First of all,14reference strains of14mycobacterium species (14NTM reference strains in part1) and35clinical NTM isolates were subcultured, which were identified by biochemical method. Then these strains were identified by simplex real-time PCR coupled with HRMA after DNA extraction and purification, rpoB gene of which was amplified and sequenced. Finally, the accuracy of this method was calculated in comparison with the biochemical method.35clinical NTM isolates included10Mycobacterium abscessus,1Mycobacterium arupense,1Mycobacterium avium,1Mycobacterium chelonae,4Mycobacterium gordonae,10Mycobacterium intracellulare,5Mycobacterium kansasii,1Mycobacterium parescrofulaceum,1Mycobacterium szulgai,1Mycobacterium ulcerans.Part3:First of all,18reference strains of18bacterial species and100clinical mycobacterial isolates were subcultured, which were identified by biochemical method. Then these strains were detected by in-house Taqman real-time qPCR after DNA extraction, and the basic performance of this method was evaluated by using the biochemical method as gold standard. Finally,50clinical specimens were detected by commercial Taqman real-time qPCR and in-house Taqman real-time qPCR respectively, and qualitative and quantitative performance of the latter were evaluated in comparison with the former, and the difference and the agreement between the two methods were also evaluated. Eighteen reference strains included Mycobacterium abscessus ATCC19977, Mycobacterium aichiense ATCC27280, Mycobacterium aurum ATCC23366, Mycobacterium avium ATCC25921, Mycobacterium chelonae ATCC35752, Mycobacterium gordonae ATCC14470, Mycobacterium intracellulare ATCC13950, Mycobacterium kansasii ATCC12478, Mycobacterium nonchromogenicum ATCC19530, Mycobacterium scrofulaceum ATCC19981, Mycobacterium smegmatis ATCC19420, Mycobacterium triviale ATCC23292, Mycobacterium africanum ATCC25420, MTB H37Rv, Satphylococcus aureus ATCC25923, Klebsiella pneumoniae ATCC700603, Escherichia coli ATCC25922, Pseudomonas aeruginosa ATCC27853. One hundred clinical mycobacterial isolates included52clinical MTC isolates and48clinical NTM isolates.ResultsPart1:Fourteen NTM reference strains were negative and differentiated as NTM, and two MTC reference strains were positive and differentiated as MTC by colloidal gold-based immunochromatographic assay. Of95clinical mycobacterial isolates,48clinical MTB isolates and5clinical M.bovis isolates were positive and identified as MTC, and2clinical MTB isolates and40clinical NTM isolates were negative and identified as NTM by colloidal gold-based immunochromatographic assay. In comparison with biochemical method as the gold standard method, sensitivity, specificity, PPV and NPV of colloidal gold-based immunochromatographic assay were96.4%,100%,100%and95.2%respectively. All16reference strains of16mycobacterium species were correctly differentiated by duplex real-time PCR. Among95clinical mycobacterial isolates,50clinical MTB isolates and5clinical M.bovis isolates were differentiated as MTC because their melt curve peaks were at88.5～89.5℃, and40clinical NTM isolates were differentiated as NTM because their melt curve peaks were at86～88℃. Sensitivity, specificity, PPV and NPV of duplex real-time PCR were all100%by using biochemical method as the gold standard. There was no significant difference between colloidal gold-based immunochromatographic assay and duplex real-time PCR (χ2=0.500,P=0.500), and there was an excellent agreement between the two methods (κ=0.957, P<0.001). Both of the two MTB clinical isolates, which were negative with colloidal gold-based immunochromatographic assay, harboured a deletion of63nucleotides from codon43to codon63in the sequence of mpt64gene.Part2:These clinical NTM isolates which belonged to the same mycobacterium species had completely different HMRA profiles. However,14NTM species could be divided into5groups according to the Tm (Melting temperature) values. For Mycobacterium chelonae and Mycobacterium abscessus (group1), the Tm values were between89.15～89.4℃. For Mycobacterium neoaurum, Mycobacterium fortuitum and Mycobacterium parafortuitum (group2), the Tm values were between89.45～89.65℃. Group3consisted of Mycobacterium avium, Mycobacterium kansasii, Mycobacterium nonchromogenicum, Mycobacterium aurum and Mycobacterium smegmatis, the Tm values of which were between89.75～90.05℃. For Mycobacterium intracellulare, Mycobacterium scrofulaceum and Mycobacterium triviale, the Tm values were between90.25～90.5℃. The Tm value of Mycobacterium gordonae (group5) was between90.55～90.75℃. According to the Tm values,27of35clinical NTM isolates were correctly distinguished to group level and4could not be identified to group level for the lack of corresponding reference strains, so the accuracy of simplex real-time PCR coupled with Tm value analysis was77.14%. Of35clinical NTM isolates,15harboured1-3mutated bases in rpoB gene.Part3:Among18reference strains,12NTM reference strains and4reference strains of conventional bacteria were negative, and two MTC reference strains were positive by in-house Taqman real-time qPCR. Of100clinical mycobacterial isolates,52clinical MTC isolates were positive and other48clinical NTM isolates were negative by in-house Taqman real-time qPCR. The analytic sensitivity of this method was3×101CFU/ml and the analytic specificity was100%. The equation of standard curve was Y=-3.365×log(X)+45.6, and R2was0.999and PCR efficiency was98.217%. For4concentrations of DNA solution (3×106～3×103CFU/ml), intra-coefficients of variability (CV) of cycle threshold (Ct) values were0.61%,0.60%,0.13%and0.40%respectively, and inter-CV were3.4%,0.29%,0.47%and0.95%respectively. Sensitivity, specificity, PPV and NPV were85.7%,94.4%,85.7%and94.4%respectively by using commercial Taqman real-time qPCR as the gold standard. For qualitative analysis of MTC in clinical specimens, there was no significant difference between the two methods (χ2=0.250, P=1.000), and there was a good agreement between them (κ=0.802, P<0.001). In-house Taqman real-time qPCR was more accurate than commercial Taqman real-time qPCR for quantitative analysis of MTC in clinical specimens.Conclusions Part1:Both colloidal gold-based immunochromatographic assay and duplex real-time PCR showed excellent performance for rapid differentiation between MTC and NTM. There was no statistical significance between the two methods and there was an excellent agreement between them. Therefore, both of them could be routinely applied for mycobacterial differentiation in clinical laboratory. Clinical MTC isolates may be incorrectly differentiated by colloidal gold-based immunochromatographic assay due to mpt64gene mutation. Duplex real-time PCR is only suitable for central laboratories because it needs real-time qPCR instrument. The choice between the two methods is mainly dependent on the laboratory conditions and the sample size.Part2:As some clinical NTM isolates haboured rpoB mutations, these isolates could not be rapidly identified by simplex real-time PCR coupled with HRMA. However,14NTM species could be divided into5groups according to Tm value analysis. Nevertheless, NTM could not be identified to species level by this method and the accuracy is moderate. This method is needed to be improved and optimized.Part3:The Taqman real-time qPCR developed in our study had higher analytic sensitivity compared with those methods using CFU/volume concentrations. Moreover, this method had higher or similar analytic specificity compared with other real-time qPCR. Thus, this method had excellent analytic sensitivity and analytic specificity. The Taqman real-time qPCR developed in our study had better reproducibility than other real-time qPCR used for MTC detection and its PCR efficiency was between90%-110%. Therefore, this method had excellent reproducibility and amplification efficiency. The Taqman real-time qPCR developed in our study had good sensitivity and specificity like most real-time qPCR used for MTC detection in clinical specimens. As this method adopted single-copy gene, rpoB gene as target gene, it was more accurate than those methods using IS6110as target sequence theoretically. CV value from our method was obviously lower than that from IS6110-based methods, when5bacterial suspensions with the same concentration were detected by our method and commercial IS6110-based real-time qPCR respectively. It demonstrated that the Taqman real-time qPCR was more accurate than other real-time qPCR methods for quantitative analysis of MTC in clinical specimens. |
