Pathology (April 2015) 47(3), pp. 250–256
MOLECULAR DIAGNOSTICS IN MICROBIOLOGY
Molecular diagnostics for tuberculosis K. M. NOOR1,2, L. SHEPHARD1
AND
I. BASTIAN1
1Mycobacterium Reference Laboratory, SA Pathology, Adelaide, SA, Australia; 2Sultan Abdul Halim Hospital,
Kedah, Malaysia
Summary The phenotypic methods of smear microscopy, culture and indirect drug susceptibility testing (DST) remain the ‘gold standard’ diagnostics for tuberculosis (TB) in 2015. However, this review demonstrates that genotypic methods are in the ascendancy. Current-generation nucleic acid amplification tests (NAATs) are important supplementary tests for the rapid direct detection of (multidrug-resistant) TB in specific clinical settings. Genotypic detection is already the preferred method of detecting rifampicin and pyrazinamide resistance. Nextgeneration NAATs able to detect about 10 colony forming units/mL of sputum could replace culture as the initial test for detecting TB. Whole genome sequencing could also plausibly replace phenotypic DST but much work is required in method standardisation, database development and elucidation of all resistance gene determinants. The challenge then will be to rollout these increasingly complex and expensive diagnostics in the low-income countries where TB is prevalent. Key words: DNA, diagnosis, molecular diagnostic techniques, sequence analysis, tuberculosis. Received 24 November, revised 2 December 2014, accepted 7 January 2015
INTRODUCTION An estimated 9.0 million new cases of active tuberculosis (TB) disease occurred worldwide in 2013.1 Early detection and treatment is crucial in reducing the resulting 1.5 million annual deaths.1 The incident cases in 2013 included an estimated 480,000 patients with multidrug-resistant (MDR) TB, of whom about 9.0% had extensively drug-resistant (XDR) TB.1 MDRTB is defined as resistance to at least isoniazid and rifampicin, the two key drugs in ‘short course chemotherapy’ for TB. XDRTB is defined as MDRTB with resistance to a fluoroquinolone and one of the injectable second-line agents (amikacin, kanamycin or capreomycin), fluoroquinolones and an injectable agent being the key drugs in an MDRTB treatment regimen. Molecular diagnosis of TB has advanced rapidly over the past two decades in an attempt to aid early detection of TB and MDRTB. Point of care tests (POCT) and assays with the ability to detect antibiotic resistance could have major impacts in lowand middle-income countries with high burdens of (MDR)TB. This review will focus on molecular assays that detect Mycobacterium tuberculosis directly in patient specimens, genotypic tests for drug resistance in M. tuberculosis, and the burgeoning applications of whole genome sequencing (WGS) in the TB field. Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000232
#
DIRECT DETECTION TESTS FOR M. TUBERCULOSIS The introduction of liquid culture for M. tuberculosis has improved tremendously the turnaround time for the laboratory diagnosis from weeks to only 10–14 days. To further reduce this turnaround time, several molecular detection methods were introduced to detect M. tuberculosis directly in the clinical samples. Currently, there are two methods which are endorsed by the World Health Organization (WHO) to be used on patient samples: the Xpert MTB/RIF assay (Cepheid, USA) and line probe assays (LPAs). Another method, TB loop-mediated isothermal amplification (LAMP), is awaiting WHO approval. These methods plus several other polymerase chain reaction (PCR) assays are reviewed below. The sensitivities and specificities of each method are outlined in Table 1 and a summary of all methods is provided in Table 2. XpertMTB/RIF assay for direct detection The assay uses real-time (rt) PCR technology to detect TB and rifampicin resistance concurrently using unprocessed clinical specimens, regardless of their smear status. The target DNA sequence is the 81 bp core region of the bacterial RNA polymerase (rpoB) gene which encodes the active site of the enzyme.2 More than 95% of all rifampicin-resistant strains contain mutations localised within this region. In addition, the rpoB core region is flanked by M. tuberculosis-specific DNA sequences. Therefore, it is possible to test for M. tuberculosis and rifampicin resistance simultaneously.2 The assay utilises molecular beacon technology to detect DNA sequences amplified by a hemi-nested rt-PCR assay.2 Extraction, amplification and detection processes occur in an automated closed cartridge system. The assay may be used as a POCT since the system is self-contained, fully-integrated, automated, requires minimal expertise, and the result is available in 2 hours. Previous nucleic acid amplification tests (NAATs) have usually been expensive, technically demanding, restricted to centralised laboratories, at risk of cross contamination, and not generally recommended in smear-negative clinical specimens.3 Xpert is able to address many of these issues except for the high cost. Line probe assays In 2008, LPAs became the first molecular method endorsed by WHO for detection of M. tuberculosis and drug resistance from smear-positive patients at risk of MDRTB. Line probe assays may be used for the diagnosis of TB, speciation of nontuberculous mycobacteria (NTM) and drug resistance
2015 Royal College of Pathologists of Australasia
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS FOR TUBERCULOSIS
Table 1
251
Sensitivity and specificity of nucleic acid tests in clinical specimens Smear positive pulmonary Sensitivity
Xpert2 LPA45 In-house NAAT6
98–100 93.4
Amplicor6 AMTD6 LAMP9
97 92–100 92.1
Smear negative pulmonary
Specificity
Sensitivity
Extrapulmonary
Specificity
Sensitivity
Specificity
>99 NA
53–95 NA
98–99.6 NA
>95 >95 98.3
27–28 93 NA
>95 >95 NA
>98 57–83 85.6 NA Sensitivity: 84–100 Specificity: 83–100 >95 40–93 >95 40–93 98.3 53.8
AMTD, amplified mycobacterium tuberculosis direct test; LAMP, loop-mediated isothermal amplification; LPA, line probe assay; NA, not applicable; NAAT, nucleic acid amplification test.
detection, and are based on the reverse hybridisation principle.4,5 Specific oligonucleotides are immobilised at known locations on a membrane strip and are hybridised under strictly controlled conditions with the biotin-labelled PCR product. The hybrids formed are detected colorimetrically. Commercially-available LPAs include the INNO-LiPA Mycobacteria (Inno-genetics, Belgium) and the GenoType MTBC (Hain Lifesciences, Germany) for mycobacterial species identification and differentiation within the M. tuberculosis complex, respectively.6 The MDRTBplus assay (Hain Lifesciences) allows direct detection of M. tuberculosis, isoniazid and rifampicin resistance from smear-positive pulmonary specimens. LPAs are designed for use in reference and intermediate-tier laboratories and can be manual or semi-automated. The manual version is quite laborious with an advertised turnaround time of 6 hours but is usually 1–2 days in most ‘real world’ settings. Like all open-system PCR assays, there is a risk of cross contamination. Separated laboratories for DNA extraction, amplification and analysis are required, as is exemplary technique.6 Other nucleic acid amplification-based techniques In-house NAATs use different targets, either DNA or RNA, genus or species specific, followed by a detection step performed by various formats. One of the most commonly used targets for identification of M. tuberculosis is the insertion sequence IS6110.6 There have been reports of false-negative results due to absence or very low numbers of IS6110 in certain strains.7,8 A meta-analysis and meta-regression of multiple published studies found the use of IS6110 as the amplification target together with nested-PCR techniques were associated with a higher diagnostic accuracy.6 Considerable expertise is Table 2
needed to run most in-house NAATs and there is a risk of contamination, limiting their use to the reference laboratory only. NAATs are also commercially available such as the COBAS Taqman MTB test (Roche Diagnostics, Switzerland) and the m2000 RealTime MTB assay (Abbott Molecular, USA), which cater for high-throughput testing in the reference laboratory. Both systems involve automated extraction, reaction preparation and manual transfer to their amplification platforms. The tests are meant for decontaminated and concentrated smear-positive respiratory samples. Loop-mediated isothermal amplification technique This assay is based on autocycling strand displacement DNA synthesis using the large fragment of DNA polymerase. The test targets the gyrB gene and the IS6110 insertion sequence. The main characteristic of LAMP is the ability to synthesise large amounts of DNA.6 The whole procedure is carried out in a single tube with the isothermal reaction held at 638C for an incubation time of 1 hour with a visual detection. With the exception of a water bath or heating block, no other laboratory equipment is necessary,6 therefore the platform is suitable for resource-poor settings.5 The LoopAMP MTBC detection kit (Eiken Chemical Company, Japan) was launched in Japan and was assessed by a WHO expert group in 2013 but was not endorsed due to insufficient evidence.8 Presently, a larger evaluation is in progress in 14 countries. Indications for performing direct molecular detection tests for TB WHO has strongly recommended the Xpert MTB/RIF assay as the initial diagnostic test in adults and children suspected of having MDRTB or HIV-associated TB.10 The assay is also
Summary of direct molecular detection tests
Test
Location
Line probe assay Modular NAAT (Xpert)
Reference/intermediate Reference/intermediate point-of-care Reference/intermediate Reference lab Reference/intermediate
In house NAAT Automated batched PCR LAMP
Throughput
Function
Complexity
Hardware cost
Cost/test
lab lab;
Moderate Low/Moderate
MTBC/NTM diagnosis DST Diagnosis and DST MTBC
Moderate Low
Moderate High
Moderate Moderate
lab
High/Moderate High/Moderate Moderate
MTBC/NTM diagnosis MTBC diagnosis MTBC diagnosis
High High Moderate
High High Moderate
Moderate Low Low
lab
Adapted from UNITAID, World Health Organization.3 DST, drug susceptibility testing; LAMP, loop-mediated isothermal amplification; MTBC, Mycobacterium tuberculosis complex; NAAT, nucleic acid amplification test; NTM, non-tuberculous mycobacterium; PCR, polymerase chain reaction.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
252
NOOR et al.
conditionally recommended as the initial diagnostic test in all adults and children suspected of having TB (acknowledging resource implications and very low-quality evidence).10 Furthermore, the assay was cleared for use on extra-pulmonary specimens, particularly CSF and lymph node specimens based on limited evidence.10 The United States (US) Centers for Disease Control have gone further in a low-burden setting recommending NAAT ‘. . .on at least one (preferably the first) respiratory specimen from each patient suspected of pulmonary TB. . .for whom the test result would alter case management or TB control activities’.11 In contrast to these aggressive US recommendations, the National Institute for Health and Care Excellence (NICE) in the United Kingdom (UK) advised that all clinical samples from TB suspects be sent for automated liquid culture.12 They limited NAAT testing on primary specimens to the following four circumstances: ‘if rapid confirmation of a TB diagnosis in a sputum smear-positive person would alter their care’; before conducting a large contact-tracing initiative; for rapid rifampicin-resistance detection in an MDRTB suspect; or on smearpositive biopsy material that has been placed mistakenly in formalin. Using UK data and prices, Hughes et al. performed a cost-effectiveness analysis of various diagnostic strategies involving smear microscopy, NAAT and culture that supported the NICE guidelines.13 The cost effective strategy at a threshold of £20,000 per quality adjusted life year (QALY) was smear microscopy followed by culture routinely. A full work-up of microscopy, NAAT and culture became cost-effective as the TB prevalence increased. The performance and cost-effectiveness of the Xpert MTB/RIF assay have been studied mainly in high-incidence low-income countries. Sohn et al. reported using the assay in a university hospital TB clinic in Montreal, Canada, mainly screening (asymptomatic) migrants with abnormal chest X-rays.14 They found the Xpert MTB/RIF to have a lower sensitivity 46% (specificity 100%) than previously reported and limited impact on time-to-diagnosis and -treatment.14 The study findings were affected by the study population and the use of non-concentrated induced sputa. The current generation of NAATs remain supplemental (not replacement) tests because culture detects 10 or fewer colony forming units (CFU)/mL of sputum whereas the estimated level of detection (LOD) for the Xpert MTB/RIF assay is 131 (106– 176) CFU/mL, which is among the more sensitive NAATs.2 Two companies, Cepheid and Abbott Molecular, have just announced next-generation NAATs with reported LODs equivalent to culture.15,16 These new assays will be truly ‘game changing’ if this test performance is confirmed in laboratory evaluations and field trials. NAATs could replace culture for the initial detection of M. tuberculosis in clinical specimens. Will an isolate even be required for phenotypic susceptibility testing?
GENOTYPIC TESTS FOR DRUG RESISTANCE IN M. TUBERCULOSIS Resistance gene mutations Some resistance genes have been well characterised such as rpoB (encoding the RNA polymerase b subunit) for rifampicin, and inhA (encoding enoyl-ACP reductase) and katG (encoding the catalase–peroxidase enzyme) for isoniazid. The 81 bp rifampicin resistance determining region (RRDR) of the rpoB gene accounts for more than 95% of all rifampicin resistance.17
Pathology (2015), 47(3), April
InhA and katG have been estimated to account for 70–90% of isoniazid resistance, however the proportion varies substantially in different countries. In a study of 96 isoniazid-resistant isolates in Myanmar, inhA and katG mutations accounted for 2.1% and 65.6% of resistance, respectively.18 Codons 90, 91 and 94 in the gyrA gene (encoding the DNA gyrase) account for 71% of the fluoroquinolone mutations as demonstrated in a study by van Doorn et al.19 EmbB codon 306 mutations are found in 30–68% of ethambutol resistant strains.20,21 Resistance genes for these and other drugs are still being identified and the mechanisms of some anti-TB agents are not well understood, including pyrazinamide-resistance mutations in the pncA gene (encoding pyrazinamidase).22 Resistance detection methodologies Commercial and in-house methodologies including GeneXpert, LPAs and gene sequencing are used to detect drug resistance. Rapid performance of molecular methods has proven a valuable supplement to phenotypic DST. However, as commercial assays do not cover all mutations, their performance may vary in different geographical settings and may require local evaluation. Xpert MTB/RIF assay for rifampicin resistance detection The Xpert MTB/RIF assay has been described above. The assay detects up to 99.5% of all rifampicin-resistant strains.10,23 Rifampicin resistance serves as a surrogate marker for MDR allowing the Xpert assay to have direct clinical impact. The advantages of the Xpert assay are summarised in Table 2. Following the assay’s endorsement by WHO in December 2010, 110 high-burden and low/middle-income countries have procured 3553 GeneXpert instruments and 8.8 million Xpert MTB/RIF cartridges under concessional pricing as of 30 September 2014.24 Line probe assays Commercially-available LPAs for the genotypic detection of drug resistance are available from Hain Lifescience (Germany), Inno-genetics (Belgium) and Nipro Corporation (Japan). LPAs have many advantages over conventional culture and phenotypic DST. These characteristics have been described above and in Table 2. Genotype MTBDRplus (Hain Lifescience) is useful in detecting rifampicin resistance and both low- and high-level isoniazid resistance. Mutations are detected in the rpoB gene, inhA and katG. Drobniewski et al. reported in seven studies the MTBDRplus assay for isoniazid resistance detection had a pooled sensitivity of 77% (95% CI 69–83%), and specificity of 99% (95% CI 97–100%).25 Although the assay is a useful tool to detect isoniazid resistance, isoniazid-resistant TB cannot be excluded.25 The pooled sensitivity for rifampicin resistance detection was 97% (95%CI 92–99) and specificity of 98% (95%CI 95–99).25 For the detection of MTRTB using both rifampicin and isoniazid the sensitivity was 92% (95%CI 75–99) and the pooled specificity was 99% (95%CI).25 The review concluded that the level of performance would allow the test to rule out MDRTB and supplement phenotypic DST. Genotype MTBDRsl assay (Hain Lifescience) can detect fluoroquinolone resistance by the identification of genetic mutations in the gyrA gene.26 For the aminoglycosides kanamycin, amikacin and the cyclic peptide capreomycin, mutations in the rrs gene are detected. For ethambutol
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS FOR TUBERCULOSIS
resistance the embB gene is examined for mutations. The MTBDRsl assay is the only ‘rapid’ commercial test for the molecular detection of resistance to second line drugs (advertised turnaround time of 6 hours) but is subject to the same ‘real world’ practicalities as other LPAs with an effective turnaround time of 1–2 days. In 2012 a WHO expert group reviewed 11 published and seven unpublished studies to evaluate the MTBDRsl assay for direct testing on clinical samples and indirectly on M. tuberculosis isolates.26 The report published in 2013 concluded the MTBDRsl assay has a high test specificity and moderate sensitivity (see Table 3). The assay could rule-in XDRTB but had insufficient performance to rule-out XDRTB and could not replace phenotypic DST.26 Hillemann et al. reported that while the assay is easily integrated into routine workflow, a local evaluation would be required to determine the prevalence of mutations which may affect assay performance.27 Low sensitivity of ethambutol resistance (38.5%) is a particular limitation of the assay.27 Pyrazinamide is important for both first- and second-line drug treatment regimens. Nipro Corporation (Japan) has produced a new LPA containing four strips that identify M. tuberculosis resistance to pyrazinamide by detecting mutations in the pncA gene, in addition to detecting rifampicin, isoniazid and levofloxacin resistance and speciation of four mycobacteria.28 The strips can be used directly on sputum or on clinical isolates. The NTM/MDR-TB strip is designed to identify M. tuberculosis, M. avium, M. intracellulare and M. kansasii and detect mutations associated with rifampicin and isoniazid in M. tuberculosis. The isoniazid, pyrazinamide and fluoroquinolone strips allow for detection of mutations in inhA, katG, pncA and gyrA genes.28 An evaluation by Mitarai et al. of the Nipro LPA showed a sensitivity and specificity of 89.7% and 96.0%, respectively, for pyrazinamide resistance on M. tuberculosis isolates.28 Sensitivity and specificity respectively for levofloxacin resistance was 93% and 100%, for rifampicin resistance was 90.6% and 100%, and for isoniazid resistance was 89.7% and 96.0%.28 Overall the Nipro LPA demonstrated excellent performance and with the advantage of detecting resistance to pyrazinamide and fluoroquinolone.
In-house amplification and sequencing In-house amplification and sequencing of resistance genes provides greater flexibility than commercial assays. Laboratories may routinely sequence inhA promoter region and katG (for isoniazid resistance), embB (for ethambutol resistance), pncA (for pyrazinamide resistance), gyrA (for fluoroquinolone resistance), rrs (for kanamycin, amikacin, and capreomycin), tlyA (associated with capreomycin resistance), and eis (promoter region associated with kanamycin resistance). Sensitivity and specificities will vary depending on geographical location and the genes targeted. Table 3
253
Sequencing shares some limitations of the commercial assays but addresses others. Sequencing may identify mutations which are known to be associated with resistance to a drug, however not all mutations and resistance mechanisms are known and there may be insufficient data to link the mutation to resistance. Another limitation of sequencing is that the mutation may lie outside the sequenced region. Mutations outside of the rpoB 81 bp region do exist but are uncommon. For example, low level rifampicin-resistance at codon 572Phe is outside the ‘hotspot’. This problem can be resolved by sequencing the entire rpoB gene.29 Practical laboratory considerations When a laboratory is choosing which molecular method to integrate into current work practices, many factors need to be considered (Table 4). The selection will depend upon the testing workload, whether direct testing from respiratory specimens and/or indirect investigation of M. tuberculosis isolates are to be performed, the skill set of the staff, and the available instrumentation including the use of generic thermocyclers and semi-automation for high throughput. Discordant phenotypic and genotypic results At present, phenotypic DST is considered the ‘gold standard’ for all anti-TB drugs (excluding WHO Groups IV and V drugs). However, for rifampicin susceptibility, molecular testing is supplanting phenotypic DST as the ‘gold standard’ providing much faster and accurate results. Discrepant rifampicin results which are phenotypically susceptible (usually by automated broth-based MGIT 960 system; Becton Dickinson, USA) and genotypically resistant have been reported by Rigouts et al. and Van Deun et al.29,30 Broth-based testing misidentifies isolates with low-level rifampicin resistance as susceptible. Mutations 511Pro, 516Tyr, 526Asn, 533Pro, 572Phe located at the ends or outside the rpoB ‘hotspot’ region have been associated with low-level resistance and poor clinical outcomes equivalent to rifampicin-resistant isolates where phenotypic and genotypic testing concord.29–31 The proportion of strains with low-level rifampicin resistance is around 10–13% for treatment failure cases,29 but has been documented at up to 22% in a Hong Kong study evaluating isolates from predominantly new cases.32 Silent mutations also occur where a mutation results in a base change but no change to the amino acid sequence and so the isolate retains susceptibility to rifampicin. However, the mutation is recognised by molecular testing. Silent mutations occur at a low rate (12 SNPs.35 They validated these predictions by sequencing 217 isolates from 168 patients involved in 11 community clusters identified by mycobacterial interspersed repetitive-unit–variable-number tandem repeat (MIRU-VNTR) genotyping. The strengths of WGS were: the ability to detect hidden outbreaks in populations where contact tracing is problematic (e.g., substance abusers); unparalleled discrimination allowing isolates with matching MIRU-VNTR profiles to be excluded from outbreak investigations by the presence of numerous SNP differences; and identification of potential ‘super-spreaders’ by inspection of the phylogenetic trees generated by WGS. Numerous studies confirm that WGS represents a significant advance over previous genotyping methods such as MIRUVNTR and IS6110 restriction fragment length polymorphism (RFLP) analysis.38,39 Unfortunately, the latency and low mutation rate of M. tuberculosis may confound even WGS! Bryant et al. estimated an average mutation rate of 0.3 SNPs per genome per year but with wide variation between epidemiologically-linked pairs.37 Without a ‘regular’ molecular clock, WGS cannot confidently confirm transmission between patients. This problem is evident even in the seminal study from Birmingham and Leicester.35 In six genotypic clusters, including one involving an African community, an epidemiological link could be found in only 25 (24%) of 103 patients. The homogeneity and low mutation rate of M. tuberculosis could prove even more problematic for WGS analyses in highprevalence high-transmission settings. A study of 1000 isolates from Samara, Russia, confirms this concern.40 In only 20 of the 35 clusters did patients with identical isolates even live in the same region. Patients with identical strains lived up to 136 km apart! Short SNP differences may not always imply direct transmission. Further complicating the use of WGS for epidemiological purposes is the increasing recognition of mixed infections with different strains of M. tuberculosis, and even a ‘cloud of diversity’ within the same strain in a single patient. For example, Sun et al. performed WGS on seven serial isolates
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS FOR TUBERCULOSIS
collected from three patients at different stages of treatment for TB or MDRTB.41 They detected up to four to five different resistant mutants in a single sputum sample. While WGS may not be an ideal genotypic marker, it is a major advance with improved discrimination that will supplant alternative methods such as MIRU-VNTR. However, some technical issues must be addressed before WGS can be widely adopted. The WGS methodologies (e.g., chemistries, read lengths, coverage, sample type) and analytical pipelines (e.g., reference genome used, filters, result reporting) must be harmonised. Whole genome studies also generate enormous datasets. To address this problem, Kohl et al. have proposed a core genome multilocus sequencing typing (cgMLST) scheme that extends MLST typing from a handful of genes to the genome level.42 They defined a standard set of 3041 genes and validated cgMLST against SNP-based WGS using 26 isolates from an outbreak in Germany. Core genome MLST may provide a standardised portable scheme requiring less computing resources. Antibiotic resistance studies for M. tuberculosis This review has already summarised various molecular assays that detect a limited number of antibiotic resistances (e.g., GeneXpert, LPA). In a single test that is becoming faster and more affordable, WGS can interrogate all antibiotic resistance and accessory genes (termed the ‘resistome’) carried by an organism. Koser et al. provided a ‘proof of concept’ performing WGS on the broth culture (positive at day 3 of incubation) from a patient with XDRTB.43 A mixed infection with two distantlyrelated Beijing strains was detected and 39 resistance genes mapped. This information could be available to clinicians within days compared with the weeks required for primary culture and indirect DST. This concept was extended in the enormous WGS study of 1000 M. tuberculosis isolates from Samara, Russia, mentioned earlier.40 Resistance genes were annotated and compensatory mechanisms restoring fitness studied. Forty-eight percent of all isolates had an MDRTB genotype; 16% had an XDR genotype. Beijing strains were noted to carry: mutations associated with no fitness cost (e.g., isoniazid resistance conferred by a katG mutation encoding p.Ser315Thr); mutations associated with compensatory mechanisms (e.g., rifampicin resistance conferred by an rpoB mutation encoding pSer450Leu associated with rpoC mutations that restore fitness); and kanamycin resistance mutations in the eis gene associated with increased virulence. Ultimately, WGS may prove more useful for resistance detection than for epidemiological investigations (for which WGS grabbed initial notoriety). Some WGS advocates even foresee resistome investigations replacing DST with scaleddown phenotypic testing continuing only to detect new resistances and as a quality control tool for the resistome databases.34 Beforehand, all of the WGS standardisation mentioned previously must occur plus a comprehensive list of resistanceconferring determinants collated.34,40 These determinants need to be validated against phenotypic testing and, more importantly, against patient outcome data. This validation will be particularly problematic for complex resistance mechanisms, such as for pyrazinamide where diverse mutations and occasional deletions that may (or may not) confer resistance are spread across the length of the pncA gene.22 Finally, this molecular, phenotypic and patient information will need to be entered in standardised formats into enormous well-curated
255
unified international data warehouses, which can be accessed securely through a web portal.44
CONCLUSION While phenotypic methods currently remain the ‘gold standard’ TB diagnostics, this review has demonstrated that genotypic testing is gaining the ascendancy. Next-generation NAATs able to detect about 10 CFU/mL of sputum could replace culture as the initial test for detecting TB. Whole genome sequencing could also plausibly replace phenotypic DST but much work is required in method standardisation, database development and elucidation of all resistance gene determinants. International organisations have done a commendable job rolling-out 3553 GeneXpert instruments and 8.8 million Xpert MTB/RIF cartridges to low-income under-resourced countries where TB is prevalent.24 Even more ingenuity in financing and testing platforms will be required to roll-out increasingly complex and expensive tests such as WGS. Acknowledgements: The authors thank Richard Lumb for his constructive comments during manuscript preparation. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Dr I. Bastian, Microbiology Infectious Diseases Directorate, SA Pathology, PO Box 14 Rundle Mall, SA 5000, Australia. E-mail: [email protected]
References 1. World Health Organization. Global tuberculosis report 2014. Geneva: WHO, 2014. WHO/HTM/TB/2014.08. http://www.who.int/tb/publica tions/global_report/en/. 2. Lawn SD, Nicol MP. Xpert1MTB/RIF assay: development, evaluation and implementation of a new rapid molecular diagnostic for tuberculosis and rifampicin resistance. Future Microbiol 2011; 6: 1067–82. 3. UNITAID, World Health Organization. Tuberculosis: diagnostics technology and market landscape. 3rd edition. Geneva: WHO, 2014. http://www. unitaid.eu/images/marketdynamics/publications/UNITAID_TB_Diagnos tics_Landscape_3rd-edition.pdf. 4. World Heath Organization. Molecular line probe assays for rapid screening of patients at risk of multidrug-resistant tuberculosis (MDR-TB) – policy statement. Geneva: WHO, 2008. http://www.who.int/tb/features_archive/ policy_statement.pdf. 5. Rossau R, Traore H, De Beenhouwer H, et al. Evaluation of the INNO-LiPA Rif. TB assay, a reverse hybridization assay for the simultaneous detection of Mycobacterium tuberculosis complex and its resistance to rifampin. Antimicrob Agents Chemother 1997; 10: 2093–8. 6. Palomino JC. Molecular detection, identification and drug resistance detection in Mycobacterium tuberculosis. FEMS Immunol Med Microbiol 2009; 56: 103–11. 7. Yuen LKW, Ross BC, Jackson KM, Dwyer B. Characterization of Mycobacterium tuberculosis strains from Vietnamese patients by Southern blot hybridization. J Clin Microbiol 1993; 31: 1615–8. 8. Sanker S, Ramamuthy M, Nandagopal B, et al. An appraisal of PCR-based technology in the detection of Mycobacterium tuberculosis. Mol Diag Ther 2011; 15: 1–11. 9. Ou X, Li Q, Xia H, et al. Diagnostic accuracy of the PURE-LAMP Test for pulmonary tuberculosis at the county-level laboratory in China. PLoS One 2014; 9: e74544. 10. World Health Organization. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Policy update. Geneva: WHO, 2014. WHO/HTM/TB/2013.16. http://apps.who.int/iris/bitstream/ 10665/112472/1/9789241506335_eng.pdf?ua ¼ 1. 11. Centers for Disease Control. Updated guidelines for the use of nucleic acid amplification tests in the diagnosis of tuberculosis. MMWR 2009; 58: 7–10. 12. National Institute for Health and Care Excellence. Tuberculosis – clinical diagnosis and management of tuberculosis, and measures for its prevention
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
256
13. 14. 15.
16.
17. 18.
19.
20.
21. 22. 23.
24. 25. 26.
27.
NOOR et al.
and control. London: NICE, 2011. http://www.nice.org.uk/guidance/cg117/ resources/guidance-tuberculosis-pdf. Hughes R, Wonderling D, Li B, Higgins B. The cost effectiveness of nucleic acid amplification techniques for the diagnosis of tuberculosis. Respir Med 2012; 106: 300–7. Sohn H, Aero AD, Menzies D, et al. Xpert MTB/RIF testing in a low tuberculosis incidence, high-resource setting: limitations in accuracy and clinical impact. Clin Infect Dis 2014; 58: 970–6. Cepheid. Press release. Cepheid, FIND & Rutgers announce collaboration for next-generation innovations to game-changing Xpert MTB/RIF test. 28 Oct 2014, cited Nov 2014. http://www.finddiagnostics.org/resource-centre/ press/141028.html. Tang N, Frank A, Pahalawatta V, et al. The new Abbott realtime MTB assay for the detection of for the detection of Mycobacterium tuberculosis. Late breaker session, 45th Union World Conference on Lung Health, Barcelona, 28 October-1 November 2014. Telenti A, Imboden P, Marchesi F, et al. Detection of rifampicin-resistance mutations in Mycobacterium tuberculosis. Lancet 1993; 341: 647–50. Valvatne H, Syre H, Kross M, et al. Isoniazid and rifampicin resistanceassociated mutations in Mycobacterium tuberculosis isolates from Yangon, Myanmar: implications for rapid molecular testing. J Antimicrob Chemother 2009; 64: 694–701. van Doorn HR, An DD, de Jong MD, et al. Fluoroquinolone resistance detection in Mycobacterium tuberculosis with locked nucleic acid probe real-time PCR. Int J Tuberc Lung Dis 2008; 12: 736–42. Plinke C, Ru¨sch-Gerdes S, Niemann S. Significance of mutations in embB codon 306 for prediction of ethambutol resistance in clinical Mycobacterium tuberculosis isolates. Antimicrob Agents Chemother 2006; 50: 1900–2. Ahmad S, Jaber AA, Mokaddas E. Frequency of embB codon 306 mutations in ethambutol-susceptible and -resistant clinical Mycobacterium tuberculosis isolates in Kuwait. Tuberculosis 2007; 87: 123–9. Miotto P, Cabibbe AM, Feuerriegel S, et al. Mycobacterium tuberculosis pyrazinamide resistance determinants: a multicenter study. MBio 2014; 5: e01819–4. World Health Organization. Xpert MTB/RIF implementation manual: technical and operational ‘how-to’; practical considerations. Geneva, WHO 2014. WHO/HTM/TB/2014.1. http://apps.who.int/iris/bitstream/ 10665/112469/1/9789241506700_eng.pdf. World Health Organization. Factsheet Tuberculosis Diagnostics Xpert MTB/RIF Test. Geneva: WHO, 2014. http://www.who.int/tb/publications/ Xpert_factsheet.pdf?ua ¼ 1. Drobniewski F, Nikolayevskyy V, Balabanova Y, et al. Diagnosis of tuberculosis and drug resistance: what can new tools bring us? Int J Tuberc Lung Dis 2012; 16: 860–70. World Health Organization. The use of molecular line probe assay for the detection of resistance to second line anti tuberculosis drugs - 2013 Expert Group Meeting Report, Geneva. Geneva: WHO, 2013. WHO/HTM/TB/ 2013.01 http://apps.who.int/iris/bitstream/10665/78099/1/WHO_HTM_ TB_2013.01.eng.pdf. Hillemann D, Ru¨sch-Gerdes S, Richter E. Feasibility of the GenoType MTBDRsl Assay for fluoroquinolone, amikacin-capreomycin, and ethambutol resistance testing of Mycobacterium tuberculosis strains and clinical specimens. J Clin Microbiol 2009; 46: 1767–72.
Pathology (2015), 47(3), April
28. Mitarai S, Kato S, Ogata H, et al. Comprehensive multicenter evaluation of a new line probe assay kit for the identification of Mycobacterium species and detection of drug-resistant Mycobacterium tuberculosis. J Clin Microbiol 2012; 50: 884–90. 29. Van Deun A, Aung KJM, Bola V, et al. Rifampin drug resistance tests for tuberculosis: challenging the gold standard. J Clin Microbiol 2013; 51: 2633–40. 30. Rigouts L, Mourad Gumusboga AB, Willem Bram de Rijk A, et al. Rifampin resistance missed in automated liquid culture system for Mycobacterium tuberculosis isolates with specific rpoB mutations. J Clin Microbiol 2013; 51: 2641–5. 31. Williamson DA, Roberts SA, Bower JE, et al. Clinical failures associated with rpoB mutations in phenotypically occult multidrug-resistant Mycobacterium tuberculosis. Int J Tuberc Lung Dis 2011; 16: 216–20. 32. Yip CW, Leung KL, Wong D, et al. Denaturing HPLC for high throughput screening of rifampicin- resistant Mycobacterium tuberculosis. Int J Tuberc Lung Dis 2006; 10: 625–30. 33. Chang KC, Yew WW, Zhang Y. Pyrazinamide susceptibility testing in Mycobacterium tuberculosis: a systematic review with meta-analyses. Antimicrob Agents Chemother 2011; 55: 4499–505. 34. Didelot X, Bowden R, Wilson DJ, et al. Transforming clinical microbiology with bacterial genome sequencing. Nat Genet Rev 2012; 13: 601–12. 35. Walker TM, Ip CLC, Harrell RH, et al. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect Dis 2013; 13: 137–46. 36. Ford CB, Lin PL, Chase MR, et al. Use of whole genome sequencing to estimate the mutation rate of Mycobacterium tuberculosis during latent infection. Nat Genet 2011; 43: 482–6. 37. Bryant JM, Schurch AC, van Deutekom H, et al. Inferring patient to patient transmission of Mycobacterium tuberculosis from whole genome sequencing data. BMC Infect Dis 2013; 13: 110. 38. Gardy JL, Johnston JC, Ho Sui SJ, et al. Whole-genome sequencing and social network analysis of a tuberculosis outbreak. N Engl J Med 2011; 364: 730–9. 39. Jamieson FB, Teatero S, Guthrie JL, et al. Whole-genome sequencing of the Mycobacterium tuberculosis Manila sublineage results in less clustering and better resolution than mycobacterial interspersed repetitive-unit– variable-number tandem repeat (MIRU-VNTR) typing and spoligotyping. J Clin Microbiol 2014; 52: 3795–8. 40. Casali N, Nikolayevskyy V, Balabanova Y, et al. Evolution and transmission of drug-resistant tuberculosis in a Russian population. Nat Genet 2014; 46: 279–86. 41. Sun G, Luo T, Yang C, et al. Dynamic population changes in Mycobacterium tuberculosis during acquisition and fixation of drug resistance in patients. J Infect Dis 2012; 206: 1724–33. 42. Kohl TA, Diel R, Harmsen D, et al. Whole-genome-based Mycobacterium tuberculosis surveillance: a standardized, portable, and expandable approach. J Clin Microbiol 2014; 52: 2479–86. 43. Koser CU, Bryant JM, Becq J, et al. Whole-genome sequencing for rapid susceptibility testing of M. tuberculosis. N Engl J Med 2013; 369: 290–2. 44. Aviles E. Requirements for managing "big data": a global database of DR molecular sequence data. Int J Tuberc Lung Dis 2014; 18 (Suppl 1): S48–9. 45. Zar H, Udwadia Z. Advances in tuberculosis 2011-2012. Thorax 2013; 68: 283–7.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS IN MICROBIOLOGY
Molecular diagnostics for tuberculosis K. M. NOOR1,2, L. SHEPHARD1
AND
I. BASTIAN1
1Mycobacterium Reference Laboratory, SA Pathology, Adelaide, SA, Australia; 2Sultan Abdul Halim Hospital,
Kedah, Malaysia
Summary The phenotypic methods of smear microscopy, culture and indirect drug susceptibility testing (DST) remain the ‘gold standard’ diagnostics for tuberculosis (TB) in 2015. However, this review demonstrates that genotypic methods are in the ascendancy. Current-generation nucleic acid amplification tests (NAATs) are important supplementary tests for the rapid direct detection of (multidrug-resistant) TB in specific clinical settings. Genotypic detection is already the preferred method of detecting rifampicin and pyrazinamide resistance. Nextgeneration NAATs able to detect about 10 colony forming units/mL of sputum could replace culture as the initial test for detecting TB. Whole genome sequencing could also plausibly replace phenotypic DST but much work is required in method standardisation, database development and elucidation of all resistance gene determinants. The challenge then will be to rollout these increasingly complex and expensive diagnostics in the low-income countries where TB is prevalent. Key words: DNA, diagnosis, molecular diagnostic techniques, sequence analysis, tuberculosis. Received 24 November, revised 2 December 2014, accepted 7 January 2015
INTRODUCTION An estimated 9.0 million new cases of active tuberculosis (TB) disease occurred worldwide in 2013.1 Early detection and treatment is crucial in reducing the resulting 1.5 million annual deaths.1 The incident cases in 2013 included an estimated 480,000 patients with multidrug-resistant (MDR) TB, of whom about 9.0% had extensively drug-resistant (XDR) TB.1 MDRTB is defined as resistance to at least isoniazid and rifampicin, the two key drugs in ‘short course chemotherapy’ for TB. XDRTB is defined as MDRTB with resistance to a fluoroquinolone and one of the injectable second-line agents (amikacin, kanamycin or capreomycin), fluoroquinolones and an injectable agent being the key drugs in an MDRTB treatment regimen. Molecular diagnosis of TB has advanced rapidly over the past two decades in an attempt to aid early detection of TB and MDRTB. Point of care tests (POCT) and assays with the ability to detect antibiotic resistance could have major impacts in lowand middle-income countries with high burdens of (MDR)TB. This review will focus on molecular assays that detect Mycobacterium tuberculosis directly in patient specimens, genotypic tests for drug resistance in M. tuberculosis, and the burgeoning applications of whole genome sequencing (WGS) in the TB field. Print ISSN 0031-3025/Online ISSN 1465-3931 DOI: 10.1097/PAT.0000000000000232
#
DIRECT DETECTION TESTS FOR M. TUBERCULOSIS The introduction of liquid culture for M. tuberculosis has improved tremendously the turnaround time for the laboratory diagnosis from weeks to only 10–14 days. To further reduce this turnaround time, several molecular detection methods were introduced to detect M. tuberculosis directly in the clinical samples. Currently, there are two methods which are endorsed by the World Health Organization (WHO) to be used on patient samples: the Xpert MTB/RIF assay (Cepheid, USA) and line probe assays (LPAs). Another method, TB loop-mediated isothermal amplification (LAMP), is awaiting WHO approval. These methods plus several other polymerase chain reaction (PCR) assays are reviewed below. The sensitivities and specificities of each method are outlined in Table 1 and a summary of all methods is provided in Table 2. XpertMTB/RIF assay for direct detection The assay uses real-time (rt) PCR technology to detect TB and rifampicin resistance concurrently using unprocessed clinical specimens, regardless of their smear status. The target DNA sequence is the 81 bp core region of the bacterial RNA polymerase (rpoB) gene which encodes the active site of the enzyme.2 More than 95% of all rifampicin-resistant strains contain mutations localised within this region. In addition, the rpoB core region is flanked by M. tuberculosis-specific DNA sequences. Therefore, it is possible to test for M. tuberculosis and rifampicin resistance simultaneously.2 The assay utilises molecular beacon technology to detect DNA sequences amplified by a hemi-nested rt-PCR assay.2 Extraction, amplification and detection processes occur in an automated closed cartridge system. The assay may be used as a POCT since the system is self-contained, fully-integrated, automated, requires minimal expertise, and the result is available in 2 hours. Previous nucleic acid amplification tests (NAATs) have usually been expensive, technically demanding, restricted to centralised laboratories, at risk of cross contamination, and not generally recommended in smear-negative clinical specimens.3 Xpert is able to address many of these issues except for the high cost. Line probe assays In 2008, LPAs became the first molecular method endorsed by WHO for detection of M. tuberculosis and drug resistance from smear-positive patients at risk of MDRTB. Line probe assays may be used for the diagnosis of TB, speciation of nontuberculous mycobacteria (NTM) and drug resistance
2015 Royal College of Pathologists of Australasia
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS FOR TUBERCULOSIS
Table 1
251
Sensitivity and specificity of nucleic acid tests in clinical specimens Smear positive pulmonary Sensitivity
Xpert2 LPA45 In-house NAAT6
98–100 93.4
Amplicor6 AMTD6 LAMP9
97 92–100 92.1
Smear negative pulmonary
Specificity
Sensitivity
Extrapulmonary
Specificity
Sensitivity
Specificity
>99 NA
53–95 NA
98–99.6 NA
>95 >95 98.3
27–28 93 NA
>95 >95 NA
>98 57–83 85.6 NA Sensitivity: 84–100 Specificity: 83–100 >95 40–93 >95 40–93 98.3 53.8
AMTD, amplified mycobacterium tuberculosis direct test; LAMP, loop-mediated isothermal amplification; LPA, line probe assay; NA, not applicable; NAAT, nucleic acid amplification test.
detection, and are based on the reverse hybridisation principle.4,5 Specific oligonucleotides are immobilised at known locations on a membrane strip and are hybridised under strictly controlled conditions with the biotin-labelled PCR product. The hybrids formed are detected colorimetrically. Commercially-available LPAs include the INNO-LiPA Mycobacteria (Inno-genetics, Belgium) and the GenoType MTBC (Hain Lifesciences, Germany) for mycobacterial species identification and differentiation within the M. tuberculosis complex, respectively.6 The MDRTBplus assay (Hain Lifesciences) allows direct detection of M. tuberculosis, isoniazid and rifampicin resistance from smear-positive pulmonary specimens. LPAs are designed for use in reference and intermediate-tier laboratories and can be manual or semi-automated. The manual version is quite laborious with an advertised turnaround time of 6 hours but is usually 1–2 days in most ‘real world’ settings. Like all open-system PCR assays, there is a risk of cross contamination. Separated laboratories for DNA extraction, amplification and analysis are required, as is exemplary technique.6 Other nucleic acid amplification-based techniques In-house NAATs use different targets, either DNA or RNA, genus or species specific, followed by a detection step performed by various formats. One of the most commonly used targets for identification of M. tuberculosis is the insertion sequence IS6110.6 There have been reports of false-negative results due to absence or very low numbers of IS6110 in certain strains.7,8 A meta-analysis and meta-regression of multiple published studies found the use of IS6110 as the amplification target together with nested-PCR techniques were associated with a higher diagnostic accuracy.6 Considerable expertise is Table 2
needed to run most in-house NAATs and there is a risk of contamination, limiting their use to the reference laboratory only. NAATs are also commercially available such as the COBAS Taqman MTB test (Roche Diagnostics, Switzerland) and the m2000 RealTime MTB assay (Abbott Molecular, USA), which cater for high-throughput testing in the reference laboratory. Both systems involve automated extraction, reaction preparation and manual transfer to their amplification platforms. The tests are meant for decontaminated and concentrated smear-positive respiratory samples. Loop-mediated isothermal amplification technique This assay is based on autocycling strand displacement DNA synthesis using the large fragment of DNA polymerase. The test targets the gyrB gene and the IS6110 insertion sequence. The main characteristic of LAMP is the ability to synthesise large amounts of DNA.6 The whole procedure is carried out in a single tube with the isothermal reaction held at 638C for an incubation time of 1 hour with a visual detection. With the exception of a water bath or heating block, no other laboratory equipment is necessary,6 therefore the platform is suitable for resource-poor settings.5 The LoopAMP MTBC detection kit (Eiken Chemical Company, Japan) was launched in Japan and was assessed by a WHO expert group in 2013 but was not endorsed due to insufficient evidence.8 Presently, a larger evaluation is in progress in 14 countries. Indications for performing direct molecular detection tests for TB WHO has strongly recommended the Xpert MTB/RIF assay as the initial diagnostic test in adults and children suspected of having MDRTB or HIV-associated TB.10 The assay is also
Summary of direct molecular detection tests
Test
Location
Line probe assay Modular NAAT (Xpert)
Reference/intermediate Reference/intermediate point-of-care Reference/intermediate Reference lab Reference/intermediate
In house NAAT Automated batched PCR LAMP
Throughput
Function
Complexity
Hardware cost
Cost/test
lab lab;
Moderate Low/Moderate
MTBC/NTM diagnosis DST Diagnosis and DST MTBC
Moderate Low
Moderate High
Moderate Moderate
lab
High/Moderate High/Moderate Moderate
MTBC/NTM diagnosis MTBC diagnosis MTBC diagnosis
High High Moderate
High High Moderate
Moderate Low Low
lab
Adapted from UNITAID, World Health Organization.3 DST, drug susceptibility testing; LAMP, loop-mediated isothermal amplification; MTBC, Mycobacterium tuberculosis complex; NAAT, nucleic acid amplification test; NTM, non-tuberculous mycobacterium; PCR, polymerase chain reaction.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
252
NOOR et al.
conditionally recommended as the initial diagnostic test in all adults and children suspected of having TB (acknowledging resource implications and very low-quality evidence).10 Furthermore, the assay was cleared for use on extra-pulmonary specimens, particularly CSF and lymph node specimens based on limited evidence.10 The United States (US) Centers for Disease Control have gone further in a low-burden setting recommending NAAT ‘. . .on at least one (preferably the first) respiratory specimen from each patient suspected of pulmonary TB. . .for whom the test result would alter case management or TB control activities’.11 In contrast to these aggressive US recommendations, the National Institute for Health and Care Excellence (NICE) in the United Kingdom (UK) advised that all clinical samples from TB suspects be sent for automated liquid culture.12 They limited NAAT testing on primary specimens to the following four circumstances: ‘if rapid confirmation of a TB diagnosis in a sputum smear-positive person would alter their care’; before conducting a large contact-tracing initiative; for rapid rifampicin-resistance detection in an MDRTB suspect; or on smearpositive biopsy material that has been placed mistakenly in formalin. Using UK data and prices, Hughes et al. performed a cost-effectiveness analysis of various diagnostic strategies involving smear microscopy, NAAT and culture that supported the NICE guidelines.13 The cost effective strategy at a threshold of £20,000 per quality adjusted life year (QALY) was smear microscopy followed by culture routinely. A full work-up of microscopy, NAAT and culture became cost-effective as the TB prevalence increased. The performance and cost-effectiveness of the Xpert MTB/RIF assay have been studied mainly in high-incidence low-income countries. Sohn et al. reported using the assay in a university hospital TB clinic in Montreal, Canada, mainly screening (asymptomatic) migrants with abnormal chest X-rays.14 They found the Xpert MTB/RIF to have a lower sensitivity 46% (specificity 100%) than previously reported and limited impact on time-to-diagnosis and -treatment.14 The study findings were affected by the study population and the use of non-concentrated induced sputa. The current generation of NAATs remain supplemental (not replacement) tests because culture detects 10 or fewer colony forming units (CFU)/mL of sputum whereas the estimated level of detection (LOD) for the Xpert MTB/RIF assay is 131 (106– 176) CFU/mL, which is among the more sensitive NAATs.2 Two companies, Cepheid and Abbott Molecular, have just announced next-generation NAATs with reported LODs equivalent to culture.15,16 These new assays will be truly ‘game changing’ if this test performance is confirmed in laboratory evaluations and field trials. NAATs could replace culture for the initial detection of M. tuberculosis in clinical specimens. Will an isolate even be required for phenotypic susceptibility testing?
GENOTYPIC TESTS FOR DRUG RESISTANCE IN M. TUBERCULOSIS Resistance gene mutations Some resistance genes have been well characterised such as rpoB (encoding the RNA polymerase b subunit) for rifampicin, and inhA (encoding enoyl-ACP reductase) and katG (encoding the catalase–peroxidase enzyme) for isoniazid. The 81 bp rifampicin resistance determining region (RRDR) of the rpoB gene accounts for more than 95% of all rifampicin resistance.17
Pathology (2015), 47(3), April
InhA and katG have been estimated to account for 70–90% of isoniazid resistance, however the proportion varies substantially in different countries. In a study of 96 isoniazid-resistant isolates in Myanmar, inhA and katG mutations accounted for 2.1% and 65.6% of resistance, respectively.18 Codons 90, 91 and 94 in the gyrA gene (encoding the DNA gyrase) account for 71% of the fluoroquinolone mutations as demonstrated in a study by van Doorn et al.19 EmbB codon 306 mutations are found in 30–68% of ethambutol resistant strains.20,21 Resistance genes for these and other drugs are still being identified and the mechanisms of some anti-TB agents are not well understood, including pyrazinamide-resistance mutations in the pncA gene (encoding pyrazinamidase).22 Resistance detection methodologies Commercial and in-house methodologies including GeneXpert, LPAs and gene sequencing are used to detect drug resistance. Rapid performance of molecular methods has proven a valuable supplement to phenotypic DST. However, as commercial assays do not cover all mutations, their performance may vary in different geographical settings and may require local evaluation. Xpert MTB/RIF assay for rifampicin resistance detection The Xpert MTB/RIF assay has been described above. The assay detects up to 99.5% of all rifampicin-resistant strains.10,23 Rifampicin resistance serves as a surrogate marker for MDR allowing the Xpert assay to have direct clinical impact. The advantages of the Xpert assay are summarised in Table 2. Following the assay’s endorsement by WHO in December 2010, 110 high-burden and low/middle-income countries have procured 3553 GeneXpert instruments and 8.8 million Xpert MTB/RIF cartridges under concessional pricing as of 30 September 2014.24 Line probe assays Commercially-available LPAs for the genotypic detection of drug resistance are available from Hain Lifescience (Germany), Inno-genetics (Belgium) and Nipro Corporation (Japan). LPAs have many advantages over conventional culture and phenotypic DST. These characteristics have been described above and in Table 2. Genotype MTBDRplus (Hain Lifescience) is useful in detecting rifampicin resistance and both low- and high-level isoniazid resistance. Mutations are detected in the rpoB gene, inhA and katG. Drobniewski et al. reported in seven studies the MTBDRplus assay for isoniazid resistance detection had a pooled sensitivity of 77% (95% CI 69–83%), and specificity of 99% (95% CI 97–100%).25 Although the assay is a useful tool to detect isoniazid resistance, isoniazid-resistant TB cannot be excluded.25 The pooled sensitivity for rifampicin resistance detection was 97% (95%CI 92–99) and specificity of 98% (95%CI 95–99).25 For the detection of MTRTB using both rifampicin and isoniazid the sensitivity was 92% (95%CI 75–99) and the pooled specificity was 99% (95%CI).25 The review concluded that the level of performance would allow the test to rule out MDRTB and supplement phenotypic DST. Genotype MTBDRsl assay (Hain Lifescience) can detect fluoroquinolone resistance by the identification of genetic mutations in the gyrA gene.26 For the aminoglycosides kanamycin, amikacin and the cyclic peptide capreomycin, mutations in the rrs gene are detected. For ethambutol
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS FOR TUBERCULOSIS
resistance the embB gene is examined for mutations. The MTBDRsl assay is the only ‘rapid’ commercial test for the molecular detection of resistance to second line drugs (advertised turnaround time of 6 hours) but is subject to the same ‘real world’ practicalities as other LPAs with an effective turnaround time of 1–2 days. In 2012 a WHO expert group reviewed 11 published and seven unpublished studies to evaluate the MTBDRsl assay for direct testing on clinical samples and indirectly on M. tuberculosis isolates.26 The report published in 2013 concluded the MTBDRsl assay has a high test specificity and moderate sensitivity (see Table 3). The assay could rule-in XDRTB but had insufficient performance to rule-out XDRTB and could not replace phenotypic DST.26 Hillemann et al. reported that while the assay is easily integrated into routine workflow, a local evaluation would be required to determine the prevalence of mutations which may affect assay performance.27 Low sensitivity of ethambutol resistance (38.5%) is a particular limitation of the assay.27 Pyrazinamide is important for both first- and second-line drug treatment regimens. Nipro Corporation (Japan) has produced a new LPA containing four strips that identify M. tuberculosis resistance to pyrazinamide by detecting mutations in the pncA gene, in addition to detecting rifampicin, isoniazid and levofloxacin resistance and speciation of four mycobacteria.28 The strips can be used directly on sputum or on clinical isolates. The NTM/MDR-TB strip is designed to identify M. tuberculosis, M. avium, M. intracellulare and M. kansasii and detect mutations associated with rifampicin and isoniazid in M. tuberculosis. The isoniazid, pyrazinamide and fluoroquinolone strips allow for detection of mutations in inhA, katG, pncA and gyrA genes.28 An evaluation by Mitarai et al. of the Nipro LPA showed a sensitivity and specificity of 89.7% and 96.0%, respectively, for pyrazinamide resistance on M. tuberculosis isolates.28 Sensitivity and specificity respectively for levofloxacin resistance was 93% and 100%, for rifampicin resistance was 90.6% and 100%, and for isoniazid resistance was 89.7% and 96.0%.28 Overall the Nipro LPA demonstrated excellent performance and with the advantage of detecting resistance to pyrazinamide and fluoroquinolone.
In-house amplification and sequencing In-house amplification and sequencing of resistance genes provides greater flexibility than commercial assays. Laboratories may routinely sequence inhA promoter region and katG (for isoniazid resistance), embB (for ethambutol resistance), pncA (for pyrazinamide resistance), gyrA (for fluoroquinolone resistance), rrs (for kanamycin, amikacin, and capreomycin), tlyA (associated with capreomycin resistance), and eis (promoter region associated with kanamycin resistance). Sensitivity and specificities will vary depending on geographical location and the genes targeted. Table 3
253
Sequencing shares some limitations of the commercial assays but addresses others. Sequencing may identify mutations which are known to be associated with resistance to a drug, however not all mutations and resistance mechanisms are known and there may be insufficient data to link the mutation to resistance. Another limitation of sequencing is that the mutation may lie outside the sequenced region. Mutations outside of the rpoB 81 bp region do exist but are uncommon. For example, low level rifampicin-resistance at codon 572Phe is outside the ‘hotspot’. This problem can be resolved by sequencing the entire rpoB gene.29 Practical laboratory considerations When a laboratory is choosing which molecular method to integrate into current work practices, many factors need to be considered (Table 4). The selection will depend upon the testing workload, whether direct testing from respiratory specimens and/or indirect investigation of M. tuberculosis isolates are to be performed, the skill set of the staff, and the available instrumentation including the use of generic thermocyclers and semi-automation for high throughput. Discordant phenotypic and genotypic results At present, phenotypic DST is considered the ‘gold standard’ for all anti-TB drugs (excluding WHO Groups IV and V drugs). However, for rifampicin susceptibility, molecular testing is supplanting phenotypic DST as the ‘gold standard’ providing much faster and accurate results. Discrepant rifampicin results which are phenotypically susceptible (usually by automated broth-based MGIT 960 system; Becton Dickinson, USA) and genotypically resistant have been reported by Rigouts et al. and Van Deun et al.29,30 Broth-based testing misidentifies isolates with low-level rifampicin resistance as susceptible. Mutations 511Pro, 516Tyr, 526Asn, 533Pro, 572Phe located at the ends or outside the rpoB ‘hotspot’ region have been associated with low-level resistance and poor clinical outcomes equivalent to rifampicin-resistant isolates where phenotypic and genotypic testing concord.29–31 The proportion of strains with low-level rifampicin resistance is around 10–13% for treatment failure cases,29 but has been documented at up to 22% in a Hong Kong study evaluating isolates from predominantly new cases.32 Silent mutations also occur where a mutation results in a base change but no change to the amino acid sequence and so the isolate retains susceptibility to rifampicin. However, the mutation is recognised by molecular testing. Silent mutations occur at a low rate (12 SNPs.35 They validated these predictions by sequencing 217 isolates from 168 patients involved in 11 community clusters identified by mycobacterial interspersed repetitive-unit–variable-number tandem repeat (MIRU-VNTR) genotyping. The strengths of WGS were: the ability to detect hidden outbreaks in populations where contact tracing is problematic (e.g., substance abusers); unparalleled discrimination allowing isolates with matching MIRU-VNTR profiles to be excluded from outbreak investigations by the presence of numerous SNP differences; and identification of potential ‘super-spreaders’ by inspection of the phylogenetic trees generated by WGS. Numerous studies confirm that WGS represents a significant advance over previous genotyping methods such as MIRUVNTR and IS6110 restriction fragment length polymorphism (RFLP) analysis.38,39 Unfortunately, the latency and low mutation rate of M. tuberculosis may confound even WGS! Bryant et al. estimated an average mutation rate of 0.3 SNPs per genome per year but with wide variation between epidemiologically-linked pairs.37 Without a ‘regular’ molecular clock, WGS cannot confidently confirm transmission between patients. This problem is evident even in the seminal study from Birmingham and Leicester.35 In six genotypic clusters, including one involving an African community, an epidemiological link could be found in only 25 (24%) of 103 patients. The homogeneity and low mutation rate of M. tuberculosis could prove even more problematic for WGS analyses in highprevalence high-transmission settings. A study of 1000 isolates from Samara, Russia, confirms this concern.40 In only 20 of the 35 clusters did patients with identical isolates even live in the same region. Patients with identical strains lived up to 136 km apart! Short SNP differences may not always imply direct transmission. Further complicating the use of WGS for epidemiological purposes is the increasing recognition of mixed infections with different strains of M. tuberculosis, and even a ‘cloud of diversity’ within the same strain in a single patient. For example, Sun et al. performed WGS on seven serial isolates
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
MOLECULAR DIAGNOSTICS FOR TUBERCULOSIS
collected from three patients at different stages of treatment for TB or MDRTB.41 They detected up to four to five different resistant mutants in a single sputum sample. While WGS may not be an ideal genotypic marker, it is a major advance with improved discrimination that will supplant alternative methods such as MIRU-VNTR. However, some technical issues must be addressed before WGS can be widely adopted. The WGS methodologies (e.g., chemistries, read lengths, coverage, sample type) and analytical pipelines (e.g., reference genome used, filters, result reporting) must be harmonised. Whole genome studies also generate enormous datasets. To address this problem, Kohl et al. have proposed a core genome multilocus sequencing typing (cgMLST) scheme that extends MLST typing from a handful of genes to the genome level.42 They defined a standard set of 3041 genes and validated cgMLST against SNP-based WGS using 26 isolates from an outbreak in Germany. Core genome MLST may provide a standardised portable scheme requiring less computing resources. Antibiotic resistance studies for M. tuberculosis This review has already summarised various molecular assays that detect a limited number of antibiotic resistances (e.g., GeneXpert, LPA). In a single test that is becoming faster and more affordable, WGS can interrogate all antibiotic resistance and accessory genes (termed the ‘resistome’) carried by an organism. Koser et al. provided a ‘proof of concept’ performing WGS on the broth culture (positive at day 3 of incubation) from a patient with XDRTB.43 A mixed infection with two distantlyrelated Beijing strains was detected and 39 resistance genes mapped. This information could be available to clinicians within days compared with the weeks required for primary culture and indirect DST. This concept was extended in the enormous WGS study of 1000 M. tuberculosis isolates from Samara, Russia, mentioned earlier.40 Resistance genes were annotated and compensatory mechanisms restoring fitness studied. Forty-eight percent of all isolates had an MDRTB genotype; 16% had an XDR genotype. Beijing strains were noted to carry: mutations associated with no fitness cost (e.g., isoniazid resistance conferred by a katG mutation encoding p.Ser315Thr); mutations associated with compensatory mechanisms (e.g., rifampicin resistance conferred by an rpoB mutation encoding pSer450Leu associated with rpoC mutations that restore fitness); and kanamycin resistance mutations in the eis gene associated with increased virulence. Ultimately, WGS may prove more useful for resistance detection than for epidemiological investigations (for which WGS grabbed initial notoriety). Some WGS advocates even foresee resistome investigations replacing DST with scaleddown phenotypic testing continuing only to detect new resistances and as a quality control tool for the resistome databases.34 Beforehand, all of the WGS standardisation mentioned previously must occur plus a comprehensive list of resistanceconferring determinants collated.34,40 These determinants need to be validated against phenotypic testing and, more importantly, against patient outcome data. This validation will be particularly problematic for complex resistance mechanisms, such as for pyrazinamide where diverse mutations and occasional deletions that may (or may not) confer resistance are spread across the length of the pncA gene.22 Finally, this molecular, phenotypic and patient information will need to be entered in standardised formats into enormous well-curated
255
unified international data warehouses, which can be accessed securely through a web portal.44
CONCLUSION While phenotypic methods currently remain the ‘gold standard’ TB diagnostics, this review has demonstrated that genotypic testing is gaining the ascendancy. Next-generation NAATs able to detect about 10 CFU/mL of sputum could replace culture as the initial test for detecting TB. Whole genome sequencing could also plausibly replace phenotypic DST but much work is required in method standardisation, database development and elucidation of all resistance gene determinants. International organisations have done a commendable job rolling-out 3553 GeneXpert instruments and 8.8 million Xpert MTB/RIF cartridges to low-income under-resourced countries where TB is prevalent.24 Even more ingenuity in financing and testing platforms will be required to roll-out increasingly complex and expensive tests such as WGS. Acknowledgements: The authors thank Richard Lumb for his constructive comments during manuscript preparation. Conflicts of interest and sources of funding: The authors state that there are no conflicts of interest to disclose. Address for correspondence: Dr I. Bastian, Microbiology Infectious Diseases Directorate, SA Pathology, PO Box 14 Rundle Mall, SA 5000, Australia. E-mail: [email protected]
References 1. World Health Organization. Global tuberculosis report 2014. Geneva: WHO, 2014. WHO/HTM/TB/2014.08. http://www.who.int/tb/publica tions/global_report/en/. 2. Lawn SD, Nicol MP. Xpert1MTB/RIF assay: development, evaluation and implementation of a new rapid molecular diagnostic for tuberculosis and rifampicin resistance. Future Microbiol 2011; 6: 1067–82. 3. UNITAID, World Health Organization. Tuberculosis: diagnostics technology and market landscape. 3rd edition. Geneva: WHO, 2014. http://www. unitaid.eu/images/marketdynamics/publications/UNITAID_TB_Diagnos tics_Landscape_3rd-edition.pdf. 4. World Heath Organization. Molecular line probe assays for rapid screening of patients at risk of multidrug-resistant tuberculosis (MDR-TB) – policy statement. Geneva: WHO, 2008. http://www.who.int/tb/features_archive/ policy_statement.pdf. 5. Rossau R, Traore H, De Beenhouwer H, et al. Evaluation of the INNO-LiPA Rif. TB assay, a reverse hybridization assay for the simultaneous detection of Mycobacterium tuberculosis complex and its resistance to rifampin. Antimicrob Agents Chemother 1997; 10: 2093–8. 6. Palomino JC. Molecular detection, identification and drug resistance detection in Mycobacterium tuberculosis. FEMS Immunol Med Microbiol 2009; 56: 103–11. 7. Yuen LKW, Ross BC, Jackson KM, Dwyer B. Characterization of Mycobacterium tuberculosis strains from Vietnamese patients by Southern blot hybridization. J Clin Microbiol 1993; 31: 1615–8. 8. Sanker S, Ramamuthy M, Nandagopal B, et al. An appraisal of PCR-based technology in the detection of Mycobacterium tuberculosis. Mol Diag Ther 2011; 15: 1–11. 9. Ou X, Li Q, Xia H, et al. Diagnostic accuracy of the PURE-LAMP Test for pulmonary tuberculosis at the county-level laboratory in China. PLoS One 2014; 9: e74544. 10. World Health Organization. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children. Policy update. Geneva: WHO, 2014. WHO/HTM/TB/2013.16. http://apps.who.int/iris/bitstream/ 10665/112472/1/9789241506335_eng.pdf?ua ¼ 1. 11. Centers for Disease Control. Updated guidelines for the use of nucleic acid amplification tests in the diagnosis of tuberculosis. MMWR 2009; 58: 7–10. 12. National Institute for Health and Care Excellence. Tuberculosis – clinical diagnosis and management of tuberculosis, and measures for its prevention
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
256
13. 14. 15.
16.
17. 18.
19.
20.
21. 22. 23.
24. 25. 26.
27.
NOOR et al.
and control. London: NICE, 2011. http://www.nice.org.uk/guidance/cg117/ resources/guidance-tuberculosis-pdf. Hughes R, Wonderling D, Li B, Higgins B. The cost effectiveness of nucleic acid amplification techniques for the diagnosis of tuberculosis. Respir Med 2012; 106: 300–7. Sohn H, Aero AD, Menzies D, et al. Xpert MTB/RIF testing in a low tuberculosis incidence, high-resource setting: limitations in accuracy and clinical impact. Clin Infect Dis 2014; 58: 970–6. Cepheid. Press release. Cepheid, FIND & Rutgers announce collaboration for next-generation innovations to game-changing Xpert MTB/RIF test. 28 Oct 2014, cited Nov 2014. http://www.finddiagnostics.org/resource-centre/ press/141028.html. Tang N, Frank A, Pahalawatta V, et al. The new Abbott realtime MTB assay for the detection of for the detection of Mycobacterium tuberculosis. Late breaker session, 45th Union World Conference on Lung Health, Barcelona, 28 October-1 November 2014. Telenti A, Imboden P, Marchesi F, et al. Detection of rifampicin-resistance mutations in Mycobacterium tuberculosis. Lancet 1993; 341: 647–50. Valvatne H, Syre H, Kross M, et al. Isoniazid and rifampicin resistanceassociated mutations in Mycobacterium tuberculosis isolates from Yangon, Myanmar: implications for rapid molecular testing. J Antimicrob Chemother 2009; 64: 694–701. van Doorn HR, An DD, de Jong MD, et al. Fluoroquinolone resistance detection in Mycobacterium tuberculosis with locked nucleic acid probe real-time PCR. Int J Tuberc Lung Dis 2008; 12: 736–42. Plinke C, Ru¨sch-Gerdes S, Niemann S. Significance of mutations in embB codon 306 for prediction of ethambutol resistance in clinical Mycobacterium tuberculosis isolates. Antimicrob Agents Chemother 2006; 50: 1900–2. Ahmad S, Jaber AA, Mokaddas E. Frequency of embB codon 306 mutations in ethambutol-susceptible and -resistant clinical Mycobacterium tuberculosis isolates in Kuwait. Tuberculosis 2007; 87: 123–9. Miotto P, Cabibbe AM, Feuerriegel S, et al. Mycobacterium tuberculosis pyrazinamide resistance determinants: a multicenter study. MBio 2014; 5: e01819–4. World Health Organization. Xpert MTB/RIF implementation manual: technical and operational ‘how-to’; practical considerations. Geneva, WHO 2014. WHO/HTM/TB/2014.1. http://apps.who.int/iris/bitstream/ 10665/112469/1/9789241506700_eng.pdf. World Health Organization. Factsheet Tuberculosis Diagnostics Xpert MTB/RIF Test. Geneva: WHO, 2014. http://www.who.int/tb/publications/ Xpert_factsheet.pdf?ua ¼ 1. Drobniewski F, Nikolayevskyy V, Balabanova Y, et al. Diagnosis of tuberculosis and drug resistance: what can new tools bring us? Int J Tuberc Lung Dis 2012; 16: 860–70. World Health Organization. The use of molecular line probe assay for the detection of resistance to second line anti tuberculosis drugs - 2013 Expert Group Meeting Report, Geneva. Geneva: WHO, 2013. WHO/HTM/TB/ 2013.01 http://apps.who.int/iris/bitstream/10665/78099/1/WHO_HTM_ TB_2013.01.eng.pdf. Hillemann D, Ru¨sch-Gerdes S, Richter E. Feasibility of the GenoType MTBDRsl Assay for fluoroquinolone, amikacin-capreomycin, and ethambutol resistance testing of Mycobacterium tuberculosis strains and clinical specimens. J Clin Microbiol 2009; 46: 1767–72.
Pathology (2015), 47(3), April
28. Mitarai S, Kato S, Ogata H, et al. Comprehensive multicenter evaluation of a new line probe assay kit for the identification of Mycobacterium species and detection of drug-resistant Mycobacterium tuberculosis. J Clin Microbiol 2012; 50: 884–90. 29. Van Deun A, Aung KJM, Bola V, et al. Rifampin drug resistance tests for tuberculosis: challenging the gold standard. J Clin Microbiol 2013; 51: 2633–40. 30. Rigouts L, Mourad Gumusboga AB, Willem Bram de Rijk A, et al. Rifampin resistance missed in automated liquid culture system for Mycobacterium tuberculosis isolates with specific rpoB mutations. J Clin Microbiol 2013; 51: 2641–5. 31. Williamson DA, Roberts SA, Bower JE, et al. Clinical failures associated with rpoB mutations in phenotypically occult multidrug-resistant Mycobacterium tuberculosis. Int J Tuberc Lung Dis 2011; 16: 216–20. 32. Yip CW, Leung KL, Wong D, et al. Denaturing HPLC for high throughput screening of rifampicin- resistant Mycobacterium tuberculosis. Int J Tuberc Lung Dis 2006; 10: 625–30. 33. Chang KC, Yew WW, Zhang Y. Pyrazinamide susceptibility testing in Mycobacterium tuberculosis: a systematic review with meta-analyses. Antimicrob Agents Chemother 2011; 55: 4499–505. 34. Didelot X, Bowden R, Wilson DJ, et al. Transforming clinical microbiology with bacterial genome sequencing. Nat Genet Rev 2012; 13: 601–12. 35. Walker TM, Ip CLC, Harrell RH, et al. Whole-genome sequencing to delineate Mycobacterium tuberculosis outbreaks: a retrospective observational study. Lancet Infect Dis 2013; 13: 137–46. 36. Ford CB, Lin PL, Chase MR, et al. Use of whole genome sequencing to estimate the mutation rate of Mycobacterium tuberculosis during latent infection. Nat Genet 2011; 43: 482–6. 37. Bryant JM, Schurch AC, van Deutekom H, et al. Inferring patient to patient transmission of Mycobacterium tuberculosis from whole genome sequencing data. BMC Infect Dis 2013; 13: 110. 38. Gardy JL, Johnston JC, Ho Sui SJ, et al. Whole-genome sequencing and social network analysis of a tuberculosis outbreak. N Engl J Med 2011; 364: 730–9. 39. Jamieson FB, Teatero S, Guthrie JL, et al. Whole-genome sequencing of the Mycobacterium tuberculosis Manila sublineage results in less clustering and better resolution than mycobacterial interspersed repetitive-unit– variable-number tandem repeat (MIRU-VNTR) typing and spoligotyping. J Clin Microbiol 2014; 52: 3795–8. 40. Casali N, Nikolayevskyy V, Balabanova Y, et al. Evolution and transmission of drug-resistant tuberculosis in a Russian population. Nat Genet 2014; 46: 279–86. 41. Sun G, Luo T, Yang C, et al. Dynamic population changes in Mycobacterium tuberculosis during acquisition and fixation of drug resistance in patients. J Infect Dis 2012; 206: 1724–33. 42. Kohl TA, Diel R, Harmsen D, et al. Whole-genome-based Mycobacterium tuberculosis surveillance: a standardized, portable, and expandable approach. J Clin Microbiol 2014; 52: 2479–86. 43. Koser CU, Bryant JM, Becq J, et al. Whole-genome sequencing for rapid susceptibility testing of M. tuberculosis. N Engl J Med 2013; 369: 290–2. 44. Aviles E. Requirements for managing "big data": a global database of DR molecular sequence data. Int J Tuberc Lung Dis 2014; 18 (Suppl 1): S48–9. 45. Zar H, Udwadia Z. Advances in tuberculosis 2011-2012. Thorax 2013; 68: 283–7.
Copyright © Royal College of pathologists of Australasia. Unauthorized reproduction of this article is prohibited.
