Rapid Genotypic Detection of Rifampin- and Isoniazid-Resistant Mycobacterium tuberculosis Directly in Clinical Specimens

Abstract

A multiplex PCR DNA strip assay (Genotype MTBDR) designed to detect rifampin (rpoB) and high-level isoniazid (katG) resistance mutations in Mycobacterium tuberculosis isolates was optimized for clinical specimens. Successful genotypic results were achieved with 36 of 38 (95%) smear-positive respiratory specimens, allowing rapid therapeutic adjustments in transmittable drug-resistant tuberculosis.

Multidrug-resistant tuberculosis (MDR TB) is defined as Mycobacterium tuberculosis resistant to rifampin (RIF) and isoniazid (INH). The spread of MDR TB has escalated worldwide (8). Drug susceptibility testing (DST) of M. tuberculosis in clinical specimens is time-consuming. INH and RIF are crucial elements of the standard treatment regimen of tuberculosis, and resistance to these drugs requires extension of therapy (4). The vast majority of RIF resistance is caused by mutations located in the 81-bp region of the rpoB gene (22). INH resistance is more complex, as the mutations conferring resistance are located in several genes and loci. INH resistance has been associated mainly with mutations in katG, inhA, ahpC, and kasA (2, 13, 15, 18). Studies investigating strains with high-level INH resistance have documented that 50 to 100% have mutations located in codon 315 of the katG gene (9, 16).

Genotype MTBDR, a commercially available multiplex PCR DNA strip assay (Hain Lifescience, Nehren, Germany), is designed to simultaneously detect the most important rpoB and katG gene mutations conferring RIF and high-level INH resistance in isolates (10). It is based on the hybridization between rpoB and katG amplicons to membrane-bound probes. The DNA strip covers five rpoB wild-type probes, four rpoB mutant probes (with D516V, H526Y, H526D, and S531L mutations), one katG wild-type probe, and two katG mutant probes (with S315T1 and S315T2 mutations). The aim of this study was to optimize and evaluate the application of the assay directly in clinical specimens. Furthermore, we compared the DNA strip results with phenotypic DST results.

The study was carried out at the International Reference Laboratory of Mycobacteriology, Statens Serum Institut, Denmark. A total of 106 NaOH-N-acetyl-l-cysteine-pretreated clinical specimens submitted for M. tuberculosis examination were selected on the basis of susceptibility test result and acid-fast bacillus smear grade. A volume of surplus material from the selected specimens was collected. The respiratory specimens consisted of sputum, tracheal secretion, bronchoalveolar lavage fluid (n = 66), and pleural fluid (n = 4). The nonrespiratory specimens were obtained from lymph node (n = 12), gastric lavage fluid (n = 11), cerebrospinal/synovial fluid/urine (n = 6), and tissue biopsy samples (n = 7). None of the patients were registered in the Registry of Human Tissue Utilization, National Board of Health (prohibits tissue use for research). Of 90 culture-positive specimens, 62% (n = 56) were smear positive and 42% (n = 38) were RIF and/or INH resistant. All specimens were processed by conventional mycobacterial procedures as previously described (11). Smears were stained with auramine-rhodamine and examined by ×200 magnification (21). Specimens were incubated on Löwenstein-Jensen slants and MGIT (Becton Dickinson), and growth of Mycobacterium tuberculosis was confirmed by Inno-LiPA species assay (InnoGenetics). DST for RIF and INH was performed using a BACTEC 460 system (Becton Dickinson) according to the manufacturer's instructions (20). Critical drug concentrations were 2.0 μg/ml for RIF and 0.1 μg/ml for INH. In addition, INH was tested at 0.4 and 2.0 μg/ml. On request, strand displacement amplification (SDA) (Becton Dickinson) was performed as described elsewhere (12). Specimens were stored at −20°C prior to analysis.

Genotype MTBDR optimization.

An aliquot (500 μl) of pretreated specimen was centrifuged (10,000 × g, 15 min). The pellet was resuspended in sterile Milli-Q water (100 μl) and heat inactivated (100°C, 20 min). DNA was extracted by ultrasonication (60°C, 15 min) followed by centrifugation (10,000 × g, 5 min). We optimized the multiplex PCR targeting rpoB, and katG, and 23S rRNA with regard to sample volume (1, 3, and 5 μl), MgCl₂ concentration (1.0, 1.5, and 2.0 mM), and the number of PCR cycles (20 and 30). Following optimization, the multiplex PCR conditions for the rest of the study consisted of a 50-μl reaction volume, 35 μl primer-nucleotide mix (kit), 10× PCR buffer without MgCl₂, 3 μl of MgCl₂ stock solution (25 mM), 2 U of FastStart Taq DNA polymerase (Roche Applied Science), and 5 μl of extracted DNA. With a Gene Amp PCR 9600 system (Applied Biosystems), the amplification profile consisted of a denaturation step (95°C, 5 min), 10 cycles of 95°C for 30 s and 58°C for 2 min, followed by 30 cycles of 95°C for 25 s, 53°C for 40 s, and 70°C for 40 s, and a final extension of 70°C for 8 min. Amplicons of 200 bp (universal control), 250 bp (rpoB), and 120 bp (katG) on an agarose gel were found in 47 of 56 (84%) smear-positive specimens. Hybridization utilizing a preprogrammed Auto-LiPA system (InnoGenetics) and interpretation of the DNA strip (Fig. 1) were performed according to the manufacturer's instructions.

FIG. 1.

Representative DNA strip patterns obtained with the optimized Genotype MTBDR assay. The positions of the oligonucleotide probes are given on the left. The target genes and specific probe lines are shown from top to bottom as follows: conjugate control; universal amplification control (23S rRNA); M. tuberculosis complex-specific control (23S rRNA); control of rpoB amplification; five rpoB wild-type probes; four rpoB mutant probes with mutations in codon 516, 526, or 531; control of katG amplification; one katG codon 315 wild-type probe; and two katG probes with mutations in codon 315. An absent or weakened wild-type probe indicates resistance, and the corresponding mutation may be identified. Samples with different strip susceptibility patterns are shown on the right. Lane 1, drug susceptible (rpoB and katG wild type); lane 2, RIF susceptible, high-level INH resistant (rpoB wild type and katG S315T1 mutation); lane 3, MDR (rpoB H526Y and katG S315T1 mutations); lane 4, MDR (rpoB S531L and katG S315T1 mutations); lane 5, MDR (rpoB S531L mutation, rpoB D516V mutation, and an absent wild-type katG).

Interpretable DNA strip readings correlated with smear grade and were obtained for 36 of 38 (95%) respiratory and 15 of 18 (83%) nonrespiratory smear-positive specimens, regardless of DST (Table 1). Genotypic results were concordant with those found with BACTEC in 54 of 57 (95%) samples (Tables 1 and 2). Discordant results could be explained by the INH resistance mutations being located elsewhere than codon 315. All 10 MDR TB specimens had results concordant with BACTEC. Mutations and wild-type mismatch (katG) correlated in 17 of 20 (85%) samples with high-level INH resistance (MIC ≥ 0.4 μg/ml) as approved by CLSI (formerly NCCLS) standard M24-A (17). The S315T1 mutation was the most prevalent, accounting for 14 of 20 (70%) samples. Fourteen of 20 (70%) specimens with high-level INH resistance were also resistant at the higher concentration (2.0 μg/ml) tested. Similarly to our findings, high-level INH resistance (MIC > 2 μg/ml) has previously been associated with 89% of isolates with mutations in codon 315 (23). Samples with low-level INH resistance (0.1 μg/ml) had wild-type patterns, indicating that this mutation was located in another gene. As the modified assay is unable to determine the proportion of resistant bacteria and is limited to the detection of RIF and INH resistance mutations in smear-positive specimens, the assay cannot replace phenotypic DST, which remains the gold standard. In a few studies, mycobacteriophage-based assays for RIF resistance detection have also shown potential for direct application with sputum samples (1, 3, 5). However, the accuracy of these methods remains to be elucidated. Similarly to our findings, other PCR-based assays utilizing TaqMan, molecular beacons, and fluorescence resonance energy transfer probes (katG codon 315, inhA, and rpoB) demonstrated high sensitivities in smear-positive specimens (6, 7, 14, 19, 24).

TABLE 1.

Performance of the optimized assay with primary drug-susceptible and -resistant respiratory and nonrespiratory specimens

Smear resulta	M. tuberculosis culture resultb	SDA resultb	No. of specimens with interpretable DNA strip reading results/no. of tested specimens (%)
			RIF and INH susceptibled		RIF and/or INH resistantd
			Respiratory	Nonrespiratory	Respiratory	Nonrespiratory
4+	+	NDc	7/7	1/1	4/4	1/1
3+	+	ND	4/4	4/4	7/7	2/3 f
2+	+	ND	6/6	2/2	4/4	3/3
1+	+	ND	3/5	2/3	1/1	0/1
−	+	ND	2/6	1/2	0/8	2/6
−	+	+	0/5	1/3
−	+	−	0/1	0/3
−	−	+	2/6 e	0/2
−	−	−	0/6	0/2
All positive			20/22 (91)	9/10 (90)	16/16 (100)	6/8 (75)
All negative			4/24 (17)	2/12 (17)	0/8 (0)	2/6 (33)
Total			24/46 (52)	11/22 (50)	16/24 (67)	8/14 (57)

^aSmear grading for acid-fast bacilli was done according to an international standard. −, negative result.

^b+, positive result; −, negative result.

^cND, not done.

^dBy BACTEC 460.

^eInterpretable DNA strip reading with a culture-negative, SDA-positive sample, probably due to the presence of nonviable bacteria.

^fOne sample (gastric lavage fluid) failed to give an interpretable DNA strip result, as the M. tuberculosis complex-specific control (23S rRNA) was lacking.

TABLE 2.

Comparison of genotypic and phenotypic drug resistance results for respiratory and nonrespiratory specimens

Method	Drug (concn [μg/ml]) or target gene/probea	Phenotypic drug resistance pattern or interpretable genotypic rpoB or katG mutation pattern for no. of samples (n)b
		n = 3c	n = 4	n = 1	n = 2	n = 1d	n = 6	n = 3	n = 4
Phenotypic DST	Rifampin (2.0)	R	R	R	R	S	S	S	S
	Isoniazid (0.1)	R	R	R	R	R	R	R	R
	Isoniazid (0.4)	R	R	R	R	R	R	R	S
	Isoniazid (2.0)	R	R	S	S	R	R	S	S
rpoB probe assay	rpoB WT1	+	+	+	+	+	+	+	+
	rpoB WT2	+/−	+	+	+	+/−	+	+	+
	rpoB WT3	+	+	+	+	+	+	+	+
	rpoB WT4	+	+	−	+/−	+	+	+	+
	rpoB WT5	+/−	+/−	+	+/−	+	+	+	+
	rpoB D516V mutation	+	−	−	−	−	−	−	−
	rpoB H526Y mutation	−	−	+	+	−	−	−	−
	rpoB H526D mutation	−	−	−	−	−	−	−	−
	rpoB S531L mutation	+	+	−	−	−	−	−	−
katG probe assay	katG WT	−	−	−	−	−	−	+	+
	katG S315T1 mutation	−	+	+	+	+	+	−	−
	katG S315T2 mutation	−	−	−	−	−	−	−	−

^aWT, wild type.

^bR, resistant by BACTEC 460; S, sensitive by BACTEC 460; +, band at the respective wild-type or mutant probe zone; +/−, weak band at the respective wild-type or mutant probe zone (for the wild type, a weakened band indicates the presence of a mixed subpopulation of drug-resistant and -sensitive bacteria); −, absent band at the respective wild-type or mutant probe zone.

^cA clear rpoB S531L mutation and a second D516V mutation (weak band) as strong as the universal control assumed to be a true mutation combined with the absence of the katG wild-type probe, indicating RIF and INH resistance.

^dA weak wild-type probe band 2 for rpoB indicated at least partial (≤1% of the bacterial population) RIF resistance in combination with an INH S315T1 mutation.

In conclusion, the optimized Genotype MTBDR assay was found to be directly applicable with smear-positive specimens. The assay is rapid (<48 h) and easy to perform and allows detection of multidrug resistance in tuberculosis patients suspected of treatment failure or reactivation of prior disease or originating from countries with high prevalences of MDR TB. A rapid tool that simultaneously detects RIF and the more prevalent high-level INH resistance may have a major impact on the future management of tuberculosis.