Evaluation of a Rapid PCR-Based Epidemiological Typing Method for Routine Studies of Mycobacterium tuberculosis

Abstract

Restriction fragment length polymorphism (RFLP) based on the insertion sequence IS6110 is used to investigate episodes of suspected transmission of infection of tuberculosis but usually takes a number of weeks from receipt of request to obtain a result. Often investigations would benefit from a more rapid method, possibly one containing an amplification step. The method employed uses a simple DNA extraction followed by a PCR step involving a single primer. Restriction enzyme analysis was performed when the patterns obtained from the PCR products were indistinguishable, especially when only single similar-size bands were obtained. The isolates used were strains of Mycobacterium tuberculosis submitted for epidemiological investigations as part of (i) possible contact-outbreak (22 episodes involving between 2 and 20 patients), (ii) possible incidents of laboratory cross-contamination (21 episodes), and (iii) possible change in drug resistance pattern or a case of reinfection (1 patient). The PCR products giving similar patterns were then subjected to restriction enzyme analysis. In conclusion it has been shown that this method is rapid, with results within 1 to 2 days of the request being received; is reproducible; and gives the same results as does RFLP. The restriction enzyme analysis stage has improved the efficiency of the technique.

DNA fingerprinting or restriction fragment length polymorphism (RFLP) based on the insertion sequence IS6110 (5) is used to investigate episodes of suspected transmission of infection of tuberculosis. The technique has proven value in both confirming and disproving transmission within groups involving families, friends, and/or work colleagues or within communities such as schools, social centers, prisons, and workplaces. False clusters, however, can occur when there is a breakdown in technique within the laboratory or in the collection of specimens on the ward or clinic leading to cross-contamination of specimens.

Very often the cluster under investigation would benefit greatly from a rapid verification, as delays could result in either further transmission or patients receiving incorrect and possibly hazardous treatment.

The present “gold standard” method, RFLP, takes a number of weeks from the receipt of the request to obtain a result. The method does not involve an amplification step; therefore, a large amount of growth is required in order to extract enough DNA to do the test; this can take anywhere from 2 to 6 weeks, and the test itself takes a minimum of 3 to 5 days. A more rapid method, possibly one containing an amplification step, is required to respond quickly to investigate possible clusters, and a variety of methods have been suggested elsewhere (2, 3, 4, 7).

The method that we have employed uses a simple DNA extraction procedure followed by a PCR step involving a single primer aimed at the insertion sequence IS6110 and is based on the method described by Wilson et al. (6). The technique, however, was not able to distinguish between products of about the same size, and so a further step involving restriction enzyme analysis of the PCR product was introduced. The results were compared to those obtained using RFLP.

Strains.

The 177 isolates used were strains of Mycobacterium tuberculosis submitted for epidemiological investigations from laboratories in the United Kingdom and Ireland. The strains were identified by phenotypic and biochemical tests (1) and/or DNA hybridization tests (Accuprobe; Genprobe, San Diego, Calif.). The cultures submitted were then presented for DNA extraction for PCR (see below) followed by subculturing onto two Lowenstein-Jensen (LJ) slopes for RFLP analysis.

The strains were submitted either as part of (i) possible contact-outbreak investigations (81 isolates), (ii) possible incidents of laboratory cross-contamination (94 isolates), or (iii) possible changes in drug resistance patterns or a case of reinfection (2 isolates).

DNA extraction.

With a 1-μl loop a small quantity of growth, equivalent to about two small colonies, was scraped from the top of the culture and placed into 100 μl of sterile distilled water in a microcentrifuge tube. Chloroform (100 μl; Sigma) was added, and the mixture was vortexed for about 10 s. The mixture was then heated at 80°C for 20 min, after which time it was held at −20°C for at least 20 min or until needed. When required, the sample was allowed to thaw but while still cold was centrifuged at 12,000 × g for 3 min in a minicentrifuge. For the PCR 10 μl of the supernatant was used.

PCR.

A single primer was used that was targeted to the inverted repeat sequence of the IS6110 (5"-GAGTCTCCGGACTCACCGG-3"), and the PCR was performed in a volume of 40 μl as described previously (6).

Briefly, the reaction conditions were as follows: an initial denaturation at 95°C for 120 s; 1 cycle of 95°C for 20 s, 45°C for 360 s, and 72°C for 120 s; 30 cycles of 95°C for 20 s, 62°C for 30 s, and 72°C for 180 s; and a final extension at 72°C for 10 min.

The PCR products were visualized on an 0.8% agarose gel with 0.5 mg of ethidium bromide per ml. The gel was run on a Bio-Rad Wide Mini-Sub cell electrophoresis system at 100 V for 30 min initially and a further 20 min if further separation was required (Fig. 1a).

FIG. 1.

Gel electrophoresis of PCR product directly after PCR (a) and after restriction enzyme digestion with HaeIII (b). Lanes 1 to 3, possible contacts; lanes 4 to 6, possible laboratory cross-contamination; lanes 7 and 8, possible contacts; lanes M, molecular weight marker (lambda DNA/HindIII); lanes B, Bioline hyperladder I (a) and IV (b). The 1,000-bp marker is indicated.

The PCR products giving similar patterns were then subjected to restriction enzyme analysis using HaeIII (Promega). It was extremely useful when the patterns gave single bands of similar sizes.

Restriction enzyme analysis.

Ten microliters of PCR product was added to 6 μl of sterile distilled water, 2 μl of restriction enzyme HaeIII (Promega), and 2 μl of corresponding buffer. The mixture was incubated at 37°C in a water bath for 1 to 2 h. The results were analyzed on a 2% (wt/vol) agarose gel with 0.5 mg of ethidium bromide per ml (Fig. 1b).

RFLP.

When the growth on the LJ slopes was sufficient, the DNA was extracted and purified and RFLP based on the insertion sequence (IS6110) was performed as described by van Embden and colleagues (5).

Reproducibility.

The reproducibility of the method was tested by performing PCR on (i) DNA extracted and stored at −20°C (both by the method described above and with purified DNA as used in RFLP), (ii) DNA extracted from LJ slopes stored at room temperature and −20°C, and (iii) DNA extracted from repeat isolates from patients.

The effect of different PCR runs on the PCR product obtained was also investigated.

Results are shown in Table 1.

TABLE 1.

Types of epidemiological investigation (all confirmed by RFLP)^a

Type of investigation	No. of episodes	No. of patients in episode (no. of patient strains that were indistinguishable)
Contact tracing
Health care worker and patients	6	2 (0), 2 (0), 2 (0), 2 (0), 3 (0), 20 (2)
Patients in a nursing home	1	2 (2)
Members of a church group	1	6 (6)
Children in a nursery school	1	5 (2)
Children in a school	2	3 (2), 6 (6)
Colleagues at work	1	2 (0)
Members of the same family	4	2 (2), 2 (2), 2 (2), 2 (2)
Students in the same house	2	2 (0), 3 (0)
Patients on a renal ward	1	8 (0)
Case of sexual transmission	1	2 (2)
Contacts with MDRTB	1	2 (2)
Contacts with INH-resistant TB	1	3 (3)
Laboratory cross-contamination	21	2 (0), 2 (2), 2 (2), 2 (2), 2 (2), 2 (2), 2 (2), 3 (3), 3 (3), 3 (3), 3 (3), 3 (3), 3 (3), 3 (2), 3 (2), 4 (4), 6 (6), 8 (6), 10 (9), 11 (11), 17 (8)
Change in drug resistance	1	1 (2 isolates)

^aMDRTB, multidrug-resistant tuberculosis; INH, isoniazid; TB, tuberculosis.

In all cases the rapid PCR-based epidemiological typing (RAPET) and RFLP gave the same conclusion.

From the receipt of culture the RAPET method produced a result within 1 to 2 days whereas the RFLP method took 5 days after a confluent growth had been grown on the subcultures.

Contact tracing.

There were 22 episodes involving the epidemiological investigation of between 2 and 20 patients as listed in Table 1.

Cross-contamination.

RAPET was also applied to 94 patient isolates from 22 episodes of possible laboratory cross-contamination. In two cases involving 8 and 17 isolates the sodium hydroxide used for decontaminating the sputum had been contaminated and in a third case (2 isolates) the contamination had occurred via a BACTEC 460 machine. Another investigation involved the possible contamination of consecutive bronchoscopy specimens. In all other cases the method of laboratory contamination remains unclear.

Change in drug resistance.

In one patient on treatment for 8 months the strain isolated had changed from being fully sensitive to being resistant to isoniazid and rifampin. The question was raised whether the strains had become resistant or the patient had been reinfected with a second strain. Both the RAPET and subsequently the RFLP showed that the DNA patterns were indistinguishable between the two isolates, suggesting that the strain had become resistant.

Reproducibility.

In all tests the pattern obtained after PCR remained constant for the strain under investigation even when the DNA was amplified in different PCR runs.

In conclusion it has been shown that this method is rapid, with results in 1 to 2 days of the request being received; is reproducible; and gives the same results as the more time-consuming RFLP method. The introduction of the restriction enzyme analysis stage has improved the efficiency of the technique, especially when the patterns obtained after PCR contained only single bands of similar sizes.

Its use is limited to small investigations rather than population surveys, but it has been extremely useful in a recent outbreak investigation of isoniazid-resistant tuberculosis cases. This investigation was instigated by the finding of three patients with an isoniazid-resistant strain of M. tuberculosis that had an indistinguishable DNA pattern (Table 1). The RAPET method has proved valuable in being able to rapidly screen a large number of isolates in order to limit the spread of the outbreak strain. To date 288 monoresistant isoniazid-resistant strains of M. tuberculosis from a 2-year period have been investigated, with the detection of 67 cases. Further isolates are being investigated. So far RFLP has been done on 40 of the 67 outbreak isolates, showing all to have the same 15-band pattern, but its progress is much slower than that of RAPET.

Overall the RAPET method gave results within 2 days of the culture being received, as very little growth was required. RFLP requires a large amount of growth in order to produce enough DNA for the test. This can take 2 to 6 weeks. The technique then takes 5 days to complete.