TEXT

Top Abstract Text References

Eight million new cases of tuberculosis (TB) occur annually in the world, producing almost 3 million deaths, and in 1993, the World Health Organization declared TB to be a global emergency (3). The decline in TB cases in industrialized countries and parts of the developing world ceased or reversed in the mid-1980s. Drug resistance has compromised the successful treatment of TB cases, and multidrug-resistant TB (i.e., having resistance to at least isoniazid and rifampin) has caused high morbidity and mortality, particularly in immunocompromised populations (2, 9). Outbreaks of both drug-sensitive and drug-resistant TB, particularly multidrug-resistant TB, have occurred in both immunocompetent and immunocompromised patients (2, 4, 7, 11). The ability to identify Mycobacterium tuberculosis complex (MTC) isolates and differentiate them from nontuberculous mycobacteria (NTM) is therefore of importance for patient management, hospital control of infection, and public health TB control services.

Several in-house and commercial systems for the rapid identification of mycobacterial species isolated from clinical specimens exist, such as the Accuprobe system (GenProbe, San Diego, Calif.), which includes tests for the identification of the MTC, Mycobacterium avium complex, M. avium, Mycobacterium intracellulare, Mycobacterium kansasii, and Mycobacterium gordonae. The system is highly specific, but a separate test must be performed for each species (8). High-performance liquid chromatography analysis of mycolic acids (6), DNA sequencing of the 16S rRNA gene (10), and 65-kDa hsp PCR restriction enzyme assays (12) have been used successfully. Novel commercial systems for the identification of multiple mycobacterial species have been described, including high-density DNA probe arrays on a "microchip" system (5, 13). The latter system has been commercially developed, but the cost of specialized analytical equipment is high.

In most laboratories, smears of putative mycobacterial cultures are prepared to confirm the presence of acid-fast bacilli, detect gross contamination, and, in experienced reference centers, make a tentative separation of MTC and NTM cultures. Unlike solid cultures, only the microscopic appearance of mycobacteria is available for liquid cultures, which is not a reliable distinguishing indicator alone. The culture would normally be identified with a combination of growth and microscopic characteristics, biochemical assays, and high-performance liquid chromatography or DNA hybridization tests (7).

A new commercial culture confirmation system which utilizes the hydrophobic character of peptide nucleic acid (PNA) probes, enabling them to penetrate the mycobacterial cell wall and bind with high specificity to MTC- and NTM-specific rRNA sequences, has been developed. The system detects binding of PNA oligomers (15 oligonucleotides) to MTC-specific and NTM-specific 23S and 16S rRNA sequences, respectively, by a fluorescence in situ hybridization (FISH) system and microscopic techniques familiar to microbiology laboratories (11a).

The aim of the study was to perform a blinded comparison of the DAKO MTC culture confirmation system (Glostrup, Denmark) on liquid mycobacterial cultures from the Organon Teknika MB/BacT 3D nonradiometric culture system (Cambridge, United Kingdom). We prospectively compared the performance of the new PNA-FISH system with conventional methodology used at the United Kingdom Public Health Laboratory Service Mycobacterium Reference Unit. We believe that this is the first study to demonstrate the use of a commercial PNA-FISH system for mycobacterial identification with liquid cultures.

Sequential positive liquid mycobacterial cultures from the continuously monitored and automated Organon Teknika MB/BacT 3D system were used in a prospective blinded analysis. The medium is based on Middlebrook 7H9 broth. Cultures were analyzed by the PNA-FISH assay and then were analyzed conventionally by macroscopic appearance after subculture, microscopic appearance of Ziehl-Neelsen-stained smears, and growth and biochemical characteristics. Identity was confirmed with DNA hybridization systems (Accuprobe; GenProbe) and thin-layer chromatography or PCR-restriction enzyme analysis (1, 8, 12). Drug susceptibility testing was performed by the resistance ratio method on solid Lowenstein-Jensen medium as previously described (1).

The methodology used was as described in the documentation supplied by the manufacturer, DAKO A/S, with modifications. Approximately 1 ml of acid-fast bacillus-positive liquid culture was transferred from the MB/BacT culture vial into a clean screw-top Microfuge tube and centrifuged at 13,000 × g for 5 min (model no. 5417C; Eppendorf, Hamburg, Germany). The supernatant was carefully removed, 100 to 200 µl of phosphate-buffered saline was added to each tube, depending on the size of deposit visible, and the contents were mixed. From this, 25 µl was added to each of two wells on the supplied two-well slide. The slide was allowed to air dry (approximately 30 min) before fixing the smears by passing the slide through a Bunsen burner flame three or four times and then immersing it in 80% (vol/vol) ethanol for 15 min. The slide was then air dried for a further 10 min. One drop of the fluoroscein isothiocyanate (FITC)-labelled MTC PNA conjugate was added to one well, and one drop of FITC-labelled NTM PNA conjugate was added to the other well. A coverslip (22 by 22 by 0.15 mm) was added to each well, and the slide was incubated in a humidity chamber for 90 min at 55°C. The wash solution was prepared and preheated to 55°C. The coverslips were removed, and the slides were immersed in wash solution for 30 min at 55°C. They were then dipped in distilled water for 30 s and air dried for 10 min. One drop of mounting fluid was added to each well, and the wells were covered with a double coverslip (24 by 60 by 0.15 mm) and incubated in a humidity chamber for 30 min at 55°C. The smears were examined with a fluorescent microscope equipped with a FITC-Texas Red dual-band filter set and a 100× oil objective (Hund) (5 min per slide) (total magnification, ×1,000). Mycobacteria were detected on the basis of green fluorescence against a red-brown background and characteristic morphology. MTC appears as a 1- to 3-µm slender, rod-shaped bacillus which may show cording. NTM species vary in morphology. The intensity of fluorescence may vary, depending on growth medium and species.

The results are given in Table 1. A total of 74 cultures were prospectively analyzed by the usual reference laboratory procedures; 47 cultures were shown to be MTB alone, one was a mixture of MTC and NTM, and 26 were NTM alone. The mean time taken to culture M. tuberculosis from all specimens and from sputum isolates was 17.6 days (range, 6 to 45 days) and 15.2 days (range, 6 to 38 days), respectively. Nontuberculous isolates were cultured for a mean of 16 days (range, 5 to 30 days). The PNA-FISH system correctly identified all MTB cultures (48 of 48), and 22 of 27 (81.5%) NTM cultures were correctly identified as NTM. Two M. tuberculosis strains which subsequently showed isoniazid resistance were also correctly identified by the PNA system, which is of importance since resistance is associated with changes in mycolic acid synthesis and bacterial surface properties. The five NTM cultures which gave no reaction with either of the two probes were identified as Mycobacterium xenopi, Mycobacterium fortuitum (2), and Mycobacterium flavescens. Overall, no result was obtained in 5 of 74 (6.8%) total cultures flagged positive for mycobacteria.

In general, the new culture confirmation system performed well, identifying all M. tuberculosis cultures analyzed. A weak signal was seen in 7 of the 48 MTC-containing cultures (14.6%), although in half the cultures, few mycobacteria were seen by direct microscopy with Ziehl-Neesen staining. Five NTM cultures were not identified by PNA-FISH. Earlier data obtained with pure cultures had suggested that the NTM probe was not complementary to rRNA from Mycobacterium marinum, M. xenopi, M. flavescens, and M. fortuitum, so these species would not be recognized by the assay (11a). The NTM probe is complementary (and so cross-reacts) to rRNA from species of Actinomyces and Rickettsia (11a). Neither probe cross-reacts with rRNA from Corynebacterium or Neisseria species or from Haemophilus influenzae, Klebsiella pneumoniae, Pseudomonas aeruginosa, Propionibacterium acnes, Streptococcus pneumoniae, Staphylococcus aureus, Moraxella catarrhalis, Escherichia coli, and Nocardia asteroides (11a).

The new system is principally targeted at the niche occupied by the Accuprobe system, which is able to identify cultures of the MTC, the M. avium-M. intracellulare complex, M. avium, M. kansasii, and M. gordonae. Neither the new assay nor the Accuprobe system is able to differentiate between members of the MTC or identify all NTM isolates. The fluorescent assay is able to identify a broader range of NTM species than the Accuprobe but cannot distinguish between species. It does permit direct visualization of mycobacteria which, in the hands of experienced reference staff, can support the provisional identification of MTC by macroscopic and microscopic appearances even when low fluorescence due to a small organism load is a problem. Routine clinical laboratories examining few mycobacterial cultures may have difficulty with tests yielding low fluorescence, which was a problem with some cultures, particularly NTM. The test, however, does allow for a second opinion in tests exhibiting borderline results, in that the slides can be reviewed. The identification of mixed mycobacterial cultures is also possible, as occurred in this study. It is less likely that smaller clinical laboratories seeing few mycobacterial cultures would be as successful, as low fluorescence was the most consistent problem with the FISH, particularly when examining NTM. The systems use comparable levels of general laboratory equipment; a luminometer is required for the Accuprobe system, and an appropriately modified FITC-Texas Red filter set needs to be fitted to the microscope to obtain reproducible results with the FISH assay. These additional costs would probably prohibit small laboratories and those in the developing parts of the world from utilizing the FISH assay. The overall times needed to perform the assays are comparable.

In conclusion, the new FISH assay offers a useful alternative to current methods for identifying MTC cultures in reference laboratories.