Use of Nucleic Acid Probes for Identification of Mycobacterium tuberculosis Directly from MB/BacT Bottles

Abstract

The feasibility of using nucleic acid probes directly from positive MB/BacT broth to identify mycobacteria was determined in this study. A total number of 2,727 specimens were cultured into the MB/BacT (Organon Teknika) automated system and on conventional Loweinstein-Jensen (LJ) slants. The Gen-Probe AccuProbe culture identification tests (DNA probes) were used on samples from bottles which were identified as positive for mycobacteria by MB/BacT. Samples of positive MB/BacT broth (0.1 ml) were used directly in the broth culture method for the DNA probes as published by Gen-Probe. Centrifugation of the contents of the bottle was not done prior to probe testing. The number of mycobacteria detected by MB/BacT and LJ was 253 (221 isolates of M. tuberculosis and 32 isolates of mycobacteria other than M. tuberculosis [MOTT]). A total of 96.4% (213 of 221) of the bottles growing M. tuberculosis produced a positive direct DNA probe result for M. tuberculosis complex. One hundred percent (16 of 16) of the bottles growing M. gordonae produced a positive direct DNA probe result for M. gordonae. A total of 3.6% (8 of 221) of the bottles growing M. tuberculosis did not yield a positive direct DNA probe result for M. tuberculosis complex. The testing of subcultures made onto solid media from the positive bottles by AccuProbe identified six of these eight M. tuberculosis isolates. Two (0.9%) M. tuberculosis isolates gave a negative result for the M. tuberculosis probe test applied on the MB/BacT broth and its subculture. The rest of the positive MB/BacT bottles growing MOTT (16 of 32) were negative for M. gordonae, M. avium, M. intracellulare, and M. kansasii probes. The sensitivity and specificity of AccuProbe for the identification of M. tuberculosis and M. gordonae directly from MB/BacT broth were 96.4 and 100% for M. tuberculosis and 100 and 100% for M. gordonae, respectively. The direct testing of positive MB/BacT broth by AccuProbe, without prior centrifugation, allows for the accurate and rapid identification of M. tuberculosis and M. gordonae.

Mycobacterium tuberculosis infection remains a public health concern especially in developing countries like Turkey. One of the most distressing problems around the world to arise from the resurgence of tuberculosis among AIDS patients and the homeless population is the increasing incidence of multiple drug resistance among M. tuberculosis isolates. These points have focused attention on the time required to report acid-fast bacillus (AFB) smear results and to isolate, identify, and determine the antimicrobial susceptibility of M. tuberculosis (4, 19). The most reliable method for diagnosis of mycobacterial infection is the culture method, while more rapid laboratory alternatives are microscopy and nucleic acid amplification methods (2, 13). There have recently been developed several automated, nonradiometric culture systems on mycobacterial culture. One of these is the MB/BacT system (Organon Teknika, Turnhout, Belgium), which relies on a continuous colorimetric CO₂ detection device to indicate the mycobacterial growth in a closed, automated system. Several reports have described the MB/BacT system as a well-automated system for detection of mycobacteria in clinical specimens compared with the BACTEC-460TB system and solid media such as Loweinstein-Jensen (LJ) and Middlebrook 7H11 agar. They reported the sensitivity and contamination rate of MB/BacT to be between 78 and 95% and 4 and 7%, respectively. The average number of days required for recovery of mycobacteria by the MB/BacT system was reported to be between 13.8 and 17.5 days (1, 12, 16, 18).

DNA probes (AccuProbe; Gen-Probe, Inc., San Diego, Calif.) can be used for the rapid identification of M. tuberculosis, M. avium and M. intracellulare, M. gordonae, and M. kansasii from culture (3, 7, 8). Several researchers have used the Gen-Probe or AccuProbe probe to identify mycobacteria from positive BACTEC-460TB, Mycobacteria Growth Indicator Tube, and Difco ESP MYCO culture system bottles. Rates of identification of M. tuberculosis, M. avium-M. intracellulare, M. kansasii, or M. gordonae from positive bottles with the probes have varied by species or by the number of mycobacterial organisms in the bottles, showing sensitivities of 47 to 100% for M. tuberculosis and 78.5 to 100% for M. avium-M. intracellulare (3, 5, 7–9, 11, 14, 15, 17). In these reports, prior to probe testing, centrifugation of 1 to 1.5 ml of positive broth from the AFB-positive cultures at 3,000 to 10,000 × g for 15 to 30 min is recommended. However, this centrifugation step would be time-consuming and labor-intensive and would lead to cross-contamination from the positive bottles. Instead of centrifugation, AccuProbe recommends selecting a 0.1-mL sample from the growth in Middlebrook 7H9 broth with turbidity equivalent to or greater than a McFarland no. 1 nephelometer standard if the probe test is applied with the broth culture method.

Few data are available concerning the use of the MB/BacT system in conjunction with DNA probes for the rapid detection and identification of mycobacteria (6, 12). In this study, we aimed to determine the feasibility of using a nucleic acid probe directly from positive MB/BacT bottle broth without prior centrifugation of the broth to identify mycobacteria.

Methods.

A total of 2,727 clinical specimens (1,776 respiratory and 951 nonrespiratory) were submitted to the Clinical Mycobacteriology Laboratory of Ege University Medical School, Izmir, Turkey, for mycobacterial culture. All nonsterile specimens were digested and decontaminated by the N-acetyl-l-cysteine (NALC)–4% NaOH method and were concentrated by centrifugation at 3,000 × g for 15 min. Sediments were resuspended in 50 ml of phosphate buffer and recentrifuged at 3,000 × g for 15 min (10). After decontamination and concentration, the sediments were neutralized with 1 N HCl, and 0.5 ml of the sediment was inoculated in a MB/BacT bottle and 0.1 ml was inoculated onto LJ slants. Sterile specimens were not decontaminated and were inoculated in one MB/BacT bottle and two LJ slants after concentration by centrifugation at 3,000 × g for 15 min. Prior to inoculation in the MB/BacT bottle, according to the manufacturer’s instructions, 0.5 ml of reconstituted MB/BacT antibiotic supplement (Organon Teknika) was added to each MB/BacT bottle for culture of nonsterile specimens while 0.5 ml of MB/BacT reconstitution fluid (Organon Teknika) alone was added to each MB/BacT bottle for culture of sterile specimens. A smear was made for fluorochrome staining and was quantified as follows: >10 AFBs per 200× field, +4; 1 to 10 AFBs per 200× field, +3; 1 to 10 AFBs per 10 200× field, +2; <10 AFBs per entire smear, +1 (15). All cultures were incubated at 35°C for 8 weeks. MB/BacT bottles were analyzed every 10 min by the system software, and the solid medium cultures were examined weekly. When a bottle flagged positive, Kinyoun staining was performed on well-vortexed MB/BacT broth. If AFB presence was confirmed on the smear, AccuProbe was tested in the well-mixed broth according to the manufacturer’s instructions, and furthermore the broth was subcultured onto a 7H11 agar plate and two LJ slants for purity and growth control. If Kinyoun smear was negative for AFB and no other contaminant microorganisms were seen, the bottle and its subcultures were reincubated at 35°C for a period of 4 weeks. They were examined visually every week, and Kinyoun staining and DNA probe testing were performed if needed. M. tuberculosis, M. avium-M. intracellulare, M. gordonae, and M. kansasii probes were used on samples from AFB-positive MB/BacT bottle broths for identification as an initial step. The AccuProbe procedure was performed once a week. Samples of broth (0.1 ml) from the positive bottles were used directly in the broth culture method for the DNA probes as published by Gen-Probe. Centrifugation of the contents of the broth was not done prior to probe testing. The specific probe was chosen on the basis of the color and Kinyoun staining of the broth in the bottle. If cord formation of AFB was seen on Kinyoun staining, only the M. tuberculosis probe was tested for this bottle. The set of M. tuberculosis, M. avium-M. intracellulare, and M. kansasii probes was tested for nonchromogenic (not pigmented) MB/BacT broth without cord formation on Kinyoun staining. Chromogenic broth, which was pigmented light yellow to orange, was tested for M. avium-M. intracellulare and M. gordonae as well as M. tuberculosis to determine mixed growth. If the AFB-positive bottle broth gave negative results at the initial attempt for the probes, the bottle was incubated at 35°C for 1 week before the next set of probe tests. If MB/BacT broth gave a negative result after the second set of probe tests, the identification procedure was performed on the colonial growth of subcultures made onto solid media from the positive bottle. All subcultures were examined for purity. The colonies on the subculture were tested with M. tuberculosis, M. avium-M. intracellulare, M. gordonae, or M. kansasii probes on the basis of the colonial morphology and pigmentation as the first step of identification. If the DNA probes tested on these colonies gave negative results, the organisms were identified by conventional biochemical tests. The positive results of M. tuberculosis and M. gordonae probes were confirmed by the conventional biochemical testing of subculture growth. The conventional biochemical tests also identified the other species of mycobacteria for which the commercial probes were not available. All conventional biochemical tests were performed according to reference 10. If no growth was detected in the MB/BacT bottle, mycobacterial isolates that grew only on the original solid media were not included in this study.

Results.

A total number of 2,727 specimens collected from 1,997 patients were cultured in the MB/BacT system and onto LJ slants. The majority of specimens (1,776) were from respiratory sources, with sputum samples accounting for 1,112 specimens. Smear examinations were positive for AFB for 131 (52% sensitivity) culture-positive specimens. The smear and culture results for all specimens with a positive smear or culture result are shown in Table 1. As shown, AFB grew from 131 (113 M. tuberculosis specimens and 18 specimens of mycobacteria other than M. tuberculosis [MOTT]) of 140 smear-positive specimens. The smear-positive and culture-negative specimens included nine specimens that were obtained from patients under treatment for M. tuberculosis infection. There were 122 smear-negative specimens that subsequently grew mycobacteria (108 M. tuberculosis isolates and 14 MOTT isolates). The number of mycobacterial isolates detected by MB/BacT and LJ was 253 (172 isolates were from respiratory specimens while 81 isolates were from nonrespiratory specimens). The isolates that were detected by solid media only were not included in this study. The average number of days required for recovery of M. tuberculosis and MOTT by the MB/BacT system was 16 and 23 days, respectively. All positive bottles that grew mycobacteria were positive when stained for AFB by Kinyoun staining. Three bottles flagged positive, and Kinyoun staining of the broth was negative for AFB. There was no growth in these bottles and their subcultures after incubation at 35°C for 4 weeks, and they were excluded from the study. Kinyoun staining yielded positive results for cord formation on 157 positive MB/BacT broth bottles, and these bottles were probed for M. tuberculosis only. Nonchromogenic broth without cord formation was recorded for 75 bottles, and they were probed for M. tuberculosis, M. avium-M. intracellulare, and M. kansasii. There were 21 positive MB/BacT bottles with chromogenic broth that grew mycobacteria (16 bottles with M. gordonae, 1 with M. szulgai, 1 with M. flavescens, and 3 with M. scrofulaceum), and they were probed for M. gordonae and M. avium-M. intracellulare. Of the 253 mycobacterial isolates, 221 (87%) were identified as M. tuberculosis (161 M. tuberculosis isolates were from respiratory specimens) and 32 (13%) were identified as MOTT (11 MOTT isolates were from nonrespiratory specimens) by AccuProbe or conventional biochemical tests. AccuProbe identified 219 (99%) of M. tuberculosis isolates and all M. gordonae (100%) isolates among those of MOTT. AccuProbe results for M. tuberculosis and M. gordonae are shown in Table 2. Two hundred thirteen (96.4%) of the 221 bottles growing M. tuberculosis produced a positive direct DNA probe result for M. tuberculosis complex, and 100% (16 of 16) of the bottles growing M. gordonae produced a positive direct DNA probe result for M. gordonae. Of the 221 bottles, 8 (3.6%) bottles did not yield a positive direct DNA probe result for M. tuberculosis complex. The testing of subcultures made onto solid media from the positive bottles by AccuProbe for M. tuberculosis identified six of these eight isolates. Two (0.9%) of the 221 isolates gave negative results for the M. tuberculosis probe test applied on the MB/BacT broth and its subculture and original solid medium culture growth. These isolates were identified by colony morphology and conventional biochemical tests applied on subculture growth. One hundred ninety-nine (90%) of 221 broth bottles yielded positive results for M. tuberculosis complex probes at the initial attempt, and the turbidity of each broth was greater than a McFarland no. 1 nephelometer standard. Twenty-two of the 221 broth bottles gave negative results at initial M. tuberculosis complex probe testing, and the turbidity of each broth was almost equivalent to or lower than a McFarland no. 1 nephelometer standard. These were retested after incubation of the bottles at 35°C for 1 week. After incubation, the turbidity achieved at least a McFarland no. 1 nephelometer standard for 18 (82%) of the 22 bottles. Fourteen of 18 (6.3% of M. tuberculosis samples) bottles gave positive results for M. tuberculosis complex after this second attempt. Eight (3.6%) of the 221 bottles were still negative at the second attempt of M. tuberculosis complex probe applied directly on MB/BacT broth. In addition to M. tuberculosis probe, M. avium-M. intracellulare, M. gordonae, and M. kansasii probes were tested for these eight bottles. None of the bottles was positive with these probes.

TABLE 1.

Smear and culture results for all specimens

Specimen	Total no.	No. of M. tuberculosis specimens	No. of MOTT specimens	No. of specimens with smear result:
				Negative		1+		2+		3+		4+
				M. tuberculosis	MOTT	M. tuberculosis	MOTT	M. tuberculosis	MOTT	M. tuberculosis	MOTT	M. tuberculosis	MOTT
Respiratorya	1,776	161	11	73	5	35	3	27		13	1	13	2
Nonrespiratoryb	951	60	21	35	9	9	5	7	2	3	2	6	3
Total	2,727	221	32	108	14	44	8	34	2	16	3	19	5

^aTypes and numbers of specimens were as follows: sputum, 1,112; bronchoalveolar lavage, 171; bronchial wash, 177; bronchial brush, 54; bronchial aspiration, 176; pleural fluid, 62; lung biopsy, 24. 

^bTypes and numbers of specimens were as follows: urine, 402; feces, 18; peritoneal fluid, 34; ascitic fluid, 35; gastric aspiration, 112; biopsy, 56; cerebrospinal fluid, 62; bone marrow, 54; synovial fluid, 84; pericardial fluid, 25; blood, 8; pus, 61. 

TABLE 2.

AccuProbe results for M. tuberculosis and M. gordonae

Sp.	Overall no.	No. (%) of positive results from medium:				No. (%) of negative results from medium:
		MB/BacT broth			Subculture medium	MB/BacT broth			All media
		Total	1st attempt	2nd attempt		Total	1st attempt	2nd attempt
M. tuberculosis	221	213 (96.40)	199 (90)	14 (6)	6 (2.70)	8 (3.60)	22 (10)	8 (0.60)	2a (0.90)
M. gordonae	16	16 (100)	16 (100)

^aIdentified by conventional biochemical tests. 

In this study, 32 MOTT isolates were grown in the MB/BacT bottles and on LJ slants. AFB was seen on Kinyoun staining from all of the MB/BacT bottles, which grew MOTT. Further identification of 32 MOTT isolates yielded 16 M. gordonae isolates and 16 other strains of MOTT; 3 of M. fortuitum, 2 of M. chelonae, 2 of M. abscessus, 2 of M. szulgai, 1 of M. terrae complex, 1 of M. flavescens, 1 of M. xenopi, 1 of M. triviale, and 3 of M. scrofulaceum. There was no M. kansasii or M. avium-M. intracellulare isolate growth. All of the M. gordonae (100%) isolates were detected in the chromogenic MB/BacT broth and were positive for the M. gordonae probe and negative for the M. avium-M. intracellulare probe performed directly on the MB/BacT broth. The rest of the MOTT (16 of 32) were negative for M. gordonae, M. avium-M. intracellulare, and M. kansasii probe, and they were identified by conventional biochemical tests. The turbidity of each broth that grew MOTT was greater than a McFarland no. 1 nephelometer standard at the time of positive signal of the bottle.

All the probe identification results were also confirmed by conventional biochemical tests that were performed on the colonies grown on solid subculture media. Mixed mycobacterial growth from the same specimens was not found in this study.

The sensitivity, specificity, positive predictive value, and negative predictive value of AccuProbe for the identification of M. tuberculosis directly on MB/BacT broth were 96.4, 100, 100, and 80%, respectively. For M. gordonae, the sensitivity and specificity of the AccuProbe identification test directly on MB/BacT broth were 100 and 100%, respectively.

Discussion.

In this study, identification of mycobacteria to species level was attempted directly from positive MB/BacT bottles by the AccuProbe assay according to the broth culture method published by Gen-Probe. We report the rates of sensitivity of AccuProbe assay on the MB/BacT broth to be 96.4 and 100% for M. tuberculosis and M. gordonae, respectively, which is similar to the 100 and 97% for M. tuberculosis and M. avium-M. intracellulare, respectively, reported by others (6). At the time of positive signal of MB/BacT bottles, the broth of the bottles has sufficient cell mass (equivalent to or greater than the McFarland no. 1 nephelometer standard) to allow for identification of the isolate by AccuProbe test from an aliquot of 0.1 ml obtained directly from the MB/BacT bottle broth. Centrifugation of the contents of the broth is not necessary to obtain positive probe reactions for 96.6% of positive MB/BacT bottles, which grow mycobacteria that can be identified with AccuProbe. Also, continued incubation of the bottles after detection is not necessary since at the time of positive signal of the MB/BacT bottles, 90% of the bottles that grow M. tuberculosis and 100% of the bottles that grow M. gordonae have sufficient cell mass for the AccuProbe assay. This percentage for bottles that grow M. tuberculosis increases to 98.2% after 1 week of incubation of the bottles at 35°C. This procedure is technically easier and more rapid than centrifugation of the broth. The choice of the specific probe to be utilized in the MB/BacT bottle is made by pigmentation of the broth, cord formation of AFB in the broth, and the prevalence of the mycobacterial strains in the region.

In summary, the automated detection of mycobacterial growth in the MB/BacT system and simplified mycobacterial identification by direct AccuProbe assay are labor-conserving methods that do not involve radioactive compounds and result in rapid and accurate mycobacterial identification from cultures of clinical specimens.