INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The Mycobacterium avium complex (MAC) remains a challenge to mycobacterial classification due to the high heterogeneity described within its components. The MAC has been formally divided into two well-recognized species: Mycobacterium avium and Mycobacterium intracellulare. Three different subspecies of M. avium have also been described: Mycobacterium avium subsp. avium, Mycobacterium avium subsp. silvaticum, and Mycobacterium avium subsp. paratuberculosis (14, 30).

Conventional cultural and biochemical tests give little information on which to separate M. avium and M. intracellulare, and therefore, the species are difficult to distinguish in a standard clinical microbiology setting (14). Recent studies suggest that DNA-based methods for the identification of MAC species may be more useful. A commercial hybridization assay (the AccuProbe system; GenProbe Inc., San Diego, Calif. [8, 22]) is the most widely used molecular system and is considered the molecular "gold standard" for the rapid identification of MAC components. Several clinical laboratories now perform identification only to the MAC level by using these commercial probes, and thus, many clinical isolates are identified as MAC regardless of their clinical relevance. Several authors have described a remarkable internal heterogeneity within the complex, suggesting that the MAC probably contains several unknown taxonomic groups (8, 13, 25, 33). A more precise knowledge of which MAC components are involved in clinical infections could give better insight into the relevance that these species have as human pathogens.

MAC components have been isolated from a wide variety of sources, including animals, humans, and the environment. M. avium subsp. silvaticum and M. avium subsp. paratuberculosis are more frequently isolated from animal sources, the latter causing Johne's disease in cattle and other animals (14). The association of M. avium subsp. avium and M. intracellulare with human diseases is often strain specific. Disseminated infections of M. avium subsp. avium have been identified more frequently than M. intracellulare infections in human immunodeficiency virus-positive (HIV⁺) patients (16). The use of highly active antiretroviral therapy has greatly improved the prognosis in HIV⁺ patients, leading to a sharp decrease in the isolation rate of M. avium subsp. avium in disseminated infections in AIDS. M. intracellulare has been isolated more frequently from HIV patients than from HIV⁺ patients, and its relationship to infectiveness is poorly characterized (7, 11). Isolation of M. intracellulare is more frequent from children with cervical lymphadenopathy and patients with pulmonary illnesses (17). Neither of these groups of patients has a recognized association with host immune dysfunction (3).

During a recent study of molecular typing of MAC clinical isolates using standardized restriction fragment length polymorphism methodology with IS1245 as a probe (32), we detected several isolates from human samples lacking IS1245. Three of those isolates were obtained from an HIV child suffering from an essential immunodeficiency (1).

All three isolates were positive with the commercial MAC-specific probe (AccuProbe system), but species-specific probes gave unclear results. This study describes further molecular methods applied to determine a more accurate identification of those isolates.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Patient profile. A 5-year-old girl was admitted to the hospital with a diagnosis of chronic multifocal osteomyelitis of unknown origin. When she was 2 years old, she complained of pain and difficulty in walking. This was investigated by bone scanning with ⁹⁹Technetium and a camera, which showed an increased uptake in the left femur, shank, and right ankle. She also had positive responses to antigen 60 (1,700 U) and Mantoux tests (8 mm of induration). Her brother had died previously from meningitis caused by Mycobacterium bovis infection, which suggested that antituberculosis treatment should be commenced. After 1 year of treatment, only partial remission was obtained with standard antimycobacterial therapy, with persistant pain and an unaltered gammagraphic ⁹⁹Tc uptake. She continued with the same treatment until 3 months before a second admission to the pediatric department. At that time, imaging of the cranial computerized axial tomography and gammagraphic studies showed multifocal illness, and the histopathological study indicated chronic osteomyelitis. Immunological studies showed that receptor I of gamma interferon was not expressed, suggesting structural and functional deficiencies (1).

Standard identification techniques. Mycobacteria were isolated from three samples obtained from frontal and maxillar origins. The isolates obtained were labeled DO67, DO68, and DO69 (frontal exudate, maxillar exudate, and maxillar biopsy specimen, respectively). They were identified as MAC members by standard biochemical methods (18) and commercial probes (AccuProbe system). Analyses of blood, gastric lavage fluid, and urine were negative. Antimycobacterial susceptibility tests were performed using the radiometric BACTEC-460 system with the following drugs and concentrations: ethambutol (4 and 8 µg/ml; Wyet-Lederle), rifabutin (0.25 and 0.5 µg/ml; Pharmacia and Upjohn), ofloxacin (2 and 8 µg/ml; Hoechst Marion Roussel), amikacin (4 and 8 µg/ml; Laboratorios Juste), clarithromycin (2 and 4 µg/ml; Abbot Laboratories), and azithromicin (16 and 32 µg/ml; Pfizer).

Bacterial strains. The following bacterial strains were used: reference strains, M. tuberculosis Mt14323, M. avium ATCC 25261^T, and M. intracellulare IWGMT (International Working Group on Mycobacterial Taxonomy) 10 (23); clinical isolates, M. avium 26C, M. avium DO22, M. avium DO64, and M. intracellulare GM51.

Molecular techniques. Mycobacteria were maintained on Löwenstein-Jensen agar slants. DNAs of the mycobacteria were purified as described previously (32). The following molecular methods, used for identification of the isolates, were applied. (i) AccuProbe system. Three kind of probes were applied, detecting the MAC, M. avium, and M. intracellulare, respectively; each test was performed by following the manufacturer's instructions and with the inclusion of controls. The MAC probe detects both species mentioned above, as well as nonspecific MAC isolates and isolates designated MAIX (24). (ii) Characterization of the 16S rRNA gene was performed by PCR amplification of a 1,500-bp fragment, followed by manual sequencing of both chains at the hypervariable region (15). (iii) PCR restriction analysis of the hsp65 gene was performed as described by Telenti et al. (28). (iv) Detection of DT1 and DT6 MAC-specific genes was undertaken by using PCR as described by Sola et al. (26). Hybridization analysis was also done. DNA digestions, as well as DT1 and DT6 probes, were prepared as described by the same authors. The probes were radioactively labeled with [-³²P]dCTP (Amersham Life Science) by using the Megaprime DNA-labeling system (Amersham Life Science). The gels were subjected to Southern blotting, transferred to Hybond N+ membranes (Amersham Life Science), and hybridized overnight at 55°C with 30% formamide. After hybridization, the filters were washed in 2× SSC (sodium saline citrate; 1× SSC is 0.15 M NaCl-0.015 M trisodium citrate) plus 0.1% sodium dodecyl sulfate for 15 min at room temperature and then in 1× SSC plus 0.1% sodium dodecyl sulfate for 15 min at 55°C. (v) PCR amplification and sequencing of the gene coding for the 32-kDa protein were performed as described previously (24). (vi) PCR amplification and sequencing were applied to study the internal transcribed spacer (ITS) of the rrnA mycobacterial operon following the procedure described by Frothingham and Wilson (12). (vii) Detection of the macrophage-induced gene (mig) by PCR was done as described previously (3).

Nucleotide sequence accession numbers. The new nucleotide sequences were submitted to GenBank and assigned accession numbers AJ306710, AJ306711, and AJ306712, corresponding to the 16S rrn gene, the ITS 1 genomic region, and the 32-kDa gene (fbpA), respectively.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Several molecular techniques were applied for the characterization of MAC clinical isolates. Commercial probes for identification of MAC members, such as the AccuProbe system, are particularly useful in a clinical mycobacteriology laboratory due to their rapidity and ease of use. By this approach, isolate DO67 was weakly positive to the M. avium-specific probe. The two remaining isolates (DO68 and DO69) were negative to both species-specific probes. It was therefore impossible to assign them to any of the well-recognized species described in the MAC.

Analysis of conserved genes. Sequencing of the 16S ribosomal DNA (rDNA) is considered essential in the description of bacterial species. Within the genus Mycobacterium, however, resolution of species by comparison of 16S rDNA sequences is not possible (33). Despite giving different reactions with species-specific AccuProbe systems (as mentioned above), our three strains all showed the same 16S rDNA sequence within the hypervariable region (Fig. 1A). This sequence was most closely related to that described by Wayne et al. (33) as belonging to M. intracellulare-like strains (Fig. 1A, second line).

View larger version (30K): [in this window] [in a new window]   FIG. 1.   Comparison of conserved genomic sequences. Dots indicate identity; dashes represent alignment gaps. Min, M. intracellulare; Mav, M. avium. (A) Comparison of the 16S rDNA species-specific hypervariable region A. The first nucleotide corresponds to Escherichia coli 16S rDNA position 109. DO67-69, clinical isolates from this work; Min 1, IWGMT 90247 (X88917); Min 2, ATCC 13950T (af059849); Min 3, m61685; Mav 1, X52918; Mav 2, ATCC 25291T (af059853). (B) Comparison of 16S-23S rDNA ITS sequences. Dots in parentheses indicate a longer sequence that is not shown. DO67-69, clinical isolates; MAC-C, Min-A to -C, and Mav-A, sequevar designations from reference 12 (superscript a) and reference 6 (superscript b).

Analysis of MAC-specific genes. Several genes have been identified as specific to members of the MAC, which may aid species differentiation within the complex. Thierry et al. (29) successfully used species-specific primers for a selective amplification of the DT6 genomic region in the M. avium genome, as well as the DT1 genomic region in the M. intracellulare genome. DT1 has also been detected in M. avium serovars 2 and 3 (26). We were unable to detect any of the fragments by PCR in the isolates under study; however, hybridization signals were obtained by using DT1 as a probe to hybridize to PvuII-digested genomic DNA. No hybridization was obtained by using DT6 as a probe (data not shown).

Analysis of other genomic regions. The ITS is a sequence located between the 3' end of the 16S and the 5' end of the 23S coding regions inside the mycobacterial ribosomal operon. It has also been used in the characterization of mycobacterial species (10, 13), particularly in the characterization of members of the MAC (6). Sequence data from the ITSs of isolates in this study showed the most sequence similarity to sequevar MAC-C (Fig. 1B) (12), representing a group of M. intracellulare variants (see Fig. 2 in reference 6).

View larger version (29K): [in this window] [in a new window]   FIG. 2.   Alignment of the sequences coding for the mycobacterial 32-kDa protein between nucleotides 753 and 893. The deduced amino acid sequences are shown in boldface single-letter code. The DO67, DO68, and DO69 sequence was used as a reference. Only nucleotides and amino acids different from the reference sequence are indicated in lines 1 to 5; dots indicate identity in nucleotides, and dashes represent alignment gaps. DO67-69, clinical isolates from this work; line 1, Mycobacterium scrofulaceum ATCC 35791 (X92568); line 2, Mycobacterium simiae NCTC 25275 (X92569); line 3, M. avium ATCC 17769 (Z33657); line 4, M. intracellulare ATCC 35761 (Z33660); line 5, MAIX clinical isolate HO312/91 (Z33663). The sequences in lines 1 to 5 were obtained from Soini et al. (24).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The main aim of this study was to use several molecular techniques for a more precise characterization of MAC clinical isolates. The MAC mycobacteria isolated from different clinical sources (DO67, DO68, and DO69) were all susceptible to the lowest concentration used for all the drugs tested: ethambutol (4 µg/ml), rifabutin (0.25 µg/ml), ofloxacin (2 µg/ml), amikacin (4 µg/ml), clarithromycin (2 µg/ml), and azithromycin (16 µg/ml). The patient was treated with ethambutol, rifabutin, and clarithromycin. A clear clinical and radiological improvement was observed after 9 months of treatment, and the patient was discharged (1).

Table 1 summarizes results obtained in the characterization of the isolates under study. GenProbe analysis showed unclear results, making the assignment of isolates to any of the recognized species described within the complex difficult. The identification of several differences between our isolates and a representative M. intracellulare strain at the 16S rDNA hypervariable sequence could explain the negative result obtained with the commercial species-specific probe (Table 1). All three MAC isolates were more closely related to M. intracellulare than to M. avium, even though they showed differences in the sequence of the 16S rDNA hypervariable region when compared to specific sequence of the M. intracellulare type strain (Fig. 1A). They can be differentiated from M. avium by the absence of the mig gene in their genomes (Fig. 3B) and also by lacking DT6 and IS1245. The absence of this insertion sequence is a rare event in M. avium (3). The isolates studied could also be distinguished from the MAIX group by their 32-kDa amino acid sequence (Fig. 2) and their susceptibility to antimycobacterial drugs (31). Our variant strain shares its 16S rDNA hypervariable sequence with a sequevar, described by an IWGMT study, that was classified as a member of the M. intracellulare species by other taxonomical approaches (33). Strains similar to our isolate have also been described by Devallois et al. (8). The hsp65 gene PCR restriction analysis electrophoretic patterns in all three isolates are identical to a representative M. intracellulare pattern (4, 9). A study by Swanson et al. of the hsp65 gene in the MAC (27) describes an allelic profile with a restriction map identical to that of our isolates (Fig. 3A). In the Swanson et al. study, that profile corresponds to strains either not reactive to any of the species-specific commercial probes or positive only to M. avium species-specific commercial probes (see Table 1, hsp65.6, in reference 27); these results are also shared by our MAC isolates (Table 1). Analysis of other genomic regions, such as DT1 and the ITS, clearly locate these isolates within the species M. intracellulare.

View larger version (46K): [in this window] [in a new window]   FIG. 3.   Ethidium bromide staining of PCR restriction analysis of hsp65 gene (A) and PCR amplification of mig gene (B). (A) hsp65 PCR restriction analysis patterns of mycobacteria obtained upon BstEII and HaeIII digestion. The sizes of fragments in base pairs are indicated on the left. L, molecular size marker (100-bp DNA ladder; Gibco BRL, Life Technologies); lane 1, M. avium DO22 clinical isolate; lane 2, M. avium ATCC 25291T; lane 3, DO67 clinical isolate; lane 4, DO68 clinical isolate; lane 5, DO69 clinical isolate; lane 6, M. intracellulare IWGMT 10 (23); lane 7, M. tuberculosis Mt14323. (B) mig PCR amplification. The size of the fragment in base pairs is indicated on the left. L, molecular size marker (1-kb Plus DNA ladder; Gibco BRL, Life Technologies); lane 1, negative control; lane 2, M. avium ATCC 25291T; lane 3, DO67 clinical isolate; lane 4, DO68 clinical isolate; lane 5, DO69 clinical isolate; lane 6, M. avium 26 C clinical isolate; lane 7, M. avium DO22 clinical isolate; lane 8, M. avium DO64 clinical isolate; lane 9, M. intracellulare GM51 clinical isolate; lane 10, M. intracellulare IWGMT 10 (23).

Differentiation of mycobacteria at the subspecies level has been based on several phenotypic characteristics, together with the host range distributions of the bacteria (2, 30). Some studies have detected a correlation of the genomic presence and distributions of insertion sequences with a particular host range in mycobacteria (19, 20, 23), indicating the usefulness of insertion sequence genomic distribution in the subspecies differentiation of M. avium. Unfortunately, only two insertion sequences have been described in M. intracellulare (GenBank accession numbers AJ011837 and L10239), and they have been poorly studied. The mycobacteria analyzed in the present work highlight the extensive internal variability present within the MAC and suggest that the isolates described here may represent a distinct subspecies of M. intracellulare.

Although M. intracellulare is usually implicated in infectious diseases in immunocompetent patients (11), this paper describes an unusual infection in an immunocompromised patient due to M. intracellulare. It is important to note the negative result we obtained by using species-specific commercial probes and the consequent inaccurate identification of our isolates. This misdirection could be averted by rapid and easy methods, such as analysis of the hsp65 gene, PCR detection of mig, or detection of DT1 or DT6 MAC-specific sequences. These second-line approaches would allow a more accurate identification of MAC isolates that show a positive signal with complex-specific commercial probes but are negative with both species-specific commercial probes. Some of those isolates could then be correctly identified as variants of either M. avium or M. intracellulare.