Genetic Diversity of Mycobacterium avium Isolates Recovered from Clinical Samples and from the Environment: Molecular Characterization for Diagnostic Purposes

Abstract

Isolation of Mycobacterium avium complex (MAC) organisms from clinical samples may occur in patients without clinical disease, making the interpretation of results difficult. The clinical relevance of MAC isolates from different types of clinical samples (n = 47) from 39 patients in different sections of a hospital was assessed by comparison with environmental isolates (n = 17) from the hospital. Various methods for identification and typing (commercial probes, phenotypic characteristics, PCR for detection of IS1245 and IS901, sequencing of the hsp65 gene, and pulsed-field gel electrophoresis) were evaluated. The same strain was found in all the environmental isolates, 21 out of 23 (91.3%) of the isolates cultured from urine samples, and 5 out of 19 (26.3%) isolates from respiratory specimens. This strain did not cause disease in the patients. Testing best characterized the strain as M. avium subsp. hominissuis, with the unusual feature that 81.4% of these isolates lacked the IS1245 element. Contamination of certain clinical samples with an environmental strain was the most likely event; therefore, characterization of the environmental mycobacteria present in health care facilities should be performed to discard false-positive isolations in nonsterile samples, mainly urine samples. Molecular techniques applied in this study demonstrated their usefulness for this purpose.

The clinical relevance of Mycobacterium avium complex (MAC) has increased in recent decades because of the rising population of immunocompromised individuals due to longer life expectancy, immunosuppressive chemotherapy, and AIDS pandemia (28). In AIDS patients, MAC is the most common nontuberculous mycobacteria causing disease, especially in developed countries, and usually produces disseminated infections (8, 16). However, MAC also causes infection in immunocompetent individuals with underlying pulmonary disease and in children with lymphadenitis or cystic fibrosis (10, 14, 16).

Mycobacterial species comprised in MAC have undergone reorganization based on biological, serological, and molecular characteristics (Table 1). The most important species causing disease in humans can be differentiated using PCR-based methods for the detection of the presence/absence of insertion sequences (IS): M. avium subsp. avium (37) presents IS901 (7, 21, 27) and low copy numbers of IS1245 (13, 17), M. avium subsp. hominissuis (25) usually harbors high numbers of IS1245 but no IS901 (25), and Mycobacterium intracellulare (32, 42) presents neither IS901 nor IS1245 (2). Other MAC species include M. avium subsp. paratuberculosis, whose involvement in Crohn's disease is still under discussion (9), and M. avium subsp. silvaticum (37).

TABLE 1.

Characteristics of the most common MAC species causing human disease

Strain	Serotypes	Virulence for birds	IS901	IS1245	hsp65 code(s)a	Reference(s)
M. avium subsp. avium	1-3	Virulent	+	+	4	25, 37
M. avium subsp. hominissuis	4-6, 8-11, 21	Partially virulent	−	+	1-3, 7-9	25
M. intracellulare	7, 12-20, 22-28	Avirulent	−	−	10-12, 14	32

^aAccording to Turenne et al. (39).

In spite of the environmental colonization and rising prevalence of disease due to members of MAC, the epidemiology of infection is still not clear. The routes of transmission have not been unambiguously identified, and human-to-human transmission has not been demonstrated. Mycobacteria belonging to MAC have also been detected in a wide range of animal species, including birds (mainly M. avium subsp. avium) (27), pigs (more frequently M. avium subsp. hominissuis), (19), and small vertebrates (11).

The environment is the most likely source of infection for humans, as M. avium is considered a ubiquitous organism and has been isolated from water (1), soil, and dust (15). Water distribution systems have been reported as a possible source of infection in hospitals, homes, and commercial buildings (1, 41); moreover, hospital water can also contaminate samples of uninfected patients, making the clinical significance of the MAC isolation unclear (38).

Current diagnosis is based on probes or PCR tests that occasionally are not able to discriminate between MAC species or subspecies. Several molecular biology techniques can be used for epidemiological studies in order to establish measures to control the dissemination of the infection. For this purpose, restriction fragment length polymorphism (RFLP) using IS1245 as a probe (40) has been widely used; however, the effectiveness of IS1245 RFLP is limited in M. avium strains harboring a low number of IS1245 copies.

Pulsed-field gel electrophoresis (PFGE) has also been shown to be a useful technique for assessing the relatedness of strains among MAC (24). Nevertheless, PFGE requires a large amount of bacterial DNA for good visualization of the patterns. Other characterization techniques include sequencing of the 65-kDa heat shock protein gene (hsp65) (29, 39) and of the internal transcribed spacers (ITS) (26).

The aim of this study was to characterize the environmental strains of MAC present in a hospital using different molecular tests (IS1245 and IS901 detection, hsp65 sequencing, and PFGE analysis) and to compare them with clinical isolates from patients from the same hospital in order to evaluate the possibility of nosocomial infections or contamination of samples.

MATERIALS AND METHODS

A total of 64 isolates belonging to MAC from a public hospital were analyzed. This hospital covers a population of 525,000 inhabitants and is located in Madrid (Spain). Forty-seven isolates were recovered from clinical samples obtained from patients from different sections of the hospital. The other 17 isolates were cultured from several environmental sources (tap water from various floors of the hospital and containers for the collection of urine samples). All the isolates were cultured in the period between 2002 and 2005.

Clinical M. avium isolates.

Forty-seven MAC isolates were recovered from 12 AIDS patients and 27 patients not infected with human immunodeficiency virus (HIV). Initially, all patients showed clinical syndromes compatible with mycobacterial diseases (pneumonia, meningitis, pericarditis, etc.) though active mycobacterial infection was afterwards confirmed in only 13 patients based on clinical criteria. The isolates were processed according to Tacquet and Tison (34) and cultured in Coletsos medium (BioMedics S.L., Tres Cantos, Madrid, Spain). Among these 47 isolates, 13 were sequential isolates cultured from identical or nonidentical sites from five patients (patient numbers 17, 18, 19, 35, and 36), and the rest were single isolates from 34 patients. Thirty-eight isolates were cultured from admitted patients, while the remaining nine isolates were obtained from nonadmitted patients. Most of the isolates were cultured from nonsterile sites (urine, n = 23; sputum, n = 19) except for five isolates (blood, n = 3; liver biopsy, n = 1; bone marrow, n = 1).

Environmental M. avium isolates.

Eleven isolates were cultured from the intermediate containers used for the collection of urine samples in disabled patients and six were from tap water taken from the different patient's rooms at the hospital. Fifty milliliters of water was collected from the water taps of the patients' rooms into sterile tubes; the intermediate containers were cleaned following the standard procedure and then rinsed with 50 ml of sterile water that was afterwards collected into sterile tubes. These samples were centrifuged at 3,500 rpm for 30 min; the pellet was decontaminated according to Tacquet and Tisson (34) and centrifuged again at 3,500 rpm. The pellet was then inoculated into Löwestein-Jensen medium supplemented with amphotericin (400 mg/liter), nalidixic acid (35 mg/liter), and lincomycin (2 mg/liter) and into Coletsos (BioMedics S.L.) medium. Media were incubated at 36°C for up to two months.

Primary identification.

Isolates were identified as MAC members by using the commercial AccuProbe MAC culture identification kit (Gen-Probe, CA) and by phenotypical characteristics: growth temperature at 25°C, 36°C, and 43°C; tellurite reduction, Tween 80 hydrolysis; nitrate reduction test; and sodium chloride tolerance (18).

PCR analysis.

Bacterial cells were suspended in 200 μl of sterile water and boiled for 10 min; then they were centrifuged, and the supernatant was used as the DNA template for the PCRs. The identification of the isolates as M. avium or M. intracellulare was carried out by PCRs aimed at specific regions of the 16S rDNA (5, 43). PCR amplification of the insertion sequences IS901 and IS1245 was performed as described previously (13, 20). The PCR and sequencing of the hsp65 gene were carried out as described previously in M. avium isolates (29, 35, 39).

For confirmation of the identity of the IS1245-negative M. avium isolates, 12 isolates were randomly selected (seven from urine samples and five from environmental sources) and subjected to PCR amplification and restriction enzyme analysis (REA) of IS1311; amplified DNA was digested with the restriction enzymes HinfI and MseI (New England Biolabs, Hitchin, Hertfordshire, United Kingdom) as described elsewhere (23). Analysis of the 16S rDNA and 16S-23S rDNA ITS sequences was performed in four isolates (two environmental cultures and two isolates from urine samples) (5, 31). PCR primers used in all reactions are described in Table 2. Positive and negative controls were included in each batch of tests.

TABLE 2.

Primers used in this study for the identification and characterization of the MAC isolates

Name or target	Directiona	Sequence (5′ to 3′)	Reference(s)
Mycgen-F/R	F	AGAGTTTGATCCTGGCTCAG	5, 43
	R	TGCACACAGGCCACAAGGGA
Mycint	F	CCTTTAGGCGCATGTCTTTA	43
	R	TGCACACAGGCCACAAGGGA
IS901	F	GCAACGGTTGTTGCTTGAAA	20
	R	TGATACGGCCGGAATCGCGT
IS1245	F	GCCGCCGAAACGATCTAC	13
	R	AGGTGGCGTCGAGGAAGA
IS1311	F	GCGTGAGGCTCTGTGGTGAA	23
	R	ATGACGACCGCTTGGGAGAC
ITS	F	TTGTACACACCGCCCGTCA	31
	R	TCTCGATGCCAAGGCATCCACC
hsp65 3′ fragment	F	CGGTTCGACAAGGGTTACAT	39
	R	ACGGACTCAGAAGTCCATGC
hsp65 5′ fragmentb	F	ACCAACGATGGTGTGTCCAT	35
	R	CTTGTCGAACCGCATACCCT

^aF, forward; R, reverse.

^bTelenti fragment.

PFGE analysis.

M. avium subsp. hominissuis isolates were typed by PFGE analysis as described before (24). Briefly, bacterial cells were cast into low-melting-point agarose plugs and lysed with lysozyme (1 mg/ml), 1% sodium dodecyl sulfate, and proteinase K (1 mg/ml). Plugs containing DNA were digested with 20 U of XbaI for 4 h (New England Biolabs). The digested DNA was loaded in a 1% agarose gel and run in Tris-borate-EDTA buffer. PFGE was performed using a CHEF-DR III System (Bio-Rad Laboratories Ltd.) at 14°C for 20 h at 200 V. Gels were stained with ethidium bromide and photographed under UV light. The PFGE patterns obtained were interpreted as proposed by Tenover et al. (36); isolates were considered closely related if differences of no more than 2 to 3 bands were observed between the obtained PFGE profiles.

RESULTS

PCR results.

All 64 isolates were identified as belonging to the Mycobacterium genus based on the results of the AccuProbe test and on the detection of specific regions of the 16S rDNA. Fifty-five isolates were identified as M. avium, and nine were identified as M. intracellulare by detection of a specific 16S rDNA fragment. The IS901 PCR gave negative results in all 55 M. avium isolates under study. IS1245 was detected in 11 M. avium isolates, all from clinical samples (Table 3, Table 4, and Table 5).

TABLE 3.

Characterization of M. avium subsp. hominissuis isolates cultured from urine samples (n = 23) from 20 patients

Patient no.	Isolate reference	IS1245	hsp65 sequevar	PFGE profilea	Infected with MACb	Comment
1	MI05/02099	−	NAc	NA	No
2	MI05/02102	−	Code 2	A*	No
3	MI05/02103	−	Code 2	A	No
4	MI05/03003	−	Code 2	A	No	HIV
5	MI05/03005	−	Code 2	A	No
6	MI05/03006	−	Code 2	A	No
7	MI05/03007	−	Code 2	A	No	HIV
8	MI05/03009	−	Code 2	A	No
9	MI05/03010	−	Code 2	A	No
10	MI05/03011	−	Code 2	A	No
11	MI05/03012	−	Code 2	A*	No	HIV
12	MI05/03014	−	Code 2	A	No
13	MI05/03015	−	Code 2	A	No
14	MI05/03017	−	Code 2	A*	No
15	MI06/00268	−	Code 2	A	No
16	MI06/00269	−	Code 2	A	No
17	MI05/02104	−	Code 2	A	No
	MI05/03002	−	Code 2	A	No
18	MI05/02100	−	NA	NA	No
	MI05/02101	−	Code 2	A	No
	MI06/00263	−	Code 2	A	No
19	MI06/00266d	+	Code 1	B	Yes	HIV
20	MI06/00272	+	Code 1	Unrelated	Yes	HIV

^a*, pattern closely related to PFGE pattern A.

^bBased on clinical criteria.

^cNA, not available.

^dPatient with MAC isolates cultured from urine, sputum, and blood samples.

TABLE 4.

Characterization of M. avium subsp. hominissuis isolates cultured from sputum samples (n = 11) from 11 patients

Patient no.	Isolate reference	IS1245	hsp65 sequevar	PFGE profile	Infected with MACb	Comment
22	MI06/00271	+	Code 1	Unrelated	No	HIV
23	MI06/03008	+	Code 2	A	No	HIV
24	MI07/01955	−	Code 2	NAa	No
25	MI07/01958	−	Code 2	NA	No
26	MI07/01959	−	Code 2	NA	No
27	MI07/01960	−	Code 2	NA	No	HIV
28	MI05/03018	−	Code 2	A	No	HIV
29	MI06/00273	+	Code 2	A	Yes
30	MI07/01954	+	Code 9	NA	Yes
33	MI06/00262	−	Code 15	Unrelated	Yes
19	MI06/00267c	+	Code 1	B	Yes

^aNA, not available.

^bBased on clinical criteria.

^cPatient with MAC isolates cultured from urine, sputum, and blood samples.

TABLE 5.

Characterization of M. avium subsp. hominissuis isolates cultured from sterile samples (n = 4) from three patients

Patient no.	Isolate reference	Sample	IS1245	hsp65 sequevar	PFGE profilea	Infected with MACb	Comment
37	MI06/00264	Liver biopsy	+	Code 3	Unrelated	Yes	HIV
39	MI05/03004	Blood	+	Code 1	Unrelated	Yes	HIV
19	MI06/00265c	Blood	+	Code 1	NA	Yes	HIV
	MI06/03278c	Blood	+	Code 1	NA	Yes	HIV

^aNA, not available.

^bBased on clinical criteria.

^cPatient with MAC isolates cultured from urine, sputum, and blood samples.

Clinical isolates.

The 23 isolates cultured from urine samples from 20 patients were identified as belonging to MAC, but only two of them harbored IS1245 (Table 3). Sequencing of the 3′ end of the hsp65 gene was performed on 21 isolates, 19 IS1245-negative isolates, and the 2 IS1245-positive isolates. The code 2 hsp65 sequevar (39), a typical sequevar of M. avium subsp. hominissuis, was obtained in all 19 isolates lacking IS1245. The two IS1245-positive isolates contained a code 1 sequevar. All of the IS1245-negative code 2 isolates were cultured from samples from patients who were later considered by clinical criteria not to be infected with MAC (Table 3). The two patients from whom the IS1245-positive isolates came were the only ones considered infected by MAC. In the first case, M. avium was also isolated from blood and sputum of the same patient; all three isolates shared the same code 1 hsp65 sequevar, and the two PFGE-typed isolates shared the same profile. Thus, this patient was thought to have a disseminated infection. In the second case, the patient had a history of repetitive M. avium isolations in urine samples obtained using a catheter; the strain isolated presented a code 1 in the hsp65 analysis, and its PFGE profile was considered unrelated to all the others obtained in the study.

Nineteen MAC isolates were cultured from sputum samples of 17 patients. Eight isolates were identified as M. intracellulare, seven of which were causing disease in immunocompetent patients, while one isolate could not be associated with any pathology. The remaining 11 isolates were identified as M. avium (Table 4), and four of these isolates were associated with disease in the patients (three of these isolates harbored IS1245 and contained different hsp65 sequevars [codes 1, 2, and 9], and one was IS1245 negative and contained a code 15 hsp65 sequevar [GenBank accession no. EU085419]). Seven M. avium isolates came from patients not infected with MAC, and five were IS1245 negative and contained a code 2 hsp65 sequevar. The other two were IS1245 positive and showed different hsp65 codes (1 and 2).

The three isolates from blood samples, two from the same patient, contained the IS1245 element and had the code 1 hsp65 sequevar. The strain coming from a liver biopsy presented the IS1245 element and had a code 3 hsp65 sequevar (Table 5). Finally, the isolate from the bone marrow sample was identified as M. intracellulare. All of the isolates cultured from sterile samples were causing disease.

Environmental isolates.

All of the 17 isolates were identified as M. avium. The IS1245 element was absent in all of them. The sequencing of the hsp65 3′ end revealed the same sequevar (code 2) in 16 strains (no signal was obtained in 1 strain).

IS1311 REA.

In the 12 IS1245-negative M. avium isolates analyzed by means of the IS1311 PCR and subsequent digestion with HinfI and MseI, the expected fragments for the M. avium 285- and 323-bp bands and 419- and 189-bp bands, respectively, were obtained; thus, this confirmed the identification of the IS1245-negative isolates as M. avium strains.

Four IS1245-negative isolates were also subjected to 16S rDNA and ITS sequencing, and identical sequences were obtained for the four strains in both fragments. The 16S rDNA sequence from the four isolates showed one single-nucleotide polymorphism (T instead of G at nucleotide 445) with the M. avium 104 strain sequence (GenBank accession CP000479). The ITS sequence obtained matched perfectly with the Mav-A sequevar described previously (6, 12).

PFGE results.

Thirty-eight isolates yielded readable patterns after digestion with the XbaI restriction enzyme: 29 from clinical samples (21 from urine samples, 6 from sputum samples, and 1 from a blood and a liver biopsy sample) (Table 3, Table 4, and Table 5) and 9 from environmental samples. Nine unrelated patterns were observed. All of the environmental isolates, 19 out of 21 isolates from urine, and 3 out of 6 from sputum yielded identical (pattern A) or closely related (A*) patterns. All of these isolates also contained a code 2 hsp65 sequevar and, with the exception of two isolates cultured from sputum samples, harbored no copies of IS1245. A second pattern (pattern B) was found in two isolates coming from sputum and urine samples from the same patient. The remaining isolates showed unrelated patterns.

DISCUSSION

This report describes the usefulness of several molecular techniques in the correct diagnosis and study of the epidemiology of MAC infections. Genetic tools, including commercial kits, for identification of MAC members are now available to mycobacteriology laboratories. Still, many clinical laboratories just identify the isolates as MAC members, as the treatment is the same for all MAC infections (4). Therefore, only incomplete epidemiological studies can be carried out. In our study, a combination of techniques was required to precisely identify the isolates to a subspecies level and to assess their relatedness.

IS1245 is considered to be highly specific to M. avium (13), so it is commonly used for identification and typing (40). However, M. avium strains lacking IS1245 were first described by Ritacco et al. (30) and were later reported as being isolated from human or animal sources (3, 19, 22); nonetheless, they were considered a rare finding and were always found in low proportions compared with the IS1245-positive strains. In most cases, characterization of IS1245-negative isolates is based only on commercial probes that might not identify the isolate unambiguously to the species-subspecies level. In our study, the identity of the M. avium strains lacking IS1245 was confirmed by 16S rDNA, ITS, and hsp65 sequencing and by IS1311 REA and represented 81.8% of all M. avium isolates cultured, regardless of the source. The hsp65 sequencing was a useful tool for confirmation of species identification of IS1245-negative M. avium isolates as previously described (4). Though PFGE and IS1245 RFLP had been shown to have similar discriminative powers, PFGE is the only applicable typing technique for these isolates.

Molecular tests demonstrated that the same M. avium strain lacking IS1245, with a code 2 hsp65 sequevar and a certain PFGE profile, pattern A, was isolated from all environmental samples, most of the urine samples (21 out of 23, or 91.3%), and several respiratory specimens (5 out of 19, or 26.3%).

The clinical significance of the isolation of MAC members from nonsterile human samples is a controversial issue. In this way, MAC-positive culture has been shown to occur in respiratory specimens from patients without clinical disease (38); this could happen due to contamination of the sample or transient or persistent colonization. Likewise, results of direct observation of mycobacteria by acid-fast staining in urine samples should be interpreted carefully.

The molecular characterization correlated with clinical significance in that the 18 patients in whose urine the IS1245-negative code 2 hsp65 sequevar was found were considered not infected by MAC by clinical criteria. On the other hand, the two M. avium isolates cultured from urine samples that harbored IS1245 and had a code 1 hsp65 sequevar came from two patients likely to have an M. avium infection: the first patient was thought to have a disseminated infection, and the second had a history of repetitive M. avium isolations in urine.

The explanation of the IS1245-negative code 2 hsp65 sequevar isolations from certain urine samples could be attributed to the collection method: in this hospital, patients collect their urine samples into sterile receptacles, but intermediate containers might be used when the patient is handicapped. These intermediate containers are washed with tap water and soap after every use but not always sterilized while used by the same patient. The environmental isolates were cultured from the sediment obtained from the intermediate containers or from the taps of the patient's rooms. Thus, it is likely that urine samples collected using these containers would be contaminated by the environmental MAC strains. MAC members have been shown to grow in biofilms, increasing their resistance to disinfection methods such as chlorination (33); the most prevalent strain found in our study might persistently colonize the water supply system of the hospital as it was isolated from water samples from different floors of the hospital during a 3-year period.

The possibility of contamination of urine samples during the collecting process in the hospital is enhanced by the fact that, in samples collected in patient's houses, the recovery rate of M. avium is lower. In this context, in the period 1996 to 2003, 3,195 urine samples from admitted patients were cultured for the presence of mycobacteria, and 79 (2.47%) were positive. In the same period, 3,486 samples from outpatients were also collected and analyzed, and only 4 (0.11%) yielded positive isolation of MAC (E. Gomez-Mampaso, unpublished data).

Unlike the urine sample results, in the sputum samples the IS1245-negative code 2 hsp65 sequevar strain represented only 5 out of the 19 MAC isolates recovered (again, they were considered as nonpathogenic). Sputum samples are collected directly into sterile receptacles, and therefore contamination of these samples is unlikely to occur; however, patients can be exposed to aerosols from shower and tap water from the bathrooms. If the water harbors M. avium, it could be a possible source of contamination.

Regarding M. intracellulare, eight isolates were recovered from respiratory samples and one from bone marrow. All except one (isolated from sputum) were found to be disease causing, highlighting the importance of this bacteria as a human pathogen.

Hospital water has been reported as a possible source of MAC infection (1, 38, 41). In these studies some degree of relationship was found between environmental M. avium and strains isolated from clinical samples (sputum, blood, or bone marrow). However, tap water was not usually added during the collection of the samples from which the cultures were recovered, and therefore contamination with environmental M. avium strains was considered unlikely. In our setting, we consider more possible a contamination of certain samples due to an incorrect collection method in some special cases of disabled patients. This hypothesis is supported by the fact that most of the patients (15 out of 20, or 75%) with positive isolations of MAC in urine were immunocompetent and were considered afterwards uninfected by MAC based on further medical analysis. Therefore, the environmental strain seemed to have low virulence since, even though it was widely distributed, it was not found to be pathogenic in any patient. However, the risk of infection by this strain cannot be ruled out, particularly in the case of HIV patients. To our knowledge, there are no studies that correlate the absence of IS1245, the hsp65 sequevar, the PFGE pattern, and virulence, and therefore further research is needed to find out if these results can be extrapolated to other settings.

In summary, these results highlight the possibility of false-positive acid-fast staining or isolations of MAC members in nonsterile clinical samples, especially in the case of urine specimens. Therefore, it is important to characterize the environmental mycobacteria present in the facilities in order to exclude false-positive isolation in samples susceptible to contamination. For this purpose a combination of molecular techniques is required to correctly characterize MAC isolates. In our study we have demonstrated the presence of a particular strain persistently colonizing the water supply system and representing a potential source of contamination of samples and of infection in patients.