Detection of Mycobacterium avium in Blood Samples of Patients with AIDS by Using PCR

Blood specimens.

Sixty-nine samples from 47 AIDS patients of the Hospital of the Universidade Federal do Rio de Janeiro, a national reference center for AIDS, and of the Public Hospitals of Rio Grande do Sul, Brazil, were analyzed. The samples were collected in the period between 1996 and 1999. Fourteen out of 47 patients had more than one aliquot (5 ml) of whole blood collected in a period of 24 h, the number of samples collected being dependent on the seriousness of disease. The samples were collected at time intervals of between 8 and 12 h. Of these patients, six had two samples collected and eight had three samples collected during this period. This was possible because the patients were hospitalized. All samples were prepared by the lysis centrifugation technique.

Each aliquot was inoculated in Löwenstein-Jensen medium, and 500 μl was used for DNA purification followed by PCR amplification. The “gold standard” used was based on the clinical diagnosis for mycobacteremia caused by M. avium. The clinical diagnosis was based on the symptoms presented by the patients and on the isolation of the organism by microbiological methods and according to therapeutic criteria (resolution of the process during the treatment).

Isolation and DNA detection procedures.

The DNA was purified with DNAzol (Gibco BRL and Life Technologies, Rockville, Md.). The DNAzol procedure developed by Chomezynski et al. (4) is based on the use of a guanidine detergent lysing solution that hydrolyzes RNA and allows the selective precipitation of DNA from a cell lysate. DNA was amplified by PCR with the primers derived from the 16S rRNA sequence, mycgen-f (5′-AGAGTTTGATCCTGGCTCAG-3′), mycgen-r (5′-TGCACACAGGCCACAAGGGA-3′), and mycav-r (5′-ACCAGAAGA CATGCGTCTTG-3′), which have been described by Böddinghaus et al. (2). A product of 1,030 bp resulting from amplification with the primers mycgen-f and mycgen-r is indicative of the genus Mycobacterium, and a smaller fragment of 180 bp amplified by the mycgen-f and mycav-r primers is specific for M. avium. The samples were tested with specific primers for Mycobacterium tuberculosis derived from IS6110 as described by Hermans et al. (7). Amplified DNA was analyzed by electrophoresis in 1.5% agarose gel, subsequently transferred over the course of 2 h to nylon membranes by using a vacuum blotter system, and hybridized with specific probes for 3 h. The probes were produced by PCR, purified by using MicroSpin S-300 HR columns (Pharmacia Biotech, San Francisco, Calif.), and labeled with peroxidase by using the ECL direct system (Amersham Pharmacia Biotech, Little Chalfont, Buckinghamshire, United Kingdom). Hybridization was detected by using a reagent that decayed to hydrogen peroxide, the substrate for peroxidase, and a second reagent that was involved in light production (luminol).

The specificity of the primers for M. avium was confirmed by using DNAs from other mycobacteria and fungi that can cause symptoms similar to those of mycobacteriosis. Human DNA was also tested. No amplification was observed with DNA from the other organisms analyzed. DNA sequencing with a Thermo Sequenase radiolabeled terminator cycle sequencing kit (Amersham Pharmacia Biotech) showed that the sequence of the amplified fragment was identical to the sequence deposited in GenBank (M29572 ) by Stahl and Urbance (17). The analytical sensitivity was determined by amplification of the DNA extracted from blood specimens spiked with different amounts of M. avium previously grown in Löwenstein-Jensen medium. The limit of detection by PCR following hybridization was approximately 40 CFU with the species-specific primers and 3,000 CFU with the genus-specific primers.

DNA amplification in blood samples.

Among the 69 blood samples analyzed, 41 were culture positive for mycobacteria and 28 were negative. Although the clinical diagnoses have confirmed mycobacterial disease in all 41 cases, 19 were characterized as infections with M. avium, 16 as infections with M. tuberculosis, and 6 as coinfections with both mycobacteria. The DNA fragment that identifies the genus Mycobacterium (1,030 bp) was observed in 29.3% of the culture-positive samples and after the hybridization step was observed in 68.3% of the samples. The DNA fragment that identifies the species M. avium (180 bp) was observed in 60% of the samples and after hybridization was observed in 84% of the samples. Detection of the fragment that identifies M. tuberculosis (245 bp) was possible in 95.4% of the M. tuberculosis-positive samples both with agarose gel and by hybridization. All culture-negative samples were also negative by the method described in this work (Table 1). The low analytical sensitivity in detecting the genus Mycobacterium is in agreement with the results obtained by Kulski and Pryce (11). Four samples (16%) were false negatives, probably because the number of bacilli in the samples was below the limit of detection. We also found one false-positive result, in which an M. tuberculosis-positive sample was detected as M. avium positive after PCR and hybridization. In the coinfection cases indicated by PCR analysis, the culture identified only M. tuberculosis, while the patients presented symptoms indicating coinfection with M. tuberculosis and M. avium. It is possible that the overgrowth of M. tuberculosis covered the M. avium colonies, since the latter are smooth and translucent while M. tuberculosis colonies are bigger and dry (15).

On the other hand, the results can be analyzed by comparison of the molecular diagnoses (PCR and hybridization) of each of the 47 patients with culture results associated with clinical diagnosis. Fourteen patients had more than one blood sample analyzed, totaling 36 samples that were individually analyzed by PCR. Thirty-three were identified as mycobacteria positive by PCR. We observed that in 14.3% of the patients (two patients), mycobacterial disease was confirmed only when more than one sample was analyzed. These results show the importance of collecting more samples in a period of 24 h, because different bacteremia peaks occur in AIDS patients. One of the patients from whom only one sample was analyzed presented a negative result by molecular biology techniques, while the culture and clinical diagnosis were positive for M. avium. It is possible that the diagnosis would have been confirmed by PCR if more than one sample had been analyzed. Another patient was positive for M. tuberculosis by culture and clinical diagnosis, but the PCR result was positive for M. avium, characterizing a false result in the PCR. Three patients coinfected with M. avium and M. tuberculosis had positive results for both mycobacteria by the method described in this work. The results obtained in this study show that when each sample is analyzed individually, 91.3% are in accordance with the gold standard and that when each patient is considered, 95.7% are in agreement (Table 2).

In this work, we show that it is possible to detect the presence of mycobacteria by PCR followed by hybridization and that sensitivity is increased when more than one blood sample is analyzed from each patient. The developed diagnostic methodology presented here is expected to be incorporated into the routine procedures of clinical analysis laboratories, mainly those belonging to the public health systems, in developing countries. The procedure is rapid (1 day) and can be of remarkable importance in diagnosing mycobacterial disease and defining the most adequate treatment for patients.