ABSTRACT

Background: Various primers are being tested for the detection of Mycobacterium tuberculosis DNA. The accuracy of the polymerase chain reaction (PCR) depends on the target sequence used and whether the test will be performed in culture or in clinical specimens. Objectives: To identify DNA sequences, specifically those commonly reported as targets for diagnosis of tuberculosis (TB), in clinical samples of M. tuberculosis strains. Method: Eighty-one clinical samples from suspected TB patients were initially processed and submitted to bacilloscopy (smear) and culture, and PCR was performed with specific primers for the following targets: IS 6110, 65 kDa, 38 kDa and MPB64. Results: Smear and culture results were negative in 24 samples, as was the PCR. The 19 samples testing smear positive, as well as the isolated strains, were 100% positive on PCR, with the exception of the 89.4% result from PCR with MPB64 primers. In the 38 smear negative and culture positive samples, PCR results were inconsistent. The primers specific for amplifying the 123 bp IS 6110 fragment gave the highest positivity (92.1%), diagnostic agreement (0.9143), co-positivity (94.7%) and co-negativity (100%). Conclusion: The IS 6110, 38 kDa, MPB64 and 65 kDa sequences were found in the genome of all M. tuberculosis strains isolated in patients from the state of Amazonas. The protocol for processing the clinical samples prior to PCR analysis and the specific primers used to amplify the 123bp IS 6110 fragment showed a greater efficiency in diagnosing pulmonary (paucibacillary) tuberculosis in comparison to published data.

Keywords: Primers/PCR. Diagnosis/Tuberculosis. Mycobacterium tuberculosis.

RESUMO

INTRODUÇÃO: Há diferentes primers sendo testados para a detecção do DNA do Mycobacterium tuberculosis. A acuidade da reação em cadeia da polimerase (PCR) depende da existência da seqüência alvo no bacilo e de os testes serem realizados em cepas isoladas ou em amostras clínicas. OBJETIVO: Verificar a presença das seqüências de DNA alvo mais relatadas na literatura para o diagnóstico da tuberculose em amostras clínicas usando como controle positivo as respectivas cepas de M. tuberculosis isoladas. MÉTODO: Oitenta e uma amostras clínicas de pacientes com suspeita de tuberculose foram submetidas à baciloscopia e cultivo. A técnica de PCR foi realizada nas amostras clínicas e cepas isoladas com primers específicos para os seguintes alvos: IS6110, 65 kDa, 38 kDa e MPB64. RESULTADOS: Em 24 amostras com baciloscopia e cultivo negativos, a PCR também foi negativa com todos os primers testados. Em 19 amostras com baciloscopia positiva e nas cepas isoladas obteve-se 100% de resultados positivos nas PCR, exceto nas PCR em amostras clínicas com os primers para a seqüência MPB64 (89,4%). Em 38 amostras com baciloscopia negativa e cultivo positivo, as PCR tiveram resultados variáveis, sendo que os primers específicos que amplificam o fragmento de 123 pb da seqüência IS6110 foram os que forneceram os maiores percentuais de positividade (92,1%), concordância diagnóstica (0,9143), co-positividade (94,7%) e co-negatividade (100%). CONCLUSÃO: As seqüências IS6110, 38 kDa, MPB64 e 65 kDa foram encontradas no genoma de todas as cepas de M. tuberculosis isoladas desses pacientes do Estado do Amazonas. O protocolo utilizado no processamento das amostras clínicas e os primers específicos utilizados para amplificação do fragmento de 123 pb da seqüência IS6110 demonstraram maior eficiência no diagnóstico da tuberculose pulmonar (paucibacilar) em comparação com a literatura.

Palavras-chave: Primers/PCR. Diagnóstico/Tuberculose. Mycobacterium tuberculosis.

Introduction

Polymerasechain reaction (PCR) is a molecular technique whose accuracy is determined bythe choice of the target DNA and the definition of the primers within the DNAsequence. The PCR method has been used as an alternative that presents highsensitivity and specificity for the rapid diagnosis of infectious diseases. However, the use of PCR in the detection of Mycobacteriumtuberculosis (Mtb) has produced varying results, especially in relation tothe sensitivity of the test(1,2,3). Various targetsequences, such as insertion sequence 6110(IS6110), 65 kDa (GroEL), 38 kDa(PhoS, CIE Ag78 or Pab) and MPB64 (23 kDa), have therefore been used. Amongthese, IS6110 is more commonly usedbecause it is a repetitive sequence in the Mtb genome. This characteristichelps increase the sensitivity of PCR over that obtained in the amplificationof single DNA sequences(4). However, the IS6110 sequence was reported to be absentfrom an Mtb strain isolated in India(5), and homologous sequences have been detected in potentially pathogenicmycobacterium strains such as M. intracellulare, M. fortuitum, M. kansasii, M. xenopi, M. malmoenseand M. chelonae(6,7). In Brazil, the existence of Mtb strains lacking IS6110 has never been reported, and someof the species mentioned above are frequently isolated in the state of Amazonas(8,9,10).

Thesefacts led us to ask the following questions regarding the use of PCR for thediagnosis of tuberculosis (TB) in the state of Amazonas: Would the sequencesreported in the literature also be found in the Mtb strains isolated in the stateof Amazonas? Which primers would be most recommended for PCR of clinicalsamples collected in the state of Amazonas?

Inan attempt to address these questions, we investigated the presence of the DNAtarget sequences most commonly reported in the literature. A set of Mtbstrains, as well as the clinical samples from which the strains had beenisolated, were analyzed.

Method

Clinicalsamples were collected from 81 suspected TB patients living in the state ofAmazonas and were sent to the Laboratory of Mycobacteriology of the Instituto Nacional de Pesquisas da Amazônia(INPA, National Research Institute of Amazônia) for study. The samples weresubmitted to acid-fast bacilli (AFB) smear microscopy and Mtb culture. Smear microscopy and culture resultswere negative in 24 samples and positive in 19 samples (multibacillarysamples). In 38 samples, the smear microscopy was negative and the culture waspositive (paucibacillary samples).

Smearmicroscopy was performed in accordance with the guidelines of the Ministério da Saúde (Ministry of Health)(11). Cultures were performed as follows: samples weretransferred to graduated 15-mL conical centrifuge tubes and an equal volume of4% NaOH solution was added. The samples were agitated and left to settle for 15minutes. Sterile distilled water was added up to a combined volume of 12 mLwith. Samples were then centrifuged at 3000 x g and the supernatant was setaside. The sediment was suspended in 2 mL of sterile distilled water,neutralized with 4% HCl solution, and 0.2-mL aliquots were transferred forculture in Löwenstein-Jensen culture medium (three tubes per sample). Theremainder of the suspension was stored for later PCR analysis. The culturetubes were incubated at 37ºC for up to 60 days. Identification of the isolatedmycobacterium strains as Mtb was performed as recommended by David et al. (12)

Suspensionsfrom the clinical samples and from the 57 strains isolated from these sampleswere analyzed using PCR in order to identify the following targets: IS6110, 65 kDa, 38 kDa and MPB64. Twodistinct pairs of primers were tested for the detection of the IS6110 region: one that promotesamplification of a 123-bp fragment(4) and another that promotes that of a 541-bp fragment(13). Amplification of nucleotide sequences of the targets65 kDa and 38 kDa was performed through two consecutive PCRs - PCR and nestedPCR - using two pairs of specific primers for each test. When the target was 65kDa(14), PCR of the external primerpair amplified a fragment of 383 bp, whereas nested PCR of the internal primerpair amplified a fragment of 155 bp. When the target was 38 kDa(15), amplified products from standard PCR and nested PCRproduced fragments of 419 bp and 322 bp, respectively. Primers for the MPB64region amplified a fragment of 240 bp(16).

DNA extraction from sample sediment was performed in accordance with the Ogusku et al. protocol(17). The DNA was purified with phenol-chloroform extraction and absolute ethanol precipitation, in accordance with the Davis et al. method(18). For DNA extraction from the 57 Mtb strains isolated from the culture samples, we used an adaptation of the Perosio et al. (19) protocol, in which approximately 1 mg of each strain suspension was transferred into a screw cap tube containing glass beads. The samples in these tubes were agitated on a vortex mixer for cell separation, and sterile distilled water was added to adjust the turbidity of the suspension to a 1.0 McFarland standard. From each suspension, a 50-mL aliquot was used for DNA extraction. To each aliquot, 50 mL lysis buffer solution (200 mM Tris-HCl pH 8,0, and 800 mg/mL Proteinase K) was added. Aliquots were incubated at 37° C overnight and subsequently at 100° C for 10 minutes, then centrifuged at 14.000 x g for 15 seconds. We used 5 mL of the supernatant to perform PCR with the various primers.

We verified the intrinsic sensitivity of strains in accordance with the protocol devised by Bollela et al. (20) in order to establish the minimum concentration of bacilli that would be detectable by PCR with the various primers. Therefore, a suspension of Mtb H37Rv was prepared and adjusted to a 1.0 McFarland standard turbidity level (equivalent to 3 x 108 bacilli/mL). This suspension was subjected to successive 1:10 dilutions in sterile distilled water - up to a concentration of 3 x 10-1 bacilli/mL. For DNA extraction, each dilution was heated at 100ºC for 10 minutes and centrifuged at 13,500 rpm for 10 minutes. We then used 5 mL of the supernatant for PCR. Intrinsic specificity of primers was verified in relation to the DNA of the following microorganisms: Mycobacterium fortuitum, Mycobacterium avium-intracellulare, Shigella sonnei, Salmonella paratyphi A, Escherichia coli and Staphylococcus aureus. The above-mentioned protocol was used for DNA extraction from these species.

The PCR for each target sequence in the DNA extracted from clinical samples and from the isolated strains was performed using a solution containing 10 mM of Tris-HCl at pH 8.3, 50 mM of KCl, 0.01% gelatin; 2 mM of MgCl2, 200 mM of dNTP (Sigma, St. Louis, MO, USA), 0.1mM of each primer, 2U of Taq DNA Polymerase (Invitrogen, Melbourne, Australia) and 5 mL of the extracted DNA (from the isolated strain, clinical sample or PCR product) - in a final volume of 50 mL. Primer sequences for the respective target DNA and amplification parameters are described in Table 1. Primers were synthesized by Invitrogen, and we used a GeneAmp PCR System 2400 thermal cycler (Applied Biosystems, Foster City, CA, USA). A positive control containing Mtb H37Rv DNA and a negative control (reaction without DNA) were incorporated into each DNA amplification session.





Amplifiedproducts were electrophoresed on a 1.5% agarose gel. The agarose gel wassubsequently stained with ethidium bromide and PCR products were displayedusing an Eagle Eye II ultraviolet transilluminator (Stratagene, La Jolla, CA,USA).

Resultswere analyzed using the Kappa index to determine concordance between the twodiagnostic methods (culture and PCR) and among the various primers. Co-positivityand co-negativity were analyzed using 2 X 2 tables.

Results

Intrinsicsensitivity of the primers tested extended to a dilution of 3 x 100, that is, 3 bacilli/mL, except for MPB64 primers,which showed lower sensitivity since positive results were only obtained at adilution of 3 x 102 bacilli/mL. Specificity of the various primers was considered high since no amplificationoccurred in DNA extracted from M. fortuitum, M. avium-intracellulare, S. sonnei, S. paratyphi A, E. coli and S. aureus.

Inthe sediment of clinical samples that were negative in smear microscopy andculture, PCR results were also negative with all primers tested. On the otherhand, all AFB smear-positive clinical samples, as well as all isolated strains,were 100% positive in PCR, with the exception of the 89.4% result from PCR withMPB64 primers. Clinical samples that were negative for AFB, as well as thestrains isolated from them, presented inconsistent PCR results depending on theprimers tested. Percentages of positivity for these samples are shown in Figure 1.





Table 2 shows the results of the concordance analysis between TB diagnosis throughculture and that made by PCR in relation to the various primers studied, aswell as the results of co-positivity and co-negativity testing, both using theisolation of Mtb as a diagnostic standard.

Concordanceof TB diagnoses using PCR with the various primers studied, as well asdiagnoses made through co-positivity and co-negativity testing, both using IS6110 primer as a diagnostic standard dueto its having the highest percentage of positivity, are shown in Table 3.

Discussion

Ourresults show that the target sequences IS6110,65 kDa, 38 kDa and MPB64 were preserved in all strains studied and that therespective primers studied can be used in the diagnosis of TB using PCR inclinical samples. In multibacillary samples, only primers for the MPB64sequence (240 bp) were less than 100% efficacious, amplifying only 89.4% ofsuch samples. In paucibacillary samples, percentages of positivity varied dependingon the primers studied, as shown in Figure 1.

Thepercentage of positivity for the amplification of the 123-bp fragment of targetIS6110 in the present study (92.1%)was higher than that reported by Nolte et al. (21), Shawar et al. (22) or Montenegro etal. (2), who obtained 57.0%, 53.0%,and 76.7%, respectively. When we used primers that amplified a 541-bp fragmentof target IS6110, we obtained 86.8%positivity. The proportions we noted were higher than those reported by Abe et al. (13), who obtained only 50% with these primers. Using 65-kDa primers,positivity after nested-PCR (65 kDa - 155 bp) was 76.3%, higher than thatreported by Pierre et al. (14), who reported only40%. As for 38-kDa primers, nested-PCR positivity was 86.8%, higher than the38.4% reported by Miyazaki et al. (15). The PCR for MPB64presented the lowest positivity (60.5%) in paucibacillary samples. Our resultwas lower than the 70% reported by Martins et al. (23), after the inclusion of the nested-PCR step. However, in the presentstudy, this step was not performed for MPB64.

The differences in positivity percentages between the present study and those inthe literature may be related to slight changes in the adopted protocols, suchas sample decontamination processes and variations in the composition of lysisbuffer solutions. These results corroborate the statement made by Butcher etal. (24) to the effect thatPCR positivity varies widely in samples that test negative in smear microscopyand positive in culture for Mtb, towhich we add that variations in PCR protocols may also affect positivity.





Concordanceof diagnosis, co-positivity and co-negativity among primers and cultures, usingMtb isolation as a standard (Table 2), shows that the primers amplifying the123-bp IS6110 fragment are highly indicated for the diagnosis of TB fromclinical samples, independently of the smear microscopy results. Althoughdiagnosis of TB was not confirmed through the use of these primers in 3 samplesproven to contain Mtb (94.7% co-positivity), no false-positive results werefound (100% co-negativity). It is important to highlight the fact that nofalse-positive results were found with any of the primers tested.

Inthe PCR analysis of the samples studied, only one of the primers specific forthe amplification of the 123-bp IS6110 fragment was unable to detect MtbDNA, which was only detected by the primers amplifying the 541-bp IS6110, the65 kDa and the 38 kDa target fragments, as shown in Table 3 (analysis ofconcordance between primer amplifications of the DNA). This occurred in apaucibacillary sample with the lowest number of isolated colonies per sample(four colonies in the three Löwenstein-Jensen culture tubes).

Multiplecopies of the IS6110 sequence arefrequently present in the Mtb genome. This results in higher PCR sensitivitywhen compared to amplifications of single copy sequences (38 kDa, 65 kDa, andMPB64), even when these are submitted to two consecutive PCR reactions, such asin the PCR-nested PCR technique. The high percentage of positivity in thesamples isolated in the state of Amazonas may indicate that the Mtb strainspresent a higher frequency of copies of this sequence in their genome, whichcan be verified through the restriction fragment length polymorphism. Thispremise, as well as the differences in PCR protocols, may explain the fact thatproportions of positivity found in the present study were higher than thosereported in the literature.

Ourresults show that the primer pairs studied are useful for the identification ofMtb genome sequences in colonies isolated from cultures, a strategy recommendedby Kontos et al. (25). Theidentification protocol using PCR within 24 to 48 hours after culture becomesmore essential when there is an insufficient number of colonies since, in sucha case, identification tests can only be performed after subculture, delayingthe identification of the mycobacterium by at least 21 days.

Despitethe small number of samples studied, we can conclude that, in the state ofAmazonas, IS6110 (123-bp) primers arethe most highly recommended for PCR in clinical samples and in Mtb strainsobtained from health care services that perform laboratory diagnosis of TB. Inorder to confirm these results, a greater number of samples and strains shouldbe analyzed using the primers studied and that other laboratories in the stateshould also begin to employ this technique.

References

1. García-Quintanilla A, Garcia L, Tudó G, Navarro M, González J, Anta MTJ. Single-tube balanced heminested PCR for detecting Mycobacterium tuberculosis in smear-negative samples. J Clin Microbiol 2000; 38:1166-9.
2. Montenegro SH, Gilman RH, Sheen P, Cama R, Caviedes L, Hopper T, et al. Improved detection of Mycobacterium tuberculosis in Peruvian children by use of a heminested IS6110 polymerase chain reaction assay. Clin Infect Dis 2003; 36:16-23.
3. Shingadia D, Novelli V. Diagnosis and treatment of tuberculosis in children. Lancet Infect Dis 2003; 3:624-32.
4. Eisenach KD, Cave MD, Bates JH, Crawford JT. Polymerase chain reaction amplification of a repetitive DNA sequence specific for Mycobacterium tuberculosis. J Infect Dis 1990; 161:977-81.
5. van Soolingen D, de Haas PW, Hermans PW, Groenen PMA, van Embden JDA. Comparison of various repetitive DNA elements as genetic markers for strain differentiation and epidemiology of Mycobacterium tuberculosis. J Clin Microbiol 1993; 31:1987-95.
6. Kent L, McHugh TD, Billington O, Dale JW, Gillespie SH. Demonstration of homology between IS6110 of Mycobacterium tuberculosis and DNAs of other Mycobacterium spp. J Clin Microbiol 1995; 33:2290-3.
7. McHugh TD, Newport LE, Gillespie SH. IS6110 homologs are present in multiple copies in Mycobacteria Other Than Tuberculosis-causing mycobacteria. J Clin Microbiol 1997; 35:1769-71.
8. Salem JI, Marója MF, Carvalho FF, Lima MO, Feuillet A. Mycobacteria other than tubercle bacilli in sputum specimens from patients in Manaus Amazonas, Brasil. Acta Amazonica 1989; 19:349-54.
9. Salem JI, Gontijo Filho PP, Frebault VL, David HL. Isolation and characterization of mycobacteria colonizing the healthy skin. Acta Leprologia 1989; 7:18-20.
10. Fandinho FCO, Salem JI, Gontijo Filho PP, Marója MF, David HL. Mycobacterial flora of the skin in leprosy. Int J Leprosy 1991; 59:569-5.
11. Campelo CL, Vieira FD, Salem JI, Telles MAS, Maia R, Jardim SBV, et al. Tuberculose - diagnóstico laboratorial - Baciloscopia. Série Telelab. Brasília: Ministério da Saúde; 2001.
12. David LH, Lévi-frébault V, Thorel MF. Mèthode de Laboratoire pour Mycobacteriologie Clinique. Paris: Institut Pasteur, 1989; 39-58.
13. Abe C, Hirano K, Wada M, Kazumi Y, Takahashi M, Fukasawa Y, et al. Detection of Mycobacterium tuberculosis in clinical specimens by polymerase chain reaction and Gen-Probe amplified Mycobacterium tuberculosis direct test. J Clin Microbiol 1993; 31:3270-4.
14. Pierre C, Lecossier D, Boussougant Y, Bocart D, Joly V, Yeni P, et al. Use of a reamplification protocol improves sensitivity of detection of Mycobacterium tuberculosis in clinical simples by amplification of DNA. J Clin Microbiol 1991; 29:712-7.
15. Miyazaki Y, Koga H, Kohno S, Kaku M. Nested polymerase chain reaction for detection of Mycobacterium tuberculosis in clinical samples. J Clin Microbiol 1993; 31:2228-32.
16. Shankar P, Manjunath N, Mohan KK, Prasad K, Behari M, Ahuja, GK. Rapid diagnosis of tuberculous meningitis by polymerase chain reaction. The Lancet 1991; 337:5-7.
17. Ogusku MM, Sadahiro A, Hirata MH, Hirata RDC, Zaitz C, Salem JI. PCR in the diagnosis of cutaneous tuberculosis. Braz J Microbiol 2003; 34:165-70.
18. Davis LG, Kuehl WM, Battey JF. Basic methods in biology molecular. Connecticut: Appleton & Lange, 1994; 16-21.
19. Perosio PM, Frank TS. Detection and species identification of mycobacteria in paraffin sections of lung biopsy specimens by the polymerase chain reaction. Am J Clin Pathol 1993; 100:643-7.
20. Bollela VR, Sato DN, Fonseca BAL. McFarland nephelometer as a simple method to estimate the sensitivity of the polymerase chain reaction using Mycobacterium tuberculosis as a research tool. Braz J Med Biol Res 1999; 32:1073-6.
21. Nolte FS, Metchok B, McGowen JE, Jr. Edwards A, Okwumabna O, Thurmond C, et al. Direct detection of Mycobacterium tuberculosis in sputum by polymerase chain reaction and DNA hybridization. J Clin Microbiol 1993; 31:1777-82.
22. Shawar RM, El-Zaatari FAK, Nataraj A, Clarridge JE. Detection of Mycobacterium tuberculosis in clinical samples by two step polymerase chain reaction and nonisotopic hybridization methods. J Clin Microbiol 1993; 31:61-5.
23. Martins LC, Paschoal IA, Von Nowakonski A, Silva SAB, Costa FF, Ward LS. Nested-PCR using MPB64 fragment improves the diagnosis of pleural and meningeal tuberculosis. Rev Soc Bras Med Trop 2000; 33:253-7.
24. Butcher PD, Hutchinson NA, Doran TJ, Dale JW. The application of molecular techniques to the diagnosis and epidemiology of mycobacterial diseases. in: Mycobacterial disease - old problems, new solutions. (Suplement). J Appl Bacteriol 1996; 81:53S-71S.
25. Kontos F, Petinaki E, Nicolaou S, Gitti Z, Anagnostou S, Maniati M, et al. Multicenter evaluation of the fully automated Bactec MGIT 960 system and three molecular methods for the isolation and the identificiation of mycobacteria from clinical specimens. Diag Microbiol Infect Dis 2003; 46:299-301.

Abbreviations used in this paper:
AFB - Acid-fast bacilli
bp - Base pairs
IS6110 - Insertion sequence 6110
Mtb - Mycobacteriumtuberculosis
PCR - Polymerase chain reaction
TB - Tuberculosis

Correspondence
Mauricio Morishi Ogusku
Instituto Nacional de Pesquisas da Amazônia - INPA
Av. André Araújo, 2936, Bairro do Aleixo
CEP 69060-001 - Manaus, AM, Brazil
Phone: 55 92 643 3058
E-mail: mmogusku@inpa.gov.br

Submitted: 5 March 2004.
Accepted, after review: 3 June 2004.
Financial support provided by: the Ministry of Health, CAPES/RENOR. Rede Brasileira de Pesquisa em TB (Rede-TB, Brazilian Tuberculosis Research Network)/Grant no. 62.0055/01-4-PACDT-Milenio.

* Study carried out in the Laboratory of Mycobacteriology at the Coordenação de Pesquisas em Ciências da Saúde (Science and Health Research Coordination Center) of the Instituto Nacional de Pesquisas da Amazônia (INPA, National Research Institute of Amazônia), Manaus, Amazonas.