Identification of Mycobacterium tuberculosis in the cerebrospinal fluid of patients with meningitis using nested PCR

Introduction

Tuberculosis (TB) remains among the leading infectious causes of death in developing nations, with Mycobacterium tuberculosis (M. tuberculosis; MT) as the species responsible for most of these deaths (1,2). According to the World Health Organization (WHO), in 2014, the estimated number of tuberculosis cases was 9 million, including 3.5 million new cases and 1.1 million human immunodeficiency virus (HIV)-positive cases. WHO reported a prevalence of 370 tuberculosis cases per 100,000 individuals in the United States (3). In Mexico and other developing nations, tuberculosis is ranked 17th among the leading causes of all deaths in the working-aged general population (4).

The two main presentations of tuberculosis are pulmonary and extra pulmonary (5). Extra-pulmonary tuberculosis results from bacterial spread to other sites in the body, and occurs in approximately 10% of patients hospitalized with pulmonary tuberculosis (6). Tuberculous meningitis (TBM) is the most severe form of the infection, which is usually caused by M. tuberculosis and poses a serious threat to human health worldwide (7). If left untreated, the mortality associated with TBM is almost 100% and delayed treatment may lead to permanent neurological damage (3,8). The conventional 'gold standard' bacteriological methods, namely direct smear and culture isolation, are hardly able to detect M. tuberculosis in the cerebrospinal fluid (CSF) of patients with TBM. The clinical severity of TBM demands a rapid diagnosis and appropriate treatment in order to improve the clinical outcome. In infections of the central nervous system, such as TBM, the bacterial load is extremely low, thus making diagnoses based on the culture or staining of CSF difficult, as the results are usually negative. Furthermore, the clinical manifestations of TBM are often non-specific, similar to those of other forms of meningitis, such as neurocysticercosis, neuroborreliosis, or viral infections; thus, its diagnosis is currently based on clinical characteristics, radiological tests, responses to treatment and in the characteristics of CFS (9–14).

Molecular tests, in particular those based on the PCR for amplification of genes specific for the infecting organism may identify a wide variety of microorganisms, including mycobacteria (15–18). The analysis of the genomic sequence of M. tuberculosis (19) and studies on comparative genomics (20,21) have identified major deletions that specifically characterize different species of the M. tuberculosis complex, and are useful in differentiating these from non-tuberculous mycobacteria (22–25). Some studies have reported the use of PCR and quantitative PCR to identify M. tuberculosis in the CSF of patients with TBM, with varying sensitivity and specificity (25,26). However, the reported methods are based primarily on primers that bind IS6110, which allow the identification of the genus of the tuberculosis complex, but does not discern between species. Thus, the aim of this study was to identify TB in cases of meningitis with clinical and laboratory evidence suggestive of TBM, and to confirm our findings with molecular tests for TB infection.

Patients and methods

Bacterial strains and DNA isolation

The mycobacterial strains used as controls were M. tuberculosis H37Rv (donated by The Pasteur Institute, Paris, France), M. bovis AN5 (35734; ATCC, Rockville MD, USA), M. avium, M. intracellulare and M. habana (clinical isolates previously identified by rRNA 16S sequencing), M. celatum (51130; ATCC) and M. marinum (clinical isolate given by Dr Ruth Parra of the National Medical Center, Mexico City, Mexico). The strains were grown in Middlebrook 7H9 broth (Difco, Franklin Lakes, NJ, USA) and DNA was isolated using the phenol-chloroform method, as previously described (27,28). DNA was precipitated with isopropanol and resuspended in 50 µl of distilled water, and a 10 µl aliquot was used for PCR amplification (100 ng of DNA for the test), as previously described (15).

Ethics statement

The study protocol was approved by the Ethics Committee of the Mexican Social Security Institute (IMSS).

Patient selection

We recruited 144 consecutive patients with suspected meningitis and neurological symptoms who were examined at the neurology services of 4 Hospitals of IMSS in Mexico City between June 2008 and February 2013. The patients were both male and female, and included children and adults. Written informed consent was obtained from all the patients or the parents/legal guardians prior to recruitment. A total of 94 cases of meningitis with clinical and laboratory evidence suggestive of TBM were included, but only 50 of these cases fulfilled the criteria for probable TBM, the other 44 cases were eliminated. As the controls, we included 50 cases with neurological disease other than TBM, such as neuropathy, neurocysticercosis, neuroborreliosis, or viral infections. CSF was extracted from these patients as part of the protocol for diagnosis; a fraction of the sample was processed immediately for DNA isolation and was stored at −70°C until analysis.

Definition of case and control

As the bacterial load in the CSF is extremely low and culture is difficult using these samples, we diagnosed the cases of TBM based on clinical findings, CSF criteria or both. The clinical criteria of meningitis included headache, fever and neck stiffness, with or without an altered consciousness. The CSF criteria were a cell count >10 cells/mm3, a protein concentration >45 mg/dl and a glucose concentration <40 mg/dl (9), either alone or in combination. TBM was defined as meningitis with typical CSF findings in conjunction with either suggestive chest X-ray abnormalities or suggestive TB lymphadenitis. We chose to use clinical and laboratory data to define the cases and the controls, as previously described (9,13,14). A TBM case was defined as a patient suffering from fever, headaches and meningismus (stiff neck), along with focal neurological deficits, behavioural changes and alterations in consciousness, with CSF characteristics of moderate lymphocytic pleocytosis, moderately elevated protein levels and hypoglycorrachia (Tables I and II) (12). In addition, all the TBM cases should respond clinically and radiologically to specific anti-tuberculosis treatment. A history of a positive tuberculin skin test, or exposure to tuberculosis, was also considered as risk for TBM. A non-TBM case was defined as a patient with a confirmed diagnosis of neuropathy other than TBM. From our examination, there were 50 patients who fullfiled the criteria for TBM (the cases) and 50 patients who fulfilled the criteria for the controls (non-TBM).

Table I Baseline characteristics of tuberculous meningitis (TBM) cases.

Table I

Baseline characteristics of tuberculous meningitis (TBM) cases.

Characteristics	TBM cases, n (%)
Age (years)
<18	21 (42)
>18	29 (58)
Gender
Male	27 (54)
Female	23 (46)
Clinical manifestation
Headache	32 (64)
Fever	19 (32)
Neck stittness	18 (36)
Vomitting	18 (36)
Abnormal behaivior	26 (52)
Unconsciousness	3 (6)
Drowsiness	11 (22)
Seizures	24 (48)
Nausea	10 (20)
Blurred vision	10 (20)

[i] TBM (cases), n=50; average age, <18 [10.8 (0.5–18)] and >18 [42.3 (19–63) years].

Table II The cerebrospinal fluid (CSF) characteristics of the tuberculous meningitis (TBM) cases and non-TMB controls.

Table II

The cerebrospinal fluid (CSF) characteristics of the tuberculous meningitis (TBM) cases and non-TMB controls.

CSF characteristics	TBM cases	Non-TBM controls
Total cell count (cell/mm3)	96.5 (17–176)	59 (12–106)
Sugar (mg/dl)	49 (15–83)	65 (9–121)
Protein (mg/dl)	151 (98–204)	74 (53–95)

Sample size calculation

The sample size was estimated with the EPI info 3.5.1 (2008) programme, aiming to achieve 90% sensitivity for our multiplex-nested PCR analysis of the CSF of patients with neurological manifestations suggestive of TBM (26,29). The sample size calculation was 20 cases and 20 controls. We recruited 50 cases and 50 controls for this study.

Multiplex and nested PCR

We used 100 ng of DNA from the M. tuberculosis strain H37Rv to standardize the test. The sequences of the primers are shown in Table III. In the initial amplification, primers MT1 and MT2 amplified the gene encoding the 32-kDa α antigen present in all described mycobacteria, whereas primers IS5 and IS6 amplified the IS6110 insertion element (30,31), and PT1 and PT2 were used to amplify the species-specific gene mtp40. Nested PCR further amplified an internal region of the mtp40 gene of M. tuberculosis (32). All reactions were performed in a final volume of 25 µl, with 10X reaction buffer, 1.25 U of Taq DNA polymerase, 0.2 mM of each deoxynucleotide, 2.5 mM MgCl2, 10 pmol of MT1 and MT2, 15 pmol of IS5 and IS6, and 20 pmol of PTI and PT2. The cycling parameters for the initial PCR were 94°C for 5 min, followed by 30 cycles of denaturation at 94°C for 1 min, annealing at 71°C for 2 min and extension at 72°C for 3 min, followed by a final extension at 72°C for 10 min. Following amplification, the PCR products were analysed by horizontal electrophoresis on 2.0% agarose gels (33) using DNA molecular marker 1 kb (Invitrogen Life Technologies, Carlsbad, CA, USA). The multiplex PCR products were subjected to nested PCR under the following conditions: 94°C for 5 min, followed by 30 cycles of 94°C for 1 min, 74°C for 2 min, and 72°C for 2 min, followed by a final extension at 72°C for 7 min.

Table III Primers used in the study.

Table III

Primers used in the study.

Gene fragment	Primer target (description)	Primer sequences 5′→3′ (foward/reverse)	Amplicon size (bp)	Authors/year/(Refs.)
p32	MT1 MT2	TTC CTG ACC AGC GAG CTG CCG CCC CAG TAC TCC CAG CTG TGC	506	Del Portillo et al, 1991, 1996 (30,31) Nava et al, 2005 (33)
IS6110	IS 5 IS 6	CGG AGA CGG TGC GTA AGT GG GAT GGA CCG CCA GGG CTT GC	984	Del Portillo et al, 1991, 1996 (30,31) Nava et al, 2005 (33)
mtp40	PT1 PT2	CGG CAA CGC GCC GTC GGT GG CCC CCC ACG GCA CCG CCG GG	396	Del Portillo et al, 1991 (30,31) Herrera and Segovia, 1996 (32)
Internal fragment of mtp40	PT3 PT4	CAC CAC GTT AGG GAT GCA CTG C CTG ATG GTC TCC GAC ACG TTC G	223	Gori et al, 1996 (35)

Laboratory sensitivity and specificity

Before testing the clinical samples, the specificity of the assay was tested using DNA from the following Mycobacterium species: M. tuberculosis H37Rv, M. bovis AN5, M. intracellulare, M. avium, M. celatum, M. habana and M. marinum, and DNA from Stap hylococcus aureus (S. aureus) and Escherichia coli (E. coli) (clinical isolate from the clinical laboratory of Pediatric Hospital of the National Medical Center, Mexico City, Mexico). The sensitivity of the assay was estimated using 10-fold dilutions of M. tuberculosis DNA, ranging from 100 ng to 0.01 fg.

Processing of the clinical samples

All CSF samples from the patients with meningeal infection were analyzed by multiplex and nested PCR, and the study was blinded. An aliquot of 500 µl of each CSF was subjected to thermal shock, and centrifuged at 17,000 × g for 15 min. The DNA was isolated from the pellet with guanidine isothiocyanate according to the procedure previously described by Chomczynski and Sacchi (27) and Chomczynski (28), and stored at −70°C until use. A 100 ng sample of DNA from each CSF sample was used for the multiplex and nested PCR. The results of PCR for the cases and controls were used to estimate the clinical specificity and sensitivity of the test.

DNA sequences and analysis

Four of the amplified nested PCR products from the clinical samples were purified using the QIAEX II Gel Extraction kit (Qiagen, Hilden, Germany), and sequenced with the CEQ 8800 Genetic Analysis system (Beckman Coulter, Inc., Brea, CA, USA), according to the manufacturer's instructions. The sequences generated were analyzed with the BLAST tool (available online at www.ncbi.nih.gov/BLAST) and aligned with the available M. tuberculosis H37Rv genome sequence database, accessible at http://genolist.pasteur.fr/Tuberculist, to estimate the degree of homology.

Statistical analysis

Measures of central tendency were the descriptive statistics used to analyze the quantitative variables. A 2×2 contingency table was used to determine the sensitivity (S), specificity (E), positive predictive value (PPV) and negative predictive value (NPV) of the diagnostic test using clinical samples (27).

Results

Multiplex and nested PCR

The sizes of the amplified fragments were sufficiently different to be distinguishable on agarose gels; the primers had no potential matches with sequences at non-specific target sites, and the optimal DNA-primer annealing temperatures were almost the same for all template-primer combinations (15). Fig. 1A shows the fragments amplified when each individual primer set was used separately. Fig. 1B, lane 3, shows the simultaneous amplification of the 3 fragments. A nested PCR was used to increase the sensitivity of the system, with a pair of primers that amplifies a fragment of 223 bp, corresponding to the internal region of the species-specific sequence of the mtp40 gene (Fig. 1C).

Figure 1 Electrophoresis of PCR and nested PCR products of Mycobacterium tuberculosis (M. tuberculosis) H37Rv DNA. (A) Amplification of p32, IS6110 and mtp40 fragments by conventional PCR. Lane 1, DNA molecular marker 1 kb; lane 2, negative control of MT1 and MT2 primers; lane 3, p32 amplification fragment; lane 4, negative control of IS5 and IS6 primers; lane 5, IS6110 amplification fragment; lane 6, negative control of PT1 and PT2 primers; and lane 7, mtp40 amplification fragment. (B) Amplification of p32, IS6110 and mtp40 fragments by multiplex PCR lane 1, DNA molecular marker 1 kb; lane 2, negative control; and lane 3, amplification products of multiplex PCR of M. tuberculosis H37Rv DNA. (C) Amplification of internal fragment of mtp40 by nested PCR. Lane 1, DNA molecular marker 1 kb; lane 2, negative control; lanes 3 and 4, M. tuberculosis H37Rv DNA.

Laboratory specificity and sensitivity of the multiplex and nested PCR system

The specificity of the assay was evaluated with DNA from different Mycobacterium strains (M. tuberculosis, M. bovis AN5, M. intracellulare, M. avium, M. celatum, M. habana and M. marinum) and DNA from S. aureus and E. coli. Following multiplex PCR amplification, M. tuberculosis DNA amplify the 3 expected bands of genus (506 bp), complex (984 bp) and species (396 bp), and M. bovis DNA amplify the 2 expected bands of 506 and 984 bp. We also tested DNA from other Mycobacterium species (M. intracellulare, M. avium, M. celatum, M. habana and M. marinum), and in all these strains, only the fragment corresponding to the p32 gene of the Mycobacterium genus was amplified. Finally, no amplification was observed with any of the primers when DNA form S. aureus and E. coli was tested (Fig. 2A).

Figure 2 Amplification of DNA from other bacteria. (A) Multiplex PCR products. (B) Nested PCR products. Lane 1, DNA molecular marker 1 kb; lane 2, negative control; lane 3, Escherichia coli (E. coli) DNA; lane 4, Staphylococcus aureus (S. aureus) DNA; lane 5, M. habana DNA; lane 6, M. celatum DNA, lane 7, M. marinum DNA; lane 8, M. avium DNA; lane 9, M. intracellulare DNA; lane 10, M. bovis AN5 DNA; and lane 11, Mycobacterium tuberculosis (M. tuberculosis) H37Rv DNA.

These results confirm that multiplex PCR is specific for M. tuberculosis. Moreover, the products obtained from the multiplex PCR were reamplified in the nested PCR, and the internal fragment of the mtp40 gene only amplified with the M. tuberculosis DNA, confirming that this system specifically identifies M. tuberculosis (Fig. 2B).

The sensitivity of the multiplex PCR assay was 10 ng of M. tuberculosis H37Rv DNA where it amplified the 3 expected products (Fig. 3A). However, the sensitivity of the nested PCR was 0.1 fg of DNA for the fragment corresponding to an internal region of the mtp40 gene (Fig. 3B).

Figure 3 Sensitivity test of the multiplex and nested PCR for detect Mycobacterium tuberculosis (M. tuberculosis) DNA. (A) Multiplex PCR products of different concentrations of M. tuberculosis H37Rv DNA. (B) Nested PCR products from multiplex PCR of different concentrations of M. tuberculosis H37Rv DNA. Lane 1, DNA molecular marker 1 kb; lane 2, negative control; lane 3, 100 ng; lane 4, 10 ng; lane 5, 1 ng; lane 6, 100 pg; lane 7, 10 pg; lane 8, 1 pg; lane 9, 100 fg; lane 10, 10 fg; and lane 11, 1 fg.

CSF sample analysis

Following the standardization of the multiplex and nested PCR system with the Mycobacterium strain, we assessed its usefulness in the CFS samples obtained from clinical cases suggestive of TBM. We analyzed 100 CSF samples from patients with diverse neurological symptoms, from 50 patients with a clinical diagnosis of TBM (cases) and 50 patients with a confirmed neurological aetiology other than TBM (controls). In the initial multiplex PCR, only one (patient 3) of the 100 CSF samples amplified a fragment of 396 bp corresponding to the mtp40 gene (Fig. 4A). In the nested PCR, the 223-bp amplification product of the mtp40 gene specific for M. tuberculosis was amplified from 53 samples (Fig. 4B). Forty-nine of these samples corresponded to the group of cases and responded to specific treatment against M. tuberculosis, thus confirming the presence of M. tuberculosis DNA in the CSF. Moreover, we had one false-negative sample in the nested PCR. The negative control run with each set of test samples produced no visible product on ethidium bromide-stained agarose gel.

Figure 4 Multiplex and nested PCR for detect Mycobacterium tuberculosis (M. tuberculosis) in cerebrospinal fluid (CSF). (A) Multiple-PCR test applied to CSF samples from patients with neurologic diseases. (B) Nested PCR products from multiple-PCR products. Lane 1, patient 16 (true negative); lane 2, patient 11; lane 3, patient 8; lane 4, patient 3; lane 5, patient 12; and lane 6, patient 17 (patient 11, 8, 3, 12 and 17 were true positive). MM, molecular marker 1 kb; C-, negative control; C+, positive control DNA from M. tuberculosis H37Rv strain. (C) Alignment of the mtp40 internal fragment with sequence of nested PCR product using PT3 and PT4 primers. The highlighted positions indicate the sequence of the mtp40 internal fragment (Query) from 2630676 to 2630827 positions, and sequence of nested PCR product (Sbjct).

DNA sequence of PCR products

We determined the DNA sequence of the amplified 223-bp fragment corresponding to the mtp40 gene from 4 positive samples to confirm that the product was specific for M. tuberculosis. The sequence alignment showed 99% homology between the nested PCR products and the corresponding region of the mtp40 gene (Fig. 4C).

Sensitivity and specificity of the test

The clinical sensitivity of this method was 98.0% and its specificity was 88.0%; the positive predictive value (PPV) was 91.0% and the negative predictive value (NPV) was 98.0% (Table IV).

Table IV Sensitivity and specificity as a diagnostic test for tuberculous meningitis (TBM).

Table IV

Sensitivity and specificity as a diagnostic test for tuberculous meningitis (TBM).

Nested PCR	%
Sensitivity	98
Specificity	92
PPV	88
NPV	98

[i] PPV, positive predictive value; NPV, negative predictive value.

Discussion

The suboptimal and often delayed results of traditional microbiological techniques used in the diagnosis of TBM underscore the need for a more rapid and more accurate diagnostic method to facilitate early treatment. Several molecular-based methods used for the diagnosis of tuberculosis in respiratory specimens have been evaluated for their applicability to the diagnosis of TBM (23). Many of these methods have failed with CSF samples, mainly due to the fact that the number of bacilli typically present in TBM is low and that amplification inhibitors are present in the CSF. Commercially available methods of nucleic acid amplification (NAA) for the direct detection of the M. tuberculosis complex have been approved in the United States for testing respiratory specimens; Ling et al, performed an extensive literature search and identified a total of 125 separate studies from 105 articles that reported NAAT results from respiratory specimens. The results showed that sensitivity and specificity estimates for commercial NAATs in respiratory specimens were highly variable, with sensitivity lower and more inconsistent than specificity. Thus, summary measures of diagnostic accuracy are not clinically meaningful. The use of different cut-off values and the use of specimens other than sputum could explain some of the observed heterogeneity (34). However, there are few NAA methods that have been approved for testing the CSF and several studies have evaluated their performance in patients with TBM (31,32).

The PCR system also has been used as an epidemiological tool in strain classification (23,24). The most widely used target sequence for the diagnosis of tuberculosis has been the IS6110 insertion element, present in a different number of copies in the genome of species of the M. tuberculosis complex (12,35) and PCR techniques based on this sequence have shown to be useful for diagnosis. However, studies have demonstrated that some M. tuberculosis strains do not carry the IS6110 sequence (36,37). Therefore, the use of a PCR method based on the detection of IS6110 for the diagnosis of M. tuberculosis may in some cases lead to false-negative results. Additionally, PCR-based diagnoses based exclusively on the IS6110 sequence would also fail to distinguish M. tuberculosis from other mycobacteria of the M. tuberculosis complex.

In the present study, we used the multiplex PCR method described by Del Portillo et al (30,31) to detect M. tuberculosis in pulmonary-type tuberculosis, with some modifications to identify M. tuberculosis in CSF samples. Additionally, we used the nested PCR described by Herrera and Segovia (32). In this study, we demonstrated that using a multiplex and nested PCR system, it is possible to simultaneously amplify the α antigen gene (p32), which identifies the Mycobacterium genus, the IS6110 insertion element and the species-specific mtp40 gene. The nested amplification of a 396-bp fragment corresponding to the mtp40 gene ensures that the results are specific for M. tuberculosis. In addition, the nested PCR allowed us to reach a limit of detection to 0.1 fg of mycobacterial DNA, which theoretically would allow the detection of one bacterium in the sample (16). The PCR amplification of the mtp40 sequence provides a sensitivity and specific method for the diagnosis of tuberculosis (35,38) and it may be useful for the identification of M. tuberculosis, as 98.5% of strains possess this gene (39). We used a multiplex-PCR method followed by a nested PCR, reaching high specificity and sensitivity.

With the multiplex PCR we were able to detect only the mtp40 gene in a single CSF sample; this may be due to the extremely low number of bacteria in the CSF. However, with the nested PCR, we detected the presence of M. tuberculosis in 53 patients with the amplification of the species-specific mtp40 gene. The products of the nested PCR were sequenced to further confirm that the amplified fragment corresponded to the mtp40 M. tuberculosis gene, and demonstrated that the sequence had a 99% homology with the mtp40 gene. Thus, we were able to detect the presence of M. tuberculosis in the CSF samples of patients with suggested TBM using a species-specific multiplex and nested PCR test.

Our test showed a sensitivity of 98.0% and a specificity of 88.0% for the diagnosis of TBM in CSF samples, which are better than those reported previously (40,41). In relation to the false-negative result, it is possible that a fraction of mtp40 gene was excised by IS6110 recombination as previously described Vera-Cabrera et al (42), which could explain the lack of genes in some M. tuberculosis strains described by Weil et al (43). Concerning the false-positive results, we suggest that they may represent cases of co-infection, where clinical diagnosis could have masked the M. tuberculosis symptoms, focusing on the diagnosis of the other disease. The test presented a reasonably good PPV (91.0%) and NPV (98.0%), which contrast with the findings reported by Lima et al (37), who applied a nested PCR to the peripheral blood samples from patients with extra pulmonary tuberculosis, with a rather low PPV (55.6%) and an acceptable NPV (92.7%).

In conclusion, the present study demonstrates that the identification of M. tuberculosis in the CSF of patients with meningitis is possible using a developed system of multiplex and nested PCR. This study supports the implementation of two molecular tests as a sensitive and specific diagnostic tool, which offers an improved alternative for the diagnosis of TBM. The multiplex-nested PCR test is a rapid, sensitive and specific tool which may have a beneficial impact on the management of patients with suspected TBM, allowing a more timely anti-tuberculosis treatment to reduce sequelae and mortality.

Acknowledgments

The present study was supported by the Instituto de Ciencia y Tecnología del Distrito Federal (ICyTDF), grant no. ICyTDF/DSBMA/350/2009 and the Instituto Mexicano del Seguro Social (IMSS), grant no. FIS/IMSS/PROT/G09/757.

References