Tuberculosis associated with Mycobacterium tuberculosisBeijing and non-Beijing genotypes: a clinical and immunological comparison

Subjects and setting

Inpatients with pulmonary tuberculosis were prospectively and consecutively enrolled from Tan Tock Seng Hospital (TTSH), and National University Hospital (NUH) from September 2004 to May 2005. Subject inclusion criteria were: (1) clinical presentation and radiographic findings considered to be strongly suggestive of pulmonary tuberculosis by the primary physicians, (2) ≥ 18 years old, (3) antituberculous treatment naïve or not longer than 48 hours, (4) able and willing to give informed consent. Patients who had a past history of tuberculosis, who were known to have human immunodeficiency virus (HIV) infection or other coexisting systemic diseases, such as chronic renal failure, diabetes mellitus, and neoplastic diseases, and pregnant or breastfeeding women were excluded. Approval for this study was granted by the Ethics Committee of the National Healthcare Group (NHG), Singapore. Written informed consent was obtained from all study subjects.

Demographic and clinical data collection

Demographic (date of birth, place of birth, sex, and race), clinical, and routine laboratory testing data were collected by interview of patients and/or from medical records. The data of BCG scar, BCG vaccination history, TB contact history, and risk factors for HIV infection were collected by interview of the subjects. The information on clinical symptoms and their durations, including cough, fever, night sweats, haemoptysis, weight loss, appetite loss, and chest pain, were first collected from medical records and confirmed through patient interviews. Routine temperature charts were examined covering the time of admission until the time of enrolment. Axillary temperature was recorded every 6 hours using a digital thermometer in both hospitals. The patient was defined having fever if there was a recorded temperature of > 37.5°C. Routine laboratory test data including total and differential white blood cell counts, hemoglobin, sputum smear, culture and drug-susceptibility test results were collected. Anemia was defined when hemoglobin < 13 g/dL in males and < 11.8 g/dL in females. All patients were screened for risk factors of HIV infection and offered HIV serology testing.

Chest X-ray (CXR) films were reviewed at each institution by a single clinician (TKL and AKYO) experienced in the diagnosis and management of tuberculosis who were blinded to M. tuberculosis genotyping results. The disease extent on CXR was classified as: (1) unilateral disease, (2) bilateral disease, (3) cavitary disease, (4) pleural effusion, and (5) miliary disease.

DNA extraction and genotyping of M. tuberculosis

Sputum specimens were collected from each subject at enrolment. Sputum samples were first decontaminated and liquefied with 1% N-Acetyl-L-Cysteine (NALC)-2% NaOH. DNA was extracted using phenol-chloroform/isoamyl alcohol, precipitated with absolute ethanol and washed with 70% ethanol according to standard protocol. DNA was finally dissolved in 50 μl distilled water.

Spoligotyping was performed using a commercial kit (Isogen Bioscience BV, Maarssen, the Netherlands) according to the manufacturer's protocol. Spoligotyping patterns were translated to 15 octal codes as suggested before [15]. Mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) typing was performed as previously described [16, 17]. Two μl of sputum DNA sample was used in each PCR reaction for these genotyping analyses. MIRU-VNTR allele number was manually converted according to amplicon size.

Isolation of plasma and PBMC

Blood samples collected in heparinized tubes were centrifuged at 300 × g for 10 minutes to separate plasma and blood cells. Plasma was recovered and further centrifuged at 14,000 rpm for 15 minutes at 4°C, and then recovered in 0.5 ml aliquots and stored in -80°C freezer. The precipitated blood cells were resuspended in appropriate volume of RPMI 1640 Media and subjected to PBMC isolation by Ficoll-Paque gradient centrifugation.

Cytokine ELISA

Blood plasma levels of IFN-γ, IL-2, IL-4, IL-13, and IL-18 were measured using quantitative ELISA kits (Bender MedSystems, Austria) according to the manufacturer's protocols and each sample was assayed in duplicate. The sensitivity of these kits were 0.66 pg/ml, 2.3 pg/ml, 0.66 pg/ml, 0.99 pg/ml, 9.2 pg/ml, respectively. A high sensitivity (0.06 pg/ml) IFN-γ kit was used to measure some samples with very low level of IFN-γ.

Quantification of cDNA by Real-Time PCR

Total RNA was extracted from 2 × 10⁶ peripheral blood mononuclear cells (PBMCs) using TRIZOL^® reagent (Invitrogen, California, USA) according to the manufacturer's protocol. cDNA was reverse transcribed from total RNA using the 1st Strand cDNA Synthesis Kit for RT-PCR (AMV) (Roche Applied Science). The cDNA of reverse transcribed cytokine genes IFN-γ, IL-2, IL-4, IL-13, and IL-18, and housekeeping gene β-actin were quantified by quantitative real-time PCR using the LightCycler™ FastStart Master Sybr^®Green I kit on the Roche LightCycler apparatus (Roche Applied Science). The quantitative real-time PCR primers and standards for these genes were obtained from Search-LC GmbH (Heidelberg, Germany) and the manufacturer's protocols were followed. Melting curve analysis was performed to confirm the identity and specificity of PCR products. All cytokine cDNA copies were normalized to the housekeeping gene β-actin.

Statistical analysis

χ² test or Fisher's exact test was used to analyze categorical variables. Student's t test was used to compare normally distributed continuous variables and those variables which were normally distributed after log-transformation. Multivariate logistic regression was used to adjust for potential confounding factors. A p value of <0.05 was considered statistically significant. Odds ratios (OR) and 95% confidence intervals (CI) were calculated.

Patient enrolment and determination of M. tuberculosisgenotypes

A total of 54 patients were enrolled in this study (35 from TTSH and 19 from NUH). One of the 54 patients was found to be HIV-infected, thus was excluded from further analysis. Of the remaining patients, 16 tested negative for HIV antibodies. Those who were not tested did not have any apparent risk factors for HIV infection nor any evidence of immune suppression noted on physical examination. Two further patients were excluded because they were unable to provide a sputum specimen for genotyping. Nine patients were excluded because they were culture negative, sputum smear negative for acid-fast bacilli and did not have M. tuberculosis detected by either spoligotyping or MIRU-VNTR typing.

Forty-two patients had culture-proven, drug-susceptible (to rifampicin, streptomycin, isoniazid, and ethambutol) pulmonary tuberculosis. Of these 42, 38 (90.5%) were smear positive. M. tuberculosis DNA was detected in the sputum sample and successfully typed by MIRU-VNTR typing in 41 cases, and by spoligotyping in 39 cases. All the cases of failure of genotyping were also smear negative. The single case that could not be typed by either method was excluded from further analysis. Therefore, the study analysis concerns 41 culture-proven tuberculosis patients with genotyping data.

Demographic and epidemiological characteristics

Clinical and radiological features

There was no significant difference in the frequency of cough between Beijing and non-Beijing infected patients (p = 0.125). However, in the 18 patients who had cough and were infected with Beijing strains, the duration of cough before the diagnosis varied widely from 14 to 1440 days (4 years) with a median of 60 days. Only 39% (7/18) of such patients were identified and diagnosed within one month. For the patients infected with non-Beijing strains, the range of the duration was much narrower, from 2 to 180 days with a median of 30 days. Much higher proportion (65%, 13/20) of the patients was identified within one month. The individual cough duration is shown in Figure 1. The difference between the two groups is statistically significant (p = 0.048), indicating that patients infected with Beijing strains presented with longer duration of cough before the diagnosis.

Fever was recorded in 9 (42.9%) of the 21 patients infected with Beijing strains and 17 (85.0%) of the 20 patients infected with non-Beijing strains [OR (95% CI), 0.13 (0.03–0.57), p = 0.005]. In addition to fever, night sweats were less frequently reported in patients infected with Beijing strains [OR (95% CI), 0.20 (0.04–0.99), p = 0.049]. To further test these associations, we performed multivariate logistic regression analysis. Independent variables included sex, age, and ethnicity, though they did not differ significantly between the two groups, as well as fever and night sweats. The inverse association between febrile response and Beijing strains remained strong after multivariate adjustment [OR (95% CI), 0.12 (0.02–0.63), p = 0.008]. However, the inverse association between night sweats and Beijing strains was no longer statistically significant after the multivariate adjustment [OR (95% CI), 0.20 (0.02–1.75), p = 0.145].

The chest radiographic abnormalities associated with infections of Beijing and non-Beijing strains were compared (Table 3). There was significantly less cavitary disease in patients infected with Beijing strains [OR (95% CI), 0.20 (0.04–0.99), p = 0.049]. No difference was observed in the frequency of unilateral and bilateral non-cavitary disease, and pleural effusion. Miliary disease was not found in the subjects.

The frequency of anemia in the entire study group was 65.9% (27/41), and there was no difference between the two groups (Table 3). Table 3 also presents total and differential counts of white blood cells of the patients. The mean counts for total white blood cells, lymphocytes, eosinophils, and basophils were not significant different in each group. The mean monocyte count was significantly higher in the non-Beijing group (mean, 901/μl) compared to the Beijing group (mean, 656/μl) (p = 0.014).

Plasma cytokine and cytokine gene expression analysis

Blood samples were collected from 40 of the 41 tuberculosis patients. Plasma IL-2, IL-4, and IL-13 were not detectable by the ELISA kits, thus not shown. The plasma levels of IFN-γ and IL-18 were similar between patients infected with Beijing and non-Beijing strains (Figure 2)

Of the 40 blood samples collected and prepared for cytokine gene expression analysis, the first 15 samples had very low RNA yields, thus the cDNA transcribed from these samples was too low to be reliably quantified. The results presented below pertain to the 25 samples in which satisfactory RNA samples were obtained. The 25 samples were from 12 patients infected with Beijing strains (8 males, 66.7%; mean age, 52, SD, 14) and 13 patients infected with non-Beijing strains (5 males, 38.5%; mean age, 48, SD, 21). These characteristics were similar between the subgroup patients and the whole group of study subjects (refer to Table 2).

The expression of cytokine genes was standardized as per 10⁵ β-actin cDNA copies. The expression level of IL-13 was too low to be detected by the quantitative real-time PCR, thus not shown. As shown in Figure 3, the cDNA copies of IFN-γ, IL-2 and IL-18 did not differ significantly between the Beijing and non-Beijing groups. The median cDNA copies for IFN-γ, IL-2 and IL-18 were respectively 232 (range, 77–691), 7 (range, 0–78), and 314 (range, 120–715) in patients infected with Beijing strains, and 228 (range, 39–491), 10 (range, 0–30), 355 (range, 81–1067) in patients infected with non-Beijing strains. The cDNA copies of IL-4 were significantly higher in patients infected with non-Beijing strains (median = 26, range, 9–87) compared to those with Beijing strains (median = 16, range, 0–68) (p = 0.018).

We also compared the IFN-γ : IL-4 cDNA copy number ratio between the Beijing and non-Beijing groups. Two patients who had no detectable IL-4 cDNA copies were assumed having 1 cDNA copy in order to calculate the ratios. Patients infected with non-Beijing strains showed significantly lower IFN-γ : IL-4 cDNA copy number ratios than those infected with Beijing strains (p = 0.01) (Figure 4).