Ex vivo mRNA expression of toll-like receptors during latent tuberculosis infection

Using purposive sampling, a total of 64 apparently healthy children and adolescents, aged between 11 and 18 years, and who have been recruited from different schools in Addis Ababa, capital city of Ethiopia (between 2014 and 2016), a country that has been reported as one of the highest TB burden settings globally [1], were included in this study. Children below the age of 11 were excluded for the sake of ethical regulations in Ethiopia, restricting blood volumes needed to conduct the proposed research. Participants had no known illness and were asymptomatic upon screening during enrolment and were categorized as latently infected with TB (N = 32), after being tested positive by tuberculin skin testing (TST) or quantiferon (QFT) and non-infected (N = 32) after being tested negative by TST or QFT as previously described [13]. An indurations size of 10 mm and above was considered as a cut-off to define tuberculin reactivity [14]. Similarly, an IFN-γ level above 0.35 International Units (IU)/mL for (TB Antigen-Nil) was considered as positive, while testing using QFT, according to the manufacturer’s instruction. Participants had no known history of contact with a known TB patient or TB disease, with the majority having history of BCG vaccination and no concomitant infections (such as parasites) (Table 1). Diagnosis of stool helminthic parasites/ova was done using the Kato-Katz technique as described in our previous study [13]. Participants who were HIV positive or those with a history of allergy or chronic infection at the time of clinical screening were excluded from the study.

Approximately 2.5 ml of venous blood was collected into PAXgene Blood RNA Kit (PreAnalytiX, QIAGEN); reverse transcribed into complementary DNA (cDNA) using Omniscript Reverse Transcription Kit (QIAGEN, Germany) and oligo (dT) primers (Promega, USA) as described previously [13] and stored at -20 °C until assayed further by quantitative real-time PCR (qRT-PCR). To analyze the relative mRNA expression of selected TLRs in whole blood, qRT-PCR was performed using QuantiTect Reverse Transcription PCR Kit (QIAGNE, Germany). The amount and purity of all cDNA samples were measured using NanoDrop 2000 Spectrophotometer (Thermo Scientific, Amersham Biosciences, UK), adjusted and normalized to a final concentration of 0.2 μg/μl and a total of 1 μg of cDNA was used as a starting template in a total reaction volume of 12.5 μl. Specific primers and florescent-labelled probes (Eurofins MWG Operon) (Table 2) were used at a final concentration of 0.5 μM and 0.3 μM, respectively to span exon-intron junctions to prevent amplification of genomic DNA and amplicons of fewer than 150 base pairs to enhance the efficiency of PCR amplification. All the PCR mixtures were prepared using Corbett Robotics (Corbett Research, Australia) and run in duplicates including non-template controls using Rotor-Gene (RG-3000) thermo-cycler (Corbett Research, Australia). The qRT-PCR cycling was carried out under the following conditions: initial Taq enzyme heat activation step at 95 °C for 15 min, followed by 40 repeats of three step cycling, denaturation step 94 °C for 10 s, primer/probe annealing step ranging between 58 °C and 60 °C for 20 s and final extension/polymerization step at 72 °C for 30 s. In all experiments, data were acquired at extension phase (on FAM/Sybr channel) and analyzed using Rotor-Gene Real-Time Analysis Software 6.0 (Corbett Research, Australia). Human ribosomal protein (HuPO) was used as a housekeeping gene and internal control and a non-template tube as a negative control throughout the experiment; yielding a PCR efficiency of 97% (See Supplementary data). The threshold cycle (CT), the PCR cycle at which the fluorescent signal of the reporter dye crosses an arbitrarily point, was set at the exponential phase of amplification and used as the quantitative endpoint of qRT-PCR as described earlier [16]. Comparative CT method (also known as 2^(−ΔΔCT) was used to describe the fold change in the relative expression of TLR genes as previously described [16]. Briefly, the non-LTBI group was considered as calibrator and the LTBI as target group to estimate the change in the CT values (ΔCT); all CT values were initially normalized to the CT values of HuPO. The computed numbers indicate the fold change in the expression of a target gene in latently infected group relative to non-infected. A similar analysis was performed to compare the relative expression of the genes across age, where children under 15 years were considered as calibrator groups and adolescents (those 15 years and above) as target groups. Furthermore, to appreciate the relative expression of TLRs localized to the cell surface (TLR-1, TLR-2, TLR-4 and TLR-6) to enodsomal (TLR-9) regions, we compared the ratio of comparative CT values of TLRs in TST positive and negative groups. This work was conducted at the Armauer Hansen Research Institute (AHRI) laboratory.

Data from qRT-PCR (2^(−ΔΔCT)) were exported to GraphPad Prism Version 7.01 for Windows (GraphPad Software, La Jolla California USA, www.​graphpad.​com) and SPSS Version 20.0 for further statistical analyses. Categorical variables were presented as counts and percentages and compared using Chi square test. For two groups, analyses of the fold change in the relative expression of the TLRs were done using non-parametric ANOVA (Kruskal-Wallis test with Dunn’s multiple comparison tests). In addition, Spearman linear correlation (r) was done to compare the correlation between the surface and intracellular TLRs. In all instances, a P-value < 0.05 was considered statistically significant.

The relative mRNA expressions of TLRs (TLR-1, TLR-2, TLR-4, TLR-6, and TLR-9), normalized to HuPO, were compared between latently infected and non-infected individuals. As the numerical value of the CT inversely relates to the amount of amplicon in the reaction [16], threshold values above 30 were considered as weak fluorescent signals. Results are presented as fold change of target gene expression in a target sample relative to a reference sample, normalized to a reference gene. With the relative gene expression set to 1 for the reference sample, the ΔΔCT is equal to 0, resulting in 2⁰, i.e., 1 [17], and thus the mean fold change greater than one was considered as up-regulation or increased expression of a particular gene.

An increased mean fold change in the mRNA expression of TLRs was observed in latently infected individuals relative to those non-infected (Fig. 1). This increase was four-fold for TLR-2 gene (mean with SEM: 4.04, 0.91 and 95% CI (2.18, 5.91)) and three-fold for TLR-6 (mean with SEM: 3.21, 1.12 and 95% CI (0.93, 5.50)) and approaching a two-fold increase for TLR-4 and TLR-9 (mean with SEM: 1.71 ± 0.95 and 95% CI (0.2, 3.7) and 1.59 ± 0.8 and 95% CI (0.07, 3.3)), respectively (p = 0.0001). On the other hand, a slight decrease in the mean fold change in the mRNA expression of TLR-1 was observed in those latently infected compared to non-infected (mean with SEM: 0.80, 0.29, and 95% CI (0.20, 1.39)) (Fig. 1a). The difference looks more apparent when the normalized ΔΔCt values are compared between LTBI infected and those non-infected (Fig. 1b).

To understand whether adolescents have distinct mRNA expression patterns compared to younger children, the mean fold change in the expression of TLRs was also compared across age, where those aged between 11 and 14 years were considered as children and those above 15 years as adolescents and no significant difference was observed between the two groups (Data not shown). Similarly, no apparent difference was noted between male and female participants (P > 0.05) (Fig. 2).

To better visualize the gene expression patterns among the study population, we did a heatmap using the ΔΔCt values of the TLR genes in LTBI and those non-infected (Fig. 3). This helps to clearly visualize the level of expressions of the genes in the study groups.

Similarly, correlation in the expression of surface TLRs with endogenously expressed markers was analysed and overall a positive correlation was observed between surface TLRs and an endogenously expressed TLR, TLR-9 (Fig. 4). A moderate correlation was also noted between the different surface TLRs (r² = 0.6, P-value < 0.0001, data not shown). A similar comparison was done for those high and low expressed receptors to see possible correlations, especially those forming heterodimers; and the ratios (TLR-2/TLR-1 and TLR-2/TLR-6) were found to be higher (7.2 ± 1.12 (95% CI: 4.9, 9) and 3 ± 0.6 (95% CI: 1.6, 4.2)), among latently infected group when compared to non-infected groups, respectively (P = 0.004) (Fig. 5).