Presence of RD149 Deletions in M. tuberculosis Central Asian Strain1 Isolates Affect Growth and TNFα Induction in THP-1 Monocytes

Central Asian Strain 1 (CAS1) is the prevalent Mycobacterium tuberculosis genogroup in South Asia. CAS1 strains carry deletions in RD149 and RD152 regions. Significance of these deletions is as yet unknown. We compared CAS1 strains with RD149 and concurrent RD149-RD152 deletions with CAS1 strains without deletions and with the laboratory reference strain, M. tuberculosis H37Rv for growth and for induction of TNFα, IL6, CCL2 and IL10 in THP-1 cells. Growth of CAS1 strains with deletions was slower in broth (RD149; p = 0.024 and RD149-RD152; p = 0.025) than that of strains without deletions. CAS1 strains with RD149 deletion strains further showed reduced intracellular growth (p = 0.013) in THP-1 cells as compared with strains without deletions, and also as compared with H37Rv (p = 0.007) and with CAS1 RD149-RD152 deletion strains (p = 0.029). All CAS1 strains induced higher levels of TNFα and IL10 secretion in THP-1 cells than H37Rv. Additionally, CAS1 strains with RD149 deletions induced more TNFα secretion than those without deletions (p = 0.013). CAS1 RD149 deletion strains from extrapulmonary sources showed more rapid growth and induced lower levels of TNFα and IL6 secretion in THP-1 cells than isolates from pulmonary sources. This data suggests that presence of RD149 reduces growth and increases the induction of TNFα in host cells by CAS1 strains. Differences observed for extrapulmonary strains may indicate an adaptation which increases potential for dissemination and tropism outside the lung. Overall, we hypothesise that RD149 deletions generate genetic diversity within strains and impact interactions of CAS1 strains with host cells with important clinical consequences.


Introduction
Molecular epidemiological studies have suggested an association between Mycobacterium tuberculosis strains and geographical locations [1]. Predominant M. tuberculosis clades from the Indian subcontinent include Central Asian strains (CAS) defined by absence of spacers 4-7 and 23-34 [2,3]. Within the CAS strains, CAS1_DEHLI (ST26) has been identified as being the most widespread (39%) in Pakistan [4]. Reasons underlying successful transmission of CAS genogroup strains in South Asia are unknown. Large sequence polymorphisms (LSPs) or regions of difference (RD) are identified as deletions or large sequence polymorphisms in M. tuberculosis strains.
Amongst M. tuberculosis isolates from Pakistan, deletions in RD149 are reported in 39.8% and concurrent RD149 and RD152 in 18.8% of CAS1 strains [5]. The impact of RD149 and RD152 deletions on M. tuberculosis biology is as yet unclear. RD149 region is known to consist of probable phage proteins (Rv1572c-Rv1585c), although the role of phage proteins has not been defined. RD152 region also consists of putative transposases (Rv1756c, Rv1764) along with virulence factor, plcD (Rv1755c) [6,7]. Both RD149 and RD152 deletions are also known to be characteristic of Beijing genogroup strains which are prevalent worldwide, thought to be highly virulent [7] and also associated with multi-drug resistance [4,8]. Clinical presentations of tuberculosis have been shown to differ depending on the infective M. tuberculosis strain group; Euro-American strains are associated with increased lung consolidation, while the meningitis caused by the East Asian/Beijing lineage was associated with younger adults, more rapid disease progression and fewer leucocytes in the cerebrospinal fluid (CSF) [9].
In vitro studies have shown that virulent M. tuberculosis Beijing clinical isolates grow more rapidly than H37Rv in murine and human macrophages as well as in THP-1 cell line model [8,10,11].
M. tuberculosis infection of macrophages has been shown to induce both proinflammatory cytokines; Tumor necrosis factor alpha (TNFa), Interleukin 2 (IL2) and Interleukin 6 (IL6) as well as downregulatory cytokine; Interleukin 10 (IL10) [12]. TNFa is essential for macrophage activation and granuloma formation [13,14]. The release of TNFa in human macrophages infected with M. tuberculosis has the dual effect of increasing antimycobacterial activity by activating macrophages and inducing granuloma formation, while excess TNFa results in host cell necrosis and dissemination of mycobacteria [15]. TNFaalso increases the expression of chemokine CCL2 (monocyte chemoattractant protein (MCP-1)) by macrophages [16]. CCL2 plays a role in granuloma formation [17] and is reported to contribute towards protection against tuberculosis in the murine model [18]. IL10 is important for regulating immunity in mycobacterial infections and while IL10 is associated with progressive disease [19], absence of IL10 results in lack of antimycobacterial immunity [20].
Previous studies have shown that virulent clinical M. tuberculosis strains grow faster and cause more necrosis than laboratory strain H37Rv in murine bone marrow-derived macrophages [10]. Beijing genotype strains induce lower levels of proinflammatory cytokines (TNFa, IL6, IL12) as compared with M. tuberculosis H37Rv in human monocyte derived macrophages [11,21]. Overall, reduced pro-inflammatory responses have been associated with increased intracellular persistence of M. tuberculosis strains [11,22,23].
It is known that RD1 encodes for highly antigenic proteins thought to play an important role in immunomodulation of the host by M. tuberculosis [24]. These and smaller deletions in M. tuberculosis may contribute to adaptation of strains for improved interactions with the host. These include a polymorphism in the pks 15/1 gene which encodes for a phenol glycolipid whereby the mutation in the glycolipid results in immunosuppression of host responses. A previous study has shown that deletion of Rv1519 in CH, an outbreak strain belonging to the CAS2 lineage resulted in a higher secretion of immunosuppressive IL10 and IL6, contributing to its persistence in the human population [25].
This study explored the biological relevance of RD149 and of concurrent RD149 and RD152 deletions on CAS1 strains. In vitro broth growth of the CAS1 strains with RD149 and concurrent RD149-152 deletions was compared with that of CAS1 strains without deletions and with H37Rv. Intracellular growth of these strains was investigated using the human acute monocytic leukemia cell line (THP-1) [26,27,28]. Mycobacterium-induced TNFa, IL6, CCL2 and IL10 secretion was also determined.

Ethics Statement
Approval for this study was taken from the Ethics Review Committee of the Aga Khan University, Pakistan.

Mycobacterial strain selection
M. tuberculosis isolates used in this study were obtained from diagnostic specimens submitted to the clinical laboratory at the Aga Khan University and stored in the strain bank. All strains were subcultured once on Middlebrook 7H10 agar before being stored in aliquots in glycerol peptone broth (GPB) on beads at 270uC. Central Asian Strain1 (ST26) strains (n = 133) from the strain bank at the Aga Khan University previously screened for RD deletions were included [5]. Of these 53 had RD149 deletions alone and 25 strains showed concurrent RD149 and RD152 deletions [5]. From this latter pool, strains with RD149 (n = 9) and with concurrent RD149 and RD152 (n = 9) were randomly selected for this study along with 3 CAS1 strains without deletions in RD149 or RD152 regions. Each strain culture was established from a fresh aliquot of each stored strain. M. tuberculosis H37Rv reference strain obtained from American Type Culture Collection (ATCC) was used as a control.

Growth in broth
Growth in broth was assessed using radiometric methods; Study strains were grown in 7H9 broth medium supplemented by 5% ADC (Difco Laboratories, Detroit, MI, USA). The strains were then inoculated onto Middlebrook 7H10 agar (Becton Dickinson, Sparks, MD, USA) incubated for a 3 further weeks. Bacterial colonies were suspended in 1 ml sterile PBS, vortexed vigorously for 30 sec to obtain a single cell suspension and matched with 0.5 McFarland (approx cell density is 1.5610 8 /ml).
The bacterial suspension was diluted (1:100) to a final concentration of 1.5610 6 /ml, 0.1 ml of the final concentration was added to BACTEC 12B medium culture vials (Becton Dickinson, USA) containing 14 C-labelled fatty acid substrate and incubated at 37uC for up to 37 days [29,30]. The growth in broth for days 0 to 37 is mentioned in Figure S1. As peak growth was observed at Day 2 for all strains measurements at this interval were used for further analysis.

Intracellular growth
M. tuberculosis were first grown in Middlebrook 7H9 broth medium supplemented by 5% ADC (Difco Laboratories, Detroit, MI, USA) at 37uC to mid-logarithmic phase (OD 620 nm = 0?5). Aliquots of mycobacteria at a stock concentration of 2610 6 CFU/ ml in 7H9 broth medium containing 15% glycerol were stored at 270uC for use in monocyte infection assays as described previously [22].
M. tuberculosis inoculum for infection was prepared as described previously [31]. Briefly, THP-1 cells (2610 5 /well) were infected with M. tuberculosis (2610 5 CFU/ml) for 3 h (T 0 ), the supernatants were collected and monolayers washed thoroughly with phosphate buffer saline (PBS) to remove extracellular adherent mycobacteria. For the T 0 wells, THP-1 cells were lysed with 100 ml of 0.5% Triton X-100 and lysate plated on 7H10 agar plates for enumeration by colony forming unit (CFUs) counts.
Fresh RPMI medium was added to the remaining wells until harvested. Cell supernatants were collected at days 1 and 3 for cytokine measurements, were filtered using 0.2 mm filters, aliquoted and stored at 280uC until used. Cell lysates were plated for CFUs at 1, 3, 5 and 7 days post-infection. The intracellular growth pattern observed for each strain is depicted in Figure S2. The largest change in growth was observed between days 0 and 3. Therefore, analysis of intracellular growth for each strain was calculated using the ratio of growth at day3/day 0 for each isolate.

Measurement of Cytokines and Chemokines
Cytokines (TNFa, IL6 and IL10) standards and monoclonal antibody pairs for capture and detection obtained from Endogen (Rockford, IL, USA) were used to obtain a dose response curve with a range of detection from 3.9-1000 pg/ml using a sandwich ELISA technique as described previously. Each well was measured in duplicate. Chemokine, (CCL2) standards and monoclonal antibody pairs for capture and detection obtained from R&D Systems (Abingdon, UK) were used to obtain a dose response curve with a range of detection from 6.25-1000 pg/ml for CCL2. All measurements were carried out according to the manufacturer's instructions and as reported previously [31,32].

Statistical analysis
Non-parametric statistical analysis was performed using the Statistical Package for Social Sciences software (SPSS version 17.0). The analyzed variables for growth indices and CFU/ml were compared using Mann-Whitney U Tests. While the analyzed variables for cytokine analysis was compared using one way ANOVA (Tukey, Post Hoc analysis) and Kruskal Wallis Test.

Results
M. tuberculosis strains were isolated from both male (n = 9) as well as female (n = 12) patients, Table 1. The median ages of male (24.5 y) and female (24 y) patients was comparable, p = 0.36 as analysed using the Mann-Whitney U Test.

CAS1 clinical strains show variable growth in broth
Growth of CAS1 strains with deletions was slower in broth (RD149, p = 0.024; and RD149-RD152, p = 0.025) than that of strains without deletions. Comparison of CAS1 strains with M. tuberculosis H37Rv showed that while CAS1 strains without deletion strains grew more rapidly in broth as compared with M. tuberculosis H37Rv (p = 0.036), growth of H37Rv was faster than that of CAS1 strains with RD149 deletions (p = 0.04), Figure 1. When assessed individually however, the growth of deletion strains showed variability; One strain with RD149; and 3 strains with concurrent RD149-RD152 deletions grew faster, whereas 6 strains with RD149 and 3 strains with RD149-RD152 grew slower than the laboratory reference strain H37Rv ( Figure 2).

Increased TNFa and IL10 induction in CAS1 deletion strains
To investigate association between the presence of deletions and the capacity of M. tuberculosis strains to elicit host cytokine activation. TNFa, IL6, IL10 and CCL2 secretion by THP-1 cells infected with CAS1 strains was measured. Significantly higher levels of TNFa induction was noted by CAS1 strains; CAS1 without deletions (p = 0.006), strains with RD149 (p = 0.005) and with concurrent RD149 and RD152 (p,0.001) as compared to M. tuberculosis H37Rv. Within CAS1 strains though, isolates with RD149 deletions induced significantly higher levels of TNFa as compared to strains without deletions, (p = 0.013), Figure 4A.
Induction of IL6 and CCL2 by the study strains was investigated. However, no differences were detected between IL6 induction by CAS1 strains and H37Rv, or between strains with and without deletions; IL6 levels induced by the study strains were

Differential intracellular growth of strains from pulmonary and extrapulmonary sources
To investigate whether the source; pulmonary or extrapulmonary had any bearing on strain virulence, growth and cytokine activation in response to strains was compared. No difference was observed in growth rates or activation of host cytokines between pulmonary or extrapulmonary strains of CAS1 or CAS1 with concurrent RD149 and RD152 deletions. In contrast, extrapulmonary strains with RD149 deletions showed increased intracellular growth (p = 0.015) and induced lower levels of TNFa (p = 0.01), CCL2 (p = 0.003) and IL6 (p = 0.008) secretion in THP-1 monocytes as compared with strains from pulmonary sources, Table 2.

Discussion
This study showing that presence of RD149 deletions in M. tuberculosis CAS1 strains is associated with reduced growth (in broth and in vitro macrophages) and with increase in proinflammatory TNFa suggests a role of RD149 in causing phenotypic diversity within these strains.
Earlier studies on growth characteristics of M. tuberculosis strains have hypothesized that faster growth of clinical strains such as Beijing in comparison to the laboratory reference H37Rv may contribute to their success in the establishment of disease [8]. Faster growth in broth seen amongst CAS1 strains without deletions as compared with H37Rv in this study may thus be contributing to the success of these strains. However, the presence of RD149 either singly or concurrent with RD152 appears to influence this advantage.
We found that the optimal time period to study intra-strain growth differences was at 3 days post-stimulation. At this time point, CAS1 strains with RD149 deletions showed reduced intracellular growth in THP-1 macrophages as compared with H37Rv, with CAS1 strains (without deletions) and with CAS1 with concurrent RD149-RD152 deletions. This difference was not present at the later 5 day time point. However, the intracellular growth difference observed at the earlier time point is important as it likely to represent early events in mycobacterium-host cell interactions such as, uptake into the phagosome followed by sorting within endosomal compartments of the host cell [33]. Early events post uptake and entry into the macrophage are key to determining phagosome biogenesis and outcome of the infection. Therefore, variations between strains which affect intracellular growth even in an interim period through retardation of the mycobacterial phagosome may well impact host immune response and therefore the outcome of infection.
All CAS1 strains investigated induced higher TNFa as well as IL10 levels in THP-1 cells than the laboratory strain H37Rv. Our data is consistent with reports of increased TNFa and IL10 induction in THP-1 monocytes by clinical strains including the Beijing genotype [8]. TNFa has been shown to play a crucial role in macrophage activation and granuloma formation and thus in controlling mycobacterial replication [34]. TNFa has also been associated with cellular damage and dissolution of the granuloma due to cellular necrosis [34]. While M. tuberculosis infection of alveolar macrophages results in TNFa dependent apoptosis [35] and pathogen killing, it may alternatively also lead to the dissemination of the bacillus due to TNFa driven necrosis [36]. It is therefore possible to hypothesize that increased induction of the down modulatory IL10 together with proinflammatory TNFa by CAS1 strains may thus contribute to skewing the host immune response towards a downmodulatory phenotype which would facilitate intra-macrophage growth of the strains.
CAS1 strains with RD149 deletions induced higher TNFa levels as compared to CAS1 strains without deletions. Clinical isolate CDC1551 strain have been shown to induce higher TNFa responses in monocytes but to persist for shorter periods than the more virulent Beijing strain [37]. Similarly, avirulent Mycobacterium sp. is reported to induce higher TNFa secretion than the virulent strains [34]. Therefore, the reduced intracellular growth and increased TNFa secretion of CAS1 strains with RD149 deletions may be suggestive of a decrease in virulence amongst the deletion strains.
Beijing strains have been shown to differ in their ability to induce cytokines in both in vitro and in animal models [22,38]. Differences between Beijing strains have been attributed to variation in the production of lipids including Phenolic glycolipids (PGL); strains with mutations in the pks 15/1 gene induce higher levels of cytokine-dependent Th1-type protective immune responses such as TNFa [7,39]. Recent studies have identified variations in the transcriptome profiles of M. tuberculosis strains from different lineages suggesting, that genetic variations may impact phenotypic patterns of clinical strains [40]. It is likely therefore that the differences observed in responses to CAS1 strains could be due to inter-strain variations perhaps even due to the presence of RD deletions.
RD149 and RD152 are associated with mobile genetic elements e.g., prophage and hot spots for IS6110 insertion, respectively [41]. The Rv 1759c gene (wag22) within RD152 region is a member of the PE family, polymorphic CG-repetitive sequences (PGRS) subfamily of glycine rich proteins which are putative virulence factors for M. tuberculosis [41]. The combined effect of RD149 and RD152 deletions may differ from that of RD149 deletion strains due to difference in strain properties which result from the variable antigenic profile of strains as a result of their differing presentations. This may subsequently affect uptake and intracellular growth of strains hence, the differential outcome of infections with RD149 strains from that of strains with RD149-RD152.
Within CAS1 strains with RD149 deletions, those from extrapulmonary sites grew faster and induced lower levels of TNFa, CCL2 and IL10 as compared with pulmonary strains. This is consistent with trends from a previous study showing that extrapulmonary M. tuberculosis strains grow faster and induce less TNFa secretion in human monocyte derived macrophages [42].
To date, there is little data on the role of RD149 and RD152 deletions in either the virulence or the epidemiology of TB. The current study comparing clinical strains with and without RD deletions suggests that observed differences in growth and host cytokine modulation are associated with deletions in RD149. We hypothesize that the variations observed in the virulence patterns of M. tuberculosis clinical isolates may be affected by the acquisition of RD149 deletions and may further be dependent on the location of the deletions on the bacterial chromosome. This is important since the data suggests that these deletions may have an impact on pathogenesis and therefore on clinical outcome of TB infection. Figure S1 Measurement of M. tuberculosis growth in broth using the BACTEC system. M. tuberculosis H37Rv was adjusted to approx cell density of 1.5610 6 bacterial cells/ml for inoculation and subsequently grown for 37 days in the radiometric BACTEC 460 TB system. To determine Colony Forming Units (CFUs), bacterial suspension from 12B medium was plated on 7H10 agar medium and incubated for 3 weeks for enumeration. Growth in broth was measured for H37Rv, CAS1 strains (without deletions), CAS1 strains: S1, S2, S5, S6, EP1, EP2, EP4, EP7 and EP10 with RD149 deletions and CAS1 strains: S3, S4, S7, S8, S9, S10, EP6, EP8 and EP9 with concurrent RD149-RD152 deletions ## denotes a growth ratio of strain calculated as growth at day 2/ initial growth at day 0.