Final Report Summary - HOMITB (Host and microbial molecular dissection of pathogenesis and immunity in tuberculosis)
Executive Summary:
The global incidence of tuberculosis (TB) is rising, with 8.8 million new cases and 2 million deaths each year. The rate of progression from infection to disease is highly variable, and approximately 90% of infected individuals never develop clinical disease. About half of the 10% of Mycobacterium tuberculosis-infected individuals who do develop overt clinical disease are diagnosed with TB within 2 years of infection and can be considered fast progressors. This 'primary' TB is particularly common in children and is often disseminated due to early haematogenous spread of the mycobacterium with miliary and/or extra-pulmonary disease.
About 50% of TB patients who develop clinical disease more than two years after infection are commonly described as cases of 'reactivation' or post-primary TB. Post-primary TB is predominantly a pulmonary disease of adults. Although neonatal immunisation with the Mycobacterium bovis-derived Bacillus Calmette-Guérin (BCG) vaccine can provide protection against childhood TB, BCG is relatively ineffective in preventing disease in adults.
Tightly controlled genetically determined interactions between the mycobacteria and different host target cells that are regulated by the host innate and adaptive immune responses dictate the outcome of mycobacterial infection in humans, ranging from an asymptomatic infection to a life-threatening disease.
At HOMITB we identified novel host genes involved in control of bacterial load, and regulation of inflammation and immune responses, and bacterial molecules involved in escape from such immune regulatory and effector mechanisms. Both targeted and genome-wide explorations were employed to identify susceptibility of genes involved either in the immunological control or pathogenesis of primary M. tuberculosis infection in children and secondary pulmonary TB in adults, as well as in the control of mycobacterial infection in the mouse model. Immunological validation of the novel host susceptibility genes were performed.
Mycobacterium host interactions in macrophages, dendritic cells and non-haematopoietic cells such as epithelial cells are studied both in human and mouse models. The role of chemokines, chemokine receptors and key intracellular signalling molecules in the outcome of mycobacterial infection is investigated. Mice with human lymphoid and myeloid lineages de novo developed after transplantation of haematopoietic stem cells are used to explore the human immune responses to infection with Mycobacterium.
HOMITB dissected the molecular and cellular immune responses, combining mice models, and human patients. In both approaches, the responses of various host target cells regulating mycobacterial growth and the mycobacterial adaptation to different host cell responses to infection were studied. The role of innate and adaptive immune responses in the control of mycobacterial growth and the regulation of inflammation was studied as well. In addition, the project investigates such responses using a third model based on mice reconstituted de novo with human haematopoietic-derived cell lineages. Finally, targeted and genome-wide explorations were employed to identify novel susceptibility genes involved in the immune control of primary M. tuberculosis infection and secondary pulmonary TB.
The advanced knowledge of molecular and cellular interactions between host and mycobacteria is crucial for the design of novel vaccination or immunotherapeutic strategies and for the discovery of better diagnostic markers.
Project Context and Objectives:
Tuberculosis (TB), an infectious disease caused by Mycobacterium tuberculosis, remains a leading public health problem worldwide. The global incidence of TB is rising, with 8.8 million new cases and 2 million deaths each year. The rate of progression from infection to disease is highly variable, and approximately 90% of infected individuals never develop clinical disease. About half of the 10% of M. tuberculosis-infected individuals who do develop overt clinical disease are diagnosed with TB within two years of infection and can be considered fast progressors. This “primary” TB is particularly common in children, and is often disseminated due to early, haematogenous spread of the mycobacterium with miliary and/or extra-pulmonary disease. About 50% of TB patients, who develop clinical disease more than two years after infection, are commonly described as cases of “reactivation” or post-primary TB. Post-primary TB is predominantly a pulmonary disease of adults, involving extensive damage to the lungs and efficient airborne transmission of bacteria. Although neonatal immunization with the current vaccine, Mycobacterium bovis-derived bacillus Calmette-Guérin (BCG), can provide protection against childhood manifestations of the disease (primary disease), BCG is relatively ineffective in preventing disease in older children and adults (reactivation disease).
Consequently, latently infected individuals run a life-long risk of reactivation to active disease. The risk is particularly important when these individuals are immuno-compromised, either as the result of HIV infection or of immunosuppressive treatment. As TB reactivation results in the most contagious, pulmonary form of the disease, preventing latently infected individuals from reactivating should constitute a major step towards the elimination of TB. Of importance, it is still not fully understood why some individuals develop active disease, whereas most apparently control the infection by an appropriate immune response.
Therefore, tightly controlled genetically determined interactions between the mycobacteria and different host target cells that are regulated by the host innate and adaptive immune responses dictate the outcome of mycobacterial infection in man ranging from an asymptomatic infection to a life-threatening disease. Thus, knowledge of such molecular and cellular interactions is crucial for the design of novel vaccination or immunotherapeutic strategies, and to the discovery of better diagnostic markers.
In most cases, the host immune system is able to control bacterial replication and to prevent the development of active disease. However, M. tuberculosis is able to reprogram its gene expression, preventing the immune response from totally eliminating the microorganism. Thus, the host becomes latently infected. The identity of these genes and how they control M. tuberculosis adaptation to the intracellular environment remain to be delineated. Infected alveolar macrophages and dendritic cells have long been known as a primary site of M. tuberculosis infection. More recently, other cell types, such as adipocytes, fibroblasts, endothelial cells, and most interestingly, epithelial cells have also been identified as cellular niches for latent M. tuberculosis. Upon infection, these cells can release a panel of antimicrobial effector molecules, cytokines, and chemokines, which govern innate immune responses and initiate specific immunity. This complex network of mediators induces activation of antimicrobial activity in macrophages and perhaps other cell types, migration of antigen laden phagocytes to the draining lymph nodes, and finally the influx of antigen-specific T lymphocytes to the site of infection. Primary infection most often leads to the formation of a well organised granuloma consisting of central macrophages with surrounding lymphocytes and fibroblasts. This cellular microenvironment provides a physical and immunological barrier for the entrapped bacilli thereby preventing multiplication and spread of the mycobacteria. Live mycobacteria may also remain in many different cell types throughout the body and undergo a dormancy phase. Factors disrupting the fine-tuned balance between the low level mycobacterial metabolism and the cellular immune response of the host are likely to increase the risk of tipping the balance in favour of the pathogen and result in reactivation of disease.
AIM
In this project we proposed to identify novel host genes involved in control of bacterial load, and regulation of inflammation and immune responses, and bacterial molecules involved in escape from such immune regulatory and effector mechanisms. Both targeted and genome-wide explorations were employed to identify susceptibility genes involved either in the immunological control or pathogenesis of primary M. tuberculosis infection in children and secondary pulmonary TB in adults as well as in the control of mycobacterial infection in the mouse model. Immunological validation of the novel host susceptibility genes were performed. In parallel, the M. tuberculosis strains collected in field studies will be typed by the most recent and discriminatory molecular typing method, based on Mycobacterial Interspersed Repetitive Units (MIRUs). Mycobacterium host interactions in macrophages, dendritic cells and non-haematopoietic cells such as epithelial cells will also be studied both in human and mouse models. The role of chemokines, chemokine receptors and key intracellular signalling molecules in response to mycobacterial infection was analyzed. Mice with human lymphoid and myeloid lineages de novo developed after transplantation of haematopoietic stem cells were used to explore the human immune responses to infection with Mycobacterium.
SPECIFICALLY
We searched for novel genes involved in the susceptibility of children to disseminated TB. By using a candidate gene approach, we investigated whether known MSMD (Mendelian Severe Mycobacterial Diseases) causing genes and new immune genes involved in the IFN-γ production and signalling pathway are important for control of childhood disseminated TB. Furthermore, new genes involved in control of adult pulmonary TB were identified by positional cloning, and functionally validated. The identification of single nucleotide polymorphisms (SNP) conferring a major effect in predisposition to adult TB was based on large association familial studies. SNPs in an interval in chromosome 8 known to predispose to pulmonary TB, and in MSMD causing genes will be genotyped. Functional characterization of alleles predisposing to TB was performed by biochemical and immunological methods.
Mycobacterial adaptation into different target cells was investigated. The mycobacterial heparin-bindin hemagglutinin (HBHA) is a protective antigen, especially well recognized by latently infected subjects. The expression of hbhA is controlled by the intracellular environment.The hbhA gene is strongly up regulated in epithelial cells but not in macrophages. The mechanisms of regulation of the hbhA gene within infected epithelial cells and other cell types was investigated. Other members of the same regulon were identified by comparing the mycobacterial gene transcriptome after infection of different host cells. The role of such genes in persistence in non-haematopoietic cells and macrophages was analyzed.
Macrophages have long been recognized as key players in the pathogenesis of TB and as sites for M. tuberculosis infection. We defined the relative role of non-haematopoietic cells, with emphasis on alveolar epithelial cells in the outcome of M. tuberculosis infection in humans, murine and humanized murine models. The host cell transcriptome of IFN-γ-stimulated macrophages and of epithelial cells infected with Mycobacterium was compared. The antigen-presenting capacity of epithelial cells infected with mycobacterium was tested. The mycobacterial transcriptome during infection of macrophages and epithelial cells stimulated with IFN-γ was analyzed. Differential expression of mycobacterial genes were related to the adaptation to diverse effector mechanisms in these cell populations.
Dendritic cells (DC) represent key components of the immune system mediating initiation of adaptive immune responses that restrict M. tuberculosis infection. However, during infection, the specific functions of DCs are modulated by mycobacterial molecules thereby affecting the ensuing adaptive responses and the outcome of the infection. We proposed to perform detailed analysis of the relative role of distinct DC populations in driving immune responses to M. tuberculosis. We dissected the molecular pathways in DC activated by M. tuberculosis infection or mycobacterial molecules and the mechanisms by which mycobacteria interfere or promote DC functionality.
Chemokines play a key role in cell migration and are therefore crucial to innate and adaptive immunity to infection and inflammation. The initial innate immune response to M. tuberculosis is triggered by Toll-like receptors (TLR) on host cells, responding to M. tuberculosis-derived pathogen-associated molecular patterns (PAMP). The interaction of PAMP with TLR triggers inflammatory cytokine release culminating in the production of inflammatory chemokines which control the migration of the leukocytes into the infected area. A mycobacterial lipoprotein (agonist of TLR2) selectively suppresses inflammatory chemokine receptor expression and down regulates inflammatory cell responses to chemokines. Interference with this TLR2/chemokine receptor axis may therefore be instrumental in the re-emergence of latent M. tuberculosis infection. We will investigate the relationship between TLR2 signalling, chemokine receptor expression and granuloma formation. In addition, the role of a chemokine decoy receptor D6 involved in the suppression/resolution of pulmonary inflammation and thus representing a good candidate for involvement in the differential pathogenesis of TB of humans and mice will be studied. For this purpose, genetic alterations in the D6 chemokine decoy receptor that are associated with pulmonary TB will be determined and tested functionally. In parallel the role of D6 in the outcome of mycobacterial infection will be tested in mouse model of tissue-specific D6 transgenic mice.
Immune and inflammatory systems are controlled by multiple soluble mediators. SOCS are a family of intracellular proteins, several of which have been shown to be regulators of cytokine homeostasis. We studied the regulation of expression and role of SOCS1 and 3 molecules during the mouse mycobacterial infection in vivo and in macrophages. The levels of SOCS proteins in cells from patients or asymptomatic infected individuals were studied.
We will use a model of BCG and M. tuberculosis infection in RAG-1-/-/γcR-/- mice reconstituted with human lymphoid and myeloid cells. Such “humanized” mice, carrying a functional human immune system, will be of considerable importance to experimentally decipher the cellular and molecular mechanisms of anti-mycobacterial immunity in humans. This model will permit the first experimental investigation of human immunity to mycobacteria.
HOMITB represents a combined effort of 9 leading groups in genetics, molecular microbiology, cellular biology, immunology and clinical practice in TB The grouping, which also includes a highly successful SME. In this proposal, we will deploy our recent findings in basic and applied research to tackle TB, using the most advanced and innovative approaches. We were the lead member of international collaborations in defining human immune response genes directly related to susceptibility to TB. This advanced concept and technology will be used to further pin-point the host genetic markers for TB. Much past effort has been focused on the role of macrophages in TB. Since non-haematopoietic cells are also infected by M. tuberculosis, we propose a novel approach to investigate the role of pulmonary epithelial cells, endothelial cells and fibroblasts in TB based on our recent research in the function of these cells. Recent understanding of the physiology and functions of DC has been a major advance in immunology. We will apply this new knowledge to TB research. The role of chemokines in TB is virtually unknown. Our strength in chemokine biology and biochemistry and the novel finding of the crucial role of bacterial products (TLR2-ligands) in modulating chemokine receptor expression led us to a highly innovative approach to explore the role of chemokines/chemokine receptors in regulating innate and adaptive immunity to M. tuberculosis. We have recently found that the SOCS family members play an important role in infection. We will apply this novel approach to explore the role of this key transcription factors in M. tuberculosis infection. We have recently constructed humanized mouse carrying a functional human immune system, in which the responses to experimental infection with BCG and M. tuberculosis will be studied. A real time method will be used to evaluate the presence of mycobacteria. This state-of-the-art technology will be employed to test the findings obtained above in a model system as close to the human disease as possible.
This closely interactive grouping with extensive complementary expertise (basic immunology, cell biology, human and mycobacterial genetics, clinical field practice, industrial facilities and knowledge) and strong track record in collaborative research created a powerful research centre to deliver highly innovative approach to tackle one of the most important diseases in the world.
Project Results:
HUMAN GENETIC BASIS OF TUBERCULOSIS
The Mendelian basis of paediatric tuberculosis
Candidate gene approach
Patients included in this part of the project are children (≤15 years) with a disseminated/extra-pulmonary form of TB. These forms include miliary TB, meningitis, peritonitis, and osteomyelitis (in particular Pott’s disease). As these forms are often bacteriologically negative, the inclusion criteria are based on a stringent combination of clinical, radiological, biological, and histological features. Two kinds of samples have been collected and used for the results presented here:
- a first sample of 32 Moroccan children with disseminated TB and available DNA that were used for hypothesis-based systematic sequencing of candidate genes and immunological studies. In a second step, we have collected an additional sample of 30 Moroccan children.
- a sample of 21 consanguineous families with at least one child with disseminated TB that were used for the hypothesis-generating GWL study.
The first goal was to sequence the genes involved in the syndrome of Mendelian susceptibility to mycobacterial diseases (MSMD), in particular IFNGR1, IFNGR2, STAT1, IL12B, IL12RB1, in our sample of 32 Moroccan children. The two receptors of IL12, IL12RB1 and IL12RB2 were fully sequenced with a coverage of 100%. No unreported mutation was found in IL12RB2. In IL12RB1, we identified one patient carrying a homozygous nonsense mutation in exon 9 of IL12RB1 leading to a premature stop codon in amino acid position 305 (K305X). The two parents are heterozygous for the mutation, and we showed that this mutation was complete loss of function. The observation of at least one IL-12RB1-deficient child among 32 patients may at first glance seem fortuitous. However, we recently replicated this observation in similar surveys in Turkey and Iran, implying that recessive mutations in IL12RB1 loci may universally be found in a substantial fraction of children with severe TB (about 1-3%). These results strongly suggest that pediatric TB may actually result from a collection of single-gene defects. Up to 3% of severe pediatric TB may be accounted for by mutations solely in the IL12RB1 gene. This work was published in 2011. No interesting candidate mutations were found in the other genes sequenced (IFNGR1, IFNGR2, STAT1, IL12, NEMO) as well as in ISG15 that we recently identified as a new MSMD gene. We are now pursuing this investigation directly by exome sequencing in order to identify new genes and maybe new pathways involved in Mendelian predisposition to TB.
Genome wide linkage scan
In order to determine which new genes are involved in susceptibility to disseminated TB in children, we have already recruited 21 consanguineous families, extracted gDNA from each member and conducted a genome-wide linkage scan using an Affymetrix 250K array. The results were analysed by the same technique that was used for 17 other consanguineous families with MSMD, a technique that has proved to be efficient. Indeed, genome-wide (GW) homozygosity mapping (HM) is a powerful method to locate rare recessive Mendelian mutations. However, statistical power can decrease dramatically in the presence of genetic heterogeneity, which is likely to increase with sample size. We thus developed a novel approach to test for linkage accounting for genetic heterogeneity in the context of GWHM. We used this novel method to analyse a GWL study conducted with 21 Moroccan consanguineous families with at least one child presenting with severe TB. The results obtained for regions involved in TB showed three interesting regions on chromosomes 4, 9 and 11.
We focused first on the chromosome 9 region as the interval contains the gene JAK2. JAK2 is a janus kinase involved in many different signalling pathways including IL-12 and IFN-γ. This renders JAK2 a good candidate gene. In one patient who had suffered from miliary tuberculosis, sequencing of the 25 exons and the flanking intron regions revealed a homozygous missense mutation at amino-acid position 417, leading to the replacement of a glycine to a serine (G417S) (Figure 1). This variation has not been previously reported as a polymorphism in databases. In addition, we have sequenced gDNA from 200 healthy controls (half of them being Moroccan) and did not find this variation. The glycine at position 417 is conserved among species and the amino acid substitution is predicted to be possibly damaging functionally, using the polyphen software. Preliminary functional studies showed an impairment of STAT4 phosporylation upon IL12 stimulation in the patient cells suggesting that this mutation may affect the IL12 response. We are currently testing the JAK2 mutation functionally using EBV-B and Saimiri-T cell lines. We will study not only the response to IL-12 and IFN-γ, but also the cellular responses to other cytokines, such as IL-6, IL-23 and IL-27. We will then characterize the mutant alleles by gene transfer. Altogether, these data strongly suggest that this variation is indeed a pathogenic, TB-predisposing mutation, rather than an irrelevant polymorphism.
Overall, our preliminary results of disseminated TB of childhood in Morocco strongly suggest that mutations in IL12RB1 may account for a detectable fraction (about 1-3%) of pediatric TB. We are pursuing our sequencing discovery strategy at the genome-wide scale thanks to the new exome-sequencing technology. Moreover, the discovery of TB-predisposing genes may also benefit from GWL approaches. We are now validating the pathogenic role of the JAK2 mutation and analyzing the other two chromosomal regions linked to pediatric TB together with exome-sequencing data. All these results are consistent with our bold model as well as the Bayesian prediction that pediatric TB results from a collection of immunologically related single-gene defects
The genetics of human adult pulmonary TB
The main objective is the fine-mapping of a region on chromosome 8 linked with pulmonary TB in a) a family-based sample with at least one offspring with TB, and b) an independent case/control sample (>300 patients with TB and >600 controls). For this study, only patients >15 years with bacteriologically confirmed pulmonary TB are included. The family of the patient is enrolled if both parents or at least one parent and one unaffected older sibling are available. This family-based sample constitutes our primary sample, which is used in both the linkage disequilibrium mapping of the chromosome 8 region and the GWA study. Otherwise, patients with bacteriologically confirmed pulmonary TB without available family members are included in the case/control sample. The primary family-based sample with available DNA that has been used in the analyses described below consists of 185 families, including a total of 302 PTB patients.
The case/control sample was used as the replication sample for the results found in our primary sample. We have now collected a sample of 290 pulmonary TB patients and 750 healthy controls (blood donors) that was used for replication studies.
We previously reported the first region linked with pulmonary TB, on chromosome 8. Based on linkage results, a 90% confidence interval was defined to establish the peak region corresponding to a 1-LOD drop from the chromosomal position recording the highest multipoint LOD score. The region was further extended to avoid partial coverage of genes at the borders, so that the final region was defined from 55257 kb to 61363 kb. We previously observed that minor allele frequencies (MAFs) and linkage disequilibrium (LD) patterns of populations from Morocco most closely resemble those observed in European populations from among populations included in the International Hapmap Project (http://www.hapmap.org). Within the linked region, SNPs were selected using Hapmap data from CEPH (Utah Residents with Northern and Western European Ancestry) (CEU) and based on tag-SNP information available on http://www.hapmap.org in order to ensure 95% coverage of CEU tag SNPs at an R2 cut-off of 0.80. These SNPs also cover 93% of CEU tag SNPs at a higher R2 cut-off of 0.90. A total of 3183 SNPs were selected and genotyped on an ultra-high throughput Illumina platform (This represents the double of SNPs that were initially scheduled in our project in order to directly perform an ultra-fine association mapping of the linked region). The final set of SNPs was filtered to remove those with a genotyping call rate (CR) < 80%, a Hardy-Weinberg equilibrium (HWE) P-value<10-6 a MAF <1%, and >10 Mendelian inconsistencies, leaving 2865 high quality SNPs. The software package PLINK, which is optimized for handling large genome-wide datasets, was used for data management and the formatting of data files for analysis programs.
The analysis was based on methods developed for family-based association studies, in particular the FBAT program. The strongest association was found at rs2176641 located near TOX under the additive model (P=0.0002). A total of 49 SNPs achieved a P-value under 0.01 (see figure), and were selected and genotyped in our replication Moroccan sample comprised of 317 independent cases and 650 independent controls. Results showed that only a single signal (based on two correlated SNPs, TOX1 and TOX2) in the proximity of the TOX gene was replicated under the very same genetic model (subjects homozygous for the rare allele are at risk of pulmonary TB) in the case/control replication sample (combined Moroccan analysis p-value for TOX1: 1.1 x 10-5). In addition, the association was much stronger (combined Moroccan analysis p-value for TOX1: 4.3 x 10-8) in the subsample consisting of the younger patients (age at onset of TB < 25 years). For the TOX1 SNP, the odds ratio of developing PTB for AA homozygous subjects vs. AG/GG subjects was estimated at 3.1 (2.0-4.9).
We also attempted to validate this main finding in another pulmonary TB family-based population from Madagascar comprised of 254 trios (two parents and one affected PTB offspring). We genotyped in this sample our two main TOX SNPs as well as 6 other SNPs in LD with those SNPs in Hapmap populations of different origin. Although the frequencies of the risk allele of TOX1 and TOX2 SNPs are lower in Madagascar (~0.15) as compared to Morocco (~0.40) we could replicate the association for the TOX2 SNP (p<0.04) in the sub-sample of patients with age at onset of TB < 25 years (table 1). Interestingly, the evidence for replication was even stronger in the other correlated SNPs which present a higher frequency in Madagascar (~0.22). The TOX gene is an interesting candidate gene as it codes for a nuclear factor (“Thymocyte selection-associated high mobility group box” protein) which has several roles in the immune response, and in particular the development of CD4+ T cells. We are currently conducting functional studies exploring the association of our TOX SNPs with a) the expression of TOX in different cell types, and b) the distribution of lymphocyte population subsets (in particular CD4/CD8). A manuscript reporting these findings is in preparation.
In addition to the positional cloning strategy of the chromosome 8 region, we are conducting a genome-wide association (GWA) study which is making possible to investigate the whole human genome for genetic variants associated with common diseases such as pulmonary TB. The genotyping was performed in our primary family-based sample, and the subjects were genotyped with the Illumina Hapmap610 including 620,901 markers (599,011 SNPs and 21,890 monomorphic markers used for CNV analysis). Of a total of 599,011 SNPs included on the chip, 566,317 SNPs met the quality control criteria. We conducted the same analyses as described for the previous Aim 2.1 leading to a total of 272 SNPs displaying a p-value <5x10-5. All these SNPs were genotyped and analyzed in the replication sample of 317 cases/650 controls, again as described above. The replication study showed that at least three regions on chromosomes 3 (2 regions) and 4 presented replications based on the same allele as compared to results from the primary sample with combined p-values ranging from 2x10-5 to 4x10-5. Replication studies are ongoing in the Madagascar sample as well as in other samples from Mexico and Columbia.
Recently published GWA studies as well as our own GWA study in Morocco did not provide any evidence for strong effect of common polymorphisms in pulmonary TB. We therefore plan to assess the role of rare coding variants in pulmonary TB (which cannot be properly investigated by GWA studies) by conducting whole exome sequencing in a subsample of pulmonary TB patients (in particular including the newly recruited patients) and exposed controls. All the variants found to be interesting (e.g. strong excess in patients as compared to controls) in this first study will be explored in the whole Moroccan sample, and, when confirmed, will be investigated by functional studies.
Our identification of susceptibility variants in the proximity of TOX represents a major breakthrough in the human genetics of Pulmonary TB, in that this finding has been replicated in Morocco and validated in Madagascar. Moreover, TOX encodes Thymocyte selection-associated high mobility group box protein, and is a key transcription factor leading to CD4 cell development in the murine model. The role of T-cells in mycobacterial infection and disease is becoming increasingly recognized. The positional cloning approach thus identified a novel PTB candidate gene implicated in processes that are under active investigation among immunologists interested in host-Mycobacterium tuberculosis interaction.
MICROBIAL ADAPTATION AND TARGET CELL RESPONSES TO MYCOBACTERIAL INFECTIONS
Responses of macrophages and epithelial cells to infection with M. tuberculosis
Studies with primary human cells and cell lines
The expression of cytokines, chemokines and molecules involved in antigen presentation after infection of A549 AECII human cell lines, mouse pulmonary epithelial primary cells and of human primary epithelial cell cultures with BCG, M. tuberculosis, or stimulated with mycobacterial products was compared to that of autologous (syngeneic) macrophages.
Alveolar epithelial type II cells (AEC II) have been proposed to be of importance in pulmonary infections. In the lung tissue, there are 10 times more AEC II than alveolar macrophages (AM) and therefore, the chance that M. tuberculosis encounters an epithelial cell is higher than encountering a macrophage during the initial steps of infection. We postulate that type II pneumocytes are an important reservoir of Mtb that could be responsible for a sustained presentation of HBHA to T lymphocytes and for a subsequent chronic stimulation of these cells that could participate to a better control of the infection.
The murine alveolar macrophage cell line MH-S derived from BALB/c mice and as AEC II we used T7 a transformed cell line on C57BL/6 genetic background. In parallel, primary AEC type II and alveolar macrophages were isolated. A method in which AECII were isolated after negative depletion of CD45+ leukocytes and CD146 endothelial cells was used for this purpose. Isolated AECII consisted of 9% of total cells. When incubated with BCG, LPS or TNF-α, type II cells produce MCP-1 and MMP-9, while alveolar macrophages produce MIP-2 and TNF-α .
The response of human AECII to mycobacterium was also studied. To determine the role of type II pneumocytes in HBHA presentation, we first investigated whether primary type II pneumocytes fresh and the A549 cell line possess surface molecules required for efficient presentation to T lymphocytes and whether these molecules are modulated by IFN-γ and TNF-α inflammatory cytokines as well as by infection with attenuated Mtb (BCG and M. tuberculosis phoP mutant SO2 strain). We first observed that fresh AECII constitutively express HLA-DR molecules as opposed to A549 cell line for which the HLA-DR molecule expression is low and not inducible. Thus, A549 cannot be used as a model for studying class II antigenic presentation. Classical co-stimulatory molecules of the B7 family (CD80, CD86 and B7-H2) were either not detectable or expressed at very low level by few cells, and this is the case for both the A549 cell line and for the human freshly isolated AECII. On the opposite, alternative co-stimulatory molecules (CD54 and CD58) are expressed at very high level by most cells. Finally stimulation with IFN-γ and/or TNF-α increases significantly the cell surface expression only of CD54 and we observed that infection with the two attenuated M. tuberculosis strains used has no significant effect on phenotypic markers expression. Thus, AECII possess HLA molecules required for antigenic presentation and also alternative co-stimulatory molecules that could help for an efficient antigenic presentation to CD4+ and CD8+ T cells.
Regarding the molecules produced secreted by AECII, we observed that high level of chemokines are secreted by AECII (MCP-1, RANTES, IL-8, IP10) after exposure to IFN-γ and TNF-α or infection with attenuated M. tuberculosis SO2, allowing the recruitment of T lymphocytes and other cells. Only IL-6 is produced at high level. No IL1-β, IL-15, IL-18, IL-23, TNF-α, IFN-α, IL-10 or IFN-γ were detected in our conditions. TGF-b is secreted at basal level by AECII. Mtb-SO2 infection increases chemokine secretion. BCG infection has no effect on chemokines/ cytokines production.
In order to investigate the possible antigenic presentation of HBHA by AECII to autologous T lymphocytes, AECII were freshly isolated from lung fragments removed surgically from cancer patients and blood was collected in parallel. Eleven patients were recruited, but their Mtb infection status was unknown. Therefore, whereas the aim of the study was to analyse the presentation of mycobacterial antigens (PPD, HBHA and ESAT-6), numerous control antigens were added in the study to validate the co-culture model as in addition we did not know whether AECII could activate or anergize T cells in a recall response. Thus, CMV-pp65 protein, EBV-LMP1 overlapping peptides, EBV-LMP2A overlapping peptides, Candida albicans-MP65 protein, tetanus toxoid from Clostridium tetani, were tested in parallel to the mycobacterial antigens as some patients might have been infected in the past either with Epstein-Barr Virus (EBV), Cytomegalovirus (CMV), C. albicans and as most of them should been immunized against tetanus. In addition, classical antigen-presenting cells, namely alveolar macrophages and blood monocytes, were tested as controls in parallel to the AECII for their capacity to present HBHA as well as the other mycobacterial and control antigens to autologous T lymphocytes.
In parallel, IFN-γ release assay (IGRA) in response to the different antigens were performed on the patient’s PBMC to determine their immune status toward the different antigens that were tested. IGRA indicated that three patients responded to PPD, three to CMV-pp65, two to C. albicans-MP65 and two to ESAT-6, whereas only one patient responded to HBHA and was therefore considered as latently infected with Mtb (LTBI subject). One patient presented in addition to his lung cancer, a cervical adenopathy that was demonstrated to be due to a Mtb infection. PBMC from this patient produced IFN-γ in response to ESAT-6 but not in response to HBHA (Figure 2, grey symbol).
We showed that human AECII are able to present antigens to autologous lymphocytes, as exemplified on Figure 2 for PPD, ESAT-6, and to a lesser extent for HBHA, and in other patients for CMV-pp65. However, the antigen presentation by macrophages was most often more efficacious.
As IFN-γ is probably not the only read-out of an efficacious antigen presentation, secretion of other cytokines than IFN-g as well as the lymphocyte proliferation were measured. TNF-α and IFN-γ showed a similar response pattern. In contrast, among several other cytokines / chemokines measured, IP10 was detected in the co-culture of HBHA- or ESAT-6-loaded AECII with lymphocytes from the TB patient, who had only a very low IFN-g response at least to HBHA.
In conclusion, we have shown that human AECII play a role in the presentation of mycobacterial antigens to autologous T lymphocytes. IFN-g secretion in the coculture is not the only interesting read-out and it should be combined at least to measurement of IP10 concentrations. A few more patients, especially selected as responding to HBHA, will be evaluated before writing the manuscript to report these data.
Alveolar epithelial cells in murine infections with mycobacterium.
We compared primary murine alveolar epithelial cells with pulmonary macrophages (PuM) in their ability to internalize and control M bovis BCG growth and their capacity as APCs. AEC internalized and controlled bacterial growth as well as presented antigen to primed T cells.
Secondly, we compared the capacity of both cell types to secrete cytokines and chemokines upon stimulation with mycobacterial products. Activated PuM and AEC displayed different patterns of secretion.
Finally, we analyzed the profile of response of AEC microbial and immune stimuli exemplified by TLR ligands and cytokines, respectively. Interestingly, TNF could stimulate the production of CCL2/MCP-1. Since MCP-1 plays a role in the recruitment of monocytes and macrophages to sites of infection and macrophages are the main producers of TNF, we speculate that both cell types stimulate each other. Another cell-cell interaction was suggested when IFNs (produced mainly by lymphocytes) were able to induce expression of chemokines (IP-10 and RANTES) by AEC involved in the recruitment of circulating lymphocytes to areas of injury, inflammation, or viral infection.
Thus, we confirm previous data on the capacity of AEC regarding internalization of mycobacteria and their role as APC, and extend the knowledge of AEC as a multifunctional cell type by assessing the secretion of a broad array of immune factors in response to several different types of stimuli.
The respiratory epithelium is a physical and functional barrier actively involved in the clearance of environmental agents and composed of type I and type II alveolar epithelial cells (AEC). Another cell important in lung defense is the pulmonary macrophage (PuM). We hypothesized that AEC play a role in transmitting signals from the airway mucosal to systemic compartments by modulating the activity of PuM. For this, we collected supernatants from unstimulated or LPS stimulated AEC. We also tested the direct effect of AEC-derived media on bacteria and found that AEC-secreted factors had a dual effect, in one hand controlling bacteria growth and on the other hand enhancing macrophage microbicidal activity.
We have shown that soluble factors released by alveolar epithelial cells (AECs) increase macrophage intracellular killing of BCG. We aimed to identify this bacteriotoxic mechanism. Murine bone marrow-derived macrophages (BMM) and pulmonary macrophages (PuM) infected with BCG. After infection, bacteria were removed and cells incubated further in media with or without IFN-γ or supernatants from AEC (AECsup). Treating BMM with IFN-γ, increased intracellular killing compared with that of cells cultured in medium alone whereas in PuM IFN-γ was ineffective. In both cell-types, IFN-γ treatment increased transcription of iNOS, IP-10 and IL-12 and secretion of IL-12 and IL-6. Transcription of suppressor of cytokine signaling (SOCS)1 was higher in PuM than BMM. SOCS1 expression was important for mediating the ineffectiveness of IFN-γ in increasing intracellular killing in PuM, as PuM from SOCS1/IFN-γ-/- mice showed enhanced killing after IFN-γ treatment. On the other hand, AECsup treatment did not affect iNOS, IP-10 nor IL-12 expression but increased Arg1 transcription and IL-6 secretion. BMM and PuM responded to IFN-γ indicating that both cell types have receptors that recognize and transmit signals delivered by IFN-γ. IFN-γ only increased intracellular killing of BCG in BMM but not in PuM. On the other hand, treatment with AECsup increased intracellular killing of BCG in both types of macrophages. Such killing was not associated with increased IL-12 or iNOS expression.
Thus, AEC supernatants but not IFN-γ augment the intracellular killing of BCG by pulmonary macrophages in a NO-independent manner.
Mycobacterial adaptation to infection
The hbhA gene is known to be differentially expressed during infection of different cell types. The aim of this part of the work is to better understand this regulation and find factors involved in the differential expression. The expression profile of mutants affected in these factors will help us to find hbhA co-regulated genes that could code for proteins involved in the persistence or latency of Mycobacteria. The basis of this approach is the study of the expression of hbhA and downstream genes potentially correlated in an operon structure by using qRT-PCR during axenic cultures or during infections of different cell types with BCG or M. tuberculosis. We developed first a method to study the expression of hbhA in in vitro growth conditions. During growth, aliquots were collected at different time points to extract RNAs. The expression of hbhA was then analyzed by quantitative PCR (qRT-PCR).
These studies showed a regulation of the hbhA relative expression during growth and evidenced also an up-regulation of the hbhA relative expression in M. tuberculosis compared to M. bovis BCG.
Upstream of the hbhA gene, a putative regulatory gene encoding a protein of the Xre family of transcriptional regulator is located. This gene is present in all sequenced mycobacterial genomes, associated to hbhA. The Xre family of transcriptional regulator has been described in the past to be involved in the repression of Bacillus subtilis prophages cycle. More recently, some members of this family have been proposed to act as transcriptional activators. The distance between this gene and hbhA is sufficiently long in the M. tuberculosis genome to contain a hbhA promoter. By using allelic exchange, the xre-like gene was deleted in the BCG genome. The expression levels of hbhA and adjacent genes in BCG-Δxre mutant strain have been compared to the levels in BCG by qRT-PCR normalised to the 16S rRNA gene during in vitro growth. During the early phase growth, we observed a relative expression decrease of hbhA in the mutant strain compared to BCG of about 17-fold. Later in the growth curve (mid-log phase), the measured decrease was around 4-fold. These amplitudes of downregulation are high, compared to amplitudes usually observed in mycobacterial transcriptomic studies. In contrast, for the two downstream genes the maximum amplitude of decrease in their expression was around 1.7-fold at the early phase growth stage. Then the transcription of these genes was even less affected by the deletion of xre-like gene, suggesting that their expression is not correlated to the hbhA expression and that the Xre-like regulator activates the expression of hbhA only. As a control, the expression of a gene located three genes upstream of hbhA was not at all affected by the deletion of xre-like. To confirm that the Xre-like regulator is involved in the activation of hbhA, we transformed BCG with a construction allowing for the overproduction of Xre-like by using an inducible promoter fused to its gene and compared the expression level of hbhA in this recombinant strain with that of non-recombinant BCG. After culture of both strains up to mid-log phase, we measured by qRT-PCR a 3.8-fold increase of hbhA expression in the recombinant strain. This confirmed that the xre-like gene located directly upstream of hbhA codes for a transcriptional regulator that is an activator of hbhA. For this reason, we decided to rename this gene “hbhR”.
We failed to obtain reproducible results by using this vector for complementation. It was most probably due to the fact that the promoter used has been demonstrated to behave randomly in the M. bovis BCG background. For this reason, we decided to complement the deleted strain by using a plasmid containing the hbhR gene under its own promoter and compared the expression level of hbhA in this recombinant strain and in BCG. After culture of both strains up to mid-log phase, we measured hbhA expression by qRT-PCR. There was no restoration of expression in the recombinant strain despite a correct expression of hbhR. This suggests that there is no trans-acting activation of hbhA complementing the deletion of hbhR. In parallel, the HbhR transcription factor fused to a histidine tag has been over-expressed in E. coli and purified for molecular analyzes, such as gel retardation assays on the hbhA promoter and on its own promoter, since many transcriptional regulators are under their own control. No binding of HbhR on the hbhA promoter has been observed in these conditions, but binding on its own promoter has been observed. From the lack of complementation, we thus conclude that the knock-in deletion of hbhR has most probably a polar effect on the expression of hbhA. This implies that hbhA expression can be driven through a promoter located upstream of hbhR. This gene is separated from the upstream predicted ORF by a 86 bp region in which we could predict a promoter with -35 and -10 boxes by using BPROM algorithm (www.softberry.com). This algorithm predicted also a consensual promoter upstream of hbhA in the proximal part of the 353 bp region separating hbhR and hbhA. Moreover, a second promoter for hbhA in the distal part of this region has been recently described to be SigF-dependent. Despite the presence of these two promoters for hbhA, our results suggest that the promoter located upstream of hbhR is mainly used for the transcription of hbhA in M. bovis BCG 1173P2 as knock-in in the hbhR gene strongly affects it. The absence of binding of purified HbhR on the hbhA promoter suggests that, in BCG, HbhR has no direct interaction for the activation of hbhA.
The same experiments were performed in the M. tuberculosis background. A H37Rv-∆hbhR mutant was generated. As for the BCG-∆hbhR, we analyzed the HBHA protein level and the hbhA gene relative expression. As observed with BCG-∆hbhR, the results showed a decrease in the HBHA protein level and a down-regulation (approximately 4-fold) of the hbhA gene expression in the mutant compared to the parental strain. As for the BCG-∆hbhR mutant, the complementation did not restore the wild-type phenotype.
To characterize the ∆hbhR mutant, we followed the expression of hbhA after infection of human epithelial type II pneumocytes (A549 cells line) by qRT-PCR. BCG-∆hbhR and its parental strain were used to infect A549 and RNA from both strains was collected after 24, 48 and 72 hours of infection. Expression levels of hbhA were measured at each time point and compared. In these conditions, hbhA transcription levels measured showed a decrease of about 8.3-fold in the mutant strain when compared to BCG after 24 hours. In both strains, the level of transcription reaches a maximum after 48 hours of infection and then decreased, but the level of transcription in the mutant strain is always lower than in BCG. This observation, as well as the observation of regulation of hbhA transcription during axenic cultures even in the absence of hbhR suggests an additional mechanism of regulation that may occur on the promoter region of hbhA. This mechanism could be relayed by another transcription factor or by SigF.
To continue our analysis of the HBHA regulation, the BCG-∆hbhR and H37Rv-∆hbhR mutants were integrated in microarray experiments to define the hbhR co-regulated genes. When we compared the ∆hbhR mutant with the parental strain, the main significant result is the down-regulation of hbhA confirming the previous results obtained by qRT-PCR. The microarray comparison between the both strains evidenced 20 up-regulated genes in the BCG ∆hbhR mutant with the parental strain. Among these genes, 43% are conserved hypothetical proteins, the others are involved in different metabolisms (cell wall, lipid, respiration).
Then, the comparison between the both strains evidenced 35 down-regulated genes in the the ∆hbhR mutant with the parental strain. Among these, hbhA was found confirming the previous results obtained by ELISA and qPCR. The others are again involved in different metabolisms (cell wall, lipid, respiration). Interstingly, most precisely the down-regulated genes, a lot of them are located into the same locus, the katG locus. The katG gene encodes for a catalase peroxidase that is involved in the conversion of isoniazid to its active form in M. tuberculosis.
CELLULAR IMMUNITY AND REGULATION OF HOST PATHOGEN-INTERACTIONS
Dendritic cells and M. tuberculosis
Phenotype and functional responses of DCs stimulated with mycobacterial molecules
During M. tuberculosis infection, dendritic cells (DCs) function is modulated by mycobacterial molecules, affecting the ensuing protective immune response and the outcome of the infection. In this project we proposed to perform detailed analysis of the relative role of distinct DC populations in driving immune response to Mtb. Thus, we dissected the molecular pathways linked to DC activation following Mtb infection or exposure to mycobacterial molecules. Moreover, we analyzed the mechanisms by which mycobacteria interfere or promote DC functionality.
We first analysed the phenotype and functional properties of DCs derived from monocytes from active Tuberculosis patients, asymptomatic infected (LTBI) and non-infected individuals. To this end we took advantage of a DC-differentiation in vitro model by exposing peripheral monocytes to interferon-α (IFN-α). IL-4 inducing the phenotype of immature DCs (IL-4-DCs), while IFN-α induces differentiation of monocytes into highly activated and partially mature DCs (IFN-α-DCs) retaining a marked phagocytic activity and exhibiting he ability to stimulate CD8(+) T cell responses against antigens from both pathogens and tumors. During this period, we found that DCs differentiated in the presence of IFN-α are fully competent for presenting apoptotic cell-associated antigens for cross-priming CD8(+) effector T cells and concomitantly for activating CD4(+) T helper cell. We also found that both IFN-α-DCs and IL-4-DCs exposed to a series of mycobacterial molecules, including Ag85B, PstS1, ESAT-6 and LAM, exhibit high levels of the activation and maturation markers, namely CD80, CD86, HLA-DR and HLA-A-B-C whereas only IFN-α-DCs are able to express high levels of the differentiation marker CD83. These findings suggest that the synergy between selected Mtb Ags and IFN-α promote DC maturation stimulating Ag-specific T cell response. The further analysis of peripheral DC subsets, including type-1 and type-2 myeloid dendritic cells (MDCs) and plasmacytoid DCs (pDCs), of both TB patients and LTBI individuals revealed that pDCs are significantly decreased in TB and AI patients with respect to healthy donors. Moreover, CD11c+/CD86+ myeloid DCs are higher in TB and AI patients than controls, whereas BDCA3+ type-2 myeloid DCs are lower in TB patients and healthy controls with respect to AI individuals. Therefore, we may assume that individuals with active or latent Mtb infection, characterized by a reduced pDC compartment in favour of myeloid DC populations, have a defect in a proper cross-presentation of mycobacterial Ags to CD8(+) T cells.
In a second set of experiments, we used expression microarrays to compare the molecular pathways activated in IFN-α-DCs and IL-4-DCs following infection either with Mtb or BCG. We found that these two agents stimulate different gene expression profiles in IFN-α-DCs and IL-4-DCs. BCG induces immune system processes genes in both type of DCs, However, in IFN-α DCs, it determines selectively a striking up-modulation of genes that regulate cytokine production during immune response, such as TLR2, IL-1, IL-6 and IL-23A. On the contrary, Mtb infection mainly induces genes related to metabolic pathways in both DC types even though a limited immune signature characterized by up-modulation of genes such as IL-6, IL-10 and IL23A occurs in IFN-α-DCs.
Of interest, IFN-α-DCs generated from PBMCs of actively or AI patients showed that a significant reduced expression of genes linked to immune response, in particular to Ag processing (LAMP-3) and Ag presentation (HLA family, CD80) as well as a downregulation of IFN-related gene IRF-4 occurs in these individuals with respect to healthy donors. The further phenotypic analysis of the same samples indicates that IFN-α-DCs derived from infected patients exhibited a reduced expression of the co-stimulatory molecules CD80 and CD86. Thus, these finding confirm that IFN-α-DCs derived from Mtb infected patients may not be able to appropriately process and present mycobacterial Ags to T cells leading to a defective immune response to the pathogen.
We also characterized microRNA pathways in IFN-α-DCs and IL-4-DCs following Mtb infection. The microRNA expression profile of IFN-α-DCs differs significantly from that of IL-4-DCs being the former characterized by the expression of microRNAs which regulate pivotal genes of the immune function and differentiation.
Infection of IRF8 KO mice with M. tuberculosis
Immune cell recruitment into lungs is pivotal in establishing protective immunity through granuloma formation and neogenesis of lymphoid structures (LS). Interferon regulatory factor-8 (IRF-8) plays an important role in host defense against M. tuberculosis , although its mechanisms driving anti-mycobacterial immunity remain unclear. IRF8 KO mice are characterized by their high susceptibility to many pathogens, including several bacterial species. IRF-8 KO and WT were infected via aerosol with Mtb Erdman strain. We found that up to two weeks following M. tuberculosis exposure, bacterial load in lungs and spleens in the two groups were similar. After this, IRF-8 KO mice developed a lethal acute TB, whereas in WT-B6 the disease is chronic stage. In IRF-8 KO lungs, pulmonary granulomas were larger and neogenesis of LS impaired. This associated with a progressive loss of infiltrating T lymphocytes and a massive prevalence of neutrophils. Moreover, our findings unravel an important role of IRF-8 in granuloma organization and the distinct pattern of cellular recruitment into the inflammatory lesion IRF-8 KO, with respect to controls, confirms the importance of the diverse recruitment of immune cells in acute versus chronic granulomatous lesions.
We characterized phenotypic and functional responses of DCs stimulated with mycobacterial molecules, including expansion of CD8(+) T cells and synthesis of IL-12, IL-23 and IFN-γ. The induction of CD4(+) Th-17 cells after immunization has been shown to contribute to protection and the generation of effector Ag-specific CD4(+) Th1, CD4 Th17 cells and CD8(+) T responses are strictly dependent on Ag presentation by DCs. The DC-mediated T cell responses to phosphate-binding transporter lipoprotein of the mycobacterial cell membrane PstS1, reported to be a strong immunomodulatory Ag of Mtb and an inducer of Th1 and Th17 responses. We found that PstS1, in contrast to other M. tuberculosis antigens such as Ag85B and HBHA, selectively activates DCs, both in vitro and in vivo, producing high levels of IL-6 and IL-23. We also investigated the ability of PstS1 to modulate CD4+ T cell responses during immunization with the Mtb candidate vaccine Ag85B and found that PstS1 induces Ag85B-specific memory CD4+ T cell proliferation and IFN-γ, IL-17 and IL-22 secretion. These effects were mediated through DC stimulation, particularly of the CD8α- subtype. To test the adjuvant activity of PstS1, we immunized mice with Pst1 and Ag85B. PstS1 protein administration during priming impaired the generation of Ag85B-specific immunity. Our data provide evidence that PstS1 may function as an immunomodulatory molecule stimulating DC, particularly the CD8α- subtype, affecting principally memory phenotype T cells and suggest the molecular rationale for exploiting its use in Mtb subunit-based vaccine strategies.
Role SOCS proteins in the outcome of infection with M. tuberculosis
SOCS1 inhibits IFN-γ secretion and thereby precludes efficient clearance of BCG and M. tuberculosis in macrophages
Protective immunity against infection with M. tuberculosis demands IFN-γ. Modulation of IFN-γ signaling is therefore of central importance in the outcome of infection with M. tuberculosis. SOCS1 The suppressor of cytokine signalling-1 (SOCS1) plays a major role in the inhibition of IFN-γ-mediated responses. We investigated the regulation and the role of SOCS1 during infection with BCG or M. tuberculosis in vitro and in vivo. We observed that infection with virulent or avirulent mycobacteria induces SOCS1 expression in vitro, in mouse and human macrophages and DCs, and in vivo in a murine model. Murine BMM, human monocyte-derived macrophages and lungs from infected mice contained higher titers of SOCS1 mRNA after infection with M. tuberculosis compared to BCG.
We then asked whether MyD88- and IRF3-mediated intracellular signalling pathways are involved in the augmented SOCS1mRNA levels after infection with BCG. Similar SOCS1 mRNA levels were found in BCG-infected IRF3-/- and WT BMM, while MyD88-/- BMM contained significantly lower SOCS1 mRNA levels than controls. SOCS1 mRNA levels were diminished in TLR2-/- BMM albeit to a lower degree than those in MyD88-/- BMM. TLR4 played a minor role if any in regulation of SOCS1 mRNA levels in BCG-infected macrophages.
Next, we investigated whether SOCS1 plays a role in intracellular mycobacterial growth control. SOCS1-/- and WT BMM were infected with BCG or M. tuberculosis. SOCS1-/- BMM showed reduced bacterial load compared to WT controls. SOCS1 gene was deleted in macrophages and neutrophils by crossing SOCS1fl/fl mice with LysM-cre SOCS1fl/fl mice. Infected LysM-cre SOCS1fl/fl BMM also showed reduced M. tuberculosis levels compared to SOCS1fl/fl controls. In agreement, increased bacterial growth was observed upon incubation of BCG or M. tuberculosis-infected BMM with Tkip, a peptide that, similar to SOCS1, binds and inhibits autophosphorylation of the JAK2 kinase and phosphorylation of the intracellular IFN-γ receptor.
Mycobacterial-infected SOCS1fl/fl LysMcre or SOCS1-/- BMM and BMDC showed increased levels of IFN-β, IFN-α, IFN-γ as well as IFN-regulated genes compared to controls. Since IFN-γ stimulates macrophage activation and controls mycobacterial growth, we studied whether the improved mycobacterial growth control in SOCS1-/- BMM is IFN-γ-dependent. For this purpose, bacterial titers in IFN-γ-/- /SOCS1-/- and IFN-γ -/- BMM were compared. IFN-γ-/- /SOCS1-/- and IFN-γ-/- BMM showed similar BCG loads. Moreover, similar bacterial levels were measured in WT, IFN-γ-/- and IFN-γ-/-/ SOCS1-/- M. tuberculosis-infected BMM, while the bacterial load in SOCS1-/- BMM was lower compared to all other groups. Altogether these results indicate that SOCS1 inhibits secretion of IFN-γ that otherwise would mediate a more effective clearance of mycobacteria in macrophages.
Since IL12 is known to be a main stimulus for IFN-γ secretion, we next investigated whether a M. tuberculosis-induced SOCS1 diminished expression of IL12 could account for the increased IFN-γ secretion in SOCS1-/- BMM. However, similar of IL12 were secreted by M. tuberculosis-infected SOCS1-/- and WT BMM, suggesting that the response to, but not the secretion of IL12 could account for the increased IFN-γ release in SOCS1-/- BMM. In fact, SOCS1-/- BMM stimulated with IL12 contained enhanced IFN-γ mRNA and protein levels compared to WT controls. Addition of anti-IL12 antibodies reduced IFN-γ mRNA levels in M. tuberculosis-infected SOCS1-/- and WT BMM, indicating that IL12 is required for the enhanced IFN-γ secretion by the infected SOCS1-/- BMM. The increased response to IL12 of SOCS1-/- BMM was associated with higher levels of the IL12 receptor (R) β1 in SOCS1-/- compared to WT BMM.
Whether SOCS1 further impedes the responses of M. tuberculosis-infected BMM to IFN-γ was then studied. Incubation of either SOCS1-/- or WT BMM with IFN-γ after infection with M. tuberculosis resulted in diminished bacterial load as compared to respective IFN-γ-untreated infected controls. In line with this, CXCL10 or iNOS mRNA levels were increased in IFN-γ-treated, infected WT or SOCS1-/- BMM cells as compared to uninfected controls or to infected cells in absence of IFN-γ-stimulation.
To investigate how SOCS1 shapes the macrophage response to M. tuberculosis at the transcriptome level, genome-wide expression analysis was performed in infected and uninfected SOCS1-/- and WR macrophages. A relevant segment of the macrophage transcriptome is remodelled after infection with M. tuberculosis. Unexpectedly, the majority of such genes overlapped with those differentially regulated due to SOCS1 deficiency, strongly suggesting a relevant role of SOCS1 in regulating the metabolic activity of macrophages during the infection.
Role of SOCS1 in the outcome of murine infections with M. tuberculosis.
Using different knock out and conditional knock down mice, we demonstrated that SOCS1 expression by macrophages impaired M. tuberculosis clearance and IFN-γ and iNOS expression in mice, when measured 7 days after infection. At this early stage of infection, when increased secretion of IFN-γ is not detected in lungs from infected WT mice, we propose that a higher level of IFN-γ secreted by SOCS1-deficient macrophages impair bacterial growth in vivo.
Instead, SOCS1 did not hamper bacterial control at later time points, when IFN-γ secretion by T cells is prominent. Thus may the explained by the capacity of infected, SOCS1-expressing macrophages to respond to IFN-γ.
We also observed an overwhelming infection-induced pulmonary inflammation in RAG1-/-/ SOCS1-/- and IFN-γ-/-/ SOCS1-/- mice. Since LysM-cre SOCS1fl/fl mice in contrast, showed no increased pathology, we conclude that SOCS1 expressing non-macrophage cells are main controllers of detrimental inflammation.
We propose that M. tuberculosis actively promotes SOCS1 expression by macrophages to counteract effective mycobacterial control at early time points after infection in vivo, before initiation of IFN-γ secretion by NK or T cells. SOCS1 facilitates mycobacterial growth by hampering macrophage IFN-γ secretion in response to infection-induced IL-12. At later time points after infection, despite SOCS1 expression, macrophages respond to IFN-γ secreted by T- or NK cells, and will not counteract bacterial control to any further extent. Instead, SOCS1 in non-macrophage cells protects mice from severe inflammation.
SOCS1 expression in TB patients
We investigated the association of SOCS with disease progression in patients with pulmonary TB. For this purpose we studied peripheral blood mononuclear cells (PBMCs) and T cells from patients with pulmonary TB (TB, n=33) and healthy endemic controls (EC, n=15). Cases were stratified into those with moderately advanced (mod-PTB) or far advanced disease (adv-PTB). IFN-γ, SOCS1 and SOCS3 gene expression was determined by RT-PCR. Statistical analysis was performed using the Mann-Whitney test. Levels of IL6 (p=0.018) and IL10 (p=0.013) were found to be elevated in PBMC supernatants from TB patients as compared with EC. SOCS1 mRNA gene expression in T cells from TB patients was increased as compared with that of EC (p=0.02). In addition, levels of SOCS1 mRNA transcripts were found to be elevated in PBMCs of adv-PTB as compared with mod-PTB (p=0.008) cases.
Our data show that raised SOCS1 levels are associated with increased disease severity in TB.
Chemokines- TLR and inflammation during infection with M. tuberculosis
Effects of heat-inactivated M. tuberculosis on the expression of chemokine receptors in macrophages and T cells
We initially investigated the effect of heat-killed M. tuberculosis (HKMTB) on the induction of Th17 cells. CD4+ T cells from BALB/c mice were polarized to Th17 cells in the presence of graded concentrations of HKMTB in vitro. HKMTB enhanced the polarisation of Th17 cells in a dose-dependent manner. Furthermore, HKMTB also induced an elevated level of RORγt (the key transcription factor of Th17).
Toll-like receptors (TLRs) represent the primary sentinels of microbial infection and have evolved to detect, and respond to, a variety of pathogen-associated molecular patterns. The TLR binding in leukocytes and in cells endogenous to the site of infection such as epithelial and endothelial cells, triggers the rapid synthesis and secretion of inflammatory cytokines and chemokines. The expression of chemokines after TLR agonist exposure is necessary for the efficient recruitment of inflammatory leukocytes to the tissue in response to localized microbial infection. Paradoxically, we found that TLR ligands administered systemically profoundly alter peripheral blood leukocyte trafficking to sites of inflammation. The TLR2 ligand bacterial lipoprotein (BLP) given intravenously as opposed to intraperitoneally significantly reduces the recruitment of leukocytes to a site of skin inflammation These results indicate that TLR ligands can modulate or suppress expression of chemokine receptor involved in cell trafficking. We have also investigated the effect of the various ligands of TLRs on the chemokine expression in murine macrophages. We used a novel approach to assess the expression and modulation of all known chemokine receptors in defined cell populations simultaneously by TagMan Low Density Array (TLDA, Applied Biosystems). B cells, M-CSF induced macrophages, Flt3L-induced dendritic cells and thioglycolate-elicited peritoneal macrophages were purified and then stimulated with a panel of TLRs. Our results show that CCR2 and CCR4 were down-regulated by nearly all the TLR ligands, whereas CCR7 was simultaneously up-regulated. These data highlight a novel role for TLR ligands, such as mycobacterial BLP, in regulating leukocyte retention and migration away from innate immune lesions via discrete constitutive and inflammatory chemokine receptor regulation.
The role of chemokine expression in granuloma formation
We induced granuloma formation in mice using Freund’s complete adjuvant, which contains killed M tuberculosis. Granuloma were formed 1-2 weeks following subcutaneous injection and the granuloma together with the draining lymph notes were harvested 4 weeks after injection. Histological slides of the tissues were made and single suspensions were prepared for FACS analysis. Our results show that CCR2 and CCR4 expression on CD4+ T cells and macrophages were significantly enhanced whereas the expression of CCR7 was decreased. We then tested the ability of these cells to migrate in vivo after intravenous injection. CD4+ T cells migrated rapidly to the lymph notes and were highly responsive to recall antigen, MTB extracts. We also attempted to block the migration of these cells with anti-TLR2 antibody. Partial suppression of the cell migration was observed but not the recall response of the cells. These results, which are being prepared for publication, show that granuloma formation is one of a potent mean by which the host heightened defence against the infection.
The role of IL-33 on sepsis
Sepsis is a systemic inflammatory condition following bacterial infection with high mortality rate and limited therapeutic options. We have now shown that IL-33, the new member of the IL-1 family, reduced mortality in mice with experimental sepsis from cecal ligation and puncture (CLP). IL-33-treated mice developed increased neutrophil influx in the peritoneal cavity and more efficient bacterial clearance than untreated mice. IL-33 reduced systemic but not local pro-inflammatory response, nor a Th1 to Th2 shift. The chemokine receptor CXCR2 is crucial for recruitment of neutrophils from circulation to the site of infection. Activation of Toll-like receptors (TLRs) in neutrophils down-regulates CXCR2 expression and impairs neutrophil migration. We also showed that IL-33 prevented the down-regulation of CXCR2 and chemotaxis induced by the activation of TLR4 and TLR2 in mouse and human neutrophils. Furthermore, we demonstrated that IL-33 reversed the LPS-induced reduction of CXCR2 in neutrophils via the inhibition of G-protein-coupled receptor kinase-2 (GRK2) expression, a serine/threonine protein kinase that induces internalization of chemokine receptors. Finally, we found that patients who did not recover from sepsis had significantly more soluble ST2 (sST2, the decoy receptor of IL-33) than those who did. Together, we provide a novel mechanism of action of IL-33 and suggest a therapeutic potential of IL-33 in sepsis. We have also investigated the role of IL-33 and TLRs in the susceptibility to opportunistic infection following recovery from sepsis. We found that in contrast to the initial sepsis, IL-33 exacerbates the opportunistic infections. This is because IL-33 promotes Type II response via the induction of IL-4 which in turn activates alternatively-activated macrophages that induce heightened levels of regulatory T cells. These novel finding is now being prepared for publication.
Genotyping of D6 chemokine receptor
The D6 gene (CCBP2), coding for a chemokine decoy receptor, is on chromosome 3 ranging from 42,825,967 to 42,883,779 bps (maps hg18/NCBI 36). There are 15 Quality control-filtered SNPs in D6 (+/- 10 kbs) from the full set of GWA SNPs (see table). The International Hapmap project reports on 48 SNPs in D6 among 30 European trios (CEU) (46 have a Minor allele frequency > 0.01). The 15 SNPs that have been genotyped allow to tag 36 of these 46 SNPs with a R2=0.8 (~ 78% coverage at R2=0.8). The association analysis was based on methods developed for family-based association studies, in particular the FBAT program. We considered additive, dominant and recessive models, and the minimum P-value was retained for each SNP (Table). The lowest p-value at p=0.009 was observed with rs7433284. Note that a simple Bonferoni correction for testing 15 SNPs (without accounting for the fact we have tested several genetic models for each SNP) provided a corrected p-value of 0.05/15 = 0.0033. Note also that over the whole set of 566,317 SNPs tested at the genome-wide level, 13,618 (~2.4%) provided a p-value lower than 0.009. In conclusion, there is no significant association of any of the 15 CCBP2 SNPs and pulmonary TB in our Moroccan sample.
HUMANIZED MOUSE MODEL OF MYCOBACTERIAL INFECTION
Non-obese diabetic (NOD)/SCID/ γcR-/- mice and BALB/c RAG2-/- /γcR-/- can be stably engrafted with human (CD34+ hematopoietic stem cells (HSC) following sub-lethal irradiation due to a lack of an adaptive immune system and innate immunity defects. In these models, human CD34+ HSCs self-renew in the bone marrow and provide long-term repopulation of mice with multi-lineage human myeloid and lymphoid cell populations. These humanized mice have been shown to be useful tools to examine different microorganisms that involve human hematopoietic cells in their life cycles.
Granuloma formation in M. tuberculosis-infected individuals is considered to constitute an important component of the protection against disease.
The widely used mouse model of tuberculosis fails to reproduce the specific features of human granulomas, which are a hallmark of mycobacterial disease. Whether humanized mice can provide a better tool to study human granuloma formation and to experimentally address the role of human immune responses during Tuberculosis was explored.
Specifically, we have studied the outcome of infection with the attenuated Mycobacterium bovis BCG in NOD/ SCID/ γcR-/- mice, reconstituted de novo with human lymphoid and myeloid cells after neonatal transplantation with human CD34+ hematopoietic stem cells.
Spleens of BCG-infected humanized mice showed increased human CD4/ CD8 T cell ratio and a widespread differentiation of T cells from naïve into CD45RA-CCR7- effector memory cells.
BCG-infected humanized mice showed increased numbers of human T cells and macrophages in lungs and livers but paradoxically contained higher number of mycobacteria in tissues as compared to non-humanized controls. This enhancement of bacterial load was abrogated after in vivo depletion of human CD4+ but not CD8+ cells, and was associated with an increase in T cell expression of CD57 and PD-1, molecules associated with senescence and inhibition of T-cell function.
Despite the lack of bacterial control, organized granulomas resembling human lesions were observed in livers and lungs from BCG-infected humanized mice but not in non-humanized infected controls. Granulomas showed a core with compact aggregates of human CD68+ macrophages and giant multinucleated cells and elevated load of bacilli, surrounded by a layer of CD3+ T cells and a fibrous capsule encasing the lesions. Few to moderate numbers apoptotic and necrotic cells were present in the core of the larger granulomas. Higher levels of human IFN-γ, and the IFN-γ-regulated, T cell chemotactic CXCL9 and CXCL10 transcripts were detected in lungs and livers from BCG-infected humanized mice compared to non-infected controls.
The formation of granulomas depended on human CD4+ but not CD8+ cells since granulomas from mice depleted of CD4+ cells were fewer, minimal and irregular, while those from mice depleted in CD8+ cells retained the size and complexity of non-treated controls.
M. tuberculosis (but not BCG)-infected non-humanized mice showed granulomatous lesions in livers and lungs. However, typical granulomas were only found in M. tuberculosis-infected humanized mice. Granulomas from M. tuberculosis-infected humanized mice displayed a strikingly increased extension of necrosis compared to those from BCG-infected humanized mice or non-humanized M. tuberculosis-infected mice. In relation with this, it has been shown that M. tuberculosis infection stimulates necrosis of infected cells while BCG does not [5]. Increased severity and larger lesions were observed in livers of humanized mice as compared to non-humanized controls. Thus, the human transplant determined the formation of granulomas in BCG-infected mice and the granuloma features in mice infected with M. tuberculosis.
In summary, humanized mice showed a defective control of BCG infection, probably due to dysfunctional T cell responses or an impaired ability of macrophages to mount microbicidal responses. Notwithstanding the impaired bacterial control, granulomas that look like those observed in human M. tuberculosis were observed. Human CD4+ but not CD8+ cells were required for granuloma formation in mycobacterial infected humanized mice and for the enhanced growth of mycobacteria in these mice.
Our data support the hypotheses that HIV-induced depletion of CD4+ T cells lead to a direct disruption of the granulomas during M. tuberculosis infection. Alike, CD4+ T cells have been shown to contribute to the organization of granulomas in MHCII-/- or CD4-/- M. tuberculosis-infected mice. However, and different to humanized mice, susceptibility to Tuberculosis increases with decreasing peripheral CD4 T cell counts in HIV infected individuals.
Granuloma initiation during mycobacterial infections has been interpreted as a host-protective event, to provide the microenvironment where specific T cells activate macrophages to contain M. tuberculosis infection. On the contrary, our results suggest that non-protective, dysfunctional T cell responses participate in granuloma formation in humanized mice. Whether such a conclusion can be generalized for the natural course in of infection requires further studies. Supportive of our interpretation, data in a zebra fish model indicated that mycobacteria can exploit the granuloma for local expansion and systemic dissemination.
To our knowledge, this is the first study of mycobacterial infections and description of granuloma formation in humanized mice.
Our data indicate that humanized mouse model can be used to study the role of cytokines, chemokines, different immune cells, host genetics and bacterial components in mycobacterial granuloma formation. We also propose the use of humanized mice as a model for other infectious and non-infectious granulomatous diseases.
Production of recombinant mycobacterial antigens
Strains for the production of the recombinant mycobacterial antigens Acr, HBHA, ESAT6, PstS1, Ag85A, HSP70 and its truncated derivatives are available. Purification and quality control procedures for the antigens Acr, Ag85A, ESAT6, PstS1, HSP70, and its truncated derivatives are established. The purification of recombinant HBHA, PPD, ESAT-6, 16kDa, PstS1, Ag85A and HSP70 was completed during this period. Several of these antigens were distributed to different partners of the project. Natural mycobacterial HBHA was prepared and distributed to different partners. For preparation of HBHA from mycobacterial source, expression vectors for the production of recombinant N- and C-terminal His-tagged M. tuberculosis HBHA in M. smegmatis were constructed, protein was expression and purified, and showed to contain methylated groups.
Ethical Issues
Thus ethical aspects of both human and animal research were considered as indicated in the Technical annex of the project. Human and animal research from this proposal conformed to accepted scientific standards. The research was conducted in accordance with all relevant local or national rules and regulations of the country where the research was carried and subject, as appropriate, to prior authorization of the project by the competent national and institutional research ethics or medico-ethical committees.
Our project involved the use of human tissues such as blood from adults and children. These will be obtained under standard procedures of informed consent from the individuals or legal guardian(s). The procedures are minimally invasive and do not require hospitalisation.
The studies proposed have been approved by the relevant authorities and will be conducted according to the local, national (following the legislation in place in each country) and EC directives. All institutional, national and for all European laboratories EU regulations and guidelines will be followed. The study were done only by trained clinical staff and poses no risk to the health or well being of the participants. All potential participants had first the purpose of the study explained to them and were asked to give written and oral assent to participation. Samples were only be taken after they have indicated that they understand the purpose of the study. In particular, the study will be conducted in accordance with the World Medical Association Declaration of Helsinki (as updated in October 2000) and Directive 2001/20/EC of the European Parliament and of the Council of 4 April 2001 on "The approximation of the laws, regulations and administrative provisions of the Member States relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use.
Regarding animal research performed at HOMITB, it conformed current legislation and regulations in the countries where the research was carried out. The research was be conducted in accordance with all relevant local or national rules and regulations of the country where the research will be carried and subject, as appropriate, to prior authorization of the project by the competent national and institutional research ethics or medico-ethical committees.
Experimental animals were used in several of the participants’ research facilities, where such experiments were subject to the review of protocols and permission of each insti¬tute’s committee for the use of experimental animals in research. In any case, proper care and humane treatment was always provided.
Potential Impact:
4. The potential impact
Scientific impact
Tightly controlled genetically determined interactions between the mycobacterial stress responses and the host innate and adaptive immune responses to infection dictate the outcome of mycobacterial infection in man ranging from an asymptomatic disease to a disseminated infection. The outcome of HOMITB has been an in depth assessment of molecular and cellular interactions crucial for the design of novel vaccination or immunotherapeutic strategies, aimed at enhancing the efficacy of the innate or adaptive immunity against infection and disease and to the design of better diagnostic markers.
We have studied
1) Host target cell responses regulating the growth of mycobacterial infection
2) Mycobacterial adaptation to different effector host cell responses to infection
3) The role of innate and adaptive immune responses in the control of mycobacterial growth and survival and the regulation of inflammation.
4) Novel human genes responsible for the susceptibility to children and adult TB.
Our main results are
1. An homozygous missense mutation at amino-acid position 417 of JAK2 was found in by genome-wide (GW) homozygosity mapping, a method to find rare recessive mendelian mutations in a child with severe Tuberculosis when screening Chromosome 9 of a sample of 21 families. Preliminary data show that STAT4 activation is diminished in IL-12 activated patient cells.
2. In a sample of 50 children from Iran and Morocco, 2 children showed loss-of-function IL12RB1 alleles resulting in complete IL-12Rβ1 deficiency. This finding may have important medical implications, as recombinant IFN-γ is an effective treatment for mycobacterial infections in IL-12Rβ1-deficient patients. Together with the finding of JAK2 mutants in this relatively small samples, this also provides additional support for the view that severe tuberculosis in childhood may result from a collection of single-gene inborn errors of immunity.
3. We previously reported the first region linked with pulmonary TB, on chromosome 8. Our identification of susceptibility variants in the proximity of TOX represents a major breakthrough in the human genetics of Pulmonary TB, in that this finding has been replicated in Morocco and validated in Madagascar. Moreover, TOX encodes Thymocyte selection-associated high mobility group box protein, and is a key transcription factor leading to CD4 cell development in the murine model.
4. Our data showed that AECII express the molecules required for T cell activation, including HLA class II molecules, which expression is usually restricted to professional antigen presenting cells; secrete cytokines and chemokines critical for the recruitment of leukocytes to the lung, and are able to present antigens, either mycobacterial or viral antigens, to autologous T lymphocytes. We confirmed therefore their role of non-professional antigen-presenting cells. We confirmed that Mtb invades and multiplies within AECII and these accessory presenting cells might participate in the chronic stimulation of anti-HBHA T cells which is observed during Mtb latent infection. AECII might therefore participate to the control of Mtb replication in latently infected subjects.
5. HbhR is a repressor in some mycobacterial species, but not in others, in which it may be even a transcriptional activator. Future work will address the full regulon of HbhR in various mycobacterial species, especially in those in which HbhR is a repressor. Within the frame of HomiTB, we found that in BCG and M. tuberculosis, HhhR appears to regulate only a few additional genes. In the future, the products of these genes will be characterised and their antigenicity will be tested in M. tuberculosis-infected mice and humans. We hypothesize that they may also be latency antigens. The work in HomiTB has led to the finding that hbhA regulation is complex and involves several activators/repressors. Future work will establish the importance of one over the others. Finally, during the HomiTB project, a BCG strain was fortuitously isolated that contained a large deletion in the chromosome, containing genes that are involved in lipid metabolism and genetic regulation. This will also be further investigated.
6. IRF8 is critical for development of dendritic cells. We here show that this transcription factor is required for control of M. tuberculosis infection in mice. IRF8-/- showed higher bacterial levels in lungs, enhanced severity of pathology and diminished recruitment of T cells to pulmonary lesions.
7. B cells, M-CSF induced macrophages, Flt3L-induced dendritic cells and thioglycolate-elicited peritoneal macrophages were purified and then stimulated with a panel of TLRs. Our iresults show that CCR2 and CCR4 were down-regulated by nearly all the TLR ligands, whereas CCR7 was simultaneously up-regulated. The secretion of chemokines was not down-regulated by the TLR agonist. These data highlight a novel role for TLR ligands, such as BLP, in regulating leukocyte retention and migration away from innate immune lesions via discrete constitutive and inflammatory chemokine receptor regulation.
8. We show that the novel cytokine, IL-33, play an important role in bacterial infection. Although its roles in type 2 immunity are well-documented, our data reveals that innate immune cells respond strongly to this cytokine and helped clarify important roles for IL-33. Mycobacterial products mediate their responses via TLRs. Understanding the mechanisms by which cytokines such as IL-33 modulate the functional outcome of TLR signaling through neutrophil activation and migration could lead to novel approaches to the control of systemic as well as local bacterial clearance. This work therefore opens up a new direction of investigation to the mechanism of host response to MTB and may lead to novel approaches to vaccine designs and immune therapy against MTB.
9. The role of SOCS1 during infection with M. tuberculosis was analysed. We propose a model in which M. tuberculosis actively promotes SOCS1 expression by macrophages to counteract effective mycobacterial control at early time points after infection in vivo, before initiation of IFN-γ secretion by NK or T cells. SOCS1 facilitates mycobacterial growth by hampering macrophage IFN-γ secretion in response to infection-induced IL-12. At later time points after infection, despite SOCS1 expression, macrophages respond to IFN-γ secreted by T- o NK cells with unaffected potency, and will not counteract bacterial control to any further extent. Instead SOCS1 in non-macrophage cells protects mice from severe inflammation. The mechanisms behind the anti-inflammatory properties of SOCS1 during M. tuberculosis remain to be studied. We suggest that the efficient stimulation of SOCS1 expression by M. tuberculosis reflects an evolutionary adaptation that may be associated to the ability of M. tuberculosis to establish a chronic infection. We also demonstrated that levels of SOCS1 in blood from patient are associated to disease severity.
10. We have studied the outcome of infection with mycobacteria in humanized mice, in which both human immune cells differentiate de novo from transplanted cord blood progenitor cells. M. bovis BCG-infected humanized mice showed organized granulomas containing a high proportion of human leukocytes, with a core containing human macrophages and giant cells, surrounded by a layer of T cells and a fibrous capsule. Granulomas from M. tuberculosis-infected humanized mice contained widespread necrosis. The formation of these structures depended on human CD4+ but not CD8+ cells, since mice depleted of CD4+ cells after administration of specific monoclonal antibodies showed scarcer, minimal and irregular granulomas. Granuloma formation in humanized mice was segregated from bacterial protection since CD4+ cell activity resulted in a higher susceptibility of humanized mice to mycobacterial infection.
Granuloma initiation during mycobacterial infections has been interpreted as a host-protective event, to provide the microenvironment where specific T cells activate macrophages to contain M. tuberculosis infection. On the contrary, our results suggest that dysfunctional CD4+ T cells participate in granuloma formation in humanized mice and may facilitate bacterial growth. Whether such a conclusion can be generalized for the natural course of infection requires further studies.
Thus, the humanized mouse model can be used to study the role of cytokines, chemokines, different immune cells, host genetics and bacterial components in mycobacterial granuloma formation. We also propose the use of humanized mice as a model for other infectious and non-infectious granulomatous diseases.
Strategic impact
Tuberculosis is one of the leading causes of death in the world. M. tuberculosis was the cause of the “White Plague” of the 17th and 18th centuries in Europe. During this period nearly 100% of the European population was infected with M. tuberculosis, and 25% of all adult deaths were caused by M. tuberculosis. Today, M. tuberculosis is responsible for more morbidity in humans than any other bacterial disease. 1.7 billion people are infected with M. tuberculosis, which amounts to approximately 1/3 of the entire world population. TB is responsible for over 2 million deaths/year. TB has declined in countries of Western and Central Europe, but the burden is still high and increasing in Eastern Europe. This rise is largely attributed to the emergence of HIV infections, which has occurred during the same period. WHO estimated that there were 484,000 new TB cases in Europe in 2001, representing 6% of the global TB burden. Furthermore, patient non-compliance to anti-mycobacterial treatment has given rise to multi-drug resistant strains of M. tuberculosis.
Live BCG is the only currently available vaccine against TB and has been inoculated in about 85% of newborns world-wide. It protects best against miliary TB and TB meningitis in children less than one year of age. However, BCG has failed to confer significant protection in older children and adults against pulmonary TB, the most common and contagious form of TB at that age. Hence, prevention of M. tuberculosis is likely to result in a great saving of societal costs in addition to improving the quality of life and life expectancy of a large part of the population. HOMITB aimed at improving public health within Europe by evaluating novel strategies for the control of TB. The outcome will contribute to the improvement of the design of specific vaccines, to the prevention of severe TB and to a better understanding of the relationships between the invading bacteria and the host immune system.
Medical impact
Uncovering genetic and environmental factors involved either in latency or in disease constitutes a major break-through in the understanding of the host-pathogen interaction. Studies within HOMITB further defined the molecular basis of immunity to M. tuberculosis by dissecting the genetic components that are altered in patients with TB and identifying novel genes that predispose or account for Tuberculosis in man. This will allow developing novel lines of research aimed at restoring impaired immunity. Immunological intervention, whether preventive or curative, will rely on a rational understanding of the pathogenesis of TB in the two important age groups. The design of novel vaccines for paediatric TB, which might be more efficient than BCG, will benefit from such studies as the defective pathways will require to be either complemented or circumvented. The design of a vaccine that might protect adults from re-activation of the mycobacterium is a formidable challenge that will also benefit from exploring the process of immunity to latent mycobacteria by genetic means. Immunological pathways that normally restrict the growth and reactivation of latent mycobacteria are currently unknown and their genetic identification will pave the way for the development of novel vaccines or other immunological interventions. Studies of mechanisms of regulation of mycobacterial gene (hbhA) within infected epithelial cells and macrophages identifying other members of the same regulon may ultimately lead to the discovery of other latency-associated antigens that may be effective immuno-protective and/or diagnostic antigens, and of understanding the mycobacterial adaptation mechanisms during a chronic infection. These studies complementary to the understanding of the host response to gene-targeted studies have focused on the role of molecules that regulate both the inflammatory response and/ or the growth of M. tuberculosis. These molecules affect the activation of the innate and adaptive immune responses, the effector mechanisms in different target cell populations and the stability of the inflammatory lesions. Our results demonstrate that targeting these genes will be important in the control of disease and mycobacterial persistence.
Technological breakthrough
Genomewide scanning for human mutations has been performed in HOMITB resulting in the identification of novel genes that account for the susceptibility of man to Tuberculosis. Genomic technology that generates data on differential expression of both mycobacterial and host genes will be used in analysing the role of different murine or human host target cells, and the role of mycobacterial molecules or immune genes in defining the outcome of infection by using mutant bacteria or transgenic mice respectively. New recombinant molecules will also be used. The humanized mouse model, based on reconstitution of immunodeficient mice with human lymphoid and myeloid cellular populations, will be used to overcome ethical constraint in the research of human immune system. We will monitor mycobacterial load dynamics using a non- invasive real time imaging method in vivo.
Dissemination activities
Dissemination has been pursued achieved through websites, publications, presentations, interviews etc. Work carried out within this project has also been disseminated by publishing in international research journals, for the benefit of the scientific community. A list of publications and presentation in different scientific meetings is imcluded in the next section. We expect a number of manuscripts to be published in the future that are in process of preparation or have already been submitted. A list of these still unpublished manuscripts is also added. Several of these published or unpublished manuscripts are co-authored by two or more participants of HOMITB.
List of Websites:
http://www.mtc.ki.se/HOMITB/index.html
The global incidence of tuberculosis (TB) is rising, with 8.8 million new cases and 2 million deaths each year. The rate of progression from infection to disease is highly variable, and approximately 90% of infected individuals never develop clinical disease. About half of the 10% of Mycobacterium tuberculosis-infected individuals who do develop overt clinical disease are diagnosed with TB within 2 years of infection and can be considered fast progressors. This 'primary' TB is particularly common in children and is often disseminated due to early haematogenous spread of the mycobacterium with miliary and/or extra-pulmonary disease.
About 50% of TB patients who develop clinical disease more than two years after infection are commonly described as cases of 'reactivation' or post-primary TB. Post-primary TB is predominantly a pulmonary disease of adults. Although neonatal immunisation with the Mycobacterium bovis-derived Bacillus Calmette-Guérin (BCG) vaccine can provide protection against childhood TB, BCG is relatively ineffective in preventing disease in adults.
Tightly controlled genetically determined interactions between the mycobacteria and different host target cells that are regulated by the host innate and adaptive immune responses dictate the outcome of mycobacterial infection in humans, ranging from an asymptomatic infection to a life-threatening disease.
At HOMITB we identified novel host genes involved in control of bacterial load, and regulation of inflammation and immune responses, and bacterial molecules involved in escape from such immune regulatory and effector mechanisms. Both targeted and genome-wide explorations were employed to identify susceptibility of genes involved either in the immunological control or pathogenesis of primary M. tuberculosis infection in children and secondary pulmonary TB in adults, as well as in the control of mycobacterial infection in the mouse model. Immunological validation of the novel host susceptibility genes were performed.
Mycobacterium host interactions in macrophages, dendritic cells and non-haematopoietic cells such as epithelial cells are studied both in human and mouse models. The role of chemokines, chemokine receptors and key intracellular signalling molecules in the outcome of mycobacterial infection is investigated. Mice with human lymphoid and myeloid lineages de novo developed after transplantation of haematopoietic stem cells are used to explore the human immune responses to infection with Mycobacterium.
HOMITB dissected the molecular and cellular immune responses, combining mice models, and human patients. In both approaches, the responses of various host target cells regulating mycobacterial growth and the mycobacterial adaptation to different host cell responses to infection were studied. The role of innate and adaptive immune responses in the control of mycobacterial growth and the regulation of inflammation was studied as well. In addition, the project investigates such responses using a third model based on mice reconstituted de novo with human haematopoietic-derived cell lineages. Finally, targeted and genome-wide explorations were employed to identify novel susceptibility genes involved in the immune control of primary M. tuberculosis infection and secondary pulmonary TB.
The advanced knowledge of molecular and cellular interactions between host and mycobacteria is crucial for the design of novel vaccination or immunotherapeutic strategies and for the discovery of better diagnostic markers.
Project Context and Objectives:
Tuberculosis (TB), an infectious disease caused by Mycobacterium tuberculosis, remains a leading public health problem worldwide. The global incidence of TB is rising, with 8.8 million new cases and 2 million deaths each year. The rate of progression from infection to disease is highly variable, and approximately 90% of infected individuals never develop clinical disease. About half of the 10% of M. tuberculosis-infected individuals who do develop overt clinical disease are diagnosed with TB within two years of infection and can be considered fast progressors. This “primary” TB is particularly common in children, and is often disseminated due to early, haematogenous spread of the mycobacterium with miliary and/or extra-pulmonary disease. About 50% of TB patients, who develop clinical disease more than two years after infection, are commonly described as cases of “reactivation” or post-primary TB. Post-primary TB is predominantly a pulmonary disease of adults, involving extensive damage to the lungs and efficient airborne transmission of bacteria. Although neonatal immunization with the current vaccine, Mycobacterium bovis-derived bacillus Calmette-Guérin (BCG), can provide protection against childhood manifestations of the disease (primary disease), BCG is relatively ineffective in preventing disease in older children and adults (reactivation disease).
Consequently, latently infected individuals run a life-long risk of reactivation to active disease. The risk is particularly important when these individuals are immuno-compromised, either as the result of HIV infection or of immunosuppressive treatment. As TB reactivation results in the most contagious, pulmonary form of the disease, preventing latently infected individuals from reactivating should constitute a major step towards the elimination of TB. Of importance, it is still not fully understood why some individuals develop active disease, whereas most apparently control the infection by an appropriate immune response.
Therefore, tightly controlled genetically determined interactions between the mycobacteria and different host target cells that are regulated by the host innate and adaptive immune responses dictate the outcome of mycobacterial infection in man ranging from an asymptomatic infection to a life-threatening disease. Thus, knowledge of such molecular and cellular interactions is crucial for the design of novel vaccination or immunotherapeutic strategies, and to the discovery of better diagnostic markers.
In most cases, the host immune system is able to control bacterial replication and to prevent the development of active disease. However, M. tuberculosis is able to reprogram its gene expression, preventing the immune response from totally eliminating the microorganism. Thus, the host becomes latently infected. The identity of these genes and how they control M. tuberculosis adaptation to the intracellular environment remain to be delineated. Infected alveolar macrophages and dendritic cells have long been known as a primary site of M. tuberculosis infection. More recently, other cell types, such as adipocytes, fibroblasts, endothelial cells, and most interestingly, epithelial cells have also been identified as cellular niches for latent M. tuberculosis. Upon infection, these cells can release a panel of antimicrobial effector molecules, cytokines, and chemokines, which govern innate immune responses and initiate specific immunity. This complex network of mediators induces activation of antimicrobial activity in macrophages and perhaps other cell types, migration of antigen laden phagocytes to the draining lymph nodes, and finally the influx of antigen-specific T lymphocytes to the site of infection. Primary infection most often leads to the formation of a well organised granuloma consisting of central macrophages with surrounding lymphocytes and fibroblasts. This cellular microenvironment provides a physical and immunological barrier for the entrapped bacilli thereby preventing multiplication and spread of the mycobacteria. Live mycobacteria may also remain in many different cell types throughout the body and undergo a dormancy phase. Factors disrupting the fine-tuned balance between the low level mycobacterial metabolism and the cellular immune response of the host are likely to increase the risk of tipping the balance in favour of the pathogen and result in reactivation of disease.
AIM
In this project we proposed to identify novel host genes involved in control of bacterial load, and regulation of inflammation and immune responses, and bacterial molecules involved in escape from such immune regulatory and effector mechanisms. Both targeted and genome-wide explorations were employed to identify susceptibility genes involved either in the immunological control or pathogenesis of primary M. tuberculosis infection in children and secondary pulmonary TB in adults as well as in the control of mycobacterial infection in the mouse model. Immunological validation of the novel host susceptibility genes were performed. In parallel, the M. tuberculosis strains collected in field studies will be typed by the most recent and discriminatory molecular typing method, based on Mycobacterial Interspersed Repetitive Units (MIRUs). Mycobacterium host interactions in macrophages, dendritic cells and non-haematopoietic cells such as epithelial cells will also be studied both in human and mouse models. The role of chemokines, chemokine receptors and key intracellular signalling molecules in response to mycobacterial infection was analyzed. Mice with human lymphoid and myeloid lineages de novo developed after transplantation of haematopoietic stem cells were used to explore the human immune responses to infection with Mycobacterium.
SPECIFICALLY
We searched for novel genes involved in the susceptibility of children to disseminated TB. By using a candidate gene approach, we investigated whether known MSMD (Mendelian Severe Mycobacterial Diseases) causing genes and new immune genes involved in the IFN-γ production and signalling pathway are important for control of childhood disseminated TB. Furthermore, new genes involved in control of adult pulmonary TB were identified by positional cloning, and functionally validated. The identification of single nucleotide polymorphisms (SNP) conferring a major effect in predisposition to adult TB was based on large association familial studies. SNPs in an interval in chromosome 8 known to predispose to pulmonary TB, and in MSMD causing genes will be genotyped. Functional characterization of alleles predisposing to TB was performed by biochemical and immunological methods.
Mycobacterial adaptation into different target cells was investigated. The mycobacterial heparin-bindin hemagglutinin (HBHA) is a protective antigen, especially well recognized by latently infected subjects. The expression of hbhA is controlled by the intracellular environment.The hbhA gene is strongly up regulated in epithelial cells but not in macrophages. The mechanisms of regulation of the hbhA gene within infected epithelial cells and other cell types was investigated. Other members of the same regulon were identified by comparing the mycobacterial gene transcriptome after infection of different host cells. The role of such genes in persistence in non-haematopoietic cells and macrophages was analyzed.
Macrophages have long been recognized as key players in the pathogenesis of TB and as sites for M. tuberculosis infection. We defined the relative role of non-haematopoietic cells, with emphasis on alveolar epithelial cells in the outcome of M. tuberculosis infection in humans, murine and humanized murine models. The host cell transcriptome of IFN-γ-stimulated macrophages and of epithelial cells infected with Mycobacterium was compared. The antigen-presenting capacity of epithelial cells infected with mycobacterium was tested. The mycobacterial transcriptome during infection of macrophages and epithelial cells stimulated with IFN-γ was analyzed. Differential expression of mycobacterial genes were related to the adaptation to diverse effector mechanisms in these cell populations.
Dendritic cells (DC) represent key components of the immune system mediating initiation of adaptive immune responses that restrict M. tuberculosis infection. However, during infection, the specific functions of DCs are modulated by mycobacterial molecules thereby affecting the ensuing adaptive responses and the outcome of the infection. We proposed to perform detailed analysis of the relative role of distinct DC populations in driving immune responses to M. tuberculosis. We dissected the molecular pathways in DC activated by M. tuberculosis infection or mycobacterial molecules and the mechanisms by which mycobacteria interfere or promote DC functionality.
Chemokines play a key role in cell migration and are therefore crucial to innate and adaptive immunity to infection and inflammation. The initial innate immune response to M. tuberculosis is triggered by Toll-like receptors (TLR) on host cells, responding to M. tuberculosis-derived pathogen-associated molecular patterns (PAMP). The interaction of PAMP with TLR triggers inflammatory cytokine release culminating in the production of inflammatory chemokines which control the migration of the leukocytes into the infected area. A mycobacterial lipoprotein (agonist of TLR2) selectively suppresses inflammatory chemokine receptor expression and down regulates inflammatory cell responses to chemokines. Interference with this TLR2/chemokine receptor axis may therefore be instrumental in the re-emergence of latent M. tuberculosis infection. We will investigate the relationship between TLR2 signalling, chemokine receptor expression and granuloma formation. In addition, the role of a chemokine decoy receptor D6 involved in the suppression/resolution of pulmonary inflammation and thus representing a good candidate for involvement in the differential pathogenesis of TB of humans and mice will be studied. For this purpose, genetic alterations in the D6 chemokine decoy receptor that are associated with pulmonary TB will be determined and tested functionally. In parallel the role of D6 in the outcome of mycobacterial infection will be tested in mouse model of tissue-specific D6 transgenic mice.
Immune and inflammatory systems are controlled by multiple soluble mediators. SOCS are a family of intracellular proteins, several of which have been shown to be regulators of cytokine homeostasis. We studied the regulation of expression and role of SOCS1 and 3 molecules during the mouse mycobacterial infection in vivo and in macrophages. The levels of SOCS proteins in cells from patients or asymptomatic infected individuals were studied.
We will use a model of BCG and M. tuberculosis infection in RAG-1-/-/γcR-/- mice reconstituted with human lymphoid and myeloid cells. Such “humanized” mice, carrying a functional human immune system, will be of considerable importance to experimentally decipher the cellular and molecular mechanisms of anti-mycobacterial immunity in humans. This model will permit the first experimental investigation of human immunity to mycobacteria.
HOMITB represents a combined effort of 9 leading groups in genetics, molecular microbiology, cellular biology, immunology and clinical practice in TB The grouping, which also includes a highly successful SME. In this proposal, we will deploy our recent findings in basic and applied research to tackle TB, using the most advanced and innovative approaches. We were the lead member of international collaborations in defining human immune response genes directly related to susceptibility to TB. This advanced concept and technology will be used to further pin-point the host genetic markers for TB. Much past effort has been focused on the role of macrophages in TB. Since non-haematopoietic cells are also infected by M. tuberculosis, we propose a novel approach to investigate the role of pulmonary epithelial cells, endothelial cells and fibroblasts in TB based on our recent research in the function of these cells. Recent understanding of the physiology and functions of DC has been a major advance in immunology. We will apply this new knowledge to TB research. The role of chemokines in TB is virtually unknown. Our strength in chemokine biology and biochemistry and the novel finding of the crucial role of bacterial products (TLR2-ligands) in modulating chemokine receptor expression led us to a highly innovative approach to explore the role of chemokines/chemokine receptors in regulating innate and adaptive immunity to M. tuberculosis. We have recently found that the SOCS family members play an important role in infection. We will apply this novel approach to explore the role of this key transcription factors in M. tuberculosis infection. We have recently constructed humanized mouse carrying a functional human immune system, in which the responses to experimental infection with BCG and M. tuberculosis will be studied. A real time method will be used to evaluate the presence of mycobacteria. This state-of-the-art technology will be employed to test the findings obtained above in a model system as close to the human disease as possible.
This closely interactive grouping with extensive complementary expertise (basic immunology, cell biology, human and mycobacterial genetics, clinical field practice, industrial facilities and knowledge) and strong track record in collaborative research created a powerful research centre to deliver highly innovative approach to tackle one of the most important diseases in the world.
Project Results:
HUMAN GENETIC BASIS OF TUBERCULOSIS
The Mendelian basis of paediatric tuberculosis
Candidate gene approach
Patients included in this part of the project are children (≤15 years) with a disseminated/extra-pulmonary form of TB. These forms include miliary TB, meningitis, peritonitis, and osteomyelitis (in particular Pott’s disease). As these forms are often bacteriologically negative, the inclusion criteria are based on a stringent combination of clinical, radiological, biological, and histological features. Two kinds of samples have been collected and used for the results presented here:
- a first sample of 32 Moroccan children with disseminated TB and available DNA that were used for hypothesis-based systematic sequencing of candidate genes and immunological studies. In a second step, we have collected an additional sample of 30 Moroccan children.
- a sample of 21 consanguineous families with at least one child with disseminated TB that were used for the hypothesis-generating GWL study.
The first goal was to sequence the genes involved in the syndrome of Mendelian susceptibility to mycobacterial diseases (MSMD), in particular IFNGR1, IFNGR2, STAT1, IL12B, IL12RB1, in our sample of 32 Moroccan children. The two receptors of IL12, IL12RB1 and IL12RB2 were fully sequenced with a coverage of 100%. No unreported mutation was found in IL12RB2. In IL12RB1, we identified one patient carrying a homozygous nonsense mutation in exon 9 of IL12RB1 leading to a premature stop codon in amino acid position 305 (K305X). The two parents are heterozygous for the mutation, and we showed that this mutation was complete loss of function. The observation of at least one IL-12RB1-deficient child among 32 patients may at first glance seem fortuitous. However, we recently replicated this observation in similar surveys in Turkey and Iran, implying that recessive mutations in IL12RB1 loci may universally be found in a substantial fraction of children with severe TB (about 1-3%). These results strongly suggest that pediatric TB may actually result from a collection of single-gene defects. Up to 3% of severe pediatric TB may be accounted for by mutations solely in the IL12RB1 gene. This work was published in 2011. No interesting candidate mutations were found in the other genes sequenced (IFNGR1, IFNGR2, STAT1, IL12, NEMO) as well as in ISG15 that we recently identified as a new MSMD gene. We are now pursuing this investigation directly by exome sequencing in order to identify new genes and maybe new pathways involved in Mendelian predisposition to TB.
Genome wide linkage scan
In order to determine which new genes are involved in susceptibility to disseminated TB in children, we have already recruited 21 consanguineous families, extracted gDNA from each member and conducted a genome-wide linkage scan using an Affymetrix 250K array. The results were analysed by the same technique that was used for 17 other consanguineous families with MSMD, a technique that has proved to be efficient. Indeed, genome-wide (GW) homozygosity mapping (HM) is a powerful method to locate rare recessive Mendelian mutations. However, statistical power can decrease dramatically in the presence of genetic heterogeneity, which is likely to increase with sample size. We thus developed a novel approach to test for linkage accounting for genetic heterogeneity in the context of GWHM. We used this novel method to analyse a GWL study conducted with 21 Moroccan consanguineous families with at least one child presenting with severe TB. The results obtained for regions involved in TB showed three interesting regions on chromosomes 4, 9 and 11.
We focused first on the chromosome 9 region as the interval contains the gene JAK2. JAK2 is a janus kinase involved in many different signalling pathways including IL-12 and IFN-γ. This renders JAK2 a good candidate gene. In one patient who had suffered from miliary tuberculosis, sequencing of the 25 exons and the flanking intron regions revealed a homozygous missense mutation at amino-acid position 417, leading to the replacement of a glycine to a serine (G417S) (Figure 1). This variation has not been previously reported as a polymorphism in databases. In addition, we have sequenced gDNA from 200 healthy controls (half of them being Moroccan) and did not find this variation. The glycine at position 417 is conserved among species and the amino acid substitution is predicted to be possibly damaging functionally, using the polyphen software. Preliminary functional studies showed an impairment of STAT4 phosporylation upon IL12 stimulation in the patient cells suggesting that this mutation may affect the IL12 response. We are currently testing the JAK2 mutation functionally using EBV-B and Saimiri-T cell lines. We will study not only the response to IL-12 and IFN-γ, but also the cellular responses to other cytokines, such as IL-6, IL-23 and IL-27. We will then characterize the mutant alleles by gene transfer. Altogether, these data strongly suggest that this variation is indeed a pathogenic, TB-predisposing mutation, rather than an irrelevant polymorphism.
Overall, our preliminary results of disseminated TB of childhood in Morocco strongly suggest that mutations in IL12RB1 may account for a detectable fraction (about 1-3%) of pediatric TB. We are pursuing our sequencing discovery strategy at the genome-wide scale thanks to the new exome-sequencing technology. Moreover, the discovery of TB-predisposing genes may also benefit from GWL approaches. We are now validating the pathogenic role of the JAK2 mutation and analyzing the other two chromosomal regions linked to pediatric TB together with exome-sequencing data. All these results are consistent with our bold model as well as the Bayesian prediction that pediatric TB results from a collection of immunologically related single-gene defects
The genetics of human adult pulmonary TB
The main objective is the fine-mapping of a region on chromosome 8 linked with pulmonary TB in a) a family-based sample with at least one offspring with TB, and b) an independent case/control sample (>300 patients with TB and >600 controls). For this study, only patients >15 years with bacteriologically confirmed pulmonary TB are included. The family of the patient is enrolled if both parents or at least one parent and one unaffected older sibling are available. This family-based sample constitutes our primary sample, which is used in both the linkage disequilibrium mapping of the chromosome 8 region and the GWA study. Otherwise, patients with bacteriologically confirmed pulmonary TB without available family members are included in the case/control sample. The primary family-based sample with available DNA that has been used in the analyses described below consists of 185 families, including a total of 302 PTB patients.
The case/control sample was used as the replication sample for the results found in our primary sample. We have now collected a sample of 290 pulmonary TB patients and 750 healthy controls (blood donors) that was used for replication studies.
We previously reported the first region linked with pulmonary TB, on chromosome 8. Based on linkage results, a 90% confidence interval was defined to establish the peak region corresponding to a 1-LOD drop from the chromosomal position recording the highest multipoint LOD score. The region was further extended to avoid partial coverage of genes at the borders, so that the final region was defined from 55257 kb to 61363 kb. We previously observed that minor allele frequencies (MAFs) and linkage disequilibrium (LD) patterns of populations from Morocco most closely resemble those observed in European populations from among populations included in the International Hapmap Project (http://www.hapmap.org). Within the linked region, SNPs were selected using Hapmap data from CEPH (Utah Residents with Northern and Western European Ancestry) (CEU) and based on tag-SNP information available on http://www.hapmap.org in order to ensure 95% coverage of CEU tag SNPs at an R2 cut-off of 0.80. These SNPs also cover 93% of CEU tag SNPs at a higher R2 cut-off of 0.90. A total of 3183 SNPs were selected and genotyped on an ultra-high throughput Illumina platform (This represents the double of SNPs that were initially scheduled in our project in order to directly perform an ultra-fine association mapping of the linked region). The final set of SNPs was filtered to remove those with a genotyping call rate (CR) < 80%, a Hardy-Weinberg equilibrium (HWE) P-value<10-6 a MAF <1%, and >10 Mendelian inconsistencies, leaving 2865 high quality SNPs. The software package PLINK, which is optimized for handling large genome-wide datasets, was used for data management and the formatting of data files for analysis programs.
The analysis was based on methods developed for family-based association studies, in particular the FBAT program. The strongest association was found at rs2176641 located near TOX under the additive model (P=0.0002). A total of 49 SNPs achieved a P-value under 0.01 (see figure), and were selected and genotyped in our replication Moroccan sample comprised of 317 independent cases and 650 independent controls. Results showed that only a single signal (based on two correlated SNPs, TOX1 and TOX2) in the proximity of the TOX gene was replicated under the very same genetic model (subjects homozygous for the rare allele are at risk of pulmonary TB) in the case/control replication sample (combined Moroccan analysis p-value for TOX1: 1.1 x 10-5). In addition, the association was much stronger (combined Moroccan analysis p-value for TOX1: 4.3 x 10-8) in the subsample consisting of the younger patients (age at onset of TB < 25 years). For the TOX1 SNP, the odds ratio of developing PTB for AA homozygous subjects vs. AG/GG subjects was estimated at 3.1 (2.0-4.9).
We also attempted to validate this main finding in another pulmonary TB family-based population from Madagascar comprised of 254 trios (two parents and one affected PTB offspring). We genotyped in this sample our two main TOX SNPs as well as 6 other SNPs in LD with those SNPs in Hapmap populations of different origin. Although the frequencies of the risk allele of TOX1 and TOX2 SNPs are lower in Madagascar (~0.15) as compared to Morocco (~0.40) we could replicate the association for the TOX2 SNP (p<0.04) in the sub-sample of patients with age at onset of TB < 25 years (table 1). Interestingly, the evidence for replication was even stronger in the other correlated SNPs which present a higher frequency in Madagascar (~0.22). The TOX gene is an interesting candidate gene as it codes for a nuclear factor (“Thymocyte selection-associated high mobility group box” protein) which has several roles in the immune response, and in particular the development of CD4+ T cells. We are currently conducting functional studies exploring the association of our TOX SNPs with a) the expression of TOX in different cell types, and b) the distribution of lymphocyte population subsets (in particular CD4/CD8). A manuscript reporting these findings is in preparation.
In addition to the positional cloning strategy of the chromosome 8 region, we are conducting a genome-wide association (GWA) study which is making possible to investigate the whole human genome for genetic variants associated with common diseases such as pulmonary TB. The genotyping was performed in our primary family-based sample, and the subjects were genotyped with the Illumina Hapmap610 including 620,901 markers (599,011 SNPs and 21,890 monomorphic markers used for CNV analysis). Of a total of 599,011 SNPs included on the chip, 566,317 SNPs met the quality control criteria. We conducted the same analyses as described for the previous Aim 2.1 leading to a total of 272 SNPs displaying a p-value <5x10-5. All these SNPs were genotyped and analyzed in the replication sample of 317 cases/650 controls, again as described above. The replication study showed that at least three regions on chromosomes 3 (2 regions) and 4 presented replications based on the same allele as compared to results from the primary sample with combined p-values ranging from 2x10-5 to 4x10-5. Replication studies are ongoing in the Madagascar sample as well as in other samples from Mexico and Columbia.
Recently published GWA studies as well as our own GWA study in Morocco did not provide any evidence for strong effect of common polymorphisms in pulmonary TB. We therefore plan to assess the role of rare coding variants in pulmonary TB (which cannot be properly investigated by GWA studies) by conducting whole exome sequencing in a subsample of pulmonary TB patients (in particular including the newly recruited patients) and exposed controls. All the variants found to be interesting (e.g. strong excess in patients as compared to controls) in this first study will be explored in the whole Moroccan sample, and, when confirmed, will be investigated by functional studies.
Our identification of susceptibility variants in the proximity of TOX represents a major breakthrough in the human genetics of Pulmonary TB, in that this finding has been replicated in Morocco and validated in Madagascar. Moreover, TOX encodes Thymocyte selection-associated high mobility group box protein, and is a key transcription factor leading to CD4 cell development in the murine model. The role of T-cells in mycobacterial infection and disease is becoming increasingly recognized. The positional cloning approach thus identified a novel PTB candidate gene implicated in processes that are under active investigation among immunologists interested in host-Mycobacterium tuberculosis interaction.
MICROBIAL ADAPTATION AND TARGET CELL RESPONSES TO MYCOBACTERIAL INFECTIONS
Responses of macrophages and epithelial cells to infection with M. tuberculosis
Studies with primary human cells and cell lines
The expression of cytokines, chemokines and molecules involved in antigen presentation after infection of A549 AECII human cell lines, mouse pulmonary epithelial primary cells and of human primary epithelial cell cultures with BCG, M. tuberculosis, or stimulated with mycobacterial products was compared to that of autologous (syngeneic) macrophages.
Alveolar epithelial type II cells (AEC II) have been proposed to be of importance in pulmonary infections. In the lung tissue, there are 10 times more AEC II than alveolar macrophages (AM) and therefore, the chance that M. tuberculosis encounters an epithelial cell is higher than encountering a macrophage during the initial steps of infection. We postulate that type II pneumocytes are an important reservoir of Mtb that could be responsible for a sustained presentation of HBHA to T lymphocytes and for a subsequent chronic stimulation of these cells that could participate to a better control of the infection.
The murine alveolar macrophage cell line MH-S derived from BALB/c mice and as AEC II we used T7 a transformed cell line on C57BL/6 genetic background. In parallel, primary AEC type II and alveolar macrophages were isolated. A method in which AECII were isolated after negative depletion of CD45+ leukocytes and CD146 endothelial cells was used for this purpose. Isolated AECII consisted of 9% of total cells. When incubated with BCG, LPS or TNF-α, type II cells produce MCP-1 and MMP-9, while alveolar macrophages produce MIP-2 and TNF-α .
The response of human AECII to mycobacterium was also studied. To determine the role of type II pneumocytes in HBHA presentation, we first investigated whether primary type II pneumocytes fresh and the A549 cell line possess surface molecules required for efficient presentation to T lymphocytes and whether these molecules are modulated by IFN-γ and TNF-α inflammatory cytokines as well as by infection with attenuated Mtb (BCG and M. tuberculosis phoP mutant SO2 strain). We first observed that fresh AECII constitutively express HLA-DR molecules as opposed to A549 cell line for which the HLA-DR molecule expression is low and not inducible. Thus, A549 cannot be used as a model for studying class II antigenic presentation. Classical co-stimulatory molecules of the B7 family (CD80, CD86 and B7-H2) were either not detectable or expressed at very low level by few cells, and this is the case for both the A549 cell line and for the human freshly isolated AECII. On the opposite, alternative co-stimulatory molecules (CD54 and CD58) are expressed at very high level by most cells. Finally stimulation with IFN-γ and/or TNF-α increases significantly the cell surface expression only of CD54 and we observed that infection with the two attenuated M. tuberculosis strains used has no significant effect on phenotypic markers expression. Thus, AECII possess HLA molecules required for antigenic presentation and also alternative co-stimulatory molecules that could help for an efficient antigenic presentation to CD4+ and CD8+ T cells.
Regarding the molecules produced secreted by AECII, we observed that high level of chemokines are secreted by AECII (MCP-1, RANTES, IL-8, IP10) after exposure to IFN-γ and TNF-α or infection with attenuated M. tuberculosis SO2, allowing the recruitment of T lymphocytes and other cells. Only IL-6 is produced at high level. No IL1-β, IL-15, IL-18, IL-23, TNF-α, IFN-α, IL-10 or IFN-γ were detected in our conditions. TGF-b is secreted at basal level by AECII. Mtb-SO2 infection increases chemokine secretion. BCG infection has no effect on chemokines/ cytokines production.
In order to investigate the possible antigenic presentation of HBHA by AECII to autologous T lymphocytes, AECII were freshly isolated from lung fragments removed surgically from cancer patients and blood was collected in parallel. Eleven patients were recruited, but their Mtb infection status was unknown. Therefore, whereas the aim of the study was to analyse the presentation of mycobacterial antigens (PPD, HBHA and ESAT-6), numerous control antigens were added in the study to validate the co-culture model as in addition we did not know whether AECII could activate or anergize T cells in a recall response. Thus, CMV-pp65 protein, EBV-LMP1 overlapping peptides, EBV-LMP2A overlapping peptides, Candida albicans-MP65 protein, tetanus toxoid from Clostridium tetani, were tested in parallel to the mycobacterial antigens as some patients might have been infected in the past either with Epstein-Barr Virus (EBV), Cytomegalovirus (CMV), C. albicans and as most of them should been immunized against tetanus. In addition, classical antigen-presenting cells, namely alveolar macrophages and blood monocytes, were tested as controls in parallel to the AECII for their capacity to present HBHA as well as the other mycobacterial and control antigens to autologous T lymphocytes.
In parallel, IFN-γ release assay (IGRA) in response to the different antigens were performed on the patient’s PBMC to determine their immune status toward the different antigens that were tested. IGRA indicated that three patients responded to PPD, three to CMV-pp65, two to C. albicans-MP65 and two to ESAT-6, whereas only one patient responded to HBHA and was therefore considered as latently infected with Mtb (LTBI subject). One patient presented in addition to his lung cancer, a cervical adenopathy that was demonstrated to be due to a Mtb infection. PBMC from this patient produced IFN-γ in response to ESAT-6 but not in response to HBHA (Figure 2, grey symbol).
We showed that human AECII are able to present antigens to autologous lymphocytes, as exemplified on Figure 2 for PPD, ESAT-6, and to a lesser extent for HBHA, and in other patients for CMV-pp65. However, the antigen presentation by macrophages was most often more efficacious.
As IFN-γ is probably not the only read-out of an efficacious antigen presentation, secretion of other cytokines than IFN-g as well as the lymphocyte proliferation were measured. TNF-α and IFN-γ showed a similar response pattern. In contrast, among several other cytokines / chemokines measured, IP10 was detected in the co-culture of HBHA- or ESAT-6-loaded AECII with lymphocytes from the TB patient, who had only a very low IFN-g response at least to HBHA.
In conclusion, we have shown that human AECII play a role in the presentation of mycobacterial antigens to autologous T lymphocytes. IFN-g secretion in the coculture is not the only interesting read-out and it should be combined at least to measurement of IP10 concentrations. A few more patients, especially selected as responding to HBHA, will be evaluated before writing the manuscript to report these data.
Alveolar epithelial cells in murine infections with mycobacterium.
We compared primary murine alveolar epithelial cells with pulmonary macrophages (PuM) in their ability to internalize and control M bovis BCG growth and their capacity as APCs. AEC internalized and controlled bacterial growth as well as presented antigen to primed T cells.
Secondly, we compared the capacity of both cell types to secrete cytokines and chemokines upon stimulation with mycobacterial products. Activated PuM and AEC displayed different patterns of secretion.
Finally, we analyzed the profile of response of AEC microbial and immune stimuli exemplified by TLR ligands and cytokines, respectively. Interestingly, TNF could stimulate the production of CCL2/MCP-1. Since MCP-1 plays a role in the recruitment of monocytes and macrophages to sites of infection and macrophages are the main producers of TNF, we speculate that both cell types stimulate each other. Another cell-cell interaction was suggested when IFNs (produced mainly by lymphocytes) were able to induce expression of chemokines (IP-10 and RANTES) by AEC involved in the recruitment of circulating lymphocytes to areas of injury, inflammation, or viral infection.
Thus, we confirm previous data on the capacity of AEC regarding internalization of mycobacteria and their role as APC, and extend the knowledge of AEC as a multifunctional cell type by assessing the secretion of a broad array of immune factors in response to several different types of stimuli.
The respiratory epithelium is a physical and functional barrier actively involved in the clearance of environmental agents and composed of type I and type II alveolar epithelial cells (AEC). Another cell important in lung defense is the pulmonary macrophage (PuM). We hypothesized that AEC play a role in transmitting signals from the airway mucosal to systemic compartments by modulating the activity of PuM. For this, we collected supernatants from unstimulated or LPS stimulated AEC. We also tested the direct effect of AEC-derived media on bacteria and found that AEC-secreted factors had a dual effect, in one hand controlling bacteria growth and on the other hand enhancing macrophage microbicidal activity.
We have shown that soluble factors released by alveolar epithelial cells (AECs) increase macrophage intracellular killing of BCG. We aimed to identify this bacteriotoxic mechanism. Murine bone marrow-derived macrophages (BMM) and pulmonary macrophages (PuM) infected with BCG. After infection, bacteria were removed and cells incubated further in media with or without IFN-γ or supernatants from AEC (AECsup). Treating BMM with IFN-γ, increased intracellular killing compared with that of cells cultured in medium alone whereas in PuM IFN-γ was ineffective. In both cell-types, IFN-γ treatment increased transcription of iNOS, IP-10 and IL-12 and secretion of IL-12 and IL-6. Transcription of suppressor of cytokine signaling (SOCS)1 was higher in PuM than BMM. SOCS1 expression was important for mediating the ineffectiveness of IFN-γ in increasing intracellular killing in PuM, as PuM from SOCS1/IFN-γ-/- mice showed enhanced killing after IFN-γ treatment. On the other hand, AECsup treatment did not affect iNOS, IP-10 nor IL-12 expression but increased Arg1 transcription and IL-6 secretion. BMM and PuM responded to IFN-γ indicating that both cell types have receptors that recognize and transmit signals delivered by IFN-γ. IFN-γ only increased intracellular killing of BCG in BMM but not in PuM. On the other hand, treatment with AECsup increased intracellular killing of BCG in both types of macrophages. Such killing was not associated with increased IL-12 or iNOS expression.
Thus, AEC supernatants but not IFN-γ augment the intracellular killing of BCG by pulmonary macrophages in a NO-independent manner.
Mycobacterial adaptation to infection
The hbhA gene is known to be differentially expressed during infection of different cell types. The aim of this part of the work is to better understand this regulation and find factors involved in the differential expression. The expression profile of mutants affected in these factors will help us to find hbhA co-regulated genes that could code for proteins involved in the persistence or latency of Mycobacteria. The basis of this approach is the study of the expression of hbhA and downstream genes potentially correlated in an operon structure by using qRT-PCR during axenic cultures or during infections of different cell types with BCG or M. tuberculosis. We developed first a method to study the expression of hbhA in in vitro growth conditions. During growth, aliquots were collected at different time points to extract RNAs. The expression of hbhA was then analyzed by quantitative PCR (qRT-PCR).
These studies showed a regulation of the hbhA relative expression during growth and evidenced also an up-regulation of the hbhA relative expression in M. tuberculosis compared to M. bovis BCG.
Upstream of the hbhA gene, a putative regulatory gene encoding a protein of the Xre family of transcriptional regulator is located. This gene is present in all sequenced mycobacterial genomes, associated to hbhA. The Xre family of transcriptional regulator has been described in the past to be involved in the repression of Bacillus subtilis prophages cycle. More recently, some members of this family have been proposed to act as transcriptional activators. The distance between this gene and hbhA is sufficiently long in the M. tuberculosis genome to contain a hbhA promoter. By using allelic exchange, the xre-like gene was deleted in the BCG genome. The expression levels of hbhA and adjacent genes in BCG-Δxre mutant strain have been compared to the levels in BCG by qRT-PCR normalised to the 16S rRNA gene during in vitro growth. During the early phase growth, we observed a relative expression decrease of hbhA in the mutant strain compared to BCG of about 17-fold. Later in the growth curve (mid-log phase), the measured decrease was around 4-fold. These amplitudes of downregulation are high, compared to amplitudes usually observed in mycobacterial transcriptomic studies. In contrast, for the two downstream genes the maximum amplitude of decrease in their expression was around 1.7-fold at the early phase growth stage. Then the transcription of these genes was even less affected by the deletion of xre-like gene, suggesting that their expression is not correlated to the hbhA expression and that the Xre-like regulator activates the expression of hbhA only. As a control, the expression of a gene located three genes upstream of hbhA was not at all affected by the deletion of xre-like. To confirm that the Xre-like regulator is involved in the activation of hbhA, we transformed BCG with a construction allowing for the overproduction of Xre-like by using an inducible promoter fused to its gene and compared the expression level of hbhA in this recombinant strain with that of non-recombinant BCG. After culture of both strains up to mid-log phase, we measured by qRT-PCR a 3.8-fold increase of hbhA expression in the recombinant strain. This confirmed that the xre-like gene located directly upstream of hbhA codes for a transcriptional regulator that is an activator of hbhA. For this reason, we decided to rename this gene “hbhR”.
We failed to obtain reproducible results by using this vector for complementation. It was most probably due to the fact that the promoter used has been demonstrated to behave randomly in the M. bovis BCG background. For this reason, we decided to complement the deleted strain by using a plasmid containing the hbhR gene under its own promoter and compared the expression level of hbhA in this recombinant strain and in BCG. After culture of both strains up to mid-log phase, we measured hbhA expression by qRT-PCR. There was no restoration of expression in the recombinant strain despite a correct expression of hbhR. This suggests that there is no trans-acting activation of hbhA complementing the deletion of hbhR. In parallel, the HbhR transcription factor fused to a histidine tag has been over-expressed in E. coli and purified for molecular analyzes, such as gel retardation assays on the hbhA promoter and on its own promoter, since many transcriptional regulators are under their own control. No binding of HbhR on the hbhA promoter has been observed in these conditions, but binding on its own promoter has been observed. From the lack of complementation, we thus conclude that the knock-in deletion of hbhR has most probably a polar effect on the expression of hbhA. This implies that hbhA expression can be driven through a promoter located upstream of hbhR. This gene is separated from the upstream predicted ORF by a 86 bp region in which we could predict a promoter with -35 and -10 boxes by using BPROM algorithm (www.softberry.com). This algorithm predicted also a consensual promoter upstream of hbhA in the proximal part of the 353 bp region separating hbhR and hbhA. Moreover, a second promoter for hbhA in the distal part of this region has been recently described to be SigF-dependent. Despite the presence of these two promoters for hbhA, our results suggest that the promoter located upstream of hbhR is mainly used for the transcription of hbhA in M. bovis BCG 1173P2 as knock-in in the hbhR gene strongly affects it. The absence of binding of purified HbhR on the hbhA promoter suggests that, in BCG, HbhR has no direct interaction for the activation of hbhA.
The same experiments were performed in the M. tuberculosis background. A H37Rv-∆hbhR mutant was generated. As for the BCG-∆hbhR, we analyzed the HBHA protein level and the hbhA gene relative expression. As observed with BCG-∆hbhR, the results showed a decrease in the HBHA protein level and a down-regulation (approximately 4-fold) of the hbhA gene expression in the mutant compared to the parental strain. As for the BCG-∆hbhR mutant, the complementation did not restore the wild-type phenotype.
To characterize the ∆hbhR mutant, we followed the expression of hbhA after infection of human epithelial type II pneumocytes (A549 cells line) by qRT-PCR. BCG-∆hbhR and its parental strain were used to infect A549 and RNA from both strains was collected after 24, 48 and 72 hours of infection. Expression levels of hbhA were measured at each time point and compared. In these conditions, hbhA transcription levels measured showed a decrease of about 8.3-fold in the mutant strain when compared to BCG after 24 hours. In both strains, the level of transcription reaches a maximum after 48 hours of infection and then decreased, but the level of transcription in the mutant strain is always lower than in BCG. This observation, as well as the observation of regulation of hbhA transcription during axenic cultures even in the absence of hbhR suggests an additional mechanism of regulation that may occur on the promoter region of hbhA. This mechanism could be relayed by another transcription factor or by SigF.
To continue our analysis of the HBHA regulation, the BCG-∆hbhR and H37Rv-∆hbhR mutants were integrated in microarray experiments to define the hbhR co-regulated genes. When we compared the ∆hbhR mutant with the parental strain, the main significant result is the down-regulation of hbhA confirming the previous results obtained by qRT-PCR. The microarray comparison between the both strains evidenced 20 up-regulated genes in the BCG ∆hbhR mutant with the parental strain. Among these genes, 43% are conserved hypothetical proteins, the others are involved in different metabolisms (cell wall, lipid, respiration).
Then, the comparison between the both strains evidenced 35 down-regulated genes in the the ∆hbhR mutant with the parental strain. Among these, hbhA was found confirming the previous results obtained by ELISA and qPCR. The others are again involved in different metabolisms (cell wall, lipid, respiration). Interstingly, most precisely the down-regulated genes, a lot of them are located into the same locus, the katG locus. The katG gene encodes for a catalase peroxidase that is involved in the conversion of isoniazid to its active form in M. tuberculosis.
CELLULAR IMMUNITY AND REGULATION OF HOST PATHOGEN-INTERACTIONS
Dendritic cells and M. tuberculosis
Phenotype and functional responses of DCs stimulated with mycobacterial molecules
During M. tuberculosis infection, dendritic cells (DCs) function is modulated by mycobacterial molecules, affecting the ensuing protective immune response and the outcome of the infection. In this project we proposed to perform detailed analysis of the relative role of distinct DC populations in driving immune response to Mtb. Thus, we dissected the molecular pathways linked to DC activation following Mtb infection or exposure to mycobacterial molecules. Moreover, we analyzed the mechanisms by which mycobacteria interfere or promote DC functionality.
We first analysed the phenotype and functional properties of DCs derived from monocytes from active Tuberculosis patients, asymptomatic infected (LTBI) and non-infected individuals. To this end we took advantage of a DC-differentiation in vitro model by exposing peripheral monocytes to interferon-α (IFN-α). IL-4 inducing the phenotype of immature DCs (IL-4-DCs), while IFN-α induces differentiation of monocytes into highly activated and partially mature DCs (IFN-α-DCs) retaining a marked phagocytic activity and exhibiting he ability to stimulate CD8(+) T cell responses against antigens from both pathogens and tumors. During this period, we found that DCs differentiated in the presence of IFN-α are fully competent for presenting apoptotic cell-associated antigens for cross-priming CD8(+) effector T cells and concomitantly for activating CD4(+) T helper cell. We also found that both IFN-α-DCs and IL-4-DCs exposed to a series of mycobacterial molecules, including Ag85B, PstS1, ESAT-6 and LAM, exhibit high levels of the activation and maturation markers, namely CD80, CD86, HLA-DR and HLA-A-B-C whereas only IFN-α-DCs are able to express high levels of the differentiation marker CD83. These findings suggest that the synergy between selected Mtb Ags and IFN-α promote DC maturation stimulating Ag-specific T cell response. The further analysis of peripheral DC subsets, including type-1 and type-2 myeloid dendritic cells (MDCs) and plasmacytoid DCs (pDCs), of both TB patients and LTBI individuals revealed that pDCs are significantly decreased in TB and AI patients with respect to healthy donors. Moreover, CD11c+/CD86+ myeloid DCs are higher in TB and AI patients than controls, whereas BDCA3+ type-2 myeloid DCs are lower in TB patients and healthy controls with respect to AI individuals. Therefore, we may assume that individuals with active or latent Mtb infection, characterized by a reduced pDC compartment in favour of myeloid DC populations, have a defect in a proper cross-presentation of mycobacterial Ags to CD8(+) T cells.
In a second set of experiments, we used expression microarrays to compare the molecular pathways activated in IFN-α-DCs and IL-4-DCs following infection either with Mtb or BCG. We found that these two agents stimulate different gene expression profiles in IFN-α-DCs and IL-4-DCs. BCG induces immune system processes genes in both type of DCs, However, in IFN-α DCs, it determines selectively a striking up-modulation of genes that regulate cytokine production during immune response, such as TLR2, IL-1, IL-6 and IL-23A. On the contrary, Mtb infection mainly induces genes related to metabolic pathways in both DC types even though a limited immune signature characterized by up-modulation of genes such as IL-6, IL-10 and IL23A occurs in IFN-α-DCs.
Of interest, IFN-α-DCs generated from PBMCs of actively or AI patients showed that a significant reduced expression of genes linked to immune response, in particular to Ag processing (LAMP-3) and Ag presentation (HLA family, CD80) as well as a downregulation of IFN-related gene IRF-4 occurs in these individuals with respect to healthy donors. The further phenotypic analysis of the same samples indicates that IFN-α-DCs derived from infected patients exhibited a reduced expression of the co-stimulatory molecules CD80 and CD86. Thus, these finding confirm that IFN-α-DCs derived from Mtb infected patients may not be able to appropriately process and present mycobacterial Ags to T cells leading to a defective immune response to the pathogen.
We also characterized microRNA pathways in IFN-α-DCs and IL-4-DCs following Mtb infection. The microRNA expression profile of IFN-α-DCs differs significantly from that of IL-4-DCs being the former characterized by the expression of microRNAs which regulate pivotal genes of the immune function and differentiation.
Infection of IRF8 KO mice with M. tuberculosis
Immune cell recruitment into lungs is pivotal in establishing protective immunity through granuloma formation and neogenesis of lymphoid structures (LS). Interferon regulatory factor-8 (IRF-8) plays an important role in host defense against M. tuberculosis , although its mechanisms driving anti-mycobacterial immunity remain unclear. IRF8 KO mice are characterized by their high susceptibility to many pathogens, including several bacterial species. IRF-8 KO and WT were infected via aerosol with Mtb Erdman strain. We found that up to two weeks following M. tuberculosis exposure, bacterial load in lungs and spleens in the two groups were similar. After this, IRF-8 KO mice developed a lethal acute TB, whereas in WT-B6 the disease is chronic stage. In IRF-8 KO lungs, pulmonary granulomas were larger and neogenesis of LS impaired. This associated with a progressive loss of infiltrating T lymphocytes and a massive prevalence of neutrophils. Moreover, our findings unravel an important role of IRF-8 in granuloma organization and the distinct pattern of cellular recruitment into the inflammatory lesion IRF-8 KO, with respect to controls, confirms the importance of the diverse recruitment of immune cells in acute versus chronic granulomatous lesions.
We characterized phenotypic and functional responses of DCs stimulated with mycobacterial molecules, including expansion of CD8(+) T cells and synthesis of IL-12, IL-23 and IFN-γ. The induction of CD4(+) Th-17 cells after immunization has been shown to contribute to protection and the generation of effector Ag-specific CD4(+) Th1, CD4 Th17 cells and CD8(+) T responses are strictly dependent on Ag presentation by DCs. The DC-mediated T cell responses to phosphate-binding transporter lipoprotein of the mycobacterial cell membrane PstS1, reported to be a strong immunomodulatory Ag of Mtb and an inducer of Th1 and Th17 responses. We found that PstS1, in contrast to other M. tuberculosis antigens such as Ag85B and HBHA, selectively activates DCs, both in vitro and in vivo, producing high levels of IL-6 and IL-23. We also investigated the ability of PstS1 to modulate CD4+ T cell responses during immunization with the Mtb candidate vaccine Ag85B and found that PstS1 induces Ag85B-specific memory CD4+ T cell proliferation and IFN-γ, IL-17 and IL-22 secretion. These effects were mediated through DC stimulation, particularly of the CD8α- subtype. To test the adjuvant activity of PstS1, we immunized mice with Pst1 and Ag85B. PstS1 protein administration during priming impaired the generation of Ag85B-specific immunity. Our data provide evidence that PstS1 may function as an immunomodulatory molecule stimulating DC, particularly the CD8α- subtype, affecting principally memory phenotype T cells and suggest the molecular rationale for exploiting its use in Mtb subunit-based vaccine strategies.
Role SOCS proteins in the outcome of infection with M. tuberculosis
SOCS1 inhibits IFN-γ secretion and thereby precludes efficient clearance of BCG and M. tuberculosis in macrophages
Protective immunity against infection with M. tuberculosis demands IFN-γ. Modulation of IFN-γ signaling is therefore of central importance in the outcome of infection with M. tuberculosis. SOCS1 The suppressor of cytokine signalling-1 (SOCS1) plays a major role in the inhibition of IFN-γ-mediated responses. We investigated the regulation and the role of SOCS1 during infection with BCG or M. tuberculosis in vitro and in vivo. We observed that infection with virulent or avirulent mycobacteria induces SOCS1 expression in vitro, in mouse and human macrophages and DCs, and in vivo in a murine model. Murine BMM, human monocyte-derived macrophages and lungs from infected mice contained higher titers of SOCS1 mRNA after infection with M. tuberculosis compared to BCG.
We then asked whether MyD88- and IRF3-mediated intracellular signalling pathways are involved in the augmented SOCS1mRNA levels after infection with BCG. Similar SOCS1 mRNA levels were found in BCG-infected IRF3-/- and WT BMM, while MyD88-/- BMM contained significantly lower SOCS1 mRNA levels than controls. SOCS1 mRNA levels were diminished in TLR2-/- BMM albeit to a lower degree than those in MyD88-/- BMM. TLR4 played a minor role if any in regulation of SOCS1 mRNA levels in BCG-infected macrophages.
Next, we investigated whether SOCS1 plays a role in intracellular mycobacterial growth control. SOCS1-/- and WT BMM were infected with BCG or M. tuberculosis. SOCS1-/- BMM showed reduced bacterial load compared to WT controls. SOCS1 gene was deleted in macrophages and neutrophils by crossing SOCS1fl/fl mice with LysM-cre SOCS1fl/fl mice. Infected LysM-cre SOCS1fl/fl BMM also showed reduced M. tuberculosis levels compared to SOCS1fl/fl controls. In agreement, increased bacterial growth was observed upon incubation of BCG or M. tuberculosis-infected BMM with Tkip, a peptide that, similar to SOCS1, binds and inhibits autophosphorylation of the JAK2 kinase and phosphorylation of the intracellular IFN-γ receptor.
Mycobacterial-infected SOCS1fl/fl LysMcre or SOCS1-/- BMM and BMDC showed increased levels of IFN-β, IFN-α, IFN-γ as well as IFN-regulated genes compared to controls. Since IFN-γ stimulates macrophage activation and controls mycobacterial growth, we studied whether the improved mycobacterial growth control in SOCS1-/- BMM is IFN-γ-dependent. For this purpose, bacterial titers in IFN-γ-/- /SOCS1-/- and IFN-γ -/- BMM were compared. IFN-γ-/- /SOCS1-/- and IFN-γ-/- BMM showed similar BCG loads. Moreover, similar bacterial levels were measured in WT, IFN-γ-/- and IFN-γ-/-/ SOCS1-/- M. tuberculosis-infected BMM, while the bacterial load in SOCS1-/- BMM was lower compared to all other groups. Altogether these results indicate that SOCS1 inhibits secretion of IFN-γ that otherwise would mediate a more effective clearance of mycobacteria in macrophages.
Since IL12 is known to be a main stimulus for IFN-γ secretion, we next investigated whether a M. tuberculosis-induced SOCS1 diminished expression of IL12 could account for the increased IFN-γ secretion in SOCS1-/- BMM. However, similar of IL12 were secreted by M. tuberculosis-infected SOCS1-/- and WT BMM, suggesting that the response to, but not the secretion of IL12 could account for the increased IFN-γ release in SOCS1-/- BMM. In fact, SOCS1-/- BMM stimulated with IL12 contained enhanced IFN-γ mRNA and protein levels compared to WT controls. Addition of anti-IL12 antibodies reduced IFN-γ mRNA levels in M. tuberculosis-infected SOCS1-/- and WT BMM, indicating that IL12 is required for the enhanced IFN-γ secretion by the infected SOCS1-/- BMM. The increased response to IL12 of SOCS1-/- BMM was associated with higher levels of the IL12 receptor (R) β1 in SOCS1-/- compared to WT BMM.
Whether SOCS1 further impedes the responses of M. tuberculosis-infected BMM to IFN-γ was then studied. Incubation of either SOCS1-/- or WT BMM with IFN-γ after infection with M. tuberculosis resulted in diminished bacterial load as compared to respective IFN-γ-untreated infected controls. In line with this, CXCL10 or iNOS mRNA levels were increased in IFN-γ-treated, infected WT or SOCS1-/- BMM cells as compared to uninfected controls or to infected cells in absence of IFN-γ-stimulation.
To investigate how SOCS1 shapes the macrophage response to M. tuberculosis at the transcriptome level, genome-wide expression analysis was performed in infected and uninfected SOCS1-/- and WR macrophages. A relevant segment of the macrophage transcriptome is remodelled after infection with M. tuberculosis. Unexpectedly, the majority of such genes overlapped with those differentially regulated due to SOCS1 deficiency, strongly suggesting a relevant role of SOCS1 in regulating the metabolic activity of macrophages during the infection.
Role of SOCS1 in the outcome of murine infections with M. tuberculosis.
Using different knock out and conditional knock down mice, we demonstrated that SOCS1 expression by macrophages impaired M. tuberculosis clearance and IFN-γ and iNOS expression in mice, when measured 7 days after infection. At this early stage of infection, when increased secretion of IFN-γ is not detected in lungs from infected WT mice, we propose that a higher level of IFN-γ secreted by SOCS1-deficient macrophages impair bacterial growth in vivo.
Instead, SOCS1 did not hamper bacterial control at later time points, when IFN-γ secretion by T cells is prominent. Thus may the explained by the capacity of infected, SOCS1-expressing macrophages to respond to IFN-γ.
We also observed an overwhelming infection-induced pulmonary inflammation in RAG1-/-/ SOCS1-/- and IFN-γ-/-/ SOCS1-/- mice. Since LysM-cre SOCS1fl/fl mice in contrast, showed no increased pathology, we conclude that SOCS1 expressing non-macrophage cells are main controllers of detrimental inflammation.
We propose that M. tuberculosis actively promotes SOCS1 expression by macrophages to counteract effective mycobacterial control at early time points after infection in vivo, before initiation of IFN-γ secretion by NK or T cells. SOCS1 facilitates mycobacterial growth by hampering macrophage IFN-γ secretion in response to infection-induced IL-12. At later time points after infection, despite SOCS1 expression, macrophages respond to IFN-γ secreted by T- or NK cells, and will not counteract bacterial control to any further extent. Instead, SOCS1 in non-macrophage cells protects mice from severe inflammation.
SOCS1 expression in TB patients
We investigated the association of SOCS with disease progression in patients with pulmonary TB. For this purpose we studied peripheral blood mononuclear cells (PBMCs) and T cells from patients with pulmonary TB (TB, n=33) and healthy endemic controls (EC, n=15). Cases were stratified into those with moderately advanced (mod-PTB) or far advanced disease (adv-PTB). IFN-γ, SOCS1 and SOCS3 gene expression was determined by RT-PCR. Statistical analysis was performed using the Mann-Whitney test. Levels of IL6 (p=0.018) and IL10 (p=0.013) were found to be elevated in PBMC supernatants from TB patients as compared with EC. SOCS1 mRNA gene expression in T cells from TB patients was increased as compared with that of EC (p=0.02). In addition, levels of SOCS1 mRNA transcripts were found to be elevated in PBMCs of adv-PTB as compared with mod-PTB (p=0.008) cases.
Our data show that raised SOCS1 levels are associated with increased disease severity in TB.
Chemokines- TLR and inflammation during infection with M. tuberculosis
Effects of heat-inactivated M. tuberculosis on the expression of chemokine receptors in macrophages and T cells
We initially investigated the effect of heat-killed M. tuberculosis (HKMTB) on the induction of Th17 cells. CD4+ T cells from BALB/c mice were polarized to Th17 cells in the presence of graded concentrations of HKMTB in vitro. HKMTB enhanced the polarisation of Th17 cells in a dose-dependent manner. Furthermore, HKMTB also induced an elevated level of RORγt (the key transcription factor of Th17).
Toll-like receptors (TLRs) represent the primary sentinels of microbial infection and have evolved to detect, and respond to, a variety of pathogen-associated molecular patterns. The TLR binding in leukocytes and in cells endogenous to the site of infection such as epithelial and endothelial cells, triggers the rapid synthesis and secretion of inflammatory cytokines and chemokines. The expression of chemokines after TLR agonist exposure is necessary for the efficient recruitment of inflammatory leukocytes to the tissue in response to localized microbial infection. Paradoxically, we found that TLR ligands administered systemically profoundly alter peripheral blood leukocyte trafficking to sites of inflammation. The TLR2 ligand bacterial lipoprotein (BLP) given intravenously as opposed to intraperitoneally significantly reduces the recruitment of leukocytes to a site of skin inflammation These results indicate that TLR ligands can modulate or suppress expression of chemokine receptor involved in cell trafficking. We have also investigated the effect of the various ligands of TLRs on the chemokine expression in murine macrophages. We used a novel approach to assess the expression and modulation of all known chemokine receptors in defined cell populations simultaneously by TagMan Low Density Array (TLDA, Applied Biosystems). B cells, M-CSF induced macrophages, Flt3L-induced dendritic cells and thioglycolate-elicited peritoneal macrophages were purified and then stimulated with a panel of TLRs. Our results show that CCR2 and CCR4 were down-regulated by nearly all the TLR ligands, whereas CCR7 was simultaneously up-regulated. These data highlight a novel role for TLR ligands, such as mycobacterial BLP, in regulating leukocyte retention and migration away from innate immune lesions via discrete constitutive and inflammatory chemokine receptor regulation.
The role of chemokine expression in granuloma formation
We induced granuloma formation in mice using Freund’s complete adjuvant, which contains killed M tuberculosis. Granuloma were formed 1-2 weeks following subcutaneous injection and the granuloma together with the draining lymph notes were harvested 4 weeks after injection. Histological slides of the tissues were made and single suspensions were prepared for FACS analysis. Our results show that CCR2 and CCR4 expression on CD4+ T cells and macrophages were significantly enhanced whereas the expression of CCR7 was decreased. We then tested the ability of these cells to migrate in vivo after intravenous injection. CD4+ T cells migrated rapidly to the lymph notes and were highly responsive to recall antigen, MTB extracts. We also attempted to block the migration of these cells with anti-TLR2 antibody. Partial suppression of the cell migration was observed but not the recall response of the cells. These results, which are being prepared for publication, show that granuloma formation is one of a potent mean by which the host heightened defence against the infection.
The role of IL-33 on sepsis
Sepsis is a systemic inflammatory condition following bacterial infection with high mortality rate and limited therapeutic options. We have now shown that IL-33, the new member of the IL-1 family, reduced mortality in mice with experimental sepsis from cecal ligation and puncture (CLP). IL-33-treated mice developed increased neutrophil influx in the peritoneal cavity and more efficient bacterial clearance than untreated mice. IL-33 reduced systemic but not local pro-inflammatory response, nor a Th1 to Th2 shift. The chemokine receptor CXCR2 is crucial for recruitment of neutrophils from circulation to the site of infection. Activation of Toll-like receptors (TLRs) in neutrophils down-regulates CXCR2 expression and impairs neutrophil migration. We also showed that IL-33 prevented the down-regulation of CXCR2 and chemotaxis induced by the activation of TLR4 and TLR2 in mouse and human neutrophils. Furthermore, we demonstrated that IL-33 reversed the LPS-induced reduction of CXCR2 in neutrophils via the inhibition of G-protein-coupled receptor kinase-2 (GRK2) expression, a serine/threonine protein kinase that induces internalization of chemokine receptors. Finally, we found that patients who did not recover from sepsis had significantly more soluble ST2 (sST2, the decoy receptor of IL-33) than those who did. Together, we provide a novel mechanism of action of IL-33 and suggest a therapeutic potential of IL-33 in sepsis. We have also investigated the role of IL-33 and TLRs in the susceptibility to opportunistic infection following recovery from sepsis. We found that in contrast to the initial sepsis, IL-33 exacerbates the opportunistic infections. This is because IL-33 promotes Type II response via the induction of IL-4 which in turn activates alternatively-activated macrophages that induce heightened levels of regulatory T cells. These novel finding is now being prepared for publication.
Genotyping of D6 chemokine receptor
The D6 gene (CCBP2), coding for a chemokine decoy receptor, is on chromosome 3 ranging from 42,825,967 to 42,883,779 bps (maps hg18/NCBI 36). There are 15 Quality control-filtered SNPs in D6 (+/- 10 kbs) from the full set of GWA SNPs (see table). The International Hapmap project reports on 48 SNPs in D6 among 30 European trios (CEU) (46 have a Minor allele frequency > 0.01). The 15 SNPs that have been genotyped allow to tag 36 of these 46 SNPs with a R2=0.8 (~ 78% coverage at R2=0.8). The association analysis was based on methods developed for family-based association studies, in particular the FBAT program. We considered additive, dominant and recessive models, and the minimum P-value was retained for each SNP (Table). The lowest p-value at p=0.009 was observed with rs7433284. Note that a simple Bonferoni correction for testing 15 SNPs (without accounting for the fact we have tested several genetic models for each SNP) provided a corrected p-value of 0.05/15 = 0.0033. Note also that over the whole set of 566,317 SNPs tested at the genome-wide level, 13,618 (~2.4%) provided a p-value lower than 0.009. In conclusion, there is no significant association of any of the 15 CCBP2 SNPs and pulmonary TB in our Moroccan sample.
HUMANIZED MOUSE MODEL OF MYCOBACTERIAL INFECTION
Non-obese diabetic (NOD)/SCID/ γcR-/- mice and BALB/c RAG2-/- /γcR-/- can be stably engrafted with human (CD34+ hematopoietic stem cells (HSC) following sub-lethal irradiation due to a lack of an adaptive immune system and innate immunity defects. In these models, human CD34+ HSCs self-renew in the bone marrow and provide long-term repopulation of mice with multi-lineage human myeloid and lymphoid cell populations. These humanized mice have been shown to be useful tools to examine different microorganisms that involve human hematopoietic cells in their life cycles.
Granuloma formation in M. tuberculosis-infected individuals is considered to constitute an important component of the protection against disease.
The widely used mouse model of tuberculosis fails to reproduce the specific features of human granulomas, which are a hallmark of mycobacterial disease. Whether humanized mice can provide a better tool to study human granuloma formation and to experimentally address the role of human immune responses during Tuberculosis was explored.
Specifically, we have studied the outcome of infection with the attenuated Mycobacterium bovis BCG in NOD/ SCID/ γcR-/- mice, reconstituted de novo with human lymphoid and myeloid cells after neonatal transplantation with human CD34+ hematopoietic stem cells.
Spleens of BCG-infected humanized mice showed increased human CD4/ CD8 T cell ratio and a widespread differentiation of T cells from naïve into CD45RA-CCR7- effector memory cells.
BCG-infected humanized mice showed increased numbers of human T cells and macrophages in lungs and livers but paradoxically contained higher number of mycobacteria in tissues as compared to non-humanized controls. This enhancement of bacterial load was abrogated after in vivo depletion of human CD4+ but not CD8+ cells, and was associated with an increase in T cell expression of CD57 and PD-1, molecules associated with senescence and inhibition of T-cell function.
Despite the lack of bacterial control, organized granulomas resembling human lesions were observed in livers and lungs from BCG-infected humanized mice but not in non-humanized infected controls. Granulomas showed a core with compact aggregates of human CD68+ macrophages and giant multinucleated cells and elevated load of bacilli, surrounded by a layer of CD3+ T cells and a fibrous capsule encasing the lesions. Few to moderate numbers apoptotic and necrotic cells were present in the core of the larger granulomas. Higher levels of human IFN-γ, and the IFN-γ-regulated, T cell chemotactic CXCL9 and CXCL10 transcripts were detected in lungs and livers from BCG-infected humanized mice compared to non-infected controls.
The formation of granulomas depended on human CD4+ but not CD8+ cells since granulomas from mice depleted of CD4+ cells were fewer, minimal and irregular, while those from mice depleted in CD8+ cells retained the size and complexity of non-treated controls.
M. tuberculosis (but not BCG)-infected non-humanized mice showed granulomatous lesions in livers and lungs. However, typical granulomas were only found in M. tuberculosis-infected humanized mice. Granulomas from M. tuberculosis-infected humanized mice displayed a strikingly increased extension of necrosis compared to those from BCG-infected humanized mice or non-humanized M. tuberculosis-infected mice. In relation with this, it has been shown that M. tuberculosis infection stimulates necrosis of infected cells while BCG does not [5]. Increased severity and larger lesions were observed in livers of humanized mice as compared to non-humanized controls. Thus, the human transplant determined the formation of granulomas in BCG-infected mice and the granuloma features in mice infected with M. tuberculosis.
In summary, humanized mice showed a defective control of BCG infection, probably due to dysfunctional T cell responses or an impaired ability of macrophages to mount microbicidal responses. Notwithstanding the impaired bacterial control, granulomas that look like those observed in human M. tuberculosis were observed. Human CD4+ but not CD8+ cells were required for granuloma formation in mycobacterial infected humanized mice and for the enhanced growth of mycobacteria in these mice.
Our data support the hypotheses that HIV-induced depletion of CD4+ T cells lead to a direct disruption of the granulomas during M. tuberculosis infection. Alike, CD4+ T cells have been shown to contribute to the organization of granulomas in MHCII-/- or CD4-/- M. tuberculosis-infected mice. However, and different to humanized mice, susceptibility to Tuberculosis increases with decreasing peripheral CD4 T cell counts in HIV infected individuals.
Granuloma initiation during mycobacterial infections has been interpreted as a host-protective event, to provide the microenvironment where specific T cells activate macrophages to contain M. tuberculosis infection. On the contrary, our results suggest that non-protective, dysfunctional T cell responses participate in granuloma formation in humanized mice. Whether such a conclusion can be generalized for the natural course in of infection requires further studies. Supportive of our interpretation, data in a zebra fish model indicated that mycobacteria can exploit the granuloma for local expansion and systemic dissemination.
To our knowledge, this is the first study of mycobacterial infections and description of granuloma formation in humanized mice.
Our data indicate that humanized mouse model can be used to study the role of cytokines, chemokines, different immune cells, host genetics and bacterial components in mycobacterial granuloma formation. We also propose the use of humanized mice as a model for other infectious and non-infectious granulomatous diseases.
Production of recombinant mycobacterial antigens
Strains for the production of the recombinant mycobacterial antigens Acr, HBHA, ESAT6, PstS1, Ag85A, HSP70 and its truncated derivatives are available. Purification and quality control procedures for the antigens Acr, Ag85A, ESAT6, PstS1, HSP70, and its truncated derivatives are established. The purification of recombinant HBHA, PPD, ESAT-6, 16kDa, PstS1, Ag85A and HSP70 was completed during this period. Several of these antigens were distributed to different partners of the project. Natural mycobacterial HBHA was prepared and distributed to different partners. For preparation of HBHA from mycobacterial source, expression vectors for the production of recombinant N- and C-terminal His-tagged M. tuberculosis HBHA in M. smegmatis were constructed, protein was expression and purified, and showed to contain methylated groups.
Ethical Issues
Thus ethical aspects of both human and animal research were considered as indicated in the Technical annex of the project. Human and animal research from this proposal conformed to accepted scientific standards. The research was conducted in accordance with all relevant local or national rules and regulations of the country where the research was carried and subject, as appropriate, to prior authorization of the project by the competent national and institutional research ethics or medico-ethical committees.
Our project involved the use of human tissues such as blood from adults and children. These will be obtained under standard procedures of informed consent from the individuals or legal guardian(s). The procedures are minimally invasive and do not require hospitalisation.
The studies proposed have been approved by the relevant authorities and will be conducted according to the local, national (following the legislation in place in each country) and EC directives. All institutional, national and for all European laboratories EU regulations and guidelines will be followed. The study were done only by trained clinical staff and poses no risk to the health or well being of the participants. All potential participants had first the purpose of the study explained to them and were asked to give written and oral assent to participation. Samples were only be taken after they have indicated that they understand the purpose of the study. In particular, the study will be conducted in accordance with the World Medical Association Declaration of Helsinki (as updated in October 2000) and Directive 2001/20/EC of the European Parliament and of the Council of 4 April 2001 on "The approximation of the laws, regulations and administrative provisions of the Member States relating to the implementation of good clinical practice in the conduct of clinical trials on medicinal products for human use.
Regarding animal research performed at HOMITB, it conformed current legislation and regulations in the countries where the research was carried out. The research was be conducted in accordance with all relevant local or national rules and regulations of the country where the research will be carried and subject, as appropriate, to prior authorization of the project by the competent national and institutional research ethics or medico-ethical committees.
Experimental animals were used in several of the participants’ research facilities, where such experiments were subject to the review of protocols and permission of each insti¬tute’s committee for the use of experimental animals in research. In any case, proper care and humane treatment was always provided.
Potential Impact:
4. The potential impact
Scientific impact
Tightly controlled genetically determined interactions between the mycobacterial stress responses and the host innate and adaptive immune responses to infection dictate the outcome of mycobacterial infection in man ranging from an asymptomatic disease to a disseminated infection. The outcome of HOMITB has been an in depth assessment of molecular and cellular interactions crucial for the design of novel vaccination or immunotherapeutic strategies, aimed at enhancing the efficacy of the innate or adaptive immunity against infection and disease and to the design of better diagnostic markers.
We have studied
1) Host target cell responses regulating the growth of mycobacterial infection
2) Mycobacterial adaptation to different effector host cell responses to infection
3) The role of innate and adaptive immune responses in the control of mycobacterial growth and survival and the regulation of inflammation.
4) Novel human genes responsible for the susceptibility to children and adult TB.
Our main results are
1. An homozygous missense mutation at amino-acid position 417 of JAK2 was found in by genome-wide (GW) homozygosity mapping, a method to find rare recessive mendelian mutations in a child with severe Tuberculosis when screening Chromosome 9 of a sample of 21 families. Preliminary data show that STAT4 activation is diminished in IL-12 activated patient cells.
2. In a sample of 50 children from Iran and Morocco, 2 children showed loss-of-function IL12RB1 alleles resulting in complete IL-12Rβ1 deficiency. This finding may have important medical implications, as recombinant IFN-γ is an effective treatment for mycobacterial infections in IL-12Rβ1-deficient patients. Together with the finding of JAK2 mutants in this relatively small samples, this also provides additional support for the view that severe tuberculosis in childhood may result from a collection of single-gene inborn errors of immunity.
3. We previously reported the first region linked with pulmonary TB, on chromosome 8. Our identification of susceptibility variants in the proximity of TOX represents a major breakthrough in the human genetics of Pulmonary TB, in that this finding has been replicated in Morocco and validated in Madagascar. Moreover, TOX encodes Thymocyte selection-associated high mobility group box protein, and is a key transcription factor leading to CD4 cell development in the murine model.
4. Our data showed that AECII express the molecules required for T cell activation, including HLA class II molecules, which expression is usually restricted to professional antigen presenting cells; secrete cytokines and chemokines critical for the recruitment of leukocytes to the lung, and are able to present antigens, either mycobacterial or viral antigens, to autologous T lymphocytes. We confirmed therefore their role of non-professional antigen-presenting cells. We confirmed that Mtb invades and multiplies within AECII and these accessory presenting cells might participate in the chronic stimulation of anti-HBHA T cells which is observed during Mtb latent infection. AECII might therefore participate to the control of Mtb replication in latently infected subjects.
5. HbhR is a repressor in some mycobacterial species, but not in others, in which it may be even a transcriptional activator. Future work will address the full regulon of HbhR in various mycobacterial species, especially in those in which HbhR is a repressor. Within the frame of HomiTB, we found that in BCG and M. tuberculosis, HhhR appears to regulate only a few additional genes. In the future, the products of these genes will be characterised and their antigenicity will be tested in M. tuberculosis-infected mice and humans. We hypothesize that they may also be latency antigens. The work in HomiTB has led to the finding that hbhA regulation is complex and involves several activators/repressors. Future work will establish the importance of one over the others. Finally, during the HomiTB project, a BCG strain was fortuitously isolated that contained a large deletion in the chromosome, containing genes that are involved in lipid metabolism and genetic regulation. This will also be further investigated.
6. IRF8 is critical for development of dendritic cells. We here show that this transcription factor is required for control of M. tuberculosis infection in mice. IRF8-/- showed higher bacterial levels in lungs, enhanced severity of pathology and diminished recruitment of T cells to pulmonary lesions.
7. B cells, M-CSF induced macrophages, Flt3L-induced dendritic cells and thioglycolate-elicited peritoneal macrophages were purified and then stimulated with a panel of TLRs. Our iresults show that CCR2 and CCR4 were down-regulated by nearly all the TLR ligands, whereas CCR7 was simultaneously up-regulated. The secretion of chemokines was not down-regulated by the TLR agonist. These data highlight a novel role for TLR ligands, such as BLP, in regulating leukocyte retention and migration away from innate immune lesions via discrete constitutive and inflammatory chemokine receptor regulation.
8. We show that the novel cytokine, IL-33, play an important role in bacterial infection. Although its roles in type 2 immunity are well-documented, our data reveals that innate immune cells respond strongly to this cytokine and helped clarify important roles for IL-33. Mycobacterial products mediate their responses via TLRs. Understanding the mechanisms by which cytokines such as IL-33 modulate the functional outcome of TLR signaling through neutrophil activation and migration could lead to novel approaches to the control of systemic as well as local bacterial clearance. This work therefore opens up a new direction of investigation to the mechanism of host response to MTB and may lead to novel approaches to vaccine designs and immune therapy against MTB.
9. The role of SOCS1 during infection with M. tuberculosis was analysed. We propose a model in which M. tuberculosis actively promotes SOCS1 expression by macrophages to counteract effective mycobacterial control at early time points after infection in vivo, before initiation of IFN-γ secretion by NK or T cells. SOCS1 facilitates mycobacterial growth by hampering macrophage IFN-γ secretion in response to infection-induced IL-12. At later time points after infection, despite SOCS1 expression, macrophages respond to IFN-γ secreted by T- o NK cells with unaffected potency, and will not counteract bacterial control to any further extent. Instead SOCS1 in non-macrophage cells protects mice from severe inflammation. The mechanisms behind the anti-inflammatory properties of SOCS1 during M. tuberculosis remain to be studied. We suggest that the efficient stimulation of SOCS1 expression by M. tuberculosis reflects an evolutionary adaptation that may be associated to the ability of M. tuberculosis to establish a chronic infection. We also demonstrated that levels of SOCS1 in blood from patient are associated to disease severity.
10. We have studied the outcome of infection with mycobacteria in humanized mice, in which both human immune cells differentiate de novo from transplanted cord blood progenitor cells. M. bovis BCG-infected humanized mice showed organized granulomas containing a high proportion of human leukocytes, with a core containing human macrophages and giant cells, surrounded by a layer of T cells and a fibrous capsule. Granulomas from M. tuberculosis-infected humanized mice contained widespread necrosis. The formation of these structures depended on human CD4+ but not CD8+ cells, since mice depleted of CD4+ cells after administration of specific monoclonal antibodies showed scarcer, minimal and irregular granulomas. Granuloma formation in humanized mice was segregated from bacterial protection since CD4+ cell activity resulted in a higher susceptibility of humanized mice to mycobacterial infection.
Granuloma initiation during mycobacterial infections has been interpreted as a host-protective event, to provide the microenvironment where specific T cells activate macrophages to contain M. tuberculosis infection. On the contrary, our results suggest that dysfunctional CD4+ T cells participate in granuloma formation in humanized mice and may facilitate bacterial growth. Whether such a conclusion can be generalized for the natural course of infection requires further studies.
Thus, the humanized mouse model can be used to study the role of cytokines, chemokines, different immune cells, host genetics and bacterial components in mycobacterial granuloma formation. We also propose the use of humanized mice as a model for other infectious and non-infectious granulomatous diseases.
Strategic impact
Tuberculosis is one of the leading causes of death in the world. M. tuberculosis was the cause of the “White Plague” of the 17th and 18th centuries in Europe. During this period nearly 100% of the European population was infected with M. tuberculosis, and 25% of all adult deaths were caused by M. tuberculosis. Today, M. tuberculosis is responsible for more morbidity in humans than any other bacterial disease. 1.7 billion people are infected with M. tuberculosis, which amounts to approximately 1/3 of the entire world population. TB is responsible for over 2 million deaths/year. TB has declined in countries of Western and Central Europe, but the burden is still high and increasing in Eastern Europe. This rise is largely attributed to the emergence of HIV infections, which has occurred during the same period. WHO estimated that there were 484,000 new TB cases in Europe in 2001, representing 6% of the global TB burden. Furthermore, patient non-compliance to anti-mycobacterial treatment has given rise to multi-drug resistant strains of M. tuberculosis.
Live BCG is the only currently available vaccine against TB and has been inoculated in about 85% of newborns world-wide. It protects best against miliary TB and TB meningitis in children less than one year of age. However, BCG has failed to confer significant protection in older children and adults against pulmonary TB, the most common and contagious form of TB at that age. Hence, prevention of M. tuberculosis is likely to result in a great saving of societal costs in addition to improving the quality of life and life expectancy of a large part of the population. HOMITB aimed at improving public health within Europe by evaluating novel strategies for the control of TB. The outcome will contribute to the improvement of the design of specific vaccines, to the prevention of severe TB and to a better understanding of the relationships between the invading bacteria and the host immune system.
Medical impact
Uncovering genetic and environmental factors involved either in latency or in disease constitutes a major break-through in the understanding of the host-pathogen interaction. Studies within HOMITB further defined the molecular basis of immunity to M. tuberculosis by dissecting the genetic components that are altered in patients with TB and identifying novel genes that predispose or account for Tuberculosis in man. This will allow developing novel lines of research aimed at restoring impaired immunity. Immunological intervention, whether preventive or curative, will rely on a rational understanding of the pathogenesis of TB in the two important age groups. The design of novel vaccines for paediatric TB, which might be more efficient than BCG, will benefit from such studies as the defective pathways will require to be either complemented or circumvented. The design of a vaccine that might protect adults from re-activation of the mycobacterium is a formidable challenge that will also benefit from exploring the process of immunity to latent mycobacteria by genetic means. Immunological pathways that normally restrict the growth and reactivation of latent mycobacteria are currently unknown and their genetic identification will pave the way for the development of novel vaccines or other immunological interventions. Studies of mechanisms of regulation of mycobacterial gene (hbhA) within infected epithelial cells and macrophages identifying other members of the same regulon may ultimately lead to the discovery of other latency-associated antigens that may be effective immuno-protective and/or diagnostic antigens, and of understanding the mycobacterial adaptation mechanisms during a chronic infection. These studies complementary to the understanding of the host response to gene-targeted studies have focused on the role of molecules that regulate both the inflammatory response and/ or the growth of M. tuberculosis. These molecules affect the activation of the innate and adaptive immune responses, the effector mechanisms in different target cell populations and the stability of the inflammatory lesions. Our results demonstrate that targeting these genes will be important in the control of disease and mycobacterial persistence.
Technological breakthrough
Genomewide scanning for human mutations has been performed in HOMITB resulting in the identification of novel genes that account for the susceptibility of man to Tuberculosis. Genomic technology that generates data on differential expression of both mycobacterial and host genes will be used in analysing the role of different murine or human host target cells, and the role of mycobacterial molecules or immune genes in defining the outcome of infection by using mutant bacteria or transgenic mice respectively. New recombinant molecules will also be used. The humanized mouse model, based on reconstitution of immunodeficient mice with human lymphoid and myeloid cellular populations, will be used to overcome ethical constraint in the research of human immune system. We will monitor mycobacterial load dynamics using a non- invasive real time imaging method in vivo.
Dissemination activities
Dissemination has been pursued achieved through websites, publications, presentations, interviews etc. Work carried out within this project has also been disseminated by publishing in international research journals, for the benefit of the scientific community. A list of publications and presentation in different scientific meetings is imcluded in the next section. We expect a number of manuscripts to be published in the future that are in process of preparation or have already been submitted. A list of these still unpublished manuscripts is also added. Several of these published or unpublished manuscripts are co-authored by two or more participants of HOMITB.
List of Websites:
http://www.mtc.ki.se/HOMITB/index.html