Final Report Summary - PNEUMOPATH (A comprehensive dissection of pneumococcal-host interactions)
Executive Summary:
PNEUMOPATH was a wide-ranging study of host-pathogen interactions during disease due to Streptococcus pneumonia. The interdisciplinary approach was made possible by the PNEUMOPATH consortium, involving academic and industrial partners, gave rise to innovative ideas and results. The project resulted in entirely new insights into the pathogenesis of pneumococcal disease, with implications for diagnosis, vaccination and treatment.
The project was founded on the premise that because pneumococcal isolates vary in their repertoire of genes, the contribution of an individual factor to the infectious process will vary according to the other host and pathogen factors that are present. Previous studies tended to consider the contribution of each virulence factor in single pneumococcal strains or host factors in isolation, but in contrast, in PNEUMOPATH a panel of pneumococcal isolates associated with carriage or disease was used to challenge a range of in vivo and in vitro models.
The project confirmed that clonal types, exhibiting different invasive disease potential, populate pneumococcal serotypes. Furthermore, even intra-clonal variants showed major differences in pneumococcal surface antigens. This observation has implications for the design of protein-based vaccines. A DNA microarray chip covering all of the PNEUMOPATH strains was created and it was used to identify genes associated with invasive disease: results with high diagnostic and treatment implications. It also was found that most S. pneumoniae virulence factor genes are present in S. oralis and S. mitis and that only a few loci have S. pneumoniae-specific virulence genes. This observation is a very significant lesson for those seeking new diagnostics.
PNEUMOPATH also investigated the genetics of susceptibility to invasive pneumococcal disease. Three loci associated with susceptibility were localised on two mouse chromosomes but none of the genes identified in the mouse showed a high association with invasive diseases caused by encapsulated bacteria in humans. In humans, two rare sequence variants disrupting two genes were discovered to have large effects on risk of IPD, when homozygous. These are by far the biggest genetic risk factors of IPD described to date.
An intriguing observation was that a pneumococcal isolate had an entirely different pattern of pathogenesis, with death occurring without sepsis. This isolate appeared to be invisible to the host, with an atypical absence of inflammatory cellular and cytokine inflammatory responses.
A combination of genomic methods identified new factors involved in adherence. This work underlined the importance of nutritional balance during adherence and identified new potential therapeutic targets. Comparison of carbon and nitrogen regulatory networks revealed several overlapping targets. Various regulatory elements were discovered to be important for growth: the global regulator CcpA affected expression of many C and N metabolic genes; two other transcriptional factors regulated arginine acquisition. One regulator was essential for pneumococcal survival and constitutes a new therapeutic target. The importance of nutritional factors in virulence was supported by the observation that better growth in vitro was linked to loss of virulence. These studies reinforce the idea that a pathogen's physiology, a long neglected research area, could be a rich area of new therapeutic targets.
Project Context and Objectives(Project Results):
Streptococcus pneumoniae is a major cause of human disease in all countries of the world. It is the main cause of bacterial pneumonia and sepsis accompanies many cases of pneumococcal pneumonia. In many countries, S. pneumoniae is the leading cause of bacterial meningitis. Issues related to anti-pneumococcal antibiotics and vaccines exacerbate the impact of burden of pneumococcal diseases.
Antibiotic resistance in S. pneumoniae is an increasing global phenomenon. Disturbingly, most penicillin resistant pneumococci also are resistant to cephalosporins. The increasing incidence of antibiotic-resistant pneumococci complicates empirical therapy of community-acquired pneumonia. The increasing prevalence of ??-lactam/macrolide resistance means more reliance on newer agents, such as fluoroquinolones, but resistance to these is emerging The incidence of multidrug resistant strains also is driving more reliance on vancomycin, yet about some pneumococcal isolates already may be vancomycin tolerant.
New pneumococcal conjugate vaccines appear to be successful they are unlikely to be a long-term solution. Although sometimes useful, pneumococcal polysaccharide vaccines are not generally effective, especially in children. The limitations of polysaccharides as vaccines led to the introduction of a polysaccharide-protein conjugate vaccine, which have been a great success, especially in reducing the incidence of pneumococcal diseases in children.
The issues outlined in the preceding paragraphs point to the need for further research. However, almost all previous studies of pneumococci in models of disease had been done with a very limited number of strains, and often only one strain. This is because over the past twenty years study of infection has tended to take a reductionist approach, with the contribution of individual virulence factors or host factors being considered in isolation This may give a distorted picture because pneumococcal isolates have differing potentials to cause diseases. Different pneumococcal isolates have differing propensities to cause invasive disease or to be associated with carriage. This suggests that the mechanisms for evoking pathogenicity differ between strains. It is very well accepted that the outcome of an infection is determined by the interplay of the collection of attributes of the host and of the pathogen. Also, it is known that pneumococcal isolates vary in the repertoire of genes that they possess and hence the contribution of an individual factor to the infectious process may vary according to the other host and pathogen factors that are present.
The overarching hypothesis of the PNEUMOPATH project was that through a systems approach to host-pneumococcal interaction, the most important and consistently involved host and pneumococcal factors would be identified. These could be targets for diagnosis, therapy, enhancement of host defence and prophylaxis. Four individual hypotheses were tested:
1. The carriage rate, invasive disease potential, epidemiology and spread of antibiotic susceptible and resistant strains depend upon the repertoire of genes and the expression of genes for individual virulence factors that will vary between pneumococcal isolates of different clonal types.
2. The influence on the infectious process of individual virulence factors depends on the pneumococcal strain. In other words, the importance of each virulence factor in pathogenesis and transmission will depend on the level of its synthesis and its structure, plus it will depend on the total gene repertoire of the pneumococcal strain and the overall pattern of gene expression.
3. The importance of individual pneumococcal virulence factors and the response of the host to these virulence factors will depend on attributes of the host, particularly the innate susceptibility of the host and the site of infection.
4. The details of key physiological processes, for example competence, biofilm formation, central carbon metabolism and energetics, will vary between pneumococcal strains. Differences in these physiological processes may influence in vivo behaviour.
It was intended that the scientific and technological objectives of the project would be achieved by testing these hypotheses.
3. Scientific and Technical Results / Foregrounds
Workpackage 1 Molecular Epidemiology and Genetic Characterisation of Pneumococcal Isolates'
Workpackage Objectives
Well-characterised, antibiotic susceptible and resistant, pneumococcal isolates, as well as the laboratory strain, TIGR4, were provided for use in other WPs. The genetic content and genetic relationships of clinical pneumococcal isolates were determined with molecular epidemiological tools, microarray technology and sequencing. This comparative genomic study was done with pneumococcal isolates from carriage and disease.
Progress Towards Objectives
All objectives were attained
Provision of strains and genome sequences. Nine pneumococcal strains were selected for use within PNEUMOPATH. These were distributed to the partners.
In collaboration with WP6, genomic sequences of the sequenced strains were assembled and put on the project webpage so that all partners within the project can access the data.
Development of a pneumococcal microarray chip from genome sequences
To be able to perform various microarray experiments with flexible and custom design, Partner 3 installed the Nimblegen platform. Nimblegen has considerable experience with prokaryotic array design, and their experience in prokaryotic genomics and the flexibility of their array formats was the deciding factors in this choice.
Several expression designs have been made during the project. For a single genome expression array (e.g. for S. pneumoniae TIGR4), on average 7-8 probes were designed for each gene, each spotted 3-4 times (4x72k format). For a multiple-genome array, ideally the pneumococcal core genome (i.e. genes common to all strains) is represented by one reference (TIGR4), after which probes for strain-specific transcripts are added for each additional strain. Initial attempts to accomplish this were undertaken using six strains.
Characterisation of the genetic content and genetic relationships of pneumococcal isolates
Partners 4 and 14 have characterised 715 invasive and carriage isolates from children, from the same region (Stockholm) and time period, using pulsed-field-gel-electrophoresis (PFGE) and multi-locus-sequence-typing (MLST). Different serotypes were associated with different invasive disease potential (IDP), with serotypes not covered by the 13-valent pneumococcal vaccine (PCV13) collectively showing a low IDP. Examples of clonal types of the same serotype with different IDPs were found. Serotype 6B isolates of common clonal complex (CC) 138 could be divided into several PFGE patterns, partly explained by number, location and type of temperate bacteriophages.
Partner 7 made an analysis of members of a 23F clone. This was done by comparative genomic hybridisation using a S. pneumoniae R6/TIGR4-specific oligonucleotide microarray, and genomic sequence data obtained from six of these isolates. The numbers indicate a high degree of conservation even between isolates from different continents, and those isolated over 15 years apart.
The microarray chip, described above, was used for the identification of genes associated with pneumococcal invasive disease potential. Strains used in this study were selected from a well-characterised collection of carriage and invasive isolates belonging to serotypes and clones with different invasive disease potential. The identification of genes was done by comparative genomic hybridisation (CGH).
To improve understanding of pneumococcal genetic relationships, Partner 8 conducted four other studies that are summarised below:
1. The impact of the 7-valent pneumococcal conjugate vaccine (PCV7) on the pneumococcal flora was mostly studied without evaluating multiple colonisation and the mechanism(s) leading to serotype replacement. The effect of a single dose of PCV7 on colonisation was studied. A group of children received one PCV7 dose just after nasopharyngeal sampling, with the control receiving no vaccine and both groups being sampled again a month later. Up to ten pneumococcal colonies were recovered per colonised child ' 1,224 isolates were serotyped and representative ones were analysed by pulsed-field gel electrophoresis.
2. A collection of atypical pneumococci isolated from healthy children was characterised. While performing surveillance studies in Oeiras, Portugal, aimed to describe the impact of pneumococcal conjugate vaccine on colonisation, an increase from 0.7% in 2003 to 5% in 2006 was observed in the prevalence of penicillin resistance (MIC 2-6 mg/L) among presumptively identified pneumococci. Although 15 of the 22 penicillin-resistant isolates obtained in 2006 were optochin-resistant, they were bile soluble and thus considered to be bona fide pneumococci. Clarification on the nature of these isolates was obtained by using a combination of phenotypic and genotypic approaches that included routine strategies for pneumococcal identification, multilocus sequence analysis (MLSA), and comparative genomic hybridisation (CGH).
3. As non-capsulated pneumococci (also referred to as non-typeable pneumococci or NTPn) are difficult to identify, as their differentiation from closely related species such as Streptococcus pseudopneumoniae and other streptococcus of the mitis group is not always straight forward, a low cost and easy assay was developed to detect and quantify NTPn in primary samples (which may contain multiple species) obtained from nasopharyngeal swabs.
4. The introduction of the seven-valent pneumococcal conjugate vaccine (PCV7) in Portugal led to extensive serotype replacement among carriers of pneumococci, with a marked decrease of PCV7-types. Although antimicrobial resistance was traditionally associated with PCV7-types, no significant changes were observed in the rates of non-susceptibility to penicillin, resistance to macrolides or multidrug resistance. Investigation of the mechanisms leading to maintenance of antimicrobial resistance despite marked serotype replacement was undertaken. Through molecular typing, 252 antibiotic resistant pneumococci recovered from young carriers in 2006 and 2007 (era of high-PCV7 uptake) were compared with collections of isolates from 2002-2003 (n=374, low-PCV7 uptake era) and 1996-2001 (n=805, pre-PCV7 era).
In Sweden, Partner 4 performed a national surveillance of invasive pneumococcal infections before and after the introduction of the conjugated pneumococcal vaccine in the childhood vaccination program in 2009. All invasive isolates from the clinical microbiological laboratories in Sweden were sent for serotyping and selected have also been characterized using molecular typing, PFGE and MLST. Clinical information about the patients has also been collected through responsible physician. A decreased incidence of invasive pneumococcal disease in Sweden was observed as well as an increase of non-vaccine types for example of serotype 22F. All data are in the process of being analysed and two manuscripts are being written, one on the situation in the largest area, Stockholm, and one comprising all data on a national level.
The CGH data confirmed that all three Streptococcus spp. share much of the accessory genome, which causes a 'smooth transition' between these species, as had been proposed previously. Moreover, most genes known as virulence factors in S. pneumoniae are also commonly found in S. oralis and S. mitis, leaving only a few loci, including the capsule cluster, the hyaluronidase, and some surface proteins as S. pneumoniae specific virulence genes.
Specific to S. oralis was a cluster that encodes a novel export system not known previously in S. pneumoniae, but which has been discussed as virulence factors in other bacteria. The genetic and biochemical analysis of this cluster and its elements has been characterised, suggesting a complex regulatory mechanism of this machinery (Becker et al, manuscript in preparation).
Workpackage 2 Host Susceptibility
Workpackage Objectives
-The overall objective was to identify new host loci conferring innate resistance (or susceptibility) to invasive pneumococcal disease.
-To use the diverse susceptibility of the different strains of mice to pneumonia to map susceptibility QTL (quantitative trait locus) by in silico methods.
-Cross comparison of the mouse pneumococcal susceptibility or resistance QTLs with human pneumococcal susceptibility or resistance loci identified through linkage and/or genome-wide association analysis of human sample sets from invasive pneumococcal disease (IPD).
-Cross species validation of genetic markers, either identified through mouse QTL analysis or linkage and association analysis in the IPD sample set, for identification of novel sequence variants that predispose to pneumococcal susceptibility or resistance.
-Identification of new human pneumococcal susceptibility or resistance genes by analysis and cross species comparison of genome-wide gene expression data derived from blood and adipose tissue of 1000 human individuals and lung tissue of different mouse strains.
-Identification of genetic variation in pathways contributing to susceptibility or resistance from IPD, using tagging SNPs across genes within candidate pathways.
Progress Towards Objectives
All objectives were attained
The susceptibility to infection with S. pneumoniae is associated with host genetic background. It is well described for humans, as well as in mice, that some individuals are less likely to develop invasive pneumococcal disease, even after confirmed exposure. In murine models there are well-described strains displaying permissive or resistant phenotypes. Despite these observations, the genetic factors conferring resistance to pneumococcal infection remain unknown. Recent advances in genetics of humans and mice and increasing availability of dense genotyping data offer a new opportunity to dissect the genetics behind complex diseases.
A panel of 26 genetically defined mouse strains from Jackson Laboratories (Jax) was phenotyped, by Partner 1, for their susceptibility to invasive pneumococcal disease after intranasal infection with serotype 2 pneumococcal strain D39. The phenotyping data were then used to look for genetic association with observed clinical outcomes.
Recently the density of SNP coverage of mouse genomes increased significantly and with this increase far greater sensitivity is obtained in genome-wide association studies. From the panel of 26 Jax-mouse strains, three significant QTLs were identified which strongly associated with resistance to pneumococcal disease. One of these QTLs was located within the Spir1 region previously identified by Denny at al (2003) as conferring resistance to pneumococcal infection in independent linkage studies using CBA/Ca x BALB/c crosses. Within the loci identified in WP2 55, potential candidate genes were located.
The strongest gene candidates were Fas and Ch25h, which play important roles in lung injury and macrophage function. Fas is expressed by neutrophils, macrophages, dendritic and epithelial cells. The myeloid-cells respond to Fas activation primarily by secretion of proinflammatory cytokines while epithelial cells undergo apoptosis. It was also reported that Fas can control Cxcl1 release by lung epithelium in MyD88-dependent manner. During resolution of lung inflammation Fas selectively targets recruited macrophages for apoptosis. The Ch25h is involved in oxysterol 25HC productions.
A polymorphism within the mouse Fas gene was also identified in this study. SNPs that separated resistant mouse strains from susceptible ones were located within Fas coding and regulatory sequence. One of those SNPs resulted in an amino acid change. A similar polymorphism was observed upstream of the Ch25h coding sequence. It was previously reported that located upstream of the Ch25h coding sequence were consensus IFN-?³ activation-site elements and INF-stimulated response elements.
Among genes located within chromosome 7 QTLs, the most interesting gene candidates, with changed expression during infection, were Nfkbib, Map4k1 and Kcnk6.
The results of the analysis allowed selection of a short list of candidate genes that may play an important role in resistance to pneumococcal disease. The claim was supported by results of other studies in mouse and humans that pointed to those proteins or loci as associated with pneumonia in humans (Fas, Nfkbib) or mice (Spir1). The candidate genes selected also have strong links to immune responses in lungs (Fas, Ch25h).
Partner 5 used bioinformatic tools to map twelve genes identified as associating with resistance to pneumococcal infection in the mouse. The coding region of each gene and 1 MB on either site was used to look for association of sequence variants (SNPs and Indels) existing in the database dbSNP, with invasive pneumococcal disease (IPD) and/or invasive disease caused by encapsulated bacteria (S. pneumoniae, N. meningitidis or H. influenzae) (IBD_encaps) in humans. The associations in the mouse were not replicated in humans, as no sequence variant in the genes or within 1 MB flanking regions of the genes in these loci showed genome wide significant or suggestive association with IPD or IBD_encaps in humans. No genome wide significant or suggestive association of non-synonymous variants (causing missense mutations, or amino acid change) with IPD or IBD_encaps was found.
Partner 5 also mapped 146 genes in the ten QTL loci that showed significant or suggestive association with susceptibility to pneumococcal pneumonia in the mouse and identified the location of each gene plus/minus 1MB. Assessment was made of whether sequence variants in dbSNP associated with IPD or IBD_encaps in humans. No sequence variant showed genome wide significant or suggestive association with IPD or IBD_encaps in humans. Two very rare variants associated with IBD_encaps. Both are in the Dcc locus on chr 18, but have very low functional impact as one is downstream of the gene and the other is intergenic. The only missense variant showing some association with IPD in humans was in the ZNF740 gene, encoding a zinc finger protein, and corresponding to the Zfp808 on chr 13 in the mouse. Thus, association of the QTLs associating with susceptibility to pneumonia in the mouse was not replicated in humans.
Workpackage 3 Animal Models
Workpackage Objectives
-Comparative virulence testing in inbred mouse strains permissive to pneumococcal infection.
-Comparative virulence testing in inbred mouse strains resistant to pneumococcal infection.
-Definition of host inflammatory responses to infection with different pneumococcal strains.
-Definition of host genetic responses to infection with different pneumococcal strains.
-Definition of genetic responses of different pneumococcal strains in innately resistant and susceptible mice and in different tissues in these mice.
-Definition of active and passive protection models for pneumococcal infection
-Evaluation in an appropriate animal model of the hypotheses developed from use of ex vivo and in vitro models in WP4 and WP5.
Progress Towards Objectives
All objectives were attained
Comparative virulence testing in inbred mouse strains either permissive or resistant to pneumococcal infection.
The susceptibility to pneumococcal disease is influenced by host genetic background. Numerous studies had showed that some individuals are able to withstand the infection while others show severe clinical signs or even mortality. The host responses to infection were studied with nine pneumococcal strains belonging to five different serotypes and six sequence types. Two mouse strains, resistant (BALB/c) and sensitive (CBA/Ca) to infection with serotype 2 D39 strain were chosen for this study. After infection with nine pneumococcal strains BALB/c mice showed resistance to 7 out of 9 pneumococcal strains. Two of the seven non-virulent strains (both 6B serotypes) were able to enter the blood stream of BALB/c animals but were later cleared. Only two strains: TIGR4 (ST4) and BS71 (ST3) were virulent in BALB/c with rapid onset, high disease severity and the experiment endpoint reached within 28h. Surprisingly this strain had very poor dissemination to the blood and the majority of the bacterial load was contained in the lungs. CBA/Ca animals showed higher susceptibility as compared to BALB/c mice. Five out of nine pneumococcal strains tested (BS71, D39, TIGR4 and both 6B strains: BHN191 and BHN418) caused 70%-100% mortality in CBA/Ca mice.
Intranasal challenge of C57/BL6 mice was performed using ten (plus additionally 2) strains of the PNEUMOPATH project. The comparative intranasal virulence assays revealed statistically significant differences in the pathogenicity between the bacterial strains. There was a significant difference in the overall survival and the bacterial burden in the blood stream. Strains of serotype 14 (one strain), 19F (three strains) were less virulent than strains of type 4 (one strain), 6B (four strains) and 3 (one strain). Two variants of the same strain D39-L and D-39-S showed a big difference in virulence where D39-L was much more virulent than D39-S. Sequence analyses within the project have identified genomic differences between these two strains.
Definition of host inflammatory and genetic responses to infection with different pneumococcal strains.
Pneumococcal infection causes strong immune and inflammatory responses with large number of recruited neutrophils being the hallmark of the infection. In PNEUMOPATH project we investigated host genetic and inflammatory responses to infection with strains from different pneumococcal serotypes. The results showed a complex picture of the interplay between the host and the pathogen. Host whole lungs gene expression, as early as 6h post infection, showed surprisingly uniformed picture with very similar gene regulation in both mouse strains. Unsurprisingly the gene expression strongly reflected lung inflammatory responses when the highest regulated genes were directly involved in pathogen recognition (CD14 among top 5 up-regulated genes) neutrophil recruitment (e.g.: Cxcl1 was the highest up-regulated gene), cytokines and interleukins production (IL4i1, IL1b, CCl4, CCl3 and CCl17 among the top 15 up-regulated genes), regulation of cytokine production (Socs3, among top 10 regulated genes) or apoptosis (up-regulation of Fas and Bcl3). This pattern of gene expression was observed for D39-, BHN191- and LgtSt215-infected BALB/c and CBA/Ca mice but not for BS71-infected mice. The BS71 strain failed to efficiently induce gene expression, with only a couple of genes showed change in expression level as compared to sham-infected control animals. A similar picture was seen when host cytokine level was measured (IL1, IL6, IL10, IL17A, INF and TNF ). At 6h time point, BS71-infected mice had all cytokines unchanged (with the exception of IL1 in CBA/Ca), however 24h post infection IL1, IL6 and TNF reached the highest level as compared to the mice infected with other pneumococcal strains. BS71-infected BALB/c mice also had IL10 and INF unchanged while infection with other strains usually caused further decrease in these cytokines at 24h.
Variation in cytokine levels was observed between animals infected with different pneumococcal strains. However the impact of these differences on animal survival was complex. Only combined analysis of all tested cytokines revealed correlation with observed clinical outcome of the infection. To summarise, our results indicated that changes in IL6, IL10, IL17, INF and TNF but not IL1 significantly correlated with animal survival. IL6 and IL10 seemed to have opposite effect since higher level of IL6 and lower of IL10 at 6h time point and decrease in IL6 with increase in IL10 at 24h correlated with better survival. Decrease in INF and TNF at 24h also seemed beneficial. The Pearson R2 in our model was moderate (r2=0.6 p<0.00001) and perhaps could be improved if more cytokine were included or different set of cytokines.
A very interesting aspect of this study was the unique phenotype of the BS71 serotype 3 strain. Its hypervirulence and its simultaneous failure to induce host responses at an early time post-infection suggests that either this strain adopted some unique hiding strategy or was able to block host signalling pathways. It was interesting to check whether this phenomenon is serotype specific. To do so BS71 was compared with another serotype 3 strain of the same sequence type: BHN35. Surprisingly BHN35 differed significantly from strain BS71. Survival of BS71-infected animals was extremely short (below 28h) despite a very poor dissemination of the bacteria into blood.
Gender differences in susceptibility to pneumococcal disease have been noticed before. In humans this results in a slight predominance in invasive disease in male with respect to females. In PNEUMOPATH, it was demonstrated that this occurs in rodents. The influence of gender on the ability to control systemic infection with S. pneumoniae was investigated in male and female mice, in both pneumonia and sepsis. Males were found to be more susceptible to invasive infection. In sepsis, male mice showed greater weight loss, more marked decrease of body temperature and a significantly higher mortality rate. In pneumonia, there were significant differences in survival, with females able to clear their infection over time.
Using a Bioplex suspension array, the cytokine and chemokine profile was determined during infection. Overall, pneumonia was characterised by a more pronounced increase in cytokines indicative of neutrophil and macrophage involvement, while in sepsis Th1 related cytokines were up-regulated. In both experimental infections male mice exhibited significant increases in several pro-inflammatory cytokines during both sepsis and pneumonia compared to female mice. Increases in IL-6, IL-12(p70), IL-17A, IFNI, RANTES, KC and G-CSF were seen during sepsis, while increases in IL-1I, IL-15, IL-17A, IL-18, TNFI, RANTES, MIP2, KC and GM-CSF during pneumonia.
Definition of genetic responses of different pneumococcal strains in innately resistant and susceptible mice and in different tissues in these mice.
The capacity to clear bacteria in the initial minutes and hours after invasion of the bloodstream has been found to correlate with clearance of the infection. Susceptible and resistant mice have been shown to have drastically different capacities to carry out these early innate events. The early events have been in depth characterised in resistant mice. Analysis of macrophage-depleted mice clearly indicated that macrophage phagocytosis is the main innate mechanism responsible for the initial clearance of pneumococci after intravenous challenge. In order to identify the cellular mechanism and define appropriate in vitro or ex vivo assays to investigate this a methodology was developed to efficiently obtain primary splenic macrophage cells. In FACS analysis these cells presented the markers characteristic of marginal zone macrophages, both when cultured from susceptible and resistant mice. When running phagocytosis assays with these cells, surprisingly both were equally efficient in ingesting non-virulent pneumococci and also showed equally low efficiency in ingestion of virulent pneumococci. Thus, despite the significant difference observed in macrophage phagocytosis in susceptible and resistant mice, once cultured in vitro, the relative primary cells did not show this phenotype. These data highlight the complexity of the innate early response to pneumococci.
Definition of active and passive protection models for pneumococcal infection.
The early events in the host responseto pneumococcal infection were investigated the by following the fate of an intravenous inoculum of pneumococci comprised of roughly equal numbers of three isogenic variants. Sequential analysis of blood samples over the first hours and days indicated that most episodes of invasive infection were monoclonal events. This finding, based on mathematical and statistical analysis, can only be reasonably explained if there is a single bacterial cell bottleneck. This allowed definition of a restricted timeframe when host clearance mechanisms removed all but one organism of the original challenge inoculum. It was shown that macrophages are the major cellular mediators of early clearance of pneumococci, with neutrophils having importance at later stages.
Choline (Cho) is an essential nutrient for growth and virulence of the pneumococcus. One of the main operons involved in Cho metabolism is the essential lic1 operon, which was unambiguously established to be induced in conditions of Cho deprivation in vitro in WP5. To test whether this Cho-based regulation was important in vivo, a mutant in the Cho-regulated promoter, unable to respond to Cho deprivation, was compared with its wildtype parent in a mouse model of nasal colonization and virulence. The ability to respond to Cho deprivation by inducing expression of lic1 was found to be important for nasal colonisation, as the mutant was out-competed by the wild-type strain in a co-infection setup (see also WP5 section).
Neuraminidases are upon the best known pneumococcal virulence factors. We have set up carriage model in different mouse strains and confirmed that intranasal administration of sialic acid increases pneumococcal carriage by a factor of 10 to 1000 depending on the mouse strain. Neuraminidase inhibitors are drugs licensed for the treatment of influenza infection. Here we confirm that neuraminidase inhibitors can be efficiently tested in different mouse strains to evaluate their efficacy for reduction of pneumococcal carriage. The carriage model thus has been confirmed to be robust and suitable to assay for the in vivo effect of carbohydrates on pneumococcal carriage and to be suitable also for the assay of the efficacy of drugs based to be used in treatment of pneumococcal carriage.
Workpackage 4 Cell Culture Models
Workpackage Objectives
The aims of the workpackage were to explore, in a standardised and inter-laboratory fashion, various in vitro cell culture models.
-Identification of the species- and strain-specific characteristics of the pneumococcal isolates with respect to adherence, invasion, transcytosis, cytotoxicity, uptake/killing, cytotoxicity and chemotaxis.
-Identification and characterisation of the species- and strain-specific pneumococcal genes involved in adherence, invasion, transcytosis, cytotoxicity, uptake, cytotoxicity and chemotaxis.
-Characterisation of the species- and strain-specific molecular response of pneumococci during adherence, invasion, transcytosis, cytotoxicity, uptake/killing, cytotoxicity and chemotaxis.
-Characterisation of the molecular response of the host during pneumococcal adherence, invasion, transcytosis, cytotoxicity, uptake/killing, cytotoxicity and chemotaxis.
-Characterisation of the pneumococcus-specific pathways involved in adherence, invasion, transcytosis, cytotoxicity and chemotaxis.
Progress Towards Objectives
All objectives were attained
Phenotypic characterization of pneumococcal isolates in various in vitro cell culture models.
Adherence to epithelial cells.To ensure standardisation between laboratories, inocula used for all in vitro assays were prepared in the same manner by all participants in this work package. Kinetics of adhesion and invasion of epithelial cells by the ten PNEUMOPATH strains was studied using two human cell lines: pharyngeal epithelial Detroit 562 and lung epithelial A549 cells. Clear adherence kinetics were observed using Detroit 562 cells, with highest levels of adherence reached after 4h for all strains. Furthermore, strain-specific adherence phenotypes were identified, but no clear correlation between adherence ability and strain origin (carriage or disease). Highest adherence levels were obtained with the non-encapsulated R6 strain, while of the encapsulated strains, the serotype 6B strains showed the greatest capacity to adhere, and the thickly encapsulated serotype 3 strain (Sp3-BS71) the lowest. Adherence levels of an additional serotype 3 strain, BHN35, were comparable to those of Sp3-BS71. Adherence kinetics of pneumococcal strains using A549 cells were similar to Detroit cells. Invasion of both epithelial cell lines by pneumococcal strains was limited, with only minor differences between strains. Given the known variation across capsular serotypes in their ability to adhere to and colonise epithelial surfaces, the observed phenotypic differences between the strains could, at least partly, be due to their difference in serotypes. To examine this in more detail, capsular switch mutants were constructed, i.e. strains that are isogenic except for the capsular type produced (TrzcisAki et al., 2003, Appl. Environ. Microbiol, 69:7364-7370), in which the serotype 4 capsule locus (cps) was removed and / or replaced by serotype 2, 6B, 14, and 19F cps loci. Adherence kinetics of these switch mutants clearly showed a similar adherence pattern between serotypes as obtained with the wild-type strains. Interestingly, the presence of a foreign cps locus did appear to confer some (energetic) cost, as adherence levels did not increase much after 1-2h.
Adherence, phagocytosis and killing by immune cells. To analyse the interaction between the ten PNEUMOPATH strains and antigen presenting cells, a standard bacterial uptake / killing assay was used with different immune cells. First, human monocytes isolated from buffy coats were differentiated into macrophages for ten days using three different growth factors: Macrophage colony-stimulating factor (M-CSF), Granulocyte / macrophage-CSF (GM-CSF), and Interleukin 3 (IL-3). Results obtained showed enormous variability between different macrophage donors, as well as between different strains. All strains were able to bind to the macrophages, most prominently after differentiation with GM-CSF and IL3. All strains were phagocytosed and showed some survival, with only very low levels for D39, Sp3-BS71, and LgSt215. Second, PMA-differentiated THP-1 cells (a human monocyte leukemia cell line) were used. Low variability during the assay and detectable differences between strains were observed, suggesting a higher reproducibility for differentiated THP-1 cells. Finally, bacterial uptake/killing assay using Raw 264.7 mouse macrophage cells were performed. Statistically significant differences in binding capability, phagocytosis and killing susceptibility of the ten PNEUMOPATH strains was observed.
Phagocytosis by murine spleen macrophages. Spleen macrophages (Sp-M?) were isolated from BALB/c mice, essentially as described before (Alatery et al. 2008, J Immunol Meth), and mature macrophages were obtained after seven days of culture in medium containing M-CSF, confirmed to be positive for CD11b, CD11c, F4/80 and SIGLEC-1 by FACS analysis. Phagocytosis assays using different pneumococcal strains showed that the unencapsulated strain was efficiently phagocytosed in vitro, while phagocytosis of the encapsulated type 2 strain D39 were at very low levels.
Cytotoxicity. The ciliary beat frequency (CBF) of rat ependymal was used to assess toxicity of selected pneumococcal strains essentially as described (Hirst et al. 2000, Infect. Immun. 68:1557-62). All strains, except serotype 3 strain BS71, were able to significantly inhibit the CBF rat epyndemal cells, reaching complete stasis within the first 1-4h of incubation. Surprisingly, the efficiency of cilia beating inhibition did not correlate with bacterial virulence as BS71 was the most virulent of all strains tested (see WP3).
Chemotaxis. The ability of selected pneumococcal strains to attract human CD4 T-lymphocytes isolated from peripheral blood of healthy human donors was assessed as described before (Kadioglu et al. 2004, Infect. Immun. 72: 2689'2697). Chemotaxis of CD4 T cells was only induced by TIGR4 and not by any of the other in vitro grown pneumococci. Interestingly, in vivo grown (mouse-passaged) D39 was much more efficient in activation of CD4 T-cells than any in vitro grown strain, with an apparent effect of the route of infection used for the mouse passage: 50% migration for intraperitoneally injected bacteria compared to 13% after intranasal infection.
Comparison of two D39 strains. Comparison of the D39 strain from Leicester (D39-L) and the D39 strain from Stockholm (D39-S) with regard to uptake/killing by Raw 264.7 macrophages cells and adhesion to A549 and Detroit epithelial cells showed a significantly lower number of macrophage-associated, as well as intracellular killing of D39-L than D39-S, as well as a significantly lower binding ability to A549, but not Detroit, epithelial cells. Furthermore, FACS analysis using anti-type 2 serum or purified antibodies suggested slight differences in the amount of their capsule. No difference in induction of host gene expression was observed between the two D39 strains.
Transcriptional response of pneumococci during adherence
To be able to examine the transcriptional response of pneumococci during adherence, a pan-genomic Nimblegen expression array was designed, covering common and strain-specific genes of all ten strains used within the consortium (see WP1). After 4 hours of adherence to Detroit 562 cells, the expression of adherent (cell-attached) bacteria were compared to the expression of non-adherent (planktonic) bacteria present in the same well to exclude effects on pneumococcal gene expression of secreted factors by the Detroit cells. Array signal were subjected to extensive data filtering to exclude signal due to cross-hybridisation of Detroit RNA (present in the fraction of cell-attached bacteria despite enrichment for microbial RNA) to pneumococcal gene probes.
Identification of pneumococcal genes affecting adherence
To identify (novel) pneumococcal genes essential for adherence, the high-throughput genome-wide negative selection screenings method TnSeq (van Opijnen et al., 2009, Nat. Methods 6:767-772) was employed, where transposon mutant libraries are subjected to a particular challenge, after which next-generation sequencing technology (NGS) is used to identify mutants that are negatively selected from the library population due to the challenge. To be able to perform an adherence TnSeq screen, generation of mutant libraries, and thus efficient transformability of a strain, is of utmost importance.
The genes identified as factors affecting adherence by both genome-wide approaches were distributed over a variety of functional categories, with predominant functional classes of 'Conserved hypothetical and hypothetical proteins' and 'Transport and binding'. KEGG pathway enrichment analysis of identified genes was performed using the web-based DAVID bioinformatics tool (Huang et al., 2009, Nature Protoc. 4:44-57). This showed significant enrichment of the ABC transporters pathway (KEGG pathway spn02010) among the genes whose expression was downregulated during adherence as well as the essential genes identified by TnSeq in R6. Enrichment of the same pathway was also observed in the group of genes with increased expression during adherence and the TIGR4 TnSeq list, but this was not significant in both cases.
Characterisation of the host response during adherence.
The host response after 4h of pneumococcal adherence was examined by real-time PCR, focusing on a selected number of innate immunity genes. Upon data normalisation to GAPDH, clear differences in expression response of Detroit cells to adherent pneumococci was observed, with the highest expression levels observed for IL8, IL6, IL1B, and CXCL2. Expression of IL10 and IL17 was below detection level, while expression of a second housekeeping gene, ACTB, did not differ between the samples. The differences in host response appeared to be somewhat related to capsular type and adherence levels, as the strain with highest adherence generally induced the strongest response. No difference was observed in the response to adherence of two serotype 3 strain.
Workpackage 5 Pneumococcal Physiology
Workpackage Objectives
The main objective of this work package was to document the central metabolism of S. pneumoniae to provide the basic knowledge required to correlate strain-to-strain variations in the adaptation to different host compartments with possible alterations resulting in fine-tuning changes in central metabolic fluxes.
For this goal, the aim was to determine carbon, nitrogen and energy fluxes under defined growth conditions; identify the connection(s) between carbon and nitrogen metabolism, and the interplay between carbon and nitrogen regulators; characterize the connection(s) between central metabolism and expression of the X-state (competence) and of key virulence determinants, e.g. capsule production and utilization of choline (Cho); when possible, evaluate the impact of predefined physiological conditions on expression of candidate strain-specific determinants of colonisation or virulence.
Progress Towards Objectives
All objectives were attained
Carbon, Nitrogen and Energy Fluxes.
All the clinical isolates and model strains retained for the PNEUMOPATH project have been systematically assessed for their growth under defined conditions. Carbon and energy fluxes were calculated from known pathway topography and stoechiometric matrix concepts. No significant correlation was established coupling a specific growth phenotype with virulence in the initial identified strains. The effect on pH evolution was characterised and seen to be a factor influencing carbon and energy fluxes with potential effects on virulence and capsule production.
A clear growth phenotype was observed with various D39 variants (see below). In all cases observed maximum growth rates were retained at a constant value, but for virulent strains, growth arrest occurs at a relatively low biomass concentration, while an avirulent strain continues to grow, reaching biomass levels at least three times higher. Metabolome analysis confirmed a modified metabolic flux distribution, clearly indicating that the metabolic sugar catabolism network is responding to the overall fitness of the strains rather than specific and local regulatory phenomena. Transcriptional analysis shows that during the exponential growth phase in which growth rates and sugar uptake rates are highly conserved in these strains, only a limited number of genes can be seen to be differentially expressed.
Carbon and Nitrogen Metabolism.
Initial physiological analysis suggested that the expression of genes associated with capsule biosynthesis (and the rate of capsule synthesis) and Cho metabolism might be under the control of the carbon catabolite regulator CcpA, a master regulator of carbohydrate metabolism in Gram-positive bacteria. In addition, consensus sites for CcpA binding (cre sites) exist upstream of the cps (coding for capsule synthesis) and lic (coding for Cho utilization) operons in all sequenced S. pneumoniae strains. A ccpA mutant, however, did not show modified expression of these genes. Thus, while capsule synthesis and Cho metabolism do appear to be linked to high rates of sugar conversion, CcpA is not directly involved in this mechanism.
Genomic Analysis of D39 strains.
Three D39 strains have been sequenced and SNPs identified compared to the original D39 sequence available in the databanks. Comparison of the avirulent D39 strain with two virulent strains showed only two common SNPs and approximately 50 SNPs specific to one or more but not all of the three strains. These are potential targets that could explain the differences in virulence phenotypes.
Repertoire of Sugar Transporters and Regulation of Sugar Transport in S. pneumoniae.
A genomic survey allowed establishment the occurrence of 21 phosphotransferase systems, seven carbohydrate uptake ABC transporters, one sodium:solute symporter and a permease, emphasising an exceptionally high capacity for uptake of carbohydrate substrates. Despite high genomic variability, combined phenotypic and genomic analysis of 20 sequenced strains assigned the substrate specificity only to two uptake systems. Systematic analysis of mutants for most carbohydrate transporters enabled a phenotype and substrate specificity to be assigned to 23 transport systems. For five putative transporters for galactose, pentoses, ribonucleosides and sulphated glycans activity was inferred, but not experimentally confirmed and only one transport system remains with an unknown substrate and lack of functional annotation. Using a metabolic approach, 80% of the 32 fermentable carbon substrates were assigned to the corresponding transporter. The complexity and robustness of sugar uptake is underlined by the finding that many transporters have multiple substrates, and many sugars are transported by more than one system.
CelR-I, a transcription factor previously implicated in virulence, was found to function as an activator of a cellobiose-utilisation gene cluster.
Connections between Carbon and Nitrogen Metabolism and Regulators.
Regulation of carbohydrate metabolism.
Regulation of Arginine Metabolism.
The Global Nutritional Regulator CodY.
The finding that CodY is an essential protein in the pathogenic strain D39 and laboratory equivalent R800 was extended to three other pathogenic strains (out of three tested), suggesting that CodY is globally essential in the species, and a potential target for treatment of pneumococcal diseases.
The two-component regulatory system CiaR-CiaH (CiaRH) affects a number of physiological processes such as development of competence, autolysis and A-lactam resistance.
Overall, results obtained in this project reveal many connections between carbon- and nitrogen-dependent transcriptional regulation.
Connection(s) between Central Metabolism and Expression of the X-state (Competence) and of Key Virulence Determinants.
CiaRH, CodY and the Regulation of X-state (Competence).
Metal Homeostasis and Virulence. Besides the documentation of iron homeostasis revealed through analysis of CodY essentiality (see above), it was shown that a number of pneumococcal genes are differentially regulated by copper, including an operon, cop, encoding the CopY regulator (Shafeeq et al., 2011, Mol Microbiol 81: 1255-1270).
Zinc (Zn++) is another important trace metal ion that has been shown to regulate the expression of several (virulence) genes in streptococci.
Workpackage 6 Integrated Bioinformatics and Data Management
Workpackage Objectives
There were three main goals in this work package, namely: deployment of a web-based semantic integration environment (S3DB), development of dedicated data processing services and development and maintenance of an Application Programming Interface (API), which allows project participants to control the sharing and public dissemination of their contributions.
Progress Towards Objectives
All objectives were attained.
One of the main goals of this work package was to develop a semantic web data management system to store the data produced by all partners during the project. Particular attention was given to the user interfaces. Different web based interfaces for data visualisation, retrieval and querying have been developed.
All project participants used the sdLink system, with good results both in scalability and usability by non-expert users. The system has a public project available for evaluation purposes (http:/kdbio.inesc-id.pt/sdlink/sample/) see Figure 3.6.1.
Presently, all data that were produced by the partners are modelled and stored in this system. Excel interfaces have been programmed to represent the database model and to facilitate the data collection and upload to the system. All the project participants can query and retrieve the data.
Figure 3.6.1: Interface to access the projects available in the system
sdLink Data Management System
During the last couple of years new standards and technologies have been developed to support the development of databases using semantic web principles. The development of sdLink was supported by the state-of-the-art technologies available for this type of systems.
The sdLink system was built on top of two main premises. First, it should take a minimal assumption approach with respect to data models, allowing for model extensibility and refinement, without enforcing severe restrictions on expressiveness. Second, it should be scalable, ensuring data availability, integrity and security.
With respect to the first assumption, sdLink follows the concept of linked data, a recommended best practice for exposing, sharing, and connecting pieces of data, information, and knowledge on the Semantic Web using URIs and RDF. Thus, data are represented following an ontology, that can be updated and which allows the definition of inference rules, an important feature for automatic data transformation, model updating and knowledge discovery.
sdLink was then built on top of well-known open technologies and was comprised of three main modules:
1) A management module for projects, users and access control. Authentication relies on the OpenID open standard, providing a decentralized mechanism, and all data about projects, users and access controls (authorization) is stored within the system itself. Figure 3.6.2 presents the new login interface where it is possible to select the OpenID provider.
2) The second module makes available endpoints for data retrieval and querying in several formats, both as REST endpoints and rich Web user interfaces, providing several views on the data. The query endpoint accepts SPARQL, with a user-friendly interface and pre-defined views and queries available. Figure 3.6.3 presents a user interface that is already available and which can be used to visualize the database concepts and the data,.
3) A third module consists of a scalable triple store, including support for transactions and a powerful query engine. This module is built on top of Virtuoso, a high performance triple store. All modules and configuration will be available as pre-configured virtual machine, suitable for both evaluation and production deployments.
Figure 3.6.2: PNEUMOPATH new login interface
Figure 3.6.3: User interface used to visualise the database concepts and the data already available
Excel spreadsheet API
Besides the visualisation interfaces, specific interfaces have been developed to upload new data to the system. As is well-known, biologists are used to collection of the data in Excel spreadsheets and this pre-requisite was determinant to this type of interface development.
During this project, the development of API for spreadsheet client applications has been started and concluded. A new API is available at http://kdbio.inesc-id.pt/~pmreis/Setup.zip and the development site maintained at http://kdbio.inesc-id.pt/mailman/listinfo/sdl-dev.
Figure 3.6.4 shows an Excel workbook including a new tab for the sdLink API. With this new resource the user can select the database concepts and generate a personalized Excel workbook for the data that must be uploaded. When selecting the project PNEUMOPATH ontology, that describes the information system, a new window is opened in order to allow the selection of the concepts to be represented at the Excel workbook. This workbook contains one concept per worksheet and in each worksheet, each column correspond to an attribute connected to the concept. In this semantic web-based system, each concept should be seen as a table and each attribute, a column in that table. Each row of the table will be filled with the values of each attribute.
Figure 3.6.4: Excel workbook with the sdLink API included
Workpackage 7 Management and Dissemination
Workpackage Objectives
-To co-ordinate the project within the consortium
-To co-ordinate contact between the consortium and the EC
-To disseminate, explain and exploit the objectives and results of the project with stakeholders such as the scientific community, public authorities and industry.
Progress Towards Objectives
All the objectives were attained.
During the lifetime of the project nine project consortium meetings were held. All partners participated in, and contributed to, all of the consortium meetings. In addition, meetings between partners took place alongside other conferences and meetings where opportunity arose.
Continuing written and verbal communication, facilitated by Partner 1, took place between the project partners in order to ensure the effective management of the project.
Following agreement from all project partners, an extension to the project from 36 to 42 months was agreed by the Commission. In addition, the re-profiling of the budget to transfer some of the budget of Partner 13 to Partner 1 to enable completion of deliverable DL3.4 and some funds from Partner 5 to Partner 11 was agreed. This re-profiling enhanced the consortium's data analysis capabilities and thus deliverables across all work packages.
Following the approval of the extension, Partners 1 and 8 made minor re-profiling of their budgets in order to appropriately finance the no-cost extension.
The project website, http://www.le.ac.uk/projects/pneumopath was developed in the first year and maintained throughout the lifetime of the project to disseminate information about the project's activities.
Potential Impact:
PNEUMOPATH was a wide-ranging study of host-pathogen interactions during disease due to Streptococcus pneumoniae undertaken in response to the need to increase the knowledge of mechanisms of pathogenesis across a much wider range of pneumococcal strains and host situations than has been done before now. A major impediment previously was that the breadth of technologies and model systems that were necessary to fulfill the objectives of the project were beyond the resources of a single group. The interdisciplinary effort that was made possible by the formation of the PNEUMOPATH consortium, involving academic and industrial partners, gave rise to a large collection of ground-breaking results and ideas.
All of the scientific workpackages produced data with good potential for impact. Studies conducted in WP1 improved our understanding of the evolution and genetic differences of S. pneumoniae and its close relatives. In particular, it may be possible to explore the genetic differences between these species to develop rapid diagnostic tests aimed to unequivocally identify S. pneumoniae in clinical samples. However, the data represent a 'double-edged sword': the data show that a highly specific test is a real possibility but they also provide a warning to scientists and health-care professionals that the closeness of the relationships with other species means that the difficulties in developing a unequivocal diagnostic test are high and that false positive results with the diagnostic targets that have been proposed elsewhere are real possibilities.
WP1 provided a DNA microarray chip covering ten pneumococcal strains. This was a valuable tool within the project that was used to identify genes associated with invasive disease, as well as genes expressed during specific conditions, but this microarray also will be a valuable resource to the wider pneumococcal research community. Within PNEUMOPATH use of the chip provided data of high value for the development of new diagnostic reagents and treatment strategies aimed to contain disease caused by virulent strains.
Within WP2, PNEUMOPATH investigated the genetics of host susceptibility to invasive pneumococcal disease. In the first study of its type applied to an infectious disease, a genome wide association study (GWAS) was undertaken using a large panel of inbred mice to take advantage of the increasing availability of dense genotyping data. The GWAS analysis identified three loci that associated with susceptibility to pneumonia. This type of work is in its infancy but nevertheless the data obtained here will provide a rich seam of new information for those scientists seeking to understand mammalian susceptibility to infection.
The in vivo studies undertaken in WP3 made important discoveries. It was found that cytokine levels could predict the outcome of pneumococcal disease in mice early in infection and were characteristic of host susceptibility. The prognostic significance of these observations should now be tested in humans. The in vivo studies also gave new insights into the process of development of pneumococcal bacteraemia, a condition that accompanies many cases of pneumococcal pneumonia and which is responsible for at least 10% of bacteraemia in Europe. It was ascertained that the action of spleen macrophages during the first minutes of infection determines the clinical outcome and that sustained bacteraemia develops from the replication of a single bacterial cell even when many thousands invade the blood.
A very useful tool for future studies to understand the pathogenesis of pneumonia was the identification of a strain (BS71) that was lethal to mice but was restricted to the lungs and did not cause a bacteraemia. The potential experimental impact of this strain derives from the fact that while much human pneumonia is not accompanied by bacteraemia, all of the animal models of pneumonia are accompanied by bacteraemia. The identification of BS71 changes this situation by giving researchers a tool to study pneumonia in isolation. The knowledge gained and the realisation of the importance of this strain illustrated the impact of the interdisciplinary PNEUMOPATH consortium: work in WP1, 3, 4, 5 and 6 contributed to knowledge of this strain.
In WP4, the interplay between different pneumococcal strains and a variety of host cells in in vitro models were examined, representing different aspects of pneumococcal pathogenesis. This enabled identification of those aspects of the pneumococcus-host interaction that are common between strains and cell types, as well as characteristics specific for certain interactions. The knowledge of these factors, especially those that are conserved between strains, will lead to the identification of new targets for diagnosis, therapy or prevention.
Several bioinformatic advances were made in WP6 that will have applicability beyond the PNEUMOPATH project, for those projects in which the organisation and querying of complex and disparately described data is required. The developments in WP6 included development of dedicated data processing services and an Application Programming Interface and implementation of a semantic web data management system. The systems were designed for use by those unfamiliar with semantic technologies and are currently available for public evaluation.
Data generated during this project were disseminated by presentation at local, national and international conferences and publication in relevant high-impact, peer-reviewed journals. Dissemination of results to the wider scientific community occurred throughout the project, as soon as it was suitable for peer-review publication and after IP considerations.
List of Websites:
http://www2.le.ac.uk/projects/pneumopath
Contact: Professor Peter Andrew, Head of Department
Department of Infection, Immunity and Inflammation
University of Leicester
University Road
Leicester
LE1 9HN
UK
Tel: +44-0116-252-2951
Fax: +44-0116-252-5030
Email: pwa@le.ac.uk
PNEUMOPATH was a wide-ranging study of host-pathogen interactions during disease due to Streptococcus pneumonia. The interdisciplinary approach was made possible by the PNEUMOPATH consortium, involving academic and industrial partners, gave rise to innovative ideas and results. The project resulted in entirely new insights into the pathogenesis of pneumococcal disease, with implications for diagnosis, vaccination and treatment.
The project was founded on the premise that because pneumococcal isolates vary in their repertoire of genes, the contribution of an individual factor to the infectious process will vary according to the other host and pathogen factors that are present. Previous studies tended to consider the contribution of each virulence factor in single pneumococcal strains or host factors in isolation, but in contrast, in PNEUMOPATH a panel of pneumococcal isolates associated with carriage or disease was used to challenge a range of in vivo and in vitro models.
The project confirmed that clonal types, exhibiting different invasive disease potential, populate pneumococcal serotypes. Furthermore, even intra-clonal variants showed major differences in pneumococcal surface antigens. This observation has implications for the design of protein-based vaccines. A DNA microarray chip covering all of the PNEUMOPATH strains was created and it was used to identify genes associated with invasive disease: results with high diagnostic and treatment implications. It also was found that most S. pneumoniae virulence factor genes are present in S. oralis and S. mitis and that only a few loci have S. pneumoniae-specific virulence genes. This observation is a very significant lesson for those seeking new diagnostics.
PNEUMOPATH also investigated the genetics of susceptibility to invasive pneumococcal disease. Three loci associated with susceptibility were localised on two mouse chromosomes but none of the genes identified in the mouse showed a high association with invasive diseases caused by encapsulated bacteria in humans. In humans, two rare sequence variants disrupting two genes were discovered to have large effects on risk of IPD, when homozygous. These are by far the biggest genetic risk factors of IPD described to date.
An intriguing observation was that a pneumococcal isolate had an entirely different pattern of pathogenesis, with death occurring without sepsis. This isolate appeared to be invisible to the host, with an atypical absence of inflammatory cellular and cytokine inflammatory responses.
A combination of genomic methods identified new factors involved in adherence. This work underlined the importance of nutritional balance during adherence and identified new potential therapeutic targets. Comparison of carbon and nitrogen regulatory networks revealed several overlapping targets. Various regulatory elements were discovered to be important for growth: the global regulator CcpA affected expression of many C and N metabolic genes; two other transcriptional factors regulated arginine acquisition. One regulator was essential for pneumococcal survival and constitutes a new therapeutic target. The importance of nutritional factors in virulence was supported by the observation that better growth in vitro was linked to loss of virulence. These studies reinforce the idea that a pathogen's physiology, a long neglected research area, could be a rich area of new therapeutic targets.
Project Context and Objectives(Project Results):
Streptococcus pneumoniae is a major cause of human disease in all countries of the world. It is the main cause of bacterial pneumonia and sepsis accompanies many cases of pneumococcal pneumonia. In many countries, S. pneumoniae is the leading cause of bacterial meningitis. Issues related to anti-pneumococcal antibiotics and vaccines exacerbate the impact of burden of pneumococcal diseases.
Antibiotic resistance in S. pneumoniae is an increasing global phenomenon. Disturbingly, most penicillin resistant pneumococci also are resistant to cephalosporins. The increasing incidence of antibiotic-resistant pneumococci complicates empirical therapy of community-acquired pneumonia. The increasing prevalence of ??-lactam/macrolide resistance means more reliance on newer agents, such as fluoroquinolones, but resistance to these is emerging The incidence of multidrug resistant strains also is driving more reliance on vancomycin, yet about some pneumococcal isolates already may be vancomycin tolerant.
New pneumococcal conjugate vaccines appear to be successful they are unlikely to be a long-term solution. Although sometimes useful, pneumococcal polysaccharide vaccines are not generally effective, especially in children. The limitations of polysaccharides as vaccines led to the introduction of a polysaccharide-protein conjugate vaccine, which have been a great success, especially in reducing the incidence of pneumococcal diseases in children.
The issues outlined in the preceding paragraphs point to the need for further research. However, almost all previous studies of pneumococci in models of disease had been done with a very limited number of strains, and often only one strain. This is because over the past twenty years study of infection has tended to take a reductionist approach, with the contribution of individual virulence factors or host factors being considered in isolation This may give a distorted picture because pneumococcal isolates have differing potentials to cause diseases. Different pneumococcal isolates have differing propensities to cause invasive disease or to be associated with carriage. This suggests that the mechanisms for evoking pathogenicity differ between strains. It is very well accepted that the outcome of an infection is determined by the interplay of the collection of attributes of the host and of the pathogen. Also, it is known that pneumococcal isolates vary in the repertoire of genes that they possess and hence the contribution of an individual factor to the infectious process may vary according to the other host and pathogen factors that are present.
The overarching hypothesis of the PNEUMOPATH project was that through a systems approach to host-pneumococcal interaction, the most important and consistently involved host and pneumococcal factors would be identified. These could be targets for diagnosis, therapy, enhancement of host defence and prophylaxis. Four individual hypotheses were tested:
1. The carriage rate, invasive disease potential, epidemiology and spread of antibiotic susceptible and resistant strains depend upon the repertoire of genes and the expression of genes for individual virulence factors that will vary between pneumococcal isolates of different clonal types.
2. The influence on the infectious process of individual virulence factors depends on the pneumococcal strain. In other words, the importance of each virulence factor in pathogenesis and transmission will depend on the level of its synthesis and its structure, plus it will depend on the total gene repertoire of the pneumococcal strain and the overall pattern of gene expression.
3. The importance of individual pneumococcal virulence factors and the response of the host to these virulence factors will depend on attributes of the host, particularly the innate susceptibility of the host and the site of infection.
4. The details of key physiological processes, for example competence, biofilm formation, central carbon metabolism and energetics, will vary between pneumococcal strains. Differences in these physiological processes may influence in vivo behaviour.
It was intended that the scientific and technological objectives of the project would be achieved by testing these hypotheses.
3. Scientific and Technical Results / Foregrounds
Workpackage 1 Molecular Epidemiology and Genetic Characterisation of Pneumococcal Isolates'
Workpackage Objectives
Well-characterised, antibiotic susceptible and resistant, pneumococcal isolates, as well as the laboratory strain, TIGR4, were provided for use in other WPs. The genetic content and genetic relationships of clinical pneumococcal isolates were determined with molecular epidemiological tools, microarray technology and sequencing. This comparative genomic study was done with pneumococcal isolates from carriage and disease.
Progress Towards Objectives
All objectives were attained
Provision of strains and genome sequences. Nine pneumococcal strains were selected for use within PNEUMOPATH. These were distributed to the partners.
In collaboration with WP6, genomic sequences of the sequenced strains were assembled and put on the project webpage so that all partners within the project can access the data.
Development of a pneumococcal microarray chip from genome sequences
To be able to perform various microarray experiments with flexible and custom design, Partner 3 installed the Nimblegen platform. Nimblegen has considerable experience with prokaryotic array design, and their experience in prokaryotic genomics and the flexibility of their array formats was the deciding factors in this choice.
Several expression designs have been made during the project. For a single genome expression array (e.g. for S. pneumoniae TIGR4), on average 7-8 probes were designed for each gene, each spotted 3-4 times (4x72k format). For a multiple-genome array, ideally the pneumococcal core genome (i.e. genes common to all strains) is represented by one reference (TIGR4), after which probes for strain-specific transcripts are added for each additional strain. Initial attempts to accomplish this were undertaken using six strains.
Characterisation of the genetic content and genetic relationships of pneumococcal isolates
Partners 4 and 14 have characterised 715 invasive and carriage isolates from children, from the same region (Stockholm) and time period, using pulsed-field-gel-electrophoresis (PFGE) and multi-locus-sequence-typing (MLST). Different serotypes were associated with different invasive disease potential (IDP), with serotypes not covered by the 13-valent pneumococcal vaccine (PCV13) collectively showing a low IDP. Examples of clonal types of the same serotype with different IDPs were found. Serotype 6B isolates of common clonal complex (CC) 138 could be divided into several PFGE patterns, partly explained by number, location and type of temperate bacteriophages.
Partner 7 made an analysis of members of a 23F clone. This was done by comparative genomic hybridisation using a S. pneumoniae R6/TIGR4-specific oligonucleotide microarray, and genomic sequence data obtained from six of these isolates. The numbers indicate a high degree of conservation even between isolates from different continents, and those isolated over 15 years apart.
The microarray chip, described above, was used for the identification of genes associated with pneumococcal invasive disease potential. Strains used in this study were selected from a well-characterised collection of carriage and invasive isolates belonging to serotypes and clones with different invasive disease potential. The identification of genes was done by comparative genomic hybridisation (CGH).
To improve understanding of pneumococcal genetic relationships, Partner 8 conducted four other studies that are summarised below:
1. The impact of the 7-valent pneumococcal conjugate vaccine (PCV7) on the pneumococcal flora was mostly studied without evaluating multiple colonisation and the mechanism(s) leading to serotype replacement. The effect of a single dose of PCV7 on colonisation was studied. A group of children received one PCV7 dose just after nasopharyngeal sampling, with the control receiving no vaccine and both groups being sampled again a month later. Up to ten pneumococcal colonies were recovered per colonised child ' 1,224 isolates were serotyped and representative ones were analysed by pulsed-field gel electrophoresis.
2. A collection of atypical pneumococci isolated from healthy children was characterised. While performing surveillance studies in Oeiras, Portugal, aimed to describe the impact of pneumococcal conjugate vaccine on colonisation, an increase from 0.7% in 2003 to 5% in 2006 was observed in the prevalence of penicillin resistance (MIC 2-6 mg/L) among presumptively identified pneumococci. Although 15 of the 22 penicillin-resistant isolates obtained in 2006 were optochin-resistant, they were bile soluble and thus considered to be bona fide pneumococci. Clarification on the nature of these isolates was obtained by using a combination of phenotypic and genotypic approaches that included routine strategies for pneumococcal identification, multilocus sequence analysis (MLSA), and comparative genomic hybridisation (CGH).
3. As non-capsulated pneumococci (also referred to as non-typeable pneumococci or NTPn) are difficult to identify, as their differentiation from closely related species such as Streptococcus pseudopneumoniae and other streptococcus of the mitis group is not always straight forward, a low cost and easy assay was developed to detect and quantify NTPn in primary samples (which may contain multiple species) obtained from nasopharyngeal swabs.
4. The introduction of the seven-valent pneumococcal conjugate vaccine (PCV7) in Portugal led to extensive serotype replacement among carriers of pneumococci, with a marked decrease of PCV7-types. Although antimicrobial resistance was traditionally associated with PCV7-types, no significant changes were observed in the rates of non-susceptibility to penicillin, resistance to macrolides or multidrug resistance. Investigation of the mechanisms leading to maintenance of antimicrobial resistance despite marked serotype replacement was undertaken. Through molecular typing, 252 antibiotic resistant pneumococci recovered from young carriers in 2006 and 2007 (era of high-PCV7 uptake) were compared with collections of isolates from 2002-2003 (n=374, low-PCV7 uptake era) and 1996-2001 (n=805, pre-PCV7 era).
In Sweden, Partner 4 performed a national surveillance of invasive pneumococcal infections before and after the introduction of the conjugated pneumococcal vaccine in the childhood vaccination program in 2009. All invasive isolates from the clinical microbiological laboratories in Sweden were sent for serotyping and selected have also been characterized using molecular typing, PFGE and MLST. Clinical information about the patients has also been collected through responsible physician. A decreased incidence of invasive pneumococcal disease in Sweden was observed as well as an increase of non-vaccine types for example of serotype 22F. All data are in the process of being analysed and two manuscripts are being written, one on the situation in the largest area, Stockholm, and one comprising all data on a national level.
The CGH data confirmed that all three Streptococcus spp. share much of the accessory genome, which causes a 'smooth transition' between these species, as had been proposed previously. Moreover, most genes known as virulence factors in S. pneumoniae are also commonly found in S. oralis and S. mitis, leaving only a few loci, including the capsule cluster, the hyaluronidase, and some surface proteins as S. pneumoniae specific virulence genes.
Specific to S. oralis was a cluster that encodes a novel export system not known previously in S. pneumoniae, but which has been discussed as virulence factors in other bacteria. The genetic and biochemical analysis of this cluster and its elements has been characterised, suggesting a complex regulatory mechanism of this machinery (Becker et al, manuscript in preparation).
Workpackage 2 Host Susceptibility
Workpackage Objectives
-The overall objective was to identify new host loci conferring innate resistance (or susceptibility) to invasive pneumococcal disease.
-To use the diverse susceptibility of the different strains of mice to pneumonia to map susceptibility QTL (quantitative trait locus) by in silico methods.
-Cross comparison of the mouse pneumococcal susceptibility or resistance QTLs with human pneumococcal susceptibility or resistance loci identified through linkage and/or genome-wide association analysis of human sample sets from invasive pneumococcal disease (IPD).
-Cross species validation of genetic markers, either identified through mouse QTL analysis or linkage and association analysis in the IPD sample set, for identification of novel sequence variants that predispose to pneumococcal susceptibility or resistance.
-Identification of new human pneumococcal susceptibility or resistance genes by analysis and cross species comparison of genome-wide gene expression data derived from blood and adipose tissue of 1000 human individuals and lung tissue of different mouse strains.
-Identification of genetic variation in pathways contributing to susceptibility or resistance from IPD, using tagging SNPs across genes within candidate pathways.
Progress Towards Objectives
All objectives were attained
The susceptibility to infection with S. pneumoniae is associated with host genetic background. It is well described for humans, as well as in mice, that some individuals are less likely to develop invasive pneumococcal disease, even after confirmed exposure. In murine models there are well-described strains displaying permissive or resistant phenotypes. Despite these observations, the genetic factors conferring resistance to pneumococcal infection remain unknown. Recent advances in genetics of humans and mice and increasing availability of dense genotyping data offer a new opportunity to dissect the genetics behind complex diseases.
A panel of 26 genetically defined mouse strains from Jackson Laboratories (Jax) was phenotyped, by Partner 1, for their susceptibility to invasive pneumococcal disease after intranasal infection with serotype 2 pneumococcal strain D39. The phenotyping data were then used to look for genetic association with observed clinical outcomes.
Recently the density of SNP coverage of mouse genomes increased significantly and with this increase far greater sensitivity is obtained in genome-wide association studies. From the panel of 26 Jax-mouse strains, three significant QTLs were identified which strongly associated with resistance to pneumococcal disease. One of these QTLs was located within the Spir1 region previously identified by Denny at al (2003) as conferring resistance to pneumococcal infection in independent linkage studies using CBA/Ca x BALB/c crosses. Within the loci identified in WP2 55, potential candidate genes were located.
The strongest gene candidates were Fas and Ch25h, which play important roles in lung injury and macrophage function. Fas is expressed by neutrophils, macrophages, dendritic and epithelial cells. The myeloid-cells respond to Fas activation primarily by secretion of proinflammatory cytokines while epithelial cells undergo apoptosis. It was also reported that Fas can control Cxcl1 release by lung epithelium in MyD88-dependent manner. During resolution of lung inflammation Fas selectively targets recruited macrophages for apoptosis. The Ch25h is involved in oxysterol 25HC productions.
A polymorphism within the mouse Fas gene was also identified in this study. SNPs that separated resistant mouse strains from susceptible ones were located within Fas coding and regulatory sequence. One of those SNPs resulted in an amino acid change. A similar polymorphism was observed upstream of the Ch25h coding sequence. It was previously reported that located upstream of the Ch25h coding sequence were consensus IFN-?³ activation-site elements and INF-stimulated response elements.
Among genes located within chromosome 7 QTLs, the most interesting gene candidates, with changed expression during infection, were Nfkbib, Map4k1 and Kcnk6.
The results of the analysis allowed selection of a short list of candidate genes that may play an important role in resistance to pneumococcal disease. The claim was supported by results of other studies in mouse and humans that pointed to those proteins or loci as associated with pneumonia in humans (Fas, Nfkbib) or mice (Spir1). The candidate genes selected also have strong links to immune responses in lungs (Fas, Ch25h).
Partner 5 used bioinformatic tools to map twelve genes identified as associating with resistance to pneumococcal infection in the mouse. The coding region of each gene and 1 MB on either site was used to look for association of sequence variants (SNPs and Indels) existing in the database dbSNP, with invasive pneumococcal disease (IPD) and/or invasive disease caused by encapsulated bacteria (S. pneumoniae, N. meningitidis or H. influenzae) (IBD_encaps) in humans. The associations in the mouse were not replicated in humans, as no sequence variant in the genes or within 1 MB flanking regions of the genes in these loci showed genome wide significant or suggestive association with IPD or IBD_encaps in humans. No genome wide significant or suggestive association of non-synonymous variants (causing missense mutations, or amino acid change) with IPD or IBD_encaps was found.
Partner 5 also mapped 146 genes in the ten QTL loci that showed significant or suggestive association with susceptibility to pneumococcal pneumonia in the mouse and identified the location of each gene plus/minus 1MB. Assessment was made of whether sequence variants in dbSNP associated with IPD or IBD_encaps in humans. No sequence variant showed genome wide significant or suggestive association with IPD or IBD_encaps in humans. Two very rare variants associated with IBD_encaps. Both are in the Dcc locus on chr 18, but have very low functional impact as one is downstream of the gene and the other is intergenic. The only missense variant showing some association with IPD in humans was in the ZNF740 gene, encoding a zinc finger protein, and corresponding to the Zfp808 on chr 13 in the mouse. Thus, association of the QTLs associating with susceptibility to pneumonia in the mouse was not replicated in humans.
Workpackage 3 Animal Models
Workpackage Objectives
-Comparative virulence testing in inbred mouse strains permissive to pneumococcal infection.
-Comparative virulence testing in inbred mouse strains resistant to pneumococcal infection.
-Definition of host inflammatory responses to infection with different pneumococcal strains.
-Definition of host genetic responses to infection with different pneumococcal strains.
-Definition of genetic responses of different pneumococcal strains in innately resistant and susceptible mice and in different tissues in these mice.
-Definition of active and passive protection models for pneumococcal infection
-Evaluation in an appropriate animal model of the hypotheses developed from use of ex vivo and in vitro models in WP4 and WP5.
Progress Towards Objectives
All objectives were attained
Comparative virulence testing in inbred mouse strains either permissive or resistant to pneumococcal infection.
The susceptibility to pneumococcal disease is influenced by host genetic background. Numerous studies had showed that some individuals are able to withstand the infection while others show severe clinical signs or even mortality. The host responses to infection were studied with nine pneumococcal strains belonging to five different serotypes and six sequence types. Two mouse strains, resistant (BALB/c) and sensitive (CBA/Ca) to infection with serotype 2 D39 strain were chosen for this study. After infection with nine pneumococcal strains BALB/c mice showed resistance to 7 out of 9 pneumococcal strains. Two of the seven non-virulent strains (both 6B serotypes) were able to enter the blood stream of BALB/c animals but were later cleared. Only two strains: TIGR4 (ST4) and BS71 (ST3) were virulent in BALB/c with rapid onset, high disease severity and the experiment endpoint reached within 28h. Surprisingly this strain had very poor dissemination to the blood and the majority of the bacterial load was contained in the lungs. CBA/Ca animals showed higher susceptibility as compared to BALB/c mice. Five out of nine pneumococcal strains tested (BS71, D39, TIGR4 and both 6B strains: BHN191 and BHN418) caused 70%-100% mortality in CBA/Ca mice.
Intranasal challenge of C57/BL6 mice was performed using ten (plus additionally 2) strains of the PNEUMOPATH project. The comparative intranasal virulence assays revealed statistically significant differences in the pathogenicity between the bacterial strains. There was a significant difference in the overall survival and the bacterial burden in the blood stream. Strains of serotype 14 (one strain), 19F (three strains) were less virulent than strains of type 4 (one strain), 6B (four strains) and 3 (one strain). Two variants of the same strain D39-L and D-39-S showed a big difference in virulence where D39-L was much more virulent than D39-S. Sequence analyses within the project have identified genomic differences between these two strains.
Definition of host inflammatory and genetic responses to infection with different pneumococcal strains.
Pneumococcal infection causes strong immune and inflammatory responses with large number of recruited neutrophils being the hallmark of the infection. In PNEUMOPATH project we investigated host genetic and inflammatory responses to infection with strains from different pneumococcal serotypes. The results showed a complex picture of the interplay between the host and the pathogen. Host whole lungs gene expression, as early as 6h post infection, showed surprisingly uniformed picture with very similar gene regulation in both mouse strains. Unsurprisingly the gene expression strongly reflected lung inflammatory responses when the highest regulated genes were directly involved in pathogen recognition (CD14 among top 5 up-regulated genes) neutrophil recruitment (e.g.: Cxcl1 was the highest up-regulated gene), cytokines and interleukins production (IL4i1, IL1b, CCl4, CCl3 and CCl17 among the top 15 up-regulated genes), regulation of cytokine production (Socs3, among top 10 regulated genes) or apoptosis (up-regulation of Fas and Bcl3). This pattern of gene expression was observed for D39-, BHN191- and LgtSt215-infected BALB/c and CBA/Ca mice but not for BS71-infected mice. The BS71 strain failed to efficiently induce gene expression, with only a couple of genes showed change in expression level as compared to sham-infected control animals. A similar picture was seen when host cytokine level was measured (IL1, IL6, IL10, IL17A, INF and TNF ). At 6h time point, BS71-infected mice had all cytokines unchanged (with the exception of IL1 in CBA/Ca), however 24h post infection IL1, IL6 and TNF reached the highest level as compared to the mice infected with other pneumococcal strains. BS71-infected BALB/c mice also had IL10 and INF unchanged while infection with other strains usually caused further decrease in these cytokines at 24h.
Variation in cytokine levels was observed between animals infected with different pneumococcal strains. However the impact of these differences on animal survival was complex. Only combined analysis of all tested cytokines revealed correlation with observed clinical outcome of the infection. To summarise, our results indicated that changes in IL6, IL10, IL17, INF and TNF but not IL1 significantly correlated with animal survival. IL6 and IL10 seemed to have opposite effect since higher level of IL6 and lower of IL10 at 6h time point and decrease in IL6 with increase in IL10 at 24h correlated with better survival. Decrease in INF and TNF at 24h also seemed beneficial. The Pearson R2 in our model was moderate (r2=0.6 p<0.00001) and perhaps could be improved if more cytokine were included or different set of cytokines.
A very interesting aspect of this study was the unique phenotype of the BS71 serotype 3 strain. Its hypervirulence and its simultaneous failure to induce host responses at an early time post-infection suggests that either this strain adopted some unique hiding strategy or was able to block host signalling pathways. It was interesting to check whether this phenomenon is serotype specific. To do so BS71 was compared with another serotype 3 strain of the same sequence type: BHN35. Surprisingly BHN35 differed significantly from strain BS71. Survival of BS71-infected animals was extremely short (below 28h) despite a very poor dissemination of the bacteria into blood.
Gender differences in susceptibility to pneumococcal disease have been noticed before. In humans this results in a slight predominance in invasive disease in male with respect to females. In PNEUMOPATH, it was demonstrated that this occurs in rodents. The influence of gender on the ability to control systemic infection with S. pneumoniae was investigated in male and female mice, in both pneumonia and sepsis. Males were found to be more susceptible to invasive infection. In sepsis, male mice showed greater weight loss, more marked decrease of body temperature and a significantly higher mortality rate. In pneumonia, there were significant differences in survival, with females able to clear their infection over time.
Using a Bioplex suspension array, the cytokine and chemokine profile was determined during infection. Overall, pneumonia was characterised by a more pronounced increase in cytokines indicative of neutrophil and macrophage involvement, while in sepsis Th1 related cytokines were up-regulated. In both experimental infections male mice exhibited significant increases in several pro-inflammatory cytokines during both sepsis and pneumonia compared to female mice. Increases in IL-6, IL-12(p70), IL-17A, IFNI, RANTES, KC and G-CSF were seen during sepsis, while increases in IL-1I, IL-15, IL-17A, IL-18, TNFI, RANTES, MIP2, KC and GM-CSF during pneumonia.
Definition of genetic responses of different pneumococcal strains in innately resistant and susceptible mice and in different tissues in these mice.
The capacity to clear bacteria in the initial minutes and hours after invasion of the bloodstream has been found to correlate with clearance of the infection. Susceptible and resistant mice have been shown to have drastically different capacities to carry out these early innate events. The early events have been in depth characterised in resistant mice. Analysis of macrophage-depleted mice clearly indicated that macrophage phagocytosis is the main innate mechanism responsible for the initial clearance of pneumococci after intravenous challenge. In order to identify the cellular mechanism and define appropriate in vitro or ex vivo assays to investigate this a methodology was developed to efficiently obtain primary splenic macrophage cells. In FACS analysis these cells presented the markers characteristic of marginal zone macrophages, both when cultured from susceptible and resistant mice. When running phagocytosis assays with these cells, surprisingly both were equally efficient in ingesting non-virulent pneumococci and also showed equally low efficiency in ingestion of virulent pneumococci. Thus, despite the significant difference observed in macrophage phagocytosis in susceptible and resistant mice, once cultured in vitro, the relative primary cells did not show this phenotype. These data highlight the complexity of the innate early response to pneumococci.
Definition of active and passive protection models for pneumococcal infection.
The early events in the host responseto pneumococcal infection were investigated the by following the fate of an intravenous inoculum of pneumococci comprised of roughly equal numbers of three isogenic variants. Sequential analysis of blood samples over the first hours and days indicated that most episodes of invasive infection were monoclonal events. This finding, based on mathematical and statistical analysis, can only be reasonably explained if there is a single bacterial cell bottleneck. This allowed definition of a restricted timeframe when host clearance mechanisms removed all but one organism of the original challenge inoculum. It was shown that macrophages are the major cellular mediators of early clearance of pneumococci, with neutrophils having importance at later stages.
Choline (Cho) is an essential nutrient for growth and virulence of the pneumococcus. One of the main operons involved in Cho metabolism is the essential lic1 operon, which was unambiguously established to be induced in conditions of Cho deprivation in vitro in WP5. To test whether this Cho-based regulation was important in vivo, a mutant in the Cho-regulated promoter, unable to respond to Cho deprivation, was compared with its wildtype parent in a mouse model of nasal colonization and virulence. The ability to respond to Cho deprivation by inducing expression of lic1 was found to be important for nasal colonisation, as the mutant was out-competed by the wild-type strain in a co-infection setup (see also WP5 section).
Neuraminidases are upon the best known pneumococcal virulence factors. We have set up carriage model in different mouse strains and confirmed that intranasal administration of sialic acid increases pneumococcal carriage by a factor of 10 to 1000 depending on the mouse strain. Neuraminidase inhibitors are drugs licensed for the treatment of influenza infection. Here we confirm that neuraminidase inhibitors can be efficiently tested in different mouse strains to evaluate their efficacy for reduction of pneumococcal carriage. The carriage model thus has been confirmed to be robust and suitable to assay for the in vivo effect of carbohydrates on pneumococcal carriage and to be suitable also for the assay of the efficacy of drugs based to be used in treatment of pneumococcal carriage.
Workpackage 4 Cell Culture Models
Workpackage Objectives
The aims of the workpackage were to explore, in a standardised and inter-laboratory fashion, various in vitro cell culture models.
-Identification of the species- and strain-specific characteristics of the pneumococcal isolates with respect to adherence, invasion, transcytosis, cytotoxicity, uptake/killing, cytotoxicity and chemotaxis.
-Identification and characterisation of the species- and strain-specific pneumococcal genes involved in adherence, invasion, transcytosis, cytotoxicity, uptake, cytotoxicity and chemotaxis.
-Characterisation of the species- and strain-specific molecular response of pneumococci during adherence, invasion, transcytosis, cytotoxicity, uptake/killing, cytotoxicity and chemotaxis.
-Characterisation of the molecular response of the host during pneumococcal adherence, invasion, transcytosis, cytotoxicity, uptake/killing, cytotoxicity and chemotaxis.
-Characterisation of the pneumococcus-specific pathways involved in adherence, invasion, transcytosis, cytotoxicity and chemotaxis.
Progress Towards Objectives
All objectives were attained
Phenotypic characterization of pneumococcal isolates in various in vitro cell culture models.
Adherence to epithelial cells.To ensure standardisation between laboratories, inocula used for all in vitro assays were prepared in the same manner by all participants in this work package. Kinetics of adhesion and invasion of epithelial cells by the ten PNEUMOPATH strains was studied using two human cell lines: pharyngeal epithelial Detroit 562 and lung epithelial A549 cells. Clear adherence kinetics were observed using Detroit 562 cells, with highest levels of adherence reached after 4h for all strains. Furthermore, strain-specific adherence phenotypes were identified, but no clear correlation between adherence ability and strain origin (carriage or disease). Highest adherence levels were obtained with the non-encapsulated R6 strain, while of the encapsulated strains, the serotype 6B strains showed the greatest capacity to adhere, and the thickly encapsulated serotype 3 strain (Sp3-BS71) the lowest. Adherence levels of an additional serotype 3 strain, BHN35, were comparable to those of Sp3-BS71. Adherence kinetics of pneumococcal strains using A549 cells were similar to Detroit cells. Invasion of both epithelial cell lines by pneumococcal strains was limited, with only minor differences between strains. Given the known variation across capsular serotypes in their ability to adhere to and colonise epithelial surfaces, the observed phenotypic differences between the strains could, at least partly, be due to their difference in serotypes. To examine this in more detail, capsular switch mutants were constructed, i.e. strains that are isogenic except for the capsular type produced (TrzcisAki et al., 2003, Appl. Environ. Microbiol, 69:7364-7370), in which the serotype 4 capsule locus (cps) was removed and / or replaced by serotype 2, 6B, 14, and 19F cps loci. Adherence kinetics of these switch mutants clearly showed a similar adherence pattern between serotypes as obtained with the wild-type strains. Interestingly, the presence of a foreign cps locus did appear to confer some (energetic) cost, as adherence levels did not increase much after 1-2h.
Adherence, phagocytosis and killing by immune cells. To analyse the interaction between the ten PNEUMOPATH strains and antigen presenting cells, a standard bacterial uptake / killing assay was used with different immune cells. First, human monocytes isolated from buffy coats were differentiated into macrophages for ten days using three different growth factors: Macrophage colony-stimulating factor (M-CSF), Granulocyte / macrophage-CSF (GM-CSF), and Interleukin 3 (IL-3). Results obtained showed enormous variability between different macrophage donors, as well as between different strains. All strains were able to bind to the macrophages, most prominently after differentiation with GM-CSF and IL3. All strains were phagocytosed and showed some survival, with only very low levels for D39, Sp3-BS71, and LgSt215. Second, PMA-differentiated THP-1 cells (a human monocyte leukemia cell line) were used. Low variability during the assay and detectable differences between strains were observed, suggesting a higher reproducibility for differentiated THP-1 cells. Finally, bacterial uptake/killing assay using Raw 264.7 mouse macrophage cells were performed. Statistically significant differences in binding capability, phagocytosis and killing susceptibility of the ten PNEUMOPATH strains was observed.
Phagocytosis by murine spleen macrophages. Spleen macrophages (Sp-M?) were isolated from BALB/c mice, essentially as described before (Alatery et al. 2008, J Immunol Meth), and mature macrophages were obtained after seven days of culture in medium containing M-CSF, confirmed to be positive for CD11b, CD11c, F4/80 and SIGLEC-1 by FACS analysis. Phagocytosis assays using different pneumococcal strains showed that the unencapsulated strain was efficiently phagocytosed in vitro, while phagocytosis of the encapsulated type 2 strain D39 were at very low levels.
Cytotoxicity. The ciliary beat frequency (CBF) of rat ependymal was used to assess toxicity of selected pneumococcal strains essentially as described (Hirst et al. 2000, Infect. Immun. 68:1557-62). All strains, except serotype 3 strain BS71, were able to significantly inhibit the CBF rat epyndemal cells, reaching complete stasis within the first 1-4h of incubation. Surprisingly, the efficiency of cilia beating inhibition did not correlate with bacterial virulence as BS71 was the most virulent of all strains tested (see WP3).
Chemotaxis. The ability of selected pneumococcal strains to attract human CD4 T-lymphocytes isolated from peripheral blood of healthy human donors was assessed as described before (Kadioglu et al. 2004, Infect. Immun. 72: 2689'2697). Chemotaxis of CD4 T cells was only induced by TIGR4 and not by any of the other in vitro grown pneumococci. Interestingly, in vivo grown (mouse-passaged) D39 was much more efficient in activation of CD4 T-cells than any in vitro grown strain, with an apparent effect of the route of infection used for the mouse passage: 50% migration for intraperitoneally injected bacteria compared to 13% after intranasal infection.
Comparison of two D39 strains. Comparison of the D39 strain from Leicester (D39-L) and the D39 strain from Stockholm (D39-S) with regard to uptake/killing by Raw 264.7 macrophages cells and adhesion to A549 and Detroit epithelial cells showed a significantly lower number of macrophage-associated, as well as intracellular killing of D39-L than D39-S, as well as a significantly lower binding ability to A549, but not Detroit, epithelial cells. Furthermore, FACS analysis using anti-type 2 serum or purified antibodies suggested slight differences in the amount of their capsule. No difference in induction of host gene expression was observed between the two D39 strains.
Transcriptional response of pneumococci during adherence
To be able to examine the transcriptional response of pneumococci during adherence, a pan-genomic Nimblegen expression array was designed, covering common and strain-specific genes of all ten strains used within the consortium (see WP1). After 4 hours of adherence to Detroit 562 cells, the expression of adherent (cell-attached) bacteria were compared to the expression of non-adherent (planktonic) bacteria present in the same well to exclude effects on pneumococcal gene expression of secreted factors by the Detroit cells. Array signal were subjected to extensive data filtering to exclude signal due to cross-hybridisation of Detroit RNA (present in the fraction of cell-attached bacteria despite enrichment for microbial RNA) to pneumococcal gene probes.
Identification of pneumococcal genes affecting adherence
To identify (novel) pneumococcal genes essential for adherence, the high-throughput genome-wide negative selection screenings method TnSeq (van Opijnen et al., 2009, Nat. Methods 6:767-772) was employed, where transposon mutant libraries are subjected to a particular challenge, after which next-generation sequencing technology (NGS) is used to identify mutants that are negatively selected from the library population due to the challenge. To be able to perform an adherence TnSeq screen, generation of mutant libraries, and thus efficient transformability of a strain, is of utmost importance.
The genes identified as factors affecting adherence by both genome-wide approaches were distributed over a variety of functional categories, with predominant functional classes of 'Conserved hypothetical and hypothetical proteins' and 'Transport and binding'. KEGG pathway enrichment analysis of identified genes was performed using the web-based DAVID bioinformatics tool (Huang et al., 2009, Nature Protoc. 4:44-57). This showed significant enrichment of the ABC transporters pathway (KEGG pathway spn02010) among the genes whose expression was downregulated during adherence as well as the essential genes identified by TnSeq in R6. Enrichment of the same pathway was also observed in the group of genes with increased expression during adherence and the TIGR4 TnSeq list, but this was not significant in both cases.
Characterisation of the host response during adherence.
The host response after 4h of pneumococcal adherence was examined by real-time PCR, focusing on a selected number of innate immunity genes. Upon data normalisation to GAPDH, clear differences in expression response of Detroit cells to adherent pneumococci was observed, with the highest expression levels observed for IL8, IL6, IL1B, and CXCL2. Expression of IL10 and IL17 was below detection level, while expression of a second housekeeping gene, ACTB, did not differ between the samples. The differences in host response appeared to be somewhat related to capsular type and adherence levels, as the strain with highest adherence generally induced the strongest response. No difference was observed in the response to adherence of two serotype 3 strain.
Workpackage 5 Pneumococcal Physiology
Workpackage Objectives
The main objective of this work package was to document the central metabolism of S. pneumoniae to provide the basic knowledge required to correlate strain-to-strain variations in the adaptation to different host compartments with possible alterations resulting in fine-tuning changes in central metabolic fluxes.
For this goal, the aim was to determine carbon, nitrogen and energy fluxes under defined growth conditions; identify the connection(s) between carbon and nitrogen metabolism, and the interplay between carbon and nitrogen regulators; characterize the connection(s) between central metabolism and expression of the X-state (competence) and of key virulence determinants, e.g. capsule production and utilization of choline (Cho); when possible, evaluate the impact of predefined physiological conditions on expression of candidate strain-specific determinants of colonisation or virulence.
Progress Towards Objectives
All objectives were attained
Carbon, Nitrogen and Energy Fluxes.
All the clinical isolates and model strains retained for the PNEUMOPATH project have been systematically assessed for their growth under defined conditions. Carbon and energy fluxes were calculated from known pathway topography and stoechiometric matrix concepts. No significant correlation was established coupling a specific growth phenotype with virulence in the initial identified strains. The effect on pH evolution was characterised and seen to be a factor influencing carbon and energy fluxes with potential effects on virulence and capsule production.
A clear growth phenotype was observed with various D39 variants (see below). In all cases observed maximum growth rates were retained at a constant value, but for virulent strains, growth arrest occurs at a relatively low biomass concentration, while an avirulent strain continues to grow, reaching biomass levels at least three times higher. Metabolome analysis confirmed a modified metabolic flux distribution, clearly indicating that the metabolic sugar catabolism network is responding to the overall fitness of the strains rather than specific and local regulatory phenomena. Transcriptional analysis shows that during the exponential growth phase in which growth rates and sugar uptake rates are highly conserved in these strains, only a limited number of genes can be seen to be differentially expressed.
Carbon and Nitrogen Metabolism.
Initial physiological analysis suggested that the expression of genes associated with capsule biosynthesis (and the rate of capsule synthesis) and Cho metabolism might be under the control of the carbon catabolite regulator CcpA, a master regulator of carbohydrate metabolism in Gram-positive bacteria. In addition, consensus sites for CcpA binding (cre sites) exist upstream of the cps (coding for capsule synthesis) and lic (coding for Cho utilization) operons in all sequenced S. pneumoniae strains. A ccpA mutant, however, did not show modified expression of these genes. Thus, while capsule synthesis and Cho metabolism do appear to be linked to high rates of sugar conversion, CcpA is not directly involved in this mechanism.
Genomic Analysis of D39 strains.
Three D39 strains have been sequenced and SNPs identified compared to the original D39 sequence available in the databanks. Comparison of the avirulent D39 strain with two virulent strains showed only two common SNPs and approximately 50 SNPs specific to one or more but not all of the three strains. These are potential targets that could explain the differences in virulence phenotypes.
Repertoire of Sugar Transporters and Regulation of Sugar Transport in S. pneumoniae.
A genomic survey allowed establishment the occurrence of 21 phosphotransferase systems, seven carbohydrate uptake ABC transporters, one sodium:solute symporter and a permease, emphasising an exceptionally high capacity for uptake of carbohydrate substrates. Despite high genomic variability, combined phenotypic and genomic analysis of 20 sequenced strains assigned the substrate specificity only to two uptake systems. Systematic analysis of mutants for most carbohydrate transporters enabled a phenotype and substrate specificity to be assigned to 23 transport systems. For five putative transporters for galactose, pentoses, ribonucleosides and sulphated glycans activity was inferred, but not experimentally confirmed and only one transport system remains with an unknown substrate and lack of functional annotation. Using a metabolic approach, 80% of the 32 fermentable carbon substrates were assigned to the corresponding transporter. The complexity and robustness of sugar uptake is underlined by the finding that many transporters have multiple substrates, and many sugars are transported by more than one system.
CelR-I, a transcription factor previously implicated in virulence, was found to function as an activator of a cellobiose-utilisation gene cluster.
Connections between Carbon and Nitrogen Metabolism and Regulators.
Regulation of carbohydrate metabolism.
Regulation of Arginine Metabolism.
The Global Nutritional Regulator CodY.
The finding that CodY is an essential protein in the pathogenic strain D39 and laboratory equivalent R800 was extended to three other pathogenic strains (out of three tested), suggesting that CodY is globally essential in the species, and a potential target for treatment of pneumococcal diseases.
The two-component regulatory system CiaR-CiaH (CiaRH) affects a number of physiological processes such as development of competence, autolysis and A-lactam resistance.
Overall, results obtained in this project reveal many connections between carbon- and nitrogen-dependent transcriptional regulation.
Connection(s) between Central Metabolism and Expression of the X-state (Competence) and of Key Virulence Determinants.
CiaRH, CodY and the Regulation of X-state (Competence).
Metal Homeostasis and Virulence. Besides the documentation of iron homeostasis revealed through analysis of CodY essentiality (see above), it was shown that a number of pneumococcal genes are differentially regulated by copper, including an operon, cop, encoding the CopY regulator (Shafeeq et al., 2011, Mol Microbiol 81: 1255-1270).
Zinc (Zn++) is another important trace metal ion that has been shown to regulate the expression of several (virulence) genes in streptococci.
Workpackage 6 Integrated Bioinformatics and Data Management
Workpackage Objectives
There were three main goals in this work package, namely: deployment of a web-based semantic integration environment (S3DB), development of dedicated data processing services and development and maintenance of an Application Programming Interface (API), which allows project participants to control the sharing and public dissemination of their contributions.
Progress Towards Objectives
All objectives were attained.
One of the main goals of this work package was to develop a semantic web data management system to store the data produced by all partners during the project. Particular attention was given to the user interfaces. Different web based interfaces for data visualisation, retrieval and querying have been developed.
All project participants used the sdLink system, with good results both in scalability and usability by non-expert users. The system has a public project available for evaluation purposes (http:/kdbio.inesc-id.pt/sdlink/sample/) see Figure 3.6.1.
Presently, all data that were produced by the partners are modelled and stored in this system. Excel interfaces have been programmed to represent the database model and to facilitate the data collection and upload to the system. All the project participants can query and retrieve the data.
Figure 3.6.1: Interface to access the projects available in the system
sdLink Data Management System
During the last couple of years new standards and technologies have been developed to support the development of databases using semantic web principles. The development of sdLink was supported by the state-of-the-art technologies available for this type of systems.
The sdLink system was built on top of two main premises. First, it should take a minimal assumption approach with respect to data models, allowing for model extensibility and refinement, without enforcing severe restrictions on expressiveness. Second, it should be scalable, ensuring data availability, integrity and security.
With respect to the first assumption, sdLink follows the concept of linked data, a recommended best practice for exposing, sharing, and connecting pieces of data, information, and knowledge on the Semantic Web using URIs and RDF. Thus, data are represented following an ontology, that can be updated and which allows the definition of inference rules, an important feature for automatic data transformation, model updating and knowledge discovery.
sdLink was then built on top of well-known open technologies and was comprised of three main modules:
1) A management module for projects, users and access control. Authentication relies on the OpenID open standard, providing a decentralized mechanism, and all data about projects, users and access controls (authorization) is stored within the system itself. Figure 3.6.2 presents the new login interface where it is possible to select the OpenID provider.
2) The second module makes available endpoints for data retrieval and querying in several formats, both as REST endpoints and rich Web user interfaces, providing several views on the data. The query endpoint accepts SPARQL, with a user-friendly interface and pre-defined views and queries available. Figure 3.6.3 presents a user interface that is already available and which can be used to visualize the database concepts and the data,.
3) A third module consists of a scalable triple store, including support for transactions and a powerful query engine. This module is built on top of Virtuoso, a high performance triple store. All modules and configuration will be available as pre-configured virtual machine, suitable for both evaluation and production deployments.
Figure 3.6.2: PNEUMOPATH new login interface
Figure 3.6.3: User interface used to visualise the database concepts and the data already available
Excel spreadsheet API
Besides the visualisation interfaces, specific interfaces have been developed to upload new data to the system. As is well-known, biologists are used to collection of the data in Excel spreadsheets and this pre-requisite was determinant to this type of interface development.
During this project, the development of API for spreadsheet client applications has been started and concluded. A new API is available at http://kdbio.inesc-id.pt/~pmreis/Setup.zip and the development site maintained at http://kdbio.inesc-id.pt/mailman/listinfo/sdl-dev.
Figure 3.6.4 shows an Excel workbook including a new tab for the sdLink API. With this new resource the user can select the database concepts and generate a personalized Excel workbook for the data that must be uploaded. When selecting the project PNEUMOPATH ontology, that describes the information system, a new window is opened in order to allow the selection of the concepts to be represented at the Excel workbook. This workbook contains one concept per worksheet and in each worksheet, each column correspond to an attribute connected to the concept. In this semantic web-based system, each concept should be seen as a table and each attribute, a column in that table. Each row of the table will be filled with the values of each attribute.
Figure 3.6.4: Excel workbook with the sdLink API included
Workpackage 7 Management and Dissemination
Workpackage Objectives
-To co-ordinate the project within the consortium
-To co-ordinate contact between the consortium and the EC
-To disseminate, explain and exploit the objectives and results of the project with stakeholders such as the scientific community, public authorities and industry.
Progress Towards Objectives
All the objectives were attained.
During the lifetime of the project nine project consortium meetings were held. All partners participated in, and contributed to, all of the consortium meetings. In addition, meetings between partners took place alongside other conferences and meetings where opportunity arose.
Continuing written and verbal communication, facilitated by Partner 1, took place between the project partners in order to ensure the effective management of the project.
Following agreement from all project partners, an extension to the project from 36 to 42 months was agreed by the Commission. In addition, the re-profiling of the budget to transfer some of the budget of Partner 13 to Partner 1 to enable completion of deliverable DL3.4 and some funds from Partner 5 to Partner 11 was agreed. This re-profiling enhanced the consortium's data analysis capabilities and thus deliverables across all work packages.
Following the approval of the extension, Partners 1 and 8 made minor re-profiling of their budgets in order to appropriately finance the no-cost extension.
The project website, http://www.le.ac.uk/projects/pneumopath was developed in the first year and maintained throughout the lifetime of the project to disseminate information about the project's activities.
Potential Impact:
PNEUMOPATH was a wide-ranging study of host-pathogen interactions during disease due to Streptococcus pneumoniae undertaken in response to the need to increase the knowledge of mechanisms of pathogenesis across a much wider range of pneumococcal strains and host situations than has been done before now. A major impediment previously was that the breadth of technologies and model systems that were necessary to fulfill the objectives of the project were beyond the resources of a single group. The interdisciplinary effort that was made possible by the formation of the PNEUMOPATH consortium, involving academic and industrial partners, gave rise to a large collection of ground-breaking results and ideas.
All of the scientific workpackages produced data with good potential for impact. Studies conducted in WP1 improved our understanding of the evolution and genetic differences of S. pneumoniae and its close relatives. In particular, it may be possible to explore the genetic differences between these species to develop rapid diagnostic tests aimed to unequivocally identify S. pneumoniae in clinical samples. However, the data represent a 'double-edged sword': the data show that a highly specific test is a real possibility but they also provide a warning to scientists and health-care professionals that the closeness of the relationships with other species means that the difficulties in developing a unequivocal diagnostic test are high and that false positive results with the diagnostic targets that have been proposed elsewhere are real possibilities.
WP1 provided a DNA microarray chip covering ten pneumococcal strains. This was a valuable tool within the project that was used to identify genes associated with invasive disease, as well as genes expressed during specific conditions, but this microarray also will be a valuable resource to the wider pneumococcal research community. Within PNEUMOPATH use of the chip provided data of high value for the development of new diagnostic reagents and treatment strategies aimed to contain disease caused by virulent strains.
Within WP2, PNEUMOPATH investigated the genetics of host susceptibility to invasive pneumococcal disease. In the first study of its type applied to an infectious disease, a genome wide association study (GWAS) was undertaken using a large panel of inbred mice to take advantage of the increasing availability of dense genotyping data. The GWAS analysis identified three loci that associated with susceptibility to pneumonia. This type of work is in its infancy but nevertheless the data obtained here will provide a rich seam of new information for those scientists seeking to understand mammalian susceptibility to infection.
The in vivo studies undertaken in WP3 made important discoveries. It was found that cytokine levels could predict the outcome of pneumococcal disease in mice early in infection and were characteristic of host susceptibility. The prognostic significance of these observations should now be tested in humans. The in vivo studies also gave new insights into the process of development of pneumococcal bacteraemia, a condition that accompanies many cases of pneumococcal pneumonia and which is responsible for at least 10% of bacteraemia in Europe. It was ascertained that the action of spleen macrophages during the first minutes of infection determines the clinical outcome and that sustained bacteraemia develops from the replication of a single bacterial cell even when many thousands invade the blood.
A very useful tool for future studies to understand the pathogenesis of pneumonia was the identification of a strain (BS71) that was lethal to mice but was restricted to the lungs and did not cause a bacteraemia. The potential experimental impact of this strain derives from the fact that while much human pneumonia is not accompanied by bacteraemia, all of the animal models of pneumonia are accompanied by bacteraemia. The identification of BS71 changes this situation by giving researchers a tool to study pneumonia in isolation. The knowledge gained and the realisation of the importance of this strain illustrated the impact of the interdisciplinary PNEUMOPATH consortium: work in WP1, 3, 4, 5 and 6 contributed to knowledge of this strain.
In WP4, the interplay between different pneumococcal strains and a variety of host cells in in vitro models were examined, representing different aspects of pneumococcal pathogenesis. This enabled identification of those aspects of the pneumococcus-host interaction that are common between strains and cell types, as well as characteristics specific for certain interactions. The knowledge of these factors, especially those that are conserved between strains, will lead to the identification of new targets for diagnosis, therapy or prevention.
Several bioinformatic advances were made in WP6 that will have applicability beyond the PNEUMOPATH project, for those projects in which the organisation and querying of complex and disparately described data is required. The developments in WP6 included development of dedicated data processing services and an Application Programming Interface and implementation of a semantic web data management system. The systems were designed for use by those unfamiliar with semantic technologies and are currently available for public evaluation.
Data generated during this project were disseminated by presentation at local, national and international conferences and publication in relevant high-impact, peer-reviewed journals. Dissemination of results to the wider scientific community occurred throughout the project, as soon as it was suitable for peer-review publication and after IP considerations.
List of Websites:
http://www2.le.ac.uk/projects/pneumopath
Contact: Professor Peter Andrew, Head of Department
Department of Infection, Immunity and Inflammation
University of Leicester
University Road
Leicester
LE1 9HN
UK
Tel: +44-0116-252-2951
Fax: +44-0116-252-5030
Email: pwa@le.ac.uk