PREVIS - Pneumococcal resistance epidemicity and virulence - an international study

EC contribution: : € 3.000.000
Duration: : 42 months
Starting date: : 01/01/2004
Funding scheme: : Specific Targeted Research Project
Keywords: : Streptococcus pneumoniae, antibiotic resistance, molecular epidemiology, pathogenicity, virulence factors, host factors, transmission, innate immunity, systems biology, vaccines, drugs
Contract/Grant agreement number: : LSHM-CT-2003-503413
Project web-site: : http://www.previs.net

Background:

Centers in five countries with different frequencies of resistance among pneumococci - Sweden (low), the United Kingdom (low), the Czech Republic (low), Iceland (middle), and Portugal (high) - have collected and characterized drug resistant and drug susceptible pneumococci causing invasive disease (sepsis and meningitis) and carriage (nasopharyngeal colonization), respectively, from the same geographical area and during the same time. The isolates have been characterized with respect to their transmissibility, antibiotic resistance and potential to cause severe disease. This has enabled knowledge on the molecular epidemiology of pneumococcal infections and carriage in different geographic areas where the frequency of resistance differs significantly. Resistance determinants in pneumococcal clones associated with high transmissibility have been studied and a genomic comparison between resistant and susceptible isolates has been performed. A genomic comparison between streptococci and sequencing of a strain of S. mitis, a frequent source of heterologous genes carrying resistance determinants in S. pneumoniae, has been done as well. Host factors important in patients with pneumococcal disease has been determined using sera from healthy individuals and patients and host-pathogen interactions affected has been elucidated using the clinical isolates collected and data obtained concerning clinical and genetic parameters. Environmental factors and antibiotic consumption data have been monitored and mathematical modeling performed and correlated to transmission of drug resistant pneumococci. A website has been created. Using comparative genomics and a test platform the project has tried to identify novel antivirulence compounds against pneumococci and other human specific pathogens.

The program of PREVIS has provided an integrated platform to study important and unexplored aspects of pneumococcal disease/pathogenesis, epidemiology/transmission and molecular mechanisms for resistance development and even if the program now has come to an end the network will continue to study invasive disease potential in a collected study from all the countries as well as pneumococcal meningitis in Europe. Hence the project has created novel interactions and collaborations that will continue even after the funding has ceased from the EU commission

Problem:

Streptococcus pneumoniae remains among the most important causes of life-threatening community-acquired diseases such as pneumonia, septicemia and meningitis, particularly in high risk groups such as young children, HIV+ individuals and the aged. The annual global mortality rate is over one million. Streptococcus pneumoniae is also the major cause of upper respiratory tract infections (URTI) such as otitis media, and URTIs are one of the most common reasons for visits to doctor´s offices and for antibiotic prescriptions. As many as 60% of healthy children attending day-care centers have been found to be colonized with pneumococci in the nasopharynx, which therefore appear to be a main reservoir for this pathogen. Drug resistant clones (DRPn) emerging from this major ecological reservoir are widely spread in Europe, threatening effective antibiotic therapy. For decades, penicillin has been the drug of choice for treating pneumococcal infections, but increasing levels of penicillin resistance, up to 50% in some areas, has resulted in the use of alternative antibiotics. However, this has resulted in the development of resistance to many of the alternative antibiotics as well, and vancomycin is the last drug of choice, especially when treating invasive pneumococcal infections caused by DRPn clones, since no resistance to this antibiotic has been observed so far. Of perhaps even greater concern, multiresistant isolates (isolates that are resistant to = 3 different classes of antibiotics) are widespread globally. Interestingly, relatively few DRPn clones have been found among healthy children in day-care centers. Are the pneumococci found among healthy children of the same genetic lineages as those pneumococci causing invasive disease? Why S. pneumoniae sometimes acts as a devastating pathogen causing a severe disease with a potential fatal outcome, while in other instances causes a non-invasive upper respiratory tract infection or is simply carried harmlessly, is not known. Are there differences in the severity and nature of diseases caused by DRPn and drug susceptible pneumococci (DSPn)? How are transmissibility and virulence affected by antibiotic resistance determinants? Properties affecting virulence of the organism, such as adherence, invasion and transmission of the bacteria, and human host factors have to be investigated and correlated to the development of resistance and to the acquisition of resistance markers. Also, antibiotic consumption may be an important factor affecting transmission and selection for resistance determinants. A better knowledge of molecular mechanisms involved in resistance and of host-pathogen interactions affecting pneumococcal infections would lead to improved intervention, prevention and treatment strategies of these common community acquired infections.

Aim:

Human pathogens emerging in the contemporary environment face two main kinds of evolutionary challenges. A. Survival and growth in the antibiotic rich milieu makes it essential that the bacteria acquire genetic traits of resistance. B. Successful drug resistant strains must also be able to compete with other members of the species for colonization, geographic spread and disease in the human host. The primary objective of this research is to examine the interplay of these two - as well as host and environmental - factors.

Expected and obtained results:

Invasive as well as nasopharyngeal isolates from children attending day-care centers have been collected in centers in four countries, with different frequencies of resistance among pneumococci, Sweden (low), the Czech Republic (low), Iceland (middle), and Portugal (high). The pneumococcal isolates, drug susceptible (DSPn) as well as drug resistant (DRPn), have been characterized with serotype, antibiotype and with molecular techniques (PFGE and MLST). The results obtained suggest that most DSPn and DRPn lineages identified are internationally disseminated and that the population structure of DSPn and/or of particular serotypes found in this population is more homogeneous than that of DRPn probably reflecting an increased capacity for microevolution among serotypes associated with drug-resistance.

Invasive isolates have also been collected in the same countries and when comparing the invasive disease isolates (from sepsis and meningitis) and carriage isolates (nasopharyngeal colonization) respectively from the same geographical area and collected during the same time period, our data are in agreement with the literature where invasive disease potential has been intimately associated with the capsular type. Some serotypes are more invasive while others are mainly associated with carriage. Invasive disease potential has been calculated for different serotypes and clones. We find that the genetic background is important in our collection, and there were clear differences among different clones of serotypes 19A and 6B for example in Portugal. The effect of the 7PCV vaccine was also studied and non-typeable pneumococci (NTPn) were characterized and the results indicate that NTPn have diverse genetic backgrounds, evolved as a distinct group of pneumococcal isolates, and their population structure is the same regardless of the clinical source and of the specific genetic background. We have analyzed the phase variants of pneumococcal isolates responsible for invasive infections in humans and found that the opaque variant was more frequent than the transparent, in agreement with previous indications from animal models, but also that there was association of particular phase variants with certain serotypes.

Several of the drug resistant and drug susceptible clones found with different invasive disease potential, have been tagged with luciferase and characterized in live animal models (mice), to study differences in virulence. We showed that not only serotype but also clonal type is important for disease outcome in mice. Also, we have found a novel bacterial factor important for colonization, virulence and the inflammatory response in mice. This bacterial factor is a pilus-like structure encoded by one of the recently described pathogenicity islands in pneumococci. We have found that this pilus is important for bacterial adhesion to epithelial cells, for virulence in mice and for the inflammatory response. Pili are present mainly in pneumococcal isolates common in carriage and in antibiotic resistance. With microarray analysis, Southern blot analysis and PCR we have found that this islet is present in internationally recognized resistant pneumococcal clones, such as the Spain9V-3 clone, and may be one reason for the successful spread of this clone around the world. However, in rare cases isolates belonging to this clone may also lack this islet, suggesting that an important virulence factor may differ between strain of the same clone by PFGE and MLST. The pilus is composed of three proteins called RrgA, RrgB, and RrgC. We have found that RrgB makes up the stalk of the pilus, and that RrgC mainly is found at the tip of the pilus and together with RrgA are ancillary proteins. We have also shown that RrgA is a major adhesion important for colonization. Pili are found in strains mainly causing carriage and antibiotic resistance in humans and we have found that the pilus might be an important factor for the successful spread of antibiotic resistant clones around the world.

The major symptoms of pneumococcal infections involve a massive inflammatory response by the host. We have investigated the role of the Toll like receptors (TLRs) in host defense against pneumococcal infection and found that the adaptor protein the myeloid differentiation protein MyD88 is absolutely essential for host defense against pneumococci. MyD88 deficient mice became hypercolonized and were much more susceptible to pneumococcal infections. The MyD88 protein seems to be important for maintaining an asymptomatic carrier stage. We found that the only TLR that played a non-redundant role in pneumococcal infections was to our surprise TLR9 recognizing CpG DNA. We also demonstrated that TLR9 deficient macrophages are defective in pneumococcal phagocytosis and killing

Neutrophils are important phagocytic cells in the innate immune response toward pneumococcal infections. We have shown that they produce so called NETs (Neturophil Extracelullar Traps) in response to pneumococcal infection. These NETs are composed of DNA, histones, and granule enzymes such as neutrophil elastase. NETs may trap but in contrast to other pathogens not kill pneumococci. Furthermore, we now show that the EndA DNase allows pneumococci to degrade the DNA scaffold of NETs and escape, and by escaping NETs the bacteria may spread from the upper airways to the lungs and from the lungs into the bloodstream during pneumonia. We have also studied the role of the polysaccharide capsule and lipoteichoic acid modification on pneumococcal interaction with NETs. We found that pneumococcal capsule of different serotypes significantly reduced trapping by NETs, but were not required for resistance to NET-killing. Pneumococci contain a dlt operon that in other Gram-positive bacteria has been shown to catalyze the incorporation of D-alanine residues into lipoteichoic acids (LTAs), thereby introducing positive charge. Genetic inactivation of dltA in non-encapsulated pneumococci rendered the organism sensitive to killing by antimicrobial components present in NETs. Also, in a murine model of pneumococcal pneumonia, the encapsulated dltA-mutant strain was outcompeted by the wild-type upon invasion into the lungs and bloodstream. This suggests a non-redundant role for LTA alanylation in pneumococcal virulence and that the reduced negative surface charge by LTA-D-alanylation contributes to the resistance of pneumococci to killing by NET components.

We have sequenced and annotated the S. mitis B6 genome resulting in an expected genome with the size of 1.14 Mb. An oligonucleotide microarray of 70mers have been designed and is currently being synthesized. The genome has an unusual arrangement when compared to the known S. pneumoniae genomes, and the sites of apparent recombination events have been confirmed again by PCR analysis followed by sequencing of these regions. A large data set on PBP2x genes from penicillin-resistant S. pneumoniae and oral streptococci has been compiled. We could show that some changes in penicillin resistant strains, i.e. the acquisition of a resistant PBP2b gene confers benefits to the strains such as a reduced capacity to autolyse. Finally, finishing of the genome of a high level beta-lactam resistant Streptococcus mitis enabled the analysis of the core genome as well as accessory gene clusters involved in antibiotic resistance in a commensal species and documented extensive gene transfer events with other species including Streptococcus pneumoniae. The results present an excellent basis for future research dedicated to the genomic plasticity of pneumococci and related bacterial species, and will further our understanding on the dissemination of antibiotic resistance genes.

The role in pneumococcal infections of so called virulence genes found in several bacterial species has been investigated, and so has the mechanisms for the vag gene, vag H. vagH of Yersinia pseudotuberculosis has been shown to be an essential virulence associated gene and the VagH homologue in E. coli, HemK is a methyl transferase transferring methyl groups to at least two known protein targets; the peptide release factors RF1 and RF2. The most promising substances from the two different screening campaigns have been tested in an animal model. Mice were infected orally with Yersinia pseudotuberculosis, and the substances were injected intraperitoneal once a day for 12 days. There were no effects of the substances on the bacterial infection, thus the optimised substances are not suitable for further drug development. However, the substances have been very useful in the work to study virulence in bacteria. Thanks to the work with the substances we have identified an unexpected connection between the vagH-gene and the well known type III secretion system. This is novel knowledge and open up new avenues regarding targets for drugs against bacterial infections.

Databases have been created in Sweden and on Iceland including data on pneumococci non-susceptible to penicillin (PRP) as well as demographic data. Comparisons have been made between the frequency of antibiotic resistance and antibiotic sales. An iterative artificial neural network (ANN) model has been developed to describe and predict the spread of PRP in space and time as a function of consumption and a number of different confounders. A web-based infrastructure for warehouseing data had been developed and a public web site is available on www. previs.net. The results of the work in PREVIS has been and will be published in peer reviewed international journals and presented on conferences.

Potential applications:

Basis for molecular diagnostics and human-pathogen interactions. Large collection of DRPn and DSPn strains from invasive and non-invasive samples together with comprehensive databank on their phenotypic and genotypic properties Pneumococcal pili. We have found that pneumococci harbour pili on their surface and that these structures are important for adhesion, virulence as well as for the inflammatory response. In collaboration with the vaccine company Novartis we are no studying probable protection in animal models since the pilus might be a potential protein vaccine candidate to be use in a multi component protein based pneumococcal vaccine.
Connection between vagH and type III secretion system. We have found an unexpected connection between vagH and the well known type III secretion system in Gram-negative bacteria. A vagH-mutant has a down-regulated type III secretion system. Also, substances that have been identified to block VagH also inhibit the type III secretion system-dependent toxicity. This opens up the possibility for a new target for Gram-negative bacteria.

Coordinator:

Prof. Birgitta Henriques Normark
Swedish Institute for Infectious Disease Control
171 82 Solna, Sweden
Tel. +46 84572300
Fax +46 8302566
E-mail: Birgitta.Henriques@smi.ki.se

European Commission