Skip to main content

Novel Approaches to Bacterial Target Identification, Validation and Inhibition

Final Report Summary - NABATIVI (Novel Approaches to Bacterial Target Identification, Validation and Inhibition)


Executive Summary:

Infectious diseases by opportunistic pathogens such as Pseudomonas aeruginosa and Burkholderia cenocepacia retain a prominent position as a major worldwide cause of morbidity and mortality in a wide range of patients. This problem has worsened with the emergence of antibiotic multi-resistant bacteria and the failure of drug discovery programmes to design antibiotics with truly novel modes of action. NABATIVI was supported within the programme FP7-HEALTH-2007-2.3.1-1 with the objective to identify and validate novel drug targets in order to select lead compounds for future development of a new class of anti-infective drugs against Gram-negative bacteria. Two approaches have been considered:

A) Discovery Phase Approach A: “From target to lead compound”. A combination of advanced genomic approaches devised to screen the entire genome of P. aeruginosa for the identification of potential targets was applied from the beginning. Essential and virulence bacterial targets (genes) have been identified and validated in clinical strains and in a sequential cascade of disease models including Caenorhabditis elegans, Drosophila melanogaster, human cell lines and mouse models. These models have provided the rationale and the proof-of-concept for the characterisation and validation of selected targets as candidates for the development of antibacterials. Amongst the different targets, we focused our attention on i) quorum sensing systems, which orchestrate the expression of different virulence factors and have the potential to substitute or complement traditional antibiotic treatment; ii) essential cellular processes including the bacterial cell wall; iii) virulence factors which were not recognised in previous discovery programmes.

Using a combination of targets both previously known and discovered during the NABATIVI project, several screening processes (in silico and functional) were used and a number of potential lead compounds identified from both synthetic and natural compound libraries. These compounds will have to be further profiled to assess their therapeutic potential. In addition, a number of extracts from natural products and peptide nucleic acids (PNAs) have been found to inhibit some of these targets. The identification of the active compounds from these is under way but may run beyond the lifespan of NABATIVI.

B) Discovery Phase Approach B: “From drugs to targets”. In this case it was the natural antimicrobial peptide protegrin I which served as starting point to discover a novel class of antibiotics with a novel mode of action. The target (gene) was identified during the drug discovery process. In particular we have identified LptD as novel drug target. LptD is an outer-membrane protein widely distributed in Gram-negative bacteria. The LptD–drug lead interaction causes a major disturbance in the outer membrane structure, possibly by interfering with its biogenesis, which requires LPS. Hit-to-lead and lead optimisation following a multidimensional optimisation strategy of ADMET properties led the novel compounds POL7001 and POL7080, which are effective against a wide range of clinical strains. Protection against lethal P. aeruginosa infection with potency superior to currently available antibiotics in preclinical studies led to POL7080, which was nominated as a clinical candidate. On 4th of March 2013 Polyphor Ltd announced the successful completion of a Phase I clinical trial demonstrating the clinical safety and tolerability of its Pseudomonas specific antibiotic POL7080. All primary study objectives were achieved in this Phase I trial. The study was a randomised, double-blind, placebo-controlled Phase I dose escalation trial assessing the safety, tolerability and pharmacokinetics of POL7080 in 52 healthy male volunteers. The study confirmed that this novel antibiotic was well tolerated by healthy volunteers and thus complements the comprehensive safety and tolerability profile of POL7080. Polyphor expects that POL7080 in Phase II trials will demonstrate its efficacy in treating lung infections caused by multi-drug resistant Pseudomonas bacteria, in line with guidance from the health authorities. On 4th of November 2013, Polyphor Ltd and Roche announced that they have entered into an exclusive worldwide license agreement to develop and commercialize Polyphor’s investigational macrocycle antibiotic POL7080 for patients suffering from bacterial infections caused by P. aeruginosa.

Overall, work by NABATIVI has opened a window for the design of novel antibiotics with novel mode of action and provided an excellent example of how academia and a small innovative company can work productively together to fill the gap left by the migration of big pharmaceuticals away from antibacterial field.

Project Context and Objectives:

Antimicrobial resistance (AMR) is an increasing, global problem for public health. We are in the midst of an emerging crisis of antibiotic resistance for microbial pathogens in the EU and throughout the world (http://www.idsociety.org/badbugsnodrugs.html). Highly virulent and increasingly AMR pathogens have been described in different contexts, including the hospital and in the community. Infections caused by resistant microorganisms often fail to respond to the standard treatment, resulting in prolonged illness and greater risk of death. The death rate for patients with serious infections treated in hospitals is about twice that in patients with infections caused by non-resistant bacteria. When infections become resistant to first-line medicines, more expensive therapies must be used. The longer duration of illness and treatment, often in hospitals, increases health-care costs and the economic burden to families and societies.

The global human burden posed by drug-resistant infections is difficult to quantify, but we have reason to fear that it may be enormous. Antibiotic-resistant germs are now found regularly in many hospitals in the EU, infecting some 4 million patients every year (focusing only on a limited group of healthcare-associated bacterial infections the burden is in the range of 2.5 million hospital days), causing 25.000 deaths and economic losses in the order of €1.5 billion due to extra healthcare costs and productivity losses (http://ecdc.europa.eu/en/publications/_layouts/forms/Publication_DispForm.aspx?ID=444).

The most dangerous multi-drug resistant (MDR) bacteria were described as the ESKAPE pathogens including Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baummannii, Pseudomonas aeruginosa and Enterobacter species. In particular, P. aeruginosa is a major and dreaded cause of infection causing a wide range of diseases in humans at various body sites including the respiratory tract, skin and soft tissues, the urinary tract, post-operative and burn wounds, brain, heart, bloodstream and cornea. Infections caused by P. aeruginosa are often life threatening and they are of particular concern for intensive care units (ICU) where ventilated patients may develop ventilator-associated pneumonia (VAP) and sepsis. Cystic Fibrosis (CF) is another disease where P. aeruginosa lung infections are frequent and often life threatening. Current treatment regimes of Pseudomonas infections either involve broad-spectrum antibiotics such as fluoroquinolones or aminoglycosides, carbapenems such as imipenem or meropenem, third generation cephalosporins such as ceftazidime, or drugs such as the monobactam aztreonam and the penicillin derivative piperacillin, either alone or in combination with tazobactam. However, frequently observed inefficacy of these treatments is linked to intrinsic resistance of P. aeruginosa, development of antibiotic resistance and/or limited penetration of antibiotics into biofilms (Gellatly, 2013).

AMR is a complex problem driven by many interconnected factors, so single isolated interventions have little impact and coordinated actions are required. According to the WHO (http://www.who.int/en/) factors that accelerate the emergence and spread of AMR include:

• Lack of a comprehensive and coordinated EU response;
• Weak or absent antimicrobial resistance surveillance and monitoring systems;
• Inappropriate use of antimicrobial medicines;
• Poor infection prevention and control practices;
• Insufficient diagnostic, prevention and therapeutic tools
• Insufficient research and development activities due to the lack of federal and private funding;

In spite of the pressing need for new drugs to treat multidrug-resistant bacterial infections, there are simply not enough new antibiotics in the pharmaceutical pipeline to keep up with the pace of emerging resistance. Of greatest concern is the departure of many large pharmaceutical companies from antibacterial drug discovery and their decreasing investment in this area of research (Spellberg, 2004). The length of time (~10 to 15 years) and huge costs (U.S. $800 million on average) associated with the taking of a new drug from the discovery phase to market (DiMasi, 2003), combined with the perceived failure of whole-genome sequence-based approaches to spur a second golden age of novel antibacterial drug classes have led many companies to prioritise other areas of research (Fernandes, 2006). For many reasons, investment in the discovery of antibacterials is not as attractive to companies as research into other novel therapeutic drugs. First, the success rate for the discovery of drugs in other therapeutic areas is four to five times higher than that for antibacterial discovery, according to the GlaxoSmithKline screening metrics (14 high-throughput screening runs are required to obtain one lead compound) (Payne, 2007). Since multidrug resistant infections are still a small proportion of the total, although serious for suffering patients, the market for antibiotics is small, and is restricted in poor countries that cannot afford the costs. Finally, repeated prescriptions are unnecessary, and the treatments are short in comparison to therapies for chronic long-term pathologies, representing a disadvantage for drug companies. Thus, the number of new antibiotics annually approved for marketing in the United States has continued to decline. Importantly, the number of large multinational pharmaceutical companies (i.e. “Big Pharma”) actively developing antimicrobial drugs also continues to decline. Only 2 new antibiotics have been approved since IDSA's 2009 pipeline status report (Boucher, 2013).

A major limitation in antibiotic development has been the difficulty associated with the identification of new structures that display the same low cytotoxicity for the host characteristic of conventional antibiotics, but at the same time have a narrow spectrum. No new classes of antibiotics were produced in the 37 years that elapsed between the introduction of nalidixic acid (a bacteriostatic quinolone marketed in 1962) and that of linezolid (Zyvox; Pfizer) (an agent used to treat infections caused by multidrugresistant Gram-positive bacteria marketed in 2000), which was followed by daptomycin (Cubicin; Cubist) in 2003 (Leeb, 2004) and, more recently, by retapamulin (Altabax/Altargo; GlaxoSmithKline) (Coates, 2002). All of the antibacterial agents that entered the market during this period were modifications of existing molecules (Boucher, 2009). Hence, physicians urgently need in their arsenals elusive antibiotics with novel structures and/or modes of action (Projan, 2002).

To overcome the inertia that currently surrounds antibiotic resistance, the European Union adopted a strategic plan on AMR based on a multifaceted approach. The EU's 6th and 7th Framework Programme fund a wide range of projects focusing on basic research, strategies for the prudent use of existing antimicrobials, development of new antimicrobials, development of point of care diagnostic tests, and vaccine development (http://ec.europa.eu/research/health/infectious-diseases/antimicrobial-drug-resistance/index_en.html).

One priority of the area of antimicrobial drug resistance within FP7-HEALTH is the discovery of novel and efficacious antimicrobials for treatment and prevention of bacterial infections. In particular, FP7-HEALTH-2007-2.3.1-1 has supported novel targets identification for drugs against MDR bacteria.

2.3. TRANSLATIONAL RESEARCH IN MAJOR INFECTIOUS DISEASES: TO CONFRONT MAJOR THREATS TO PUBLIC HEALTH

2.3.1. Anti-microbial drug resistance including fungal pathogens

2.3.1-1: Novel targets for drugs against Gram negative bacteria.

The objective is to identify and validate novel drug targets in order to select lead compounds, which may be derived from natural sources or from synthetic compounds, for future development of a new class of anti-infective drugs against Gram-negative bacteria. Significant industrial involvement, particularly by SMEs, is foreseen in this topic.

Funding scheme: Collaborative projects (Small or medium-scale focused research projects with maximum EC contribution of € 6,000,000/project).

NABATIVI is a cooperative project supported within the programme FP7-HEALTH-2007-2.3.1-1. NABATIVI was specifically designed to provide key answers and mobilised the relevant know-how and resources needed to generate new antimicrobial drugs by bypassing current antimicrobial mechanisms of resistance in Gram-negative bacteria. The research project was mainly focused on P. aeruginosa as a model pathogen, since it is an important, intrinsically resistant Gram negative bacterium responsible for high infection rates in humans within the hospital environment, has a completely sequenced genome and is highly amenable to genetic manipulation. However, the ubiquitousness of this organism had made the new discoveries potentially applicable to other important human pathogens such as B. cenocepacia, which was the secondary focus of this proposal.

NABATIVI brings together a multidisciplinary team including nine leading European research groups with complementary expertise to develop new drugs which could be exploited for the prevention and treatment of infections caused by P. aeruginosa and other gram-negative bacteria. The collaboration incorporates top-level expertise in the virulence of P. aeruginosa and related bacteria, in molecular microbiology and genetics, genomics, biochemistry, structural biology, drug development, analytical chemistry, structural biology, cell biology, bioinformatics, clinical research, epidemiology, and high-technologies related to the high-throughput screenings of compound libraries.

Six academic groups and three small innovative companies have worked productively together to fill the gap left by the migration of big pharmaceuticals away from antibacterial development.

Project Results:

Main Scientific & Technological Results

The objective of NABATIVI was to identify and validate novel drug targets in order to select lead compounds for future development of a new class of anti-infective drugs against Gram-negative bacteria. This objective impacts on the increasing emergence and spread of antimicrobial drug resistant pathogens in Europe and the rest of the world and conforms to the aim of the FP7-HEALTH to contribute to international efforts addressing global health problems, such as antimicrobial drug resistance. One priority of the area of antimicrobial drug resistance within FP7-HEALTH is the discovery of new and urgently needed antibiotics.

The proposed work was organised into ten research work packages (WPs), one dissemination and exploitation WP and the project management WP. The research work included three major steps to identify novel drugs against multi-resistant P. aeruginosa and B. cenocepacia:

Target Identification (WP1, WP2, WP3)

Applied a combination of advanced genomic approaches devised to screen the entire genome of P. aeruginosa for the identification of potential targets. Target identification was carried out using a combination of P. aeruginosa clinical epidemic and dominant strains including those resistant to antibiotics. The main molecular genetic approaches used in the identification of targets were:

• Improved insertional mutagenesis for the identification of genes involved in virulence and hence establishment and maintenance of P. aeruginosa infection.
• Signature Tagged Mutagenesis (STM) for the identification of genes expressed during infection and hence required for the adaptation of the pathogen to the host.
• Screening of conditional antisense RNA libraries for the identification of genes essential for growth.

The use of these three strategies was paramount and it provided a unique picture on the complexity of host/pathogen interaction of P. aeruginosa during infection and lead to the identification of the physiological role of the target gene(s).

Target Validation (WP4, WP5, WP6)

Used a multifaceted approach, by a sequential cascade of models including biofilm, human cell lines wild type or carrying CFTR mutations, the nematode worm Caenorhabditis elegans, Drosophila melanogaster and mouse models both for acute and chronic infection, either to provide the rationale and the proof-of-concept for the characterisation and validation of selected targets as candidates for the development of antibacterials.

This project also revealed the role in virulence of a significant number genes of unknown function in the genome of P. aeruginosa as well as invaluable insights into their conservation across different clinical isolates of this organism. Target validation was carried out in a collection of P. aeruginosa clinical strains of different origins and searched in other gram negative bacteria such as, B. cenocepacia strains through multiple genome sequencing and comparative analysis. This provided a panel of targets with a global relevance and can be considered a critical step in terms of providing solid rationale to further proceed with the project towards the target inhibition.

Target Inhibition (WP7, WP8, WP9, WP10)

Developed two alternative, yet complementary, strategies aimed to attack the problem of P. aeruginosa infection:

• Novel drugs from the screening of chemical libraries against specifically selected targets, and the characterisation of the mode of action of novel drugs with antibacterial activity discovered by some members of the partnership.
• Novel antisense inhibitors of bacterial targets by peptide nucleic acids (PNA) delivery.

The combined efforts resulted in a panel of new agents for the treatment of P. aeruginosa infections in patients with cystic fibrosis and other patient risk groups and a robust network of technologies that could also be applied for the development of anti-microbial drugs for other bacterial pathogens, especially other Gram-negative bacteria, such as B. cenocepacia.



WP1 - Search for novel virulence functions by the screening of transposon libraries.

The Gram-negative P. aeruginosa is a clinically relevant opportunistic pathogen highly resistant to most classes of antibiotics (Rehm, 2008). Both acute and chronic infections can be established in a wide range of patients. The conditions present in acute infections force pathogens to injure or kill the host, with multi-organ failure sometimes occurring in hours or days (A). In contrast, chronic infections occur without injury and in the presence of biofilm structures—a population of microorganisms that aggregates on a matrix—that develop over days or weeks, and bacterial genetic variants may grow in the biofilms (B) (Bragonzi, 2010a). P. aeruginosa virulence factors required for the initiation of acute infections are selected against during chronic infection. Thus, acute and chronic bacterial infections are different at cellular and organ levels and may require distinct treatments.

The genomes of P. aeruginosa strains are larger than those of most sequenced bacteria (Stover, 2000). The genome size varies between 6 to 7 Mbp with over 5.500 ORFs. A significant number (8.4%) of P. aeruginosa genes are predicted to be involved in regulation, which at the time of publication of the genome was the largest fraction of regulators among sequenced bacterial genomes. Around 32% of the genes in the P. aeruginosa PAO1 genome have no homology to any previous reported sequences and only 6.7% have experimentally demonstrated functions. This implies that there is likely a large number of genes to be discovered, that are involved in the virulence in this organism, and which will be coding for novel antimicrobial targets. These targets could be exploited for the development of novel therapies.

Two genomic approaches including functional genomics through insertion mutagenesis (strategy A: acute virulence) and signature-tagged mutagenesis (STM) technology (strategy B: chronic virulence), has been used for the identification of novel targets in P. aeruginosa relevant in acute and chronic infection. In particular, the Tn5 mutagenesis approach used to generate transposon libraries is considered a powerful tool to examine a large number of mutants. Using this technology, it is possible to generate 60,000 mutants for each strain needed to cover a genome of 6.5 Mb (Liberati, 2006). The approach has identified novel genes involved in virulence and antibiotic resistance in microbial pathogens. An improved version of the Tn5 mutagenesis is the STM approach. STM uses transposons tagged with unique oligonucleotides for the simultaneous analysis of a large collection of mutants in vivo with a limited number of animals (Potvin, 2003). The approach has been used to identify novel P. aeruginosa genes involved in the adaptation of this pathogen to the host.

The strategy A was based on a multistep-driven screening for genes which mutation results in the attenuation of virulence. Using a mutagenesis approach by the transposon Tn5, the University of Nottingham generated a total of 57,360 mutants in P. aeruginosa PAO1. The target screening was performed at different sequential levels. Firstly, mutants were individually tested for reduced swarming as well as pyocyanin and protease production. Secondly, a selection of those having pleiotropic phenotypes were further examined for attenuation in Caenorhabditis elegans, Drosophila melanogaster and, after a further shortlisting, for reduced cytotoxicity on respiratory cell lines (WP4). Finally, a total of 13 mutants resulting from these sequential screenings, including a selection of genes not attenuated in virulence used as controls, were verified as potential virulence targets using a combination of a mouse acute infection model, together with cell invasion assays and IL-8 release employing the A549 alveolar epithelial cell assay (see WP4) (manuscripts in preparation).

The screening of the transposon library resulted in the selection of 72 open reading frames (ORFs) with unique Tn5 insertions and attenuated in virulence. The predicted proteins encoded by these ORFs revealed that these genes are involved in various cellular functions spanning seventeen functional classes, including energy, amino acid and nucleotide metabolisms, transcriptional and post-transcriptional regulation, motility and adaptation, secretion and transport of small molecules, enzymes as well as conserved hypothetical proteins of unknown functions. There was a wide spread distribution for the predicted functional category of the mutated genes, with the highest abundance corresponding to 24 genes with unknown function followed at a distance by 7 genes involved in transport of small molecules. To address the global relevance of the target genes, a comparative protein sequence analysis was carried out against six sequenced genomes (PA14, LESB58, PA7, 2192, C3719 and PACS2) of highly pathogenic P. aeruginosa and against E. coli K12 and B. cenocepacia J2315 in order to check presence and conservation of our candidates in other P. aeruginosa genomes and in other bacterial species (see WP5). The majority of the selected targets are conserved in all the six P. aeruginosa genomes analysed.

The mutants validated in the cell culture, D. melanogaster and C. elegans and animal model (see WP4) were analysed in silico in order to obtain information on their patentability, biochemical activity, 3-D structure (WP8), and conservation among clinical isolate (WP5). This resulted in the selection of 5 novel targets to be further investigated for their pathogenicity. Since the use of Tn5 mutagenesis can result in the inactivation of more than one gene located in the vicinity of those mutated, it is important to make defined mutations in the specific genes identified in the screening to ensure that the correct target has been identified. The defined mutants were further validated in different model systems as described above. Results of this screening approach will be published soon.

The strategy B was based on the identification of genes involved in the adaptation of P. aeruginosa to the host. A novel STM positive (Pos-STM) approach in P. aeruginosa was reported for the first time by Università Vita e Salute (Bianconi, 2011). The latter study was undertaken to screen for bacterial maintenance, as opposed to elimination (Potvin, 2003). The study was undertaken on the premise that loss-of function mutations in P. aeruginosa isolates from CF patients enhance fitness and sustain chronic infection.

A novel STM screening based on positive selection by using P. aeruginosa PAO1 STM library as source of mutants (Potvin, 2003) and the validated agar beads mouse model of P. aeruginosa chronic infection was designed and published in PLoS Pathogen (Bianconi, 2011). The source of mutants was a P. aeruginosa PAO1 STM mini-Tn5 library constructed previously (Potvin, 2003). Thus, Pos-STM screening of a collection of 7968 P. aeruginosa PAO1 mutants in a murine model of chronic airway infection identified 16 mutants. The mini-Tn5 insertion of each Pos-STM mutant was mapped by sequencing. The 15 insertion sites mapped within a specific P. aeruginosa gene while the remaining one was mapped in an intergenic region (PA0436-PA0437). Interestingly, the previous report showed that eight different intergenic regions were mutated in the 96-month isolate from CF patients (Smith, 2006). The 15 mini-Tn5 inserted genes encoded proteins from almost all functional classes: hypothetical, unknown, unclassified proteins (PA2972, PA4842, PA5028), motility and attachment (PA0410-pilI, PA0499, PA1077-flgB, PA4554-pilY1), putative enzymes (PA1856), transport of small molecules (PA0890-aotM, PA2252, PA4887), amino acid biosynthesis and metabolism (PA0895-aruC), energy metabolism (PA2998-nqrB), secreted factors (PA3478-rhlB), chaperones and heat shock proteins (PA5053-hslV). The phenotypes associated to these Pos-STM mutations reflect alterations in diverse aspects of P. aeruginosa biology which include lack of swimming and twitching motility, lack of production of the virulence factors such as pyocyanin, biofilm formation, and metabolic functions. RhlB and pilY1 specific loci were identified independently by both Pos-STM and Neg-STM approach, supporting previous hypothesis in P. aeruginosa that virulence factors essential for acute infection are lost when P. aeruginosa establishes long-term chronic infection. The reliability and robustness of Pos-STM approach is supported by the high proportion of mutations found in Pos-STM genes of longitudinally P. aeruginosa isolates from CF patients as verified in WP5. This approach generated a list of genes whose inactivation increased the colonisation and persistence in chronic airways infection in murine model similarly to clinical strains isolated from patients with CF.

Animal studies were carried out in strict accordance with the Ministry of Health guidelines for the use and care of experimental animals. This study was approved by the San Raffaele Scientific Institute (Italy), University of Tuebingen (Germany) and University of Notthingam (UK) Institutional Animal Care and Use Committee (IACUC). All efforts were made to minimize the number of animals used and their suffering.



WP2 - Novel essential functions by the screening of antisense libraries

Along with virulence functions (WP1), NABATIVI aimed at identifying novel targets of P. aeruginosa with an essential role for its survival. For this specific purpose, University of Milan adopted in WP2 a genetic technique called “shotgun antisense screening” that was developed in the Gram-positive bacterium Staphylococcus aureus and, following a period of limited success in Gram-negative bacteria (like P. aeruginosa), has recently been used effectively in the well-known model bacterium Escherichia coli. To further improve the method and also target low expressed essential genes, we included in our protocol some variant steps that were expected to overcome the non-stringent regulation of the promoter carried by the expression vector used for the shotgun antisense libraries. Our antisense screenings identified 33 growth-impairing single-locus genomic inserts that allowed us to generate a list of 27 “essential-for-growth” genes: five were “classical” essential genes involved in DNA replication, transcription, translation, and cell division; seven were already reported as essential in other bacteria; and 15 were “novel” essential genes with no homologs reported to have an essential role in other bacterial species. Interestingly, the essential genes in our panel were suggested to take part in a broader range of cellular functions than those currently targeted by extant antibiotics, namely protein secretion, biosynthesis of cofactors, prosthetic groups and carriers, energy metabolism, central intermediary metabolism, transport of small molecules, translation, post-translational modification, non-ribosomal peptide synthesis, lipopolysaccharide synthesis/modification, and transcription regulation.

This study also identified 43 growth-impairing inserts carrying multiple loci targeting 105 genes, of which 25 have homologs reported as essential in other bacteria.

Taken together, our results show the feasibility of antisense technology in P. aeruginosa for identifying novel essential genes. Because of its supposed inefficiency, this approach has been neglected in Gram-negative bacteria for several years, and was only recently recovered in E. coli. By comparison with this previous work, the WP2 results strongly suggest that our modification of the antisense strategy can broaden the class variety of the identified essential genes. We expect that our methodology could be well suited for antisense-mediated searches of essential genes in other Gram-negative bacterial species. The results of our antisense screenings will be published soon and become available as source of novel antibacterial targets for P. aeruginosa.

WP2 also aimed at the characterization of novel essential functions in P. aeruginosa. For this purpose, we considered hits that were found by the antisense screenings. In particular, we focused on the PA2873 gene product that was annotated as a hypothetical membrane protein endowed with a periplasmic region harbouring a structural domain belonging to the transglutaminase-like superfamily, a group of archaeal, bacterial and eukaryotic proteins homologous to animal transglutaminases. In the study recently published in PLoS ONE (Milani, 2012), University of Milano shows that the periplasmic portion of the PA2873 protein, which we named TgpA, does possess transglutaminase activity in vitro. This is the first report of transglutaminase activity in P. aeruginosa. In addition, we have provided strong evidences that TgpA plays a critical role in the viability of P. aeruginosa and is a good candidate as antibacterial target. The characterization of other novel essential functions from the “antisense panel” is in progress.

For the refinement of raw data, a genome-wide approach, as was our antisense screening, necessarily relies on the information about protein function and localization present in the genomic database. However, it is very common that in the sequenced bacterial genomes the percentage of proteins annotated as “hypothetical” (i.e. only predicted in silico and no experimental evidence of in vivo expression) is very high. Likewise, very frequently the subcellular localization confidence is based on pure algorithms. Therefore, as a support of the genome-wide approach of this and other WPs (e.g. WP1), we performed a multiview proteomic analysis of P. aeruginosa to get hints about target expression and localization. In this context, University of Milano developed an innovative proteomic technique for the analysis of the envelope district (Vecchietti, 2012)

Finally, as antibacterial targets, we did not focus only on proteins but also on small RNAs (sRNAs), which several studies suggest to play key roles in regulating cellular processes linked to pathogenesis and essential functions. The number of such regulatory molecules previously identified and annotated in P. aeruginosa was relatively low, considering its genome size, phenotypic diversity and adaptability. The apparent low proportion of sRNAs in P. aeruginosa could reflect either a real paucity of regulatory sRNAs or the limited number of genome-wide searches that have been performed in this species. Therefore, we accomplished a deep-sequencing approach to explore the P. aeruginosa complement of sRNAs (Ferrara, 2012).



WP3 - Identification and characterisation of targets inhibited by peptidomimetic antimicrobials

At the time of this application, a novel compound class derived from natural antimicrobial peptide protegrin I was discovered as potent antibiotic against Gram-negative P. aeruginosa by University of Zürich and Polyphor Ltd (Srinivas, 2010). The target (gene) and mode of action of novel peptidomimetic antimicrobials was elucidated during the drug discovery process within the NABATIVI project.

Photoaffinity labelling experiments were designed to test the hypothesis that the peptidomimetic antibiotics bind to the outer membrane (OM) protein LptD in P. aeruginosa PAO1. A petidomimetic with a photoaffinity labelling reagent (PAL-1, with an MIC against PA PAO1 of 0.05 µg/ml) was designed for this purpose. Photoaffinity labelling with PAL-1 consistently revealed one major photo-labeled protein, confirmed to be LptD (PA0595) by in-gel protease digestion/LC-ESI-MS-MS analysis and Western blot. When the photoaffinity labelling was repeated in the presence of a 100x excess of the antibiotic (L27-11), the labelled band disappeared from the blot, demonstrating a direct competition between L27-11 and PAL-1 for binding to the target. Furthermore, when the experiment was repeated using the resistant PAO1RES1 mutant (containing a mutated lptD gene), photolabelling of LptD was not detected, indicating that the mutated LptD protein is no longer able to bind the antibiotic with comparable affinity. These results provide firm support for the conclusion that the peptidomimetic antibiotics bind with high specificity to LptD in the OM of intact PA cells.

Electron microscopy studies were carried out in an effort to detect morphological changes to cells caused by the antibiotics. For this purpose, P. aeruginosa PAO1 cells grown for 3-5 h in the presence of antibiotic were examined in thin sections by transmission electron microscopy (EM). After fixation many apparently intact cells showed remarkable accumulations internally of membrane-like material, an effect not seen in cells grown normally without the antibiotic, nor in cells of the resistant PAO1RES1 mutant grown in the presence of the antibiotic. Similar accumulations of extra membranous material were reported in E. coli cells depleted of lptD, and in other bacteria exposed to antimicrobial peptides. These results provide a first indication that the peptidomimetics have an impact on outer membrane biogenesis, resulting in the appearance of a membrane-stress-like response.

To confirm the essentiality of ostA, a conditional mutant was constructed and grown in the presence of either glucose or rhamnose. ostA mutant grew in the presence of 0.5 % rhamnose similarly to the wild type but was unable to grow in 0.5 % glucose on both plates and liquid media as expected for conditional mutant with an essential gene under the control of the rhamnose promoter. Depletion experiment using liquid media also showed that the ostA mutant was unable to grow after subculturing into fresh medium supplemented with glucose instead of rhamnose. To further confirm that ostA is essential for the viability, the cells at the 11 hour time point of the growth were stained with the Baclight Live/Dead stain and examined by the phase contrast and fluorescent microscopy. When grown under inducting conditions (0.005 % rhamnose) the ostA mutant cells were green/alive similar to the wild type, while under inhibiting conditions (0.005 % glucose and 0.001 % rhamnose) the majority of the ostA mutant cells were red/dead, thus proving that ostA is important not only for growth but also for the survival of Pseudomonas.

Down-regulation of lptD expression should produce effects on lipid A modification that are similar to those induced by the antibiotic. To test this, lipid A was prepared from the conditional PAO1 mutant grown under permissive conditions and analysed by ESI-MS. The MS spectrum showed mostly the same lipid A species as seen from the wt PAO1 (coloured red below). Under limiting growth conditions, however, bacterial growth was slower and the extracted lipid A showed accumulation of the same lipid A species seen when the wt strain is exposed to antibiotic (blue lipid A). Thus, down-regulation of lptD leads to the accumulation of LPS forms that have not been processed by PagL, just like in the case of antibiotic-treated cells.

In summary, we have gained substantial new evidence that the peptidomimetic antibiotics exert their antimicrobial effect by interacting with LptD and inhibiting the transport of LPS to the cell surface. This is the first example of antibiotics targeting LptD and having a new mode of action, representing a valuable tool to combat emerging resistance to the currently used treatments. The results have been published in Science (Srinivas, 2010) and commented in Science Translational Medicine (Bragonzi, 2010).



WP4 - Characterization of the novel targets by a cascade of different model systems

The identification of targets carried out mainly in vitro by genome-wide screenings (see WP1-2) does not imply that they are relevant for their pathogenesis in vivo. A pathogen can experience several radically different “host environments” at various stages of infection (acute or chronic). In the context of infections caused by P. aeruginosa virulence is mediated by a wide variety of trans-acting regulators that sense the environment and the physiological state of the cell and adjust the transcription of specific genes changing their phenotype significantly. The idea of a core set of virulence factors common to all infection models from plants and insects to humans is thus unlikely. Therefore, we need a multifaceted approach, by using a sequential cascade of models to provide the rationale and the proof-of-concept for the characterisation and validation of a selected bacterial target as candidate for the development of antibacterials.

To address this question we have used a new validation strategy based on a cascade of in vitro and in vivo virulence models including airways cells, C. elegans, D. melanogaster and mouse models of acute and chronic lung infection. In WP1, the University of Nottingham constructed a Tn5 mutant library in P. aeruginosa PAO1. Among a total of 57,360 mutant strains, 404 mutants were found to be attenuated in virulence factors including swarming, pyocyanin and protease production. These mutants showing pleiotropic phenotypes were tested in a second stage screening by the University of Zurich. This was based on the use of C. elegans and D. melanogaster as infection hosts to identify attenuated mutants. The aim was to further shortlist these mutants prior to evaluate their cytotoxicity and virulence in eukaryotic cells lines and murine models. Indeed, although rodents are the first choice for understanding infectious diseases in human, screening a large amount of targets in mouse models is unfeasible.

From a total of 404 transposon mutants tested, 108 were found to be attenuated in at least one of the infection models. This screening and the identification of the insertion sites to eliminate mutants redundancy reduced the numbers of mutants with attenuated virulence in at least one non-mammalian model to 72 target genes. A Genomic Target Database (GTD) has been developed by scoring the target genes for virulence traits, phenotype and pathogenicity in different model systems. Next, we selected five mutant strains attenuated in all the models to be exploited for the inhibition stage of the NABATIVI project (see WP8). All these five mutants have an unknown functions and are potentially patentable. In addition, eight mutant strains attenuated in one, both or none of the C. elegans and D. melanogaster disease models for publication purpose. These mutants show a proof of concept of the validation strategy.

Stable ko specific mutants were generated (see WP1) and further validation tests included: (i) additional phenotyping including biofilm formation, ability to produce elastase and pyoverdine, (ii) C. elegans and D. melanogaster killing curves, (iii) cell cytotoxicity, (iv) invasion and (v) cytokine production assays. All the 13 selected mutants showed a reduction in surface motility, biofilm formation, and their ability to produce elastase and pyoverdine. The impact of the total 13 selected mutations on D. melanogaster and C. elegans survival was confirmed for the majority of the target genes repeating the infections with stable ko mutants. In accordance with the previous outcomes of the sequential screening cascade approach Eberhard Karls Universitat and Università Vita-Salute San Raffaele challenged A549 alveolar epithelial cells with the mutants in order to validate the attenuation of the candidates in terms of cytotoxicity, invasion capacity and secretion of the pro-inflammatory cytokine IL-8. Six out of the eight mutants selected for publication and three of five mutants selected for inhibition purpose were attenuated in their virulence in immortalized cells.

Finally, Università Vita-Salute San Raffaele tested P. aeruginosa selected mutants for lethality in the mammalian host using a murine model of acute lung infection, by intra-tracheal injection of planktonic bacterial cells. While wild type PAO1 was totally lethal within 36h, the lethality of some mutants was significantly lower and temporally shifted. Five out of eight mutants for publication purpose and two out of five mutants chosen for the inhibition stage were significantly attenuated in inducing mortality when compared to wild type PAO1-L. It should be noticed that some mutants attenuated in C. elegans were not attenuated in the mouse model. Our data show that identification of virulence genes carried out mainly in vitro does not imply that they are relevant for their pathogenesis in vivo, suggesting a host-specific response to P. aeruginosa and indicating the necessity to test a selection of mutants in the mouse model, as the nearest to the human host.

Acute P. aeruginosa infection in immuno-compromised patients or chronic persistent lung infection in patients with CF are different at the cellular and organ levels and thus may require distinct treatments. To mimic the environmental conditions present in the lung of CF patients, a mouse model of chronic lung infection, by intra-tracheal injection of agar-bead encased bacteria, was established in mice by Eberhard Karls Universitat and Università Vita-Salute San Raffaele (Bragonzi, 2010b). Five targets selected for the inhibition step were tested. Results showed that one of the target mutants, attenuated in all the previous screenings, was highly reduced in lethality in the murine model of chronic infection, thus representing the most promising candidate for drug development. In addition these results confirm that a multifaceted bottleneck approach, by using different screening systems to provide the rationale and the proof-of-concept, is essential for the characterization and validation of a large number of bacterial target candidates.

In summary, by using a genomic approach devised to screen the entire P. aeruginosa genome for novel virulence genes (WP1) and a multifaceted approach (WP4) based on a sequential cascade of models for the validation step, we have identified several novel virulence genes. These targets have been further investigated for their function and for their inhibition (WP7), representing interesting targets for innovative anti-virulence approach to P. aeruginosa infections.



WP5 - Validation of the targets in strains of various clinical origin

One limitation of many identified targets described during the past three decades is that the identified targets showed sequence variability, and although they induced protection against homologous strains, they failed to induce protection against heterologous ones. Clinical relevance means that the new targets should be adequate to cover a spectrum of pathogens faced by a physician, thus providing the opportunity for the generation of new antibacterials. To address this need, NABATIVI exploited genomic sequences of P. aeruginosa strains from different origins and other gram-negative bacteria and use this information to select validated targets as clinically relevant. Thus, the aim of this work package was to validate the targets selected during the project in strains of various clinical origin by multiple genome sequence and comparative analysis.

The genome-wide screenings for the putative sensitive bacterial targets in NABATIVI has been carried out mainly on the laboratory reference strain P. aeruginosa PAO1, and at some extent on P. aeruginosa dominant strain PA14. As first step, to determine the relevance of our panel as therapeutic targets, Università Vita-Salute San Raffaele carried out comparative analysis in order to check whether 226 targets selected in WP1, WP2 and WP3 were present in six sequenced genomes of P. aeruginosa (PA14, LESB58, PA7, 2192, C3719 and PACS2). For homology analysis the BLASTP program present on the NCBI website was used. The analysis was done by aligning the amino acid sequence of PAO1 against the other six genomes of P. aeruginosa. We generated a target database containing information about the class of confidence and the functional classification of the proteins, the strain in which the target has been identified and the percentage of identity of the protein in each strain compared to PAO1. The database of targets has been grouped by ranking the following scores:

• The degree of conservation;
• The presence in all six sequenced strains;
• The presence of ortholog genes in the same strain;
• The target confidence class.

Although the screening for novel targets was carried out in P. aeruginosa as model bacterium for studying opportunistic pathogens, any identified targets has been searched and validated in other gram-negative bacteria, such as Burkholderia cenocepacia and Escherichia coli. Some Burkholderia strains and species have been determined to be transmissible between patients either in hospitals or clinics or through contact between CF patients outside clinical settings. E. coli is known as one of the most common pathogens that cause nosocomial infections, together with P. aeruginosa and Staphylococcus aureus. A comparative genome analysis of the selected targets was performed in B. cenocepacia J2315 and E. coli K12 strains with the goal to identify all the target that are common or specific for one or the other strain, therefore providing an overall complete list of universal encoded potential targets. Furthermore, the targets were also analysed to verify the absence of similar proteins in humans.

All the results obtained in this WP and in WP4 were used to generate a genomic target database (GTD) of 72 selected targets. An exemplificative picture of this database is shown below. Sequences analysis of virulence genes identified by insertional mutagenesis and validated by a sequential different models outlined in WP4 showed that these genes were conserved and not mutated thus representing excellent candidate targets for inhibition.

It has been recently shown that bacterial functions needed for acute and generalizing types of infections, such as those occurring in immunosuppressed patients are selected against during the progression toward localised chronic infections (Smith, 2006). Indeed, many targets that would be considered for antibacterial development, may be expressed at low levels or absent from the majority of isolates during chronic infection. While these targets may be useful to eliminate the initially infecting strains, it is not clear whether their inhibition will control or promote progression towards the chronic infection. Thus, conservation of targets in phenotypically and genotypically different P. aeruginosa strains generated during the course of chronic infection, is required to develop an effective antimicrobial against P. aeruginosa. To address this issue, Università Vita-Salute San Raffaele evaluated the panel of targets, selected during the course of NABATIVI project, in terms of genetic diversity and frequency of recombination. This was intended to provide clinical validation of selected targets and solid rationale to further proceed with the project towards the targets inhibition. To this end, twenty-five sequential P. aeruginosa isolates from 6 patients with CF were chosen from the strains collection of the Hannover CF center (Germany), including early strains isolated at the onset of chronic colonisation and late clonal strains collected over a period of up to 16.3 years. In addition, the collection was implemented with strains isolated from respiratory specimens of 345 CF patients recruited in one of the largest centre in Europe (Cystic Fibrosis Center of Verona, Italy) for an epidemiological study aimed to establish the potential presence of epidemic strains. Phenotypic and genotypic analysis have been carried out in the strains of these collections. In details, the isolates were characterised for the presence of specific phenotypic traits associated with the acute phases of infection (twitching and swimming motility, production of proteases, siderophore and pyocyanin, haemolysis) or chronic phases of infection (mucoidy, LasR).

To address the issue of target conservation, Università Vita-Salute San Raffaele set out to sequence 16 target genes which were selected by Pos-STM (see WP1). Sequences analysis of the genes by comparison between early and late strains revealed single base-pair synonymous and non-synonymous mutations in seven of 16 genes in six different clonal lineages. Non-synonymous mutations were found in five genes from five different clonal lineages of CF patients. A standard computational method for predicting the effect of each non-synonymous mutations on protein function suggested that the non-tolerated changes in three genes are likely to affect protein function. This approach generated a list of genes whose inactivation increased the colonisation and persistence in chronic airways infection.



WP6 - Exploitation of known targets

Previous studies performed by NABATIVI participants have revealed the existence of a number of potential antibacterial targets in P. aeruginosa and B. cenocepacia. Amongst these are quorum sensing (QS) systems which comprise highly attractive novel therapeutic targets, which have the potential to substitute or complement traditional antibiotic treatments of chronic diseases (Hentzer, 2003). The QS response orchestrates the expression of a cocktail of virulence factors in a population density-dependent manner in P. aeruginosa as well as in B. cenocepacia. Therefore blocking of the QS cascades in these Gram-negative bacteria is a very promising approach to attenuate their pathogenicity.

The aim of this work package was to develop bioreporter systems using known and new targets which can be exploited for high throughput screening (HTS). Initial work focused on the exploitation of QS-related targets in Gram-negative bacteria, particularly the N-acyl homoserine lactone (AHL)-dependent QS system operating in B. cenocepacia and the Pseudomonas quinolone signal (PQS)- dependent QS system of P. aeruginosa.

To develop a platform for the screening of compounds interfering with PQS signalling in P. aeruginosa a biosensors was constructed by University of Nottingham utilising promoter regions of lasI, lasR, rhlI, rhlR, pqsA and pqsR. The biosensors were sent to Actar-KDevExploratory to screen for compounds, which could potentially block the synthesis of QS signal molecules (see WP7). For the screening of compounds interfering with octanoyl-homoserine lactone (C8-HSL)-dependent QS in B. cenocepacia biosensors suitable for mass screen of compound libraries were constructed by University of Zurich utilizing GFP-based biosensors in different genetic backgrounds. The resulting strains were then characterized with respect to their potential to screen for compounds interfering with QS in B. cenocepacia.

Another task in this work package was the construction of whole cell reporter screens for compounds that inhibit OstA of P. aeruginosa, the target of compound POL7001, which is investigated by Polyphor Ltd. To this end a conditional ostA mutant, in which the natural ostA promoter has been exchanged for a rhamnose-inducible one, was constructed. Growth experiments in which a culture of this strain is shifted to a medium lacking arabinose unequivocally demonstrated an essential role of ostA for cell viability (see WP3 for details). We have performed a transcriptome analysis of the conditional mutant under permissive and non-permissive conditions and identified differentially regulated genes when ostA expression is abolished or the wild type strain is treated with POL7001. These promoters will be valuable for the construction of biosensors to screen and further analyse OstA inhibitors.

A third aim of this work package was the identification of novel antibacterial targets in B. cenocepacia. Employing a bioinformatics approach we determined that the core genome of the order Burkholderiales consists of 649 genes. All but two of these identified genes were located on chromosome 1 of B. cenocepacia. Although many of these 649 core genes have been shown to be essential in other bacteria, we were also able to identify a number of novel essential genes present mainly, or exclusively, within this order. To test the essentiality of the identified genes conditional knock-down mutants in which the native promoter of the target gene or operon has been exchanged with a rhamnose-inducible promoter (as described above for OstA) were constructed. In this approach, approximately 300 bp fragments spanning the 5’ region of a targeted gene were cloned into pSC200 and the resulting recombinant plasmids were subsequently transferred into the model strain B. cenocepacia H111 by triparental mating. Essentiality of target genes was examined by growing the strains in medium containing either rhamnose (permissive) or glucose (non-permissive). All conditional mutants grew well in the presence of rhamnose but were unable to grow in the presence of glucose on agar plates or in liquid medium. Using this approach the essentiality of some of the core genes, including the known essential genes infB, gyrB, ubiB, and valS, as well as the so far uncharacterized genes BCAL1882, BCAL2769, BCAL3142 and BCAL3369 could be confirmed experimentally. These latter genes have not previously been shown to be essential and thus represent novel targets for the development of anti-Burkholderia drugs.

NABATIVI has identified a handful of top targets which are currently under patent consideration. One of the approaches used was to identify compounds which could prevent the production of these targets and hence attenuate the virulence of P. aeruginosa. For this purpose, the University of Nottingham generated a collection of biosensors based on the construction of CTX lux-based chromosomal fusions to the promoter of each of these targets so that their expression could be monitored through the production of light. This was done to enable the identification of compounds which could switch off the expression of the genes coding for these targets using a high throughput approach. These biosensors also enabled the monitoring of the expression of these targets through growth.



WP7 - Screening of natural and chemical compound libraries and lead compound generation

The objectives of this WP were to develop appropriate HTS assays and then screen collections of natural and synthetic compound libraries to select inhibitors of targets identified and validated by partners in NABATIVI consortium. The WP7 output was a wide panel of small molecules, both hit and lead compounds, expected to be further optimized and well-suited to enter a portfolio of candidates for pre-clinical trials.

WP7 was successfully completed with vHTS and HTS assays on known target from QS (see WP6) and 4 novel virulence targets selected, identified and validated by NABATIVI consortium in WP1. In addition, HTS assays were developed and applied to select inhibitors of two essential targets identified and validated by NABATIVI consortium in WP2. Additional screening have been initiated of natural compound and plant extract libraries for growth inhibition activity on P. aeruginosa PAO1 to enhance the possibility to identify hits endowed with antibiotic potential.

HTS bioluminescence assays were developed for screening of QS inhibitors based on P. aeruginosa. 12,500 synthetic compounds, 480 natural purified compounds and 800 plant extracts of libraries were screened on PA14 biosensors, PA14 and PAO1. Using structural data on targets of interest, 25,000 original synthetic compounds from Actar-KDevExploratory, 20 million available products from Ambinter.com and 150,000 natural compound structures from Greenpharma were screened with in silico methods (WP8). Starting with 597 actives anti-QS synthetic and 13 natural anti-QS compounds, finally 21 synthetic compounds and 8 natural compounds were confirmed as anti-QS hits. 8 scaffolds from anti-QS screening were selected with good potential in IP and freedom to operate in SAR. 42 extracts reduced QS expression, which activity guided fractionation is still under way. 7 natural compounds and 8 extracts with growth inhibiting activity on PAO1 strain were identified. The scaffolds were optimized and champion compounds were tested in available phenotypic assays, by LCMS to identified direct effect on QS signal molecule production.

Example of results from compound screening. Results expressed as remaining activity of compounds tested for inhibition of QS genes, QS molecule levels and virulence factors controlled by QS.

Four novel targets from P. aeruginosa GTD identified and validated by NABATIVI consortium were proceed in screening assay development. HTS biosensor bioluminescence assay was developed for screening. Proteins of two novel targets were successfully purified. DSF screening assay were established for one. 5000 compounds of synthetic library and 100 natural compounds were screened on the novel targets. 41 actives were selected and 15 hits identified.

Finally, 21 compounds, selected by vHTS (WP8) for targeting two essential targets identified in WP2, were assayed. For six of these compounds that showed significant growth inhibition activities on PAO1, protocols of optimization are underway.



WP8 - Structural characterization of the targets and virtual HTS of compound libraries

This WP aimed at identifying lead inhibitors of selected targets by virtual HT screenings (vHTS) in a rational drug discovery approach. To obtain 3D target structures for vHTS, two strategies were adopted in parallel: (i) determination of protein 3D structure by crystallography and (ii) homology modeling. Unfortunately, the first strategy was unsuccessful for the targets that were considered. Therefore, homology 3D modeling was performed whenever possible. Furthermore, Greenpharma established a procedure to score the relevancy of 3D structures of target proteins starting from their amino-acid sequence; the druggability and patentability of targets were also considered. This procedure was applied to 332 P. aeruginosa proteins.

Importantly, experimental inputs from Università Vita e Salute, University of Nottingham, University of Tuebingen and University of Milan were crucial to prioritise and balance the novel targets for vHTS between essential and virulence factors. A well-established target such as DNA gyrase (new inhibitory site), was also selected to possibly accelerate the discovery of leads.

Ten vHTS campaigns were undertaken by Greenpharma using 25,000 original synthetic compounds from Actar-KDevExploratory, 20 million products from Ambinter.com and 150,000 natural compound structures available in Greenpharma database.

Several candidate inhibitors were identified by vHTS and assayed for activity in WP7. Structure-Activity Relationship (SAR) was explored for several compounds to prepare hit-to-lead optimisation. This consists in designing (computer-aided) similar derivatives from hits, yet with a certain degree of diversity between them, in order to generate a maximum of SAR data.

In parallel with 3D structure-based vHTS, chemoinformatic tools were used to design two libraries with top-level chemical diversity to maximise the probability to get hits in the assays performed in WP7. One library consisted of natural compounds from plants and micro-organisms (480 pure molecules); the other contained 7000 molecules from synthetic or hemisynthetic compounds. These libraries were used in in vitro screening without target structural bias (it is complementary to the protein-based approach) to explore thoroughly the possibility to inhibit select targets. This screening yielded 12 inhibitors of PqsA, an important gene controlling QS and responsible for biofilm formation and virulence factors. Two most potent hits were retained for further chemical improvements. Evaluations of these new derivatives are on-going.



WP9 - Development of novel antisense antimicrobials

One alternative strategy of NABATIVI was to explore antisense technology for targeted inhibition of gene expression, and to exploit this new knowledge for discovery of novel peptide nucleic acid (PNA) based antibiotics against P. aeruginosa.

Studies of PNA oligomers and PNA-peptide conjugates in E.coli and S.aureus had very clearly shown that insufficient bacterial uptake is the major limiting factor for the efficacy of PNA antibacterials (Good, 2001). The use of carrier peptides such as the KFFKFFKFFK peptide can to some extent alleviate this problem in E. coli but is far from the optimal solution. However, it was not known which carrier peptides could be effective P. aeruginosa. Thus for discovery of novel antimicrobial agents against P. aeruginosa a variety of both gene and sequence targets as well as delivery agents was explored in the WP. The starting point iwas the essential genes acpP and ftsZ and the KFFKFFKFFK peptide carrier already validated in E. coli, as well as the biofilm associated psqA gene identified in key pathways of QS.

University of Copenhagen has designed and synthesized more than 70 PNA-peptide conjugates using a variety of carrier peptides and mRNA target sequences, and characterized these in terms of antibacterial activity against P. aeruginosa as well as in terms of mechanism of action. This screening and optimization resulted in identification of four PNA-peptide conjugates (PNA3969, PNA3984, PNA4071 & PNA4073) (targeting acpP and ftsZ, respectively) exhibiting potent antibacterial activity (MIC values of 0.6 – 2 µM) against three medically relevant P. aeruginosa strains (PA01, PA14 & LESB58), and all showed (at least partly) bactericidal activity.

On the basis of lower toxicity to eukaryotic cells in culture one of these (PNA3969) was selected for in vivo experiments in a lung infection mouse model, and data from this study demonstrate in vivo antibacterial activity of this PNA compared to untreated control or a similar sequence mismatch control PNA (PNA3993). A PNA-peptide conjugate (PNA4064) down regulating pqsA was also discovered and characterized, showing to down regulate quorum sensing signals and inhibit biofilm formation.

In addition, PNA peptide conjugates targeting mRNA of genes selected by NABATIVI consortium (PNAs 4359 & 4360) have been synthesized and are being tested in in vitro and in vivo models (manuscript in preparation).

In conclusion this work package has identified an antisense antimicrobial lead compound showing potent antibacterial activity in vitro as well as in vivo activity against P. aeruginosa.



WP10 - Efficacy and safety of selected antimicrobials

The long-term goal of NABATIVI is to develop a new class of anti-bacterial drugs which are active against sensitive and multi-drug resistant P. aeruginosa and other gram-negative strains. Thus, the main objective of this WP was to test the activity of selected antimicrobials against a panel of multi-drug resistant P. aeruginosa strains. POL7001, a potent antibiotic acting specifically against P. aeruginosa with a new mode of action (see WP3) was chosen for this purpose. The efficacy of POL7001 was demonstrated in vitro by assessing its antimicrobial activity against a selected panel of sensitive and multi-drug resistant (MDR) P. aeruginosa clinical isolates from CF patients. In particular, the Minimun Inhibitory Concentration (MIC), the lowest concentration of an antimicrobial that inhibits the visible growth of a microorganism after overnight incubation, was measured. MICs are important in clinical microbiology and laboratories to monitor the activity of new antimicrobial agents and to confirm resistance of microorganisms to an antimicrobial agent. The MICs of POL7001 and other antibiotics clinically used to treat P. aeruginosa infections, the so called reference antibiotics, were determined against the strains collection of the Hannover CF center (see WP5). MICs were determined for POL7001, ciprofloxacin, meropenem, ceftazidime, colistin, gentamicin, tobramycin, and imipenem. MICs for POL7001 were particularly low, ranging between < 0.0005 - 0.125 µg/mL for all isolates, indicating a potent in vitro activity of POL7001. For the same isolates, MICs of reference antibiotics were comparable or much higher (ranges 0.03->8 µg/mL). Over time, many of the P. aeruginosa isolates from the patients became resistant to two or more antibiotics while remaining sensitive to POL7001. There was no difference in activity of POL7001 against mucoid and non-mucoid or hypermutable isolates, which tend in general to become resistant to several antibiotic classes. This is in agreement to the new mode of action identified for POL7001.

Next, Polyphor Ltd evaluated the safety of selected antimicrobials. For this purpose, the in vitro ADMET properties of POL7001 were measured. POL7001 was not cytotoxic against COS 7 and Hela cell lines at concentrations of 100mg/ml. In addition, POL7001 was not hemolytic against red blood cells at 100mg/ml indicating no potential for toxicity against mammalian cells. The metabolism stability of POL7001 was measured in presence of liver microsomes from different species. Microsomes are valuable tools for investigating the metabolism of compounds (enzyme inhibition, clearance and metabolite identification) in vitro. POL7001 is metabolically stable towards liver microsomes, indicating that the compound will not be readily metabolised in the liver. The stability of POL7001 was also measured in presence of plasma from different species. POL7001 was stable towards enzymatic degradation in the plasma, indicating that the compound will remain in the blood stream after administration to exert its antimicrobial activity. In summary, the ADMET properties of POL7001 are favourable and support further development of this novel antibiotic.

To show in vivo efficacy of POL7001, the compound was tested in murine models of acute lung infection induced by P. aeruginosa. Mice were infected with bacteria by intra-tracheal surgery and treated either with POL7001 or ciprofloxacin as control antibiotic. Different routes of administration were tested, in order to reproduce standard clinical procedures. In particular, the antibiotics were administered either subcutaneously (s.c.) or intra-tracheally (i.t.) at different doses and at different time points post infection. In all studies POL7001 showed comparable or significantly better efficacy than ciprofloxacin, especially when administered intra-tracheally.

The in vivo efficacy of POL7001 was tested also in a chronic infection model with bacteria embedded in agar beads. These immobilising agents provide the microaerobic/anaerobic conditions that allow bacteria to grow in the form of microcolonies, similarly to the growth in the mucus of CF patients. After the infection, mice were treated once daily for six days by s.c. administration of POL7001 or ciprofloxacin. After 6 days of infection, mice treated s.c. with POL7001 displayed a slight but significant decrease in bacterial load Interestingly, ciprofloxacin did not show a significant efficacy in this model. Considering that bacteria in this microenvironment are particularly difficult to treat, the efficacy results obtained with POL7001 are encouraging.

POL7001 reduced P. aeruginosa bacterial counts in murine models of acute and chronic respiratory infection. The efficacy studies in vivo support the pulmonary administration as a possible therapeutic approach for the treatment of lung infections with POL7001.

Potential Impact:

Improving human health and impact on the economic cost of antimicrobial resistance. P. aeruginosa infections are a social burden and have a strong impact on the budgets of the National Health Systems. Due to the resistance to antibiotic treatment, two-thirds of nosocomial infections would be preventable by the development of new antibacterial principles. A major achievement of NABATIVI project was the discovery of a novel class of antibiotics against Gram-negative Pseudomonas with a novel mode of action. Inspired by the natural antimicrobial peptide protegrin I, a hit-to-lead optimisation program followed by a multidimensional strategy to optimize ADMET properties and in vivo efficacy in murine septicemia infection models, led to the identification of POL7080, which was nominated as a clinical candidate. POL7080 and its analogues represent a new antibiotic class (PEMdrug antibiotic) with a novel mode of action and an excellent safety and efficacy profile. The target gene (LptD) was identified during the drug discovery process. Efficacy of POL7001, a close analogue of POL7080, in murine models of pneumonia was confirmed. On November 4, 2013 Polyphor Ltd and Roche announced that they have entered into an exclusive worldwide license agreement to develop and commercialize POL7080 for patients suffering from bacterial infections caused by P. aeruginosa. Antimicrobial resistance represents a major threat to public health worldwide, leading to 25,000 deaths and related costs of over €1.5 billion in healthcare expenses and productivity losses in the European Union alone each year. P. aeruginosa accounts for one in every 10 hospital-acquired infections in the US and is listed as one of the six most dangerous drug-resistant microbes. Over 15% of P. aeruginosa isolates were resistant to at least three classes of antibiotics and close to five percent were resistant to all five classes under surveillance. Development of a new antibiotic from this class will enable the prescription of an efficacious drug to patients suffering from serious lung infections by P. aeruginosa where other existing antibiotics exhibit no or insufficient activity due to multidrug resistance. The development and validation of new therapies fostered by NABATIVI results are expected to impact on health promotion and prevention as well as to contribute significantly to the sustainability of efficient health care systems. In addition, the improved therapeutic strategies promoted by the NABATIVI results are expected to impact on the budgets of the National Health Systems lowering significantly the actual economic cost of the antibiotic resistance. Every treatment that substantially lowers P. aeruginosa prevalence in CF, as well as exerts beneficial effects on the patient's clinical status, is cost-effective compared with conventional antibiotic therapy for chronically infected CF patients. The reduction in the number of infections caused by P. aeruginosa would lead to lower number of hospital admissions with great consequences not only for patient quality of life, but also in indirect costs, with a decrease in the number of lost work days.

Reinforcing the competitiveness and boosting the innovative capacity of European health-related SMEs and businesses. A general aim of FP7-HEALTH Theme 1 was “increasing the competitiveness and boosting the innovative capacity of European health-related industries and businesses”. As reported in I.1 Policy Context – Work Programme FP7-HEALTH Theme 1, “in line with the European strategy on life sciences and biotechnology and the Lisbon objectives, this Theme help increase the competitiveness of European health care biotechnology and medical technology sectors where research intensive SMEs are the main economic drivers and pharmaceutical industries”. As mentioned above, we witness the departure of many large pharmaceutical companies from antibacterial drug discovery for several economic reasons. In the case of NABATIVI, innovative SMEs filled the gap left by the migration of big pharmaceuticals out of the antibacterial sector and were the main economic drivers in the field. NABATIVI outcomes impacted positively the competitiveness, innovative capacity, and businesses of the involved SMEs relative to the wide market of antibiotics and provided them with the chance to become leaders in the field of antibiotic discovery.

Integrated multidisciplinary research at European level and exploitation of post-genomic information: Another aim of FP7-HEALTH Theme 1 was “to stimulate and sustain multidisciplinary basic biomedical research where large scale collaboration at the EU level is essential to exploit the full potential of post-genomic information to underpin applications to human health” (I.1 Policy Context – Work Programme FP7-HEALTH Theme 1). The integrated steps which supported the strategy of NABATIVI spanned different disciplines such as molecular microbiology and genetics, genomics, chemistry, biochemistry, structural biology, cell biology, bioinformatics, clinical research, epidemiology, and high-technologies related to the high-throughput screenings of compound libraries. Furthermore, the identification of novel targets was based on genetic technologies which deeply rely on the post-genomic information. For this configuration, the NABATIVI project matched also the objectives of the topic HEALTH-2007-2.1.2-5 (Multidisciplinary fundamental genomics and molecular biology approaches to study basic biological processes relevant to health and diseases) for the study of basic biological processes in bacterial such essential and virulence functions, which was relevant to the discovery of new and urgently needed antibiotics.

To support the high content of multidisciplinary research required for objective achievements NABATIVI project gathers academic and enterprise scientists from six European Countries bringing together multidisciplinary expertise. Therefore, the project promoted the integration of European Excellence in the field of antimicrobial drug resistance. The cooperative aspect in this proposal has undoubtedly facilitated such efforts devoted to the fight against the antibacterial drug resistance. Furthermore, the consolidation of NABATIVI partnership during the development of the project will lead to other collaborative research programmes between the partners to further progress the science in the field basing on NABATIVI outcomes, such as, for instance, projects focused on i) improvement and optimization of the lead compounds to increase their efficacy, ii) improved genetic technologies for genome-wide detection of virulence and essential genes, iii) genetic tools for defining and tracking the mode of action on novel antibacterial agents iv) process improvement of high-throughput screenings of compound libraries and antibacterial drug design.

List of Websites:

Project website address: www.nabativi.org

Alessandra Bragonzi
Università Vita e Salute /Fondazione Centro San Raffaele
Via Olgettina 58,
20132 Milano, Italy
Tel: +39 02 2643 4189
Fax: +39 02 2643 5183
E-mail: bragonzi.alessandra@hsr.it