Skip to main content

Preventing community and nosocomial spread and infection with MRSA ST 398 - instruments for accelerated control and integrated risk management of antimicrobial resistance

Final Report Summary - PILGRIM (Preventing community and nosocomial spread and infection with MRSA ST 398 - instruments for accelerated control and integrated risk management of antimicrobial resistan

Executive summary:

The spread of infections in hospitals poses a considerable risk to patients and therefore is of public health concern. Prevention and control measures are urgently needed. Methicillin-resistant staphylococcus aureus (MRSA) is of increasing concern both as community-and hospital-acquired infection. The European Union (EU)-funded project PILGRIM focused on one strain, the MRSA ST398, an animal-adapted, zoonotic, resistant pathogen that causes colonisation and infection in humans in community and health care settings. MRSA ST398 was the most relevant strain to study because it is animal-adapted allowing us to use animal models and it is able to infect humans offering a unique opportunity to investigate colonisation, host range and virulence factors. Up to 39 % of pigs may carry MRSA ST398 in affected countries and human MRSA cases are 35 times as likely as controls to be in contact with pigs. As MRSA ST398 occurs in the community, as well as in hospitals with a high potential for further spread in countries with approved prevalence and emergence throughout Europe timely action is required to prevent a major public health impact.

The central hypothesis of PILGRIM research was that deeper understanding of factors affecting pathogen-host interaction of resistant bacteria at the molecular as well as the population level will lead to new and more effective control measures against nosocomial infection. The project was conducted by 12 partners based in Belgium, the Czech Republic, Denmark, The Netherlands, Switzerland and the United Kingdom.

PILGRIM was structured into six Work packages (WPs), 4 scientific and 2 administrative:

Epidemiological and physiological studies of MRS ST398 were conducted using conventional observational studies (WP1) and strain typing (WP3) and also using well comparable animal models (WP2, WP4) in order to better understand the specific ecology of zoonotic resistant bacteria.

The main activities during the first reporting period consisted of field research on the epidemiology and transmission of MRSA ST398 in livestock and community settings. Results allowed establishing the transmission risks between animal and human populations as well as the environment. In parallel, lab-based research was conducted to investigate molecular and pathogenic characteristics of MRSA ST398.

An in vivo porcine model for MRSA colonisation studies was developed. The model is based upon perinatal transmission of MRSA. The study also provided epidemiological information on the dynamics of MRSA transmission in pig production systems as it showed that MRSA lineage ST398 is efficiently transmitted from contaminated sows to their offspring. The in vivo porcine model was then used to evaluate the pathogenicity of specific MRSA ST398 clones, to elicitate host specificity and to quantify the effect of bacterial antagonism. An in vitro corneocyte model was developed and used to identify pathogenicity determinants.

Also, a diagnostic test for MRSA CC398 was developed using a novel Multiplex PCR test which allows cheap, rapid and accurate detection of CC398 (hsdS) and methicillin resistance (mecA) genes simultaneously. This invention can be applied by newer technology (microarray or deep-stick) to develop a solid-based assay for MRSA-typing. The annotated sequence of MRSA ST398 was released for public use.

In terms of control methods, work was conducted to identify novel technical solutions to reduce MRSA contamination in hospital and farm environments. Antimicrobial agents for surface MRSA decontamination were identified and their efficacy was investigated. The use of these technologies was evaluated in vivo in farm-style chambers. Suitability for prevention and control of MRSA ST398 transmission in health-care settings was established and ranked. A mathematical model was developed to simulate the spread of MRSA ST398 within pig populations and its transmission to humans.

A concept for a Technology testing platform (TTP) was developed to provide Small and medium-sized enterprises (SMEs) with validation possibilities for products targeted at the control of MRSA in healthcare environments. The TTP was implemented by PILGRIM partners with the aim to provide protocols for product testing in a "natural" MRSA infected environment. The TTP was made commercially available and was used by 4 companies. It was demonstrated that there is a need for standardised testing protocols in the industry. The TTP supported industry partners by making available testing and evaluation protocols that allow for comparison between products in terms of effectiveness and costs.

Project context and objectives:

The spread of infections in hospitals poses a considerable risk to patients and therefore is of public health concern. Prevention and control measures are urgently needed. MRSA, also known as ?super bug?, is of increasing concern both as community-and hospital-acquired infection.

MRSA is one of the most significant antibiotic-resistant pathogens worldwide and poses a huge economic burden not only for healthcare institutions but for the society at large. Surveillance data show that the proportion of MRSA in European Member States is variable and some countries are witnessing major epidemics. Most cases in humans are nosocomial infections but community acquired MRSA carrier rates are increasing, also in children. Child health was identified as an overarching issue of strategic importance within the EU. MRSA also poses a huge economic burden not only for healthcare institutions but for the society at large. In addition to direct health care and personal costs, MRSA causes indirect costs impacting on the gross domestic product. The latter has been estimated for the United Kingdom to range between EUR 4.5 and 16.2 billion annually due to lost work or output. MRSA carriage is a risk factor for disease. Therefore, understanding the epidemiology and spread of MRSA, and eliminating carriage, are the key weapons in the fight against MRSA.

There is considerable difference in MRSA prevalence between European Member States ranging from less than 1 % to above 50 %. Furthermore, 12 out of 29 countries reported a significant increase in the proportion of MRSA within the last years. Until recently, MRSA in low-prevalence countries was most frequently related to patients returning after admission in foreign hospitals. Today, however, there is a new, ?national? source of MRSA: livestock. Contact with livestock has been described as a risk factor for MRSA-carriage in Dutch pig farmers and veterinarians. These strains were all non-typable (NT) by standard Pulsed field gel electrophoresis (PFGE) due to the presence of a novel Deoxyribonucleic acid (DNA) methylation enzyme. Staphylococcal protein A (spa) typing of these strains showed that they belonged to a number of closely related spa-types. Up to 39 % of pigs may carry MRSA ST398 in affected countries and human MRSA cases are 35 times as likely as controls to be in contact with pigs. Amongst those with indirect or direct contact with livestock, Health care workers (HCW) may form a small but important subgroup, since they could spread MRSA ST398 leading to nosocomial infections.

Similar to humans, animals can be asymptomatic carriers of MRSA on the skin and mucosal surfaces, and transmit the bacteria to humans. A survey in the Netherlands demonstrated a high prevalence (39 %) of MRSA ST398 among slaughter pigs. The prevalence of MRSA was lower in Denmark, with only 1-2 % of slaughter pigs found to be positive. Very little is known about risk factors for the occurrence of MRSA ST398 in animals. The use of tetracyclines is a potential risk factor since all ST398 strains are resistant to tetracyclines, the antibiotic class most commonly used in pig farming.

Methicillin resistance genes are encoded on mobile genetic elements called staphylococcal cassette chromosomes (SSCmec) and have moved into about ten different lineages of S. aureus. Strains of MRSA isolated from pigs and from people handling pigs are all closely related. Their special character is evident as they are not typeable using the regular PFGE protocols because their DNA resists digestion by the SmaI restriction enzyme routinely used in PFGE typing of S. aureus strains. MRSA ST398 is only rarely found among isolates from human beings and it is likely that MRSA ST398 represents a clone or biotype of S. aureus that over evolutionary time has adapted itself successfully to pigs (and other animals, but not humans). However, nothing is known presently regarding the genetic basis of S. aureus biotypes. Comparative sequence analyses between typical human and typical animal lineages of S. aureus have not been published. Questions related to host specificity and adaptation have not yet been addressed.

MRSA were previously detected using culture and antibiotic resistance tests that typically took 3-5 days, delaying control measures and leading to overuse of antibiotics of last resort. New molecular tests based on light-cycling PCR can detect MRSA in 4 hours are becoming available but are relatively expensive and require specialist equipment in the laboratory. To identify a specific lineage of MRSA such as ST398 requires identification and pure culture of the MRSA, and then spa or MLST typing which requires specialist equipment and expertise and many days Ideally, a rapid specific test for ST398 would be performed in situ in front of the patient or at the pig farm, as well as being inexpensive and simple to use. Such a test would need to exploit patented dip-stick test technology targeting specific sequence or proteins.

The species S. aureus includes various biotypes or ecotypes that have adapted to specific hosts. This has been clear since the 1970s, when strains from poultry, dogs and pigeons and the majority of bovine, human and porcine strains were classified as belonging to different biotypes. More recent studies based on genetic comparison between strains have confirmed that rabbit, bovine and human S. aureus isolates are not clonally related suggesting that host-dependent factors may have evolved independently in these species. Differently from other S. aureus lineages, the host-range of ST398 appears to be unusually broad. Despite its recent recognition, ST398 has been cultured from the nasal cavity of pigs, cattle and humans, from ventilator-associated pneumonia in hospital patients, skin infections in farmers, pigs and dogs, sinusitis and wound infections in horses. Another S. aureus clone that seems to infect distinct hosts is ST22, which has been known for many years as a typical nosocomial MRSA but recently has emerged as a veterinary pathogen in companion animals.

Adherence is an essential step of both colonisation and infection. Study of bacterial adherence requires a compromise between the complexity of host-pathogen interactions in vivo, and the need for simple in vitro assays to assess adherence properties of large numbers of strains. Manual counting by direct microscopy was originally used to quantify in vitro bacterial adherence to epithelial cells or corneocytes. This difficult and time-consuming approach was replaced by the use of fluorescein-labelled probes and computerised image analysis. Recent studies have shown that the ability of staphylococci to adhere to human and animal corneocytes correlates with the host-range of staphylococcal species. These studies indicated that in vitro assays based on corneocyte adherence can be employed to investigate the host-range of distinct MRSA clones.

Animal models enable host-microbe interactions to be studied under controlled conditions. Ideally, such models should mimic the real problem being studied. Many models have been developed of S. aureus disease but few deal with colonisation. Mice have been widely used. However, nasal carriage in mice was short term unless streptomycin was applied. More recently, bacterial and host factors implicated in murine nasal carriage of MRSA were studied. It was demonstrated that a constant decrease in signal and concluded that S. aureus did not multiply actively in the nasal cavity of the mice. Because human MRSA strains are not good colonisers of pigs, the ability of MRSA ST398 to naturally colonise pigs made the ST398-pig interaction a very useful model for understanding host specific colonisation and methods for control of nosocomial infection.

Porcine models of staphylococcal colonisation and infection have been developed both in conventionally raised pigs and in gnotobiotes. Piglets are naturally colonised at birth by bacteria, including staphylococci from the vulvar flora of the sow. Attempts to implant staphylococcal clones on piglets by inoculating them onto the skin post-partum resulted in only brief persistence of the exogenous organisms. The hypothesis that implantation of MRSA in the vulva shortly prior to parturition would be necessary to achieve long-term colonisation.

Findings from animal models about bacterial and host factors related to colonisation and infection by S. aureus cannot be fully extrapolated to humans because of major physiological host differences. Only one human colonisation model was ever developed. Previously, the role of the host factors in acquiring persistent or transient S. aureus nasal carriage were investigated using nasal inoculation studies in healthy human volunteers. Results showed that most human carriers preferentially carry one lineage of S. aureus, strongly suggesting a specific host component associated with carriage. The specific interactions between S. aureus and the human nose are currently unknown.

A significant difficulty in the development of nosocomial infection control technology is the lack of protocols to assess efficiency of decontamination and decolonisation methods. A majority of products have not been validated sufficiently. Studies in humans are inhibited by ethical and organisational constraints and experimental animal models have thus far failed to find non-human species that are naturally colonised by MRSA. The MRSA ST398-colonised pig provides an opportunity to develop appropriate models to assess control strategies in chambers inhabited by colonised animals and test novel, safe methods for environmental decontamination and maintenance of decolonisation.

The EU-funded project PILGRIM investigated MRSA. MRSA occur in hospitals throughout Europe. It can also be found in animals from where it can spread to humans and through the community, and sometimes reach health care settings. In order to develop measures and tools for reducing infections caused by MRSA in humans, PILGRIM aimed to contribute to better understand their epidemiology and spread in different environments. The PILGRIM project was designed to provide novel control measures for the accelerated identification and control of resistant bacteria initially emerging from animals in order to prevent and eradicate community and nosocomial infections. The central PILGRIM hypothesis was that deeper understanding of factors affecting pathogen-host interaction of resistant bacteria at the molecular as well as the population level will lead to new and more effective control measures against nosocomial infection.

PILGRIM focused on research using an animal-adapted strain, i.e. MRSA ST398, which required immediate attention due to its recent rapid spread and impact on public health. PIGLRIM work provided new control measures taking into account new evidence on factors impacting on transmission, colonisation, host range and virulence of resistant nosocomial strains.

The PILGRIM project was organised into four research WPs, one dissemination and exploitation WP and the project management WP. The research work was organised around methodological approaches, including epidemiological and physiological studies, in vivo and in vitro experiments as well as molecular-genetic analysis:

WP1 was to provide quantification of the transmission of MRSA ST398 between animals and humans and the risk for health care settings. The data collected in this WP were used in WP4.

WP2 provided in vivo and in vitro models to study factors relevant for the colonisation, host range and virulence of MRSA. WP2 interlinked with WP3 and provided the animal models used in WP4.

WP3, in close interaction with WP2, identified key genetic components of the specific interaction between S. aureus ST398 and its hosts. Findings were fed back to WP2 to test identified factors using knock-out technology. Additionally, WP3 identified specific genes for the development of new rapid tests to identify specific MRSA strains and potentially for bed-side testing.

WP4 developed novel intervention technology based on decolonisation and environmental sanitation using photocatalysis and electrochemically generated charged metal ion solutions as well as antagonism. It was with WP5 to provide the technical components of TTP.

WP5 integrated findings to provide and communicate a range of recommendations and guidelines for risk management, including the proposition of novel control strategies. The TTP from WP4 was made available to companies for validation of existing and new solutions targeted at nosocomial infections under standardised and realistic conditions.

WP6 was to assure efficient scientific and administrational management and effective internal and external cooperation.

Project results:

Note: The foregrounds and scientific and technological results are reported by WP and tasks:

WP1: Epidemiology and ecology of ST398 in farms and community environments

At the start of the PILGRIM project available data on LA-MRSA (ST398-MRSA) mainly came from the Netherlands since the problem was first recognised by one of the PILGRIM partners. Combined human and veterinary research by a national consortium was already on-going in the Netherlands. As a consequence some of the research planned for PILGRIM was not repeated for the Netherlands, but comparable methods were used in Belgium and Denmark. In general, the topic of livestock as a source of multi-drug resistance become so critical that PILGRIM researched was paralleled by other national and international research initiatives.

- to determine the incidence of colonisation and mode of spread of MRSA ST398 in pigs and from pigs to humans by direct and indirect contact with pigs or pigfarmers / pigfarm workers;
- to determine the reproductive number of MRSA ST 398 among humans in the community and healthcare settings to guide future infection control strategies.

T1-1: Longitudinal epidemiology and transmission of MRSA ST398 in animals and humans

Task 1-1a: Transmission of MRSA ST398 between animals - Results PILGRIM:
Sows and their offspring were sampled at varying intervals during a production cycle. Overall MRSA prevalence of sows increased from 33 % before farrowing to 77 % before weaning. Overall MRSA prevalence of piglets was above 60 % during the entire study period. The recurrent finding of MRSA in the majority of individuals indicates true colonisation or might be the result of contamination. Transmission rates were estimated using a Susceptible infectious susceptible (SIS) model, which resulted in values of the reproduction ratio (R0) varying from 0.24 to 8.08. Transmission rates were higher in pigs treated with tetracyclins and lactams compared to untreated pigs implying a selective advantage of MRSA CC398 when these antimicrobials are used. Furthermore, transmission rates were higher in pre-weaning pigs compared to post-weaning pigs which might be explained by an age-related susceptibility or the presence of the sow as a primary source of MRSA CC398. Finally, transmission rates increased with the relative increase of the infection pressure within the pen compared to the total infection pressure, implying that within-pen transmission is a more important route compared to between-pen transmission and transmission through environmental exposure. In conclusion, our results indicate that MRSA CC398 is able to spread and persist in pig herds, resulting in an endemic situation. Transmission rates are affected by the use of selective antimicrobials and by the age and location of pigs. Future research: Role of other livestock animals, especially calves and poultry. Most importantly methods to prevent the spread of ST398 MRSA between or from animals need to be investigated such as: antimicrobial stewardship, environmental cleaning and dust reduction, use of probiotics, implementing ?infection control measures, such as cohorting of animals

Task 1.1.b: Transmission of MRSA ST398 to and between animal care takers - Results PILGRIM:
Persistent carriage of livestock associated MRSA is highly prevalent among pig farmers and their employees in Belgium, Denmark and The Netherlands (100, 56 and 60 %, respectively). Household members seem to have a lower risk (29, 5 and 6 %, respectively), possibly due to the low transmissibility of this livestock associated clone outside the pig stable. Specific risk moments, and the exact role of the environment remains to be elucidated, but seen the rate of contamination the home environment may have a potential role in transmission (9 / 12 farm home environments MRSA positive = 75 %). Future research: Role of the environment and of companion animals as a key factor into the transmission of ST398. Furthermore, the environmental contamination in the wider range around farms as a source of ST398 for community spread.

Task 1.1.c Vertical transmission of MRSA ST398 We investigated the role of pig trade in the transmission of MRSA CC398 between farms using PFGE, a highly discriminatory method for strain typing. PFGE analysis of 57 MRSA isolates from a retrospective study in the Netherlands and a prospective study in Denmark provided molecular evidence that the strains present in 5 of the 8 recipient farms were indistinguishable from those occurring in the corresponding supplying farm. Our molecular typing data support the conclusions by Broens and others (2011) by providing molecular evidence that MRSA CC398 can be transmitted by pig trade. This notion has important implications for the development of intervention strategies for control and prevention of MRSA CC398 in pig farming. Future research: Analysis of alternative transmission routes is needed to identify other risk factors and ultimately to design effective measures to control transmission of MRSA CC398 between farms.

T1-2: Transmissibility of ST398 in humans in the non-farming and healthcare environment

Task 1-2a: Non-farming setting - Results PILGRIM:
Veterinarians are at high risk to acquired LA-MRSA when they are exposed to pigs, more particularly to live pigs, rather than other animal species. LA-MRSA prevalence in veterinarians was higher in Belgium but not in Denmark compared to their LA-MRSA prevalence found in the general population. Veterinarians can become persistent LA-MRSA carriers for several months up to a year and a large proportion could be considered as long term carriers. LA-MRSA has a limited but not negligible capacity of transmission inside the household in a non-farm environment. Interestingly, veterinarian's partners were the only relative to become, persistent nasal carriers. Future research: Role and impact of the veterinarians as a vehicle of introduction of LA-MRSA into the community and comparable research into other professional groups in contact with ST398 MRSA positive livestock (e.g. slaughterhouse, transportation)

Task 1-2b: Follow-up of chronic ST398 carriers - Results PILGRIM:
The overall carriage rates of MRSA and MSSA CC398 among adult individuals with no history of livestock exposure but residing in communities with high densities of pig production were 0.4% and 0.9%, respectively, with important geographic variations. In Belgium, the carriage rate of MRSA CC398 was above 6-fold higher than those observed in the cohorts from Denmark and the Netherlands, which correlates with the distribution of pig production holdings positive for MRSA CC398.

Future research

We are currently investigating whether the emergence of S. aureus CC398 isolates in the community is due to coincidental spread from the animal agriculture setting or if these isolates have evolved adaptations to the human host that facilitate human-to-human transmission.

Task 1-2c Healthcare setting - Results PILGRIM:
The prevalence of MRSA ST 398 in nursing homes was 0 %. Two (0.7 %) HCWs carried non-livestock-associated MRSA-types, namely t031 and t324. That result was unexpected and probably due to inclusion bias. Therefore a different approach, namely an anonymously web-based short questionnaire spread by mail was chosen. The prevalence rate of contact with pigs and / or veal calves was significantly higher in the web-based questionnaire (6.7 % versus 3.7 %). We conclude that MRSA carriership of NT-MRSA and contact with pigs and / or veal calves does not appear frequently among HCWs in nursing homes. The rate of HCWs with contact with pigs and / or veal calves is not complete pointed out in this study, but there seems not to be a large problem with regard to the risk for residents in nursing homes at this moment. Future research: Changes in transmissibility and virulence of ST398, may lead to an increasing risk for patients in healthcare settings. Furthermore, the number of HCWs that are or become colonised with ST398 MRSA will increase and many of these are probably not decolonisable. The Dutch national guidelines will allow in the future these HCWs to work under restricted circumstances and thereby offer a unique opportunity for research into the transmission of ST398 MRSA. Conclusions PILGRIM WP1 contributed to novel insights into the epidemiology of ST398 MRSA in animals and humans but furthermore raised new questions for future research. It is evident that only combined research and efforts of farmers, veterinarians and human medicine will be able to (better) control this emerging problem.

WP2: Comparative host-range and colonisation of MRSA ST398 and other MRSA clones

The aim of WP2 was to elucidate the comparative host range and colonisation properties of ST398 and other MRSA clones by means of in vitro and in vivo experimental models. The work summarised in this report has resulted in two scientific publications in epidemiology and infection (Moodley and Guardabassi 2010, vol. 15, pp. 1-7) and veterinary microbiology (Moodley et al. 2011, vol. 152, pp. 420-423). Two additional manuscripts are presently in preparation and will be submitted for publication in 2012.

Comparative host range using in vitro, pig and human models

Despite S. aureus is generally regarded as a host-specific organism, there is increasing evidence that specific clonal lineages such as ST398 have an extended-host range and are able to adapt to both humans and animals. This notion was confirmed by the results of the experiments in WP2, where two of the most common S. aureus lineages in pigs (ST398 and ST433) and three lineages typically associated to humans (ST8 (USA300), ST22 (EMRSA-15), and ST36 (EMRSA-16)) were investigated for their ability to adhere to human and pig corneocytes (T1-1a) and to colonise pigs (T1-1c). In the corneocyte adherence assay, the pig-associated ST433 strain and the human-associated ST22 and ST36 strains showed significantly greater adhesion to porcine and human corneocytes, respectively (p < 0.0001). On the contrary, ST8 and ST398 did not display preferential host binding patterns, indicating that these two lineages may have an increased ability to adapt to heterologous hosts. Surprisingly, ST398 was outcompeted by ST8 in the pig experiment, revealing an unexpected ability to colonise pigs by this human-associated clone. The broad host spectrum and ability to adapt to humans of ST398 was confirmed by the human experiment. In the last 10 years in Dutch people an increase in the number of MRSA is observed. This increase is primarily due to MRSA of sequence type 398 (ST398) and concerns in most cases people who have been in contact with farm animals or livestock. We hypothesised that it could very well be that ST398 MRSA is frequently contaminating not colonising the nose of especially farm workers who are intensively in contact with livestock. We inoculated healthy human volunteers with a mixture of the human S. aureus strain 1036 (ST931, CC8) and the calf S. aureus strain 5062 (ST398, CC398). We studied survival by follow-up cultures over 21 days. A rapid decrease in bacterial load was observed for both strains in the first days after inoculation. The human strain 1036 was eliminated faster (median 14 days; range 2 - 21 days) than the animal strain 5062 (median 21 days; range 7 - 21 days) but this difference was not significant (p = 0.065) yet the bacterial loads were significantly higher for the animal strain on day 7 and day 21 (resp. p = 0.012 and p = 0.015). Two main trends in elimination of the inoculated strains were observed. One group (4 / 14) showed elimination of both strains within 21 days, while the others did not (10 / 14). Interestingly this second group can be subdivided in two additional groups for in half of them (5 / 10) no differences in bacterial counts between both strains until the end of follow-up was observed. On the contrary in the remaining half (5 / 10 the ST398 strain was out performing the human S. aureus strain. In conclusion, MSSA ST398 strain 5062 of calf origin (spa-type t034) is capable of persisting in the human nose for 21 days.

In vivo colonisation

In vivo research on S. aureus has largely relied on the use of murine models in the past. However, the mouse is not an ideal organism for modelling S. aureus colonisation since it has previously been shown that S. aureus does not easily grow and multiply in the nasal cavity of mice. In task T2-2a, we developed an efficient model to colonise pigs (T2-2a). We chose pigs as model organisms due to the fact that S. aureus is a natural coloniser of these animals and porcine skin has anatomical and physiological similarities to human skin. Colonisation was obtained by implantation of MRSA in the vagina of pregnant sows shortly before farrowing, which resulted in the delivery of stably colonised piglets. MRSA carriage was shown to persist over a period of 28 days, suggesting that the piglets were truly colonised. The model, which is relatively easy to perform and does not require invasive procedures, was successfully employed to study host-specificity of different S. aureus lineages in task T2-1c. In the future it may be a useful tool to test the efficacy of MRSA decolonisation and decontamination control strategies in pig farming.

In task T2-2b, we investigated MRSA antagonism by maternal vaginal flora. This study developed a gnotobiotic pig model to study colonisation kinetics of MRSA and potential bacterial interference by Coagulase negative staphylococci (CNS). Groups of two-week-old piglets were atraumatically inoculated either with MRSA only or with two alternative inoculation patterns with MRSA or selected CNS (Staphylococcus warnerii, S. xylosus and S. chromogenes) preceding each inoculation by two days. Animals were swabbed periodically and bacterial counts compared over a period of 23 days from nasal mucosae and skin behind the ears and from the sacrum, using MRSA selective media. Inoculation of MRSA on piglet skin resulted in spontaneous colonisation and populations increased similarly for MRSA and CNS until day 32. However, a reduction in final counts on day 37 was greater for MRSA, suggesting that bacterial interference from these CNS may have occurred. Overall, these studies demonstrate that gnotobiotic piglets provide a model suitable for bacterial interference studies, which should be further explored as an important alternative in the control of MRSA.

The aim of Task T2-2c was to study in vivo gene expression during colonisation: These experiments would identify the likely gene expression pathways necessary for successful colonisation, and provide useful information about targets for interventions blocking colonisation. The strategy involved two sources of colonising MRSA; firstly, swabbing naturally colonised pigs at UCPH, and the second was swabbing artificially colonised gnotobiotic piglets at RVC. The swabs were placed immediately in RNA stabilising buffer, and Ribonucleic acid (RNA) extracted. Contaminating eukaryotic messenger RNA (Mrna) was removed using the MicrobeEnrich method (Invitrogen, United Kingdom). As expected, the concentration of bacterial mRNA obtained was low, and the MessageAmp Bacteria Kit from Ambion was used to amplify the signal. Total mRNA was labelled with fluorescent probes and hybridised to a custom designed 62-strain whole genome S. aureus microarray constructed in WP3. Although controls using in vitro derived mRNA gave excellent results, we were unable to generate good quality data from in vivo derived mRNA, likely due to insufficient quantity and quality of nucleic acid. We expect future improvements to amplification technologies in the near future will make this experiment feasible. In the meantime, results generated in WP3 provided excellent CC398 targets for future intervention therapies to reduce colonisation. Previous studies have hypothesised that the emergence and spread of MRSA ST398 in pigs may be driven by the use of tetracycline and zinc oxide.

In T2-2d, we assessed the effects of tetracycline and zinc on pig nasal colonisation and on transmission from MRSA-positive to MRSA-negative animals housed within the same pen. Significantly higher nasal MRSA counts were observed in zinc-treated (p=0.015) and tetracycline-treated (p=0.008) pigs compared to non-treated animals. However, transmission of MRSA from positive to negative animals housed within the same pen was not influenced by exposure to these agents. These results indicate that feed supplemented with tetracycline or zinc increases the numbers of MRSA ST398 in the nasal cavity of pigs. Although it is reasonable to assume that higher levels of MRSA in the nasal cavity may facilitate animal-to-animal spread, the epidemiological significance of these results is uncertain since the efficiency of MRSA spread within farms may not be related to the numbers of nasal MRSA in individual pigs. However, it cannot be excluded that prolonged use of these agents over time may result in a cumulative effect and have an impact on MRSA transmission by increasing the levels of environmental contamination.

WP3: Molecular biology of host-pathogen interaction ST398 is predominantly associated with colonisation of pigs, but the isolates can colonise and spread to humans and also cause infection. This is in contrast to the major human lineages of S. aureus which colonise and infect humans but are rarely found in pigs. The specific interaction between S. aureus and these different mammalian hosts was unknown. It is important to understand these interactions in order to predict how these strains might spread and evolve. The first aim of this WP was to identify key genetic components of the specific interaction between S. aureus ST398 and its pig and human hosts, as well as to reduce colonisation and subsequent infection in pigs and humans using a vaccine or drug requires validated targets.

The second aim is therefore to develop this information to a level that could be exploited for development of vaccines or drugs to control colonisation in pigs and humans Identification of ST398 requires culture, and specialised molecular methods only available at dedicated research laboratories or regional expert laboratories with extensive experience. This holds back our understanding of epidemiology, spread, diagnosis and treatment, and ability to test intervention strategies. The third aim was to develop rapid and inexpensive diagnostics to identify ST398 in the laboratory and potentially in the field.

T3-1: Identify key components of specific MRSA ST398 mammalian host interaction

In this task, the relevant technology was developed multi-strain microarrays, bioinformatic analysis tools ? and appropriate strains were collected in order to compare comprehensively isolates of ST398 with differing colonisation and infection capacities in a range of countries. The project identified that there were two major types of ST398. Firstly, the pig clade, which is highly successful at colonising pigs, and can colonise humans but generally does not spread between humans. The pig clade causes infection in humans. We identified the best marker of the pig clade as being resistance to tetracycline antibiotics, and these isolates also frequently carried resistance to trimethoprim and a bacteriophage phi6. The second clade is the human clade, which is less prevalent. It colonises humans and spreads between humans. The best marker of this clade is the bacteriophage phi3, and in some countries there is a strong association with bacteriophage phi7. Significantly, we identified several isolates of ST398 that belonged to the pig clade but that appeared to have spread between humans, and this was associated with acquisition of the bacteriophage phi7. This occurred in Denmark. Thus we are watching the abundant ST398 pig clade evolve to spread between humans and this is of great concern. Our data also shows that each country has differences in their own isolates, so that although there is substantial variation, there is also some geographical variation. This may be useful in the longer term to identify differences in the ability to spread, cause disease or resist antibiotics between countries, and also to track how isolates evolve and spread between countries. Importantly, the key differences between the pig and human clades are consistent between countries, as well as continents we included isolates from Europe, as well as the Americas and the Carribean. We used complex bioinformatics analysis and data from microarray and sequencing projects to identify the differences between isolates colonising and infecting animals versus humans. The initial hypothesis was that the major differences would be in surface proteins that bind to host tissue. However, our findings did not wholly support this argument. Suprisingly, human and animal isolates of S. aureus have a wide variety of these surface proteins, but differ in their combinations of these proteins, rather than the proteins themselves. This is unexpected as the binding targets for these proteins in different hosts are expected to be different. Nevertheless, we identified one key protein, that we call here proteinP, that showed important differences between hosts.

T3-2: Prove the role of candidate microbial genes in host-specific interactions The knockout strain of proteinP was constructed in the ST398 background. The mutant strain and its matching wild-type strain have been compared in a gnotobiotic piglet colonisation model. The experiment is completed and data is being analysed now. The results are promising, and this is essential background knowledge in order to take this protein forward as a target for therapeutics.

T3-3: Develop diagnostic tests

A gene specific for the ST398 group was identified, which is a variant of the type I restriction modification specificity subunit gene, hsdS. In this project we developed a PCR test for this gene. This test accurately assigned 100% of isolates as ST398 (n=1032) or not (n=275) despite originating from a range of hosts. The test could be combined with a PCR test for the mecA gene to identify the isolate was MRSA. The test could also be adapted to the real-time PCR format, allowing the test to be automated and with potentially integrated into commercially available platforms in standard diagnostic microbiology laboratories or in portable molecular diagnostic devices for use in the field. A patent on this finding was filed with an international filing data of June 2010 by St George?s University of London, Statens Serum Institute and University of Copenhagen. We are in talks with potential commercial partners. We also searched for potential surface targets for an antibody based dip-stick type test for ST398. As outlined above, there were few targets except one, proteinP. This will be followed for its potential applicability. In addition, dipstick tests are increasingly able to detect molecular targets due to advances in technology to amplify the signal of a small amount of specific DNA. The hsdS gene variant is a promising candidate for such testing.

Summary of remaining gaps and future research needs

1. The potential for pig clades to evolve and spread to humans is concerning, and needs to be followed using appropriate molecular methods such as those that we have described. This can identify where this evolution is happening, and whether we can or should seek to control it. The strong association between pig clade ST398 and tetracycline resistance genes is a strong indicator that antibiotic policies are selecting for these MRSA with consequent impact on human health. These factors need urgent research data to support the reduction in antibiotic use in pig farming. The fact that S. aureus surface proteins that bind to host tissue were not the major differences between animal and human isolates was unexpected. The differences were more closely aligned with strategies for evading the immune response. Little is known about the immune response in humans and animals to S. aureus, but this could be a potential area for research in order to develop better strategies for reducing colonisation and infection.
2. ProteinP is a potential target for intervention therapies, vaccines or as a drug target to reduce colonisation in pigs. It needs investment to develop this potential.
3. Commercialisation will require studies to show usefulness of the test which may or may not be funded by commercial partners. Investment to develop a dip-stick test will be needed to bring this to the level ready to commercialise.

WP4: Defining intervention strategies and developing control measures

WP4 aimed to define intervention strategies and develop control measures for MRSA ST398 making use of results relating to colonisation from other WPs. At the beginning of the project, MRSA ST398 was already well-recognised for its ability to colonise pigs and other farm animals and there were considerable data showing that it could spread from farms into associated human populations. Because it was adapted to pigs, it was appropriate to develop porcine models of colonisation, infection and intervention with this organism. A model was developed firstly in gnotobiotic piglets because previous studies had demonstrated that they were appropriate for the investigation of colonisation by pathogenic staphylococci, and because their germfree status enabled interaction and possible bacterial interference of selected staphylococcal clones with MRSA ST398 to be studied in the absence of other microbes. Intervention models were subsequently developed using conventionally reared piglets naturally colonised with MRSA ST398 in farm style chambers so that the effects of active environmental decontamination on MRSA colonisation could be studied. The selected decontamination systems were provided by two SME companies (Aguacure Ltd and Virobuster GmbH) and one Pilgrim partner (ICT, Prague). They included photocatalytic paint on the chamber walls, air cleansing with ultraviolet subtype C (UVC) radiation, and misting with electrochemically generated charged metal ion solutions. When exposed to UV subtype A (UVA) radiation photocatalytic paint was known to have an antimicrobial and self-cleansing effect. UVC radiation was known to be able to kill bacteria suspended in the air. The charged metal ion solutions had already been shown to have potent antimicrobial activity. Demonstration of the efficacy of these interventions in vitro and ex vivo within the laboratory, and in a farm style chamber with naturally MRSA colonised pigs was expected to provide avenues for their use in MRSA contaminated environments. The studies also aimed to develop methodology for the provision of a TTP which could be offered for development and testing of other forms of disinfection technology in WP5.

T4-1: Efficacy assessment of novel antimicrobial agents for MRSA-decontamination

The antimicrobial agents and processes were tested in vivo and ex vivo on porcine skin at RVC within a custom made box designed and constructed at ICT. The UV subtype A (UVA) field delivered within the box was tested and standardised. Photocatalytic coatings and paints supplied and developed by ICT were then assessed by RVC for activity against single and mixed porcine isolates of MRSA ST398 spa types t011, t034 and t108, and MRSA ST9 t889, a novel lineage associated with pig farming. Inactivation of the MRSA bacteria on the tested surfaces both in suspension and after air drying was demonstrated. The most efficient candidate surfaces were white photocatalytic facade paints suitable for use in the subsequent studies of environmental decontamination in farm style chambers. An air sterilisation device based on the activity of UVC radiation on air pumped through a cylinder was developed by ICT. Initial tests on this at RVC using the box indicated that it would be costly and difficult to employ in the planned farm style chamber decontamination studies. A commercially available device manufactured by Virobuster GmbH was found to be a suitable replacement with known efficacy against S. aureus. This was provided and maintained in subsequent studies by Virobuster. A misting device was provided and installed in the box by Aguacure Ltd. Preparations of the Aguacure charged metal ion solutions delivered as a mist were tested and shown to have bactericidal efficacy against the MRSA. The most effective of these was chosen for further study in the farm style chambers.

T4-2: Decontamination and decolonisation efficiency testing

Environmental decontamination was studied in two farm style chambers each occupied by three two-week old piglets, naturally colonised with MRSA ST398. In one chamber, the ability of photocatalytic paint, the air cleansing device and misting with the metal ion Aguacure solution, to reduce or eliminate environmental contamination with the MRSA and coliform bacteria was determined. Each method was effective against both MRSA and coliform bacteria when applied over a five-day period. Recontamination of the chamber with MRSA from the piglets occurred following each treatment period except after misting with Aguacure solution, when MRSA was eliminated from both the environment and the piglets. In the second chamber, no special decontamination processes were carried out but the chamber was kept clean. Lower levels of piglet colonisation were present and less contamination occurred; decolonisation of these piglets occurred spontaneously over a seven-week period. A study of decolonisation of MRSA ST398 colonised pigs using antimicrobial drugs had been planned, but was not carried out after it had been shown that decolonisation occurred without treatment in piglets held in clean and hygienic conditions. The technology and expertise developed in T4-1 and T4-2 provided a model that was suitable for exploitation as a TTP. This is being marketed to industry in WP5.

T4-3: Development of a model of MRSA antagonism in vivo

Three groups of gnotobiotic piglets, born by aseptic caesarean section, were housed in sterile plastic isolators. They were inoculated on skin in the sacral region with MRSA ST398 spa-type t011 and with staphylococcal antagonists (502A and J94) that had demonstrated antagonism against ST398 in well diffusion assays. Different inoculation patterns were examined in the three groups of piglets: A. MRSA only at 14-days of age (control group, n=4), B. MRSA day 14 followed by 502A at day 16 (n=5) and C. MRSA day 14 followed by J94 at day 16 (n=4). Piglets were subsequently sampled at day 16 (before the second inoculation), and periodically for 23 days. There was no preferential interference of strains 502A and J94 against MRSA ST398.

T4-4: Assessing the relative efficiency of interventions at population level

A framework was developed to model two processes, namely the spread of MRSA ST398 within a human population exposed to a farm with carrier pigs and the spread of MRSA ST398 within a pig farm consisting of different compartments. The framework was parameterised using data from the literature and from WP1 of PILGRIM. The model was validated and a sensitivity analysis was conducted to establish robustness. Possible intervention scenarios were explored based on the most sensitive parameters in the model to inform suitable intervention scenarios. This included results from experimental studies conducted as part of Task 4.4. Exposure to pigs carrying MRSA ST398 was shown to be the most important factor promoting spread in the human community. Summary of remaining gaps and future research needs It is clear that decontamination systems can be installed in farm style accommodation and the pilot studies carried out in PILGRIM indicate that these can accelerate processes of environmental cleansing and decolonisation in piglets that are naturally colonised with MRSA ST398. Further studies are now required to determine whether these effects can be achieved on farms, and to examine the need for and effects of combining different decontamination processes. The failure to demonstrate antagonisation of MRSA by the staphylococcal antagonists, 502A and J94, and the observation that colonisation of porcine skin with these bacteria may have enhanced the growth of MRSA ST398 requires further study. These findings may indicate that MRSA ST398 is particularly well adapted to porcine skin. Future studies of potential bacterial antagonism may need to employ pig adapted antagonists. Further work is required to expand the validity of the models developed to examine the spread of MRSA ST398 within human populations exposed to farms with carrier pigs and the spread of MRSA ST398 within a pig farm consisting of different compartments. Additional studies of the transmissibility of MRSA ST398 from livestock to humans are required particularly in relation to the difference between MRSA strains, and farm husbandry practices.

Potential impact:

Strategic impact: The PILGRIM project has provided three types of outputs, namely scientific evidence, technology, and engagement

1. Reinforced research integration: PILGRIM research has provided an integrated research platform for the investigation of transmission dynamics and colonisation processes for MRSA ST398 specifically and MRSA in general. The research approaches, study designs, experimental protocols and molecular methodologies developed as part of the PILGRIM work programme have been successfully applied in the wider context of research into antimicrobial resistance mechanisms. The impact has been achieved through a large number of publications and presentations (see other section of the final report and list of publications). PILGRIM has provided an integrated research and evaluation platform to investigate the epidemiology of resistant bacteria emerging in the community and causing nosocomial infections. The creation of integrated human-veterinary research network has provided the basis for accelerated control of newly emerging resistant strains with zoonotic potential. Through TTP, research has been integrated in the evaluation process of new technology targeted at the control of resistant bacteria in healthcare settings. This has allowed assessing new control strategies using a validated, evidence-based approach, and the integration has accelerated the assessment and implementation of cost-effective control strategies.

2. Identification of new targets and development of novel control instruments: Detailed studies of MRSA ST398 host-pathogen interaction have been executed by the PILGRIM consortium. These have not only provides scientific evidence of factors influential on colonisation and host range, but have also led to the identification of targets for intervening with S. aureus colonisation and infection. PILGRIM has additionally provided new instruments for rapid identification of MRSA lineages using PCR technology. This outcome in combination with patented dip-stick test technology (patented by UCPH, SGUL and SSI) may lead to rapid, cheap, bed-side testing. The use of this technology could impact on the cost of MRSA screening. More tests can be conducted in shorter time contributing to early detection and accelerated control of nosocomial MRSA infections. Additionally, PILGRIM developments have provides a TTP resolving the bottle neck which currently prevents new products to be tested and approved. The TTP particularly supports SMEs who do not have the means to develop realistic testing environments of their own. The TTP also assures evidence-based evaluation of new technology and accelerated release of cost-effective products to the market.

3. Policy and societal impact: Scientific evidence provided by PILGRIM research has been translated into guidelines and recommendations for healthcare workers, persons specifically at risk of exposure to MRSA ST398 and the public in general. This has been achieved by public and professional engagement. Dissemination activities have been specifically targeted at the different audiences. Multipliers such as professional organisations were used to achieve maxi-mum outreach. A focus was put on Member States currently affected by MRSA ST398-related health issues, but instruments would also be applicable in a wider context. General awareness of society towards new resistant bacteria and nosocomial infection has been promoted. Impact on policy and society has been assured in PILGRIM through dissemination activities targeted at policy level. An approach coordinated across Europe is still required as colonised persons can carry resistant bacteria into previously not-affected regions. The majority of new and emerging pathogens are of zoonotic nature. Similarly, new resistant strains are likely to emerge from animal reservoirs. PILGRIM provides not only a scientific framework for investigating such pathogens but also discusses their relevance and increases awareness and understanding in society. This is relevant in the context of compliance with control strategies and risk management in general. The dissemination report contains detailed information of the dissemination activities executed.

List of websites: