Preventive Veterinary Medicine 118 (2015) 189–195
Contents lists available at ScienceDirect
Preventive Veterinary Medicine journal homepage: www.elsevier.com/locate/prevetmed
Starting from the bench—Prevention and control of foodborne and zoonotic diseases Kitiya Vongkamjan a , Martin Wiedmann b,∗ a b
Department of Food Technology, Prince of Songkla University, Hat Yai 90112, Thailand Department of Food Science, Cornell University, Ithaca, NY 14850, USA
a r t i c l e
i n f o
Article history: Received 3 February 2014 Received in revised form 30 October 2014 Accepted 2 November 2014 Keywords: Foodborne and zoonotic pathogens Subtyping Molecular epidemiology
a b s t r a c t Foodborne diseases are estimated to cause around 50 million disease cases and 3000 deaths a year in the US. Worldwide, food and waterborne diseases are estimated to cause more than 2 million deaths per year. Lab-based research is a key component of efforts to prevent and control foodborne diseases. Over the last two decades, molecular characterization of pathogen isolates has emerged as a key component of foodborne and zoonotic disease prevention and control. Characterization methods have evolved from banding pattern-based subtyping methods to sequenced-based approaches, including full genome sequencing. Molecular subtyping methods not only play a key role for characterizing pathogen transmission and detection of disease outbreaks, but also allow for identification of clonal pathogen groups that show distinct transmission characteristics. Importantly, the data generated from molecular characterization of foodborne pathogens also represent critical inputs for epidemiological and modeling studies. Continued and enhanced collaborations between infectious disease related laboratory sciences and epidemiologists, modelers, and other quantitative scientists will be critical to a One-Health approach that delivers societal benefits, including improved surveillance systems and prevention approaches for zoonotic and foodborne pathogens. © 2014 Elsevier B.V. All rights reserved.
1. Scope of microbial foodborne diseases in the US Diseases caused by foodborne pathogens are a major public health concern in the US. The US Centers for Disease Control and Prevention (CDC) recently reported estimates of foodborne illness in the US (CDC, 2011a), which indicated that about 47.8 million cases of foodborne illnesses, most of them representing gastrointestinal symptoms, occur annually in the US. Approximately 1 in 6 Americans thus experience a foodborne illness episode each year. For approximately 80% of these cases (38.4 million cases), the causative agent is unknown; on the other hand, 20% of
∗ Corresponding author. Tel.: +1 607 254 2838; fax: +1 607 254 4868. E-mail address: [email protected] (M. Wiedmann). http://dx.doi.org/10.1016/j.prevetmed.2014.11.004 0167-5877/© 2014 Elsevier B.V. All rights reserved.
cases (9.4 million cases) are estimated to be caused by 31 known pathogens. Overall, foodborne illnesses were estimated to result in 127,000 cases of hospitalization and 3037 deaths each year in the US. Of the hospitalization cases and deaths, over 50% of the cases are estimated to be due to unknown causative agents. The top five pathogens responsible for foodborne illnesses include Norovirus, followed by four bacterial pathogens (i.e., Salmonella, Clostridium perfringens, Campylobacter spp., Staphylococcus aureus). Salmonella is the leading cause of hospitalizations (19,336 annually) and deaths (378 annually), while Listeria monocytogenes is responsible for the second most foodborne deaths (255 annually) due to bacterial pathogens. The World Health Organization (WHO) asserts that food and water-borne diarrhoeal illnesses present a “growing public health problem” which is estimated to affect 1 in 3
190
K. Vongkamjan, M. Wiedmann / Preventive Veterinary Medicine 118 (2015) 189–195
people annually worldwide, and result in 2.2 million deaths globally each year. The majority of these deaths (1.9 million) occur in children less than 5 years of age (WHO, 1999). Diseases transmitted through food thus are of importance for both developed and developing countries with many of the key bacterial pathogens attributed to foodborne illnesses in the US also commonly causing disease in developing countries. While some progress in reduction of foodborne diseases has been made, usually successes have been limited to reducing transmission of a specific pathogen through a specific commodity or food chain, often in only one country or region. For example, while considerable progress has been made to reduce L. monocytogenes prevalence in readyto-eat (RTE) deli meats in the US since about 1990 (FSIS, 2010), the number of reported listeriosis cases in many European countries have increased in recent years (Goulet et al., 2008). In addition, since 2010 a number of listeriosis outbreaks in the US have been linked to produce (CDC, 2013), including the most deadly listeriosis outbreak in almost 90 years in the US, which caused 33 deaths and was linked to consumption of contaminated cantaloupe in 2011 (CDC, 2011b). Consequently, our thesis is that long term progress in reducing foodborne illness burden worldwide will require a combination of laboratory approaches and field-based and epidemiological studies, including training of future professionals that can bridge these disciplines. 2. Laboratory-based subtyping techniques and their application to detect disease outbreaks and characterize transmission routes One key area where the importance and impact of combining appropriate (and advanced) laboratory methods with epidemiological and field studies has been well established is in foodborne disease surveillance. This section will briefly describe some of the methods that are used for subtyping (strain differentiation) of foodborne pathogens, with a focus on bacterial pathogens, and then will detail how these methods are being applied in the context of national and international foodborne disease surveillance systems. The learnings from these surveillance systems that utilize laboratory methods for informing and facilitating epidemiological surveillance will likely provide insights that will facilitate better integration of laboratory methods and epidemiology in various areas of public health and related research, which will facilitate broad improvements in public health and food safety. 2.1. Phenotypic and genotypic subtyping techniques Molecular subtyping methods can be classified into both phenotypic and genotypic methods (Wiedmann, 2002). Briefly, subtyping methods allow for differentiation of pathogen isolates into distinct subtypes or strains; these data are useful for determination of the relatedness of bacterial isolates and to further establish linkages between illnesses occurring in an outbreak and sources or transmission routes of the pathogen. While the term “subtype” refers to isolates that share common genetic characteristics (as determined by genotypic methods), the term
“strain” refers to isolates that share common phenotypic characteristics (as determined by phenotypic methods); these two terms are often used interchangeable though (Wiedmann, 2002). To detect disease outbreaks, a typing method should have sufficient discriminatory power to distinguish epidemiologically related isolates from epidemiologically unrelated isolates. Phenotypic methods, e.g., phage-typing, may not only be plagued by poor reproducibility, but also often show insufficient discriminatory power to be valuable in epidemiological studies or for outbreak surveillance (Graves et al., 2003). Nevertheless, a number of phenotypic subtyping methods have played important roles in foodborne, animal and human disease surveillance in the past. For example, wide-spread use of Salmonella serotyping has been essential not only for surveillance, but also for improving our understanding of host specificity, diversity, and transmission pathways for this important zoonotic pathogen (Shi et al., 2013). Genotypic subtyping methods can be classified as “banding-based” and “sequence-based” methods; methods representing both of these categories have also been used widely for characterization, identification, and typing of bacterial isolates, including to assist in surveillance and epidemiological investigation and to define clones, i.e., groups of genetically similar isolates descending from a common ancestor. Among the “banding-based” subtyping methods, pulsed-field gel electrophoresis (PFGE) is often considered the gold standard typing method, including for a number of foodborne bacterial pathogens where PFGE shows a high discriminatory power for most strains (e.g., L. monocytogenes, many Salmonella serotypes) (Graves et al., 2003). Other banding pattern based typing methods that have commonly been used to characterize foodborne pathogens include ribotyping as well as REP-PCR (Jerˇsek et al., 1999; Sauders et al., 2003). Sequence-based subtyping methods include multi-locus variable number tandem repeat analysis (MLVA) and multilocus sequence typing (MLST) (Sauders et al., 2003; Chenal-Francisque et al., 2011; Murphy et al., 2007). While MLST is a powerful method for studying the population genetics of bacteria, it typically has limited discriminatory ability and thus is less commonly used for surveillance. MLVA, a typing method based on the number of tandem repeats, or copy units, at multiple variable-number tandem repeat (VNTR) loci within the genome, has been suggested as an alternative typing method, to PFGE, for some bacterial pathogens, e.g., Salmonella and Escherichia coli (STEC) (Larsson et al., 2009; Hyytia-Trees et al., 2010). In many cases MLVA provides improved discriminatory power over PFGE, particularly for highly clonal pathogens. MLVA has been successfully used in outbreak investigation worldwide, for example, investigation of Salmonella Typhimurium infections in Denmark (Petersen et al., 2011), as well as outbreak associated with Escherichia coli O157:H7 contamination in spinach in California in 2006 (Cooley et al., 2007). 2.2. Public health impact of molecular subtyping methods and their application for foodborne disease surveillance While subtyping based surveillance systems for foodborne and zoonotic pathogens have long been recognized
K. Vongkamjan, M. Wiedmann / Preventive Veterinary Medicine 118 (2015) 189–195
as important tools for controlling foodborne diseases and protecting public health, the development of molecular subtyping methods has been critical to widespread application of subtyping based surveillance systems for infectious diseases. The most important and widely used subtype based surveillance system, is the PulseNet system, which was initiated in 1996 by the US CDC as a national surveillance network for foodborne infections (Swaminathan et al., 2001; Gerner-Smidt et al., 2006). PulseNet initially focused on surveillance for E. coli O157:H7, but was soon expanded to include other bacterial foodborne pathogens. PulseNet is a network of public health and food regulatory agency laboratories that uses PFGE, with standardized protocols, for molecular subtyping. PFGE patterns obtained from local public health and food regulatory agency laboratories can be rapidly compared through the PulseNet PFGE database maintained by the CDC. PulseNet allows for rapid detection of clusters of foodborne disease, which is often an indication of an outbreak. Overall, rapid detection of outbreaks can reduce the outbreak scope by allowing for rapid responses and traceback. Rapid source traceback can then allow for appropriate actions, product recalls, establishment closures etc., that limit the outbreak scope and hence the number of disease cases and sometimes deaths linked to a given outbreak. Outbreaks can be identified and solved sooner which will reduce the number of illnesses and the economic losses. The economic benefits of subtype based surveillance systems in general and PulseNet in particular have also been well established. For example, one study (Elbasha et al., 2000) assessed the societal costs and benefits of the PulseNet system, using subtype-based identification of outbreaks associated with E. coli O157:H7 infections after eating hamburgers as a model. This study suggested that the PulseNet system would have recovered all costs for start-up and 5 years of operation, if only 5 cases were averted by subtype-based facilitation of more rapid detection of an O157:H7 outbreak and the associated accelerated recall of the product responsible for the outbreak (Elbasha et al., 2000). The success of US PulseNet also led to international expansion of the PulseNet system; “PulseNet International” is a global network facilitated by the US CDC, comprising of National and regional laboratory networks that work closely together to track foodborne disease outbreaks, worldwide, through sharing and comparison of subtyping data, using the same general system and structure that was developed initially for the US PulseNet system (Swaminathan et al., 2006). Currently, over 60 countries/regions from Canada, Europe, Latin America & Caribbean, Middle East, and Asia Pacific are members of PulseNet International. The development of partnerships and establishment of efficient collaborations among the networks throughout the world has considerable potential to further improve global surveillance of foodborne diseases. Global subtype based surveillance systems are essential not only to facilitate improved control of infectious diseases throughout the world, but also to facilitate improved detection of outbreaks that occur in multiple countries and continents. This is essential as increases in international trade and food imports and exports (Jiang et al., 2013) and international travel can result in higher
191
risk of contracting foodborne illness through cross-border transmission. Subtype-based surveillance systems not only facilitate more rapid detection of outbreaks, but also are essential to improve the effectiveness of case control studies that are aimed at detecting likely outbreak sources. Specifically, subtype data allow for improved case definitions (e.g., “individuals in the US infected with L. monocytogenes subtype X between date Y and Z”), increasing the power of case control studies. In addition, subtyping can be used to confirm outbreak sources through subtype characterization of pathogen isolates from an implicated food source. Public health benefits of subtyping methods extend beyond rapid detection of disease outbreaks and support of outbreak investigations. Subtype data can also identify new and emerging pathogens and can provide general insights into the epidemiology of infectious diseases. For example, analysis of subtype data for human clinical L. monocytogenes isolates indicated that a considerably higher number of human listeriosis cases may occur in clusters than previously assumed (Sauders et al., 2003); these data highlighted the importance of conducting epidemiological follow-up investigations even of small human listeriosis clusters. Subtyping methods also have been shown to make critical contributions to improving our understanding of the transmission of foodborne pathogens outside human hosts. A number of studies have used subtyping methods to track the spread and sources of L. monocytogenes in foods and food processing plants (Hoffman et al., 2003; Vongkamjan et al., 2013; Norton et al., 2001). Many studies have shown that specific subtypes of L. monocytogenes can persist over time in the environment of a given processing plant (Norton et al., 2001; Ferreira et al., 2014), including for more than 10 years (Orsi et al., 2008b). These findings have been critical for the development and implementation of environmental pathogen control plans aimed at reducing L. monocytogenes contamination of RTE foods from environmental sources. Similarly, subtyping data have also helped identify Salmonella persistence in food processing plants as an important source of Salmonella contamination of RTE food products. Overall, subtyping data thus can provide valuable information for development and implementation of control measures that reduce incidence of pathogen contamination in finished products and hence prevent human foodborne illness cases. 2.3. Use of rapid whole genome sequencing (WGS) based subtyping While, as detailed above, banding pattern based subtyping methods, such as PFGE, have been extremely valuable for surveillance and subtyping of different bacterial pathogens, a number of shortcomings of these methods have become apparent over the years (Lienau et al., 2011). For example, PFGE typing as well as many other typing methods show limited discriminatory power for some highly clonal pathogens, such as B. anthracis, or specific groups within a given pathogen (e.g., certain Salmonella or L. monocytogenes serotypes). In addition, PFGE typing and most other banding pattern based subtyping methods do not allow for phylogenetic inference of isolate
192
K. Vongkamjan, M. Wiedmann / Preventive Veterinary Medicine 118 (2015) 189–195
relationships. This in particular can cause practical problems as isolates that are epidemiologically linked may still differ in their banding patterns, e.g., by 3 bands for PFGE (Tenover et al., 1995); determining whether isolates that differ by up to 3 bands are closely related (and share a recent common ancestor) or distinct is not possible without use of additional sequence-based subtyping methods (Gilmour et al., 2010). With the rapid advances in “next generation” genome sequencing methods, Whole Genome Sequencing (WGS) of bacteria (as well as other pathogens) has emerged as a new powerful tool with considerable promise for applications in food safety and public health (Medini et al., 2008; FDA, 2013). Current benchtop instrumentation such as the Ion Torrent PGM (Life Technologies) and the MiSeq (Illumina, San Diego) allow WGS to be completed in hours to days with cost for sequencing a bacterial genome as low as
Contents lists available at ScienceDirect
Preventive Veterinary Medicine journal homepage: www.elsevier.com/locate/prevetmed
Starting from the bench—Prevention and control of foodborne and zoonotic diseases Kitiya Vongkamjan a , Martin Wiedmann b,∗ a b
Department of Food Technology, Prince of Songkla University, Hat Yai 90112, Thailand Department of Food Science, Cornell University, Ithaca, NY 14850, USA
a r t i c l e
i n f o
Article history: Received 3 February 2014 Received in revised form 30 October 2014 Accepted 2 November 2014 Keywords: Foodborne and zoonotic pathogens Subtyping Molecular epidemiology
a b s t r a c t Foodborne diseases are estimated to cause around 50 million disease cases and 3000 deaths a year in the US. Worldwide, food and waterborne diseases are estimated to cause more than 2 million deaths per year. Lab-based research is a key component of efforts to prevent and control foodborne diseases. Over the last two decades, molecular characterization of pathogen isolates has emerged as a key component of foodborne and zoonotic disease prevention and control. Characterization methods have evolved from banding pattern-based subtyping methods to sequenced-based approaches, including full genome sequencing. Molecular subtyping methods not only play a key role for characterizing pathogen transmission and detection of disease outbreaks, but also allow for identification of clonal pathogen groups that show distinct transmission characteristics. Importantly, the data generated from molecular characterization of foodborne pathogens also represent critical inputs for epidemiological and modeling studies. Continued and enhanced collaborations between infectious disease related laboratory sciences and epidemiologists, modelers, and other quantitative scientists will be critical to a One-Health approach that delivers societal benefits, including improved surveillance systems and prevention approaches for zoonotic and foodborne pathogens. © 2014 Elsevier B.V. All rights reserved.
1. Scope of microbial foodborne diseases in the US Diseases caused by foodborne pathogens are a major public health concern in the US. The US Centers for Disease Control and Prevention (CDC) recently reported estimates of foodborne illness in the US (CDC, 2011a), which indicated that about 47.8 million cases of foodborne illnesses, most of them representing gastrointestinal symptoms, occur annually in the US. Approximately 1 in 6 Americans thus experience a foodborne illness episode each year. For approximately 80% of these cases (38.4 million cases), the causative agent is unknown; on the other hand, 20% of
∗ Corresponding author. Tel.: +1 607 254 2838; fax: +1 607 254 4868. E-mail address: [email protected] (M. Wiedmann). http://dx.doi.org/10.1016/j.prevetmed.2014.11.004 0167-5877/© 2014 Elsevier B.V. All rights reserved.
cases (9.4 million cases) are estimated to be caused by 31 known pathogens. Overall, foodborne illnesses were estimated to result in 127,000 cases of hospitalization and 3037 deaths each year in the US. Of the hospitalization cases and deaths, over 50% of the cases are estimated to be due to unknown causative agents. The top five pathogens responsible for foodborne illnesses include Norovirus, followed by four bacterial pathogens (i.e., Salmonella, Clostridium perfringens, Campylobacter spp., Staphylococcus aureus). Salmonella is the leading cause of hospitalizations (19,336 annually) and deaths (378 annually), while Listeria monocytogenes is responsible for the second most foodborne deaths (255 annually) due to bacterial pathogens. The World Health Organization (WHO) asserts that food and water-borne diarrhoeal illnesses present a “growing public health problem” which is estimated to affect 1 in 3
190
K. Vongkamjan, M. Wiedmann / Preventive Veterinary Medicine 118 (2015) 189–195
people annually worldwide, and result in 2.2 million deaths globally each year. The majority of these deaths (1.9 million) occur in children less than 5 years of age (WHO, 1999). Diseases transmitted through food thus are of importance for both developed and developing countries with many of the key bacterial pathogens attributed to foodborne illnesses in the US also commonly causing disease in developing countries. While some progress in reduction of foodborne diseases has been made, usually successes have been limited to reducing transmission of a specific pathogen through a specific commodity or food chain, often in only one country or region. For example, while considerable progress has been made to reduce L. monocytogenes prevalence in readyto-eat (RTE) deli meats in the US since about 1990 (FSIS, 2010), the number of reported listeriosis cases in many European countries have increased in recent years (Goulet et al., 2008). In addition, since 2010 a number of listeriosis outbreaks in the US have been linked to produce (CDC, 2013), including the most deadly listeriosis outbreak in almost 90 years in the US, which caused 33 deaths and was linked to consumption of contaminated cantaloupe in 2011 (CDC, 2011b). Consequently, our thesis is that long term progress in reducing foodborne illness burden worldwide will require a combination of laboratory approaches and field-based and epidemiological studies, including training of future professionals that can bridge these disciplines. 2. Laboratory-based subtyping techniques and their application to detect disease outbreaks and characterize transmission routes One key area where the importance and impact of combining appropriate (and advanced) laboratory methods with epidemiological and field studies has been well established is in foodborne disease surveillance. This section will briefly describe some of the methods that are used for subtyping (strain differentiation) of foodborne pathogens, with a focus on bacterial pathogens, and then will detail how these methods are being applied in the context of national and international foodborne disease surveillance systems. The learnings from these surveillance systems that utilize laboratory methods for informing and facilitating epidemiological surveillance will likely provide insights that will facilitate better integration of laboratory methods and epidemiology in various areas of public health and related research, which will facilitate broad improvements in public health and food safety. 2.1. Phenotypic and genotypic subtyping techniques Molecular subtyping methods can be classified into both phenotypic and genotypic methods (Wiedmann, 2002). Briefly, subtyping methods allow for differentiation of pathogen isolates into distinct subtypes or strains; these data are useful for determination of the relatedness of bacterial isolates and to further establish linkages between illnesses occurring in an outbreak and sources or transmission routes of the pathogen. While the term “subtype” refers to isolates that share common genetic characteristics (as determined by genotypic methods), the term
“strain” refers to isolates that share common phenotypic characteristics (as determined by phenotypic methods); these two terms are often used interchangeable though (Wiedmann, 2002). To detect disease outbreaks, a typing method should have sufficient discriminatory power to distinguish epidemiologically related isolates from epidemiologically unrelated isolates. Phenotypic methods, e.g., phage-typing, may not only be plagued by poor reproducibility, but also often show insufficient discriminatory power to be valuable in epidemiological studies or for outbreak surveillance (Graves et al., 2003). Nevertheless, a number of phenotypic subtyping methods have played important roles in foodborne, animal and human disease surveillance in the past. For example, wide-spread use of Salmonella serotyping has been essential not only for surveillance, but also for improving our understanding of host specificity, diversity, and transmission pathways for this important zoonotic pathogen (Shi et al., 2013). Genotypic subtyping methods can be classified as “banding-based” and “sequence-based” methods; methods representing both of these categories have also been used widely for characterization, identification, and typing of bacterial isolates, including to assist in surveillance and epidemiological investigation and to define clones, i.e., groups of genetically similar isolates descending from a common ancestor. Among the “banding-based” subtyping methods, pulsed-field gel electrophoresis (PFGE) is often considered the gold standard typing method, including for a number of foodborne bacterial pathogens where PFGE shows a high discriminatory power for most strains (e.g., L. monocytogenes, many Salmonella serotypes) (Graves et al., 2003). Other banding pattern based typing methods that have commonly been used to characterize foodborne pathogens include ribotyping as well as REP-PCR (Jerˇsek et al., 1999; Sauders et al., 2003). Sequence-based subtyping methods include multi-locus variable number tandem repeat analysis (MLVA) and multilocus sequence typing (MLST) (Sauders et al., 2003; Chenal-Francisque et al., 2011; Murphy et al., 2007). While MLST is a powerful method for studying the population genetics of bacteria, it typically has limited discriminatory ability and thus is less commonly used for surveillance. MLVA, a typing method based on the number of tandem repeats, or copy units, at multiple variable-number tandem repeat (VNTR) loci within the genome, has been suggested as an alternative typing method, to PFGE, for some bacterial pathogens, e.g., Salmonella and Escherichia coli (STEC) (Larsson et al., 2009; Hyytia-Trees et al., 2010). In many cases MLVA provides improved discriminatory power over PFGE, particularly for highly clonal pathogens. MLVA has been successfully used in outbreak investigation worldwide, for example, investigation of Salmonella Typhimurium infections in Denmark (Petersen et al., 2011), as well as outbreak associated with Escherichia coli O157:H7 contamination in spinach in California in 2006 (Cooley et al., 2007). 2.2. Public health impact of molecular subtyping methods and their application for foodborne disease surveillance While subtyping based surveillance systems for foodborne and zoonotic pathogens have long been recognized
K. Vongkamjan, M. Wiedmann / Preventive Veterinary Medicine 118 (2015) 189–195
as important tools for controlling foodborne diseases and protecting public health, the development of molecular subtyping methods has been critical to widespread application of subtyping based surveillance systems for infectious diseases. The most important and widely used subtype based surveillance system, is the PulseNet system, which was initiated in 1996 by the US CDC as a national surveillance network for foodborne infections (Swaminathan et al., 2001; Gerner-Smidt et al., 2006). PulseNet initially focused on surveillance for E. coli O157:H7, but was soon expanded to include other bacterial foodborne pathogens. PulseNet is a network of public health and food regulatory agency laboratories that uses PFGE, with standardized protocols, for molecular subtyping. PFGE patterns obtained from local public health and food regulatory agency laboratories can be rapidly compared through the PulseNet PFGE database maintained by the CDC. PulseNet allows for rapid detection of clusters of foodborne disease, which is often an indication of an outbreak. Overall, rapid detection of outbreaks can reduce the outbreak scope by allowing for rapid responses and traceback. Rapid source traceback can then allow for appropriate actions, product recalls, establishment closures etc., that limit the outbreak scope and hence the number of disease cases and sometimes deaths linked to a given outbreak. Outbreaks can be identified and solved sooner which will reduce the number of illnesses and the economic losses. The economic benefits of subtype based surveillance systems in general and PulseNet in particular have also been well established. For example, one study (Elbasha et al., 2000) assessed the societal costs and benefits of the PulseNet system, using subtype-based identification of outbreaks associated with E. coli O157:H7 infections after eating hamburgers as a model. This study suggested that the PulseNet system would have recovered all costs for start-up and 5 years of operation, if only 5 cases were averted by subtype-based facilitation of more rapid detection of an O157:H7 outbreak and the associated accelerated recall of the product responsible for the outbreak (Elbasha et al., 2000). The success of US PulseNet also led to international expansion of the PulseNet system; “PulseNet International” is a global network facilitated by the US CDC, comprising of National and regional laboratory networks that work closely together to track foodborne disease outbreaks, worldwide, through sharing and comparison of subtyping data, using the same general system and structure that was developed initially for the US PulseNet system (Swaminathan et al., 2006). Currently, over 60 countries/regions from Canada, Europe, Latin America & Caribbean, Middle East, and Asia Pacific are members of PulseNet International. The development of partnerships and establishment of efficient collaborations among the networks throughout the world has considerable potential to further improve global surveillance of foodborne diseases. Global subtype based surveillance systems are essential not only to facilitate improved control of infectious diseases throughout the world, but also to facilitate improved detection of outbreaks that occur in multiple countries and continents. This is essential as increases in international trade and food imports and exports (Jiang et al., 2013) and international travel can result in higher
191
risk of contracting foodborne illness through cross-border transmission. Subtype-based surveillance systems not only facilitate more rapid detection of outbreaks, but also are essential to improve the effectiveness of case control studies that are aimed at detecting likely outbreak sources. Specifically, subtype data allow for improved case definitions (e.g., “individuals in the US infected with L. monocytogenes subtype X between date Y and Z”), increasing the power of case control studies. In addition, subtyping can be used to confirm outbreak sources through subtype characterization of pathogen isolates from an implicated food source. Public health benefits of subtyping methods extend beyond rapid detection of disease outbreaks and support of outbreak investigations. Subtype data can also identify new and emerging pathogens and can provide general insights into the epidemiology of infectious diseases. For example, analysis of subtype data for human clinical L. monocytogenes isolates indicated that a considerably higher number of human listeriosis cases may occur in clusters than previously assumed (Sauders et al., 2003); these data highlighted the importance of conducting epidemiological follow-up investigations even of small human listeriosis clusters. Subtyping methods also have been shown to make critical contributions to improving our understanding of the transmission of foodborne pathogens outside human hosts. A number of studies have used subtyping methods to track the spread and sources of L. monocytogenes in foods and food processing plants (Hoffman et al., 2003; Vongkamjan et al., 2013; Norton et al., 2001). Many studies have shown that specific subtypes of L. monocytogenes can persist over time in the environment of a given processing plant (Norton et al., 2001; Ferreira et al., 2014), including for more than 10 years (Orsi et al., 2008b). These findings have been critical for the development and implementation of environmental pathogen control plans aimed at reducing L. monocytogenes contamination of RTE foods from environmental sources. Similarly, subtyping data have also helped identify Salmonella persistence in food processing plants as an important source of Salmonella contamination of RTE food products. Overall, subtyping data thus can provide valuable information for development and implementation of control measures that reduce incidence of pathogen contamination in finished products and hence prevent human foodborne illness cases. 2.3. Use of rapid whole genome sequencing (WGS) based subtyping While, as detailed above, banding pattern based subtyping methods, such as PFGE, have been extremely valuable for surveillance and subtyping of different bacterial pathogens, a number of shortcomings of these methods have become apparent over the years (Lienau et al., 2011). For example, PFGE typing as well as many other typing methods show limited discriminatory power for some highly clonal pathogens, such as B. anthracis, or specific groups within a given pathogen (e.g., certain Salmonella or L. monocytogenes serotypes). In addition, PFGE typing and most other banding pattern based subtyping methods do not allow for phylogenetic inference of isolate
192
K. Vongkamjan, M. Wiedmann / Preventive Veterinary Medicine 118 (2015) 189–195
relationships. This in particular can cause practical problems as isolates that are epidemiologically linked may still differ in their banding patterns, e.g., by 3 bands for PFGE (Tenover et al., 1995); determining whether isolates that differ by up to 3 bands are closely related (and share a recent common ancestor) or distinct is not possible without use of additional sequence-based subtyping methods (Gilmour et al., 2010). With the rapid advances in “next generation” genome sequencing methods, Whole Genome Sequencing (WGS) of bacteria (as well as other pathogens) has emerged as a new powerful tool with considerable promise for applications in food safety and public health (Medini et al., 2008; FDA, 2013). Current benchtop instrumentation such as the Ion Torrent PGM (Life Technologies) and the MiSeq (Illumina, San Diego) allow WGS to be completed in hours to days with cost for sequencing a bacterial genome as low as
