PCR to predict risk of airborne disease Jon S. West 1 , Simon D. Atkins 1 , Jean Emberlin 2 and Bruce D.L. Fitt 1 1 Rothamsted Research, Harpenden, Hertfordshire, AL5 2JQ, UK 2 National Pollen and Aerobiology Research Unit, University of Worcester, Henwick Grove, Worcester, WR2 6AJ, UK Plant, animal and human diseases spread by microscopic airborne particles have had major economic and social impacts during history. Special air-sampling devices have been used to collect such particles since the 19th century but it has often been impossible to identify them accurately. Exciting new opportunities to combine air sampling with quantitative PCR to identify and count these particles are reviewed, using crop pathogen examples. These methods can be used to predict the risk of unexpected outbreaks of airborne diseases by identifying increases in pathogen inoculum or genetic changes in pathogen populations that render control ineffective. The predictions can provide guidance to policymakers, health professionals or the agricultural industry for the development of strategies to minimise the risk of severe pandemics. The need to predict risk of airborne disease Many plant, animal and human diseases spread by air- borne particles have had major economic and social impacts during history. For example, the 19th century potato famine in Ireland, which was caused by potato late blight, resulted in mass migration to the USA [1,2] and the 2001 foot and mouth disease outbreak in the UK [3,4] caused £8 billion of damage to the rural economy [5]. The influenza pandemic of 1918 resulted in 21–50 million deaths and influenza and whooping cough continue to cause epidemics in humans of all ages [6–8]. These lessons from history explain why concerns about the global spread of the H5N1 influenza virus among birds are understand- able [9]. Airborne biological particles, or bioaerosols, in- clude particles of biological origin or activity that can affect living things through infectivity, allergenicity, toxicity or other processes [10]. Such particles can cause disease directly (e.g. epidemics of infectious diseases [1–7]) or indirectly (e.g. non-infectious diseases, such as asthma or hay fever, caused by hypersensitivity to allergens). Biological particles that cause disease are often invisible to the naked eye but can be seen under a microscope. In the 19th century, Pasteur’s classic work demon- strated that certain diseases were caused by such airborne particles, rather than being generated spontaneously [11]. Over the past 150 years, many different methods have been used to collect airborne particles. However, identification of the particles causing disease has relied traditionally on microscopy, culture of the organisms on artificial media or immunological detection methods [11–15]. These methods are extremely time-consuming and require considerable expertise to identify the organisms accurately. Specialised containment facilities are needed if pathogenic organisms are to be cultured. Furthermore, microscopy cannot dis- tinguish airborne inoculum of pathogens when different species have visually identical spores and many air-borne organisms cannot grow on artificial media (<1% of bacteria can be cultured [16]). There is a need to predict the risk of severe epidemics started by airborne inoculum because epidemic severity varies in time and space and unexpected epidemics can result from unusual weather or sudden genetic changes in pathogen populations. However, it has proven difficult to use airborne inoculum measurements to predict the sever- ity of epidemics because of difficulties in identifying the pathogen species rapidly and accurately by traditional methods. Identification of genetic changes in populations, for example, for fungicide or herbicide resistance, has been even more difficult and time-consuming, if not impossible. Since the introduction of PCR [17], advances in molecular diagnostics (Table 1) have now made it possible to detect, identify and accurately quantify airborne inoculum rapidly. Furthermore, it can also provide data about changes in abundance of specific pathogen genes, such as those for resistance to pesticides. Additionally, it can give information about several airborne pathogens or aller- gens or several different genetic changes in one population simultaneously. There are thus new opportunities to com- bine molecular diagnostics with air-sampling to accurately forecast the risk of severe disease initiated by bioaerosols. This review will focus on the potential use of PCR to predict the risk of airborne disease, or to understand epidemics retrospectively, with particular reference to the control of fungal diseases of arable crops. It is timely to review this topic because there are now new threats owing to climate change, which is predicted to increase the severity of some plant diseases and to change the timing of crop growth stages [18–20]. Other reviews have focussed on airborne human diseases rather than crop diseases and have not considered disease forecasting [14]. There is a real risk that bioaerosols will spread diseases that are cur- rently rare or absent so that they establish and become endemic in new countries. For example, the pathogens Puccinia graminis [21] and Cercospora beticola [22] (Table 2) might be spread to the UK. Air dispersal is a major route for crop pathogens to reach new territory; some fungal spores can travel thousands of kilometres to cause disease outbreaks that were not predicted previously [23]. Therefore, methods to sample air in combination with appropriate diagnostic methods are of increasing relevance. The principles for measuring airborne inoculum are generic, with relevance not only to agriculture but also to human and animal health and biosecurity [24]. Review Corresponding author: West, J.S. (jon.west@bbsrc.ac.uk). 380 0966-842X/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.tim.2008.05.004 Available online 1 July 2008