Evolution and emergence of Bordetella in humans Ottar N. Bjørnstad 1,3 and Eric T. Harvill 2,3 1 Departments of Entomology and Biology Penn State University, 501 ASI Building, University Park, PA 16803, USA 2 Department of Veterinary Sciences, 124 ASI Building, University Park, PA 16803, USA 3 Center for Infectious Disease Dynamics, Penn State University, 501 ASI Building, University Park, PA 16803, USA Two highly infectious bordetellae, Bordetella pertussis and B. parapertussis, have emerged in historical times as co-dominant in human populations. Both of these cause acute disease (whooping cough), whereas their progenitor, B. bronchiseptica, is of variable virulence in a wide variety of animals. The remarkably close phyloge- netic relatedness of these three bordetellae and the two independent jumps to humans provide a unique opportunity to examine the evolution and genetics involved in the emergence of acute human pathogens. We hypothesize that the more virulent strains in humans reflects how acutely infectious pathogens might be favored in communities with large contact networks. Furthermore, we suggest that the differential expression of the various virulence factors by the two human pathogens can be explained by immune- mediated competition between the strains. The evolu- tionarily favored strategies of both of the human bordetellae result in immunizing infections and acute epidemics. Introduction The term ‘whooping cough’ refers to the sound of the desperate inspiration of air between paroxysmal coughing episodes of children infected with Bordetella pertussis or B. parapertussis. Although the symptoms are highly distinctive and diagnostic, the first medical chronicle of whooping cough was recorded in 1578 [1]. Since this disease description is relatively recent in the history of respiratory diseases, it has been hypothesized that whooping cough has recently emerged or increased in virulence. B. pertussis and B. parapertussis are highly transmissible, Gram-negative coccobacilli that colonize human respiratory tracts and transmit through the aerosolized droplets produced by coughing. The disease is acute, with severe coughing that can progress to vomiting, convulsions, coma and death [2]. B. bronchiseptica – their close evolutionary progenitor – infects and can cause diseases in a wide range of mammals from marsupials through to ungulates, rodents and carnivores [3]. Surveys of domestic animals have revealed high (25–100%) seroprevalence in cats [4], dogs [5] and pigs [6]. Although human infections are rare, clinical records are steadily accumulating. Risk groups to B. bronchiseptica appear to be infants [7], and individuals in close contact with animals [8] or who are immunologi- cally compromised [9]. A fascinating aspect of Bordetella biology is how virulent, acute pathogens have evolved at least twice from the animal progenitor. Phylogenetic analyses (Box 1) reveal that B. pertussis and B. parapertussis represent independent lineages that cause similar illness (except for the lack of lymphocytosis associated with the pertussis toxin that is expressed only by B. pertussis). A B. parapertussis-like strain has been isolated from sheep; however, this appears to represent a separate evolutionary clade [1,3,10] (Box 1). The two whooping cough-causing strains, therefore, appear to be specific to the human host. There are several testaments to the success of the bordetellae as pathogens of humans. First, current estimates places the annual number of infected children at w50 million [11]. Second, the basic reproductive ratio (the average number of secondary infections directly resulting from one infection in a completely susceptible population) of whooping cough is w15, making it one of the most contagious directly transmitted human patho- gens [12]. Third, before the successful development of a vaccine, it was one of the most important childhood infections with a mean age of infection of 5 years [13]. From an epidemiological point of view, however, the great success in terms of transmissibility might have come with a cost in terms of a reduced infectious period. B. bronchiseptica persists, sometimes for life, in the nasal cavity of its wildlife hosts [14]. By contrast, the human bordetellae can only infect humans transiently with a latent period of 7–10 days and an infectious period of 3 weeks or more. Although B. pertussis and B. parapertussis cannot persist within an individual, the severe cough that results from disease provides for efficient transmission that enables epidemiological per- sistence within human populations. The benefit of heightened transmissibility at the cost of shortened infectious period could reflect a recently identified trade- off between invasion speed and persistence [15]. Epidemiology: invasion–persistence trade-offs There are two fundamental epidemiological quantities: the transmission rate (the per time unit rate at which susceptible individuals are being infected) and the length Corresponding author: Bjørnstad, O.N. (onb1@psu.edu). Available online 28 June 2005 Opinion TRENDS in Microbiology Vol.13 No.8 August 2005 www.sciencedirect.com 0966-842X/$ - see front matter Q 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.tim.2005.06.007