The dynamics and persistence of infectious diseases cannot be understood without considering the role of ecological and genetic heterogeneities that influence parasite transmission dynamics (Rand et al. 1995; Hagenaars et al. 2004). Unlike the relatively simple homogeneous populations described earlier, populations of wild primates are stratified by age, sex, social rank, or clumped spatially due to naturally fluctuating resources or habitat fragmentation. Further complications arise when pathogens can infect multiple host species, requiring that researchers consider transmission heterogeneities among multiple host species and the consequences for parasite spread and persistence. Here, we briefly examine three factors that should be important for patterns of disease spread in free-living primates: spatial heterogeneity (including landscape features and metapopulation dynamics), host social system, and parasites capable of infecting multiple host species. Several approaches have been developed to examine how different sources of ecological heterogeneity influence disease spread, including metapopulation models, mixing matrices, individual-based models, and social network theory. Not surprisingly, advances gained by modeling approaches have rapidly outpaced field and experimental work. Thus, empirical studies in natural systems are badly needed to identify which heterogeneities are likely to be most relevant in wild primate populations, and how control strategies might be implemented in response to pathogens in heterogeneous environments (Chapter 7).
4.4.1 Spatial heterogeneity: landscape features and metapopulation dynamics
In many other wildlife systems, spatially explicit models have been used to understand the influence of landscape ecology and host dispersal patterns on the spread of newly introduced diseases across a geographic region (Shigesada and Kawasaki 1997; Russell et al. 2004). Perhaps the best examples include rabies infecting foxes (Murray et al. 1986) and raccoons (Smith et al. 2002), where transmission is highly local and host movement is affected by natural barriers like rivers or mountain ranges (see Box 3.3). Consideration of these factors requires information on the spatial configuration of host populations, rates of local and long-distance host dispersal, and potential natural barriers to host movement. Spatial simulations could point to sites for implementing physical barriers or intensive vaccination efforts to slow or stop pathogen spread (Russell et al. 2005). Detailed records of habitat use, spatial distributions, and between-group contact necessary for such simulations already exist for several wild primate species (Waser 1976; Kappeler 1998b; Di Fiore 2003; Dias and Strier 2003), and these can be augmented by gene flow estimates derived from molecular data (Gagneux et al. 2001). In other cases, monitoring data that track the spatial spread of novel pathogens like Ebola virus can be used to parameterize models, and thus used to predict where new outbreaks might occur and how fast the pathogen will spread in populations of susceptible hosts.
Beyond the details of landscape features, models have also been used to examine disease spread in the context of metapopulation processes more generally (Hess 1996; Carlsson-Graner and Thrall 2002; McCallum and Dobson 2002; Park et al. 2002). In the case of primates, metapopulations (defined as a group of populations or patches between which dispersal can occur) might arise from naturally patchy habitats or the subdivision of host populations into social groups. Loss of suitable habitat caused by forest fragmentation and other habitat changes can further isolate primate individuals or groups into remaining patches, as documented for primate species such as Cercopithecus mitis, Procolobus badius, and Macaca silenus (Lawes et al. 2000; Singh et al. 2002; Galat-Luong and Galat 2005).
Insights from metapopulation models point to the joint roles of two key processes on pathogen establishment and persistence: (a) within-patch dynamics and (b) local colonization and extinction (Hess 1996; Grenfell and Harwood 1997; Carlsson-Graner and Thrall 2002; Gog et al. 2002). One consequence of metapopulation dynamics is that subdividing a host population into smaller units can increase the critical community size required for pathogen persistence (Park et al. 2002). Thus, local population sizes might be too small for pathogens to persist, and limited movement among patches could further reduce pathogen spread at the entire population level (Hess 1996; Gog et al. 2002). Other models show that host movement among local patches can be crucial to re-colonization following local extinction, allowing hosts to escape to areas not yet affected by parasites, while also facilitating the spatial spread of alleles determining host resistance and pathogen infectiousness (Hassell et al. 1991; Hess 1996; Grenfell and Harwood 1997; Thrall and Burdon 1997). Metapopulation approaches and concepts have tremendous importance for examining the role of habitat fragmentation and isolation in host-pathogen dynamics, including in primates (Cowlishaw and Dunbar 2000). In the context of disease and primate conservation, these issues are addressed in more detail in Chapter 7.
Primates are generally social animals, and as such they might experience greater infectious disease risk through increased local density, close proximity, or higher contact rates among host individuals (Anderson and May 1979; Arneberg 2002, see Chapters 3 and 6). The details of host social systems will determine how diseases spread through populations; pathogens spread within groups through a network of social and mating contacts and between groups through dispersal. Patterns of transmission will also depend on the type of contact and characteristics of interacting individuals. For example, infections are more likely to spread from mother to dependent offspring, or between preferred mating partners, than between individuals that avoid one another at food resources or sleeping sites. Information on the frequency of pairwise contacts can often be extracted from existing data sources on primates, such as grooming matrices or records of group composition and intergroup movements (Sugiyama 1971; Pusey and Packer 1987; Rowell 1991; Isbell and VanVuren 1996).
Several modeling approaches have been applied to capture heterogeneity in patterns of social contact, focusing primarily on the spread of contagious infections in human populations. One strategy is to group individuals into classes (e.g. social status, kinship, or sexual activity) and describe contacts among classes in terms of a "mixing matrix," where the entries in each of the cells describe the frequency distribution of contacts per unit time (Blower and McLean 1991). The most important insight gained from these models is that the pattern of contacts between different activity classes has a major impact on parasite spread (Jacquez et al. 1988). Specifically, a high degree of mixing within an activity class results in a more rapid initial spread but a lower population-wide prevalence, as compared to a higher degree of mixing among activity classes. Despite their importance in human epidemiology, mixing matrices have not been applied widely to animal social and mating systems because detailed information for their construction (contact rates within and among social classes or mating groups) has generally not been available. In the context of a female-bonded primate species, this approach could be applied by developing matrices that measure contact rates among females within and across matrilines, among males, and among males and females.
Stimulated in part by increasing computational power, agent-based or individualbased modeling approaches have been increasingly applied to problems in epidemiology to simulate more realistic contact patterns (Keeling 1999a; Koopman et al. 2002). These models essentially assume that individual animals interact with one another using simple local rules for group formation, within-group contact, and among-group dispersal (see Grimm and Railsback 2005 for more details on individual based models in ecology). For example, Thrall et al. (2000) used individual-based models to show how the spread of an STD in a polygynous host was influenced by variance in male mating success and migration of females among mating groups (Box 4.5). Other individual-based models have been applied to understand patterns of disease spread in social insects (Naug and Camazine 2002; Pie et al. 2004). These models showed that division of labor, limited worker activity, and spatial separation of units within a colony could slow or diminish disease outbreaks. Although these simulation-based approaches can provide insights into the consequences of heterogeneities in behavior, they are relatively data-hungry in terms of the number of traits, and detailed measures of these traits, that are required for model parameterization.
STDs are increasingly recognized as an important parasite group with potentially large impacts on host reproduction and evolution (Smith and Dobson 1992; Lockhart et al. 1996). The characteristics and dynamics of STDs differ from many other infectious diseases. STDs have smaller host ranges, longer infectious periods, and are less likely to cause host mortality or induce protective host immunity (Oriel and Hayward 1974; Smith and Dobson 1992; Lockhart et al. 1996). Characteristics of many STDs also cause their dynamics to differ from other directly transmitted parasites. In particular, STDs tend to persist as endemic (rather than epidemic) infections, with transmission relatively unaffected by increased host density or crowding. They have been described as a unique class of pathogens well adapted to persisting in small, low density host populations (Smith and Dobson 1992), although their presence in large populations is not theoretically precluded. Animals with promiscuous mating systems (or species in which females engage in frequent extra-pair copulations) are predicted to experience a greater risk of acquiring STDs. However, empirical patterns illustrating potential links between host mating behavior and infectious disease risk have not been well documented in mammals or other vertebrates.
The dynamics of most STDs in humans requires consideration of heterogeneities in sexual activity (Anderson and May 1991). For this reason, population models developed to predict the dynamics and control of HIV, syphilis, gonorrhoea and other STDs have focused on human sexual contact patterns (Anderson et al. 1988, 1989; Boily and Masse 1997; Hethcote and Yorke 1984; Garnett et al. 1997). Mathematical models that incorporate heterogeneity in mating behavior show that STD transmission increases with increasing variance in partner exchange rates, and that highly promiscuous individuals ("super-spreaders") can facilitate STD persistence even when the mean number of sexual partners is low (Anderson and May 1991). Consistent with models that predict a higher risk of infection among more promiscuous subgroups, surveys of HIV and other STDs in human populations show that prevalence increases with increasing numbers of sexual partners per year (reviewed in Anderson and May 1991). One might expect this generalization to apply to wild mammals with polygynous mating systems, with variance in male mating success at the population level being proportional to increased transmission of STDs.
Using an individual-based simulation model of polygynous mating systems, Thrall et al. (2000) examined how variance in male mating success (i.e. mating skew) affects the spread of STDs, and how this interacts with longevity and the migration of females among mating groups. Their model assumed that males varied in their attractiveness to females, that females had only one mate per breeding season, and that females could change groups between breeding seasons. Two mating system parameters were examined: variation in male mating success (degree of polygyny) and variation in female fidelity to males (dispersal to new groups between mating systems). When females moved frequently among groups, variance in male mating success had a weaker effect on prevalence of infection in females. When intergroup movement was limited, parasites spread more rapidly and reached higher prevalence in groups with more females (i.e. greater polygyny) (Fig 4.9).
A notable outcome of the model by Thrall et al. (2000) was that equilibrium STD prevalence was significantly greater in females than in males. When variance in male mating success was high, many males remained unmated, lowering the equilibrium prevalence among males relative to females. Using published data on two sexually transmitted retroviruses in wild primate populations (SIV and STLV), Nunn and Altizer (2004) found support for the prediction that STD prevalence is higher in females than in males among
(a) Social groups, each with one male
(a) Social groups, each with one male
Was this article helpful?