Spillover, hybridization, and persistence in schistosome transmission dynamics at the human–animal interface - pnas.org
Significance
The threat to public health that is presented by zoonotic spillover of pathogens from animal reservoirs is predicted to increase with rapid anthropogenic changes and global trends such as migration and changing land use. Schistosomiasis currently infects more than 220 million people worldwide, and the multihost Schistosoma spp. system within Africa is a key example of where spillover of animal parasites into human populations has enabled the formation of viable hybrid parasite genotypes. Our study demonstrates how zoonotic spillover and complex interactions between pathogen species, such as parasite hybridization, may have implications such as resilience to current disease control strategies and may facilitate the spread of tropical diseases such as schistosomiasis beyond their original geographical boundaries.
Abstract
Zoonotic spillover and hybridization of parasites are major emerging public and veterinary health concerns at the interface of infectious disease biology, evolution, and control. Schistosomiasis is a neglected tropical disease of global importance caused by parasites of the Schistosoma genus, and the Schistosoma spp. system within Africa represents a key example of a system where spillover of animal parasites into human populations has enabled formation of hybrids. Combining model-based approaches and analyses of parasitological, molecular, and epidemiological data from northern Senegal, a region with a high prevalence of schistosome hybrids, we aimed to unravel the transmission dynamics of this complex multihost, multiparasite system. Using Bayesian methods and by estimating the basic reproduction number (R0), we evaluate the frequency of zoonotic spillover of Schistosoma bovis from livestock and the potential for onward transmission of hybrid S. bovis × S. haematobium offspring within human populations. We estimate R0 of hybrid schistosomes to be greater than the critical threshold of one (1.76; 95% CI 1.59 to 1.99), demonstrating the potential for hybridization to facilitate spread and establishment of schistosomiasis beyond its original geographical boundaries. We estimate R0 for S. bovis to be greater than one in cattle (1.43; 95% CI 1.24 to 1.85) but not in other ruminants, confirming cattle as the primary zoonotic reservoir. Through longitudinal simulations, we also show that where S. bovis and S. haematobium are coendemic (in livestock and humans respectively), the relative importance of zoonotic transmission is predicted to increase as the disease in humans nears elimination.
There is a growing acknowledgment of the profound threat that zoonoses (diseases transmitted between animals and humans) pose to human health worldwide, with animal reservoirs presenting diverse complications to elimination efforts for many existing diseases, as well as novel threats in the form of newly emerging diseases (1, 2). Global trends such as increased migration, altering agricultural practices, and a changing climate are all predicted to enhance the potential for human and animal populations to encounter new infectious agents, thereby also increasing opportunities for coinfection by multiple pathogen species within the same host (3⇓⇓⇓–7). Such mixed infections can lead to exchange of genetic material between the coinfecting agents, generating new pathogen genotypes. In the case of helminth parasites which sexually reproduce, this can occur through heterospecific (between-species) mate pairings (8, 9), which can lead to the formation of hybrid offspring (9, 10). Hybridizations, as well as subsequent introgressions (the introduction of single genes or chromosomal regions from one species into another through repeated backcrossing), represent an additional source of genetic variation that may drive parasite evolution, with potential implications including increased host and geographical range, altered pathology, resistance to drug therapy, and, ultimately, persistence in the face of elimination efforts (9). The presence of interspecific hybridizations and introgressions brings an increased level of complexity to disentangling the transmission dynamics of multihost systems, requiring multimodal and original approaches and necessitating data from across disciplines and scales—from the molecular to the population level (11).
The significance of zoonotic transmission and parasite hybridization is exemplified by the case of schistosomiasis. Schistosomiasis is a neglected tropical disease estimated to infect over 220 million people, over 90% of whom reside in sub-Saharan Africa (12⇓–14). The disease is caused by parasites of the Schistosoma genus, which have a complex life cycle that includes sexual reproduction in a mammalian definitive host and indirect transmission via a freshwater snail intermediate host. The schistosome species Schistosoma haematobium (the causative agent for urogenital schistosomiasis in people) has been demonstrated to form viable hybrids and introgressions with closely related schistosome species of the larger Haematobium group which infect livestock, notably Schistosoma bovis, Schistosoma curassoni, and Schistosoma mattheei (causative agents of intestinal schistosomiasis in livestock) (15⇓⇓⇓⇓–20). Hybrids between S. haematobium and S. bovis have been reported in human patients across sub-Saharan Africa, including Senegal, Niger, Côte d'Ivoire, and Mali (21⇓–23), as well as in the recent outbreaks of schistosomiasis in Corsica (24, 25). One recent study in northern Senegal, a region where both S. haematobium and S. bovis are coendemic, indicated that ongoing pairing between these two species leads to creation of hybrids in human hosts and is occurring here via zoonotic spillover of S. bovis from a livestock reservoir to people who are simultaneously infected with S. haematobium (19). Spillover is defined as transmission, from a reservoir to a defined target host, of a pathogen that cannot normally be sustained within the target host population (26, 27), while reservoirs are defined as one or more populations in which a pathogen can be maintained and from which infection is transmitted to a defined target population (28). While molecular data from the study in Senegal indicated that S. bovis cannot be maintained by the human population (hence fitting the definition for a spillover pathogen), onward transmission of hybrids was shown to occur within the human population via backcrossing and introgressions, leading to a complex array of observed miracidia genotypes (miracidia being the first larval schistosome stage which hatches from eggs shed by an infected host) from human specimens.
The public health impact of zoonotic spillover is fundamentally determined by the force of infection from the reservoir species and the potential for onward transmission and persistence within the human population (29), with identification of key hosts and reservoirs also critical to our understanding of multihost systems. These aspects of transmission are crucial for determining appropriate targets for interventions, and ultimately the feasibility of disease elimination, yet are often little-studied and poorly understood.
Given that formation of F1 hybrids requires a host to be simultaneously infected with S. haematobium and S. bovis, attention is naturally focused on geographic localities where circulation of these two closely related schistosome species may be sympatric. It has remained unclear whether humans can act as maintenance hosts for these hybrid parasites after they have been generated by the initial cross-species pairing, or if ongoing zoonotic spillover is required for hybrids to persist in the human population. The fact that hybrid schistosomes identified in the Corsican outbreak were found to have been imported from Senegal (25) highlights both the need to understand the zoonotic potential of S. bovis and, crucially, to evaluate the potential for hybridization to facilitate spread and establishment of schistosomiasis beyond its original geographical boundaries.
Here we combine mathematical modeling and statistical approaches with molecular and epidemiological data from a recent study in northern Senegal (19) in order to evaluate the spillover and multihost dynamics within the Haematobium group hybrid schistosome system. We examine the relative importance of zoonotic spillover in maintaining transmission both at current levels of endemicity and in scenarios where the human disease may be nearing elimination and characterize the relative role of the livestock species involved in zoonotic transmission. Model structure was informed by the insights from the molecular data and parameterized using Bayesian approaches, with output then applied to existing frameworks for classifying multihost systems and zoonotic disease threats (26, 27, 29). This included estimation of the basic reproduction number, or R0, within each host species. Given the lack of literature on key aspects of S. bovis biology necessary for fitting the model to observable data, we also describe a Bayesian approach for inferring the relationship between worm burden and fecal egg count using postmortem studies of naturally occurring infections. We explore the potential for density-dependent effects in this relationship and consider the implications of such effects for macroparasite transmission dynamics in general.
With current trends in anthropogenic activities predicted to drive infectious disease emergence at the global scale, integrative methods for characterizing pathogen transmission events at the animal–human interface and evaluating the consequences of novel pathogen interactions, including hybridizations, have been identified as key scientific challenges (9, 30, 31). As we seek to understand increasingly complex disease dynamics, our work will therefore be of relevance not only for the control of schistosomiasis but also for a growing number of disease systems in our rapidly changing world.
Results
The framework for the mathematical model (summarized in the schematic shown in Fig. 1) was based on the multihost, multiparasite transmission cycle for Haematobium group schistosomes proposed in ref. 19, which was informed by molecular and epidemiological findings from surveys of human, snail, and livestock populations. Key aspects of the system which are incorporated into the model include zoonotic transmission of S. bovis from the livestock reservoir to the human population, interactions between worm genotypes including formation of F1 S. bovis × S. haematobium hybrids in the human host via bidirectional cross-species pairings, and onward transmission of hybrid genotypes via backcrossing of F1 worms.
Schematic of the multihost, multiparasite transmission model: S. haematobium, S. bovis, and Haematobium group hybrids in human and livestock populations. Parameter description and categorization of worm genotypes in definitive hosts are detailed in SI Appendix, Tables S1 and S2.
Definitive hosts are humans (H), cattle (C), sheep (S), and goats (G), (d ∈{H, C, S, G}). Here, livestock can be infected with S. bovis, and humans can be infected with five categories of worm genotypes: S. haematobium (Sh), S. bovis (Sb), two categories of first-generation or F1 hybrids (F1a being the product of a pairing between a male S. haematobium and a female S. bovis; F1b being the product of a pairing between a female S. haematobium and a male S. bovis), and later-generation/introgressed S. haematobium × S. bovis hybrids (Hyb).
The model tracks the mean number of worms (md,j) of each of these schistosome genotypes j (where j ∈{Sh, Sb, F1a, F1b, Hyb}), within definitive host population d. In longitudinal simulations it is assumed that praziquantel (the drug used for mass drug administration, which is the mainstay of schistosomiasis control efforts) is equally efficacious against all worm genotypes.
Spillover Dynamics.
The median posterior estimates of the mean worm burden for each schistosome genotype in the human population (mH,i) are given in Table 1, together with the estimated transmission rates from the larval pool to the human population. Confidence intervals are given as 95% Bayesian credible intervals (BCI).
Estimated mean worm burden in the human population (mH,j) of each distinguished genotype (j) and estimated transmission rate from larval pool Lj to human population (AH,j)
Estimates indicate that the majority of overall worm burden in human hosts is composed of S. haematobium (median estimated percentage of worm burden 73.1%; 95% BCI 70.3 to 75.7%) and the majority of the hybrid worm burden are the later-generation/introgressed Hyb genotype (median 96.76%; 95% BCI 95.32 to 97.88%).
The estimated mean worm burden of S. bovis in humans (0.07; 95% BCI 0.03 to 0.14) corresponds to a very low percentage (0.17%; 95% BCI 0.07 to 0.34%) of the total Haematobium group worm burden in the human population being zoonotically acquired and, accordingly, the median estimated transmission rate from the S. bovis larval pool to the human population (AH,Sb; Table 1) was several orders of magnitude lower than the corresponding transmission rates for other genotypes. Here, transmission rates from the larval pool to humans were estimated to be similar for all genotypes derived from miracidia shed by the human population (AH,Sh, AH,Hyb,, AH,F1a, and AH,F1b), noting wider CIs for AH,F1a and AH,F1b due to the lower number of miracidia these estimates were based on (19).
Estimation of Basic Reproduction Number for Hybrid Schistosomes (R0H,Hyb).
In this framework the basic reproduction number, R0, is defined as the average number of mated female offspring produced by one mated adult female schistosome in the absence of constraints on population growth (32). The estimation of R0 can therefore be used to evaluate the capacity of S. haematobium × S. bovis hybrids to be maintained in the human population in the absence of zoonotic spillover. If the average Hyb female does not replace herself in the next generation (R0H,Hyb < 1), this genotype category could not be maintained without the creation of novel hybrids via zoonotic spillover.
The basic reproduction number for the hybrid genotype Hyb (R0H,Hyb) within the human population was estimated to be 1.76 (95% BCI 1.59 to 1.99; Table 2), comparable to the value estimated for S. haematobium (1.90; 95% BCI 1.74 to 2.14).
Estimated within-species basic reproductive number (R0) for Schistosome genotypes in human and livestock populations and estimated overall basic reproductive number for S. bovis under current conditions with all host species involved in transmission and in simulated conditions where transmission from a host population is eliminated
Projected Impact of Removing Zoonotic Transmission: Longitudinal Simulations under Varying Treatment Coverage Levels.
At all treatment coverage levels simulated (50% and 90% shown in Fig. 2; 75% coverage, SI Appendix, Fig. S10) with no zoonotic transmission of S. bovis, transmission of the Hyb genotype was predicted to persist after 15 y (Fig. 2). This indicates that even under the increased pressure exerted by higher levels of mass drug administration (MDA) hybrids can be maintained within the human population without zoonotic transmission of S. bovis to the human population and generation of novel hybrids.
Longitudinal simulations under assumed current level of MDA with praziquantel (A: 50% annual coverage school-aged children) and under enhanced coverage (B: 90% annual coverage school-aged children). Mean worm burden in human population of S. haematobium, Hyb, F1 hybrids, and S. bovis under current level of zoonotic transmission and with no zoonotic transmission. Median shown as solid lines, dashed lines show 95% CIs.
In the absence of zoonotic transmission, predicted mean burden of F1 hybrids and S. bovis declines rapidly to zero, as would be expected (Fig. 2). However, this is predicted to have minimal impact on the overall worm burden and lead to only a slight decrease in the burden of the later generation Hyb hybrids. At current treatment coverage levels (assumed 50% coverage of school-aged children based on World Health Organization preventative chemotherapy databank for Senegal; ref. 33) the overall mean worm burden after 15 simulated years was predicted to be 38.70 (95% CI 31.95 to 46.84) with zoonotic transmission and 39.12 in the absence of zoonotic transmission (95% CI 32.30 to 47.22), corresponding to a median 1% difference in worm burden between the two scenarios (95% CI 0.23 to 2.07%).
Although spillover of S. bovis was estimated to contribute a very small proportion to the overall worm burden in the human population at current levels of MDA, the relative importance of ongoing zoonotic transmission from a stable reservoir in terms of worm burden and genotype composition in the human population was predicted to increase considerably under scenarios of increasing MDA coverage. This is illustrated in Fig. 3, where it can be seen that, at current levels of zoonotic spillover under MDA coverage levels of 75% and 90%, the percentage of mean worm burden that is composed of hybrids would be predicted to increase. This effect was most marked at 90% treatment coverage, with the proportion of mean worm burden that is composed of hybrids predicted to reach more than 50% after 15 simulated years (median estimate 51.90%, 95% CI 41.17 to 67.36).
Longitudinal simulations: Predicted impact of varied MDA coverage on proportion of worm burden in the human population that are hybrids under current level of zoonotic transmission (Left) and with no zoonotic transmission (Right).
Spillover, hybridization, and persistence in schistosome transmission dynamics at the human–animal interface - pnas.org
Comments
Post a Comment