Host migration and environmental temperature influence avian haemosporidians prevalence: a molecular survey in a Brazilian Atlantic rainforest

Introduction

The avian haemosporidians of the genera Plasmodium and Haemoproteus are vector-borne parasites that infect a wide range of host species (Hellgren, Pérez-Triz & Bensch, 2009; Ricklefs et al., 2014) and have frequently switched to new host species and new environments throughout their evolutionary history (Ricklefs, Fallon & Bermingham, 2004; Valkiūnas, 2005; Ricklefs et al., 2014; Ellis et al., 2019). These intracellular parasites reproduce sexually in different dipteran vectors: mosquitoes (Culicidae) are vectors of Plasmodium, and biting midges (Ceratopogonidae) and hippoboscid flies (Hippoboscidae) are vectors of Haemoproteus (Parahaemoproteus) and Haemoproteus (Haemoproteus), respectively (Valkiūnas, 2005; Santiago-Alarcon, Palinauskas & Schaefer, 2012). Concerning their vertebrate hosts, Plasmodium and Haemoproteus (Parahaemoproteus) infect birds of various orders, while Haemoproteus (Haemoproteus) is more specific and infects birds of the order Columbiformes and some sea birds (Work & Rameyer, 1996; Valkiūnas, 2005; Padilla et al., 2006; Levin & Parker, 2012; Levin et al., 2012).

Haemosporidians are closely connected to their hosts in interaction with outcomes ranging from sublethal effects on the host fitness (Ortego et al., 2008; Knowles, Palinauskas & Sheldon, 2010) to the decline and extinction of populations (van Riper et al., 1986; Atkinson et al., 1995, 2000). These parasites can exert selective pressure on host populations through effects on reproductive success, lifetime, and survival (Hamilton & Zuk, 1982; Scott, 1988; Spencer et al., 2005; Asghar et al., 2015; Ricklefs et al., 2016). Therefore, identifying the geographical distribution, host preferences, and infection prevalence of these parasites may help the development of appropriate management strategies to promote biodiversity conservation efforts worldwide.

Haemosporidian parasite prevalence, distribution, and richness vary widely across host species and can be affected by several factors. Prevalence may increase with age of host, since older individuals tend to have higher infection risk as a result of accumulated exposure to parasites or potentially immunosenescence (Atkinson et al., 1995; Ricklefs et al., 2005; Wood et al., 2007; Eastwood et al., 2019). However, this relationship is not consistent among studies and is not always observed, which may be indicative that it depends on the parasite and host species studied (Wood et al., 2007; Antonini et al., 2019). Shape and height of birds nest may also influence haemosporidian prevalence, nest height must be associated with the spatial feeding preferences of vectors when seeking hosts, while its shape must determine the birds’ exposure to vectors (Cerný, Votýpka & Svobodová, 2011; Fecchio et al., 2011; González et al., 2014; Lutz et al., 2015; Matthews et al., 2015).

A migration strategy is another factor that may also influence haemosporidian prevalence in birds, since migratory species are exposed to different vectors and parasites as a consequence of their habitat change during their annual cycle (Waldenström et al., 2002; Hellgren et al., 2013; Ricklefs et al., 2017; Slowinski et al., 2018; Pulgarín-R et al., 2019; Soares, Latta & Ricklefs, 2019). This relationship does not occur evenly between parasites and host species, as can be seen in studies that have observed that migratory birds can be infected by parasites from their wintering grounds (Waldenström et al., 2002) or have little or no haemosporidian lineage shared with resident birds (Ricklefs et al., 2017; Pulgarín-R et al., 2019). Likewise, the prevalence may be higher in migratory bird breeding (Hellgren et al., 2013) or wintering areas (Pulgarín-R et al., 2019).

Environmental factors can also be related to prevalence in a bird community. Seasonality, which involves temperature and rainfall, may influence vector infection dynamics and hence the haemosporidian transmission (Medeiros et al., 2016; Ferreira Junior et al., 2017; Hernández-Lara, González-García & Santiago-Alarcon, 2017a), and was even considered an important driver of host specialization (Fecchio et al., 2019). However, seasonality, in the same way as the host life-history traits, is not always linked to haemosporidian prevalence, not even in the tropics (Ishtiaq, Bowden & Jhala, 2017). Moreover, as seasonality is also associated with the breeding period of birds, a relationship could be observed between these two variables influencing the haemosporidian prevalence, since during bird’s breeding season may be an increase in adult susceptibility to infections, due to changes in its behavior and immunity (Drobney, Train & Fredrickson, 1983; Richner, Christe & Oppliger, 1995; Ardia, 2005). Besides that, juveniles born throughout this time may also be more susceptible to infections (Cosgrove et al., 2008; Møller, 2010; Santiago-Alarcon et al., 2011; Ferreira Junior et al., 2017; Rodrigues et al., 2020). Therefore, many factors interact to determine the parasite prevalence in birds, revealing distinct results in studies testing the same relationships in different communities. This makes the knowledge of these interactions in different ecosystems valuable, to better understand the various factors that interact and influence the dynamics of parasite-host infection in natural environments.

Here, we tested whether the probability of an individual being infected with haemosporidian parasites is related to (1) its individual-level traits (i.e., age and breeding condition), which may influence the host immune defense mechanisms and individual exposure to vectors; (2) its species-specific traits (i.e., diet, foraging strata, period of activity, species body weight, migratory status, and nest shape), which may be linked to differential vector exposure; and (3) climate factors (i.e., temperature and rainfall), which might influence in vector abundance and richness. These hypotheses were tested in an area of Brazilian Restinga, a poorly sampled phytophiosionomy for hemoparasites, and therefore with a high potential to house new haemosporidian lineages.

Materials & methods

Study site

The study was carried out in Barreira do Inferno Rocket Launch Center of the Brazilian Air Force (acronym in Portuguese, Centro de Lançamento Barreira do Inferno–CLBI; 5°55′S 35°9′W), a protected area of ~1,800 ha located in Parnamirim, State of Rio Grande do Norte, northeastern Brazil (Fig. 1). The area is in a tropical coastal vegetation region, named Restinga, which is a type of vegetation associated with the Atlantic Rainforest Domain (for more details about the study site, see Rodrigues et al. (2020)). Following the local climatic data (Fig. S1), we determined as rainy season the period that goes from March to August and, as dry season the period from September to February.

Results

Discussion

What does influence avian haemosporidian prevalence? Here, we found that migratory birds were less likely to be infected with Haemoproteus (Parahaemoproteus) when compared to resident birds and that the probability of infection by Plasmodium was positively influenced by temperature. By observing avian haemosporidians in a diverse region, and exploring how ecological variables are related to parasite infection probability in wild birds, we can compare these interactions we have found here with patterns of interactions already observed in avian communities in different contexts, to add knowledge that allows us to better understand infectious diseases in wild birds.

Migratory behavior of birds from CLBI had a significant and inverse association with the probability of infection by Haemoproteus (Parahaemoproteus), but there was no effect of migration on the probability of infection by Plasmodium. The current knowledge of the host specificity of the parasite lineages predicts that Haemoproteus parasites tend to be more host-specific than Plasmodium parasites (Ishtiaq et al., 2007, 2010; Dimitrov, Zehtindjiev & Bensch, 2010). Therefore, it is possible that migratory birds are not suitable hosts for Parahaemoproteus lineages with which they geographically overlap during their annual cycle. This agrees with the study presented by Hellgren et al. (2007) showing that Haemoproteus and Leucocytozoon had a significant affiliation to a single resident bird fauna, while Plasmodium lineages showed a higher degree of infecting both resident and migratory bird species. If that is true in the studied bird community, migrants may be less likely to become infected by Haemoproteus in CLBI when compared to resident birds, due to the greater specificity this parasite presents concerning its local hosts. Because Plasmodium lineages are usually more generalist, they can probably infect migrants as well as resident birds, which would explain why we did not find the same influence of migratory behavior on the probability of infection by these parasites. If, on the one hand, the migration has the potential to increase the exposure of birds to parasites by concentrating individuals at breeding, overwintering or, migratory stopover sites (Waldenström et al., 2002), on the other hand, migration could make possible the escape of birds from habitats where parasite transmission stages have accumulated, or selective removal of infected hosts during movements (Hall, Altizer & Bartel, 2014).

Alternatively, our results might be related to the hypothesis that the selection experienced by migratory birds in their breeding and wintering areas resulted in greater investment in immune defense (Møller & Erritzøe, 1998). If a bird individual is negative for haemosporidian infections, it might be because either the host individual is not infected, the parasite is dormant in tissues and not found in the bloodstream (Valkiūnas, 2005), or that it occurs in such low intensities in the avian blood that it is not detected by PCR screening. This former situation could be a cue that the bird was able to fight the infection, reducing its parasitemia. In that case, migrants having greater investment in immune defense could also have low parasitemia when infected.

Despite the absence of data about mosquitoes in CLBI to allow us to correctly evaluate the vector-parasite-host relationship, the positive association observed between temperature and Plasmodium prevalence may be related to the effects of temperature on the vectors of this parasite. For avian haemosporidians, the temperature is commonly described as an important abiotic factor influencing the parasite development and vector breeding opportunities (Beier, 1998; Santiago-Alarcon, Palinauskas & Schaefer, 2012; Medeiros et al., 2016; Mordecai et al., 2019). It has been demonstrated that the development of different malaria parasites in vectors can be influenced by the climate and is generally hampered by low increments in temperature (LaPointe, Goff & Atkinson, 2010; Zamora-Vilchis, Williams & Johnson, 2012). Temperature also determines the rate at which mosquitoes develop into adults, the frequency of their blood-feeding, and the rate at which parasites are acquired (Patz et al., 2000). Garamszegi (2011) has shown that a 1 °C increase in global temperature led to a two- to three-fold increase in the average prevalence of Plasmodium in birds. It is also demonstrated that studies made with samples from years and localities where temperature anomalies were strongly expressed generally detected higher Plasmodium prevalence than surveys based on samples that were less affected by temperature anomalies (Garamszegi, 2011). Sehgal et al. (2011), in a study conducted on Olive Sunbirds (Cyanomitra olivacea) in West and Central Africa, also showed an association between higher temperatures and elevated Plasmodium prevalence, with data indicating that the maximum temperature of the warmest month was the most important indicator for elevated malaria prevalence. In contrast, Zamora-Vilchis, Williams & Johnson (2012), in a study conducted in Australia, demonstrated that in areas with high temperatures the birds had a higher prevalence of Haemoproteus, and relationships for Leucocytozoon and Plasmodium were also positive but not statistically significant. Given that, in a warmer climate, the vector abundance may increase, and the transmission of vector-borne diseases must be higher. On the other hand, in our study the infection probability by Parahaemoproteus lineages was not influenced by temperature and, it is possible that biting midges, which act as their vectors, do not have their reproduction and development as closely related to climatic factors as Plasmodium vectors. Immature biting midges require a certain amount of free water or moisture, being able to develop in a wide range of habitats that meet that criterion like pools, streams, marshes, bogs, beaches, swamps, tree holes, irrigation pipe leaks, saturated soil, animal dung, and even rotting fruit and other vegetation (Mellor, Boorman & Baylis, 2000). With so many possible breeding sites, Parahaemoproteus vectors are probably present in the community during the whole year and, even if there is some reduction in their abundance due to changes in temperature and rainfall (Mellor, Boorman & Baylis, 2000), it may be less evident for biting midges than for mosquitoes.

Neither age nor breeding condition explained the probability of infection by haemosporidians. Although many studies have evidenced that birds age and/or breeding condition may influence haemosporidian prevalence (Wood et al., 2007; Ferreira Junior et al., 2017; Hernández-Lara, González-García & Santiago-Alarcon, 2017b; Eastwood et al., 2019), others have failed to detect such an association (Ricklefs et al., 2005; Matthews et al., 2015). There was also no association between prevalence and rainfall or, except for migratory behavior, any tested species-specific traits (diet, foraging strata, period of activity, species body weight, and nest shape). It is true that many studies have shown that different host-traits and abiotic factors are important determinants in a host-parasite interaction (Ricklefs et al., 2005; Wood et al., 2007; Medeiros et al., 2016; Ferreira Junior et al., 2017; Hernández-Lara, González-García & Santiago-Alarcon, 2017a; Ishtiaq, Bowden & Jhala, 2017; Fecchio et al., 2017b; Eastwood et al., 2019). However, there are many variations in these studies’ results, and some of them fail to detect these interactions. Based on the mixed results found in these studies, it is possible that the relationship between species-specific traits as well as individual-level traits and the risk of infection by haemosporidian parasites might be location-dependent. It is important to highlight that several factors might be working together to determine such variations we see in all these different studies, including the host species that, as we observed, had a small influence on the infection probability by Plasmodium. That small influence of the species on Plasmodium prevalence could be related to factors that were not tested in our studies, like phylogeny or co-infection with other hemoparasites. Studying haemosporidian infections only at the community level reduces our ability to detect if some species are more susceptible to the most common Plasmodium lineages in CLBI, for example. That is why species-specific studies are also important and allow us to identify some relationships that may be overshadowed in a bird community (Rodrigues et al., 2020).

The CLBI harbors a diverse community of avian haematozoan lineages distributed among 63% (44/70) of the bird species sampled in this study and we estimated an overall parasite prevalence of ~22%. Estimates for the prevalence of haemosporidian parasites in bird communities from Brazil indicate a great variation both among different ecosystems and between different sites in the same ecosystems. The estimated prevalence in Cerrado varied from 21% to 42% (Belo et al., 2011; Fecchio et al., 2013; Lacorte et al., 2013). In other Brazilian habitats it has also been observed a great variation in prevalence estimates, e.g., 17.4% (Fecchio et al., 2017c) to 21.7% (Svensson-Coelho et al., 2013) in Amazonian Region; 38.5% (Lacorte et al., 2013) to 42% (Ferreira Junior et al., 2017) in Seasonally Dry Tropical Forest; and 12.4% to 39.6% in Atlantic Forest (Ribeiro et al., 2005; Sebaio et al., 2010; Lacorte et al., 2013). This considerable variation in prevalence among studies is evidence that we still have many aspects of this complex, spatially variable parasite–host system to understand. The avian haemosporidian parasite–host community in northeast Brazil adds to our understanding of the distribution and diversity of avian haemosporidian parasites and examines the ecological factors that influence host susceptibility. However, we acknowledge that much remains to be investigated in the parasite-host relationship in Restinga and suggest that future studies use information from blood smears and mixed infections to extend the ability to detect and identify haemosporidians in this bird community.

Conclusions

In this first exploration of avian haemosporidian parasites in a largely unexplored region of Brazil, we could demonstrate that this environment harbors a high diverse community of Plasmodium and Haemoproteus parasites. We established that migration can be an important factor predicting the prevalence of Haemoproteus, but not Plasmodium, in hosts. Thus, in CLBI, Haemoproteus lineages infect preferably resident birds and should be more difficult to disperse into new environments. The other individual- and species-level traits were not important in determining the probability of infection by Plasmodium or Haemoproteus in our study, which indicates a great variation of the influence of these factors on haemosporidian prevalence in different communities. The temperature, but not the rainfall, seems to predict the Plasmodium prevalence in this bird community. This result raises the possibility that ongoing climate change will impact the dynamics of Plasmodium transmission, a subject that should be explored in future studies. The higher number of birds infected by Haemoproteus lineages than by Plasmodium is an uncommon finding in Brazil and led us to suggest that northeast Brazil must have a different haemosporidian infection dynamics when compared to other studied regions of the country. Further investigations in northeast biomes and sampling of the haemosporidian vectors are needed to better understand the transmission dynamics and to elucidate the factors promoting higher levels of Haemoproteus infection in birds of this region.

Supplemental Information