Variation in ontogeny and strength of immune defence mechanisms can be integrally related to variation in life-history strategies and determined by trade-offs during development. However, little is known about the ontogeny of immune function in wild birds, especially in altricial birds and in a comparative context across altricial species with diverse life-history strategies.
In this study, we examined the ontogeny of constitutive immunity in a group of 22 passerine species sampled in tropical Venezuela and north temperate Arizona.
Our results show activity of constitutive components of the immune defence at 1–3 days posthatching and an increase in immune activity with age. Interspecific variation in immune activity at hatching was mainly explained by extrinsic factors mediated by the mother (egg size and egg temperature), suggesting an important role of maternal effects on offspring immunity at hatching. In contrast, the increase in agglutination activity with age suggests that immune function in older nestlings reflects intrinsic development. The increase in immune activity was greater in species that hatched with lower initial levels, and was somewhat negatively related to growth rate across species.
Our results suggest slower intrinsic development of immune function may be compensated by larger maternal contributions. Slower intrinsic development of immune function, in turn, may reflect a trade-off with faster somatic growth. Our study highlights the importance of both maternal (extrinsic) and endogenous (intrinsic) contributions to variation in immune function across altricial species that may reflect an important axis of developmental strategies.
The ability to resist infections early in life is expected to be under intense selection pressure because infection, as well as other extrinsic sources of stress during development, may have severe impacts on fitness (reviewed by Monaghan 2008). Yet, development of the immune system is costly (Klasing & Leshchinsky 1999) and can trade-off with development of other internal systems or rate of development (Fair, Hansen & Ricklefs 1999; Soler et al. 2003; Brommer 2004; Pitala et al. 2010). Such trade-offs during development may be related to variation in life history and developmental strategies (Sheldon & Verhulst 1996; Norris & Evans 2000). Yet, variation in the rate of intrinsic development of immune function and possible trade-offs with other components of developmental strategies, such as offspring growth rates, has not been compared across species.
In birds, development of the immune system begins during the embryonic and posthatching periods (Apanius 1998; Starck & Ricklefs 1998; Fellah, Jaffredo & Dunon 2008). The immune function of juveniles was traditionally assumed to be little developed in the first weeks posthatching and instead developing later in life (Apanius 1998). Thus, the nestling period is considered a critical window of vulnerability to pathogens. However, some protection during this early period is provided by maternally transmitted immune components in the egg (e.g. Grindstaff, Brodie & Ketterson 2003; Boulinier & Staszewski 2008; Hasselquist & Nilsson 2009). Maternally transferred antibodies can have long-lasting effects on the development of the humoral immune system through, for example, an educational effect (e.g. Grindstaff et al. 2006; Reid et al. 2006), an imprinting effect (e.g. Lemke, Coutinho & Lange 2004) or a blocking effect (e.g. Staszewski et al. 2007; Staszewski & Siitari 2010). Genetic differences in transfer of antibodies to the egg yolk can cause differential early offspring immunity (Abdel-Moneim & Abdel-Gawad 2006; Hamal et al. 2006). Extrinsic influences on mothers (i.e. exposure to pathogens) may also influence vertical transmission of immune components to offspring (Staszewski et al. 2007; Hasselquist & Nilsson 2009). Finally, relative egg size (corrected for adult size) may contribute differing resource levels to embryos and early nestlings (Krist 2011). Thus, variation among species in maternal contributions to eggs (Addison et al. 2009) may affect differential early expression of immune function, although this possibility has not been studied.
Maternal effects may not be restricted to egg constituents, because mothers also may influence the development of nestling immune function by providing a suitable environment for embryo development through parental care (Rahn & Ar 1974; Saino, Calza & Møller 1997). Indeed, parental behaviour (i.e. nest attentiveness) plays an important role in keeping embryo thermal conditions within or outside the optimal range for development (Martin et al. 2007; Martin & Schwabl 2009). This parentally induced variation in embryo temperatures can thereby influence development and performance of the immune system (Ardia, Pérez & Clotfelter 2010; Martin, Arriero & Majewska 2011b; DuRant et al. 2012). Yet, whether these maternal effects (eggs, thermal environment) interact with endogenous development of immune function differently among species remains unstudied.
Recent evidence shows intrinsic development of immune function through endogenous production of specific antibodies (Grindstaff et al. 2006; Reid et al. 2006; King, Owen & Schwabl 2010) and activity of constitutive components within the first weeks after hatching in species with an altricial mode of development (Mauck et al. 2005; Palacios et al. 2009). This intrinsic development is contrary to the widespread idea of limited endogenous production of immune components during the first period of posthatching life. Documentation of the extent of early endogenous development of immune components in wild birds is rare but needed, and variation among species is unstudied. Ultimately, our understanding of the interplay of maternal (eggs, parental care) vs. endogenous influences on immune function during ontogeny of postnatal offspring, and its variation across species, is still limited.
The domestic chicken has served as a model species for the study of immune function in vertebrates (e.g. Leslie 1975; Fellah, Jaffredo & Dunon 2008), but it may not provide a suitable model for the broad diversity of wild birds. Domestic chickens are maintained in a controlled environment, whereas wild birds are faced with substantial variation in environmental stressors (e.g. food, temperature, predation risk, parasite pressure). These stressors might affect trade-offs during development of immune function. Moreover, ontogeny of immune function in a precocial species, like chickens, may be quite different from species with an altricial mode of development given the very different developmental trajectories of these two modes. Thus, a comparative study of both maternal influences and endogenous production during the ontogeny of immune function among wild altricial bird species can provide important insight into the extent and types of variation that may occur in nature.
In this study, we examined interspecific variation in immune function at hatch and during postnatal development to explore the potential roles of maternal vs. endogenous effects. In particular, we conducted a comparative field study of the ontogeny of immune function relative to parental care, egg size and nestling growth rate among nestlings of a phylogenetically diverse group of 22 passerine species, including 18 spp. sampled in a tropical site in South America and 4 spp. sampled in a temperate site in North America. Tropical bird species have longer incubation periods than related temperate species, and tropical species show extensive variation in incubation periods (e.g. Martin et al. 2007). Thus, by including species from the tropics and the temperate zone, we have a set of species with broad natural variation in developmental periods.
Materials and methods
The study was conducted between 2005 and 2009 at two field sites in North and South America. The temperate site was located in the Mogollon Rim in north-central Arizona, USA (34°N, 111°W latitude) and consisted of snowmelt drainages of a high-elevation mixed forest (2400 m). The tropical site was a near-cloud tropical forest (1350–2000 m) in the northern Andes of Venezuela (09°42′N, 69°42′W). The site was at Yacambú National Park, a mountainous area composed mainly of primary forest with mean annual rainfall of 2047 mm with maximum between May and July (see Martin et al. 2007 for more details on the field sites). The set of species in our study includes passerine birds that use forest montane habitat, but exhibit broad variation in developmental periods (see Fig. 1 for species list and Table S1, Supporting information for life-history details).
Measurement of Immune Parameters
As estimation of constitutive immune defence, we used the activity of natural antibodies and complement proteins (Ochsenbein & Zinkernagel 2000; Lee 2006), following the protocol of Matson, Ricklefs & Klasing (2005) scaled down to the amount of plasma available in our samples (between 15–25 μL). We measured agglutination of rabbit erythrocytes (which provides a measure of natural antibody titres) and lytic activity (that reflects complement protein activity and natural antibodies). Immune activity was scored as –log2 of the highest dilution exhibiting agglutination (Nab) or lysis (complement). We collected blood samples from nestlings (ca. 30–75 μL) by venipuncture at two standardized stages of development, 1–3 days posthatching (hereafter hatching), and when the primary feathers broke their sheaths (hereafter pin break), which usually occurs between 7 and 11 days after hatching. As most of the agglutination activity observed with this technique is attributed to endogenously produced IgM (Matson, Ricklefs & Klasing 2005), we used the difference between Nab titres at hatching and pin break as an estimation of endogenous production of natural antibodies. The amount of blood collected from nestlings followed the recommendations of animal care (below 10% of total blood volume, or < 1% of body mass). Blood samples were kept in coolers in the field and centrifuged within 8 h to separate the plasma. Plasma samples were stored frozen at −20 °C until analyses were performed. Comparisons between Nab titres at hatching and pin break were made using samples from the same breeding season to account for any differences between years.
Data were obtained as part of a long-term project from 1988 to 2009 at the Arizona site and during the breeding seasons of 2002–2008 in Venezuela. Nests were found and monitored following standardized protocols (see e.g. Martin et al. 2007). Embryonic temperatures were measured by placing a thermistor in the centre of one egg and measuring temperature experienced by the embryo every 12 s or 24 s in 3–9 nests per species over 5 days per nest, on average (Martin et al. 2007). Egg size ranged from 1·3 to 7·8 g, but much of this variation reflects differences in adult sizes; egg size is strongly related to adult size across these species (Martin et al. 2006). The more relevant character for analyses here is relative egg size, which is the residual egg size after correcting for adult size. These data were taken from Martin (2008) because those analyses included larger numbers of species that better characterize the relationship between egg size and adult size. Relative egg size then reflects whether eggs are relatively large or small for the body size of a species (Martin 2008). Incubation temperate ranged from 31·7 to 36·1 °C, with an average of 35·0 °C. Incubation period was quantified as the difference in days between the last egg laid and the last egg hatched and ranged from 12·4 to 27·2 day, with an average of 15·3 day. Nestling period was quantified as the difference in days between the last egg hatched and the last nestling fledged (Table S1, Supporting information). Nestlings were measured every 2 days, and nestling growth rates were estimated based on the growth rate constant, K, which is independent of size and standardized across species (Martin et al. 2011a).
Statistical models were fitted by generalized least squares (GLS) incorporating phylogenetic information (Fig. 1 and Appendix S1, Supporting information) with a Martin's and Hansen's correlation structure (Paradis 2006). For analytical purposes, we calculated averages per brood and species for the immune and developmental parameters. We computed the change in natural antibody titres from hatch to pin break using species averages instead of data from different broods, as we assume that the averages will be closer to the true species means. To control for allometric effects on physiological measurements, log10-transformed nestling body mass was included in the analyses. Sample size differed slightly between analyses, as not all the variables could be obtained for all the species. We computed phylogenetic autoregressions to estimate parameters of character evolution and the proportion of variance explained by phylogenetic associations among species. We used Moran's autocorrelation index to test for the absence of phylogenetic autocorrelation on the immune parameters (Diniz-Filho 2001). All analyses were carried out in R 2.11.1 (R Development Core Team 2010) using the package ‘ape’ (Paradis 2006). We used the software MESQUITE (Maddison & Maddison 1997) to construct the phylogenetic tree.
Interspecific differences in the ability to agglutinate foreign erythrocytes (i.e. natural antibody titres) at hatching were explained by a phylogenetically controlled model that included two maternal influences: relative egg size and egg temperature (relative egg size: t19 = 4·171, P < 0·01; egg temperature: t19 = 2·690, P = 0·016; Fig. 2). Sampling site (Arizona/ Venezuela) was not significant (site: t19 = 0·663, P = 0·517). Thus, maternal influences were important for immune expression of hatchlings.
Natural antibody titres also developed endogenously and increased significantly with age irrespective of sampling location (age: t21 = −4·788, P < 0·001; site: t21 = 0·442, P = 0·663). Immune function activity was positively correlated between hatch vs. pin break across species when controlling for phylogeny (r = 0·55, P < 0·001, N = 22 spp.).
The extent of the increase in agglutination scores from hatching to pin break was inversely related to levels recorded during the first days posthatching (r = −0·798, P < 0·001, N = 22 spp.). In other words, nestlings of species that hatched with lower natural antibody levels experienced greater increases in the production of natural antibodies during the first weeks after hatching. To account for the possibility that this result was a phenomenon of regression to the mean (Kelly & Price 2005), we examined whether interspecific variation in natural antibody titres exceeded intraspecific variation. Immunological data of 123 broods sampled at hatch or pin break showed significant differences between species (hatching Nab: F20,38 = 8·51, P < 0·001; pin break Nab: F20,58 = 4·52, P < 0·001), and a significant proportion of the variation among broods was explained by species (hatching Nab: 0·73, pin break Nab: 0·49). This indicates that natural antibody levels differ significantly between species even while accounting for within-species variation, and species that had high values at hatching had correlated higher values at pin break (see above).
Given that natural antibody levels at hatching was explained by maternal effects (relative egg size and egg temperature), we examined their possible contribution to the change to pin break, along with nestling growth rate, while controlling for nestling body mass. We found the change in natural antibody titres from hatching to pin break was negatively related to nestling growth rate (Fig. 3a) and relative egg size (Fig. 3b), when nestling mass and phylogeny were taken into account (growth rate: t18 = −2·794; P = 0·014; relative egg size: t18 = −3·142; P = 0·007; nestling body mass: t18 = −3·695; P = 0·002). The change in natural antibody titres was not explained by egg temperature (t18 = −0·627; P = 0·54) nor sampling location (t18 = −0·461; P = 0·65). Moreover, growth rate was not confounded with body mass in our study species (t18 = −1·001, P = 0·333), as was shown more generally across more songbird species (Martin et al. 2011a). The species to the far left in Fig. 3a is Mionectes olivaceus, which has quite slow growth (Martin et al. 2011a) and, thus, extends the growth rate axis. Removing this species from the analyses results in a nonsignificant negative trend between growth rate and change in natural antibody titres (t17 = −0·753; P = 0·465), although maternal effects remain significant (relative egg size: t17 = −2·319; P = 0·037).
Egg temperature drops out in part because it is strongly correlated with relative egg size (see Martin 2008). If we remove relative egg size from the model, then nestling growth rate remains significant and egg temperature shows a marginally significant positive relationship (growth rate: t18 = −2·345; P = 0·034; egg temperature: t18 = 1·993; P = 0·066; nestling body mass: t18 = −3·435; P = 0·004).
The activity of complement proteins was very low at both developmental stages in which we sampled the nestlings. However, we also observed an apparent increase in lytic activity of nestlings with age; only 7% of the broods sampled at hatching showed lytic activity, while 15% of the broods sampled at pin break exhibited lytic activity. Due to the low lytic activity detected in nestlings in general, this immune parameter was not considered in analyses.
To test for phylogenetic autoregressions in the immune parameters, we calculated Moran's autocorrelation index and tested the null hypothesis of no correlation derived from the phylogeny. We found no phylogenetic signal on natural antibody titres and a very small proportion of the variation among species explained by the phylogeny (hatching Nab: P = 0·398; pin break Nab: P = 0·646).
Our results show that immune activity at hatching and endogenous production of constitutive components vary extensively across species of passerine birds. The random nature of lymphocyte generation has been speculated to yield a fixed rate of immunological development across species (Martin, Weil & Nelson 2006). However, recognition of the variation in immunological development is important because it can influence differential vulnerability to pathogens and highlight that species differ in their developmental strategies for immune function.
Maternal antibodies may contribute to early immune function in altricial birds, but persist only a few days in nestling plasma in passerine birds (Nemeth, Oesterle & Bowen 2008; King, Owen & Schwabl 2010; but see Garnier et al. 2012 for long-lived sea birds). However, maternal effects mediated by parental care behaviour may also have an important role in promoting offspring immune function. The results of our study suggest that a significant part of the variation in innate immune activity at hatching is related to both maternal contributions. Egg size is a major determinant of maternal nutrients that influence offspring quality in many species (Williams 1994; Krist 2011), and our study shows that it was an important predictor of immune activity during the first days after hatching. Similarly, egg temperature during incubation is strongly associated with female incubating activity (Martin 2002; Martin et al. 2007) and has recently been associated with immune activity within and across species (Ardia, Pérez & Clotfelter 2010; Martin, Arriero & Majewska 2011b; DuRant et al. 2012). We also found that egg temperature was strongly associated with immune function at hatching (Fig. 2b), and a suggestion that warmer egg temperatures were associated with greater endogenous development of innate immune function. Thus, increasing maternal contributions to egg size and embryonic temperature are associated with higher immune activity within the first days after hatching, with potential effects after the initial stages (Fig. 2).
Natural antibodies are part of the first line of defence against infections, and have been classified as constitutive components of both the innate and adaptive immune defence (Ochsenbein & Zinkernagel 2000; Matson, Ricklefs & Klasing 2005). They are present in serum at low concentrations and are mainly constituted by the IgM isotype, although small amounts of IgG (the homologous IgY in birds) and IgA have also been described (Ochsenbein & Zinkernagel 2000). Although some natural antibodies may be from maternal origin in nestlings, maternal IgM is absorbed into embryonic circulation at very low concentrations, and, thus, most of the agglutination activity observed in nestling plasma is attributed to endogenously produced IgM (Matson, Ricklefs & Klasing 2005; Hamal et al. 2006). The low lytic activity observed in all nestlings indicates that the complement cascade is still at low efficiency during the first weeks after hatching (Mauck et al. 2005; but see Palacios et al. 2009).
Our results show that antibody titres at hatching and pin break were correlated across species, which suggests that species have some intrinsically determined base of immune function that remains correlated during development. However, species differed in the relative extent of endogenous change in immune function; the relative change in innate immune activity was greater for species with low initial levels. The relative change in immune function was related in part to maternal contributions to relative egg size, where greater contributions (larger relative egg size) led to increased immune function at hatch (Fig. 2a) but smaller endogenous changes (Fig. 3b). One possible reason for such effects may be that maternal effects drive investment in immune function development during the nestling stage, whereby nestlings of species with low maternal investment invest more in immune development. Alternatively, species with slower endogenous development of innate immune function may attempt to offset this slow development with increased maternal contributions. In either of these cases, greater investment in immune development during the nestling period may come at the expense of somatic growth, as suggested by the negative trend between growth rate and change in immune function (Fig. 3a). However, the loss of significance when the slowest species was removed leaves the importance of this trade-off unclear and deserving of further study across more species. Nonetheless, a trade-off between immunity and growth rate has been observed in several studies at the intraspecific level (Soler et al. 2003; Brommer 2004), which underscores the possibility of a trade-off across species. Existence of such a trade-off sets up the possibility that selection imposed by nest predation risk for faster nestling growth rates (e.g. Martin et al. 2011a) then causes a slower rate of endogenous development of immune function that is possibly offset through increased maternal contributions. In short, the initial level and subsequent rate of endogenous development of immune function may reflect a delicate balance between protecting developing juveniles from pathogens against rate of growth to minimize other environmental sources of mortality.
In conclusion, the variation among species in initial immune activity vs. rate of endogenous development in immune activity is relatively unexplored (although see Mauck et al. 2005 for an intraspecific example), but may reflect an important axis of developmental strategies. Our study highlights the importance of both maternal (extrinsic) and endogenous (intrinsic) contributions to variation in immune function across altricial species that are integrated with developmental strategies, including growth rate. Thus, an approach to ontogenetic development of immune function that integrates it with broader developmental and maternal strategies is needed in the future.
This work would not have been possible without the help of numerous field assistants who searched for nests and helped collecting the data reported here, in particular L. Biancucci, R. Ton, A. Cox and W. Goulding. We thank Prof. C. Bosque, INPARQUES and the Yacambu National Park staff for support with field logistics in Venezuela, and the Coconino National Forest staff in Arizona. We are grateful to Douglas Emlen and Creagh Breuner from the University of Montana for logistical support during laboratory analyses. We thank M.G. Palacios for helpful comments on an earlier version of the manuscript. This work was supported by National Science Foundation grants for work in Venezuela and Arizona, and the United States Geological Survey Global Climate Research Program for Arizona. EA was supported by a postdoctoral fellowship from the Spanish Ministry of Science and Education. The work adhered to recommended practices for the use of wild birds in research and was carried out under permission from University of Montana's Animal Care committee (059-10TMMCWRU), US Fish and Wildlife Service (MB791101-1) and Venezuelan authorities (FONACIT: DM/0000237; INPARQUES: INP-005-2004; and Ministerio del Ambiente 01-03-03-1147). Any use of trade names is for descriptive purposes only and does not imply endorsement by the US Government.