Prenatal and postnatal parental effects on immunity and growth in ‘lactating’ pigeons
Laboratoire Ecologie et Evolution, UMR7625 UPMC CNRS ENS, Université Pierre et Marie Curie, 7 quai Saint-Bernard, 75005 Paris, France
Université de Versailles Saint-Quentin-en-Yvelines UVSQ, 55 avenue de Paris, 78035 Versailles, France
Correspondence author. Laboratoire Ecologie et Evolution, Université Pierre et Marie Curie, UMR 7625, Bât. A, 7ème étage, case 237, 7 quai Saint-Bernard 75252, Paris, France. E-mail: email@example.com
Correspondence author. Laboratoire Ecologie et Evolution, Université Pierre et Marie Curie, UMR 7625, Bât. A, 7ème étage, case 237, 7 quai Saint-Bernard 75252, Paris, France. E-mail: firstname.lastname@example.org
1. Parental effects can have profound consequences on offspring phenotype. Still, little is known about the relative influence of prenatal versus postnatal parental effects of parasite exposure of parents on offspring traits.
2. In this study, we investigated the respective role of a prenatal and a postnatal immune challenge of parent feral pigeons (Columba livia) on offspring humoral immunity, growth and survival. We used a cross-fostering design and antigen injections in biological and foster parents. Feral pigeons are particularly suitable for studying the effects of parental immune challenges because they can affect the phenotype of their young through the transmission of prenatal antibodies in the egg and postnatal antibodies in the ‘crop milk’, a substance produced in the crop of both parents.
3. Results show that a prenatal immune challenge of biological parents with keyhole limpet haemocyanin (KLH) antigen decreased the humoral response against KLH of nestlings injected at 14 days of age. In contrast, a postnatal immune challenge of foster parents with KLH enhanced the humoral response of 1-year-old juveniles exposed to a second KLH injection, but only when these juveniles had received their first injection at 3 days of age.
4. No effect on nestling and juvenile response to another antigen (NDV) was observed, indicating that the changes in humoral responses were specific to the KLH injected in parents. In addition to this, prenatal and postnatal parental immune challenges had an interaction effect on fledging body mass, but no effect on juvenile survival.
5. This study shows that pre- and postnatal exposure to antigens in parents has contrasted effects on offspring humoral response and growth. Moreover, it shows that the timing of an early exposure to antigens in nestlings has important effects on their specific humoral response.
6. This study thus suggests that pre- and postnatal parental effects have distinct roles in shaping the phenotype of the offspring on different time scales and calls for further investigations on the potential adaptive role of combined parental effects. Moreover, it suggests that pigeon milk has positive effects on offspring humoral immunity and thus could potentially have a similar immune role as mammalian milk.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Parental effects are now recognized as a major source of transgenerational phenotypic plasticity that has short-term and long-term consequences on offspring ontogeny and fitness (Mousseau & Fox 1998). In the context of ecological immunology, maternal exposure to parasites at the time of breeding can profoundly affect maternal effects in different ways. For instance, an immune challenge with artificial antigens can indirectly affect parental effects as it is now well established that mounting an immune response is costly and can negatively impact the intensity of parental investment (Bonneaud et al. 2003). In addition, an immune challenge can directly affect parental effects by the transfer of specific antibodies from mothers to offspring, which are known to affect several aspects of offspring phenotype (Grindstaff, Brodie & Ketterson 2003; Boulinier & Staszewski 2008; Hasselquist & Nilsson 2009). Indeed, maternal antibodies can confer a transient protection of the young against parasites and can therefore have positive effects on offspring survival (Heeb et al. 1998). They also have profound effects on offspring immune ontogeny, through positive influences on the intensity of the humoral response and on the ontogeny of the immune system (educational effects, Gasparini et al. 2006; Grindstaff et al. 2006; Reid et al. 2006), but also through transient blocking effects on the young immune response, because high levels of maternal antibodies can suppress the stimulation of the immune system (blocking effect, Staszewski et al. 2007; Staszewski & Siitari 2010). Furthermore, maternal antibodies have been shown to positively affect offspring growth, likely through reallocation of energy of offspring from costly immune response to growth (Robison, Stott & DeNise 1988).
Maternal antibodies can be transmitted through two main routes: before the birth of the young (prenatal antibody transmission) or after the birth (postnatal antibody transmission) (Boulinier & Staszewski 2008). Prenatal antibodies are mainly Immunoglobulins Y (IgY) contained in avian eggs and IgG transmitted through the placenta in mammals (Chucri et al. 2010). Birds are ideal models to study the prenatal transmission of antibodies because the mother can potentially modulate the amount of antibodies transferred in the egg yolk (Saino et al. 2002; Hasselquist & Nilsson 2009). Such prenatal antibodies have been shown to affect several fitness parameters in offspring (reviewed in Grindstaff, Brodie & Ketterson 2003; Hasselquist & Nilsson 2009). In contrast, mammals are ideal models to study the postnatal transmission of antibodies through the colostrum and breast milk (e.g. Marquez et al. 2003). Such postnatal antibodies provide offspring with protection against diseases in humans (Van de Perre 2003; Hanson 2007), domestic animals (reviewed in Grindstaff, Brodie & Ketterson 2003) and potentially in wild animal populations (Graham et al. 2010). Colostrum and milk of some mammals like humans or rabbits are mainly composed of immunoglobulins A (IgA), which are believed to play a central role in gut protection against local infections (Van de Perre 2003) and potentially in immune function priming (Hanson 1998). Indeed in mammals, antibodies transmitted postnatally seem to have a parallel but distinct role compared with antibodies transmitted prenatally (Morshed et al. 1993). However, few studies have been able to disentangle the respective role of pre- and postnatal antibodies in animals and their evolutionary consequences in wild populations.
Both prenatal and postnatal maternal effects are indeed likely to play a role in the evolution of natural systems (Biard, Surai & Møller 2006). To better understand the evolutionary consequences of the prenatal and postnatal effects of parental exposure to parasites, we need to investigate the direct and indirect consequences of an immune challenge in parents on offspring phenotype. Columbidae, such as feral pigeons Columba livia (Fig. 1), are ideal models to address this question. Indeed, parent and offspring free-living feral pigeons share the risk of exposure to the same deleterious parasites (such as Chlamydia psitacci and Trichomonas), which are responsible for a high nestling and subadult mortality (30% and 50%, respectively, Johnston & Janiga 1995). In this context, the transmission of maternal antibodies to the young may have strong positive effects on offspring fitness by conferring a protection against such parasites through direct protective effects of maternal antibodies and/or through an enhancement of offspring immune defence (Heeb et al. 1998; Kallio et al. 2006; Nemeth & Bowen 2007, but see Addison, Ricklefs & Klasing 2010).
Moreover, adult pigeons feed their chicks with a lipid-rich substance produced in their crop, named crop milk (Johnston & Janiga 1995). It is known to contain nutrients, minerals and growth factors (Shetty et al. 1992), but also immune active substances such as carotenoids (Eraud et al. 2008) and immunoglobulins (Engberg et al. 1992). This model offers us a unique chance to disentangle the effects of a prenatal immune challenge (inducing the transmission of prenatal egg antibodies through the yolk), and of a postnatal immune challenge (inducing the transmission of postnatal antibodies through the crop milk) on offspring phenotype. Although the production of milk is quite an exception in birds (it is known only in Columbids: Goodwin 1977; emperor penguins: Prevost 1962 and flamingos: Studer-Thiersch 1967), it is suspected that other altricial birds feeding their young by oral regurgitation (like vultures or swifts) could also potentially transfer salivary antibodies to their offspring (Apanius 1998).
In this study, we aimed to unravel the respective importance of prenatal and postnatal parental effects on growth, humoral immune response and survival of chicks following an immune challenge. To this end, we injected selected adult feral pigeons with an antigen (keyhole limpet haemocyanin, KLH) and swapped eggs between nests, allowing us to obtain chicks with only a prenatal parental immune challenge, only a postnatal parental immune challenge, both postnatal and prenatal or no parental immune challenge. We then recorded the short-term and long-term effects of these parental treatments on offspring humoral immune response against two antigens (KLH and NDV), as well as on their growth and survival. Because we did not directly manipulate parental antibodies in eggs or milk, the effect of parental antibodies on chick humoral response may be confounded by parental effects that affect the whole offspring immune system (Grindstaff et al. 2006). To ensure that effects observed on the humoral response were mediated by specific parental antibodies, we considered not only the humoral response of offspring against KLH, but also their humoral response against NDV, an antigen to which parents were not exposed. If the effects mediated by maternal antibodies are specific to the antigen injected in parents, we expect to find different effects of parental immunization on offspring humoral response against KLH and NDV antigens.
Materials and methods
Feral pigeons usually lay two to six clutches per year. They lay two eggs 1 day apart, which hatch 18 days after laying. Hatching success is between 75 and 90% (Johnston & Janiga 1995). Hatchlings are immediately fed with crop milk, a protein and lipid-rich substance secreted in the crop of both parents (Goodwin 1977). The crop milk is progressively mixed with cereals to feed the growing young, and production of crop milk by parents usually stops around 10–12 days after hatching (Johnston & Janiga 1995).
Parental Immune Treatment
In February 2009, 50 females and 67 males feral pigeons (C. livia) from two different populations were captured and kept in 10 outdoor aviaries in similar conditions and fed ad libitum with a mix of maize, wheat and peas, with mineral and vitamin supplements. Birds were kept in captivity for 3 months for acclimation to obtain naturally representative pigeon physiology and behaviour (e.g. Pascual, Fryday & Hart 1999). After acclimation, all individuals were weighed to the nearest g, tarsus length was measured to the nearest mm, and body condition was calculated as the residual of the regression of body mass (log transformed) on tarsus length (log transformed) (Jakob, Marshall & Uetz 1996). Aviaries did not differ according to sex ratio (z115 = 0·082, P = 0·94), population of origin (z115 = −1·14, P = 0·25) or body condition (t115 = −1·23, P = 0·22).
In the ‘Antigen-injected’ group, 57 adults (24 females and 33 males) from five aviaries were injected subcutaneously with 50 μg of KLH (an artificial antigen commonly used in immunology, Hasselquist et al. 1999) and with 50 μL of a solution of inactivated bacteria Chlamydia psittaci (for another study). In the ‘Sham-injected’ group, 60 other adults (26 females and 34 males) from the five other aviaries were injected with a neutral saline solution (phosphate-buffered saline). Individuals of the same reproductive pair received the same immune treatment. Antigen- and sham-injected groups did not differ according to sex ratio (z115 = 0·13, P = 0·89), population of origin (z115 = −0·33, P = 0·74), body condition (t115 = −1·25, P = 0·21) or melanin-based coloration (t115 = 0·84, P = 0·40) (feral pigeons display a variable melanin-based coloration that is linked to various immune traits, e.g. Jacquin et al. 2011). One month later, a second injection of antigens or saline was performed. One week after the second injection, KLH-injected adults had a higher circulating anti-KLH antibody level (mean ± SE: 4·69 ± 0·13) than sham-injected adults (mean ± SE: 1·54 ± 0·09) (effect of injection group on anti-KLH antibody level: F1,112 = 405·2, P < 0·001), and females started to lay eggs. Between June and September, each female laid from 0 to 3 clutches. Nests were visited each day to identify the first egg from the second egg laid. The day of a clutch completion, females were weighed and their body condition at the time of egg-laying was determined. A blood sample was then taken to assess females’ blood anti-KLH antibody levels at the time of egg-laying.
Experimental Groups of Chicks
Between June and September, 68 clutches were laid. The day of a clutch completion, the first egg of the clutch was collected for antibody assay. The second egg of each clutch was cross-fostered with another nest with a similar laying date to create experimental groups of chicks differing by prenatal and postnatal parental immune treatments (immune treatment of the biological and foster parents). All chicks came from the second egg of the clutches and were raised alone. Seventy-eight percentage of the second eggs hatched successfully (53 from 68 eggs), and parental treatment did not affect this proportion (effect of parental treatment on hatching success: z66 = 1·26, P = 0·20). Three chicks died soon after hatching; 50 chicks survived until fledging and were used in this study. In the ‘Pre– Post–’ group, biological and foster parents were sham-injected (no prenatal and no postnatal parental immune challenge, N = 13). In the ‘Pre− Post+’ group, biological parents were sham-injected, whereas foster parents were antigen-injected (only postnatal parental immune challenge, N = 16). In the ‘Pre+ Post−’ group, biological parents were antigen-injected, whereas foster parents were sham-injected (only prenatal parental immune challenge, N = 11). In the ‘Pre+ Post+’ group, both biological and foster parents were antigen-injected (both prenatal and postnatal parental immune challenges, N = 10). These experimental groups of chicks did not differ according to sex ratio (mean = 0·56 ± 0·07; χ2 = 2·15; d.f. = 3, P = 0·54), hatching date (mean hatching date = 24/7/09 ± 5j; χ2 = 0·86; d.f. = 3; P = 0·83) or body mass at hatching (mean = 16·8 g ± 0·49; χ2 = 0·68; d.f. = 3; P = 0·57).
Crop Milk Sampling
Three days after hatching, we collected approximately 0·5 g of crop milk from the crop of young chicks using a 5-mm-diameter plastic catheter attached to a 2-mL syringe (Jannssens et al. 1999) to assess its antibody content [specific anti-KLH antibody level and natural antibody (NAb) level]. Crop milk is provided by both parents, and we could not distinguish crop milk coming from foster fathers or mothers. Crop milks and egg yolks were diluted 1 : 1 and homogenized in phosphate-buffered saline. An equivalent volume of chloroform was added, and the supernatant was used for antibody assays after homogenization and centrifugation (6 min at 6000 g).
Nestling Humoral Response
To assess the effect of a parental immune challenge on the immune response against KLH in young nestlings, we injected all nestlings with 25 μg of KLH, which is the same antigen as injected in parents. To assess the specificity of parental effects on offspring humoral response, we also challenged nestlings with 0·05 mL of vaccine against Newcastle disease virus (NDV; Colombovac PMV, Fort Dodge), an antigen to which parents were not exposed. As we aimed at comparing the intensity of the humoral response between our experimental groups, we injected all nestlings with KLH and NDV antigens.
As little is known about the dynamics of antibody response in young pigeons, we injected half of the nestlings at the age of 3 days (early injection: N = 24) and half of the nestlings at the age of 14 days (late injection: N = 26) to maximize our chance of triggering an effective immune response in young squabs. Early- and late-injected nestlings were equally distributed in each experimental group (χ2 = 0·088, d.f. = 3, P = 0·99). We measured the antibody level 11 days after the antigen injection. We took 0·5 mL of blood from the brachial vein at the age of 14 days for nestlings injected at 3 days and at the age of 25 days for nestlings injected at 14 days. We measured anti-KLH and anti-NDV antibody level in nestling plasma using ELISA assay (see after). Sex of nestlings was molecularly determined from blood samples following Griffiths et al. (1998).
Juvenile Humoral Response
In 2009, the chicks were separated from their parents when they reached the age of 38 days (juvenile feral pigeons usually leave the nest at 28 to 32 days of age, Johnston & Janiga 1995) and kept in separate cages with food ad libitum for 1 year. We monitored survival during their first year of life. In March 2010, 33 juveniles from the 50 chicks had survived (mean mortality rate ± SE = 34% ± 7%). This high mortality rate over the first year of life was likely due to infections in captivity (e.g. C. Psitacci), but the exact causes of death could not be determined with certainty in most cases. However, this mortality rate remains lower than mortality rate observed in wild juvenile feral pigeons (between 43% and 56%, Johnston & Janiga 1995). To assess the long-term impact of parental treatments, we measured the body mass of these 1-year-old juveniles and exposed them to another KLH (50 μg) and NDV (0·05 mL of vaccine) antigen challenge. Fourteen days later, we took 1 mL of blood from the brachial vein and measured the amount of anti-KLH and anti-NDV antibodies in the plasma. The measured humoral response should correspond to a secondary response to a repeated challenge, as juveniles received a first injection of these antigens as nestlings.
Anti-KLH Antibody Assay
Anti-KLH antibody levels in female plasma, egg yolk and crop milk extracts were determined using a sandwich ELISA following the technique adapted from Hasselquist et al. (1999). Briefly, high-binding plates (Greiner Bio-One, Solingen, Germany) were coated overnight at 5 °C with 100 μL of 40 μg mL−1 KLH in 50 mm carbonate buffer. Wells were blocked with 200 μL of 3% milk powder (Regilait Bio, Saint-Martin-Belle-Roche, France), and plasma (dilution 1 : 500 for adults and 1-year-old juveniles, 1 : 200 for chicks), yolk extract (dilution 1 : 1000) or crop milk extracts (dilution 1 : 10) were then distributed. After incubation and washing, 100 μL of rabbit-anti-pigeon IgG conjugated to horseradish peroxidase (Nordic Immunology, Eindhoven, Netherlands) was added. OPD (Sigma-Aldrich, St Louis, USA) was then added, and the reaction was stopped after 10 min by adding 50 μL HCl (1 m). Plates were then read at 490 nm in a microplate reader (Bio-Rad Laboratories, UK). As a standard, a mixture of several pigeon samples was measured in serial dilutions to calculate a relative antibody concentration after calibrations with this standard. A panel of samples from another experiment showed a high repeatability of antibody values within plates (93%, F45,46 = 13·94) and between plates (81%, F30,31 = 38·32). Anti-KLH IgY relative concentrations were log transformed and called anti-KLH antibody levels throughout the study. We were not able to rerun ELISA analyses to determine anti-KLH IgA levels in crop milks because of low sample amounts.
Specific Anti-NDV Antibody Assay
We used a PMV-1 monoclonal antibody-blocking enzyme-linked immunosorbent assay developed for birds (ELISA kit; SVANOVA Biotech, Sweden) to measure anti-NDV antibody plasma level in the plasma of nestlings and 1-year-old juveniles, following the kit instructions (http://svanova.se/filearchive/Insert%20NDV-Ab_04.pdf). Optical density (OD) values of test plasma were compared with the OD value of a kit NDV-negative control using a spectrophotometer (Microplate Reader 680; Bio-Rad, Marnes-la-Coquette France). The amount of NDV antibodies was assessed as the percentage of inhibition (PI) following the formula: [(ODnegative control − ODtest plasma) × 100]/ODnegative control. Large PI values indicate large amounts of specific antibodies directed against NDV in the plasma sample. A panel of samples from another experiment showed a high repeatability of antibody values within plates (85%, F29,32 = 12·07) and between plates (92%, F18,23 = 27·4). PI values were log transformed for statistical analyses and named anti-NDV antibody levels in the study.
Natural Antibody Assay in Crop Milk
Contrary to specific antibodies produced by the acquired arm of immunity, NAbs are part of the innate immunity and can bind to various antigens, leading to complement enzyme cascade and to cell lysis (Carroll & Prodeus 1998). NAbs are thus believed to constitute a first non-specific line of defence against parasites and to act in interaction with the acquired arm of immunity (Carroll & Prodeus 1998). As crop milk is believed to provide a protection against parasites in feral pigeons (Brand 1989), we measured both its specific antibody level and its NAb level, as a way to explore whether innate components of immunity could also be transferred to offspring through the crop milk. For this, we used a hemagglutination assay measuring the total level of natural IgG, IgM and IgA (Matson, Ricklefs & Klasing 2005) in crop milk extracts. Briefly, 25 μL of successive dilutions (from 1/2 to 1/128) of crop milk extracts was distributed in microplates (Greiner Bio-One); 25 μL of a 1% cell suspension of fresh rabbit blood (Agro-bio, La Ferté Saint-Aubin, France, supplied as 50% whole blood- 50% Alsever) was then added to each well. After agitation and incubation (90 min at 37 °C and 20 min at 18 °C), a scan was performed and a visual score of hemagglutination was attributed to each well following the description in Matson, Ricklefs & Klasing (2005). The log2 of the last dilution exhibiting agglutination was then used as a semi-quantitative measure of NAb level in crop milk.
Body mass of chicks was monitored daily from the age of 1 day to 30 days, then again at the ages of 34 and 38 days, allowing us to construct accurate growth curves for each individual. Body mass growth of nestlings was described using the logistic growth curves in the form: , where W = body mass, t = age, A = asymptote (final body mass at fledging), K = growth rate constant (slope of the linear regression between mass and time in the beginning of the growth), ti = the inflexion point of the curve (Newbrey & Reed 2009). As ti and K are usually highly correlated (Newbrey & Reed 2009), we only studied K (body mass growth rate) and A (final body mass) parameters. Estimation of the final body mass from a growth curve is more accurate than a direct estimation of body mass at the end of growth, because body mass is likely to fluctuate throughout time when fledglings reach independence (Johnston & Janiga 1995).
First, we wanted to test whether the parental immune treatments triggered an effective antibody transmission to the chicks. We started by testing whether prenatal parental antigen injection had an effect on antibody levels of females, eggs and plasma of newly hatched chicks using generalized mixed linear models with female identity as random effect (to take into account the non-independence of samples coming from the same mother). Then, we tested the effects of the postnatal parental treatment on the amount of anti-KLH IgG level and NAb level in the crop milk using a mixed model with foster parent identity as random effect to take into account the non-independence of crop milk produced by the same foster parents. To test the effects of prenatal and postnatal parental immune treatments on the humoral response and growth of nestlings and 1-year-old juveniles, we first used linear mixed models and included biological and foster parent identity and nest identity as random effects. We then compared each full mixed model with a similar full linear model without random effect as recommended by Zuur et al. (2009). Because all mixed models had higher or similar AICcs compared with linear models without random effect (ΔAICc <2), models without parent identity random effect were preferred. Similarly, mixed models with a random effect of the aviary had higher AICcs than models without aviary random effect. Models without aviary random effect were therefore preferred (Zuur et al. 2009). Prenatal parental immune treatment, postnatal parental immune treatment, injection age and their double interactions were included as explanatory variables. We started with full models with all explicative variables and successively dropped terms with P-values >0·10. In case of a significant effect, we conducted post-hoc analyses using Wilcoxon tests (as sample sizes were small) to determine which experimental group differed from the others. All tests were conducted using the R software (R Development Core Team 2010). Significance levels were set to 0·05 and tests were two-tailed.
Antibody Transmission from Parents to Chicks
Immune challenge had no significant effect on mothers’ body mass or body condition at the time of egg-laying (mixed model: body mass: mean = 298 ± 4 g, t25 = 1·04, P = 0·31; body condition: t25 = 1·66, P = 0·11). Circulating levels of anti-KLH antibodies at the time of egg-laying were higher in antigen-injected females than in sham-injected females (mixed model, effect of antigen injection on anti-KLH antibody level in females: estimate = 1·26 ± 0·43, t22 = 2·91, P = 0·0081), whereas time since injection had no significant effect on the circulating antibody level of females (mixed model, effect of time since injection on anti-KLH antibody level: t15 = −0·22, P = 0·82). Anti-NDV antibody levels of females were not related to their immune treatment (mixed model, t25 = −0·27, P = 0·79). Moreover, injected females produced eggs with higher anti-KLH antibody levels than sham-injected females (mixed model: estimate = 1·21 ± 0·29, t24 = 4·16, P < 0·001) and had chicks with higher anti-KLH antibody levels at 3 days of age (mixed model: estimate = 0·74 ± 0·25, t25 = 2·86, P = 0·008). This shows that prenatal parental antigen injection led to an effective transmission of anti-KLH antibodies from the biological mother to newly hatched chicks through the egg yolk.
However, antigen-injected foster parents produced crop milk containing anti-KLH antibody levels similar to sham-injected foster parents (mixed model: t20 = 0·59, P = 0·56). Anti-KLH antibody levels detected in the crop milk with the ELISA technique were very low compared with levels found in egg yolk and female plasma (mean relative anti-KLH levels: crop milk: 16·2 ± 2·1; egg yolk: 24220 ± 4966; plasma of 3-days-old chicks: 930 ± 241). In contrast, NAb levels of crop milk produced by antigen-injected parents tended to be higher than NAb levels of crop milk produced by sham-injected parents (mixed model: estimate = 2·80 ± 1·49, t25 = 1·88, P = 0·072).
Effects of Parental Immune Challenge on Nestling Humoral Response
Injection age, prenatal parental treatment and their interaction had a significant effect on nestling anti-KLH antibody levels after antigen injection (Table 1). Prenatal parental antigen injection significantly decreased anti-KLH humoral response for nestlings injected at 14 days of age (Wilcoxon post-hoc test: W = 29, P = 0·0059), but not for nestlings injected at 3 days of age (Wilcoxon post-hoc test: W = 57·5, P = 0·48) (Fig 2a). Postnatal parental treatment, as well as other interactions, had no significant effect on nestling anti-KLH antibody level (all P-values >0·25). Injection age, prenatal parental treatment and their interaction, as well as postnatal parental treatment, had no significant effect on anti-NDV antibody levels (all P-values >0·30, Table 1, Fig 2b).
Table 1. Influence of parental immune challenges against KLH on nestling humoral response against a first challenge of KLH and NDV, and on juvenile humoral response against a second challenge of KLH and NDV. Significant values are bold
Nestling humoral response against KLH
Nestling humoral response against NDV
KLH, keyhole limpet haemocyanin.
−1·22 ± 0·25
0·012 ± 0·32
Prenatal parental challenge
−1·01 ± 0·27
−0·011 ± 0·34
Prenatal parental challenge × injection age
0·89 ± 0·38
0·092 ± 0·49
Juvenile humoral response against KLH
Juvenile humoral response against NDV
−2·67 ± 0·76
0·012 ± 0·21
Postnatal parental challenge
−0·087 ± 0·74
−0·004 ± 0·20
Postnatal parental challenge × injection age
2·27 ± 1·01
0·22 ± 0·28
Effects of Parental Immune Challenge on Juvenile Humoral Response
Anti-KLH antibody level of 1-year-old juveniles after a second injection of antigens was affected by injection age, postnatal parental immune challenge and their interaction (Table 1). Postnatal antigen injection in foster parents tended to increase the anti-KLH humoral response of juveniles, but only if they had received their first injection of antigens at 3 days of age (Wilcoxon post-hoc test: W = 60, P-value = 0·057). Postnatal parental treatment had no significant effect on juvenile humoral response if they received their first injection at 14 days of age (Wilcoxon post-hoc test: W = 19, P-value = 0·34) (Fig 3a). Prenatal parental treatment and other interactions had no effects on KLH antibody level (all P-values >0·60). Anti-NDV antibody level (Fig 3b) and body mass of 1-year-old juveniles were not affected by injection age, postnatal or prenatal parental treatments, or by their interactions (all P-values >0·10).
Effects of Parental Immune Challenge on Chick Growth
The interaction between prenatal and postnatal immune treatments did marginally affect body mass growth rate (Table 2). Post-hoc tests revealed that the body mass growth rate of chicks with both prenatal and postnatal parental immune challenges was significantly lower (mean ± SE = 0·26 ± 0·02) than the growth rate of chicks with no parental immune challenge (mean ± SE = 0·32 ± 0·01; Wilcoxon post-hoc test: W = 33, P = 0·050), only prenatal parental immune challenge (mean ± SE = 0·31 ± 0·02) (W = 26, P = 0·043), and marginally lower than chicks with only postnatal immune challenge (mean ± SE = 0·31 ± 0·01) (W = 44, P = 0·061) (Fig 4a). In addition to this, the interaction between prenatal and postnatal parental treatments did significantly affect final body mass (Table 2). The final body mass of chicks with both prenatal and postnatal parental immune challenges was significantly higher (mean ± SE = 304·6 g ± 6·9) than final body mass of chicks with only a postnatal parental challenge (mean ± SE = 280·6 g ± 6·3) (Wilcoxon test: W = 135, P = 0·003) (Fig 4b). Injection age and other interactions did not affect any growth parameter (all P-value >0·2) and were therefore removed from the models. Visualization of the mean body mass growth curves confirms that chicks having received both prenatal and postnatal antibodies (red-dashed curve on Fig 4c) have a delayed body mass growth, but attain a similar or higher final body mass compared with chicks from the other groups (Fig 4c).
Table 2. Influence of parental immune challenges on body mass growth parameters
Body mass growth rate
Final body mass
−0·002 ± 0·022
−14·6 ± 11·9
−0·004 ± 0·019
−13·6 ± 10·8
Prenatal × postnatal parental challenge
−0·053 ± 0·030
38·5 ± 16·6
Influence of Parental Immune Challenge on Survival
Survival of chicks until 1 year old was not affected by prenatal or postnatal parental treatment, by injection age or by their interactions (generalized linear model, all P-values >0·27).
Effects of a Prenatal Immune Challenge in Biological Parents on Nestling Immunity
Prenatal and postnatal parental immune challenges had different effects on the immune phenotype of young feral pigeons. KLH injection of biological parents decreased the humoral response against KLH in young nestlings injected at 14 days of age, whereas injection of foster parents had no effect. Several mechanisms could explain this result. For instance, the prenatal immune challenge in biological parents could impair their body condition and/or modify the quality or the hormonal content of their eggs (Sheldon & Verhulst 1996), causing negative effects on offspring quality and humoral response. However, maternal body condition and body mass of newly hatched chicks were not affected by the prenatal parental immune challenge. Moreover, nestling humoral response against NDV (an antigen not injected in parents) was not affected by the prenatal parental immune challenge. This strongly suggests that maternal anti-KLH antibodies transmitted through the yolk had a specific blocking effect on the humoral response of nestlings as reported in several previous studies (reviewed in Boulinier & Staszewski 2008).
This blocking effect was only detected for nestlings injected at 14 days of age and not for nestlings injected at 3 days of age. This is surprising because a stronger blocking effect is expected for early immune challenges in nestlings compared with late immune challenges, because of the elimination of maternal antibodies in chick plasma across time (Staszewski et al. 2007). However, persistence of maternal antibodies in chick plasma is usually longer in altricial species (e.g. 27 days in pigeons, Gibbs et al. 2005) compared with precocial species (e.g. 14 days in chickens, Apanius 1998), potentially explaining the persistence of a strong blocking effect in nestlings injected at 14 days in this study. The lower blocking effect observed in nestlings injected at 3 days of age could be due to their immature immune system compared to older nestlings injected at 14 days of age (Apanius 1998), making the effect of prenatal parental treatment less detectable. In line with this, the mean anti-KLH antibody level of nestlings injected at 3 days was two times lower (mean ± SE = 0·55 ± 0·09) than the anti-KLH antibody level of nestlings injected at 14 days (mean ± SE = 1·39 ± 0·18), though the intensity of response against NDV was not affected by injection age.
Effects of a Postnatal Immune Challenge in Foster Parents on Nestling Immunity
In contrast, the postnatal parental immune challenge did not affect the humoral response of young nestlings. This is consistent with the fact that crop milk of columbids contains mostly IgAs that are believed to play a local protective role in the gut of young nestlings (Engberg et al. 1992). However, it is still unclear whether such antibodies cross the gut barrier of the young (Apanius 1998). Postnatal transmission of antibodies is believed to be particularly important in altricial species with similar postnatal feeding to pigeons, as it could limit the transmission of infections through regurgitation of infected materials (Brand 1989; Kocianova, Rehacek & Lisak 1993). Similar mechanisms could have evolved in other bird species through the transmission of antibodies in the saliva (Apanius 1998). In some mammals, like humans, antibodies transmitted postnatally are also mainly IgAs and are believed to provide local protection against gut infections (Van de Perre 2003). It remains however to be tested whether crop milk could play a protective role similar to human colostrum during the first days of life of nestling pigeons.
In this study, the ELISA technique showed that crop milk contains very small amounts of specific anti-KLH IgY, and the level of anti-KLH IgYs in the crop milk was not related to the immune treatment of parents. This could be due to the fact that crop milk contained few IgYs compared with IgAs (Engberg et al. 1992). We were not able to test this hypothesis by running new ELISA analyses to determine anti-KLH IgAs levels because of low sample amounts, but the hemagglutination assay permitted us to detect significant levels of NAbs (IgY, IgM and IgA subclasses) in the crop milk and showed potentially that antigen-injected parents tended to produce a crop milk containing more NAbs than sham-injected parents. There is some evidence that parasite exposure may enhance the level of circulating NAbs in wild bird populations (e.g. De Coster et al. 2010). It is thus possible that the antigen exposure influenced the level of NAbs in parents and/or the intensity of NAb transfer in the crop milk. Although the nature of antibodies detected in the crop milk remains to be determined, this study suggests that pigeons are able to transmit antibodies to their offspring through the crop milk when confronted with parasites, opening the interesting possibility that postnatal parental effects could influence offspring immune phenotype depending on local parasite conditions (Hanson 1998).
Effect of a Postnatal Immune Challenge in Foster Parents on Juvenile Immunity
In accordance with this, an immune challenge of foster parents enhanced the humoral response against a second injection of KLH in 1-year-old juveniles, but only for juveniles having received their first antigen injection at 3 days of age. This effect could be mediated by a differential investment in parental care of antigen-injected foster parents and different quality of offspring (e.g. Velando, Drummond & Torres 2006). However, this is not likely because the body mass of 1-year-old juveniles and their humoral response against another antigen (NDV) were not affected by postnatal treatment of parents. This result should be taken with caution given the low sample sizes, but it suggests that the postnatal parental challenge had a specific effect on the juvenile secondary humoral response against KLH, maybe through educational effects of postnatal antibodies contained in the crop milk (e.g. Hanson 1998). A possible scenario is that antigen-injected foster parents would provide crop milk with higher levels of antibodies (NAbs or specific anti-KLH IgAs) or other immune active factors (e.g. antioxidants, Eraud et al. 2008) to their nestlings compared with sham-injected foster parents. These immune factors could reach the blood stream of the young and affect the development of their immune response after an additional challenge later in life. This scenario is speculative, but it is consistent with a recent study showing an up-regulation of antioxidant and immune genes in the crop of lactating pigeons (Gillespie et al. 2011). It opens the interesting possibility that crop milk could play a comparable role to breast milk of mice and humans, which has a long-term instructive role on the development of the humoral immune system of young (e.g. Hanson 1998; reviewed in Hasselquist & Nilsson 2009).
However, this effect was significant only for offspring having received their first injection at 3 days and not for offspring first injected at 14 days. As parents stop providing crop milk to their nestlings after 10–12 days (Johnston & Janiga 1995), this suggests that a concomitant exposure to antigens and to parental immune factors would be necessary for an educational effect to arise, for instance through the formation of maternal antibody–antigen complexes (Hasselquist & Nilsson 2009). Further studies are now necessary to test this hypothesis, for instance by directly feeding young birds with antibodies and exposing them to antigens at different ages.
Whatever the underlying mechanisms, this study shows that prenatal and postnatal parental immune challenges have distinct effects on the humoral response of offspring against an antigen injected in parents. Another important result is that the timing of nestling exposure to antigens has important consequences on offspring humoral response, even 1 year after the antigen exposure, in interaction with parental effects.
Juvenile Growth and Survival
However, the timing of injection did not affect growth parameters, suggesting that parental effects are more important than parasite exposure in shaping the growth phenotype of offspring. Indeed, prenatal and postnatal parental immune challenges had a combined effect on offspring body mass growth. Chicks with both prenatal and postnatal parental immune challenges had a lower body mass growth rate than other chicks, but tended to have a higher final body mass at fledging. Body mass at fledging is often correlated to survival and recruitment probability in wild bird populations (e.g. Krementz et al. 1989; Monros, Belda & Barba 2002), and heavier adult feral pigeons have a higher reproductive success (Johnston & Janiga 1995). This result therefore suggests that the combination of prenatal and postnatal parental effects may have positive effects on offspring fitness. In this study, the survival chance of 1-year-old juvenile feral pigeons was, however, not affected by parental treatment, but the differences in survival may have been masked by the favourable environmental conditions encountered in captivity (lack of predators and high food availability). Future research in free-living populations is now needed to test the effects of parental immune challenges on survival in natural conditions and ultimately on fitness.
In conclusion, this study shows that prenatal and postnatal parental effects may modulate in concert the offspring phenotype regarding humoral immunity and growth and calls for a closer examination of the respective role of prenatal and postnatal parental effects in natural populations, which may have been overlooked in previous studies. Although more work is necessary to assess the evolutionary implications of such parental effects, future studies should now consider both prenatal and postnatal parental effects induced by parasite exposure as potential important factors shaping offspring phenotype across different time scales.
All experiments and protocols were conducted under the approbation of the French Veterinary Services of the DSV77 (authorization No 77-05). We thank M. Toresila and the SACPA Company for pigeon captures. We are grateful to G. Leboucher, P. Lenouvel, F. Péron, A.-C. Prévot-Julliard, B. Decencière and to F. Tisseront, A. Aguilar and H. Tuphile for great help at different stages of this study. This work was supported by grants from the Region Ile-de-France (Sustainable Development Network R2DS No 2008-07). L. Jacquin was supported by a grant from the French Ministry of Research.