†Author to whom correspondence should be addressed. Present address: Department of Biology, University of Turku, FIN-20024, Turku, Finland. E-mail: firstname.lastname@example.org
1Immune defence has recently been viewed as a life-history trait that shows trade-offs with other life-history traits. However, studies exploring correlations between different components of immune defence and other life-history traits are scarce.
2In this study, two measures of immune function, body size, and development time were studied in the Mediterranean Field Cricket, Gryllus bimaculatus.
3We found no differences between the sexes, but differences in the correlation between the measure of immune function, development time, and body size. In both sexes, encapsulation rate was negatively correlated with body size and development time, whereas lytic enzyme activity was positively correlated with these two traits. Furthermore, encapsulation rate and lytic activity were themselves negatively correlated.
4These results indicate that there is no single trait that can be termed ‘immunocompetence’ and that an understanding of trade-offs between the different types of immune function and life-history traits will require a more detailed analysis of the physiological basis of the traits.
An area of very active research in evolutionary biology is that of immune ecology, especially the importance of trade-offs between immune function and other life-history traits (for reviews see e.g. Zuk & Stoehr 2002; Schmid-Hempel 2003). Life-history theory assumes that immune defence is a trait whose costs are traded off against some other fitness component. Thus, trade-offs are to be expected between immune function and life-history traits such as growth and reproduction (Sheldon & Verhulst 1996). Allocation decisions between reproductive effort, development time and immune defence are hypothesized to be targets of optimizing selection, favouring individuals that allocate their resources in a manner that maximizes their lifetime reproductive success.
There is some evidence that trade-offs exist between immunity and other traits in insects. For example, in the bumble-bee Bombus terrestris, increased foraging effort reduces encapsulation response (König & Schmid-Hempel 1995) and experimental activation of the immune system reduces life span during starvation (Moret & Schmid-Hempel 2000). In Drosophila melanogaster, immune response to parasitoids reduces resistance to starvation (Hoang 2002) and, in males, increased sexual activity reduces the rate at which bacteria are cleared by the immune system (McKean & Nunney 2001). Kraaijeveld & Godfray (1997) found that Drosophila melanogaster that were resistant to the parasitoid Asobara tabida had lower larval competitive ability, and similar results have been obtained using the parasitoid Leptopilina boulardi (Fellowes et al. 1998). This lower competitive ability is associated with reduced rate of larval feeding (Fellowes et al. 1999a). Other studies in Drosophila (Fellowes et al. 1998a,b, 1999b; Fellowes & Godfray 2000) have demonstrated direct costs of immunity. These costs include reduced female size and fecundity, decreased tolerance to desiccation, and susceptibility to other parasitoids.
Among evolutionary ecologists, immunocompetence has been traditionally defined as a quantitative measure indicating host capacity to resist infection in the absence of parasitism (e.g. Siva-Jothy 1995; Sheldon & Verhulst 1996; Schmid-Hempel & Ebert 2003). A general assumption in the literature is that immunocompetence can be regarded as a single trait, or as a complex in which the components are positively correlated so that an individual that has a high immune function when measured using one assay method will display a high immune function measured on another assay (e.g. Folstad & Karter 1992). However, it is also possible that the cost of mounting an immune response to one type of challenge may impair the ability to mount a defence against another type of challenge. For example, an individual that has a high encapsulation rate will have a high lysozyme activity if the hypothesis of a general immunocompetence ability is correct, but a relatively low lysozyme activity if there is a trade-off between these two types of immunocompetence.
Males of the Mediterranean Field Cricket, Gryllus bimaculatus De Geer, produce calls that act as signals for mate attraction and repulsion of rivals (Simmons 1988). Female G. bimaculatus prefer the courtship song of males with higher encapsulation rate (Rantala & Kortet 2003) and males showing higher encapsulation rate are more successful in male–male competition (Rantala & Kortet 2004). Moreover, males with higher encapsulation rate are morphologically more symmetrical (Rantala, Ahtiainen & Suhonen 2004). The present study addresses three questions raised above: first, is immune function negatively correlated with the life-history trait, development time? Because there is frequently a positive correlation between development time and body size (Roff 2000), we also analyse the relationship between immune function and body mass. The second question we address is whether there is a positive or negative correlation between two measures of immunocompetence, the encapsulation rate against a novel antigen and the haemolymph concentration of an antibacterial enzyme, lysozyme. Finally, we ask if there is a difference between the immunocompetence of male and female G. bimaculatus.
Materials and methods
The crickets used in this experiment were the first laboratory generation of wild animals collected from Costa del Sol in Southern Spain in October 2001. They were maintained at 29 ± 1 °C with ad libitum food and water under a 12:12 h light/dark photoperiod. Experimental crickets were derived from a bulk laboratory stock as freshly hatched larvae and maintained individually in covered plastic containers (1 l). Standard laboratory rodent pellets and fresh cabbage were provided ad libitum, and water was available via a cotton plugged tube. Containers were checked daily to determine the stage of development of the larvae. The larval development time was defined as the number of days between hatching and eclosion. To ensure virginity, the sexes were physically (but not acoustically) isolated from other individuals. Immune assays were performed on eighth day after eclosion. Before the immune assays, we weighed the fresh body mass of crickets to the nearest 0·01 g. Body weight shows strong correlations with linear measures of size (see Simmons 1986) and was therefore used as the index of adult size in this study.
encapsulation rate assay
To measure encapsulation rate, crickets were chilled on ice for 20 min after which a 2-mm long piece of nylon monofilament (diameter 0·18 mm, rubbed with sandpaper) was inserted through a puncture in the pleural membrane between the second and third sternite. Crickets were then placed into individual vials and kept at a constant temperature (28 ± 1 °C) for 5 h to allow an immune response. The implant was then removed and the removed monofilament was photographed from three different angles under a light microscope with a digital videorecorder. These pictures were analysed using the Image Pro program. The degree of encapsulation was analysed as grey values of reflecting light from implants. As a measure of encapsulation rate, we used the average grey values of three video pictures. The data were transformed so that the darkest grey values correspond to the highest encapsulation rate. This transformation was done by subtracting the observed grey values from the control grey value (clear implant) (see Rantala & Kortet 2003).
After the encapsulation rate measurements, we collected 10 µl haemolymph from each cricket from the puncture at their abdomen. Lysozyme activity against Micrococcus lysoideikticus was assayed turbidometrically using methods similar to Rantala & Kortet (2003). Some 200 µl of 0·35 mg ml−1 freeze-dried M. lysodeikticus buffered (pH 6·4) solution was mixed with 50 µl 1:4 buffered haemolymph (pH 6·4) was placed in a plastic multicuvette (Labsystems cliniplate, Labsystems, Finland). The optical density of the mixture was measured at 492 nm at 20 °C in one minute intervals for 30 min with a plate reader (Multiskan Plus, Labsystems). All samples were analysed in random order. Lytic activity was expressed as a total change in optical density. To control for the typical decline in turbidity due to settling we used six blank controls in each plate. The mean change in optical density of blank samples was subtracted from each experimental reading.
To examine the overall pattern of covariation, we calculated all pairwise correlations between the traits measured (Table 1). Only body mass significantly covaried with sex, while correlations between all other traits were highly significant (Table 1, Fig. 1) Elimination of the three obvious, and statistically significant, outliers does not alter the results. Body mass correlated positively with development time and lytic activity but negatively with encapsulation rate. Similarly, development time correlated positively with lytic activity but negatively with encapsulation rate. Finally, lytic activity covaried negatively with encapsulation rate. This last result suggests that immunocompetence cannot be equated with a single measure of immune function but rather that there is a trade-off between types of immune function. The changing sign of the correlation between development time and the two measures of immunocompetence also suggest that trade-offs with life-history traits depend upon the type of immune function assayed. Finally, the lack of covariation with sex suggests that immunocompetence is not a function of sex in G. bimaculatus.
Table 1. Pairwise correlations between all variables measured in G. bimaculatus. Correlations are shown above the diagonal, probabilities (with Bonferroni correction for multiple comparisons) and sample sizes (in parentheses) below diagonal
Sex was coded as 0,1. Though this variable is not normally distributed, its correlation with other traits can be viewed from a regression perspective in which Sex is the independent variable. In this case the bivariate normality assumption is not necessary and the correlation coefficient is equivalent to the square root of the coefficient of determination and the same statistical tests apply as for the correlation coefficient (thus, strictly we refer to the covariation of sex with other traits rather than sex being correlated with other traits).
The relationships between traits may be confounded by correlations with other traits: therefore, to take into account the possible confounding effects of sex and body mass we used stepwise multiple regression with encapsulation rate and lytic activity as the response variables and development time, sex, body mass and all interactions as predictor variables (this multiple regression approach is a general linear model analysis using both continuous and categorical [Sex] variables). Both forward and backwards regression gave the same final model, which, in the case of encapsulation rate, included only body mass and the interaction between body mass (Bmass) and development time (Dtime) (R = 0·67, n = 208, Encapsulation rate = 54·00 +81·96(Dtime) − 1·34(Dtime)(Bmass), P < 0·0001 for included terms). The final model for lytic activity was simpler with only body mass and development time being retained (R = 0·40, n = 207, lytic activity = 0·0129 + 0·0013(Dtime) + 0·0548(Bmass), P < 0·0001 for included terms). These results confirm the pairwise results indicating that sex itself does not play a role in determining immunocompetence in G. bimaculatus. The inclusion of both development time and body mass in the final models indicates that both of these traits covary with immunocompetence. To better visualize the combined influence of these two traits we used the final multiple regression models to construct a contour surface of immunocompetence as a function of body mass and development time (Fig. 2). Onto this surface we have plotted the observed combinations of body size and development time. Note that the observed correlation between the two traits (body mass, development time) and immunocompetence is dependent upon the positive covariation between the two traits (e.g. a high encapsulation rate could occur for a small development time only if there were also a large body mass, which is unlikely).
In summary, with respect to the three questions posed in the introduction: (1) a trade-off between immune function and the life-history trait, development time depends upon the measure of immune function used: it occurs in encapsulation rate but not lytic activity, in which there is a positive correlation, (2) there is a negative correlation between the two measures of immune function and (3) there is no effect of sex per se on immune function.
Trade-offs between traits are a fundamental component of life-history theory (Roff 2002). A trade-off between traits associated with growth and development and immune function may be expected because substantial nutritional and energetic demands are associated with immune activation and the maintenance of an efficient immune system (Lochmiller & Deerenberg 2000). There is some evidence for a genetic trade-off between such life-history traits and immune function in insects. For example, Koella & Boete (2002) found that selection for late pupation increased encapsulation response in mosquitoes. By contrast, Yan, Severson & Christensen (1997) found that resistant mosquitoes matured earlier and were smaller than susceptible mosquitoes. In this study we found that crickets that had the shortest development time also had higher encapsulation rates as adults. However, individuals that had long development times and were relatively large had higher lytic activity. These results support the hypothesis of a phenotypic trade-off between encapsulation rate and body size. On the other hand, the positive correlation between development time and lytic activity demonstrates that the situation is more complex and that it will undoubtedly be necessary to take into account the physiological underpinnings of the relationship between the type of immune function and the type of life-history trait under study.
In our previous studies of G. bimaculatus, using males matched for body size, we found that females preferred the courtship song of males with high encapsulation rate (Rantala & Kortet 2003), that males with higher encapsulation rate were more successful in male–male competition (Rantala & Kortet 2004) and that the most symmetrical males had the strongest encapsulation rate (Rantala et al. 2004). Studies of larval competition in G. bimaculatus have shown that competition is asymmetric, with some apparently ‘good’ competitors achieving large adult size while other ‘poor’ competitors remain small (Simmons 1987). However, in the present study all males were raised individually with ad libitum food in a standard environment, and thus differences in size and development time cannot be the result of asymmetric competition nor differences in breeding conditions.
Large body size has often been found to be advantageous in sexual selection (reviewed in Andersson 1994), and specifically in G. bimaculatus male body size has been shown to have important implications for its competitive ability and mating success (Simmons 1986; Bateman, Gilson & Ferguson 2001). However, large size cannot be used as a surrogate measure for general immunocompetence because we found that larger males had higher antibacterial activity (lytic activity) but lower encapsulation rate than small males. In insects body fat synthesizes antimicrobial peptides (e.g. Hetru, Hoffman & Bulet 1998) and may limit the rate of synthesis. It is possible that larger males have more fat in their body, thereby enabling them to maintain higher antibacterial activity (lytic activity). However, the reason why larger, longer developing males had lower encapsulation rate is less clear. What is clear is that immunocompetence is not in itself a single trait and immune defence is complex. This is consistent with studies in vertebrates, which have found that immune defence is complex, and different immune assays do not always correlate with one another (e.g. Luster et al. 1993; Keil, Luebke & Pruett 2001). Most importantly, our study suggests that in G. bimaculatus, choosing males on the basis of male size and/or courtship song will not increase the chance of mating with more ‘immunocompetent’ males. Instead, by choosing males on the basis of body size, females will prefer males that have longer development time and have higher antibacterial activity (lytic activity), whereas by choosing males on the basis of courtship song they will favour males that have higher encapsulation rate but lower lytic activity.
In this study, encapsulation rate and antibacterial activity (lysozyme-like) were negatively correlated, indicating that there may be a trade-off between these two immune responses, like we found in previous studies with this species (Rantala & Kortet 2003). This is consistent with the findings of Moret & Schmid-Hempel (2001), who found that effectiveness of melanization response and the antibacterial response were negatively associated in bumble-bees. Furthermore, Cotter, Kruuk & Wilson (2004) found a negative genetic correlation between phenoloxidase (PO) activity and antibacterial activity in the Egyptian Cotton Leafworm, Spodoptera littoralis, suggesting a genetic trade-off between encapsulation rate and antibacterial activity. The negative association between different components of the immune system makes it difficult to use the term immunocompetence in any general sense: rather the term must always be used in reference to a particular immune response as shown here for G. bimaculatus.
In many vertebrate species, males suffer more from parasitic infections and also tend to have reduced immune responses compared with females (Poulin 1996; Zuk & McKean 1996; Møller et al. 1998, 1999). However, Sheridan et al. (2000) found no sex differences in parasite infections among arthropod hosts. Studies in Gryllus texensis (Adamo et al. 2001) and Acheta domesticus (Gray 1998) have found that male crickets are more susceptible than females to infection with the bacteria Serratia liquefaciens. Consistent with these studies, many other studies on sex differences in arthropod immunity also show lower male immune defence (e.g. Rheins & Karp 1985; Nigam et al. 1997; Radhika et al. 1998; Wedekind & Jacobsen 1998; Kurtz et al. 2000; Kurtz & Sauer 2001; Rolff 2001; Siva-Jothy et al. 2001). However, in the grasshopper Melanoplus sanguinipes, there are no sex differences in resistance to the fungus Beauvieria bassiana (Gillespie & Khachatourians 1992) and there was no sex difference in PO in Acheta domestica, although females had significantly higher haemocyte counts than males (da Silva et al. 2000). There was no sex difference in melanotic encapsulation of mite feeding tubes in four species of damselflies (Yourth et al. 2001). By contrast, in Mealworm Beetle, Tenebrio molitor, males had higher PO activity than females (Siva-Jothy & Thompson 2002). In this study we did not find sex differences in encapsulation rate and in lytic activity. Thus our study does not support the hypothesis that females are more ‘immunocompetent’ than males. On the other hand, since we did not use real disease/parasitoid in our experiment, we do not know whether male and female crickets differ in their actual disease/parasite resistance (see Adamo 2004), and the sexes might differ in immune functions that we did not measure. Furthermore, our study was done in laboratory conditions which might have effects to sex differences in immune function. For example, in the damselfly Lestes forcipatus Rambur, under field conditions, sex and host size had no effect on encapsulation rate of water mite feeding tubes, but under controlled lab conditions both sex and (in males) condition were significant predictors of the encapsulation rates when challenged with Sephadex beads (Yourth, Forbes & Smith 2002a,b).
In summary, this study has shown that immunocompetence is not a single trait and that the costs of mounting a defence against one sort of challenge may involve both trade-offs with life-history traits, such as development time, and with other immune functions. Furthermore, we found no differences between the sexes in either measure of immune function. The present study has demonstrated correlations between various components of immune function, body size and development time. The question that remains to be addressed is the extent to which these correlations represent paths of causation: this will require further research into the physiological basis of immune function and its connection with development.
We thank K. J. Emerson, S. Gershman, K. Fedorka, D. Lewkiewicz, A. Stoehr, R. Tinghitella, J. Trexler and M. Zuk, who gave fruitful comments on the manuscript. Special thanks to J. Ahtiainen, J. Lampela, R. Kortet, S. Koistinen, J. Tuusa, J. Valkonen, L. Vainio and S. Väänänen for their assistance in the laboratory. This study was supported by the Academy of Finland to MJR (Project: 202624).