Sex differences in immunity are often observed, with males generally having a weaker immune system than females. However, recent data in a sex-role-reversed species in which females compete to mate with males suggest that sexually competitive females have a weaker immune response. These findings support the hypothesis that sexual dimorphism in immunity has evolved in response to sex-specific fitness returns of investment in traits such as parental investment and longevity, but the scarcity of data in sex-reversed species prevents us from drawing general conclusions. Using an insect species in which males make a large but variable parental investment in their offspring, we use two indicators of immunocompetence to test the hypothesis that sex-biased immunity is determined by differences in parental investment. We found that when the value of paternal investment was experimentally increased, male immune investment became relatively greater than that of females. Thus, in this system, in which the direction of sexual competition is plastic, the direction of sex-biased immunity is also plastic and appears to track relative parental investment.
Parasites and pathogens are ubiquitous, and infection by these organisms comes at great costs to hosts, thus creating strong pressures for hosts to invest in immunity. Although the negative covariance between fitness and infection is expected to reduce variation in immunity [1,2], levels of immunocompetence vary markedly across individuals within a population . The proposition that fluctuations in both pathogen prevalence and diversity favour temporal and spatial variation in host immune traits [4,5] does not explain the common observation of females exhibiting greater levels of immune function than males [6,7]. If we are to understand variation in immunocompetence, then the diversity in life histories found in invertebrates makes them ideal systems in which to test the role of sex-specific life-history traits on the maintenance of variance in immunity.
The sexes typically differ in how fitness is maximized, and fitness is likely to restrict or trade off with immunity . Given the costs of mate-attracting displays and ornaments, these condition-dependent traits are expected to have greater expression in healthy individuals . Thus, most studies of sex-biased immunity involving sexual selection have focused on females obtaining indirect (genetic) benefits by preferring male traits that either signal high immunocompetence or, under the related immunocompetence handicap hypothesis, signal that the male is of such high quality (condition) that he can tolerate the costs (including immune suppression) of expressing said traits [9–12]. The signalling immunocompetence hypothesis predicts a positive correlation between immunity and the expression of secondary sexual traits. However, some studies have failed to find any such relationship , or, even more perplexing, have found contrasting relationships between sexually selected traits and immune measures . Thus, a recurrent challenge in ecoimmunology is determining how to interpret the often contradictory results attained when studying multiple immune parameters.
One possibility is that sexually dimorphic immunocompetence results from differential investment by the sexes in the pathways that are responsible for the immune response. Because of life-history differences between the sexes, specifically the relatively large investment by females in each gamete, female fitness is expected to be more dependent on longevity than male fitness owing to the time required to obtain the necessary resources for their relatively large gametic investment. If we assume that greater immunocompetence leads to increased longevity, females are predicted to invest more in immunity than males. Thus, differences in life history may explain greater investment in immunity by females . Yet the sexes differ in a number of life-history traits, and it is unclear whether greater immunocompetence in females is a result of the large gametic investment—the defining trait of females—or some other life-history trait .
Contrary to the above-mentioned predictions, some studies have found that females are less immunocompetent than males  and that immunocompetence is plastic; when fitness-limiting resources are manipulated, sex-biased immunity is observed, but the direction of this bias varies . These exceptions can be explained if sex differences in immunity result not from the large investment by females in zygotes, but from all relative contributions by the sexes in their offspring . This hypothesis is testable if changes in relative investment by the sexes can be shown to be tracked by sex differences in immunity.
Relative investment in offspring varies in some katydids (Orthoptera: Tettigoniidae) in which males feed their mates [19,20]. When food in the environment is limited, male investment in each food gift remains unchanged but becomes of greater value to females , because the relative investment by males in individual offspring (eggs) increases as investment by females decreases. As a result, females compete for access to males (role reversal) . Using the katydid Kawanaphila nartee, we used two commonly studied invertebrate immune parameters to test the hypothesis that the direction of sex-biased immunity mirrors the direction of sexual competition and relative parental investment. We repeated field experiments  that decreased food availability in field enclosures (while controlling insect density and sex ratio), and induced reversal in mating roles. We quantified immunocompetence for each sex under the different mating roles. For high-food enclosures, where males were the competitive sex and would invest less in offspring (relative to females), we predicted higher relative female immunity. For low-food enclosures, where the direction of sexual competition (mating roles) and relative investment in offspring was expected to reverse, we predicted higher relative male immunity.
(a) The system
During mating, male K. nartee transfer a spermatophore to their partners, consisting of an ampulla that inseminates, while the female consumes a protein-rich spermatophylax nuptial gift . When food (pollen) is abundant, more males are capable of producing these gifts, resulting in increased male–male competition (more males calling to attract females) , and females are more discriminating of mates [19,22]. By contrast, when food is limited, mating roles are reversed [19,22]. This reversal tracks a reversal in relative parental investment  as predicted by theory , and relatively more male-derived nutrients are invested in zygotes.
We collected final instar K. nartee in King's Park, Perth, Australia, from 29 August to 2 September 2011. Individuals were housed communally in same-sex groups. Fresh apple pieces and pollen from a health food store were provided ad libitum, and colonies were misted with water daily.
(b) Field study testing the hypothesis that immunity tracks reproductive investment
Seven to 14 days following adult eclosion, 160 individuals of each sex were evenly distributed among eight cuboidal fibreglass-screened field enclosures with 0.75 m linear dimensions, each placed over a clump of flowering kangaroo paws. Four high-food enclosures had supplemental food provided from a grass tree flower stalk that we coated in honey and store-bought bee pollen. Four low-food enclosures received a bare grass tree flower stalk. The location of the low and high enclosures was randomized and they were evenly spaced throughout the approximately 100 m2 experimental area. Enclosures were left undisturbed for 8 days. On the 9th and 13th day, we surveyed the enclosures from dusk through the first 2 h of the night (the period during which most mating occurs ). We observed each of the high- and low-food enclosures twice each night for a total of 16 observation periods per treatment. During each observation period, we noted whether there were males calling (determined by observing movement of stridulatory forewings), the number of males calling, whether any mating pairs were observed, the number of mating pairs observed, whether females were observed rejecting males by preventing mounting (female choice), the number of female choice encounters observed, whether males were observed dismounting a female without transferring a spermatophore (male choice) and the number of male choice encounters observed. These behavioural observations confirmed that individuals in the high-food enclosures were exhibiting typical mating roles, whereas the low-food enclosures exhibited mating-role reversal (electronic supplementary material, table S1 ). On the 15th night, all individuals were collected and brought into the laboratory for immune assays.
(c) Laboratory study investigating the direct effect of diet on immunity
Seven to 12 days following adult eclosion, males and females were taken from their communal same-sex housing described above and put into individual containers in an environmental chamber under a 12 L : 12 D cycle at 20°C and one of the following treatments for 14 days: mating opportunity with high food; mating opportunity with low food; no mating opportunity with high food; and no mating opportunity with low food. Because food availability affects mating opportunity in K. nartee [19,23], we manipulated mating opportunity in order to distinguish between the effects of diet and mating opportunity on immunity. All individuals were provided with water ad libitum. Each day within the first 2 h following sunset, mating-opportunity individuals were presented with a virgin of the opposite sex for either 1 h or until the first copulation attempt; no-mating-opportunity individuals were kept in isolation for the duration of the study. The mating treatments were realistic as some females in nature lose in sexual competition and fail to obtain matings . Animals in the mating opportunity treatment were not allowed to mate, because mating has been shown to affect immunity , and thus mating could have created or exaggerated differences between the two mating opportunity treatments. High-food individuals were fed ad libitum, and low-food individuals were fed ad libitum on alternate days. Similar methods were used to induce sex-role reversal by changing the perceived mating opportunity  and resource availability  in K. nartee. On the 15th day, immune assays were performed.
(d) Immune assays
We used two commonly studied immune parameters in invertebrates: phenoloxidase activity (POA) and encapsulation . Phenoloxidase is a catalytic enzyme involved in melanization and antimicrobial activity in invertebrates, and it assists in defending the host against a variety of pathogens . In measuring the activity rate of phenoloxidase, we can obtain an estimate of how efficiently an individual could potentially encapsulate a foreign particle and resist microbial infection. Encapsulation is a phenoloxidase-dependent immune response and is a widely used defence in invertebrates against a range of macroparasites . Before haemolymph sampling and implantation, needles and monofilaments were sterilized in 70% ethanol.
To estimate POA, a 2 µl haemolymph sample was taken from between the second and third abdominal segment of each individual. If less than 2 µl of haemolymph was collected from an individual, then we combined 1 µl from that individual with 1 µl from another individual of the same treatment. As a result, in some instances, the number of haemolymph samples analysed for a particular treatment was less than the number of individuals sampled. Haemolymph samples were diluted 1 : 50 in phosphate-buffered saline and vortexed. Samples were immediately placed in a −80°C freezer where they remained for one week prior to performing the assay.
We placed 45 µl of each sample into two wells of a microplate to obtain two replicates from each individual. Because phenoloxidase is cytotoxic, it exists in the haemolymph in its inactive form prophenoloxidase. To activate prophenoloxidase, we added 45 µl of the proteolytic enzyme chymotrypsin to each well and promptly placed the plate into a plate reader (M5 Spectramax microplate reader, Molecular Devices, Sunnyvale, CA) to incubate for 20 min at 25°C. We then added 90 µl of dopamine to each well and the change in absorbance was measured every minute for 30 min at 490 nm. Control wells containing chymotrypsin and dopamine, but no haemolymph, did not show a change in absorbance. Enzyme activity was measured as the slope of the reaction curve of absorbance on time.
To estimate encapsulation, we inserted a 2 mm monofilament (0.7 mm diameter) into the wound created during haemolymph extraction. We left monofilaments in for 24 h, because during our preliminary trials we found that in 24 h all individuals began to encapsulate the monofilament, but there remained variation in the degree of encapsulation . We tied a knot at the end of each monofilament to ensure non-destructive removal. After 24 h, monofilaments were recovered and placed into 70% ethanol. A total of 11 individuals that did not survive the 24 h encapsulation period were removed from the analysis (n = 5 females, 6 males). All monofilaments were stored in ethanol until 1 h before photographing, at which time they were placed in distilled water to rehydrate. We took two photographs of each monofilament using a Leica DFC290 HD camera (Leica Microsystems, Heerbrugg, Switzerland) mounted on a dissecting scope and analysed using ImageJ v. 1.45 software. We calculated the mean greyscale value of the pixels in each image to determine the degree of encapsulation. We averaged the greyscale values from the pictures of each monofilament.
(g) Statistical analyses
For both the field and laboratory study, phenoloxidase and encapsulation data were log-transformed in order to meet assumptions of normality. For the field data, we performed an ANOVA with logEncapsulation and logPOA as dependent variables, and sex, food and their interaction as factors. For the laboratory data, we included mating opportunity as an additional factor. To test for differences in immune investment within the sexes, we used independent t-tests. We corrected for multiple testing by adjusting the significance level using the false discovery rate method ; in these instances we report corrected p-values. A correlation analysis of the phenoloxidase replicates for each individual found that phenoloxidase measures were highly repeatable (p < 0.001). All statistical analyses were performed using SPSS v. 17.
(a) Field study
As predicted, individuals—male or female—that assumed the sexually competitive role exhibited relatively lower levels of immunity. We found a significant effect of food (mating role) on both POA (F1,78 = 21.68, p < 0.001, n = 79) and encapsulation (F1,79 = 29.18, p < 0.001, n = 80). Overall, there was no significant difference in immunity between the sexes (POA: F1,78 = 2.26, p = 0.137; encapsulation: F1,79 = 0.288, p = 0.593), but a significant interaction between food and sex (POA: F1,78 = 22.39, p < 0.001; figure 1a; encapsulation: F1,79 = 23.76, p < 0.001; figure 1b), such that high-food females had both greater POA and encapsulation than low-food females, whereas the two did not differ for males (female: POA: t35 = −6.51, p < 0.001, n = 37; encapsulation: t36 = −7.079, p < 0.001, n = 38; males: POA: t40 = 0.055, p = 0.957, n = 42; encapsulation: t40 = 0.383, p = 0.704, n = 42).
(b) Laboratory study
There was no significant effect of diet on POA (F1,185 = 1.49, p = 0.222, n = 192), encapsulation (F1,124 = 0.001, p = 0.983, n = 132), nor their interactions with sex or mating opportunity (table 1). Because neither diet nor its interactions had an effect on immunity, we removed diet from the model and examined the effect of mating opportunity and sex on immunity. We found a significant interaction between the effects of mating opportunity and sex on phenoloxidase (F1,185 = 36.18, p < 0.001): males with mating opportunities expressed significantly higher levels of POA compared with males without mating opportunities, whereas females exhibited a non-significant trend to the converse (figure 2). Additionally, we found an effect of mating opportunity on encapsulation in both males and females (females: F1,64 = 8.18, p = 0.006; males: F1,60 = 9.44, p = 0.003; figure 2); for both sexes, individuals with no mating opportunities exhibited a greater encapsulation response.
Our field experiment demonstrated that immune investment is relatively lower in the sex—male or female—that is engaged in high levels of sexual competition. When food is abundant female K. nartee are both choosy and more immunocompetent than males (figure 1). By contrast, when food is scarce, nuptial gifts provided by males increase in value, and hungry females increase their mating rates and compete for mates. Our results show that under such role reversals, relative (although not absolute) immune investment increases in males (figure 1). Our laboratory experiment confirmed that this was not a direct result of food stress via changes in quantity of food [32,33]. It is unlikely that our low-food field treatment failed to create a condition of food scarcity as we observed reversals in the mating roles in low-food cages comparable with those reported in similar experimental manipulations at this site .
The direction of sexual competition in K. nartee appears to be driven by changes in relative parental investment, mediated by food availability. The relative investment by males in offspring increases as resources derived from the female's general diet decrease, and more male-derived nutrients are allocated to eggs . This change in investment is due exclusively to a change in reproductive effort of females, because diet does not affect energy expended by males in offspring investment, as was found in a previous study of the same metapopulation of K. nartee . This is consistent with our result that female (but not male) immunity changes under food deprivation, causing a reversal in the direction of sexual competition. Reproduction and immunity are predicted to trade off with one another [15,34,35]. Thus, a diet-induced reduction in female mating opportunities (as the number of males supplying gifts decreased), and subsequent increases in, for example, female courtship and competitive behaviour—rather than direct dietary constraints on allocation to immunity —may have caused the decline in immunity in low-food females from the field.
The costs of courtship and mating on immunity in invertebrates have been well demonstrated [24,34,35]. The katydids in our laboratory study remained virgins, so it is unlikely that they incurred any costs of copulation per se. However, there could have been courtship costs. Such costs have been found for both sexes in a number of species. For example, male drumming wolf spiders, Hygrolycosa rubrofasciata, with increased rates of drumming (signal used to attract females), had lower encapsulation and antimicrobial responses than controls . Thus, one possible cost to immunity incurred by the males in our laboratory study was the large portion of total daily energy used to signal (call) to potential mates . Males in the mating opportunity treatment would be more immunocompetent because they would have spent less time calling, and therefore had more resources available for the immune response; this hypothesis is supported by the POA, but not by the encapsulation assay (figure 2).
For females, the effects of multiple mating on immunity have been investigated , but no study has distinguished between the costs to female immunity resulting from copulation versus other mating interactions, such as male harassment. Although the females in our laboratory study were virgins and therefore were likely to be receptive to mating, male harassment is undoubtedly costly to females, as has been demonstrated in other insects [36,37]. Females in the low-mating-opportunity treatment had both greater POA and a stronger encapsulation response, as would be predicted if male harassment in the mating opportunity treatment caused a depressed female immune response.
As is common in studies of immunity , results of our two immune parameters did not correspond in males from our laboratory study. Because phenoloxidase is a circulating enzyme involved in the activation of multiple immune components (including encapsulation), POA provides an estimate of an individual's ‘potential’ immunocompetence. By contrast, our encapsulation assay introduced a foreign object (simulating an invading macroparasite), therefore allowing us to assess the ‘realized’ immune response . Thus, one possible interpretation for the discrepancy between results from our phenoloxidase and encapsulation assays is that the two represent the difference between constitutive levels of an immune enzyme (phenoloxidase), and the actual efficacy of the immune response (encapsulation). If life history predicts that females invest more in immunity than males , perhaps it follows that males invest disparately in ‘potential’ and ‘realized’ immunity. We note, however, that these are just two estimates of the total immune response, which includes multiple systemic and behavioural components.
Our results support the hypothesis that immune investment tracks relative investment in offspring rather than other traits that may covary with the direction of sexual competition. These findings are consistent with previous suggestions of immune investment being a plastic trait, and suggest that sexually dimorphic immunity results from differential allocation of resources rather than fixed physiological or life-history differences between the sexes. Our findings point to differences in relative parental investment being a general explanation for sex-biased immunity.
The field and laboratory data can be found in the electronic supplementary material (table S1, field behavioural observations) and the Dryad digital data depository (doi:10.5061/dryad.bj287).
This work was supported by NSERC Discovery Grants to D.T.G. and R. L. Baker and NSERC IPS, Toronto Zoo IPS and L'Oreal–UNESCO scholarships to C.M.V. Travel was supported through an Entomological Society of Canada Graduate Research-Travel Scholarship to C.M.V.
We thank M. Beveridge, I. Dadour, S. Iaschi, K. McNamara and L. Simmons for laboratory space and procedural guidance. I. Dadour, N. Kreitals and S. Voss provided field assistance and equipment. We are indebted to I. Dadour for providing laboratory space and for his hospitality during our visit. This work was approved by the King's Park Board and we thank S. Easton for providing on-site assistance. We thank L. Holman, L. Rowe, L. Simmons and two anonymous reviewers for valuable comments on the manuscript.
- Received February 10, 2014.
- Accepted June 25, 2014.
- © 2014 The Author(s) Published by the Royal Society. All rights reserved.