Effects of testosterone and corticosterone on immunocompetence
in the zebra finch
Mark L. Robertsa,⁎, Katherine L. Buchananb, Dennis Hasselquistc, Matthew R. Evansa
aCentre for Ecology and Conservation, School of Biosciences, University of Exeter, Cornwall Campus, Penryn, Cornwall TR10 9EZ, UK
bCardiff School of Biosciences, Main Building, Park Place, Cardiff University, Cardiff CF10 3TL, UK
cAnimal Ecology, Lund University, Ecology Building, Sölvegatan 37, 22362 Lund, Sweden
Received 13 February 2006; revised 4 September 2006; accepted 5 September 2006
Available online 17 October 2006
The original immunocompetence handicap hypothesis (ICHH) suggested that testosterone has a handicapping effect in males by both
promoting the development of sexual signals and suppressing immune function. A modified version, the stress-linked ICHH, has recently
proposed that testosterone is immunosuppressive indirectly by increasing production of corticosterone. To test both the original and stress-
mediated versions of the ICHH, we implanted male zebra finches taken from lines selected for divergent maximum stress-induced levels of
corticosterone (high, low and control) with either empty or testosterone-filled implants. Their humoral and cell-mediated immune responses were
then assessed by challenge with diphtheria:tetanus vaccine and phytohemagglutinin respectively. We found no effect of the hormone
manipulations on either PHA or tetanus antibody responses, but found a significant interaction between titers of both testosterone and
corticosterone on diphtheria secondary antibody response; antibody response was greatest in individuals with high levels of both hormones. There
was also a significant interactive effect between testosterone treatment group and corticosterone titer on body mass; the body mass of males in the
elevated testosterone treatment group decreased with increasing corticosterone titer. These results suggest that, contrary to the assumption of the
stress-mediated version of the ICHH, high plasma levels of corticosterone are not immunosuppressive, but are in fact immuno-enhancing in the
presence of high levels of plasma testosterone. Equally, the central assumption of the ICHH that testosterone is obligately immunosuppressive is
also not supported. The same individuals with the highest levels of both hormones and consequently the most robust antibody response also
possessed the lowest body mass.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Corticosterone; Testosterone; Glucocorticoid; Zebra finch; Immunocompetence; Stress; Immunocompetence handicap hypothesis; PHA; Diphtheria:
The immunocompetence handicap hypothesis (ICHH)
suggests that testosterone (T) serves a dual role in signal
expression and immune function (Folstad and Karter, 1992).
High levels of Tresult not only in full signal expression but also
a concomitant reduction in immunocompetence. Therefore,
only high quality males can afford to fully express sexual traits
because only they will be able to resist or tolerate parasite/
pathogen attack. One of the key assumptions of the ICHH is the
immunosuppressive nature of T or any biochemical substance
that is related to sexual signal expression (Folstad and Karter,
1992). However, several correlational and manipulative studies
carried out since the ICHH was first proposed have only
produced equivocal evidence to support the ICHH in its
simplest form (with T being the immunosuppressive agent; see
Roberts et al., 2004). It is generally agreed that T does have
immunosuppressive characteristics in mammals (Grossman,
1985), but studies in birds have yielded contradictory results.
Several studies in which T has been experimentally manipulated
have found the hormone to be immunosuppressive in birds
(Buchanan et al., 2003; Casto et al., 2001; Duffy et al., 2000;
Hormones and Behavior 51 (2007) 126–134
⁎Corresponding author. Max Planck Institute for Ornithology, Vogelwarte
Radolfzell, Schlossallee 2, 78315 Radolfzell, Germany.
E-mail address: email@example.com (M.L. Roberts).
0018-506X/$ - see front matter © 2006 Elsevier Inc. All rights reserved.
Owen-Ashley et al., 2004; Peters, 2000), whereas other studies
have not found any such effect (Hasselquist et al., 1999).
Indeed, some studies have found a positive effect of T on
immunity (Evans et al., 2000). A recent meta-analysis that
examined the effects of testosterone on immunity in vertebrates
confirmed this ambiguity; only when considering ectoparasites
on manipulated lizards did T have a demonstrably negative
effect on a measure of immunity (Roberts et al., 2004).
In some recent studies, T has been found to correlate (both
in birds, corticosterone (CORT) (e.g. Evans et al., 2000;
Klukowski et al., 1997; Owen-Ashley et al., 2004; Parker et
al., 2002; Schoech et al., 1999). The stimulation of the
hypothalamus–pituitary–adrenal axis in response to stress
leads to the secretion of CORT, which is involved in the
mobilization of energy stores (glucose), the shutdown of
digestive processes and increasing the peripheral blood supply
(Buchanan, 2000; Sapolsky et al., 2000; Silverin, 1998). CORT
dispersalbehavior and increasesforaging activity (Breuneretal.,
1998; Silverin, 1998; Wingfield et al., 1997). T increases
circulating CORT in several avian species (Casto et al., 2001;
Duffy et al., 2000; Evans et al., 2000; Owen-Ashley et al., 2004;
Poiani et al., 2000), and there is good evidence to suggest that
CORT is immunosuppressive (e.g. Buchanan, 2000; Harvey et
al., 1984; Råberg et al., 1998; Sapolsky et al., 2000; Wingfield et
al., 1997; but see Svensson et al., 2002). In a study on the house
sparrow (Passer domesticus), Evans et al. (2000) found that
experimentally increased T impaired antibody production.
covaried with T), T was found to enhance immunocompetence.
The authors of this study suggested a modification to the original
ICHH; rather than being immunosuppressive directly, T-related
immunosuppression occurs indirectly through an increase in
CORT. T itself may be immuno-enhancing, again indirectly
through its effect on behavior leading to more dominant
individuals (with higher T) gaining greater access to dietary
resources and therefore being in better condition (Evans et al.,
2000). Therefore, the fact that T and CORT may interact with
each other requires that the levels of both hormones be
manipulated independently to distinguish between the immuno-
suppressive effect (if any) of the two hormones.
Several studies have reported a negative effect of elevated
testosterone on avian body condition (defined as body mass in
relation to body size) (Clotfelter et al., 2004; Mougeot et al.,
2004; Ros, 1999; Wikelski et al., 1999), and other studies have
found increased CORT levels to be negatively related to body
condition (Breuner and Hahn, 2003; Hood et al., 1998;
Kitaysky et al., 2001; Pereyra and Wingfield, 2003; Perfito et
al., 2002; Schwabl, 1995; Sockman and Schwabl, 2001). In
some cases, however,no effectof either elevated CORTor Thas
been found on body condition (Lormee et al., 2003; Lynn et al.,
2003; and Alonso-Alvarez et al., 2002 Buttemer and Astheimer,
2000; respectively), and in a few studies, testosterone appeared
to have a positive effect (Briganti et al., 1999; Chastel et al.,
2005). It is therefore unclear whether either hormone has an
effect (positive or negative) on body mass. An additional aim of
this experiment was to elucidate the effect of elevated
testosterone and corticosterone on avian body condition defined
as body mass corrected for skeletal size.
The study reported here was conducted to ascertain the effect
of manipulating levels of T and CORT on individual
immunocompetence and general body condition. If the original
version of the ICHH (that assumes that T is directly
immunosuppressive) is to be supported, a high T treatment
group should have the poorest immune response, and the males
with the lowest T should have the highest immune response,
regardless of CORT levels. However, if the modified version of
the ICHH that identifies glucocorticoids as the immunosup-
pressive agent is to be supported, then males with a low peak
(stress) CORT response should have a more robust immune
response than males with a high peak CORT response,
regardless of testosterone titers. Although the stress-linked
version of the ICHH is based upon measurements of baseline
levels of plasma corticosterone, there is good evidence to
suggest that baseline and stress-induced levels of CORT
positively covary (e.g. Cockrem and Silverin, 2002; Romero
and Wingfield, 1998; Schoech et al., 1999). Moreover, there is
considerable evidence for immunosuppressive effects mediated
through Type II receptors, which are only activated at elevated
(as opposed to basal) levels of CORT (Sapolsky et al., 2000).
Populations of zebra finch (Taeniopygia guttata) were
selected for low, high and control levels of plasma corticoster-
al., 2006). Males from the G4 generation were housed under
short day photoperiod (to remove any confounding effect of
and individuals from each corticosterone line were allocated
equally into two testosterone treatment groups — high and low
T. These males were then challenged by phytohemagglutinin
injection to test their cell-mediated immunity and diphtheria:
tetanus injection to test their humoral response. In this way, the
levels of both hormones could be artificially fixed so that
Materials and methods
Since 1999, three replicate lines of zebra finch (2×low, control and high
CORT; 6 lines in total) were selected for contrasting levels of maximum stress-
induced CORT in response to a mild stressor (Evans et al., 2006). A significant
difference in plasma corticosterone levels was observed between the lines in the
expected direction from the F2 generation of selection, with selection pressure
exerted on each generation. There was a downward trend in corticosterone over
generation regardless of selection line; little difference has existed in changes in
corticosterone titer between the low lines and controls, but the high lines have
shown a significant realized heritability of 20–25%. There was no correspond-
ing change in testosterone over the generations, and no significant differences
exist between the corticosterone lines in testosterone (Evans et al., 2006). Birds
in each line were housed together in a large aviary giving c1m3per breeding pair
and maintained at an ambient temperature of 20–24 °C (Jones and Slater, 1999).
The humidity was maintained between 50 and 70%, and the rooms sprayed with
water two or three times per day. The birds were provided with ad libitum seeds
(foreign finch mixture, Haiths Ltd., Cleethorpes, Lincolnshire, UK), Chinese
millet sprays, mineralized grit, water and cuttlefish bone. The finches were
127M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
provided with c10g of a 3:1 mixture of ‘nectarblend’ (Haiths Ltd., Cleethorpes,
Lincolnshire, UK) and egg biscuit food (Haiths Ltd., Cleethorpes, Lincolnshire,
UK), and either lettuce or cucumber daily with a cod liver oil supplement in the
seed weekly. An excess of nesting baskets and boxes was provided. Full details
of the selection methodology, characteristics and housing can be found in Evans
et al. (2006). We have taken every care to minimize the welfare implications of
this project, and all work was carried out under license by the UK Home Office
and after local ethical review.
Hormone sampling and assay characteristics
Six weeks post fledging, all birds were blood sampled. Blood samples for
CORT (100 μl) were taken from the brachial vein after 20 min holding in a cloth
bag, after a pilot study revealed that peak CORT response occurs in the zebra
finch after 20 min of a stressful stimulus (Evanset al., 2006). This is the standard
capture-restraint protocol used in studies analyzing the CORT response
(Wingfield, 1994). The blood was centrifuged at 11,000×g for 15 min and the
plasma frozen. Corticosterone concentrations were measured after extraction of
20 μl aliquots of plasma in diethyl ether, by radioimmunoassay (Wingfield et al.,
1992) using anticorticosterone antiserum (code B21-42 and B3-163, Esoterix
Inc. Endocrinology, CA) and [1,2,6,7-3H]-corticosterone label (Amersham,
UK). The interassay coefficient of variation was 15.7%, and the intraassay
coefficient of variation 3.1%. The mean extraction efficiency was 72%. The
assay was run with 50% binding at 134 pg/tube, and the detection limit (for
7.3 μl aliquots of extracted plasma) was 1.76 nmol l−1.
Blood samples for testosterone assay were taken immediately upon capture,
and the plasma obtained and stored in an identical manner to the CORTsamples.
Testosterone concentrations were measured in plasma samples by direct
radioimmunoassay using anti-testosterone antiserum (code 8680-6004, Biogen-
esis, UK) and [125I]-testosterone label (code 07-189126, ICN, UK) (Parkinson
and Follett, 1995). The assay was run with 50% binding at 11.0 pg/tube and a
detection limit of 0.068 nmol l−1for the 20 μl plasma volumes that were run in
the assay. The interassay coefficient of variation was 15.5%, and the intraassay
coefficient of variation was 2.2%.
Implantation of testosterone
Following experimental trials with birds outside the selection lines, a 7 mm
implant size (Dow Corning medical grade tubing RX-50: inner diameter
0.76 mm, outer diameter 1.65 mm) was deemed most appropriate to give a high
plasma T level, within the natural physiological range. Both the males used in
the pilot study and the main experiment were housed under a short day
photoperiod (18 h dark:6 h light) to limit the production of endogenous T.
Although castration may have limited the production of endogenous T more
effectively, such a procedure on a small passerine may well have resulted in
morbidity and possible mortality. In addition, the procedure may affect an
individual's condition and CORT titer, thus introducing greater variation and
possible confounding elements into the experimental design. Furthermore,
reduced photoperiod has been shown to cause a decrease in testicular mass
relative to increased photoperiod in zebra finches (Bentley et al., 2000).
The mainexperiment consistedof 72 adultmalezebra finches taken fromthe
selected lines. Twelve birds were taken from each of the 3 replicate (×2) CORT
lines, and these were randomly subdivided into 6 empty-implanted and 6
testosterone-implanted birds. The implants in the high T groups were packed
with crystalline T (Sigma T-1500), while empty implants were inserted into the
control group birds. The silastic implants were inserted into an incision made in
the skin of the neck. This was then closed by application of a tissue adhesive
(Nexaband, Abbott Labs, Chicago, US) and where necessary by surgical suture.
The implantation process took a total of 2 days to complete, and both CORTand
T groups were balanced across these 2 days.
Cell-mediated immune response
The males' cell-mediated immunity was challenged by phytohemagglutinin
(PHA) injection 8–9 days after implantation (balanced across treatment groups).
Louis, MO) intradermally into the left wing web. Each male received 30 μl of a
et al., 1993). A spessimeter (Alpas.r.l. Milan) was usedto measure the wing web
before injection (as a control measurement), and at 24 h after injection (to the
nearest 0.01 mm), to measure the wing web swelling in response to the mitogen.
males were then weighed to the nearest 1 g using a Pesola spring balance, and
their right tarsus lengths measured to the nearest 0.1 mm with digital calipers.
Three days after the PHA injections, the males were blood sampled. One
hundred microliters of blood was collected in heparinized capillary tubes and
centrifuged at 11,000×g for 15 min. The plasma was removed and stored at
−20 °C and subsequently tested for cross-reaction to diphtheria:tetanus as a
control measurement. Three days after this, the primary humoral immune
response of the males was tested. Each male was immunized with 100 μl of
diphtheria:tetanus vaccine (Aventis Pasteur, Swiftwater, PA) by intraperitoneal
injection. Twelve days later, blood samples were taken from the brachial vein.
Twenty one days after the initial injection, the males were inoculated again with
diphtheria:tetanus vaccine to test the secondary humoral response. Eight days
after this, the birds were again blood sampled, and the samples split between
testing for antibodies and for testosterone assay. In total, three blood samples
were taken from each manipulated male. Antibody levels were determined by
the use of ELISA (see Hasselquist et al., 1999; Owen-Ashley et al., 2004 for a
full description of the methodology employed).
The degree of swelling exhibited by the wing web injected with PHA (the
difference between pre- and 24-hour post-injection) was used as the dependent
variable in Restricted Maximum Likelihood models (ReMLs). This form of
Generalized Linear Model (GLM) has been widelyused before in similar studies
(e.g. Evans et al., 2006; Peters, 2000; Peters et al., 2004). The maximal model
TtiterandpeakCORTtiter asfixedeffects, andCORTlinereplicateas arandom
term. A minimal model was derived by stepwise deletion by removal of non-
significant terms that did not significantly increase the residual deviance of the
model. The residuals of the models conformed to a normal distribution and were
homoscedastic so the models were run with a Gaussian distribution in Genstat 6.
The same procedure was followed for the models containing control,
primary and secondary levels of anti-diphtheria and anti-tetanus antibodies as
response variables. Therefore, in total, there were six separate models with a
measure of humoral immunity as the dependent variable. The residuals of the
models were checked for homoscedasticity and normality and where necessary
the response variables were transformed appropriately.
Finally, the residuals from a reduced major axis regression of mass against
tarsus length were used as a dependent variable with all other variables
mentioned included as explanatory variables (again with replicate line as the
random term) to ascertain whether body condition as defined as body mass
controlling for skeletal size was affected by testosterone and/or corticosterone
manipulations and titers. The residualsderived from a leastsquares regression of
mass against tarsus length are not thought to be an appropriate measure of
condition (see Darlington and Smulders, 2001; Green, 2001), so we used
reduced major axis regression as recommended by Green (2001). This measure
of body condition was included in the immunity models as an explanatory
variable. The residuals of the model conformed to a normal distribution and
were homoscedastic, so a Gaussian model was appropriate for this model.
In all of the above models, all 2-way interactions were included between
treatment groups and hormone titers.
Plasma hormone levels
There was a significant difference in the expected direction
in maximum stress-induced CORT levels between the CORT
selection lines (Wald=96.96, df=2, p<0.001; see Fig. 1).
128 M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
There was also a significant difference between the testoster-
one treatment groups, again in the expected direction (mean
control=2.91 nmol l−1, SE±0.52, mean T-treated=12.67 nmol
l−1, SE±1.25; Wald=63.64, df=1, p<0.001).
Table 1 gives the Wald statistics divided by their respective
degrees of freedom (as well as significance) for all the
explanatory variables included in all models containing immune
responses and body condition as the dependent variables.
There was no significant effect of any of the independent
variables on the size of wing web swelling at 24 h after PHA
injection (CORT line: Wald=0.30, df=2, p=0.86; testosterone
group: Wald=0.94, df=1, p=0.33; p>0.05 for all covariates).
The only measure of humoral immunity that was significantly
affected by the hormone treatments was the secondary antibody
response against diphtheria. There were significant interactions
between CORT titer and T titer (Wald=4.28, df=1, p=0.04;
Fig. 2), and between CORT selection line and T titer
(Wald=8.89, df=2, p=0.01; Fig. 2). Fig. 2 shows that the
highest antibody response was exhibited by the males with the
highest levels of both CORTand T, but there is a suggestion that
high levels of T were immunosuppressive at low levels of
CORTin the high CORTselection line. This is indicated as dark
shading (low antibody response) at low levels of CORT and
high levels of T (Fig. 2(c)). To investigate these interactions
further, we tested whether corticosterone and testosterone titers
positively covaried with secondary antibody response to
diphtheria and what relationship testosterone titer had with the
same antibody response, in each CORT line separately. We
found a non-significant, positive interaction between both
hormone titers and antibody response in the low and high lines,
and a strongly significant positive relationship in the control line
(Wald=11.51, df=1, p<0.001). Testosterone titers had a non-
significant negative relationship with antibody response in the
low line; a significant positive relationship in the control line
(Wald=13.99, df=1, p<0.001); and a significant negative
relationship with antibody response in the high line
(Wald=4.75, df=1, p=0.029). However, it should be borne in
mind that such results are more prone to a Type I error, and
obviously rely on smaller sample sizes, than when relationships
are inferred from models containing data from all treatment
The other covariates and interactions included in the models
had no significant effect on humoral immunity (p>0.05 in all
Given that in previous avian studies testosterone implanta-
tion has had a significant effect on both PHA response and
antibody response to diphtheria:tetanus injection (Casto et al.,
2001; Duffy et al., 2000; Owen-Ashley et al., 2004), we carried
out power analyses to ascertain whether our sample sizes were
too small to detect any significant differences in immune
response between testosterone treatment groups. In all cases, the
sample sizes obtained were far greater than both ours and the
sample sizes used in the studies that did find an effect of
testosterone on immunity (see Table 2). Indeed, the most similar
study to ours that implanted testosterone and immunized with
diphtheria:tetanus as well as injected with PHA found
Fig. 1. Predicted means of log10 body mass controlling for skeletal size (a) and
mean levels of maximum stress-induced plasma corticosterone (b) in replicate
lines of zebra finch (n=24 per line) selected for divergent levels of corticos-
terone±SE. The lines differed significantly in CORT level (Wald=96.96,
df=2, p<0.001) and body mass (Wald=16.99, df=2, p<0.001).
Effects of body condition and hormone treatment groups and titers on immune response and condition in zebra finches
Explanatory variablesResponse variables
Control PrimarySecondary ControlPrimary Secondary
CORT selection line
T treatment group
CORT line* T titer
CORT line* T treatment group
T treatment group* CORT titer
CORT titer* T titer
Variation explained by each explanatory variable is expressed by Wald statistics divided by its respective degrees of freedom (*p<0.05, **p<0.01, ***p>0.001).
129 M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
significant differences between treatment groups with samples
sizes of 6 birds in each group (Owen-Ashley et al., 2004). For
this reason, we suggest that our sample sizes were sufficiently
large to detect any differences.
There was a significant effect of CORT selection line such
that birds in the high line had greater mass (after controlling for
skeletal size) than controls and low CORT birds (Wald=16.99,
df=2, p<0.001; see Fig. 1) (see Table 1).
In addition, an interaction between CORT titer and T
treatment group also had a significant effect on body mass
(Wald=7.96, df=1, p=0.005; see Fig. 3). T implanted birds
showed a greater decrease in mass with increasing CORT
production than empty implanted birds. As with the significant
interactions reported for secondary antibody response to
diphtheria, we tested the nature of the relationship between
CORT titer and body mass within each testosterone treatment
group. We found that CORT titer was not a significant predictor
of mass in the empty-implanted (control T) group, but had a
significant, negative relationship with body mass in the high T
group (Wald=6.46, df=1, p=0.011).
From the results of the present experiment, it appears that
corticosterone and testosterone have an interactive effect on
immunity. Although no effect of either manipulated hormone
was found on cell-mediated immunity as measured by PHA
response, there was a significant, positive relationship between
the plasma levels of both hormones and the birds' secondary
antibody response to diphtheria; but this was only the case when
the other hormone was also at high levels in the blood plasma.
Whereas in the low and control lines there was a positive
relationship between T level (at high levels of CORT) and
secondary diphtheria antibody response; in the high line there
was a negative relationship at low levels of CORT. There was
nevertheless a positive relationship between Tand antibody titer
at high levels of CORT. As can be seen in Fig. 2, when all
interactions are included in the predicted model, the general
positive interactive effect the two hormones have on antibody
Fig. 2. Fitted model of the relationship between testosterone titer, corticosterone titer and diphtheria secondary antibody response in males selected for (a) low, (b)
control and (c) high maximum stress-induced levels of plasma corticosterone (n=24 per line). Antibody response is expressed as a percentage of the standard and is
log10 transformed. Lighter shades represent higher antibody responses. The interactions between CORTand testosterone titer (Wald=4.28, df=1, p=0.04) and CORT
line and testosterone titer (Wald=8.89, df=2, p=0.01) were significant.
Results from power analyses showing sample sizes required to give significant
differences between T treatment groups (α=0.05) in immune response given the
effect sizes obtained, at a power of 80%
Immune test T treatment
Sample size required
in each treatment
Primary tetanus Control 6.6
Fig. 3. Relationship between corticosterone titer, body mass and testosterone
treatment group. Triangles and solid trend line represent empty-implanted birds
(n=36); circles and dashed trend line represent testosterone-implanted birds
(n=36). The interaction was significant (Wald=7.96, df=1, p=0.005).
130M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
response is evident. In the males selected for high levels of
CORT, testosterone appears to be immunosuppressive, but only
in the presence of low levels of plasma corticosterone.
The results confirmed that the CORT lines differed
significantly in maximum stress-induced levels of CORT and
that the testosterone treatment groups differed significantly in T
in the expected direction. Little is known about wild male zebra
finch T levels, and these may not be relevant to our captive
selected finch stocks, but the manipulations changed the
testosterone-implanted males' T levels to be relatively high
compared to the levels of the empty-implanted males. These
levels however remained within physiological limits as the
levels were at the upper limit of the range of T titers previously
recorded in this generation of zebra finches.
Males from the high CORT lines had significantly greater
mass (when controlling for tarsus length) than males from the
low CORT lines. There was a significant, negative relationship
between actual CORT plasma levels and body mass in the high
T treatment group.
In general, individuals with both the highest CORT and T
plasma levels exhibited the greatest secondary antibody
response against diphtheria. This is contrary to our under-
standing that these hormones are immunosuppressive, and as
this is a manipulative experiment, contrary to the predictions of
the ICHH. There was a suggestion from the results that, in the
presence of low levels of CORT, T may be immunosuppressive,
but only in individuals selected for high levels of CORT (Fig. 2
(c)). In the other CORT lines, testosterone and corticosterone
titers together were positively related to antibody response. This
may be an idiosyncratic characteristic of the high CORT line or
may suggest that birds with genes for a high maximum stress-
induced CORT response are immunosuppressed by high levels
of T if CORT levels are relatively low. Nevertheless, the overall
effect of both hormones on humoral response is positive when
both hormones are at high levels in the blood plasma. Previous
studies have generally found either T or CORT to be
immunosuppressive (Buchanan et al., 2003; Casto et al.,
2001; Duffy et al., 2000; Owen-Ashley et al., 2004; Peters,
2000; Evans et al., 2000; Råberg et al., 1998; Wingfield et al.,
1997 respectively). In contrast, a few studies have found these
hormones to exhibit immuno-enhancing properties (T: Evans et
al., 2000; Peters, 2000; CORT: Svensson et al., 2002). This is
the first experiment to our knowledge that has found a
significant, positive interactive effect of both hormones on
antibody response. Testosterone may have a positive effect on
immunocompetence by increasing aggressive behavior and
therefore dominance ranking; this may lead to greater access to
dietary requirements, and consequently more resources are
made available for immune defense (Evans et al., 2000). The
birds in this study were housed together and may have
competed for food; however, they were fed ad libitum.
Therefore, it is not clear how greater dominance caused by
testosterone treatment affected dietary intake. An alternative
explanation is that, rather than being able to obtain more or
better quality food, males with high testosterone may be better
able to utilize the food they ingest for immune function.
Exogenous testosterone administration has been found to
increase plasma carotenoid concentrations via upregulation of
carotenoid-binding lipoproteins (McGraw et al., 2006). Car-
otenoid concentration in the blood is in turn known to have a
positive effect on immune function and antioxidant defenses
(Alonso-Alvarez et al., 2004; Blount et al., 2003; Faivre et al.,
2003; McGraw and Ardia, 2003).
Elevated CORT levels result in an increase in foraging
behavior and mobilization of glucose reserves (Breuner et al.,
1998; Buchanan, 2000; Silverin, 1998; Wingfield et al., 1997),
therefore an increase in both hormones may result in an excess
of resources readily accessible for producing and regulating an
antibody response. Moreover, peak CORT levels (as opposed to
basal levels) and acute stress have also been found to enhance
immune response in several studies (e.g. Cocke et al., 1993;
Dhabhar and McEwen, 1999; Persoons et al., 1995; see
Dhabhar, 2002 for a review). Thus, a combination of both
high testosterone and high peak levels of corticosterone in the
blood may instigate either directly or indirectly a potent
Several studies have found that artificially increasing
testosterone correspondingly increases CORT levels (Casto et
al., 2001; Duffy et al., 2000; Evans et al., 2000; Owen-Ashley et
al., 2004). By increasing T in individuals with naturally low
levels of T, an increase in CORT may be a predictable
phenomenon, partly because the methodology employed to
increase T is probably stressful and should increase CORT
levels (see Moore et al. 2004) and partly because increasing Tin
males with naturally low levels of the androgen may in itself be
stressful for the individual concerned and consequently raise
CORT levels, quite apart from the implantation procedure itself.
Individuals with naturally low levels of testosterone may well
be stressed by suddenly finding themselves with high (often
above physiological) levels of the hormone; this does not
necessarily mean that the two hormones naturally positively
covary. If the two hormones do positively correlate in free-
living birds (and there are some data to support this, see
Johnsen, 1998; Mateos, 2005), females would be expected to
choose males with both high T (for their sexual signals) and
concomitantly high CORT. These males would have the greatest
humoral immune response, so females wouldbe choosing males
with the highest immunocompetence. This interpretation is
consistent with Hamilton and Zuk's (1982) parasite theory.
However, it should be borne in mind that we measured peak
levels of CORT in this study; although basal levels of CORT
may positively covary with T, it does not necessarily imply any
relationship between peak CORT and T. There are examples of
the two hormones negatively covarying within individuals in
multiple taxa (in birds: Parker et al., 2002; in reptiles: Knapp
and Moore, 1997; Lance et al., 2003; Moore et al., 2001; but see
Moore et al., 2000; in mammals: Sankar et al., 2000). The
important question that needs to be addressed is how do both
peak and basal CORT and testosterone naturally interact within
free-living individuals. This point iscrucial in understanding the
results of laboratory-based studies on immunocompetence and
endocrine systems in birds.
There was no effect of either hormone on the cell-mediated
immune response or on several measures of humoral immunity.
131 M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
Power analyses suggested that the sample sizes of our treatment
groups needed to be much larger to find a significant effect of T
treatment given the mean differences obtained; however,
significant effects of testosterone treatment on PHA and
antibody responses have been found in previous studies on
much smaller sample sizes (e.g. Casto et al., 2001; Duffy et al.,
2000; Owen-Ashley et al., 2004). In addition, other studies have
found no effect of testosterone treatment on the cell-mediated
response (Buchanan et al., 2003; Greenman et al., 2005).
Therefore, although we cannot definitely conclude that this null
result is not merely due to a small sample and effect size, it
seems more likely that this is a real result. The important point
to note is that, although no significant result was obtained,
neither hormone had an immunosuppressive effect on either the
cell-mediated response or on antibody production.
Body mass was negatively related to CORT levels, but only
in the high T treatment group. In addition to this, the males in
the high CORT lines were on average in better condition than
males from the low CORT lines when all the independent
parameters were accounted for in the model. These results
would suggest that both high T and high CORT in the same
individual will have a negative effect on general body mass. But
high CORTwith relatively lower levels of Twill have a positive
effect on mass. The alternative explanation is that differences
between the CORT lines may be an artefact of the selection
process, and rather than CORT per se affecting condition in this
way, it is possible that another determinant of condition may
have been incidentally selected. Nevertheless, this explanation
does not account for the negative effect of both high Tand high
CORT titer together on body condition. Several studies have
found a negative effect of T on body condition or body mass
(Clotfelter et al., 2004; Mougeot et al., 2004; Ros, 1999;
Wikelski et al., 1999), as well as a similar effect for CORT
(Sockman and Schwabl, 2001), so this result should not be too
surprising. However, given that males with simultaneously the
highest T and highest CORT titers exhibited the greatest
humoral immunity, it seems counter-intuitive to find that males
in the high testosterone treatment group with the highest CORT
titers males were also in the poorest condition, particularly if
these individuals had the highest immunocompetence due to
greater access to resources caused by high levels of both
hormones. The reasons for this are unclear, but it is possible that
these individuals quickly expend any excess resources on
energy demanding processes such as immune defense and
general locomotor activity. Low body mass will result if high
plasma CORT levels lead to glucose being quickly broken down
(therefore more resources will rapidly be made available for
immune defense). Additionally, elevated levels of both CORT
and T have been found to increase basal metabolic rate and
locomotor activity (Breuner et al., 1998; Buchanan et al., 2001;
Lynn et al., 2000; Wikelski et al., 1999). In addition, the positive
relationship we found between antibody production and
simultaneously high levels of both hormones may have nothing
to do with body mass or condition per se, but instead be related
to direct as well as indirect effects of the hormones in terms of
increasing carotenoid availability and stimulating antibody
Our study suggests that testosterone-mediated signaling may
not be costly in terms of immunosuppression, but may
detrimentally affect body condition. Therefore, poorer quality
males that elevate testosterone for sexual signaling may pay the
price in times of environmental hardship when food is scarce.
Higher quality males may be able to elevate testosterone
without losing body mass or alternatively be able to resist the
negative effects of low body mass. Our results certainly suggest
that the T-mediated costs of sexually selected traits do not occur
through direct immunosuppression but may occur through
metabolic (e.g. Buchanan et al., 2001) or dominance/aggression
costs (e.g. Rohwer and Ewald, 1981).
Only one measure of immunity exhibited any relationship
with either CORT or T, suggesting that these hormones do not
consistently affect all aspects of an individual's immune
response. It is important to note that the birds used in this
study were selected for maximum stress-induced levels of
CORT, and it was this measure of CORT interacting with T that
showed a relationship with the secondary antibody response
against diphtheria. Basal levels of CORT were not measured,
and these may well have had an effect that was missed.
Nevertheless, this experiment demonstrates that, contrary to the
generally accepted consensus, CORT and T are not always
associated with immunosuppression.
We would like to thank the staff of the Animal Unit at the
University of Stirling for their assistance with the selection
program and to Alistair Dawson for his advice on the
implantation procedure; we are also grateful to A.R. Goldsmith
for access to the RIA facilities at the University of Bristol. All
work was conducted under Home Office license PPL 60/2584.
MLR was funded by a studentship from NERC, and the
selection program under MRE and KLB was funded at various
times by the Royal Society, ASAB, NERC and the University of
Stirling, and DH was funded by the Swedish Research Council
for the Environment, Agricultural Sciences and Spatial Plan-
ning (Formas), the Carl Trygger Foundation, the Swedish
Research Council (VR) and the Crafoord Foundation. One
anonymous referee and Julio Blas provided very helpful
comments on an initial version of the manuscript.
Alonso-Alvarez, C., Ferrer, M., Figuerola, J., Veira, J.A.R., Estepa, J., Torres,
L.M., 2002. The effects of testosterone manipulation on the body condition
of captive male yellow-legged gulls. Comp. Biochem. Physiol., A 131,
Alonso-Alvarez, C., Bertrand, S., Devevey, G., Gaillard, M., Prost, J., Faivre,
B., Sorci, G., 2004. An experimental test of the dose-dependent effect of
carotenoids and immune activation on sexual signals and antioxidant
activity. Am. Nat. 164, 651–659.
Bentley, G.E., Spar, B.D., MacDougall-Shackleton, S.A., Hahn, T.P., Ball, G.F.,
2000. Photoperiodic regulation of the reproductive axis in male zebra
finches, Taeniopygia guttata. Gen. Comp. Endocrinol. 117, 449–455.
Blount, J.D., Metcalfe, N.B., Birkhead, T.R., Surai, P.F., 2003. Carotenoid
modulation of immune function and sexual attractiveness in zebra finches.
Science 300, 125–127.
132 M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
Breuner, C.W., Hahn, T.P., 2003. Integrating stress physiology, environmental
change, and behaviour in free-living sparrows. Horm. Behav. 43, 115–123.
Breuner, C.W., Greenberg, A.L., Wingfield, J.C., 1998. Noninvasive corticos-
terone treatment rapidly increases activity in Gambel's white-crowned
sparrows. Gen. Comp. Endocrinol. 111, 386–394.
Briganti, F., Papeschi, A., Mugnai, T., Dessi-Fulgheri, F., 1999. Effect of
testosterone on male traits and behaviour in juvenile pheasants. Ethol. Ecol.
Evol. 11, 171–178.
Buchanan, K.L., 2000. Stress and the evolution of condition-dependant signals.
Trends Ecol. Evol. 15, 157–160.
Buchanan, K.L., Evans, M.R., Goldsmith, A.R., Bryant, D.M., Rowe, L.V.,
2001. Testosterone influencesbasal metabolic rate in male house sparrows: a
new cost of dominance signalling? Proc. R. Soc. Lond., Biol. 268,
Buchanan, K.L., Evans, M.R., Goldsmith, A.R., 2003. Testosterone, dominance
signalling and immunosuppression in the house sparrow, Passerdomesticus.
Behav. Ecol. Sociobiol. 55, 50–59.
Buttemer, W.A., Astheimer, L.B., 2000. Testosterone does not affect basal
metabolic rate or blood parasite load in captive male white-plumed
honeyeaters Lichenostomus penicillatus. J. Avian Biol. 31, 479–488.
Casto, J.M., Van Nolan Jr., Ketterson, E.D., 2001. Steroid hormones and
immune function: experimental studies in wild and captive dark-eyed juncos
(Junco hyemalis). Am. Nat. 157, 408–420.
Chastel, O., Barbraud, C., Weimerskirch, H., Lormee, H., Lacroix, A., Tostain,
O., 2005. High levels of LH and testosterone in a tropical seabird with an
elaborate courtship display. Gen. Comp. Endocrinol. 140, 33–40.
Clotfelter, E.D., O'Neal, D.M., Gaudioso, J.M., Casto, J.M., Parker-Renga,
I.M., Snajdr, E.A., Duffy, D.L., Nolan, V., Ketterson, E.D., 2004.
Consequences of elevating plasma testosterone in females of a socially
monogamous songbird: evidence of constraints on male evolution? Horm.
Behav. 46, 171–178.
Cocke, R., Moynihan, J.A., Cohen, N., Grota, L.J., Ader, R., 1993. Exposure to
conspecific alarm chemosignals alters immune responses in BALB/c mice.
Brain Behav. Immun. 7, 36–46.
Cockrem, J.F., Silverin, B., 2002. Variation within and between birds in
corticosterone responses of great tits (Parus major). Gen. Comp.
Endocrinol. 125, 197–206.
Darlington, R.B., Smulders, T.V., 2001. Problems with residual analysis. Anim.
Behav. 62, 599–602.
Dhabhar, F.S., 2002. A hassle a day may keep the doctor away: stress and the
augmentation of immune function. Integr. Comp. Biol. 42, 556–564.
Dhabhar, F.S., McEwen, B.S., 1999. Enhancing versus suppressive effects of
stresshormoneson skin immunefunction.Proc. Natl. Acad.Sci. U.S. A.96,
Duffy, D.L., Bentley, G.E., Drazen, D.L., Ball, G.F., 2000. Effects of
testosterone on cell-mediated and humoral immunity in non-breeding
adult European starlings. Behav. Ecol. 11, 654–662.
Evans, M.R., Goldsmith, A.R., Norris, S.R.A., 2000. The effects of testosterone
on antibody production and plumage colouration in male house sparrows
(Passer domesticus). Behav. Ecol. Sociobiol. 47, 156–163.
Evans, M.R., Roberts, M.L., Buchanan, K.L., Goldsmith, A.R., 2006.
Heritability of corticosterone response and changes in life history traits
during selection in the zebra finch. J. Evol. Biol. 19, 343–352.
Faivre, B., Gegoire, A., Preault, M., Cezilly, F., Sorci, G., 2003. Immune
activation rapidly mirrored in a secondary sexual trait. Science 300, 103.
Folstad, I., Karter, A.J., 1992. Parasites, bright males, and the immunocompe-
tence handicap. Am. Nat. 139, 603–622.
Green, A., 2001. Mass/length residuals: measures of body condition or
generators of spurious results? Ecology 82, 1473–1483.
Greenman, C.G., Martin, L.B., Hau, M., 2005. Reproductive state, but not
testosterone, reduces immune function in male house sparrows (Passer
domesticus). Physiol. Biochem. Zool. 78, 60–68.
Grossman, C.J., 1985. Interactions between the gonadal steroids and the
immune system. Science 227, 257–261.
Hamilton, W.D., Zuk, M., 1982. Heritable true fitness and bright birds: a role for
parasites? Science 218, 384–387.
Harvey, S.,Phillips, J.G., Rees, A., Hall, T.R., 1984.Stress and adrenal function.
J. Exp. Zool. 232, 633–645.
Hasselquist, D., Marsh, J.A., Sherman, P.W., Wingfield, J.C., 1999. Is avian
humoral immunocompetence suppressed by testosterone? Behav. Ecol.
Sociobiol. 45, 167–175.
Hood, L.C., Boersma, P.D., Wingfield, J.C., 1998. The adrenocortical response
to stress in incubating Magellanic penguins (Spheniscus magellanicus). Auk
Johnsen, T.S.,1998.Behaviouralcorrelates of testosteroneand seasonalchanges
of steroids in red-winged blackbirds. Anim. Behav. 55, 957–965.
Jones, A.E., Slater, P.J.B., 1999. The zebra finch. In: Poole, T. (Ed.), UFAW
Handbook on the Care and Management of Laboratory Animals. Black-
Kitaysky, A.S., Kitaiskaia, E.V., Wingfield, J.C., Piatt, J.F., 2001. Dietary
restriction causes chronic elevation of corticosterone and enhances stress
response in red-legged kittiwake chicks. J. Comp. Physiol., B 171,
Klukowski, L.A., Cawthorn, J.M., Ketterson, E.D., Nolan Jr., V., 1997. Effects
of experimentally elevated testosterone on plasma corticosterone and
corticosteroid-binding globulin in dark-eyed juncos (Junco hyemalis).
Gen. Comp. Endocrinol. 108, 141–151.
Knapp, R., Moore, M.C., 1997. Male morphs in tree lizards have different
testosterone responses to elevated levels of corticosterone. Gen. Comp.
Endocrinol. 107, 273–279.
Lance, V.A., Elsey, R.M., Butterstein, G., Trosclair III, P.L., 2003. Rapid
suppression of testosterone secretion after capture in male American
alligators (Alligator mississippiensis). Gen. Comp. Endocrinol. 135,
Lochmiller, R.L., Vestey, M.R., Boren, J.C., 1993. Relationship between protein
nutritional status and immunocompetence in northern bob-white chicks.
Auk 110, 503–510.
Lormee, H.,Jouventin,P., Trouve, C., Chastel,O., 2003. Sex-specific patterns in
baseline corticosterone and body condition changes in breeding red-footed
boobies Sula sula. Ibis 145, 212–219.
Lynn, S.E., Houtman, A.M., Weathers, W.W., Ketterson, E.D., Nolan, V., 2000.
Testosterone increases activity but not daily energy expenditure in captive
male dark-eyed juncos, Junco hyemalis. Anim. Behav. 60, 581–587.
Lynn, S.E., Hunt, K.E., Wingfield, J.C., 2003. Ecological factors affecting the
adrenocortical response to stress in chestnut-collared and McCown's
longspurs (Calcarius ornatus, Calcarius mccownii). Physiol. Biochem.
Zool. 76, 566.
Mateos, C., 2005. The subordination stress paradigm and the relation between
testosterone and corticosterone in male ring-necked pheasants. Anim.
Behav. 69, 249–255.
McGraw, K.J., Ardia, D.R., 2003. Carotenoids, immunocompetence, and the
information content of sexual colours: an experimental test. Am. Nat. 162,
McGraw, K.J., Correa, S.M., Adkins-Regan, E., 2006. Testosterone upregulates
lipoprotein status to control sexual attractiveness in a colorful songbird.
Behav. Ecol. Sociobiol. 60, 117–122.
Moore, I.T, Lerner, J.P, Lerner, D.T., Mason, R.T., 2000. Relationships between
annual cycles of testosterone, corticosterone, and body condition in male
red-spotted garter snakes, Thamnophis sirtalis concinnus. Physiol. Biochem.
Zool. 73, 307–312.
Moore, I.T., Greene, M.J., Mason, R.T., 2001. Environmental and seasonal
adaptations of the adrenocortical and gonadal responses to capture stress in
two populations of the male garter snake, Thamnophis sirtalis. J. Exp. Zool.
Mougeot, F., Irvine, J.R., Seivwright, L., Redpath, S.M., Piertney, S., 2004.
Testosterone, immunocompetence, and honest sexual signalling in male red
grouse. Behav. Ecol. 15, 930–937.
Owen-Ashley, N.T., Hasselquist, D., Wingfield, J.C., 2004. Androgens and the
Immunocompetence Handicap Hypothesis: unravelling direct and indirect
pathways of immunosuppression in song sparrows. Am. Nat. 164, 490–505.
Parker, T.H., Knapp, R., Rosenfield, J.A., 2002. Social mediation of sexually
selected ornamentation and steroid hormone levels in male junglefowl.
Anim. Behav. 64, 291–298.
Parkinson, T.J., Follett, B.K., 1995. Thyroidectomy abolishes testicular cycles
of Soay rams. Proc.R. Soc. Lond., Biol. 259, 1–6.
Pereyra, M.E., Wingfield, J.C., 2003. Changes in plasma corticosterone and
133 M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134
adrenocortical response to stress during the breeding cycle in high altitude Download full-text
flycatchers. Gen. Comp. Endocrinol. 130, 222–231.
Perfito, N., Schirato, G., Brown, M., Wingfield, J.C., 2002. Response to acute
stress in the Harlequin duck (Histrionicus histrionicus) during the breeding
season and moult: relationships to gender, condition, and life-history stage.
Can. J. Zool. 80, 1334–1343.
Persoons, J.H.A., Berkenbosch, F., Schornagel, K., Thepen, T., Kraal, G., 1995.
Increased specific IgE production in lungs after the induction of stress in
rats. J. Allergy Clin. Immunol. 95, 765–770.
Peters, A., 2000. Testosterone treatment is immunosuppressive in superb fairy-
wrens, yet free-living males with high testosterone are more immunocom-
petent. Proc.R. Soc. Lond., Biol. 267, 883–889.
Peters, A., Denk, A.G., Delhey, K., Kempenaers, B., 2004. Carotenoid-based
bill colour as an indicator of immunocompetence and sperm performance in
male mallards. J. Evol. Biol. 17, 1111–1120.
Poiani, A., Goldsmith, A.R., Evans, M.R., 2000. Ectoparasites of house
sparrows (Passer domesticus): an experimental test of the immunocompe-
tence handicap hypothesis and a new model. Behav. Ecol. Sociobiol. 47,
Råberg, L., Grahn, M., Hasselquist, D., Svensson, E., 1998. On the adaptive
significance of stress-induced immunosuppression. Proc.R. Soc. Lond.,
Biol. 265, 1637–1641.
Roberts, M.L., Buchanan, K.L., Evans, M.R., 2004. Testing the immunocom-
petence handicap hypothesis: a review of the evidence. Anim. Behav. 68,
Rohwer, S., Ewald, P.W., 1981. The cost of dominance and advantage of
subordination in a badge signaling system. Evolution 35, 441–454.
Romero, L.M., Wingfield, J.C., 1998. Seasonal changes in adrenal sensitivity
leucophrys gambelii). Comp. Biochem. Physiol. 119C, 31–36.
Ros, A.F.H., 1999. Effects of testosterone on growth, plumage pigmentation,
and mortality in black-headed gull chicks. Ibis 141, 451–459.
Sankar, B.R., Maran, R.R.M., Sudha, S., Govindarajulu, P., Balasubramanaian,
K., 2000. Chronic corticosterone treatment impairs Leydig cell 11 beta-
hydroxysteroid dehydrogenase activity and LH-stimulated testosterone
production. Horm. Metab. Res. 32, 142–146.
Sapolsky, R.M., Romero, M., Munck, A.U., 2000. How do glucocorticoids
influence stress responses? Integrating permissive, suppressive, stimulatory,
and preparative actions. Endocr. Rev. 21, 55–89.
Schoech, S.J., Ketterson, E.D., Van Nolan Jr., D., 1999. Exogenous
testosterone and the adrenocortical response in dark-eyed juncos. Auk
Schwabl, H., 1995. Individual variation of the acute adrenocortical response to
stress in the white- throated sparrow. Zoology 99, 113–120.
Silverin,B., 1998. Stress responses in birds. Poult. Avian Biol. Rev. 9, 153–168.
Sockman, K.W., Schwabl, H., 2001. Plasma corticosterone in nestling American
kestrels: effects of age, handling stress, yolk androgens, and body condition.
Gen. Comp. Endocrinol. 122, 205–212.
Svensson,E.I., Sinervo,B., Comendant, T., 2002.Mechanistic andexperimental
analysis of condition and reproduction in a polymorphic lizard. J. Evol. Biol.
metabolism, testosterone and corticosterone in white-crowned sparrows.
J. Comp. Physiol., A 185, 463–470.
Wingfield, J.C., 1994. Hormone–behavior interactions and mating systems in
male and female birds. In: Short, R.V., Balaban, E. (Eds.), The Difference
Between the Sexes. Cambridge Univ. Press, London, pp. 303–330.
Wingfield, J.C., Vleck, C.M., Moore, M.C., 1992. Seasonal changes of the
adrenocortical response to stress in birds of the Sonoran desert. J. Exp. Zool.
Wingfield, J.C., Breuner, C., Jacobs, J., 1997. Corticosterone and behavioural
responses to unpredictable events. In: Harvey, S., Etches, R.J. (Eds.),
Perspectives in Avian Endocrinology. Journal of Endocrinology Ltd.,
134 M.L. Roberts et al. / Hormones and Behavior 51 (2007) 126–134