Arch. Environ. Contam. Toxicol.
43, 338 –344 (2002)
DOI: 10.1007/s00244-002-1200-9
Thyroid Hormone Suppression and Cell-Mediated Immunomodulation in
American Kestrels (Falco sparverius) Exposed to PCBs
J. E. Smits,1 K. J. Fernie2, G. R. Bortolotti,3 T. A. Marchant3
1
Toxicology Centre and Veterinary Pathology, University of Saskatchewan, 44 Campus Drive, Saskatoon, Saskatchewan S7N 5B3, Canada
2
Environment Canada, 867 Lakeshore Road, Burlington, Ontario L7R 4A6, Canada
3
Department of Biology, University of Saskatchewan, 112 Science Place, Saskatoon, Saskatchewan S7N 5E2, Canada
Received: 10 October 2001/Accepted: 5 April 2002
Abstract. Exposure to environmental contaminants can induce
physiological changes in animals through various mechanisms.
One manifestation of subclinical toxicity from polychlorinated
biphenyl (PCB) exposure is the disruption of normal immune
function described in numerous species, including American
kestrels (Falco sparverius). In 1998, 152 mature male and
female kestrels were fed either a mixture of Aroclor 1248:
1254:1260 (approximately 7 mg/kg kestrel/day) through their
food items, or control diets. Offspring produced by 50 breeding
pairs (thus, half received in ovo PCB exposure only) were also
studied. Total and differential white blood cell counts, the
phytohemagglutinin (PHA) skin response, as well as thyroid
hormone levels were tested in vivo in nonbreeding adults (1998
only) and nestlings (1998 and 1999). In 1999, nestlings came
from three parental groups; adults exposed in 1998, birds
produced by PCB-exposed parents, and unexposed birds. In
1998, directly exposed males but not females had increased
total white blood cell counts driven by lymphocytosis, plus a
decreased heterophil-to-lymphocyte ratio relative to controls.
PCB-exposed birds had a significantly greater response to PHA
than did controls, with sex as a significant factor and plasma
triiodothyonine (T3) as a significant covariate. Levels of T3
were significantly depressed in PCB-exposed birds of both
sexes. The 1999 nestlings (F1 generation with respect to PCB
exposure) did not show any effect of parental treatment group
on the PHA skin response, yet T3 remained as a significant
covariate. Immunological effects are discussed in light of the
antibody-mediated immunotoxicity found in the same birds and
reported previously.
Polychlorinated biphenyls (PCBs) continue to persist as envi-
ronmental contaminants due to their lipophilic nature and their
ability to bioaccumulate and biomagnify in the food chain
(Hoffman et al. 1996). These are among the numerous com-
pounds that are causing concern because of their potential as
Correspondence to: J. E. Smits; email: judit.smits@usask.ca
A R C H I V E S O F
Environmental
Contamination
a n d Toxicology
© 2002 Springer-Verlag New York Inc.
endocrine disruptors in (Ankley and Giesy 1998; Ankley et al.
1998). Contaminant-induced embryonic mortality and malfor-
mation, as well as reduced reproductive success, have been
well described in birds exposed to dietary PCBs (Fernie et al.
2001a, 2001b; Gilbertson et al. 1991; Hoffman et al. 1996;
Kozie and Anderson 1991; Schwartz et al. 1993).
In addition to the well-studied reproductive effects of PCBs,
the possibility of immunotoxic effects is a growing concern.
For example, mass mortalities from viral infections in seals
were associated with immunosuppression likely due to PCB
exposure (Ross 1995). PCBs have been documented to cause
immune modulation in avian (Fairbrother 1994; Smits and
Bortolotti 2001) and mammalian species (Poland and Knutson
1982; Ross 1995). In nestling herring gulls (Larus argentatus)
and Caspian terns (Sterna caspia) exposed to persistent organo-
chlorine contaminants on the Great Lakes, T cell–mediated
immunosuppression was expressed through a decreased re-
sponse to the phytohemagglutinin (PHA) skin test (Grasman et
al. 1996). In addition, immune function is proving to be a
sensitive indicator of contaminant exposure when compared to
other measures of animal health (Smits et al. 1996a, 1996b).
Increasingly, immunotoxicology is being applied as a monitor-
ing tool for evaluating the impacts of environmental contami-
nants on wildlife (Bishop et al. 1998; Ross 1995; Smits et al.
1996a, 1999). Immune function has been compromised in
waterfowl (Fowles et al. 1997b), upland game birds (Grasman
and Scanlon 1995), and passerines (Smits and Williams 1999)
exposed to a range of environmental pollutants.
Because American kestrels are at the top of the food chain
and are common over a wide geographical range, they are a
convenient model to study effects of xenobiotics (Wiemeyer
and Lincer 1987). Studies have shown kestrels to be suscepti-
ble to PCB toxicity; experimental exposure caused moderate
hepatic enzyme (CYP1A) induction in adults (Elliot et al.
1997), and severe hepatic, thyroid, splenic, and bursal pathol-
ogy in nestlings (Hoffman et al. 1996). The latter findings are
consistent with immunomodulation.
Measures of immune function in birds have comprised a
routine component of studies of pathology and physiology in
domestic fowl (e.g., Chu et al. 1995; Goto et al. 1978; Laudert
et al. 1993; Van der Zijpp 1983). However, under field condi-
Immunomodulation in PCB-Exposed Kestrels
tions, there are a limited number of nonlethal tests that can be
used in vivo in wild birds, two popular exceptions being the
PHA skin test and antibody production against sheep red blood
cells (Grasman and Scanlon 1995; Smits et al. 1999). The PHA
skin test provides a measure of the proliferative response
potential of circulating T lymphocytes to an injected mitogen
(Goto et al. 1978; Hungerford et al. 1959). An evaluation of
immune function can be achieved more comprehensively by
concurrent assessment of the status of thyroid hormones. Nor-
mal thyroid hormone levels are necessary for adequate devel-
opment, maintanence and function of both the antibody- and
cell-mediated immune responses (Erf 1993; Cremaschi et al.
2000; Klecha et al. 2000). Because of the recognized potential
for PCBs to cause immune modulation and thyroid hormone
disruption, we assessed plasma thyroxine (T4) and triiodothy-
ronine (T3) together with the T lymphocyte–mediated immune
response. As part of an overall health assessment, we con-
ducted routine white blood cell (WBC) counts and differentials
as well.
Materials and Methods
This study was conducted at the Avian Science and Conservation
Centre of McGill University (Quebec, Canada) using an outbred
colony of American kestrels of known pedigree. In 1998 birds were
fed PCBs that had been injected into their food items; in 1999 indi-
viduals exposed in the previous year were tested again without addi-
tional dietary inputs (Fernie et al. 2001a, 2001b). In March 1998,
nonbreeding birds and future breeding pairs (n ⫽ 152) were randomly
assigned to PCB or control groups. They were housed in large flight
pens segregated by sex and treatment. Birds were maintained at
ambient temperature, and their diet was exclusively day-old cockerels,
as it had been for several generations. Exposure of kestrels to PCBs
began on 18 March and continued for 120 days for nonbreeders or
until the hatching of the first egg for breeding pairs. A mixture of 1:1:1
Aroclor 1248:1254:1260 suspended in food grade safflower oil was
injected intracranially into the cockerels. The food of control birds
received the equivalent dose of safflower oil. Kestrels were fed ad
libitum and were assumed to be eating 1.5 cockerels per day, although
this varied among individuals. The dosage offered to the birds was
calculated to deliver approximately 7 mg PCB/kg bird/day and re-
sulted in mean plasma levels of 3.8 ␮g/g wet weight at the end of the
dosing period (measured in male kestrels) (Droulliard et al. 2001).
Contaminant levels in the food items and the PCB mixture used in this
study were similar to the body burdens that have been found in small
mammals around PCB-contaminated sites in Canada (Environment
Canada unpublished data). PCB residues peaked after 80 days of
feeding on the Aroclor-contaminated diets and remained relatively
constant during the final 80 –120-day dosing period. Following cessa-
tion of the PCB dosing, tissue levels of individual PCBs began to
decline at different rates depending on congener chlorine substitution
patterns and chemical hydrophobicity. In general, labile PCBs were
completely cleared from animal tissues following 340 days of elimi-
nation, whereas clearance of persistent PCBs during this time varied
from 98% clearance to less than 50% clearance depending on the
chemical hydrophobicity (Drouillard et al. 2001).
On 21 April 1998, 25 pairs of both PCB-exposed and control birds
were put into individual, outdoor breeding pens (2.5 ⫻ 1.5 ⫻ 2.5 m).
The exposure to PCBs for these breeding adults continued for approx-
imately 100 days until the eggs in each nest started to hatch. Therefore
the fledglings produced by these birds (n ⫽ 32 from 18 control nests,
n ⫽ 14 from 15 PCB-exposed nests) had only been exposed in ovo.
Analysis of eggs revealed mean total PCB residues of 34.1 ␮g/g
339
(Fernie et al. 2001a) on a wet-weight basis, which is comparable to
levels found in free-ranging wildlife species (Bosveld and Van den
Berg 1994).
Blood samples were collected by jugular venipuncture using a
heparin-rinsed 1-ml syringe. Blood for the WBC counts was collected
on 1 July 1998; the PHA tests were conducted over the following 4
days. The blood smears, made on glass slides at the time of collection,
were stained with Wright-Giemsa and examined by light microscopy
using a 40⫻ objective. Total WBC count (WBC ⫻ 109/L) was esti-
mated based on the average number of leucocytes in five fields. The
WBC count was standardized to the number of red blood cells (RBCs);
to control for variability in smear thickness, only the fields with an
evenly distributed monolayer of RBCs were counted (described in
Tella et al. 2000). The proportions of heterophils, lymphocytes, mono-
cytes, eosinophils, and basophils were determined from a sample of
100 leukocytes counted at the feathered edge of the smear. Counts
were log-transformed for statistical analyses. To provide a frame of
reference for the blood cell counts, we present data on wild American
kestrels during the breeding season from north central Saskatchewan,
Canada, in 1997 (see Tella et al. 2000).
The T cell proliferative potential, one aspect of cell-mediated im-
mune function, was tested in vivo in nonbreeding adults (1998 only),
and all nestlings at 22–28 days of age (i.e., around fledging) (1998 and
1999) using the PHA skin test. The response of T lymphocytes was
based on the original techniques described by Goto et al. (1978), as
modified by Smits et al. (1999). Briefly, a 0.5–1 cm patch of skin on
the mid-patagium of the right wing was gently plucked or trimmed
clean of feathers. The bare skin was swabbed with alcohol and 50 ␮g
PHA in 50 ␮l of sterile phosphate buffered saline (PBS) was injected
subcutaneously with a 27-gauge needle. Patagium thickness was mea-
sured to 0.01 mm, using a digital micrometer (Mitutoyo 0 –1", Tokyo)
in 1998 and a gauge micrometer (Dyer OD gage 0.01 mm, Dyer,
Lancaster, PA) in 1999 immediately prior to the injection and 24 h
after injection. The mean of two to four measurements of patagium
thickness was recorded, and the PHA response was the calculated
difference between the 24-h and 0-h skin thickness.
In late July 1998, breeding birds and their offspring were transferred
to communal flight pens with the nonbreeding adults. In the spring of
1999, the adult PCB and control birds from 1998 were randomly
paired and bred again. The offspring from 1998 were bred with
experienced, control adults. No additional PCBs were fed to kestrels in
1999. At this time, whole-carcass residue samples from a small sub-
group of birds euthanized for this purpose showed that the readily
metabolizable congeners were completely eliminated from animals 1
year postexposure, whereas persistent PCB congeners exhibited 44 –
98% losses during the same time period (Drouillard et al. 2001).
The PHA response of nestlings produced in 1999 was tested in the
manner described. We randomly selected one nestling of each sex
(determined by sexually dimorphic feather coloration) per nest when
possible. Treatment groups were designated according to the female
parent’s PCB status; dams that had eaten PCBs in 1998 (PCB diet),
dams that had been exposed in ovo in 1998 (ovo exp), and dams that
had never been fed PCBs (control) (Smits and Bortolotti 2001).
As this study was part of a larger project on endocrine-disrupting
potential of PCBs, levels of corticosterone, total triiodothyronine (T3)
and thyroxine (T4) were also measured. On 1 July 1998 (just before
PCB exposure ended for nonbreeders) a 0.5-ml sample of blood was
taken from the jugular vein of each bird for hormone extraction, and
smears were taken for WBC differential counts. The skin test was
conducted on 4 July. In 1999, the blood sample was taken at the same
time as the PHA injection, between 1 and 4 July. The hormones listed
were examined as potential covariates of the PHA response during
statistical analysis.
T3 and T4 levels in plasma samples were determined by a radioim-
munoassay (RIA) specific for each hormone. Antisera and purified
hormones used for standards were purchased from Sigma (St. Louis,
MO), whereas radiochemicals were purchased from Amersham
340
(Oakville, ON) or NEN (Mandel Scientific, Guelph, ON). Other re-
agents and materials were purchased from VWR Canada (Edmonton,
AB) or Sigma. In all assays, samples were measured in duplicate, and
internal control samples containing known amounts of the hormones
were included to ensure that the accuracy of each assay was high and
that between-assay and within-assay variability (i.e., precision) was
acceptable (% coefficient of variation ⬍ 10%). Displacement curves
produced by serial dilutions of samples were parallel with each hor-
mone standard curve, indicating that the RIAs described below were
suitable for measurement of circulating hormone levels in kestrels. For
all RIAs, the minimum detection limit was defined as a hormone
concentration less than the ED80 of the standard curve.
Thyroid hormone RIAs were conducted using procedures that de-
termine total T3 and total T4 levels (Chopra 1972), as described in
instructions from the supplier of the antisera. The relatively small
plasma volume available in this study precluded measurement of free
thyroid hormone levels in kestrel plasma. The T3 and T4 antisera were
diluted so as to obtain specific binding of 25–30%; nonspecific binding
was low (⬍ 10%) in the RIAs. Incubation of diluted antiserum (200
␮l) with samples or standards (100 ␮l total) and approximately 15,000
CPM of 125I-hormone (200 ␮l) was performed at room temperature for
24 h. A gamma counter (Titretech, Huntsville, AL) was used to
measure radioactivity. A second-antibody technique was then used to
separate bound and free hormone. For assays of samples collected in
1998, 125I-hormones were purchased from Amersham. Subsequent T3
and T4 RIAs were conducted using 125I-T3 and 125I-T4, respectively,
prepared according to Kjeld et al. (1975). Internal control samples
indicated that precision and variability of the RIAs were equivalent
and not dependent on the origin of the 125I-hormones. Based on
information from the supplier, the T3 and T4 antisera utilized in these
RIAs display a high degree of hormone specificity. The T3 antiserum
has negligible cross-reactivity (⬍ 0.5%) with T4 and related molecules
(diiodo-L-thyronine, diiodo-L-tyrosine, and moniodo-L-tyrosine). The
T4 antiserum displays only 5% cross-reactivity with T3 and ⬍ 0.5%
cross-reactivity with diiodo-L-tyrosine, moniodo-L-tyrosine, and D-
thyroxine. Plasma samples were diluted 4- to 10-fold prior to use in the
thyroid hormone assays and measured at the midrange of the standard
curves. The minimum detection limit was 80 pg per ml of plasma for
the T3 RIA and 250 pg per ml of plasma for the T4 RIA.
Statistical analyses were performed using SPSS (Norušis 1993). A
general linear model (GLM) was used to test responses of PCB versus
control groups, the criteria for significance being p ⬍ 0.05. The PHA
skin test carried out on nonbreeding adults and nestlings in 1998 was
analyzed with a GLM using sex, treatment, and age category (fledg-
lings and adults) as fixed factors and T3 as a covariate. Mass was
examined as a potential covariate also, using a subset of the data; but
it proved to be nonsignificant (F1,73 ⫽ 0.007, p ⫽ 0.933) and was
excluded to expand the sample size. In 1999, only sex and treatment
were relevant factors, and T3 was again used as a covariate.
Results
Total WBC counts from nonbreeding kestrels sampled during
early July 1998 were significantly increased in PCB-exposed
males (F1,24 ⫽ 7.206, p ⫽ 0.013) but not females (F1,20 ⫽
0.564, p ⫽ 0.46). This was largely driven by significantly
higher numbers of lymphocytes (F1,24 ⫽ 14.654, p ⫽ 0.001),
which in turn resulted in the heterophil to lymphocyte (H:L)
ratio being significantly decreased in the PCB-treated males
(F1,24 ⫽ 10.491, p ⫽ 0.003) (Table 1).
In 1998 T4 levels showed a significant effect due to age, as
chicks had higher concentrations than adults (F1,73 ⫽ 5.282,
p ⫽ 0.024), but not sex (F1,73 ⫽ 0.012, p ⫽ 0.914) or treatment
(F1,73 ⫽ 0.534, p ⫽ 0.467) (Table 2). However, T3 levels were
J. E. Smits et al.
significantly lower in males than females (F1,80 ⫽ 11.449, p ⫽
0.001), and lower in PCB-exposed than control birds (F1,80 ⫽
16.081, p ⬍ 0.001), and lower in adults than nestlings (F1,80 ⫽
8.404, p ⫽ 0.005). In 1999, considering only the nestlings, T4
levels in PCB-exposed kestrels were lower than controls
(F1,43 ⫽ 4.365, p ⫽ 0.043), but there was no sex (F1,43 ⫽
0.743, p ⫽ 0.393) effect; T3 no longer varied with treatment
(F1,97 ⫽ 2.112, p ⫽ 0.149) or sex (F1,97 ⫽ 0.583, p ⫽ 0.447)
(Table 2).
In 1998, PCB birds had a significantly greater response to
PHA than controls (F1,79 ⫽ 4.518, p ⫽ 0.037), with concurrent
significant effects for sex (F1,79 ⫽ 1.851, p ⫽ 0.006) and T3
(F1,79 ⫽ 1.950, p ⫽ 0.005) but not for age (F1,79 ⫽ 0.665, p ⫽
0.095) (Figure 1A). There were no significant interactions. The
estimated marginal means (⫾ SE) for PHA skin responses as
calculated by the GLM were 1.639 ⫾ 0.069 and 1.886 ⫾ 0.091
for control and PCB-treated birds, respectively.
The 1999, nestlings, i.e., F1 generation with respect to PCB
exposure, did not show any effect of parental treatment group
(F2,49 ⫽ 0.127, p ⫽ 0.881) or sex (F1,49 ⫽ 0.052, p ⫽ 0.820).
However, the PHA skin response still correlated positively
with T3 (F1,49 ⫽ 4.407, p ⫽ 0.041) (Figure 1B) and almost with
mass (F1,49 ⫽ 3.883, p ⫽ 0.054). The estimated marginal
means (⫾ SE) for PHA skin responses as calculated by the
GLM were 1.421 ⫾ 0.129 for controls, 1.512 ⫾ 0.185 for PCB
diet, and 1.360 ⫾ 0.271 for ovo-exp. Because the level of
exposure for F1 birds from in ovo– exposed mothers is likely
very low, we reanalyzed the data based on two treatments only:
nestlings with parents directly exposed to PCBs the previous
year, and those with parents receiving no direct exposure (n ⫽
16 and 32, respectively). The results were identical to analyses
based on three treatment groups.
Discussion
As is seen in many avian species (Campbell 1995), there is a
wide variation in the normal leukogram of kestrels. There is no
established range of normality for this species. In comparison
with the leukogram values of the control birds as well as the
range found in wild birds, it appears that the WBC values of the
PCB-exposed birds were normal. This supports the clinical
observations that they were free from significant disease. Over-
all, the WBC count was substantially lower in the captive
compared to the wild kestrels, which may be expected because
birds in the wild are exposed to a much broader range of
infectious organisms, fluctuations in prey supply, and natural
and weather-related stressors that are not encountered in a
controlled, captive environment. For American kestrels, lym-
phocytes normally tend to be more numerous than heterophils,
and females have higher total leukocyte counts than males.
The total and differential leukocyte counts revealed a sex-
specific pattern in wild and control birds, but this largely
disappeared in the PCB-exposed kestrels (Table 1). Males
exposed to PCBs had conspicuously higher WBC counts than
control males. This was largely driven by a lymphocytosis,
although heterophils were marginally higher in the PCB group.
Lymphocytosis is normally stimulated by chronic antigenic
exposure associated with specific types of infectious agents.
Heterophilia is generally associated with acute infections and
Immunomodulation in PCB-Exposed Kestrels 341

Table 1. Hematological variablesa (mean ⫾ SE) from American kestrels exposed to dietary PCBs or safflower oil vehicle (controls) or from
wild, free-ranging kestrels

Lymphocytes
Total ⫻ 109/L Heterophils Heterophil/
n WBCs n (⫾ SE) n ⫻ 109/L n Lymphocyte
Females
PCB diet 11 11.4 ⫾ 0.91 11 5.3 ⫾ 0.27 11 4.8 ⫾ 0.66 11 0.90 ⫾ 0.128
Control 11 11.1 ⫾ 1.70 11 5.2 ⫾ 0.75 11 3.9 ⫾ 0.62 12 0.72 ⫾ 0.063
Wild 84 20.0 ⫾ 1.21 81 9.0 ⫾ 0.58 72 5.5 ⫾ 0.31 73 0.74 ⫾ 0.064
Males
PCB diet 13 9.3 ⫾ 1.12 13 4.4 ⫾ 0.47* 13 4.2 ⫾ 0.75* 13 0.99 ⫾ 0.106*
Control 13 5.9 ⫾ 0.58 13 2.2 ⫾ 0.22 13 3.1 ⫾ 0.36 13 1.47 ⫾ 0.104
Wild 48 11.4 ⫾ 0.73 46 5.8 ⫾ 0.51 34 4.4 ⫾ 0.42 36 0.96 ⫾ 0.138
a
Blood samples from captive birds (PCB diet and control) were collected during the time pairs were raising nestlings; those from the wild birds
were collected during incubation.
*Values in the PCB-diet group are different from those in the control group (p ⬍ 0.05).

Table 2. Plasma thyroid hormone levels (mean ng/ml ⫾ SE) in adult and nestling American kestrels from PCB-exposed and control groups
(see text for treatment details)
Nestlings Adults
Treatment n Male ⫻ ⫾ SE n Female ⫻ ⫾ SE n Male ⫻ ⫾ SE n Female ⫻ ⫾ SE
1998
T3 PCB 7 0.854 ⫾ 0.011 7 1.108 ⫾ 0.092 8 0.498 ⫾ 0.045 12 0.748 ⫾ 0.066
Control 13 1.012 ⫾ 0.097 18 1.259 ⫾ 0.093 8 0.870 ⫾ 0.068 13 0.969 ⫾ 0.087
T4 PCB 7 5.882 ⫾ 0.620 7 5.764 ⫾ 0.420 7 4.928 ⫾ 0.449 12 4.719 ⫾ 0.334
Control 13 5.899 ⫾ 0.471 18 6.325 ⫾ 0.397 5 5.535 ⫾ 0.262 12 4.340 ⫾ 0.353
1999
T3 PCB diet’98 10 1.875 ⫾ 0.337 11 1.737 ⫾ 0.154
In ovo PCB 8 1.460 ⫾ 0.109 8 2.100 ⫾ 0.086
Control 34 2.013 ⫾ 0.137 30 2.155 ⫾ 0.209
T4 PCB diet’98 4 4.088 ⫾ 1.049 7 4.051 ⫾ 0.771
In ovo PCB 5 7.454 ⫾ 2.021 4 6.055 ⫾ 1.029
Control 16 7.822 ⫾ 1.165 15 6.907 ⫾ 0.858

social or environmental stressors. Environmental contaminants,
such as insecticides (Mandal et al. 1986), crude oil (Leighton
1986), and some organochlorines (Grasman et al. 2000), have
been found to stimulate a relative heterophilia in birds, but
findings are variable. Some studies report depressed (Gross and
Siegel 1983) or no effect (Grasman et al. 2000) on heterophil
numbers in exposed birds.
The captive kestrels in this study were clinically healthy
before and after these blood collections, so it is unlikely that
they were suffering from an infectious disease. The males but
not the females exposed to PCBs had significantly depressed
corticosterone levels (p ⬍ 0.05, Marchant et al. unpublished
data) along with higher levels of circulating lymphocytes. It is
possible that the normal physiological increase in serum con-
centrations of glucocorticosterone related to capture and han-
dling during experimental manipulations, which results in sup-
pression of the immune reaction, was responsible for
moderating the production and release of T lymphocytes in the
control birds. Because the PCB-treated males were not mount-
ing the normal corticosterone-related stress response (i.e.,
lower circulating WBC counts), this may have allowed in-
creased production and release into circulation of T lympho-
cytes. The finding of higher numbers of T cells in circulation
would be compatible with the increased PHA skin response
seen in the birds exposed to PCBs, as they are the main effector
cells in this reaction.
The PHA skin test response demonstrated that in nestling
kestrels T lymphocytes are mature enough to respond to mito-
genic stimulation and to promote cytokine-mediated local mo-
lecular and vascular reactions that allow the perivascular ac-
cumulation of mononuclear cells (Goto et al. 1978; Kean and
Lamont 1994), which are collectively responsible for the mea-
surable skin swelling. At hatching, chickens have phenotypi-
cally mature T cells (CD4⫹ and CD8⫹) that can bind mito-
gens, but they show no proliferative response, nor can they
secrete cytokines including interleukin-2 (IL-2) (Lowenthal et
al. 1994). By 4 days of age, their T cell proliferative response
is approximately 60% that of mature birds, whereas IL-2 pro-
duction is approximately 32%. In an altricial species, such the
kestrel, this delay in T cell function is likely to be longer than
in precocial species, such as domestic chickens, although the
ontogeny of the kestrel immune system is not yet known. In
3-week-old nestlings from this study, the PHA response was
not different from that of adult birds (greater than 1 year old)
although age category approached significance in the GLM.
Because of dietary as well as other challenges, immunological
development must occur rapidly in young carnivorous birds,
342
Fig. 1. (A) A scatterplot depicting the increased phytohemagglutinin
skin response of kestrels directly (adults) and indirectly (nestlings
exposed in ovo) exposed to PCBs compared with control birds in 1998.
Solid circles and dashed line ⫽ PCBs, open circles and solid line ⫽
controls. (B) A scatterplot demonstrating the relationship between
plasma T3 levels and the PHA skin response in nestlings from three
categories of parental exposure in 1999. Solid circles and dashed
line ⫽ dams fed PCBs in 1998; solid triangles and dotted line ⫽ dams
exposed in ovo to PCBs; open circles and solid line ⫽ dams not
exposed to PCBs)
but comparisons of immunological development between these
and precocial birds are not well documented.
The positive relationship between plasma T3 concentrations
and the PHA response of PCB-exposed males during the year
of direct exposure persisted the following year, although there
was no longer any effect of maternal exposure to PCBs. The
influence of chronic stress on thyroid hormones and resultant
effects on the immune response (Cremaschi et al. 2000), as
well as interaction between the thyroid axis and immune func-
tion has been well studied in mammals (Klecha et al. 2000;
Perez-Castro et al. 1998). Reduced serum levels of thyroid
hormones were accompanied by lower antibody production in
stressed mice (Cremaschi et al. 2000). Klecha et al. (2000)
demonstrated decreased cell-mediated and humoral immune
responses in animals with low thryoid hormone levels. Simi-
larly, Perez-Castro et al. (1998) has shown the T cell– depen-
dent antibody response in rats to be inhibited subsequent to
J. E. Smits et al.
suppression of the thyroid hormone release. Influences of thy-
roid hormones on immune function have been studied in avian
species as well. Fowles et al. (1997a) found that in mallard
ducks with depressed thyroid hormone levels, cytotoxic T cell
activity was suppressed, but thyroid status had no effect on T
cell– dependent antibody response, macrophage phagocytosis,
or natural killer cell activity. Decreased thyroid hormones in
chickens lead to an increased in vitro T cell proliferative
response to mitogenic stimulation (Williamson et al. 1990),
and increased T3 depressed T cell responses (Erf and Marsh
1989). Thus, the relationship between thyroid hormones and
immune response is far from predictable even without the
complication of PCB exposure.
As Rolland (2000) pointed out in her recent review of
contaminant-induced alterations in thyroid hormone levels in
wildlife, experimental and field studies have produced diver-
gent findings. No predictable pattern or consistent response is
emerging from different avian studies in which Ah-inducing
chemicals are being examined. This may be because of differ-
ences in experimental protocols and, for field research, unmea-
sured contaminants or confounding biological factors. In our
study, T3 proved to be more sensitive to suppression by PCBs
than was T4. These findings concur with those of Fowles et al.
(1997b) in which Aroclor caused depression of serum T3 in
mallards (Anas platyrhynchos). The mechanism by which
PCBs are interfering with circulating thyroid hormone levels
cannot be determined from the current study.
Modulation of immune function results from increased levels
of hormones (glucocorticoids) produced as a generic response
to social, physical, or contaminant-related stress, but expres-
sion of the shift in immune function is not predictable. In an
immunotoxicity study on Japanese quail (Coturnix coturnix),
experimental corticosterone treatment that caused immunosup-
pression of the cell-mediated immune response to PHA did not
decrease antibody response (Grasman and Scanlon 1995). In
contrast, in mallards (A. platyrhynchos) injected with the syn-
thetic corticosteroid dexamethazone, antibody titers were de-
creased while natural killer cell activity was increased (Fowles
et al. 1993). In the PCB-exposed birds, females had higher
antibody levels (Smits and Bortolotti 2001) but no difference in
plasma corticosterone relative to controls. The PCB exposed
males had decreased corticosterone levels, decreased antibody
production, and an increased PHA response. This may repre-
sent a PCB-induced, sex-specific shift in T cell subpopulations,
favoring the expression of T helper 1 (Th1) over T helper 2
(Th2) development. This shift would favor the activation of
cell-mediated immunity at the expense of the humoral immune
response (Rook et al. 1994).
In mammals immunized with sheep RBCs to stimulate a
humoral immune response, peak antibody production occur
simultaneously with peak cortisol level precipitated by in-
creased activity in the hypothalamus (Besedovsky and Sorkin
1977). Work by this group describes influences between the
immune response and hypothalamic-pituitary-adrenal (HPA)
axis. Several cytokines (interleukin-1, tumor necrosis factor,
interleukin-6) are important mediators of the HPA axis re-
sponse to activation of the immune system in vivo (Besedovsky
et al. 1991). They cause secretion of hypothalamic corticotrop-
in-releasing hormone, which leads to adrenocorticotropic hor-
mone release by the pituitary and subsequent glucocorticoid
production. This in turn enhances Th2 activity, which supports
Immunomodulation in PCB-Exposed Kestrels
antibody-mediated immunity. In highly stressed humans, the
cell-mediated but not humoral immune response is suppressed
(Rook et al. 1994). Thus, the suppression of glucocorticoids
indirectly enhances Th1 activity by blocking corticosteroid-
driven Th1 T cell suppression (Daynes et al. 1991). These
findings are consistent with the antibody-mediated (Smits and
Bortolotti 2001) and cell-mediated immune responses seen in
the PCB-exposed male kestrels. The mechanism through which
PCBs may be interacting with this system is different for males
and females. Exposure to PCBs suppressed plasma T3 levels in
both sexes, whereas their immune responses were either similar
(CMI) or opposite (antibody response) to each other. Cortico-
sterone, which plays a pivotal role in the HPA axis as well as
the immune response, was suppressed in males only (Marchant,
unpublished data).
The PHA skin test provides an indication of the responsive
potential of cell-mediated immunity in preparation for targeted
function. Clonal proliferation is integral to the cell-mediated
immune response, but it is important to recognize that this test
alone does not provide insight into the later functional capacity
of those cells. The role of well-differentiated T cells is to seek
and destroy foreign antigens, which is accomplished by recruit-
ing, through cytokine release, macrophages or B cells as re-
quired to help perform this function. In the wild, birds face
numerous environmental, infectious, physical, and nutritional
challenges. An apparently subtle change in the capacity of the
immune system under captive conditions could be exacerbated
in birds facing multiple stressors.
Through these studies, it is evident that dietary PCBs cause
immunomodulation of both cell-mediated and humoral (Smits
and Bortolotti 2001) immune responses in kestrels, and both
are strongly sex linked. PCBs will exist as environmental
toxicants for many years to come and are known to have a
negative impact on reproductive performance in high-trophic-
level avian species (e.g., Fernie et al. 2001a, 2001b; Baron et
al. 1995; Boudewijn and Dirksen 1995). The results of this
study can be incorporated into future pursuits of immunology
and other health-related issues in wild birds at risk of exposure
to environmental contaminants, as well as captive birds under
experimental conditions.
Acknowledgments. We thank Ian Ritchie for his excellent manage-
ment of the kestrel colony and David Bird for generous access to the
colony at the Avian Science Centre, University of McGill. K. J. Fernie
was supported by the Isabel Maria López Martı́nez Memorial Schol-
arship. We thank R. D. Dawson, J. L. Tella, and M. G. Forero for their
help in collecting blood samples from wild birds and K. Skelton for
analyzing the blood smears. Funding was provided by the Canadian
Network of Toxicology Centres (to J. E. G. Smits and G. R. Bortolotti)
and the Natural Sciences and Engineering Research Council of Canada
(to G. R. Bortolotti).
References
Ankley GT, Giesy JP (1998) Endocrine disruptors in wildlife: a weight
of evidence perspective. In: Kendall R, Dickerson R, Suk W,
Giesy J (eds) Principles and processes for assessing endocrine
disruption in wildlife. SETAC, Pensacola, FL, pp 349 –368
Ankley G, Mihaich E, Stahl R, Tillit D, Colburn T, et al. (1998)
343
Overview of a workshop on screening methods for detecting
potential (anti-) estrogenic/androgenic chemicals in wildlife. En-
viron Toxicol Chem 17:68 – 87
Baron MG, Galbraith H, Beltman D (1995) Comparative reproductive
and developmental toxicology of PCBs in birds. Comp Biochem
Physiol C 112:1–14
Besedovsky H, Sorkin E (1977) Network of immune-endocrine inter-
actions. Clin Exp Immunol 27:1–12
Besedovsky HO, del-Rey A, Klusman I, Furukawa H, Monge-Arditi
G, Kabiersch A (1991) Cytokines as modulators of the hypothal-
amus-pituitary-adrenal gland axis. J Steroid Biochem Molec Biol
40:613– 618
Bishop CA, Boermans HJ, Ng P, Campbell GD, Struger J (1998)
Health of tree swallows (Tachycineta bicolor) nestlings in pesti-
cide-sprayed apple orchards in Ontario, Canada. Immunological
parameters. J Toxicol Environ Health A 55:531–559
Bosveld ATC, Van Den Berg M (1994) Effects of polychlorinated
biphenyls, dibenzo-p-dioxins,
Boudewijn TJ, Dirksen S (1995) Impact of contaminants on the
breeding success of the cormorant Phalacrocorax carbo sinensis
in the Netherlands. Ardea 83:325–338 and dibenzofurans on fish-
eating birds. Environ Rev 2:147–166
Campbell TW (1995) Avian hematology and cytology, 2d ed. Iowa
State University Press, Ames, IA, pp 3–19
Chopra IH (1972) A radioimmunoassay for measurement of thyroxine
in unextracted serum. Clin Endocrinol Metab 34:938 –947
Chu Q, Wu W, Cook ME, Smalley EB (1995) Induction of tibial
dyschondroplasia and suppression of cell-mediated immunity in
chickens by Fusarium oxysporum grown on sterile corn. Avian
Dis 39:100 –107
Cremaschi GA, Gorelik G, Klecha AJ, Lysionek AE, Genaro AM
(2000) Chronic stress influences the immune system through the
thyroid axis. Life Sci 67:3171–3179
Daynes RA, Meikle AW, Araneo BA (1991) Locally active steroid
hormones may facilitate compartmentalization of immunity by
regulating the types of lymphokines produced by helper T cells.
Res Immunol 142:40 – 45
Drouillard KG, Fernie K, Smits JE, Bortolotti GR, Bird DM, Norstrom
RJ (2001) Bioaccumulation and toxicokinetics of 42 PCB conge-
ners in American kestrels (Falco sparverius). Environ Toxicol
Chem 20:2514 –2522
Elliott JE, Kennedy SW, Lorenzen A (1997) Comparative toxicity of
polychlorinated biphenyls to Japanese quail (Coturnix c. japonica)
and American kestrels (Falco sparverius). J Toxicol Environ
Health 51:67–75
Erf GF (1993) Immune development in young-adult C.RF-hyt mice is
affected by congenital and maternal hypothyroidism. Proc Soc
Biol Med 204:40 – 48
Erf GF, Marsh JA (1989) Effect of dietary triiodothyronine on mixed-
lymphocyte responsiveness in young male chickens. Dev Comp
Immunol 13:177–186
Fairbrother A (1994) Immunotoxicology of captive and wild birds. In:
Kendall R, Lacher TE (eds) Wildlife toxicology and population
modelling. Lewis Publishers, Boca Raton, FL, pp 251–261
Fernie KJ, Smits JE, Bortolotti GR, Bird DM (2001a) Reproductive
success of American kestrels exposed to dietary polychlorinated
biphenyls. Environ Toxicol Chem 20:776 –781
Fernie KJ, Smits JE, Bortolotti GR, Bird DM (2001b) In ovo exposure
to polychlorinated biphenyls: reproductive effects on second-gen-
eration American kestrels. Arch Environ Contam Toxicol 40:544 –
550
Fowles J, Fairbrother A, Fix M, Schiller S, Kerkvliet NL (1993)
Glucocorticoid effects on natural and humoral immunity in mal-
lards. Dev Comp Immunol 17:165–177
344
Fowles J, Fairbrother A, Kerkvliet NL (1997a) Effects of induced
hypo- and hyperthyroidism on immune function and plasma bio-
chemistry in mallards (Anas playrhynchos). Comp Biochem
Physiol 118:213–220
Fowles J, Fairbrother A, Trust KA, Kerkvliet NL (1997b) Effects of
Aroclor 1254 on the thyroid gland, immune function and hepatic
cytochrome P450 activity in mallards. Environ Res 75:119 –129
Gilbertson M, Kubiak T, Ludwig J, Fox G (1991) Great Lakes embryo
mortality, edema and deformities syndrome (GLE-MEDS) in co-
lonial fish-eating birds: similarity to chick edema disease. J Toxi-
col Environ Health 33:455–520
Goto N, Kodama H, Okada K, Fujimoto Y (1978) Suppression of
phytohemagglutinin skin response in thymectomised chickens.
Poultry Sci 57:246 –250
Grasman KA, Scanlon PF (1995) Effects of acute lead ingestion and
diet on antibody and T-cell mediated immunity in Japanese quail.
Arch Environ Contam Toxicol 28:161–167
Grasman KA, Scanlon P, Fox GA (2000) Geographic variation in
hematological variables in adult and prefledgling herring gulls
(Larus argentatus) and possible associations with organochloring
exposure. Arch Environ Contam Toxicol 38:244 –253
Grasman KA, Fox G, Scanlon P, Ludwig J (1996) Organochlorine-
associated immunosuppression in pre-fledgling Caspian terns and
herring gulls from the Great Lakes: an ecoepidemiological study.
Environ Health Perspect 104(suppl 4):829 – 842
Gross WB, Siegel HS (1983) Evaluation of the heterophil/lymphocyte
ratio as a measure of stress in chickens. Avian Dis 27:972–979
Hoffman DJ, Melancon MJ, Klein PN, Rice CP, Eisemann JD, Hines
RK, Spann JW, Pendleton GW (1996) Developmental toxicity of
PCB 126 (3,3⬘,4,4⬘,5-pentachlorobiphenyl in nestling American
kestrels (Falco sparverius). Fund Appl Toxicol 34:188 –200
Hungerford DA, Donelly AJ, Nowell PC, Beck S (1959) The chro-
mosome constitution of a human phenotypic intersex. Am J Hum
Genet 11:215–235
Kean RP, Lamont SJ (1994) Effect of injection site on cutaneous
basophil hypersensitivity response to phytohemagglutinin. Poultry
Sci 73:1763–1765
Kjeld JM, Kuku SF, Diamant L, Fraser TR, Joplin GF, Mashiter K
(1975) Production and storage of [125I]thyroxine and [125I]triiodo-
thyronine of high specific activity. Clin Chim Acta 61:381–389
Klecha AJ, Genaro AM, Lysionek AE, Caro RA, Coluccia AG, Cre-
maschi GA (2000) Experimental evidence pointing to the bidirec-
tional interaction between the immune system and the thyroid
axis. Int J Immunopharmacol 22:491–500
Kozie KD, Anderson RK (1991) Productivity, diet and environmental
contaminants in bald eagles nesting near the Wisconsin shoreline
of Lake Superior. Arch Environ Contam Toxicol 20:41– 48
Laudert E, Sivanandan V, Halvorson D (1993) Effect of an H5N1
avian influenza virus infection on the immune system of mallard
ducks. Avian Dis 37:845– 853
Leighton FA (1986) Clinical, gross and histological findings in herring
gulls and Atlantic puffins that ingested Prudhoe crude oil. Vet
Pathol 23:254 –263
Lowenthal JW, Connick TE, McWaters PG, York JJ (1994) Develop-
ment of T cell immune responsiveness in the chicken. Immunol
Cell Biol 74:115–122
J. E. Smits et al.
Mandal A, Chakraborty S, Lahiri P (1986) Hematological changes
produced by lindane (␥-HCH) in six species of birds. Toxicology
40:103–111
Norušis M (1993) The SPSS guide to data analyses for SPSS/PC⫹.
Prentice Hall, Englewood Cliffs, NJ
Perez-Castro C, Paez PM, Keller EA, Artz E (1998) The specific
neuroendocrine pro critically involved in an adequate T-depen-
dent immune response. Med Buenos Aires 58:189 –193
Poland A, Knutson J (1982) 2,3,7,8-Tetrachlorodibenzo-p-dioxin and
related halogenated aromatic hydrocarbons: examination of the
mechanism of toxicity. Ann Rev Pharmacol Toxicol 22:517–554
Rolland RM (2000) A review of chemically-induced alterations in
thyroid and vitamin A status from field studies of wildlife and fish.
J Wildl Dis 36:615– 635
Rook GAW, Hernandez-Pando R, Lightman SL (1994) Hormones,
peripherally activated prohormones and regulation of the Th1/Th2
balance. Immunol Today 15:301–303
Ross PS (1995) Seals, pollution and disease: environmental contami-
nant-induced immunosuppression. PhD thesis, Universitiet Utre-
cht, Netherlands
Schwartz TR, Tillitt DE, Feltz KP, Peterman PH (1993) Determination
of mono-and non-o,o⬘-chlorine substituted polychorinated biphe-
nyls in aroclors and environmental samples. Chemosphere 26:
1443–1460
Smits JE, Bortolotti GR (2001) Antibody mediated immunotoxicity in
American kestrels (Falco sparverius) exposed to polychlorinated
biphenyls. J Toxicol Env Health A 62:101–110
Smits JE, Williams TD (1999) Validation of immunotoxicology tech-
niques in passerine chicks using oil sands tailings water. Ecotoxi-
col Environ Safety 44:105–112
Smits JEG, Blakley BR, Wobeser GA (1996a) Immunotoxicity studies
in mink (Mustela vison) chronically exposed to dietary bleached
kraft pulp mill effluent. J Wildl Dis 32:199 –208
Smits JEG, Haines DM, Blakely BR, Wobeser GA (1996b) Enhanced
antibody responses in mink (Mustela vison) exposed to dietary
bleached-kraft pulp mill effluent. Environ Toxicol Chem
15:1166 –1170
Smits JE, Bortolotti GR, Tella JL (1999) Simplifying the phytohemag-
glutinin skin testing technique in studies of avian immunocompe-
tence. Func Ecol 13:567–572
Tella JL, Bortolotti GR, Dawson RD, Forero MG (2000) The T-cell-
mediated immune response and return rate of fledgling American
kestrels are positively correlated with parental clutch size. Proc
Royal Soc Lond B 267:891– 895
Van der Zijpp AJ (1983) The effect of genetic origin, source of antigen
and dose of antigen on the immune response of cockerels. Poultry
Sci 62:205–211
Wiemeyer SN, Lincer JL (1987) The use of kestrels in toxicology. In:
Bird DM, Bowman R (eds) The ancestral kestrel, raptor research
reports no. 6. Raptor Research Foundation and McGill University,
pp 165–178
Williamson RA, Davison TF, Payne LN (1990) Effects of thyroid
hormones on humoral and cell mediated immune response in the
domestic fowl (Gallus domesticus). Dev Comp Immunol 14:305–
318