Magnification and toxicity of PCBs, PCDDs, and PCDFs in upriver-migrating Pacific salmon.
ABSTRACT The depletion of lipids associated with pre-spawning migration of Pacific salmon has the potential to magnify concentrations of hydrophobic organic contaminants (HOCs), which elevates risk of toxic effects. We present data from a field study of sockeye salmon (Oncorhynchus nerka) migrating to spawn in Great Central Lake, BC, which demonstrate that pre-spawning migration causes a magnification of PCB, PCDD, and PCDF concentrations in female gonads (1.9-2.5-fold), female soma (3.4-5.6-fold), and male soma (5.6-9.7-fold). We further develop a model of prespawning migration chemical magnification for sockeye salmon stocks as a function of migration distance. This model is shown to be consistent with available empirical data on pre-spawning magnification and predicts magnification factors ranging between 1.4 and 7.9 in gonad and between 1.6 and 10.4 in soma in seven Pacific salmon stocks in British Columbia. Post-migration (prespawning) toxic equivalent dioxin concentrations in roe were measured to be approximately 3 pg/g lipid in salmon from the Great Central Lake sockeye stock and estimated to range between 1.5 pg/g lipid for the shortest-migrating stocks and 7 pg/g lipid for the longest-migrating stocks. Concentrations in certain stocks approach or exceed the concentration of 3 pg/g lipid associated with 30% egg mortality in Oncorhynchus mykiss. This indicates the potential for population-level effects of current contaminant levels. It also suggests that historic contaminant concentrations, which were greater than current concentrations, may have contributed significantly to the decline of certain Pacific salmon stocks in British Columbia.
- SourceAvailable from: AlberTinka Murk[show abstract] [hide abstract]
ABSTRACT: During the development of an embryo into a juvenile, the physiology and behavior of a fish change greatly, affecting exposure to and uptake of environmental pollutants. Based on experimental data with sole (Solea solea), an existing bioaccumulation model was adapted and validated to calculate the development of concentrations of persistent organic pollutants in the tissue of developing fish. Simulation revealed that toxic tissue concentrations of pollutants with log octanol-water partition ratio (K(OW)) > 5 peak at the moment when the larvae become free-feeding, when the lipid reserves are depleted. This may explain the delayed effects observed in fish early-life-stage experiments with exposed eggs. In the field, eggs can be exposed through maternal transfer to adult pollutant tissue concentrations, which will increase in the larva to peak tissue concentrations, exceeding those of the adult fish. The results demonstrate the risk of underestimating the effects of lipophilic persistent organic pollutants with log K(OW) > 5 in short-term, early-life-stage fish tests and underscore the importance of maternal transfer as an exposure route in the field situation.Environmental Toxicology and Chemistry 04/2012; 31(6):1381-90. · 2.62 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Fish and piscivorous bird eggs collected in 2003 from Lake Maggiore (Italy), were analysed for PCB and DDT contamination. Lake Maggiore has been severely polluted by DDT through production of the pesticide within its catchment. Although agricultural application of DDT was banned in Italy in the 1978, industrial production continued until 1996, with enough contamination of water and soil for serious bioaccumulation in the lake biota. PCB and DDT concentrations in a whitefish (Coregonus macrophthalmus Nusslin 1882) were seen to be dependent on season and fish age, but not on sex. The average increase of the lipid-normalised concentration of DDTs and PCBs was two-fold across season and also across age, resulting in an overall increase of four fold. The seasonal variation was related to the eco-physiological cycle of the fish and to the contamination dynamic of the lake, while the effect of the fish age was explained on the base of biomagnification-related mechanisms. A fugacity model was applied to predict the age-dependent bioaccumulation potential of PCBs, whose concentrations were rather stable in recent years in the lake. Predicted values for compounds with negligible biotransformation were in good agreement with experimental data (calculated vs. experimental mean difference of 14%), and a relationship between the increase of experimental age-dependent concentration and K ow was observed. The good correspondence between the predicted and the measured values for most PCB congeners confirmed the general inability of fishes to biotransform these compounds. On the contrary, the importance of biotransformation processes was recognised in birds; eggs of a fish eating bird (Podiceps cristatus) from the same area selectively bioaccumulated p,p′-DDE. For PCBs, congener 149 appears to be completely metabolized by the bird species, and congeners 95, 101, 132, 151 and 174 were reduced as well. The role of the meta–para free position on at least one phenyl ring of PCB congeners in biotransformation processes was confirmed.Water Air and Soil Pollution 04/2012; 197(1):193-209. · 1.75 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: We assessed the transportation tendency of dioxins and predict locations at high risk for dioxin pollution. A new parameter, the compartment distribution coefficient DC, was created to account for the tendency of dioxins to preferentially accumulate in particular compartments. It was obtained by a model using levels of polychlorinated dibenzo-p-dioxins/dibenzofurans (PCDD/Fs) in four countries: Japan, the United States, the United Kingdom, and Australia. The comparison with the temporal and spatial variation of DC indicated whether the location release or long-range transportation caused the changes. This study showed that PCDD/Fs have the greatest tendency to remain in soil among studied media. A higher DC value in Australia may indicate that this location is a potential future reservoir source of dioxins. KeywordsPolychlorinated dibenzo-p-dioxins/dibenzofurans (PCDD/Fs)–Compartment distribution coefficient–Transportation tendency–Reservoir sourcesEnvironmental Chemistry Letters 04/2012; 9(2):267-271. · 1.62 Impact Factor
Magnification and Toxicity of PCBs,
PCDDs, and PCDFs in
Upriver-Migrating Pacific Salmon
A D R I A N M . H . D E B R U Y N ,†
M I C H A E L G . I K O N O M O U ,‡A N D
F R A N K A . P . C . G O B A S *, †
School of Resource & Environmental
Management, Simon Fraser University,
Burnaby, British Columbia, Canada V5A 1S6, and
Contaminant Sciences, Institute of Ocean Sciences, Fisheries
and Oceans Canada (DFO), 9860 West Saanich Road,
Sidney, British Columbia, Canada V8L 4B2
The depletion of lipids associated with pre-spawning
migration of Pacific salmon has the potential to magnify
which elevates risk of toxic effects. We present data
from a field study of sockeye salmon (Oncorhynchus nerka)
migrating to spawn in Great Central Lake, BC, which
demonstrate that pre-spawning migration causes a
magnification of PCB, PCDD, and PCDF concentrations in
spawning migration chemical magnification for sockeye
salmon stocks as a function of migration distance. This
model is shown to be consistent with available empirical
data on pre-spawning magnification and predicts
magnification factors ranging between 1.4 and 7.9 in
gonad and between 1.6 and 10.4 in soma in seven Pacific
salmon stocks in British Columbia. Post-migration (pre-
measured to be approximately 3 pg/g lipid in salmon
from the Great Central Lake sockeye stock and estimated
to range between 1.5 pg/g lipid for the shortest-
migrating stocks and 7 pg/g lipid for the longest-migrating
stocks. Concentrations in certain stocks approach or
exceed the concentration of 3 pg/g lipid associated with
30% egg mortality in Oncorhynchus mykiss. This indicates
the potential for population-level effects of current
concentrations, which were greater than current con-
of certain Pacific salmon stocks in British Columbia.
Anadromous Pacific salmon are an extreme example of
energetic investment in reproduction. Maturing adults stop
an osmotically unfavorable freshwater environment, and
migrate long distances inland to spawning grounds (1, 2).
Females invest heavily in gonadal development, with the
mature ovaries and roe comprising about 20% of the fresh
weight of the spawning adult. Males invest in conspicuous
secondary sexual characteristics (3). The arduous migration
and associated physiological changes are accomplished at
the cost of 50-90% of the adult’s somatic energy stores (3).
protein weight declines by up to 50% (2, 4).
The severe depletion of lipid associated with spawning
migration may expose Pacific salmon to elevated internal
concentrations of hydrophobic organic contaminants (HOCs).
If HOCs accumulated during the salmon’s marine growth
phase cannot be rapidly transformed or eliminated during
upriver migration, the loss of lipid content can cause a
contaminants in the organism’s tissues. In the single
published study of this phenomenon (5), lipid-normalized
concentrations of total PCBs and DDT were reported to
increase in flesh and gonads of sockeye salmon (Oncorhyn-
chus nerka) migrating along the Copper River, AK. These
increases were on the order of 2-5-fold, concomitant with
a modest depletion of muscle lipid from 5.5% to 2.2%. Other
studies of migrating salmon have reported much greater
declines in lipid content (e.g., ref 4), and thus some stocks
may experience greater increases in HOC concentrations.
Reported HOC concentrations in wild, pre-migration
Pacific salmon are near thresholds for some toxic effects (6).
It is therefore possible that lipid depletion during migration
elevates these concentrations to levels that pose a threat to
developing embryos, which are generally more sensitive to
HOCs than adult salmonids (7, 8). Certain Pacific salmon
since the 1960s (9) when HOC concentrations were greater
than current levels. It is possible that HOCs contributed to
declines in certain Pacific salmon populations in British
Columbia. Understanding the relationship between HOC
concentrations and the health of salmon populations is
important in the light of growing concern over the rising
levels of several HOCs with a dioxin-like toxicity such as
polybrominated diphenyl ethers (PBDEs) (10, 11).
In this paper, we present (i) a field study of the effects of
lipid depletion on the internal concentrations of polychlo-
rinated biphenyls (PCBs), dibenzo-p-dioxins (PCDDs), and
of chemical magnification in soma and gonads of upriver-
data and the model to assess the risk of dioxin-like toxicity
in several British Columbia salmon stocks.
Uptake from food and elimination to feces are the major
hydrophobic organic contaminants (HOCs) (12), but these
gills is relatively unimportant because of the low concentra-
tions in water (14), and gill elimination half-lives for very
and PCDFs are on the order of years to decades (15-17).
Half-lives for metabolic transformation of many HOCs in
fish are also long and on the order of a few months to many
years for even the most metabolizable PCBs, PCDDs, and
PCDFs (18, 19). The duration of spawning in Pacific salmon
is typically on the order of weeks to months and is thus too
or metabolic transformation for many chemicals. With
respect to such nonmetabolizable, slowly eliminated HOCs
* Corresponding author phone: (604)291-5928 or (604)268-6813;
fax: (604)291-4968; e-mail: email@example.com.
†Simon Fraser University.
‡Institute of Ocean Sciences, Fisheries and Oceans Canada.
Environ. Sci. Technol. 2004, 38, 6217-6224
10.1021/es049607w CCC: $27.50
Published on Web 09/14/2004
2004 American Chemical SocietyVOL. 38, NO. 23, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY96217
and assuming that there is no significant uptake of HOCs
are effectively closed systems, and internal HOC concentra-
of ovaries and roe. This growth acts to dilute HOCs already
present in the gonads and promotes the redistribution of
on the kinetics of this internal redistribution (20). Chemical
by a kinetic (unsteady-state) model. In this model, the total
mass of chemical in each compartment i of the fish (Mi) can
be expressed as the product of lipid equivalent-normalized
concentration (Ci) and the mass of equivalent lipid, which
weight of compartment i (Wi): Mi) CiEiWi. Concentrations
the highest storage capacity for HOCs. Because organisms
and tissues can vary in lipid content, lipid-normalized
concentrations are a convenient measure to express the
relative chemical and toxicological activity of HOCs among
organisms. Lipid equivalent-normalized concentrations are
Net passive movement of a chemical (e.g., between soma
than wet weight-based concentrations (21). When lipid
content is low, however, nonlipid organic matter can
contribute substantially to the storage capacity of a tissue
(ash e 2% in whole fish; carbohydrate < 1%; 2, 4). We can
calculate the lipid equivalent fraction of each tissue i as Ei
fractions (unitless), and (Zprotein/Zlipid) is the ratio of sorptive
protein resembles lipid in its ability to sorb HOCs, is
capacity of lipids.
Figure 1 illustrates a simple model of the distribution of
chemical between a migrating (non-feeding and non-
defecating) fish and the ambient water. For this model, we
consider two internal compartments: gonad (including
gametes) and soma. The purpose of this model is to forecast
concentrations in these two compartments throughout
migration and thus to estimate the exposure of early life
stage salmon to HOCs accumulated by the adult fish.
A dynamic mass balance for each compartment can be
in soma, and
in gonad. CS and CG are the lipid equivalent normalized
concentrations (g/kg equivalent lipid) in the soma and
gonads, respectively, and CWis the chemical concentration
in the water (g/L). The subscripts S, G, and W refer to soma,
gonad, and ambient water, respectively. W is wet weight,
and E is lipid content of the soma or gonad. kWS is a
clearance rate for uptake of chemical from ambient water to
soma (L/d). kSWis a first-order rate constant for elimination
of chemical from soma to ambient water (d-1). kSGand kGS
soma and gonad (d-1). kSMand kGMare first-order rates of
chemical biotransformation in the soma and gonad, respec-
tively. If, for simplicity, we assume that the concentration in
water (CW) is constant throughout migration, we can obtain
the rate of change of lipid equivalent-normalized concentra-
product d(WiEiCi)/dt and rearranging:
where MGis WGEGCGand
where MS is WSESCS. Here dW/W dt is the relative rate of
rate of change of lipid equivalent fraction of soma or gonad;
these terms describe the change in size and composition of
the two tissues over time as a fraction of the tissue’s weight
or lipid equivalent content. Upriver migration is a period of
relative rates of change are typically on the order of several
percent per day.
It is possible to simplify this model further because some
of the rate constants are very slow relative to dW/W dt and
dE/E dt. Respiratory elimination rate constants (kSW) can be
PCDDs, and PCDFs are highly hydrophobic substances (log
KOW g 6) and, therefore, have kSW in adult salmon on the
order of 10-3-10-4d-1(15-17). Branchial uptake rate
constants (kWS) are expected to be much higher (23), but
HOC concentrations in water are very low, so uptake from
water may be assumed to be negligible on the time scale of
migration. Rates of metabolic transformation (kSMand kGM)
to about 10-3d-1for some PCDDs and PCDFs (18, 19).
If rates of internal redistribution (kSG and kGS) are also
very slow relative to dWi/Widt and dEi/Eidt, a fish behaves
essentially as two isolated internal compartments, and each
final lipid equivalent-normalized Ciwill be a function only
of changes in Wiand Ei:
which can be integrated to give post-/pre-migration mag-
nification factors (MFs) in gonad and soma:
kWSCW+ kGSWGEGCG- (kSW+ kSG+ kSM)WSESCS(1)
d(WGEGCG)/dt ) kSGWSESCS- (kGS+ kGM)WGEGCG (2)
are rates of growth and rates of change of lipid equivalent content
of soma and gonad.
(kSW+ kSG+ kSM+
62189ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 23, 2004
which expresses the degree of chemical magnification that
has taken place as a result of migratory lipid depletion.
If kSG and kGS are very large relative to dWi/Wi dt and
dEi/Eidt, the chemical is quickly distributed between soma
and gonads, and lipid equivalent concentrations in soma
and gonads are equal (i.e., CG) CS). In that case, MF will be
a function of changes in W and E in both compartments:
(see Supporting Information for complete derivations). For
intermediate values of kSGand kGS, the internal distribution
of chemical will fall between these two extremes, and final
described in eqs 3 and 4.
Materials and Methods
during migration, we collected sockeye salmon migrating to
Great Central Lake (elevation 82 m) on Vancouver Island,
BC. Nine pre-migration fish (6 females, 3 males) were
captured by seining in Barkley Sound between June 14 and
June 22, 1995. Five post-migration fish (2 females, 3 males)
were captured in Robertson Creek about 1 km upstream of
Great Central Lake, between October 2 and October 7, 1995.
The latter fish had migrated 37 km up Alberni Inlet and 26
The duration of this migration for an individual fish was
estimated to be 3 weeks. All fish were age 4-5, averaging 61
muscle, liver (both sexes), and roe (females) were taken for
analysis of moisture content, lipid content, and concentra-
tions of PCBs, PCDDs, and PCDFs.
and spiked; a mixture of13C-labeled PCDDs, PCDFs, and
PCBs (internal standards as supplied by Cambridge Isotope
Laboratories; Andover, MA) was then mixed with Na2SO4in
and extracted with CH2Cl2/hexane (1:1 v/v). The extract was
reduced to a few milliliters by rotary evaporation and
up by (i) gel permeation chromatography; (ii) silica gel
silica); (iii) activated alumina chromatography; and (iv)
carbon fiber chromatography. Four fractions were collected
from the carbon fiber column. Fraction I contained the di-
ortho-PCBs, fraction II contained the mono-ortho-PCBs,
fraction III contained the coplanar-PCBs, and fraction IV
contained the PCDDs and PCDFs. Each fraction was con-
13C-labeled method performance standards prior to instru-
mental analysis. Details on the procedures, solvents, and
standards used are discussed elsewhere (24).
HOC analyses were conducted by gas chromatography/
high-resolution mass spectrometry (GC-HRMS). The GC-
HRMS system used was VG AutoSpec-S (Micromass,
Manchester, UK) HRMS equipped with a HP 5890 series II
The GC column used was DB-5 (60 m × 0.25 mm i.d., 0.1 µm
injected in splitless mode, at an injector temperature of 282
°C. The temperature program used under constant pressure
(ca. 140 kPa for a 60 m column) was as follows: for the
coplanar- and mono-ortho-PCBs, hold at 80 °C for 2 min, 8
°C/min to 150 °C, 4 °C/min to 285 °C; for the di-ortho-PCBs:
hold at 80 °C for 2 min, 8 °C to 150 °C, 4 °C to 300 °C, hold
for 2 min. The GC/HRMS interface temperature was 285 °C,
analyses were performed with the HRMS operating in the
power and acquiring data under SIM conditions. Two ions,
were calculated using mean relative response factors de-
termined from calibration standard runs. The instrumental
analyses conditions (for all analyses; PCBs and PCDD/Fs)
and the criteria used for identification and quantitation are
described in detail elsewhere (24).
Lipid determinations were performed gravimetrically.
Approximately 5 g (wet weight) tissue samples were ho-
mogenized by grinding with anhydrous Na2SO4, exracted
using 100 mL of 1:1 DCM:hexane, reduced by turboevapo-
ration to a few milliliters, dried at 40 °C overnight, and
subsequently weighed. Moisture content was determined
by drying the sample in a vented oven at 105 °C for 48 h and
weighing the sample before and after drying.
Model Parameterization and Scenarios. To explore the
we parameterized eqs 6 and 7 in three ways: (i) we used
parameters measured in the present study to predict
magnification factors (MFs, the ratio of post-migration to
pre-migration HOC concentrations) for Great Central Lake
sockeye (this allowed the model to be tested against
(ii) we obtained complete sets of parameters from the
literature to predict MFs for several well-studied stocks of
Pacific salmon; and (iii) we used a more extensive survey of
the literature to develop general relationships between the
model parameters and migration distance and then used
these general relationships to predict how MFs vary with
migration distance among all stocks of sockeye salmon. We
then compared the predicted MFs for Great Central Lake
sockeye to measured MFs for PCB, PCDD, and PCDF
congeners, and we compared the predictions of the general
model to MFs measured here and reported elsewhere.
We obtained pre- and post-migration tissue weight and
proximate composition (lipid, protein, and water content)
parameters from the literature for 14 stock-years of Pacific
salmon, including 9 stock-years of sockeye (O. nerka; 3-5,
(O. keta; 31), and chinook (O. tshawytscha; 32, 33) salmon.
Ref 4 reported the most complete set of parameters,
salmon. Because Pacific salmon store about half their lipid
in visceral, subcutaneous, and skeletal deposits (34, 35), we
re-analyzed Gilhousen’s data (4) to establish a relationship
between proximate composition of muscle tissue (reported
model) (Table S1 in the Supporting Information).
Specific rates of change of tissue weight and lipid
equivalent fraction were calculated from pre- and post-
migration values and the reported duration of migration,
assuming first-order kinetics: dWi/Widt ) ln(Wi,post/Wi,pre)/
t; dEi/Eidt ) ln(Ei,post/Ei,pre)/t. Sufficient data were available
in the literature to calculate MFs from stock-specific pa-
rameters for 7 stock-years of females and 5 of males. We
distance on tissue weight and proximate composition
parameters among all 14 stock-years and used these rela-
that were not correlated with migration distance (MD), we
were species-specific (e.g., mean wet weight varies greatly
among species), we parameterized eqs 6 and 7 as a function
of migration distance for sockeye salmon only. Sockeye
salmon exhibit the greatest range of migration distances,
VOL. 38, NO. 23, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY96219
and an excellent collection of model input parameters is
available for this species.
kGS) are unknown, we examined a range of possible values.
to pre-migration concentrations) that would be predicted if
internal transport rates were very fast (i.e., eq 7 assumingCG
) CS) or essentially nil (i.e., eq 6 assuming that gonads and
soma can be viewed as isolated compartments). These
provide the boundary conditions for the range of MFs that
may be observed. We also estimated a range of likely values
for kSG and kGS by assuming that blood is the major
transporting medium for soma-gonad exchange. Rate
constants for advective transport in blood can be ap-
proximated by kSG) QBEB/VSESand kGS) QBEB/VGEG, where
QBis the daily flux of blood through the gonad, estimated as
a function of gonad size (37), and EBis the lipid equivalent
during migration (37). QBincreases with gonad volume VG
as the gonads develop; EB, ES, and EGdecrease as the lipid
contents of blood and other tissues decline; and kSGand kGS
are therefore relatively constant throughout migration (cal-
culated ranges: kSG) 0.5-5 d-1; kGS) 10-40 d-1). Ref 36
found that their pharmacokinetic model was improved by
for resistance to diffusion. We therefore considered a range
of estimated values of kSG and kGS (0.1 and 0.01 of the
Toxicology. Finally, we assessed the toxicological sig-
nificance of post-migration concentrations of HOCs mea-
sured in Great Central Lake sockeye salmon and predicted
by our model for several other Pacific salmon stocks. We
calculated taxon-specific toxic equivalent concentrations
ortho- and coplanar) PCBs for post-migration salmon roe
based on toxic equivalency factors (TEFs) from ref 38. To do
this, measured congener concentrations were multiplied by
their TEFs to determine a toxic equivalent concentration
(TEC) for each congener. The TECs were then summed
assuming an additive mode of toxic action to calculate the
We then compared total TEQ values to a roe concentration
of 0.30 pg/g wet weight (8) or 3 pg/g lipid associated with
30% egg mortality in O. mykiss, assuming that differences in
This toxicological end point was selected because of the
similiarity in the mode of exposure (i.e., maternal transfer)
between wild and laboratory manipulated fish. In studies
hydrocarbons through injection, TEQ concentrations of 0.3
pg/g wet weight caused approximately 5% egg mortality in
curve dose-response curve is less steep than the dose-
far below the threshold effect concentrations for sac fry
mortality produce significant levels of egg mortality.
Results and Discussion
Field Study. Table 2 and Figure 2 demonstrate that post-
to Great Central Lake were significantly greater than pre-
sockeye salmon captured in Barkley Sound (pre-migration)
exhibited total PCB concentrations (∑PCB) of 17-27 ng/g
lipid, whereas ∑PCB concentrations in fish captured near
the end of the spawning migration were 28-145 ng/g lipid.
The total PCB body burden of post-migration Great Central
Lake sockeye was 99% of the pre-migration total burden of
or transformationduring migration. For ∑PCDDs, the post-
TABLE 1. Parameters Used To Model Effect of Migration Distance (MD, 1000s of km) on Magnification of HOC Concentrations in
Upriver-Migrating Sockeye Salmon
present studyvalue or expression used
2.32 ( 0.313 (1.58-2.63)
2.53 ( 0.410 (1.68-2.92)
1.99 ( 0.199 (1.77-2.42)
2.29 ( 0.407 (1.57-2.87)
0.130 ( 0.0351 (0.088-0.174)
0.0707 ( 0.0143 (0.0568-0.0905)
0.361 ( 0.0744 (0.268-0.493)
0.0514 ( 0.0205 (0.0332-0.0890)
0.113 ( 0.0343 (0.0647-0.178)
0.119 ( 0.0283 (0.0773-0.172)
0.0199 ( 0.00980 (0.0093-0.040)
0.0207 ( 0.0126 (0.0116-0.0505)
0.134 ( 0.0408 (0.0510-0.155)
0.0160 ( 0.0080 (0.0040-0.020)
0.0723 ( 0.0268 (0.0410-0.105)
0.0258 ( 0.00320 (0.0210-0.0275)
0.143 ( 0.0388 (0.0647-0.208)
0.152 ( 0.0259 (0.117-0.202)
0.0479 ( 0.0120 (0.0332-0.0707)
0.0489 ( 0.0138 (0.0359-0.0804)
0.177 ( 0.0327 (0.111-0.195)
0.0411 ( 0.00760 (0.0322-0.0509)
0.110 ( 0.0290 (0.0867-0.150)
0.0502 ( 0.0040 (0.0456-0.0536)
2.30 ( 0.221
2.87 ( 0.359
2.03 ( 0.0739
2.22 ( 0.226
-0.0701MD + 0.181 (r2) 0.51; p ) 0.047)
-0.0318MD + 0.0949 (r2) 0.76; p ) 0.024)
0.0623MD + 0.0707 (r2) 0.48; p ) 0.026)
0.0597MD + 0.0779 (r2) 0.77; p ) 0.0042)
0.0641MD + 0.0998 (r2) 0.40; p ) 0.049)
0.0538MD + 0.115 (r2) 0.74; p ) 0.0059)
0.280 ( 0.140
10 0.0725 ( 0.0141
0.0773 ( 0.0142
0.0118 ( 0.0008
0.0121 ( 0.0056
0.149 ( 0.0297
0.103 ( 0.0267
100.114 ( 0.0122
0.120 ( 0.0127
0.0518 ( 0.0027
0.0529 ( 0.006
0.192 ( 0.0286
0.156 ( 0.0234
aValues are mean ( SD (range in parentheses) of n reported stock-year means, including the present study.bWTis total body weight; WS)
WT- WG.cEstimated from regression.
TEQ )∑(TECi) )∑(TEFi‚Ci)(8)
62209ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 23, 2004
and for ∑PCDFs, 68% of the pre-migration burden was
present in fish after migration, indicating a greater capacity
to PCBs. Congener-specific data are presented in the Sup-
Lipid-normalized MFs were on the order of 1.9-2.5 for
which may reflect differences in their metabolizability.
Observed MFs for PCDDs and PCDFs were generally lower
levels declined substantially during migration. This caused
a magnification of the lipid-normalized concentration of
these contaminants in the fish.
This observed magnification of HOC concentrations in
Great Central Lake sockeye salmon was accompanied by a
by 18% on average, partly offset by a 1.6-fold increase in
on average (testes were not weighed but are <5% of total
weight; 4). Somatic lipid content declined by 84% in both
Model Predictions. Literature data (Figure 3, Table 3)
somatic wet weight declined during upriver migration by
24% on average (all species combined), partly offset by a
salmon show the greatest decline in somatic weight (mean
30%) and the greatest increase in gonad weight (mean 2.8-
little through migration. Figure 3 and Table 1 document a
but post-migration values are uniformly low (∼2%) among
TABLE 2. Concentrations of ∑PCBs, ∑PCDDs, and ∑PCDFs
Measured in Tissues of Upriver-Migrating Great Central Lake
soma (muscle) ∑PCBF
27.05 ( 8.2
24.27 ( 6.58 234.77 ( 133.66 9.7
10.72 ( 5.0856.29 ( 0.57
8.37 ( 1.0875.16 ( 43.59
12.97 ( 5.6644.24 ( 13.21
12.44 ( 6.3269.69 ( 21.31
16.63 ( 4.2364.90 ( 28.93
19.42 ( 4.65 84.90 ( 47.15
13.07 ( 4.4231.66 ( 3.12
11.06 ( 1.8245.93 ( 7.93
10.97 ( 7.74 41.33 ( 20.15
11.32 ( 3.94 56.28 ( 24.56
20.99 ( 2.99 52.57 ( 0.42
4.04 ( 1.827.87 ( 0.37
8.31 ( 2.1220.16 ( 3.17
150.23 ( 56.61 5.6
a∑PCB include 39 congeners, ∑PCDD include 7 congeners, and
post-migration (Robertson Creek) to pre-migration (Barkley Sound)
FIGURE 2. Observed magnification factors for individual PCBs
of male (open symbols) and female (closed symbols) Great Central
Lake sockeye salmon. Lines are magnification factors predicted
using measured parameters for this stock.
FIGURE 3. Mean wet weight (W, kg), lipid fraction (L, g g-1), and
protein fraction (P, g g-1) of soma and gonad for stock-years of
Pacific salmon, before and after spawning migration (see text for
literature sources). Values are for male (open symbols) and female
and chinook (diamonds) salmon. Values for male (+) and female
(×) sockeye from the present study are also shown. Diagonals are
1:1 (no difference between pre- and post-migration values).
VOL. 38, NO. 23, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY96221
species, among stock-years within species, and between
sexes, representing declines of up to 93% in lipid content or
protein declined by as much as 54%. Somatic water content
increased through migration, largely offsetting the loss of
or body shape.
Farther-migrating species and stocks of Pacific salmon
experience the most severe depletion of lipid stores (2, 40,
reported in Table 1; illustrated in Figure 4 in the Supporting
Information) indicated that farther-migrating stocks (e.g.,
Stuart Lake, BC) begin migration with somatic lipid content
nearly 3-fold greater than shorter-migrating stocks. Chum
and pink salmon, with the shortest migrations, exhibit pre-
migration lipid contents that are even lower than those of
the shortest-migrating sockeye stocks (Figure 3). The single
stock-year of chinook salmon described in the literature has
between ocean and spawning ground, but the duration of
for the energetic demand of migration (2, 42). These
observations illustrate that a greater energetic expenditure
Parameterizing eq 7 with the measured, stock-specific
parameters for Great Central Lake sockeye predicts lipid-
normalized MFs for females of 4.4 in soma and 3.1 in gonad
(Table 3). Males experience a greater depletion of somatic
lipid and are predicted to have a higher somatic MF of
by parameterizing eqs 6 and 7 for a range of pre-spawning
migration distances. Somatic MFs are predicted to increase
with migration distance in both male and female salmon
according to both models, and gonadal MFs are predicted
to increase unless soma-gonad exchange (kSG and kGS) is
Figure 5 shows that for the three stocks for which empirical
MFs are available, the observed MFs are in reasonable
agreement with the model predictions (from eq 7), which
and soma. The only exception is for the MFs for male Great
reflects the unusually low post-migration somatic lipid
to the value of 2% typical for other stocks and assumed by
the model. Observed MFs are in better agreement with MFs
predicted by eq 7, which assumes fast internal HOC
distribution between soma and gonads, than by eq 6, which
assumes a slow distribution. This is consistent with observa-
tions that lipid-normalized concentrations in mothers and
eggs tend to be similar (20, 37) and suggests that eq 7 is an
adequate model to estimate MFs in Pacific salmon stocks.
MFs predicted by eq 7 for sockeye stocks for which
empirical HOC data are not available, using stock-specific
1.6 (for soma) and 1.4 (for gonad) for the 1983 Thompson
River pink salmon females (270 km migration distance) to
10.4 and 7.9 for 1957-1958 Stuart Lake sockeye salmon
females (1016 km migration distance). The general model
for sockeye likewise predicts an increase in MFs with
migration distance among stocks (Figure 5). Sockeye stocks
with relatively short migrations (<100 km) are predicted to
experience 4-5-fold increases in lipid-normalized HOC
concentrations in the soma, while increases of up to 8-fold
may apply to farther-migrating stocks (Figure 5).
It is important to note that the predictions of our general
model should be taken only to reflect the general trend and
magnitude of MFs expected for sockeye salmon. Energetic
besides migration distance (e.g., river slope, flow regime;
Toxicological Implications. Reproduction in many spe-
cies is accompanied by a rapid depletion of lipid stores that
exposes the reproducing adult and the offspring to elevated
concentrations of HOCs (5, 44-46) and hence an elevated
risk of toxic effects (47-49). To approximate the risk of PCB,
PCDD, and PCDF concentrations to developing sockeye
salmon embryos, TEQs calculated from measured and
TABLE 3. Changes in Body Composition during Upriver Migration for Some Well-Studied Stocks of Pacific Salmon and the
Resulting Magnification of HOCs Predicted for Each Stock
factor decline in parameters
predicted magnification factorsa
allstock-year (ref) speciesMD sexgonad
Chitose River 1951 (28)
Columbia River 1908 (32, 33)
Thompson River 1983 (31)
present study sockeye
Pick Creek 1996 (3)
Adams Lake 1958 (4)
Chilko Lake 1959 (4)
Stuart Lake 1957-1958 (4)
aMFs calculated by assuming rapid soma-gonad exchange (equilibrium between tissues).bTestes weight and lipid content assumed similar
to other stocks to calculate somatic MF.
62229ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 23, 2004
predicted concentrations of PCBs, PCDDs, and PCDFs in
tissues of pre- and post-migration salmon are presented in
of spawning was dominated by concentrations of 2,3,4,7,8-
PeCDD, which provided 40% of the TEQ, and 1,2,3,7,8-
PeCDD, which provided 26% of the TEQ. Figure 6 illustrates
that while the TEQ in pre-migration sockeye salmon was
less than the 3 pg/g lipid concentration associated with 30%
egg mortality, concentrations in post-migration sockeye
salmon roe are predicted to approach, and in some cases
exceed, this threshold. The TEQ in post-migration Great
Central Lake sockeye roe was observed to be 2.9 ( 0.21 pg/g
and Stuart Lake sockeye stocks are all predicted to exceed
3 pg/g lipid.
in the population dynamics of salmon (e.g., ref 50), it is
possible that the combined burden of dioxin-like contami-
nants has an effect on recruitment of some Pacific salmon
stocks. Large reductions in Pacific salmon populations have
with high concentrations of PCBs, PCDDs, and PCDFs
worldwide. A number of stressors (including harvesting,
climate effects, ocean currents, habitat destruction, and
in Pacific salmon populations in the last few decades. This
study indicates that concentrations of HOCs with dioxin-
like toxicity could also have played a significant role. It is
important to clarify the historic and current effects of HOCs
on Pacific salmon stocks as concentrations of several new
relatively new HOCs add to current toxicological burdens of
historic HOCs with the potential to affect Pacific salmon
populations in the future.
We thank Ian Birtwell for providing fish for the pilot study.
Thanks to Marc Trudel and Glenys Webster for helpful
comments. This work was supported by grants and fellow-
ships from the Natural Sciences and Engineering Research
Council of Canada.
FIGURE 5. Variation in magnification factors (ratios of post- to pre-migration tissue concentrations of HOCs) among stocks of Pacific
salmon, as a function of the total distance each stock migrates in fresh water to spawn. Lines are predictions of a steady-state model
(solid line ) no exchange between soma and gonad, kSG) 0; dashed line ) equilibrium between soma and gonad, kSG. 0), circles are
0.01 of calculated values). Model predictions assume no metabolic or diffusive loss of total chemical through migration. Also shown are
measured values for soma (+) and gonad (×) in sockeye migrating to Lower Fish Lake, AK (ref 5: 410 km, ∑PCBs and DDT), Gluskie Creek,
BC (Gray et al., unpublished manuscript: 1200 km, ∑PCBs), and Great Central Lake, BC (present study: 63 km, ∑PCBs, ∑PCDDs, and
TABLE 4. Concentrations of Dioxin-TEQs (pg/g wet weight and pg/g lipid) in Tissues of Upriver-Migrating Great Central Lake
tissue sexpre-migration MFa
0.13 ( 0.10
0.11 ( 0.05
0.05 ( 0.02
0.07 ( 0.03
0.11 ( 0.01
0.08 ( 0.00
0.13 ( 0.04
0.16 ( 0.09
0.36 ( 0.07
0.30 ( 0.05
2.03 ( 1.23
1.79 ( 0.82
1.12 ( 0.45
1.32 ( 0.41
0.86 ( 0.22
8.03 ( 0.79
13.60 ( 3.63
5.89 ( 2.90
14.59 ( 5.28
2.88 ( 0.21
aMF (magnification factor) is the ratio of post-migration (Robertson Creek) to pre-migration (Barkley Sound) values.
FIGURE 6. Observed dioxin toxic equivalent concentration (TEQ,
pg g-1lipid) in roe of Great Central Lake sockeye salmon before
(coastal) and after (spawning) migration. Dashed line is the lowest
with 30% mortality in eggs of Oncorhynchus mykiss. Grey bars are
post-migration TEQs predicted for several farther-migrating stocks
for one stock each of (from left to right) chum, pink and chinook
VOL. 38, NO. 23, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY96223
Supporting Information Available
Additional text, equations, tables, and figure. This material
is available free of charge via the Internet at http://
(1) Groot, C., Margolis, L., Eds. Pacific Salmon Life Histories; UBC
Press: Vancouver, 1991.
(2) Brett, J. R. Physiological Ecology of Pacific Salmon; Groot, C.,
Margolis, L., Clarke, W. C., Eds.; UBC Press: Vancouver, 1995.
(3) Hendry, A. P.; Berg, O. K. Can. J. Zool. 1999, 77, 1663.
(5) Ewald, G.; Larsson, P.; Linge, H.; Okla, L.; Szarzi, N.Arctic 1998,
(6) Arkoosh, M. R.; Collier, T. K. Human Ecol. Risk Assess. 2002, 8,
(7) Walker, M. K.; Spitsbergen, J. M.; Olsson, J. R.; Petersen, R. E.
Can. J. Fish. Aquat. Sci. 1991, 48, 875.
(8) Giesy, J. P.; Jones, P. D.; Kannan, K.; Newsted, J. L.; Tillitt, D.
E.; Williams, L. L. Aquat. Toxicol. 2002, 59, 35.
(9) Ruckelshaus, M. H.; Levin, P.; Johnson, J. B.; Kareiva, P. M.
Annu. Rev. Ecol. Syst. 2002, 33, 665.
2002, 36, 1886-1892.
(11) Hale, R. C.; Alaee, M.; Manchester-Neesvig, J. B.; Stapleton, H.
M.; Ikonomou, M. G. Environ. Int. 2003, 29, 771-770.
(12) Qiao, P.; Gobas, F. A. P. C.; Farrell, A. P. Arch. Environ. Contam.
Toxicol. 2000, 39, 369.
(13) Higgs, D. A.; Macdonald, J. S.; Levings, C. D.; Dosanjh, B. S.
Physiological Ecology of Pacific Salmon; Groot, C., Margolis, L.,
Clarke, W. C., Eds.; UBC Press: Vancouver, 1995.
(14) Gobas, F. A. P. C.; Mackay, D. Environ. Toxicol. Chem. 1987, 6,
Appl. Pharmacol. 1977, 39, 329.
(16) Niimi, A. J.; Oliver, B. G. Can. J. Fish. Aquat. Sci. 1983, 40, 1388.
(17) Fitzsimmons, P. N.; Fernandez, J. D.; Hoffman, A. D.; Butter-
worth, B. C.; Nichols, J. W. Aquat. Toxicol. 2001, 55, 23.
(18) Niimi, A. J. Sci. Total Environ. 1996, 192, 123.
J. Fish. Aquat. Sci. 1999, 56, 1389.
(20) Russell, R. W.; Gobas, F. A. P. C.; Haffner, G. D. Environ. Sci.
Technol. 1999, 33, 416.
(21) Mackay, D.; Patterson, S. Environ. Sci. Technol. 1982, 16:655a.
(22) Gobas, F. A. P. C.; Wilcockson, J. B.; Russell, R. W.; Haffner, G.
D. Environ. Sci. Technol. 1999, 33, 133.
(23) Gobas, F. A. P. C. Ecol. Modell. 1993, 69, 1.
(24) Ikonomou, M. G.; Fraser, T. L.; Crewe, N. F.; Fischer, M. B.;
Rogers, I. H.; He, T.; Sather, P. J.; Oamb, R. L. Can. Tech. Rep.
Fish. Aquat. Sci. 2001, 2389, 1.
(25) Idler, D. R.; Clemens, W. A. Int. Pac. Salmon Fish. Commission
Prog. Rep. 1959, 6, 1.
(26) Thurston, C. E.; Newman, H. W. Fish. Ind. Res. 1962, 2, 15.
(27) Williams, I. V.; Fagerlund, U. H. M.; McBride, J. R.; Strasdine,
Prog. Rep. 1977, 37, 1.
(28) Igarashi, H.; Zama, K. Bull. Jpn. Soc. Sci. Fish. 1953, 18, 618.
(29) Ando, S.; Hatano, M.; Zama, K. Comp. Biochem. Physiol., Part
B: Biochem. Mol. Biol. 1985, 80, 303.
1991, 45, 47.
(31) Williams, I. V.; Brett, J. R.; Bell, G. R.; Traxler, G. S.; Bagshaw,
J.; McBride, J. R.; Fagerlund, U. H. M.; Dye, U. M.; Sumpter, J.
P.; Donaldson, E. M.; Bilinski, E.; Tsuyuki, H.; Peters, M. D.;
Choromanski, E. M.; Cheng, J. H. Y.; Coleridge, W. L. Int. Pac.
Salmon Fish. Commission Bull. 1986, 23, 1.
(32) Greene, C. W. J. Biol. Chem. 1919, 39, 435.
(33) Greene, C. W. Physiol. Rev. 1926, 6, 201.
(34) Robinson, J. S.; Mead, J. F. Can. J. Biochem. 1970, 48, 837.
(35) Phleger, C. F.; Laub, R. J.; Benson, A. A. Lipids 1989, 24, 286.
Toxicol. Chem. 1998, 17, 2422.
(37) Patton, S.; Crozier, G. F.; Benson, A. A. Nature 1970, 225, 754.
P. D.; Walker, M. K.; Zabel, E. W.; Peterson, R. E. Environ. Sci.
Technol. 2003, 37, 3864.
(39) Wright, P. J.; Tillit, D. E. Aquat. Toxiocol. 1999, 47, 77.
(40) Bernatchez, L.; Dodson, J. J. Can. J. Fish. Aquat. Sci. 1987, 44,
(41) Leonard, J. B. K.; McCormick, S. D. Can. J. Fish. Aquat. Sci.
1999, 56, 1159.
(42) Healey, M. C. J. Fish Biol. 2001, 58, 1545.
(43) Hinch, S. G.; Rand, P. S. Can. J. Fish. Aquat. Sci. 1998, 55, 1821.
1996, 92, 27.
(45) Wania, F. Environ. Sci. Pollut. 1999, 6, 11.
(46) MacDonald, R.; MacKay, D.; Hickie, B. Environ. Sci. Technol.
2002, 36, 456A.
Appl. Pharmacol. 1994, 126, 45.
(49) Polischuk, S. C.; Norstrom, R. J.; Ramsay, M. A. Environ. Pollut.
2002, 118, 29.
J. S.; Ladle, M. Fish. Res. 2003, 62, 111.
Received for review March 12, 2004. Revised manuscript
received July 12, 2004. Accepted July 16, 2004.
62249ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 23, 2004