Environ. Sci. Technol. 2003, 37, 468-474
Polychlorinated Dibenzo-p-dioxin and
Dibenzofuran Concentration Profiles
in Sediments and Flood-Plain Soils
of the Tittabawassee River,
Michigan
KLARA HILSCHEROVA,†
K U R U N T H A C H A L A M K A N N A N , * ,‡
HARUHIKO NAKATA,†
NOBUYASU HANARI,§
NOBUYOSHI YAMASHITA,§
PATRICK W. BRADLEY,†
JOHN M. MCCABE,|
ALLAN B. TAYLOR,| AND JOHN P. GIESY†
National Food Safety and Toxicology Center, Department of
Zoology, Institute for Environmental Toxicology,
Michigan State University, East Lansing, Michigan 48824,
Wadsworth Center, New York State Department of Health,
and Department of Environmental Health and Toxicology,
School of Public Health, State University of New York, Empire
State Plaza, P.O. Box 509, Albany, New York 12201-0509,
National Institute of Advanced Industrial Science and
Technology, 16-1 Onogawa, Tsukuba, Japan, and Waste
Management Division, Michigan Department of
Environmental Quality, Lansing, Michigan 48909
Concentrations of polychlorinated dibenzo-p-dioxins
(PCDDs) and dibenzofurans (PCDFs) in sediments and floodplain soils collected along the Tittabawassee River in
Michigan ranged from 102 to 53 600 pg/g, dry wt. Mean PCDD/
PCDF concentrations in downstream sediment and soil
were from 10- to 20-fold greater than those found at locations
upstream of Midland, Michigan. Concentrations of PCDD/
PCDF in sediments and flood-plain soils from the
Tittabawassee watershed were comparable to those
found in industrialized areas such as the Housatonic and
lower Passaic Rivers in the U.S. Concentrations of PCDDs/
PCDFs in soil and sediment were not correlated with
total organic carbon (TOC) in sediments or soils. OCDD
and 2,3,7,8-TeCDF were the predominant congeners
in sediment/soil collected from locations downstream of
Midland, Michigan. Principal component analysis of the PCDD/
PCDF congener profile suggested the presence of
sources originating from a mixture of chlorophenol and
other chlorinated compound production. Mass balance
analysis of TCDD equivalents (TCDD-EQs) derived from H4IIEluc bioassay of sediment extracts and 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents (TEQs) estimated from
instrumental analysis suggested that PCDDs/PCDFs were
the major dioxin-like compounds present in sediments. A
* Corresponding author present address: Wadsworth Center, New
York State Department of Health, Empire State Plaza, P.O. Box 509,
Albany, NY 12201-0509; phone: 518-474-0015; fax: 518-473-2895;
e-mail: kkannan@wadsworth.org.
† Michigan State University.
‡ Wadsworth Center and State University of New York.
§ National Institute of Advanced Industrial Science and Technology.
| Michigan Department of Environmental Quality.
468
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 3, 2003
significant correlation existed between bioassay-derived
TCDD-EQs and instrumentally measured TEQs (r 2 ) 0.94).
Introduction
Water and sediment quality in Saginaw Bay and its tributaries
in Michigan has been demonstrably impacted by historical
industrial activities within the Saginaw Bay watershed (1).
The U.S. Environmental Protection Agency (EPA) has identified the Saginaw River and Bay as an Area of Concern (1) as
part of the Assessment and Remediation of Contaminated
Sediments Program (ARCS). The Saginaw River and Bay
receives discharges from 87 industrial facilities and 127
wastewater treatment plants, including those of the cities of
Flint, Saginaw, Bay City, and Midland in Michigan. The
Tittabawassee River is the largest tributary to the Saginaw
River (Figure 1). It drains 5426 km2 and contributes nearly
50% of the tributary flow to the Saginaw River. The Tittabawassee River watershed is mainly woodland (41%) with
scattered agricultural areas (37%). The Chippewa and Pine
Rivers are two of the major tributaries to the Tittabawassee
River. The city of Midland is the major industrial and
population center on the Tittabawassee River. The Michigan
Department of Public Health has issued fish advisories on
the basis of elevated levels of polychlorinated dibenzo-pdioxins (PCDDs) and dibenzofurans (PCDFs) and polychlorinated biphenyls (PCBs) found in fish tissues in the
Tittabawassee and Saginaw Rivers downstream of Midland.
The objectives of this study were to elucidate spatial
differences in concentrations of PCDDs and PCDFs in
sediment and flood-plain soils collected along the Tittabawassee River. Sediments collected upstream of Midland and
in the Chippewa and Pine Rivers were used as references for
comparison. Absolute and relative (congener profile) concentrations and 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents (TEQs) were evaluated to describe possible sources and
potential for effects. In addition, the H4IIE-luc bioassay was
used to determine total dioxin-like activity (TCDD-EQs) in
sediments. A mass balance analysis of dioxin-like activity
derived from instrumental (i.e., TEQs) and bioassay (i.e.,
TCDD-EQs) analyses was used to test for the presence of
other dioxin-like compounds that can bind to the aryl
hydrocarbon receptor (AhR). Earlier studies have demonstrated that the H4IIE-luc bioassay coupled with instrumental
analysis is useful in integrated assessments of dioxin-like
activity in sediments (2-4). The instrumental and bioanalytical approaches provide different and complementary
information. Whereas instrumental analyses are useful for
identifying the compounds of interest and evaluating concentrations of specific compounds, they provide little
information regarding the integrated biological relevance of
a complex mixture of compounds associated with environmental samples such as sediment. Where appropriate,
bioassay-directed fractionation and mass balance analyses
are powerful tools for characterizing the causative agents
responsible for bioassay responses observed. Studies have
indicated that organic extracts of sediments elicited significant AhR-mediated responses in vitro, although the chemical
concentrations often did not explain the activities observed
in bioassays (2, 5, 6). Empirical bioassay results and mass
balance analyses can suggest the magnitudes of the contributions of target organic compounds to total AhR-mediated
activity. Thus, bioassay-based toxicity identification and
evaluation (TIE) and mass balance analyses are important
approaches in the assessment of sediment contamination
10.1021/es020920c CCC: $25.00
 2003 American Chemical Society
Published on Web 01/01/2003
FIGURE 1. Map of the Tittabawassee River and its tributaries in Michigan. Sampling area is shaded.
as the sediment extracts can contain many potentially AhRactive compounds that were not analyzed by instrumental
methods.
Materials and Methods
Sample Collection. Surface sediment (0-10 cm) and floodplain soil samples were collected from more than 100
locations along the Tittabawassee River from August to
October 2001, using a stainless steel Ponar grab sampler.
Equal masses of individual sediment samples from certain
locations were pooled to obtain composite sediments.
Individual samples within the selected composite samples
were also analyzed to examine the extent of variability in
PCDD/PCDF concentrations. Further, sediments collected
from river transects were analyzed and are represented as
transect sediments. Transect samples are composites of Ponar
grab samples from several locations along a transect perpendicular to the bank of the river. Transect sediments
include both surface sediments collected using a Ponar
sampler (0-10 cm) and subsurface (40-60 cm) sediments
collected using a PVC corer. In addition, flood-plain soils
collected along the river were also analyzed. For comparison,
corresponding samples of composite and transect sediments
and flood-plain soils were collected from the Pine and
Chippewa Rivers upstream of Midland (Figure 1). In this
study, “upstream” refers to those samples collected upstream
of Midland. Sediment samples were homogenized thoroughly
prior to analysis. Pebbles and twigs were removed, and
portions of sediments were analyzed for moisture content,
total organic carbon (TOC), and total organic matter (TOM).
TOC was determined by dry combustion microcarbon
analysis, and TOM was determined by loss on ignition. TOC
and TOM were analyzed at the Soil and Plant Nutrient
Laboratory at Michigan State University, East Lansing, MI.
Chemical Analysis. Concentrations of 17 2,3,7,8-substituted PCDDs and PCDFs were analyzed by modifications of
previously described methods (7-9). Sediment and soil
samples (40-50 g wet) were homogenized with anhydrous
sodium sulfate, Soxhlet extracted using 400 mL of toluene,
and concentrated to 10 mL. Extracts were treated with
activated copper to remove sulfur. Portions of sediment
samples were dried at 80 °C for the determination of moisture
content. Two-milliliter aliquots of extracts were taken for in
vitro bioassays using H4IIE-luc cells to estimate total dioxin-
like activity. 13C-labeled PCDDs and PCDFs were added as
internal standards to the remaining extracts before they were
passed through 6 g of multilayer silica gel (100-200 mesh)
column packed in a glass column (10 mm i.d.). The silica gel
column was prepared by packing 2 g of silica gel followed
by 2 g of 40% acidic silica and then 2 g of silica gel. The silica
gel bed was washed with 100 mL of hexane prior to loading
of the sample extract. Extracts were then eluted with 10%
dichloromethane in hexane. Samples were concentrated and
passed through 1 g of activated-carbon-impregnated silica
gel column (Wako Pure Chemical Industries, Tokyo, Japan).
The first fraction eluted with 150 mL of hexane was discarded,
and the second fraction eluted with 200 mL of toluene was
concentrated and analyzed by high-resolution gas chromatography and high-resolution mass spectrometry (HRGC/
HRMS). A Hewlett-Packard 6890 GC interfaced with a JEOL
JMS-700 HRMS was used for the determination of heptaand octa-chlorinated congeners. A Hewlett-Packard 6890
Series II gas chromatograph coupled with a HRMS Micromass
Autospec-Ultima instrument was used for the determination
of tetra- through hexa-chlorinated congeners. The HRMS
was operated in electron-impact, selected-ion-monitoring
mode at a resolution R > 10 000 (10% valley). Separation was
achieved using a DB-5 and a DB-17 column (J&W Scientific;
each 0.25 mm i.d. × 60 m length). The mass spectrometer
was operated at an EI energy of 70 eV, and the ion current
was set at 600 µA. PCDD/PCDF congeners were monitored
by selective ion monitoring (SIM) at the two most intense
ions at the molecular ion cluster. Further details of the
instrumental analysis are presented elsewhere (7). Calculated
concentrations were reported as less than the limit of
detection if either the observed isotope ratio was not within
(20% of the theoretical ratio or the peak area was not greater
than the specified threshold (3 times the noise). The detection
limits of individual PCDD and PCDF congeners varied from
0.1 to 1.5 pg/g, dry wt. Quality-assurance and quality-control
protocols include analysis of matrix spike (MS) and matrix
spike duplicates (MSD), replicate samples, standard reference
materials, and procedural blanks. Mean recoveries of 13C2378-TCDF, 13C-12378-PeCDF, 13C-123678-HxCDF, 13C1234678-HpCDF, 13C-2378-TCDD, 13C-12378-PeCDD, 13C123478-HxCDD, 13C-1234678-HpCDD, and 13C-OCDD spiked
to 62 samples of sediments were 86 ( 17, 87 ( 19, 82 ( 17,
82 ( 23, 89 ( 16, 106 ( 39, 83 ( 16, 90 ( 26, and 65 ( 24%,
VOL. 37, NO. 3, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
469
TABLE 1. Mean and Range of 2,3,7,8-Substituted PCDD and PCDF Concentrations in Sediments from Upstream (Tributary Rivers)
and Downstream Locations of the Tittabawassee River
sediment
n
composite sediment
16
transect sediment (overall)
14
transect (upper)
7
transect (lower)
6
flood-plain soil
7
individual sediment
12
composite sediment
3
transect sediment
4
flood-plain soil
3
individual sediment
2
TOC
(%)
PCDDs
(pg/g, dry wt)
PCDFs
(pg/g, dry wt)
Downstream
0.64
1.11
(0.36-1.3)
(0.61-2.24)
0.41
0.71
(0.11-1.27)
(0.19-2.19)
0.37
0.64
(0.26-0.71)
(0.45-1.22)
0.43
0.74
(0.11-1.27)
(0.19-2.19)
3.2
5.4
(1.2-6.5)
(2.1-11.1)
0.89
1.53
(0.11-4.05)
(0.19-6.98)
2400
(560-7590)
1480
(160-5710)
1260
(222-5710)
1470
(160-4250)
14800
(2730-38400)
4380
(330-19900)
3790
(1030-11400)
3140
(88-17700)
1560
(88-4360)
4830
(140-17700)
10600
(2390-17200)
6280
(188-17100)
Upstream
0.72
(0.58-0.95)
0.76
(0.41-1.28)
4
(3.6-4.4)
0.42
(0.23-0.6)
120
(92-140)
97
(35-200)
320
(120-450)
116
(72-160)
59
(27-120)
22
(9.2-47)
46
(7.2-67)
145
(103-186)
respectively. Reported concentrations were corrected for the
recoveries of corresponding 13C-labeled congeners.
In vitro Bioassay Analysis. For all of the samples, raw
extracts, after Soxhlet extraction, were assessed for AhRmediated activity by use of the in vitro H4IIE-luc recombinant
cells. Solvent was transferred to hexane prior to dosing of the
cells. Test and control wells were dosed with appropriate
extract or solvent. For selected samples, raw extracts were
treated with concentrated sulfuric acid for at least 12 h and
then assessed for AhR-mediated activity. The procedures
applied to conduct the bioassays have been described in
detail previously (6, 10, 11). Luciferase and protein assays
were conducted after 72 h of exposure. Sample responses,
expressed as mean relative luminescence units (RLU) over
three replicate wells, were converted to relative response
units, expressed as a percentage of the maximum response
observed for 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD; %
TCDD max) standard curves generated on the same day.
Potencies of samples relative to TCDD were estimated. Mass
balance analysis (or potency balance analysis) was used to
examine whether the known composition of a sample
(identified by instrumental analysis) can account for the
magnitude or potency of biological response observed.
Further details of the bioassay procedures and mass balance
analysis are presented elsewhere (2, 5, 6, 10).
Data were analyzed statistically by principal component
analysis (PCA). For PCA, concentrations of PCDD/PCDF
isomers were normalized to percentages of the total concentrations. PCA was performed using a statistical package
supplied by Esumi Co. Ltd., Tokyo, Japan.
Results and Discussion
Chemical Analysis. Mean concentrations of 17 2,3,7,8substituted PCDD and PCDF congeners in sediment collected
from the Tittabawassee River downstream of Midland were
2400 and 3790 pg/g, dry wt, respectively (Table 1). Concentrations of total PCDDs/PCDFs in surface transect sediments
were comparable to those found in composite sediment
samples. No significant difference in PCDD concentration
was observed in sediment collected from surface (0-10 cm)
or subsurface (40-60 cm) transects although mean PCDF
concentrations were greater in subsurface layers (4830 pg/g)
470
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 3, 2003
TOM
(%)
1.23
(1-1.65)
1.32
(0.71-2.21)
6.9
(6.1-7.5)
0.72
(0.4-1.04)
than in surface layers (1560 pg/g). The greater concentration
of PCDFs in subsurface layers is due to an elevated
concentration in one of the subsurface sediment samples
collected at the lower (downstream) sections of the river
(17 700 pg/g, dry wt). Nevertheless, the mean concentration
of PCDFs was greater than that of PCDDs even when the
outlier was removed. Concentrations of PCDDs/PCDFs were
6-10-fold greater in flood-plain soils than in sediments
collected along the Tittabawassee River. The greatest PCDD/
PCDF concentration (53 600 pg/g, dry wt) was found in a
flood-plain soil collected from the middle sections of the
river (SS5). Total PCDD and PCDF concentrations in floodplain soil and sediment samples collected upstream of
Midland in the Tittabawassee River and in the Chippewa
and Pine Rivers were 10-20-fold less than those collected
downstream of Midland (Table 1). These results suggest the
existence of potential sources near Midland. Previous studies
have reported the occurrence of PCDDs/PCDFs arising from
historical industrial operations at Dow Chemical Company
in Midland (12, 13).
No spatial gradient in PCDD/PCDF concentrations was
evident, although the midsections of the river contained
greater concentrations than did the upper or lower sections
(Figure 2, Supporting Information). Concentrations of total
PCDDs/PCDFs were not significantly correlated with organic
carbon content in sediment or soils (p > 0.05) (Figure 3,
Supporting Information). The greater total concentrations
of PCDDs/PCDFs in certain locations could be explained by
the hydrological and geological characteristics (e.g., sedimentation) of the river.
Concentrations of total PCDDs and PCDFs in Tittabawassee River sediments were greater than those reported for
Housatonic River sediments in Massachusetts, which ranged
from 160 to 5400 pg/g, dry wt, except in one sample that
contained a concentration of 82 000 pg/g, dry wt (14).
Concentrations of PCDDs/PCDFs in surficial sediments from
the lower Passaic River and Newark Bay, New Jersey, ranged
from 370 to 24 000 pg/g, dry wt (15). Total concentrations
of PCDDs/PCDFs in sediments from the rivers in the United
Kingdom ranged from 446 to 9310 pg/g, dry wt. In Korea,
concentrations of PCDD/PCDF ranged from 102 to 6490 pg/
g, dry wt, whereas concentrations reported for sediments in
FIGURE 5. Principal component analysis of PCDD/PCDF congener
profiles in soil/sediments collected along the Tittabawassee River
and in representative sources.
FIGURE 4. Contributions of individual PCDD/PCDF congeners to
total PCDD/PCDF concentrations in soil and sediments collected
upstream and downstream locations in the Tittabawassee River.
1, 2378-TCDF; 2, 12378-PeCDF; 3, 23478-PeCDF; 4, 123478-HxCDF; 5,
123678-HxCDF; 6, 123789-HxCDF; 7, 234678-HxCDF; 8, 1234678-HpCDF;
9, 1234789-HpCDF; 10, OCDF; 11, 2378-TCDD; 12, 12378-PeCDD; 13,
123478-HxCDD; 14, 123678-HxCDD; 15, 123789-HxCDD; 16, 1234678HpCDD; 17, OCDD.
Japan ranged from 377 to 15 800 pg/g, dry wt (7, 9, 16-18).
In general, the concentrations of PCDDs/PCDFs measured
in sediments and flood-plain soils from the Tittabawassee
River were similar to those reported in industrialized areas
(15).
Concentrations of PCDFs were greater than those of
PCDDs in approximately 50% of the sediments/soils collected
from the Tittabawassee River. Ratios of concentrations of
PCDDs to PCDFs in sediments collected from the Tittabawassee River ranged from 0.32 to 1.82. In contrast, sediments
collected from the Pine and Chippewa Rivers (upstream
reference locations) contained 6-16-fold greater concentrations of PCDDs compared to PCDFs. Profiles of PCDD and
PCDF congeners in sediment and soil from sampling
locations upstream and downstream of Midland are presented in Figure 4. OCDD accounted for 30-60% of the total
PCDD/PCDF concentrations in downstream sediment/soils
and 60-80% in upstream samples. Sediment/soil samples
from upstream locations contained relatively greater proportions of OCDD and lesser proportions of PCDFs than
those from the downstream sampling locations. Among PCDF
congeners, 2378-TCDF was the most abundant congener in
downstream sediments accounting for 12-25% of the total
PCDD/PCDF concentration. This was followed in order by
PeCDF > HxCDF > OCDF > HpCDF. The distributions of
PCDD/PCDF isomers in the majority of sediment/soil from
the Tittabawassee River were similar to those from New
Bedford Harbor and the lower Passaic River (15).
Different sources of PCDDs/PCDFs are characterized by
different congener and homologue patterns (15, 19). Furthermore, differences in the physicochemical (mobility,
solubility, etc.) and biological (biodegradation, bioaccumulation, etc.) properties can alter the congener profiles. The
profiles of PCDD and PCDF congeners in sediment from the
Tittabawassee River downstream of Midland were all similar,
suggesting the presence of a single major source. The pattern
of relative concentrations of PCDD/PCDF congeners in soils
and sediment collected downstream of Midland was different
from that in samples collected from upstream reference
locations. A large proportion of OCDD and HpCDD has been
suggested to be due to the sources originating from chlorophenol-related inputs (20). Greater proportions of TCDFs
suggest sources originating from PCB mixtures, chlorobenzenes, the chlor-alkali process, or the incineration of PCBs
and PVC (19-23). Total concentrations of PCBs in sediments
from the Tittabawassee River were less than 150 ng/g, dry
wt (24). This suggests that PCBs were not likely to be a source
of PCDD/PCDFs, but rather, that other sources such as
chlorophenol and chlorobenzene production, incinerationrelated activities, and/or chlor-alkali processes are likely
sources of PCDFs found in Tittabawassee River sediments
collected below Midland.
The profiles of relative concentrations of PCDD/PCDF
congeners were characterized by use of principal component
analysis (PCA). The relative orderings of profiles of concentrations of PCDD/PCDF congeners, normalized to the total
concentration of PCDD and PCDF, in sediments or soils from
the Tittabawassee River and some representative source
materials are plotted as a function of the first two principal
components (Figure 5). The first and second principal
components explained 31 and 16%, respectively, of the
original variance of the normalized congener dataset. The
first principal component (PC1) contained greater positive
eigenvectors for congeners 123478- and 234678-HxCDF and
12378-PeCDF, whereas the vectors were negative for OCDD
and 1234678-HpCDD. The second principal component (PC2)
contained positive eigenvectors for congeners 123789HxCDF, OCDF, 2378-TCDD, and 12378-PeCDD, whereas a
negative vector was obtained for 2378-TCDF. In a plot of
PC1 versus PC2, the pattern for Aroclors 1242 and 1248,
graphite sludge, and PVC/PCB pyrolysis, ordered them toward
the right side of the graph, whereas pentachlorophenol (PCP),
sodium pentachlorophenoate, and municipal solid waste
incineration related sources were placed toward the left
(Figure 5). The PCDD/PCDF profiles of Aroclors, graphite
sludge, and chlorobenzenes are dominated by PCDFs,
whereas those in PCP or other chlorophenol-related sources
are dominated by PCDDs, particularly HpCDD and OCDD
(15, 19-23). PCDD/PCDF homologue profiles in tri- and
tetrachlorobenzenes were in the order TeCDF > PeCDF >
HxCDF, which resembled the order of PCDF profiles found
in sediments. Whereas homologue profiles of PCDD/PCDF
have been reported for chlorobenzenes (22), profiles of
individual isomers have not been reported. Because of the
lack of isomer-specific information on PCDD/PCDF profiles
in chlorobenzenes, only homologue profiles were plotted as
a function of principal components. PCA of homologue
profiles of tri- and tetrachlorobenzenes exhibited a greater
negative loading of PC2 as a result of the lack of OCDD and
HpCDD in these mixtures. Tri- and tetrachlorobenzenes (not
shown) were ordered toward the right near Aroclor 1242.
The majority of the soil and sediment samples collected along
the Tittabawassee River were positioned at the intersection
of the two principal components intermediate between the
patterns of PCP, Aroclors 1242 and 1248, and graphite sludge.
Therefore, PCDDs/PCDFs in Tittabawassee River sediments/
VOL. 37, NO. 3, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
471
TABLE 2. Concentrations of Instrumentally Derived TEQs and
Bioassay-Derived TCDD-EQs in Sediments/Soils from the
Tittabawassee Rivera
sample
n
TEQs (EC20)
(pg/g, dry wt)
TCDD-EQs (EC20)
(pg/g, dry wt)
Downstream
composite sediment 16 390 (25-1260) 560 (137-1860)
transect sediment
14 310 (4-1980)
340 (17-1430)
flood-plain soil
7 800 (250-1350) 1060 (290-2550)
composite sediment
transect sediment
flood-plain soil
a
Upstream
3 5.7 (1.7-13)
4 1.6 (0.37-4)
3 4.3 (1.3-7.4)
54 (40-63)
73 (35-128)
190 (58-260)
Values in parentheses indicate the range.
flood-plain soils are likely to have come from these sources
and others. Soils/sediments from locations upstream of
Midland exhibited greater negative loading on PC1 because
of their greater proportions of HpCDD and OCDD. Although
the patterns of relative concentrations of PCDD/PCDF
congeners alone are insufficient to determine the exact source
of PCDDs/PCDFs (as PCDD/PCDFs undergo congenerspecific degradation in soil/sediments and volatilization in
the environment), the results of the principal components
analysis of relative proportions of congeners, combined with
the absolute concentrations, indicate that the Midland area
is the source of PCDDs/PCDFs found in sediments collected
from the Tittabawassee River.
Bioassay Analysis. Both raw extracts after Soxhlet extraction and acid-treated extracts of several sediment and soil
samples were examined for their ability to elicit AhR-mediated
activity in vitro using H4IIE-luc cells. All of the tested sediment
extracts, both raw and acid-treated, elicited significant AhRmediated activity. TCDD-EQs were calculated on the basis
of the amount of sample needed to produce a response
equivalent to EC20 - EC80 of TCDD. The estimate of a range
is more appropriate than the estimate of a single point
because of the nonparallelism of the dose-response curves
(the lower the range, the more parallel the curves) (25). The
values on the basis of the response equivalent to EC20 were
used for statistical comparisons between instrumental and
bioassay results. TCDD-EQs estimated on the basis of an
EC20 are presented (Table 2). EC20 values are used because
all of the samples reached 20% of the efficacy and, therefore,
extrapolations were not needed to calculate EC20 concentrations. Concentrations of TCDD-EQs in sediments ranged from
17 to 1860 pg/g, dry wt, whereas those in flood-plain soils
ranged from 290 to 2550 pg/g, dry wt. Concentrations of
TCDD-EQs in upstream sediment were 10-100-fold less than
those in downstream sediments. Similarly, TCDD-EQs in
upstream soils were approximately 5-10-fold less than those
in downstream sediment samples.
On the basis of the results of instrumental analysis of
PCDD/PCDF, concentrations of 2,3,7,8-tetrachlorodibenzop-dioxin equivalents (TEQs) were calculated using H4IIEluc cell line-specific relative potencies (REPs-EC20) (26). The
use of REPs instead of toxic equivalency factors (TEFs) was
necessary to perform mass balance analyses of TCDD-EQs
derived using the H4IIE cell bioassay and TEQs calculated
on the basis of instrumental analysis. Concentrations of TEQs
in sediments from the Tittabawassee River ranged from 4 to
1980 pg/g, dry wt (Table 2). The mean concentration of TEQs
in flood-plain soils was 800 pg/g, dry wt. As observed for
TCDD-EQs, concentrations of TEQs in upstream soils and
sediments were 50 to 100-fold less than those in downstream
soils and sediments. A significant correlation existed between
concentrations of TCDD-EQs derived from H4IIE-luc bioassay and TEQs estimated from instrumental analysis
472
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 3, 2003
FIGURE 6. Relationship between instrumentally derived TEQs and
bioassay-derived TCDD-EQs in soils and sediments collected along
the Tittabawassee River and its tributaries: (A) raw extracts, (B)
acid-treated extracts.
(p < 0.05) (Figure 6). In general, concentrations of TCDDEQs were greater than those of TEQs, although the difference
in magnitude of the concentrations between TCDD-EQs and
TEQs was less than 2-fold. Because of the variability in
bioassay results, analytical errors within the assays, and
variations in relative potency values, it has been suggested
that a 2-fold difference is the minimum required to conclude
that the concentrations of TEQs and TCDD-EQs are not
equivalent (27). These results suggest that the target compounds, PCDDs/PCDFs, are the major sources of dioxin-like
activity in sediments/soils from the Tittabawassee River and
that little, if any, activity is due to other compounds such as
PCBs or PAHs present in sediments. This conclusion is
supported by the observation that the concentrations of PCBs
in soil/sediments were relatively small (<150 ng/g, dry wt;
24), not likely to contribute significant amounts of PCDD/
PCDFs to the sediments or soils, and also not likely to
contribute to the total concentration of TEQs as measured
in the bioassay. TEQs contributed by non-, mono-, and diortho-PCBs in flood-plain soils collected along the Tittabawassee River ranged from 0.12 to 11 pg/g, dry wt (mean )
1.2 pg/g, dry wt) (Data from the Environmental Response
Division, Michigan Department of Environmental Quality,
2002). This is 100-1000 times less than the TEQs contributed
by PCDD/PCDFs in sediments/soils. This provides additional
evidence that PCBs are not a major source of TEQs in
sediments/soils from the Tittabawassee River.
PAHs can also contribute to AhR-mediated activity in
sediments measured in the bioassay (2, 5). Therefore,
sediment extracts were also treated with concentrated sulfuric
acid to remove PAHs, if any, and then used in bioassays. The
response magnitudes induced by acid-treated sediment
extracts were similar to those induced by nontreated extracts.
This suggests that the contribution of acid-labile compounds
including PAHs to the dioxin-like activity was minor in the
sediments. However, the concentrations of TCDD-EQs were
greater than those of TEQs in flood-plain soils collected from
reference locations. Treatment of these upstream soil samples
with sulfuric acid resulted in a considerable reduction in
dioxin-like activity, which suggests that PAHs or other acidlabile compounds are the major contributors to dioxin-like
activity in upstream soils (Figure 7).
Despite the fact that OCDF was assigned a TEF of 0, PCDFs
were the major contributors to TEQs, accounting for, on
average, >90% of the total TEQs in sediments and soils
collected from the Tittabawassee River downstream of
FIGURE 7. Comparison of concentrations of bioassay-derived TCDD-EQs in flood-plain soils before and after acid treatment. Locations
SS3, SS4, and SS10 were collected from reference locations on the Tittabawassee River, upstream of Midland.
FIGURE 8. Contributions of individual PCDD/PCDF congeners to total TEQs in soils and sediments collected along the Tittabawassee River.
TABLE 3. Relative Contributions (%) of PCDDs and PCDFs to
Total TEQs in Sediment and Soils Collected from the
Tittabawassee River
location
PCDFs
PCDDs
downstream transect
upstream transect
downstream composite
upstream composite
downstream flood-plain soil
upstream flood-plain soil
97.6
59.5
97.2
87.6
92.9
57.5
2.4
40.5
2.8
12.4
7.1
42.5
Midland (see Table 3 and Figure 8). Among PCDFs, 23478PeCDF was the major contributor to TEQs in sediments,
accounting for 40-50% of the TEQs, followed in order by
12378-PeCDF, 2378-TeCDF, and 123478-HxCDF (Figure 7).
These four congeners collectively accounted for 85-90% of
the total TEQs.
These results suggest that the concentrations of PCDDs/
PCDFs in sediments and flood-plain soils collected downstream of Midland were comparable to those observed in
several other industrialized regions. OCDD and TCDF were
the predominant congeners in soil and sediments, which
suggests a unique pattern. Concentrations of PCDD/PCDF
in sediment/soil were not correlated with TOC. Bioassayderived TCDD-EQs were similar to those derived from
instrumental analysis of PCDDs/PCDFs, which suggests that
PCDDs/PCDFs were the critical contaminants that contribute
to dioxin-like activity in sediments from the Tittabawassee
River.
The present study has not addressed the issue of the risks
of PCDDs/PCDFs to humans or wildlife. Risk is a factor of
both exposure and hazard (toxicity). An assessment of the
risks posed by PCDDs/PCDFs would need to consider the
bioavailability of each congener as well as other factors that
would affect the potential exposure. The Agency for Toxic
Substances and Disease Registry in the United States (28)
has adopted interim policy guidelines for PCDDs/PCDFs in
residential soils near or on hazardous waste sites. When
concentrations of TEQs exceed 50 pg of TEQ/g, dry wt,
evaluation of site-specific factors such as pathway analysis
and soil cover are needed. Actions to be taken when soil
concentrations exceed 1000 pg/g, dry wt, include surveillance
and research of exposure investigations. In this study, TEQ
concentrations in 42 of 61 soil/sediment samples were greater
than 50 pg/g, dry wt. More than 90% of the sediment samples
collected downstream of Midland contained TEQ concentrations greater than 50 pg/g, dry wt. Fifteen of the 61 samples
contained TEQ concentrations greater than 1000 pg/g, dry
wt. These TEQs estimates might be an underestimation
because these are based on H4IIE-bioassays rather than WHO
guidelines, which have greater TEF values. According to these
guidelines additional reconnaissance work would be warranted. Remedial options should be based on sound data on
the relative risks now and in the future (29).
Acknowledgments
We thank Rick Lundgren of Michigan Department of
Environmental Quality for assistance in sampling design,
VOL. 37, NO. 3, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
473
technical support, and field work and Great Lakes Protection
Fund for funding this study.
Supporting Information Available
Figures showing spatial trends in concentrations of total
PCDDs/PCDFs in surface sediment and flood-plain soils and
relationship between total PCDD/PCDF concentrations and
TOC in transect and composite sediments and flood-plain
soils collected along the Tittabawassee River and its tributaries. This material is available free of charge via the Internet
at http://pubs.acs.org.
Literature Cited
(1) U.S. EPA. Assessment of Sediments in the Saginaw River Area
of Concern, September 28, 1995. Prepared by Science Applications International Corporation. EPA Contract No. 68-D3-0030.
(2) Khim, J. S.; Villeneuve, D. L.; Kannan, K.; Koh, C. H.; Giesy, J.
P. Environ. Sci. Technol. 1999, 33, 4199-4205.
(3) Hilscherova, K.; Machala, M.; Kannan, K.; Blankenship, A. L.;
Giesy, J. P. Environ. Sci. Pollut. Res. 2000, 7, 159-171.
(4) Khim, J. S.; Lee, K. T.; Villeneuve, D. L.; Kannan, K.; Giesy, J. P.;
Koh, C. H. Arch. Environ. Contam. Toxicol. 2001, 40, 151-160.
(5) Kannan, K.; Villeneuve, D. L.; Yamashita, N.; Imagawa, T.;
Hashimoto, S.; Miyazaki, A.; Giesy, J. P. Environ. Sci. Technol.
2000, 34, 3560-3567.
(6) Hilscherova, K.; Kannan, K.; Kang, Y.-S.; Holoubek, I.; Machala,
M.; Masunaga, S.; Nakanishi, J.; Giesy, J. P. Environ. Toxicol.
Chem. 2001, 20, 2768-2777.
(7) Yamashita, N.; Kannan, K.; Imagawa, T.; Villeneuve, D. L.;
Hashimoto, S.; Miyazaki, A.; Giesy, J. P. Environ. Sci. Technol.
2000, 34, 3560-3567.
(8) Kannan, K.; Kober, J. L.; Kang, Y.-S.; Masunaga, S.; Nakanishi,
J.; Ostaszewski, A.; Giesy, J. P. Environ. Toxicol. Chem. 2001, 20,
1878-1889.
(9) Im, S. H.; Kannan, K.; Matsuda, M.; Giesy, J. P.; Wakimoto, T.
Environ. Toxicol. Chem. 2002, 21, 245-252.
(10) Villeneuve, D. L.; Blankenship, A. L.; Giesy, J. P. Environ. Toxicol.
Chem. 2000, 19, 2835-2843.
(11) Villeneuve, D. L.; Khim, J. S.; Kannan, K.; Giesy, J. P. Aquat.
Toxicol. 2000, 54, 125-141.
(12) U.S. EPA. Dow chemical wastewater characterization study:
Tittabawassee River sediments and native fish. U.S. Environmental Protection Agency, Region V, Westlake, OH, 1986.
(13) U.S. EPA. Risk assessment for dioxin contamination at Midland,
Michigan. U.S. Environmental Protection Agency, Region V,
Chicago, IL, 1988.
474
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 3, 2003
(14) Eitzer, B. D. Environ. Sci. Technol. 1993, 27, 1632-1637.
(15) Wenning, R. J.; Harris, M. A.; Ungs, M. J.; Paustenbach, D. J.;
Bedbury, H. Arch. Environ. Contam. Toxicol. 1992, 22, 397413.
(16) Rose, C. L.; McKay, W. A.; Ambidge, P. F. Chemosphere 1994,
29, 1279-1292.
(17) Im, S. H.; Kannan, K.; Giesy, J. P.; Matsuda, M.; Wakimoto, T.
Environ. Sci. Technol. 2002, 36, 3700-3705.
(18) Sakurai, T.; Kim, J. G.; Suzuki, N.; Nakanishi, J. Chemosphere
1996, 33, 2007-2020.
(19) Kannan, K.; Watanabe, I.; Giesy, J. P. Toxicol. Environ. Chem.
1998, 67, 135-146.
(20) Masunaga, S.; Takasuga, T.; Nakanishi, J. Chemosphere 2001,
44, 873-885.
(21) Wakimoto, T.; Kannan, N.; Ono, M.; Tatsukawa, R.; Masuda, Y.
Chemosphere 1988, 17, 743-750.
(22) Swami, K.; Narang, A. S.; Narang, R. S.; Eadon, G. A. Chemosphere
1992, 24, 1845-1853.
(23) Rackwitz, D. Formation of polychlorinated dibenzo-p-dioxins
and furans by pyrolysis of PCB and PVC in a fire in an electric
melting plant/decontamination and deposition of the refuse.
In Dioxin ′91; Research Triangle Park, NC, 1991; Abstract PD5,
p 109.
(24) Michigan Department of Environmental Quality. Report of trace
metal and organic residues in sediment from the Tittabawassee
River watershed. Lansing, MI, 2002.
(25) Villeneuve, D.; Khim, J. S.; Kannan, K.; Giesy, J. P. Environ.
Toxicol. 2002, 17, 128-137.
(26) Brown, D. J.; Chu, M.; Overmeire, I. V.; Chu, A.; Clark, G. C.
Organohalogen Compd. 2001, 53, 211-214.
(27) Sanderson, J. T.; Giesy, J. P. Functional Response Assays in
Wildlife Toxicology. In Encyclopedia of Environmental Analysis
and Remediation; Meyers, R. A., Ed.; John Wiley and Sons: New
York, 1998; pp 5272-5297.
(28) De Rosa, C. T.; Brown, D.; Dhara, R.; Garrett, W.; Hansen, H.;
Holler, J.; Jones, D.; Jordan-Izaguirre, D.; O’Connor, R.; Pohl,
H.; Xintaras, C. Dioxin and dioxin-like compounds in soil, Part
1: ATSDR interim policy guideline. In Agency for Toxic
Substances and Disease Registry; U.S. Department of Health and
Human Services: Atlanta, GA, 1997; Vol. 13, pp 759-768.
(29) NRC. Risk Management Strategy for PCB-Contaminated Sediments; National Academy Press: Washington, DC, 2001; p 432.
Received for review September 4, 2002. Revised manuscript
received November 12, 2002. Accepted November 13, 2002.
ES020920C