Archives of Environmental Contamination and Toxicology

, Volume 58, Issue 3, pp 772–782

CYP1A Expression in Caged Rainbow Trout Discriminates Among Sites with Various Degrees of Polychlorinated Biphenyl Contamination

Authors

    • Department of BiologyUniversity of Kentucky
  • J. Scott McClain
    • Department of ZoologyMiami University
    • Monsanto Regulatory Creve Couer Campus
  • James T. Oris
    • Department of ZoologyMiami University
  • David J. Price
    • LFUCG, Division of Water QualityTown Branch Laboratory
  • Wesley J. Birge
    • Department of BiologyUniversity of Kentucky
  • Adria A. Elskus
    • Department of BiologyUniversity of Kentucky
    • U.S. Geological Survey, Maine Toxicology SectionUniversity of Maine
Article

DOI: 10.1007/s00244-009-9368-x

Cite this article as:
Brammell, B.F., McClain, J.S., Oris, J.T. et al. Arch Environ Contam Toxicol (2010) 58: 772. doi:10.1007/s00244-009-9368-x

Abstract

It has become increasingly apparent that resident fish can develop resistance to chemicals in their environment, thus compromising their usefulness as sentinels of site-specific pollution. By using a stream system whose resident fish appear to have developed pollutant resistance (Brammell et al., Mar Environ Res 58:251–255, 2005), we tested the hypothesis that the pollutant-inducible biomarker, cytochrome P4501A (CYP1A), as measured in field-caged juvenile rainbow trout (Oncorhynchus mykiss), would reflect relative pollution differences between reference and polychlorinated biphenyl (PCB)-contaminated sites. Trout were caged in the Town Branch/Mud River system (Logan County, KY), a stream system undergoing remediation for PCBs. Fish were held in remediated (Town Branch), unremeditated (Mud River), and reference sites for 2 weeks during spring 2002. At the end of this period, gill and hepatic CYP1A expression were measured. To evaluate the relative PCB exposure of caged trout and provide a reference point against which to calibrate CYP1A response, PCB levels were quantified in sediments from each site. Hepatic CYP1A expression in caged trout clearly detected the presence of PCBs in the Town Branch/Mud River stream system. Sediment PCB levels and hepatic CYP1A expression in caged trout produced identical pollution rankings for the study sites. Gill CYP1A expression, although suggestive of site differences, was not statistically different among sites. Unlike resident fish, which failed to show site differences in hepatic CYP1A expression in this waterway (Brammell et al. 2005), caged fish proved to be a sensitive discriminator of relative PCB contamination in this system. In summary, we determined that CYP1A expression in caged fish reflected relative in situ pollutant exposure. The exposure paradigm confirmed that 2 weeks was a sufficient caging period for evaluating CYP1A response in this species at these temperatures (13–19°C). In addition, these studies demonstrate that tissue-specific CYP1A expression can provide insights into likely routes of exposure. We conclude that CYP1A expression in caged trout is a reliable and inexpensive first-pass determination of relative environmental pollutant exposure and bioavailability in aqueous systems.

Biomarkers are an effective monitoring tool in toxicology, allowing researchers to assess the biologic response to pollutants in both aquatic and terrestrial organisms (Bucheli and Fent 1995). Unlike measurements of contaminants in tissues or sediments, which provide no information on biologic response, alterations in biomarkers indicate that chemicals present in a system are biologically available and active. The xenobiotic metabolizing enzyme, cytochrome P-4501A (CYP1A), is a widely used biomarker of exposure to planar aromatic hydrocarbon compounds, including environmental contaminants, such as polychlorinated biphenyls (PCBs) (Bucheli and Fent 1995; Elskus and Stegeman 1989; Whyte et al. 2000). CYP1A is a phase I inducible cytochrome P-450 isoform in fish (Rand and Petrocelli 1985) that is sensitive to waterborne pollutants. Although endogenous levels of the enzyme are relatively low, strong and rapid (within hours) induction of CYP1A mRNA and catalytically active protein occurs in response to exposure to promoting compounds (Stegeman and Hahn 1994). Induced CYP1A mRNA and protein eventually return to basal levels after exposure has been terminated and the compounds have been metabolized or eliminated. This type of induction response supports the use of CYP1A as an ideal biomarker of short-term exposure to appropriate substrates, such as PCBs.

Rainbow trout are excellent candidates for biomonitoring studies because they are amenable to caging (Blom et al. 1998; Oikari 2006) and sensitive to CYP1A inducers (Buhler and Wang-Buhler 1998; Blom et al. 1998). Exposure to industrial mixtures of PCBs is known to induce CYP1A expression in rainbow trout (Melancon and Lech 1983). Rainbow trout CYP1A protein concentrations and activity levels in liver, gill, and other tissues have also been validated as biomonitoring tools (Tuvikene et al. 1996; Fenet et al. 1998; McClain et al. 2003). Thermal restrictions prevent rainbow trout survival in most southeastern United States streams during the warmest summer months, therefore eliminating the possibility of inadvertently introducing a nonnative species. Rainbow trout are also widely cultured and thus are both readily available and acclimatized to confinement in relatively small areas.

The Town Branch/Mud River system in southwestern Kentucky has been contaminated since the 1960s with PCBs—including Aroclor 1260, a PCB mixture that induces CYP1A in fish—from a local manufacturing plant (Commonwealth of Kentucky 1997; Focardi et al. 1995). Only certain sections of this waterway have been remediated, thus providing an optimal system for investigating the ability to apply a stream biomonitoring study of CYP1A expression in caged fish to detect and discriminate among sites with varying levels of PCB contamination.

The objective of this study was to assess the presence of biologically available contaminant concentrations in the Town Branch/Mud River system using CYP1A expression in caged trout as a biomonitoring tool. We hypothesized that relative expression of CYP1A in caged trout would reflect relative levels of sediment PCBs among the study sites. CYP1A expression in gill and liver was measured to compare the relative sensitivity of these tissues and to evaluate the possibility of aqueous routes of exposure to CYP1A inducers, as would be indicated by CYP1A gene expression in gill tissue. In addition, we compared CYP1A protein expression levels, catalytic activity, and mRNA expression in liver to evaluate the relative sensitivity of the techniques as biomonitoring metrics in this in situ rainbow trout model. PCB analysis of sediment was conducted to evaluate PCB residue content in this system. A few studies have measured both in situ PCB concentrations and the biologic response of fish caged in PCB-contaminated freshwater sites (Blom et al. 1998; Otto and Moon 1996). The present study was intended to provide a basis for a reliable exposure paradigm to the biomonitoring field by demonstrating the successful use of a sentinel fish model to detect biologically active PCB contamination in a remediated stream site.

Materials and Methods

Materials

7-Ethoxyresorufin and resorufin were obtained from Molecular Probes (Eugene, OR). The monoclonal antibody made against scup (Stenotomus chrysops) CYP1A protein, MAb 1-12-3, was a generous gift from Dr. John Stegeman (Woods Hole Oceanographic Institution). Cy 5-conjugated Affinipure goat antimouse immunoglobulin (Ig) G was obtained from Jackson Immunoresearch Laboratories (West Grove, PA), and precast polyacrylamide gradient gels were obtained from Invitrogen (Carlsbad, CA). Nitrocellulose membrane (0.45 μm) was obtained from Schleicher and Schull (Keene, NH). The Bio-Dot SF Microfiltration Apparatus was obtained from Bio-Rad (Hercules, CA). Tri-Reagent was from Sigma (St. Louis, MO), DNAse I and QuantumRNA classic internal standard were from Ambion (Austin, TX), and cDNA synthesis kit was from Amersham Biosciences (Uppsala, Sweden). All other reagents were obtained from Sigma, Fisher Scientific (Fair Lawn, NJ), or Invitrogen.

Animals

Juvenile (12–13 months old) rainbow trout (O. mykiss) (weight range 20–100 g) were obtained from Wolf Creek National Fish Hatchery in Jamestown, KY, and held in the laboratory for 7 weeks before caging. Fish were held in a 948 l (250 gallon) flow through tank at 9°C, 14:10 light-to-dark cycle, and fed Purina Trout Chow ad libitum every other day.

Study Site: Town Branch/Mud River

The Town Branch/Mud River system (Fig. 1) has been contaminated since the 1960s with PCBs from a local manufacturing plant. For >20 years PCBs were released into a lagoon behind the plant that leaked waste containing high concentrations of PCBs (≤332,500 ppm) into Town Branch approximately 8 km (5 miles) upstream of its confluence with the Mud River (Commonwealth of Kentucky 1997). Sediment PCB concentrations of 280 ppm (dry sediment, clay-silt fraction) (Table 1) (Birge 1988) were documented in Town Branch in 1986 and led to the remediation efforts on Town Branch that began in 1997. Removal of contaminated sediments from both the streambed and floodplain of Town Branch was completed in July 2001 (M. Mills, Kentucky Division of Water, personal communication). No remediation has been conducted in the Mud River section downstream of the Town Branch confluence, and relatively high levels of PCBs continue to be found in sediments in this area (5.9 ppm dry sediment, clay-silt fraction) (Price and Birge 1999). We chose caging sites in the remediated section of Town Branch, the unremediated Mud River, and two reference sites located upstream of these areas that were expected to be free of measurable PCBs in the sediment.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-009-9368-x/MediaObjects/244_2009_9368_Fig1_HTML.gif
Fig. 1

Location of study sites in the Town Branch/Mud River system near Russellville, KY

Table 1

Water-quality parameters recorded during the 2-week caging experiment at the Town Branch/Mud River study sites during April and May 2002

Site

Water temperature (°C)

pH

Day 0

Days 14 and 15

Town Branch reference

13–14

16.5–17.5

7.50

Town Branch remediated

14–16

17.0–18.8

7.86

Mud River reference

15

18

7.64

Mud River unremediated

19

18.5

8.23

Cage Design

Cages were octagonal, 61 cm (24 inches) in diameter, 31.5 cm (12 inches) in height, and constructed of 1/2 inch polyvinyl chloride and heavy-duty garden mesh sewn together with galvanized wire and plastic locking ties. Cages were completely submerged at a depth of approximately 1 m and tethered by way of a nylon rope to an iron fence that was post driven into the streambed.

Caging Experiment

Juvenile rainbow trout were caged at four sites (10–12 fish/cage and 5 cages/site) in the Town Branch/Mud River system (Fig. 1) during April and May 2002, a time when water temperatures and water levels are optimal for cage deployment of this species. Each cage represents one replicate. Cages were positioned in four sites: (1) the remediated section of Town Branch (TB remediated), (2) in a clean section of Town Branch upstream from the original source of PCB contamination (TB reference), (3) in the Mud River just below the confluence with the contaminated section of Town Branch (MR unremediated), and (4) in a clean section of the Mud River well above the confluence with Town Branch (MR reference). Fish were held in cages for 14 days in the Mud River and for 15 days in the Town Branch sites. Trout held in the laboratory and killed on day 16 served as laboratory-held controls. Temperature measurements were recorded at each site on days 0, 14, and 15, and pH measurements were recorded on days 14 and 15 after cage deployment (Table 1). There are no United States Geographical Survey stream gages at these sites, and no additional information on either rainfall or temperature is available.

On days 14 and 15, trout were removed from cages and immediately killed on site. Fish were weighed, and liver and gill tissue were removed and flash frozen in liquid nitrogen (−192°C). Carcasses were kept on wet ice until returned to the laboratory (within 36 h) and stored at −80°C for later analysis of gut contents. All fish were analyzed individually.

Liver Microsomal Protein Isolation

Livers were removed from liquid nitrogen, weighed, and subsectioned. Aliquots for RNA analysis (0.07–0.2 g) were refrozen in liquid nitrogen. Aliquots for microsome preparation were immediately homogenized in 10 volumes (weight to volume) of ice-cold 50 mM Tris buffer (pH 7.4). Microsomal fractions were obtained by differential centrifugation as previously described (Stegeman 1979). The final 100,000×g microsomal pellets were resuspended in 50 mM Tris buffer (pH 7.4) containing 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM dithiothreitol (DTT), and 20% glycerol at a 1:1 ratio (liver weight-to-resuspension buffer volume). Microsomal samples were stored in liquid nitrogen until analyzed for catalytic activity and CYP1A protein content (within 3 weeks).

RNA Isolation and Analysis

Total RNA was isolated from 150-mg medial sections of gill or liver tissue using Tri-Reagent (Chomczynski and Sacchi 1987). Total RNA was dissolved in molecular-grade water and quantified by spectroscopy at 260 nm. A 5-μg aliquot of total RNA was treated with 1 U DNAse I for 10 minutes at 37°C to remove genomic DNA. DNAse was rendered inactive before cDNA synthesis with 1 μl 20-mM solution of EDTA. The same 5-μg aliquot of total RNA was then reverse transcribed with Moloney murine leukemia virus reverse transcriptase and random hexamer primers as part of the first-strand cDNA synthesis kit. A 2-μL aliquot of a 15-μL cDNA synthesis reaction was polymerase chain reaction (PCR) amplified for 26 cycles using the protocol and primers for the target gene, CYP1A, in rainbow trout as described in McClain et al. (2003). An 18S universal ribosomal RNA primer (QuantumRNA Classic 18S internal standard) was used to coamplify (in a separate reaction) a reference transcript for all samples. The 18S PCR product (amplicon) expression value was used to normalize all CYP1A raw values by calculating CYP1A expression as a ratio of (measured CYP1A amplicon) ÷ (measured 18S amplicon) for each sample. This ratio of CYP1A/18S was determined for each sample and is referred to as a normalized expression value. All precautions for verifying template integrity and quality control of the mRNA quantization were the same as described in McClain et al. (2003).

Catalytic and Protein Assays

Ethoxyresorufin-O-deethylase (EROD) activity, a reaction catalyzed predominantly by CYP1A (Buhler and Wang-Buhler 1998), was measured fluorometrically in 48-well CoStar plates using the method of Hahn et al. (1996). The reaction mixture contained 50 mM Tris [pH 7.4], 0.1 M NaCl, 2 μM 7-ethoxyresorufin, and 100–300 ug microsomal protein in a final volume of 200 μl. The reaction was initiated by the addition of 1.34 mM NADPH, run at room temperature (28–29°C), and production of the fluorescent product, resorufin, measured for 13 minutes. The linear portion of the curve was used to evaluate the rate of the reaction. All EROD assays were run in triplicate on a PerSeptive Biosystems Cytofluor Series 4000 multiwell plate fluorometer.

Microsomal protein was measured fluorometrically using the method described by Lorenzen and Kennedy (1993) with bovine serum albumin as the standard. All protein assays were run in triplicate.

Immunoblotting Procedures

CYP1A protein was quantified by immunoblotting using a Bio-Dot SF microfiltration slot-blot apparatus (BIO RAD, Hercules, CA). Twenty micrograms of microsomal protein diluted in 200 μL buffer (20 mM Tris, 0.5 M) was loaded into each well and vacuum transferred onto a nitrocellulose membrane (0.45 μM). The membrane was incubated in TBS–5% milk at 4°C overnight to block nonspecific binding, followed by incubation with MAb 1-12-3, a monoclonal antibody that recognizes CYP1A in multiple vertebrate species (Stegeman and Hahn 1994) as described (Elskus et al. 1999). CYP1A signal was detected using Cy 5-conjugated Affinipure goat antimouse IgG as the secondary antibody, and blots were scanned at 633 nm excitation/670 nm emission using a Typhoon 8600 scanner (Molecular Dynamics/Amersham, Sunnyvale, CA) and quantified using Image Quant (Molecular Dynamics/Amersham, Sunnyvale, CA). Liver microsomes from trout treated with the CYP1A model inducer, ß-naphthoflavone, were loaded in seven concentrations ranging from 0.1 to 7 μg protein diluted in 200 μL buffer to evaluate linearity of the CYP1A signal on each blot. All samples were run at least in triplicate.

Stomach Content Analysis

Fish carcasses were removed from the −80°C freezer in the laboratory and allowed to thaw. At least five caged fish from each site were dissected, and stomach and digestive tract contents were examined to determine if caged fish were eating.

Condition Factor Assessment

We measured the condition factor (CF) of the laboratory fish, from which the caged fish were drawn, as a representation of the starting CF of the caged fish.

Sediment Collection and Extraction for PCB Analysis

Sediment samples were collected from caging sites in fall 2002, approximately 5 months after completion of the caging study. Samples were collected from the upper 5 to 10 cm of sediment in acetone-rinsed glass jars and placed on ice. Wet-sediment extractions of PCBs were performed according to United States Environmental Protection Agency (USEPA) SW-846 Method 3540C (1997). Detection limits for sediment samples ranged from 8 to 17 μg PCB/kg sediment. Spiked sediment recoveries were also analyzed using Aroclor 1248 recoveries and averaged 97.75 ± 8.98%. Reported values were corrected for recovery.

Statistical Treatment of Data

Statistical analyses were performed using either SYSTAT Version 10 (SYSTAT, Chicago, IL) or Statistica 99 (StatSoft, Tulsa, OK), version 5.5. End points were measured in individual fish; the data were averaged per cage; and each cage was considered a single replicate. The laboratory-held fish were housed together in one tank and thus considered one replicate; therefore, they were excluded from statistical analyses. All data were transformed (log 10) before analysis. All CYP1A data were analyzed using one-way analysis of variance, and differences among means were tested using Bonferroni test. All differences were considered significant at p ≤ 0.05.

Results

PCB Analysis

PCB concentrations in Mud River reference sediment were lower than detection. Somewhat unexpectedly, PCB levels in the Mud River unremediated sediment were fairly low. In stark contrast to our expectations, PCB sediment concentrations in the Town Branch remediated section were quite high (Table 2).
Table 2

Preremedition and postremediation PCB concentrations (μg/g dry weight) in sediment samples collected from Town Branch/Mud River study sites in 1987 and 2002 (mean ± SD [n])

Site

Preremediation (1987)a

Postremediation (2002)

Mean total PCB concentration (ppm dry weight)

Mean total PCB concentration (ppm dry weight)

TB reference

0.08 ± 0.03 (5)

0.033 ± 0.213 (2)

TB remediated

184.3 ± 121 (5)

45.7 ± 7.44 (4)

MR reference

0.06 ± 0.02 (3)b

BD (2)

MR unremediated

2.36 ± 1.35 (5)

0.086 ± 0.047 (2)

BD below detection (<0.008 ppm); MR Mud River; TB Town Branch

aPreremediation data from Birge (1988)

bPCBs were only detected in three of five samples collected at this site

Liver CYP1A Expression

All three measures of hepatic CYP1A expression reflected the relative degree of sediment PCB contamination in Town Branch. Liver CYP1A mRNA levels (Fig. 2A) in fish caged in the remediated portion of Town Branch were significantly increased (8-fold) compared with levels in caged Town Branch reference fish. In contrast, no differences were observed in hepatic CYP1A mRNA expression between fish caged at the Mud River unremediated and reference sites. All means within a cage observed in either gill or liver tissue are displayed in Table 3. The mean of each cage was treated as a single replicate.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-009-9368-x/MediaObjects/244_2009_9368_Fig2_HTML.gif
Fig. 2

Relative CYP1A mRNA levels, CYP1A protein levels, and EROD activity (pmol resorufin/min/mg microsomal protein) in liver of rainbow trout caged in the Town Branch/Mud River system for 2 weeks during April and May 2002. Values represent means ± SE, n = number of cages. Means with the same letter are not significantly different at p < 0.05. Laboratory-held fish represent a single pool of five fish and were not included in the statistical analysis

Table 3

Mean hepatic EROD activity (pmol resorufin/min/mg microsomal protein) values for rainbow trout caged in the Town Branch/Mud River system for 2 weeks during April and May 2002

Site

Cage no.

EROD (SD)

n

Liver protein (SD)

n

Liver CYP1A mRNA (SD)

n

Gill CYP1A mRNA (SD)

n

pmol/min/mg

TB reference

1

2

29.5 (13.5)

2

0.313 (0.147)

  

3

54.6 (22.4)

2

2

33.5 (43.6)

2

0.574 (0.573)

  

2

90.2 (4.63)

3

7

16.2 (19.3)

6

0.210 (0.0536)

    

4

4

20.3 (19.0)

2

0.315 (0.0214)

4

5.23 (8.97)

  

5

1

48.39 (NA)

1

0.228 (NA)

1

8.4 (NA)

  

TB remediated

1

6

40.3 (58.0)

6

0.802 (1.27)

1

80.7 (NA)

1

73.1 (NA)

2

5

79.3 (62.0)

3

3.67 (3.83)

1

19.8 (NA)

1

82.6 (NA)

3

5

55.8 (20.8)

5

2.33 (3.81)

1

64.1 (NA)

1

87.5 (NA)

4

5

58.9 (37.3)

5

0.830 (0.826)

2

39.2 (34.1)

1

90.4 (NA)

5

5

48.6 (31.3)

5

0.906 (0.772)

  

1

92.8 (NA)

MR reference

1

8

2.8 (1.1)

1

0.162 (0.0304)

3

0.675 (0.0183)

5

78.6 (17.2)

2

9

3.2 (0.9)

2

0.150 (0.0647)

2

0.652 (0.0210)

  

MR unremediated

1

7

6.3 (3.8)

6

0.130 (0.0760)

2

0.910 (0.459)

5

59.2 (30.8)

2

4

8.3 (7.9)

3

0.170 (0.0152)

3

2.00 (2.43)

  

3

2

14.2 (5.4)

1

0.017 (NA)

    

4

3

23.5 (8.3)

3

0.330 (0.151)

    

Lab fish

NA

10

16.8 (11.2)

9

0.243 (0.089)

5

36.4 (17.2)

5

72.0 (15.7)

N number of cages; NA not available/appropriate

Levels of hepatic CYP1 catalytic activity (Fig. 2B), measured as EROD, were significantly higher in trout caged in the remediated sections of Town Branch compared with EROD levels in fish caged in the Town Branch reference site as well as both Mud River sites. EROD activity in trout caged in the unremediated section of the Mud River was significantly higher than EROD activity in trout caged in the Mud River reference section, but this did not differ from EROD levels in Town Branch reference fish. Fish caged at the Mud River reference site had lower EROD activity levels than fish caged at any other site.

Hepatic CYP1A protein levels (Fig. 2C) in fish caged at the remediated section of Town Branch were significantly increased (5-fold) compared with CYP1A levels in fish caged in Town Branch reference sites. Fish caged in the unremediated section of the Mud River had CYP1A protein levels similar to those in fish caged at the reference sites.

Gill CYP1A Expression

Unlike liver CYP1A, gill CYP1A levels did not differ significantly among Town Branch remediated versus reference sites, the only sites with sufficiently high “n” values for statistical analysis (Fig. 3). However, the general trend was similar to that of the hepatic response. Gill CYP1A mRNA levels in fish from the Town Branch remediated section tended to be higher relative to gill CYP1A mRNA levels in reference site fish (Fig. 3). The great variance in these data, combined with the small sample size (n = 5) likely contributed to the lack of significance. A posthoc power analysis (Sokal and Rohlf 1981) of these data indicated a sample size of 19 fish would be needed to detect a 50% change in gill CYP1A mRNA levels among fish from these sites.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-009-9368-x/MediaObjects/244_2009_9368_Fig3_HTML.gif
Fig. 3

Relative CYP1A mRNA levels in gill tissue of rainbow trout caged in the Town Branch/Mud River system for 2 weeks during April and May 2002. Values represent means ± SE for n pools of 5 fish per pool. Fish were pooled on a per-cage basis. All means are similar at p < 0.05

Correlation Analysis

Matched pair correlation analysis was conducted using a matched pairs Student t test between each combination of assays for CYP1A EROD, protein, and mRNA values (alpha = 0.05; n = 15, 12, 15, respectively). These analyses showed a strong correlation between the assays for CYP1A mRNA and catalytic activity (r = 0.90), between CYP1A mRNA and protein concentrations (r = 0.82), and between CYP1A protein concentrations and catalytic activity (r = 0.96).

CF

The CF of the laboratory fish from which the caged fish were drawn was considered a fair representation of the starting CF of the caged fish. The Fulton-type CF [W(g)/L(cm)3] × 100,000 (Anderson 1983) of trout caged in all sites was similar and ranged from 0.79 (±0.07) to 0.86 (±0.21) mean (±SD). The CF of laboratory-held trout from which the caged fish were drawn was significantly higher 1.04 (±0.04) than that of the caged fish.

Fish Survival

Seventeen of the 20 cages deployed in the experiment were recovered, and cage loss was likely caused by 2 high-flow events that occurred during the 2-week caging period. Some mortality of caged fish did occur. Overall survival of caged fish at sites ranged from approximately 42% in the Mud River unremediated section to approximately 73% in the Town Branch remediated section. Fish survival appeared to relate directly to the flow regime at individual cages, with fish in cages exposed directly to swift current during high-flow events experiencing higher mortality than fish caged in slower flow areas. Water-quality parameters were similar among sites (Table 1), and were well within the range of tolerance of rainbow trout (Myrick and Cech 2000; Wilkie and Wood 1994), making it unlikely that pH or temperature confounded the results of this experiment. In addition, the fish densities present in the cages in this study were lower than the majority of those used in rainbow trout caging studies discussed in a recent review (Oikari 2006), suggesting that mortality caused by excessive crowding should not have been an issue.

Stomach Content Analysis

Analysis of the stomach contents of caged rainbow trout in this study showed the presence of a limited number of invertebrates. Nearly all fish examined had some prey items in their stomach, and there was no significant differences observed between sites. However, the stomach contents of nearly all fish were sparse, consisting of only a few very small (5- to 10-mm) invertebrates, an indicator of limited feeding during the exposure period.

Discussion

The results of this study demonstrate the utility of caged fish in biomarker-based pollution evaluations. CYP1A expression in field-caged rainbow trout reflected relative sediment PCB concentrations within a stream: Fish caged in the Town Branch remediated section (high sediment PCBs) had CYP1A levels that were higher than those in fish caged in the Town Branch reference site (very low sediment PCBs), and EROD activity levels were higher in fish caged at the Mud River unremediated site (very low sediment PCBs) versus fish caged at the Mud River reference site (nondetectable sediment PCBs). Liver CYP1A proved to be more sensitive than gill CYP1A in detecting site contamination. Although this study was not designed to distinguish among routes of exposure, evidence of limited feeding, together with CYP1A expression in gills, suggest that fish in this stream system were exposed to PCBs directly through the water column. The results of this study illustrate the utility of caging studies and demonstrate that CYP1A inducers are present and bioavailable in the Town Branch/Mud River system. Importantly, this study demonstrates that CYP1A expression in caged trout reliably reflected relative PCB concentrations in contaminated waterways, can help to determine routes of exposure, and is an inexpensive first-pass determination of relative environmental pollution in aqueous systems.

The results of this study demonstrate the ability to deploy a biomonitoring study of CYP1A to detect bioavailable PCBs, or other planar hydrocarbons, at a site where the use of CYP1A in resident fish fails to indicate the presence of contamination. Previous work in our laboratory found no differences in hepatic CYP1 catalytic activity in resident fish species collected from the contaminated and reference sites in the Town Branch/Mud River system, despite the presence of high PCB levels in resident fish tissues from the impacted sites (Brammell et al. 2005). It is likely that the Town Branch/Mud River resident fish have become nonresponsive to CYP1A induction by PCBs. Laboratory exposure experiments with resident fish as well as previous studies by others (Celander and Forlin 1995; Elskus et al. 1999) demonstrate the loss of CYP1A inducibility in fish residing in habitats chronically impacted with these and other CYP1A-inducing chemicals. The present study highlights the advantages of using caged fish in addition to the collection of resident fish tissues in biomarker studies.

The relative levels of CYP1A induction observed in caged rainbow trout were comparable with those reported in other studies of fish caged in polluted areas. For hepatic EROD, our observation of a 2.5-fold increase in rainbow trout caged at a PCB-remediated site relative to caged controls agreed well with the 2-fold induction of EROD activity reported in juvenile rainbow trout caged for 2 weeks (12–15°C) at a site contaminated by industrial and municipal effluents (Fenet et al. 1998). Our results were also similar to the 8-fold EROD induction observed in juvenile rainbow trout caged for 3 weeks (4.7–10.5°C) in a harbor contaminated by poorly treated industrial and municipal sewage as well as oil from commercial fishing vessels (Tuvikene et al. 1996). For hepatic CYP1A protein, the 5.5-fold induction of CYP1A protein levels in fish caged at the remediated site was somewhat higher than the 2.4-fold and 2.1-fold increase in hepatic CYP1A protein in largemouth bass and channel catfish, respectively, after 7- and 14-day caging experiments (Haasch et al. 1993). Sediment PCB levels in that river system (5–20 ppm dry weight (Haasch et al. 1993) were somewhat lower than those in the Town Branch remediated site (46 ppm dry weight). The differences in relative CYP1A protein induction among these studies may be due to species and/or exposure differences. For CYP1A mRNA, McClain et al. (McClain et al. 2003) reported 4- and 11-fold induction of hepatic CYP1A mRNA in juvenile rainbow trout caged for 8 days (15–20°C) at a site contaminated by creosote as well as a mixture of other industrial and agricultural contaminants relative to fish caged at a control site. We found an 8.2-fold increase in hepatic CYP1A mRNA in trout caged in the Town Branch remediated section relative to levels in caged reference fish. Although direct comparisons of CYP1A expression among different field studies are difficult to make because exposure conditions are necessarily diverse, relative comparisons among studies are useful for indicating the range of CYP1A responses likely to be observed in the field. Together these studies demonstrate the utility and sensitivity of CYP1A expression—measured either as mRNA, protein, or catalytic activity—as an indicator of pollutant presence and bioavailability.

Although gill CYP1A mRNA can be increased in response to contaminant exposure through either aqueous or dietary routes, aqueous exposure is known to result in greater gill CYP1A expression relative to dietary exposure (Van Veld et al. 1997). Gill CYP1A expression, in conjunction with evidence that caged fish consumed little food, indicates that caged fish were exposed to contaminants in part, if not mainly, by way of an aqueous route. Although PCBs are strongly hydrophobic, with dissolved water concentrations on the order of pg/L to ng/L, aqueous exposure to PCBs can nevertheless be a significant route of PCB exposure to aquatic organisms (Schrap and Opperhuizen 1990; Randall et al. 1998). The high ventilation rate of fish makes them particularly efficient at extracting hydrophobic organic pollutants from the water column (McKim et al. 1985). The decrease in CF and the limited stomach contents of caged fish during the experiment relative to laboratory controls further indicated that dietary exposure was not a major route of exposure. Exposure by way of aqueous routes, possibly coupled with exposure to suspended sediment during two high-flow water events occurring during the experiment, seems the most likely route of contaminant exposure to the caged fish.

We found a high degree of correlation among levels of hepatic CYP1A mRNA, catalytic activity, and CYP1A protein in fish caged for 2 weeks in a system contaminated with various concentrations of PCBs, a result consistent with the temporal nature of CYP1A induction by chlorinated inducers. Rainbow trout aqueously exposed to 3,4,3′,4′-tetrachlorobiphenyl (PCB 126) exhibited EROD activity that was significantly increased compared with controls even after a 2-week recovery and depuration period, whereas EROD levels in trout exposed in an identical manner to benzo[a]pyrene, a nonchlorinated inducer did not exhibit activity (Jönsson et al. 2006). In scup (S. chrysops) exposed to 2,3,7,8-tetrachlorodibenzofuran, CYP1A mRNA, EROD, and CYP1A protein concentrations all remained increased and highly correlated during a 2-week period after a single intraperitoneal injection (Hahn and Stegeman 1994), likely reflecting the resistance of halogenated inducers to metabolism and excretion (White et al. 1997). This contrasts with the shorter duration of CYP1A mRNA induction after exposure to rapidly metabolized, nonhalogenated inducers, such as polynuclear aromatic hydrocarbons (PAHs). Kloepper-Sams and Stegeman (1989) reported that CYP1A mRNA levels in killifish returned to control levels 5 days after a single intraperitoneal injection of the PAH ß-naphthoflavone, whereas CYP1A protein and EROD activity levels remained increased for 13 days. Similarly, Levine and Oris (1999) reported that in trout aqueously exposed to the PAH, BaP, and hepatic CYP1A mRNA concentrations returned to control levels within 72–120 h, whereas EROD activity remained increased through 120 h. Because sediment concentrations of PAHs are not known for the sites examined in this study, their relative contribution to the observed CYP1A induction in the caged fish cannot be determined. However, the temporal pattern of CYP1A mRNA induction by the chlorinated compounds previously described as well as our analyses demonstrating that PCBs are present in this system, supports our conclusion that increased CYP1A levels in caged trout likely reflect their direct exposure to PCBs and the bioavailability of PCBs in this stream system.

CYP1A expression was significantly lower in fish caged in the Mud River sites than in our laboratory held controls, an observation that may reflect a nonchemically mediated alteration of CYP1A in response to the transfer of fish from the laboratory to the field or the presence of CYP1A inducers in laboratory fish chow. There is evidence that rainbow trout CYP1 activity may undergo an adaptation period after transfer of trout to new environments. Huuskonen et al. (1995) reported that EROD and benzo(a)pyrene hydroxylase activities (another CYP1-catalyzed reaction) decreased 64% and 76%, respectively, during a 2-week period in the spring after the transfer of rainbow trout from hatchery tanks to cages in a lake. It is important to note that despite the observed decrease in biotransformation enzyme activity after transfer to new environments, previous reports indicate that the enzyme systems are still inducible during the adaptation period (Lindstrom-Seppa and Oikari 1990). The other possibility is that there were CYP1A-inducing compounds in the fish chow (Maule et al. 2007). Nonetheless, our findings clearly indicate that CYP1A-inducing compounds were present in both Town Branch and the Mud River, demonstrating the robustness of this caging model and the necessity of pertinent reference locations when using a live animal model in biomonitoring studies.

Confinement coupled with two periods of relatively high water flow likely induced stress in caged fish during the course of this study. Stress hormones, including cortisol, are known to modulate CYP1A enzyme activity in both mammals (Xiao et al. 1995) and cultured fish hepatocytes (Devaux et al. 1992; Celander et al. 1995). In a study examining the effects of caging stress on EROD, tilapia exposed to PCB 126 had significantly higher hepatic and head kidney EROD activity after a 2-h confinement relative to unconfined, PCB-treated fish (Quabius et al. 2002). Jorgensen et al. (2001) demonstrated that chronically high cortisol levels resulted in a downregulation of BaP-induced CYP1A protein levels in Arctic char (Salvelinus alpinus), although both protein and EROD levels were still induced significantly more than in untreated controls. Although stress hormones do modulate CYP1A, this modulation does not appear to either mask exposure effects or to threaten CYP1A’s usefulness as a biomarker. Thus, although the results observed in our study may have been affected by stress, they were likely affected equally among the sites, with little relative effect on the outcome of the expression analysis.

Although the CF of caged fish in this study was significantly lower than the CF of laboratory-held fish, it was similar to or greater than the CF observed in healthy wild salmonids, including rainbow trout (Ensign and Strange 1990). This indicates that the condition of caged fish was not of the normal range observed in salmonids. Moreover, complete lack of feeding (starvation) in rainbow trout for periods up to 3 weeks has been reported to enhance CYP1A response to inducers (Vigano et al. 1993). Based on these studies, it is highly unlikely that inadequate feeding compromised the results of the present experiment.

In summary, we found CYP1A expression in caged trout to reliably reflect PCB exposure and relative PCB sediment content in contaminated and reference waterways. Two weeks appears to be an optimal exposure time for using CYP1A expression in caged rainbow trout as a biomonitoring tool. Monitoring CYP1A can also help determine route of exposure because it was possible to measure biomarker response in both liver and gill while evaluating feeding of the fish during the exposure period. Moreover, we found the use of caged fish to be preferable to resident fish for biomarker-based pollution evaluations because CYP1A expression in caged rainbow trout successfully discriminated among varying levels of contamination (this study) in sites where CYP1A expression in resident fish was homogenous (Brammell et al. 2005). We conclude that CYP1A expression in caged trout is a sensitive, reliable, and accessible supplement to chemical analysis that can be used for first-pass determination of relative environmental pollution and pollutant bioavailability in aqueous systems.

Acknowledgements

The authors thank J. Gray at the Wolf Creek National Fish Hatchery for providing rainbow trout for this project. Sincere thanks are also extended to X. Arzuaga and T. Sagar for providing valuable field assistance and to K. Powers for assistance in the laboratory. The authors also thank C. McCollum (City of Russellville) and several landowners for providing assistance in obtaining access to study sites. We gratefully acknowledge the support and collaboration of the United States Geologic Survey and the University of Kentucky Research Foundation, Grant Agreement No. 01HQGR0133, through the Kentucky Water Resources Research Institute.

Copyright information

© Springer Science+Business Media, LLC 2009