Use of fish gill cells in culture to evaluate the cytotoxicity and photocytotoxicity of intact and photomodified creosote

Cell line and culture media

The cell line from rainbow trout gills, RTgill-W1, was developed in this laboratory [15]. This cell line, when tested as outlined in Clemons et al. [16], was found to lack inducible 7-ethoxyresorufin O-deethylase (EROD) activity, which is a measure of cytochrome CYP1A catalytic activity (N.C. Bols, unpublished data). Cells were cultured in 75-cm² culture flasks at 22°C in Leibovitz's L-15 medium supplemented with 10% fetal bovine serum. The source of the tissue culture supplies and a description of the subcultivation procedure have previously been presented in detail [17, 18].

A modification of the basal medium, L-15, was used for exposure of cells to creosote and UV radiation. This was necessary because treatment of conventional growth media with UV radiation generates toxicants that appear to arise from UV modification of medium components such as vitamins and aromatic amino acids. A description of the constituents of the modified L-15 medium, which was designated L-15/ex, and preparation procedures have been outlined in Schirmer et al. [7].

UV radiation exposure

Cells were irradiated at room temperature in an atmosphere of air and in the presence of tissue culture plate lids as described in Schirmer et al. [7]. The UV irradiation was done with two UV-B fluorescent lamps (Southern N.E. Ultraviolet, Branford, CT, USA) at a photon fluence rate of 1.4 μmol/m²/s UV-B. This photon fluence rate, which approximates an energy fluence rate of 53 μW/cm², previously has been shown to be environmentally relevant [5, 7]. In addition to UV-B (290–320 nm), the UV-B lamps emitted some visible (400-700 nm) and some UV-A (320–400 nm) radiation. However, with a visible to UV-A to UV-B ratio of 5:1.5:1, compared to 200:10:1 in natural sunlight [11], the spectrum was weighted toward its UV-B component. Irradiation was measured with an InstaSpec™II photodiode array spectroradiometer calibrated with a 1-kW quartz halogen lamp (Oriel, Stratford, CT, USA). The values represent the photon fluence rates at the surface of the medium in the wells. A 500-μl/well aliquot of L-15/ex previously has been shown to have little discernible effect on these fluence rates [7]. The duration of irradiation was 2 h for all experiments.

Preparation of creosote in L-15/ex

Liquid-phase creosote was provided by Carbochem (Mississauga, ON, Canada) and was found to have a density of approximately 1.0 g/ml. For application to cultured fish cells, creosote was dissolved in L-15/ex by a technique recommended by Tadokoro et al. [19] for dissolving creosote in water. This technique facilitated the solubilization of creosote in a manner that closely reflected the solubilization process in the environment. A volume of 70 μl (70 mg) of liquid creosote was added to 1 L of sterile L-15/ex in a sterilized amber glass storage bottle with a Teflon®-lined cap (VWR Canlab, Mississauga, ON, Canada) and a sterilized, Teflon-coated magnetic stirrer (VWR Canlab). This nominal creosote concentration was chosen because at some creosote treatment and storage sites the maximum aqueous concentration of PAHs was in this range [20]. To fully protect the solution from light, the bottle was wrapped in aluminum. The solution was stirred at room temperature and the solubilization process was monitored after stirring for 1 h, 1 d, 3 d, and 7 d. At the end of each stirring period, the solution was allowed to settle for 1 h to allow undissolved particles to settle to the bottom of the bottle before 150 ml was carefully removed for application to cell cultures, and for extraction and chemical analysis.

Extraction and chemical analysis of creosote

A sample volume of 120 ml creosote in L-15/ex was extracted as described by Bestari et al. [21] with minor modifications. Briefly, each sample was extracted three times with methylene chloride (Fisher Scientific, Nepean, ON, Canada). The three extracts were combined in an Erlenmeyer flask, which was equipped with a glass wool-plugged funnel containing a layer of sodium sulfate to facilitate the removal of water. The Erlenmeyer flask was then swirled several times in the presence of sodium sulfate followed by passage of the entire extract into a round-bottomed flask, which was equipped with a glass wool-plugged funnel and contained a few glass beads. The methylene chloride was evaporated with a rotary evaporator at 28°C to concentrate the extract to approximately 5 ml, followed by a stream of nitrogen to achieve dryness. The final extracts were obtained by adding 5 ml of methylene chloride to the round-bottomed flasks. The extracts were stored in 5-ml amber glass vials at −20°C until chemical analysis. Naphthalene and fluoranthene could be recovered from spiked samples to 82 and 90%, respectively, which was similar to the recovery rates reported by Bestari et al. [21]. The deviations between two extractions of the same stock solution were less than 12% for all compounds.

For chemical analysis, 3 μl of extract was injected into a Hewlett Packard 5890 Gas Chromatograph (GC) equipped with an HP7673A autosampler and flame ionization detector (FID) (Hewlett Packard, Mississauga, ON, Canada). The temperature was held at 35°C for 2 min, followed by a 15°C/min increase to 165°C, followed by a 30°C/min increase to 300°C and a constant temperature of 300°C for 10 min. The carrier gas was helium with a flow rate of 30 ml/min. The injector temperature of the GC was 200°C, and the FID temperature was 300°C. The deviations between repeated measurements of the same creosote sample were less than 10% for all chemicals.

Chemical analysis focused on PAHs because they can account for 85% of the compounds in creosote [14, 20]. However, a few other compounds were analyzed as well. These included the heterocyclics dibenzofuran and carbazole, and the aromatic hydrocarbons benzene, toluene, and xylene. The detection limits of the GC method for each of the compounds are given in Table 1.

Cytotoxicity and photocytotoxicity of creosote

Confluent monolayers of RTgill-W1 cells in 48-well tissue culture plates were used to study the cytotoxicity and photocytotoxicity of creosote. Confluent cultures were achieved by plating 50,000 cells per well and allowing them to grow for 3 d. At confluency, the culture medium was removed and each well was rinsed once with 500 μl of L-15/ex. After the rinse, all wells received increasing volumes of creosote in L-15/ex together with fresh L-15/ex to yield a total volume per well of 500 μl. To determine cytotoxicity, cells were exposed to creosote for 6 h in the dark before the alamar Blue and CFDAAM cytotoxicity assays were carried out as described below. For photocytotoxicity, cells were exposed to creosote for 6 h in the dark before being UV irradiated for 2 h and measured with the alamar Blue and CFDA-AM cytotoxicity assays.

Cytotoxicity and photocytotoxicity of photomodified creosote

In order to obtain photomodified creosote, aliquots of creosote in L-15/ex were placed in glass jars and UV irradiated for 2 h. Glass jars were used to reduce losses due to adsorption. The jars were covered with 48-well plate lids to achieve photon fluence rates that were similar to those obtained in UV-irradiated 48-well tissue culture plates. The protocol that followed the preparation of photomodified creosote was identical to that described above for the cytotoxicity and photocytotoxicity of creosote but used photomodified creosote solution instead of intact creosote.

Alamar Blue and CFDA-AM cytotoxicity assays

The alamar Blue (Immunocorp. Science, Montreal, PQ, Canada) and CFDA-AM (Molecular Probes, Eugene, OR, USA) indicator dyes were used in combination as described in Schirmer et al. [7]. Briefly, alamar Blue and CFDA-AM were prepared together in L-15/ex to give final concentrations of 5% v/v and 4 μM, respectively. After removal of the creosote solutions, 150-μl aliquots of the dyes were added to each well and, 2 h later, fluorescence was quantified with a CytoFluor 2350 (PerSeptive Biosystems, Burlington, ON, Canada) at respective excitation and emission wavelengths of 530 (±30) and 595 (±35) nm for alamar Blue, and 485 (±22) and 530 (±30) nm for CFDA-AM.

Data analysis

Analysis of raw data. The fluorescence readings in wells that contained creosote in L-15/ex were expressed as a percentage of the readings in control wells that received L-15/ex only. Before these calculations, fluorescence readings for wells without cells were subtracted from the experimental and control values with cells. Dose–response data followed a sigmoid relationship and were analyzed by nonlinear regression using the curve-fitting routine of SigmaPlot (Jandel Scientific, San Rafael, CA, USA). Data were fitted to the logistic function as previously described [8] with

(1)

where y(d) is the % of control at the nominal creosote dose d, g is a slope parameter, and EC50 is the nominal creosote concentration that produces 50% cellular activity compared to the control.

Predicting the cytotoxicity and photocytotoxicity of creosote. Predictions on the effects on cellular activity were made for the cytotoxicity and photocytotoxicity of creosote. These predictions were based on the chemical creosote analysis in this report and on the PAH toxicity data described in Schirmer et al. [8, 9]. In accordance with the toxicity data, naphthalene and fluoranthene were used as model compounds for, respectively, the cytotoxicity and photocytotoxicity of the hydrocarbons in creosote.

Cytotoxicity. Three assumptions were made to predict the potential of creosote solutions to be directly cytotoxic. First, all aromatic hydrocarbons and heterocyclics were assumed to contribute to the direct cytotoxicity of creosote through their accumulation in cell membranes, independent of their structure [8]. Second, water solubility and lipophilicity were regarded as the fundamental properties of a PAH contributing to the accumulation of a PAH in cell membranes. Because these two properties are interrelated and work in opposite directions, the potential ability of a PAH to accumulate in cell membranes was expressed as the product of its water solubility limit times its octanol/water partition coefficient. This is referred to as the membrane accumulation factor (MAF). Inasmuch as naphthalene was the most potent PAH to act through direct cytotoxicity and the one most likely to act in this way in the environment, the MAF of each PAH was expressed relative to the MAF for naphthalene to give a naphthalene equivalent membrane accumulation factor (NEMAF) (Table 2). The concentration of each PAH could be converted into a naphthalene equivalent concentration (NEC) by multiplying its concentration by its NEMAF. Third, all compounds were assumed to act in membranes additively and in a manner as previously shown for naphthalene [8]. Thus, the total NEC of a creosote solution was simply the sum of the NECs for each component identified and quantified in creosote (Table 3). Finally, to predict the effect (% of control) on cells due to cytotoxic compounds in a creosote solution, NECs were applied to the logistic functions obtained previously for the direct cytotoxicity of naphthalene [8]. These functions were

(2)

(3)

With these functions, where d is the sum of NECs that would be expected at each of the applied nominal creosote concentrations, entire dose–response curves were predicted.

Photocytotoxicity. To predict the photocytotoxicity of creosote, previously obtained fluoranthene equivalent factors (FEFs [9]) were used to convert the concentrations of strictly photocytotoxic PAHs into fluoranthene equivalent concentrations (FECs; Table 4). As for the cytotoxic compounds, additivity was assumed. Thus, to predict the effect on cells in % of control due to photocytotoxic compounds in creosote, the sum of FECs in each creosote sample was applied to the logistic functions obtained previously for the immediate photocytotoxicity of fluoranthene [9]. These functions were

(4)

(5)

With these functions, where d is the sum of FECs that would be expected at each of the applied nominal creosote concentrations, entire dose–response curves were predicted.

Statistical analysis. The cytotoxicity and photocytotoxicity of creosote, which were predicted as described above, were compared to the observed values of cytotoxicity and photocytotoxicity to RTgill-W1 cells in both the alamar Blue and the CFDA-AM assay. Four observed values of cytotoxicity or photocytotoxicity were available for each creosote concentration in each of the four creosote samples and stemmed from four replicate wells. Each of these replicate wells contained alamar Blue and CFDA-AM in mixture and was read twice with the CytoFluor to measure either alamar Blue or CFDAAM. These observed values, which were expressed as % of control wells that contained no creosote, were plotted along the y axis.

For predicted values of cytotoxicity or photocytotoxicity, only one value was available for each creosote concentration in each of the four creosote samples because predictions were based on the concentrations of the chemicals in the creosote. These predicted values, expressed as % of control, were plotted along the x axis.

To describe the goodness of fit between the observed and the predicted values, three regression analysis parameters had to be taken into account. These were the coefficient of determination (r²), the slope, and the y intercept. For a perfect match between the observed and the predicted values, these parameters would have to be one for r², one for the slope, and zero for the intercept.

Preparation and chemistry of creosote solutions

The amount of creosote that could be dissolved in L-15/ex was dependent on the time of stirring. After 1 d of stirring, all priority PAHs were found at higher concentrations in L-15/ ex than after 1 h of stirring, with the exception of naphthalene (Table 1). After 3 d, some PAHs had further increased in concentration whereas others slightly decreased. In contrast, all PAH concentrations decreased between days 3 and 7 of stirring. This was also reflected in the total amount of priority PAHs detected, which was lowest after 7 d with 3,300 μg/L and highest after 1 and 3 d with 4,823 μg/L and 4,785 μg/L, respectively (Table 1). However, with the exception of naphthalene, the relative contribution of each PAH to the total mass detected remained relatively constant for each time point with acenaphthene, fluorene, and phenanthrene being the main contributors (Table 1).

Peak concentrations were observed after 3 d of stirring for the other aromatic hydrocarbons and for the heterocyclics (Table 1). Among the heterocyclics, dibenzofuran was more abundant than carbazole whereas among the aromatic hydrocarbons, 2-methyl-naphthalene + indole were the main contributors.

If all detected compounds were added together, the highest total mass was found to be 8,077 μg/L after 3 d of stirring, followed by 6,904 μg/L, 5,830 μg/L, and 5,571 μg/L after 1 d, 1 h, and 7 d of stirring, respectively. Thus, with the method used to solubilize creosote and with taking 31 compounds into account, 7.9 to 11.5% of the 70 μg/L creosote originally added to L-15/ex could be accounted for.

Cytotoxicity of creosote solutions

After RTgill-W1 cell cultures had been exposed to creosote solutions for 6 h in the dark, the two fluorescent dyes, alamar Blue and CFDA-AM, indicated cytotoxic responses. With both indicator dyes, fluorometric readings were greatly reduced at the highest nominal creosote concentration, which was 70 mg/L (Fig. 1). For the creosote sample that was obtained after 1 d of stirring, cytotoxicity was observed also at the second highest creosote concentration, 50 mg/L. For this creosote sample, EC50 values were 53 mg/L for alamar Blue and 52 mg/L for CFDA-AM and were the lowest observed for the cytotoxicity of the creosote solutions (Fig. 1). In contrast, EC50 values were highest for the creosote sample that had been stirred for 1 h before dosing. The EC50 values were 71 mg/L and 77 mg/L, respectively, for the alamar Blue and the CFDA-AM cytotoxicity assay (Fig. 1). Differences that were observed between the alamar Blue and CFDA-AM cytotoxicity assays were minor in the creosote samples that had been stirred between 1 h and 3 d. However, for the 7-d creosote sample, the alamar Blue assay revealed a higher level of cytotoxicity for 70 mg/L than did the CFDA-AM assay (Fig. 1).

The observed and predicted abilities of creosote solutions to be directly cytotoxic were compared (Fig. 2 and Table 5). For predicted ability, the concentrations of all identified compounds in the creosote solutions were converted to NECs and summed to express each nominal creosote dose for each creosote solution as an NEC (Tables 2 and 3). The NEC of each nominal creosote dose for each creosote sample was subsequently applied to the appropriate logistic functions to yield dose–response curves of predicted direct cytotoxicity (Fig. 2). Generally, the predicted curves were too shallow at concentrations at which cytotoxicity occurred, leading to an underestimation of the direct cytotoxicity by the creosote solutions. Thus, it became apparent that not all cytotoxicity could be accounted for by NECs. The lack of fit between the predicted and observed values, particularly at high creosote concentrations, was also reflected by the regression analysis (Table 5). The values for ^r2 were below 0.82 for most cases and/or the slope or y intercept varied greatly from their ideal values of one or zero, respectively.

Photocytotoxicity of creosote solutions

When RTgill-W1 cells were UV irradiated immediately after being exposed to creosote for 6 h in the dark, the two fluorescent dyes, alamar Blue and CFDA-AM, indicated more pronounced cytotoxic responses than dark controls. This photocytotoxicity was observed at concentrations above 1 mg/L. Except for minor differences, the two fluorescent dyes yielded similar dose–response curves for the creosote solutions that were obtained after 1 h to 3 d of stirring (Fig. 3). The EC50 values were lowest with the 1- and 3-d creosote samples but were slightly higher in the creosote solution that was obtained after 1 h of stirring (Fig. 3). For the 7-d creosote sample, the alamar Blue cytotoxicity assay yielded a dose-response curve and an EC50 value that were comparable to the dose-response curves and EC50 values obtained either with alamar Blue or CFDA-AM for the other three creosote samples. However, the CFDA-AM assay was less sensitive in the 7-d creosote sample, and resulted in an EC50 value that was three times above the EC50 value for alamar Blue (Fig. 3).

Previously, several PAHs were found to be strictly photocytotoxic, and fluoranthene was established as a reference compound for this mode of toxic action [9]. Among the photocytotoxic PAHs, four were found in all creosote samples. These PAHs were anthracene, fluoranthene, pyrene, and benzo[a]anthracene (Tables 1 and 4). When the concentrations of these compounds in the stock solution of 70 mg/L creosote in L-15/ex were corrected by their appropriate FEF [9], FECs were obtained (Table 4). The sum of FECs in each creosote sample and each nominal creosote dose was subsequently applied to the appropriate logistic functions to yield dose–response curves of predicted photocytotoxicity. If these predicted dose–response curves were compared to the observed photocytotoxicity data, a good agreement between the two was seen (Fig. 4). Linear regression analysis revealed high r² values along with values that were close to one and zero for the slopes and y intercepts, respectively (Table 5). Therefore, the four photocytotoxic PAHs, anthracene, fluoranthene, pyrene, and benzo[a]anthracene, were concluded to fully account for the observed photocytotoxicity of creosote to RTgill-W1 cells.

Cytotoxicity of photomodified creosote solutions

When RTgill-W1 cells were exposed for 6 h in the dark to creosote solutions that previously had been UV irradiated for 2 h, cytotoxicity continued to be detected for all four creosote solutions and with both assays of cellular activity (Fig. 5). However, in contrast to the cytotoxicity of unirradiated (intact) creosote, the two assays gave quantitatively different results. For the alamar Blue assay, cytotoxicity was seen at nominal creosote concentrations above 1 mg/L. However, for the CFDA-AM assay, nominal creosote concentrations of 5 mg/L and above were necessary to elicit an appreciable response (Fig. 5). Consequently, the EC50 values for the CFDA-AM assay were between 3.2 and 4.5 times higher than those for the alamar Blue assay.

Photocytotoxicity of photomodified creosote solutions

When RTgill-W1 cells were UV irradiated for 2 h immediately after being exposed to photomodified creosote for 6 h in the dark, photocytotoxicity continued to be detected for all four creosote solutions and with both assays of cellular activity (Fig. 6). In general, the alamar Blue and the CFDA-AM assays showed responses that were similar to those obtained for the cytotoxicity of photomodified creosote, although toxicity was much more pronounced for the two endpoints. For the alamar Blue assay, EC50 values were comparable to the EC50 values that were obtained for the photocytotoxicity of creosote and ranged from 2.1 mg/L for the creosote sample obtained after 1 d of stirring to 4.3 mg/L for the creosote sample obtained after 1 h of stirring (Fig. 6). For the CFDA-AM assay, EC50 values were 2.2 to 3.3 times above those obtained for the alamar Blue assay and ranged from 4.6 mg/L for 1 d of stirring to 11 mg/L for 7 d of stirring (Fig. 6).

Comparison of cytotoxicity and photocytotoxicity of intact and photomodified creosote

The toxic potency of creosote to RTgill-W1 cells was strongly influenced by UV radiation. The most potent combination was the simultaneous presence of creosote and UV radiation, the photocytotoxicity of creosote. Compared to creosote alone, toxicity was increased 35 and 23 times in the alamar Blue and the CFDA-AM assays, respectively, in the creosote sample that was solubilized for 3 d (Table 6). If the same creosote sample was photomodified before exposure of the cells in the dark, toxicity was 7.6 (alamar Blue) and 2.0 (CFDA-AM) times that of intact creosote, but was lower than the photocytotoxicity of either creosote or photomodified creosote (Table 6). Similar ratios were found for the creosote samples that were obtained after 1 h, 1 d, and 7 d of solubilization. The average increase in toxicity compared to intact creosote for all creosote samples was 27 ± 5.9 and 19 ± 5.7 times for alamar Blue and CFDA-AM in the photocytotoxicity of creosote, 8.4 ± 1.1 and 2.3 ± 0.4 times for alamar Blue and CFDA-AM in the cytotoxicity of photomodified creosote, and 20 ± 4.0 and 8.5 ± 1.9 times for alamar Blue and CFDAAM in the photocytotoxicity of photomodified creosote.

The cytotoxicity and photocytotoxicity of intact and photomodifed creosote caused different changes in cell morphology as judged by phase-contrast microscopy. For the cytotoxicity of intact and photomodified creosote, cells often were enlarged and their nuclei more distinct than in control cultures. However, membrane blebs were not observed and cells remained attached to the culture surface. In contrast, for the photocytotoxicity of intact and photomodified creosote at the highest concentrations, many cells rounded up and detached from the growth surface during the 2 h of UV irradiation. At lower concentrations many cells remained attached but had blebs.

Preparation and chemistry of creosote solutions

Creosote was solubilized in this study without the aid of a carrier solvent. Although the applied nominal creosote concentration was relatively high, chemical analysis showed compounds to be present in the final creosote solution at concentrations that well reflected those found in groundwater at different wood treatment sites [22, 23]. Compared to available data on the composition of the original creosote used in this study [24], creosote in L-15/ex was similar for major components such as phenanthrene, pyrene, and anthracene. However, naphthalene was found at lower concentrations than expected from the original creosote composition. This confirms studies by Tadokoro et al. [19], who found that the applied solubilization procedure led to creosote solutions that well reflected the original creosote with the exception of low-boiling chemicals. Alternatively, the relatively low naphthalene concentrations found in our study could be due to losses during the storage of the creosote before its application in the experiments described in this report.

Cytotoxicity of creosote solutions

The cytotoxicity elicited by creosote was similar in many ways to the direct cytotoxicity of individual low molecular weight PAHs [8]. Similar morphologic changes were observed in the cell cultures exposed in the dark to either creosote or individual low molecular weight PAHs. Both the alamar Blue and the CFDA-AM assay gave similar levels of impairment, which indicates that the two assays likely were measuring the same damage. This damage would be due to the specific inhibition of membrane-integrated enzymes or the general disturbance of membrane properties [8, 25]. In contrast, toxicity due to metabolic activation was unlikely to have occurred in the current experiments because of the absence of cytochrome CYP1A in the RTgill-W1 cell line, and because cytotoxicity occurred within such a short period.

High nominal creosote doses were necessary to elicit a cytotoxic response in RTgill-W1 cells. Comparisons to other studies that have used creosote are difficult because of the different ways in which creosote has been applied to test organisms and because the composition of creosote can vary significantly between manufacturers. However, mouse embryo cultures were affected by petroleum creosote at concentrations similar to those that were effective on RTgill-W1 cells [26]. Moreover, the supplementation of culture medium with a rodent hepatic S9 fraction did not modify the embryotoxicity of petroleum creosote.

Additivity has been an adequate model in the past for explaining the toxicity of mixtures of several creosote components, but in the current study such a model only partly described the cytotoxicity of the creosote solution. In the past, benzene, naphthalene, acenapthene, and 1-chloronaphthalene were shown to act additively in inhibiting mitochondrial respiration in vitro [27]. Naphthalene, acenaphthene, phenanthrene, and anthracene were slightly less than additive at equitoxic concentrations in eliciting acute toxicity to Daphnia magna, but additivity was still thought a sufficient model [28]. Presumably, in the current study, additional compounds were dissolved in the creosote solution that were not analyzed for chemically. These additional compounds could cause an underestimation of the cytotoxic potential of the creosote solutions in several ways.

The simplest cause of underestimation would be that one or more of the additional compounds was directly cytotoxic. In favor of this possibility is the fact that the chemistry of creosote is extremely complex, with up to 200 different compounds [14], whereas only 15 to 16 chemicals were identified in the current study and considered in the additive model. The phenols are one chemical class to consider. Phenols are frequently found in creosote [14], and have been shown to be cytotoxic to cultured fish cells [29] as well as to fish [30]. Another chemical class of potential importance is the polycyclic nitroaromatic hydrocarbons. In studies on the in vitro embryotoxicity of petroleum creosote, Iyer et al. [26] suggested that polycyclic nitroaromatic hydrocarbons were major contributors to the direct embryotoxicity. A different explanation for the underestimation is that the additional compounds might not be directly cytotoxic, but they could potentiate the cytotoxic actions of other compounds. Only further research with other complex solutions of aromatic hydrocarbons and more complete chemical analysis will allow these possibilities to be distinguished.

Photocytotoxicity of creosote solutions

The photocytotoxicity of creosote appeared to arise from photosensitized reactions that led to the formation of reactive oxygen species. This was because the responses of cells to UV irradiation and creosote solutions were similar to the responses to UV irradiation and individual PAH congeners, which appeared to be photocytotoxic through the production of reactive oxygen species [7, 9]. The similar responses were the morphologic changes of cells and the decline in cellular activities that occurred to the same extent in either the alamar Blue or CFDA-AM assays. The photosensitized reactions potentially could be due to intact PAHs or products that arise from the photomodification of the original sensitizing compound. Alternatively, photomodification products could be directly cytotoxic. However, if cytotoxic photomodification products should indeed arise, they appear to be less important in the presence of UV radiation. This is because if the creosote solution was photomodified before being added to the cells in the dark, toxicity was less than for the photocytotoxicity of creosote and the two fluorescent indicator dyes showed significantly different results. One exception in the photocytotoxicity of creosote is seen in the sample that was solubilized for 7 d. The different responses in the alamar Blue and CFDAAM assays hint at the formation of oxidized products whose impairment of the mitochondrial electron transport system occurs before general membrane damage. Such oxidized products could have formed spontaneously during the prolonged stirring process.

The photocytotoxicity of creosote solutions could be fully accounted for by four PAHs, using previously developed FEFs [9]. These four PAHs were anthracene, fluoranthene, pyrene, and benzo[a]anthracene. In a study on the ahototoxicity of several coal tar constituents to guinea pig skin, Kochevar et al. [31] found anthracene, fluoranthene, and pyrene to be most potent, whereas phenanthrene, chrysene, and carbazole were not phototoxic. Other researchers have found relatively high phototoxic potencies for anthracene, fluoranthene, pyrene, or benzo[a]anthracene in a variety of biological systems [3]. However, it has not been shown previously that these four compounds can account for the photocytotoxicity of a complex mixture such as creosote. This is particularly intriguing as two of the above congeners, namely fluoranthene and pyrene, are prevalent not only in creosote-contaminated aquatic environments [20, 23] but also in air ambient to aluminum plants, wood heating sources, and rural or urban areas, as well as in contaminated fresh water and sediments [32].

The application of FEFs in our study was done under the assumption that the photocytotoxic compounds act in an additive fashion. Research on the phototoxicity of defined mixtures has just begun. Hatch and Burton [33] recently showed that fluoranthene and anthracene were additive in their phototoxicity to benthic organisms. This is in agreement with the proposed similar mode of phototoxic action for PAHs, the generation of reactive oxygen species upon absorption of UV radiation. Hence, higher concentrations of a phototoxic compound or mixtures of phototoxic compounds will give rise to higher concentrations of reactive oxygen species and consequently enhanced toxicity.

Cytotoxicity of photomodified creosote solutions

Photomodifying creosote through UV irradiation must have generated cytotoxic products because irradiation caused the creosote solutions to be more cytotoxic. Previously, fluoranthene was shown to be extensively photomodified under similar experimental conditions [7], and because fluoranthene has a half-life in solar radiation that is intermediate among the detected PAHs [34], significant modifications of other PAHs in creosote can be expected as well. Most commonly, the photomodification of a single PAH leads to a complex mixture that contains, among other classes of compounds, quinones and diols [12, 13]. For example, 9,10-phenanthrenequinone (PHEQ) was found to be the major component and primary toxicant in phenanthrene-containing solutions that were exposed to actinic radiation [12]. This PHEQ was unique in that it not only was resistant to further photooxidation but also in that it did not support photosensitized reactions. With phenanthrene being the most abundant PAH in all creosote samples, PHEQ thus likely contributes to the cytotoxicity of photomodified creosote in the current study.

The alamar Blue assay was more sensitive than the CFDAAM assay in detecting an inhibition of cellular activity by the photomodified creosote solutions. The EC50s were three- to fourfold lower with alamar Blue. This indicates that photomodification products had specific actions in mitochondria. Such an impairment could be due to compounds that structurally mimic components of the electron transport chain. For example, quinones can receive electrons from NADH that are normally donated to ubiquinone. Alamar Blue and natural receptors would then compete for electrons with the quinones of the photomodified creosote solution, which potentially leads to a diminished reduction of alamar Blue and thus diminished fluorescent readings. The reduced quinone can pass on its electron to oxygen, forming superoxide radical and the parent quinone. The quinone can then repeat the cycle while the superoxide radical is converted into hydrogen peroxide by superoxide dismutase or into hydroxyl radical in the presence of iron by the Fenton reaction [35]. Thus, quinones can lead to yields of oxygen radicals that exceed the capabilities of cellular antioxidant defense mechanisms and result in a more general damage to cell membranes and proteins. In cultured fish cells, this general damage due to oxidative stress has been measured with CFDA-AM in this report and with Neutral Red in previous reports [36, 37].

Photocytotoxicity of photomodified creosote solutions

As with the photocytotoxicity of the original creosote solution, the photocytotoxicity of photomodified creosote appears to arise from photosensitized reactions. Support for this comes from the fact that in combination with UV irradiation both the intact and photomodified creosote solutions caused similar changes in cellular morphology and a level of impairment that was the same with either the alamar Blue or CFDAAM assays. These photosensitized reactions would have been initiated by parent compounds that remained after the photomodification process or by photomodification products that themselves act as photosensitizers. One such photomodification product that has recently been identified is 2-hydroxyanthraquinone [13].

Comparison of cytotoxicity and photocytotoxicity of intact and photomodified creosote solutions

The results of this study indicate that assessing the environmental impact of creosote is complex. Ultraviolet irradiation consistently increased the toxicity of creosote to fish cells in culture and, at least in the case of the photocytotoxicity of creosote, this was attributable to PAHs. A positive association in fish between high PAH exposures near creosote-contaminated sites and the induction of cytochrome P4501A1 [38] or the occurrence of liver neoplasms [39] has been reported and constitute other ways in which creosote can be toxic to aquatic organisms. For the UV radiation exposures, the highest toxicity was seen when creosote and the fish gill cells were present simultaneously. This photocytotoxicity of creosote led to a 35-fold increase in toxic potency. Exposure conditions in which photocytotoxicity would be expected to occur in the environment are near the wastewater efflux of wood treatment facilities. Wood treatment and storage facilities are major contributors of creosote entry into the environment [20]. Furthermore, the toxicity of creosote was enhanced when creosote was photomodified before exposure to the cell culture system. This cytotoxicity of photomodified creosote led to an eightfold increase in toxic potency and implies that UV irradiation does not necessarily result in the toxicity deactivation of complex mixtures. Borthwick and Patrick [40] tested the deactivation of creosote in salt water due to exposure to direct sunlight and measured a half-life of less than 1 week. This study indicates that prolonged UV radiation exposures will likely lead to a significant decrease in toxic potency under static conditions. However, if the supply of creosote-contaminated water is continuous, as would be expected at wood treatment sites, deactivation due to UV irradiation is not likely to occur, but in contrast, can lead to a further activation of toxicity.