Evaluation of Liposome−Water Partitioning for Predicting

Oct 27, 2009 - very robust partitioning phase; and (iii) Klipw values agreed very well with literature ... coefficients (3-5) and aqueous solubilities...
1 downloads 0 Views 274KB Size
Environ. Sci. Technol. 2009 43, 8854–8859

Evaluation of Liposome-Water Partitioning for Predicting Bioaccumulation Potential of Hydrophobic Organic Chemicals STEPHAN A. VAN DER HEIJDEN* AND MICHIEL T. O. JONKER Institute for Risk Assessment Sciences, Utrecht University, P.O. Box 80177, 3508 TD Utrecht, The Netherlands

Received July 28, 2009. Revised manuscript received October 6, 2009. Accepted October 19, 2009.

Considering the importance of bioaccumulation factors (BAFs) in risk assessment of chemicals and the ethical issues and complexity of the determination of these factors in standard tests with living organisms, there is a need for alternative approaches for predicting bioaccumulation. In this study, liposome-water partitioning coefficients as determined by using solid-phase microextraction (SPME) were evaluated for the cause of assessing bioaccumulation potential of hydrophobic organic chemicals (HOCs). To this end, the SPME method was mapped (in terms of mass balance, mode of spiking, kinetics, and reproducibility) and validated against literature data. Furthermore, the robustness of liposomes as partitioning phase was investigated (in terms of chemical loading, and pH and ionic strength of the medium), and finally liposome-water partition coefficients (Klipw) determined for polycyclic aromatic hydrocarbons (PAHs; 4.5 < logKow < 7.2) were compared with literature BAF values for several aquatic species. The results indicated that (i) SPME is a valid, fast, and reproducible method for measuring Klipw values; (ii) liposomes provide a very robust partitioning phase; and (iii) Klipw values agreed very well with literature PAH BAF values. SPME-derived Klipw values therefore seem a very promising predictor of bioaccumulation potential of HOCs. By including model- or in vitroderived biotransformation rates, bioaccumulation potential estimates might be converted into surrogate BAFs, thereby extending the applicability of Klipw values to metabolizable chemicals and species with more advanced biotransformation capacity.

Introduction For proper environmental risk assessment of hydrophobic organic chemicals (HOCs) insight in their bioaccumulation potential is imperative. This notion is recognized by the recently enacted European regulation on registration, evaluation, and authorization of chemicals (REACH), which requires bioaccumulation assessment of all bioaccumulative compounds (i.e., those having a logarithmic octanol-water partition coefficient [logKow] of >3) with a yearly production of more than 100 t (1). However, in vivo bioaccumulation testing of the 3025 chemicals meeting these criteria (2) may not be feasible due to associated high costs, experimental * Corresponding author e-mail: [email protected]; phone: +31 30 2535018; fax: +31 30 2535077. 8854

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 23, 2009

difficulties, and ethical considerations. Consequently, alternative methods for predicting the bioaccumulation potential of these chemicals should be explored. For this cause, experimentally derived chemical descriptors such as bulk solvent-water partition coefficients (3-5) and aqueous solubilities (6), and theoretically derived ones, i.e., linear solvation energy relationships (7), have been used as predictors for bioaccumulation factors (BAFs). However, the relationships between these descriptors and actual BAF values are still uncertain for increasingly hydrophobic chemicals (8). For some time now, artificial phospholipid vesicles (liposomes) have been used by several researchers for studying membrane-water partitioning of organic chemicals (9-17), as membranes are considered the major target site for partitioning and toxicity of these chemicals in organisms. Though the aforementioned studies predominantly addressed mechanistic issues, one (16) also suggested a good agreement between bioaccumulation and liposome-water partitioning, hence hinting at the possible use of liposomewater partition coefficients (Klipw) for predicting bioaccumulation potential. Liposome-water partitioning has been investigated by using several different methods, but not all of them are suitable for very hydrophobic compounds and some suffer from experimental difficulties (see Table S1 for a summary of the methods and their disadvantages). Considering Table S1, as well as methodological ease of performance and insensitivity to artifacts (i.e., third-phase effects or nonequilibrium conditions), the solid-phase microextraction (SPME) partitioning method (13, 16, 18) was selected from these methods and evaluated in the present study for the purpose of predicting bioaccumulation potential of HOCs. The first step in the evaluation included the investigation of method characteristics and performance by studying the mass balance of test systems, the comparability of Klipw data measured in systems dosed by conventional spiking (16) or by solid-phase dosing (19), equilibration kinetics in test systems spiked in either fashion, and method reproducibility as indicated by replicated Klipw measurements over a two-year period. Second, the method was validated against another method reported in the literature (9). Third, the robustness of liposomes as partitioning phase was explored in terms of chemical loading of liposomes and the degree of alkaline phospholipid hydrolysis to simulate liposome deterioration. Also, the dependence of the results on experimental parameters, such as incubation temperature, ionic strength of the medium, and type of phospholipid comprising the liposomes was studied. Fourth, Klipw values measured in this study for polycyclic aromatic hydrocarbons (PAHs) were compared to literature BAF values for ectothermic aquatic species to finally judge the method’s potential for assessing actual bioaccumulation.

Materials and Methods Chemicals, SPME Fibers, and Liposomes. The organic solvents acetone, cyclohexane, n-hexane (all Pestiscan grade), acetonitrile (HPLC grade), and methanol (Super gradient grade) were obtained from Lab-Scan (Dublin, Ireland). The chlorobenzenes 1,2,3,5-tetrachlorobenzene (97.2%), 1,2,4,5tetrachlorobenzene (99.1%), pentachlorobenzene (99.8%), and hexachlorobenzene (99.6%) were purchased from Riedelde Hae¨n (Seelze, Germany). PCB-52 (>98%) was obtained from Promochem (Wesel, Germany). The polycyclic aromatic hydrocarbons (PAHs) phenanthrene, anthracene, fluoranthene, pyrene, benz[a]anthracene, chrysene, benzo[e]pyrene, benzo[b]fluoranthene, benzo[k]fluoranthene, benzo[a]pyrene, ben10.1021/es902278x CCC: $40.75

 2009 American Chemical Society

Published on Web 10/27/2009

zo[ghi]perylene, dibenz[ah]anthracene, and indeno[1,2,3-cd]pyrene (all >98%) were supplied by Aldrich (Steinheim, Germany). Internal standards used were PCB-31 (99%; Dr. Ehrenstorfer) for chlorobenzenes and 2-methylchrysene (99.2%; Community Bureau of Reference, Geel, Belgium) for PAHs. Other chemicals used were calcium chloride, sodium azide, hydrochloric acid (37%), and sodium hydroxide (Merck, Darmstadt, Germany). Poly(dimethylsiloxane) (PDMS)coated disposable SPME fiber (glass core diameter 110 µm, 30 µm thick coating) was supplied by Polymicro Technologies (Phoenix, AZ). Prior to use, the fiber was cut into pieces of desired length, which were rinsed with methanol (3×) and Millipore water (3×), respectively. Dimyristoylphosphatidylcholine (DMPC; alkyl chain length (ACL) C14:C14) and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine (POPC; ACL C16:C18) were obtained from Avanti Polar Lipids (Alabaster, AL). Liposomes were produced at Eawag (Du ¨bendorf, Switzerland) as described in ref 20, in a buffer of 50 mg/L of sodium azide and 0.01 M calcium chloride. The diameter of the vesicles was typically in the range of 100 nm. The phospholipid content of the final suspensions was determined by total phosphate minus inorganic phosphate measurements (courtesy of AUA laboratories, Eawag, Switzerland). Liposomes were transported in a cooled package to our lab in Utrecht. Liposome-Water Partitioning. In the typical approach, POPC-based liposomes were used as they, unlike DMPCbased ones, reside in a liquid crystalline state at a standard incubation temperature of 20 °C, thus representing the dominant lipid phase in biological membranes. The liposome dispersion was weighed into an amber glass bottle of either 50 or 100 mL and the bottle was subsequently filled with Millipore water containing 50 mg/L of sodium azide and 0.01 M calcium chloride. Next, SPME fibers were added, after which the test system was spiked with a mixture solution of 13 PAHs in acetone. Spike volumes were maximally 0.2% of the volume of the test system. Thus-prepared bottles were shaken in the dark for 4 weeks at 120 rpm and 20 °C. After the equilibration period, SPME fibers were sampled, wiped with a wet tissue, and swiftly transferred to amber-colored autosampler vials filled with acetonitrile. Finally, internal standard was added. In the Supporting Information, the deviations from the above approach are described for the experiments aimed at studying the mass balance, solid-phase dosing, equilibration time, method reproducibility, similarities with literature Klipw data, and the effects of hydrolysis, liposome loading, calcium concentration, temperature, and phospholipid type. Klipw values were calculated according to the following equation: Klipw )

Clip (Qspike - Qf - Qw)/Mlip ) Cw Cf /Kf

in which the aqueous concentration (Cw) was calculated by dividing the concentration in the fiber (Cf; measured) by the fiber-water partition coefficient (Kf; for PAHs adopted from refs 16, 21, 22, depending on the temperature and ionic strength of the medium; for the reference compounds values were determined in the present studyssee Table S6). The concentration in liposomes (Clip) was derived by subtracting the chemical amount in the fiber (Qf) and the aqueous phase (Qw) from the total (spiked) chemical amount (Qspike) and subsequent division by the liposome mass (Mlip). Instrumental Analysis. PAHs were analyzed by HPLC as described in ref 16. Chlorobenzenes were analyzed by gas chromatography with electron capture detection (GC-ECD). The system consisted of a Fisons AS 800 autosampler, a Fisons HRGC 8000 GC oven, and a Carlo Erba ECD 40 electron capture detector with an ECD 400 controller (all: Milan, Italy). Extracts were injected on-column on a deactivated uncoated

precolumn (1.5 m × 0.53 mm), connected to a fused-silica DB-5.625 column (30 m × 0.25 mm; d.f. 0.25 µm; J&W Scientific, Folsom, CA). Helium was used as carrier gas. Comparison of Klipw Values with Reported BAFs. A database query was performed in ECOTOX (23) to obtain literature BAF values for PAHs in aquatic species. Data were critically evaluated, i.e., only lipid-normalized, steady-state, lab-determined BAFs for species known to hardly metabolize PAHs (mostly invertebrates) were selected. The resulting data set contained BAF values measured for fish eggs and larvae (24), the marine amphipod Rhepoxinius abronius (25), the freshwater isopod Asellus aquaticus (26), and the freshwater amphipod Pontoporeia hoyi (27). In addition, a recently published data set was included in which BAF values for Lumbriculus variegatus at 5, 13, and 25 °C were reported (21). To correct for temperature differences between the presently measured Klipw values for POPC (i.e., Kpopcw) and literature BAF values, the former were adjusted to the temperatures applied in the literature studies based on a Van’t Hoff plot. For all PAHs, a similar graph was produced by plotting the natural logarithm of the Kpopcw values measured at 20 and 30 °C (normal logarithmic values are presented in Table S2) against the reciprocal absolute testing temperatures and assuming a linear relationship, as also observed for partitioning of more polar compounds to liposomes (28) and for PAHs to animal lipids (21). From the slopes and intercepts of the resulting lines and the reciprocal absolute temperature derived from the literature BAF studies, Kpopcw values were calculated for the respective temperatures at which the BAFs were measured. Although this approach relies on two (replicated) data points per individual PAH only, as well as the assumption that linear extrapolation to lower temperatures is possible, it should be mentioned that the logKpopcw adjustments were always small (generally