Article pubs.acs.org/est
Partitioning of Polychlorinated Biphenyls into Human Cells and Adipose Tissues: Evaluation of Octanol, Triolein, and Liposomes as Surrogates Cristina L. Quinn,†,‡ Stephan A. van der Heijden,† Frank Wania,‡ and Michiel T. O. Jonker*,† †
Institute for Risk Assessment Sciences, Utrecht University, P.O. Box 80177, 3508 TD Utrecht, The Netherlands Department of Physical and Environmental Sciences, University of Toronto Scarborough, 1265 Military Trail, Toronto, Canada M1C 1A4
‡
S Supporting Information *
ABSTRACT: Whereas octanol, triacylglycerides, and liposomes have all been proposed as surrogates for measuring the affinity of hydrophobic organic contaminants to human lipids, no comparative evaluation of their suitability exists. Here we conducted batch sorption experiments with polyoxymethylene passive samplers to determine the partition coefficients at 37 °C of 18 polychlorinated biphenyls (PCBs) from water into (i) triolein (Ktriolein/water), (ii) eight types of liposomes (Kliposome/water), (iii) human abdominal fat tissues (KAFT/water) from seven individuals, and (iv) human MCF-7 cells cultured in vitro (Kcell/water). Differences between KAFT/water among individuals and between Kliposome/water among liposome types were very small and not correlated to structural attributes of the PCBs. Similarly, the length and degree of saturation of the phospholipid carbon chains, the headgroup, and the composition of the liposome did not affect the partitioning of PCBs into the studied liposomes. Whereas Kliposome/water values were similar to literature values of Koctanol/water adjusted to 37 °C, they both were lower than KAFT/water and Kcell/water by a factor of 3 on average. Partitioning of PCBs into triolein on the other hand closely mimicked that into human lipids, for which triolein is thus a better surrogate than either octanol or liposomes. Previously published polyparameter linear free energy relationships for partitioning from water into storage lipids and liposomes predicted the measured partition coefficients with a root-mean-square error of less than 0.15 log units, if the chosen equations and solute descriptors do not allow chlorine substitution in the ortho-position to influence the prediction. By guiding the selection of (i) a surrogate for the experimental determination and (ii) a method for the prediction of partitioning into human lipids, this study contributes to a better assessment of hydrophobic organic contaminant bioaccumulation in humans.
■
INTRODUCTION Hydrophobic organic compounds (HOCs), such as polychlorinated biphenyls (PCBs), are an important class of toxicants that accumulate predominantly in lipid tissues.1 A quantitative understanding of the partitioning of HOCs to lipids is therefore critical. When estimating the bioaccumulation potential of organic contaminants, octanol−water equilibrium partition coefficients (Koctanol/water) are typically used to represent the partitioning of chemicals from the aqueous phase to the storage (triacylglyceride) (Kstorage lipid/water) and membrane lipid compartments (Kmembrane lipid/water) of organisms.2−5 Recently, however, the suitability of this approach for aquatic organisms has been questioned.6 Furthermore, it has been reported that the sorption capacity of membrane lipids differs from that of storage lipids3 and that not all lipid homogenates of aquatic organisms have the same partitioning properties.6 For humans the situation is unclear, since measured coefficients for partitioning into human lipid tissues are not available. It is important, however, to know whether Koctanol/water accurately © 2014 American Chemical Society
predicts uptake into human lipid tissue; if not, the human bioaccumulation potential of HOCs may not be assessed correctly. This study sought to assess how well partitioning into octanol describes the actual, experimentally determined partitioning of PCBs to human tissues. Additionally, partitioning to human tissues was compared to partitioning into triolein and liposomes; phases that have been suggested previously as alternative surrogates for storage and membrane lipids, respectively.4,7 Even though this is, to our knowledge, the first study to investigate partitioning of HOCs into real human tissue, partitioning into surrogates has been studied before. Over the last 30 years, liposome−water partition coefficients (Kliposome/water) have been measured for both neutral5,7−9 and Received: Revised: Accepted: Published: 5920
January 7, 2014 April 25, 2014 April 28, 2014 May 7, 2014 dx.doi.org/10.1021/es500090x | Environ. Sci. Technol. 2014, 48, 5920−5928
Environmental Science & Technology
Article
times through a Millex-LG 0.20-μm hydrophilic PTFE filter (Merck Millipore, Billerica, MA, USA). Liposome suspensions were stored at 4 °C and used within 1 week. Eight differently composed liposomes were prepared in total in order to evaluate the influence of phospholipid chain length, degree of unsaturation, headgroup, and phospholipid mixing on the sorption of PCBs to liposome bilayers (Supporting Information (SI) Table S1). Phospholipids were selected based on their abundance in human cells. The two most common head groups of the phospholipids present in the membranes of female fat cells are phosphatidylcholine (PC) and phosphatidylethanolamine (PE),12 while palmitic (16:0), palmitoleic (16:1), stearic (18:0), and oleic (18:1) acid are some of the most commonly found fatty acids in human phospholipids,13 with the numbers in parentheses denoting the number of carbon atoms and double bonds in the fatty acid chain, respectively. The phospholipid types were also selected such that biologically relevant thermodynamically stable (i.e., liquid crystalline) bilayers could be formed at 37 °C, while still representing fatty acids of varying chain length and degree of saturation. For comparison purposes, it is preferable to prepare liposomes with only one type of fatty acid tail, but bilayers for all of the saturated fatty acids of interest would be in the ripple or gel phase at 37 °C, which show sorption behavior different from those in the liquid crystalline state.14 Furthermore, although liposomes produced from the pure PE selected (i.e., POPE, see SI Table S1) should reside in a bilayer state at 37 °C, as the lamellar−hexagonal transition does not ideally occur below 71 °C,15 more stable liposomes were prepared by mixing them with increasing ratios of PC liposomes.16 Comparison of pure POPE liposomes with 3 liposome mixtures of varying contributions of PC and PE (SI Table S1) should identify any destabilizing effect, if present. Human Tissues. MCF-7 cells from ATCC (Middlesex, UK) were selected as representative human cells. The MCF-7 cell line was originally derived from breast cancer cells from a 69-year-old Caucasian female. The cells were cultured in a humidified atmosphere at 37 °C/5% CO2 over a period of 2.5 months in order to obtain a sufficient number of cells. The culture medium was Dulbecco’s modified Eagle’s Medium, supplemented with 5% fetal calf serum, 1% penicillin/ streptomycin, 1% sodium pyruvate, 1% glutamine, and 1 μg/ mL insulin. Cells were stored at −20 °C until enough cells had been produced. All cells were combined and freeze-dried prior to the partitioning experiments. Seven abdominal fat tissues (AFT-A through AFT-G) were obtained from female patients aged 40−70, who undertook elective abdominoplastic surgery. Tissue subsamples free of skin were freeze-dried and subsequently homogenized. All handling of the AFTs was performed under strict biohazard conditions and approved by the medical ethical commission of the University Medical Centre of Utrecht. Partitioning Experiments. The partitioning of PCBs into each of the lipid systems was measured using batch sorption experiments with POM passive samplers, according to a method described previously.6,8 POM was used as passive sampler, because, contrary to other common sampler materials, it allows rigorous cleaning required to remove any lipid residue sticking to the sampler surface. The partitioning method has previously been validated for studying sorption to lipids in terms of linearity of isotherms, sorption competition, spiking approach, lipid and aqueous solubility limits, stability of sorption behavior of lipids, equilibration time, and ionic
ionogenic2,10 organic chemicals. Liposomes are vesicles composed of phospholipid molecules, which self-arrange to form a bilayered vesicle. The bilayer arrangement is driven by the hydrophilic headgroup of the phospholipid molecule and the hydrophobic tail consisting of two fatty acids.11 There are dozens of combinations of headgroup and fatty acid tails, but so far no systematic evaluation of the effect of phospholipid type on Kliposome/water for phospholipids in the liquid-crystalline phase has been conducted. Also, to our knowledge, measured Ktriolein/water or Kliposome/water values have not yet been compared with measured Kstorage lipid/water or Kmembrane lipid/water values, hence, the actual suitability of both surrogate phases has yet to be determined. Finally, the existing literature on PCB partitioning to triolein and liposomes is not consistent, since reported partition coefficients are smaller than, greater than, or similar to Koctanol/water.4,7 The main goal of this study was to evaluate the validity of using various surrogate phases (i.e., octanol, liposomes, and triolein) to represent PCB partitioning into human storage lipids and cells. To this end, partitioning of a series of 18 PCBs into three different lipid systems was investigated: (i) commercially available triolein, (ii) five different types of commercially available phospholipids and three mixtures thereof, representing the phospholipids that are most abundant in human cells, and (iii) lab-cultured human cells and surgically removed human abdominal fat tissues (AFT), as systems representing real human lipid tissues. Furthermore, the ability of polyparameter linear free energy relationships (ppLFERs) to predict partitioning of PCBs into human lipids was evaluated.
■
MATERIALS AND METHODS Chemicals. PCBs (congeners with IUPAC numbers 18, 28, 52, 66, 72, 77, 101, 103, 118, 126, 138, 153, 155, 156, 169, 170, 180, 187, and 209) were obtained from AccuStandard (New Haven, CT, USA) or LGC Standards (Wesel, Germany) and were >98% pure. Solvents used were methanol (Supra gradient grade), acetonitrile (HPLC-S grade), and acetone and hexane (both Pesti-S grade); they were purchased from BioSolve (Valkenswaard, The Netherlands). Chloroform (HPLC grade) was obtained from LabScan (Dublin, Ireland). Triolein (> 99%) was obtained from Sigma-Aldrich (St. Louis, MO, USA) and phospholipids were purchased from Avanti Polar Lipids, Alabaster, AL, USA. Chemicals used for cell culturing were Dulbecco’s modified Eagle’s Medium, fetal calf serum, penicillin/streptomycin, sodium pyruvate, and glutamine, which were all supplied by Life Technologies (Bleiswijk, The Netherlands), as well as insulin, that was obtained from SigmaAldrich (Zwijndrecht, The Netherlands). Polyoxymethylene (POM) was purchased as a sheet (thickness 77 μm) from CS Hyde Company (Lake Villa, IL, USA). Rectangular pieces were cut to the appropriate mass (4 or 20 mg, depending on the system), washed (via shaking with hexane (30 min), acetonitrile (2 × 30 min), and methanol (2 × 30 min)), and air-dried prior to use. Preparation of Liposomes. Liposome vesicles were prepared according to an adapted version of the procedure described by Khemtémourian et al.:12 each phospholipid solution in chloroform was transferred to a round-bottom flask and evaporated to dryness under nitrogen gas. The dried phospholipids were resolubilized in aqueous solution (50 mg/L NaN3 and 0.01 M CaCl2 in Millipore water), while keeping the suspension under a nitrogen flow. The resulting dispersion underwent 10 freeze−thaw cycles and was then extruded 10 5921
dx.doi.org/10.1021/es500090x | Environ. Sci. Technol. 2014, 48, 5920−5928
Environmental Science & Technology
Article
Table 1. Chlorine Substitution Patterns, Number of Ortho Chlorine Substitutions, and Various Partition Coefficients of the PCB Congeners Studieda PCB 18 28 52 66 72 77 101 103 118 126 138 153 155 156 169 170 180 187
2,2′,5trichloro 2,4,4′trichloro 2,2′,5,5′tetrachloro 2,3′,4,4′tetrachloro 2,3′,5,5′tetrachloro 3,3′,4,4′tetrachloro 2,2′,4,5,5′pentachloro 2,2′,4,5′,6pentachloro 2,3′,4,4′,5pentachloro 3,3′,4,4′,5pentachloro 2,2′,3,4,4′,5′hexachloro 2,2′,4,4′,5,5′hexachloro 2,2′,4,4′,6,6′hexachloro 2,3,3′,4,4′,5hexachloro 3,3′,4,4′,5,5′hexachloro 2,2′,3,3′,4,4′,5heptachloro 2,2′,3,4,4′,5,5′heptachloro 2,2′,3,4′,5,5′,6heptachloro
no. of ortho-Cl
Lb
log Koctanol/water at 25 °Cc
log Koctanol/water at 37 °Cd
log KAFT/watere
log Kcell/water
log Kliposome/waterf
log Ktriolein/water
2
7.48
5.63
5.48
5.72 ± 0.04g
5.63 ± 0.07
5.27 ± 0.06g
5.63 ± 0.04
1
7.904
6.05
5.87
6.28 ± 0.06
6.21 ± 0.06
5.70 ± 0.05g
6.12 ± 0.04
2
8.144
6.22
6.04
6.49 ± 0.07g
6.46 ± 0.06
5.90 ± 0.03
6.35 ± 0.05
1
8.716
6.56
6.40
6.69 ± 0.05g
6.65 ± 0.06
6.18 ± 0.04
6.60 ± 0.04
1
8.422
6.35
6.20
6.72 ± 0.05g
6.69 ± 0.06
6.20 ± 0.04
6.65 ± 0.05
0
9.205
6.63
6.47
6.82 ± 0.05
6.75 ± 0.08
6.32 ± 0.06g
6.72 ± 0.05
2
8.868
6.82
6.68
7.06 ± 0.05g
7.02 ± 0.08
6.42 ± 0.03
6.91 ± 0.04
3
8.429
6.76
6.60
7.07 ± 0.06
7.08 ± 0.06
6.40 ± 0.03
6.89 ± 0.04
1
9.396
7.00
6.83
7.23 ± 0.05
7.18 ± 0.05
6.56 ± 0.04
7.09 ± 0.04
0
9.884
7.05
6.89
7.34 ± 0.05
7.26 ± 0.05
6.74 ± 0.04
7.18 ± 0.04
2
9.772
7.34
7.18
7.57 ± 0.05
7.47 ± 0.04
6.94 ± 0.04
7.41 ± 0.06
2
9.587
7.34
7.15
7.68 ± 0.04
7.62 ± 0.05
6.96 ± 0.07
7.54 ± 0.04
4
8.715
7.29
7.12
7.77 ± 0.07g
7.66 ± 0.05
7.23 ± 0.03
7.73 ± 0.04
1
10.2
7.37
7.20
7.76 ± 0.05
7.68 ± 0.05
6.97 ± 0.04
7.56 ± 0.04
0
10.609
7.41
7.24
7.76 ± 0.06g
7.58 ± 0.06
7.25 ± 0.06g
7.78 ± 0.05
2
10.577
7.66
7.49
8.10 ± 0.05
8.04 ± 0.05
7.44 ± 0.06g
7.92 ± 0.04
2
10.415
7.66
7.49
8.11 ± 0.05
8.05 ± 0.05
7.37 ± 0.04
7.95 ± 0.04
7.39
8.22 ± 0.05
8.16 ± 0.06
7.55 ± 0.04
3
9.864
7.57
g
7.97 ± 0.05
a Partition coefficients are those for n-hexadecane−air (L), octanol−water (Koctanol/water), human abdominal fat tissue−water (KAFT/water), human MCF-7 cells−water (Kcell/water), liposomes−water (Kliposome/water), and triolein−water (Ktriolein/water). KAFT/water and Kcell/water values are lipidnormalized. bTaken from ref 27. cTaken from ref 21. dData from ref 21 adjusted for temperature using relationship from ref 22. eAverage of seven different abdominal fat tissues. fAverage of eight different liposome samples. gIndicates statistically significant differences between samples.
strength of the medium.6,8 Briefly, a lipid phase was weighed into a 50- or 100-mL amber-colored glass bottle and filled with a known amount of aqueous solution (0.01 M CaCl2 and 50 mg/L NaN3 in Millipore water) after which POM was added. Fifty μL of a cocktail solution of the PCBs in acetone (concentration approximately 3.5 mg/L per congener) were then spiked into the bottles. The bottles were sealed with glass stoppers and shaken on a reciprocal shaker at 150 rpm for 4 weeks at 37 °C. Afterward, the POM pieces were retrieved from each bottle, rinsed with Millipore water, thoroughly wiped with a wet tissue, and transferred to GC vials with 0.45 mL of acetonitrile. While POM retrieval and transfer was done at room temperature, the bottles were kept at 37 °C until immediately preceding retrieval such that the equilibrium would not be disturbed. All vials were stored in the freezer until analysis. Prior to the analysis, each vial was left at room temperature for 1 day, received 50 μL of analytical internal standard (PCB-209), and was vortexed for 2 min. For each lipid type, 4 or 5 replicate systems were prepared, plus 2 or 3 blanks (i.e., no PCBs added). All measurements were corrected for blanks.
Lipid Quantification. The phospholipid concentration in the liposome dispersions was determined gravimetrically. Lipids in MCF-7 cell and AFT material were quantified using the Bligh and Dyer method.17 The averaged lipid content of the 2− 4 replicates measured for each sample was used for lipid normalization of the AFT− and cell−water partition coefficients. The lipid contents measured are provided in SI Table S2. Instrumental Analysis. PCB analyses were done on a TRACE GC Ultra equipped with a Triplus autosampler and an electron capture detector (all Thermo Scientific, Waltham, MA, USA). Samples were injected on-column on a polar guard column, connected to a Zebron ZB-5Msi analytical column of 30 m × 0.25 mm with a 0.25-μm film thickness (Phenomenex, Torrance, CA, USA). After injection of a sample, the column temperature was kept at 80 °C for 1 min before increasing it first to 190 °C at 15.0 °C/min, then to 250 °C at 3.0 °C/min, and subsequently to 300 °C at 15.0 °C/min, where it was held for 5 min, before initiation of the cooling cycle. Chromatograms were processed manually using Chrom-Card 2.4.1 (Thermo Fisher Scientific Inc., Rodano Milan, Italy). 5922
dx.doi.org/10.1021/es500090x | Environ. Sci. Technol. 2014, 48, 5920−5928
Environmental Science & Technology
Article
Figure 1. Comparison of various partition coefficients with log Koctanol/water. The displayed partition coefficients are those between human MCF-7 cells and water (A), abdominal fat tissues of seven human individuals (AFT-A to AFT-G) and water (B), triolein and water (C), and eight different liposomes and water (D). Partition coefficients for cells and fat tissues are lipid-normalized. Panels C and D also include data from the literature. Koctanol/water values in panels A and B apply to 37 °C, whereas in panels C and D they apply to either 25 (literature data measured at 20−26.5 °C) or 37 °C (present data and literature data measured at 37 °C). The black 1:1 lines correspond to perfect agreement between the partition coefficients.
Data Analysis. Partition coefficients to the different materials were calculated by a mass balance approach described previously18 (see SI section 1 for details), using POM−water partition coefficients (KPOM/water) determined at 37 °C according to a method published before19 (see SI Table S4). Partition coefficients for AFT (KAFT/water) and cells (Kcell/water) were lipid-normalized by dividing the derived concentrations in the AFT samples and cells by the materials’ lipid fractions. Lipid normalization is not necessary for triolein (Ktriolein/water) and liposomes (Kliposome/water) as these materials are considered to be pure lipids.
All partition coefficients were compared with Koctanol/water. Despite an abundance of measurements, accurate and consistent Koctanol/water values for PCBs are difficult to obtain.20 In the main manuscript we use the Koctanol/water values reported by Sabljic et al.21 adjusted to 37 °C using log Koctanol/water (T ) = log Koctanol/water (25°C) − ΔUOW /(ln(10) ·R ) · (1/310.15K − 1/298.15K ) 5923
(1)
dx.doi.org/10.1021/es500090x | Environ. Sci. Technol. 2014, 48, 5920−5928
Environmental Science & Technology
Article
where ΔUOW is the internal energy of phase transfer between octanol and water in J/mol. Values for ΔUOW were taken from Schenker et al.22 when available or else calculated using22 ΔUOW = −23·MM − 16 200
SI Table S10). Thus, these differences probably are a result of experimental variation rather than an actual mechanism. In Figure 1, lipid-normalized Kcell/water and K AFT/water measured at 37 °C are plotted against the Koctanol/water values by Sabljic et al.21 adjusted to 37 °C (see Figure 1A and B, respectively). Koctanol/water underpredicts PCB partitioning into AFTs by a factor of 1.5−7.3 depending on the congener or a factor of 3.2 on average (Figure 1B). Likewise, PCB partitioning to lipids in cells is underpredicted by the Koctanol/water by a factor of 1.4−5.8 or a factor of 2.8 on average (Figure 1A). It should be noted that cells contain sorption phases for PCBs other than lipids;29 in particular proteins have been reported to sorb hydrophobic chemicals.30 The lipidnormalized Kcell/water values in Table 1 and Figure 1A may thus be overestimating the actual partitioning to cell lipids, especially considering the relatively low lipid content of the cells (13% of dry weight). Because the protein quantity and protein type of the cell material are unknown, the degree of overestimation can, however, not be quantified. On the other hand, the AFTs contained as much as 95−98% lipid (on a dry weight basis), so that KAFT/water values can be assumed to closely reflect partitioning to actual human (storage) lipids. Human bioaccumulation models often make use of Koctanol/water values at 25 °C, as the availability of 37 °C data is limited and temperature adjustments, as performed here, are not standard procedure and suffer from unknown and uncertain ΔUOW values. Therefore, the Kcell/water and KAFT/water measured at 37 °C were also compared with the Koctanol/water at 25 °C. The comparison shows slightly better agreement, because the Koctanol/water at 25 °C are somewhat higher (SI Figure S1). However, KAFT/water and Kcell/water are still on average higher by a factor of 2.2 (1.1−4.9) and 2.0 (1.0−3.9), respectively. Section 2 in the SI highlights that Koctanol/water values for the PCBs are quite uncertain and any comparison between lipid partitioning and octanol is thus dependent upon which set of Koctanol/water values is used. Also, it should be noted that partition coefficients should strictly be compared at the same temperature. Overall and despite the fact that the relationship between log KAFT/water and log Koctanol/water is linear and has only limited scatter, Figures 1 and SI 1 indicate that Koctanol/water at either 25 or 37 °C underpredicts partitioning of PCBs into both human adipose tissues and cell lipids and therefore presumably also into human lipids. The differences are relatively small, however, and in the context of human exposure assessment their importance has to be judged relative to other sources of uncertainty such as the variability in the lipid content of different individuals and the common assumption that only lipids contribute to the body’s sorption capacity.31 For many hydrophobic substances the uncertainty in the Koctanol/water value probably exceeds the observed discrepancies between Koctanol/water and KAFT/water.32 Measured PCB Partitioning into the Surrogate Materials Triolein and Liposomes. Table 1 lists log Ktriolein/water and log Kliposome/water values measured for the 18 PCBs (see SI Tables S4 and S5 for individual values for each replicate). Small, yet significant, differences between liposomes were found for PCBs 18, 28, 77, 187, 169, and 170, though again no consistent trend with the degree of chlorination or ortho-chlorination, molecular weight, or hydrophobicity of the PCB congeners was apparent (see SI Table S10). The partition coefficients for the other PCBs were not significantly different between liposomes, indicating that neither the length nor
(2)
where MM is the molecular mass of the congener. While eq 2 has poor predictive power (R2adj = 0.13), ΔUOW is relatively small and when used over a small temperature range (12 K), even a highly uncertain ΔUOW does not introduce undue uncertainty into the temperature-adjusted Koctanol/water. The SI (Section 2) additionally contains comparisons with the Koctanol/water values at 25 °C by Sabljic et al.,21 as well as those by Hawker and Connell,23 and the final adjusted values by Schenker et al.22 Statistical analysis to determine significant differences between AFTs and liposomes was accomplished using the one-way ANOVA function in Microsoft Excel. Polyparameter Linear Free Energy Relationships. The experimentally determined partition coefficients were further compared with Kstorage lipid/water and Kmembrane lipids/water values predicted by ppLFERs by Geisler et al.24 and Endo et al.,3 respectively. These authors presented two different versions of the ppLFER equations: Type 1: log K = eE + sS + aA + bB + vV + c
(3)
Type 2: log K = lL + sS + aA + bB + vV + c
(4)
Type 1 was proposed by Abraham et al.25 and type 2 was proposed by Goss.26 Both can be used to predict the partitioning of neutral organic compounds between any two condensed phases. The descriptors e, s, a, b, v, and l quantify the different capacities of the two condensed phases for interacting with a solute, which itself is characterized by a set of descriptors E, S, A, B, V, and L. The descriptor E describes the solute’s excess molar refraction, S is its dipolarity/polarizability, A and B are the H-bond acidity and basicity, V represents its molar volume, and L is the logarithm of its hexadecane−air partition coefficient. Two sets of solute descriptors exist for PCBs.27,28 The more recent data set by Van Noort et al.28 updates the descriptor values for E, S, B, and V, but retains the values for A and L from Abraham and Al-Hussaini27 in an attempt to reduce the discrepancy between ppLFER predictions and Koctanol/water observed for ortho-substituted PCBs. Because there were no ortho-substitution-related effects observed in this study (see below), predictions of Kstorage lipid/water and Kmembrane lipids/water for PCBs were obtained using the ppLFER of type 1 and the descriptors from Abraham and Al-Hussaini.27
■
RESULTS AND DISCUSSION Measured PCB Partitioning into Human Adipose Tissues and Cells. The averaged measured lipid-normalized log KAFT/water and lipid-normalized log Kcell/water values for the PCBs are listed in Table 1 (see SI Tables S3 and S4 for individual values for each replicate). Within the measurement uncertainty, the various partition coefficients for the seven abdominal fat tissues were statistically identical for most PCBs. There were, however, small, but statistically significant differences in the measured KAFT/water for PCBs 18, 52, 66, 72, 101, 155, and 169. Whether or not statistical differences were observed between individuals was not correlated to the degree of chlorination or chlorination in the ortho-position, molecular weight, or hydrophobicity of the PCB congeners (see 5924
dx.doi.org/10.1021/es500090x | Environ. Sci. Technol. 2014, 48, 5920−5928
Environmental Science & Technology
Article
Figure 2. Comparing (A) lipid-normalized log Kcell/water against the average log Kliposome/water; (B) lipid-normalized log Kcell/water against log Ktriolein/water; (C) lipid-normalized log KAFT/water against the average log Kliposome/water; and (D) lipid-normalized log KAFT/water against log Ktriolein/water. The black 1:1 lines correspond to perfect agreement between the partition coefficients.
Koctanol/water, Jabusch and Swackhamer7 found that Ktriolein/water was almost identical to Koctanol/water for congeners with 4.7 < Koctanol/water < 7.2 and about a factor of 2 greater than Koctanol/water in case of some very hydrophobic hexa- and heptachlorobiphenyls. However, Jabusch and Swackhamer’s7 Koctanol/water values deviate from the Koctanol/water from Sabljic et al. at 25 °C. Compared to the latter values, the Ktriolein/water values from Jabusch and Swackhamer differ by a factor of 0.3− 3.6. In other words, when compared against the same Koctanol/water data set, our measurements show a similar pattern (relative difference with Koctanol/water) as those by Jabusch and Swackhamer7 (see Figure 1C). Chiou4 reported Ktriolein/water for PCB congeners at 20 °C to be greater than Koctanol/water (taken from refs 33 and 34) when log Koctanol/water < 5 and to be smaller than Koctanol/water when log Koctanol/water > 5, i.e., the relationship between their Ktriolein/water and Koctanol/water has a slope
