Contents lists available at SciVerse ScienceDirect
Toxicology Letters
jou rn al h om epa ge: www.elsev ier .com/ locate / tox le t
he dynamics of accumulation of PCBs in cultured adipocytes vary with the cellipid content and the lipophilicity of the congener
ophie Boureza, Carine Van den Daelena, Soazig Le Layb, Jacques Poupaertd, Yvan Larondellea,ean-Pierre Thomée, Yves-Jacques Schneidera, Isabelle Dugail c, Cathy Debiera,∗
Institut des Sciences de la Vie, UCLouvain, B-1348 Louvain-la-Neuve, BelgiumLUNAM Université, INSERM U1063, Angers, FranceUniversité Pierre et Marie Curie – Paris 6, UMR S 872, Paris F-75006, FranceLouvain Drug Research Institute, UCLouvain, B-1200, Woluwe-Saint-Lambert, BelgiumLaboratoire d’Ecologie animale et d’Ecotoxicologie, Université de Liège, Allée du 6 Août 15, B-4000 Liège, Belgium
i g h l i g h t s
Accumulated amounts of PCBs in cells differ according to the in vitro model.Accumulated levels of PCBs were strongly correlated with lipid levels in cells.Dynamics of accumulation importantly differed according to the PCB congener.The log P was a major parameter governing PCB dynamics across the plasma membrane.
r t i c l e i n f o
rticle history:eceived 31 July 2012eceived in revised form4 September 2012ccepted 26 September 2012vailable online 16 November 2012
eywords:ccumulation dynamics
a b s t r a c t
Lipophilic pollutants such as polychlorinated biphenyls (PCBs) accumulate in high amounts in the adi-pose tissue. Recent epidemiological studies correlate their presence in fat cells to possible alterationsin the regulation of lipid metabolism. The factors governing their accumulation dynamics, storage andrelease in/from fat cells remain however unclear. Several in vitro models of cultured adipocytes can beused to address these questions. Nevertheless, the cell culture system as well as the PCB congener mayinfluence the behavior of such pollutants toward adipocytes and thus the results obtained. In the presentstudy, we compared the accumulation of 3 PCB congeners (PCB-28, -118 and -153) during a 4-h periodin two common models of cultured adipocytes (mouse embryonic fibroblasts (MEFs) differentiated into
CBsdipocyte modelsriglycerides
adipocytes and differentiated 3T3-L1 cells). The results show that adipocytes from different models accu-mulate significantly different amounts of a same pollutant added at the same initial concentration in theculture medium. These amounts were strongly correlated to the amounts of triglycerides stored in cells.Moreover, the dynamics of accumulation varied between the three congeners, PCB-28 entering the cellsmore rapidly than the two other congeners. The lipophilicity of these molecules, shown by the partitioncoefficient (log P) appears to be a major parameter governing their uptake dynamics in fat cells.
. Introduction
Polychlorinated biphenyls (PCBs) are persistent environmental
ollutants that tend to accumulate in lipid-rich tissues of the orga-isms they contaminate, due to their highly lipophilic character.he adipose tissue therefore constitutes one of the most significant
Abbreviations: PCBs, polychlorinated biphenyls; MEFs, mouse embryonic fibrob-asts.∗ Corresponding author at: Institut des Sciences de la Vie, UCLouvain, Croix du
internal reservoirs of such organic pollutants (Barouki and Clément,2009; Mullerova and Kopecky, 2007). Data from recent epidemi-ological studies suggest a role for these chemicals in the basicmechanisms controlling the regulation of the energetic balance(Dirinck et al., 2011; Kim et al., 2011). PCBs and other lipophilicpollutants could indeed potentially be involved in the obesity epi-demic (Dirinck et al., 2011; Lee et al., 2011). Moreover, it has beenshown that during periods of body weight loss, PCB concentrationincreases in the adipose tissue due to lipid mobilization (Kim et al.,
2011). During such a physiological state, PCBs are also released intothe blood circulation, potentially exposing the individual to thevarious known adverse health effects of these pollutants (Chevrieret al., 2000; Irigaray et al., 2006; Kim et al., 2011). There is thus
growing interest regarding the behavior of lipophilic pollutantsuch as PCBs toward adipocytes.
To date, little information is available about the mechanismsnvolved in the storage and release of such lipophilic pollutantsn/from the adipose tissue. Even if numerous in vitro models ofultured adipocytes have been extensively used to study the mech-nisms involved in adipogenesis, their use for purposes such ashe characterization of the accumulation dynamics, storage andelease of lipophilic toxicants in/out of this endocrine tissue isess described. Important questions such as whether adipocytesriginating from different in vitro models accumulate pollutantsimilarly or not, has never been addressed to our knowledge. Inddition, it is unclear if the type of PCB congener investigated influ-nces its behavior toward adipocytes.
In the present study, we evaluated and compared theynamics of accumulation of a cocktail of 3 PCB congeners,,4,4′-trichlorobiphenyl (PCB-28), 2,3′,4,4′,5-pentachlorobiphenylPCB-118) and 2,2′,4,4′,5,5′-hexachlorobiphenyl (PCB-153) addedt the same concentration in the medium of two in vitro modelsf adipocytes: mouse embryonic fibroblasts (MEFs) differentiatednto adipocytes and the very commonly used 3T3-L1 preadipocyteell line. We followed the incorporation of these three pollutants indipocytes over a 4-h period and observed differences of accumu-ation profiles. Considering each congener separately, the amountsccumulated in both in vitro culture models from the same initialmounts added in the medium differed importantly according tohe model. The extent of PCB uptake was highly correlated to themounts of triglycerides in the fat cells. Considering each in vitroodel individually, we observed notable differences in the accu-ulation profiles of each PCB congener, although these pollutants
nly differ by the number and position of chlorine atoms on theiphenyl backbone.
. Materials and methods
.1. Cell cultures
.1.1. 3T3-L1 cell line3T3-L1 cells (kind gift of Dr. J. Pairault, Paris, FR) were cultured in 6-well plates
n high-glucose Dulbecco’s modified Eagle’s medium (DMEM, Invitrogen, Carls-ad, CA) supplemented with 10% fetal calf serum (FCS, Invitrogen) until 2 daysost-confluence (day 0). Adipocyte differentiation was induced as described in Leay et al. (2006) by adding isobutylmethylxanthine (IBMX (250 �M), dexametha-one (1.25 �M), insulin (250 nM) (all from Sigma–Aldrich, St. Louis, MO), penicillin100 U/ml, (Invitrogen)) and streptomycin (0.1 mg/ml, (Invitrogen)) for 2 days. Theells were then cultured for 2 more days in the medium just described but withoutBMX and dexamethasone and with lower insulin concentration (100 nM). For theest of the differentiation process, insulin was also suppressed from the mediumhat was then replaced every 48 h until experiments.
.1.2. Mouse embryonic fibroblasts (MEFs)MEFs were prepared as described in Le Lay et al. (2006). Adipocyte differentia-
ion was induced during 2 days as described for 3T3-L1 cells except for the initialoncentrations of IBMX (500 �M) and insulin (870 nM). For the rest of the differenti-tion process, IBMX and dexamethasone were removed and rosiglitazone (0.5 �M)as added to the differentiation inducer cocktail every 48 h until experiments.
.2. Cell treatments
Once the differentiation achieved, the cells from both adipocyte models werencubated during 4 h (or up to 24 h for the experiments where PCB recov-ry was controlled) at 37 ◦C–10% CO2 with a cocktail of 3 PCB congeners,,4,4′-trichlorobiphenyl (PCB-28); 2,3′ ,4,4′ ,5-pentachlorobiphenyl (PCB-118) and,2′ ,4,4′ ,5,5′-hexachlorobiphenyl (PCB-153) (Dr. Ehrenstorfer GmbH, Ausburg, DE).CBs were added in the medium of the cells as an ethanolic solution to reach anal concentration of 50 ng/ml for each PCB in the medium of the cells. Control cellseceived the ethanol vehicle alone that did not represent more than 0.5% (v/v). Cyto-
oxicity of the PCBs was assessed by measuring the activity of lactate dehydrogenaseLDH) released into the extracellular medium of treated cells, using the Cytotoxicityetection Kit (Roche diagnostics, Mannheim, DE) according to the manufacturer’s
nstructions, and using 1% (v/v) Triton-X 100 (Sigma–Aldrich) as full toxicity control.ell morphology was also daily observed by phase contrast microscopy.
tters 216 (2013) 40– 46 41
2.3. Determination of PCB concentration
After PCB exposure, the culture media were removed and cells were collectedin lysis buffer (35 mM sodium dodecyl sulfate, 10 mM EDTA, 60 mM Tris buffer;pH = 7.2). Five milliliters of n-hexane (Burdick & Jackson Brand, Muskegon, MI) wereadded to the samples in an EPA vial (Alltech, Lokeren, BE) for liquid–liquid extrac-tion of the PCBs. After a thorough 5 min shaking, the samples were stored at 4 ◦Cuntil further analysis. All prepared samples (media and cells) were then purifiedand analyzed as described in Debier et al. (2003). Quantification was performed bycomparison with external standards of the analyzed components in a certified cal-ibration mixture (Ultra Scientific, North Kingstown, RI and Dr. Ehrenstorfer), usinga linear calibration curve for each PCB congener (concentration ranging from 1 to150 pg/�l). The PCB recovery was calculated on the basis of the concentration ofthe surrogate standard (IUPAC 112, Dr. Ehrenstorfer) (50 pg/�l), which was addedto the samples at the beginning of the clean-ups.
2.4. Protein assay
Protein concentration was determined using the Bicinchoninic Acid ProteinAssay kit (Sigma–Aldrich), with a bovine serum albumin (BSA, Sigma–Aldrich) cali-bration curve.
2.5. Determination of triglyceride concentration and fatty acid profiles
Aliquots of 200 �l of all cell lysates were transferred into microtubes followedby the addition of 200 �l of 0.1 M KOH in methanol (Labscan, Gliwice, PO) in a 70 ◦Cwater bath with gentle shaking for 1 h. Samples were thoroughly vortexed for 10 severy 20 min. After saponification, the samples were centrifuged at 17,000 × g for10 min. Total released glycerol was then measured by using a clinical kit accord-ing to the manufacturer’s recommendations (Free glycerol FS, DiaSys, Holzheim,DE). Glycerol being also present in the backbone of membrane phospholipids, thisfraction was quantified by gas chromatography for each cell condition (Schneideret al., 2012). The quantity of glycerol corresponding to the phospholipids was thensubtracted from the total amount quantified by the glycerol enzymatic kit aftersaponification. The fatty acid profiles of cell triglycerides were analyzed by gaschromatography as described in Schneider et al. (2012).
2.6. Physico-chemical parameters of PCBs
Physico-chemical parameters were calculated using the molecular propertymodule present in Chem3D Ultra 12.0 (CambridgeSoft, Cambridge, MA). Structureswere generated from ChemDraw (CambridgeSoft) files and energy-minimized firstin a water periodic box (2 nm × 2 nm × 2 nm) using the OPLS molecular mechanicsmethod as implemented in Hyperchem Professional 8.0 (Hypercube, FL) and fur-ther refined using the semi-empirical AM1 method down to a gradient of 0.01 usingthe Polak–Ribiere conjugate gradient method. Molecular surfaces representing theelectrostatic potential contours of the three congeners were then acquired from theabove resulting structures and displayed from Chem3D Ultra 12.0. The grid was setat 60 × 60 × 60 voxels. The isopotential value was set at 1.0000 eV.
2.7. Statistical analysis
The statistical analysis of the data was performed by SAS Institute Inc. Software.Comparisons between cell groups were made by two- and one-way ANOVA. Datawere considered statistically significant at p-values <0.05.
3. Results and discussion
3.1. In vitro models of adipocytes
In the present study, two in vitro models of adipocytes (MEFsdifferentiated into adipocytes and 3T3-L1 cells) were used to com-pare their potential to accumulate PCBs. All cells were cultured anddifferentiated according to their proper required experimental con-ditions, as described in the Materials and methods (Sections 2.1.1and 2.1.2). Once differentiated into adipocytes, obvious differencesof triglyceride content appeared in adipocytes from both in vitromodels (Fig. 1A and B). Subsequent quantification of neutral lipidsin those cells confirmed that they were indeed characterized bysignificantly different triglyceride levels (p < 0.05): MEFs accumu-lated 20.7 ± 3.5 ng triglycerides/�g proteins and 3T3-L1 adipocytes
stored up to 51.4 ± 6.0 ng triglycerides/�g proteins (Fig. 1C).
The main fatty acids encountered in both cell models wereeven chained saturated, monounsaturated and polyunsaturatedfatty acids (Table 1). In addition, some odd chain fatty acids were
42 S. Bourez et al. / Toxicology Letters 216 (2013) 40– 46
F lasts
i g of toc
iTCtamoa
TFibe
ig. 1. Phase contrast photographies of differentiated (A) mouse embryonic fibrobn MEF and 3T3-L1 adipocytes. Results are expressed as ng of triglycerides per �onditions (p < 0.05).
dentified in 3T3-L1 as also observed by Roberts et al. (2009).he following seven fatty acids, i.e. C14:0, C16:0, C18:0, C14:1,16:1, C18:1 (n − 9) and C18:1 (n − 7), accounted for up to 97% ofotal quantified triglyceride fatty acids in MEFs and 98% in 3T3-L1dipocytes. The fact that both cell models accumulated the same
ajor fatty acids can be explained by the high lipogenic activity
f cultured adipocytes, which essentially store fat by de novo fattycid biosynthesis from non lipid substrates, mainly carbohydrates,
able 1atty acid profiles of triglycerides in both in vitro cell models. Results are expressedn ng fatty acids/�g triglycerides (TG) and show similar fatty acid compositionsetween both models but significantly different amounts of most fatty acidsxpressed per unit of fat.
MEFs 3T3-L1
Fatty acids Mean (ng/�g TG) SD Mean (ng/�g TG) SD p Value
a Significantly different fatty acid profiles between the adipocyte models.
(MEFs) and (B) 3T3-L1 adipocytes. Bar, 50 �m. (C) Cell triglyceride quantificationtal cellular proteins and show significantly different accumulation levels in both
present in high concentration in the culture medium (Collins et al.,2011). However, it is important to notice that, even if the cells accu-mulated the same major fatty acids during their differentiation intoadipocytes, the concentrations of most fatty acids expressed perunit of cell triglycerides significantly differed between the models(p < 0.05) (Table 1). Taken together, these results show that the totalamounts of triglycerides and the proportion of each fatty acid perunit of fat importantly differ between both models.
3.2. Control of PCB recovery
PCBs being highly lipophilic molecules, it was important to ver-ify that they did not significantly adsorb to the plastic materialsused in our assays (polystyrene), before launching incorporationstudies of PCBs into adipocytes. We therefore controlled the recov-eries and summed up the amounts of PCBs quantified in the mediaand cells at each time point of the incubation. Total quantities ofPCBs in all experimental units corresponded to the initially addedamounts in the medium. Percentages of recovery in all assays werecomprised between 90 and 125% (Fig. 2), showing that there wasno significant loss of PCBs by adsorption on culture materials dur-ing the incubation period (p > 0.05). The sequestration of PCBs inthe medium rather than on the plastic materials resulted from thepresence of 10% fetal bovine serum (FBS) in the culture media,rich in BSA and diverse lipoproteins. Indeed, the data obtainedwith non FBS-containing media revealed adsorptions rates of PCBsonto the plastics of over 20% of the initially added amounts withinonly 2–3 min of incubation and up to 75% in 24 h of incubation.Such results are actually not surprising seeing that in vivo, PCBsare known to be transported in the bloodstream through theirtight association with plasma proteins, especially lipoproteins and
serum albumin (Becker and Gamble, 1982; Hjelmborg et al., 2008;Matthews et al., 1984; Spindler-Vomachka et al., 1984). Our datathus show that a significant percentage of FBS in the cell culturemedium prevents the adsorption of PCBs onto the plastic of the
S. Bourez et al. / Toxicology Letters 216 (2013) 40– 46 43
Fig. 2. Percentage of recovery of PCBs-28, -118 and -153 after 24 h of incubationwith cultured adipocytes. Results represent the percentage of the initial amountsaed
cic
3a
iiiMw1PMta
srPoalaoetPsPMutbi
eddts3dcpdc
Fig. 3. Accumulation rates of PCB-28 (A), -118 (B) and -153 (C) added during 4 h as acocktail (final concentration in the medium of 50 ng/ml each) on MEFs differentiatedinto adipocytes (black circles) and differentiated 3T3-L1 cells (white circles). Results
dded in the medium that were recovered at each time of the incubation during thexperiment and show that no significant losses of PCBs occurred on plastic materialsuring incubations (p > 0.05).
ell culture materials. This means that in such culture conditions,t is possible to conduct in vitro toxicological studies with PCBs onultured cells without having to use glass materials.
.3. Differences of PCB accumulation according to the testeddipocyte model
The extent of accumulation of each PCB congener differedmportantly according to the in vitro model used. Indeed, after 4 h ofncubation, the percentage of accumulation in cells of comparablenitial amounts of PCB-28 added in the medium ranged from 23% in
EFs up to 81% in 3T3-L1 adipocytes (Fig. 3A). Similar observationsere made with PCB-118 for which the accumulation rates were
3% in MEFs and 82% in 3T3-L1 adipocytes (Fig. 3B) as well as forCB-153 for which the quantified intracellular levels were 13% inEFs and 77% in 3T3-L1 adipocytes (Fig. 3C). For each congener,
he extent of uptake was always significantly higher in 3T3-L1dipocytes than in MEFs (p < 0.05).
As shown in Section 3.1, intracellular triglyceride contentsignificantly differed between the adipocyte models. Pearson’s cor-elation coefficients (r) showed that the amounts of accumulatedCB-28, -118 and -153 were highly correlated to the cellular levelsf triglycerides (r = 0.96; 0.98 and 0.94 for each PCB respectively,ll p < 0.01), suggesting a direct role for the quantity of intracellularipids in the storage potential of adipocytes. Indeed, when the totalmounts of PCBs were expressed per unit of triglycerides after 4 hf incubation, differences of accumulated levels between the mod-ls became non significant (p > 0.05), respectively 22.9 ± 2.7 mmolotal PCBs/mol triglycerides for MEFs and 18.8 ± 2.0 mmol totalCBs/mol triglycerides for 3T3-L1 adipocytes. When PCBs were con-idered separately, no significant differences were observed forCBs-118 and -153 (p > 0.05) expressed per unit of fat betweenEFs and 3T3-L1 adipocytes. PCB-28 concentrations expressed per
nit of fat were however significantly higher in MEFs as comparedo 3T3-L1 adipocytes (p < 0.05). This observation points out a proba-le physico-chemical related property of each pollutant, governing
ts behavior in fat cells, as discussed in Section 3.4.An important observation to be drawn from these results is that,
ven if the global amount of fat present in adipocytes was a crucialriving force for PCB accumulation, the fatty acid profile of the cellsid not appear to influence the extent of PCB accumulation. Indeed,he amounts of fatty acid expressed per unit of triglycerides showedignificant differences between the models (Table 1 and Section.1) whereas total PCB levels, also expressed per unit of triglyceri-es did not. This means that, in the classical experimental growth
onditions for adipocytes where no exogenous fatty acids were sup-lemented in the medium, the composition of cellular triglyceridesid not significantly influence the accumulation levels of PCBs inells. However, the precise role and contribution of certain families
are expressed as the percentage of the initially added amounts in the medium thatwere quantified in cells after 4 h of incubations with the pollutants and show thatsignificantly higher levels of each congener accumulated in 3T3-L1 adipocytes ascompared to MEFs.
of fatty acids to the storage capacity of adipocytes for lipophilic
pollutants like PCBs certainly warrants more investigation beforeconclusions can be drawn. This issue is of particular interest consid-ering that it has become common to enrich our diets with diversefatty acids such as the omega-3 fatty acid series (Ruxton et al.,
4 logy Letters 216 (2013) 40– 46
2to9icipibft(to
3c
htdtii2f3abfrei1
tcaocbb(cI52pitla5h
t
Fig. 4. Accumulation rates of PCB-28 (black circles), -118 (white circles) and -153(black triangles) added during 4 h as a cocktail (final concentration in the medium of50 ng/ml each) on MEFs differentiated into adipocytes (A) and differentiated 3T3-L1
TP
4 S. Bourez et al. / Toxico
004), which will have a direct impact on the fatty acid composi-ions of the tissues. We have recently shown that the lipid dropletsf adipocytes were the major intracellular targets for PCBs, as up to8% of the added PCBs in the medium were found to be accumulated
nto these organelles (Bourez et al., 2012). A significant modifi-ation of the fatty acids composing the lipid droplets could thusnfluence their capacity of storing lipophilic pollutants. In addition,olyunsaturated fatty acids are known to modulate membrane flu-
dity and thus its permeability. A study on the gram-negative soilacterium, Ralstonia eutropha, has shown that alterations in theatty acid composition and fluidity of the cell membranes affectedhe accumulation of a tetrachlorobiphenyl congener in the bacteriaKim et al., 2002). It would thus be interesting to further considerhe possible impact of membrane fatty acids on the incorporationf PCBs in adipocytes.
.4. Differences of PCB accumulation according to the type ofongener
Although we have recently shown that PCBs-28, -118 and -153ave the intracellular lipid droplets as a common final accumula-ion target (Bourez et al., 2012), their kinetics of uptake importantlyiffered in the present study, especially regarding PCB-28. Indeed,his congener entered the cells more rapidly than the two othersn both adipocyte models. The effect was however more markedn MEFs. After only 1 h of incubation in MEF derived adipocytes,0 ± 5% of PCB-28 had accumulated in cells against only 6 ± 2%or PCB-118 and 5 ± 2% for PCB-153 (Fig. 4A) (p < 0.05). In theT3-L1 adipocytes, which contain more triglycerides, we observedccumulation rates reaching 49 ± 5% for PCB-28 after 1 h of incu-ation with the pollutants, which was significantly higher thanor PCBs-118 and -153 for which the accumulated amounts wereespectively of 32 ± 3% and 28 ± 1% (Fig. 4B) (p < 0.05). In both mod-ls, the rate of entrance of PCBs in adipocytes during the first hour ofncubation respected the following order: PCB-28 > PCB-118 > PCB-53.
The number and/or the position of the chlorine substituents onhe biphenyl backbone of a PCB congener determine its physico-hemical properties, which in turn may govern its dynamics ofccumulation in adipocytes. It appears that the dynamics of passagef exogenous molecules across the cell membrane can indeed beorrelated with parameters such as the molecular size (as reflectedy the molecular weight), the molecular volume (as appreciatedy Connolly’s volume), the lipophilicity as measured by the log Plog of the partition coefficient n-octanol:water) (Table 2), and theapacity to be involved in hydrogen bondings (Lipinski et al., 2001).n this context, the contribution offered by Lipinsky and his rule of
has been inspirational in the field of drug development (Leeson,012; Lipinski et al., 2001). This rule indeed describes molecularroperties important for a drug’s pharmacokinetic behavior, which
s in general largely conditioned by passive diffusion with regardo absorption and distribution. This rule can be summarized as fol-ows: to be efficiently absorbed in cells, a molecule should have
molecular weight lower than 500 g/mol, a log P not greater than
, no more than five hydrogen bond donors and no more than 10ydrogen bond acceptors.
Among the three congeners considered in the present study,hey all satisfy three out of four of Lipinsky’s requirements:
able 2hysico-chemical properties of the three PCBs tested.
PCB congeners Molecular weight (g/mol)
PCB-28 257.54
PCB-118 326.44
PCB-153 360.88
cells (B). Results are expressed as the percentage of the initially added amounts inthe medium that were quantified in cells after 4 h of incubations with the pollutantsand show that PCB-28 entered cells more rapidly than PCB-118 and -153.
molecular weight and number of hydrogen bond donors andacceptors. However, only PCB-28 marginally satisfies the demandregarding log P, PCBs 118 and 153 exhibiting values higher than 5.In both adipocyte models, PCB-28 entered the cells more rapidlythan the two others. Due to their higher molecular weight, theirbigger size and especially, their higher lipophilicity, PCBs-118 and-153 might have a higher initial tendency to remain trapped in thehydrophobic cell membrane core than PCB-28. A particularity ofadipocytes as compared to other cell types is their extremely highcontent in triglycerides, which are stored in lipid droplets occupy-ing the vast majority of the cytoplasm. This important intracellularhydrophobic pool is most probably the driving force enabling tosuccessfully attract PCBs out of the cell membrane. In the case ofPCBs-118 and -153, this transfer might occur more slowly, as a con-sequence of their more pronounced temporary sequestration in thecell membrane.
The more triglycerides present in the cells, the higher theirattraction/accumulation capacity for PCBs and thus the higher theaccumulation levels of PCBs reached after 4 h of incubation uponaddition of a same dose of pollutants in the medium (as shown
in Fig. 4A and B). In 3T3-L1 adipocytes, characterized by highertriglyceride contents than MEFs (Fig. 1), the dynamics of uptake ofPCBs-118 and -153 were slower than the one of PCB-28 at the begin-ning of the experiment. However, after 4 h of incubation, the cells
Molecular volume (A3) Lipophilicity (log P)
179.95 5.16208.06 6.39223.42 7.01
S. Bourez et al. / Toxicology Le
FP
attcP2pa
cdvd5mdhrmemoc1APhz2ba
amounts of PCBs added in the medium (Table 3A) or with theremaining quantities present in the medium (Table 3B). Altogether,these results show that adipocytes have an enormous capacity ofconcentrating intracellular amounts of lipophilic pollutants such as
Table 3Ratios between PCBs-28, -118 and -153 concentration in the cells after 4 h of incu-bation and the initial amounts added in the medium (A) and between PCBs-28, -118and -153 concentration in the cells and the corresponding medium after 4 h of incu-bation (B). The results show that the concentrations of PCBs in the cell layer aremuch higher than in those in the medium.
A PCB-28 PCB-118 PCB-153
AMEFs 95x 55x 55x3T3-L1 330x 330x 300x
ig. 5. Electrostatic potential distribution on molecular structures of PCB-28 (A),CB-118 (B) and PCB-153 (C).
ccumulated around 80 ± 3% of each PCB congener, which meanshat PCBs-118 and -153 have caught up. On the contrary, in MEFs,he accumulation percentage was much less consistent for all PCBongeners (23 ± 3% of PCB-28, 13 ± 2% of PCBs-118 and 13 ± 1% ofCB-153) and remained lower for PCB-118 and -153 than for PCB-8, even after 4 h of incubation. Taken together, these results clearlyoint out the significant role of triglycerides on the dynamics ofccumulation of PCBs in fat cells.
The electrostatic potentials were acquired for the three PCBongeners on the basis of AM1 energy-minimized structures asescribed Section 2.6. These maps around the PCB molecules showolumes that are electron-rich (figured in blue) and electron-eficient (figured in red) (Fig. 5A for PCB-28, 5B for PCB-118 andC for PCB-153). Electrostatic potentials highlight the ability of aolecule to engage hydrogen bonds with external hydrogen bond
onors or not. Electron-rich zones are more likely to engage intoydrogen bonds as acceptors and can therefore be considered asather hydrophilic whereas electron-poor zones are converselyore lipophilic. While all three congeners showed, to various
xtents, some electron-rich volumes reflecting their common aro-atic character (pi orbital overlapping), PCB-28 was however the
nly structure that exhibited a more concentrated electronegativeontinuum between the phenyl cycles relative to both others. PCBs-18 and -153 showed more extended electron-deficient lobes.lthough the global shape of the electronegative continuum ofCB-153 was rather similar to that of PCB-28, that of PCB-153 wasowever importantly surrounded by much larger electro-deficientones making PCB-153 less prone to hydrogen bondings than PCB-
8. The latter was therefore more prone to accept hydrogen bondsy hydrophilic solvents than both other congeners were. This prob-bly accounted for its lower hydrophobicity as also reflected by its
tters 216 (2013) 40– 46 45
lower log P value as compared to PCB-118 and PCB-153 (Table 2).The lower log P and molecular volume of PCB-28 thus explain itsbetter diffusibility through cellular membranes as compared tothe two other more highly chlorinated congeners. PCBs-28 and-118, two mono-ortho substituted PCBs, shared a rather similargeneral structure (Fig. 5A and B). However, there was a major dif-ference between them as only PCB-28 presented an electronegativecontinuum between the phenyl rings whereas a clear electro-deficient zone was apparent between the phenyl cycles of PCB-118.This made this congener less prone to hydrogen bondings andcontributed to its higher log P value. Regarding PCB-153, whichexhibited the slowest accumulation rate in both adipocyte mod-els, it was clearly the most bulky congener of all three via the moreperpendicular position of the phenyl rings due to the presence oftwo chlorine atoms in the ortho position. Moreover, marked largerelectro-deficient zones characterized this molecule, rendering PCB-153 less prone to hydrogen bondings and thus more lipophilic(Fig. 5C), as confirmed by its higher log P value. Taken together,these calculations support the fact that the accumulation profilesof these three pollutants in fat cells were highly dependent on thelipophilicity parameters (governed by their ability to get involvedin hydrogen bondings and their log P values).
3.5. Adipocytes ability to concentrate PCBs present in thesurrounding medium
The concentration ratios of PCBs within the cells were calcu-lated by comparing their concentrations in the cell layer after 4 h ofincubation to their concentration in the medium at the beginning ofthe experiments (time = 0 h). The cell layer of one experimental unit(10 cm2) was estimated at 5 ± 2 �l, by independently scrapping thecell layer of 20 experimental units and calculating the mean ± SDvolume obtained. The corresponding growth medium volume was2 ml. Results are shown in Table 3A. After 4 h of incubation withthe pollutants, 3T3-L1 adipocytes concentrate the pollutants up to300 times as compared to their final concentration in the mediumat the beginning of the experiment. Even in MEFs, where lipidcontents were lower, the adipocytes concentrated PCB-28 up to100 times and more than 50 times for PCBs-118 and -153 as com-pared to the initial medium concentration. The great ability of fatcells to accumulate and concentrate high levels of lipophilic pol-lutants is even more pronounced when concentration ratios werecalculated between the cell layer and the corresponding growthmedium after 4 h of incubation instead of the initial concentrationin the medium, especially in 3T3-L1 cells (Table 3B). In MEFs, asabout 80% of the added PCBs were still present in the mediumafter 4 h of incubation, concentration ratios did not differ impor-tantly depending on whether they were calculated with the initial
BMEFs 140x 60x 60x3T3-L1 2000x 1900x 1900x
4 logy Le
PcHmchmsortac
tpohcteaaaicttvstwPiti
C
A
tFcfUWStatmts
der, Y.-J., Larondelle, Y., 2012. Conversion of t11t13 CLA into c9t11 CLA in Caco-2
6 S. Bourez et al. / Toxico
CBs. Seeing their low molecular weight and pronounced lipophilicharacter, they most probably enter the cells by passive diffusion.owever, the strict definition of passive diffusion is a net move-ent of molecules through the membrane in the direction of their
oncentration gradient. This is not in line with the results describedere, as we show that the intracellular concentrations of PCBs wereuch higher than those in the medium. However, the progres-
ive sequestration of the pollutants inside the hydrophobic coref lipid droplets (Bourez et al., 2012), essentially made of triglyce-ides and esterified cholesterol, leads to a continuous depletion ofhe PCBs present in the cytosol, which in turn entertains a favor-ble concentration gradient from the outside to the inside of theells.
To conclude, in vitro culture systems are very useful toolso study the uptake potential of fat cells for lipophilic com-ounds and to evaluate both qualitative and quantitative aspectsf their kinetics and toxicity in the adipose tissue. However, weave shown that for a same amount of pollutants added to theells, the accumulated quantities differed importantly betweenhe adipocytes (MEFs differentiated into adipocytes and differ-ntiated 3T3-L1 cells). It is therefore important to take this intoccount when selecting an in vitro model for experiments. Themounts of accumulated pollutants were directly linked to themounts of triglycerides in the cells. As a consequence, contam-nation levels of lipophilic pollutants in cells with different fatontents would not be comparable. In addition, we have shownhat the dynamics of entry of PCBs inside the cells are related tohe lipophilicity (log P) of the congener investigated. Indeed, log Palues higher than 5 are not in favor of an entry in cells by pas-ive diffusion and thus such molecules rather have the tendencyo stay trapped in the plasma membrane. However in adipocytes,hich are characterized by extremely high levels of triglycerides,
CBs seem to be forced out of the membranes to then accumulaten the hydrophobic lipid droplets inside the cells. Yet, the higherhe lipophilicity of the molecule, the slower this phenomenons.
onflict of interest statement
The authors declare that there is no conflict of interest.
cknowledgements
We are very grateful to F. Lasnier and X. Le Liepvre fromhe “Centre de Recherches Biomédicales des Cordeliers”, Paris,rance, for technical assistance in diverse assays and for theell cultures. We would also like to acknowledge M. Louvetrom the “Laboratoire d’Ecologie animale et d’Ecotoxicologie”,niversité de Liège, Belgium, for her help in PCB analyses.e also thank A. Joly and E. Mignolet from the “Institut des
ciences de la Vie”, UCLouvain, Belgium, for their help inhe fatty acid analyses. We also greatly appreciated the helpnd advice of Prof. B. Govaerts and especially C. Rasse from
he “Support en méthodologie et calcul statistique”, Institut
ultidisciplinaire pour la modélisation et l’analyse quantita-ive, UCLouvain, Belgium, in the statistical analyzes of thistudy.
tters 216 (2013) 40– 46
References
Barouki, R., Clément, K., 2009. Adipose tissue: a reservoir for persistent organicpollutants. Obésité 4, 156–157.
Becker, M.M., Gamble, W., 1982. Determination of the binding of 2,2′ ,4,4′ ,5,5′-hexachlorobiphenyl by low-density lipoprotein and bovine serum albumin.Journal of Toxicology and Environment Health 9, 225–234.
Bourez, S., Le Lay, S., Van den Daelen, C., Louis, C., Larondelle, Y., Thomé, J.-P., Schnei-der, Y.-J., Dugail, I., Debier, C., 2012. Accumulation of polychlorinated biphenylsin adipocytes: selective targeting to lipid droplets and role of caveolin-1. PLoSONE 7, e31834.
Chevrier, J., Dewailly, E., Ayotte, P., Mauriege, P., Despres, J.P., Tremblay, A., 2000.Body weight loss increases plasma and adipose tissue concentrations of poten-tially toxic pollutants in obese individuals. International Journal of Obesity andRelated Metabolic Disorders: Journal of the International Association for theStudy of Obesity 24, 1272–1278.
Collins, J.M., Neville, M.J., Pinnick, K.E., Hodson, L., Ruyter, B., van Dijk, T.H., Rei-jngoud, D.-J., Fielding, M.D., Frayn, K.N., 2011. De novo lipogenesis in thedifferentiating human adipocyte can provide all fatty acids necessary for matu-ration. Journal of Lipid Research 52, 1683–1692.
Debier, C., Pomeroy, P.P., Dupont, C., Joiris, C., Comblin, V., Le Boulenge, E., Larondelle,Y., Thomeı̌, J.-P., 2003. Quantitative dynamics of PCB transfer from mother to pupduring lactation in UK grey seals Halichoerus grypus. Marine Ecology ProgressSeries 247, 237–248.
Dirinck, E., Jorens, P.G., Covaci, A., Geens, T., Roosens, L., Neels, H., Mertens, I., VanGaal, L., 2011. Obesity and persistent organic pollutants: possible obesogeniceffect of organochlorine pesticides and polychlorinated biphenyls. Obesity 19,709–714.
Hjelmborg, P.S., Andreassen, T.K., Bonefeld-Jørgensen, E.C., 2008. Cellular uptakeof lipoproteins and persistent organic compounds—an update and new data.Environmental Research 108, 192–198.
Irigaray, P., Ogier, V., Jacquenet, S., Notet, V., Sibille, P., Mejean, L., Bihain, B.E., Yen,F.T., 2006. Benzo[a]pyrene impairs b-adrenergic stimulation of adipose tissuelipolysis and causes weight gain in mice—a novel molecular mechanism of tox-icity for a common food pollutant. The FEBS Journal 273, 1362–1372.
Kim, I.S., Beaudette, L.A., Cassidy, M.B., Lee, H., Trevors, J.T., 2002. Alterations in fattyacid composition and fluidity of cell membranes affect the accumulation of PCBcongener 2,2′ ,5,5′-tetrachlorobiphenyl by Ralstonia eutropha H850. Journal ofChemical Technology & Biotechnology 77, 793–799.
Kim, M.-J., Marchand, P., Henegar, C., Antignac, J.-P., Alili, R., Poitou, C., Bouillot, J.-L., Basdevant, A., Le Bizec, B., Barouki, R., Clément, K., 2011. Fate and complexpathogenic effects of dioxins and polychlorinated biphenyls in obese subjectsbefore and after drastic weight loss. Environmental Health Perspectives 119,377–383.
Lee, D.-H., Steffes, M.W., Sjodin, A., Jones, R.S., Needham, L.L., Jacobs Jr., D.R., 2011.Low dose organochlorine pesticides and polychlorinated biphenyls predict obe-sity, dyslipidemia, and insulin resistance among people free of diabetes. PLoSONE 6, e15977.
Leeson, P., 2012. Drug discovery: chemical beauty contest. Nature Reviews DrugDiscovery 481, 455–456.
Le Lay, S., Hajduch, E., Lindsay, M.R., Le Lièpvre, X., Thiele, C., Ferré, P., Parton, R.G.,Kurzchalia, T., Simons, K., Dugail, I., 2006. Cholesterol-induced caveolin targetingto lipid droplets in adipocytes: a role for caveolar endocytosis. Traffic 7, 549–561.
Lipinski, C.A., Lombardo, F., Dominy, B.W., Feeney, P.J., 2001. Experimental and com-putational approaches to estimate solubility and permeability in drug discoveryand development settings. Advanced Drug Delivery Reviews 46, 3–26.
Matthews, H.B., Surles, J.R., Carver, J.G., Anderson, M.W., 1984. Halogenated biphenyltransport by blood components. Fundamental and Applied Toxicology 4,420–428.
Mullerova, D., Kopecky, J., 2007. White adipose tissue: storage and effector site forenvironmental pollutants. Physiological Research 56, 375–381.
Roberts, L.D., Virtue, S., Vidal-Puig, A., Nicholls, A.W., Griffin, J.L., 2009. Metabolicphenotyping of a model of adipocyte differentiation. Physiological Genomics39, 109–119.
Ruxton, C.H.S., Reed, S.C., Simpson, M.J.A., Millington, K.J., 2004. The health benefitsof omega-3 polyunsaturated fatty acids: a review of the evidence. Journal ofHuman Nutrition and Dietetics 17, 449–459.
Schneider, A.-C., Beguin, P., Bourez, S., Perfield II, J.W., Mignolet, E., Debier, C., Schnei-
cells and inhibition by sterculic oil. PLoS ONE 7, e32824.Spindler-Vomachka, M., Vodicnik, M.J., Lech, J.J., 1984. Transport of 2,4,5,2′ ,4′ ,5′-
hexachlorobiphenyl by lipoproteins in vivo. Toxicology and Applied Pharma-cology 74, 70–77.
