|Home | About | Journals | Submit | Contact Us | Français|
Natural killer (NK) cells are a subset of lymphocytes capable of killing tumor cells, virally infected cells and antibody-coated cells. Dibutyltin (DBT) dichloride is an organotin used as a stabilizer in polyvinylchloride (PVC) plastics and as a deworming product in poultry. DBT may leach from PVC water supply pipes and therefore poses a potential risk to human health. We previously reported diminished NK cells lysis of tumor cells following exposure to DBT in serum-free cell culture medium. However, under in vivo conditions, circulating cells will be exposed to DBT in the presence of 100% plasma; thus we investigated whether serum supplementation and incubation time modulates DBT effects on NK cell killing and the accumulation of DBT in freshly isolated NK cells, to determine whether a serum-free model accurately predicts possible effects of DBT on human NK cells under in vivo conditions. Lytic function was decreased by approximately 35% at an intracellular DBT (DBTi) concentration of 200μM and nearly complete loss of lytic function was observed at DBTi above 300μM for one h. However, an intracellular concentration of 50μM DBT, achieved over 24 h of exposure in 50% serum, reduced lytic function by 50%. Thus, conditions that reflect prolonged contact with circulating DBT, in the presence of serum, suggest that NK cell activity is decreased at lower DBTi. These data indicate that the model is useful in predicting potential human effects of relatively low DBTi concentrations.
Dibutyltin (DBT) is used as a stabilizer in the production of polyvinylchloride (PVC) plastics (Roper, 1992; Yamada et al., 1993) and as a deworming agent in poultry. DBT has been reported to leach from PVC pipes into finished drinking water (Sadiki et al., 1996), into beverages stored in PVC pipes during manufacture (Forsyth et al., 1992a, b) and as a contaminant in some poultry products (Epstein et al., 1991). Additionally, DBT contaminates seafood (Rantakokko et al. 2006; Ueno et al., 1999) and measurable levels of butyltins have also been found in human blood and liver (Kannan et al., 1999; Nielsen and Strand, 2002; Whalen et al., 1999). As mandated by the Safe Drinking Water Act, the U.S. Environmental Protection Agency included DBT on the Candidate Contaminant List 2 because of likely human exposure via drinking water and the potential to adversely affect the immune system.
Exposure to DBT has been reported to disrupt normal embryonic development and has been associated with a variety of birth defects in rats including cleft mandible and fused ribs (Ema et al., 1996; Noda et al., 1993). Exposure has also been associated with atrophy of the thymus and pancreatitis in rats (Merkord et al., 1997; Pieters et al., 1994). Animal studies also indicate that organotins, including DBT and tributyltin (TBT), diminish immune function, including natural killer (NK) cell activity (Seinen et al., 1977; Vos et al., 1990).
NK cells are a subset of lymphocytes that are capable of killing tumor cells, virally infected cells and antibody-coated cells. NK cells are capable of killing tumor cells without prior sensitization, putting them in the forefront of lymphocyte defense against tumor cells and virally infected cells (Lotzova, 1993, Vivier et al., 2004). The earliest and possibly predominant defense against tumor cells has been attributed to NK cells (Lotzova, 1993; O'Shea and Ortaldo, 1992; Trinchieri, 1989). NK cells are defined by the absence of the T-cell receptor/CD3 complex and by the presence of CD56 and/or CD16 on the cell surface. They are responsible for limiting the spread of blood borne metastases, as well as limiting the development of primary tumors (Hanna, 1980; Kiessling and Haller, 1978). NK cells also play a central role in immune defense against viral infection as evidenced by greatly increased incidence of viral infection seen in individuals where the NK subset of lymphocytes is completely absent (Biron et al., 1989; Fleisher et al., 1982). Therefore, any chemical that can affect the ability of NK cells to recognize and lyse a target cell, could potentially increase the risk of tumors and viral infections.
In vitro exposure of human NK cells to DBT in serum-free medium dramatically inhibits their lytic function (Whalen et al., 1999, 2002a, b). This decrease in lytic function was accompanied by a decreased capacity of NK lymphocytes to bind tumor targets when the cells were exposed to DBT for 24 h. Further studies reported that ATP levels as well as levels of several functionally important proteins were decreased in NK cells exposed to DBT in serum-free medium (Catlin et al., 2005; Dudimah et al., 2007a; Odman-Ghazi et al., 2003; Whalen and Loganathan, 2001).
Published reports (Harrill et al., 2005; Mundy et al., 2004; Seibert et al., 2002) suggest that the presence or absence of serum (or serum albumin) may influence intracellular accumulation of xenobiotics under in vitro conditions, and that intracellular chemical concentrations may provide a useful metric for risk assessment. Thus, the goal of the current study was to determine the accumulation of DBT in NK cells under conditions of no serum, 10% serum, and 50% serum, as well as the effects of serum on the ability of NK cells to lyse tumor cells. This information is essential in evaluating the risk that environmental levels of DBT may pose to immune function.
Highly purified NK cells were isolated from buffy coat preparations (American Red Cross, Portland, OR) using a rosetting procedure. Buffy coats were mixed with 1 ml of RosetteSep human NK cell enrichment antibody cocktail (StemCell Technologies, Vancouver, British Columbia, Canada) per 30 ml of buffy coat. The mixture was incubated for 40 min at room temperature (~25°C) with periodic mixing. Following the incubation, 5 ml of the mixture was layered onto 4 ml of Ficoll-Hypaque (1.077 g/ml) (MP Biomedicals, Irvine, CA) and centrifuged at 1200 x g for 30 min. The cell layer was collected and washed twice with phosphate buffered saline, pH 7.2, and cultured in basic medium (RPMI-1640 supplemented with 2mM l-glutamine, 50 U penicillin G and 50 μg streptomycin/ml) plus 10% heat-activated bovine calf serum (BCS), at 1 million cells/ml for up to 16 h before exposure to DBT. The resulting cell preparation was >95% CD16+, CD56+, 0% CD3+ by fluorescence microscopy and flow cytometry (Whalen et al., 2002a).
14C-DBT was produced by RTI International (Research Triangle Park, NC) and was generously donated by Dr Tom Burka, National Institute of Environmental Health Sciences, National Institutes of Health. DBT was purchased from Sigma-Aldrich (St Louis, MO). Stock solutions of 14C-DBT and DBT were made in dimethylsulfoxide (DMSO) (Sigma-Aldrich) and diluted into the appropriate cell culture media. Final DMSO concentrations were the same in all cultures across all concentrations of DBT.
Pooled human type AB serum (Cellgro, Mediatech, Inc., Herndon, VA), without heat inactivation and from a single lot, was used in all experiments. Purified NK cells were separated by centrifugation from BCS-supplemented basic medium and transferred to basic medium containing 0.5% gelatin, 10% human serum, or 50% human serum. NK cells (at a concentration of 1.5 million cells/ml) were then exposed to varying concentrations of 14C-DBT (DBT in the medium; DBTm) from 0.5 to 10μM for 1 h. Concentrations were chosen based on previous in vitro studies and because levels as high as 0.3μM DBT have been measured in human blood (Catlin et al., 2005; Dudimah et al., 2007a; Whalen et al., 1999, 2002a, b). Additionally, NK cells were exposed to the following: 5μM 14C-DBT for 5, 10, 30, 60, and 120 min; 5μM 14C-DBT for 24 h; 1μM 14C-DBT for 24 h; and 5μM 14C-DBT for 1 h followed by 24 h in DBT-free medium. Accumulation of 14C-DBT under these conditions was measured as described below. NK cells were also exposed to unlabeled DBT under the same conditions as described for 14C-DBT and then tested for the effect on lytic function as described below. In all experiments, cell viability was assessed after incubation by 14C-DBT trypan blue dye exclusion; none of the treatments affected cell viability.
Following the exposures, the cells were centrifuged and the supernatant was removed and placed in an appropriately labeled scintillation vial. The cell pellet was then resuspended in 500 μl of medium and centrifuged. The supernatant was removed and added to the scintillation vial containing the initial supernatant. This process was repeated. Following the wash step, 100 μl of 0.1M NaOH was added to the pellet and the dissolved pellet was transferred to a scintillation vial. The total radioactivity added to the cells at a given DBT concentration was determined by adding that same amount of 14C-DBT to a separate scintillation vial and counting it with the samples. Results are given as μM DBT accumulated in the cells (DBTi). The total volume of the cells was estimated using the number of NK cells that were present and multiplying that number by the estimated volume of an individual NK cell (approximately 7.35 × 10−10 ml, assuming a diameter of 15μM). Unless stated otherwise, all results are presented as the mean ± SEM.
The ability of NK cells to lyse tumor cells was measured using a 51Cr release assay (Whalen, 1997; Whalen et al., 1999). The target (T) cell in all cytotoxicity assays was the NK-susceptible K562 (human chronic myelogenous leukemia) cell line. K562 cells were incubated with 51Cr (Perkin-Elmer Life Sciences, Boston, MA) in 0.2–0.5 ml of BCS for 1–1.5 h at 37°C in air/CO2 (19:1). Following this incubation, the target cells were washed twice with serum-free medium. NK cells (100 μl) were added to the wells of round-bottom microwell plates. Dilutions of NK effector (E) cells were prepared to produce final E:T cell ratios of 12:1 (1.2 × 105 E/100 μl), 6:1 (0.6 105/100 μl) and 3:1 (0.3 × 105/100 μl); each ratio was tested in triplicate. Target cells were added (1 × 104 in 100 μl) to each well of the microwell plate and the plate was centrifuged at 300 × g for 3.5 min and incubated for 2 h at 37°C (air/CO2, 19:1). After incubation, a 0.1-ml aliquot of the supernatant was collected and counted for radioactivity for 60 sec in a Packard COBRA gamma radiation counter (Packard Instrument Co., Meriden, CT). Target lysis was calculated as follows: 100 × [(test c.p.m − spontaneous c.p.m.)/maximum c.p.m. − spontaneous c.p.m.)]. Maximum release was determined by adding 100 μl of 10% Triton X-100 to wells containing only target cells; spontaneous release was determined by counting the supernatant of wells containing target cells in medium alone.
Data from experiments that evaluated effects of serum, DBTm or incubation time on DBTi accumulation and NK cell lytic activity were analyzed by comparing the slopes of dose-response curves, using the General Linear Model (GLM; SAS, Cary, NC). A contrast statement was used within the GLM to test the null hypothesis that Ho: B1 = B2 = B3, where B = the slope of the line for each serum concentration. This approach was chosen to evaluate treatment effects over the range of independent variables by determining whether the slopes of the lines differed, rather than a point-by-point analysis using ANOVA. However, ANOVA (SAS) was used when it was necessary to compare mean values. When ANOVA indicated a statistically significant treatment effect, individual post hoc comparisons were made using the least squares means t-test. Differences between groups was considered significant when the value of p was < 0.05.
Intracellular accumulation of DBT (DBTi) in each medium was tested in cells from the same donor to eliminate differences due to variables other than the composition of the medium and was repeated with cells from three separate donors. There was no significant binding of the DBT to the plastic tubes at any of the concentrations of DBT or at any concentration of serum. Cell viability was not affected by any treatment. The slopes of the DBT accumulation curves after 1 h of incubation in the presence of 0.5–10μM 14C-DBT differed, indicating a significant effect of serum supplementation (Fig. 1). Comparison of DBTi curves at 0, 10% and 50% serum supplementation indicated that DBTi was similar at 0 and 10% and 0 and 50%, and significantly different (p < 0.01) at 10% and 50%. The results suggest that DBTi accumulation was less at a serum concentration approaching those in vivo, compared with those typically used in tissue culture. However, the lack of statistically significant differences between serum- and 40.5% gelatin-supplemented cultures are difficult to interpret, and may indicate that 1 h of incubation was not sufficient to determine the effects of serum supplementation on DBT uptake.
A time curve of DBT accumulation at 5μM DBT from 10 to 120 min is shown in Figure 2. Intracellular concentrations of DBT in NK cells incubated in serum-free medium were 165 ± 9μM after 10 min and 396 ± 31μM after 120 min. In medium containing 10% serum, DBTi concentrations were 146 ± 9μM and 318 ± 16μM after 10 and 120 min, respectively. The slopes of the curves for 0 and 10% serum indicated a significant effect of incubation time (p < 0.0001). However, in the presence of 50% serum, DBTi was 175 ± 17μM after 10 min and 185 ± 9μM after 120 min, and the slope of the curve was not different from zero; DBTi peaked at 30 min, and did not increase thereafter. The slopes of the curves for 0 and 10% serum supplementation indicated a significant difference in uptake in cultures compared with medium supplemented with 50% serum (p < 0.01), and no significant difference between the 0 and 10% serum.
The effect of serum concentration on intracellular accumulation of DBT after 24 h in culture was also examined (Fig. 3). Incubation with 1μM DBT in medium supplemented with 0, 10% and 50% serum produced DBTi of 145 ± 11μM, 84 ± 8μM, and 50 ± 4μM, respectively. Cells cultured with 5μM DBT in 0, 10% or 50% human AB serum-containing medium accumulated 496 ± 33μM, 310 ± 30μM, and 214 ± 10μM, respectively (Fig. 3A). Slopes of the accumulation curves indicated a significant effect of serum supplementation (p < 0.0001), which varied among DBTm (p < 0.0001); comparison of 1 and 5μM DBTm slopes indicated a significant difference in DBTi (p < 0.001). The presence of 10% serum in the medium decreased uptake by 38% and 50% serum decreased uptake by 57% compared with 0% serum. These results indicated that as the concentration of serum in the medium increased there was a significant decrease in the accumulation of DBT in the NK cells at a given DBTm. Furthermore, DBTi was similar at a given DBTm/serum concentration at 1 and 24 h (Fig. 3B).
To determine whether accumulated DBT remains in the cell or is lost to DBT-free culture medium, cells were incubated for 1 h in medium containing 5μM DBT and 0, 10% or 50% serum. Cells were then washed and recultured in medium supplemented with the same serum concentration as before. Intracellular concentrations at 24 h were 168 ± 36, 64 ± 11, and 31 ± 4μM, respectively, after culture in serum-free, 10 and 50% serum (Fig. 4). DBTi was greater in NK cells recultured in serum-free medium than in medium supplemented with 10% (p < 0.003) or 50% (p < 0.0002) serum.
To determine if radiolabeled and nonlabeled DBT were equally effective at diminishing the ability of NK cells to destroy K562 tumor cells, serum-free cultures were exposed to 5 and 10μM 14C-DBT and 12C-DBT for 1 h or to 0.5 and 1μM 14C-DBT and 12C-DBT for 24 h, washed, and cultured with target cells. Exposure to 12C- and 14C-labeled DBT significantly and equally decreased NK cell activity (data not shown).
Figure 5 shows the effects of various exposures to DBT in serum-free and serum-supplemented culture medium on lytic activity of NK cells, expressed as percent of control. There was an overall effect (p < 0.01) of DBT exposure for 1 h at all serum concentrations; at the highest DBTm, NK activity was suppressed by 95%. Slopes did not differ between serum-free and 10% serum-containing medium; a trend was observed when the slopes of the zero and 50% serum curves (p = 0.054) and the 10 and 50% serum curves (p = 0.056) were compared (Fig. 5A).
The effects of a 24-h exposure to 0.25, 0.5, and 1μM DBT in the three types of medium are shown in Figure 5B. In the absence of serum, the slope of the inhibition curve for all concentrations of DBTm was similar; that is, all concentrations of DBTm significantly reduced NK cell activity. In contrast, there was a significant effect of DBTm in the presence of 10% (p = 0.0062) or 50% (p = 0.001) serum. The slopes of the curves differed significantly between 0 and 50% (p < 0.0001) and 10 and 50% serum (p = 0.001). These results indicate that the presence of serum in the medium reduced the inhibitory effect of DBT on NK cell activity, and that 50% serum supplementation had a greater protective effect than did 10% serum.
Data presented in Figure 4 indicate that the presence of serum in DBT-free medium increased the efflux of DBT from the intracellular compartment into the medium when cells were cultured for an additional 24 h. Cells subjected to the same loading/DBT-free culture scheme were incubated with target cells to determine whether culture in the absence of DBT for 24 h affected DBT-mediated suppression of NK cell activity and to assess the modulating effect, if any, of serum supplementation. There was an overall effect of serum concentration (p < 0.0001) on inhibition of NK cell activity by DBT (Fig. 5C), but no significant difference in NK cell activity of cells cultured for 24 h in serum-free or 10% serum-supplemented media. However, significantly less suppression of NK cell activity occurred in cells cultured for 24 h in medium supplemented with 50% serum (Fig. 5C). Reducing DBTi by culture in the presence of physiological serum concentrations partially reversed suppressed lytic activity.
Figures 6A and 6B plot the relationship between intracellular concentrations of DBT and the loss of lytic function at 1 h and 24 h, respectively. These plots indicate that a complete loss of lytic function occurs at 1 h when the intracellular concentration of DBT reaches about 400μM. However, after 24 h, a concentration of between 50 and 100μM was associated with complete loss of lytic function.
Recent emphasis on the development of high throughput screening methods, and the use of in vitro models to evaluate mode and mechanism of action, underscore the importance of determining whether culture conditions affect either the experimental outcome or interpretation of results. Traditional techniques often include supplementation of the culture medium with 10% serum, and other models exclude serum as a supplement to prevent binding of the test substance to serum proteins. However, published reports (Harrill et al., 2005; Mundy et al., 2004; Seibert et al., 2002) suggest that the presence or absence of serum (or serum albumin) may influence intracellular accumulation of xenobiotics under in vitro conditions. The goal of the current study was to determine whether serum supplementation affected accumulation and loss of DBT by NK cells and the effects of intracellular concentrations of DBT on target cell lysis.
DBT decreased the lytic function of human NK cells when they were exposed at concentrations ranging from 0.5 to 10μM for 1–24 h in serum-free medium (Catlin et al., 2005; Dudimah et al., 2007a; Whalen et al., 1999, 2002a, b). The lower concentrations used in the studies were similar to those detected in human blood samples (levels as high as 0.3μM; Whalen et al., 1999). However, circulatory transport of DBT occurs in the presence of 50–60% serum. It was therefore essential to determine whether the presence of serum affects DBTi concentration in NK cells as the first step in clarifying how well in vitro DBTm concentrations predict DBTi accumulation and subsequent adverse biological effects in vivo.
Intracellular concentrations of DBT increase rapidly (40- to 100-fold compared with DBTm) and are a function of DBT and serum concentrations in the culture medium. DBTi reaches a maximum relatively soon after DBT is added (Fig. 3B) but probably not as a result of a saturation event, as DBTi was greater at 5 than at 1μM DBTm. The mechanism(s) that limits DBTi at a given DBTm is unknown; however it is independent of time, but not of physiological levels of serum. Serum supplementation reduced intracellular accumulation of DBT, suggesting that binding to extracellular (serum) proteins decreases movement across the cell membrane. Nevertheless, concentrations of DBT achieved after as few as 30 min of culture in 50% serum (approximately 200μM, Fig. 2) was well within the range of DBTi that reduced target cell killing by nearly 40% (see Fig. 6A). Serum also appeared to enhanced the movement of DBT out of cells cultured in DBT-free medium (Fig. 4), suggesting that DBT may be exported, actively or passively, and is bound by serum proteins and therefore unavailable for reuptake. However, it is also possible that the presence of serum in DBT-free medium activated metabolism of DBT to labeled metabolites that were then transported out of the cell. In either case, the presence of serum under DBT-free conditions appears to have significantly decreased intracellular concentrations of DBT. We have previously shown that the magnitude of suppressed NK cell activity induced by 1-h exposure to DBT was similar in cells evaluated immediately after exposure and those cultured for up to 6 days in the absence of DBT (Whalen et al., 2002a), although those experiments did not determine whether DBTi concentrations were maintained over time. However, in the previous study, cells were recultured in serum-free medium, and movement of DBT from the intracellular compartment was therefore not influenced by events that, in the current study, led to lower DBTi. Serum effects on DBT accumulation were less dramatic than those reported with 1μM PDBE-47; PDBE-47 accumulation was diminished from about 80-fold in serum-free medium to about 20-fold when the exposure medium contained 10% serum (Mundy et al., 2004). PDBE-47 is a very lipophilic compound with a Kow of 6.5 (European Union, 2001), whereas DBT is far less lipophilic, with a reported Kow of 0.05 (Government of Canada, 1993). The less lipophilic nature of DBT may explain why its accumulation in cells was less affected by the presence of serum than was that of PDBE.
Previous serum-free in vitro studies indicate that equivalent inhibition of NK target cell lysis occurs at lower medium concentrations when incubation time is increased from 1 h to 24 h; cytotoxicity was reduced by approximately 80% following 1 h exposure to 10μM DBT or 24 h exposure to 1μM (Catlin et al., 2005; Dudimah et al., 2007a; Whalen et al., 1999). The current study indicates that duration of exposure influences the magnitude of effects and establishes that DBTi is associated with decreased function following brief or extended exposure. For example, a DBTi of 200μM decreased lytic function by approximately 35% and DBTi above 300μM resulted in nearly complete loss of lytic function within 1 h (Fig. 6A). However, an intracellular concentration of 50μM DBT, achieved over 24 h of exposure to 1μM DBT in 50% serum, reduced lytic function by 50% (Fig. 6B). Human studies indicate that DBT exposure occurs at relatively low levels, over an extended period (Kannan et al., 1999; Nielsen and Strand, 2002; Whalen et al., 1999) and the results for 24 h, 1μM exposure from the current study indicate that significant reduction of lytic function occurs with longer exposure times, at lower DBTi, than suggested by the one h exposure model. Additional work must be done to further refine the relationship between exposure time, DBTi concentration, targeting of biomolecules and loss of NK cell activity. However, our results do suggest that reduced NK cell cytotoxic function is not a threshold-type effect in which NK cell activity is permanently reduced when a critical level of DBT has accumulated. Rather, defects in cellular metabolism or lytic protein synthesis or expression appear to accumulate over time, thus reducing cellular function.
Preferential accumulation of DBT within NK cells suggests that intracellular components bind DBT. TBT, a close structural relative of DBT, binds with high affinity to ATP synthase (Aldridge, 1976; Matsuno-Yagi and Hatefi, 1993; von Ballmoos et al., 2004) and decreases ATP levels in human NK cells (Dudimah et al., 2007b). DBT, as does TBT, appears to interfere with the function of ATP synthase (Cain et al., 1977), decreases ATP in NK cells (Dudimah et al., 2007a) and may associate with sulfhydryl-containing cellular components such as dihydrolipoate (Cain et al., 1977), suggesting that DBT may affect NK cell function by interfering with energy metabolism. Recent studies indicate that a 24-h exposure to 1.5μM DBT in serum-free medium reduced the levels of the essential NK cytolytic proteins, granzyme B and perforin, by 62 and 49%, respectively (Catlin et al., 2005).
In summary, the current study suggests that, under in vivo conditions (i.e., in the presence of serum), DBTi accumulation will be attenuated, and less will accumulate than predicted by the serum-free 24 h in vitro model. However, intracellular accumulation in medium containing 1 or 5μM DBT and 50% serum was sufficient to dramatically reduce target cell killing. These studies also demonstrated that that longer contact time with DBT had a greater effect on cell function, at lower intracellular concentrations. Thus, both DBTi concentration and exposure time are important factors in reducing NK cell activity. Depuration studies (DBT exposure for 1 h followed by incubation for 24 h in DBT-free medium) indicated that a portion of accumulated DBT is lost when cells are cultured in the absence of DBT, and that an inverse relationship exists between medium serum concentration and the proportion lost, although binding to intracellular components is not entirely reversible. Extrapolating our results to in vivo exposure, the data suggest that as body burdens decrease, DBT levels in previously exposed cells may decrease rather rapidly. Nevertheless, our results suggest that even as body burdens fall, DBT may accumulate in cells at levels sufficient to compromise NK cell function. Finally, the inhibitory effects of serum on DBT and PBDE-47 (Mundy et al., 2004) and other studies that have evaluated intracellular vs. media concentrations of xenobiotics (Harrill et al., 2005) clearly indicate that medium concentrations of xenobiotics do not necessarily predict the actual dose to the cell. Our studies support this conclusion, in that we found a 40- to 80-fold difference between intracellular versus medium concentrations. However, our studies indicate that exposure time, and thus cumulative effects, are critical factors that may ultimately determine the relationship between intracellular concentration and cell function.
U.S. Environmental Protection Agency Contract (#EP06D000338); and National Institute of Health Grant (2S06GM-08092-28).
We thank Drs Bill Mundy, Raymond Pieters, Marsha Ward, Linda Birnbaum, and MaryJane Selgrade for their review of the manuscript and helpful suggestions.