0013-7227/91/1296-3321$03.00/0 Endocrinology Copyright ^ 1991 by The Endocrine Society

Vol. 129, No. 6 Printed in U.S.A.

pH-Dependent Cytotoxicity of Contaminants of Phenol Red for MCF-7 Breast Cancer Cells* LEIGH H. GRADY, DAN J. NONNEMAN, GEORGE E. ROTTINGHAUS, AND WADE V. WELSHONS Department of Veterinary Biomedical Sciences (L.H.G., D.J.N., W. V. W.) and the Veterinary Medical Diagnostic Laboratory (G.E.R.), University of Missouri, Columbia, Missouri 65211

mercial sources, the concentration for half-maximal cytotoxicity (TD50) in dose-responses after 4 h at pH 8.0 showed TDBo values of 2 and 6 Mg/ml> while the estrogenic activities, as half-maximal stimulation of estrogen-dependent proliferation, were identical at 2 Mg/111!- Both the cytotoxic and estrogenic activities could be removed from the phenol red by extraction with diethyl ether. A number of contaminants of the commercial phenol red were detected by reverse phase Cis HPLC. Cytotoxicity and estrogen bioassays of each of the HPLC fractions indicated that the pHdependent cytotoxicity was separate from the estrogenic activity and confirmed that, neither activity was associated with the phenol red itself. Although to date only MCF-7 cells have been affected, the pH-dependent cytotoxicity of phenol red may need to be considered as a factor in many studies of physiological response and pH that have been performed in tissue culture with phenol red present, especially those that have involved alkaline pH (>7.4) and low cell density. Even very brief elevations of pH in phenol red-containing media and salts solutions may produce detrimental effects on MCF-7 cells. This cytotoxicity for a hormone-dependent breast cancer cell line may have application in anticancer therapy. In view of the hormonal and now complicated cytotoxic activities present in contaminants of this pH indicator, the use of phenol red that has not been highly purified should be carefully considered, especially in endocrine studies in tissue culture. {Endocrinology 129: 3321-3330, 1991)

ABSTRACT. The pH indicator phenol red (phenolsulfonphthalein) is present in most tissue culture media. Contaminants of this indicator have shown substantial estrogenic activity for estrogen-dependent cells in culture, including the human breast cancer-derived MCF-7 cell line. In the course of other studies, we observed that brief (1- to 4-h) incubations of these cells at 37 C in serum-free medium (Hanks' or Earle's Balanced Salts Solution) could be toxic to MCF-7 cells when the pH was increased above 7.4, but only if phenol red (10 Mg/ml) was present in the medium. Because damaged/killed cells detached from the substratum (>98% of detached cells stained with trypan blue), we used DNA assay of the cells remaining after treatment and wash (98% of the remaining cells were dye excluding) to further assess cytotoxicity. The MCF-7 cells were more susceptible to the cytotoxicity at lower cell densities, so further characterization of phenol red cytotoxicity was performed at cell densities of 1-10 jug DNA/2-cm2 well, or approximately 40,000-400,000 cells/ml medium. In the pH range of 7.0-8.2, 50% cell death was observed in the presence of phenol red at pH as low as 7.6-7.7, with nearly 100% of the cells killed by pH 8.0. Little effect was seen in phenol red-free medium at any part of the tested pH range or in medium that contained phenol red at pH < 7.4. In time-course studies of cytotoxicity at pH 8.0 (phenol red, 10 ng/ ml), greater than 50% cell damage could be observed after less than 1 h, and little cell recovery was observed if the pH was restored to 7.4. For phenol red samples from two major com-

P

HENOL RED (phenolsulfonphthalein) is widely used in tissue culture as a pH indicator and has been used in the culture of the estrogen-dependent MCF-7 human breast cancer-derived cell line, since the line was established (1). Recently, the phenol red present in standard medium has been found to be estrogenic for these and other cell lines (2-5). This unrecognized estrogenic activity has probably been of utility in maintaining the estrogen-dependent phenotype of cell lines in tissue Received July 26, 1991. Address all correspondence and requests for reprints to: Wade V. Welshons, Department of Veterinary Biomedical Sciences, W116 Veterinary Medicine, University of Missouri, Columbia, Missouri 65211. * This work was supported by Missouri Agricultural Experimental Station Project 257, NIH Grant CA-50354, and USDA Grant MO00860. A preliminary report of this work was presented at the 73rd Annual Meeting of The Endocrine Society and is listed as Ref. 15.

culture, because estrogen-dependent proliferation is lost after culture for more than a few months in estrogenfree, phenol red-free medium by MCF-7 cells (6-8) and by several other cell lines (9-12). The estrogenic activity associated with phenol red has been shown to reside in a number of contaminants of commercial preparations of the indicator and is not due to the indicator itself (13, 14). Separate from the issue of estrogenic activity, we recently observed pH-dependent toxicity for MCF-7 cells by phenol red during the course of experiments to measure the turnover rate of the estrogen receptor protein in cells in serum-free medium. Drift in the calibration of the CO2 sensor of the incubator used for these experiments reduced the CO2 concentration from 5% eventually to 3%, which increased the steady state pH of serum-

3321

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

3322

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED

free media and salts solutions in that incubator. Cytotoxicity was eventually noted in estrogen-treated (phenol red-treated) cells, but not in the estrogen-free (phenol red-free) cell controls, and this soon led to the connection between alkaline pH and cytotoxicity by phenol red (15). This pH-dependent cytotoxicity of a common component of tissue culture media is of interest because phenol red is a factor in most studies that use the estrogendependent MCF-7 cell line, and because the cytotoxicity could be a confounding factor in many other studies of cell physiology and elevated pH in tissue culture. We examined the pH-dependent cytotoxicity of phenol red with respect to time course, pH course, and cytotoxicity dose-response of samples of the indicator from two major commercial sources. We also wanted to determine whether the cytotoxic activity and the estrogenic activity in phenol red were separable. By using HPLC and bioassay, we have shown that a contaminant of commercial preparations of phenol red is responsible for the pHdependent cytotoxicity to MCF-7 cells, and that this cytotoxic contaminant is distinct from the estrogenic contaminants in phenol red.

Materials and Methods Materials Phenol red (sodium salt) was obtained from Sigma Chemical Co. (St. Louis, MO) and Gibco BRL (Grand Island, NY). Powdered Hanks' Balanced Salts Solution, Earle's Balanced Salts Solution, Minimum Essential Medium (MEM) with nonessential amino acids (based on Earle's or Hanks' salts), as well as 17/3-estradiol, tamoxifen, HEPES, bovine insulin, calf thymus DNA type I, Hoechst dye 33258, Dextran T-70 (average mol wt, 77,800), streptomycin sulfate, and penicillin-G were obtained from Sigma (St. Louis, MO), "cell culture tested" when available. HPLC grade methanol and water were obtained from Burdick and Jackson (Muskegon, MI). Bovine calf serum was obtained from Cell Culture Laboratories (Cleveland, OH), and activated charcoal was purchased from Mallinckrodt (Paris, KY). Lyophilized trypsin was obtained from Gibco BRL. All other chemicals were reagent grade. Cell culture MCF-7 cells were obtained from Dr. V. C. Jordan, University of Wisconsin-Madison. MCF-7 cells were incubated in 5% CO2 at 37 C and maintained in T-75 or T-150 flasks in MEM with phenol red (10 Mg/ml; Gibco) containing nonessential amino acids, 10 mM HEPES, insulin (6 ng/ml), penicillin (100 U/ml), streptomycin (100 Mg/ml), and 5% charcoal-stripped calf serum (maintenance medium). For subculture, flasks were rinsed with calcium-magnesium-free Hanks' plus 25 mM HEPES (CMFH) and 1 mM EDTA, then incubated for 5 min at room temperature in a minimal volume of 0.25% trypsin in CMFH and 0.5 mM EDTA. The cells were then washed, suspended and dispersed in medium, and seeded after cell counting by hemocytometer. Dye exclusion was determined in trypan blue (0.07-0.2% final concentration).

Endo • 1991 Vol 129 • No 6

Charcoal-stripping procedure Dry acetone-washed activated charcoal at 4 g/liter and Dextran T-70 at 0.04 g/liter were suspended in 10 mM Tris-HCl, pH 8.0. One volume of serum was added to the drained pellet (200 x g; 5 min) from 1 vol of the charcoal suspension, and incubated for 30 min at 0 C with the charcoal, with resuspension every 10 min. After 1200 X g for 15 min, the supernatant serum was treated two more times with fresh charcoal pellets as described. The final, three times stripped serum was adjusted to pH 7.4 and sterilized through membrane filters of pore size down to 0.45 /im. Cytotoxicity assay Cells were seeded in 24-well multiwell plates (16-mm diameter wells) at 10,000-60,000 cells/ml-well and grown for 3-7 days in maintenance medium (containing 10 Mg/ml phenol red) before experiments, except where described. Cytotoxicity assays were performed at cell densities of 1-10 ng DNA/2-cm2 well, or approximately 40,000-400,000 cells/ml medium. Both the cytotoxic and estrogenic contaminants in phenol red appear to be less water soluble than the indicator itself, so concentrated stocks of phenol red (10 mg/ml) were prepared in absolute ethanol, and care was taken to permit complete dissolving of components in aqueous solutions, especially before sterile filtration, which would remove any undissolved material. Phenol red or the various fractions were dried from aliquots of their stocks and reconstituted in the appropriate volume of test medium (i.e. no carrier solvent). The wells were washed once with Hanks' Balanced Salts Solution and 25 mM HEPES (HBSS) before test media were added, and except where indicated, the cultures were treated for 4 h at 37 C. After treatment, the wells were washed twice with 1 ml HBSS to remove dead cells (see Results), and the remaining (live) cells in the well were assayed by DNA content. As noted in Results, only damaged, trypan blue-staining cells were detached in the course of these assays. Therefore, a reduction in the number of remaining cells was used as the quantitative index of cytotoxicity. Estrogenic activity bioassay Estrogen-dependent proliferation of MCF-7 cells was used to determine estrogenic activity (7). Test media were prepared from phenol red-free (estrogen-free) maintenance medium to which were added the phenol red or fractions (from solvent extraction or HPLC) of the phenol red. Typically, 10,000 cells were seeded per well of a 24-well plate on day 0 in phenol-free (estrogen-free) medium, fed on day 1 with the same medium, and then treated for days 3-6 with the test medium, with daily medium changes. On day 7, the wells were washed twice with 1 ml HBSS, and each well was then assayed for DNA content. DNA assay The contents of each well were sonicated in 1 ml buffer, either HBSS diluted 1:9 with water or 10 mM EDTA and 58 mM potassium phosphate after dissolving the cells (16). DNA was measured fluorometrically in an aliquot of the sonicate using Hoechst dye 33258 according to the method of Labarca and Paigen (17). Calf thymus DNA (in the same buffer used for sonication) was used as the standard after calibration by

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED absorbance at 254 nm, assuming 20 absorbance U for 1 mg DNA/ml. Ether extraction of phenol red Ten milligrams of phenol red were suspended in 1 ml distilled water and extracted three times with 0.5 ml diethyl ether, with vigorous vortexing. The phases were separated by centrifugation for 5 min at 700 x g. The organic phases were pooled from the extractions, dried under nitrogen, and redissolved in 1 ml absolute ethanol. Excess ether in the aqueous phase was removed under vacuum of 22 in. of mercury for approximately 1 h (the original aqueous volume was retained). Both fractions were stored at 4 C. HPLC Chromatographic separations were performed with a PerkinElmer series 4 liquid chromatograph (Norwalk, CT) equipped with a Rheodyne (Cotati, CA) model 7125 loop injector valve (500-Ail capacity), an 8-Atm, 10-mm X 250-mm reverse phase C]8 preparative column (Rainin Dynamax Macro-HPLC; Woburn, MA), and a Perkin-Elmer LC235 diode array detector set at 240 and 210 nm. Phenol red at 10 mg/ml was dissolved in either absolute ethanol (for sample volume l00 ^1) a n d chromatographed using isocratic separation with 75% methanol in water as the mobile phase at a flow rate of 4 ml/min. Preparative HPLC for bioassays Two milligrams of phenol red (500-^1 injection volume) were loaded onto the column, and 2-ml fractions were collected. Each fraction, therefore, contained components equivalent to those present in the 2 mg phenol red or 1 mgeq/ml effluent. For cytotoxicity assay or estrogen bioassay, aliquots of fractions were dried under a sterile hood and reconstituted in medium to the desired equivalent concentration of phenol red, at 10 figeq/ ml medium for the cytotoxicity assay and 40 ^geq/ml medium (for greater sensitivity) for the estrogen bioassay. This permitted quantitative comparison of the activity within the individual fractions to the total activity in the whole commercial phenol red.

Results pH-dependent cytotoxicity by phenol red MCF-7 cells were damaged by alkaline medium, but only when the pH indicator phenol red was present. This is illustrated in Fig. 1. MCF-7 cells in complete medium before treatment are shown in Fig. la. After incubation for 4 h in HBSS at pH 8.0 containing phenol red at the standard 10 Mg/ml, nearly all of these cells were damaged (Fig. lb). Cells incubated in HBSS at pH 8.0 without phenol red showed no cell damage (Fig. lc), and no cell damage was seen in the presence of phenol red at pH 7.4 (not shown). Only the combination of high pH and phenol red was toxic to the cells. pH-dependent cytotoxicity by phenol red was observed in HBSS, in Earle's Balanced Salts Solution plus 25 mM HEPES (HBSS), in

3323

CO2 and air incubators, as well as in Hanks'- and Earle'sbased MEM (not shown). The presence of serum in medium reduced, but did not eliminate, pH-dependent cytotoxicity by phenol red (not shown); cells at low cell density were particularly susceptible. Trypan blue dye exclusion was determined in cells treated in parallel to those described in Fig. 1. Control cells incubated in complete medium and cells that were incubated in HBSS plus phenol red at pH 7.4 for 4 h were 98% dye excluding. In addition, greater than 98% of cells that were incubated in HBSS at pH 8.0 without phenol red excluded the trypan blue. When cells were incubated in alkaline HBSS with phenol red, many of the cells appeared damaged, and those cells that appeared damaged (as in Fig. lb) also failed to exclude trypan blue. If the well was washed, the cells in the wash showed less than 2% dye exclusion, while the cells that remained in the well were 98% dye excluding. Most of the cells that were damaged by phenol red at pH 8.0 (Fig. lb) were removed by two successive washes of the well, leaving most of the intact cells attached to the well (Fig. Id). Because the simple wash procedure was effective at separating damaged from intact cells, we used DNA assay of cells remaining after the washes to measure the cytotoxicity in subsequent experiments. However, this was always checked by microscope before washing the treated cells, and in all experiments, only damaged cells were detached; that is, the various treatments did not detach any noticeable number of the refractile intact cells. pH course

The pH dependence of the cytotoxicity is illustrated in Fig. 2. MCF-7 cells were incubated for 4 h in HBSS at pH values ranging from 7.0-8.2, with and without phenol red present at 10 ng/m\, the concentration of the indicator present in most current formulations of HBSS or MEM. Cells incubated without phenol red were not affected by pH. However, in the presence of phenol red, cytotoxicity was observed beginning at pH 7.6, with approximately 50% of the cells killed between pH 7.67.7 (Fig. 2); at pH 8.2, less than 5% of the cells remained intact after incubation. At higher cell densities, cytotoxicity was less pronounced, but was still evident when pH was elevated above pH 7.4 (see Fig. 3). Based on these data, pH 8.0 was chosen for subsequent experiments. Time course We next examined the time course of the cytotoxicity to MCF-7 cells and whether the cells were able to recover from the cytotoxicity after exposure to phenol red at high pH (Fig. 3). Control cells were incubated for up to 4 h in phenol red-free HBSS at pH 8.0, and only a small number of cells were lost (Fig. 3, PhR-free, pH 8), com-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED

3324

Endo • 1991 Voll29«No6

FIG. 1. pH-dependent cytotoxicity by phenol red. a, MCF-7 cells in complete maintenance medium, b, The same field of cells after 4-h incubation at 37 C in HBSS at pH 8.0 containing phenol red before any washes; most cells appear to be damaged, c, Cells in a neighboring well that were incubated for 4 h in HBSS at pH 8.0 without phenol red, followed by two washes, as described; the cells appear undamaged and are attached to the well, d, The well shown in b after washing; most of the damaged cells have been removed, while most of the intact cells have remained. Magnification, X140.

.JO-...1

Phenol

the remaining cells once the cells were switched to nontoxic medium; however, there was also little recovery, at least within the first 3 h.

IRed-Free

Dose responses for cytotoxicity and estrogenic activity

DNA 5Well With Phenol Red Ctrl 7.0

7.2

7.4

7.6

7.8

8.0

8.2

pH of Incubation Buffer

FIG. 2. pH dependence of cytotoxicity by phenol red. Cells were incubated for 4 h at 37 C in HBSS at the indicated pH, with or without phenol red at 10 fig/ml. The pH ranged from 7.0-8.2. Control wells (Ctrl) were incubated in complete serum-containing medium prepared with Hanks'-based MEM, pH 7.4, with or without phenol red. After incubation, the wells were washed twice with HBSS (pH 7.4), and the DNA content that remained was determined as described. Values are the mean ± SEM (n = 2 wells).

pared to cells in complete medium (0 Hours starting point). Treated cells were incubated for up to 4 h in phenol red-containing HBSS at pH 8.0 (w/PhR, pH 8). At low cell density (Fig. 3A), more than half of the cells were killed by 1 h. However, cell damage was less rapid at higher cell density (Fig. 3B) than at lower density (Fig. 3A). To study recovery, cells were exposed to phenol red at pH 8.0 (cytotoxic medium) for 1 h, and then changed to pH 7.4 medium (nontoxic) for the rest of the incubation. There was some variation among experiments (for example compare recovery in Fig. 3, A and B), but overall there was little continuing cytotoxicity to

Dose-responses for phenol red cytotoxicity were determined in HBSS at pH 8.0. Phenol red samples obtained from two commercial sources (one lot each) were tested, and both samples showed pH-dependent cytotoxicity. However, the activities were different, with one showing a concentration for half-maximal toxicity (TD50) of about 2 i^g/ml and the other of approximately 6 Mg/ml (Fig. 4A, sources B and A, respectively). Both samples of phenol red damaged more than 93% of the cells at 10 /ng/ml (at pH 8.0), the concentration to which cells are exposed in maintenance medium or HBSS. Although the cytotoxic activities varied between samples, the phenol red indicator content of the two commercial samples was identical, as measured by spectrophotometry (Welshons, W. V., and K. S. Engler, unpublished) (4, 18) (absorbance at 430 and 560 nm). The estrogenic activities of the two phenol red samples were the same (Fig. 4B), as measured by estrogen-dependent proliferation of MCF-7 cells. Both samples of phenol red were able to sustain rapid proliferation of the cells at 10 fig/m\ (Fig. 4B), with a half-maximal response at 2 ixg/m\. Fractionation of phenol red

To address whether the cytotoxicity was due to the phenol red itself or a contaminant, an aqueous solution of phenol red was extracted with diethyl ether (see Ma-

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED

H9 DNA per 0.3 Well n 0.2

3325

PhR-free, pH 8

Control

Medium pH 7.4

Source A

.::::-4

HBSS pH 7.4 , pH 8

2 3 Hours at 37°C

0.1

1

10

100

Cone. Phenol Red (ug/ml) B PhR-free, pH 8 -. HBSS pH 7.4 Medium pH 7.4

w/ PhR, pH 8

1

2

3

4

Hours at 37°C FIG. 3. Time course of phenol red cytotoxicity. After seeding MCF-7 cells in maintenance medium and incubation for attachment and growth, the medium was removed and replaced with maintenance medium based on Hanks' salts for 1-h equilibration in air at 37 C, so that subsequent manipulations, including the incubations, could be performed in air at controlled pH. The medium was then changed to treatment medium at intervals, so that all incubations were completed at the same time. • - • , Cells incubated in HBSS at pH 8.0 with phenol red for up to 4 h and represents the cytotoxicity time course. O- -O, The time course at the same pH, but without the phenol red present. The additional dashed lines indicate cells incubated for 1 h under cytotoxic conditions, followed by further incubation in phenol red-free complete medium at pH 7.4 (A) or in HBSS at pH 7.4 with phenol red (A) to assess continued cytotoxicity and/or recovery. A, Time course at low cell density. B, Time course at higher cell density. Values are the mean ± SEM (n = 4 wells).

terials and Methods). MCF-7 cells were then incubated at pH 8.0 in HBSS containing increasing concentrations of whole commercial phenol red, ether-extracted phenol red, or the ether-extractable material itself, reconstituted in HBSS. The cells incubated with the ether-extracted phenol red (purified phenol red) were not damaged, while the cells incubated with either the whole phenol red or the ether extract of the indicator were damaged with the same dose responses (Fig. 5), both of which were alkaline pH dependent (not shown). The cytotoxic component of both of the commercial phenol red preparations proved to be an ether-extractable contaminant. The HPLC chromatograms of whole commercial phenol red and the aqueous and ether fractions are illustrated in Fig. 6. Ether extraction of the whole corn-

18 16 14

E2Max

1 9 ---R50% -. DNA 10 p er . 8 P Control Well 6 Source A 4 Source B 2 0 0.1 1 10 100 0.01 Cone. Phenol Red (ug/ml)

f

FIG. 4. Phenol red samples from two main commercial sources (source A and source B) were tested in dose-responses to determine their relative cytotoxicities (A) and estrogenic activities (B), as described in Materials and Methods. A, MCF-7 cells were incubated for 4 h at 37 C in HBSS at pH 8.0 containing phenol red at the indicated concentration. A, Control represents cells incubated in phenol red-free HBSS at pH 8.0. B, Four-day assay for estrogen-dependent proliferation in medium containing the indicated concentration of phenol red. B, Control represents growth in phenol red-free (estrogen-free medium); E2 Max is growth in 0.1 nM 17/8-estradiol (concentration for maximal growth); R 50% represents half-maximal proliferation. Values are the mean ± SEM (n = 4 wells).

mercial phenol red (Fig. 6A) removed virtually all of the impurities (Fig. 6C), leaving purified phenol red indicator in the aqueous fraction (Fig. 6B). For convenience in describing the contaminants, we labeled the regions of the major peaks of the profiles (see Fig. 6A). Phenol red itself eluted in region B, and the major contaminant eluted in region C. The major contaminant peak was then followed by three peaks in region D, two peaks in region E that were often unresolved, another three peaks in region F, and a final pair in region G. Exact elution times varied, but elution of the two peaks in region G was generally completed in 10 min. HPLC/bioassay of cytotoxicity and estrogenic activities Phenol red and its contaminants were fractionated by HPLC to determine whether the cytotoxic and estrogenic activities might involve the same compounds, or whether

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

3326

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED Control.-. Aqueous Phase

Phenol Red or Equivalents (ug/ml) FIG. 5. Ether extraction of cytotoxicity. Phenol red in water was extracted with diethyl ether, as described in Materials and Methods. The ether-extracted phenol red and the ether extract were tested for cytotoxicity (4 h; 37 C; pH 8.0) compared to the whole commercial phenol red. The cytotoxic activity was quantitatively extracted into the ether phase. Values are the mean ± SEM (n = 4 wells).

the two activities were separable. Figures 7 and 8 show the results of each bioassay performed on the same series of fractions. In Fig. 7, the dotted line at approximately 3.5 ng DNA/well indicates the cells present in the absence of cytotoxicity (pH 8, PhR-free), while the solid line at about 0.8 /xg DNA/well indicates the cytotoxicity caused by 4-h incubation of the cells with the whole commercial phenol red at pH 8.0 (pH 8, with PhR). Cytotoxicity by HPLC fractions was restricted to the

Endo«1991 Vol 129 • No 6

two fractions that eluted between 3.5 and 4.5 min (Fig. 7); cytotoxicity was indicated by a reduction in the remaining cell DNA values below the control level of 3.5 Mg/well. Comparing the location of the two cytotoxic fractions to the HPLC chromatogram superimposed over the bioassay (Fig. 7) indicated that the cytotoxicity was associated with the major contaminant of the phenol red (the region C marked on Fig. 6, A and C), at least as determined by UV absorbance. By comparison to the cytotoxicity of the whole phenol red (solid line at ~0.8 fig DNA/well), this single region appeared to account for all of the cytotoxicity present in the whole phenol red. As in all other experiments, this cytotoxicity was strictly alkaline dependent (not shown). Figure 8 contains the results of the estrogen bioassay of the same fractions assayed above for the pH-dependent cytotoxicity. The solid line at approximately 2.2 ^g DNA/well represents cells present after 4 days of control cell growth in the absence of estrogens (Ctrl), while the broad line at about 4.8 iig DNA/well represents cell growth after 4 days of maximal estrogen-stimulated cell proliferation (E Max). Most of the HPLC fractions contained no estrogenic activity and did not stimulate cell growth above the control level. However, substantial estrogenic activity was observed in fractions that eluted between 4.5-9 min and appeared as three peaks (Fig. 8).

FIG. 6. A, The HPLC scan of whole phenol red, marked for six regions, B-G. B, After extraction, the aqueous phase contained only one peak, the purified phenol red (region B). C, The HPLC profile of the organic phase showed the contaminant peaks seen in the whole phenol red, while the major phenol red peak was removed. The split in the major contaminant peak in region C is an apparent off-scale absorbance artifact of the equipment used. All scans show from 0-10 min after injection.

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED

3327

FIG. 7. Cytotoxicity assay of HPLC profile of phenol red 0.5-min fractions, as described in Materials and Methods, assayed at 10 figeq/ml. The UV absorbance profile at 240 nm is superimposed over the bioassay for comparison. The only cytotoxicity was associated with the two fractions that eluted from 3.5-4.5 min. Because of the compressed shape of the dose-response curves, the cytotoxic activity in the fraction from 4-4.5 min is not linearly related to that from 3.5-4 min and is much less than may appear. pH 8, PhR-free: Control incubation in HBSS at pH 8.0 without phenol red; pH 8, with PhR: cytotoxicity of whole unfractionated phenol red at pH 8.0,10 \igl ml. E and e represent the locations of major and minor estrogenic activities, respectively, in the same fractions (from Fig. 8). Values are the mean ± SEM (n = 3 wells).

1 2

3

The strongest estrogenic activity eluted between 6.5-7.5 min, with additional activity eluting from 4.5-6 min and 8-9 min. There was little or no significant estrogenic activity associated with the two fractions between 3.54.5 min that contained the cytotoxicity. The cytotoxic contaminant, therefore, was distinct from the estrogenic contaminants. Phenol red indicator itself, which eluted between 2-3.5 min, showed neither cytotoxicity nor estrogenic activity in these assays (Figs. 7 and 8). Estrogen independence of cytotoxicity A signal property of MCF-7 cells is that they are estrogen responsive. To determine whether the pH-dependent cytotoxicity might involve an estrogenic pathway, cells were exposed to phenol red at pH 8.0 in the presence of active concentrations of either 17/3-estradiol (0.1 nM) or the antiestrogen tamoxifen (1 /XM). If cytotoxicity involved an estrogenic (or antiestrogenic) pathway, then the presence of estrogen (or antiestrogen) would modify the development of the pH-dependent toxicity. However, cytotoxicity was observed at pH 8.0 in phenol red-containing HBSS, regardless of the addition of estrogen or tamoxifen (Fig. 9), and cells incubated at pH 8.0 in phenol red-free medium, regardless of the addition of hormone or antihormone, were not affected (Fig. 9). In addition, the cytotoxicity (with or without estradiol or tamoxifen) was observed in control cultures that had been maintained estrogen free (phenol red free) for 3 days before the experiment (not shown) as well as in cells estrogenized by culture in phenol red-containing

4

5

6

7

8 9 Minutes

10 11 12 13 14 15 16

medium before assay (Fig. 9). Therefore, neither prior treatment with estrogen nor an estrogenic mechanism seem to be involved in the pH-dependent toxicity of a contaminant(s) of phenol red in MCF-7 cells.

Discussion A contaminant of commercial preparations of phenol red can be rapidly toxic to MCF-7 cells when the pH of the medium is elevated only slightly above 7.4. The pHdependent cytotoxicity is distinct from the estrogenic activity associated with phenol red that has been described previously (2-5). Because the estrogen-dependent breast cancer-derived MCF-7 cells are widely used in endocrine studies in tissue culture, the cytotoxicity of contaminants of phenol red could dramatically affect experiments with these cells if the pH of the medium is allowed to rise, even if only temporarily. Phenol red is used routinely as a pH indicator in cell culture media, and pH-dependent cytotoxicity by phenol red could be a confounding factor in many studies of cell physiology and elevated pH in tissue culture. These may include the areas of elevated pH and cell proliferation (19-21), pH-dependent effectiveness of some cytotoxins (22, 23), and pH-dependent effectiveness of some anticancer agents investigated in tissue culture (24-26). The common wisdom that recently thawed cells may be particularly sensitive to elevated pH could also be related to the alkaline-dependent cytotoxicity in phenol red. The cytotoxic and estrogenic activities of commercial

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED

3328

FIG. 8. Estrogen bioassay of HPLC profile of phenol red, as described in Materials and Methods, at 40 /igeq/ml medium. Higher equivalent concentrations of fractions were used to increase the sensitivity of the bioassay, since estrogenic activity was distributed across several peaks. The major peak of estrogenic activity eluted from 6.5-7.5 min, with minor activities eluting at 4.5-6.0 and 8.0-9.0 min. Because of the compressed shape of the dose-response curves, the activity in the major peak is perhaps 10 or more times the activity in the minor peaks. The UV absorbance profile at 240 nm is superimposed over the bioassay for comparison. Ctrl, Control incubation in estrogen-free medium; E max, maximal estrogen-stimulated proliferation; R 50%, half-maximal stimulation of response; T, the location of cytotoxicity in the same series of fractions (from Fig. 7). Values are the mean ± SEM (n = 4 wells).

Phenol Red-Free

10-

EMax

1

2

3

With Phenol Red

i

8DMA 6 - | per Well 42 Control

pH 8

+E2 +Tam

Endo-1991 Voll29«No6

pH 8

+E2 +Tam

FIG. 9. Estrogen independence of cytotoxicity. MCF-7 cells were incubated for 4 h at 37 C in complete maintenance medium (Control), phenol red-free HBSS at pH 8.0 ( • under Phenol Red-Free), or HBSS at pH 8.0 containing phenol red at 10 Mg/ml {M under With Phenol Red). The incubations were either not augmented (pH 8), augmented with 0.1 nM 17/3-estradiol (+E2), or augmented with 1 nM of the antiestrogen tamoxifen (+Tam). Cytotoxicity was neither increased nor decreased by the hormones. Values are the mean ± SEM (n = 4 wells).

phenol red are due to contaminants that can be easily separated from the phenol red itself by ether extraction, although it has been reported that the estrogenic contaminants can reappear in the purified phenol red under some conditions (13). HPLC fractionation of commercial phenol red showed that the fractions that contained the cytotoxicity were distinct from the fractions associated with the estrogenic activity. Our finding of several estrogenic fractions (one major estrogenic peak, with two minor estrogenic peaks eluting shortly before and after the major peak) corresponds to the multiple activities that competed with 17/3-estradiol for binding to the estrogen receptor (13, 14). Our studies extend the demon-

4

5

6

7

8

9

10 11 12 13 14 15 16

stration of estrogenic activity to two more of these multiple peaks, in addition to the main estrogenic activity, which has been identified as a bisphenolic phenyl sulfonate (14). Cytotoxicity was less pronounced in serum-containing media or at higher cell densities, and this may be due to several factors. Reduced local pH at higher cell densities may buffer the cytotoxic effects, and metabolism of the compound(s) by the cells may play a role in detoxification. Rapid metabolism of the main estrogenic contaminant has been suggested by its low potency in vivo (0.10.2% that of 170-estradiol) compared to cell-free determinations of relative binding affinity for estrogen receptors (50% that of 17/3-estradiol) (14). The cytotoxic contaminants) may also bind to serum proteins when serum is present, as do phenol red and the estrogenic activity (4, 18), lowering the free concentration of the compound^) in serum-containing medium relative to that in serum-free medium. The cytotoxic and estrogenic contaminants in phenol red are less water soluble than is the indicator, and concentrated aqueous stocks may not contain fully dissolved estrogenic and/or cytotoxic activities (see Materials and Methods). The mechanism for pH-dependent cytotoxicity is unknown. The structure of the cytotoxic component (isolation in progress), if pH sensitive as is the structure of phenol red, could shift from a charged to an uncharged species with increased pH, providing a mechanism for pH-dependent entry into the cell. Because of the rapidity with which cytotoxicity developed, and the open washedout appearance of the damaged cells, it is interesting to

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED consider whether the pH-dependent cytotoxicity may be somehow related to the alkaline pH-dependent inhibition of Na + and K+ channels by calcium or other ions (2729), which has been described in some detail. However, we have no direct information on this. Alkaline-dependent toxicity for a cancer-derived cell line could involve the elevated internal pH associated with rapidly proliferating cells (19, 20). However, both estrogen-treated and estrogen-free cells, which are proliferating more rapidly and less rapidly, respectively, appeared equally sensitive to the pH-dependent cytotoxicity, and this may not be consistent with a mechanism that involves relative internal pH. Phenol red has been used intentionally as the pH indicator of choice and inadvertently as an estrogen for estrogen-dependent MCF-7 cells, since the cell line was established (1). Since several cell lines have been shown to develop estrogen-independent phenotypes when cultured long term in estrogen-free (phenol red-free) medium (6-12), the estrogenic contaminants in phenol red may be said to have aided the development of estrogendependent cell lines in tissue culture. Phenol red is also a convenient estrogen to use in tissue culture, because the estrogens that fortuitously contaminate phenol red are more easily displaced from estrogen receptors than are stronger estrogens, such as 17/?-estradiol. Because of this, the estrogens associated with the indicator have not interfered with the assay of estrogen receptors by [3H] estradiol binding. The alkaline-dependent cytotoxicity we have shown is in addition to the estrogenic activity (2) and possible glucocorticoid activity (30) present in the indicator. MCF-7 cells, a hormone-dependent human breast cancer-derived cell line, were very sensitive to this pHdependent cytotoxicity. We are currently examining the sensitivity of a number of other cell lines (Welshons, W. V., L. H. Grady, and G. E. Rottinghaus, in preparation). At present, of seven cells tested [including one estrogen receptor-positive (T47D) and two estrogen receptor-negative (MDA-MB-231 and BT-20) breast cancer-derived cell lines], only MCF-7 cells have been affected. The reason for this is not evident; however, more light may be shed on this question by identifying the structure of the cytotoxic contaminant or by further testing of cell types, both of which are in progress. At any rate, the possibility of specific toxicity for a breast cancer cell line may have an anticancer application, and this merits further investigation. The impact on a great many cell lines that are in use remains to be determined, and this pH-dependent cytotoxicity could be a significant factor in many kinds of studies. In summary, commercial preparations of phenol red contain a contaminant that can be toxic to MCF-7 cells in modestly alkaline buffers and media. This contami-

3329

nant is distinct from the estrogenic contaminants reported previously. Because of the contaminants in phenol red that have hormonal and now complicated cytotoxic activities, it should be considered whether to omit the phenol red or use purified phenol red that is free of contaminants, and this is particularly important in endocrine studies. For estrogen-dependent cell lines, this would also require supplementing the medium with a sufficient concentration of an estrogen to maintain the estrogen-dependent phenotype; for MCF-7 cells in our lab this would be equivalent to 0.1 nM 17/3-estradiol for peak growth rates.

References 1. Soule HD, Vazquez J, Long A, Albert S, Brennan M 1973 A human cell line from a pleural effusion derived from a breast carcinoma. J Natl Cancer Inst 51:1409-1413 2. Berthois Y, Katzenellenbogen JA, Katzenellenbogen BS 1986 Phenol red in tissue culture media is a weak estrogen: implications concerning the study of estrogen-responsive cells in culture. Proc Natl Acad Sci USA 83:2496-2500 3. Hubert J-F, Vincent A, Labrie F 1986 Estrogenic activity of phenol red in rat anterior pituitary cells in culture. Biochem Biophys Res Commun 141:885-891 4. Welshons WV, Wolf MF, Murphy CS, Jordan VC 1988 Estrogenic activity of phenol red. Mol Cell Endocrinol 57:169-178 5. Glover JF, Irwin JT, Darbre PD 1988 Interaction of phenol red with estrogenic and antiestrogenic action on growth of human breast cancer cells ZR-75-1 and T-47-D. Cancer Res 48:3693-3697 6. Katzenellenbogen BS, Kendra KL, Norman MJ, Berthois Y 1987 Proliferation, hormonal responsiveness, and estrogen receptor content of MCF-7 human breast cancer cells grown in the short-term and long-term absence of estrogens. Cancer Res 47:4355-4360 7. Welshons WV, Jordan VC 1987 Adaptation of estrogen-dependent MCF-7 cells to low estrogen (phenol red-free) culture. Eur J Cancer Clin Oncol 23:1935-1939 8. Clarke R, Brunner N, Katzenellenbogen BS, Thompson EW, Norman MJ, Koppi C, Paik S, Lippman ME, Dickson RB 1989 Progression of human breast cancer cells from hormone-dependent to hormone-independent growth both in vitro and in vivo. Proc Natl Acad Sci USA 86:3649-3653 9. Murphy CS, Meisner LF, Wu SQ, Jordan VC 1989 Short- and long-term estrogen deprivation of T47D human breast cancer cells in culture. Eur J Cancer 25:1777-1788 10. Murphy CS, Pink JJ, Jordan VC 1990 Characterization of a receptor-negative, hormone-nonresponsive clone derived from a T47D human breast cancer cell line kept under estrogen-free conditions. Cancer Res 50:7285-7292 11. Daly RJ, Darbre PD 1990 Cellular and molecular events in loss of estrogen sensitivity in ZR-75-1 and T-47-D human breast cancer cells. Cancer Res 50:5868-5875 12. van den Berg HW, Martin J, Lynch M 1990 High progesterone receptor concentration in a variant of the ZR-75-1 human breast cancer cell line adapted to growth in oestrogen free conditions. Br J Cancer 61:504-507 13. Bindal RD, Carlson KE, Katzenellenbogen BS, Katzenellenbogen JA 1988 Lipophilic impurities, not phenolsulfonphthalein, account for the estrogenic activity in commercial preparations of phenol red. J Steroid Biochem 31:287-293 14. Bindal RD, Katzenellenbogen JA 1988 Bis (4-hydroxyphenyl) [2(phenoxysulfonyl)phenyl]methane: isolation and structure elucidation of a novel estrogen from commercial preparations of phenol red (phenolsulfonphthalein). J Med Chem 31:1978-1983 15. Welshons WV, Grady LH, Nonneman DJ, Rottinghaus GE, Specific pH-dependent toxicity, as well as estrogenic activity, of contaminants of phenol red for hormone-dependent human breast cancer cells. 73rd Annual Meeting of The Endocrine Society,

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

3330

pH-DEPENDENT CYTOTOXICITY OF PHENOL RED

Washington DC, 1991 (Abstract 572) 16. West DC, Sattar A, Kumar S 1985 A simplified in situ solubilization procedure for the determination of DNA and cell number in tissue cultured mammalian cells. Anal Biochem 147:289-295 17. Labarca C, Paigen K 1980 A simple, rapid and sensitive DNA assay procedure. Anal Biochem 102:344-352 18. Kragh-Hansen U, Moller JV, Sheikh MI 1972 A spectrophotometric micromethod for the determination of binding of phenol red to plasma proteins of various species. Pfluegers Arch 337:163-176 19. Soltoff SP, Cantley LC 1988 Mitogens and ion fluxes. Annu Rev Physiol 50:207-223 20. Moolenaar WH 1986 Effects of growth factors on intracellular pH regulation. Annu Rev Physiol 48:363-376 21. Zetterberg A, Engstrom W 1981 Mitogenic effect of alkaline pH on quiescent, serum-starved cells. Proc Natl Acad Sci USA 78:43344338 22. Sandvig K, Olsnes S 1982 Entry of the toxic proteins abrin, modeccin, ricin, and diphtheria toxin into cells. II. Effect of pH, metabolic inhibitors, and ionophores and evidence for toxin penetration from endocytotic vesicles. J Biol Chem 257:7504-7513 23. Langone JJ, Borsos T 1978 Increased susceptibility of tumor cells

24. 25. 26. 27. 28. 29. 30.

Endo«1991 Voll29«No6

and chicken erythrocytes to lysis by antibody and complement after treatment with aminoethylisothiouronium bromide hydrobromide (AET). J Immunopharmacol 1:43-59 Rottinger EM, Smith BD, Weidenmaier W 1986 pH-Induced modification of BCNU toxicity on glial cells. Radiother Oncol 5:23-27 Palcic B, Skov KA, Skarsgard LD 1980 Effect of reducing agents on misonidazole cytotoxicity. In: Brady LW (ed) Radiation Sensitizers. Masson, New York, pp 438-440 Koch CJ, Howell RL 1982 Misonidazole: inter-related factors affecting cytotoxicity. Int J Radiat Oncol Biol Physiol 8:693-696 Garty H, Asher C, Yeger O 1987 Direct inhibition of epithelial Na+ channels by a pH-dependent interaction with calcium, and by other divalent ions. J Membr Biol 95:151-162 Jacobsen C, Mollerup S, Sheikh MI 1990 Ca2+ and pH regulation of K+ channels in membrane vesicles of rabbit proximal tubule. Am J Physiol 258:F1634-F1639 Huang WH, Askari A 1984 Regulation of (Na+ + K+)-ATPase by inorganic phosphate: pH dependence and physiological implications. Biochem Biophys Res Commun 123:438-443 Picard D, Yamamoto KR 1987 Two signals mediate hormonedependent nuclear localization of the glucocorticoid receptor. EMBO J 6:3333-3340

The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 19 December 2014. at 02:24 For personal use only. No other uses without permission. . All rights reserved.

pH-dependent cytotoxicity of contaminants of phenol red for MCF-7 breast cancer cells.

The pH indicator phenol red (phenolsulfonphthalein) is present in most tissue culture media. Contaminants of this indicator have shown substantial est...
2MB Sizes 0 Downloads 0 Views