Teratogenesis, Carcinogenesis, and Mutagenesis 10:449-462 (1990)

Evaluation of the Genotoxicity of Gentian Violet in Bacterial and Mammalian Cell Systems Anane Aidoo, Ning Gao, Robin E. Neft, Henry M. Schol, Bruce S. Hass, Toni Y. Minor, and Robert H. Heflich Department of Health and Human Services, Food and Drug Administration, Division of Genetic Toxicology, National Center for Toxicological Research, Jefferson. Arkansas Previous studies indicate that gentian violet (GV), a triphenylmethane dye used in agriculture and human medicine, is a clastogen in vitro and a carcinogen in chronically exposed mice and rats. Data on its genotoxic activity, however, have been incomplete and partly contradictory. Mutagenesis and DNA damage experiments were conducted to re-evaluate the genotoxic potential of GV in both bacterial and mammalian cell systems. GV was mutagenic in Salmonella typhimurium tester strains TA97 and TA104, but there was little mutagenic activity detected in strains TA98 and TAIOO. A rat liver homogenate fraction (S9) tended to increase mutagenicity. The major microsomal metabolites of GV, pentamethylpararosaniline and N,N,N',N'-tetramethylpararosaniline were less mutagenic in TA97 and TAlM, while N,N,N',N-tetramethylpararosaniline was a weak mutagen in Salmonella. GV was not mutagenic in Chinese hamster ovary (CHO) cell strain CHO-K1-BH4, and was a questionable mutagen in CHO-AS52 cells. While GV produced DNA damage as measured by sedimentation of nucleoids derived from B6C3FI mouse lymphocytes treated in vitro, no damage was found in lymphocytes isolated from mice dosed with GV. GV was also a weak producer of gene amplification in an SV40-transformed Chinese hamster cell line. The results indicate that GV is a point mutagen in bacteria; however, since similar exposure conditions produced weak mutagenic activity in mammalian cells, GV may be carcinogenic by virtue of its clastogenic activity. Key words: CHO-Kl-B&, CHO-AS52, B6C3FZmouse lymphocytes, Salmonella typhimurium, DNA damage, mutagenesis

INTRODUCTION

Gentian violet (GV) or crystal violet is a mixture of aminophenylmethane dyes [ 1,2]. Hexamethylpararosaniline is the predominant component of the mixture (about

96% in commercial preparations), with the remainder being mainly methyl violet or Ning Gao is now at the Drug Inspection Institute of Inner Mongolia, Huhhot, PRC.

0 1990 Wiley-Liss, Inc.

450

Aidoo et al.

pentamethylpararosaniline. GV has been used in human and veterinary medicine to control bacterial and fungal skin infections [2,3], and it is widely used by blood banks to prevent the spread of Chagas disease, which is transmitted through blood transfusion. In addition, GV has been added to livestock feed to prevent fungal growth [4]. GV is also a common laboratory stain. Human exposure may occur as a result of the various uses of this dye. GV is toxic and carcinogenic in chronically exposed B6C3FI mice and Fischer 344 rats [5,6]. Information concerning its genotoxicity in short-term assays, however, is both incomplete and partially contradictory. Various studies indicate that GV can damage DNA [7,8] and that GV is a potent in vitro clastogen [9,10]. In contrast, GV was not clastogenic in an in vivo study [ 111, and results Concerning its activity in the Salmonella typhimurium reversion assay have been inconclusive. In a series of published reports, positive [ 12,131, negative [ 11,141 and questionable positive [ 151 responses have been found in Salmonella. Fujita [ 121 and Thomas and MacPhee [ 141, however, found GV mutagenic in Escherichia coli. There are no published studies on the ability of GV to induce gene mutations in mammalian cells. This report re-evaluates the mutagenicity of GV in the Salmonella typhimurium mammalian microsome test, and utilizes two tester strains, TA97 and TA104, that have not previously been used for assaying this compound. In this way, we hoped to broaden the sensitivity of the Salmonella assay to detect potentially mutagenic GV damage that may have been missed in previous analyses. The major demethylated products of GV and metabolism, pentamethylpararosaniline, N,N,N’,N’-tetramethylpararosaniline, N,N,N’,N”-tetramethylpararosaniline [ 161, were also tested for bacterial mutagenicity. In addition, GV was assayed for mutagenicity in two Chinese hamster ovary (CHO) cell strains, CHO-K1-BH4 and CHO-AS52. These cell lines possess different mutational targets, and can potentially delineate both gene and chromosomal mutations. While CHO-K l-BH4 cells measure gene mutations at the hypoxanthine guanine phosphoribosyl transferase (hprt)locus, a hemizygous X-linked gene, AS52 cells measure mutations using the chromosomally integrated bacterial xanthine phosphoribosyl transferase (xprt or gpt) locus. The gpt locus is autosomal and appears to permit recovery of multilocus deletion mutations as well as gene mutations [ 171. To compare the in vivo and in vitro DNA-damaging potential of GV, spleen lymphocytes from B6C3F1 mice were exposed both in vivo and in vitro to various concentrations of GV. The DNA damage produced was then determined by nucleoid sedimentation [ 181. In order to further investigate the effects of GV on DNA, SV40-transformed Chinese hamster embryo C060 cells were exposed in vitro to GV and molecular analysis was performed in this model system for gene amplification [ 19,201. MATERIALS AND METHODS

Chemicals/Reagents Stock solutions of GV (Aldrich, Milwaukee, WI [manufacturer’s assay 97% dye] and Hilton-Davis, Cincinnati, OH [99% hexamethylpararosaniline]), methylmethane sulfonate (MMS, Eastman, Rochester, NY), triphenylmethane (Fluka, Ronkonkoma, NY), 7,12-dimethylbenzanthracene(DMBA, Aldrich), benzo(a)pyrene (BP, Aldrich), pentamethylpararosaniline (Hilton-Davis), N,N,N’,N’-tetramethylpararosaniline (HiltonDavis), N,N,N’,N’’-tetramethylpararosaniline(Hilton-Davis), and ethidium bromide (EB, Sigma Chemical Co., St. Louis, MO) were prepared fresh in dimethyl sulfoxide.

Gentian Violet Genotoxicity

451

S9 was prepared from male Sprague-Dawley rats pretreated with Arocolor 1254 by the method of Ames et al. [21]. Salmonella/Mammalian Microsome Mutagenicity Assay

Reversions were measured in Sulmonc~llutyphimurium tester strains TA97, TA98, TA100, and TA104 using the plate-incorporation assay of Maron and Ames [22]. For assays employing S9 activation, 2 ml of molten top agar was mixed with 500 pl of S9 mix (containing 50 p l of liver homogenate), 100 pl of an overnight tester strain culture, and 100 pl of the test chemical or solvent. For assays performed in the absence of S9, bacteria and test chemical were mixed with 2.5 ml of molten top agar. The mixtures were poured into plates containing Vogel's minimal salts agar with 2% glucose, and incubated at 37°C for 48 hours. MMS and BP were included as positive controls. All assays were conducted in triplicate and each chemical was tested on a minimum of two separate occasions with comparable results. Laboratory procedures were performed under yellow safety lights. Mammalian Cell Mutagenicity Assays

Stock cultures of CHO-K1-BH4 and AS52 cells (both obtained from Dr. A.W. Hsie, Galveston, TX) were maintained as previously described [23]. Forward mutation assays at the hprt locus in CHO-K1-BH4 cells and the gpt locus in AS52 cells were conducted following the general procedures described in Machanoff et al. [24] and Gao et al. [23]. Cultures, established at 1 >: lo6cells/lOO-mmdish on the previous day, were exposed to 0- 1.5 pg/ml of GV for S hours in serum-free nutrient mixture F12. Assays were conducted in the presence and the absence of an S9 activation system containing 400 pg/ml of S9 protein prepared from rats pretreated with Aroclor 1254. A sufficient number of cultures was treated with each test concentration of GV to ensure that a minimum of 1 X lo6 cells survived the considerable toxicity of the compound. After treatment, cells were washed with Ca+ +/Mgt +-free phosphate-buffered saline (PBS) and allowed to recover overnight in fresh F12 medium with 5 % fetal bovine serum (FBS). The cells were then plated in F12 medium with 5% FBS and incubated for 7 days with three passages to permit expression of induced mutants. Parallel cultures of 400 cells/60-mm dish were also established for cytotoxicity determination. At the end of the phenotypic expression period, 2 X lo6 cells were plated at a concentration of 2 X lo5 celIs/lOO-mrn dish in hypoxanthine-free F12 containing 5% dialyzed FBS and 10 p M 6-thioguanine. Cloning efficiency cultures, grown in medium without 6-thioguanine, had 200 cells/60-mm plate. All cultures were incubated for 7 days before colonies were fixed, stained and counted. Results were expressed as 6-thioguanineresistant mutants/106 clonable cells. Nucleoid Sedimentation Analysis

DNA damage caused by GV exposure was measured in spleen lymphocytes from B6C3Fi mice using a modification of the nucleoid sedimentation procedures of Cook et al. [18]. For in vivo experiments, lymphocytes were isolated by lympho-paque centrifugation according to the method of Aidoo et al. [25] from animals pretreated for 1 hour with 0- 10 pg/ml GV administered by injection into the tail-vein. Animals treated with MMS were used as positive controls. For in vitro studies, lymphocytes isolated from untreated animals were exposed to GV (0- 1.O pg/ml) in serum-free RPMI 1640 medium for 1 hour. MMS and EB were used as positive controls, with EB exposure

452

Aidoo et al.

occurring during nucleoid sedimentation. Cells were washed with PBS and resuspended in PBS at a concentration of 2 to 10 X lo6 cells/ml. A 50 p1 sample was then added to 150 p1 of lysing solution (2.5 M NaC1, 0.133 M EDTA, 2.67 mM Tris, and 0.67% Triton X-100, pH 8.0) layered on top of a 15-30% neutral (pH 8.0) sucrose gradient containing 1.95 M NaCl, 0.01 M Tris, 0.001 M EDTA, and 0. I kg/ml Hoechst 33258 dye (Calbiochem, La Jolla, CA). When EB was used, sucrose gradients contained the indicated amount of EB and the Hoechst dye was omitted. Following a 30-minute lysis in the dark at room temperature, the gradients were centrifuged at 25,000 rpm for 75 minutes at 20°C in an SW41 rotor (Beckman Instruments, Palo Alto, CA). Six gradients were run per rotor with one gradient serving as a reference (solvent control). The position of the nucleoids in the gradients was visually determined with a UV light. Results were expressed as the ratio of the distance traveled by nucleoids in the sample gradients to the reference gradient. Gene Amplification Analysis Cell culture and treatment. C060 cells (kindly provided by Dr. S. Lavi, Tel Aviv, Israel) were maintained at 37"C, 5% CO;? in Dulbecco's modified Eagle medium (DMEM) containing penicillin (100 U/ml), streptomycin (100 pg/ml), and 10% FBS. All medium components were from GIBCO (Grand Island, NY). First, 5 X lo5 cells were seeded into 75 cm2 tissue culture flasks. Twenty-four hours later the cells were treated with 0.02, 0.05, 0.125 pg/ml GV, DMSO (vehicle control), or 0.1 pg/ml DMBA (positive control) in serum-free DMEM for 5 hours. Following exposure, the cells were rinsed with PBS and 15 ml of fresh DMEM with serum was added to each flask. At 4 days post-treatment, DNA was isolated from the cells according to the method of Beland et al. 1261. Slot blotting. DNA samples (0, 1, 5, and 10 pg) were applied to duplicate nylon membranes (Hybond-N; Amersham, Arlington Heights, IL) using a Bio-Rad Bio-Dot SF slot blot apparatus. The nylon membranes were pre-hybridized in 25 mM KH2P04, pH7.5,5 x salinesodiumcitrate (SSC), 5 x Denhardt's solution, 50 pg/ml heat-denatured salmon sperm DNA, 50% formamide, and 1% sodium dodecyl sulfate (SDS) for 2 hours at 42°C. Dextran sulphate was then added to a concentration of 10% and the blots were hybridized overnight at 42°C with a radiolabeled probe prepared from SV40 DNA (Bethesda Research Laboratories, Gaithersburg, MD). The probe was labeled with [a-'2P]dCTP (3,000 Ci/mmol; ICN Biomedical, Costa Mesa, CA) to a specific activity of 2 X lo8 cpmlpg using a random primed DNA labeling kit (Boehringer Mannheim, Indianapolis, IN). The slot blots were washed successively in 2 x SSC, 0.1% SDS at 42"C, 0.2 x SSC, 0.1% SDS at 55"C, and 2 changes of 0.2 X SSC, 0.1% SDS at 42°C. Autoradiography was performed by exposure to Kodak XAR-5 film with intensifying screens at - 70°C for 5 days. The resulting hybridization signals were quantified with a Zeineh scanning densitometer (model SLR-2D/1D; Biomed Instruments, Fullerton, CA). Following this procedure, the membranes were stripped and reprobed with "P-labeled human p-actin cDNA 1271 (gift of Dr. D. Peffley, Memphis, TN). P-Actin DNA sequences are not amplified following exposure of C060 cells to GV, and therefore they can be used as an internal control for the amount of cellular DNA bound to the filters.

Gentian Violet Genotoxicity

453

RESULTS

Au et al. [9] and Bonin et al. [ 131 showed that GV genotoxicity results vary from batch to batch; therefore, we tested GV from two sources, Aldrich and Hilton-Davis. As shown in Table I, GV mutagenicity in these batches was similar. Based on the magnitude of the responses, their reproducibility, and their dose-responsiveness in the presence of S9, GV appeared to be mutagenic in strain TA97 with and without S9 activation and in strain TA104 with S9. The strongest mutagenic responses were found in strain TA97 in the presence of S9. GV produced dose responsive increases in mutagenicity up to a concentration of approximately 0.5 pg/plate where in excess of 200 revertants/ plate over the solvent controls were induced. Without S9,O. 1-0.5 pg/plate of GV produced increased numbers of revertantdplate in TA97, but no dose-response was seen. GV also produced a,positive, dose-responsive increase in reversions in strain TA104 in the presence of S9, reaching a maximum of more than 300 revertantdplate over the solvent controls at a concentration of 5 pgiplate. Weak mutagenic responses were found in TA100, while GV was essentially non-mutagenic in TA98 and in TA104 without S9. At higher doses and when no S9 was present, GV was sufficiently toxic to kill a large proportion of the TA97 cells and fewer mutant colonies were observed. However, at the same doses in the presence of S9, GV was apparently detoxified and the mutant yield was increased. Since S9 generally increased the mutagenicity of GV in TA97 and GV was mutagenic in TA104 only in the presence of S9., the mutagenic activities of three previously identified demethylated metabolites of GV were examined. Triphenylmethane was also included as a negative control. As shown in Table I, pentamethylpararosaniline and N,N,N’,N‘-tetramethylpararosanilinewere weakly positive in strains TA97 and TA104, with assays performed in the presence of S9 generally producing higher responses. N,N,N’,N”-Tetramethylpararosaniline was also a weak mutagen, while triphenylmethane was generally negative in all tester strains. The mutagenic activity of GV in CHO cell strains is shown in Table 11. At the concentrations tested, the results were negative in CHO-K1-BH4 cells, while 66 mutants per lo6 clonable cells were obtained i.n AS52 cells at a GV concentration producing a high level of toxicity. This positive response, however, was not consistently found in subsequent experiments. Toxicity was apparent in CHO cells, both with and without metabolic activation, although it was more evident in assays conducted without S9. GV-treated cells, especially AS52 cells, became tenaciously attached to the plates such that it required longer trypsinization times to dissociate them during subculture. When observed with an inverted microscope, the exposed cells appeared larger and more spindle-shaped than control cells. No DNA damage was detected by nucleoid sedimentation analysis of lymphocytes isolated from B6C3F1 mice exposed to different concentrations of GV (Fig. 1). Concentrations of GV higher than 8 mg/kg were fatal to the mice during the 1 hour treatment period. In cells exposed to GV in vitro, DNA damage was produced, even at low concentrations of the chemical (Fig. 1). To confirm that nucleoid sedimentation under these conditions would detect damage produced in vivo by a known DNAdamaging agent, B6C3F1 lymphocytes were exposed to specific concentrations of MMS both in vivo and in vitro. As shown by Figure 2, DNA damage was detected in both systems. In nucleoid sedimentation analysis, the degree of supercoiling or relaxation of the DNA structure is used as a measure of DNA damage [18]. Therefore, to make

43 t 10 4 4 2 04 56 2 03 44 2 02 55 2 05 4 8 ? 03 49 ? 01 260 5 42 331 2 18 350 2 05 374 ? 31 378 ? 24 401 ? 33 354 5 48

230 2 296 2 301 2 331 2 217 2 279 2 216 2

0" 0.1 0.25 0.5 1 .o 2.5 5.0

Pentamethylpararosaniline

02 33 14 19 17 06 22

49 42 43 45 39 57 49 53

281 ? 04 479 ? 22 5 1 6 ? 29 515 It 12 453 2 50 470 2 32 273 ? 16 6 2 ? 27

240 2 10 373 2 23 349 2 47 381 2 20 294 2 18 38 ? 02 Toxic Toxic

Oa 0.1 0.25 0.5 1.0 2.5 5.0 10.0

Gentian violet (Hilton Davis)

?

11

2 II 2 04

&

06 f 06 2 05 2 01

2 04

06 10 10 01 10 ? 06 2 04 2 01

2 2 2 2 2

28 23 27 24 30 33 34 25

137 2 06 175 2 19 241 ? 27 374 ? 21 3 4 4 ? 44 269 t 18 210 2 35 Toxic

136 2 19' 216 t 09 208 ? 13 2 0 6 ? 17 57 2 16 98 2 24 32 2 04 Toxic

0" 0.1 0.25 0.5 1.0 2.5 5.0 10.0

Gentian violet (Aldnch)

-s9

+s9

Compound

- S9

258 -t 27 259 t 20 262 2 26 250 ? 35 242 ? 45 237 ? 44 239 ? 20

09 29 22 20 25 28 25 22

243 ? 35 235 t 15 235 t 37 227 ? 06 211 ? 00 195 ? 30 201 ? 09 56 2 51 2 53 2 51 ? 51 ? 59 2 68 2

?

?

04 05 08 04 09 08 04

t 03

2 04

242 ? 29 237 5 49 291 2 37 301 2 23 298 2 36 339 2 13 339 ? 43 233 ? 06

34 34 22 38 45 66 52 40 268 t 20 264 ? 8 263 ? 16 256 k 36 291 2 15 294 t 08 283 t 27 179 ? 25

+s 9

49 ? 04 60 ? 06 66 2 06 63 2 09 63 2 13 76 2 14 81 ? 07 79 ? 11

TAlOO

09 04 -t 05 2 10 ? 03 2 07

-s 9

250 2 222 ? 228 ? 260 ? 308 2 351 2 351 2 305 2

+s 9

Revertantslplate

226 2 11 223 2 17 232 2 24 233 t 09 261 ? 19 291 ? 24 269 2 16 185 t 18

TA98

Dose (Fgiplate)

TA97

TA 104

609 2 29 6 0 4 2 50 637245 607 ? 0 8 682-t 18 6 5 6 2 17 782245

23 20 31 47 -t 25 t 55 2 25 627 543 561 634 596 680 736

(cotztinued)

367221 390f32 383?07 498 f 29 5 1 5 2 13 602247 771 2 5 3 649254

16 22 13 18 23 55 51 33 2 2 2 2 2 2 2 2

333 315 335 325 342 406 358 323

2 2 2 2

4 4 3 2 16 455 2 10 540k32 618250 585 2 32 737? 22 755 2 3 0 492261

37 32 31 08 28 ? 28 2 36 2 31

+s9 2 2 2 2 2

-s 9

441 389 421 406 472 335 349 269

TABLE I. Mutagenicity of Gentian Violet, Some of Its Metabolites, and Control Compounds in the SuZmonellulMicrosomeTest*

671 !C 18 40 f 10

209 f 17

= not tested. aSolventcontrol, 100 p1DMSOIplate. bMean + standard deviation (n = 3).

*-

Benzo(a)Dyrene Methylrnkihanesulfonate

258 t 04 233 f 32 227 t 15 224 t 35 249 t 22 222 t 19 212 !C 36 283 f 16

195 t 07 197 f 13 182 t 16 187 ? 11 200 t 19 207 t 15 167 t 24 208 f 16

0" 0.1 0.25 0.5

Triphenylmethane

l(p1)

5.0

2.5 5.0 10.0

1.o

159 t 15 182 t 21 198 f 22 230 i 32 211 & 07 2Oi & 03 184 t 23 71f16

188 t 03 167 k 20 192 t 10 215 i 26 131 f 20 i29 _r i8 99 t 13 Toxic

0" 0.1 0.25 0.5 1.o 2.5 5.0 10.0

N,N,N',N"-tetramethylpararosaniline

42 t 03 35 t 01 43 f 12 39 i 06 43 t 06 3 7 f 10 38 t 03 39 f 06

23 t 03 2 6 f 05 18 k 06 29 t 06 24 t 12 25 t 05 19 t 03 27f04

1 5 9 t 15 156 t 23 156 f 49 230 i 06 231 f 21 277 f 07 211 t 21 247 t 66

188 f 03 162 t 17 150 f 44 175 t I 1 199 f 10 113 t 02 125 t 15 147 t 16

23 f 30 19 t 06 28 f 06 24 f 07 26 f 03 27 f 07 27 t 07 31 f 05

+s9

+s 9

-

208 t 27

979 t 164 -

-

06

10

19 13 12 06 38 17

1,390 f 79

204 ? 38 205 f 170 t 06 1 8 8 t 192 f 09 194 f 190 t 04 194 t 221 t 08 218 t 216 f 36 245 f 239 i 23 179 t 189 t 09 232 f

36 f 03 47 t 04 41 f 07 44 k 07 4 0 k 11 33 t 05 42 t 10 47 t 07

27 13 13 31 32 26 27 23

-

212 t 27 169 t 14 169 f 02 1 8 8 " 19 219 f 21 287 f 83 262 t 12 215 f 18

202 i 18 171 f 03 187 t 09 230 f 06 201 t 04 237 f 15 221 ? 32 199 t 21

30 t 02 39 f 07 28 f 01 29 !C 06 30 2 02 36 k 04 36 f 14 34 t 05

212 t 185 t 192 f 251 t 218 f 237 t 266 t 303 i

TA 100 202 t 18 221 t 19 220 t 06 246 i I 1 222 f 25 214 f 20 260 t 18 241 f 21

-s9

Revertantsiplate

30 t 02 40 f 15 36 i 06 34f I1 33 k 06 35 f 09 32 f 06 53 t 05

TA98 -s9

+s9

TA97 - S9

0" 0.1 0.25 0.5 1 .o 2.5 5.0 10.0

Dose (pgiplate)

N,N,N',N'-tetramethylpararosaniline

Compound

TABLE I. Continued

12 35 09 72 29 53 48 35

57 16 02 26 31 24 32 10

501 t 3 1 581 2 08 556 2 3 0 544k31 538254 5 2 7 2 16 5 1 7 2 13 479f 33

5 2 4 2 10 517+ 10 547 t 15 565 t 04 612222 581 2 4 0 6 4 2 2 12 386?33

5 4 2 t 10 546 2 56 541 f 50 602 f 46 581 t 25 609 t 26 642 f 11 560 k 26

+s 9

1,361 t 50 1,568 t 83 -

542 t 560 k 578 t 510 f 518 k 511 C 522 C 498 -t

458 t 394 t 382 t 399 f 412i 419 t 427 f 285 2

458 f 418 k 420 f 4442 504 k 467 2 490 t 311 2

52 06 21 20 22 48 11 103

TA 104 -s 9

Aidoo et al.

456

TABLE 11. Mutagenicity and Cytotoxicity of Gentian Violet (Aldrich) in CHO-Kl-BH, and CHO-AS52 Cells* Cloning efficiency Compound

CHO-K I

AS52

CHO-K1

0 0.02 0.04 0.05 0.06 0.08 0.10 0.125 0.25 0.50 1.oo

I00

100 144 98

17 20 23 8 1 3 2 6 -

Gentian violet ( - S9)

Gentian violet (

+ S9)

131 133 96 85 53 25 11 2