oine

Color Yes, Toxicity No: SystematicApproaches to Meeting this ChallengeBy Harold S. Freeman, Olney Medalist. North Carolina State University, Raleigh

While the efficacy of a new chemical sub-stance is critical to its commercial success,few researchers would argue against theproposition that the major barrier to thecommercialization of a technically inter-esting new compound often lies with itstoxicologicai properties. The right to in-troduce a new chemical product to themarketplace requires its manufacturer todemonstrate minimal human health andenvironmental risk in its production andintended end use.

With regard to textile dyes, safety isoften demonstrated by conducting a bat-tery of tests designed to assessgenotoxicity and ecotoxicity. Thegenotoxicity (adverse interactions be-tween DNA and chemical substancesleading to heritable changes in cells oforganisms) of dyes used for textile and

A B S T R A C T

This paper provides an overview of studies con-ducted at the North Carolina State UniversityCollege ofTextilessince the inception ofthe dyechemistry program in 1984. Design of new tex-tile dyes to replace existing dyes and dye inter-mediates were studied with the goal of mini-mizing human health and environmental risks.Initially, correlations between chemical struc-tures and environmental toxicity were estab-lished and the resulting information used indesign studies. These studies resulted innonmutagenic replacements for genotoxic aro-matic amines used in azo dye development andmetal compiexed dyes based on benign alterna-tives to priority pollutants. Key technical prop-erties of newly developed dyes and summariesof certain toxicologicai test methods and regu-latory policies used to ensure dye applicationsafety are presented.

Key Terms

Dyes

Environmental IssuesGenotoxicityHealth IssuesMutagenicitySynthesis

related applicationscame to the forefrontwhen it was discoveredthat certain commonlyused azo dye intermedi-ates were either humanor animal carcinogens.^This recognition led towidespread testing ofazo dyes and their inter-mediates for mutage-nicity and/or carcinoge-nicity and to theassessment of methodsused.^'^ While thesestudies were mainlyaimed at risk assessment, they were alsoused to design replacements for certaintoxic compounds. The present paper pro-vides an overview of our laboratory stud-ies pertaining to the environmental chem-istry of azo dyes, by far the major class oftextile dyes.

B A C K G R O U N D

Test Methods

Genotoxicity

The results described in this paper were ob-tained mainly from studies using the stan-dard Ames test or Prival modification toassess genotoxicity, or from the use ofCeriodaphnia or Lemna minor to assessaquatic toxicity.^"'° The Ames test has beenused as a cost effective "first look" at thecarcinogenic potential of organic com-pounds. Specifically, this test is used todetermine whether a compound is mu-tagenic, a property believed to be associ-ated with the early steps of caronogenesis.The Ames test uses strains of Salmoriellabacteria that are sensitive to frameshift andbase-pair substitution mutations and un-able to grow (replicate) in the absence ofhistidine. Exposures of 5a/mone//a bacteriato a test compound in which the bacteriarepiicate to levels at least twice the back-

^ r

'''^'-^^SOjNa

NHz

rST

N

— NH2

v_I

1+

0Azo reductase

H 2 N -

sos^

—^^^v^TNH2

B«RDdine

' - ' ^

>

Fig. 1. Reductive cleavage of C.I. Direct Red 28 (structure 1) usingan azo reductase enzyme system.

ground count lead to the designation ofthe test compound as mutagenic. This testis also conducted in the presence of en-zyme systems (S9) capable of producingmetabolites of the test compound. In thisway, an assessment of the potential for thetest compound to function as an indirect-arting mutagen is possible.

The development of the Prival modifi-cation of the Ames test grew out of aneed to enhance the reductive-cleavageof azo dyes during the metabolic process.This was important because azo dyes suchas Congo Red (C.I. Direct Red 28; struc-ture 1) are not direct-acting mutagensand thus often give a nonmutagenic re-sponse in the Ame5 test, despite beingderived from the well-known human car-cinogen benzidine. The Prival modificationuses hamster liver S9 in lieu of rat liver S9,the former being richer in azo reductaseenzymes and demonstrating the impor-tance of azo group reductive cleavage (Fig.1) in the mutagenesis process.

The choice of test methods for evalu-ating carcinogenicity is important. Datafrom a variety of published methods havebeen assessed for utility.^ The resultingdata for 97 chemicals, mainly azo dyes,associated with the dye industry producedthe following conclusions:

16

AATCC REVIEW

Voineymedal

h Less than one-third of the 97 compounds reported hadbeen adequately tested to permit their carcinogenicity to bejudged with a high degree of confidence.

• Test methods involving urinary bladder implantation, re-peated injections, or very few test animals are inappro-priate for evaluating chemical carcinogenicity.

• The purity of test compounds must be sufficient to at-tnbute the test results to the compound in question andnot an impurity

• The preferred testing involves at least two species of ani-mals and non-conflicting data; a minimum group of 25animals of each sex, with a sufficient number survivingabout two-thirds of their expected life span; and chronicapplication of test compound at a range of doses up toand including a level giving a biological effect.

Aquatic Toxicity

In keeping with the standard method recommendationsfor examining water and wastewater, Lemna minor(duckweed) was used in our studies to assess the aquatictoxicity of a group of commercially-used metal complexdyes and their iron (Fe) analogs.''^ Under ideal nutrientconditions, duckweed plants can double their mass in lessthan two days. ^ In solutions containing high levelsof nu-trients, duckweed is known to be resistant to environ-mental pollutants, including metal ions. While duckweedappears to have the ability to adapt to sub-lethal concen-trations of pollutants over a period of time, it has beenshown to be more sensitive to metal toxicity than otheraquatic species."

B A S I C R E S E A R C H S T U D I E S

Structure-Property Relationships

Genotoxicity

Following the recognition that certain aromatic aminesused in azo dye synthesis caused bladder cancer, a widevariety of chemicals were evaluated in animal studies. Theresults showed that aromatic amines and azo compoundsof the type shown in Fig. 2 are carcinogenic.^ In this regard,it is generally agreed that the ultimate carcinogen arisesfrom the metabolic conversion of these compounds to elec-trophilic species (Fig. 3) that interact with electron-rich sitesin DNA to cause adverse effects on protein synthesis. It isalso clear that ring substituents that enhance hydrophobiccharacter increase carcinogenic potential. Consequently,adding a methyl group to 2-naphthyiamine (structure 2; R= H) or mefa-phenylenediamine (structure 4) enhancescarcinogenicity

As a follow up to the carcinogenicity data assessmentstudy, correlations between dye structure and carcinogenic-ity data were established.^-^^ In a companson of hydropho-bic azo dyes containing amino groups in the para- or ortho-position, it was found that the para-isomers were

DECEMBER 2004

(6) R - H, COCHj

Fig. 2, Examples of carcinogenic aromatic amines and azo compounds.

Eii/>me Ox. A -

OH

CongugationAr—

/OSO3©

R

Ar—N©

R

Fig. 3. Nitrenium ion formation from the metabolism of aromatic amines andrelated azo dyes.

Fig. 4. Proposed reaction pathways for structure type 9 nitrenium ion metabo-lites of azobenzene isomers.

Fig. 5. Representative carcinogenic (structure 12) and noncarcinogenic (struc-ture 13) water-soluble monoazo dyes.

carcinogenic while the ortrto-isomers were not. To account for theseresults, the chemistry in Fig. 4 was proposed. It is believed that thetwo types of isomers produce nitrenium ions (structure 9) that ei-ther interact with DNA (para-isomer) or undergo intramolecular cy-clization (oz-t/io-isomer) to produce adducts (structure 10) andbenzotriazoles (structure 11) respectively

A correlation of carcinogenicity data with the structures of hy-drophilic azo dyes was also presented. ^ Water-soluble dyes of struc-ture type 12 were carcinogenic, while structure type 13 dyes (Fig.5) were not. In this case, the nature of the reductive-cleavage prod-

WWW.AATCC.ORG

17

Fig, 6. Representative azo (structures 14-15) and formazan (structures16) dye ligands used in metallized dye studies.

ments has led tothe search for al-ternatives to Crions for produc-ing lightfast andwashfast aciddyes. In this re-gard, metals suchas Al and Fe havereceived signifi-cant attention,with azo andformazan struc-tures (Fig, 6) serv-ing as the mainbase colorants

Fig. 7. Examples of nongenotoxic sulfonated azo dye intermediates.

HiN

Fig. 8. Examples of nongenotoxic hydrophobic azo dye intermediates aris-ing from design studies.

clear from studiespertaining to theaquatic toxicity ofthe resultant dyesthat the Fe com-plexes are gener-ally less toxic thanthe Cr counter-parts, however,the chemicalmakeup of theligand also playsa role in the ob-served toxicity.

ucts was the determining factor in theobserved carcinogenicity data. Whereasdye structure 12 produces a iipophilicamine, dye structure 13 does not, mai<-ing aromatic amine genotoxicity an im-portant consideration in azo dye design.

Aquatic Toxicity

The need to address toxicity concernsassociated with dyes containing metalsdesignated as priority pollutants was asignificant research area.'" This designa-tion includes Cr " , which was commonlyused by commerciai dyers in the post-metallization of mordant dyes to achievehigh fastness properties on wool,'^ Mor-dant dyes have since been replaced inpart by pre-metallized dyes for wool andnylon, placing the metallization process inthe hands of dye manufacturers. How-ever, the uncertainty surrounding the fateof metallized dyes in aquatic environ-

Molecular Design

Genotoxicity

The focus of studies in this area was thedevelopment of approaches to the de-sign of viable alternatives to azo dyes andintermediates determined to begenotoxic. Design considerations emerg-ing from the study of structure-propertyrelationships were implemented. Themost important pertain to the propertiesof the raw materials employed in dye syn-thesis and the reductive cleavage prod-ucts of the parent dyes. With regard toboth points, it is clear that sulfonatedamines such as structures 17-19 (Fig. 7)are nongenotoxic, unlike their sulfonicacid free counterparts. While the designof azo dyes producing sulfon.npH reduc-tive-cleavage products 1approach to addressing c.cerns, it is not universoil

the one hand, increasing the number ofsulfonic acid groups (-SO.Hj has a delete-rious effect on washfastness, and on theother hand, water-soluble dyes are unsuit-able for the coloration of hydrophobic fi-bers. With this in mind, approaches to thedesign of nongenotoxic hydrophobic dyeintermediates were undertaken.

Interesting approaches to developingnongenotoxic hydrophobic azo dye pre-cursors have involved the introduaion ofheteroatoms, bulky alkyl or alkoxygroups, or bridging groups into the struc-tures of aromatic amines.^''^^ This led tonongenotoxic dye intermediates such asstructures 20-24 (Fig. 8), While the rea-son for the removal of genotoxicity, es-pecially mutagenicity is unclear in eachcase, it is believed that bulky groupsplaced ortho to the amino groups inter-fere with nitrenium ion formation. Di-amine structures 21, 23, and 24 havebeen used to make nonmutagenic red toblack azo dyes (e.g., structures 25 and26; Fig. 9) for cotton, while amine struc-ture 20 and diamine strurture 22 havebeen used as dye precursors for non-tex-tile applications.^^'^^ Diamine structure21 and related analogs have also beenused as alternatives to benzidine and itshomologs; e.g., dichlorobenzidine, in or-ganic pigment synthesis (e.g., structures

18

AATCC REVIEW

More recently the mutagenic proper-ties of diaminophenylene homologs(struaures 29, Fig. 10) were assessed. °While all diamines except diamino-quarterphenyl (DAQP) were mutagenic inTA98 with S9 enzyme activation, all fourdiamines were negative in TA100 with orwithout S9 activation. It is also clear fromFig. 10 that the mutagenicity of these di-amines increased in the order diamino-terphenyl (DATR n=3) > diaminobiphenyl(DABP, n=2) > diaminophenylene (DAP,x=1) > DAQP, at 0-600 ig dose levels. At700 [xg. the mutagenicity of DABP andDATP was essentially the same. S9 acti-vation was not required for DATP to ex-hibit mutagenicity in TA98, nnaking thisdiamine a direct-acting mutagen viabase-pair substitution.

Following the synthesis and testing ofCongo Red homologs (struct, -es 30, Fig.11), mutagenicity data, prodLiced using

2QD4

oineymedal

Prival preincubation assay, for CongoRed (x=2) and the two phenylene ho-mologs (x=3, 4) at low (0-60 ng) andhigh (0-1,000 yig) dose ranges werestudied. At low doses, only Congo Redwas mutagenic in both strains, althoughits activity was most noticeable at the 50ng dose level. In TA98, the DATP-basedCongo Red homolog was more mu-tagenic than the parent dye, whiie theDAQP-based homolog wasnonmutagenic. Beyond the 50 ig doselevel, the mutagenicity of Congo Red in-creased appreciably beyond that of theDATP homolog, while the DAQP ho-

NH2 on M i l

Fig. 9. Examples of nongenotoxic azo dyes (structures 25-26) and pig-ments (structures 27-28) arising from molecular design studies.

moiog remained nonmutagenic in bothstrains (Fig, 11).

Aquatic Toxicity

Except for dye structure 32, theunmetallized forms of each azo dye hadEC50 > 1,000 mg/L. This low level of tox-icity is probably due to the low watersolubility of dyes structures 14, 15, and31. The Fe complexes of structures 14, 15(R=NHAc), and 31 had lower aquatic tox-icity than the corresponding commercialCr and Co complexes. On the otherhand, there was no apparent benefit ans-ing from substituting Fe for Co in metal-

lizing dye struc-ture 15 (R=H). Inthese studies, anEC5o> 300 mg/Lwas regarded asgood and EC5o>1,000 mg/L asexcellent.

An exampleof the data gen-erated is pro-vided in Fig, 12.The graph showsthe frond count(number of"leaves" on theduckweed plant)at the end ofthree, six, andseven days asthe dye concen-tration was var-ied. Fresh dyesolutions were

29

Dose-Response curve

50

introduced on days three and six to simu-late the periodic release of effluents froma manufacturing plant. It is clear from thegraph that frond counts above day onelevels were produced at the end of dayseven indicating plant growth. Althoughnot shown in the graph, aquatic toxicitytesting of ozone-decolorized dye solu-tions showed no change in the patternobserved with colored solutions, in thatthe Fe complexes were generally less toxicthan their commercial prototypes follow-ing ozone treatment.

In the case of formazan dyes, theunmetallized dyes had EC50 > 300 mg/Land only one Fe-complexed dye gave anunacceptably low EC50. This dye, struc-ture 16whereZ = SO2NH2andY = H, isthe only dye in the group that lacks asubstituent in the para-position of thediazo component. The presence of a -Cl,-NO2, or -SO2NH2 group in the para-po-sition led to a significant increase in EC50.In contrast to its adverse effect on mu-tagenicity, the -NO2 group did not havea similar effect on aquatic toxicity. In faa,the 1:2 Fe-complex having Z = SO2NH2and Y = NO2 had EC50 > 400 mg/L be-fore ozone treatment and > 200 mg/Lfollowing ozone treatment. Whileozone-decolorizing experiments did notinvolve dye concentrations above 200mg/L, our calculations projected an EC50> 600 mg/L for this N02-substituted dye.Interestingly, our studies revealed a cor-relation between the pH of azo dye so-lutions following ozone treatment andaquatic toxicity, with those solutions pro-ducing a pH below 3.3 exhibiting higher

aquatic toxicity (EC50 < 200mg/L). A similar correlationbetween the aquatic toxicityof the decolorized formazandye solutions and pH wasnot observed, even thoughthe pH dropped as low as3.0 in one case.

•x=1

•)t=2

•x=3

•x=4

200 400 600

Dose(Ljg)fTA98.+S9)

800

Fig, 10. Structures and mutagenicity data for diaminophenyiene homologs used in these studies.

DECEMBER 2004

T H E R O A D A H E A D

In the absence of a crystalball to help point the wayfonA/ard in the field of colorchemistry, one can simplynote and take advantage ofa few clues provided in the

WWW.AATCC.ORG

19

oineymedaS

Nil

SO3N11

Dose-Response curve

180016001400

2 1200ra 1000I 8001? 600

.x=4

200 400 600 aoo 1000 1200

Dose (ug) (TA98, +S9)

Dose-Response curve

1400

x=3

x=4

200 400 600 800

Dose(ug)frA100, +S9)

1000 1200

Fig. 11, Structures, TA98 data, and TA 100 data for Congo Red (x=2)and two homologs, using the Prival test.

Me

Oil

O2N

SO2NH231

SOjNa32

9 0 mgEJIOmg• 50nig• 100mgD 200 mg• 300 mg• 1000 mg

Fig, 12. Fe-complex ligand structures 31 and 32 and aquatic toxicity data forFe-complex structure 15 (X = NHAc) before ozone treatment.

form of nationaland internationalregulatory policies.

First, it is clearthat regulations,such as the Ger-man Ordinancethat bans the saleof goods contain-ing dyes derivedfronn a specificgroup of genotoxicaromatic amines,are here to stay.^^Consequently, at-tention must begiven to the selec-tion and safe han-dling of intermedi-ates used in dyemanufacturing andto the nature ofpotential metabo-lites from enzymeinteractions withsynthetic dyes. It isevident that theentire global com-munity will soon bewithin reach ofsuch policies.

Secondly, re-strictions on thenumber and quanti-

ties of chemi-cais permittedin R&D labora-tories requirestrategies thatminimize theimpact of re-duced chemi-cal inventories.The answerlies in the judi-cious use ofcom p u t e r -aided design( m o l e c u larm o d e l i n g )methods thatpermit theprediction oft e c h n i c a lproperties of

contemplated dye structures, thus re-ducing the number o^ ornpounds re-quiring synthesis before a viable productis obtained.^^"^^

The search for metal-free lightfast dyesmust continue, as those based on prior-ity pollutants will eventually fade into thesunset. "* Finally, we must find a way toattract the brightest available studentsback to color chemistry and train them totackle the challenging opportunities weface. Without them, the suggestion thatwe are part of a sunset industry will be-come a reality faster than we would liketo think possible. It would be a tragedyfor this to happen in an environment suchas ours, where leading-edge technologiesin the hands of talented researchers couldhelp to open new frontiers in dye chem-istry and applications.

Acknowledgments

This work was supported through fund-ing from the National Textile Center, Ciba-Geigy, Clariant, Dainippon Ink, and IBM.The contribution ofthe following gradu-ate students and research associates isacknowledged with pleasure: Jin-SeokBae, Heather A. Bardole, NatachaBerthelon, Julie P. Clemmons, Russell D.Cox, Laura C. Edwards, James F. Esancy,Michelle K. Esancy, Ahmed El-Shafie,Mohamed El-Zayatie, Zhimin Hao, DavidHinks, Whei-Neen Hsu, Sung-Dong Kim,Jason Lye, Mary E. Mason, Stanley A.Mclntosh, Sachin B. Nayar, MonthonNakpathom, Christopher A. Odilora,James C. Posey Jr., Abraham Reife,Stephen D. Shaw, Potjanart Suwanruji,Bonita L. Richardson,'Jolanta Sokolowska,Malgorzata Szymczyk, David 0.Ukponmwan, Jinlong Wang, and CynthiaS. Wiliiard. Without their efforts, this pa-per and the recognition associated with itwould not have been possible. A key col-laboration with Larry D. Claxton of theU.S. Environmental Protection Agencywas also invaluable.

References1. Weisburger, E., Cancer—The Outlaw Cell.

Chap. 6, edited by R. E. LaFond, AmericanChemical Society, 1978, pp73-86.

2. lARC Monographs on the E\ 'uationoftheCarcinogenic Risk ofChemn: - tQ f^/jgp vol,8, International Agency for ?searc'h on

20

AATCC REVIEW DECEMBER 2004

oineymedal

I Cancer, Lyon, France, 1975. ppl-357.

3. lARC Monographs on the Evaluation of theCarcinogenic Risk of Chemicals to Man.Vol. 16, International Agency for Researchon Cancer, Lyon, France, 1978, ppi-377.

4. lARC Monographs on the Evaluation of theCarcinogenic Risk of Chemicals to Man.Vol. 27, International Agency for Researchon Cancer, Lyon, France, 1982, pp38-188.

5. Longstaff, E., Dyes and Pigments, Vol, 4,No. 4, 1983, pp243-304.

6. Maron, D. and B. N. Ames, Mutation Re-search. Vol. 113, No. 3-4,1983, ppl 73-215.

7. Ctaxton, L. D,, et al.. Mutation Research.Vol. 189, No. 2, 1987, pp83-91,

8. Prival, M. J., etal.. Mutation Research. Vol.136, No. 1, 1984, pp33-47.

9. Annual Book of ASTM Standards. Vol. ,Philadelphia, Pa., 1993a.

10. Pederal Register, Vol. 50, No. 188, 1985,pp39,331-39,333.

11. Wang, W., Environmental Research. Vol,52, No, 1, 1990, pp7-22.

12. Skillicorn, P., W. Spira, and W, Journey,Duckweed Aquaculture: A New AquaticEarming System for Developing Countries,TheV^orld Bank, Washington, D.C., 1993.

13. Gregory, P., Dyes and P/gments, Vol. 7, No.1, 1986, pp45-56.

14. NIOSH Pocket Guide to Chemical Hazards.NIOSH Publication No. 90-117, 1990.

15. Welham, A. C. Journal of the Society ofDyers and Colourists. Vol. 102, No. 4,1986,pp126-131.

16. Czajkowski, W. and M. Szymczyk, Dyesand Pigments. Vol. 37, No, 3, May 1998,ppl 97-204.

17. Bardole, H. A., H. S. Freeman, and A. Reife,Textile Research Journal. Vol. 68, No, 2,February 1998, pp141-149.

18. Freeman, H. S., etal.. Textile Chemist andColorist. Vol. 27, No. 2, February 1995,PP13-16.

19. Reife, A., E. Weber, and H. S. Freeman,Chemtech. Vol, 27, No. 10, October 1997,pp17-25,

20. Lehmann, U.and M. Frick, German PatentNo. 19.546.600. 1996.

21. Aniiker, R., Review of Progress in Colora-tion. Wol. 8, 1977, PP60-72.

22. Freeman, H. S,, etal., Chemtech.Mol. 21,No. 7, July 1991, pp438-445.

23. Shahin, M, M., et al.. Mutation Research.Vol. 79, No. 4, 1980, pp289-306.

24. Esancy, J. F., H. S. Freeman, and L. D.Claxton, Mutation Research. Vol. 238, No,1, 1990, pp1-22.

25. Delia Porta, G. and T. A, Dragani, CancerLetters. Vol. 14, No, 3, December 1981,pp329-36

26. Bae, J-S, and H. S. Freeman, AATCC Review.Vol. 1, No. 9, September 2001, pp67-70.

27. Bae, J-S. and H. S. Freeman, Eibers and

Polymers. Vol. 3, No. 4, December 2002,ppl 40-146.

28. Hinks, D.. et a\.. Dyes and Pigments. Vol.48, No. 1, January 2001, pp7-14.

29. Sokolowska-Gajda, J., D. Hinks, and H. S.Freeman, Dyes and Pigments. Vol. 48, No.1, January 2001, pp15-28.

30. Wang, J., H. S. Freeman, and L D. Claxton,Proceedings. Colorchem '04 Conference,2004.

31. ETAD Information Notice No, 6, TextileChemists and Colorists. Vol. 28, No, 4,Apnl 1996, ppl 1-13.

32. Lye, J., et al,. Proceedings. Colour Science'98 Conference, 1999, pp9-25,

33. Lye, J., D. Hinks, and H. S, Freeman, Com-putational Chemistry and Chemical Engi-neering, edited by G. Cisneros, et al,.World Scientific Publications, Singapore,1997.

34. Hutchings, M. G.. Dyesand Pigments, Vol.29, No. 2, 1995, pp95-101,

35. Lye, J., etal,. Textile Research Journal. Vol.69, No. 8, August 1999, pp583-590.

Address

Harold S. Freeman, Campus Box 8301,North Carolina State University, Raleigh,N.C, 27695; telephone 919-515-6552;fax 919-515-3057; email haroldjreeman©ncsu.edu.

, , scalefor color change

This internationally acceptedscale is used to evaluate thedegree of shade change due

to chemical and physical changesto whichfabrics are subjected. Order #8359

VISA. MASTERCARD OR AMERICAN EXPRESS ACCEPTED

P.O. Box 12215,Research Triangle Park,

N.C.27709Tel: 919-549-8141Fax: 919-549-8933

e-mail:orders@aatcc.org

DECEMBER 2GD4 WVVW.AATCC.ORG