NOD :,!\IM~' .-(0 N f J011111V Ds.~\ j0 S)jI,),jj!j,3tIo.) jo o(p .(Oj s,)!.IojI,.rCCjL:I J;*.)J,)j;!,,, ali) LIj S.D.1111M.M.1d U011LI -GijJVlj -iql jo tiosmuduio3 L, ~L!; JOCICLI ZJU )d!1 01 alitimi, m -j! pi'noqj o I oamposla ppil slu'l u! ,~!Ilo- :ol-aul m,", Zq C)FIC Inq SlIcumoo ot!-- S'l Sluo -,ou Ino pail-tw.) q;mvosaj oisec, ei..,. ;3 IJUIJACCI~)J OUJ SOJ~)IdUI03 ~[.TOM S11.11 jo tiolluallcl-nd )Ell S~-r , (oz.-t, d) Ul (:)SAO OUds"JI-03 w[OLLIS JO UO!1-1, -110110U.11 9111 JOJ P--,CWlt0A;),l) 1.3a.13T030.1d, --;i1 XCI I)OUT111C, 0 SlIoTj3l,-.jj 110 S,~UG-.3eajq A 'ILI JQ II!AXlI:)j(I0(j U11IS JqIlOEU JO ILU!j ZRIJ ;o eq Q G 'XIIAllOr IE,,()l oill ol uo!jrrj!.z-,ur)o -).%7jvj:m zncm-,570 01 U0Tj3eJj zmljou spunodwoo polujol j0 ~Llll 0~,10Ug JO S-j1:,)nj!)SU03 arp wo-TJ ofil-19oup.1co -)"n ~-iv.T,Clos 0l Fo.Tlp)jO.[d J.) 11mat )q) u! ~.-ulj S;q-, u! uvw-aq jo ;3~Junj .1 ja C)UTZ)AU-71 Suz.; s.mov.-r.- %::is a-,nv,,,i p -).I Ow .10 -)!;1 jv~!! P-,UIIIS7~ SEA\ Ii ',ou!"31") m.-)3 I: c -1,).z SlIcanixidtili Pud ,1-.-zmu~ qj c1 )M) 3(l S-LU Z2jOLUS J,) JO I)L;;; j3rj1j0--) JJ3.11p ,30TJ-j Z)---joUjS -,,,I O.k% tumnaill!c1a oql S- j p", Ul VIO U,',)Cl Suq 0JUSU013,U03 -J%OUIS ajj0.lU:,')I0 JO UOTjUL'OTj,)IMj oq.L 3-11:11 )'JC,1!1!3 3110.1 ZZ'13 Ul J),)lj!jU,')j)l ij.,3lLEI-)Ejz) ,q c-1 (auoju pals,;j uotIm ."-)Ilpo.t(l smuclsqns) upis j-.-;itoTxi ol su~'!)uupauo S.!uuiiad atij, - sanowri u!~Is aoripoad pinoo qzmIm aiEsuaptioa wiclus u! juasoad zoouv)qcjn!, Sluo oti-, aiam 2-punodwoz) 3swil jr,-.-, vaildmisse 3141 u0 'Slupolmu ajolus .1atflo Zpql LU(J.'j SUMV -,UOIJC)-e.Tr 01.21119 0161 SlJ;)-,1j!jSU03 PUU I',-IVCI 01[l 0JUJILICIOU03 ILIC)Tqm sampacIoad uoilcuoil3e.~j Juw'isap o1 sauomtogul atil Ilu uj 'uo.;*.0-1 t!,).-,q suLl siGrildw.) uTem aqj snt(j, 3 TI3A00.T,1j*q-.%j ail) auln3ilaud 11T sonfoluuIc OT1,3SO0.1al~,Iq apill pa-B .1 od azil s(filo.1, pz)jupm 0.%\l 011.11 .it) ~!ik:jqo of .1 ,;.0) (If ',moloutoA .11:111j!)Ard tio 1:)I: oll.mloplic'a tit JO;,.')*,I.ll:!,~.,.',';;"q,.,.-..(.11,1 u.).no, Ili'j t3--: -wollum.mr.) Jo .11 W li*,.l: 0.; IcInop )MM SI JAO'kt; 'S"Ill (IJIS; I: fluq) alolu 51 Ln U!.J.:; Cll,)Iu 110 10 I.MjI,; ;;L(,, ,uqj S)!pqrqoad atil iq palimifidwoo uoaq SUI-1 Z;MJ0j 0ATJUjjjUI:Il)) UF aju~ajopzincI wjotug z)loq.NN )ql jo jl;z)l,"ojolq 0111 o1 suollac-Ij .1ulnoll.;VC1 jo UrjllI--cj!.(j(IO3 n(lI3'lULIi3.It;3 3AIJUID-1 JCJ o(I'L 666L joqWGAON 6 eiqwnloo 14SII!JI3 10 GDUIAOJd jol luownoop oo.LVS net-oll by it 11.1,; brcil necosimi-y to 1c:0 11w fractit'lls mA ()Illv i1i'lividird1v IwL in At Ow appliu;ltioll ul-Weibli'll oi~.A ril'tition t"It-tors fm- tull)('111. produotioll LroahncnL hY frnctim:s to ni'I'let.; lov )ms givvil tm'.11'- ilu;i "JO ild') the :iclkm (,I* lhr~;(- maWrials. A Y llseful s'orit-s of tcsts, li:ts hrcii caj-ricd ouL in so!:w 1:11-lui-:1toric., WIWI,' I'llicc, have been givun a sjll;~,Je initiating d(~!,L! of in acti\,(b I'ArIl such as dini(Ah 'v t bctizanthracune (W-1-11,A) before treati-ticiA with fractimis, in Order to study the proillolin" effects Of OIL: material. The methods used in aAnnilAs to concentrnte the carcinogenic constittietiLs of siiio!;c condensate have covered a wido spectrum of tcchniqucs 1)uL fall broadly into three catc,,orics: (a) the separation of snickc condensate into acid, phenolic, basic and nelitrat fraction. The carty at Ifarrogn',U -lid all the rose,--rch in the United Statcs (Stcdmaii 0 at. , Bock, \%'%-nder and lloffil-lann) was carried out using neutrat fr.,cLion; (Appendix 1) (b) the separaLion of smoke condensate into a number of sub-fractions by counter current disLribUti0l) I)Ot\\'L:Cll I Silgle Solvent pntir. .,itc and thc SEITA group in Fr4nce used this approach: Ilarrog (C) the separaticn of smoke condensatc into sub-fractions by a successive seric-s of simple bilrtc-ral scpari--tions bet%%-C'.cn a number Of differel'L solven', pairs; (App,-ndices III aijd IV). D,,~Laits of the ti-,eLhods of frac-tionation used by the different research groups are given in the Appendices and the main C011CILISiOnS Of thiS work are sunimarised below. Sepiration of niouse skin c.-ircinogens into a single frnctioll The emerging fact. that the constitticnts of smokc condensate which contribute to the prodtict;on of skin tinnours may affect only single or different collibin- ations of stages in the carcinogenic process ninke it not surprising that only a D few fractionition methods have been succussful in concentriting Substanti-n-Ity ail the biological activity into a single fraction. Only Lwo research groups Olarrotpt,c and )Ianibur-) using ipproach (e) above were successful in achievilllr an efficietiL system of concentration. FOllr CMUZICti-C-11. 1)"OCC.~;FVS V.'i'l'(' StUdWd 11 11:- IT0!`;!'.(,: (1) Separahon or water insoluble fractimi (C). ON ON BATCo document for Province of BritiSh Columbia 9 November 1999 (2) rish'ibUtion b~!I.ween 11)(111mlot (F) alld ((.'). (3) C3'C toll c..,:-i 1) (~ ((*,) alirl sulphoxi0e (L). (4) c),cl('IJL':I:,_M0 (QI) and caffcine/formic -LUid (MP). The efficiencY of these separations (Fig, 13) wns- tested in two sequclices Iltilising th),cc of Lbe four sepni-ations in each scrics, of tests. (a) A C ) G ) L(G) The final fraction L(G) being by %veight: of the original. smoke condc!lIsntc! and containing 80','o of the acti%..-'.Ly of WSC. TI-is material is Fnalo"ous to the Gcrinan fraction V. (b) A C G ---~ (R 4 P) G Tho final fi-iction (R+P)G 31' by %veight of the ori inal smoke , -L 11 gl condensatc and containing 77.5% of the activity of WSC. From thcse results it is reasonable Lo conclude that fracLion (FqP)LG prepared using ffl fGur separations: _ A ~' C o G L (G) (R+P)LG NVOUld L-Miain over 7V,'. of the activitv in 2~*L' of the weight of the orilinp! condensatc. ExperinicoLs Nvere also carried out to show thal, tiiese fractionation PrOCedUres are valid foz the concentration of Oic majority of mouse sk-in paintinq carcino,cns into sing1c, fractions irrespocLive of the type of condensate examined. The same tuinorij;enic ra',ios were cMqincd for smoke ccudensate, fraction G (257.) and fraction (R~P) w/w) prepared from nitrate-treated flue-cured ci,,areUes (lo%v activity) and small cignrs (high activity) as compared with,condensate and fractiotis obtained frorn standard flue-curod cigarettes. This demonstration of the validity of the procedures for the concenti-ation of the majoriLy of mouse shin carcinogons of WSC in P+ P)G was an important, part of the basis for project PSCI, in which attempts were made to provide direct evidence of the ability of XVSC to produce both mouse skin turnours and also lung tumou:-s in rats. Thus 'alLhough it is not possible to prodLICO squarnious cell carcinarna in rats by intratrachcat injection of WSC, given Lhe right con- ditions a hi-li proportion of animals will develop lung tumours by twice weekly intubation over 9 mondis with (R+P)G from standard flue-cured cigarettes. Also, this tun,, tuniour production appeared to follow a dose- response. The German tearn using two similar separations (Fig 11): M "trilwors 1""ethnnot (.Il~ and C~) Un (2) Distribution bcL%vccn cycloboxinc (IV) and nitromethane (V). Uli C7 \ CN BATCo document for Province of British Columbia 9 November 1999 prodilucd a fraclion (V) 6.5',,: by wciiht of C011(j(.11S;1L,2 ;:11,1 7&r, or thc ,cLiviLy of \VSC. The of differen', Of CM'6:10,`~!11 fi:0111 Wh(-,!V 51IM!,V C01101,11S-ate In the early years of smoking ind licalth rescarch, R vms thought, that the unsobstibAcd polycyclic aromatic hydrocarbons of the benz(,1)pyrcnc t~ype were responsible for nearly all the activity of whole sniole cond(InsaLe. lio%%,cvcr, the early c\pcriniunts soon showed that concentration of all the polycyctic aromatic ,,[c fraction did t,ct necessarily concentrate all the hydrocarbons into a sing activity into that fraction. Either no!-P.7'~rH carcinogeric mztLeriil was !ost in the separation process used (adsorption chromatography procedures used by all groups (Aj)!)cndLx I) and counter currenl extraction methods tcsod by the French M)d at Harrol',PtLu (!,qipendix 11)) or appeared in a fraction o,-her than the otie %vhicli contaired ail tho PArH. The first indication that mouse skin carcinogens other than unsubstituted PArI1 mighL be responsiMe'for a szibstantilal part of smoke condensate tui-nor-onic activity I)%, prinlary Carcinogens was iven after the testin, of the fractions obtal-.ic.d 1-),, 11,,o sub-fracti-,na~ion of ncitrial fraction by colu-nn (App~,r)-Cli.,: Y) !~y Si,~dinnn c' al. (Fig 2). This v.-as inter confiri-ned, by thC SUb- sequout %vorl: of all groups working in the,fietd, but perhaps most convinci. in the co-,,,n1.cr currcrit cx!r,-1~-1iO!1 Of fl'a-,LiOll frO111 11--'Utral frac,:!;J~l (Appendix I Fig 2, Experiment 1.1. 3.11) and cyclohexane fraction G (AI-1-pendixIl Fig 7, betwcon cyclohexanc ind aqucous mcthanoL carried out at Harrogate. The result of these tests demonstrated that the primary mouse skin activitv was dL.,e to substances which felt into three r 4 _.:0L1pS; the substances VJiL111. the same portiLion coofficiciA range 0~ A ) as the PArI1 giving the highest 'IcLiviiv, sul)stanccs of tov.'er K.7., vallic cau-7ing less activily and matelial of higher K.:~ vat!le causing, stilt lower activity. Thu French counter cui-,-cnL schcrie separated two active fractions - fraction ill substances of the same I~A value range as the unsubstituted PArH, and fraction IN7 containing PAM- free material (A hi,,her K value. The most-recent methods of separation at Harrogate (Appendix IV) and in Gerniany (Appendix III) also resulted in the production of an active fraction conLainin, alt Z the PArIl (fraction P(G) at Harrogate and fi-action IX at Hamburg) and a larger 1"ArIl-frec fraction (IR(G) and VIII respectively). These results are consistent with the hN .pothesis that the primary carcinogens of smoko irc, ro!ycyclic aromatic ill. Maill, 10 t dit--ivati% C;i nii,; lj,:I~:yoLl~ (`,ic analogues. Fig 1-1 shows the distribution of activity in fractions obtainod by the difforciii procedures. BATCo document for Province of British Columbia 9 November 1999 5- Allhou:-,h 0-1- viophasis or fractiowltion h;is bc('11 ()1j t1;(. c()JIcE,j~ carcilr':-,'~Ils of cond"nsatc, coll.--ider"ib1c, (-vid(-')(-c 11:1!; 1"-('11 WhiCh 111,'1t nintcri2!s of differc:M Iqws ore prosci-it and in OAF; I-L-SI)OCI most. worl.(-r.z: hm-C conuentr:itc-d oil the iden!ificntiun of fraCtions CU!Wdllill," L'IT11011- I)I'011h1fill"" SUlJ3i,-111CL'S. The very carly work on the preparation of neutral fraction donlonst,ratcd thiA a si-n-jifiennt proportion of !hc pronioLors could bc concen t rated into the %vcnlzl\, acidic or plic!ziot fracLion. Stedman ct al. later showed promoting activity !n the aqueous methanot SOIL1131C material of neutral fraction. These results were consistent with tho Harrop-ote findiiigs which showed removal of promotors by aqueous nic-Lhanot in diL-crunt separatioi-.s into fraction-j F, H and G2a. Similarly the rrench team showed Hint the proniotors of 11"SC were nicst Soluble ir, the i-nr-~tli:~,,iol/.-icctojiiti-ilc, ph-lise of their solvent s'ystem and most activit\ wa.s found in material :so,,itcd in fraction G2a of the Harrogate counter current extr.-iction scp~tration. Thus on solubility grounds, W~' formic leid./w3ter v.-Ollid bC CNI)CCtOd to extract the p.,onicting niaterial of fraction G from c%-clohexane ol,ition into fraction R(G). Fig 13 givc-.s the fractions from thc different Separation r,1'0','C(1UrOS ~Vhich hav,~, been shown to contain promotin, matcrial by the. classical mouse shin Lust usinc:r, DMBA as an initiati ng agent or by analogy can be assunied to contain proinoLors. Rccc:it I y, P;NTL li~is applied :mifficinatical i-nodels assuniing a mu!ti--tage mech- anism for carchiogenesis to the results of Lestin~: fraction R and fraclion P boLn - and in rccombin,iLion. fie Showcd Ilint the I-,,zt mo6cl for at[ thc Mouc - - . :I dati. is one in which both fraction IR and fraction P stron.-ly affect one stage and frn(-tion P also affects ano!hor stago bu~ to a snialler degree. This model would be consistent wit,h the hypothcSis that the activity of fraction P was due to complete carcinogens (i.e. PA rl-l) wbereas fraction I? contained small amounts of Initiating subsl-unccs boostcd substanUally by co-carcinocrens. The rostills of testing fraction 'R and frncticii P from condcnsatcs produced from cigars, standard f1tic-cured cigarettes and nitrate treated flue-cured cip-arettes sti-est that Lh.3 differences of the activities of the ori,inat condensates are due Lo the prosenec of different proportions of the two types of carcinogen. Thus, It seems probabie that carcinogenic inateriat in fraction R is responsible for the hilicr acrivil.y of cigar smoke condensate as compared with condensate from standard ei.-arettes, whereas the reduction of activity of standard cigarette condonsate as a result of treatin,, the tobacco with nitrate is duc tar,ely to a reduction in fraction P although fraction IR rnigiit also be slightly reduced in activity as a result of thc treatment. Q:) Cr\ BATCo document for Province of British Columbia 9 November 1999 cowl III!; i,; Is (1) The mouse sl;h, carcinogcns of \N,.(3C arc i:, v.-a(cr, can ),C C.%:I-IcLCd from ;ollltioll it) a(ILICOUS by cy! can yk-tolleN.-Ilm. with 1 sulphw.'du, bc extracted from C ni!ronicllhanc- or a ciffuine/forinic icid soluLion. (2) A Scrics of bilatcral ex1ractions LFing the above s,,Ivcnt pairs result in the concentration of over 7C" of (he original activity of WSC into a sin-Ac fraction (R+P)LG representing 2% by weight. r -)f t)je original co:idcnsate. (3) These procedures are valid irruspcctive of the typ(. of condensate lised. (.1 The unsubstituted PArl! aic: not the only primary niouse skin carcino- ge:is present in wSC, Pis all groups have PrOdUCLd %Uivc iractioni vilich do riot contain these compounds. The activity of thc-sc so-caLlect free fractions m,,i%, be due to the I)rcscnce of alkyl df,rivativos of PArII (of7en more carcino,,enic Lhin the p,-,i-cnL compounds) or biologica!ty ac!;,,,c hoteropoiyeyelic arc.ina~ic compounuls. 5) Th-r- niou~;c shin ca-roino,ons present in WSC are not of a unirorni typc but may act as conipletc ca rcinogens, as initiators, promotors, or in~fbitors. On the basis of a mujLj!;La,-,c Iv.-pothosis f(,-,- carcinogenesis, the liarro,,ate %vork- surr"ests that fractionation can separate materials wh;ch affect different. stag-cs of the proccess and thaL diffcrenucs in actEvity of NN'SC are clue to the presence of differcnt proportions cf at least two typcs of carcinogen. ON 01 \ BATCo document for Province of BritiSh Columbia 9 November 1999 APPYNDII: THY OF NEUTBAL FRACTION 0YWHOLI. SMONi' CONDENSATE' h1 louse. skill bioassnys Carried out in tile VCI)' cal-ly 19GOr, shov"Cd that Lll(! tunicrigenic activity of whole sinoke coiidcn5:.-.te resides principally ill th neuti-al fl-ILCHO'"I reinaining after ren)ovat ofacids, phenols and bases (Fig 1, Pi-cparation of neutrat fraction) and this work established neutral fraction at tile Linic as the main material used for sub-fractionaLion in identification studies of the priniary carcino-enics of sn;oke. Some of the early work at Harrogate was c,-l-ried out oil tile sub-fractionation of neutral fraction ind the two Ai-ncrican groups ;n the field (Wynder et id. and Stedman et at.) have carried out most of their rescarch oil Inis material. Column absorption chromatography has been tl~c rnain tcehniquc used ill these separations, One of tile first fractionation e.,;periments (1.1. 3.2 Final Analyses F 1531) to be carried out Pt Mirrognte in-.-olved the ;4pplicatiori of ncuLral fraction ir potroleurn C-ther to an 111111)ina column fotlov.-cd by cluLion with petrolrum ether, benzene and methanol (Stage 2, Fig, 2). The combined peLroleun. Oiler and benzene cluates contained most of the lon- chain aliphatic hYdro- carbons and all the polycyclic al-oniatic hydrocarbons and the more po,f:r constitucii'Ls were renim-cc! 1b,\ the IlIctilanol. The hydrocarbon fraction (IIC) was sh""vil it-, two (1. 1.3.2 and 1.1.3.11) to retail! fl-C-1-n 0;-.0 11111f to two ti)irdS or tile. activity of neutral fractiun and the methanol notaiat illintis hydrocarbon fracti-i (NF - JIC) to account for the remaininf- one third of 030 2 ctivity. Counter curront extracUoii was used to separaw the hydrocarbon fraction into thi-ce- stib-frac!.-Ions - IJC2a, IICI and HC21), according to distribution between acueous methanot and cyclohexane (Stagre 3, i~a 2). Fraction IICI representing 0. 2570 by Nveight of WSC contained all the unsub- stitutod polycyclic aromaLic h%-drocnrbons and was shown to contain about ha!f the activitv of the hydrocarbon fraction. IIC2a ard IlC2b in recombin ation - the PAOI-froe fracHon - showed a siniilar acLivity to JICI, denionstrating that other primary carcinogens arc present in snnokc condensate than the Unsub- stituted PArIl. The results are however consistent with the working hypothesis as N-h elc rocyc: tics tend to be strongly absorbed on solid absorbants and might be expected to be cluted from alumina with methanol. Also they are more soluble in iqueous methanot and in the counter current regime any of those substancos eluted by petroleum ether/benzene might be expected to be prescnt in fraction HC2a. MeViylated and other alkylatcd derivatives of PAM (several of which have been den-tonsti-atcd in smoke) would be expected to be ciutcd with b--nzene from atumina and as their partition coefficients are higher than (lie unsubstituted PAM, these coinpounds vrould'bc expected to be present in HC21) Stedman ct nt. (rig 3) used silien scl as tho ahsorbnnt for* sopn rat ing netitral fy.ietion and "iRCr swl'i.!! Chlti()il v.-Hi, peLroleuiii L,LIIL'I' Q~) (PE), betizenc ill pcLrolcuni t-ther (BP), belizello (B), other (E) -ind methanol (M). Each fraction was partitioned bctween vitlicr O.V001IL-Vine and diniethvi ON CN (Z) CXD BATCO document for Province of BritiSh Columbia 9 November 1999 solpllo~Jdc or pctrolcuill 01-1- ,in([ ;iquooos m(Al-inot itilo 10 sub-fr:iction!~. hloor". pailllii);~ c\11"i-inicnis -;howed th:it. Ilic icLi% NY wa!; Outud with Oit- 1.1cirl.- (a)p"l-cil'! %M11 L'OlCr (]',I)) Mid OW Ot;),.'I- PAI'll VJIh (D), or rellwillcd oil Lh(,, siliei onLit Rtriplwd wilh lilethallot. It sllov,'ll '11.1o that ilic actl%iLy fultowed Lle I)A)'I[ ill HIC Sta,"c hiLo tli(- DAI;`0 (20 and 22) or ill the case of tile mdh:tnot llltU thC 11-101-L' ),~A',U' (fr.jction 2G) as iniglit. be eNpC!CLcd if HoL-C were present ill the ori,init neut,rat fraction. Hoffinann and Wynder (Fig 4) applying colrinn absorpLion chromatography to neutral fraction on silica get obLaincd eight. fracLions following serial cluLion wiLli N-licxane (frictions A,, A.? and 11), carbon tetrachloride (C), carbon tetraclitorido./benzenc (D), bclizene (E), etbyl acetate (F) and nieth3,1 alcohol. Mousc paintin, experiments showed LUMOrigenic activity Lo be present in only two fractions A and B - thc fractioiis conLaining tile PArH. No acLi%,i,y was 2 fourd, contrary to !'.,,c finding-3 of the other kvo groups, in Lho final frection (G), a result, which might be explained if some N--heLrocycli(,s are responsible for sonie of tile activity of t)cutrat fractioti ind %vere not elu,,cd in the fractions, as halogenated sok-enis such as carl.,on tetrachloride -are known to break tip tllcsc! sub-stanecs whe-i adsorbed on silica. A similar loss of activity was found in the sepir3tion of fraction L on silica in Stage 5 of the Harrugate fractionation schenic given in Fig 10. 11 offir.-ann 3nd Wyncler (Fig 5) uSed a inodificJ procedure using fluorisit Lu concenLrqte the. activity of fi-ictions A., and P. hAo a sjl)(',lc- This niateriat has b,-cn Llic main sLarting point in tile %,cry extensive studies carried out bv these auLhors to Identify [he primary carcinogens of smoke. FracLian B1 aniountim, to 0. 6,: of %VSC was the only sub-fraction froin the separations given in Fig 5 to be found active and was shovm to contain PAr14, N- and 0- hetcro aromatic compounds together with some of Lheir pyrolysis product.s. Uli 01 \ ON BATCo document for Province of British Columbia 9 November 1999 0 0 0 (D 0 (D 0 h 0 0 z 0 CD Cr (D O/Zq9- z9 ol A Il,hole Smoke Condensate 2 N H-11 I E, I i c r Bas~c Ncu"rit Fraction 8.5% + !-~Icnollc + A c c F ra cti on. s 2 N KOH Ethcr NF Ncutml Fraction + AciCtic Frac.'Ions A q 'Na 11 C o. Einer 3 Acidic Fraction Phcnolic rraction 3.5% 12 "0 Fl!:ure I Prciparation of ne-utra, frac'icn 0 0 CL 0 T I 0 < 5' 0 CD 0 % (A :r 0 0 z 0 CD Cr (D Stac-c I Basic Fraction 8.6% Sta"c 2 A x1rhole S-nr,,kc Conrycnsatc i " !-~ ~- - T- I Ac;nic Fznzction Phenolic Fraction 3. "1 .12 C. -o Pet.lEther Bcn7!c.,nc HC -Hvdr~carbon Fraction . . Aq. McUianol C3'ctol'-'cx'-Izlc ITCI rraction NA = 2. 5 to G. 1 0. 25r, Stage 3 (CCD 3) A /o ITC2a Fraction ITC 2b Fraction I~--I. 0 - -3- 0 - q. I-,,' cc: ]Fraction IV. 14. 5~:, i. 0 - oc 1 rraction' ill I Fraction 112 Fraction 1131 3-;0 0 . 6,7C, 1 < 0. 15 15 j..5 >1. 5-) 0 0 a 0 C (D T 0 < CD 0 0 0 Cr E, z 0 CD Cr CD Stage 2 Stage 3(CCD3) Fraction % G % WSC K range. A Friction , 7" G % WSC K range A 19 9 S 0 1 A, Whole S--noke Corder-sote 100% Water B Wnter 8olubic C Insoluble 'Fraction 52. 07b T---i -nction -15. 3176 Aq. MeCH Cyclohexane F Aq. "Methanol G Cvc1ohexane Friction 15.1' Friction 24.5'~, Solvents Aq. Mc',Ihanol Cyclohexane I I - I I F raction G2A Fraction G213 I Fr.1cLion GI I Fr3ction G2C I - Fraction G2D 10. S~.; 17.2% Q.. 8% 10.1% 2. 4 0 "IJ 2.48% 2.6417, 4.21% 9. 2G "C 0 - 0.19 0.8 - 2.5 - 13. 8---X- 13. 3 cr- OP - G. 1 6. 1 Frnction G,2a Fraction GI I Fraction'G2b I I 19.8% 8. 71rO 58. G% 4. 9 70 2.10/0 14.370 0 - 2.5 2. 7, - 6. 1 6.1 - cC- Fraction G2 rruction G2a G2b 19. 2% V%ISC Figure 7 Fractionation scho-me involving simple and cottriter current extraction and PtAhwell) 0 0 CL 0 CD _h 0 0 < (D 0 0 0 z 0 < 0 3 Cr 0 Izard et at. CCD! solvents Fmction SC) CCD2 solients Metha', ~6'1-acc tin i tri I c 1:1 Fraction I Fraction IT > 2, 2, 4 ' tri-ne,"hylpentane TAIP) Fraction III Fraction IV 4. 70,1,~ 14.570 F raction Wa Fraction I'Vi~Fra_-Eon Fil-c! 4.8'~ 6. 07, < Aq. Methanol Cyclohmane F ractior, 11 rraction H.. Frac- t 10-, 113 5.0% o . C- 70 Fraction Fraction 2A Fraction 2B Fraction I Fraction 2C Fraction 2D (-r WSC) 78% 3.5% 2. G'/o 4.2% 9. 37c Fraction Fraction 2a F r a cc-, i I Fraction 2a. ~17;sc) S1. 5f,; 2. Gfc' CCD3 solvents Aquvous Methanol 1:9 Cyclohexane Whitehead a,-0 7othwell Flzurc 8 Comparable fractionsobtalned in Lhe counter current extriction schemcs of Izard et al. (Fitg. 6) an' d Whitehead and RoLliv.-o-It (rig. 7). C 6 H99LSO I APPE'NDIX III THE FBACTIONATION Or WSC 13Y SEMPLE SOLVENT EXTRACTION ~Jost of tlic Nvork carried out in the fracLionation of condcnsate in Germany by Elincifliorst and his co-woriicrs and at Ilarro;,,--itc by T.R.C. has been by a ~'erics of bilaturaL scparaLions between various pairs, of organic solvents. Figs 9, 10 and 11 show the main stages in the separations used by each g---oup. Elmenliarst and Grimmer (Fig 10) distributcd NNISC bLrWCCII CYC"(0heXane and aqueous methanot and produced I cyclohexane fraction 111 (26% V,-Iw WSC) which contained all the PArIl and certainly ncarly all the acLivity of the original smoke condcnsate, aithough this fraction was not tested scparately. This material was put through SLage 2 of the separation and distribu'Lod betveen cyclohexane and niLroniethaiie to produce an active nitroinc-Lhanc fiaction V (6. 81~~ v.,/v., WSC). This friction (V) was sho%vii Lo produce ticarly the saine Pumber ot tumour bcaring animals (TBA) as the original, condcnsate slid the recombination of both the PArll-frec fractions (11 and IV) showed only a low activity. In the Lhird sta.- .,e of the German separation the nitromethane fraction (V) was separated by column chromatography on silica get to yield a PA01-rich "ra. -1-ion (%I) ellitCd WWI bC!1ZCUc acid cyclo!;,,exane and a PArl-11--frec material (VII) I strippcd froin the column with nictlianot. Fraction VI was considerabty less active than the nitromethanc fracLion (2.61" and 9.8% TBA) and alffiou-,ii fraCLion VII was not te-sted alone, it waS shown to contain some of L,',.c activity of fraction V as the recombination of all the PArli-free fractions after Stage 3 (11, IV and VII) gave a higher activi!y (10.31~ TBA and 14.817:, TBA) than the recombination after Stage 2 (i. e. rl and IV, 5.87o TBA and 7. 8TD TBA) at the two levels of treatment. The separation produces sonic loss of carcinogc.nic material, 'as a rc coin bi na,ion of all the fractions showed a lower actMLy at each of the levels tested (3.2. 9% TBA and 26. 3' TDA) as compared wiLh the originat condensate (21. 9c'-- TDA and 30. 6% TBA). The Harrogate separation (Fig 10), developed at the same time, used three vary similar separations to the German procedures together with two additional stages (I and 3) to produce'fraction N (1. 6% w/w WSC) into which all the PAr.1 of condensate were concentrated giving a material directly comparable with the German fraction VI (1. 1% w/iv WSC). Subsequent mouse skin testing showed that taking account of the different doses used-for treatment, these final fractions produced by both separation schemes had Lhc same activity but this activity was considerably less than the activity of the whole condensate. The separation of Elinenhorst and Grimmer using the distribution between cvclo- in Stn,,c 2 of tho ilarro-,itt. s~:Jiuljic (Fig 10) following the cxtracLiun of water soluble sul)m.-Ilices from WSC. Histor- QY1 lently, this separation into water soluble and woLor insoluble components was t_^j used ill early \...ork by J.K. Whit-I.Cad in a search for unsniLurated aliplizitic alld ON Y ON CZ) BATCo document for Province of British Columbia 9 November 1999 S t.LC(011L'.S ill tjl)acco smoke and )1:1.S YA'lillt'd I)CC:1LlSV HIC i;lSOtUblC- matel-kil, fraction C. ll;lj provcd a WsL ni;ilcriztl in various biolo~,'u;ij proic-cls it l1n)-rogaLe, - i.e. it is tile sinipic-A incLhod of rcn~ovingl the nicotine froin silloke ccndunsaLc. Stage 3 Of tl'-(' SL!l)al':l1j0;, SMV thO IT1,11OV:1.1 of Iolig, Chain nlipllltic compo'nents I)V 1(`d1lLti0!l with urca, le:tving a non-adcll!c- tive rnalci-iil (frarLion J) v.,hich in St:ige 4 %vas distrihL114-0 IJC~W('Cll C3_C:10hV%,-111C and dinieth This step w;~s awitogous 1.3 Stage 2 of thu yl sulplioxide (MISO). .,es of both separclUOIIS LlSing SiliCLL German procedure as were the final star- adsorption. All stages of the programme have been exhiustivcly icsted by mouse skin biuassay (Experiments 1.1.3.3 - 13) in which the fractions were applied both alone and in recombination. The water soluble constituents (fraction B) were shown to be inactive to mouse shin and 95(,~L Of the activity was recovered in fraction C. The cyc!ollexanc fraction (G) h:.s been tested mank, Iiiiies with tile recovery of 70 - 80~-,~ of the total acUvit.v of the Original WSC, which represents a much higglicr activiLy in terms of initiation. as fraction F was shown to be virtually inactive wherl tcsted alone but restored tile apparent lost activity in recombination wir) fraction G. Promotor activitY was also sio%vn to be present in fraction F in the classicat promotor c.\j-)Cl.in-.ert, 1. 1. 5. 2. Fractions 11 and K "vere S'lown to have lov: activiLics when tCEte6 alone, atihou,h Experiment 1.1.5.2 in which tile materizil was applied three tinies a %vlaCk following, in initiating dose of DMBA ind-caLcd a small pron-,oti-,,..- nctivit.,; in fraction 11. The initiating activity of fraction G was recovered in fractions J and L, or when fraction G %vas distributed directly between cycillohexanc and D,',.TS0 (i.e. Stage 3 in Fig 10 was omitted) to produce fractions K(G) and L(G), all tile activity of G %vas, recoverod in L(G). This IaLtor fraction was c-xact!y analogous to the nitromethane fraction (V) of the German separation. As stated abrve, the final stages of both separations using silica gel absorption succeeded in concentrating the unstbs!ituted PArH into fractions V1 and N respectively which both showed the same considerably lower activity than the parent fractions or the original NNW. The apparent loss of activity was only very parLiatly recovered in the non-PAM fractions (V11 and M). An alLernntive separation of the nitroinctliane fraction (V) has rccently been published by Dontcnwill ct .11. (P 440) in which the Gel-man group used Seplildex gei to remove tile biologically inactive polar compou ,nds in fraction 171 and to scparato tile renlaillin, material (VIT 'I in'.0 two ahnoct VIll, and a final V..%rIl-rich fv%uLiull IX. PAI-11-free f) Cr*1 CrIl BATCo document for Province of British Columbia 9 November 1999 Iii Fraction Vf conuiined nll ilic. polar compoundn includin~~ phonolic substances and nicotine alkaloids; renminin', in friction V. These materints 11-v v removed during the firsL stages of the llurro,,;aLe s(!)),iration in fractions P, ind F. Mouse skin tc-,;Ls slio\v(-d this fractio;i to b-2 inactive. Thp rcnininill" Imitcrial (fraction V11), a "purificd" nitroincthane fr;ic,ion, was not lested alone or in reconihination so it is not possible to state \v c h r a y oss of 11 L active niaLerlal occurred at this sLagc. The larger sub-fraction contsined only loa, molecular weight unsubstituted PA r1j, i.e. 'IPA r11 free" and all the long cliain constituents of fraction VII such as aliphatic bydrocarbons, alkyl PAM Lerjx:aes, fatty acid esters, aromatic hydru- carbons wilh reduced ring systerns. Mousc skin tests showed it to have a activit.y. The final fraction IX containing all the unsub-stituted PA rI1 of four or more rings and ropresentin.- 0.4- by weight of WSC was responsible for t\\,c th irds of the parcnt nitromethanc fraction and one h.,.If of the activity of the original condenc:Lt, A study of [lie activity of individual fracLions and in recombination did not sho'N' a significant loss due to the fractioll,-LLiOn mothod. (--4 BATCo document for Province of British Columbia 9 November 1999 Stage I Stage 2 Stage 3 Aq. MeOH II Aq. Methanol Fraction 76.3% Cyclolicyane IV Cyclohcxane FracLion 1S.9% Nitromethane V Nitromethane Friction 6. 6% Silica get cliromntc,,r,-Iply P-3 etution 1 Alothanol VII 11csidual Friction 5.8% Ist cluLion Benzene/Cyclohe-,are Tionze C/ Cvclohuxanc SUb-li-iction gure 9 Fractionation involving simple solvent extraction and sitica get ch F! romatography. (Elmenhorst and Grimmer) I Whole Smoke Condcnsatc 10 0(,~ Cyclohexane III - Cvclohexanc Fraction 25.970 PZ99~so 1 0 0 CL 0 0 (D 0 11 0 0 h 0 0 3 Cr E, (D Z 0 Cr CD A Whole Smoke Condensate 1007o W.1 to I- c Sta- I B Water Soluble C Water Insoluble Fraction 52.0c/(, Fraction 45.3% Aq. MeOH Cyclohexane Stage 2 F Aq. Methanol G Cvclohexane Fraction 15.1% Fraction 24.5% Adduction with urca in methnnol and extraction Star,e 3 H Urea adductOte J Non-adductable with benzene Fraction 15. 10/0 FmcLicn 17.9% Cyctohcxane/DAISO Stage 4 K Cyctohexane Soluble L DNTSO Soluble DMSO insoluble Fraction 5.3% Fraction 8.8% Silica Get absorption from benzene solution Stage .5 Benzene Extract of Sitica, GO M Methanol Eluate N Benzene EZuatc 3.0% 1. G Figure 10 Fractionation involving simple solvent extraction, urca adduction and silica get chromaLographir. P?Z99C - 0 1 (Whitehead and Rothwcit) 0 0 a 0 (D 3 rq. 0 0 0 0 F 3 Cr z 0 (D 1 Whole Smoke Condensate 100% Aq. MeOH I Stage 1 11 Aq. Methanol Fraction 77.2% Stage 2 Stage 3 Sta-e 4 Cyclohexanc I III Cvc[ohexane Fraction 24.9% Cyclohexane Nitromethane IV Cyclohexane V Nitromethane Fraction 18.4% Fraction 6.5% Sephadex Get Column V1 Methanol VII n-Hexane Fraction 2.9% Fraction 3.6% Second Sephadex Get Column VIII Anthracene ix PAM Fraction Fraction 3.3% 0 . 3 OCIO Figure 11 Fractionation Involving simple solvent extraction and Sephadex get chromatography. (Dontenwitt et at.) 5 ?Z99 APPENDIX IV THE SEPARA'r1ON OF CAFFEINE COMPLEXING SUBSTANCES FROM WSC Most of the fractionation procedures used to concentrate the mouse skin carcino- gens of WSC into a single fraction were based on the working hypothesis that their components fell into one or other of the Acmicat categories: the PArH and their alkyl substituted derivatives or their aromatic heterocyclic analogues. A common property of these substances is the formation of loose molecular complexes with caffeine and its analogues, and advantage has been taken of this property to develop a useful fractionation stage at Harrogate. Early work was based on an observation that PArli/purinc complexes will migrate under the influence of an electrical potential towards the cathode, and this electrical technique was used to separate the PAr1I from fraction 11" (Appendix III, Fig 10) into a fraction P(NT) which represents 0.25% w/w WSC. This sub-fracLion proved the only active one but the method was abandoned as it was shown that the solid support (silica get) tended to irreversibly adsorb sonic of the non-PArli material and some heterocyclic compounds were unstable. to silica got adsorption. 11ov.,ever, if the evidence provided by experiments showing that at least 70170 of radioactively labelled PArII added to WSC can be recovered in fraction P(N) is SOUnd, then it is litiely that the polycyciic aromatic hydrocarbons alone are responsible for not more than 10% of the mouse skin tuniorig,enicity of WSC. Later work- showed that virtually all of the many PArH and N-heterocyclic COMI-)OUnds (PHctC) tested could be easily dissolved in a solution of caffeine in 90 Z formic acid/watcr and could be extracted from organic solvent's with this solutio-n. It was also demonstrated that basic PIIctC can be removed from a mixture of both types of compound with 90% formic acid atone. Two fraction- aLion schemes for the separation of PArH and ba:sic PIictC from cyclohexane fraction G are shown in Fig 12. Scheme I separates all the caffeine complexing constituents into (R+P)G (4.5% w/w WSC) leavin'cr a PArIf and PlietC-free residual fraction Q(G) and the basic PHetC are extracted from (R+P)G into fraction R(G) leaving the PArI-I and other caffeine comptcxing constituents in fraction P(G). Scheme 2 is an alternative form of separation into the three fractions R(G), P(G) and Q(SG). The fractions from these separations have been tested in Experiments 1.1.3.20 - 25 and it was shown that (R+P)G and R(G) + P(SG) had the same activity as the parent material, fraction G, and therefore a high proportion of the activity of WSC. The residue Q showed no activity by itself but both Q(G) and Q(SG) when recom- bined had a clear antagonistic effect; the recombined fraction only having 5070 of the activity expected, on a simple sum relationship. Chemical analysis on the 01\ fraction Q used in the bioassays demonstrated traces of caffeine left from the CN separation procedure. Experiment 1.1.3.29 showed that caffeine deliberately added to fraction G and fraction (R+P)G had a similar inhibiting effect. Thus it Cl\ BATCo document for Province of BritiSh Columbia 9 November 1999 ii seems likely that this material (fraction Q), expected on theoretical grounds to be inactive, is in fact not tumori,,cnic. Confirmatory evidence is given in the fact that fraction P(SG) and fraction S(G) (equivalent to P(SG) + Q(SG) and pre- pared without caffeine) had the same activity. - The PArfl-rich fraction P was found to retain nearly half the activity of fracLion G, and fraction R(G) which could be expected to be active if it contained carcinogenic PlletC did in fact retain just under one fifth of the activity of G. A much more important result, however, was obtained on testing the recombined fractions R(G) and P(SG), when significant synergism was demonstrated. When compared with R(G) and P(S'-) alone the activity was 41% more than wculd have been expected on a simple addition basis. The relationship of the two fractions were further investigated in Experinients 1.1.3.24 and 1.1.3.25. In the first experiment, fraction R(G) and P(SG) when tested both alone and In recombination at a number of dose levels and in different proportions Pive results which clenrly confirmed the previous work which indicated that cnrcino.-cns of different types were present in the two fractions. An analysis of the results usin,, mathematical models based on a multistage hypothesis for carcinogenesis showed that P(SG) affects both of t%,.,o stages of the cancer process and R(G) only one. These results are consistent with the hypothesis that the activity of P(SG) is largely due to PArII which would affect all stages of the careinocrenic process and that R(G) contains predominantly promotors acting on the later sta""cs together with small amounts of initiator material (possibly PlIetC) jusL sufficient to trig,er off the first phase. However, in experiments where dosing is con- Linuous, statistical analysis will not. determine the temporal order of the stacres affected by particular fractions and thus on the evidence of Experiment 1.1.3.24 alone it is not possible to distinguish whether the activity of R(G) is due pre- dominantly to carcinogens of the "initiatim," or "promoting" type. Thus, it is important Lo consider other factors in order to interpret the results more fully. The work of both American groups demonstrated that most of the turnour promoting constiLuents of condensates could be concentrated in the weakly acidic or pScnol fraction, i.e. highly polar material. Stedman et at. further demon- strated promoting activity in the aqueous methanol soluble material of neutral. fraction. These results were consistent with the Harrogate findings which showed removal of promotors by aqueous meLhanoL in different separations into fractions F, H and G2a. Similarly the French team showed that the promotors of WSC were most soluble in the metfianol/ad'eton--itrile pha'se of their solvent system a;id most activity was found in sub-fraction 112 which would be expected to contain material isolated in fraction G 2a of the Harrogate counter current extraction separation. Thus on solubility grounds, 90% aqueous formic acid would be expected to extract the promoting material of fraction G from cyclo- hexane solution into fraction 1R(GJ. Qn QN ON _1 J NJ BATCo document for Province of BritiSh Columbia 9 November 1999 0 0 CL 0 CD 0 CD 0 0 0 Cr E, z 0 < CD Cr CD Scheme 1 Stages'l and 2 G Cyclohexane Fraction 24.517o C clohexane/Caffeine Formic Acid y Sta-c 3 Q(G) Cyclohexane I~esidual Fraction 17.9% Stage 4 R(G) Formic Acid Fraction 2.2% (R + P)G Caffeine Conip!cxing Fraction 3. 5 P(G) Caffeine Complexing Fraction Scheme 2 G Cyclohexane Fraction 2 4. 5',/, Foxt.- Acid/Toluene R(G) Formic Acid Fraction S(G) Toluene Fraction 2. 2C/; 20.6% Cyctohexane/Caffeine Formic Acid Q (SG) Cyclohexane P(SG) Caffeine Res idual Cornplexing Fraction Fr:icLion 17.8% 2. 3 Figure 12 Extraction procedures for romoval of caffeine complexes from Cyclobexane Fraction. (Rotbwell and Whitehead) 8 ezqqf5Oj -6 0 0 CL 0 (D -h 0 T 0 < , 5 0 CD 0 h U) 0 0 z 0 < (D 3 Cr (D (0 to A Whole Smoke Condensate 100% Water B Water Soluble Fraction C Water Insoluble F ractlon 52.0% 45. i3'1~0' Aq. ~IcOfl/Cyclohexane I I F Aq. Methanol Fraction G Cyc :ane Fraction 15.1% 24 5% Cyctohcxane/DA1ISO K (G) DMSO Insoluble L(G) DIVISO Soluble 15.0% 7.570 Cyclohexane/Caffeine Formic Acid Q(Ld) Cyclohexane Residual (R+P)LG Caffeine Complex + Fraction Fornile Acid-Soluble 4.5% .2. 0-Y0 Formic Acid/Cyclohexane R(LG) Formic Acid Soluble 1.00"') P(LG) Formic Acid Insoluble 1.0% Figure 13 Methods of concentration of mouse skin carcinogens of wbole smoke condensate by methods developed at Harrogate 6 ?Zpqf5O, 13) > 0 0 -3 0 0 M 0 (A 0 0 E* -3 Cr F), to z 0 (D 3 Cr -4 w Method of separation Active fractions Active fractions nactive fractions Fractions containing containing PArH PArff-free promotors Neutral fractioi) I'mparlitioll (Fig 1) Neutral fraotion (3.3%) Bas I C (8.rj%) Phenot(12%) Acidic Phenotic (12%) liarrogat,& (F h, 2) HC1 (0. 2%) HC2 (4.4%) (NF-HC) (20.8%) Stedman et at. (Fig 3) (BP-DMSO) (0. 7%) (M--,AW) (1. 5-,,) PE, E, M-PE (B-DMSO) (1. 570) Hoffmann and Wynder A 2 A,, C, D, E, F, G (Fig 3) B (2. 0%) Counter current extraction Izard et a.1. (Fig 6) 111 (4.7%) IV (14.5%) 112 (5%) HArrogate (rig 7) S G1 (2.6%) G2A (2.4%) 33 (52%) F (15%,) G2 B (2. 5%) P (15%) G2 a (4.9%) G2 C (4.2%) G29 (9.30/0). G2b (14.3%) Solvent extrnction Elmenhorst and GrImmar 111(25.9%) 11 (76. 3L7,) (Fig 9) V (6. 8%) I-V (18. 901') V1 (1. 1%) Vil (5. 8%) Harrogate (Fig 10) C (45. 30/0) B (52. 0%) F (15.1%) G (24. 5%) 5. 1'7,) J (17. 9%) if (15. 1%) L (5.3%) K (8.8%) N 1.6%) M (3. 0%) OfZP)j"C /cont O, 0 0 CL 0 3 0 0 0 0 Cr 7 :5' (0 z 0 Cr to CD Metbod of separation Active fractions containing PAM Solvent extmction (continued) Donten%vllt et at. (Fig 11) 111 (24.9%) V (6. 5%) Ix (0. 39%) Caffeine complexing Harrogate (Fig 12) C (45.3%) G (24. 5%) (R+P)G (3. 5%) P(G) (1. 3%) Active fractions PA rH-f ree VIII (3.3%) R(G) (2.2%) Figure 14 Distribution of mouse skin activity In fractions of WSC ,Ezqqf~nf Fractions containing Inactive fractions promotors IT (77.2%) IV (18.4%) VI (2. 9%) B (52. 0%) F (15. 1%) Q(G) (17.8%) F (15. 1%) R (G) (2. 2%)