Document mmJb0o4KJnMB2pMVdY3E9NLY0

i Estrogenic Activity of DDT Analogs and Polychlorinated Biphenyls Joel Bitman* and Helene C. Cecil Became of the geometric similarity of DDT to the phenyls, compounds which arc environmental pol synthetic estrogen, stilbestrol, DDT and 52 related lutants of industrial origin, were cstrogenically compounds were tested in a sensitive estrogen assay active. Phcnolphthalol and phenolphlhalein, com in rats. Estrogenic activity was evaluated using the pounds which arc used as laxatives in drug prepara 18-hr glycogen response of the immature rat uterus. tions, were also esirogcnically active since they con Diphcnylmclhanc, diplv nylcthanc, and triphcnyl* tain the appropriate p,/?'-dihydroxy structures. methane compounds wer: active when a p~ or /j'-posi Stereo models indicated that /,p'-dihydroxy com i tion was unoccupied or occupied by an hydroxy or pounds of the active nuclei would have internudear methoxy group. Halide or alkyl groups in the im distances o' the hydroxyl groups which would ap positions rendered the compounds inactive. Poly proximate 'hose of the natural steroidal estrogens chlorinated biphenyls and polychlorinated tri and the synthetic slilbene estrogens. n 1945 Solmssen published an excellent and comprehen (p-chlorophenyl)ethylene; p.p'-DDMU, l-chloro-2,2-bis- I sive 117-pagc review of the synthetic estrogens and the relation between their structure and activity. In the (p-chlorophcnyl)ethylcne; and p,p'-DDA, 2,2-bis(p-chlorophenyljacetic acid. 1 25 years since Solmsscn's review, DDT, a chlorinateTdahbyledrIoI: o.p'-DDT, l,l,l-trichloro-2-(p-chloropheny|)-2- carbon with a geometric similarii v to the synthetic estrogens, (o-chlorophenyl)ethane; o,p'-DDE, l,l-dichloro-2-(p-cl loro- has been widely used throughout the world for pest control. phenyl),2-(o-chlorophenyl)cthylene; o,p'-DDMU, l-ch!oro- The recent demonstration by Welch et at. (1969) of the estro- 2-(p-chlorophenyl), 2-(o-chlorophenyl)ethylene; o,p'-DDD, ; genic activity of o.p'-DDT and i>ur own investigations (Bit- l, 1 -dichloro - 2 -(p-chlorophenyl), 2 - (o - chlorophcnyl)ethane; 1 man et a/., 1968) have prompted js to investigate a series of m, p'-DDD, l,l-dichloro-2-(p-chlorophcnyl),2-(;n-chioro- DDT analogs, homologs, and structurally related compounds phenyl)ethar e and p,p'-Mcthoxychlor, 1,1 ,l-lrichloro-2,2-bis- in an attempt to determine relatic nships of structure to estro (p-mcthoxyphenyl)elhanc. genic activity. Table HI: Compound 20, 1,1-Diphcnylmcthanc; 21, 1,1- ' METHODS Dichloro-I,l-diphcnylmeihanc; 22, l,l-bis(p-bromophcnyl)methane; 23, I-(phenyl)-I-(p-hydroxyphcnyl)mcthane; 24, i We used the sensitive 18-hr glycogen response of the rat l,l-bis(p-hydroxyphenyl)mcthane; 25, 1-f/vhydroxyphcnyl)- i uterus as a measure of estrogenic activity (Bitman et al., l-Cp-methoxyphenyl)mcthane; 26, l-(phcnyl)-l-(p-mclhoxy- 1965). The potency of active compounds is reported in o-hydroxyphcnyl)methane; 27, l.l-bis(o-hydroxyphenyl)- terms of the minimal subcutaneous dose which will increase methane; 28, l,l-bis(()-hydroxy-t-chlorophenyl)mcthane; glycogen to a level significantly different from"control. The 29, l-(phenyl)-l-(p-chlorophenyl)-l-methanol; 30, l.I-bis(p- 18-hr glycogenic response is illustrated in Figure 1, in which chlorophenyl)-l-methanol; 31, 2,2'-dihydroxybenzophenone; , the dose-response curve for o,p'-DDT is represented. The 32, 2,4-dihydroxvbcnzophenone; 33, 4,4'-dihydroxybenzo- i steeper response line for glycogen, as compared to uterine phenone; and 34, 2,2'-dihydroxy-4,4'-dimethoxybcnzophe- weight, is readily apparent. none. Test substances were dissolved in olive oil or an aqueous Sources of the compounds used in this study were: Nos. < ethanol solution and injected subcutaneously at a screening 4, 5, 16--Rohm and Haas, Philadelphia; 9, 13--Dr. G. F. 1 do*e rate of 8 mg per rat. Immature female Wistar rats Fries, U.S. Dept. Agr., Beltsville; J7--E. K. Du Pont de (21-23 days old; 36-48 g) were killed 18 hr after the injection; Nemours & Co., Inc., Wilmington; 18--Sigma Chemical Co., i uteri were quickly excised, weighed, and analyzed for glycogen St. Louis; 24, 27, 35, 42--K & K Laboratories, Inc., Plain- by the enthrone procedure (Seifter et al., 1950). Substances view, N.Y.; 38--Eastman Kodak Co., Rochester; 43-53-- f showing activity were tested further at dosage levels to 0.05 Monsanto Co., St. Louis. All other compounds were pur i mg. Statistical comparisons were made using Student's t test chased from the Aldrich Chemical Co., Inc., Milwaukee. ; with correction for unequal group size. Purity, as given by the manufacturers, was better than 99%. I ABBREVIATIONS ' RESULTS AND DISCUSSION ' ` Tabic I; p,p`-DDT, l,l,l-trichloro-2,2-bis(i-chlorophcnyl)ethanc; Tetrachloro-DDT, I,l,l,2-tetrachloro-2,2-bis- The natural estrogens are steroids which contain a phenolic ring A and an oxygen function at the Cn position, while the (y>-chlorophcnyl)cthane; p.p'-DFDT, I,l,l-trichlaro-2,2-bis- synthetic estrogens, which are stilbene derivatives, corsinin 0>-fluorophenyl)cihane; p.p'-Perthane, l,l,l-trichloro-2.2- two phenolic rings (Figure 2). It is apparent that active ! bis(/-cthylphenyl)cthane; p.p'-Kelthane, 1,1,1 -trichloro-2,2- estrogcnicity is dependent upon the presence of at least one bisf/xhlorophenyljclhane; p.p'-DDTF, I,l,l-trifluoro-2,2- phenolic hydroxy ring structure. In most estrogen tests these bis(i-chlorophcnyl)ethane; pj>'-DDD, l,l-dichioro-2,2-bis- compounds are active in the microgram or submicrogram (p-chlorophcnyljethane; /m'-DDE, l,l-dichloro-2,2-bis- range. , In contrast to this, the chlorinated hydrocarbon pesticides related to DDT are only active as estrogens in milligram Animal Husbandry Research Division, Beltsville, Md. 20705 To whom correspondence sh yuld be addressed. amounts, a 1000-fold difference. The DDT analogs arc not phenolic, but they may give rise to aromatic phenolic sub stitution during metabolic conversions in the animal. The 1.1QJ J. AGR. FOOD CHEM., VOL. I*. NO. 6, 1970 * DSW 332404 STLCOPCB4078980 Figure 1. Doae-response relationship: treine weight, glycogen, and tf.p'-DDT DDT analogs arc compounds of the diphcnyleihane type (Figure 2). Other analogs tested were compounds of the diphcnylmcthane or triphcnylmethane series. We have also examined polychlorinated biphenyls and polychlorinated triphenyis, compounds which have become increasingly im plicated as cnvironmenlal pollutants of industrial origin. Wc have not included in this study any steroids, synthetic estrogens, or anti-cstrogens of the stilbene structure, and have excluded almost all compounds of the coumarin, isoflavone, anthracene, and phenanthrene type. DIPHENYLETHANE COMPOUNDS - '-Positions Occupied by Halide or Alkyl. The com pounds evaluated in Table I are diphcnyleihane derivatives in TaHe I. Diphenytettiane Compounds with p,p'-Positions Occupied by Halide or Alkyl Groups No. X Na. 1a H --CCT, p.p'-DDT 2a a --ca, Tetrachloro-DDT JF h --CCIi 4 CH.CH, H --ca, P,p'- DFDT P,p'-Perthane 5a a oh --ca. p.p'-Kehhane h --CF, p^'-DDTF 7a h --cHa, p.p'-DDD "a ... --ca, p*'-DDE a ... --cHa p.p'-DDMU >o a h --COOH p.p'-DDA * M.E.D. - minimum effwetiv* dose. * I - inactive. Activity M.E.D.* 4 4 I* P P P P P P P which the p,n,-positions arc occupied by halide or alkyl groups. Almost all were d.evoid of estrogenic activity; p,p'- DDT (empd I) and tclrachloro-DDT (empd 2) exhibited a slight glycogenic response. It appears that halide or alkyl substitutions in the p,p'-positions were stable, and during me tabolism in the animal body, little if any p,p '-phenolic hydroxy compounds are produced. p~ or p,p' Position Occupied by --H or--OCH,. When one of the para positions of the aromatic ring is substituted by a hy drogen or methoxy group, the compound exhibits estrogenic activity (Table II). Potency is of a low order of magnitude, being approximately 1000 times less active than compounds of the stilbene series, but is similar in potency to coumarin and isoflavone estrogens (BickofT el a/., 1960). The minimum eifective dose (MED) of diethylstilbestrol which elicited a glycogen response was 0.1 pg, as compared to the most active compound of Table II, o^j'-DDT, empd 11, whose MED was 0.25 mg. The phenolic character of the natural and synthetic estro gens has demonstrated the dependency of estrogenicity upoi the presence of a phenolic structure. The aromatic rings of the active compounds of Table If are open, i.e., they have a p- or p'-position occupied by --H and may give rise to phenolic substitution during metabolism. There also appears to be a requirement for the ethane chain to be inert, i.e, either the tri- chloroethane (-CH-CCIi) or the vinyl halide group (>C CCIj) must also be present (empds II, 12, 13). Thus, empds 14, 15, and 19, containing more reactive 2-carbon chain con figurations, are inactive, even though one of the aromatic rings could be hydroxylated to the phenolic structure. We have concluded that rapid in vivo metabolism of these compounds is responsible for their lack of activity. ' In the stilbestrol series (Solmssen, 1945) and in the coumarin series (BickofT et at., 1960) of estrogens, p,/>'-dimethoxy com pounds are less active than comparable p^'-dihydroxy com pounds. In the chlorinated diphenylethane series (Tabic II), the p,p'-methoxy compound, methoxychlor (empd 17) was O STLCOPCB4078981 TAh IL . Dtplwytetoe Compmnds wttk p- or p'-Poeltkm Occupied by --H or --OCHi Ro Nat P Onopi P' 0 11 12 H H '-- ci a 1) H a 14 H a IS . H a 16 17 18 H OCH, OCH, H OCH, OCH, 19 H H a a a a m-CJ H H oH'-a o-CI M.E.D. minimum effective dose * 1 inactive. R .H.. ... H H H H H H R' --CC1, --ecu --CHC1 --CHC1, --CHC1, --CC1, --CCI, --CHO Name 0,p'-DDT o.p'-DDE <?,p'-DDMU o.p'-DDD m,p'-DDD I,l,I-Tnchloro-2,2-bis(phenyl)e(hene p.p'-Methoxychlor Tech. Methoxychlor (p.p'- + o,p'-) 2,2-Bis(o-chlorophenyl)acetaldchyde Activity Mg 0.2? 4 8 I* P 1 4 1 1* approximately as active as other compounds which might give rise to phenolic hydroxy substitution on metabolism. Tech nical methoxychlor (cmpd 18), which may contain an o,p'methoxychlor, was four times mjre active than pure p,p'~ melhoxychloc. DIPHENYLMETHANE COMPOUNDS; BENZOPHENONE COMPOUNDS A series of diphenylmethane compounds was examined to determine structural correlates of estrogenic activity (Table III). Active compounds contained either one or two p-hydroxy or p-methoxy groups (cmpds 23, 24, 25, 32, 33, 34). ----- ------------------------------------- ----------------------- i _. _ Table III. Diphenylmethane Compounds; Benzophenone Compounds R 0p'@-c-< >p 9 p<^c-p Group* Activity Ne. P P' R R' Mg Diphenylmethane derivatives 20 H 21 H 22 Br 23 H 24 OH 23 OCH, 26 OCH, 27* H 28* H 29 H 30 a 31 H H H Br OH OH OH H H H a a H HH aa HH HH H\ H H 'H H 'H HH HH H OH H OH ... ... P I* P 2 1 4 P P I P P P Benzophenone derivatives 32 H 33 OH 34 OCH, OH OH OCH, ... ... .. 2 2 4 M.E.D. - minimum effective dose. * I - inactive. *bis(o-hydroxyphenylpnethane. * bit(o-hydroxy, m-dilorophenybmethane. The most ictive compound was p.p'-dihydroxydiphi nylmethane which elicited a glycogenic response at the 1 mg dose level. Solmssen (1945) reported activity for this compourd at the 100 mg level, but the differences in bioassay proced jres could explain part of this difference in result. In the diphenylethane series, compounds with a p-hydrcgen and a stable ethane chain were metabolized to active estrogens, probably containing a p-hydroxy structure. In contrast, di phenylmethane compounds with a p-hydrogen were not active, probably being metabolized rapidly at the methane linkage and excreted from the body. Benzophenone derivatives, which contain the more stable ketone structure at the methane carbon, were active if a p-hydroxy was present (Table HI). DIPHENYL PROPANE COMPOUNDS; TRIPHENYL METHANE COMPOUNDS Two diphenylpropane compounds were active (cmpds 35 and 36), the p,p'-hydroxy compound exhibiting much greater activity than a p,p'-dimcthoxy compound. Dihydroxy di phenyl propane (cmpd 35) was as active as o,p'-DDT. Solms sen (1945) found that this compound was active at a 100 mg dose level. Since the p,p'-dihydroxy structure appeared to be the struc ture conferring activity, phenolphthalol, a phenyl substituted diphenylmethane compound containing p,/'-dihydroxy groups____ was tested. Phenolphthalol was as potent as any compound of the types studied. Ring closure, as in phenolphthalein, resulted in a 20-fold loss in potency. These compounds are not known to have estrogenic activity and are extensively used as laxatives in a number of drug preparations. In Table IV two miscellaneous derivatives which bear some relation to closed ring diphenylmethane structures are in cluded: fluorene and 9,10-dimethylanthracene. Both of these compounds were inactive when tested at dose levels up to 8 mg per rat. BIPHENYL AND TRIPHENYL COMPOUNDS ' Two hydroxy biphenyl compounds were active but only at the 4 and 8 mg dose levels (Table V). In a series of poly chlorinated biphenyls, the compounds containing up to 48% chlorine were active. A: judged from glc chromatograms, DSW 332406 -- -1 1110 J. AOR. FOOD CHEM- VOL. 18, NO. 6, 1970 STLCOPCB4078982 -JOCP3 O-O: H-H: CITtAMI *X iiaX ItllHNI till naX 4-* c OtPHINYinOTAMI -c-p. -?- MFHKNU (THANK 0-0: H-H: *.4* to.*X MKHKNUMETHANK p aiX mX TttFHENYlMETHANK A@H>* smNvt 0-0: AtoS .*X H- H: kmX P*-; TltTHENYl AtoI l4jX IS.0X P m9 TKIPHCNYl A TO g Tjt V.tX 0-0: AtoC ail H-H: 7joX AtoC mil ml Figure X Structural formulae of estrogenic ampomde these products are crude mixtures containing a number of compounds. A polychlorinated triphenyl containing 42% chlorine was found to be more active, at a 1 mg dose level. CORRELATIONS DETWEEN CHEMICAL STRUCTURE AND BIOLOGICAL ACTIVITY Schucler (1946) and his coworkers (Fisher et a!., 1952; Keasling and Schueler, 1950) have theorized t/sat a rather large, rigid, lipoid soluble molecular structure with two active hydrogen-bond forming groups located at an optimum dis tance of 14.5 A units from each other would be estrogenic. They further staled that potency is decreased as the distance between groups is decreased or increased. While DDT possesses a relatively large, rigid, lipoid soluble molecular constitution, it does not present active hydrogen atoms at the hypothesized optimum distance of 14.5 A, how ever. The presence of the electronegative chlorine atoms in the pj> '-orientations would prohibit the existence of active hydrogen. If these p.p' chlorine atoms were metabolized to groups possessing active hydrogen, the possibility of estrogen action would exist. The general lack of estrogenic activity of p,p'~DDT analogs suggests that such metabolism does not occur readily in the biological situations studied thus far. Conversely, the activity of o,p'~DDT raises interesting theoretical relationships between chemical constitution and estrogenic activity. The o,p' chlorine atoms are not at the hypothesized optimum distance. The exact nature of the active estrogen structure arising from o,p'-DDT, if it is not i7'-DDT itself, might provide important information relating to the spatial configuration of an active estrogen. Dreiding Stereomodels were constructed of many of the active structures to determine whether consistent stereochemi cal factors were present. Intemuclear distances were mea*urcd and were given in Figure 2 for both O to O atoms of assumed dihydroxy compounds, and for the H to H distance of the hydroxyl groups. Interatomic distances for the Dreid ing models of extrane and stilbene were found to be much No. 35 36 37 TaMe IV. Dlphenylpropane and Triphenylmethane Compounds Formula Name Activity M.E.D.* mg 2,2-Bis(p-hy- 0.25 droxyphenyl)propane CM,CI CHjCt CH, 1,3-Dtchloro- 2,2-bis(pmethoxy, m-methyF phenyl)propane 4 Phendphthaio) 0.2 3* 'tf5- . Phenol- 4 phlhaldn 39 Fluorene 1* 40 9,10-Dimethylanthra-- CM, cent M.E.D. - minimum effective dose, i1 -- inactive. I* STLCOPCB4078983 X - TaMe V. Biphenyl and Tripbeayl Coapoonda are only slightly smaller than corresponding bond distances in natural and synthetic estrogens. > ! i Na. Nam Activity 3LE.D.* Mg The structural observations regarding estrogenic ac.ivity in the compounds studied indicated that activity is conferred when a p- orp '-position is unoccupied (--H), or is substituted by --OH or --OCH*. Halide, or alkyl groups, occupying,the p,p'-positions render the compounds estrogenically inactive. 41 a^'-Biphenol 2,2'-Dihydroxydiphenyl 4 A stable ethane chain was found to be necessary for uctivity, 42 ^/'-Biphenol 4,4 '-Dihydroxydiphcnyl 8 e.g., the trichloroethane or the inert vinyl halide group; if Polychlorinated Biphenyl (PCB) 43 PCB Aroclor 1221 21% Chlorine 44 PCB Aroclor 1232 32% Cl either C of the ethane chain bears an oxygen function (alco 8 8 hol, atdctiyde, or acid), the compound is metabolized and no estrogenic activity is observed. Some polychlorinated bi 45 PCB Aroclor 1242 42% Cl 8 phenyl and triphenyl compounds exhibited estrogenic uctivity. 46 PCB Aroclor 1248 48% Cl 8 Measurements of internuclear distances of Dreidin; steric 47 PCB Aroclor 1254 54%C1 48 PCB Aroclor 1260 60% a 49 PCB Aroclor 1262 62% Cl | 30 PCB Aroclor 1268 68%a P models indicated that active sites would be 9-11 A apart, a P P range similar to those found in natural and synthetic estro P gens. Quantitatively similar estrogenic activity was obtained t 1; 51 PCB Aroclor 4465 60% PCB, 40% potychlorinaled triphenyl (PCT), with a series of diphenylmethane or triphenylmethane deriva tives which contained p-OH functions. Correlations of struc 65% Cl 32 PCT Aroclor 5442 42% Cl P I ture will activity suggest that the active estrogens derived 53 PCT Aroclor 5460 60% Cl P from o,p'-analogs of DDT are p-phcnolic metabolites. * M.E.D. - minimum effective Joee. *1 - inactive. I LITERATURE CITED - smaller than the 14.5 A quoted by Keasling and Schueler (1950). The 10.9 A we found agree closely with the X-ray crystallographic data of Norton et at. (1963, 1964), who found 10.95 A for 175-estradiol. This discrepancy in interatomic dif ferences may be related to the improved accuracy of the cur rent atomic models, when compared to those used in 1950. The diphenylethane, diphenyimelhane, diphenylpropane, triphenylmethane, biphenyl, and triphenyl compounds all have interatomic distances of 9.4 to 10.3 A for the most likely O to O substitutions. The H to H internuclear distances of the hydroxyl groups range from 9.1 to 11.i A in these compounds. Bkkoff, E. M., Livingston, A. L., Book, A. N., Arch. Biochem. 88, 262 (I960). Bilman, J , Cecil, H. C., Harris, S. J., Fries, C. F., Science 162, 371 (1968). Bilman, .!., Cecil. H. C, Mench, M. L., Wrenn, T. R., Endocrin ology 75, 63 (1965). Fisher, A. L.. Keasling, H. H., Schueler, F. W.,Proc. Soc. I'xp. Bio. Med. 8., 439 (1952). Keasling, H. H.. Schueler. F. W,,J. Amer. Pharm. Ass. 39,87(1950), Norton, D. A., Kartha, G., Lu, C. T,, Acta Cryst. 16, 89 (1963). Horton, D. A.. Kartha, G,, Lu. C. T.. Acta Cryst. 17, 77 (1964). Schueler, F. W., Science 103, 221 (1946). Seifter, S,, Dayton, S., Novic, B., Muntwyler, E., Arch. Biochem. 25, 191 (1950). Solmssen, U. V.. Chem. Bee. 37, 481 (1945). Welch, R. M., Levin, W,, Conney, A. H.f Toxicol. Appl. Pharmacol. 14, 358 (1969). Both the O--O and H--H internuclear distances, therefore, Receivedfoe review June 25, 1970. Accepted August 13, 1970. i\ ' '^ i <\ V DSW 332408 STLCOPCB4078984