Acknowledgments. --The aiitht~r.:ire iiiclcbtetl to Dr. Stanford L. Steelman, D r . I)olore* I'atarielli, Dr. .I Brooks, and their conorkers for the endocrine datu. Zearalenone \\as provided for these studies by the Comniercial Solvent5 Corporation through arrangement5
m:itir t>\ 111. .\I
13achm:tri. 'l7Iic :111t11(1r> alho C ~ ) I ( I I P I ~: t p 1 ) 1 ~ t I ( 1 1 1 t o I)r. JJ-. .JoIle\ sild hi, absociate\ for the h! drogeii:ttion*, t o 3Ir. R. S. Bo(>$and his associ:ite< f o r the microanalyses, and t o Dr. B. Arimi aiid ro\\ orkerq for determination of the ninr .pectr:t. t
Relationship between Lipophilic Charac:ter and Henioly tic. Activity of Testosterone and Testosterone Ester5
The relationship between lipophilic character and hemolytic activii y was studied iti a series of testosterone coinpounds. The lipophilic character was first expressed by means of the chromatographic &,, value, which was shown to be related to the partition coefficient between the mobile and t,he stationary phase of a chromatographic. system. The R, value was determined, for each compound, in two different chromatographic: systems, respectively, containing in the mobile phase XenCO or 1leOH. I n both cases there was a highly significant parabolic relationship between R, valuer and hemolytic activity. On the other hand this result could be expected from the linear relationship between the K, values, respectively, determined with XesCO or NeOH in the mobile phase. The lipophilic character of the test compounds was also expressed by ineatis of the T values. There was a linear relationship between the K values calculated from oct,anol-H,O partition coefficients and the chromatographic R,,, values. A parabolic relationship w-as therefore shown alw between T value3 and hemolytic activity. The present results support preview findings, which suggest that the correlation between peuetratioii of organic ('ompounds through biological membranes and partition coefficient is not affected kj!. the iiatiuy of the phases involved in the det,ermination of the partition data.
The hemolytic activity of some neutral steroids was shown by Tateno and IGlbourne' and Palmer.? Weissmann and Keiser3 in a series of 35 neutral steroids and bile acids pointed out that while the 3-p-H configuration was associated with hemolytic properties, the 5-a-H compounds were inactive and only a few A-4,; steroids were active. Testosterone, which is a A-4,5 steroid, was found practically i ~ i a c t i v e , while ~ , ~ I'almer? observed some activity. Segaloff described the hemolytic activity in vitro and in vivo of some testosterone ester?. The hemolytic activity was interpreted as it consequence of the insertion of the steroids at the lipid-itqueous interface a t the surface of the On the other hand, 110 direct relationship was found between hemolytic activity and water solubility." However the acetate derivatives were found t o be more :tctive than their parent c o m p o u ~ i d s . ~Bri enhanced hemolytic activity was also found in the AcOH and EhOH esters of some steroidal sapogenins.6 This could suggest the influence of the lipophilic character of the molecules, a t least in a series of ester derivatives. Hansch, et al. ,' found very good correlat'ions betwrcii partition coefficient and penetration of organic compounds through biological membraiics. 111 prcvious p a p m it was possible to show that rwersed-phitsc, tlc is ;I suitable method for the determiiiatiori of the lipophilic cli:iract~~rof drugs, as esprr. d by t h e R , 37alii~.~ I . 'l'aienvantl E , I ) . Kilhoiirne, I > r o r . S o r . l:r)j. B i o / . , 8 6 , 168 ( l ! i % ) , ( 2 ) I
/im(lie2Co)=
Km(.\leOH)
3 4 )
-0.148
+ 0.728 + 0.288 x
+ + Rm(~e?co) = 1.502 + 1,561
X,,,(iqeoH) = 0.496 0.394 105 BK = I ,289 0.766
log BK
Km(.NIezCO)
T
n
14
t
S
0.993 0.051
14 0.964 0.119 14 0.981 0 . 118 14 0.622 0.431 14
0.954 0 . 1 7 3
14
0..594
0.448
14
0.940
u.1s::
- 1.723
RmZ(hle.CO)
+ 0.53i l o g B R = 0.08T + 2.716 - 1 . 020 log BIZ = 1.192 + 0.209 x
G
10g BR = 0.922
-
R~(>I~oH)
I
Em(31eOR)
S 9
Km?(JleOH)
log B1L = 1 ,155 -+.0 . 172 - 0.1ti9 X'i
T
14 0.577 0.450 14 0.944 0.189
Discussion The present data confirm the existence of a parabolic relationship between lipophilic character and penetration of molecules through biological membranes. The negative sign associated with the Rm2term of eq 5 , 7, and 9 means that the hemolytic activity of testosterone compounds increases and decreases as the R , and K values change and pass through an optimum. The lack of any significant diff ererice between the correlation coefficients provided by eq 5, 7, and 9 shows that the nature of the phases involved in the determination of the partition data does not affect the results. This is explained on the one hand by the very high correlation between R, values calculated with ;\le2C0 and AIeOH, respectively, in the mobile phase and, on the other, by the high correlation between R , and T values. These results support the findings of Collander13 who pointed out that the penetration of organic compounds into Nitella cells was equally correlated with Et20H 2 0 and olive oil-H20 partition coefficients. In fact, C ~ l l a n d e r 'also ~ ~ found a linear relationship betweeii the partition coefficients obtained with two different ( 1 2 ) S . Draper and H. Smitll, "Applied Regression Analysis", Wiley, S e i v York, S . Y., 1966, 11 26. (1:31 (a) R.Collander, A c t r t I ' h p i u l . Scund., 7, -120 (1954); (b) R.Collandcr, ? / ) i d . ,lS, :3FS ( l W 7 ) .
I
51
I
I
I
I
58 62 66 70 Methanol concentration
1
I
I
I
74
78
82
86
Figure 2.-Linear relationship betueen K, values aiid methanol concentration in the mobile phase. Each point i-epresenth the mean of 8-12 values. The compounds are indicated as i n Figure 1.
sets of solvents. More recently Iwasa, et aZ.,14found a very high correlation betmeen T values and R , values, a further support of Collander's findings. The correlation between R , values and H values is very good, except for compound 12. This is the phenylpropionate ester which from its n value seems to be more lipophilic than expected from its R , values. This could be explained on the basis of some group interactions which interfere with the postulate of the additivity of T valu e ~ .I n~ the ~ side chain of the phenylpropionate ester a C6HjCH, group is present. In benzylpenicillin, where the same C6H&H2 group is seen, it was possible to point out, in agreement with an analogous result of Bird and Marshall,15 a similar disagreement between R , and H value.8a However, a t the present time insufficient data are available for any conclusion. The present work seems to suggest that the lipophilic character of the testosterone derivatives plays a major role in determining the disruption of the erythrocytes membrane. Seeman16 distinguished between nonspecific and specific hemolysins. The nonspecific hemolysins exert their hemolytic effect because of their properties of surface-active or lipid-soluble compounds; in very low concentrations they stabilize the erythrocytes from hypotonic hemolysis. Specific hemolysins were coilsidered to be those compounds which interact with a specific membrane component; they do not show the (14) J. Iwesa, T.Fujita, and C . Hanscli, J . .\fed. Chem., 8 , 150 (1065). (15) .i.E. Birdand.4. C . Marshall, Biochem. Z'linrmocol.. 16,227.5 (196i). (16) 1'. Seeman, {bid., 16, 1 7 6 i (1966).
:how stabilizing effect a t any concentration. I J tliih ~ sense vitamin .I.l i phenothiazine t r a n q u i l i ~ c r . , 9 1oc;il ~~~~ :mbthetics,18aalcohols, and steroid;.1811:ire rionspecific.
hemolysins.
T n particular Swniati'hl~ ronsidm1d t rionsperific hem015 siii. I'olyrric~ i i i i t i biotics and haponiris, on tht>othcr h:iii(l, : i i ~cxiniplr~-ot .pecific ht.molj m i . l l i i i is i n agrernieiit ~ i t h this p r e w i t datri, nhich show the influencf of the lipophilic character 011 thtl hemolytic :Lctivity of nonspecific hemolysin+ -uch ;I+ testosterone ebteri The lipophilic churactrr should wert its etfect d s o 111 the c of specific hthinolybiiii. tcroric to bt.
:I
- 7
Steroidal Heterocycles.
XI1l.'d 4cu,5-Eyoxy-5~androst-2-eno[2,3-d]isoxazoles and Related Conipouncls
Several compounds described iii this paper, which were prepared as p u t of a continuing program in thew laboratories involving the preparatioii arid reactions of steroidal heterocycles,Ia were found to inhibit adrenal cortical function. Jt'e have used them in attempts :it molecular modification which has played such an important role in the developmerit of better nrid safer tlrugs. htidrosta-%,G-dieno[%,3-r2]isosnzol-li~-ol(la) 011 t,rratnient with either peracrtic or perphthalic acid i r i C16H6,consistently \.ielded a inisturr of 4cr,,ky)oxy-.5aandrost-%-en0[%,~-rl]isoxazol-liB-ol (2aj with starting material in a ratio of approximately 1:2. Ho~vever. when a solution of isosazole la in CHCl, or CHaCl211 treated x i t h either permaleic acid4 or m-chloroperbenzoic acid5 in the presence of a small amount of pyridine, 2a was obtained in SO-90c{4 yield. r , -2-rai io [ 2 , :-;-(I 11r w t nic I I t of 1CY, .i-r.posj.-.i ~~-:iridi*ost isoxnzol-li@-ol actltntc (2b) wit.11 aq A\Ir2CO--HnS0, yit~ldrd 5a-androst-:! -t21io [ % , 3 - t 7 ] i s o s : ~ z o l ~ - ~ ~ , ~ ~ l i ~ ~
111 la) I'aper M I : T. C . Miller, J . ~ f d ~ r o r !C~' hdu .n ~ . .3, 338 r l O t i t i ) : (I), . \ l b a n r Medical Center Hospital, Albany. S . Y. (2) I t . 0. Clinton. A. J . l l a n s o n . F. \r.Stonner, 11. C:. S t w m a n n , t1. (;. ('!iristiansen, R . I,. Clarke. J. H. Aokrrrnan, I ) . F. Page. .I. I\'. I)ean, \V, 13. ])irkinson, a n d C y . Carahatraw, J . 9 m r r . C l i r m . S O C . , 83, 1478 (1961); I:. ('liancr. "Amino .\rids, I'rotPinv, a n < ] ('nncer Iliochemisiry," .J. T. I
