STEROID EXCRETION IN A CASE OF ADRENOCORTICAL CARCINOMA
V. 45-PREGNENETRIOL-38,17a,20a*
BY H. HIRSCHMANN AND FRIEDA B. HIRSCHMANN
(From the Department of Medicine, School of Medicine, Western Reserve University, and the Lakeside Hospital, Cleveland)
(Received for publication, June 19, 1950)
The examination of the 36 fraction of non-ketonic compounds from the urine of a boy with an adrenocortical tumor revealed the presence of three substances which appeared to be different from any steroids hitherto described (1). One of these has recently been identified as 45-pregnene- triol-33, 16a,20a (2).1 Since then 13 mg. of a second of these substances (Compound C) were accumulated which allowed the elucidation of its structure as 45-pregnenetriol-38, 17a, 20a.
The reported analysis of the acetate of Compound C (I) was considered to be indicative of a diacetate of a compound C20H3203 ± CH2 (1). Fur- ther analysis of this product has given values in excellent agreement with the composition of the higher homologue (C25H3805). The dried parent compound proved to be hygroscopic and after exposure to the atmosphere analyzed for the hemihydrate of C21H3403. It seemed probable, there- fore, that the analytical data obtained for the dried preparation which are low in carbon reflected a residuum of water rather than a composition dif- ferent from C21H&O3 for the anhydrous substance. Final and convinc- ing proof for the correctness of the C21 formula came from the synthesis of Compound C which is reported below. Examination of the infra-red spectrum of the diacetate showed an absorption peak at 2.78 u, which is characteristic of an O-H bond. It was concluded that Compound C was a triol with one inert and two reactive hydroxyl groups. This in turn suggested the presence of one olefinic double bond, if the compound
* This investigation has been supported by grants from the Bourne Fund and from the American Cancer Society on the recommendation of the Committee on Growth of the National Research Council.
1 The stereochemical nomenclature employed conforms to current usage, which frequently differs from that in the references cited. Details concerning the use of the terms a and @ for most of the relevant positions were specified in the preceding paper (2). In addition, it should be noted that 3-chlorosteroids are now designated in conformity with Shoppce (3), and 17a-hydroxy-D-homosteroids are renamed in view of the evidence presented by Goldberg et al. (4). The 20-hydroxysteroids of the 17-iso series which have not been related with certainty to those of the normal series (5) are designated “a” and “B” in accordance with Fieser and Fieser (6).
C21H3O3 is tetracyclic. The most likely locations of an unreactive hy- droxyl group in a urinary steroid are the 116 and 17@ positions. In gen- eral substitutions at these sites have a definite effect on the optical rota- tion of the molecule and can be distinguished by the sign of their contribution. Since compound I was found to be strongly levorotatory ([M]D = - 331°), the presence of the dextro-displacing 118-hydroxyl group (7) was considered to be unlikely. Moreover, the molecular rota-
CH, C=0
OH
ACO
CH3 1
17
LiA1H4 Ac20
CH3 1
HCOAc
AcOCH
OH
OH
Ac0
Ac0
V
[M]p =- 339º
I
[M] =- 1130
NaOH
CH3 1
CH3
HỌOH
HOCH
OH
ОН
HO
II
HO
HI04
HI04
VI
0
HO
III
tion of the isolated diacetate differs from that of allopregnanetriol-33, 17a,- 20x diacetate (8, 9) by 203°, a value compatible with the shift observed for reduction of a 5-6 double bond if measurements are made in alcohol or acetone (10, 2).2 It seemed possible, therefore, that the isolated com-
2 In discussing the correlation between location of the double bond and the shift of rotation on reduction (2) we unfortunately were unaware of the recent simultane- ous publications of Barton and Rosenfelder (11) and of Wintersteiner and Moore (12) which give satisfactory reference data for the 6-7 double bond. The values are significantly different from those observed on reduction of the 5-6 double bond. (A[M];at. - 46 = 441° for 46-cholestenol-36 acetate in chloroform.)
pound was A5-pregnenetriol-38, 17a,20a. In order to test this hypothesis, the triol (II) was treated with periodic acid. The reaction product (III) in carbon disulfide showed an absorption maximum at 5.74 u, as has been observed with numerous 17-ketosteroids (13). The absorption in the finger-print region of infra-red radiation was in very close agreement with that given by a reference sample of dehydroisoandrosterone. Moreover, benzoylation yielded a sparingly soluble derivative that showed no de- pression of its melting point when mixed with authentic dehydroisoan- drosterone benzoate. The identification of the oxidation product (III) as dehydroisoandrosterone establishes in the starting compound (II) the nature and configuration of the ring system, the location of the double bond, the locations of the three hydroxyl groups, and the @ configuration of the substituent at C-3, but leaves the spatial arrangement of the hydroxyl groups at C-17 and C-20 undetermined.
While the optical rotation of the diacetate agrees well with the value calculated for the 17a,20a configuration, it does not differ sufficiently from the expected value of 17-iso-45-pregnenetriol-36, 178, 20”(” diacetate3 to permit the elimination of this structure with assurance. In order to prove the configuration of the isolated compound, it was related through synthesis to a starting compound with known orientation of the 17-hy- droxy group, 45-pregnenediol-36,17a-one-20. This substance was avail- able through isolation from the ketonic fraction of the same urine (18). Its configuration at C-17 has been determined by Reichstein and his collaborators (19, 20) who obtained it by synthesis. The 3-monoacetate (IV) of this compound was reduced with lithium aluminum hydride. This method when applied to other 20-ketosteroids (9, 2) has given a mix- ture of both 20-epimeric alcohols in which the @ isomer appeared to pre- dominate. In this case again two products were obtained which were separated after acetylation. The isomeric diacetates (I and V) were con- verted to the free triols (II and VI) which were degraded with periodic
3 A product to which this structure has been assigned has been described by Bu- tenandt et al. (14). Recent experiments of Salamon (15), however, strongly indicate that this material is a molecular compound of the diacetates of 45-pregnenetriol- 38, 17a,208 and of a structural isomer, possibly 45-pregnenetriol-38, 16a, 17a. These findings would explain the conversion of Butenandt’s product to 17-iso-45-pregnenol- 36-one-20 acetate on heating with zinc (14), if it is assumed that the 3,20-diacetate in the mixture is the actual starting material (16, 17). It is less apparent, how- ever, how Salamon’s results are to be reconciled with the observations of Fieser and Huang-Minlon (17) who obtained 45-pregnenol-38-one-20 acetate on repeating Bu- tenandt’s experiments. In spite of this uncertainty, it seems warranted to consider Salamon’s new diacetate (melting at 185°) (15) as authentic 17-isoallopregnanetriol- 38,17,20”B” diacetate and to use its molecular rotation (-67º in acetone) to esti- mate that of the diacetate of 17-iso-45-pregnenetriol-36, 178,20”(” (about -286° in alcohol or acetone).
acid. In both series the reaction product was dehydroisoandrosterone (III). It follows that both triols are formed without rearrangement to D-homosteroids and therefore represent the two diastereoisomers of 45- pregnenetriol-38,17a,20. The physical constants of the lower melting diacetate (V) and of its parent compound leave little doubt that this triol is the isomer identical with the 45-pregnenetriol-33, 17a,200 synthesized by Butenandt, Schmidt-Thomé, and Paul (14). Since its diacetate has been reduced (21) to the diacetate of Reichstein’s Compound J (allopreg- nanetriol-38,17a,200), the reference compound for the 200 configuration in 17-hydroxysteroids (6), it follows that our higher melting reaction prod- uct, the diacetate I, must have the 20a orientation. It is of interest
1300
1200
1100
1000
900 cm
-1
A
100
100
B
90
90
B
80
Per cent Transmission
80
70
70
A
60
60
A
50
50
40
40
30
30
20
20
10
10
0
B
8.0
9.0
10.0
11.0 ℃
that it was obtained as the main product of the reduction. This com- pound proved to be identical with the isolated substance, as was shown by a comparison of the melting points, of the rotations, and most con- vincingly of their infra-red spectra (Fig. 1). The free triols, like the di- acetates, showed no depression of their melting points on admixture. This identity fully confirms the results of the degradation of Compound C and establishes its structure as 45-pregnenetriol-33, 17a, 20a.
These findings complete the identification of seven substances isolated from this tumor urine which were not previously known to occur in hu- man urine. Only one of these, 45-androstenetriol-36, 16a, 178 (22), has since been reported to occur in low concentrations in normal urine as well (23), but the excretion of small amounts of the other substances (with the exception of 17a-methyl-45-D-homoandrostenediol-38, 17aß-one-17 which
most probably is an artifact (18)) by normal men and women is considered not only possible but even probable. We have felt justified, therefore, in considering the adrenal tumor as the most probable site of origin of these compounds only if their urinary titers greatly exceed their (for the most part as yet hypothetical) normal levels. The compounds described thus far were obtained in amounts so large that their presence in similar concentration in normal urine is highly improbable. Although the average yield of 45-pregnenetriol-36, 17a,20a (1.3 mg. per liter of urine) was considerably lower, we do not doubt its adrenal origin, since the same urine had yielded much larger amounts of the closely related 45- pregnenediol-33, 17a-one-20 (18). Since several 20-ketosteroids that were administered to man were shown (24-27) to be converted to 20a-hydroxy- steroids, a metabolic interrelationship between 45-pregnenediol-33,17a- one-20 and 45-pregnenetriol-33,17a,20a scems a very likely explanation of the presence of both compounds in the same urine extract.
45-Pregnenetriol-38, 17a, 20a agrees with the other compounds4 reported for this case in the presence of the 5-6 double bond and in the @ con- figuration of the substituent at C-3. The examination of the acetylated non-ketonic material that failed to precipitate with digitonin has re- vealed little which would alter the view that the 45-30-stenols represent the predominating type of steroid excretion products in this case. The bulk of the crystalline material obtained from this fraction consisted of the diacetate of 45-androstenediol-38, 178 (which forms a rather soluble digitonide). We also obtained small amounts of pregnancdiol-3x, 20a diacetate (1.5 mg. of diol per liter of urine) and of four unidentified com- pounds. One of these, the acetate of a benzenoid tricnol, has been de- scribed briefly (28); the others were obtained in traces and have been characterized only by their melting points (131-133.5°, 209.5-210.5°, and 226-228°, respectively). It would appear from the data obtained thus far that the shift in excretion products from the normally predominant 3a-stanols to the 46-36-stenols is more general in this adrenal tumor case than in any other of which we are aware. Hypotheses have been set forth (1, 29) to explain this shift, but as yet none has received sufficient experimental support to permit its adoption (29).
EXPERIMENTAL5
Isolated 45-Pregnenetriol-33, 17a, 20a 3,20-Diacetate (I)-The isolation of the diacetate (Compound C acetate) from the remaining extracts (22)
4 45- Androstenetriol-38, 16a, 178 (22), 45-androstenediol-38, 178 (1), 45-pregnenedi- ol-38,20x (1), 46-pregnenediol-36,17a-one-20 (18), 17a-methyl- 45-D-homoandrostene- diol-38, 17aß-one-17 (18), 45-pregnenetriol-38, 16a,20x (2), dehydroisoandrosterone (18), 38-chloro-45-androstenone-17 (18).
6 All melting points reported are corrected. Compounds were dried for analysis and rotation at 110° unless noted otherwise. The rotations were measured in 95
of the patient’s urine followed the procedure described (1) except that our standard size columns (r2h/w = 10 cm.3 per gm., where r = radius, h = height of the column, and w = weight of material) were found to be preferable for rechromatographing the crude diacetate from alumina to the shorter ones used before. The final product melted at 203-206°.
Analysis-C24H3605. Calculated. C 71.25, H 8.97
C26H 3905. Found. ” 71.76, ” 9.18
” 71.74, ” 9.15
Rotation-[a]} = - 79° (c = 0.38)
The main absorption maxima in the accessible finger-print region of infra- red radiation (Fig. 1, Curve B) were at 13.87, 12.50, 12.32, 11.345, 11.07, 10.86, 10.63, 10.37, 10.28, 10.05, 9.69, 9.51, 9.26, 9.09, 9.00, 8.80, 8.57, 8.37, 7.63, and 7.28 u. The ester peaks were at 8.02 (broad) and 5.74 ;, and the hydroxyl peak at 2.78 u. The material studied was an aliquot of the product analyzed and was dried under the same conditions. The hy- droxyl peak, therefore, cannot be ascribed to solvent retention.
Isolated A5-Pregnenetriol-38,17a,20a (II)-The isolated diacetate (I) was hydrolyzed as described before (1). The triol was recrystallized from 95 per cent alcohol and from acetone and melted with some browning at 222-225°. A sample was dried at 110° in vacuo for 3 hours and sent to Dr. Huffman for analysis. On redrying at 100° in vacuo for 1 hour it lost 2.10 per cent of its weight but regained it rapidly on exposure to the at- mosphere (most of it within 5 minutes). The rehydrated product was analyzed. The first set of figures (a) is based on the weight of the sample before redrying, the second (b) on that of the dried sample and on a water weight which is corrected for the weight increase on exposure to air.
Analysis — (a) C20H3203 .¿ H2O. Calculated. C 72.90, H 10.10 C21 H3403. ¿ H2O. ” “73.43, ” 10.27 Found. ” 73.34, ” 10.32
(b) C20 H3203. Calculated. C 74.96, H 10.07 C21H3403. ” “75.40, ” 10.25 Found. ” 74.91, ” 10.29
Lack of material prevented the analysis of a more rigorously dried speci- men. Since the hydrogen percentage in contrast to the carbon percent- age of the compound is essentially independent of the state of hydration, it is under the circumstances probably a more reliable indicator of the composition of the anhydrous steroid.
Degradation of Isolated 45-Pregnenetriol-30,17a,20a (II)-2.6 mg. of
per cent ethanol, the infra-red spectra in 1 per cent solutions in carbon disulfide. The spectrometer and cell were as previously described (2) except that the resolu- tion has been raised to its maximal value.
compound II in 1.2 cc. of methanol were treated with 0.15 cc. of a 5 per cent aqueous solution of periodic acid. After the mixture had stood at room temperature for 18 hours, it was distributed between ether and water. The ether layer was washed with a sodium carbonate solution and with water and evaporated. A solution of the residue in carbon di- sulfide showed no significant difference in its infra-red transmission spec- trum from that of dehydroisoandrosterone. Treatment of a pyridine solution (1 cc.) of the oxidation product with 0.04 cc. of benzoyl chloride for 17 hours at room temperature yielded 3.4 mg. of benzoate which was recrystallized three times from acetone. It melted at 254-258° with de- composition and at 254.5-258.5° when mixed with authentic dehydroiso- androsterone benzoate melting at 254.5-257º.
Reduction of 45-Pregnenediol-30,17a-one-20 Acetate (IV)-Compound IV was isolated from the tumor urine extracts by the old procedure (18) except that a 1:1 mixture of magnesium silicate and celite (30) was sub- stituted for alumina in chromatographing. A solution of 31.9 mg. of ace- tate IV in 25 cc. of dry ether was added to a stirred solution of 67 mg. of lithium aluminum hydride in 3 cc. of ether. Stirring under anhydrous conditions was continued for 30 minutes, when the excess reagent was decomposed by the addition of an ether-alcohol mixture. The reaction product was distributed between ether and dilute hydrochloric acid. The ether phase was washed and yielded 28.7 mg. of dry residue which was acetylated by treatment with 2 cc. of pyridine and 1 cc. of acetic anhydride at room temperature for 17 hours. The crude acetates (34.1 mg.) were isolated in the usual manner (2) and proved to be free of ke- tonic starting material (IV), as they gave no color with alkaline m-dini- trobenzene (18). Two recrystallizations from methanol furnished 13.9 mg. (39 per cent) of 45-pregnenetriol-36,17a,20a diacetate (I) which crystallized in elongated plates melting at 202-206°. Continued recrys- tallization raised the melting point to 204-206°. There was no depres- sion on admixture of the isolated diacetate (m.p. 203-206°).
Analysis-C25H3805. Calculated, C 71.74, H 9.15; found, C 71.70, H 9.21 Rotation-[a]”’ = - 81º (c = 0.42)
The infra-red spectrum is given in Curve A of Fig. 1.
The first mother liquor obtained in the preparation of diacetate I was chromatographed on alumina. The earlier eluates obtained with a 1:1 mixture of petroleum ether and benzene and with benzene (and the cor- responding fractions obtained by rechromatographing later eluates) were crystallized from dilute ethanol and from petroleum ether-acetone and gave 7.0 mg. (20 per cent) of needles which showed incomplete fusion at 153-154° and melted after partial resolidification at 158°. The 45-preg-
nenetriol-36,17a,200 3,20-diacetate (V) was dried for analysis, rotation, and spectrographic measurements at 80° in vacuo.
Analysis-C25HI3g05. Calculated, C 71.74, H 9.15; found, C 71.82, H 9.19 Rotation-lal” = - 27° (c = 0.54)
The melting points recorded in the literature are 152-153º (uncorrected) (14), 159-160° (corrected) “nach leichtem Sintern bei 155°” (21) and 151- 152° (17). The published rotations [x] = - 36° (alcohol) (14) and -34.7° (acetone) (21), while in good agreement with each other, show rather large differences from the rotations (8, 31, 9) of the reduction prod- uct, allopregnanetriol-33, 17a,200 3,20-diacetate. The value found by us for the 45 compound is more in harmony with the standard value for this transformation (A[M ], 5- allo - 45 for 33-acetates in alcohol or acetone) and therefore may not be in error. The main maxima in the accessible finger-print region of infra-red radiation are at 13.87, 12.495, 12.32, 12.03, 11.35, 11.075, 10.66, 10.32, 10.21, 10.08, 9.67, 9.31, 8.79, 8.59, 8.31, 8.035, 7.78, 7.59, and 7.28 p.
45-Pregnenetriol-38,17a,20a (II)-16.2 mg. of synthetic diacetate (I), 32 mg. of sodium hydroxide, and 5 cc. of 80 per cent ethanol were heated under a reflux for 1 hour. The resulting triol (12.3 mg.) was isolated by ether extraction and crystallized from 95 per cent ethanol in elongated plates which melted at 221-224° with some browning. Admixture of Compound C caused no depression. The dried analytical specimen was analyzed without taking precautions against the uptake of water.
Analysis-C21 II3403 .¿ II20. Calculated, C 73.43, H 10.27; found, C 73.41, H 10.35
45-Pregnenetriol-38,17a, 200 (VI)-6.8 mg. of diacetate V were hy- drolyzed as described above. The free triol crystallized from acetone in heavy blocks that melted at 224-228° with decomposition. Butenandt et al. (14) reported a melting point of 227° (uncorrected) for their prep- aration.
Degradation of 45-Pregnenetriol-33,17a,20a (Synthetic) (II) and of 45- Pregnenetriol-36,17a,200 (VI)-6.5 mg. of triol II and 2.0 mg. of triol VI were oxidized separately with periodic acid as described above. The product of each reaction when dissolved in carbon disulfide gave an infra- red spectrum in good accord with that of authentic dehydroisoandros- terone. The reaction product derived from triol II was benzoylated and gave a derivative melting at 254-258° with decomposition and at 254- 258° in admixture with dehydroisoandrosterone benzoate. It was dried for analysis at 100° in vacuo.
Analysis-C26H32O3. Calculated, C 79.55, H 8.22; found, C 79.54, H 8.29
SUMMARY
Compound C isolated from the non-ketonic digitonin-precipitable frac- tion obtained from the urine of a boy with an adrenocortical tumor has the structure of A5-pregnenetriol-33, 17a,20a. It has been degraded with periodic acid to dehydroisoandrosterone and synthesized from 45-preg- nenediol-30,17a-one-20 acetate with lithium aluminum hydride. The known 45-pregnenetriol-38,17a,200 was also obtained, though in lesser amounts, from this reduction. The isolated triol is believed to be a metabolite of 45-pregnenediol-36,17a-one-20. A brief account is given of the compounds isolated from the non-ketonic fraction which did not precipitate with digitonin.
The authors wish to express their sincere thanks to Dr. E. W. D. Huff- man, Denver, Colorado, for the microanalyses reported in this paper.
BIBLIOGRAPHY
1. Hirschmann, H., and Hirschmann, F. B., J. Biol. Chem., 157, 601 (1945).
2. Hirschmann, H., and Hirschmann, F. B., J. Biol. Chem., 184, 259 (1950).
3. Shoppee, C. W., J. Chem. Soc., 1138, 1147 (1946).
4. Goldberg, M. W., Sice, J., Robert, H., and Plattner, P. A., Helv. chim. acta, 30, 1441 (1947).
5. Klyne, W., and Barton, D. H. R., J. Am. Chem. Soc., 71, 1500 (1949).
6. Fieser, L. F., and Fieser, M., Experientia, 4, 285 (1948).
7. von Euw, J., and Reichstein, T., Helv. chim. acta, 25, 988 (1942).
8. Steiger, M., and Reichstein, T., Helv. chim. acta, 21, 546 (1938).
9. Plattner, P. A., Heusser, H., and Feurer, M., Helv. chim. acta, 31, 2210 (1948).
10. Barton, D. H. R., J. Chem. Soc., 1116 (1946).
11. Barton, D. H. R., and Rosenfelder, W. J., Nature, 164, 316 (1949); J. Chem. Soc., 2459 (1949).
12. Wintersteiner, O., and Moore, M., Nature, 164, 317 (1949); J. Am. Chem. Soc., 72, 1923 (1950).
13. Jones, R. N., Williams, V. Z., Whalen, M. J., and Dobriner, K., J. Am. Chem. Soc., 70, 2024 (1948).
14. Butenandt, A., Schmidt-Thomé, J., and Paul, H., Ber. chem. Ges., 72, 1112 (1939).
15. Salamon, I., Helv. chim. acta, 32, 1306 (1949).
16. Shoppee, C. W., J. Chem. Soc., 1671 (1949).
17. Fieser, L. F., and Huang-Minlon, J. Am. Chem. Soc., 71, 1840 (1949).
18. Hirschmann, H., and Hirschmann, F. B., J. Biol. Chem., 167, 7 (1947).
19. Fuchs, H. G., and Reichstein, T., Helv. chim. acta, 24, 804 (1941). Hegner, P., and Reichstein, T., Helv. chim. acta, 24, 828 (1941).
20. von Euw, J., and Reichstein, T., Helv. chim. acta, 30, 205 (1947).
21. Prins, D. A., and Reichstein, T., Helv. chim. acta, 23, 1490 (1940).
22. Hirschmann, H., J. Biol. Chem., 150, 363 (1943).
23. Marrian, G. F., and Butler, G. C., Biochem. J., 38, 322 (1944).
24. Buxton, C. L., and Westphal, U., Proc. Soc. Exp. Biol. and Med., 41, 284 (1939). Venning, E. H., and Browne, J. S. L., Endocrinology, 27, 707 (1940).
25. Pearlman, W. H., and Pincus, G., Federation Proc., 5, 79 (1946).
26. Horwitt, B. N., Dorfman, R. I., Shipley, R. A., and Fish, W. R., J. Biol. Chem., 155, 213 (1944).
27. Mason, H. L., J. Biol. Chem., 172, 783 (1948). Sarett, L. H., J. Am. Chem. Soc., 71, 1175 (1949).
28. Hirschmann, H., Federation Proc., 5, 137 (1946).
29. Mason, H. L., in Pincus, G., Recent progress in hormone research; The proceed- ings of the Laurentian Hormone Conference, New York, 3, 103 (1948).
30. Lieberman, S., and Dobriner, K., J. Biol. Chem., 161, 269 (1945).
31. Sutter, M., Meystre, C., and Reichstein, T., Helv. chim. acta, 22, 618 (1939).