CA1292467C - Estradienolone derivatives in sustaining pregnancy - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE

A newly isolated steroid having the following formula:

Description

~2i~

BACI~;RQU~QF TH~lNV~NTlQ~I
Pregnancy is a complex mechanism, many aspects of which are poorly understood including premature delivery, fetal development and the mechanism of parturition. Previously it was shown by one of us that the total androgen level when measured by competitive binding to sex hormone-binding globulin (SHBG) in pregnancy serum was ten to fifteen times higher than in non-pregnancy serum (Murphy, BEP: Recent Progress in Hormone Research 25:563,1969). While the total SHBG-bound activity in the serum of non-pregnant women could be accounted for by steroids known to bind to SHBG, these steroids accounted for only approximately half of the SHBG-bound material measured in the serum of pregnant women. These results are given in Table 1.

Comparison of SHBG-bound activity (Testosterone equivalents) with known levels of steroids in the plasma of non-pregnant and pregnant women Steroids in plasma Relative Plasma concentration - ng/ml known to bind to binding SHBG Non-pregnant women Pregnant women . _ Reported T Reported T
levelsequivalents levels equivalents . _ . . .. .

2 0 Testosterone 100 0.43 0.43 1.2 1.2 Androstenedione 1.4 1.8 0.025 2.49 0.035 Dehydroepiandrosterone 2.5 5.0 0.125 3.6 0.09 Dihydrotestosterone 300 0.15 0.45 0.18 0.54 5-Androstene 3~,1713-diol 130 0.68 0.85 0.65 0.85 5alpha-Androstane-3alpha, 200 0.114 0.228 0.199 0.398 17~-diol 5alpha-Androstan~- 250 0.515 1.28 0.704 1.76 313,1713-diol Estradiol 60 0.08 0.05 16 9.6 . _ Total calculated 3.4 14.5 Total measured 2.0 32.5 L6~

This discrepancy sug~ested that substances that are still un-characterized might be bound to SHBG is pregnancy serum. Since available evidence indicates that there is a positive relationship between the affinity of steroids to specific steroid-binding plasma proteins such as SHBG and CBG and the biologicai potency of those steroids, il was reasoned that the unknown material bound to SHBG in pregnancy serum probably plays a significant physlological role in pregnancy.
The cause of premature deiivery is unknown. Prematurity is associated with higher risks for the newborn, in terms of infection, temperature 10 regulation, hyperbilirubinemia, as well as problems associated with incomplete maturation of the lungs, central nervous system, gastrointestinal tract and kidneys.
Premature delivery can in some instances be delayed by bed rest and certain medications, but those available have various undesirable side effects.
Therefore, medication preventing premature birth with minimal side effects would be highly desirable.
SUMMARY OF THE INVENTION
In accordance with the present invention, there is provided a newly isolated steroid found mainly in the serum and urine of pregnant women and in 2 O the placenta, said steroid having the following formula:

OH

'0~

This compound as well as its functional derivatives is believed to play a role in ~he maintenance of pregnancy, fetal deveiopment and in lhe mechanism of parturition. Isolation and characterization techniques are detailec5 in the following description.
IN TH~I)RAWINGS
Figure I represents the chromatographic elution pattern of SHBG-bound material obtained during preliminary studies.
Figure 2 represents the more detailed chromatographic elution pattern of SHBG-bound material at various gestational stages.
Figure 3 represents the concentration of peak 2 (17B-hydroxy-1 ,5-estradiene-3-one) in maternal serum at various gestational stages.
Figure 4 represents the co-elution of 3H-5alpha-pregnane-3,20-dione, 3H-progesterone and 2-methoxyestrone with peaks 1 a, ~i b and 3 respectively on HPLC as well as the elution volume of peak 2 (1713-hydroxy-1 ,5-estradiene 3-one).
Figure 5 r0presents the chemical structures of 5alpha-pregnane-3,20-dione methyloxirne, progesterone methyloxlme, 1 9-nor-testosterone methyloxime trimethylsilyl ether and 2-methoxyestrone methyloxime trimethylsilyl ether.
Figure 6 represents th0 mass spectra of 5alpha-pregnane-3,20-dione methyloxime and of the methyloxlme derivative of the isolated steroid corresponding to chromatographic peak 1 a.
Figure 7 represents the mass spectra of progesterone methyloxime and of the methyloxime derivative of the isolated steroid corresponding to chromatographic peak 1 b.

Z~6~7 Figure 8 represents the mass spectra of ~ methoxyestrorle methyloxime trimethylsilyl ether and of the methyloxime trimethylsilyi eth~r derivative of the isolated steroid corresponding to chromatographic peak 3.
Figure 9 represents the mass spectra of 1 9-nor-testosterone methyloxime trimethylsilyl ether and of the methyloxime trimethylsilyl ether derivative of the isolated steroid corresponding to chromatographic peak 2.
Figure 10 represents the binding of the compound corresponding to peak 2 to SHBG under physiological conditions (i.e. equilibrium dialysis of undiluted serum at 37C).
Figure 11 represents the ultraviolet absorplion spectrum in ethanol of peak 2 (1 7B-hydroxy-1 ,5-estradiene-3-one).
Figure 12 represents the elution pattern of a purified sample of the compound corresponding to peak 2 before and after exposure to 0.1 N NaOH.
DETAlLi--D DL~PTION OF THE INVENliQ~I
In preliminary studies of the chromatographic elution patterns (Sephadex'B' LH-Z0 column 31 x 0.8 cm; solvent system: - methylene chloride:
heptane:methanol 50:50:1) of serum of pregnant women, it was observed that a large amount of SHBG-bound material eluted in the region of 3H-androstenedione (Figure 1). This material, which was barely detectable in non-pregnancy serum, 20 could not be accounted for by steroids which were already known to bind to SHBG.
Androstenedione could not account for this material because the levels of androstenedione are low during pregnancy, and since the relative binding affinity to SHBG is only 1% that of testosterone. All recognized ligands of SHBG, such as ~alpha-dihydrotestosterone, testosterone, androstanediols, androstenediols and estradiol, could not be considered since they all possess longer retention times.
When mor0 detailed studies were conducted, it was found that the ~292~L67 chromatographic pattern (Shephadex~ LH-20 column 60 x 0.9 cm; solvent system: - methylene chloride:heptane:methanol 50:50:1) of the above-mentioned unidentified material resulted in four major peaks that were arbitrarily designated as peaks 1 a, 1 b, 2 and 3 (Figure 2). The first two peaks,1 a and 1 b, eluted earlier than androstenedione (dotted lines in Figure 2) while the last two, peaks 2 and 3, eluted later.
Gas chromatography-mass spectrometric analysis, in addition to data regarding the elution volumes, binding characteristics, serum levels, UV
absorption, solubility data and chemical conversion, led to the conclusion that the 10 four peaks corresponded to two weakly bound substances, namely 5alpha-pregnane-3,20-dione (peak 1 a) and progesterone (peak 1 b), and two strongly bound substances, 2-methoxyestrone (peak 3) and a new steroid, 1 713-hydroxy-1,5-estradiene-3-one (peak 2).
GENERAL QBSERVATIONS RELATED TO 171~-HYDROXY-1 .5-ESTRADIENE-3-ONE
(ESTRADIENOLONE) The leve!s of estradienolone (peak 2) and the other peaks are shown in Table 2. The levels were maasured in terms of testosterone equivalents (TE) since ~he identities of the compounds were initially unknown. In other words, the concentration corresponding to each peak was expressed as the amount of 2 0 non-radioactive testosterone required to displace the same amount of tritiated testosterone as the unknown material. Thus the concentration of the unknown substances measured as TE will depend on the affinity of that substance for SHBG
as well as the amount present in the sample. Since any substance capable of displacing 3H testosterone from its specific binding sites on SHB(3 has been measured, the material eluting as each peak may correspond to either a substance which has a high affinity for SHBG or to a substance which has only a low affinity 2~

but which is present in large amounts.
TA~
Levels of unknown SHBG ligands according to labour and gestational age (mean ~ SD) Concentration - ng/ml or ng/gm, Testosterone Equivalents Compartment Gestational Type age peak 2 peak (1 a + 1 b) peak 3 (weeks) -Maternal12-24 Nolabour3.5i4.4(12)~ O.3(11) 1.6i1.0(11) serum l O 30 -38 All 2.9 i 2.2 (17) 2.5 i 1.2 (14) 3.3 + 2.2 (9) No labour~ r3.6 i 1.5 (12) 2.6 i 1.0 (10) 3.3 + 1.9 (6) Spont.I Lo.8i 05 (5) 2.4~1.8(4) 3.3~2.0 (3) labour ~ , 39 - 42 All 1.8 i 0.6 (12) 2.1 i 1.2 (9) 3.2 i 2.0 (9) No labour1.8 i 0.7 (4) NA~4 NA
Spont.1.1 iO.1 (3) NA NA
labour . . . _ _ _ Cord Serum 22 None 1.5 (1) 1.7 (1) 1.8 (1) 38-42-o All UDa (6) 2.7 i 1.2 (6) 1.3 ~ 1.0 (1 ) No labourUD (1) 1.6 (1) 0.8 (1) Spont. orUD (5) 3.0 ~ 1.2 (5) 1.4 i 0 7 (5) ind. Iabour Placenta 14 No labour29 (1) 63 (1) 54 (1) 16 No labour82 (2) 55 (2) 41(2) 36 No labour3.6 (1) 58 (1) 13 (1) 39 - 42 All UD (3) 45 i 5.2 (3) UD (3) No labourUD (1) 44 (1) UD (1) Spont. UD (1) 41(1) UD (1) labour Ind. Iabour UD (1) 51 (1) UD (-I) P ~0.05 P ~ 0.01 Numbers in parenteses denote number of samples analysed Not analysed due to insufficient number of samples Undetectable ~- Pooled serum samples were used The novel steroid may be isolated from urine and peripheral venous blood of pregnant women at 12-42 weeks gestation as well as from placenta.
Chromatographic analysis of maternal serum at various stages of pregnancy demonstrated that the mean (+ S.D.) concentration of estradienolone remained high through pregnancy, then decreased significantly close to term (Figure 3). Thus, at 12 to 24 weeks gestation, a mean concentration of 3.5 i 4-4ng per ml TE was observed while a concentration of 3.6 i 1.5 ng per ml TE was observed at 30-38 weeks gestation in the absence of labour. In the last weeks ofgestation (39-41 weeks) a significant decrease in the mean concentration of estradienolone (1.8 ~ 0.7 ng per ml TE) was observed in the absence of labour.
In mothers who were in premature spontaneous-onset labour, (30-38 weeks gestation) the concentrations of estradienolone were significantly lower (0.8 i 0.5 ng per ml TE) than those of mothers of the same gestational agenot in labour.
Thus a lower concentration of estradienolone was associated with premature labour and pregnancy at term (impending labour). Although there were considerable variations in concentration arnong individuals in the early stages of pregnancy, (0.4 to 14 ng per ml TE at 12 to 24 weeks gestation), thesevariations were less at 30 38 and 39-~2 weeks ~estation.
Urlne analyses performed at week 33 of pregnancy demonstrated that rnost of the estradienolone was excreted from the body mainly in the sulfated form.
Studies of fetal tissues showed that estradienolone is not derivéd from the fetus itself. In fact, analysis of the Ihree main steroid-producing fetal tissues (adrenal, ovary, testis) and the steroid-metabolizing tissue (liver), obtained at 16-18 weeks gestation showed little or no trace of the novel compound.

However the concentration of estradienolone in placenta was rnuch higher than that in maternal serum, suggesting that this compound is of placental origin. This compound was undetectable in term placenta and very low in premature placenta. These observations imply that toward term, the placenta metabolizes estradienolone more quickly or loses its ability to synthesize it thereby suggesting an involvement in the initiation of labour.
ISOLATION QF THE 1 7~1Yi~ROXY-1 ~5-ESTRA~ lE-~-ONi~
Extraction and fractionation of various sample~
Blood samples were collected from pregnant women at different 10 gestational ages and centrifuged at 4 QC and the sera stored at -20QC until analysed. Placental samples obtained at therapeutic abortion and at delivery were stored at -20C until analysed. Urine samples (24 hour) wera collected from pregnant women at 28-38 weeks gestation and were stored at -20QC until analysis. All pregnant patients from whom the samples ware collected were healthy women attending their local hospital clinic.
To a suitable aliquot of serum, urine or minced tissue, small amounts (4-8,000 cpm each) of 3H-androstenedione and 3H-lestosterone were added as radioactive markers for the determination of the elution pattern and to calculate recoveries. The sample was then extracted 3 times with 4 volumes of ethyl 20 acetate and the combined extracts were evaporated to dryness under air.
Fractionations were done routinely using 2 successive chro-matographies i.e. an initial purification on a short Sephadex~ LH-20 coiumn (10 x 0.9 cm), followed by a more detailed procedure on a long Sephadexg' LH-20 column (60 x 0.9 cm). The solvent system used was methylene chioride:heptane:methanol 50:50:1. For the initial purification, ten fractions of 4 mL were collected to 40 mL or untii the 3H-tPstosterone peak was eluted. More ~Z~ 67 polar steroids were retained in the column. An aliquot from each fraction was counted for radioactivity to locate the tracer peaks. The pooled fractions were then subjected to more detailed chromatography; 1 mL fractions were collected up to 50 mL and suitable aliquots were taken from each fraction for counting to locate the radioactive steroid peaks and for assay to detect SHBG-bound material.
ss~ procedure used for the detection and quantitation of estradienolane Substances bound to SHBG were detected and quantitated using a radiotransinassay (i.e. a competitive protein-binding assay employing an indigenous protein) as described by Murphy in Recent Prog. Horm. Res. 25:563, 10 1969. Briefly, the binding protein was human late pregnancy serum diluted 1:100 in phosphate-gelatin buffer (0.2M phosphate buffer, pH 6.5; 0.5 mg/mL
gelatin), the tracer was 3H-testosterone, and the adsorbent was dextran-coated charcoal suspension (0.125 g charcoal, 12.5 mg dextran in 100 mL 0.2M
phosphate buffer, pH 6.5). Since non-radioactive testosterone was used as the standard, the results were expressed as testosterone equivalents.
For the preparation of the protein-tracer solution, 0.1 mL of late pregnancy serum was transferred into a flask containing 3H-testosterone (12,000 cpm or 3 ~LCi). The volume was made up to 10 mL with 0.2M of the phosphate-gelatin buffer described above. This solulion was prepared 20 immediately before each assay.
0.1 mL of protein-tracer solution was added to duplicate tubes containing the unknown samples and to those containing increasing amounts of testosterone (0,50,100,400,800,1200 pg) to serve as standards. The tubes were incubated at 45QC for 5 min, then at 4QC for 30 min, after which 1.0 mL of dextran-charcoal suspension was added. After 5 min, the rack containing the tubes was shaken on an automatic shaker for 1 min, the tubes centrifuged at 4QC, Z~67 2000 rpm for 5 min and returned to the 4g bath. Using a diluter containing Econofluor~, th0 supernatant 0.5 mL plus 2 mL Econofluor~g was pipelted into 5 mL counting vials. After mixing, the vials were counted to 10,000 counts or 10 min. Since the time of contact with the adsorbent is critical, care was taken to keep the procedure as uniform as possible; each batch consisted of a standard curve plus 24 sample tubes.
ANALYSIS BY (~;AS CHROMATOGRAPHY-MASS SPECTROMETRY (GC-MS) Although further purification of the unknown SHBG-bound material was attempted from maternal serum and placenta, GC-MS analysis showed that l0 they were heavily contaminated with lipids. Since this was difficult to remove, the material purified from hydrolysed pregnancy urine was used.
a) Purification of the urine samples Four 24-h urine samples were collected from different women at 28-38 weeks gestation and frozen until digestion. One liter of urine was taken from each of the four collections and the pH was adjusted to 5.2 with 25% acetic acid. To each lit0r of urine 4S ml 2M acetate buffer, pH 5.2, 40 ml Glusulase~
and a few drops of chloroform were added. (Glusulase~9, an enzyme preparation of the intestinal juice of the snail Helix pomatia, approximately 10,000 units sulphatase and 90,000 units l~-D-glucuronidase activ7ty per mL, was obtained 20 from Dupont Pharmaceuticals, Wilmlngton, Delaware). The flasks were mixed well and incubated at 38C for 50 h. To the resulting digested urine samples containing the combined unconjugated, glucuronide and sulfate fractions, 2 x 1 o6 cpm of 3H-estrone was added. The urine was then extracted using reversed phase C18 cartridges (9 mm x 10 mm id; Sep-Pak~9, Waters Associates Inc, Mississauga, Ontario) as described by Heikkinen et al in Clin. Chem. 27:1186, 1981. Seventy ml urine was passed through each column. The methanol eluates, ~ ~2~7 containing more than 90% of the added 3i-l-estrone, were evaporated to dryness using a rotor evaporator.
To the pooled dry urine extract,1 x 106 cpm of 3H-androstenedione was added and the extract was purified on Sephadex~ LH-20 columns. Fractions corresponding to peaks (1 a,1 b, 2 and 3) were then purified by high performance liquid chromatography (HPLC) using gradient elution.
The HPLC column used (Econosphere~ silica, 5 jl particle size, 25 cm long x 4.6 mm id) was purchased from Alltech Associates Inc., Deerfield, IL.
The remaining components of the HPLC system were designed by Gilson Medical 10 Electronics (Mandel Scientific Co., Rockwood, Ontario). The set-up consisted of two model 302 single piston pumps, model 811 dynamic mixer, model 7125 Rheodyne injection valve, model 111 -HPLC detector, a model 202 fraction collector and an Apple ll plus gradient controller.
All analyses were performed at room temperature with a flow rate of 1.0 mUmin. For injection, samples were dissolved in 35 ,uL methylene chloride and 25 ~lL of the resulting solution was used for injection. Tracers of trltiated 5alpha-pregnane-3,20-dione and progesterone were added to the samples before injection to serv0 as markers.
Using a gradient of 2% to 30% B (98% to 70% A) (where A was 2 o methylene chloride and B was hexane:e~hanol 90:10),1 mL fractions were collected for 60 min. One aliquot of each fraction was counted for radioactivity while another was assayed for SHBG-bound material. Fractions corresponding to each SHBG-bound peak obtained by the HPLC of peak 1a, peak lb, peak 2 and peak 3 were subjected to GC-MS. The elution profiles on HPLC of 3H-5alpha-pregnane-3,20-dione, 3H-progesterone and non-radioactive 2-methoxyestrone are compared with those of isolated peaks in Fig. 4.

b) GC-MS analysis of the isolated samples The GC MS analysis procedure used was that dascribed by Pike et al (J. Chromalogr. 306:39,1984). The samples were converted to their methyloxime, trimethylsilyl (TMS) ether derivatives. The derivatized samples were purified on a Lipidex'~ 000 column, diluted to ~i 00 u L, and 1 -1.5 u L was injected. Four hydrocarbon standards (C-24,28,32,36) were added to each sample to help determine the methylene unit value of each peak.
The analyses were performed on a Hewlett-Packard 7620A gas chromatograph fitted with a modified Van den Berg injector. The GC column consisted of 30 m x 0.2 mm, film thickness 0.25 llm, DBI fused silica open tubular column (J and W Scientific, Cordova, CA). The velocity of ~he carrier gas (helium) was 40 cm/sec. The injection block and transfer line temperature were set at 300Q C. This was interfaced to a vacuum generator MM-1 6 low resolution magnetic sector instrument. The ionizing energy was 70 eV and the ion source temperature was 200QC for 4 min, then increased to a final temperature of 300QC at a rate of 4QC per min.
Contaminants identified in the sample preparations could be readily ruled out since their affinity to SHBG and their concentrations in maternal serum were too low to account for the SHBG-bound peaks. Their presence can be 20 explalned by the expected high levels of pregnane metabolites in hydrolysed pregnancy urine.
On the basis of both GC elution and MS data, the identities of peaks 1 a, 1 b and 3 (whose tentative identities were established by elution volume on Sephadex~9 LH-20 and HPLC, binding characteristics and ma~ernal serum levels) were confirmed as 5alpha-pregnane-3,20-dione, progesterone and 2-methoxy-estrone respectively.

The chemical structures of the derivatives and the origin of some major ions which help explain the rnass spectra are shown in Fig. 5. The rnass spectra obtained for the methyloxime derivatives of standard ~alpha-pregnane-3,20-dione and the steroid isolated from peak 1 a are shown in Fig. 6.
Th0 mass spectrum of the sampl0 shows some background ions, nnost probably originating from silicone grease which is a common contaminant. The mass spectra obtained for the methyloxime derivative of progesterone and of the s1eroid corresponding to peak 1 b are shown in Fig. 7. The major component was identified as progesterone, although there were some low mass ions present which lO suggest the presence in low concentrations of a 1 7-hydroxy C21 -steroid. Since its levels were very low and since the affinity of known C21 steroids for SHBG are low, it is very unlikely to measure as SHBG-bound material. The compound represented by peak 3 was identified as 2-methoxyestrons. In Fig. ~, the mass spectra of methyloxime-trimethylsilyl ether of standard 2-methoxyestrone and of the steroid isolated from peak 3 are compared.
Finally, the compound represented by peak 2 was identified as estradienolone (17epsilon-hydroxy- estradiene-3-one). The mass spectra for the methyloxime-trimethylsilyl elher of 1 9-nor-testosterone and of the estradienolone isolated from peak 2 are shown in Fig. 9. The mass spectrum of the 20 19-nor-testosterone, al~hough it showed some similarity, was not id0ntical to that of the isolated compound, estradienolone. However, no other standard having a more similar mass spectrum was found in a large collection of steroid reference spectra.

The facts that peak 2 was extractable with organic solvents, had chromatographic properties similar to those of several classical steroids such as progesterone and androstenedione, was dialyzable, was bindable to SHBG, and was heat stable strongly suggested that the material eluting as peak 2 is steroidal in nature.
The non-polar character of the steroid was suggested by its relatively low elution volume on Sephadex~ LH-20 and silica gel HPLC.
Although all the known metabolites of androgens, estrogens and 10 progesterone were analyzed for their relative elution volumes, binding characteristics and serum levels, none of them could account for peak 2. Thus the material eluting as peak 2 is a substance which has not been isolated previously.
The strong binding of peak 2 to SHBG under physiological conditions (i.e. equilibrium dialysis of undiluted serum at 37C) was shown by the preferential accumulation of peak 2 with raspect to testosterone in the compartment containing native SHBG as compared with the compartment containing denatured SHBG (Figure 10).
Since peak 2 binds poorly to human corticosteroid-blnding globulin, guinea pig progesterone-binding globulin, but strongly to SHBG, it is very likely 20 that it is structurally related to other SHBG-bound steroids, such as testosterone and estradiol.
Peak 2 is not a lipoidal derivative of a more polar steroid, as shown by a similar elution pattern of peak 2 before and after saponification of pregnancy serum. Thus it is most likely an unconjugated steroid of low polarity.
In addition, peak 2 behaved like an estrogen, as suggested by its solubility in alkali when it was partitioned between methylene chloride and 0.1 N

2~67 NaOH. Serum, urine and placental peak 2 were soluble in 0.1N NaOH, which strengthens the argument that the material eluting as peak 2 is tha same in all 3 compartments. The solubility of this non polar steroid in 0.1 N NaOH may be explained by a rearrangement of the steroid to form a steroid with a phenolic ring A, i.e. estradiol.
The GC-MS analysis of peak 2 showed that its structure is 17epsilon-hydroxyestradiene-3-one. Proof of the complete structure was not possible by GC-MS since a reference steroid was not available.
Since a 17alpha-hydroxy structure would bind only weakly to SHBG, l0 and since peak 2 binds strongly to SHBG, the configuration at C17 must be B.
The ultra-violet absorption spectrum of peak 2 dissolved in ethanol showed an absorption maximum at 232 nm (Fig. 11), which indicates that it has a 1-ene-3-one structure, i.e. a double bond at C1-2.
Since a 1,4-diene structur0 would actually be estradiol and since a synthetic compound of similar structure, 5-estrene-3B,1 7B-diol, has been shown to bind strongly to SHBG, it was concluded that the most likely position for the second double bond is at C5-6. Also, this neutral structure has the same number of carbon, hydrogen, and oxygen atoms as estradiol and would be expected to rearrange in 0.1 N NaOH to form estradiol, which would account for the 20 solubili~y of peak 2 in NaOH.
On the basis of the above properties of peak 2, the following compiete structure was deduced:

z~

OH
0~
1 7B- Hydroxy-1 ,5(6)-estradiene-3-one Further confirmation of this structure was obtained by re-arrangement of peak 2 after exposure to 0.1 N NaOH to give estradiol (Fig. 12).
The identity of the rearranged product as estradiol was affirmed by its co-elution with 3H-estradiol, its immunoreactivity to a specific antibody raised against estradiol and by its specific binding to SHBG.
The use of functional 1713-hydroxy-1,5-estradiene-3-one C-17 derivatives as pregnancy-sustaining agents and including hydroxy caproates, 10 acetates, ethynyls, propionates, ethanates, cytraonates, methyls, decanoates, phenpropionates and 9-alpha-fluoros, also falls within the scope of the present invention .
Also within the scope of the present invention is a pharmaceutical composition for sustaining pregnancy in women showing signs of early labor.
This composition comprises a 1 7~-Hydroxy- l ,5(6)-estradiene-3-one or any functional derivalive thereof in association with a suitable pharmaceutical carrier.

Claims (2)

1. In substantially pure form, a compound having the following formula:

2. A composition for sustaining pregnancy in women showing signs of early labor, said composition comprising a compound having the following formula:

or any functional derivative thereof in association with a suitable pharmaceutical carrier.