| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Studies |
Department of Oncology, Haukeland University Hospital (J.G., H.B., P.E.L.), N-5021 Bergen; and the Department of Oncology, University Hospital Trondheim (S.L.), N-7006 Trondheim, Norway; and Chiroscience Ltd. (J.L.G.), Cambridge, United Kingdom
Address all correspondence and requests for reprints to: P. E. Lønning, Department of Oncology, Haukeland University Hospital, Department of Oncology, N-5021 Bergen, Norway.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Commercial AG (Orimeten, Novartis, Basel, Switzerland)
exists as a racemate (DL-AG; Fig. 1
). Previous studies suggest considerable
differences between the two enantiomers of AG in their affinities for
different enzymes of the P-450 superfamily. In vitro
observations showed dexaminoglutethimide (D-AG) to be
38-fold more potent in suppressing the aromatization of androstenedione
to estrone (E1) compared to L-AG (6), but to
express about 2.5 times the inhibitory activity of the
L-form with respect to the cholesterol side-chain cleavage
enzyme (7).
|
Whether the enzyme induction and the influence on E1S metabolism are caused by a single or both enantiomers of AG is currently not known. As aromatase inhibition is caused by D-AG alone, a finding that the D-form may be responsible for the induction of E1S metabolism may encourage the development of D-AG as a therapeutic drug. To test this hypo-thesis, we evaluated the influence of D-AG vs. DL-AG on estrogen disposition in 12 postmenopausal women suffering from metastatic breast cancer in a blinded, cross-over design study. Plasma levels of E1, estradiol (E2), and E1S were measured by highly sensitive RIAs, and the clearance rate of E1 and E1S were determined after tracer injections of [3H]E1S and [14C]E1 before and after 4 weeks of treatment with each drug regimen.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The study was designed to enroll a sufficient number of patients
to allow statistical comparisons in a group of at least 12 patients.
Thirteen postmenopausal women considered for treatment with an
aromatase inhibitor for stage IV breast cancer were eligible for this
study. Tumor characteristics and demographic data of the patients are
given in Table 1. Due to intolerable
side-effects (fever and skin rash), the investigators decided to
withdraw 1 patient (no. 12) from the study after 2 weeks of treatment,
leaving 12 patients for statistical analysis. The median age of all
participants was 75 yr (range, 60–81 yr), and the mean weight was 69.5
kg (range, 56–78 kg). All patients gave their written informed consent
before participation, and the study was approved by the regional
ethical committee.
|
The protocol was designed as a double blind, cross-over study. DL-AG and dexaminoglutethimide (D-AG) were supplied by Chiroscience Ltd. (Cambridge, UK). Patients were all randomized to receive either DL-AG (1000 mg daily) or D-AG (500 mg daily) for a period of 4 weeks each and were then crossed over to the alternative treatment regimen for another 4 weeks without any wash-out period. This design was justified by previous work establishing the half-life of AG to be about 7 h in postmenopausal women (12). The treatment was DL-AG (250 mg, four times daily) or D-AG (125 mg, four times daily). On the days before radiotracer injections, the last drug dose was administered at 2200 h, and the first medication during the tracer experiments was given 2 h after the radiotracer injection. Cortisone acetate was administered as 50 mg, two times daily, for the first 2 weeks of treatment, followed by 25 mg, two times daily (13). After the 8-week study period was completed, all patients continued treatment with AG racemate (250 mg, two times daily) and cortisone acetate (25 mg, four times daily) until disease progression occurred.
Sample collection
Fasting blood samples for plasma hormone measurements were obtained at 0800 h before initiation of treatment and on days 28 and 56 during treatment. All samples were collected in sodium-heparinized vials before drug intake and before tracer injections. Plasma was separated by centrifugation and was stored at -20 C until analysis.
Investigational protocol
The steroid injection technique, blood-sampling protocol, and measurement of labeled estrogens in plasma were previously described (10). The pharmacokinetic calculations of the MCRs as well as the production rates of E1 and E1S were performed as described elsewhere (11).
Materials
All solvents were of analytical or spectrophotometrical
grade and were obtained from Merck (Darmstadt, Germany), except for
ethanol, which was obtained from A/S Vinmonopolet (Oslo, Norway).
Radiolabeled [6,7-3H]E1S (40–60 Ci/mmol) for
recovery calculation and for bolus injections was obtained from
DuPont-New England Nuclear (Boston, MA). E1
antiserum (E17-94) was purchased from Endocrine Sciences (Calabasas
Hills, CA), the E2 antibody (ER 150, Sorin Biomedica,
Saluggia, Italy) was obtained from Sodiag (Losone, Switzerland),
and the tracer [2,4,6,7-3H]E1 (101 Ci/mmol)
was purchased from Amersham Life Sciences (Aylesbury, UK).
Estradiol-6-(O-carboxymethyl)oximino-2-(2-[125I]iodohistamine),
(
2000 Ci/mmol) was obtained from Amersham International (Little
Chalfont, UK), Sephadex LH-20 from Pharmacia (Uppsala, Sweden), and
sulfatase (S-9754) from Sigma Chemical Co. (London, UK). Sodium
borohydride was purchased from Fluka Chemie (Buchs, Switzerland).
[4-14C]Estrone (57 Ci/mol) was obtained for the bolus
injections from New England Nuclear Products (Boston, MA).
Estrogen measurements
Plasma estrogen levels were measured by methods reported
previously (14, 15, 16). To measure plasma E1 and
E1S, [3H]E1S (
1000 cpm)
dissolved in methanol was added to test tubes and evaporated to
dryness. Two milliliters of plasma were added, and the samples were
allowed to equilibrate overnight. Free (unconjugated) estrogens were
extracted with ether (three times, 5 mL each time). The ether extracts
were removed, evaporated to dryness, and reconstituted in
dichlormethane-ethyl acetate-methanol (97:5:1, vol/vol/vol) before
purification of the E1 fraction on LH-20 columns (1.6 mL).
The separated fractions for E1 were then evaporated to
dryness and reconstituted in methanol. Estrone was measured by RIA
using the tracer [2,4,6,7-3H]estrone and the antibody
E17-94. The results were corrected for the recovery of E1
(85%) based on repeated analyses in our laboratory (15).
After ether extraction, ethanol (12 mL) was added to the water fraction, and the sample was vortexed and centrifuged for 15 min at 600 x g. The ethanol fraction was removed and dried, and the residue was reconstituted in 2 mL acetate buffer (0.2 mol/L; pH 5) containing sulfatase (S-9754) to a final concentration of 0.2 mg/mL. Hydrolysis was allowed for 48 h at 37 C, followed by extraction of free estrogens and purification of the E1 fraction on an LH-20 column as outlined above. E1 was subsequently converted to E2 with use of sodium borohydride dissolved in 0.05 mol/L NaOH to a final concentration of 1 mg/mL in NaOH-methanol (1:10, vol/vol). After incubation at 37 C for 15 min, the methanol was evaporated, and borohydride was neutralized by adding 0.5 mL acetate buffer (0.2 mol/L; pH 3.0). E2 was extracted by ether and purified on LH-20 columns as outlined above. The E2 fraction was evaporated to dryness and reconstituted in 1 mL methanol. A 500-µL aliquot was obtained for recovery calculation. From the residual 500 µL, 150-µL aliquots were obtained in duplicate, and the E2 concentration was measured by RIA using estradiol-6-(O-carboxymethyl)oximino-2-(2-[125I]iodohistamine) as the tracer and the antibody ER 150 (Sorin).
For the measurement of E2, plasma (0.2 mL) was extracted with 3 mL ether, and E2 concentrations were measured directly by RIA without recovery calculation as described previously (14, 15).
The intraassay coefficient of variation was below 7% for all measurements, and the sensitivity limit for E1, E2, and E1S was 6.3, 2.1, and 2.7 pmol/L, respectively (15, 16).
Measurement of plasma androgens
Plasma levels of androstenedione and testosterone were measured with commercial RIA kits provided by Diagnostic Systems Laboratories (Webster, TX) and Diagnostic Products Corp. (Los Angeles, CA).
Determination of AG in plasma samples
Plasma levels of AG were measured in fasting blood samples
obtained on days 0, 14, 28, 42, and 56. Although the measurements for
the second treatment started only 2 weeks after completing the first
treatment, the half-life of AG (7 h) suggests that all of the drug
administered during the first treatment had been completely cleared
from the body by the time these measurements were made. Heparinized
vials were collected, centrifuged, and the plasma was stored at -20 C
until analysis. All samples from one patient were analyzed in the same
batch. After a protein precipitation, liquid chromatography with mass
spectrometric detection was used to determine plasma AG levels. The
method has been validated over the range 0.25–10.0 µg/mL. Samples
above this concentration had to be diluted into this range. The
analysis has been carried out at Covance Laboratories (Madison,
WI).
Evaluation of side-effects
Side-effects such as sedation, dizziness, nausea, and skin rash were evaluated before commencing therapy and on each visit during treatment (after 2 and 4 weeks on each regimen) by the same clinician.
Statistics
In a previous study, we found plasma estrogen levels to be well fitted to a log-normal distribution (15). Thus, plasma hormone levels as well as clearance and production rates are given as their geometric mean value with 95% confidence intervals of the mean. In addition, values obtained in the three test situations were compared using the Friedman test. If this revealed a significant difference between the test situations, values obtained during treatment with DL-AG and D-AG were compared using the Wilcoxon matched pair signed rank test. All P values are expressed as two-tailed. In cases with hormone levels suppressed below the sensitivity limits of the assays, the value of the sensitivity limit was used for statistical analysis.
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
and Fig. 2. In addition, plasma levels of
E1, E1S, and androstenedione as well as
E1 and E1S plasma clearance and production
rates measured before and during the two treatment options are shown in
Fig. 3.
|
|
|
Plasma levels of E1 were found suppressed from a pretreatment value of 41.7 pmol/L [95% confidence interval (CI), 29.6–58.8) to 7.7 pmol/L (95% CI, 6.1–9.8) and 7.4 pmol/L (95% CI, 6.3–8.6) during treatment with DL-AG and D-AG, corresponding to suppression to 18.6% and 17.7% of pretreatment levels, respectively. Noteworthy, plasma E1 levels were suppressed below the sensitivity limit of the assay in seven patients during treatment with DL-AG and in four patients during D-AG therapy. E2 was equally suppressed by both treatment regimens from a pretreatment value of 15.4 pmol/L (95% CI, 11.4–20.9) to 3.6 pmol/L (95% CI, 3.0–4.3 for both regimens) during treatment (mean suppression to 23.4% of the pretreatment level). Plasma E1S fell from a pretreatment value of 242.7 (95% CI, 132–447) to 45.2 pmol/L (95% CI, 29–70) during treatment with DL-AG and to 36.3 pmol/L (95% CI, 26–52) during D-AG treatment (18.6% and 15.0% of pretreatment levels). No statistically significant difference between estrogen levels obtained during treatment with D-AG and DL-AG was found.
Plasma clearance and production rates of E1 and E1S
Treatment with DL-AG increased the clearance rate of E1S from a mean of 5.9 L/h (95% CI, 3.9–9.0) before treatment to 14.0 L/h (95% CI, 10.0–19.7), whereas D-AG increased the clearance rate of E1S to 10.0 L/h (95% CI, 7.3–13.7; P = 0.001; mean increase, 136.3% and 68.7%, P = 0.001, respectively). The production rate of E1S decreased from a pretreatment value of 1.44 nmol/h (95% CI, 0.94–2.21) to 0.64 nmol/h (95% CI, 0.41–1.02) and 0.36 nmol/h (95% CI, 0.23–0.57) during treatment with DL-AG and D-AG, respectively (P < 0.005; 44.7% and 25.2% of pretreatment levels). The observed differences between the E1S clearance rates and the E1S production rates during treatment with DL-AG and D-AG both reached a level of statistical significance (P < 0.05).
The plasma clearance rate of E1 increased from a
pretreatment value of 44.8 L/h (95% CI, 37.3–53.8) to 58.5 L/h (95%
CI, 50.0–68.4) and 50.6 L/h (95% CI, 42.5–60.1) during treatment
with DL-AG and D-AG, respectively
(P < 0.05; mean increase of 30.5% and 12.9%,
respectively). The production rate of E1 decreased from a
pretreatment value of 1.87 nmol/h (95% CI, 1.18–2.95) to 0.45 nmol/h
(95% CI, 0.32–0.64) during DL-AG treatment and to 0.37
nmol/h (95% CI, 0.29–0.49) during D-AG treatment
(P = 0.001; 24.3% and 20.0% of pretreatment values,
respectively). The observed changes in the clearance and production
rate of E1 (for details see Table 2
) paralleled in general
the findings made for E1S, but the difference between
values obtained during treatment with D-AG and
DL-AG did not reach a level of statistical significance
(P > 0.10 for both).
Plasma androgen levels
Plasma levels of androstenedione and testosterone determined
before and during treatment are given in Table 2. Treatment with
D-AG suppressed plasma androstenedione levels significantly
(to a mean of 1.3 nmol/L), whereas androstenedione levels during
treatment with DL-AG (mean, 2.5 nmol/L) were unchanged
compared with pretreatment values (mean, 2.2 nmol/L; P
< 0.05 comparing the three test situations and P <
0.05 comparing the two on-treatment situations). No changes in plasma
testosterone values were observed during either treatment.
Plasma levels of AG
Morning plasma levels of AG were 6.50 µg/mL for the total racemate (95% CI, 5.16–8.20) and 2.11 µg/mL (95% CI, 1.67–2.68) during treatment with dexaminoglutethimide, respectively (P < 0.005).
Side-effects
Side-effects observed during treatment with DL-AG and
D-AG are reported in Table 3.
In general, side-effects appeared during the first 2 weeks of treatment
with the first drug schedule, with no significant difference in the
frequency of side-effects between D-AG and
DL-AG treatments.
|
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
AG is well recognized for being a nonspecific drug causing many
biochemical effects. It is known as an unselective inhibitor of several
P-450-dependent enzymes involved in steroid formation, such as
desmolase as well as 11ß-, 18-, and (probably) 21-hydroxylase,
requiring glucocorticoid administration in concert (17). The drug
causes several side-effects on the central nervous system,
i.e. dizziness, sedation, and nausea, as well as skin rash
and fever. Whether these effects are caused by one or both of the
enantiomers is not known. AG is also a potent enhancer of P-450 mixed
function hydroxylases (8), and in previous studies we demonstrated
DL-AG to cause an increase in the plasma clearance rate of
E1S by more than 100%, probably by induction of
16
-hydroxylation (10, 18), one of the main metabolic pathways of
estrogens in man (19).
In the present trial, treatment with DL-AG (1000 mg) and D-AG (500 mg) suppressed plasma levels of E1 by at least 80%, whereas plasma E2 levels decreased by 76.6% during both treatment options. This suppression is somewhat better than what has been reported in previous studies (20, 21). Although highly sensitive RIAs were used, we found plasma levels of E1 in several patients to be suppressed below the detection limit of the assay during both treatment options. Thus, the suppression of plasma E1 levels may be underestimated. Plasma levels of E1S decreased during treatment with both regimens by 80–85%. Notably, patients with low pretreatment values of E1S showed a smaller percent suppression than patients with higher pretreatment E1S levels. No significant difference between plasma levels of any of the estrogens during treatment with D-AG and DL-AG was seen.
The plasma clearance rate of E1S increased 2-fold during
treatment with D-AG and 3-fold during treatment with
DL-AG compared to pretreatment values. Meanwhile, the
production rate of E1S decreased more extensively during
treatment with the D-enantiomer compared with
DL-AG. The reason for this discrepancy is probably the drop
in plasma androstenedione levels during treatment with
D-AG. Changes in plasma androgen levels during treatment
with AG have been discussed in detail previously (17, 22). When
administered without glucocorticoids, AG causes a profound rise in
plasma 17-hydroxyprogesterone and androstenedione, probably due to
11ß- and 21-hydroxylase inhibition followed by substrate accumulation
(17). When AG is administered with glucocorticoid substitution, a
significant suppression of 
5-steroids occurs, whereas
4-steroids, in general, remain unchanged (17). In the
present trial, both treatment options were administered with similar
doses of a glucocorticoid. The finding that treatment with
D-AG suppressed plasma levels of androstenedione, contrary
to treatment with DL-AG, may be due to a more profound
inhibition of the adrenal 11ß-hydroxylase (and probably
21-hydroxylase) with use of the racemic compound, suggesting that a
lower glucocorticoid dose may be required to balance this effect during
treatment with the D-AG regimen. This drop in plasma
androstenedione levels may explain a lower production rate of
E1 and E1S during treatment with
D-AG, supporting the hypothesis of Dowsett and Harris (23)
that changes in plasma androstenedione levels may influence plasma
estrogen levels in patients receiving treatment with aromatase
inhibitors. Thus, a more profound reduction in estrogen production
rates during treatment with D-AG may counteract a more
profound enhancement of the E1S clearance rate during
treatment with DL-AG, resulting in similar plasma estrogen
levels during treatment with the two regimens.
Previous publications showed AG to be a phenobarbital-like inducer of cytochrome P-450-dependent enzymes (8, 9). Although several studies have suggested a difference between the enantiomers of AG concerning enzyme-inhibiting properties (6, 24, 25), little is known about chiral effects on enzyme induction in man. In vitro and animal investigations on barbiturates have provided conflicting evidence about whether one or both enantiomers are responsible for the induction of P-450-dependent enzymes, as some studies have suggested stereoselective enhancement of P-450-dependent mixed function oxidases with the D-isomer being the biological active part (26), whereas others have reported both stereoisomers to be active in this respect (27). The finding in this study that both regimens enhanced E1S metabolism but this effect was more profound with the DL-AG racemate suggests both enantiomers of AG to be involved in the enhancement of plasma E1S clearance. Noteworthy, previous results have shown AG to enhance warfarin metabolism in a dose-dependent manner when administered in the 250-1000 mg/day dose range (9), suggesting that a drug dose of 500 mg may be inferior to 1000 mg daily with respect to enzyme induction. Whether L-AG is active as an enzyme inducer on its own or acts indirectly via interaction with the metabolism of the D-enantiomer, as has been shown for other drugs (28), is currently unknown. Although the finding that AG levels during treatment with DL-AG were 3-fold and not 2-fold higher than plasma levels obtained during treatment with D-AG could be consistent with such a hypothesis, this observation should be interpreted carefully. First, plasma levels were measured in single samples obtained in the morning and thus may not be fully representative of mean 24-h plasma samples. Noteworthy, the half-life of the AG enantiomers has not been measured separately. Secondly, AG is known to induce its own metabolism (29), but whether this is effected by enhancing the metabolism of both enantiomers is not known.
AG is known to cause several side-effects, the most important being neurological, such as sedation, and skin rash. These side-effects occur mainly during the first 2 weeks after initiation of treatment (17). Although animal experiments have suggested both enantiomers to be responsible for the neurological side-effects (30), with the exception of one previous small study (31), this hypothesis has not been evaluated in humans to date. In theory, by using the single D-isomer in a lower dosage (500 mg daily) compared to the standard dosage of the racemate (1000 mg daily) this may reduce neurological side-effects even if they are caused by both enantiomers. However, we found treatment with D-AG (500 mg daily) to cause side-effects at the same frequency and intensity as treatment with DL-AG (1000 mg daily). These findings contrast with the observations made by others (31) that even higher doses of D-AG (900 mg daily) are well tolerated.
Our results show both treatment options to be equally effective in suppressing plasma estrogen levels and to cause the same profile of side-effects. Although L-AG is known to be ineffective as an aromatase inhibitor, our findings do not suggest any major clinical advantage from eliminating this enantiomer from the drug.
In conclusion, our finding of a higher E1S clearance during treatment with DL-AG compared to D-AG alone suggests both the D- and L-forms of AG to influence E1S metabolism. Considering E1S is an important estrogen source for the tumor cell, obviously a reduction in the plasma level of this hormone by enhancing its metabolism is an interesting approach to achieve maximal estrogen suppression. This observation together with the finding that suppression of plasma androstenedione influences estrogen levels in patients during treatment with a potent aromatase inhibitor suggest that future strategies with combined drug treatment regimens attaching E1S metabolism or suppressing androstenedione levels in addition to aromatase inhibition could be beneficial to achieve maximal estrogen suppression in breast cancer patients.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
2 Research fellow of the Norwegian Cancer Society.
Received February 20, 1998.
Revised May 1, 1998.
Accepted May 6, 1998.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
This article has been cited by other articles:
 |
|
 |
|
  K. Yacobi, A. Tsafriri, and A. Gross Luteinizing Hormone-Induced Caspase Activation in Rat Preovulatory Follicles Is Coupled to Mitochondrial Steroidogenesis Endocrinology, April 1, 2007; 148(4): 1717 - 1726. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. E. Lonning, E. Bajetta, R. Murray, M. Tubiana-Hulin, P. D. Eisenberg, E. Mickiewicz, L. Celio, P. Pitt, M. Mita, N. K. Aaronson, et al. Activity of Exemestane in Metastatic Breast Cancer After Failure of Nonsteroidal Aromatase Inhibitors: A Phase II Trial J. Clin. Oncol., June 11, 2000; 18(11): 2234 - 2244. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
