O R I G I N A L
`
`A R T I C L E
`
`E n d o c r i n e
`
`R e s e a r c h
`
`Clinical and Biochemical Consequences of CYP17A1
`Inhibition with Abiraterone Given with and without
`Exogenous Glucocorticoids in Castrate Men with
`Advanced Prostate Cancer
`
`Gerhardt Attard, Alison H. M. Reid, Richard J. Auchus, Beverly A. Hughes,
`Amy Mulick Cassidy, Emilda Thompson, Nikhil Babu Oommen, Elizabeth Folkerd,
`Mitch Dowsett, Wiebke Arlt,* and Johann S. de Bono*
`
`The Institute of Cancer Research (G.A., A.H.M.R., A.M.C., E.F., M.D., J.S.d.B.), Sutton, Surrey SM2 5NG,
`United Kingdom; The Royal Marsden National Health Service Foundation Trust (G.A., A.H.M.R., E.T.,
`N.B.O., E.F., M.D., J.S.d.B.), Sutton, Surrey SM2 5PT, United Kingdom; Division of Metabolism,
`Endocrinology, and Diabetes (R.J.A.), Department of Internal Medicine, University of Michigan, Ann
`Arbor, Michigan 48109; and Centre for Endocrinology, Diabetes, and Metabolism (B.A.H., W.A.), School
`of Clinical and Experimental Medicine, University of Birmingham, Birmingham B15 T22, United Kingdom
`
`Context: Abiraterone acetate is a small-molecule cytochrome P450 17A1 (CYP17A1) inhibitor that
`is active in castration-resistant prostate cancer.
`
`Objective: Our objective was to determine the impact of abiraterone with and without dexameth-
`asone treatment on in vivo steroidogenesis.
`
`Design and Methods: We treated 42 castrate, castration-resistant prostate cancer patients with
`continuous, daily abiraterone acetate and prospectively collected blood and urine before and
`during abiraterone treatment and after addition of dexamethasone 0.5 mg daily.
`
`Results: Treatment with single-agent abiraterone acetate was associated with accumulation of
`steroids with mineralocorticoid properties upstream of CYP17A1. This resulted in side effects,
`including hypertension, hypokalemia, and fluid overload, in 38 of 42 patients that were generally
`treated effectively with eplerenone. Importantly, serum and urinary androgens were suppressed
`by more than 90% from baseline. Urinary metabolites of 17-hydroxypregnenolone and 17-hy-
`droxyprogesterone downstream of 17␣-hydroxylase remained unchanged. However, 3␣5␣-17-
`hydroxypregnanolone, which can be converted via the backdoor pathway toward 5␣-dihydrotes-
`tosterone, increased significantly and correlated with levels of the major 5␣-dihydrotestosterone
`metabolite androsterone. In contrast, urinary metabolites of 11-deoxycortisol and active gluco-
`corticoids declined significantly. Addition of dexamethasone to abiraterone acetate significantly
`suppressed ACTH and endogenous steroids, including 3␣5␣-17-hydroxypregnanolone.
`
`Conclusion: CYP17A1 inhibition with abiraterone acetate is characterized by significant suppres-
`sion of androgen and cortisol synthesis. The latter is associated with a rise in ACTH that causes raised
`mineralocorticoids, leading to side effects and incomplete 17␣-hydroxylase inhibition. Concomi-
`tant inhibition of 17,20-lyase results in diversion of 17-hydroxyprogesterone metabolites toward
`androgen synthesis via the backdoor pathway. Addition of dexamethasone reverses toxicity and
`could further suppress androgens by preventing a rise in substrates of backdoor androgen
`synthesis. (J Clin Endocrinol Metab 97: 507–516, 2012)
`
`ISSN Print 0021-972X ISSN Online 1945-7197
`Printed in U.S.A.
`Copyright © 2012 by The Endocrine Society
`doi: 10.1210/jc.2011-2189 Received July 30, 2011. Accepted October 31, 2011.
`First Published Online December 14, 2011
`
`* W.A. and J.S.d.B. are joint senior authors.
`Abbreviations: An, Androsterone; AR, androgen receptor; CRPC, castration-resistant pros-
`tate cancer; CYP17A1, cytochrome P450 17A1; DHEA, dehydroepiandrosterone; DHEAS,
`dehydroepiandrosterone sulfate; DOC, 11-deoxycorticosterone; 3␣5␣-17HP, 3␣5␣-
`17OH-pregnanolone; LC-MS/MS, liquid chromatography/tandem mass spectrometry;
`17OHP, 17-hydroxyprogesterone; PSA, prostate-specific antigen; THALDO, 3␣,5␤-tetra-
`hydroaldosterone; THS, tetrahydro-11-deoxycortisol.
`
`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`jcem.endojournals.org
`
`507
`
`WCK1092
`Wockhardt Bio AG v. Janssen Oncology, Inc.
`IPR2016-01582
`
`1
`
`

`

`508
`
`Attard et al.
`
`CYP17A1 Inhibition by Abiraterone
`
`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`Prostate cancer is the second most common cause of male
`
`cancer-related death in the Western world (1). Treat-
`ment-naive prostate cancer is usually a hormone-driven dis-
`ease, with a response to castration observed in more than
`90% of patients. The median duration of response is 18
`months. Overwhelming evidence now confirms that in a sig-
`nificant proportion of patients, relapse with castration-resis-
`tant prostate cancer (CRPC) occurs secondary to reactiva-
`tion of androgen receptor (AR) signaling, including by serum
`androgens from nongonadal sources (2). Cytochrome P450
`17A1 (CYP17A1) is a key enzyme in cortisol synthesis via its
`17␣-hydroxylase activity and plays a central role in androgen
`biosynthesis with its 17,20-lyase activity catalyzing the conver-
`sion of 17-hydroxypregnenolone to the main androgen precur-
`sor dehydroepiandrosterone (DHEA) (3, 4). CYP17A1 is ex-
`pressed in the gonads but also at extragonadal sites including
`the prostate (5–7) where it might contribute to intracrine
`hormone synthesis.
`Abiraterone is a rationally designed, small-molecule in-
`hibitor of CYP17A1 (8, 9). The specificity of abiraterone
`for inhibition of 17,20-lyase vs. 17␣-hydroxylase is low
`(1.4-fold, IC50 ⫽ 2.9 nM compared with 4 nM) (10), and
`treatment with abiraterone acetate would therefore be ex-
`pected to cause a rise in ACTH with a consequent increase
`in 11-deoxycorticosterone (DOC) and corticosterone,
`mimicking the effects observed in patients with congenital
`adrenal hyperplasia due to inactivating CYP17A1 muta-
`tions (3). To date, it has proven difficult to develop a small-
`molecule therapeutic that specifically inhibits only the
`17,20-lyase activity of CYP17A1 (11).
`When administered to noncastrate men, abiraterone ac-
`etate resulted in suppression of testosterone with a subse-
`quent LH surge that overcame inhibition of gonadal testos-
`terone synthesis (12). We subsequently reported that
`continuous inhibition of CYP17A1 with oral abiraterone ac-
`etate in chemotherapy-naive, CRPC patients was safe and
`significantly suppressed serum androgens and estrogens (6,
`13). Importantly, we and others reported significant antitu-
`mor activity in phase I/II trials with single-agent abiraterone
`acetate after multiple previous lines of hormone therapy, in-
`cluding ketoconazole, and in chemotherapy-treated patients
`(13–15). Recently, abiraterone acetate was given regulatory
`approval for the treatment of men with advanced prostate
`cancer progressing after docetaxel after a phase III study
`showed that combined treatment with abiraterone acetate
`and prednisone conferred a survival benefit when compared
`with prednisone alone (16).
`We and others previously used RIA to measure serum
`androgens. Androstenedione and DHEA sulfate (DHEAS)
`were suppressed to below the lower limits of detection (2
`ng/dl and 15 ␮g/dl, respectively), and DHEA declined
`3-fold; however, because cross-reactivity with abiraterone
`
`was observed in this assay, the detection of DHEA in pa-
`tients on abiraterone acetate could be artifactual (6, 15).
`A liquid chromatography/tandem mass spectrometry (LC-
`MS/MS) assay was used to measure serum testosterone
`levels that declined to below the lower limit of sensitivity
`(1 ng/dl) in all patients (6, 15). However, these studies have
`incompletely dissected the biochemical consequences of
`treatment with abiraterone acetate. We here report the
`first detailed, mass spectrometry-based analysis of the ste-
`roidogenic effects of CYP17A1 inhibition in samples
`taken from medically castrated patients treated with sin-
`gle-agent abiraterone acetate and with the combination of
`abiraterone acetate and dexamethasone.
`
`Patients and Methods
`
`Patients
`All the patients included in this prospectively planned analysis
`were enrolled into the continuous daily study of abiraterone acetate
`performed at the Royal Marsden Hospital (RMH), London, UK
`(COU-001, www.clinicaltrials.gov identifier NCT00473512). All
`patients were castrate (serum testosterone ⬍ 50 ng/dl) for the du-
`ration of the study (ongoing treatment with a LHRH analog), had
`an Eastern Co-operative Oncology Group (ECOG) performance
`status of 0 or 1, had not received previous treatment with chemo-
`therapy or radionuclides, and had progressive disease as defined by
`prostate-specific antigen (PSA) Working Group I (17). The ethics
`reviewcommitteesoftheRMHapprovedthisstudy,andallpatients
`gave informed consent. The antitumor activity and safety data ob-
`served in this study were reported previously (6, 13).
`
`Treatment and procedures
`Abiraterone acetate powder was administered in four 250-mg
`capsules to fasted patients daily in 28-d cycles. Safety evaluations
`were conducted at baseline, weekly for the first two cycles and
`every cycle thereafter and included a physical examination and
`complete blood count, clotting, serum creatinine, electrolytes,
`and liver function tests. All adverse events were graded according
`to the U.S. National Cancer Institute common toxicity criteria
`version 3.0. Toxicity related to elevated mineralocorticoid levels
`was managed with the selective mineralocorticoid receptor an-
`tagonist eplerenone (50 –200 mg/d), and treatment with gluco-
`corticoids to suppress ACTH was only used if mineralocorticoid
`antagonism did not reverse these toxicities. Spironolactone was
`not used because it is an agonist for wild-type AR (18). Abi-
`raterone acetate was continued until PSA progression as defined
`by PSA Working Group I (17). Patients were then given the op-
`tion to continue abiraterone acetate in combination with dexa-
`methasone 0.5 mg daily, which as the standard glucocorticoid
`preparation for treating CRPC at our institution (19) was used
`in preference to other steroids to allow the evaluation of antitu-
`mor activity. Eplerenone was discontinued after initiation of
`dexamethasone when toxicities related to mineralocorticoid ex-
`cess resolved. Patients continued treatment on abiraterone ace-
`tate and dexamethasone until PSA, radiological (20), or clinical
`progression; withdrawal of consent; or death (6, 13).
`
`2
`
`

`

`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`jcem.endojournals.org
`
`509
`
`Patients were given the option to consent to additional blood
`draws for the evaluation of serum androgens and plasma ACTH
`weekly for the first two cycles, on d 1 every cycle thereafter until
`progression on single-agent abiraterone acetate and after addi-
`tion of dexamethasone to abiraterone acetate, weekly for the first
`cycle and monthly thereafter. Patients were also given the option
`to consent to and provide a 24-h urine sample for steroid me-
`tabolite analyses before treatment, after a minimum of 28 d con-
`tinuous dosing (d 1, cycle 2 or 3) and on d 1, cycle 6 or 7 (Sup-
`plemental Fig. 1, published on The Endocrine Society’s Journals
`Online web site at http://jcem.endojournals.org).
`
`Measurement of ACTH and steroids
`Plasma ACTH was measured by the RMH Academic Bio-
`chemistry Laboratories with a solid-phase two-site sequential
`chemiluminescent immunometric assay (LKAC1) using an IM-
`MULITE 1000 autoanalyzer (Siemens Heathcare Diagnostics
`Products Ltd., Surrey, UK). The analytical sensitivity was 9 pg/
`ml. The intraassay precision and interassay precision was 6.1 and
`9.4%, respectively, at 51 pg/ml. Serum testosterone, DHEAS,
`and androstenedione were measured using a modified LC-
`MS/MS assay developed by Esoterix (Calabasas Hills, CA) with
`lower limits of sensitivity of 0.05 ng/dl for testosterone, 0.1 ng/dl
`for DHEAS, and 0.1 ng/dl for androstenedione.
`Twenty-seven urinary steroids were measured by gas chro-
`matography/mass spectrometry in selected-ion-monitoring
`mode after solid-phase extraction and derivatization as de-
`scribed previously (21, 22). For simplification, the 17-deoxy-
`21-carbon steroids that represent metabolites of DOC, cor-
`ticosterone, and 18-hydroxycorticosterone are described as
`mineralocorticoid precursor metabolites. Similarly, the 17-hy-
`droxy-21-carbon steroids derived from active glucocorticoids
`are summarized as glucocorticoid metabolites. The sum of the
`urinary glucocorticoid metabolites ␣-cortol, ␤-cortol, ␣-cor-
`tolone, and ␤-cortolone is reported as cortols plus cortolones.
`The 19-carbon steroids represent the sum of androgen and an-
`drogen precursor metabolites (Table 1).
`
`Statistics
`The median time on treatment, defined as the time from start
`until discontinuation of abiraterone acetate treatment or addition
`of dexamethasone, for the intention-to-treat population with cen-
`soring of patients who did not have progressive disease was calcu-
`lated using the Kaplan-Meier method. The significance of the dif-
`ference between pretreatment vs. on-treatment urinary steroid
`levels and levels on-treatment vs. after addition of dexamethasone
`was determined using the sign test (calculated using Stata version
`10.1). This nonparametric method tested the null hypothesis that
`the median of the differences of paired data is 0. Correlations of
`nonparametric data are defined by the Spearman correlation coef-
`ficient (r) calculated using Prism statistical software (version 5;
`GraphPad, San Diego, CA). An r of 1 is a perfect correlation. All P
`values are two sided, and a result was considered significant when
`the P value was ⬍0.05.
`
`Results
`
`Single-agent abiraterone acetate is associated
`with a syndrome of secondary mineralocorticoid
`excess
`Forty-two patients were treated for a total of 10,359 d
`(1479 wk) with single-agent 1000-mg abiraterone acetate.
`
`This cohort was presented previously when five patients
`continued on treatment (13), but now none of these pa-
`tients remain on single-agent abiraterone acetate. Abi-
`raterone acetate was discontinued prematurely because of
`rapid disease progression in four patients and for other
`reasons in five patients (two for deranged liver function
`tests, one due to acute-onset idiopathic polyneuropathy,
`another due to worsening asthma and eosinophilia, and
`the fifth for a diagnosis of colorectal carcinoma requiring
`chemotherapy). Three patients required addition of dexa-
`methasone to abiraterone acetate due to side effects of
`secondary mineralocorticoid excess. The remaining 30 pa-
`tients continued treatment on abiraterone acetate until a
`confirmed rise in PSA, at which time dexamethasone was
`added to evaluate reinduction of sensitivity as reported
`previously (Supplemental Fig. 1) (6, 13). The median time
`on treatment calculated using the Kaplan-Meier method
`was 231 d (95% confidence interval ⫽ 122–307).
`We here present in detail the clinical manifestations and
`management of side effects attributable to mineralocorti-
`coid excess secondary to CYP17A1 inhibition in patients
`receiving abiraterone acetate without exogenous gluco-
`corticoids. Four of 42 patients received abiraterone ace-
`tate with no clinical evidence of mineralocorticoid excess
`(and no rise in PSA) for 38, 65, 253, and 392 d. Dexa-
`methasone was added before an attempt with eplerenone
`due to intractable migrainous headaches in two patients:
`in one on d 72 in the presence of grade 3 hypokalemia (2.8
`mmol/liter), grade 2 hypertension, and grade 1 lower limb
`edema and in the other on d 58 in the presence of grade 1
`hypokalemia. The other 36 patients developed clinical ev-
`idence of a syndrome of mineralocorticoid excess and
`were treated with eplerenone that was initiated after a
`median of 28 d (range, 6 –387) (Fig. 1A). Thirty-five pa-
`tients had hypokalemia at initiation of eplerenone (Fig.
`1B), whereas one had normal serum potassium, grade 1
`hypertension, and grade 1 lower limb edema (Fig. 1C). In
`addition to hypokalemia, 10 patients had a raised blood
`pressure, five patients had a raised blood pressure and
`grade 1 lower limb edema, and five patients had grade 2
`lower limb edema. One patient had grade 3 lower limb
`edema, grade 3 pulmonary edema, and a serum potassium
`of 3.4 mmol/liter (Fig. 1C) that did not resolve with epler-
`enone and required addition of dexamethasone 0.5 mg
`daily to control his symptoms. Eplerenone up to a maxi-
`mum daily dose of 200 mg (Fig. 1D) resolved the clinical
`syndrome of mineralocorticoid excess in the other 35
`patients.
`
`Continuous daily abiraterone acetate leads to
`ACTH-driven mineralocorticoid excess
`We and others have previously used RIA to identify a
`significant increase in serum DOC, corticosterone, and
`
`3
`
`

`

`510
`
`Attard et al.
`
`CYP17A1 Inhibition by Abiraterone
`
`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`TABLE 1.
`Percent change in urinary steroid metabolite excretion with continuous daily abiraterone acetate and
`after addition of dexamethasone 0.5 mg once daily
`
`Baseline vs. single-agent
`abiraterone acetate (n ⴝ 21)
`% change from
`baseline
`Range
`
`Median
`
`P value
`
`Abiraterone acetate vs. after
`addition of dexamethasone (n ⴝ 7)
`% change from single
`agent abiraterone
`Range
`Median
`
`P value
`
`Urinary 17-deoxy-21-carbon steroids
`Pregnenolone metabolites
`Pregnenediol (5-PD)
`Progesterone metabolites
`Pregnanediol (PD)
`Mineralocorticoid precursor metabolites
`Tetrahydro-DOC (TH-DOC)
`5␣-Tetrahydro-DOC (5␣-TH-DOC)
`Tetrahydrocorticosterone (THB)
`Tetrahydro-11-dehydrocorticosterone (THA)
`5␣-Tetra-11-dehydrocorticosterone (5␣-THA)
`5␣-Tetrahydrocorticosterone (5␣-THB)
`Mineralocorticoid metabolites
`3␣5␤-tetrahydro-aldosterone (THALDO)
`
`Urinary 17-hydroxy-21-carbon steroids
`17-Hydroxypregnenolone metabolites
`5-Pregnenetriol (5-PT)
`17-hydroxyprogesterone (17OHP)
`metabolites
`Pregnanetriol (PT)
`17-OH-Pregnanolone (17HP)
`3␣5␣-17-OH-pregnanolone (3␣5␣-17HP)
`Pregnanetriolone
`11-Deoxycortisol metabolites
`Tetrahydro-11-deoxycortisol (THS)
`Glucocorticoid metabolites
`Tetrahydrocortisol (THF)
`5␣-Tetrahydrocortisol (5␣-THF)
`Cortisol
`Cortols and cortolones
`Tetrahydrocortisone (THE)
`
`249 to 4400
`
`842 ⬍0.0001 ⫺99 to ⫺77
`
`381 to 6106
`
`1919 ⬍0.0001 ⫺100 to ⫺62
`
`342 to 8769
`⫺63 to 3535
`662 to 8091
`1159 to 9658
`349 to 4581
`659 to 14222
`
`3069 ⬍0.0001 ⫺99 to ⫺67
`882 ⬍0.0001 ⫺100 to ⫺61
`2997 ⬍0.0001 ⫺100 to ⫺62
`4488 ⬍0.0001 ⫺100 to ⫺62
`1819 ⬍0.0001 ⫺99 to ⫺64
`3317 ⬍0.0001 ⫺100 to ⫺84
`
`⫺100 to 49
`
`⫺40
`
`0.0118 ⫺92 to 267
`
`⫺93
`
`⫺91
`
`⫺93
`⫺92
`⫺95
`⫺94
`⫺96
`⫺99
`
`⫺21
`
`0.0156
`
`0.0156
`
`0.0156
`0.0156
`0.0156
`0.0156
`0.0156
`0.0156
`
`⬎0.9999
`
`⫺87 to 330
`
`⫺15
`
`0.1892 ⫺99 to ⫺68
`
`⫺95
`
`0.0156
`
`⫺85 to 215
`⫺75 to 314
`⫺60 to 589
`⫺75 to 800
`
`0.3833 ⫺98 to ⫺53
`⫺36
`9 ⬎0.9999 ⫺99 to 0
`0.0002 ⫺100 to ⫺50
`100
`0.1892 ⫺100 to ⫺33
`29
`
`⫺93 to 71
`
`⫺73 ⬍0.0001 ⫺100 to ⫺59
`
`⫺97 to 2
`⫺96 to 27
`⫺100 to 138
`⫺89 to ⫺30
`⫺96 to 16
`
`⫺86 ⬍0.0001 ⫺99 to ⫺84
`⫺85 ⬍0.0001 ⫺100 to ⫺87
`⫺44
`0.0015 ⫺100 to ⫺100
`⫺68 ⬍0.0001 ⫺100 to ⫺80
`⫺80 ⬍0.0001 ⫺99 to ⫺83
`
`⫺82
`⫺81
`⫺99
`⫺80
`
`⫺90
`
`⫺94
`⫺97
`⫺100
`⫺94
`⫺95
`
`0.0156
`0.0313
`0.0156
`0.0156
`
`0.0156
`
`0.0156
`0.0156
`0.0313
`0.0156
`0.0156
`
`Urinary 19-carbon steroids
`Metabolites of androgens and androgen
`precursors
`DHEA
`Androsterone (An)
`Etiocholanolone (ETIO)
`
`⫺97 ⬍0.0001 ⫺100 to 633
`⫺99 to ⫺35
`⫺94 ⬍0.0001 ⫺94 to 100
`⫺98 to ⫺39
`⫺65
`00.0118 ⫺75 to 90
`⫺100 to 179
`A significance in the difference between metabolites is denoted by a P value ⬍ 0.05 calculated using the sign test.
`
`⫺43
`⫺39
`25
`
`⬎0.9999
`⬎0.9999
`0.6875
`
`11-deoxycortisol levels after treatment with abiraterone
`acetate (6, 15). This increase was associated with a sig-
`nificant rise in ACTH (median, 660% increase; range,
`283-1416% increase; P value ⬍ 0.0001, sign test) from a
`median of 17 pg/ml (range, ⬍9 –50 pg/ml) before treat-
`ment to 124 pg/ml (range, 46 –370 pg/ml) on treatment
`(n ⫽ 26). To further study the effect of abiraterone acetate,
`we have used gas chromatography/mass spectrometry to
`comprehensively study urinary steroid metabolites in 24-h
`urine collections before treatment and after one or two
`cycles (28 –56 d) of single-agent abiraterone acetate.
`Twenty-one patients consented to these analyses and pro-
`vided 24-h urine samples. All urinary mineralocorticoid
`
`precursor metabolites (upstream of CYP17A1) increased
`markedly on treatment; however, the metabolite of aldo-
`sterone, 3␣,5␤-tetrahydroaldosterone (THALDO) de-
`clined (Fig. 2A). The median of the sum of urinary min-
`eralocorticoid metabolites excluding THALDO rose 26-
`fold after one or two cycles of treatment from 847 ␮g/24
`h (range, 388-1503 ␮g/24 h) to 22,752 ␮g/24 h (range,
`7729 –75535 ␮g/24 h). Changes in individual metabolites
`and their significance calculated using the sign test are
`reported in Table 1. In contrast, and explaining the rise in
`ACTH, all metabolites of active glucocorticoids declined
`significantly (Table 1) with a 5-fold decrease in the median
`of the sum of metabolites from 9086 ␮g/24 h (range,
`
`4
`
`

`

`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`jcem.endojournals.org
`
`511
`
`FIG. 1. Syndrome of mineralocorticoid excess in patients (n ⫽ 38) treated with single-agent abiraterone acetate. A, Scatter plot showing number
`of patients and days after starting single-agent abiraterone acetate that eplerenone was commenced (n ⫽ 36; vertical red line marks median value
`of 28 d) (log2 scale); B, scatter plot showing number of patients and serum potassium level at initiation of eplerenone (n ⫽ 36); C, clinical
`manifestation of mineralocorticoid excess at start of eplerenone (n ⫽ 36) or dexamethasone (n ⫽ 2). Each column represents a patient, and
`shaded boxes signify presence of toxicity. Patient numbers are random. Patients 37 and 38 (red box) represent the two patients administered
`dexamethasone before a trial of eplerenone, and patient 14 (red arrow) represents the one patient who required dexamethasone despite a trial
`with eplerenone. D, Number of patients and dose of eplerenone that resolved the clinical syndrome of mineralocorticoid excess (n ⫽ 35).
`
`4645–17633 ␮g/24 h) before treatment to 2154 ␮g/24 h
`(range, 426-7416 ␮g/24 h) after one or two cycles of abi-
`raterone treatment (Fig. 2B). A one-way ANOVA showed
`no evidence of a difference in mean urinary excretion of min-
`eralocorticoid precursor metabolites and the dose of epler-
`enone required to control toxicity [F(5,15) ⫽ 0.27; P ⫽
`0.9214]. We also evaluated urinary steroid metabolites
`from nine patients who continued single-agent abi-
`raterone acetate for at least five or six cycles (140 –168 d)
`and did not observe a significant difference in urinary me-
`tabolites compared with after one or two cycles (Fig. 2).
`
`Abiraterone significantly suppresses urinary
`androgen metabolites and serum androgens
`The median of the sum of urinary 19-carbon steroids,
`i.e. androgen metabolites, in castrate men before starting
`abiraterone acetate was 735 ␮g/24 h (range, 127-6755
`␮g/24 h). After 28 –56 d treatment with single-agent abi-
`raterone acetate, urinary androgen metabolites were all
`significantly suppressed (Fig. 2C and Table 1) with a 20-
`fold decrease to 37 ␮g/24 h (range, 6 – 896 ␮g/24 h). There
`was no significant further change from after one or two
`cycles to after five or six cycles of single-agent abiraterone
`acetate (n ⫽ 9) (Fig. 2C).
`We then measured circulating androgen levels using an
`ultrasensitive LC-MS/MS assay (Esoterix) in samples col-
`lected before and on treatment and at progression. Ninety-
`
`three on-treatment samples from 32 patients were avail-
`able for analysis. Serum testosterone declined to a median
`of 0.26 ng/dl with a range of ⬍0.05– 0.90 ng/dl (decline to
`⬍0.05 ng/dl in five of 32 patients). Serum DHEAS de-
`clined to a median of 0.2 ng/dl with a range of ⬍0.1–9.4
`ng/dl (⬍0.1 ng/dl in 14 of 32 patients). Serum androstene-
`dione declined to a median of 0.32 ng/dl with a range of
`⬍0.1–1.58 ng/dl (⬍0.1 ng/dl in three of 19 patients). Im-
`portantly, there was no rise in serum testosterone or
`DHEAS in the 11 patients evaluated at disease progression
`on abiraterone acetate using these ultrasensitive assays, in
`contrast to previously reported changes at progression on
`ketoconazole (23).
`
`CYP17A1 inhibition with abiraterone is associated
`with increased substrates of the backdoor
`pathway of DHT synthesis
`Urinary metabolites of 17-hydroxypregnenolone and
`17-hydroxyprogesterone (17OHP) did not change signif-
`icantly with abiraterone acetate, but there was a signifi-
`cant increase in the 17OHP metabolite 3␣5␣-17OH-preg-
`nanolone (3␣5␣-17HP) from a median of 4 ␮g/24 h
`(range, 1–13.6 ␮g/24 h) to 8 ␮g/24 h (range, 1.7–36.1
`␮g/24 h) on abiraterone acetate (median intra-patient
`change ⫽ 100%, P value for significance of rise ⫽ 0.0002,
`sign test) (Table 1 and Fig. 2B). Interestingly, 3␣5␣-17HP
`levels on abiraterone acetate showed a significant corre-
`
`5
`
`

`

`512
`
`Attard et al.
`
`CYP17A1 Inhibition by Abiraterone
`
`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`FIG. 3. Scatter plot showing correlation between urinary An and
`urinary 3␣5␣-17HP (n ⫽ 30).
`
`lation with androsterone (An), the major metabolite of 5␣-
`reduced androgens including DHT (Spearman r ⫽ 0.73;
`95% confidence interval ⫽ 0.49–0.87; P ⬍ 0.0001) (Fig. 3).
`In contrast, tetrahydro-11-deoxycortisol (THS), the main
`metabolite of 11-deoxycortisol, declined significantly in all
`patients from a median of 117 ␮g/24 h (range, 62–238 ␮g/24
`h) before treatment to 29 ␮g/24 h (range, 6–158 ␮g/24 h)
`(median intra-patient change ⫽ ⫺73%; P value for signifi-
`cance of decline ⬍ 0.0001, sign test) (Table 1).
`
`Addition of exogenous glucocorticoids to
`abiraterone acetate reverses mineralocorticoid
`excess and is associated with increased
`suppression of glucocorticoid precursor
`metabolites
`Addition of oral dexamethasone (0.5 mg/d) to single-
`agent abiraterone acetate resulted in a decrease (within
`⬍14 d) in plasma ACTH to below the lower limit of sen-
`sitivity (10 pg/ml) in all six patients studied, consistent
`with previous reports of resolution of side effects of min-
`eralocorticoid excess on addition of exogenous glucocor-
`ticoids (6, 13–15). Twenty-four hour urine samples for
`evaluation of steroid metabolite excretion after addition
`of dexamethasone to abiraterone acetate were available
`from nine patients. The median of the sum of urinary min-
`eralocorticoid precursor metabolites declined to 555
`␮g/24 h (range, 135-4343 ␮g/24 h) after addition of dexa-
`methasone, which was lower than pretreatment levels.
`However, the sum of urinary mineralocorticoid precursor
`metabolites remained higher than before treatment in
`three of nine patients (Fig. 2A). All endogenous glucocor-
`ticoid precursor metabolites were significantly suppressed
`after addition of dexamethasone including 3␣5␣-17HP
`that declined to a median of 1 ␮g/24 h (range, 0 –3.4 ␮g/24
`h) (Fig. 2B and Table 1).
`The median of the sum of androgen metabolites after
`addition of dexamethasone (nine patients) was slightly
`
`FIG. 2. Changes in urinary mineralocorticoid metabolites (aldosterone
`metabolite separated from mineralocorticoid precursor steroids) (A),
`glucocorticoids (active glucocorticoid metabolites separated from
`glucocorticoid precursors) (B), and androgen metabolites (C) in castrate
`men with advanced prostate cancer pretreatment on treatment with
`abiraterone acetate (during cycle 2 or 3 and cycle 6 or 7) and after
`addition of dexamethasone. Mean and 5th and 95th percentile values
`are shown. Abbreviations for urinary steroids are given in Table 1 and
`Ref. 22. A significant difference between levels before treatment vs. on
`abiraterone acetate and between on abiraterone acetate vs. after
`addition of dexamethasone for each metabolite as reported in Table 1
`is denoted by asterisks: *,P value 0.05–0.01; **, P value 0.009–0.001;
`***, P value ⬍ 0.001.
`
`6
`
`

`

`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`jcem.endojournals.org
`
`513
`
`FIG. 4. Decline in serum testosterone (ng/dl) on treatment with abiraterone acetate and after addition of dexamethasone due to PSA progression.
`Testosterone concentration is presented as a log10 scale and is interrupted at 1 ng/dl. The lower limit of sensitivity of the assay is 0.05 ng/dl. Pt,
`Patient.
`
`lower than on abiraterone acetate alone (25 ␮g/24-h;
`range, 5– 62 ␮g/24 h), but the changes in individual me-
`tabolites were not significant (Table 1). Serum testoster-
`one and DHEAS were also measured in eight patients tak-
`ing combination abiraterone acetate and dexamethasone.
`Serum testosterone was lower in all patients on addition of
`dexamethasone compared with on single-agent abi-
`raterone acetate: on addition of dexamethasone, serum
`testosterone declined to below 0.05 ng/dl in four of eight
`patients (Fig. 4). Serum DHEAS was below 0.1 ng/dl in five
`of eight patients on abiraterone acetate alone, and in the
`other three patients, a decline was observed on addition of
`dexamethasone.
`
`Discussion
`
`In this report, we present a planned analysis of prospec-
`tively collected samples on the clinical and biochemical
`consequences of CYP17A1 inhibition. To comprehen-
`sively evaluate changes in mineralocorticoid, glucocorti-
`coid, and androgen production in men with CRPC treated
`with abiraterone acetate, we used mass spectrometry-
`based measurements of urinary steroids and serum andro-
`gens. We confirmed a marked rise in urinary metabolites
`of steroids upstream of CYP17A1 that are able to exert
`
`mineralocorticoid activity (Figs. 2A and 5A). These ste-
`roids also have weak glucocorticoid properties, and due to
`their high levels, abiraterone acetate can be safely admin-
`istered as a single agent with no clinical manifestations of
`glucocorticoid deficiency (6, 24). Urinary aldosterone me-
`tabolite excretion was unchanged, which is in keeping
`with our previous finding of unchanged serum aldoste-
`rone levels employing a RIA (6); this could be explained by
`suppression of the renin-angiotensin-aldosterone system
`and consequently down-regulated aldosterone synthase
`expression due to the abiraterone-induced accumulation
`of steroids with mineralocorticoid properties. Contrary to
`previous reports by us and others of raised serum 11-de-
`oxycortisol (6, 15), we here report 4-fold suppression of
`THS, the urinary metabolite of 11-deoxycortisol. It is
`likely that previously used RIA gave falsely high readings
`in the presence of highly up-regulated levels of steroid
`precursors such as DOC due to cross-reactivity issues.
`Predictably, all urinary 21-carbon steroids, mineralo-
`corticoids and glucocorticoids alike, were suppressed on
`addition of oral dexamethasone 0.5 mg/d to abiraterone
`acetate due to loss of ACTH drive (Fig. 5B). However, as
`shown in Fig. 2A, mineralocorticoid levels in some pa-
`tients on abiraterone actetate and dexamethasone were
`higher than before treatment. This observation could ex-
`
`7
`
`

`

`514
`
`Attard et al.
`
`CYP17A1 Inhibition by Abiraterone
`
`J Clin Endocrinol Metab, February 2012, 97(2):507–516
`
`19-carbon steroids (androgens & metabolites)
`Dehydroepi-
`androsterone
`DHEA
`Androstenedione
`Et, An
`
`DHEA sulfate
`(DHEAS)
`DHEA
`
`Testosterone
`Et, An
`
`Androstanedione
`An
`
`5α-Dihydrotestosterone
`An
`
`An
`Androstanedione
`
`17-hydroxy-21-carbon steroids
`
`17α-OH-pregnenolone
`5-PT
`
`17α-OH-progesterone
`17HP, PT
`
`**
`
`11-deoxycortisol
`THS
`
`5α-pregnane-
`17α-ol-3,20-dione
`
`Cortisol
`THF, 5αTHF, cortols
`
`5α-pregnane-
`3α,17α-diol-20-
`one
`3 5αα
`-17HP
`
`*
`
`Androstanediol
`An
`Androsterone
`An
`
`Aldosterone
`THALDO
`
`Cortisone
`THE, cortolones
`
`A
`17-deoxy-21-
`carbon steroids
`
`cholesterol
`
`pregnenolone
`5-PD
`
`progesterone
`PD
`
`11-deoxycorticosterone
`THDOC, 5αTHDOC
`
`corticosterone
`
`THB,
`5αTHB,
`THA, 5αTHA
`18-OH-corticosterone
`
`cholesterol
`
`B
`17-deoxy-21-
`carbon steroids
`
`pregnenolone
`5-PD
`
`progesterone
`PD
`
`11-deoxycorticosterone
`THDOC, 5αTHDOC
`
`corticosterone
`
`THB,
`5αTHB,
`THA, 5αTHA
`18-OH-corticosterone
`
`19-carbon steroids (androgens & metabolites)
`Dehydroepi-
`androsterone
`DHEA
`Androstenedione
`Et, An
`
`DHEA sulfate
`(DHEAS)
`DHEA
`
`Testosterone
`Et, An
`
`Androstanedione
`An
`
`5α-Dihydrotestosterone
`An
`
`An
`Androstanedione
`
`17-hydroxy-21-carbon steroids
`
`17α-OH-pregnenolone
`5-PT
`
`17α-OH-progesterone
`17HP, PT
`3α
`
`**
`
`11-deoxycortisol
`THS
`
`5α-pregnane-
`17α-ol-3,20-dione
`
`Cortisol
`THF, 5αTHF, cortols
`
`5α-pregnane-
`3α,17α-diol-20-
`one
`5α-17HP
`3α
`
`*
`
`Androstanediol
`An
`Androsterone
`An
`
`Aldosterone
`THALDO
`
`Cortisone
`THE, cortolones
`
`FIG. 5. Steroid biosynthesis pathway showing physiological consequences of CYP17A1 inhibition by abiraterone acetate 1000 mg/d (A) and after
`addition of dexamethasone 0.5 mg/d (B). Urine metabolites are given in italics and in boxes. *, Conversions dependent on CYP17A1 that are
`inhibited by abiraterone acetate. Abbreviations for urinary steroids are given in Table 1 and Ref. 22.
`
`plain the 15% increased prevalence of mineralocorticoid-
`related side effects in patients treated with abiraterone
`acetate and prednisone compared with prednisone alone
`(16). The observation that there was no significant differ-
`ence between urinary metabolites du