J. Med. Chem. 1993,36, 1619-1629
1619
16&( [18F]Fluoro)estrogens: Systematic Investigation of a New Series of Fluorine-18-LabeledEstrogens as Potential Imaging Agents for Estrogen-Receptor-PositiveBreast Tumors Henry F. VanBrocklin,+J*gKathryn E. Carlson,t John A. Katzenellenbogen,'-t and Michael J. Welch? Division of Radiation Sciences, Mallinckrodt Institute of Radiology, Washington University Medical School, 510 South Kingshighway, St. Louis, Missouri 63110, and Department of Chemistry, University of Illinois, 1209 West California Street, Urbana, Illinois 61801 Received January 14, 1993
In order to understand the structural features that might lead to an estrogen receptor (ER) based breast tumor imaging agent with improved uptake characteristics, we have synthesized several new analogs of 16/3-fluoroestradiol (@FES)and studied their tissue distribution in immature rata. The compounds we prepared were llB-rnethoxy-@FES(7a), 11/3-ethyl-@FES(7b), 17a-ethynyl-flFES (&IC), 17a-ethynyl-l1/3-methoxy-~FES @a),and 11/3-ethyl-l'la-ethynyl-j3FES (ab). All of the analogs exhibit good affinity for ER, ranging at 25 "C from 10 to 460, with estradiol equal to 100. Measurement of their octanol/water partition coefficients by an HPLC method allowed us to estimate their level of nonspecific binding and thereby to predict their binding selectivity indices (BSI, i.e., the ratio of their ER-specific to nonspecific binding); the BSI values of three fluorinesubstituted analogs exceed that of estradiol. These ligands have been labeled in the 168 position with fluorine-18 by the nucleophilic displacement of an a-disposed trifluoromethanesulfonate by [l8F1fluorideion. Reduction with lithium aluminum hydride produced the estradiol series ([lsF]7a-c), while treatment with lithium trimethylsilylacetylide afforded the ethynylated series ( P F ]8a-c). The synthesis time was 85min for [l8F1-7a-cand 120min for [l8F1-8a-c,with radiochemical yields ranging from 16 to 43%, and effective specific activities being 90-2900 Ci/mmol(3.3-107 TBq/mmol). In tissue distribution studies in immature female rata, all of the labeled analogs demonstrated ER-selective uptake in the principal target tissues, the uterus and the ovaries, and also in organs with lower titers of ER, the secondary target sites kidney, thymus, fat, and muscle. Although factors other than specific and nonspecific binding obviously affect the tissue distribution of these 16/3-fluoroestrogens,we find that their ER-specific uptake by both the principal and the secondary target tissues correlates with their BSI values at a high level of statistical significance in most cases. The ethynylated-116-methoxyanalog [lsF]-8a had high selectivity (uterus to blood ratio) after 3 h and exhibited the highest uterine uptake (percent injected dose/gram) of any fluorine-substituted estradiol ligand we have studied to date. This compound has been chosen for more detailed studies (to be described elsewhere), including clinical trials in human patients diagnosed with primary breast cancer.
Introduction The assessment of estrogen receptor (ER)concentration in human breast carcinoma has significant clinical applications for the selection of effective therapeutic regimens.l Currently, an invasive biopsy coupled with an in vitro assay are the principal means by which receptor content is ascertained. Shortcomings in this assay and the search for noninvasive methods have prompted the investigation of in vivo receptor quantitation using radiolabeled estrogem2 The most promising in vivo agent to date is 16a-[18F]fluoroestradiol(FES)? In human studies, FES demonstrated favorable uptake in both primary and metastatic breast lesion^.^ A strong correlationwas shown in primary tumors between the estrogen receptor concentration measured by in vitro assay and the tumor uptake of labeled FES measured by positron emission tomography (PET).4a * Address correspondence to John A. Katzenellenbogen, Department of Chemistry, University of Illinois, 1209 Wec +
Washington University Medical School.
t University of Illinois.
Present address: Center for Functional Imaging, LBL, University of California, 1 Cyclotron Rd., MS 55/121, Berkeley, CA 94720. 8
0022-2623/93/1836-1619$04.00/0
Effective ER imaging agents must possess (a) high specific activity, (b) high receptor affinity and binding selectivity,and (c)appropriate distribution and clearance characteristicsof labeled rnetabolite~.~*6 In our attempts to improve upon the in vivo distribution behavior of FES, we have prepared a number of estradiol analogs bearing additional substituents in the 11sposition (methoxy and ethyl) and at the 17a position All members of this series were initially labeled with fluorine at the 16a position, because the 16a epimer binds to the estrogen receptor with higher affinitythan the 16sepimer (relative binding affinities (RBA) are estradiol = 100; l6a-fluoroestradiol = 54; 16@-fluoroestradiol=12).3b As we have described,7!8 these 11s- and l7a-substituted 16a-fluoroestrogens cover a range of estrogen receptor and nonspecificbinding affmities,and in tissue distributionstudies in immature rats, most show efficient and selective receptor-mediated uptake into the major target site, the uterus. However, despite these favorable results, two factors have led us to make a further investigation of fluoroestrogens, which has extended now into the corresponding 16sfluoro epimeric series. First, although we found in the lea-fluoroestrogenseries that there was a reasonably good correlation of target tissue uptake selectivitydirectly with 0 1993 American Chemical Society
VanBrocklin et al.
1620 Journal of Medicinal Chemistry, 1993, Vol. 36, No. 11
Scheme I h/d
LALi
A THP
3dc
~~~~
':LOA 21 MOOPH
1 TIOH
io
U E
Table I. Decay-CorrectedRadiochemical Yield Ranges for 16aand 16B-Fluoroestra-3.178-diols range of yield ( % ) ligand 16a-l8Fn 160-18F fluoroestradiol (FES) (7c) 11-47 17-41 116-methoxy-FES (7a) 12-35 7-42 7-20 18-35 116-ethyl-FES (7b) 17a-ethynylfluoroestradiol (FEES) (8c) 2-21 16-35 110-methoxy-FEES (8a) 3-13 24-44 110-ethyl-FEES (8b) 1-6 19-30 Data are taken from refs 3, 7, and 8. (I
exclusively the 16a-hydroxy diastereomer. The stereochemical assignments at C-16 were inferred from their I"? 'H-NMR spectra: The 16a-protons of the corresponding previously reported 168-hydroxy derivatives appear as triplets at 4.0 ppm, while the 168-protons in the 16ahydroxy derivatives 3a-c resonate at 4.4ppm and appear as do~blets.~*.~ Acid hydrolysisof the THP ethers provides estrogen receptor binding affinity and inversely with the corresponding 3,16a-dihydroxyestrones 4a-c. The 3lipophilicity,5this correlation did not hold in all and 16-hydroxyl groups are simultaneouslytriflated (5aSecond, in an in uivo titration study of l6a-fluoroestradiol c) with triflic anhydride and 2,6-lutidine to activate the itself: we found that the uptake by a receptor-rich target 16~-alcohol toward displacement and protect the 3-phenol. tissue such as the uterus was limited by blood flow and Treatment of the bistriflates 5a-c with 1equiv of n B u tissue permeability characteristics, and thus may not NF in THF produces the 16&fluoroestrone 3-triflates 6adirectly reflect a compound's affinity for the estrogen c. Reduction of the C-17 ketone with LiAlH4 proceeds receptor, nor its potential for efficient, selective uptake stereoselectivity to yield exclusivelythe 3,178-diols7a-c. by tissues and tumors that are less receptor rich. The 168-fluor0group and the 13-methyl group hinder the For these reasons, we have extended our investigation attack of LAH on the j3 face of the molecule, thereby of fluorine-substitutedestrogens into the l6fl-fluoroseries. directing attack on the a face, providing the desired 178In certain cases, we have investigated the extent of hydroxy derivatives with very high stereoselectivity. receptor-mediated uptake not only by the principal target Reduction with LAH also cleaves the phenolic triflates, tissues, the uterus and the ovaries, but also by secondary leaving the free phenol upon workup. target tissues, e.g. kidney, thymus, fat, and muscle, that Synthesis of Fluorine-Substituted Estrogens 8ais, those tissues with lower estrogen receptor content in c. The synthesis of the 17a-ethynyl-16j3FES (Scheme I) which uptake should not be flow limited; the uptake analogs 8a-c parallels the 16O-FES synthesis with the efficiencyof the radiolabeled estrogens by these secondary exception that lithium trimethylsilylacetylide is added in target tissues may better reflect the estrogen receptor place of LAH to the 16/3-fluoroestrone 3-triflates. The binding characteristics of the estrogens and their potential nucleophilic addition of the acetylide to the C-17 carbonyl for uptake by human breast tumors. In this study, we is highly stereoselective and proceeds from the a face, due have prepared six estradiol analogs in the l6Sfluoro to steric hinderance of the j3 face, producing only the epimeric series with substituents in the 1 l P and 17a desired l7j3-hydroxy derivatives. Removal of the tripositions. All of the ligands have good affinity for the methylsilyl group and the 3-0-triflate under aqueousbase estrogen receptor and show receptor-mediated uptake in and conditionsyieldsthe l7a-ethynyl-16/3-fluoro-ll-protio estrogen receptor-rich target tissues in immature rats, and 11-substituted estradiols 8a-c. in some cases also in secondary target tissues. One Synthesis of Fluorine-18-Labeled Estrogens. The compound in particular, 17a-ethynyl-16/3-fluoro-11/3fluorine-18-labeled 16/3-fluoroestrogens were synthesized methoxyestradiol(16,9-fluoromoxestrol),appears to have along the same pathway as the unlabeled analogs. [l8F1unusually favorable target tissue uptake efficiencies and Fluoride ion was produced by the proton bombardment selectivity, and has been selected for further study.1° of an enriched H P O target.12 The water was azeotropically removed in the presence of nBQNOH. The residue Results was resolubilized in dry THF and added to the bistriflate Synthesis of Fluorine-Substituted Estrogens 7aprecursors 5a-c. The 16a-triflate moiety was displaced c. The llp-methoxy- and llp-ethyl-substituted 16@by [lsF1fluoride with gentle heating. Subsequent reducfluoroestradiol analogs7a and 7b were prepared from their tion of the (2-17 keto group with LAH or C-17 ethynylation respective 118-substituted estrones la and lb. The 118and base deprotection afforded the 16/3-fluoroestradiols substituted estrones were synthesized from l-dehydroanPFI-7a-c and ['8F1-8a-c. All six labeled estradiols were drenosterone as reported by Pornper.' The ll-unsubstipurified by semipreparativenormal-phase HPLC. Radiotuted 16~-fluoroestradio17cwas prepared from e ~ t r o n e . ~ HPLC and radio-TLC chromatograms indicate the formation of the desired 178-hydroxy derivatives to the The estrone analogs la-c were converted to their exclusion of the 17a epimer in all six compounds. This respective 3,16a-bis(trifluoromethanesulfonates)(triparallels and confirms the selectivity seen with the flates) in four steps, as shown in Scheme I. The phenolic unlabeled compounds. functions were protected as tetrahydropyranyl (THP) Total synthesis and purification time for the 168ethers. The a-disposed hydroxy group was introduced at fluoroestradiols ['8F1-7a-c and the ethynylated 168(2-16 by direct hydroxylation of the enolate with the fluoroestradiols [18F]-8a-c was 85 and 120 min, respecmolybdenum peroxide species MoOPHll (MoOsPy. tively,from the end of bombardment. The decay-corrected HMPA). This reaction proceeds stereoselectivity,yielding
&-
1619-Fluoroestrogens
Journal of Medicinal Chemistry, 1993, Vol. 36, No. 11 1621
Table 11. Relative Binding Affinities, Nonspecific Binding Coefficients, and Binding Selectivity Indices for the Estrogen Receptor Ligands ER SBPa AFPa log Pb NSBC BSId 25 "C 100 3.26 1.00 100 100 100 estradiol (ES) ll&methoxyestradiol 86 2.72 0.57 151 1.72 0.26 118-ethylestradiol 1360 3.9 2.10 648 40.8 1.21 272 3.42 1.18 231 1.81 3.51 17a-ethynylestradiol 17a-ethynyl-ll&methoxyestradiol 185 3.01 0.76 243 0.071 0.023 ll~-ethyl-17a-ethynylestradiol 946 4.28 2.86 331 41.2 0.347 16b-FES (712) 12 2.81 8.9 20 0.63 19 llfl-methoxy-l6@-FES(7a) 13 2.35 0.39 34 0.087 0.062 llðyl-lG@-FES(7b) 253 3.72 1.60 158 0.55 0.079 17a-ethynyl-l6@-FES(812) 28 3.30 1.04 27 0.589 22 l7a-ethynyl-l1~-methoxy-16j3-FES (8a) 78 2.87 0.67 116 0.037
