TABLE 1

Summary of compound characteristics for AMG 221 metabolites

Rat or dog liver microsomal metabolism of synthetic mono-oxidized metabolites (M1–M7 and M10) were used to generate secondary metabolites. Urine and bile from a rat dosed separately with [¹⁴C]M1 discerned M1-derived in vivo metabolites. Values for human 11βHSD1 inhibition measured in vitro using transfected CHO cells are from Li et al. (2011). Metabolites are sorted by ascending molecular weight and by functional groups. Fragmentation of AMG 221 is shown in Fig. 1. Mass spectra are in Supplemental Figs. S1 to S34. Metabolite hydroxylation sites are as denoted in Fig. 1.

ID	t_R	Stereo- and Regiochemistry	Rational Precursor	Structure	Summary of Compound Characteristics
	min
M8	40.7	9-Ene^{^a}	AMG 221 or M4/M6/M7		[MH]⁺ at m/z 265. Synthetic standard available. Radioactivity recovery from excreta was too low to be quantified. MS features (from standard): a loss of 2 Da from parent drug, propylene thiazolidinone moiety fragment (171 Da), and intact norbornyl fragment at m/z 95. IC₅₀ for h11βHSD1 inhibition in CHO cells was 63% of the inhibitory potency of AMG 221.
B20	27.9	9-Ene + 4-exo-OH^{^a}	M8/M1		[MH]⁺ at m/z 281. Radioactivity recovery from excreta was 0.8 ± 0.4% of dose. MS features: product ions at m/z 263 and 111 correspond to a single hydroxylation on the norbornane ring. Fragments at m/z 171 and 129 indicate an ene (M8) functional group. Formed from M1 in vitro and in vivo.
B22	29.9	9-Ene + 5-exo-OH^{^a}	M8/M3		[MH]⁺ at m/z 281. Radioactivity recovery from excreta was 0.1 ± 0.1% of dose. Retention time and MS data matched an in vitro metabolite of M3 and a dog urine metabolite. MS/MS features (spectrum from dog urine): a hydroxylated norbornane (m/z 111) and unsaturated propylene thiazolidinone moiety fragment at m/z 171.
M10	33.3	4-Keto^{^a}	M1/M2		[MH]⁺at m/z 281. Synthetic standard available. Radioactivity recovery from excreta was 9.1 ± 6.0% of dose. MS features: Fragment ions at m/z 239, 173, and 97 correspond to neutral loss of 42 Da (propylene), loss of the intact thiazolidinone moiety, and cleavage of the thiazolidinone ring, respectively. Formed in vitro and in vivo from M1, the ketone at the 4 position carbon could be reversibly generated by M1 or M2. M10 formed B6, B10, B9a, B9b, and B13 in vitro. M10 IC₅₀ for h11βHSD1 inhibition in CHO cells was 99% of the inhibitory potency of AMG 221.
M1	30.3	4-Exo-OH^{^a}	P/M10		[MH]⁺ at m/z 283. Epimer of M2. Synthetic standard available. Radioactivity recovery from excreta was 24.2 ± 3.2% of dose, which made M1 the most abundant excreted metabolite. The fragment ion at m/z 173 corresponds to an intact thiazolidinone moiety, and m/z 111 corresponds to hydroxylation of the norbornane ring. M1 formed B20, M10, B6, B10, B4, B8a, B8b, and B13 in vitro. Urine or bile metabolites of [¹⁴C]M1 in rats were as follows: B20, M10, M2, B29, B6, B10, B4, B9a, B9b, B13, M11, B30, B14, B17, and B2 in vivo. M1 IC₅₀ for human 11βHSD1 inhibition in CHO cells was 80% of the inhibitory potency of AMG 221.
M2	30.9	4-Endo-OH^{^a}	P/M10		[MH]⁺ at m/z 283. Epimer of M1. Synthetic standard available. Radioactivity recovery from excreta was 4.7 ± 1.6% of dose. MS features: the fragment ion at m/z 173 corresponds to an untransformed thiazolidinone moiety, and m/z 111 corresponds to hydroxylation of the norbornane ring. M2 formation in vivo from M1 indicates the racemization of M1 via the M10 metabolite. M2 formed only metabolite B5 in vitro. M2 IC₅₀ for 11βHSD1 inhibition in CHO cells was 67% of the inhibitory potency of AMG 221.
	min
M3	33	5-Exo-OH^{^a}	P		[MH]⁺ at m/z 283. Synthetic standard available, endo epimer standard not available. Radioactivity recovery from excreta was 3.2 ± 5.4% of dose. MS features: the fragment ion at m/z 173 corresponds to an untransformed thiazolidinone moiety, and m/z 111 corresponds to hydroxylation in the norbornane ring. M3 formed B22, B7, B11, B12, and B13 in vitro. M3 IC₅₀ for h11βHSD1 inhibition in CHO cells was much lower than 4-hydroxy regioisomers at 4% of the inhibitory potency of AMG 221.
M4	34.6	9(R)-11-OH^{^a}	P		[MH]⁺ at m/z 283. Epimer of M6. Synthetic standard available. Radioactivity recovery from excreta was 2.6 ± 0.6% of dose. The fragment ion at m/z 189 dehydrated to m/z 171 and corresponded to methyl hydroxylation on the isopropyl group attached to the thiazolidinone ring. The absence of a 225 ion characteristic of M7 defines the regiochemistry of the hydroxylation. m/z 95 represents the untransformed norbornane. M4 was metabolized to B24, B6, B10, B4, B5, and B7 in vitro. M4 IC₅₀ for h11βHSD1 inhibition in CHO cells was 14% of the inhibitory potency of AMG 221.
M5	36.2	5-Endo- or 2-OH	P		[MH]⁺ at m/z 283. No standard was available for this metabolite. Radioactivity recovery from excreta was 1.0 ± 0.1% of dose. MS features: a fragment ion at m/z 173 corresponded to an untransformed thiazolidinone moiety, and the ion at m/z 111 corresponded to hydroxylation in the norbornane ring. C5 (endo epimer of M3) versus C2 hydroxylation regioisomers cannot be distinguished (see Discussion).
M6	37.5	9(S)-10-OH^{^a}	P		[MH]⁺ at m/z 283. Epimer of M4. Synthetic standard available. Radioactivity recovery from excreta was 0.8 ± 0.7% of dose. MS features: the fragment ion at m/z 189 corresponds to hydroxylation on the isopropyl group attached to the thiazolidinone ring, and m/z 95 represents the untransformed norbornane ring. m/z 171 is neutral loss of water from m/z 189 fragment. The absence of a 225 ion characteristic of M7 defines the regiochemistry of the hydroxylation. M6 IC₅₀ for h11βHSD1 inhibition in CHO cells was 12% of the inhibitory potency of AMG 221.
M7	38.5	9-OH^{^a}	P		[MH]⁺ at m/z 283. Synthetic standard available. Radioactivity recovery from excreta was 0.9 ± 0.6% of dose. MS features: fragment ion at m/z 225 corresponds to hydroxylation on the methine (9) carbon of the isopropyl group attached to the thiazolidinone ring. The neutral loss of 58 Da is attributed to a six-center rearrangement (retro aldol reaction) that affords acetone (C₃H₆O). M7 IC₅₀ for 11βHSD1 inhibition in CHO cells was 5% of the inhibitory potency of AMG 22.
B24	25.8	9(R)-11-CO₂H^{^a}	M4		[MH]⁺ at m/z 297. Epimer of B26. Radioactivity recovery from excreta was too low to be quantified. MS features: product ions at m/z 203 correspond to an addition of 30 Da to the isopropyl thiazolidinone moiety. Loss of water (m/z 279) and loss of 46 Da (H₂CO₂, m/z 251) indicate a carboxylic acid. The fragment ion at m/z 95 indicates an intact norbornane ring. This metabolite is a carboxylic acid metabolite that was formed in vitro from M4.
	min
B26	35.7	9(S)-10-CO₂H^{^a}	M6		[MH]⁺ at m/z 297. Epimer of B24. Radioactivity recovery from excreta was 0.5 ± 0.5% of dose. MS features (MS/MS of metabolite in dog urine): product ions at m/z 203 correspond to an addition of 30 Da to the isopropyl thiazolidinone moiety. Loss of water (m/z 279) and loss of 46 Da (H₂CO₂, m/z 251) indicate a carboxylic acid. The fragment ion at m/z 95 indicates an untransformed norbornane ring. This metabolite is a carboxylic acid metabolite that was produced in vitro from M6.
B29	5.8		M1	No structure assigned	Apparent [MH]⁺ at m/z 297. The structure of this minor polar metabolite, observed in rat urine only by LC-MS at 5.8 min, was not defined by MS/MS. This metabolite was also formed from M1 in vivo.
B6	9.2	4-Keto and 9(R)-11-OH^{^a}	M10/M4		[MH]⁺ at m/z 297. Radioactivity recovery from excreta was 0.1 ± 0.1% of dose. MS features: fragment ions at m/z 189 and 109 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring and addition of 14 Da to the norbornane ring, respectively. No loss of acetone to m/z 239 was observed, which defined hydroxylation on the methyl group. Incubation of M1, M4, or M10 in the presence of RLM all resulted in the formation of B6.
B10	14.1	4-Keto and 9(S)-10-OH^{^a}	M10/M6		[MH]⁺at m/z 297. Radioactivity recovery from excreta was 2.1 ± 0.2% of dose. MS features: fragments at m/z 189 and 109 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring and addition of 14 Da to the norbornane ring, respectively. No loss of acetone to m/z 239 was observed, which defined hydroxylation on the methyl group. Incubation of M1, M6, or M10 in vitro resulted in the formation of B6.
B15	20.7	4-Keto and 9-OH^{^a}	M10/M7		[MH]⁺ at m/z 297. Radioactivity recovery from excreta was too low to be quantified. MS features: fragments at m/z 189 and 109 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring and addition of 14 Da to the norbornane ring, respectively. The fragment ion at m/z 239, from a neutral loss of acetone (58 Da), indicated hydroxylation on the methine carbon of the isopropyl group (an M7-like fragmentation).
B21	28.1	4-Keto and norbornyl OH	M10/M3 or M5		[MH]⁺ at m/z 297. Radioactivity recovery from excreta was 0.6 ± 0.1% of dose. The MS trace indicated a coeluting dihydroxy metabolite (MH⁺ 299). MS features: the fragment at m/z 125 indicated an addition of 30 atomic mass units to the norbornyl ring, and m/z 173 indicated no transformations on the thiazolidinone moiety. Genesis from M10 or M3 or M5.
B1	5.2	Di-OH-undefined, polar		No Structure	Apparent [MH]⁺ at m/z 299. Radioactivity recovery of polar metabolite was too low to be quantified. An apparent [MH]⁺ with low abundance and high background in the matrices analyzed.
B3	6.6	Di-OH (norbornyl and 10- or 11-OH)	M4/M6		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 0.4 ± 0.6% of dose. A polar metabolite with MS/MS data available only from dog plasma. Fragments at m/z 189 and 111 indicated hydroxylations in the thiazolidinone and norbornyl moieties, respectively. Dehydration of MH⁺ and no ion from loss of acetone (−56 Da) indicated hydroxylation of one of the isopropyl methyl groups.
	min
B4	7.2	Di-OH (4-exo OH and 9(R)-11-OH^{^a}	M1/M4		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 1.2 ± 0.8% of dose. MS features: fragments at m/z 189 and 171 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring. A fragment ion at m/z 111 is consistent with a single hydroxylation on the norbornane ring. B4 was formed from M1 in vivo and from M1 and M4 in vitro.
B5	8.5	Di-OH (4-endo OH and 9(R)-11-OH^{^a}	M2/M4		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 0.2 ± 0.1% of dose. MS features: fragments at m/z 189 and 171 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring. A fragment ion at m/z 111 is consistent with a single hydroxylation on the norbornane ring. B5 was formed from M2 and M4 in vitro.
B7	9.9	Di-OH (5-exo OH and 9(R)-11-OH)^{^a}	M3/M4		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 0.1 ± 0.0% of dose. MS features: fragments at m/z 189 and 171 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring. A fragment ion at m/z 111 is consistent with a single hydroxylation on the norbornane ring. B7 was formed from M3 and M4 in vitro.
B8a	11.5	Di-OH (4-exo OH and 9(S)-10-OH)^{^a}	M1/M6		[MH]⁺ at m/z 299. Isobaric peaks B8a and B8b coeluted and were differentiated by mass spectrometry. Radioactivity recovery from excreta for the sum of 8a and 8b was 2.9 ± 0.3% of dose.
					MS features: fragments at m/z 189 and 171 correspond to hydroxylation on the isopropyl substituent of the thiazolidinone ring. A fragment ion at m/z 111 is consistent with a single hydroxylation on the norbornane ring. B8a was formed from M1 and M6 in vitro.
B8b	11.5	Di-OH (4-exo OH and 9-OH)^{^a}	M1/M7		[MH]⁺ at m/z 299. Isobaric peaks B8a and B8b coeluted and were differentiated by mass spectrometry. Radioactivity recovery from excreta for the sum of 8a and 8b was 2.9 ± 0.3% of dose.
					The major fragment ion from loss of acetone at m/z 241 (225 + 16 Da) indicated M7 regiochemistry in the isopropyl hydroxy. Accordingly, B8b was formed from M1 and M7 in vitro.
B9a	12.8	Di-OH (4-endo OH and 9(S)-10-OH)^{^a}	M2/M6		[MH]⁺ at m/z 299. Isobaric peaks B9a and B9b coeluted and were differentiated by mass spectrometry. Radioactivity recovery from excreta for the sum of 9a and 9b was 0.1 ± 0.1% of dose.
					MS features: the fragment at m/z 189 corresponds to hydroxylation on the isopropyl substituent of the thiazolidinone ring. The fragment ion at m/z 111 is consistent with a single hydroxylation on the norbornane ring.
					B9a was formed from M1 in vivo and M6 and M10 in vitro, which inferred that M10 was reduced to M2 to produce M9a in vitro.
B9b	12.8	Di-OH (4-endo OH and 9-OH)^{^a}	M2/M7		[MH]⁺ at m/z 299. Isobaric peaks B9a and B9b coeluted and were differentiated by mass spectrometry. Radioactivity recovery from excreta for the sum of 9a and 9b was 0.1 ± 0.1% of dose.
	min
					The major fragment ion from loss of acetone at m/z 241 (225 + 16 Da) indicated M7 regiochemistry in the isopropyl hydroxy. Accordingly, B9b was formed from M1 in vivo and from M10 and M7 in vitro, which inferred that M10 was reduced to M2 to produce M9b in vitro.
B11	15.7	Di-OH (5-exo OH and 9(S)-10-OH)^{^a}	M3/M6		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 0.9 ± 0.6% of dose. Structural assignment was based upon the molecular ion detected at 15.1 min in dog and rat urine and MS/MS data from the same peak formed from incubation of M6 or M3 with rat liver microsomes. MS features: the fragment at m/z 189 corresponds to hydroxylation on the isopropyl substituent of the thiazolidinone ring. The fragment at m/z 111 is consistent with a single hydroxylation on the norbornane ring.
B12	16.8	Di-OH (5-exo OH and 9-OH)^{^a}	M3/M7		[MH]⁺at m/z 299. Radioactivity recovery from excreta was 2.4 ± 4.8% of dose. The MS/MS data were acquired from an in vitro peak at the correct retention time. The fragment ion at m/z 241 indicated a hydroxylation on the isopropyl methine carbon, similar to M7, with the other hydroxylation on the norbornane. The in vitro experiments showed that M3 and M7 are precursors of B12.
B13	19.2	Di-OH (4-exo-OH and 5-exo-OH)^{^a}	M1/M3		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 0.1 ± 0.1% of dose. MS features: the fragment m/z 173 indicated an intact thiazolidinone moiety; therefore, hydroxylations are on the norbornane ring. The in vitro experiments show that M1 and M3 are precursors of B13.
B23	31.5	Di-OH (9-OH and 10 or 11-OH)	M4 or M6/M7		[MH]⁺ at m/z 299. Radioactivity recovery from excreta was 0.1 ± 0.1% of dose. Structure is based on MS/MS data acquired from dog urine and that showed a fragment at m/z 205, which indicated two hydroxylations in the propyl thiazolidine moiety and a peak at m/z 225 showing the loss of the dihydroxy propyl group as hydroxy acetone (74 Da), similar to the fragmentation of M7. The in vitro experiments showed that M7 was a precursor of B23.
M11	19.9	Tri-OH (all norbornyl)	B13		[MH]⁺ at m/z 315. Radioactivity recovery from excreta was 9.3 ± 4.5% of dose. The 48-Da mass shift indicates three sequential hydroxylations. Product ions of m/z 315 at m/z 297, 279, 269, and 251 indicate that all three hydroxylations occurred on the norbornane ring. Fragment ions at m/z 173 and 97 indicated an intact propyl thiazolidinone moiety. This is the only metabolite that indicated there is a third possible hydroxylation site in the norbornane ring. This metabolite was formed from M1 in rats in vivo.
B30	23.2	4-Exo-OSO₃H^{^a}	M1		[MH]⁺ at m/z 363. Radioactivity recovery from excreta was 6.0 ± 0.8% of dose. MS features: a neutral loss of 80 to 283 Da indicated a hydroxysulfate conjugate. The fragment at m/z 173 indicated that the propyl thiazolidinone was intact and that m/z 111 was consistent with hydroxylation on the norbornane. This metabolite was observed in bile from M1-dosed rats.
	min
B25	36.6	Norbornyl OSO₃H	Norbornyl OH		[MH]⁺ at m/z 363. Radioactivity recovery from excreta was 0.4 ± 0.8% of dose. MH⁺ at m/z 363 and an M⁺2 ¹⁴C-isotope peak at m/z 365 indicated an addition of 96 Da to AMG 221 and is consistent with a hydroxysulfate conjugate. CID spectra gave the expected loss of 80 Da to m/z 283. Fragments at m/z 173 and 131 were consistent with an unmodified propyl thiazolidinone moiety.
B14	20.6	4-Exo-OH glucuronide (M1)^{^a}	M1		[MH]⁺ at m/z 459. Radioactivity recovery from excreta was 0.2 ± 0.2% of dose. MS/MS data from rat were of low quality. The glucuronide at this retention time in dog urine afforded MS/MS data, which showed an intact propyl thiazolidinone moiety at m/z 173. This metabolite was produced by M1-dosed rats.
B17	25.2	4-Endo-OH glucuronide (M2)^{^a}	M2		[MH]⁺ at m/z 459. Radioactivity recovery from excreta was 0.2 ± 0.4% of dose. This glucuronide metabolite was also produced by M1-dosed rats. MS data and MS^3 data from M1-dosed rat bile revealed the m/z 173 ion, which indicated an intact propyl thiazolidinone moiety. The aglycone is proposed to be M2, as formed by racemization of M1 via intermediate ketone M10.
B18	27.8	OH- glucuronide	M4/M6		[MH]⁺ at m/z 459. Radioactivity recovery from excreta was 0.1 ± 0.2% of dose. MS/MS data from rat were of low quality. The glucuronide at this retention time in dog urine afforded MS/MS data and showed a hydroxylated propyl thiazolidinone aglycone fragment at m/z 189 and an intact norbornane ion at m/z 95. The absence of an m/z 225 ion from loss of acetone ruled out M7 as a precursor of the glucuronide.
B28	4.8			No structure	[MH]⁺ unknown. Radioactivity recovery from excreta was 3.3 ± 0.6% of dose. No molecular ion was apparent. High MS background and poor ionization precluded molecular weight assignment.

CHO, Chinese hamster ovary; ID, identification; t_R, retention time; P, parent (AMG 221).
↵a Complete stereo- and regiochemistry inferred by chemical synthesis (M8, M10, M1, M2, M3, M4, M6, M7) or by weight of evidence from available data (19 metabolites).