Role of aldo-keto reductase family 1 (AKR1) enzymes in human steroid metabolism.
Journal: 2014/August - Steroids
ISSN: 1878-5867
Abstract:
Human aldo-keto reductases AKR1C1-AKR1C4 and AKR1D1 play essential roles in the metabolism of all steroid hormones, the biosynthesis of neurosteroids and bile acids, the metabolism of conjugated steroids, and synthetic therapeutic steroids. These enzymes catalyze NADPH dependent reductions at the C3, C5, C17 and C20 positions on the steroid nucleus and side-chain. AKR1C1-AKR1C4 act as 3-keto, 17-keto and 20-ketosteroid reductases to varying extents, while AKR1D1 acts as the sole Δ(4)-3-ketosteroid-5β-reductase (steroid 5β-reductase) in humans. AKR1 enzymes control the concentrations of active ligands for nuclear receptors and control their ligand occupancy and trans-activation, they also regulate the amount of neurosteroids that can modulate the activity of GABAA and NMDA receptors. As such they are involved in the pre-receptor regulation of nuclear and membrane bound receptors. Altered expression of individual AKR1C genes is related to development of prostate, breast, and endometrial cancer. Mutations in AKR1C1 and AKR1C4 are responsible for sexual development dysgenesis and mutations in AKR1D1 are causative in bile-acid deficiency.
Relations:
Content
Citations
(32)
References
(107)
Diseases
(1)
Chemicals
(5)
Genes
(3)
Organisms
(1)
Processes
(1)
Affiliates
(2)
Similar articles
Articles by the same authors
Discussion board
Steroids 79: 101016/j.steroids.2013.10.012.

Role of Aldo-Keto Reductase Family 1 (AKR1) Enzymes in Human Steroid Metabolism

1. Enzymes of steroid metabolism

Imbalance in the biosynthesis and inactivation of steroids can lead to development of disease, including hormonally dependent cancer, of the breast, prostate, endometrium and ovary [1], and to diseases such as benign prostatic hyperplasia, endometriosis, cholestasis, neonatal liver failure, neurological disorders [26] and malformation or differentiation of the genitalia [7]. The metabolism of steroids involves phase I and phase II enzymes and has an important role in human health and disease. Phase I steroid biosynthetic enzymes include: monooxygenases from the cytochrome P450s superfamily (CYP), and oxidoreductases from the short-chain dehydrogenase reductase (SDR) and the aldo-keto reductase (AKR) superfamilies. An outlier are the 5α-reductases which belong to the polyprenol reductase family. The phase II conjugating enzymes include the sulfotransferases (SULTs), UDP glucuronosyl transferases (UGTs) and catechol-O-methyl transferases (COMT).

The SDR and AKR enzymes catalyze NAD(P)H dependent oxidoreductions and act at the C3, C5, C11, C17 and C20 positions of the steroid nucleus and side-chain and function as 3α/β-hydroxysteroid dehydrogenases (HSDs), 5β-reductases, 11β-HSDs, 17β-HSDs and 20α-HSDs, respectively [8]. HSDs from the SDR and AKR superfamilies differ in their protein folds, stereochemistry of hydride transfer and kinetic mechanism but share a common catalytic mechanism that has been evolutionary conserved [8,9]. Human enzymes from the AKR1 family catalyze either the biosynthesis of steroid hormones, bile acids and neurosteroids, or their inactivation and thus regulate activity and action of these mediators [8]. In addition, these enzymes act on conjugated steroids and synthetic steroids used as therapeutics [1013].

2. Human AKR1 enzymes implicated in steroid metabolism

Human AKR1 enzymes implicated in steroid metabolism are members of the AKR1C and AKR1D subfamilies. AKR1C enzymes (AKR1C1-AKR1C4) function in vivo as 3-keto-, 17-keto- and 20-ketosteroid reductases to form 3α/β, 17β- and 20α-hydroxy-metabolites to varying extents and thus metabolize a broad spectrum of natural and synthetic therapeutic steroids [14], Figure 1. These enzymes are expressed in different tissues, while AKR1C4 is mainly liver specific [14]. AKR1C enzymes share a high percentage of amino-acid sequence identity that ranges from 84% to 98%. In particular AKR1C1 and AKR1C2, differ by only seven amino-acid residues. A large number of crystal structures of human AKR1C enzymes in complex with different ligands have been solved (37 as of May 5, 2013) and are deposited in the Protein Data Bank. AKR1C enzymes share the protein fold in common with other members of the superfamily, which is the triose phosphate isomerase (TIM) barrel characterized by an alternating arrangement of α-helix and β-strand to form a β-barrel. At the back of the barrel three large loops exist which define substrate specificity. In each of the AKR1 proteins the cofactor binds in a conserved manner across the lip of the β-barrel, while the steroid lies perpendicular to the cofactor. The conserved catalytic tetrad resides at the base of the barrel and consists of Asp50, Tyr55, Lys84 and His117, where Tyr55 acts as a general acid-base [8]. These cytosolic enzymes use nicotinamide adenine dinucleotides as cofactors, which are bound in an extended anti-conformation thus facilitating 4-pro-R-hydride transfer to the acceptor group [8]. These enzymes preferably act as reductases in vivo since they have Kd values for NADPH in the nanomolar range, there is a high NADPH/NADP ratio within cells and NADPH inhibits the NAD dependent oxidation reactions at low μM concentrations [1517].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f1.jpg
Reactions catalyzed by human AKR1 enzymes

Reduction of Δ-3-ketosteroids are catalyzed by AKR1D1 to form 5β-reduced steroids or by SRD5A1/SRD5A2 to form 5α-reduced steroids. 5α/β-Reduced steroids are further reduced by AKR1C enzymes to form tetrahydrosteroids of defined stereochemistry (A). 17-Ketosteroid reduction is catalyzed mainly by AKR1C3 to form 17β-hydroxysteroids and 20-ketosteroid reduction is catalyzed mainly by AKR1C1 to form 20α-hydroxysteroids (B).

Extensive comparison of the steroid binding sites of the individual isoforms has elucidated the structural basis of substrate specificity. For example ketosteroid substrates can bind in reverse orientation so that the D-ring binds at the base of the pocket instead of the A-ring. In addition stereochemistry of the hydroxysteroid product can be inverted by either the ketosteroid being pushed across the steroid pocket by steric forces or the ketosteroid can be bound upside down so that a different face of the steroid is presented to the cofactor (Figure 2) [18, 19].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f2.jpg
Binding orientations of ketosteroids within the active sites of human AKR1 enzymes

The A-ring of 3-ketosteroid substrates can orient towards the nicotinamide ring of the cofactor to form 3α and 3β-hydroxy metabolites (A); they can bind in reverse orientation so that the D-ring binds at the base of the pocket and thus form 20α- (B); and 17β-hydroxy (C) metabolites. Different stereochemistry of the hydroxysteroid product is achieved by the ketosteroid being pushed across the steroid pocket by steric forces (Leu54 versus Val54).

In contrast to AKR1C enzymes the only human member of the AKR1D subfamily, AKR1D1, catalyzes the NADPH dependent reduction of the C4-C5 double bond in Δ ketosteroids to form A/B cis-ring junction (Figure 1). This configuration introduces a 90° bend into the planar steroid and introduces amphipathic properties into the sterol which is essential for emulsification of fats and cholesterol [20]. The AKR1D1 enzyme has the typical (α/β)8 barrel structure of the AKR superfamily with cofactor and the steroid substrate bound at the C-terminal end of the β-sheets [21]. AKR1D1 has a conserved catalytic tetrad: Tyr58, Lys87 and Glu120 and Asp53 [22] which differs from catalytic tetrad of AKR1C members, as His is replaced by Glu. A mechanism of AKR1D1 catalyzed 5β-reduction has been proposed, in which Tyr58 acts as a general acid and Glu120 facilitates enolization of the Δ double bond by acting as a superacid [23]. AKR1D1 has broad substrate specificity; it catalyzes the reduction of C18, C19, C21 and C27 Δ-ketosteroids and some synthetic drugs [23, 24]. Thus only one enzyme is required for the 5β-reduction of all Δ ketosteroids. AKR1D1 is expressed mainly in liver as might be expected by its role in bile acid biosynthesis [25, 26].

All AKR1 enzymes catalyze a sequential ordered bi bi kinetic mechanism in which there is an obligatory requirement for the cofactor to bind before the steroid substrate can bind to form the central complex. Following the chemical step, steroid product then leaves followed by the oxidized cofactor. Different steps of the reaction can be rate-limiting depending upon the substrate. The chemical step or the release of either product may contribute to rate determination. However, with substrates with high turnover numbers the rate of release of the NADP cofactor places an upper limit on the rate of reaction. The release of the NADP cofactor is a two step process in which an isomerisation event occurs to convert a E.NADP tight complex to form a E.NADP loose complex [2729].

2.1. AKR1 enzymes in androgen metabolism

In peripheral tissues, AKR1C enzymes act as 3-keto and 17-ketosteroid reductases to catalyze either the formation of the most potent androgen 5α-dihydrotestosterone (5α-DHT) or its inactivation to yield the inactive androgens 5α-androstane-3α,17β-diol (3α-Diol) or 5α-androstane-3β,17β-diol (3β-Diol) (Table 1, Figure 3). AKR1C3, also known as 17β-HSD type 5, acts as 17-ketosteroid reductase and has the highest catalytic efficiency of the AKR1C enzymes for the reduction of Δ-androstene-3,17-dione to form testosterone [30, 31]. LNCaP-cells stably transfected with AKR1C3 can recapitulate this reaction [32]. The conversion of Δ-androstene-3,17-dione to testosterone and its subsequent reduction to 5α-DHT by SRD5A1 and SDR5A2 is known as the “classical pathway” of 5α-DHT formation [33]. AKR1C3 also reduces 5α-androstane-3,17-dione to 5α-DHT which by-passes testosterone (this is known as the alternative pathway to DHT); and reduces androsterone to 3α-Diol (a backdoor precursor of 5α-DHT). Thus all pathways to the potent androgens testosterone and 5α-DHT proceed through AKR1C3. AKR1C2 and AKR1C4 also catalyze the reduction of 5α-pregnane-3,20-dione to yield 3α-hydroxy-5α-pregnane-20-one (allopregnanolone) which is the precursor of androsterone (Table 2) [34]. In this manner AKR1C2 and AKR1C4 enzymes play a critical role in the “backdoor pathway” of androgen formation in which 5α-reduction occurs at the levels of the pregnanes [33, 35].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f3.jpg
Human AKR1 enzymes involved in androgen biosynthesis via the classical, the alternative and the backdoor pathways

In prostate androgens can be synthetized from DHEA or de novo from cholesterol, where AKR1C3 catalyzes activation of androstenedione to testosterone, the reverse reaction is catalyzed by HSD17B2. Testosterone is further activated by SRD5A1 and/or SDR5A2 to form the most potent androgen 5α-dihydrotestosterone (5α-DHT), which stimulates proliferation via AR. This is the classical pathway. 5α-DHT can be formed also by an alternative pathway from 5α-androstane-3,17-dione by the action of AKR1C3. Enymes AKR1C1 and AKR1C2 catalyze inactivation of 5α-DHT to form the less potent 3α-Diol and 3β-Diol with high affinity for proapoptotic ERβ. Recent data suggest that 5α-DHT may also be formed from cholesterol via pregnenolone, progesterone, 5α-DHP, allopregnanolone and androsterone by the so called «backdoor« pathway». AKR1C2 and AKR1C4 catalyze reduction of 5α-DHP and AKR1C3 catalyzes the reduction of androsterone to from 3α-Diol. 3α-Diol is oxidized back to 5α-DHT by the action of several oxidases from the SDR superfamily including HSD17B6.

Table 1

Kinetic characteristics of the AKR1C enzymes for metabolism of androgens

EnzymeSubstrateProductKM (μM)kcat (min)kcat/KM (min mM)Reference
AKR1C15α-DHT (R)3α-Diol0.2[16]
5α-DHT (R)3β-Diol0.6[16]
Androsterone (S)21.00.29[14]
5α-androstane-3,17-dione (S)6.80.570[14]
AKR1C25α-DHT (R)3α-Diol6.63.0458[16]
5α-DHT (R)3β-Diol11.00.219[16]
5α-DHT (R)3α-Diol?4.63.8820[31]
5α-DHT (S)4.33.1714[16]
5α-androstane-3,17-dione (S)6.31.4222[14]
AKR1C35α-DHT (R)3α-Diol0.1[16]
5α-DHT (R)3β-Diol0.1[16]
5α-DHT (R)?200.313[17]
Δ-Androstene-3,17-dione (R)Testosterone13.40.965[30]
Δ-Androstene-3,17-dione (R)Testosterone6.60.224[31]
Androsterone (S)9.00.442[14]
5α-androstane-3,17-dione (S)5.00.356[14]
AKR1C45α-DHT (R)3α-Diol4.4[16]
5α-DHT (R)3β-Diol1.2[16]
5α-androstane-3,17-dione (S)1.41.81243[14]
AKR1D1Androstenedione (F)0.96.06720[23]
Testosterone (F)2.78.43120[23]
The numbers have been rounded out to keep the information clear. R, radiometric assay, S, spectrophotometric assay, F, fluorimetric assay. Products identified by radiometric assay are shown.

Steckelbroeck et al., 2004 [16]; radiometric assay: 100 mM phosphate buffer pH 7.0; 2.3 mM NADPH, 4% ACN, 37°C, 41.25 μM concentration of 5α-DHT;

Penning et al., 2000 [14]; spectrophotometric assay: 100 mM phosphate buffer, pH 7.0; 200 μM NADPH, 4% ACN, 25 °C;

Byrns et al., 2008 [31]; radiometric assay: 100 mM phosphate buffer pH 7.0, 200 μM NADPH, 4% EtOH, 37 °C;

Sharma et al., 2006 [30]; radiometric assay: 100 mM phosphate buffer pH 7.0, 2.5 mM NADPH, 2 % EtOH, 37°C;

Penning and Bryns 2009 [17]; radiometric determination

Chen et al., 2011 [23]; fluorimetric assay: 100 mM phosphate buffer, pH 6.0; 15 μM NADPH, 4% acetonitrile, 37°C.

Table 2

Kinetic characteristics of the AKR1C enzymes for metabolism of progesterone and 5α-pregnanes

EnzymeSubstrateProductKM (μM)kcat (min)kcat/KM (min mM)References
AKR1C1Progesterone (R)20α-hydroxy-4-pregnen-3-one5.70.9210[30]
Progesterone (R)20α-hydroxy-4-pregnen-3-one1.90.6300[11]
Progesterone (S)1818980[107]
5α-DHP (S)1.12.62 400[34]
20α-hydroxy-5α-pregnane-3-one (S)0.70.6860[34]
3α-hydroxy-5α-pregnane-20-one (S)2.04.42 200[34]
AKR1C2Progesterone (R)20α-hydroxy-4-pregnen-3-one7.70.230[13]
5α-DHP (S)0.60.5700[34]
5α-DHP (R)3α-hydroxy-5α-pregnane-20-one1.70.2135[108]
5α-DHP (R)20α-hydroxy-5α-pregnane-3-one3.40.272[108]
20α-hydroxy-5α-pregnane-3-one (S)0.50.5960[34]
3α-hydroxy-5α-pregnane-20-one (S)3.10.380[34]
AKR1C3Progesterone (R)20α-hydroxy-4-pregnen-3-one2.81.0370[30]
Progesterone (R)20α-hydroxy-4-pregnen-3-one200.630[17]
Progesterone (R)20α-hydroxy-4-pregnen-3-one5.60.230[11]
5α-DHP (S)-0.06-[34]
20α-hydroxy-5α-pregnane-3-one (S)2.10.52250[34]
3α-hydroxy-5α-pregnane-20-one (S)?na?[34]
AKR1C45α-DHP (S)0.61.22 000[34]
20α-hydroxy-5α-pregnane-3-one (S)1.11.31200[34]
3α-hydroxy-5α-pregnane-20-one (S)0.90.6670[34]
The numbers have been rounded out to keep the information clear. R, radiometric assay, S, spectrophotometric assay, F, fluorimetric assay. Products identified by radiometric assay are shown.

Sharma et al., 2006 [30]; radiometric assay: 100 mM phosphate buffer pH 7.0, 2.5 mM NADPH, 2% EtOH, 37 °C;

Beranič et al., 2011 [11]; radiometric assay: 100 mM phosphate buffer pH 6.5; 2.3 mM NADPH; 4% ACN, 37 °C;

Beranič et al., 2012 [13]; radiometric assay: 100 mM phosphate buffer pH 6.5; 100 μM NADPH; 4% ACN, 37 °C;

Nishizawa et al., 2000 [106]; spectrophotometric assay: 0.1 M phosphate buffer, pH 6.5, 80 μM NADPH, 37 °C;

Higaki et al., 2003 [34]; spectrophotometric assay: 100 mM phosphate buffer, pH 7.0; 0.1 mM NADPH, 25 °C

Tranger et al., 2002 [107]; radiometric assay: 100 mM phosphate buffer pH 7.0; 2 mM NADPH; 5% DMSO, 0.5 mg/ml BSA, 37 °C;

Penning and Bryns 2009 [17]; radiometric determination.

The inactivation of 5α-DHT is also catalyzed by AKR1C enzymes. AKR1C2 has the highest catalytic efficiency as a 3-ketosteroid reductase and catalyzes reduction of 5α-DHT to 3α-Diol. COS-1, PC-3 and LNCaP cells transfected with AKR1C2 were able to recapitulate this reaction [15, 16]. On the other hand, AKR1C1 forms mainly 3β-Diol, a potent agonist of ERβ which exerts anti-proliferative effects. Liver specific AKR1C4 forms 3α-Diol with the highest kcat among the AKR1C enzymes [16]. AKR1C3 is much less efficient as a 3-ketosteroid reductase and converts 5α-DHT mainly to 3α-Diol [16].

In liver, circulating testosterone is metabolized either via 5α-reductase or 5β-reductase to form 5α-DHT and 5β-DHT, respectively. Subsequent 3-ketosteroid reduction by the AKR1C enzymes yields four stereoisomeric tetrahydrosteroids which can be further conjugated by phase II enzymes. AKR1D1 has a high catalytic efficiency for the reduction of Δ-androstene-3,17-dione and testosterone to form 5β-androstane-3,17-dione and 5β-DHT, respectively (Table 1) [23]. It is noteworthy that some Δ ketosteroids show distinct substrate inhibition with AKR1D1 e.g. testosterone which can be explained by two binding modes for the steroid; productive and non-productive. In the former instance testosterone binds in the steroid cavity so that it is bound perpendicular to the cofactor and the 3-ketone is tethered to the catalytic tyrosine. In the latter instance testosterone lies across the steroid cavity and lies parallel to the cofactor where it blocks access to the channel by the other steroids [21]. Further reduction of 5β-DHT to etiocholanone is catalyzed by the AKR1C enzymes, where AKR1C4 is the most efficient enzyme, followed by AKR1C2, AKR1C3 and AKR1C1 [36].

2.2. AKR1 enzymes in estrogen and progesterone metabolism

AKR1C enzymes also act as 17-ketosteroid reductases to reduce the weak estrogen, estrone to form the potent estrogen, 17β-estradiol in peripheral tissues (Figure 4) [14]. AKR1C3 is the most efficient enzyme for this reaction (KM = 9 μM, kcat = 0.068 mM; kcat/KM = 76 min mm). In MCF-7 cells stably transfected with AKR1C3 it was found that the enzyme converts 0.1 μM estrone to 17β-estradiol faster than it catalyzes either the conversion of Δ-androstene-3,17-dione to testosterone or the conversion of progesterone to 20α-hydroxy-pregn-4-ene-3-one [37]. Other AKR1C enzymes have much lower activities for 17β-estradiol formation [14].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f4a.jpg
An external file that holds a picture, illustration, etc.
Object name is nihms537519f4b.jpg
Human AKR1 enzymes implicated in progesterone and estrogen metabolism

A) In peripheral tissues including breast and endometrium AKR1C enzymes catalyze reduction of progesterone which has high affinity for the progesterone receptor to from the the less potent progestagen 20α-hydroxy-pregn-4-ene-3-one. AKR1C enzymes also reduce 5α-DHP to yield allopregnanolone, 20α-hydroxy-5α-pregnane-3-one, and 5α-pregnane-3α,20α-diol. In the CNS allopregnanolone acts as a positive allosteric modulator of GABAA receptor, while in breast cancer 5α-preganes stimulate cell proliferation. B) Among AKR1 enzymes AKR1C3 is involved in biosynthesis of the most potent estrogen, estradiol from estrone, the reverse reaction is catalyzed by HSD17B2.

AKR1 enzymes also have important roles in metabolism of progesterone (Table 2, Figure 4). AKR1C1, but also AKR1C2 and AKR1C3 catalyze progesterone reduction to form the less potent progestagen 20α-hydroxy-pregn-4-ene-3-one [13, 14]. In liver and some peripheral tissues (e.g. placenta and myometrium) AKR1D1 catalyzes reduction of progesterone to 5β-dihydroprogesterone (5β-DHP) while the 5α-reductases SRD5A1 and SRD5A2 form 5α-dihydroprogesterone (5α-DHP) [23, 38, 39, 40]. Further metabolism of 5α/β pregnanes proceeds by AKR1C enzymes [34]. 5α-Pregnane metabolites stimulate proliferation of breast cancer cell lines [41] while 5β-pregnanes have been implicated in human parturition [40, 41].

2.3. AKR1 enzymes in neurosteroid metabolism

5α-Pregnane metabolites formed from progesterone by the action of SRD5A1 potentiate the action of γ aminobutyric acid receptor A (GABAA) agonists [42]. Thus in the absence of GABA they are ineffectual but in the presence of GABA they can potentiate chloride ion channel opening. Allopregnanolone is a potent allosteric modulator of GABAA, while 5α-pregnane-3α,20α-diol is less neuroactive [42]. All AKR1C enzymes act on 5α-pregnanes (Table 2) [34]. AKR1C1, AKR1C2 and AKR1C4 have high catalytic efficiencies for 5α-DHP reduction, where AKR1C1 preferentially forms 20α-hydroxy-5α-pregnane-3-one and AKR1C2 preferentially forms allopregnanolone. Additionally, AKR1C1 catalyzes the reduction and thus inactivation of allopregnanolone by forming 5α-pregnane-3α,20α-diol. AKR1C3 has low kcat values for the reduction of 5α-DHP and 20α-hydroxy-5α-pregnane-3-one [34, 43] The kinetic data thus suggest that in brain AKR1C2 catalyzes formation of the potent neurosteroid allopregnanolone, while AKR1C1 is involved in its inactivation [43]. AKR1C1 can regulate the cellular concentrations of allopregnanolone by preventing its formation from progesterone and by catalyzing allopregnanolone inactivation. Recently it has been shown that allopregnanolone also acts as N-methyl-D-aspartate (NMDA) receptor antagonist in rat [44], modulates NMDA receptor action, and thus induces LHRH release [45].

2.4. AKR1 enzymes in bile acid biosynthesis

AKR1D1 and AKR1C4 are predominately expressed in liver [14, 25, 26] where they play essential and sequential roles in the biosynthesis of bile acids (Figure 5) [20]. These enzymes act in concert, where AKR1D1 catalyzes 5β-reduction of bile acid precursors 7α-hydroxy-4-cholestene-3-one and 7α,12α-dihydroxy-4-cholestene-3-one to yield 7α-hydroxy-5β-cholestan-3-one and 7α,12α-dihydroxy-5β-cholestan-3-one, respectively and AKR1C4 catalyzes the further reduction of 3-keto group [20, 23]. The later reactions proceed with a higher catalytic efficiency than double bond reduction (Jin and Penning, unpublished data). Subsequent steps in bile acid biosynthesis include side chain oxidation and truncation, and conjugation of the bile acid with taurine or glycine [20].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f5.jpg
A sequential action of AKR1D1 and AKR1C4 in bile acid metabolism

In liver AKR1D1 and AKR1C4 catalyze sequential reactions in biosynthesis of chenodeoxycholic and cholic acids from precursors 7α-hydroxy-4-cholestene-3-one and 7α,12α-dihydroxy-4-cholestene-3-one. AKR1D1 catalyzes the first step, formation of 7α-hydroxy-5β-cholestan-3-one and 7α,12α-dihydroxy-5β-cholestan-3-one, while AKR1C4 catalyzes reduction of 3-keto group to 3α-hydroxy-group.

2.5. AKR1C enzymes in the metabolism of steroid conjugates

AKR1C enzymes catalyze the reduction of carbonyl groups to hydroxyl groups in the steroid molecule to enable further conjugation by sulfation or glucuronidation by phase II enzymes. It is generally accepted that phase I and phase II enzymes act sequentially. Recently AKR1C enzymes were shown to act on conjugated steroids (Table 3) [10], which suggested an alternative route where the steroid is conjugated first at the 17β-position and then reduced at the 3-position. Among AKR1C enzymes, AKR1C4 has by far the highest catalytic efficiency for reduction of 5α-DHT-17β-glucuronide or sulfate. AKR1C3 has a low kcat for this reaction. By contrast, AKR1C1 and AKR1C2 prefer sulfate to bulky glucuronide conjugates where AKR1C2 is the more efficient enzyme [10]. AKR1C1, AKR1C2 and AKR1C3 form mainly 3β-hydroxy-5α-androstane-17β-glucuronide whereas AKR1C4 preferentially forms the corresponding 3α-stereoisomer. AKR1C1 and AKR1C4 thus show unchanged stereochemical preference when compared to the reduction of the unconjugated 5α-DHT, forming 3β- and 3α-metabolites, respectively. However, the stereospecificity of AKR1C2 and AKR1C3 is reversed with 5α-DHT-17β-glucuronide forming 3β-hydroxy-5α-androstane-17β-glucuronide instead of 3α-hydroxy-5α-androstane-17β-glucuronide [10]. As longer incubation times were required to measure the weak 3-ketosteroid reductase activity of AKR1C3 for the glucuronide conjugate, the observed change in stereoselectivity may be related to its epimerase activity [10].

Table 3

Kinetic characteristics of the AKR1C enzymes for metabolism of 5α-DHT glucuronides and sulfates

EnzymeSubstrateProduct
3α: 3β alcohol*
KM (μM)kcat (min)kcat/KM (min mM)
AKR1C15α-DHT1: 34.20.7180
5α-DHT-G1: 716.10.533
5α-DHT-S1: 35.20.8140
AKR1C25α-DHT20:12.92.0680
5α-DHT-G1: 54.10.259
5α-DHT-S6: 14.21.0240
AKR1C35α-DHT1.6: 1< 0.2
5α-DHT-G1: 2.3< 0.2
5α-DHT-S1: 3< 0.2
AKR1C45α-DHT3.6: 1< 0.23.1>15000
5α-DHT-G8: 1< 0.23.3>16000
5α-DHT-S1: 1< 0.23.1>15000
The numbers have been rounded out to keep the information clear.

Jin et al. 2009 [10]; fluorimetric assay: 100 mM phosphate buffer, pH 6.0; 15 μM NADPH, 4% acetonitrile, 37 °C. Products were identified by liquid chromatography mass spectrometry.

2.6. AKR1 enzymes in metabolism of synthetic steroids

Many synthetic steroids contain the Δ or 3-keto, 20-keto-steroid structural motif and are anticipated to be substrates for the AKR1 enzymes. In support of this contention AKR1 enzymes are involved in the metabolism of tibolone and the progestins, norethynodrel and dydrogesterone (Table 4, Figure 6), as well as the synthetic glucocorticoids budesonide and flunisolide [1113, 18]

An external file that holds a picture, illustration, etc.
Object name is nihms537519f6.jpg
Synthetic steroids metabolized by AKR1 enzymes

AKR1C enzymes are involved in the metabolism of tibolone, the progestins dydrogesterone and norethynodrel and synthetic glucocorticoids budesonide and flunisolide. Tibolone and norethynodrel are metabolized to 3β-hydroxy-metabolites by AKR1C1 and AKR1C2 and to 3α-hydroxy-metabolites by AKR1C4. Dydrogesterone is metabolized by AKR1C1 and AKR1C2 to form 20α-hydroxy-metabolite. AKR1D1 and AKR1C4 metabolize glucocorticoids budesonide and flunisonide to 5β-dihydro and 3α/β-hydroxy-metabolites.

Table 4

Kinetic characteristics of the AKR1 enzymes for reduction of synthetic steroids

EnzymeSubstrateProductKM (μM)kcat (min)kcat/KM (min mM)KI (μM)References
AKR1C1Tibolone (F)3β-hydroxy tibolone0.80.91 200[12]
Norethynodrel (F)3β-hydroxynor ethynodrel2.31.0450[12]
Dydrogesterone (R)20α-OHD*2.42.8812001.9[11]
AKR1C2Tibolone (F)3β-hydroxy tibolone0.912.714 6000.7[12]
Norethynodrel (F)3β-hydroxynor ethynodrel1.58.25 6002.3[12]
Dydrogesterone (R)20α-OHD*1.67.04 4801.3/4.4*[13]
AKR1C3Tibolone (F)≫3β-hydroxy tibolone
Norethynodrel (F)3β-hydroxynor ethynodrel< 0.2[12]
Dydrogesterone (R)20α-OHD*lalala0.5[11]
AKR1C4Tibolone (F)≫ 3α-hydroxy tibolone1.01.81800[12]
Norethynodrel (F)≫ 3β-hydroxynor ethynodrel0.42.76400[12]

*20α-OHD, 20α-hydroxy, 9β,10α-pregna-4,6-diene-3-one

Jin et al., 2012 [12]; fluorimetric assay: 100 mM phosphate buffer pH 7.0, 12 mM NADPH, 0.4% MetOH, 37 °C;

Beranic et al. 2011 [11]; radiometric assay: 100 mM phosphate buffer pH 6.5, 100 μM NADPH, 4% ACN, 37 °C

Ki values for oxidation of 1-acenaphthenol and reduction of progesterone, la, low activity. Products identified by either radiometric assay or liquid chromatography mass spectrometry are shown.

Tibolone is a prodrug used for the treatment of post-menopausal symptoms. Its tissue specific effects which make it a “selective tissue estrogen activity regulator” is likely mediated by AKR metabolism. This synthetic steroid maintains bone without the undesirable side effects on the endometrium and breast. AKR1C1 and AKR1C2 form 3β-hydroxytibolone and are responsible for pro-drug activation in peripheral tissues where the 3β-hydroxymetabolite activates ERα/β. The formation of the 3β-hydroxymetabolite is preferred over the 3α-hydroxymetabolite because the presence of the 5(10)-ene and 17α-ethinyl group enables the steroid to be flipped over at the active-site relative to 5α-DHT. By contrast the mainly liver specific AKR1C4 forms the 3α-hydroxy-metabolite which is the major metabolite found in the circulation [18]. The same outcome is observed for the norethynodrel suggesting that the presence of the 7α-methyl group in tibolone does not determine the stereochemcial outcome.

The progestin dydrogesterone is used for hormone-replacement therapy, for treatment of endometriosis, menstrual disorders, corpus luteum insufficiency and habitual or threatened abortions [46]. Dydrogesterone is metabolized to the active 20α-hydroxy-metabolite, 20α-hydroxy-9β,10α-pregna-4,6-diene-3-one by AKR1C1 and AKR1C2 enzymes and acts as μM inhibitor of AKR1C3 [11, 13].

The inhaled synthetic corticosteroids budesonide and flunisolide used in anti-inflammatory drug therapy in asthma patients can be metabolized by AKR1D1 and AKR1C enzymes. These steroids also act as inhibitors of AKR1 enzymes. Interestingly, budesonide was a more potent inhibitor of the AKR1D1 and AKR1C4 enzymes in the liver than flunisolide suggesting that this inhibition might contribute to the undesirable systemic effects of budesonide [24]. More extensive studies on the roles of AKR1 enzymes on the metabolism of synthetic steroids are still required. It is anticipated that single nucleotide polymorphisms (SNPs) in the AKR1C1 and AKRID1 genes are likely to influence the metabolism of different classes of therapeutic steroids (see section 5).

2.1. AKR1 enzymes in androgen metabolism

In peripheral tissues, AKR1C enzymes act as 3-keto and 17-ketosteroid reductases to catalyze either the formation of the most potent androgen 5α-dihydrotestosterone (5α-DHT) or its inactivation to yield the inactive androgens 5α-androstane-3α,17β-diol (3α-Diol) or 5α-androstane-3β,17β-diol (3β-Diol) (Table 1, Figure 3). AKR1C3, also known as 17β-HSD type 5, acts as 17-ketosteroid reductase and has the highest catalytic efficiency of the AKR1C enzymes for the reduction of Δ-androstene-3,17-dione to form testosterone [30, 31]. LNCaP-cells stably transfected with AKR1C3 can recapitulate this reaction [32]. The conversion of Δ-androstene-3,17-dione to testosterone and its subsequent reduction to 5α-DHT by SRD5A1 and SDR5A2 is known as the “classical pathway” of 5α-DHT formation [33]. AKR1C3 also reduces 5α-androstane-3,17-dione to 5α-DHT which by-passes testosterone (this is known as the alternative pathway to DHT); and reduces androsterone to 3α-Diol (a backdoor precursor of 5α-DHT). Thus all pathways to the potent androgens testosterone and 5α-DHT proceed through AKR1C3. AKR1C2 and AKR1C4 also catalyze the reduction of 5α-pregnane-3,20-dione to yield 3α-hydroxy-5α-pregnane-20-one (allopregnanolone) which is the precursor of androsterone (Table 2) [34]. In this manner AKR1C2 and AKR1C4 enzymes play a critical role in the “backdoor pathway” of androgen formation in which 5α-reduction occurs at the levels of the pregnanes [33, 35].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f3.jpg
Human AKR1 enzymes involved in androgen biosynthesis via the classical, the alternative and the backdoor pathways

In prostate androgens can be synthetized from DHEA or de novo from cholesterol, where AKR1C3 catalyzes activation of androstenedione to testosterone, the reverse reaction is catalyzed by HSD17B2. Testosterone is further activated by SRD5A1 and/or SDR5A2 to form the most potent androgen 5α-dihydrotestosterone (5α-DHT), which stimulates proliferation via AR. This is the classical pathway. 5α-DHT can be formed also by an alternative pathway from 5α-androstane-3,17-dione by the action of AKR1C3. Enymes AKR1C1 and AKR1C2 catalyze inactivation of 5α-DHT to form the less potent 3α-Diol and 3β-Diol with high affinity for proapoptotic ERβ. Recent data suggest that 5α-DHT may also be formed from cholesterol via pregnenolone, progesterone, 5α-DHP, allopregnanolone and androsterone by the so called «backdoor« pathway». AKR1C2 and AKR1C4 catalyze reduction of 5α-DHP and AKR1C3 catalyzes the reduction of androsterone to from 3α-Diol. 3α-Diol is oxidized back to 5α-DHT by the action of several oxidases from the SDR superfamily including HSD17B6.

Table 1

Kinetic characteristics of the AKR1C enzymes for metabolism of androgens

EnzymeSubstrateProductKM (μM)kcat (min)kcat/KM (min mM)Reference
AKR1C15α-DHT (R)3α-Diol0.2[16]
5α-DHT (R)3β-Diol0.6[16]
Androsterone (S)21.00.29[14]
5α-androstane-3,17-dione (S)6.80.570[14]
AKR1C25α-DHT (R)3α-Diol6.63.0458[16]
5α-DHT (R)3β-Diol11.00.219[16]
5α-DHT (R)3α-Diol?4.63.8820[31]
5α-DHT (S)4.33.1714[16]
5α-androstane-3,17-dione (S)6.31.4222[14]
AKR1C35α-DHT (R)3α-Diol0.1[16]
5α-DHT (R)3β-Diol0.1[16]
5α-DHT (R)?200.313[17]
Δ-Androstene-3,17-dione (R)Testosterone13.40.965[30]
Δ-Androstene-3,17-dione (R)Testosterone6.60.224[31]
Androsterone (S)9.00.442[14]
5α-androstane-3,17-dione (S)5.00.356[14]
AKR1C45α-DHT (R)3α-Diol4.4[16]
5α-DHT (R)3β-Diol1.2[16]
5α-androstane-3,17-dione (S)1.41.81243[14]
AKR1D1Androstenedione (F)0.96.06720[23]
Testosterone (F)2.78.43120[23]
The numbers have been rounded out to keep the information clear. R, radiometric assay, S, spectrophotometric assay, F, fluorimetric assay. Products identified by radiometric assay are shown.

Steckelbroeck et al., 2004 [16]; radiometric assay: 100 mM phosphate buffer pH 7.0; 2.3 mM NADPH, 4% ACN, 37°C, 41.25 μM concentration of 5α-DHT;

Penning et al., 2000 [14]; spectrophotometric assay: 100 mM phosphate buffer, pH 7.0; 200 μM NADPH, 4% ACN, 25 °C;

Byrns et al., 2008 [31]; radiometric assay: 100 mM phosphate buffer pH 7.0, 200 μM NADPH, 4% EtOH, 37 °C;

Sharma et al., 2006 [30]; radiometric assay: 100 mM phosphate buffer pH 7.0, 2.5 mM NADPH, 2 % EtOH, 37°C;

Penning and Bryns 2009 [17]; radiometric determination

Chen et al., 2011 [23]; fluorimetric assay: 100 mM phosphate buffer, pH 6.0; 15 μM NADPH, 4% acetonitrile, 37°C.

Table 2

Kinetic characteristics of the AKR1C enzymes for metabolism of progesterone and 5α-pregnanes

EnzymeSubstrateProductKM (μM)kcat (min)kcat/KM (min mM)References
AKR1C1Progesterone (R)20α-hydroxy-4-pregnen-3-one5.70.9210[30]
Progesterone (R)20α-hydroxy-4-pregnen-3-one1.90.6300[11]
Progesterone (S)1818980[107]
5α-DHP (S)1.12.62 400[34]
20α-hydroxy-5α-pregnane-3-one (S)0.70.6860[34]
3α-hydroxy-5α-pregnane-20-one (S)2.04.42 200[34]
AKR1C2Progesterone (R)20α-hydroxy-4-pregnen-3-one7.70.230[13]
5α-DHP (S)0.60.5700[34]
5α-DHP (R)3α-hydroxy-5α-pregnane-20-one1.70.2135[108]
5α-DHP (R)20α-hydroxy-5α-pregnane-3-one3.40.272[108]
20α-hydroxy-5α-pregnane-3-one (S)0.50.5960[34]
3α-hydroxy-5α-pregnane-20-one (S)3.10.380[34]
AKR1C3Progesterone (R)20α-hydroxy-4-pregnen-3-one2.81.0370[30]
Progesterone (R)20α-hydroxy-4-pregnen-3-one200.630[17]
Progesterone (R)20α-hydroxy-4-pregnen-3-one5.60.230[11]
5α-DHP (S)-0.06-[34]
20α-hydroxy-5α-pregnane-3-one (S)2.10.52250[34]
3α-hydroxy-5α-pregnane-20-one (S)?na?[34]
AKR1C45α-DHP (S)0.61.22 000[34]
20α-hydroxy-5α-pregnane-3-one (S)1.11.31200[34]
3α-hydroxy-5α-pregnane-20-one (S)0.90.6670[34]
The numbers have been rounded out to keep the information clear. R, radiometric assay, S, spectrophotometric assay, F, fluorimetric assay. Products identified by radiometric assay are shown.

Sharma et al., 2006 [30]; radiometric assay: 100 mM phosphate buffer pH 7.0, 2.5 mM NADPH, 2% EtOH, 37 °C;

Beranič et al., 2011 [11]; radiometric assay: 100 mM phosphate buffer pH 6.5; 2.3 mM NADPH; 4% ACN, 37 °C;

Beranič et al., 2012 [13]; radiometric assay: 100 mM phosphate buffer pH 6.5; 100 μM NADPH; 4% ACN, 37 °C;

Nishizawa et al., 2000 [106]; spectrophotometric assay: 0.1 M phosphate buffer, pH 6.5, 80 μM NADPH, 37 °C;

Higaki et al., 2003 [34]; spectrophotometric assay: 100 mM phosphate buffer, pH 7.0; 0.1 mM NADPH, 25 °C

Tranger et al., 2002 [107]; radiometric assay: 100 mM phosphate buffer pH 7.0; 2 mM NADPH; 5% DMSO, 0.5 mg/ml BSA, 37 °C;

Penning and Bryns 2009 [17]; radiometric determination.

The inactivation of 5α-DHT is also catalyzed by AKR1C enzymes. AKR1C2 has the highest catalytic efficiency as a 3-ketosteroid reductase and catalyzes reduction of 5α-DHT to 3α-Diol. COS-1, PC-3 and LNCaP cells transfected with AKR1C2 were able to recapitulate this reaction [15, 16]. On the other hand, AKR1C1 forms mainly 3β-Diol, a potent agonist of ERβ which exerts anti-proliferative effects. Liver specific AKR1C4 forms 3α-Diol with the highest kcat among the AKR1C enzymes [16]. AKR1C3 is much less efficient as a 3-ketosteroid reductase and converts 5α-DHT mainly to 3α-Diol [16].

In liver, circulating testosterone is metabolized either via 5α-reductase or 5β-reductase to form 5α-DHT and 5β-DHT, respectively. Subsequent 3-ketosteroid reduction by the AKR1C enzymes yields four stereoisomeric tetrahydrosteroids which can be further conjugated by phase II enzymes. AKR1D1 has a high catalytic efficiency for the reduction of Δ-androstene-3,17-dione and testosterone to form 5β-androstane-3,17-dione and 5β-DHT, respectively (Table 1) [23]. It is noteworthy that some Δ ketosteroids show distinct substrate inhibition with AKR1D1 e.g. testosterone which can be explained by two binding modes for the steroid; productive and non-productive. In the former instance testosterone binds in the steroid cavity so that it is bound perpendicular to the cofactor and the 3-ketone is tethered to the catalytic tyrosine. In the latter instance testosterone lies across the steroid cavity and lies parallel to the cofactor where it blocks access to the channel by the other steroids [21]. Further reduction of 5β-DHT to etiocholanone is catalyzed by the AKR1C enzymes, where AKR1C4 is the most efficient enzyme, followed by AKR1C2, AKR1C3 and AKR1C1 [36].

2.2. AKR1 enzymes in estrogen and progesterone metabolism

AKR1C enzymes also act as 17-ketosteroid reductases to reduce the weak estrogen, estrone to form the potent estrogen, 17β-estradiol in peripheral tissues (Figure 4) [14]. AKR1C3 is the most efficient enzyme for this reaction (KM = 9 μM, kcat = 0.068 mM; kcat/KM = 76 min mm). In MCF-7 cells stably transfected with AKR1C3 it was found that the enzyme converts 0.1 μM estrone to 17β-estradiol faster than it catalyzes either the conversion of Δ-androstene-3,17-dione to testosterone or the conversion of progesterone to 20α-hydroxy-pregn-4-ene-3-one [37]. Other AKR1C enzymes have much lower activities for 17β-estradiol formation [14].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f4a.jpg
An external file that holds a picture, illustration, etc.
Object name is nihms537519f4b.jpg
Human AKR1 enzymes implicated in progesterone and estrogen metabolism

A) In peripheral tissues including breast and endometrium AKR1C enzymes catalyze reduction of progesterone which has high affinity for the progesterone receptor to from the the less potent progestagen 20α-hydroxy-pregn-4-ene-3-one. AKR1C enzymes also reduce 5α-DHP to yield allopregnanolone, 20α-hydroxy-5α-pregnane-3-one, and 5α-pregnane-3α,20α-diol. In the CNS allopregnanolone acts as a positive allosteric modulator of GABAA receptor, while in breast cancer 5α-preganes stimulate cell proliferation. B) Among AKR1 enzymes AKR1C3 is involved in biosynthesis of the most potent estrogen, estradiol from estrone, the reverse reaction is catalyzed by HSD17B2.

AKR1 enzymes also have important roles in metabolism of progesterone (Table 2, Figure 4). AKR1C1, but also AKR1C2 and AKR1C3 catalyze progesterone reduction to form the less potent progestagen 20α-hydroxy-pregn-4-ene-3-one [13, 14]. In liver and some peripheral tissues (e.g. placenta and myometrium) AKR1D1 catalyzes reduction of progesterone to 5β-dihydroprogesterone (5β-DHP) while the 5α-reductases SRD5A1 and SRD5A2 form 5α-dihydroprogesterone (5α-DHP) [23, 38, 39, 40]. Further metabolism of 5α/β pregnanes proceeds by AKR1C enzymes [34]. 5α-Pregnane metabolites stimulate proliferation of breast cancer cell lines [41] while 5β-pregnanes have been implicated in human parturition [40, 41].

2.3. AKR1 enzymes in neurosteroid metabolism

5α-Pregnane metabolites formed from progesterone by the action of SRD5A1 potentiate the action of γ aminobutyric acid receptor A (GABAA) agonists [42]. Thus in the absence of GABA they are ineffectual but in the presence of GABA they can potentiate chloride ion channel opening. Allopregnanolone is a potent allosteric modulator of GABAA, while 5α-pregnane-3α,20α-diol is less neuroactive [42]. All AKR1C enzymes act on 5α-pregnanes (Table 2) [34]. AKR1C1, AKR1C2 and AKR1C4 have high catalytic efficiencies for 5α-DHP reduction, where AKR1C1 preferentially forms 20α-hydroxy-5α-pregnane-3-one and AKR1C2 preferentially forms allopregnanolone. Additionally, AKR1C1 catalyzes the reduction and thus inactivation of allopregnanolone by forming 5α-pregnane-3α,20α-diol. AKR1C3 has low kcat values for the reduction of 5α-DHP and 20α-hydroxy-5α-pregnane-3-one [34, 43] The kinetic data thus suggest that in brain AKR1C2 catalyzes formation of the potent neurosteroid allopregnanolone, while AKR1C1 is involved in its inactivation [43]. AKR1C1 can regulate the cellular concentrations of allopregnanolone by preventing its formation from progesterone and by catalyzing allopregnanolone inactivation. Recently it has been shown that allopregnanolone also acts as N-methyl-D-aspartate (NMDA) receptor antagonist in rat [44], modulates NMDA receptor action, and thus induces LHRH release [45].

2.4. AKR1 enzymes in bile acid biosynthesis

AKR1D1 and AKR1C4 are predominately expressed in liver [14, 25, 26] where they play essential and sequential roles in the biosynthesis of bile acids (Figure 5) [20]. These enzymes act in concert, where AKR1D1 catalyzes 5β-reduction of bile acid precursors 7α-hydroxy-4-cholestene-3-one and 7α,12α-dihydroxy-4-cholestene-3-one to yield 7α-hydroxy-5β-cholestan-3-one and 7α,12α-dihydroxy-5β-cholestan-3-one, respectively and AKR1C4 catalyzes the further reduction of 3-keto group [20, 23]. The later reactions proceed with a higher catalytic efficiency than double bond reduction (Jin and Penning, unpublished data). Subsequent steps in bile acid biosynthesis include side chain oxidation and truncation, and conjugation of the bile acid with taurine or glycine [20].

An external file that holds a picture, illustration, etc.
Object name is nihms537519f5.jpg
A sequential action of AKR1D1 and AKR1C4 in bile acid metabolism

In liver AKR1D1 and AKR1C4 catalyze sequential reactions in biosynthesis of chenodeoxycholic and cholic acids from precursors 7α-hydroxy-4-cholestene-3-one and 7α,12α-dihydroxy-4-cholestene-3-one. AKR1D1 catalyzes the first step, formation of 7α-hydroxy-5β-cholestan-3-one and 7α,12α-dihydroxy-5β-cholestan-3-one, while AKR1C4 catalyzes reduction of 3-keto group to 3α-hydroxy-group.

2.5. AKR1C enzymes in the metabolism of steroid conjugates

AKR1C enzymes catalyze the reduction of carbonyl groups to hydroxyl groups in the steroid molecule to enable further conjugation by sulfation or glucuronidation by phase II enzymes. It is generally accepted that phase I and phase II enzymes act sequentially. Recently AKR1C enzymes were shown to act on conjugated steroids (Table 3) [10], which suggested an alternative route where the steroid is conjugated first at the 17β-position and then reduced at the 3-position. Among AKR1C enzymes, AKR1C4 has by far the highest catalytic efficiency for reduction of 5α-DHT-17β-glucuronide or sulfate. AKR1C3 has a low kcat for this reaction. By contrast, AKR1C1 and AKR1C2 prefer sulfate to bulky glucuronide conjugates where AKR1C2 is the more efficient enzyme [10]. AKR1C1, AKR1C2 and AKR1C3 form mainly 3β-hydroxy-5α-androstane-17β-glucuronide whereas AKR1C4 preferentially forms the corresponding 3α-stereoisomer. AKR1C1 and AKR1C4 thus show unchanged stereochemical preference when compared to the reduction of the unconjugated 5α-DHT, forming 3β- and 3α-metabolites, respectively. However, the stereospecificity of AKR1C2 and AKR1C3 is reversed with 5α-DHT-17β-glucuronide forming 3β-hydroxy-5α-androstane-17β-glucuronide instead of 3α-hydroxy-5α-androstane-17β-glucuronide [10]. As longer incubation times were required to measure the weak 3-ketosteroid reductase activity of AKR1C3 for the glucuronide conjugate, the observed change in stereoselectivity may be related to its epimerase activity [10].

Table 3

Kinetic characteristics of the AKR1C enzymes for metabolism of 5α-DHT glucuronides and sulfates

EnzymeSubstrateProduct
3α: 3β alcohol*
KM (μM)kcat (min)kcat/KM (min mM)
AKR1C15α-DHT1: 34.20.7180
5α-DHT-G1: 716.10.533
5α-DHT-S1: 35.20.8140
AKR1C25α-DHT20:12.92.0680
5α-DHT-G1: 54.10.259
5α-DHT-S6: 14.21.0240
AKR1C35α-DHT1.6: 1< 0.2
5α-DHT-G1: 2.3< 0.2
5α-DHT-S1: 3< 0.2
AKR1C45α-DHT3.6: 1< 0.23.1>15000
5α-DHT-G8: 1< 0.23.3>16000
5α-DHT-S1: 1< 0.23.1>15000
The numbers have been rounded out to keep the information clear.

Jin et al. 2009 [10]; fluorimetric assay: 100 mM phosphate buffer, pH 6.0; 15 μM NADPH, 4% acetonitrile, 37 °C. Products were identified by liquid chromatography mass spectrometry.

2.6. AKR1 enzymes in metabolism of synthetic steroids

Many synthetic steroids contain the Δ or 3-keto, 20-keto-steroid structural motif and are anticipated to be substrates for the AKR1 enzymes. In support of this contention AKR1 enzymes are involved in the metabolism of tibolone and the progestins, norethynodrel and dydrogesterone (Table 4, Figure 6), as well as the synthetic glucocorticoids budesonide and flunisolide [1113, 18]

An external file that holds a picture, illustration, etc.
Object name is nihms537519f6.jpg
Synthetic steroids metabolized by AKR1 enzymes

AKR1C enzymes are involved in the metabolism of tibolone, the progestins dydrogesterone and norethynodrel and synthetic glucocorticoids budesonide and flunisolide. Tibolone and norethynodrel are metabolized to 3β-hydroxy-metabolites by AKR1C1 and AKR1C2 and to 3α-hydroxy-metabolites by AKR1C4. Dydrogesterone is metabolized by AKR1C1 and AKR1C2 to form 20α-hydroxy-metabolite. AKR1D1 and AKR1C4 metabolize glucocorticoids budesonide and flunisonide to 5β-dihydro and 3α/β-hydroxy-metabolites.

Table 4

Kinetic characteristics of the AKR1 enzymes for reduction of synthetic steroids

EnzymeSubstrateProductKM (μM)kcat (min)kcat/KM (min mM)KI (μM)References
AKR1C1Tibolone (F)3β-hydroxy tibolone0.80.91 200[12]
Norethynodrel (F)3β-hydroxynor ethynodrel2.31.0450[12]
Dydrogesterone (R)20α-OHD*2.42.8812001.9[11]
AKR1C2Tibolone (F)3β-hydroxy tibolone0.912.714 6000.7[12]
Norethynodrel (F)3β-hydroxynor ethynodrel1.58.25 6002.3[12]
Dydrogesterone (R)20α-OHD*1.67.04 4801.3/4.4*[13]
AKR1C3Tibolone (F)≫3β-hydroxy tibolone
Norethynodrel (F)3β-hydroxynor ethynodrel< 0.2[12]
Dydrogesterone (R)20α-OHD*lalala0.5[11]
AKR1C4Tibolone (F)≫ 3α-hydroxy tibolone1.01.81800[12]
Norethynodrel (F)≫ 3β-hydroxynor ethynodrel0.42.76400[12]

*20α-OHD, 20α-hydroxy, 9β,10α-pregna-4,6-diene-3-one

Jin et al., 2012 [12]; fluorimetric assay: 100 mM phosphate buffer pH 7.0, 12 mM NADPH, 0.4% MetOH, 37 °C;

Beranic et al. 2011 [11]; radiometric assay: 100 mM phosphate buffer pH 6.5, 100 μM NADPH, 4% ACN, 37 °C

Ki values for oxidation of 1-acenaphthenol and reduction of progesterone, la, low activity. Products identified by either radiometric assay or liquid chromatography mass spectrometry are shown.

Tibolone is a prodrug used for the treatment of post-menopausal symptoms. Its tissue specific effects which make it a “selective tissue estrogen activity regulator” is likely mediated by AKR metabolism. This synthetic steroid maintains bone without the undesirable side effects on the endometrium and breast. AKR1C1 and AKR1C2 form 3β-hydroxytibolone and are responsible for pro-drug activation in peripheral tissues where the 3β-hydroxymetabolite activates ERα/β. The formation of the 3β-hydroxymetabolite is preferred over the 3α-hydroxymetabolite because the presence of the 5(10)-ene and 17α-ethinyl group enables the steroid to be flipped over at the active-site relative to 5α-DHT. By contrast the mainly liver specific AKR1C4 forms the 3α-hydroxy-metabolite which is the major metabolite found in the circulation [18]. The same outcome is observed for the norethynodrel suggesting that the presence of the 7α-methyl group in tibolone does not determine the stereochemcial outcome.

The progestin dydrogesterone is used for hormone-replacement therapy, for treatment of endometriosis, menstrual disorders, corpus luteum insufficiency and habitual or threatened abortions [46]. Dydrogesterone is metabolized to the active 20α-hydroxy-metabolite, 20α-hydroxy-9β,10α-pregna-4,6-diene-3-one by AKR1C1 and AKR1C2 enzymes and acts as μM inhibitor of AKR1C3 [11, 13].

The inhaled synthetic corticosteroids budesonide and flunisolide used in anti-inflammatory drug therapy in asthma patients can be metabolized by AKR1D1 and AKR1C enzymes. These steroids also act as inhibitors of AKR1 enzymes. Interestingly, budesonide was a more potent inhibitor of the AKR1D1 and AKR1C4 enzymes in the liver than flunisolide suggesting that this inhibition might contribute to the undesirable systemic effects of budesonide [24]. More extensive studies on the roles of AKR1 enzymes on the metabolism of synthetic steroids are still required. It is anticipated that single nucleotide polymorphisms (SNPs) in the AKR1C1 and AKRID1 genes are likely to influence the metabolism of different classes of therapeutic steroids (see section 5).

3. AKR1 enzymes control concentrations of receptor ligands

Human AKR1 enzymes catalyze the conversion of either weak ligands to form potent ligands for nuclear receptors or they are involved in the elimination of these ligands. In this manner AKR1 enzymes control the levels of potent ligands that can occupy and trans-activate nuclear receptors within endocrine target tissues. These receptors bind to DNA response elements as homo or heterodimers, recruit co-activators or co-repressors and thus regulate gene transcription [4, 14, 47]. Additionally, AKR1 enzymes also control biosynthesis of neurosteroids and modulate activity of GABA receptor A which act as a ligand-gated ion channel [14, 16, 42] and the ligand-gated and voltage dependent NMDA receptor [44, 45].

3.1. Androgen receptor

AKR1 enzymes control concentrations of active androgens and regulate the occupancy and trans-activation of androgen receptor (AR) in the prostate. AKR1C3 catalyzes the reduction of Δ-androstene-3,17-dione to form the potent androgen testosterone (Kd = 10 M for AR) and the reduction of 5α-androstane-3,17-dione to yield the most potent androgen 5α-DHT (Kd =10 M for the AR). On the other hand, AKR1C2 catalyzes inactivation of 5α-DHT to yield the less active 3α-Diol (Kd = 10 M for the AR) [15, 16]. The reverse reaction, which involves oxidation of 3α-Diol to 5α-DHT is catalyzed by the SDR members, mainly HSD17B6 also designated as SDR9C6 [48]. These reactions thus represent molecular switches that can control ligand occupancy of the AR and likely influence prostate cancer progression.

3.2. Estrogen receptors

AKR1 enzymes control concentrations of estrogens and regulate occupancy and trans-activation of estrogen receptors α and β (ERα and ERβ). AKR1C3 catalyzes reduction of estrone (Kd = 0.3 nM for ERα and ERβ) to yield the more potent estrogen 17β-estradiol, (Kd = 0.1 nM for ERα and ERβ) and thus controls activation of ERα and ERβ [37]. The reverse reaction is catalyzed by HSD17B2 also designated as SDR9C2 [49, 50]. Thus these reactions represent molecular switches that can finely tune and control ligand occupancy of the ER. AKR1C1 also regulates occupancy and trans-activation of ERβ as it catalyzes reduction of 5α-DHT to yield 3β-androstanediol (3β-Diol) a potent agonist (Kd = 10 M for ERβ) [16]. These reactions are important in estrogen sensitive tissues, such as breast and endometrium where occupancy of ERα is thought to be proproliferative and occupancy of ERβ is thought to be antiproliferative., The ERβ agonist 3β-Diol may also have important role in androgen target tissues, e.g. prostate where it promotes cell apoptosis. ERβ is also thought to play a role in the etiology of lung cancer in women [51].

3.3. Progesterone receptors

AKR1 enzymes also control local concentrations of progesterone and thus regulate occupancy and trans-activation of progesterone receptors A and B (PR-A and PR-B), both encoded by the PGR gene. AKR1C1, as well as AKR1C2 and AKR1C3, catalyze inactivation of progesterone (Kd = nM for the PR) to the less potent 20α-hydroxyl-pregn-4-ene-3-one [13, 14] (Kd = μM for the PR), while the oxidation back to progesterone is catalyzed by HSD17B2 (SDR9C2) [52]. These reactions thus act as molecular switches to regulate the PR in progesterone target tissues e.g. endometrium and breast.

3.4. GABAA Receptor

AKR1C2 catalyzes reduction of 5α-DHP to yield the neurosteroid allopregnanolone, a potent allosteric modulator of GABAA (Kd =10M), while AKR1C1 catalyzes its inactivation to form the less neuroactive 5α-pregnane-3α,20α-diol [34]. The reverse reactions are catalyzed by SDR enzymes RODH4, HSD17B6 and RDHL also designated as SDR9C8, SDR9C6 and SDR9C4, respectively [53].

3.5. Orphan Receptors

AKR1D1 and AKR1C4 have crucial roles in bile acid formation and thus regulate levels of bile acid ligands, which activate the orphan nuclear receptors the farnesoid X receptor (FXR) and the pregnane X receptor (PXR). Bile acids may also act as signaling molecules to the G protein coupled bile acid receptor (GPBAR1) [4]. In this manner they regulate their own homeostasis, and prevent the accumulation of toxic allo-bile acids in the liver which leads to liver injury and/or liver fibrosis [47]. Bile acids also stimulate the expression of inflammatory modulators (IL-8 and COX-2), and ensuing oxidative stress and DNA damage has been associated with gastro-intestinal cancers, esophagus adenocarcinoma and hepatocellular carcinoma [4]. Additionally 5β-DHP produced by the action of AKR1D1 is a ligand for the constitutive androstane receptor (CAR) and PXR, which regulate the expression of CYP3A4, which is one of the major hepatic enzymes involved in the metabolism of most drugs [54]. Activation of these receptors also affects bile acid homeostasis, but also glucose and energy metabolism [4, 47].

3.1. Androgen receptor

AKR1 enzymes control concentrations of active androgens and regulate the occupancy and trans-activation of androgen receptor (AR) in the prostate. AKR1C3 catalyzes the reduction of Δ-androstene-3,17-dione to form the potent androgen testosterone (Kd = 10 M for AR) and the reduction of 5α-androstane-3,17-dione to yield the most potent androgen 5α-DHT (Kd =10 M for the AR). On the other hand, AKR1C2 catalyzes inactivation of 5α-DHT to yield the less active 3α-Diol (Kd = 10 M for the AR) [15, 16]. The reverse reaction, which involves oxidation of 3α-Diol to 5α-DHT is catalyzed by the SDR members, mainly HSD17B6 also designated as SDR9C6 [48]. These reactions thus represent molecular switches that can control ligand occupancy of the AR and likely influence prostate cancer progression.

3.2. Estrogen receptors

AKR1 enzymes control concentrations of estrogens and regulate occupancy and trans-activation of estrogen receptors α and β (ERα and ERβ). AKR1C3 catalyzes reduction of estrone (Kd = 0.3 nM for ERα and ERβ) to yield the more potent estrogen 17β-estradiol, (Kd = 0.1 nM for ERα and ERβ) and thus controls activation of ERα and ERβ [37]. The reverse reaction is catalyzed by HSD17B2 also designated as SDR9C2 [49, 50]. Thus these reactions represent molecular switches that can finely tune and control ligand occupancy of the ER. AKR1C1 also regulates occupancy and trans-activation of ERβ as it catalyzes reduction of 5α-DHT to yield 3β-androstanediol (3β-Diol) a potent agonist (Kd = 10 M for ERβ) [16]. These reactions are important in estrogen sensitive tissues, such as breast and endometrium where occupancy of ERα is thought to be proproliferative and occupancy of ERβ is thought to be antiproliferative., The ERβ agonist 3β-Diol may also have important role in androgen target tissues, e.g. prostate where it promotes cell apoptosis. ERβ is also thought to play a role in the etiology of lung cancer in women [51].

3.3. Progesterone receptors

AKR1 enzymes also control local concentrations of progesterone and thus regulate occupancy and trans-activation of progesterone receptors A and B (PR-A and PR-B), both encoded by the PGR gene. AKR1C1, as well as AKR1C2 and AKR1C3, catalyze inactivation of progesterone (Kd = nM for the PR) to the less potent 20α-hydroxyl-pregn-4-ene-3-one [13, 14] (Kd = μM for the PR), while the oxidation back to progesterone is catalyzed by HSD17B2 (SDR9C2) [52]. These reactions thus act as molecular switches to regulate the PR in progesterone target tissues e.g. endometrium and breast.

3.4. GABAA Receptor

AKR1C2 catalyzes reduction of 5α-DHP to yield the neurosteroid allopregnanolone, a potent allosteric modulator of GABAA (Kd =10M), while AKR1C1 catalyzes its inactivation to form the less neuroactive 5α-pregnane-3α,20α-diol [34]. The reverse reactions are catalyzed by SDR enzymes RODH4, HSD17B6 and RDHL also designated as SDR9C8, SDR9C6 and SDR9C4, respectively [53].

3.5. Orphan Receptors

AKR1D1 and AKR1C4 have crucial roles in bile acid formation and thus regulate levels of bile acid ligands, which activate the orphan nuclear receptors the farnesoid X receptor (FXR) and the pregnane X receptor (PXR). Bile acids may also act as signaling molecules to the G protein coupled bile acid receptor (GPBAR1) [4]. In this manner they regulate their own homeostasis, and prevent the accumulation of toxic allo-bile acids in the liver which leads to liver injury and/or liver fibrosis [47]. Bile acids also stimulate the expression of inflammatory modulators (IL-8 and COX-2), and ensuing oxidative stress and DNA damage has been associated with gastro-intestinal cancers, esophagus adenocarcinoma and hepatocellular carcinoma [4]. Additionally 5β-DHP produced by the action of AKR1D1 is a ligand for the constitutive androstane receptor (CAR) and PXR, which regulate the expression of CYP3A4, which is one of the major hepatic enzymes involved in the metabolism of most drugs [54]. Activation of these receptors also affects bile acid homeostasis, but also glucose and energy metabolism [4, 47].

4. Expression of human AKR1 genes in disease

As AKR1 enzymes play pivotal roles in regulating steroid levels changes in gene expression are anticipated to play roles in disease pathogenesis. The AKR1C1-AKR1C4 genes are located on chromosome 10p15-p14, and the AKR1D1 gene is located on chromosome 7q32-q33. AKR1C1-AKR1C4 genes comprise 12 exons and AKR1D1 has 9 exons. Three alternatively spliced protein encoding mRNA variants were predicted for AKR1C2 and AKR1C3 and four for AKR1D1 genes in the NCBI database (Table 5). AKR1C2 transcript variants 1 and 2 translate the same full-length 323 amino acid protein, while variant 3 encodes a shorter, probably inactive protein of 139 amino acids that retains the conserved catalytic tetrad but has a truncated coenzyme binding site. Similarly, the AKR1C3 transcript variants 1 and 2 encode the same full length 323 amino acid protein but variant 3 translates to a protein of only 138 amino acids. AKR1D1 transcript variants encode four protein isoforms, variant 1 encodes the active protein of 326 amino acids, while variant 2 encodes a protein of 285 amino acids and has a compromised β-barrel structure. Variant 3 encodes a 290 amino acid protein which lacks the C-terminal flexible loop and the last helix [55]. Variant 4 encodes a short protein of 96 amino acids and consists only of exons 1 through 4. It is not known if all these truncated variants are transcribed into proteins and whether these protein isoforms would have any activity. In many instances these isoforms are predicted to be inactive since they are devoid of functional components of the protein [55].

Table 5

Single nucleotide polymorphisms within the AKR1 genes

GeneGeneTranscript InfoTranscriptgSNPcSNP
IDSize (bp)totalSYNNSYNMAF
Protein (aa)M/N/F>0.01
AKR1C11645{"type":"entrez-nucleotide","attrs":{"text":"NM_001353.5","term_id":"56121816","term_text":"NM_001353.5"}}NM_001353.51384
323
406521750/5/0none
AKR1C21646{"type":"entrez-nucleotide","attrs":{"text":"NM_001354.5","term_id":"359751395","term_text":"NM_001354.5"}}NM_001354.5V1, 3621
323
884521933/0/0Phe46Tyr, 0.051
{"type":"entrez-nucleotide","attrs":{"text":"NM_205845.2","term_id":"359751396","term_text":"NM_205845.2"}}NM_205845.2V2, 3521
323
884521933/0/0Phe46Tyr, 0.051
{"type":"entrez-nucleotide","attrs":{"text":"NM_001135241.2","term_id":"359751397","term_text":"NM_001135241.2"}}NM_001135241.2V3 1098
139
207371423/0/0Phe46Tyr, 0.051
AKR1C38644{"type":"entrez-nucleotide","attrs":{"text":"NM_003739.5","term_id":"359806986","term_text":"NM_003739.5"}}NM_003739.5V1. 1251
323
505671847/2/0His5Gln, 0.428
Arg66Gln, 0.024
Glu77Gly, 0.047
Lys104Asn, 0.132
Pro180Ser, 0.012
Arg258Cys, 0.037
{"type":"entrez-nucleotide","attrs":{"text":"NM_001253908.1","term_id":"359806989","term_text":"NM_001253908.1"}}NM_001253908.1V2, 1228
323
1515651647/2/0Arg66Gln, 0.024
Glu77Gly, 0.047
Lys104Asn, 0.132
Pro180Ser, 0.012
Arg258Cys, 0.037
{"type":"entrez-nucleotide","attrs":{"text":"NM_001253909.1","term_id":"359806997","term_text":"NM_001253909.1"}}NM_001253909.1V3 1064
138
14827917/2/0His5Gln, 0.428
Arg66Gln, 0.024
Glu77Gly, 0.047
Lys104Asn, 0.132
AKR1C41109{"type":"entrez-nucleotide","attrs":{"text":"NM_001818.3","term_id":"325652082","term_text":"NM_001818.3"}}NM_001818.31192
323
58639830/1/0Gly135Glu, 0.027
Ser145Cys, 0.104
Leu311Val, 0.104
AKR1D16718{"type":"entrez-nucleotide","attrs":{"text":"NM_005989.3","term_id":"300116269","term_text":"NM_005989.3"}}NM_005989.3V1 2710
326
8553827/0/0none
{"type":"entrez-nucleotide","attrs":{"text":"NM_001190906.1","term_id":"300116270","term_text":"NM_001190906.1"}}NM_001190906.1V2 2587
285
8553626/0/0none
{"type":"entrez-nucleotide","attrs":{"text":"NM_001190907.1","term_id":"300116272","term_text":"NM_001190907.1"}}NM_001190907.1V3 2627
290
8553424/0/1none

gSNP, SNP in gene region; cSNP, SNP within coding region; SYN, synonimous; NSY, nonsynonimous; M, missense; N, nonsense; F, frame shift; MAF, minor allele frequency; V1, V2, V3= protein variants

In diseased tissues the expression of AKR1 genes and mRNA splicing is likely to be altered leading to different transcript variants, protein isoforms and thus changed protein levels and enzymatic activities [55]. Altered AKR1 expression has been reported in hormone dependent diseases of the prostate, breast, as well as in endometrial cancers.

4.1. AKR1 enzymes and prostate cancer

AKR1C2 and AKR1C3 genes are abundantly expressed in normal prostate [14] (Table 6). Early transcript measurements on a limited number of cases showed reduction in expression of AKR1C1, AKR1C2 and AKR1C3 in primary prostate cancer versus the associated normal tissue in 8/13, 11/13 and 6/13 cases, respectively) [56]. In another study AKR1C2 transcript levels were elevated in prostate cancer epithelial cells when compared to normal epithelial cells [15]. In later studies, up-regulation of AKR1C1, AKR1C2 and AKR1C3 was seen in epithelial cells of prostate cancer as compared to normal epithelial cells [57]. Further studies revealed that AKR1C3 was the most up-regulated steroidogenic enzyme in castration-resistant prostate cancer (CRPC) and soft-tissue metastases [5860]. While expression of AKR1C1 and AKR1C2 was increased in CRPC as compared to primary PC [58], this increase failed to reach statistical significance in all studies [59, 61]. Microarray analysis in 218 prostate adenocarcinoma samples [62, 63] showed that the AKR1C1, AKR1C2, AKR1C3 and AKR1C4 genes were up-regulated in 17.5%, 10.7%, 21.4% and 25% of 28 CRPC metastatic tumors cases, respectively [63]. This altered expression of AKR1C genes was associated with a marked decrease in overall survival of CRPC patients [63]. The recent microarray study by Mitsiades et al. [64] showed marked increased AKR1C3 expression in metastatic PC (in 19 metastatic samples from 8 noncastrate and 11 castrate patients) as compared to normal prostate (29 samples) and primary PC (131 samples). Also AKR1C1, AKR1C2 and AKR1C4 were up-regulated in metastatic PC where expression of all AKR1C genes was co-regulated and negatively associated with expression of AR targeted genes [64].

Table 6

Differential expression of AKR1 genes in prostate cancer

GeneMethodSamples/Stage of diseaseRegulationReference
AKR1C1
RT-PCRPrimary PC versus normal prostate (13 samples)Decreased in 8 of 13Ji et al. [56]
RT-PCREpithelial cells of primary PC versus epithelial cells of normal prostateIncreasedBauman et al. [57]
Real-time PCRCRPC versus primary PCUnchanged
Unchanged
Mongomery et al. [59]
Locke et al. [61]
Microarray analysisCRPC versus primary PCIncreasedStanbrough et al. [58]
Microarray analysisCRPC (28 samples)Up-regulated in 18%Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]
AKR1C2
RT-PCRPrimary PC versus normal prostate (13 samples)Decreased in 11 of 13Ji et al. [56]
PC epithelial cells versus normal prostate epithelial cellsIncreasedLanisnik Rizner et al. [15]
RT-PCREpithelial cells of primary PC versus epithelial cells of normal prostateIncreasedBauman et al. [57]
Real-time PCRCRPC versus primary PCUnchanged
Unchanged
Mongomery et al. [59]
Locke et al. [61]
Microarray analysisCRPC versus primary PCIncreasedStanbrough et al. [58]
Microarray analysisCRPC (28 samples)Up-regulated in 11%Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]
IHCPC paraffin sections90 % of sections; 77/86 Chinese patients, 28/31 American patientsHuang et al. [69]
AKR1C3RT-PCRPrimary PC versus normal prostate (13 samples)Decreased in 6 of 13Ji et al. [56]
RT-PCREpithelial cells of primary PC versus epithelial cells of normal prostateIncreasedBauman et al. [57]
Real-time PCRCRPC versus primary PCUpregulatedMongomery et al. [59]
Locke et al. [61]
Microarray analysisCRPC versus primary PCUpregulatedStanbrough et al. [58]
Microarray analysisCRPC (28 samples)Up-regulated in 21 %Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]
IHCPrimary PC paraffin sections77% to 82%Nakamura et al. [66]
Fung et al. [67]
Wako et al. [68]
AKR1C4Microarray analysisCRPC (28 samples)Up-regulated in 25 %Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]

Studies on prostate cancer cell lines (LNCaP, DuCaP, VCaP) and xeonografts of human CRPC revealed that in steroid-depleted serum both AKR1C3 and AR are upregulated [60, 64, 65] and a similar response is seen with the potent anti-androgen enzalutamide. Down regulation of AKR1C3 is seen in the presence of the potent synthetic androgen R1881 [64]. Collectively, these studies show that at the transcript level AKR1C3 is consistently overexpressed in CRPC and that it is upregulated by androgen deprivation and this is part of an adaptive response of the tumor to make its own androgens in the presence of castrate levels of circulating androgens.

Immunohistochemical studies revealed positive AKR1C3 immunoreactivity in 77% to 82% of primary prostate adenocarcinoma [6668] where staining correlated with clinical stage [66, 68]. AKR1C2 was observed in 89.5% to 90.3% of prostate cancer samples [69] and expression correlated with disease status, tumor grade, cigarette smoking and expression of AR [69].

The published data (Table 6) also suggest that in the transition from metastatic PC to CRPC the individual AKR1C genes may be induced by inflammatory and proliferative agents: such as interleukin IL-6 (AKR1C2 [69]), activin A (AKR1C3 [70]) and parathyroid hormone related protein (PTHrP) (AKR1C1-AKR1C3 [71]). In this way induction of AKR1C genes may contribute to local androgen formation and prostate cancer relapse after androgen deprivation therapy. Measurement of intratumoral androgens by LC-MS/MS shows that in metastatic tumors the ratio of T: 5α-DHT increases 30-fold to show that the tumor is more dependent upon T than 5α-DHT and that this may be driven by AKR1C3 [59]. Depending on the substrate availability AKR1C2-AKR1C4 enzymes also have roles in the so called “backdoor” pathway of 5α-DHT formation [33, 35, 72] where AKR1C2 and AKR1C4 catalyze reduction of 5α-DHP to allopregnenolone and AKR1C3 catalyzes reduction of androsterone to 3α-Diol, which is oxidized to 5α-DHT predominately by the SDR member, HSD17B6 (SDR9C6) [48, 72].

4.2. AKR1 enzymes and breast cancer

In human breast, AKR1C1 and AKR1C3 are the main isoforms expressed [14] (Table 7). Decreased expression of AKR1C1, AKR1C2 and AKR1C3 in breast cancer tissue versus adjacent normal control tissue has been reported by semiquantitative RT-PCR in 11 samples and by Affymetrix gene chip analysis in 43 samples [7375]. Also Ji et al. [73] reported decreased expression of AKR1C1 and AKR1C2 in 24 paired breast cancer samples at the mRNA level and unchanged expression of AKR1C3. Recent microarray analysis in 463 invasive breast carcinoma samples [63, 76] showed that genes AKR1C1 and AKR1C3 are up-regulated in 12.3% and 11.1% of 81 PAM50 basal cases of breast cancer (the PAM50, minimal gene set is used for classification of breast cancer subtypes, [77]) while their expression was not affected in luminal or HER2 groups. This data suggests that AKR1C1 and AKR1C3 have roles in more aggressive, triple negative (ER, PR, HER2) basal-like subtype of breast cancer, where the changes in these genes decrease overall survival of these patients [63].

Table 7

Differential expression of AKR1 genes in breast cancer

GeneMethodSamples/Stage of diseaseRegulationReference
AKR1C1
RT-PCRBC versus adjacent normal prostate (11 samples)DecreasedLewis et al. [73]
Microarray analysisBC versus adjacent normal prostate (43 samples)DecreasedNi et al. [74]
Real-time PCRBC versus adjacent control tissue (24 samples)DecreasedJi et al. [75]
Microarray analysisInvasive breast cancer (463 samples, 81 basal cancer)Up-regulated in 12% of basal casesThe Cancer Genome Atlas, 2012 [76]
AKR1C2
RT-PCRBC versus adjacent normal prostate (11 samples)DecreasedLewis et al. [73]
Microarray analysisBC versus adjacent normal prostate (43 samples)DecreasedNi et al. [74]
Real-time PCRBC versus adjacent control tissue (24 samples)DecreasedJi et al. [75]
AKR1C3
RT-PCRBC versus adjacent normal prostate (11 samples)DecreasedLewis et al. [73]
Microarray analysisBC versus adjacent normal prostate (43 samples)DecreasedNi et al. [74]
Real-time PCRBC versus adjacent control tissue (24 samples)UnchangedJi et al. [75]
Microarray analysisInvasive breast cancer (463 samples, 81 basal cancer)Up-regulated in 11% of Basal cases (81)The Cancer Genome Atlas, 2012 [76]
IHCBC versus adjacent control tissue (50 samples)56% Breast cancer
92% Adjacent control tissue
Han et al. [78]
In situ hybridizationBC versus adjacent control tissue (669 paraffin sections)Increased expressionOduwole et al. [79]

AKR1C3 protein was detected in 56% of breast cancer specimens and 92% of nonmalignant adjacent tissues of 50 patients by immunohistochemistry [78]. On the other hand, Oduwole et al. [79] found increased AKR1C3 expression in breast tumor specimens as compared to adjacent normal tissue by in situ hybridization of 669 paraffin sections originating from pre- and post-menopausal women, where AKR1C3 over-expression correlated with poor survival. In another study high AKR1C3 expression in ER patients correlated with significantly increased risk of relapse later than 5 years after diagnosis [80]. These studies included patients with invasive ductal carcinoma and invasive lobular carcinoma..

The alteration in expression of AKR1C1 and AKR1C3 genes in breast cancer may also affect progesterone, estrogen and androgen metabolism and activation of the corresponding nuclear receptors. Increased expression of AKR1C1 and AKR1C3 in triple negative cancers may be related to enhanced metabolism of progesterone to 20α-hydroxy-preg-4-ene-3-one, which may prevent activation of membrane bound progesterone receptors mPR which are expressed in PR negative basal-like type of breast cancer [81]. Increased levels of AKR1C3 may affect local production of pro-proliferative estradiol, which may drive proliferation of cancerous cells via membrane-bound estrogen receptor GPER [82, 83]. AKR1C3 may also have a role in the production of testosterone, and both anti-proliferative as well as proliferative effects of androgens have been observed in breast cancer [84, 85]. Triple negative patients are currently treated with chemotherapy only, thus it is critical to find new drug targets for this molecular subtype of breast cancer [76]. Being involved in estrogen and androgen formation AKR1C3 may represent appropriate target for a subgroup of these patients.

4.3. AKR1 enzymes and endometrial disease

In human uterus the expression of AKR1C genes is relatively low, where AKR1C2 and AKR1C3 represent the major AKR1 genes expressed [14]. AKR1C genes are also expressed in endometrial cancer (EC) and ectopic endometrium of endometriosis patients (Table 8). In endometrial cancer increased AKR1C3 mRNA levels have been seen in some patients, but overall no significant difference in AKR1C3 expression in cancer versus the adjacent control endometrium was observed [86-89]. There was, however, a trend for a higher expression of AKR1C3 in G2/G3 cancers [88]. Also AKR1C1 and AKR1C2 were not significantly differentially expressed in cancer of the endometrium versus adjacent control tissue and both higher and lower levels were seen [89]. Recent microarray analysis in 373 endometrial adenocarcinoma samples [63, 90] showed that the AKR1C1 and AKR1C2 genes were up-regulated in only 3.6% of cases while AKR1C3 and AKR1C4 were up-regulated in 6.3% and 4.5% of grade 3 cases, respectively. Altered expression and activity of AKR1C1-AKR1C4 enzymes thus seems to have roles only in individual patients with more aggressive grade 3 endometrial cancer where these enzymes may modulate progesterone, estrogen and androgen metabolism. AKR1C1-AKR1C3 may decrease levels of protective progesterone and AKR1C3 may increase levels of mitogenic estradiol and more probably testosterone. The unpublished data suggests that AKR1C3 does not play an important role in estrogen biosynthesis in endometrial cancer as estrogens are mainly derived from estrone-sulfate. By contrast AKR1C3 probably has a role in androgen biosynthesis (Hevir and Lanišnik Rižner, unpublished data). Androgens have been shown to stimulate endometrial stromal migration and survival [91] and AR is up-regulated in 9.8% of endometrial adenocarcinoma grade 3 samples [63, 90].

Table 8

Differential expression of AKR1 genes in endometrial cancer and endometriosis

GeneMethodSamples/Stage of diseaseRegulationReference
AKR1C1Real time PCREC and adjacent control tissue (25 samples)No differenceŠmuc and Lanišnik Rižner [87]
Real time PCREC and adjacent control tissue (47 samples)No differenceSinreih et al. [89]
Microarray analysisEC (373 samples)Increased in 4% of G3 samplesThe Cancer Genome Atlas, 2013 [90]
AKR1C2Real time PCREC and adjacent control tissue (47 samples)No differenceSinreih et al. [89]
Microarray analysisEC (373 samples)Increased in 4% of G3 samplesThe Cancer Genome Atlas, 2013 [90]
AKR1C3Real time PCREC and adjacent control tissue (16 samples)No differenceLanišnik Rižner et al. [86]
(25 samples)Šmuc and Lanišnik Rižner [87]
(47 samples)Sinreih et al. [89]
Real time PCRTrend for higher expression G2/G3Cornell et al. [88]
Microarray analysisEC (373 samples)Increased in 6% of G3 samplesThe Cancer Genome Atlas, 2013 [90]
IHCEC and hyperplastic endometriumIncreasedIto et al. [92]
IHCEC, hyperplastic endometriumDecreasedZakharov et al. [93]
AKR1C1Real-time PCROvarian endometriosis (24 samples) versus contol endometrium (10 samples)IncreasedŠmuc and Lanišnik Rižner [87]
Real-time PCROvarian endometriosis (31 samples) versus normal endometrium (28 samples)IncreasedHevir et al. [94]
AKR1C2Real-time PCROvarian endometriosis (31 samples) versus normal endometrium (28 samples)IncreasedHevir et al. [94]
IHCOvarian endometriosis (18 sections) Control endometrium (9 sections)Increased in stromal cellsHevir et al. [94]
AKR1C3Real-time PCROvarian endometriosis (24 samples) versus contol endometrium (10 samples)IncreasedŠmuc and Lanišnik Rižner [87]
Real-time PCROvarian endometriosis (31 samples) versus normal endometrium (28 samples)IncreasedHevir et al. [94]
IHCOvarian endometriosis (18 sections) Control endometrium (9 sections)UnchangedHevir et al. [94]

At the cellular level, Ito et al. [92] reported higher AKR1C3 staining in endometrial hyperplasia and EC, but showed no experimental data for the stained sections. Later weaker AKR1C3 staining in hyperplastic endometrium and cancerous endometrium versus the normal proliferative endometrium was reported [93], but this data may be biased by different menopausal status of the case and control group.

AKR1C genes are also expressed in ectopic endometrium of endometriosis patients. Significantly higher mRNA levels of AKR1C1 and AKR1C2 were seen in ovarian endometriosis versus normal endometrium of healthy women [94]. The increased expression of AKR1C3 was borderline significant in ovarian endometriosis [94]. Immunohistochemical staining showed no significant differences in AKR1C3 in ovarian endometriosis versus control endometrium, but a significant increase was seen in AKR1C2 total scores and AKR1C2 scores in stromal cells [94]. Higher levels of AKR1C1 and AKR1C2 in endometriotic tissue may contribute to enhanced metabolism of the protective progesterone to 20α-hydroxy-pregn-4-ene-3-one, and the concurrent action of SRD5A1 may lead to the formation of the pro-proliferative 5α-pregnanes [39] (N. Beranič and T. Lanišnik Rižner, unpublished data). This suggests that the AKR1C enzymes, and especially AKR1C1 and AKR1C2, may be associated with pathogenesis of ovarian endometriosis. The expression of AKR1 genes in peritoneal and deep infiltrating endometriosis has not yet been studied.

4.1. AKR1 enzymes and prostate cancer

AKR1C2 and AKR1C3 genes are abundantly expressed in normal prostate [14] (Table 6). Early transcript measurements on a limited number of cases showed reduction in expression of AKR1C1, AKR1C2 and AKR1C3 in primary prostate cancer versus the associated normal tissue in 8/13, 11/13 and 6/13 cases, respectively) [56]. In another study AKR1C2 transcript levels were elevated in prostate cancer epithelial cells when compared to normal epithelial cells [15]. In later studies, up-regulation of AKR1C1, AKR1C2 and AKR1C3 was seen in epithelial cells of prostate cancer as compared to normal epithelial cells [57]. Further studies revealed that AKR1C3 was the most up-regulated steroidogenic enzyme in castration-resistant prostate cancer (CRPC) and soft-tissue metastases [5860]. While expression of AKR1C1 and AKR1C2 was increased in CRPC as compared to primary PC [58], this increase failed to reach statistical significance in all studies [59, 61]. Microarray analysis in 218 prostate adenocarcinoma samples [62, 63] showed that the AKR1C1, AKR1C2, AKR1C3 and AKR1C4 genes were up-regulated in 17.5%, 10.7%, 21.4% and 25% of 28 CRPC metastatic tumors cases, respectively [63]. This altered expression of AKR1C genes was associated with a marked decrease in overall survival of CRPC patients [63]. The recent microarray study by Mitsiades et al. [64] showed marked increased AKR1C3 expression in metastatic PC (in 19 metastatic samples from 8 noncastrate and 11 castrate patients) as compared to normal prostate (29 samples) and primary PC (131 samples). Also AKR1C1, AKR1C2 and AKR1C4 were up-regulated in metastatic PC where expression of all AKR1C genes was co-regulated and negatively associated with expression of AR targeted genes [64].

Table 6

Differential expression of AKR1 genes in prostate cancer

GeneMethodSamples/Stage of diseaseRegulationReference
AKR1C1
RT-PCRPrimary PC versus normal prostate (13 samples)Decreased in 8 of 13Ji et al. [56]
RT-PCREpithelial cells of primary PC versus epithelial cells of normal prostateIncreasedBauman et al. [57]
Real-time PCRCRPC versus primary PCUnchanged
Unchanged
Mongomery et al. [59]
Locke et al. [61]
Microarray analysisCRPC versus primary PCIncreasedStanbrough et al. [58]
Microarray analysisCRPC (28 samples)Up-regulated in 18%Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]
AKR1C2
RT-PCRPrimary PC versus normal prostate (13 samples)Decreased in 11 of 13Ji et al. [56]
PC epithelial cells versus normal prostate epithelial cellsIncreasedLanisnik Rizner et al. [15]
RT-PCREpithelial cells of primary PC versus epithelial cells of normal prostateIncreasedBauman et al. [57]
Real-time PCRCRPC versus primary PCUnchanged
Unchanged
Mongomery et al. [59]
Locke et al. [61]
Microarray analysisCRPC versus primary PCIncreasedStanbrough et al. [58]
Microarray analysisCRPC (28 samples)Up-regulated in 11%Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]
IHCPC paraffin sections90 % of sections; 77/86 Chinese patients, 28/31 American patientsHuang et al. [69]
AKR1C3RT-PCRPrimary PC versus normal prostate (13 samples)Decreased in 6 of 13Ji et al. [56]
RT-PCREpithelial cells of primary PC versus epithelial cells of normal prostateIncreasedBauman et al. [57]
Real-time PCRCRPC versus primary PCUpregulatedMongomery et al. [59]
Locke et al. [61]
Microarray analysisCRPC versus primary PCUpregulatedStanbrough et al. [58]
Microarray analysisCRPC (28 samples)Up-regulated in 21 %Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]
IHCPrimary PC paraffin sections77% to 82%Nakamura et al. [66]
Fung et al. [67]
Wako et al. [68]
AKR1C4Microarray analysisCRPC (28 samples)Up-regulated in 25 %Taylor et al. [62]
Microarray analysisMetastatic PCIncreasedMitsiades et al. [64]

Studies on prostate cancer cell lines (LNCaP, DuCaP, VCaP) and xeonografts of human CRPC revealed that in steroid-depleted serum both AKR1C3 and AR are upregulated [60, 64, 65] and a similar response is seen with the potent anti-androgen enzalutamide. Down regulation of AKR1C3 is seen in the presence of the potent synthetic androgen R1881 [64]. Collectively, these studies show that at the transcript level AKR1C3 is consistently overexpressed in CRPC and that it is upregulated by androgen deprivation and this is part of an adaptive response of the tumor to make its own androgens in the presence of castrate levels of circulating androgens.

Immunohistochemical studies revealed positive AKR1C3 immunoreactivity in 77% to 82% of primary prostate adenocarcinoma [6668] where staining correlated with clinical stage [66, 68]. AKR1C2 was observed in 89.5% to 90.3% of prostate cancer samples [69] and expression correlated with disease status, tumor grade, cigarette smoking and expression of AR [69].

The published data (Table 6) also suggest that in the transition from metastatic PC to CRPC the individual AKR1C genes may be induced by inflammatory and proliferative agents: such as interleukin IL-6 (AKR1C2 [69]), activin A (AKR1C3 [70]) and parathyroid hormone related protein (PTHrP) (AKR1C1-AKR1C3 [71]). In this way induction of AKR1C genes may contribute to local androgen formation and prostate cancer relapse after androgen deprivation therapy. Measurement of intratumoral androgens by LC-MS/MS shows that in metastatic tumors the ratio of T: 5α-DHT increases 30-fold to show that the tumor is more dependent upon T than 5α-DHT and that this may be driven by AKR1C3 [59]. Depending on the substrate availability AKR1C2-AKR1C4 enzymes also have roles in the so called “backdoor” pathway of 5α-DHT formation [33, 35, 72] where AKR1C2 and AKR1C4 catalyze reduction of 5α-DHP to allopregnenolone and AKR1C3 catalyzes reduction of androsterone to 3α-Diol, which is oxidized to 5α-DHT predominately by the SDR member, HSD17B6 (SDR9C6) [48, 72].

4.2. AKR1 enzymes and breast cancer

In human breast, AKR1C1 and AKR1C3 are the main isoforms expressed [14] (Table 7). Decreased expression of AKR1C1, AKR1C2 and AKR1C3 in breast cancer tissue versus adjacent normal control tissue has been reported by semiquantitative RT-PCR in 11 samples and by Affymetrix gene chip analysis in 43 samples [7375]. Also Ji et al. [73] reported decreased expression of AKR1C1 and AKR1C2 in 24 paired breast cancer samples at the mRNA level and unchanged expression of AKR1C3. Recent microarray analysis in 463 invasive breast carcinoma samples [63, 76] showed that genes AKR1C1 and AKR1C3 are up-regulated in 12.3% and 11.1% of 81 PAM50 basal cases of breast cancer (the PAM50, minimal gene set is used for classification of breast cancer subtypes, [77]) while their expression was not affected in luminal or HER2 groups. This data suggests that AKR1C1 and AKR1C3 have roles in more aggressive, triple negative (ER, PR, HER2) basal-like subtype of breast cancer, where the changes in these genes decrease overall survival of these patients [63].

Table 7

Differential expression of AKR1 genes in breast cancer

GeneMethodSamples/Stage of diseaseRegulationReference
AKR1C1
RT-PCRBC versus adjacent normal prostate (11 samples)DecreasedLewis et al. [73]
Microarray analysisBC versus adjacent normal prostate (43 samples)DecreasedNi et al. [74]
Real-time PCRBC versus adjacent control tissue (24 samples)DecreasedJi et al. [75]
Microarray analysisInvasive breast cancer (463 samples, 81 basal cancer)Up-regulated in 12% of basal casesThe Cancer Genome Atlas, 2012 [76]
AKR1C2
RT-PCRBC versus adjacent normal prostate (11 samples)DecreasedLewis et al. [73]
Microarray analysisBC versus adjacent normal prostate (43 samples)DecreasedNi et al. [74]
Real-time PCRBC versus adjacent control tissue (24 samples)DecreasedJi et al. [75]
AKR1C3
RT-PCRBC versus adjacent normal prostate (11 samples)DecreasedLewis et al. [73]
Microarray analysisBC versus adjacent normal prostate (43 samples)DecreasedNi et al. [74]
Real-time PCRBC versus adjacent control tissue (24 samples)UnchangedJi et al. [75]
Microarray analysisInvasive breast cancer (463 samples, 81 basal cancer)Up-regulated in 11% of Basal cases (81)The Cancer Genome Atlas, 2012 [76]
IHCBC versus adjacent control tissue (50 samples)56% Breast cancer
92% Adjacent control tissue
Han et al. [78]
In situ hybridizationBC versus adjacent control tissue (669 paraffin sections)Increased expressionOduwole et al. [79]

AKR1C3 protein was detected in 56% of breast cancer specimens and 92% of nonmalignant adjacent tissues of 50 patients by immunohistochemistry [78]. On the other hand, Oduwole et al. [79] found increased AKR1C3 expression in breast tumor specimens as compared to adjacent normal tissue by in situ hybridization of 669 paraffin sections originating from pre- and post-menopausal women, where AKR1C3 over-expression correlated with poor survival. In another study high AKR1C3 expression in ER patients correlated with significantly increased risk of relapse later than 5 years after diagnosis [80]. These studies included patients with invasive ductal carcinoma and invasive lobular carcinoma..

The alteration in expression of AKR1C1 and AKR1C3 genes in breast cancer may also affect progesterone, estrogen and androgen metabolism and activation of the corresponding nuclear receptors. Increased expression of AKR1C1 and AKR1C3 in triple negative cancers may be related to enhanced metabolism of progesterone to 20α-hydroxy-preg-4-ene-3-one, which may prevent activation of membrane bound progesterone receptors mPR which are expressed in PR negative basal-like type of breast cancer [81]. Increased levels of AKR1C3 may affect local production of pro-proliferative estradiol, which may drive proliferation of cancerous cells via membrane-bound estrogen receptor GPER [82, 83]. AKR1C3 may also have a role in the production of testosterone, and both anti-proliferative as well as proliferative effects of androgens have been observed in breast cancer [84, 85]. Triple negative patients are currently treated with chemotherapy only, thus it is critical to find new drug targets for this molecular subtype of breast cancer [76]. Being involved in estrogen and androgen formation AKR1C3 may represent appropriate target for a subgroup of these patients.

4.3. AKR1 enzymes and endometrial disease

In human uterus the expression of AKR1C genes is relatively low, where AKR1C2 and AKR1C3 represent the major AKR1 genes expressed [14]. AKR1C genes are also expressed in endometrial cancer (EC) and ectopic endometrium of endometriosis patients (Table 8). In endometrial cancer increased AKR1C3 mRNA levels have been seen in some patients, but overall no significant difference in AKR1C3 expression in cancer versus the adjacent control endometrium was observed [86-89]. There was, however, a trend for a higher expression of AKR1C3 in G2/G3 cancers [88]. Also AKR1C1 and AKR1C2 were not significantly differentially expressed in cancer of the endometrium versus adjacent control tissue and both higher and lower levels were seen [89]. Recent microarray analysis in 373 endometrial adenocarcinoma samples [63, 90] showed that the AKR1C1 and AKR1C2 genes were up-regulated in only 3.6% of cases while AKR1C3 and AKR1C4 were up-regulated in 6.3% and 4.5% of grade 3 cases, respectively. Altered expression and activity of AKR1C1-AKR1C4 enzymes thus seems to have roles only in individual patients with more aggressive grade 3 endometrial cancer where these enzymes may modulate progesterone, estrogen and androgen metabolism. AKR1C1-AKR1C3 may decrease levels of protective progesterone and AKR1C3 may increase levels of mitogenic estradiol and more probably testosterone. The unpublished data suggests that AKR1C3 does not play an important role in estrogen biosynthesis in endometrial cancer as estrogens are mainly derived from estrone-sulfate. By contrast AKR1C3 probably has a role in androgen biosynthesis (Hevir and Lanišnik Rižner, unpublished data). Androgens have been shown to stimulate endometrial stromal migration and survival [91] and AR is up-regulated in 9.8% of endometrial adenocarcinoma grade 3 samples [63, 90].

Table 8

Differential expression of AKR1 genes in endometrial cancer and endometriosis

GeneMethodSamples/Stage of diseaseRegulationReference
AKR1C1Real time PCREC and adjacent control tissue (25 samples)No differenceŠmuc and Lanišnik Rižner [87]
Real time PCREC and adjacent control tissue (47 samples)No differenceSinreih et al. [89]
Microarray analysisEC (373 samples)Increased in 4% of G3 samplesThe Cancer Genome Atlas, 2013 [90]
AKR1C2Real time PCREC and adjacent control tissue (47 samples)No differenceSinreih et al. [89]
Microarray analysisEC (373 samples)Increased in 4% of G3 samplesThe Cancer Genome Atlas, 2013 [90]
AKR1C3Real time PCREC and adjacent control tissue (16 samples)No differenceLanišnik Rižner et al. [86]
(25 samples)Šmuc and Lanišnik Rižner [87]
(47 samples)Sinreih et al. [89]
Real time PCRTrend for higher expression G2/G3Cornell et al. [88]
Microarray analysisEC (373 samples)Increased in 6% of G3 samplesThe Cancer Genome Atlas, 2013 [90]
IHCEC and hyperplastic endometriumIncreasedIto et al. [92]
IHCEC, hyperplastic endometriumDecreasedZakharov et al. [93]
AKR1C1Real-time PCROvarian endometriosis (24 samples) versus contol endometrium (10 samples)IncreasedŠmuc and Lanišnik Rižner [87]
Real-time PCROvarian endometriosis (31 samples) versus normal endometrium (28 samples)IncreasedHevir et al. [94]
AKR1C2Real-time PCROvarian endometriosis (31 samples) versus normal endometrium (28 samples)IncreasedHevir et al. [94]
IHCOvarian endometriosis (18 sections) Control endometrium (9 sections)Increased in stromal cellsHevir et al. [94]
AKR1C3Real-time PCROvarian endometriosis (24 samples) versus contol endometrium (10 samples)IncreasedŠmuc and Lanišnik Rižner [87]
Real-time PCROvarian endometriosis (31 samples) versus normal endometrium (28 samples)IncreasedHevir et al. [94]
IHCOvarian endometriosis (18 sections) Control endometrium (9 sections)UnchangedHevir et al. [94]

At the cellular level, Ito et al. [92] reported higher AKR1C3 staining in endometrial hyperplasia and EC, but showed no experimental data for the stained sections. Later weaker AKR1C3 staining in hyperplastic endometrium and cancerous endometrium versus the normal proliferative endometrium was reported [93], but this data may be biased by different menopausal status of the case and control group.

AKR1C genes are also expressed in ectopic endometrium of endometriosis patients. Significantly higher mRNA levels of AKR1C1 and AKR1C2 were seen in ovarian endometriosis versus normal endometrium of healthy women [94]. The increased expression of AKR1C3 was borderline significant in ovarian endometriosis [94]. Immunohistochemical staining showed no significant differences in AKR1C3 in ovarian endometriosis versus control endometrium, but a significant increase was seen in AKR1C2 total scores and AKR1C2 scores in stromal cells [94]. Higher levels of AKR1C1 and AKR1C2 in endometriotic tissue may contribute to enhanced metabolism of the protective progesterone to 20α-hydroxy-pregn-4-ene-3-one, and the concurrent action of SRD5A1 may lead to the formation of the pro-proliferative 5α-pregnanes [39] (N. Beranič and T. Lanišnik Rižner, unpublished data). This suggests that the AKR1C enzymes, and especially AKR1C1 and AKR1C2, may be associated with pathogenesis of ovarian endometriosis. The expression of AKR1 genes in peritoneal and deep infiltrating endometriosis has not yet been studied.

5. Genetics of human AKR1 enzymes

Inherited or SNP allelic variants in AKR1 genes may affect the metabolism of exogenous and endogenous steroids and may also contribute to the development of pathophysiological processes.

5.1. Inherited Mutations

Inherited mutations in AKR1D1 have been studied in patients with bile-acid deficiency that develops in infants and is characterized by reduced formation of bile acids and accumulation of Δ-3-keto and 5α-reduced (allo-bile acids) which are hepatotoxic. This deficiency is exacerbated by the inability to from sufficient bile acids to feed-back and inhibit the rate determining step in the pathway, 7α-hydroxylase (CYP7A1), and thus the formation of allo-bile acids will be further elevated [9599]. Defects in bile acid synthesis are rare but can lead to cholestasis and neo-natal liver failure and can be lethal if untreated. In addition to missense mutations at Arg50* and 511 T deletion leading to truncated inactive proteins, seven point mutations have been identified in patients with 5β-reductase deficiency: Leu106Phe, Pro133Arg, Gly223Glu, Pro198Leu, Asp241Val, Arg261Cys and Arg266Glu [9597, 99, 100]. Except Pro133Arg, these mutations reside in domains highly conserved across AKR1D1 homologues and are not involved in catalysis, coenzyme or substrate binding [21, 101]. Mutant Pro133Arg was prepared in a recombinant form and showed a 12-fold decreased catalytic efficiency for testosterone reduction [101]. Expression of the mutated enzymes in HEK293 cells revealed lower expression levels of Leu106Phe, and Arg261Cys probably due to decreased stability, and lower enzymatic activity and expression levels of both Pro198Leu and Gly223Glu. The estimated Km value for Pro133Arg was 10-fold higher and Vmax about 2-fold lower as compared to the wild type enzyme for the bile acid precursor, 7α-hydroxy-cholesten-3-one [101]. The mutants thus showed lower 5β-reductase activity as compared to wild type which is in agreement with clinical observations of bile acid deficiency [98, 101].

Recently several inherited mutations in the AKR1C2 and AKR1C4 genes have been identified in two Swiss families which were associated with disordered sexual dysgenesis (DSD). Proper sexual development in male fetus requires synthesis of 5α-DHT which is crucial for differentiation of male external genitalia. Genetic males with incompletely developed (ambiguous) genitalia usually have disorders in the classical pathway of androgen biosynthesis which proceeds from cholesterol via pregnenolone to form dehydroepiandrosterone and Δ-androstene-3,17-dione [33, 102]. In this pathway AKR1C3 converts Δ-androstene-3,17-dione to testosterone, and in the alternative pathway it converts 5α-androstane-3,17-dione to 5α-DHT [102]. However, mutations in AKR1C2 and AKR1C4 in these patients suggest that in fetal development the backdoor pathway [33, 102], which involves the conversion of progesterone into 5α-DHT plays a crucial role in androgen biosynthesis. In this pathway, AKR1C2 and AKR1C4 catalyze the reduction of 5α-pregnane-3,20-dione to yield allopregnanolone which is the precursor of androsterone. Androsterone is converted to 3α-Diol by AKR1C3 which is the immediate precursor of 5α-DHT. In the first family identified with DSD, missense AKR1C2 muations: Ile79Val, His906Glu, and Asn300Thr were identified in the affected individuals. In addition to these AKR1C2 mutations these individuals also carried a mutation causing aberrant splicing of the AKR1C4 gene (Val29_Lys84del). In the second family recombination between AKR1C1 and AKR1C3 was detected and a missense AKR1C2 mutation His222Gln was identified. The mutants Ile79Val, His90Gln, Asn300Thr and His222Gln have been prepared in recombinant forms or over-expressed in COS-1 cells and showed reduced enzymatic activities [103]. The authors observed AKR1C2 and AKR1C4 expression in fetal testes thus they concluded that these enzymes are needed for fetal 5α-DHT synthesis via the backdoor pathway and normal development of the genitalia [103].

5.2. Population Genetics

A large number of synonymous and nonsynonymous single nucleotide polymorphisms (SNPs) have been identified in AKR1 genes (Table 5). However, SNP databases are not always complete and accurate. Based on a bioinformatics approach it has been predicted that 8.3% of biallelic, coding SNPs in the dbSNP database represent single nucleotide differences (SNDs) [104]. Among all AKR1 SNPs included in the SNP database there are currently only 5 SNPs with a minor allelic frequency (MAF) greater than 0.05. These correspond to Phe46Tyr in AKR1C2 with a MAF = 0.051; His5Gln and Lys104Asn in AKR1C3 with MAF of 0.428 and 0.132, respectfully; and Ser145Cys and Leu311Val in AKR1C4, both with MAF of 0.104. So far SNPs in AKR1C1, AKR1C2 and AKR1C4 have been studied at the protein level and Phe46Tyr has been associated with disease [9597, 99, 103, 104].

In AKR1C2 Takahashi et al. functionally characterized 11 nonsynonymous SNPs known at the time but later bioinformatics studies showed that only 5 of these represent real SNPs: Phe46Tyr, Vall11Ala, Lys179Gln, Lys185Gln and Arg258Cys [104]. Phe46Tyr had a 2.5-fold reduced enzymatic activity versus wild-type enzyme (Vmax) and two SNPs, Lys185Glu and Arg258Cys had significantly lower apparent Km values, when the reduction of 5α-DHT to 3α-Diol was followed [105]. The authors observed that the frequency of Phe46Tyr SNP parallelled the risk for prostate cancer in different populations (individuals of African descent 15%, Europeans 5.9%, and undetected in the Asian population). Lower AKR1C2 activity in the Phe46Tyr carriers may increase 5α-DHT levels and may thus stimulate prostate cell proliferation. Homology modeling explained the decreased activity in the Phe46Tyr variant since this changed the interaction between the enzyme and coenzyme and could lead to weak cofactor binding [106]. For the other AKR1C1/2 SNPs (Val111Ala, Lys179Gln, Lys185Gln and Arg285Cys) it was rationalized that they are unlikely to affect enzyme activities as they are located on the periphery, and distant from the active site.

5.1. Inherited Mutations

Inherited mutations in AKR1D1 have been studied in patients with bile-acid deficiency that develops in infants and is characterized by reduced formation of bile acids and accumulation of Δ-3-keto and 5α-reduced (allo-bile acids) which are hepatotoxic. This deficiency is exacerbated by the inability to from sufficient bile acids to feed-back and inhibit the rate determining step in the pathway, 7α-hydroxylase (CYP7A1), and thus the formation of allo-bile acids will be further elevated [9599]. Defects in bile acid synthesis are rare but can lead to cholestasis and neo-natal liver failure and can be lethal if untreated. In addition to missense mutations at Arg50* and 511 T deletion leading to truncated inactive proteins, seven point mutations have been identified in patients with 5β-reductase deficiency: Leu106Phe, Pro133Arg, Gly223Glu, Pro198Leu, Asp241Val, Arg261Cys and Arg266Glu [9597, 99, 100]. Except Pro133Arg, these mutations reside in domains highly conserved across AKR1D1 homologues and are not involved in catalysis, coenzyme or substrate binding [21, 101]. Mutant Pro133Arg was prepared in a recombinant form and showed a 12-fold decreased catalytic efficiency for testosterone reduction [101]. Expression of the mutated enzymes in HEK293 cells revealed lower expression levels of Leu106Phe, and Arg261Cys probably due to decreased stability, and lower enzymatic activity and expression levels of both Pro198Leu and Gly223Glu. The estimated Km value for Pro133Arg was 10-fold higher and Vmax about 2-fold lower as compared to the wild type enzyme for the bile acid precursor, 7α-hydroxy-cholesten-3-one [101]. The mutants thus showed lower 5β-reductase activity as compared to wild type which is in agreement with clinical observations of bile acid deficiency [98, 101].

Recently several inherited mutations in the AKR1C2 and AKR1C4 genes have been identified in two Swiss families which were associated with disordered sexual dysgenesis (DSD). Proper sexual development in male fetus requires synthesis of 5α-DHT which is crucial for differentiation of male external genitalia. Genetic males with incompletely developed (ambiguous) genitalia usually have disorders in the classical pathway of androgen biosynthesis which proceeds from cholesterol via pregnenolone to form dehydroepiandrosterone and Δ-androstene-3,17-dione [33, 102]. In this pathway AKR1C3 converts Δ-androstene-3,17-dione to testosterone, and in the alternative pathway it converts 5α-androstane-3,17-dione to 5α-DHT [102]. However, mutations in AKR1C2 and AKR1C4 in these patients suggest that in fetal development the backdoor pathway [33, 102], which involves the conversion of progesterone into 5α-DHT plays a crucial role in androgen biosynthesis. In this pathway, AKR1C2 and AKR1C4 catalyze the reduction of 5α-pregnane-3,20-dione to yield allopregnanolone which is the precursor of androsterone. Androsterone is converted to 3α-Diol by AKR1C3 which is the immediate precursor of 5α-DHT. In the first family identified with DSD, missense AKR1C2 muations: Ile79Val, His906Glu, and Asn300Thr were identified in the affected individuals. In addition to these AKR1C2 mutations these individuals also carried a mutation causing aberrant splicing of the AKR1C4 gene (Val29_Lys84del). In the second family recombination between AKR1C1 and AKR1C3 was detected and a missense AKR1C2 mutation His222Gln was identified. The mutants Ile79Val, His90Gln, Asn300Thr and His222Gln have been prepared in recombinant forms or over-expressed in COS-1 cells and showed reduced enzymatic activities [103]. The authors observed AKR1C2 and AKR1C4 expression in fetal testes thus they concluded that these enzymes are needed for fetal 5α-DHT synthesis via the backdoor pathway and normal development of the genitalia [103].

5.2. Population Genetics

A large number of synonymous and nonsynonymous single nucleotide polymorphisms (SNPs) have been identified in AKR1 genes (Table 5). However, SNP databases are not always complete and accurate. Based on a bioinformatics approach it has been predicted that 8.3% of biallelic, coding SNPs in the dbSNP database represent single nucleotide differences (SNDs) [104]. Among all AKR1 SNPs included in the SNP database there are currently only 5 SNPs with a minor allelic frequency (MAF) greater than 0.05. These correspond to Phe46Tyr in AKR1C2 with a MAF = 0.051; His5Gln and Lys104Asn in AKR1C3 with MAF of 0.428 and 0.132, respectfully; and Ser145Cys and Leu311Val in AKR1C4, both with MAF of 0.104. So far SNPs in AKR1C1, AKR1C2 and AKR1C4 have been studied at the protein level and Phe46Tyr has been associated with disease [9597, 99, 103, 104].

In AKR1C2 Takahashi et al. functionally characterized 11 nonsynonymous SNPs known at the time but later bioinformatics studies showed that only 5 of these represent real SNPs: Phe46Tyr, Vall11Ala, Lys179Gln, Lys185Gln and Arg258Cys [104]. Phe46Tyr had a 2.5-fold reduced enzymatic activity versus wild-type enzyme (Vmax) and two SNPs, Lys185Glu and Arg258Cys had significantly lower apparent Km values, when the reduction of 5α-DHT to 3α-Diol was followed [105]. The authors observed that the frequency of Phe46Tyr SNP parallelled the risk for prostate cancer in different populations (individuals of African descent 15%, Europeans 5.9%, and undetected in the Asian population). Lower AKR1C2 activity in the Phe46Tyr carriers may increase 5α-DHT levels and may thus stimulate prostate cell proliferation. Homology modeling explained the decreased activity in the Phe46Tyr variant since this changed the interaction between the enzyme and coenzyme and could lead to weak cofactor binding [106]. For the other AKR1C1/2 SNPs (Val111Ala, Lys179Gln, Lys185Gln and Arg285Cys) it was rationalized that they are unlikely to affect enzyme activities as they are located on the periphery, and distant from the active site.

6. Future Directions

AKR1 enzymes play important roles in metabolism of androgens, estrogens, progesterone, glucococorticoids, neurosteroids, conjugated steroids, and in the biosynthesis of bile acids. Although AKR1 activities towards representative substrates from these steroid classes have been studied the list is far from comprehensive. A more complete understanding of substrate specificities of AKR1 enzymes and the products they generate may reveal additional physiological and/or pathophysiological roles for these enzymes.

Synthetic steroids found in oral contraceptives, hormone replacement therapies and steroid anti-inflammatory drugs are metabolized by AKR1C enzymes. In this manner AKR1C enzymes may affect treatment of a particular disease or condition and susceptibility to therapy. For instance; the prodrug tibolone is metabolized by AKR1C enzymes to active compounds and inhaled corticosteroids inhibit AKR1D1 and AKR1C4, which may disrupt endogenous steroid metabolism in liver. Other synthetic steroids, e.g. dydrogesterone, and the glucocorticoids budesonide inhibit AKR1 enzymes and may thus interfere with the normal physiological roles of these enzymes. The ability of AKR1C enzymes to metabolize other steroid based drugs needs to be examined.

Steroid preferences of AKR1 enzymes are characterized by determination of their kinetic properties in vitro, where catalytic efficiencies (kcat/KM values) are compared, and products of the reactions are identified, preferably by liquid chromatography mass spectrometry. However, in human tissues cellular levels of particular isoforms should be taken into consideration, as they determine the physiological significance of a particular enzyme in that context. It is thus of paramount importance to determine the cellular levels of AKR1 enzymes in diseased versus normal tissue and how these levels are regulated. It is likely that different ratios of AKR1 enzyme expression may tune ligand availability for nuclear and membrane bound receptors.

The physiological and pathophysiological roles of AKR1 enzymes are currently not known in every instance. A functional genomics approach would contribute enormously. After studying enzymatic activities of recombinant AKR1 enzymes (substrate specificity, catalytic efficiency, product profiling), their activities in model or transfected cell lines, and their ability to alter trans-activation of nuclear receptors should be examined. Proof-of-principle experiments in human cell lines using siRNA or sh-RNA approaches are needed to clarify the pathophysiological roles of AKR1 enzymes.

Altered regulation of AKR1 gene transcription, including epigenetic modifications, transcriptional regulation (induction, repression, alternative splicing) should be further investigated. AKR1C genes are up-regulated by androgen deprivation in prostate cancer but it is not known how AKR1C genes are regulated by hormone deprivation and whether this would be seen in other malignancies.

Acknowledgments

This work was supported by grants 1R01-DK47015, 1R01-CA90744 and P30-ES013508 to T.M.P. from the National Institutes of Health, J3-4135 and SLO-USA grants from Slovenian Research Agency, and a Fulbright Grant from Council of International Exchange of Scholars to T.L.R.

Institute of Biochemistry, Faculty of Medicine, University of Ljubljana, Slovenia
Center of Excellence in Environmental Toxicology, Department of Pharmacology, Perelman School of Medicine, University of Pennsylvania, Philadelphia PA, USA
Tea Lanišnik Rižner: is.jl-inu.fm@renzir-kinsinal.aet; Trevor M. Penning: ude.nnepu@gninnep
Corresponding authors: Tea Lanišnik Rižner, Institute of Biochemistry, Faculty of Medicine University of Ljubljana, Vrazov trg 2, SI 1000 Ljubljana Slovenia, Phone: +386-1543-7657, Fax: +386-1543-7640, is.jl-inu.fm@renzir-kinsinal.aet. Trevor M. Penning, Center of Excellence in Environmental Toxicology, Department of Pharmacology, Perelman School of Medicine University of Pennsylvania, 1315 BRBII/III, 421 Curie Blvd, Philadelphia, PA 19104-6160, Phone: +215-898-9445, Fax: +215-573-0200, ude.nnepu@gninnep
Tea Lanišnik Rižner: is.jl-inu.fm@renzir-kinsinal.aet; Trevor M. Penning: ude.nnepu@gninnep
Publisher's Disclaimer

Abstract

Human aldo-keto reductases AKR1C1-AKR1C4 and AKR1D1 play essential roles in the metabolism of all steroid hormones, the biosynthesis of neurosteroids and bile acids, the metabolism of conjugated steroids, and synthetic therapeutic steroids. These enzymes catalyze NADPH dependent reductions at the C3, C5, C17 and C20 positions on the steroid nucleus and side-chain. AKR1C1-AKR1C4 act as 3-keto, 17-keto and 20-ketosteroid reductases to varying extents, while AKR1D1 acts as the sole Δ-3-ketosteroid-5β-reductase (steroid 5β-reductase) in humans. AKR1 enzymes control the concentrations of active ligands for nuclear receptors and control their ligand occupancy and trans-activation, they also regulate the amount of neurosteroids that can modulate the activity of GABAA and NMDA receptors. As such they are involved in the pre-receptor regulation of nuclear and membrane bound receptors. Altered expression of individual AKR1C genes is related to development of prostate, breast, and endometrial cancer. Mutations in AKR1C1 and AKR1C4 are responsible for sexual development dysgenesis and mutations in AKR1D1 are causative in bile-acid deficiency.

Keywords: cancer, bile-acids, hydroxysteroid dehydrogenase, inherited mutations, neurosteroids, synthetic steroids
Abstract
Highlights

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Footnotes

References

  • 1. Henderson BE, Feigelson HSHormonal carcinogenesis. Carcinogenesis. 2000;21:427–33.[PubMed][Google Scholar]
  • 2. van der Sluis TM, Meuleman EJ, van Moorselaar RJ, Bui HN, Blankenstein MA, Heijboer AC, et al. Intraprostatic testosterone and dihydrotestosterone. Part II: concentrations after androgen hormonal manipulation in men with benign prostatic hyperplasia and prostate cancer. BJU Int. 2012;109:183–8.[PubMed]
  • 3. Lanišnik Rižner TEstrogen metabolism and action in endometriosis. Mol Cell Endocrinol. 2009;307:8–18.[PubMed][Google Scholar]
  • 4. Baptissart M, Vega A, Maqdasy S, Caira F, Baron S, Lobaccaro JMA, et al Bile acid: From digestion to cancer. Biochimie. 2013;95:504–517.[PubMed][Google Scholar]
  • 5. Sundaram SS, Bove KE, Lovell MA, Sokol RJMechanisms of disease: Inborn errors of bile acid synthesis. Nat Clin Pract Gastroenterol Hepatol. 2008;5:456–68.[Google Scholar]
  • 6. Hardoy MC, Serra M, Carta MG, Contu P, Pisu MG, Biggio GIncreased neuroactive steroid concentrations in women with bipolar disorder or major depressive disorder. J Clin Psychopharmacol. 2006;26:379–384.[PubMed][Google Scholar]
  • 7. Miller WL, Auchus RJThe molecular biology, biochemistry and physiology of human steroidogenesis and its disorders. Endocrine Rev. 2011;32:81–151.[Google Scholar]
  • 8. Penning TMHuman hydroxysteroid dehydroganses and pre-receptor regulation: insights into inhibitor design and evaluation. J Steroid Biochem Mol Biol. 2011;125:46–56.[Google Scholar]
  • 9. Bennett MJ, Schlegel BP, Jez JM, Penning TM, Lewis MStructure of 3α-hydroxysteroid/dihydrodiol dehydrogenase complexed with NADP+ Biochemistry. 1996;35:10702–11.[PubMed][Google Scholar]
  • 10. Jin Y, Duan L, Lee SH, Kloosterboer HJ, Blair IA, Penning TMHuman cytosolic hydroxysteroid dehydrogenases of the aldo-ketoreductase superfamily catalyze reduction of conjugated steroids: implications for phase I and phase II steroid hormone metabolism. J Biol Chem. 2009;284:10013–22.[Google Scholar]
  • 11. Beranič N, Gobec S, Lanišnik Rižner TProgestins as inhibitors of the human 20-ketosteroid reductases, AKR1C1 and AKR1C3. Chem Biol Interact. 2011;191:227–233.[PubMed][Google Scholar]
  • 12. Jin Y, Duan L, Chen M, Penning TM, Kloosterboer HJMetabolism of the synthetic progestogen norethynodrel by human ketosteroid reductases of the aldo-keto reductase superfamily. J Steroid Biochem Mol Biol. 2012;129:139–144.[Google Scholar]
  • 13. Beranič N, Brožič P, Brus B, Sosič I, Gobec S, Lanišnik Rižner TExpression of human aldo-keto reductase 1C2 in cell lines of peritoneal endometriosis: Potential implications in metabolism of progesterone and dydrogesterone and inhibition by progestins. J Steroid Biochem Mol Biol. 2012;130:16–25.[PubMed][Google Scholar]
  • 14. Penning TM, Burczynski ME, Jez JM, Hung CF, Lin HK, Ma H, et al Human 3α-hydroxysteroid dehydrogenase isoforms (AKR1C1-AKR1C4) of the aldo-keto reductase superfamily: functional plasticity and tissue distribution reveals roles in the inactivation and formation of male and female sex hormones. Biochem J. 2000;351:67–77.[Google Scholar]
  • 15. Lanišnik Rižner T, Lin HK, Peehl DM, Steckelbroeck S, Bauman DR, Penning TMHuman type 3 3a-hydroxysteroid dehydrogenase (aldo-keto reductase 1C2) and androgen metabolism in prostate cells. Endocrinology. 2003;144:2922–2932.[PubMed][Google Scholar]
  • 16. Steckelbroeck S, Jin Y, Gopishetty S, Oyesanmi B, Penning TMHuman cytosolic 3α-hydroxysteroid dehydrogenases of the aldo-keto reductase superfamily display significant 3β-hydroxysteroid dehydrogenase activity. J Biol Chem. 2004;279:10784–10795.[PubMed][Google Scholar]
  • 17. Penning TM, Bryns MCSteroid hormone transforming aldo-keto reductases and cancer. Ann N Y Acad Sci. 2009;1155:33–42.[Google Scholar]
  • 18. Steckelbroeck S, Jin Y, Oyesanmi B, Kloosterboer HJ, Penning TMTibolone is metabolized by the 3α/3β-hydroxysteroid dehydrogenase activities of the four human isozymes of the aldo-keto reductase 1C subfamily: inversion of stereospecificity with a Δ5(10)-3-ketosteroid. Mol Pharmacol. 2004;66:1702–11.[PubMed][Google Scholar]
  • 19. Jin Y, Penning TMMolecular docking simulations of steroid substrates into human cytosolic hydroxysteroid dehydrogenases (AKR1C1 and AKR1C2): insights into positional and stereochemical preferences. Steroids. 2006;71:380–91.[PubMed][Google Scholar]
  • 20. Russel DW, Setchell KDRBile acid biosynthesis. Biochemistry. 1992;31:4737–4749.[PubMed][Google Scholar]
  • 21. Di Constanzo L, Drury JE, Penning TM, Christianson DWCrystal structure of human liver Δ3-ketosteroid 5β-reductase (AKR1D1) and implications for substate binding and catalysis. J Biol Chem. 2008;283:16830–16839.[Google Scholar]
  • 22. Jez JM, Bennett MJ, Schlegel BP, Lewis M, Penning TMComparative anatomy of the aldo-keto reductase superfamily. Biochem J. 1997;326:625–636.[Google Scholar]
  • 23. Chen M, Drury JE, Penning TMSubstrate specificity and inhibitor analyses of human steroid 5b-reductase (AKR1D1) Steroids. 2011;76:484–490.[Google Scholar]
  • 24. Jin YActivities of aldo-keto reductase 1 enzymes on two inhalded corticosteroids: implications for the pharmacological effects of inhaled corticosteroids. Chem Biol Interact. 2011;191:234–238.[Google Scholar]
  • 25. Iyer RB, Binstock JM, Schwartz IS, Gordon GG, Weinstein BI, Southern ALHuman hepatic cortisol reductase activities: enzymatic properties and substrate specificities of cytosolic cortisol Δ4–5-β reductase and dihydrocortisol-3-α-oxidoreductase(s) Steroids. 1990;55:495–500.[PubMed][Google Scholar]
  • 26. Kondo KH, Kai MH, Setoguchi Y, Eggertsen G, Sjoblom P, Setoguchi T, et al Cloning and expression of cDNA of human Δ3-oxosteroid 5β-reductase and substrate specificity of the expressed enzyme. Eur J Biochem. 1994;219:375–363.[PubMed][Google Scholar]
  • 27. Ratnam K, Ma H, Penning TMThe arginine 276 anchor for NADP(H) dictates fluorescence kinetic transients in 3α-hydroxysteroid dehydrogenase, a representative aldo-keto reductase. Biochemistry. 1999;38:7856–64.[PubMed][Google Scholar]
  • 28. Jin Y, Penning TMMultiple steps determine the overall rate of the reduction of 5α-dihydrotestosterone catalyzed by human type 3 3α-hydroxysteroid dehydrogenase: implications for the elimination of androgens. Biochemistry. 2006;45:13054–63.[Google Scholar]
  • 29. Cooper WC, Jin Y, Penning TMElucidation of a complete kinetic mechanism for a mammalian hydroxysteroid dehydrogenase (HSD) and identification of all enzyme forms on the reaction coordinate: the example of rat liver 3α-HSD (AKR1C9) J Biol Chem. 2007;282:33484–93.[PubMed][Google Scholar]
  • 30. Sharma KK, Lindqvist A, Zhou XJ, Auchus RJ, Penning TM, Andersson SDeoxycorticosterone inactivation by AKR1C3 in human mineralocorticoid target tissues. Mol Cell Endocrinol. 2006;248:79–86.[PubMed][Google Scholar]
  • 31. Byrns MC, Steckelbroeck S, Penning TMAn indomethacin analogue, N-(4-chlorobenzoyl) melatonin, is a selective inhibitor of aldo-keto reductase 1C3 (type 2 3α-HSD, type 5 17β-HSD, and prostaglandin F synthase) a potential target for the treatment of hormone dependent and independent malignancies. Biochem Pharmacol. 2008;75:484–493.[Google Scholar]
  • 32. Byrns MC, Mindnich R, Duan L, Penning TMOverexpression of aldo-keto reductase 1C3 (AKR1C3) in LNCaP cells diverts androgen metabolism towards testosterone resulting in resistance to the 5α-reductase inhibitor finasteride. J Steroid Biochem Mol Biol. 2012;130:7–15.[Google Scholar]
  • 33. Auchus RJThe backdoor pathway to dihydrotestosterone. Trends Endocrinol Metab. 2004;15:432–438.[PubMed][Google Scholar]
  • 34. Higaki Y, Usami N, Shintani S, Ishikura S, El-Kabbani O, Hara ASelective and potent inhibitors of human 20α-hydroxysteroid dehydrogenase (AKR1C1) that metabolizes neurosteroids derived from progesterone. Chem Biol Interact. 2003;143–144:503–513.[PubMed][Google Scholar]
  • 35. Sharifi N, Auchus RJSteroid biosynthesis and prostate cancer. Steroids. 2012;77:719–726.[PubMed][Google Scholar]
  • 36. Jin Y, Mesaros C, Blair IA, Penning TMStereospecific reduction of 5β-reduced steroids by human ketosteroid reductases of the AKR (aldo-keto reductase) superfamily: role of AKR1C1-AKR1C4 in the metabolism of testosterone and progesterone via the 5β-reductase pathway. Biochem J. 2011;437:53–61.[Google Scholar]
  • 37. Byrns MC, Duan L, Lee SH, Blair IA, Penning TMAldo-keto reductase 1C3 expression in MCF-7 cells reveals roles in steroid hormone and prostaglandin metabolism that may explain its over-expression in breast cancer. J Steroid Biochem Mol Biol. 2010;118:177–187.[Google Scholar]
  • 38. Traish AM5α-Reductases in human physiology: an unfolding story. Endocr Pract. 2012;18:965–75.[PubMed][Google Scholar]
  • 39. Mitchell BF, Mitchell JM, Chowdhury J, Tougas M, Engelen SME, Senff N, et al Metabolites of progesterone and the pregnane X receptor: A novel pathway regulating uterine contractility in pregnancy? Am J Obstet Gynecol. 2005;192:1304–1315.[PubMed][Google Scholar]
  • 40. Sheehan PM, Rice GE, Moses EK, Brennecke SP5β-Dihydroprogesterone and steroid 5β-reductase decrease in association with human parturition at term. Mol Hum Reprod. 2005;11:495–501.[PubMed][Google Scholar]
  • 41. Wiebe JPProgesterone metabolites in breast cancer. Endocr Relat Cancer. 2006;3:717–738.[PubMed][Google Scholar]
  • 42. Belelleli D, Lambert JJNeurosteroids: endogenous regulators of the GABAA receptor. Nature Rev Neuroscience. 2005;6:565–575.[PubMed][Google Scholar]
  • 43. Usami N, Yamamoto T, Shintani S, Higaki Y, Ishikura S, Katagiri Y, Hara ASubstrate specificity of human 3(20)α-hydroxysteroid dehydrogenase for neurosteroids and its inhibition by benzodiazepines. Biol Pharm Bull. 2002;25:441–445.[PubMed][Google Scholar]
  • 44. González-Usano A, Cauli O, Agustí A, Felipo VHyperammonemia alters the modulation by different neurosteroids of the glutamate-nitric oxide-cyclic GMP pathway through NMDA- GABAA - or sigma receptors in cerebellum in vivo. J Neurochem. 2013;125:133–43.[PubMed][Google Scholar]
  • 45. Giuliani FA, Yunes R, Mohn CE, Laconi M, Rettori V, Cabrera RAllopregnanollone induces LHRH and glutamate release through NMDA receptor modulation. Endocrine. 2011;40:21–6.[PubMed][Google Scholar]
  • 46. Lanišnik Rižner T, Brožič P, Doucette C, Tirek-Etienne T, Mueller-Vieira U, Sonneveld E, et al Selectivity and potency of the retroprogesterone dydrogesterone in vitro. Steroids. 2011;76:607–615.[PubMed][Google Scholar]
  • 47. Kakizaki S, Takizawa D, Tojima H, Horiguchi N, Yamazaki Y, Mori MNuclear receptors CAR and PXR: therapeutic targets for cholestasis liver disease. Front Bioscience. 2011;16:2988–3003.[PubMed][Google Scholar]
  • 48. Bauman DR, Steckelbroeck S, Williams MV, Peehl DM, Penning TMIdentification of the major oxidative 3α-hydroxysteroid dehydrogenase in human prostate that converts 5α-androstane-3α,17β-diol to 5α-dihydrotestosterone: a potential therapeutic target for androgen-dependent disease. Mol Endocrinol. 2006;20:444–58.[PubMed][Google Scholar]
  • 49. Gangloff A, Garneau A, Huang Y-W, Yang F, Lin S-XHuman oestrogenic 17β-hydroxysteroid dehydrogenase specificity: enzyme regulation through an NADPH-dependent substrate inhibition towards the highly specific oestrone reduction. Biochem J. 2001;356:269–276.[Google Scholar]
  • 50. Lu M-L, Huang Y-W, Lin S-XPurification, reconstitution, and steady-state kinetics of the trans-membrane 17β-hydroxysteroid dehydrogenase 2. J Biol Chem. 2002;277:22123–22130.[PubMed][Google Scholar]
  • 51. Nilsson S, Koehler KF, Gustafsson J-ADevelopment of subtype-selective oestrogen receptor-based therapeutics. Nat Rev Drug Discov. 2011;10:778–92.[PubMed][Google Scholar]
  • 52. Wu L, Einstein M, Geissler WM, Chan HK, Elliston KO, Andersson SExpression cloning and characterization of human 17β-hydroxysteroid dehydrogenase type 2, a microsomal enzyme possessing 20 α-hydroxysteroid dehydrogenase activity. J Biol Chem. 1993;17:12964–9.[PubMed][Google Scholar]
  • 53. Lee SA, Belyaeva OV, Wu L, Kedishvili NYRetinol dehydrogenase 10 but not retinol/sterol dehydrogenase(s) regulates the expression of retinoic acid-responsive genes in human transgenic skin raft culture. J Biol Chem. 2011;286:13550–60.[Google Scholar]
  • 54. Moore LB, Parks DJ, Jones SA, Bledsoe RK, Consler TG, Stimmel JB, et al Orphan nuclear receptors constitutive androstane receptor and pregnane X receptor share xenobiotic and steroid ligands. J Biol Chem. 2000;275:15122–15127.[PubMed][Google Scholar]
  • 55. Barski OA, Mindnich R, Penning TMAlternative splicing in the aldo-keto reductase superfamily: Implications for protein nomenclature. Chem-Biol Interact. 2013;202:153–158.[Google Scholar]
  • 56. Ji Q, Chang L, VanDenBerg D, Stanczyk FZ, Stolz ASelective reduction of AKR1C2 in prostate cancer and its role in DHT metabolism. The Prostate. 2003;54:275–289.[PubMed][Google Scholar]
  • 57. Bauman DR, Steckelbroeck S, Peehl DM, Penning TMTranscript profiling of the androgen signal in normal prostate, benign prostatic hyperplasia, and prostate cancer. Endocrinology. 2006;147:5806–5816.[PubMed][Google Scholar]
  • 58. Stanbrough M, Bubley GJ, Ross K, Golub TR, Rubin MA, Penning TM, et al Increased expression of genes converting adrenal androgens to testosterone in androgen-independent prostate cancer. Cancer Res. 2006;66:2815–2825.[PubMed][Google Scholar]
  • 59. Montgomery RB, Mostaghel EA, Vessella R, Hess DL, Kalhorn TF, Higano CS, et al Maintenance of intratumoral androgens in metastatic prostate cancer: a mechanism for castration-resistant tumor growth. Cancer Res. 2008;68:4447–4454.[Google Scholar]
  • 60. Hofland J, van Weerden WM, Dits NFJ, Steenbergen J, van Leenders GJLH, Jenster G, et al Evidence of limited contributions for intratumoral steroidogenesis in prostate cancer. Cancer Res. 2010;70:1256–1264.[PubMed][Google Scholar]
  • 61. Locke JA, Guns ES, Lubik AA, Adomat HH, Hendy SC, Wood CA, et al Androgen levels increased by intratumoral de novo steroidogenesis during progression of castration-resistan prostate cancer. Cancer Res. 2008;68:6407–6415.[PubMed][Google Scholar]
  • 62. Taylor BS, Schultz N, Hieronymus H, Gopalan A, Xiao Y, Carver BS, et al Integrative genomic profiling of human prostate cancer. Cancer Cell. 2010;18:11–22.[Google Scholar]
  • 63. [PubMed]
  • 64. Mitsiades N, Sung CC, Schultz N, Danila DC, He B, Eedunuri VK, et al Distinct patterns of dysregulated expression of enzymes involved in androgen synthesis and metabolism in metastatic prostate cancer tumors. Cancer Res. 2012;72:6142–52.[Google Scholar]
  • 65. Pfeiffer MJ, Smit FP, Sedelaar JP, Schalken JASteroidogenic enzymes and stem cell markers are upregulated during androgen deprivation in prostate cancer. Mol Med. 2011;17:657–64.[Google Scholar]
  • 66. Nakamura Y, Suzuki T, Nakabayashi M, Endoh M, Sakamoto K, Mikami Y, et al In situ androgen producing enzymes in human prostate cancer. Endocrine-Related Cancer. 2005;12:101–107.[PubMed][Google Scholar]
  • 67. Fung KM, Samara ENS, Wong C, Metwalli A, Krlin R, Bane B, et al Increased expression of type 2 3α-hydroxysteroid dehydrogenase/type 5 17β-hydroxysteroid dehydrogenase (AKR1C3) and its relationship with androgen receptor in prostate carcinoma. Endocrine-Related Cancer. 2006;13:169–180.[PubMed][Google Scholar]
  • 68. Wako K, Kawasaki T, Yamana K, Suzuki K, Jiang S, Umezu H, et al Expression of androgen receptor through androgen-converting enzymes is associated with biological aggressiveness in prostate cancer. J Clin Pathol. 2008;61:448–454.[PubMed][Google Scholar]
  • 69. Huang KH, Chiou SH, Chow KC, Lin TY, Chang HW, Chiang IP, et al Overexpression of aldo-keto reductase 1C2 is associated with disease progression in patients with prostate cancer. Histopathology. 2010;57:384–394.[PubMed][Google Scholar]
  • 70. Hofland J, van Weerden WM, Steenbergen J, Dits NFJ, Jenster G, de Jong FHActivin A stimulates AKR1C3 expression and growth in human prostate cancer. Endocrinology. 2012;153:5726–34.[PubMed][Google Scholar]
  • 71. Downs TM, Burton DW, Araiza FL, Hastings RH, Deftos LJPTHrP stimulates prostate cancer cell growth and upregulates aldo-keto reductase 1C3. Cancer Lett. 2011;306:52–59.[PubMed][Google Scholar]
  • 72. Mohler JL, Titus MA, Bai S, Kennerley BJ, Lih FB, Tomer KB, et al Activation of the androgen receptor by intratumoral bioconversion of androstanediol to dihydrotestosterone in prostate cancer. Cancer Res. 2011;71:1486–1496.[Google Scholar]
  • 73. Lewis MJ, Wiebe JP, Heathcote JGExpression of progesterone metabolizing enzyme genes (AKR1C1, AKR1C2, AKR1C3, SRD5A1, SRD5A2) is altered in human breast carcinoma. BMC Cancer. 2004;4:27.[Google Scholar]
  • 74. Ni IBP, Zakaria Z, Muhammad R, Abdullah N, Ibrahim N, Emran NA, et al Gene expression patterns distinguish breast carcinoma from normal breast tissue: The Malaysian context Pathology. Research and Practice. 2010;206:223–228.[PubMed][Google Scholar]
  • 75. Ji Q, Aoyama C, Nien YD, Liu PI, Chen PK, Chang L, et al Selective loss of ALR1C1 and AKR1C2 in breast cancer and their potential effect on progesterone signaling. Cancer Res. 2004;64:7610–7617.[PubMed][Google Scholar]
  • 76. Cancer Genome Atlas NetworkComprehensive molecular portraits of human breast tumours. Nature. 2012;490:61–70.[Google Scholar]
  • 77. Parker JS, Mullins M, Cheang MC, Leung S, Voduc D, Vickery T, et al Supervised risk predictor of breast cancer based on intrinsic subtypes. J Clin Oncol. 2009;27:1160–1167.[Google Scholar]
  • 78. Han B, Li S, Song D, Poisson-Pare D, Liu G, Luu-The V, et al Expression of 17β-hydroxysteroid dehydrogenase type 2 and type 5 in breast cancer and adjacent non-malignant tissue: A correlation to clinicopathological parameters. J Steroid Biochem Mol Biol. 2008;112:194–200.[PubMed][Google Scholar]
  • 79. Oduwole OO, Li Y, Isomaa VV, Mantyniemi A, Pulkka AE, Soini Y, et al 17β-hydroxysteroid dehydrogenase type 1 is and idependent prognostic marker in breast cancer. Cancer Res. 2004;64:7604–7609.[PubMed][Google Scholar]
  • 80. Jansson AK, Gunnarsson C, Cohen M, Sivik T, Stal O17β-hydroxysteroid dehydrogenase type 14 affects estradiol levels in breast cancer cells and is a prognostic marker in estrogen-receptor positive breast cancer. Cancer Res. 2006;66:11471–11477.[PubMed][Google Scholar]
  • 81. Dressing GE, Alyea R, Pang Y, Thomas PMembrane progesterone receptors (mPR) mediate progestin induced antimorbidity in breast cancer cells and are expressed in human breast tumors. Horm Canc. 2012;3:101, 112.[PubMed][Google Scholar]
  • 82. Prossnitz ER, Barton MThe G protein coupled estrogen receptor GPER in health and disease. Nat Rev Endocrinol. 2011;7:715–726.[Google Scholar]
  • 83. Wang C, Lv X, Jiang C, Davis JSThe putative G-protein coupled estrogen receptor agonist G-1 suppresses proliferation of ovarian and breast cancer cells in a GPER-independent manner. Am J Transl Res. 2012;4:390–402.[Google Scholar]
  • 84. Sasano H, Suzuki T, Miki Y, Moriya TIntracrinology of estrogens and androgens in breast carcinoma. J Steroid Biochem Mol Biol. 2008;108:181–5.[PubMed][Google Scholar]
  • 85. Secreto G, Venturelli E, Meneghini E, Carcangiu ML, Paolini B, Agresti R, et al Androgen receptors and serum testosterone levels identify different subsets of postmenopausal breast cancers. BMC Cancer. 2012;12:599.[Google Scholar]
  • 86. Lanišnik Rižner T, Šmuc T, Rupreht R, Šinkovec J, Penning TMAKR1C1 and AKR1C3 may determine progesterone and estrogen ratios in endometrial cancer. Mol Cell Endocrinol. 2006;248:126–135.[PubMed][Google Scholar]
  • 87. Šmuc T, Lanišnik Rižner TAberrant pre-receptor regulation of estrogen and progesterone action in endometrial cancer. Mol Cell Endocrinol. 2009;301:74–82.[PubMed][Google Scholar]
  • 88. Cornel KMC, Kritwagen RFPM, Delvoux B, Visconti L, Van de Vijver KK, Day JM, et al Overexpression of 17β-hydroxysteroid dehydrogenase type 1 increases the exposure of endometrial cancer to 17β-estradiol. J Clin Endocrinol Metab. 2012;97:E591–E601.[PubMed][Google Scholar]
  • 89. Sinreih M, Hevir N, Rižner TLAltered expression of genes involved in progesterone biosynthesis, metabolism and action in endometrial cancer. Chem Biol Interact. 2013;202:210–7.[PubMed][Google Scholar]
  • 90. Cancer Genome Atlas Research NetworkIntegrated genomic characterization of endometrial carcinoma. Nature. 2013;497:67–73.[Google Scholar]
  • 91. Marshall E, Lowrey J, MacPherson S, Maybin JA, Collins F, Critchley HO, et al In silico analysis identifies a novel role for androgens in the regulation of human endometrial apoptosis. J Clin Endocrinol Metab. 2011;96:E1746–55.[Google Scholar]
  • 92. Ito K, Utsunomiya H, Suzuki T, Saitou S, Akahira J, Okamura K, et al 17β-hydroxysteroid dehydrogenases in human endometrium and its disorders. Mol Cell Endocrinol. 2006;248:136–40.[PubMed][Google Scholar]
  • 93. Zakharov V, Lin HK, Azzarello J, McMeekin S, Moore KN, Penning TM, et al Suppressed expression of type 2 3α/type 5 17β-hydroxysteroid dehydrogenase (AKR1C3) in endometrial hyperplasia and carcinoma. Int J Clin Exp Pathol. 2010;3:608–17.[Google Scholar]
  • 94. Hevir N, Vouk K, Sinkovec J, Ribič-Pucelj M, Rižner TLAldo-keto reductases AKR1C1, AKR1C2 and AKR1C3 may enhance progesterone metabolism in ovarian endometriosis. Chem Biol Interact. 2011;191:217–26.[PubMed][Google Scholar]
  • 95. Lemonde HA, Custard EJ, Bouquet J, Duran M, Overmars H, Scambler PJ, et al Mutations in SRD5B1 (AKR1D1, the gene encoding Δ3-oxosteroid 5β-reductase, in hepatitis and liver failure in infancy. Gut. 2003;52:1494–1499.[Google Scholar]
  • 96. Gonzales E, Cresteil D, Baussan C, Dabadie A, Gerhardt MF, Jacquemin ESRBD5B1 (AKR1D1) gene analysis in Δ(4)-3-oxosteroid 5β-reductase deficiency: evidence for primary genetic defect. J Hepatol. 2004;40:716–718.[PubMed][Google Scholar]
  • 97. Ueki I, Kimura A, Chen HL, Yorifuji T, Mori J, Itoh S, et al SRD5B1 gene analysis needed for the accurate diagnosis of primary 3-oxo-Δ-steroid 5β-reductase deficiency. J Gastroenterology and Hepatology. 2009;24:776–785.[PubMed][Google Scholar]
  • 98. Mindnich R, Drury JE, Penning TMThe effects of disease associated point mutation on 5b-reductase (AKR1D1) enzyme function. Chem Biol Interact. 2011;191:250–254.[Google Scholar]
  • 99. Seki Y, Mizuochi T, Kimura A, Takahashi T, Ohtake A, Hayashi SI, et al Two neonatal cholestasis patients with mutations in the SRD5B1 (AKR1D1) gene: diagnosis and bile acid profiles during chenodeoxycholic acid treatment. J Inherit Metab Dis. 2013;36:565–73.[PubMed][Google Scholar]
  • 100. Zhao J, Fang LJ, Setchell KD, Chen R, Li LT, Wang JSPrimary Δ3-oxosteroid 5β-reductase deficiency: two cases in China. World J Gastroenterol. 2012;18:7113–7.[Google Scholar]
  • 101. Drury JE, Mindnich R, Penning TMCharacterization of disease-related 5β-reductase (AKR1D1) mutations reveals their potential to cause bile acid deficiency. J Biol Chem. 2010;32:24529–24537.[Google Scholar]
  • 102. Auchus RJNon-traditional metabolic pathways of adrenal steroids. Rev Endocr Metab Disord. 2009;10:27–32.[PubMed][Google Scholar]
  • 103. Flueck C, Meyer Boeni M, Pandey AV, Kempna P, Miller WL, Schoenle EJ, et al Why boys will be boys: Two pathways of fetal testicular androgen biosynthesis are needed for male sexual differentiation. Am J Hum Genetics. 2011;89:201–218.[Google Scholar]
  • 104. Musumeci L, Arthur JW, Cheung FSG, Hoque A, Lippman S, Reichardt JKVSingle nucleotide differences (SNDs) in the dbSNP database may lead to errors in genotyping and haplotyping studies. Human Mutation. 2009;31:67–73.[Google Scholar]
  • 105. Takahashi RH, Grigliatti TA, Reid RE, Riggs KWThe effects of allelic variation in aldo-keto reductase 1C2 on the in vitro metabolism of dihydrotestosterone. J Pharmacol Experimental Therapeutics. 2009;329:1032–1039.[PubMed][Google Scholar]
  • 106. Arthur JW, Reichardt JKVModeling single nucleotide polymorphisms in the human AKR1C1 and AKR1C2 genes: implications for functional and genotyping analyses. PLOS One. 2010;5:e15604.[Google Scholar]
  • 107. Nishizawa M, Nakajima T, Yasuda K, Kanzaki H, Sasaguri Y, Watanabe K, et al Close kinship of human 20α-hydroxysteroid dehydrogenase gene with three aldo-keto reductase genes. Genes Cells. 2000;5:111–25.[PubMed][Google Scholar]
  • 108. Trauger JW, Jiang A, Stearns BA, LoGrasso PVKinetics of allopregnanolone formation catalyzed by human 3α-hydroxysteroid dehydrogenase type III (AKR1C2) Biochemistry. 2002;41:13451–9.[PubMed][Google Scholar]
Collaboration tool especially designed for Life Science professionals.Drag-and-drop any entity to your messages.