Type 5 17beta-hydroxysteroid dehydrogenase/prostaglandin F synthase (AKR1C3): role in breast cancer and inhibition by non-steroidal anti-inflammatory drug analogs.
Journal: 2009/February - Chemico-Biological Interactions
ISSN: 1872-7786
Abstract:
Aldo-keto reductase (AKR) 1C3 catalyzes the NADPH-dependent reduction of Delta(4)-androstene-3,17-dione to yield testosterone, reduction of estrone to yield 17beta-estradiol and reduction of progesterone to yield 20alpha-hydroxyprogesterone. In addition, it functions as a prostaglandin (PG) F synthase and reduces PGH(2) to PGF(2)alpha and PGD(2) to 11beta-PGF(2). Immunohistochemistry showed that AKR1C3 is over-expressed in invasive ductal carcinoma of the breast. Retroviral expression of AKR1C3 in MCF-7 breast carcinoma cells shows that each of the assigned reactions occur in a breast cell microenvironment. Steroid and prostaglandin conversions were monitored by radiochromatography. Prostaglandin conversion was validated by a second method using HPLC coupled to APCI-MRM/MS. The combined effect of the AKR1C3 catalyzed 17- and 20-ketosteroid reductions will be to increase the 17beta-estradiol:progesterone ratio in the breast. In addition, formation of PGF(2) epimers would activate F prostanoid receptors and deprive PPARgamma of its putative anti-proliferative PGJ(2) ligands. Thus, AKR1C3 is a source of proliferative signals and a potential therapeutic target for hormone-dependent and -independent breast cancer. Two strategies for AKR1C3 inhibition based on non-steroidal anti-inflammatory drugs were developed. The first strategy uses the Ullmann coupling reaction to generate N-phenylanthranilate derivatives that inhibit AKR1C enzymes without affecting PGH(2) synthase (PGHS) 1 or PGHS-2. The second strategy exploits the selective inhibition of AKR1C3 by indomethacin, which did not inhibit highly related AKR1C1 or AKR1C2. Using known structure-activity relationships for the inhibition of PGHS-1 and PGHS-2 by indole acetic acids we obtained N-(4-chlorobenzoyl)-melatonin as a specific AKR1C3 inhibitor (K(I)=6.0muM) that does not inhibit PGHS-1, PGHS-2, AKR1C1, or AKR1C2. Both strategies are informed by crystal structures of ternary AKR1C3.NADP(+).NSAID complexes. The identification of NSAID analogs as specific inhibitors of AKR1C3 will help validate its role in the proliferation of breast cancer cells.
Relations:
Content
Citations
(17)
References
(31)
Diseases
(1)
Drugs
(2)
Chemicals
(5)
Organisms
(1)
Processes
(1)
Anatomy
(1)
Affiliates
(1)
Similar articles
Articles by the same authors
Discussion board
Chem Biol Interact 178(1-3): 221-227

Type 5 17β-Hydroxysteroid Dehydrogenase/Prostaglandin F Synthase (AKR1C3): Role In Breast Cancer and Inhibition by Nonsteroidal Antiinflammatory Drug Analogs

1. Introduction

In post-menopausal women, the proliferation of hormone dependent breast cancer cells is driven by the local production of estrogen [1,2]. Aromatase inhibitors have shown that inhibition of estrogen production is an effective strategy for the treatment and prevention of hormone dependent breast cancer. In addition to aromatase, the reductive 17β-hydroxysteroid dehydrogenases (17β-HSDs) are essential for the production of 17β-estradiol from Δ-androstene-3,17-dione (androstenedione, Scheme 1A)[1,2]. For instance, reduction of androstenedione by a 17β-HSD will yield testosterone, which can then be aromatized to 17β-estradiol (the most potent endogenous estrogen). The primary 17β-HSD involved in this reaction in peripheral tissues is believed to be aldo-keto reductase (AKR) 1C3 (type 5 17β-HSD)[3,4]. If aromatase converts androstenedione to estrone, this is then reduced to 17β-estradiol. The reductive 17β-HSDs generally associated with this reaction are types 1, 7, and possibly 12, but there is evidence that AKR1C3 can do this reaction as well [4]. Given the success of aromatase inhibitors against breast cancer and the contribution of the reductive 17β-HSDs to 17β-estradiol formation, there is considerable interest in developing inhibitors of these enzymes.

An external file that holds a picture, illustration, etc.
Object name is nihms259789f2.jpg

(A) Routes to 17β-estradiol in the breast. The other reductive 17β-HSDs include types 1, 7, & possibly 12; the oxidative isoforms include types 2 and 4. (B) Proposed role for AKR1C3 in prostaglandin signaling.

AKR1C3 performs several reactions that may lead to the development of breast cancer. Homogenous recombinant AKR1C3 catalyzes the reduction of androstenedione to testosterone, progesterone to 20α-hydroxyprogesterone and, to a lesser extent, estrone to 17β-estradiol [3]. These reactions have also been observed when AKR1C3 was transfected into human embryonic kidney (HEK-293) cells [4,5]. The combined effect of these activities in breast cancer cells would be to increase the 17β-estradiol to progesterone ratio, and hence increase estrogen receptor (ER) α and decrease progesterone receptor (PR) signaling, but this remains to be tested. While the increase in estrogenic signaling will clearly contribute to the development of breast cancer [1,2], the relationship between progesterone signaling and breast cancer is not as well understood. Genes encoding aromatase, her-2/neu, prostaglandin H2 synthase (PGHS) 2 and matrix metalloproteinases are suppressed by PR activation, so progesterone inactivation by AKR1C3 could further contribute to the development of breast cancer in vivo [6,7].

AKR1C3 is also involved in the reduction of prostaglandins, which could generate hormone-independent proliferative signals (Scheme 1B). Purified recombinant AKR1C3 stereospecifically and efficiently converts prostaglandin (PG) H2 to PGF and PGD2 to 9α,11β-PGF2 [8,9]. Of the known endogenous substrates, AKR1C3 exhibits the highest catalytic efficiency towards the prostaglandins, particularly PGD2. The PGF2 isomers bind to the F prostanoid receptor and induce MAPK signaling cascades that lead to cell proliferation [10]. In addition, by removing PGD2, AKR1C3 prevents its spontaneous dehydration and rearrangement to form the anti-proliferative and anti-inflammatory J2 series prostaglandins, including 15-deoxy-Δ-PGJ2 (15dPGJ2). 15dPGJ2 covalently reacts with a cysteine residue in the ligand-binding domain of PPARγ, resulting in its activation [11]. It also reacts with residues in the DNA-binding domains of NFκB and ERα, preventing them from binding to DNA [12,13]. The resulting increase in PPARγ-dependent and decrease in NFκB-dependent and ERα-dependent gene transcription is predicted to inhibit the proliferation of breast cancer cells.

We have been exploring the role of AKR1C3 in breast cancer and developing non-steroidal anti-inflammatory drug (NSAID) analogues as selective inhibitors of AKR1C3. We will describe work from our lab and others showing that AKR1C3 is expressed in breast cancer. It will also describe our recent work using an AKR1C3 over-expressing MCF-7 hormone dependent breast cancer cell line to examine the roles of AKR1C3 in steroid hormone and prostaglandin signaling [14]. Finally, we will discuss our work developing NSAID analogues as isoform specific inhibitors of AKR1C3 [15,16].

2. AKR1C3 is expressed in human breast cancer

Using semi-quantitative RT-PCR, we have shown that the normal breast expresses very high levels of AKR1C3 relative to other tissues [3]. Using immunohistochemistry with an isoform specific antibody, we observed that AKR1C3 expression is even higher in the tumor of a patient with ER and PR positive invasive ductal carcinoma, as compared to surrounding normal tissue [17]. Work from Sasano's group has shown an 18-fold increase in the median AKR1C3 mRNA levels in breast cancer patients as compared to those without [18]. They have also detected AKR1C3 with immunohistochemistry in 53% of breast carcinomas [2]. Another RT-PCR analysis of 669 breast cancer cases by Oduwole et al found significantly higher expression of AKR1C3 in breast tumors than in normal tissue [19]. They also found that patients with the highest levels of AKR1C3 expression had a worse overall prognosis. Jansson et al observed that patients with ER tumors that overexpress AKR1C3 had a higher rate of late recurrence [20]. These results provide evidence for an association between AKR1C3 expression and breast cancer, which raises the question of how AKR1C3 might alter breast cancer cell signaling and proliferation.

3. AKR1C3 catalyzes steroid hormone reduction reactions in MCF-7 cells

In order to explore the potential of AKR1C3 to contribute to proliferative signaling in breast cancer, we developed an MCF-7 cell line that over-expresses AKR1C3 using a pLNCX retroviral vector (MCF-7-AKR1C3 cells). Using this cell line, we explored the effects of AKR1C3 expression on the metabolism of radiolabeled steroid hormones (Table 1)[14]. Parental cells primarily metabolized [C]-androstenedione to 5α-androstanedione and almost no 17β-HSD activity was observed. The MCF-7-AKR1C3 cells exhibited much higher 17β-HSD activity and converted over 20% and 10% of 0.1 and 5 μM [C]-androstenedione into testosterone, respectively, after 24 h. 5α-DHT was also formed as a minor metabolite.

Table 1

Effect of AKR1C3 expression on the metabolism of steroid hormones and prostaglandins by MCF-7 cells
% of Compound Metabolized by Cells at 6 h*
SubstrateConcentration (μM)Parental MCF-7MCF-7-AKR1C3
[C]-Androstenedione0.111 (1.9)21 (7.5)
54.6 (1.1)6.9 (2.8)
[C]-Progesterone0.119 (1.2)49 (27)
55.4 (1.0)19 (13)
[C]-Estrone0.117 (17)82 (82)
53.2 (3.2)5.2 (5.2)
[H]-PGD20.125 (0)37 (14)
521 (0)35 (7.5)
Values in parentheses represent the percentage of the radioactivity associated with the primary product formed by AKR1C3 (testosterone from androstenedione; 20α-OH-progesterone from progesterone; 17β-estradiol from estrone; 9α,11β-PGF2 from PGD2).

The parental MCF-7 cells metabolized [C]-progesterone primarily to non-polar products, possibly glucuronides, and low levels of 5α- and 20α-reduced products were also observed. The MCF-7-AKR1C3 cells exhibited significantly more 20α-HSD activity. 20α-Hydroxyprogesterone and 20α-hydroxy-5α-dihydroprogesterone accounted for 27% and 7% of the total radioactivity, respectively, when 0.1 μM [C]-progesterone was incubated for 6 h. The MCF-7-AKR1C3 cells also exhibited very high levels of 20α-HSD activity with 5 μM [C]-progesterone as substrate, where 20α-hydroxyprogesterone accounted for 50% of the radioactivity by 24 h.

Surprisingly, at 0.1 μM, [C]-estrone was the best of the substrates assayed. The parental MCF-7 cells exhibited high endogenous 17β-HSD activity towards estrone: 17β-estradiol accounted for 17% of the radioactivity by 6 h with no other products formed. When MCF-7-AKR1C3 cells were incubated with 0.1 μM [C]-estrone, 17β-estradiol was formed much faster, accounting for over 80% of the radioactivity by 6 h. However, AKR1C3 expression had almost no effect on the metabolism of 5 μM [C]-estrone. These results suggest that AKR1C3 acts as a high affinity low capacity enzyme towards estrone in vivo.

We examined whether the increased conversion of estrone to 17β-estradiol in MCF-7-AKR1C3 cells would confer a proliferative advantage (Table 2)[14]. Parental cells have a modest proliferative response to picomolar concentrations of both estrone and 17β-estradiol. Expression of AKR1C3 significantly increases the responsiveness of these cells to both estrogens. The similar response of parental cells to estrone and 17β-estradiol suggests that a rapid equilibrium is established, in which picomolar concentrations of estrogens are inter-converted by endogenous reductive and oxidative 17β-HSDs. The expression of AKR1C3 provides an additional reductive 17β-HSD that shifts the equilibrium in favor of 17β-estradiol, increasing the proliferative response of these cells.

Table 2

Proliferation of MCF-7 Cells in Response to Estrogens
Cell Number (% of DMSO Treated Control)
CompoundConcentrationParental MCF-7MCF-7-AKR1C3
Estrone1 pM102 ± 3112 ± 8
10 pM117 ± 12172 ± 27
33 pM131 ± 4204 ± 14
17β-Estradiol1 pM112 ± 4167 ± 14
10 pM134 ± 4213 ± 27
33 pM138 ± 5218 ± 12
Values expressed as means ± S.D. of three experiments performed in triplicate.

4. AKR1C3 catalyzes PGD2 reduction in MCF-7 cells

We used the MCF-7-AKR1C3 cells to examine the ability of AKR1C3 to convert [H]-PGD2 to 9α,11β-PGF2 [14]. In cell-free media, [H]-PGD2 disappeared, accompanied by the formation of PGJ2 and 15dPGJ2. In the presence of MCF-7 cells, [H]-PGD2 disappeared at a similar rate, but the levels of PGJ2 and 15dPGJ2 observed were much lower, with much of the radioactivity remaining in the aqueous fraction. This observation is consistent with Michael addition to the reactive cyclopentenone ring of the J prostaglandins by glutathione and other hydrophilic cellular nucleophiles. When MCF-7-AKR1C3 cells were used, the rate of [H]-PGD2 disappearance was slightly higher, accompanied by the formation of a PGF2 product. At 24 h, this PGF2 product accounted for 35% of the radioactivity in the media of the MCF-7-AKR1C3 cells treated with 0.1 μM [H]-PGD2. Because the TLC separation used for the radiometric assay did not conclusively distinguish 9α,11β-PGF2 from PGF, we also examined metabolism of unlabeled PGD2. In this approach, we used a method that employed pentafluorobenzyl derivatization of the prostaglandins, separation with chiral normal phase HPLC and analysis with electron capture APCI-MRM/MS [14,21]. Using this method, we confirmed that the reduction of PGD2 in MCF-7-AKR1C3 cells was stereospecific, forming only 9α,11β-PGF2. Prostaglandin levels were quantified by reference to deuterated prostaglandin internal standards and were consistent with the levels observed radiometrically.

5. Development of NSAID analogues as selective AKR1C3 inhibitors

5.1 NSAID screening

Given its potential role in breast cancer, AKR1C3 represents a promising therapeutic target. Specific inhibitors of AKR1C3 would also provide useful tools to explore its contribution to intercellular signaling pathways. We have focused our efforts on developing NSAID analogues as inhibitors of AKR1C3, based on our earlier observation that NSAIDs are potent inhibitors of rat 3α-HSD (AKR1C9)[22]. The presence of crystal structures of AKR1C3 bound to flufenamic acid and indomethacin allows for rational design of inhibitors based on these two structures [23]. Inhibitors of AKR1C3 should not inhibit the related peripheral AKR1C1 and AKR1C2 isoforms or the PGHS enzymes if they are to dissect the roles of AKR1C3 in steroid hormone and prostaglandin signaling. AKR1C1 is primarily a 20-ketosteroid reductase and catalyzes the reduction of progesterone at the 20 position, while AKR1C2 is primarily a 3-ketosteroid reductase and reduces 5α-dihydrotestosterone at the 3 position [3].

We examined a series of NSAIDs for their inhibition of the reduction of 9,10-phenanthrenequinone by AKR1C1, 1C2, and 1C3 (Table 3)[16]. We observed inhibition of the AKR1C enzymes at pharmacologically relevant concentrations with all of the NSAIDs used. These results raise the possibility that AKR1C inhibition might play a role in the pharmacological properties of these drugs. It is interesting to observe that salicylic acid, which rapidly forms from the deacetylation of aspirin in vivo, inhibits AKR1C1 and AKR1C2 at concentrations that are two orders of magnitude lower than those required to inhibit PGHS [24]. The inhibition of AKR1C1 by salicylic acid and aspirin has been recently reported by Dhagat et al, but their inhibition of AKR1C2 and AKR1C3 had not been reported [25].

Table 3

Inhibition of the AKR1C catalyzed reduction of 9,10-phenanthrenequinone by NSAIDs [16]
IC50 values (μM)Preference for AKR1C3 vs. AKR1C2Preference for AKR1C3 vs. AKR1C1
CompoundAKR1C1AKR1C2AKR1C3
Indomethacin96502.32242
Sulindac106.63.41.92.9
Meclofenamic Acid2.91.70.72.44.1
Zomepirac>5023400.6>1.25
Ibuprofen291.99.90.192.9
Naproxen2801.21.40.86200
Flurbiprofen513.27.80.416.5
Aspirin7.11612000.0130.0059
Salicylic Acid2.77.47700.00960.0035

There was surprising variability between AKR1C isoforms in their inhibition by different NSAIDs. Meclofenamic acid was a potent inhibitor of all three AKR1C isoforms [16]. Salicylates were potent inhibitors of AKR1C1 and AKR1C2, but only weakly inhibited AKR1C3 (unpublished data). Arylpropionic acids exhibited a strong preference for inhibition of AKR1C2 over AKR1C1, with an intermediate level of AKR1C3 inhibition [16]. Indomethacin inhibited AKR1C3 much more potently than AKR1C1 and AKR1C2. These observations suggested routes to AKR1C3 selective inhibitors.

5.2 N-Phenylanthranilic acids

We initially focused our efforts on the N-phenylanthranilic acid family of NSAIDs [15]. These compounds are among the most potent inhibitors of AKR1Cs and can be simply synthesized using the Ullman coupling reaction, in which any halogenated benzoic acid can be coupled to any aniline in the presence of Cu. In addition, there are well-established structure-function relationships for PGHS inhibition, which indicate minor structural changes that would abolish their PGHS inhibition (Scheme 2A). These changes include A-ring substitution and movement of the carboxylic acid from the ortho position. We observed that mefenamic acid is a potent inhibitor of AKR1C3 and its two closely related family members, AKR1C1 and AKR1C2 (Table 2). The inhibition of the AKR1C enzymes occurred at much lower concentrations than those required to inhibit their putative PGHS targets. We then screened a series of N-phenylanthranilic acid derivatives with A-ring substitution and/or movement of the carboxylic acid to the para position for their inhibition of the AKR1C and PGHS enzymes. It was found that all of the compounds tested were potent competitive inhibitors of the three AKR1C isoforms (KI < 10 μM) with very little or no inhibition of PGHS-1 and PGHS-2 (IC50 > 200 μM, Table 4). Two of the compounds we tested, 4-carboxy-2′,4′-dinitrodiphenylamine and 4-benzoyl-benzoic acid exhibited a slight preference for inhibition of AKR1C3 over the other two isoforms, but none of the compounds provided enough selectivity between AKR1C isoforms to serve as an isoform-specific molecular probe.

An external file that holds a picture, illustration, etc.
Object name is nihms259789f3.jpg

Predicted structure activity relationships of (A) N-phenylanthranilic acid derivatives and (B) indomethacin derivatives.

Table 4

Inhibition of the AKR1C and PGHS enzymes by N-phenylanthranilic acid derivatives [32]
InhibitorIC50 values (μM)

AKR1C1AKR1C2AKR1C3COX-1COX-2
Mefenamic Acid4.0±0.50.96±0.10.39±0.0933225
KI = 0.81KI = 0.22KI = 0.30*KI = 3.90
5-Methyl-N-phenylanthranilic Acid3.2±0.11.2±0.111±0.4430≫ 1,000
KI = 0.88KI = 0.53KI = 7.3
4-Carboxy-2′,4′-dinitrodiphenylamine7.1±0.83.7±0.20.40±0.06≫ 500ND**
KI = 2.7KI = 1.3KI = 0.38
4-Chloro-N-phenylanthranilic Acid8.2±0.91.2±0.13.0±0.2> 1,000≫ 1,000
KI = 3.10KI = 0.41KI = 2.9
4-Nitro-N-phenylanthranilic Acid9.2±0.52.1±0.42.0±0.3≫ 250ND
KI = 3.5KI = 0.75KI = 1.9
4-Carboxy-2-aminodiphenylamine49±8.926±3.211±1.5∼ 1,000≫ 1,000
KI = 19KI = 9.4KI = 10
4-Chloro-N-(4-toly)-anthranilic Acid8.3±0.50.56±0.171.5±0.4240≫ 1,000
KI = 3.1KI = 0.20KI = 1.4
4-Benzoyl-benzoicAcid33±4.437±2.52.0±0.2≫ 1,000≫ 1,000
KI = 12KI = 13KI = 1.9
KI values for AKR1C3 were computed assuming a KM of 2 mM for 1-acenaphthenol.
N.D., not determined.

Clues to the development of an AKR1C3 selective inhibitor based on N-phenylanthranilic acids are revealed by the comparison of the crystal structures of AKR1C2·NADP·ursodeoxycholate (PDB 1IHI) and AKR1C3·NADP·flufenamic acid (PDB 1S2C, Scheme 2A)[23]. When the two active sites of the ternary complexes are overlaid, it is apparent that ursodeoxycholate and flufenamic acid bind differently. The A-ring of flufenamic acid can tolerate bulky substituents because these would project into the steroid binding cavity of both AKR1C enzymes. Thus substituents to this ring are unlikely to provide AKR1C isoform specific inhibitors. By comparison, the B-ring of flufenamic acid projects into a unique sub-pocket present in AKR1C isoforms that is smaller and more hydrophobic for AKR1C2 than for AKR1C3 due to substitution of Phe 319, Phe 118, and Leu 308 in AKR1C2 with Tyr 319, Ser 118, and Ser 308 in AKR1C3. Polar substituents to the B-ring of N-phenylanthranilic acids and similar structures would occupy this subpocket and likely lead to isoform specific inhibitors.

5.3 Indole acetic acids

We then focused our attention on indomethacin, as this compound provided the best selectivity for AKR1C3 inhibition of the NSAIDs examined [16]. We used structure-activity relationships to guide the development of an AKR1C specific inhibitor based on indomethacin (Scheme 2B). The inhibition of the PGHS enzymes by indomethacin is largely dependent on the methyl group at the 2 position, which fits tightly into a small sub-pocket found in both PGHS isoforms, resulting in functionally irreversible inhibition [26]. Without this interaction, 2-desmethyl-indomethacin can be easily competed out of the active site of the PGHS enzymes by arachidonic acid. In addition, substituents to the acetic acid group limit PGHS-1 inhibition and can reduce the inhibition of PGHS-2 [27-29]. The crystal structure of the AKR1C3·NADP·indomethacin complex (PDB 1S2A)[23], provides structure-activity relationships for the inhibition of the AKR1C enzymes. The acetic acid group of indomethacin extends into the same sub-pocket of AKR1C3 that is occupied by the B-ring of flufenamic acid, suggesting that the addition of substituents to this region would further increase the selectivity for AKR1C3 over AKR1C2. The 2-methyl group is in an open region of the active site and does not interact with any residues, suggesting that its removal or modification would not effect AKR1C inhibition.

One compound that was predicted to retain its inhibition of AKR1C3 but not to inhibit either the PGHS enzymes or the other AKR1C isoforms was N-(4-chlorobenzoyl)-melatonin (CBM). This compound lacks the 2-methyl group that is critical for PGHS inhibition, and also has a reverse methyl amide in place of the carboxylic acid of indomethacin (Figure 1A). The synthesis of this compound and inhibition of PGHS-1 and PGHS-2 had been described by Kalgutkar et al, who observed that it did not inhibit either PGHS isoform [29]. We screened CBM for inhibition of the AKR1C enzymes and found that it was a potent and selective inhibitor of AKR1C3 [16]. It inhibited the reduction of 9,10-phenanthrenequinone by AKR1C3 with an uncompetitive pattern of inhibition and a KI of 3.4 μM, while it inhibited AKR1C1 and AKR1C2 by less than 10% at the highest concentration tested (60 μM, Figure 1B). CBM inhibited the reduction of androstenedione by AKR1C3 with a competitive pattern of inhibition and a KI of 6.0 μM (Figure 1C). The differences in the pattern of inhibition of AKR1C3 by CBM with 9,10-phenanthrenequinone (turned over with high catalytic efficiency) and androstenedione (turned over with low catalytic efficiency) suggests that CBM binds differently during the inhibition of the reduction of these two substrates. The most likely explanation is that for androstenedione, CBM binds E·NADPH and blocks the binding of this substrate, but for 9,10-phenanthrenequinone it binds to E·NADP product complex and prevents the release of cofactor. It therefore appears that CBM is a selective inhibitor of AKR1C3 that could be useful to further investigate its roles in the development of breast cancer.

An external file that holds a picture, illustration, etc.
Object name is nihms259789f1.jpg

N-Chlorobenzoyl melatonin (CBM) exhibits uncompetitive inhibition patterns for the reduction of 9,10-phenanthrenequinone by AKR1C3 but competitive inhibition patterns for the reduction of androstenedione by AKR1C3 [16]. (A) Chemical structure of CBM. (B-C) Lineweaver–Burke plots demonstrating the pattern of inhibition by CBM for the reduction of (B) 9,10-phenanthrenequinone (PQ) or (C) androstenedione catalyzed by AKR1C3 [16].

5.1 NSAID screening

Given its potential role in breast cancer, AKR1C3 represents a promising therapeutic target. Specific inhibitors of AKR1C3 would also provide useful tools to explore its contribution to intercellular signaling pathways. We have focused our efforts on developing NSAID analogues as inhibitors of AKR1C3, based on our earlier observation that NSAIDs are potent inhibitors of rat 3α-HSD (AKR1C9)[22]. The presence of crystal structures of AKR1C3 bound to flufenamic acid and indomethacin allows for rational design of inhibitors based on these two structures [23]. Inhibitors of AKR1C3 should not inhibit the related peripheral AKR1C1 and AKR1C2 isoforms or the PGHS enzymes if they are to dissect the roles of AKR1C3 in steroid hormone and prostaglandin signaling. AKR1C1 is primarily a 20-ketosteroid reductase and catalyzes the reduction of progesterone at the 20 position, while AKR1C2 is primarily a 3-ketosteroid reductase and reduces 5α-dihydrotestosterone at the 3 position [3].

We examined a series of NSAIDs for their inhibition of the reduction of 9,10-phenanthrenequinone by AKR1C1, 1C2, and 1C3 (Table 3)[16]. We observed inhibition of the AKR1C enzymes at pharmacologically relevant concentrations with all of the NSAIDs used. These results raise the possibility that AKR1C inhibition might play a role in the pharmacological properties of these drugs. It is interesting to observe that salicylic acid, which rapidly forms from the deacetylation of aspirin in vivo, inhibits AKR1C1 and AKR1C2 at concentrations that are two orders of magnitude lower than those required to inhibit PGHS [24]. The inhibition of AKR1C1 by salicylic acid and aspirin has been recently reported by Dhagat et al, but their inhibition of AKR1C2 and AKR1C3 had not been reported [25].

Table 3

Inhibition of the AKR1C catalyzed reduction of 9,10-phenanthrenequinone by NSAIDs [16]
IC50 values (μM)Preference for AKR1C3 vs. AKR1C2Preference for AKR1C3 vs. AKR1C1
CompoundAKR1C1AKR1C2AKR1C3
Indomethacin96502.32242
Sulindac106.63.41.92.9
Meclofenamic Acid2.91.70.72.44.1
Zomepirac>5023400.6>1.25
Ibuprofen291.99.90.192.9
Naproxen2801.21.40.86200
Flurbiprofen513.27.80.416.5
Aspirin7.11612000.0130.0059
Salicylic Acid2.77.47700.00960.0035

There was surprising variability between AKR1C isoforms in their inhibition by different NSAIDs. Meclofenamic acid was a potent inhibitor of all three AKR1C isoforms [16]. Salicylates were potent inhibitors of AKR1C1 and AKR1C2, but only weakly inhibited AKR1C3 (unpublished data). Arylpropionic acids exhibited a strong preference for inhibition of AKR1C2 over AKR1C1, with an intermediate level of AKR1C3 inhibition [16]. Indomethacin inhibited AKR1C3 much more potently than AKR1C1 and AKR1C2. These observations suggested routes to AKR1C3 selective inhibitors.

5.2 N-Phenylanthranilic acids

We initially focused our efforts on the N-phenylanthranilic acid family of NSAIDs [15]. These compounds are among the most potent inhibitors of AKR1Cs and can be simply synthesized using the Ullman coupling reaction, in which any halogenated benzoic acid can be coupled to any aniline in the presence of Cu. In addition, there are well-established structure-function relationships for PGHS inhibition, which indicate minor structural changes that would abolish their PGHS inhibition (Scheme 2A). These changes include A-ring substitution and movement of the carboxylic acid from the ortho position. We observed that mefenamic acid is a potent inhibitor of AKR1C3 and its two closely related family members, AKR1C1 and AKR1C2 (Table 2). The inhibition of the AKR1C enzymes occurred at much lower concentrations than those required to inhibit their putative PGHS targets. We then screened a series of N-phenylanthranilic acid derivatives with A-ring substitution and/or movement of the carboxylic acid to the para position for their inhibition of the AKR1C and PGHS enzymes. It was found that all of the compounds tested were potent competitive inhibitors of the three AKR1C isoforms (KI < 10 μM) with very little or no inhibition of PGHS-1 and PGHS-2 (IC50 > 200 μM, Table 4). Two of the compounds we tested, 4-carboxy-2′,4′-dinitrodiphenylamine and 4-benzoyl-benzoic acid exhibited a slight preference for inhibition of AKR1C3 over the other two isoforms, but none of the compounds provided enough selectivity between AKR1C isoforms to serve as an isoform-specific molecular probe.

An external file that holds a picture, illustration, etc.
Object name is nihms259789f3.jpg

Predicted structure activity relationships of (A) N-phenylanthranilic acid derivatives and (B) indomethacin derivatives.

Table 4

Inhibition of the AKR1C and PGHS enzymes by N-phenylanthranilic acid derivatives [32]
InhibitorIC50 values (μM)

AKR1C1AKR1C2AKR1C3COX-1COX-2
Mefenamic Acid4.0±0.50.96±0.10.39±0.0933225
KI = 0.81KI = 0.22KI = 0.30*KI = 3.90
5-Methyl-N-phenylanthranilic Acid3.2±0.11.2±0.111±0.4430≫ 1,000
KI = 0.88KI = 0.53KI = 7.3
4-Carboxy-2′,4′-dinitrodiphenylamine7.1±0.83.7±0.20.40±0.06≫ 500ND**
KI = 2.7KI = 1.3KI = 0.38
4-Chloro-N-phenylanthranilic Acid8.2±0.91.2±0.13.0±0.2> 1,000≫ 1,000
KI = 3.10KI = 0.41KI = 2.9
4-Nitro-N-phenylanthranilic Acid9.2±0.52.1±0.42.0±0.3≫ 250ND
KI = 3.5KI = 0.75KI = 1.9
4-Carboxy-2-aminodiphenylamine49±8.926±3.211±1.5∼ 1,000≫ 1,000
KI = 19KI = 9.4KI = 10
4-Chloro-N-(4-toly)-anthranilic Acid8.3±0.50.56±0.171.5±0.4240≫ 1,000
KI = 3.1KI = 0.20KI = 1.4
4-Benzoyl-benzoicAcid33±4.437±2.52.0±0.2≫ 1,000≫ 1,000
KI = 12KI = 13KI = 1.9
KI values for AKR1C3 were computed assuming a KM of 2 mM for 1-acenaphthenol.
N.D., not determined.

Clues to the development of an AKR1C3 selective inhibitor based on N-phenylanthranilic acids are revealed by the comparison of the crystal structures of AKR1C2·NADP·ursodeoxycholate (PDB 1IHI) and AKR1C3·NADP·flufenamic acid (PDB 1S2C, Scheme 2A)[23]. When the two active sites of the ternary complexes are overlaid, it is apparent that ursodeoxycholate and flufenamic acid bind differently. The A-ring of flufenamic acid can tolerate bulky substituents because these would project into the steroid binding cavity of both AKR1C enzymes. Thus substituents to this ring are unlikely to provide AKR1C isoform specific inhibitors. By comparison, the B-ring of flufenamic acid projects into a unique sub-pocket present in AKR1C isoforms that is smaller and more hydrophobic for AKR1C2 than for AKR1C3 due to substitution of Phe 319, Phe 118, and Leu 308 in AKR1C2 with Tyr 319, Ser 118, and Ser 308 in AKR1C3. Polar substituents to the B-ring of N-phenylanthranilic acids and similar structures would occupy this subpocket and likely lead to isoform specific inhibitors.

5.3 Indole acetic acids

We then focused our attention on indomethacin, as this compound provided the best selectivity for AKR1C3 inhibition of the NSAIDs examined [16]. We used structure-activity relationships to guide the development of an AKR1C specific inhibitor based on indomethacin (Scheme 2B). The inhibition of the PGHS enzymes by indomethacin is largely dependent on the methyl group at the 2 position, which fits tightly into a small sub-pocket found in both PGHS isoforms, resulting in functionally irreversible inhibition [26]. Without this interaction, 2-desmethyl-indomethacin can be easily competed out of the active site of the PGHS enzymes by arachidonic acid. In addition, substituents to the acetic acid group limit PGHS-1 inhibition and can reduce the inhibition of PGHS-2 [27-29]. The crystal structure of the AKR1C3·NADP·indomethacin complex (PDB 1S2A)[23], provides structure-activity relationships for the inhibition of the AKR1C enzymes. The acetic acid group of indomethacin extends into the same sub-pocket of AKR1C3 that is occupied by the B-ring of flufenamic acid, suggesting that the addition of substituents to this region would further increase the selectivity for AKR1C3 over AKR1C2. The 2-methyl group is in an open region of the active site and does not interact with any residues, suggesting that its removal or modification would not effect AKR1C inhibition.

One compound that was predicted to retain its inhibition of AKR1C3 but not to inhibit either the PGHS enzymes or the other AKR1C isoforms was N-(4-chlorobenzoyl)-melatonin (CBM). This compound lacks the 2-methyl group that is critical for PGHS inhibition, and also has a reverse methyl amide in place of the carboxylic acid of indomethacin (Figure 1A). The synthesis of this compound and inhibition of PGHS-1 and PGHS-2 had been described by Kalgutkar et al, who observed that it did not inhibit either PGHS isoform [29]. We screened CBM for inhibition of the AKR1C enzymes and found that it was a potent and selective inhibitor of AKR1C3 [16]. It inhibited the reduction of 9,10-phenanthrenequinone by AKR1C3 with an uncompetitive pattern of inhibition and a KI of 3.4 μM, while it inhibited AKR1C1 and AKR1C2 by less than 10% at the highest concentration tested (60 μM, Figure 1B). CBM inhibited the reduction of androstenedione by AKR1C3 with a competitive pattern of inhibition and a KI of 6.0 μM (Figure 1C). The differences in the pattern of inhibition of AKR1C3 by CBM with 9,10-phenanthrenequinone (turned over with high catalytic efficiency) and androstenedione (turned over with low catalytic efficiency) suggests that CBM binds differently during the inhibition of the reduction of these two substrates. The most likely explanation is that for androstenedione, CBM binds E·NADPH and blocks the binding of this substrate, but for 9,10-phenanthrenequinone it binds to E·NADP product complex and prevents the release of cofactor. It therefore appears that CBM is a selective inhibitor of AKR1C3 that could be useful to further investigate its roles in the development of breast cancer.

An external file that holds a picture, illustration, etc.
Object name is nihms259789f1.jpg

N-Chlorobenzoyl melatonin (CBM) exhibits uncompetitive inhibition patterns for the reduction of 9,10-phenanthrenequinone by AKR1C3 but competitive inhibition patterns for the reduction of androstenedione by AKR1C3 [16]. (A) Chemical structure of CBM. (B-C) Lineweaver–Burke plots demonstrating the pattern of inhibition by CBM for the reduction of (B) 9,10-phenanthrenequinone (PQ) or (C) androstenedione catalyzed by AKR1C3 [16].

Summary

AKR1C3 likely plays important roles in the development of hormone-dependent, and possibly hormone-independent, breast cancer. It is highly expressed in normal breast and upregulated in breast cancer, where its expression is associated with a worse prognosis [3,17,19]. Its expression in MCF-7 cells leads to increased synthesis of 17β-estradiol and 9α,11β-PGF2 and elimination of progesterone [14]. Consistent with a role in the development of hormone-dependent breast cancer, the increased synthesis of 17β-estradiol due to AKR1C3 expression confers a proliferative advantage to these cells.

Given that AKR1C3 may contribute to breast cancer cell proliferation, it represents a potential target for the treatment of breast cancer. Interestingly, AKR1C3 is potently inhibited by NSAIDs, which are protective against breast cancer [16,30]. While these compounds have traditionally been thought to work through inhibition of PGHS-2 activity, they also exhibit anti-cancer effects independent of PGHS [31]. AKR1C3 represents a potential alternative target for the anti-cancer effects of these drugs. A specific inhibitor of AKR1C3 might retain the anti-cancer effects of NSAIDs, without the gastro-intestinal toxicity and cardiovascular side-effects that limit their use.

By examining the inhibition of AKR1Cs by different classes of NSAIDs and NSAID analogues, we have made considerable progress in understanding how these compounds might lead to isoform specific AKR1C inhibitors. N-Phenylanthranilic acids are potent, but non-selective, inhibitors of all of the AKR1Cs. We continue to develop derivatives of these NSAIDs that will be selective for AKR1C3. Salicylates are excellent inhibitors of AKR1C1 and AKR1C2, and may provide a starting point for the development of a selective inhibitor of AKR1C1. Indomethacin inhibits AKR1C3 at much lower concentrations than those required for inhibition of AKR1C1 and AKR1C2. One indomethacin analogue, CBM, also does not inhibit the PGHS enzymes and provides a molecular probe for examining the contribution of AKR1C3 to breast cancer signaling and proliferation.

Acknowledgments

Supported by R01-CA90744 and P30-ES013508 awarded to T.M.P. MCB was funded by NIH training grants T32-DK007314-25 and T32-HD007305-22.

Address all correspondence to: Dr. Trevor M. Penning, Department of Pharmacology, University of Pennsylvania School of Medicine, 130C John Morgan Bldg, 3620 Hamilton Walk, Philadelphia, PA 19104-6084, Phone: (215)-898-9445, ude.nnepu.dem.liam@gninnep, Fax: (215)-573-2236

Abstract

Aldo-keto reductase (AKR) 1C3 catalyzes the NADPH dependent reduction of Δ-androstene-3,17-dione to yield testosterone, reduction of estrone to yield 17β-estradiol and reduction of progesterone to yield 20α-hydroxyprogesterone. In addition, it functions as a prostaglandin (PG) F synthase and reduces PGH2 to PGF and PGD2 to 11β-PGF2. Immunohistochemistry showed that AKR1C3 is over expressed in invasive ductal carcinoma of the breast. Retroviral expression of AKR1C3 in MCF-7 breast carcinoma cells shows that each of the assigned reactions occur in a breast cell microenvironment. Steroid and prostaglandin conversions were monitored by radiochromatography. Prostaglandin conversion was validated by a second method using HPLC coupled to APCI-MRM/MS. The combined effect of the AKR1C3 catalyzed 17- and 20-ketosteroid reductions will be to increase the 17β-estradiol : progesterone ratio in the breast. In addition, formation of PGF2 epimers would activate F prostanoid receptors and deprive PPARγ of its putative anti-proliferative PGJ2 ligands. Thus, AKR1C3 is a source of proliferative signals and a potential therapeutic target for hormone dependent and independent breast cancer. Two strategies for AKR1C3 inhibition based on non-steroidal anti-inflammatory drugs were developed. The first strategy uses the Ullmann coupling reaction to generate N-phenylanthranilate derivatives that inhibit AKR1C enzymes without affecting PGH2 synthase (PGHS) 1 or PGHS-2. The second strategy exploits the selective inhibition of AKR1C3 by indomethacin, which did not inhibit highly related AKR1C1 or AKR1C2. Using known structure activity relationships for the inhibition of PGHS-1 and PGHS-2 by indole acetic acids we obtained N-(4-chlorobenzoyl)-melatonin as a specific AKR1C3 inhibitor (KI = 6.0 μM) that does not inhibit PGHS-1, PGHS-2, AKR1C1, or AKR1C2. Both strategies are informed by crystal structures of ternary AKR1C3•NADP•NSAID complexes. The identification of NSAID analogs as specific inhibitors of AKR1C3 will help validate its role in the proliferation of breast cancer cells.

Keywords: Aldo-keto reductase, steroid hormone metabolism, prostaglandin metabolism, indomethacin, N-phenylanthranilic acids
Abstract

Footnotes

Conflict of Interest Statement: The authors declare that there are no conflicts of interest.

Footnotes

References

  • 1. Labrie F, Luu-The V, Lin SX, Simard J, Labrie C, El Alfy M, Pelletier G, Belanger AIntracrinology: role of the family of 17β-hydroxysteroid dehydrogenases in human physiology and disease. J Mol Endocrinol. 2000;25:1–16.[PubMed][Google Scholar]
  • 2. Suzuki T, Miki Y, Nakamura Y, Moriya T, Ito K, Ohuchi N, Sasano HSex steroid-producing enzymes in human breast cancer. Endocr Relat Cancer. 2005;12:701–720.[PubMed][Google Scholar]
  • 3. Penning TM, Burczynski ME, Jez JM, Hung CF, Lin HK, Ma H, Moore M, Palackal N, Ratnam KHuman 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]
  • 4. Dufort I, Rheault P, Huang XF, Soucy P, Luu-The VCharacteristics of a highly labile human type 5 17β-hydroxysteroid dehydrogenase. Endocrinology. 1999;140:568–574.[PubMed][Google Scholar]
  • 5. 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]
  • 6. Carnevale RP, Proietti CJ, Salatino M, Urtreger A, Peluffo G, Edwards DP, Boonyaratanakornkit V, Charreau EH, Bal de Kier JE, Schillaci R, Elizalde PVProgestin effects on breast cancer cell proliferation, proteases activation, and in vivo development of metastatic phenotype all depend on progesterone receptor capacity to activate cytoplasmic signaling pathways. Mol Endocrinol. 2007;21:1335–1358.[PubMed][Google Scholar]
  • 7. Hardy DB, Janowski BA, Chen CC, Mendelson CRProgesterone receptor inhibits aromatase and inflammatory response pathways in breast cancer cells via ligand-dependent and ligand-independent mechanisms. Mol Endocrinol. 2008;22:1812–1824.[Google Scholar]
  • 8. Matsuura K, Shiraishi H, Hara A, Sato K, Deyashiki Y, Ninomiya M, Sakai SIdentification of a principal mRNA species for human 3α-hydroxysteroid dehydrogenase isoform (AKR1C3) that exhibits high prostaglandin D2 11-ketoreductase activity. J Biochem (Tokyo) 1998;124:940–946.[PubMed][Google Scholar]
  • 9. Suzuki-Yamamoto T, Nishizawa M, Fukui M, Okuda-Ashitaka E, Nakajima T, Ito S, Watanabe KcDNA cloning, expression and characterization of human prostaglandin F synthase. FEBS Lett. 1999;462:335–340.[PubMed][Google Scholar]
  • 10. Sales KJ, Milne SA, Williams AR, Anderson RA, Jabbour HNExpression, localization, and signaling of prostaglandin F receptor in human endometrial adenocarcinoma: regulation of proliferation by activation of the epidermal growth factor receptor and mitogen-activated protein kinase signaling pathways. J Clin Endocrinol Metab. 2004;89:986–993.[PubMed][Google Scholar]
  • 11. Shiraki T, Kamiya N, Shiki S, Kodama TS, Kakizuka A, Jingami Hα,β-Unsaturated ketone is a core moiety of natural ligands for covalent binding to peroxisome proliferator-activated receptor γ J Biol Chem. 2005;280:14145–14153.[PubMed][Google Scholar]
  • 12. Straus DS, Pascual G, Li M, Welch JS, Ricote M, Hsiang CH, Sengchanthalangsy LL, Ghosh G, Glass CK15-deoxy-Δ-prostaglandin J2 inhibits multiple steps in the NF-κB signaling pathway. Proc Natl Acad Sci U S A. 2000;97:4844–4849.[Google Scholar]
  • 13. Kim HJ, Kim JY, Meng Z, Wang LH, Liu F, Conrads TP, Burke TR, Veenstra TD, Farrar WL15-Deoxy-Δ-prostaglandin J2 inhibits transcriptional activity of estrogen receptor-α via covalent modification of DNA-binding domain. Cancer Res. 2007;67:2595–2602.[PubMed][Google Scholar]
  • 14. Byrns MC, Duan L, Lee SH, Blair IA, Penning TMAldo-keto reductase 1C3 catalyzes 17β-estradiol and 11β-prostaglandin F2 formation and progesterone elimination in MCF-7 breast cancer cells and promotes estrogen-dependent cell proliferation. 16 June 2008; The Endocrine Society's 90th Annual Meeting; abstract # P2-P43. [PubMed]
  • 15. Bauman DR, Rudnick SI, Szewczuk LM, Jin Y, Gopishetty S, Penning TMDevelopment of nonsteroidal anti-inflammatory drug analogs and steroid carboxylates selective for human aldo-keto reductase isoforms: potential antineoplastic agents that work independently of cyclooxygenase isozymes. Mol Pharmacol. 2005;67:60–68.[PubMed][Google Scholar]
  • 16. 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 hormone independent malignancies. Biochem Pharmacol. 2008;75:484–493.[Google Scholar]
  • 17. Lin HK, Steckelbroeck S, Fung KM, Jones AN, Penning TMCharacterization of a monoclonal antibody for human aldo-keto reductase AKR1C3 (type 2 3α-hydroxysteroid dehydrogenase/type 5 17β-hydroxysteroid dehydrogenase); immunohistochemical detection in breast and prostate. Steroids. 2004;69:795–801.[PubMed][Google Scholar]
  • 18. Shibuya R, Suzuki T, Miki Y, Yoshida K, Moriya T, Ono K, Akahira J, Ishida T, Hirakawa H, Evans DB, Sasano HIntratumoral concentration of sex steroids and expression of sex steroid-producing enzymes in ductal carcinoma in situ of human breast. Endocr Relat Cancer. 2008;15:113–124.[PubMed][Google Scholar]
  • 19. Oduwole OO, Li Y, Isomaa VV, Mantyniemi A, Pulkka AE, Soini Y, Vihko PT17β-hydroxysteroid dehydrogenase type 1 is an independent prognostic marker in breast cancer. Cancer Res. 2004;64:7604–7609.[PubMed][Google Scholar]
  • 20. Jansson AK, Gunnarsson C, Cohen M, Sivik T, Stal O17β-hydroxysteroid dehydrogenase 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]
  • 21. Lee SH, Rangiah K, Williams MV, Wehr AY, DuBois RN, Blair IACyclooxygenase-2-mediated metabolism of arachidonic acid to 15-oxo-eicosatetraenoic acid by rat intestinal epithelial cells. Chem Res Toxicol. 2007;20:1665–1675.[PubMed][Google Scholar]
  • 22. Penning TM, Talalay PInhibition of a major NAD(P)-linked oxidoreductase from rat liver cytosol by steroidal and nonsteroidal anti-inflammatory agents and by prostaglandins. Proc Natl Acad Sci U S A. 1983;80:4504–4508.[Google Scholar]
  • 23. Lovering AL, Ride JP, Bunce CM, Desmond JC, Cummings SM, White SACrystal structures of prostaglandin D2 11-ketoreductase (AKR1C3) in complex with the nonsteroidal anti-inflammatory drugs flufenamic acid and indomethacin. Cancer Res. 2004;64:1802–1810.[PubMed][Google Scholar]
  • 24. Aronoff DM, Boutaud O, Marnett LJ, Oates JAInhibition of prostaglandin H2 synthases by salicylate is dependent on the oxidative state of the enzymes. Adv Exp Med Biol. 2003;525:125–128.[PubMed][Google Scholar]
  • 25. Dhagat U, Carbone V, Chung RP, Matsunaga T, Endo S, Hara A, El Kabbani OA salicylic acid-based analogue discovered from virtual screening as a potent inhibitor of human 20α-hydroxysteroid dehydrogenase. Med Chem. 2007;3:546–550.[PubMed][Google Scholar]
  • 26. Prusakiewicz JJ, Felts AS, Mackenzie BS, Marnett LJMolecular basis of the time-dependent inhibition of cyclooxygenases by indomethacin. Biochemistry. 2004;43:15439–15445.[PubMed][Google Scholar]
  • 27. Kalgutkar AS, Crews BC, Rowlinson SW, Marnett AB, Kozak KR, Remmel RP, Marnett LJBiochemically based design of cyclooxygenase-2 (COX-2) inhibitors: facile conversion of nonsteroidal antiinflammatory drugs to potent and highly selective COX-2 inhibitors. Proc Natl Acad Sci U S A. 2000;97:925–930.[Google Scholar]
  • 28. Kalgutkar AS, Marnett AB, Crews BC, Remmel RP, Marnett LJEster and amide derivatives of the nonsteroidal antiinflammatory drug, indomethacin, as selective cyclooxygenase-2 inhibitors. J Med Chem. 2000;43:2860–2870.[PubMed][Google Scholar]
  • 29. Kalgutkar AS, Crews BC, Saleh S, Prudhomme D, Marnett LJIndolyl esters and amides related to indomethacin are selective COX-2 inhibitors. Bioorg Med Chem. 2005;13:6810–6822.[PubMed][Google Scholar]
  • 30. Arun B, Goss PThe role of COX-2 inhibition in breast cancer treatment and prevention. Semin Oncol. 2004;31:22–29.[PubMed][Google Scholar]
  • 31. Su B, Diaz-Cruz ES, Landini S, Brueggemeier RWSuppression of aromatase in human breast cells by a cyclooxygenase-2 inhibitor and its analog involves multiple mechanisms independent of cyclooxygenase-2 inhibition. Steroids. 2008;73:104–111.[Google Scholar]
  • 32. Bauman DR, Steckelbroeck S, Penning TMThe roles of aldo-keto reductases in steroid hormone action. Drug News Perspect. 2004;17:563–578.[PubMed][Google Scholar]
Collaboration tool especially designed for Life Science professionals.Drag-and-drop any entity to your messages.