Context: Prostaglandins (PGs) E2 and PGF2α are produced in the endometrium and are important for menstruation and fertility. Dysmenorrhea is associated with increased production of PGF2α relative to PGE2, and the opposite is true for menorrhagia. The pathways leading to PGE2 biosynthesis are well described, but little is known for PGF2α. Aldoketoreductase (AKR)-1C3, the only PGF synthase identified in the human, cannot explain the production of PGF2α by endometrial cells. AKR1B1 appears to be an alternate candidate with promising therapeutic value.
Objective: The objective of the study was to address whether AKR1B1 (gene ID 231) is a functional PGF2α synthase in the human endometrium and a valid therapeutic target for menstrual pain.
Design: The design of the study was basic laboratory analyses to identify gene expression and protein levels associated with PGF2α production in endometrial tissues and endometrial cells from cycling women aged between 23 and 52 yr undergoing biopsies or hysterectomy for diverse gynecological disorders.
Results: AKR1B1 is expressed at a high level during the menstrual cycle during the secretory phase and in both epithelial and stromal cells, whereas AKR1C3 was found only in epithelial cells. Purified recombinant AKR1B1 protein, gene silencing, and transient transfection experiments all concur to demonstrate that this enzyme is a functional PGF synthase. Ponalrestat, a specific inhibitor developed to block AKR1B1 activity, reduced PGF2α production in response to IL-1β in both cultured endometrial cells and endometrial explants.
Conclusions: The human aldose reductase AKR1B1 currently associated with diabetes complications is also a highly functional PGF synthase responsible for PGF2α production in the human endometrium and a potential target for treatment of menstrual disorders.
Prostaglandins (PGs) are primary regulators of female reproductive function (ovulation, uterine receptivity, implantation, and parturition) and associated pathologies including endometrial carcinomas, menorrhagia, dysmenorrhea, endometriosis, and premature labor. PGE2 and PGF2α are the main prostanoids produced in the human endometrium (1, 2). PGs are synthesized from arachidonic acid (AA) and converted to PGG2 and PGH2 by either of two isoforms of PGH synthase (PGHS), the constitutive cyclooxygenase (COX)-1 or the inducible COX-2 encoded by two distinct genes (3). PGH2 produced by COXs is the common precursor of all PGs generated by specific terminal synthases such as PGF synthase (PGFS) for PGF2α and PGE synthase for PGE2. There are three known PGE synthases, and the expression and localization of the inducible microsomal PGES (mPGES-1) were studied in the human endometrium and found present in stromal, epithelial, and endothelial cells throughout the menstrual cycle (4). Despite its demonstrated role as the PG responsible for dysmenorrhea (menstrual pain), no data are available on PGFS activity and its localization in relation with PGF2α production within the human endometrium.
PGF2α is a bioactive lipid belonging to the eicosanoid family (5). Its biosynthesis occurs via reduction of PGH2 by a 9,11-endoperoxyde reductase (6). Several PGFSs have been identified in animals (7), but aldoketoreductase (AKR)-1C3 is the only one currently identified in humans (8). AKR1C3 (gene ID 8644), an aldoketoreductase of the 1C family generally associated with hydroxysteroid dehydrogenase (HSD) activity, has been studied primarily for its type V 17β-HSD activity and in this respect was found in the human endometrium (9).
In the bovine endometrium, we have shown a strong association between AKR1B5, recently renamed as bos taurus AKR1B1 (gene ID 317748), and PGF2α production, which is a new function for this enzyme previously known for its 20α-HSD and glucose metabolism activities (7). The human and bovine AKR1B1 belong to the AKR1B family and share 86% identity or homology. The human AKR1B1 (gene ID 231), also known as aldose reductase, is highly expressed in the placenta for glucose metabolism and in the eye and kidney for osmotic regulation (10).
We have found that AKR1B1 expression was associated with PGF2α production in human endometrial cell lines (11) and in decidualized stromal cells (12). However, in vivo expression of AKR1B1 within the human endometrium and its ability to act as a PGFS to produce PGF2α remain to be demonstrated. In a recent study, Talbi et al. (13) used high-throughput DNA array to investigate the expression of more than 54,600 genes in the human endometrium throughout the menstrual cycle. Among the full list of genes available online at GeoProfile, COX-2 and AKR1B1 were shown to be modulated during the menstrual cycle and increased in association with IL-1β, thus supporting the work of Rossi (14), who showed up-regulation of AKR1B1 mRNA by IL-1β in human endometrial stromal cells using cDNA microarray analysis.
Therefore, we studied the expression of both AKR1B1 and AKR1C3 at the mRNA and protein levels in the human endometrium across the menstrual cycle. We also investigated their association with the production of PGF2α and response to inflammatory IL-1β using human endometrial biopsies and cell lines.
Materials and Methods
Reagents
RPMI 1640 culture medium, Superscript II reverse transcriptase, TRIzol, Lipofectamine 2000, pCR3.1, and pEF6/V5 TA cloning vectors were from Invitrogen (Burlington, Ontario, Canada). TAQ DNA polymerase and PCR buffer were from Pharmacia (Montréal, Québec, Canada), and QuantiTectTM Syber Green PCR kit were from QIAGEN (Mississauga, Ontario, Canada). RiboMax polymerase kit were from Promega (Madison, WI), and Qiaquick gel extraction kit were from QIAGEN. All oligonucleotide primers were chemically synthesized using ABT 394 synthase (PerkinElmer, Foster City, CA). [α-32P]dCTP radioactivity was from PerkinElmer Life Sciences (Markham, Ontario, Canada). Bright Star-Plus nylon membrane and UltraHyb solution were from Ambion Inc. (Austin, TX). Goat polyclonal antibody (pAB) to AKR1C3 were from Abcam Inc. (Cambridge, MA), and rabbit COX pABs were from Dr. S. Kargman (Merck, Québec, Québec, Canada). Recombinant AKR1B1 was produced as described previously (11). Biotinylated secondary antibodies (goat antirabbit IgG or rabbit antigoat IgG) were from Dako Diagnostic of Canada, Inc. (Mississauga, Ontario, Canada). Vectastain Elite ABC kit was from Vector Laboratories Inc. (Burlingame, CA). Goat antirabbit horseradish peroxidase-conjugated IgG was from Jackson ImmunoResearch Laboratories (West Grove, PA). Enhanced chemiluminescence system (Renaissance) was from NEN Life Science Products (Boston, MA). AA and mPGES-1 pABs were from Cayman Chemicals (Ann Arbor, MI), and recombinant human IL-1β was from Research and Diagnostic Systems (Minneapolis, MN). Aldose reductase inhibitor (ARI) Ponalrestat (Statil) was from Tocris Bioscience (Ellisville, MO).
Endometrial tissue collection and cell lines
Endometrial tissue was obtained by biopsy or hysterectomy. Biopsies were collected with an endometrial curette (Pipelle) from women aged 25–50 yr undergoing gynecological investigation for infertility or menorrhagia. Hysterectomies originated from cycling women aged between 29 and 53 yr presenting with chronic pelvic pain, leiomyomas, and menorrhagia. The endometrium was isolated by gentle scraping of the internal lining of intact portions of the uterus selected by the hospital pathologist. Inclusion criteria included regular menstrual cycles (21–35 d) and absence of hormonal treatment in the 3 months before biopsy collection. Endometrial samples were dated according to the stated last menstrual period confirmed by histological examination using the criterion of Noyes et al. (15). Samples with conflicting dating were discarded. The research protocol was approved by the Ethics Committee on Human Research of the Centre Hospitalier Universitaire de Québec, and informed consent was obtained before collection. Tissue samples were washed and processed differentially for explant culture and RNA and protein analysis. The human endometrial stromal and epithelial cell lines HIESC and HIEEC were generated and characterized in our laboratory (11).
Assessment of COX-2, AKR1B1, and AKR1C3 mRNAs by quantitative RT-PCR (qRT-PCR)
For qRT-PCR, RNA from endometrial biopsies (n = 48) was prepared in TRIzol reagent (Invitrogen) and samples stored at −80 C. qRT-PCR was performed using a Roche Light Cycler (Roche, Mississauga, Ontario, Canada) and QuantiTectTM Syber Green PCR kit (QIAGEN) in duplicate as described previously (16). The following primers were used: COX-1 forward, 5′-CCCACCAGTTCTTCAAAACTTC-3′ (positions 749–770), reverse, 5′-CCTGGTACTTGAGTTTCCCATC-3′ (positions 898–877); COX-2 forward, 5′-ATGAGTACCGCAAACGCTTTAT-3′ (positions 1486–1507), reverse, 5′-GAATGGTGCTCCAACTTCTACC-3′ (positions 1682–1661); AKR1B1 forward, 5′-CTACCTTATTCACTGGCCGACT-3′ (positions 395–416), reverse, 5′-GTTGGAGATGCCAATAGCTTTC-3′ (positions 557–536); and AKR1C3 forward, 5′-TCTCTAAAGCCAGGTGAGGAAC-3′ (positions 430–451), reverse, 5′-TACTTGAGTCCTGGCTTGTTGA-3′ (positions 620–599). The 18S RNA served as an internal control to normalize the expression of each gene.
Immunohistochemistry and evaluation of immunostaining
Thin tissue sections (3 μm) from endometrial biopsies taken at different periods of the menstrual cycle (n = 18; six periods and three samples per period) were fixed in 4% paraformaldehyde and prepared as paraffin-embedded sections as described previously (16). The staining was evaluated by three blinded observers using the following subjective scoring system: absent (1+), weak (2+), moderate (3+), or intense (4+). Individual scores for each slide were averaged and expressed as relative expression level.
Endometrial explants culture
Explants were prepared as follows; endometrial samples were washed several times in PBS solution (DPBS; Wisent, Montréal, Québec, Canada), cut into 1- to 2-mm3 pieces with a scalpel blade, washed, and placed in 5 ml RPMI 1640 (without Phenol Red) containing 2% steroid-stripped fetal bovine serum, antimycotic, and penicillin-streptomycin, and incubated overnight at 37 C in a humidified atmosphere of 5% CO2-95% air. The next day, the explants were washed three times with RPMI 1640 and plated on 24 well-plates. Each well contained 1 ml of RPMI 1640 containing 10 ng/ml of IL-1β in the presence or absence of 100 μm Statil and cultured at 37 C for 6 h. At the end of the incubation, the medium was harvested by centrifugation for measurement of PGF2α, whereas explants tissue (pellet) were suspended in 200 μl lysis solution (20 mm Tris-HCl, pH 7.5; 1 mm dithiothreitol; 1% sodium dodecyl sulfate; 1 mm phenylmethylsulfonyl fluoride) and homogenized three times (5–10 sec) using a handheld homogenizer (BF20, 7 × 95 mm saw tooth; Troemner, Thorofare, NJ). Subsequently 600 μl of methanol was added to the homogenate to precipitate protein as described previously (17) for analysis by Western blotting. Proteins were quantified by image analysis using the AlphaImager 2000 software (Alpha Innotech Corp., San Leandro, CA).
AKR1B1-specific small interfering RNA (siRNA)
Specific siRNAs for AKR1B1 were designed using an optimization program (18) with T7 RNA polymerase (forward, 5′-aaattgttgagcaggagacggctatagtgag-tcgtattacc-3′ and the reverse, 5′-aagccgtctcctgctcaacaactatagtgagt cgtattacc-3′) according to the procedure of Donzé and Picard (19) using a RiboMax kit (Promega). The resulting siRNA products were purified by ethanol precipitation, and 100 ng/ml was used for transfection of cells grown in six- or 24-well plates using Lipofectamine 2000 (Invitrogen).
Northern and Western blots
Northern and Western blot analyses were performed as we described previously (16).
Cell transfection and treatments
Immortalized human endometrial stromal (HIESC) and epithelial (HIEEC) cells were produced and used as we described recently (11). Briefly, HIESC and HIEEC cells were grown to confluence, and knockdown with specific siRNAs and knock-in transfection with AKR1B1 or AKR1C3 cDNAs in pCR3.1 expression vectors were performed with Lipofectamine 2000 (Invitrogen) for 4 h. After 36 h, cells were treated for 24 h with IL-1β (1 ng/ml) or AA 10 μm in RPMI 1640 medium without serum. Cells were grown in 24-well plates for Western and six-well plates for Northern analyses. At the end of the treatment period, the culture medium was recovered and stored at −20 C until evaluation of PGF2α production.
AKR1B1 enzyme activity
AKR1B1 recombinant protein was overexpressed in Escherichia coli (BL21), purified, and enzymatic activity determined as we described for bovine AKR1B5 (7). The production of PGF2α was confirmed by thin-layer chromatography (TLC) using silica plates. Migration was performed in ethyl acetate (110:50:20) water-saturated solvent, and detection was achieved by spraying phosphomolybdic acid 10% (vol/vol) in methanol followed by heating on a plate at 120 C for 10 min.
Prostaglandin F2α
PGF2α was measured by enzyme immunoassay, and acetylcholinesterase-linked PGF2α tracer (Cayman) was used as described previously (20).
Statistical analysis
Data are presented as the mean ± sem. Statistical significance was assessed by one-way ANOVA using GraphPad Prism 5 (San Diego, CA). If the null hypothesis was rejected, Tukey’s multiple comparison was used as a post hoc test to find the critical difference between pairs of treatment means. In all the experiments, confidence level was set at 95% to determine the significance of difference (P < 0.05).
Results
Gene expression of COX-1, COX-2, AKR1B1, and AKR1C3
Analysis of mRNA expression was performed by quantitative RT-PCR (Fig. 1) in endometrial biopsies collected at different periods of the menstrual cycle. The results were variable among samples, a result coherent with their involvement with inflammation responses, but showed that mRNA was expressed for all four genes tested. COX-1 mRNA expression was globally higher in the secretory than in the proliferative phase. COX-2 mRNA levels were highest at the early secretory phase. AKR1C3 and AKR1B1 mRNA were present throughout the cycle without significant variation.
Expression of COX-1, COX-2, AKR1B1, and AKR1C3 mRNA during the menstrual cycle. Endometrial biopsies (n = 30) were obtained at different days of the menstrual cycle and total RNA was extracted. Target mRNAs were evaluated by qRT-PCR as described in Materials and Methods. Data are presented at six periods of the menstrual cycle. Results are expressed as mRNA quantity relative to 18S rRNA for each enzyme. Bars represent the mean + sem for each group (n = 3–5 for each bar). The same extracts were used for all enzymes tested.
Immunohistochemical analysis of AKR1B1 and AKR1C3 protein
Immunohistochemical staining for COX-1, COX-2, AKR1B1 and AKR1C3 was performed in endometrial samples collected at different phases of the menstrual cycle and evaluated by subjective analysis (Fig. 2, A and B). COX-1 protein was present at all periods except the late proliferative phase (Fig. 2B). COX-2 followed a similar pattern in epithelial and stromal compartments and was highest in the mid- and late-secretory phase (Fig. 2B). AKR1C3 protein staining was intense and fairly constant throughout the cycle in epithelial cells but completely absent in the stromal compartment (Fig. 2, A and B). AKR1B1 protein was present in luminal and glandular epithelial cells as well as in stromal cells of the endometrium (Fig. 2A). Higher expression was found in the early proliferative and mid-late secretory compared with other phases of menstrual cycle when both epithelial and stromal cell compartments were considered simultaneously (Fig. 2B).
A, Immunohistochemical analysis of AKR1B1 and AKR1C3 protein expression in human endometrium during the menstrual cycle. Representative endometrial samples labeled with AKR1C3 or AKR1B1 antibodies. On the left we show background coloration in the presence of preimmune serum. For each enzyme, immunohistochemistry (IHC) stainings are shown at the midproliferative (MP) or midsecretory (MS) period. Note the abundant expression of AKR1B1 in luminal and glandular epithelium and stroma during the secretory phase and absence of AKR1C3 in the stromal compartment. B, Subjective quantitation of PG biosynthetic enzymes in the human endometrium. Endometrial biopsies (n = 18) were incubated with specific antibodies against AKR1B1, AKR1C3, COX-1, and COX-2. IHC staining intensity levels were quantitated subjectively by three different observers. Staining in epithelial and stromal cells was considered separately and intensity scored between 1 (no staining) and 4 (clear heavy staining). Results represent the mean ± sem of the scores of the three independent observers.
Down-regulation of AKR1B1 expression by siRNA
Gene and protein expression of AKR1B1 was tested in the absence or presence of IL-1β, a known regulator of PG production, and correlated with the production of PGF2α in cultured endometrial stromal cells. Northern and Western blot analysis of stromal cell line HIESC treated with IL-1β showed that AKR1B1 mRNA and protein were significantly reduced after transfection of the cells with AKR1B1 siRNA (Fig. 3A). The same observation was found for the epithelial cell line HIEEC (results not shown). AKR1B1 siRNA did not affect COX-2 protein level after treatment with IL-1β (Fig. 3B). The decrease in AKR1B1 by specific siRNA knockdown was associated with a significant reduction in PGF2α production (P < 0.05) (Fig. 3C).
AKR1B1 silencing. HIESC2 cells were transfected with a selective AKR1B1 siRNA to down-regulate AKR1B1 gene expression and compared with nontransfected cells at the level of mRNA, protein, and PGF2α production. A, Comparison of AKR1B1 expression at the mRNA (Northern) and protein levels (Western blot). B, Effect of AKR1B1 siRNA on COX-2 and AKR1B1 protein expression in presence or not of IL-1β (1 ng/ml). C, PGF2α accumulation in culture medium of cells transfected with AKR1B1 siRNA in presence or not of IL-1β.
PGFS activity of AKR1B1
AKR1B1 recombinant protein was found to oxidize the reduced form of nicotinamide adenine dinucleotide phosphate at 10 nmol/min · mg in the presence of 40 μm PGH2 as monitored by absorbance at 340 nm (Fig. 4A). The conversion of PGH2 into PGF2α was confirmed by TLC (Fig. 4A). AKR1B1 PGFS activity within cells in vitro was confirmed by gain of function and compared with AKR1C3. After transfection with full-length AKR1B1 or AKR1C3 cDNA HIESC exhibited increased levels of the corresponding proteins (Fig. 4B). Treatment of the transfected cells with 10 μm AA results in a higher production of PGF2α with AKR1B1 than AKR1C3 (Fig. 4C). Together, these observations confirm AKR1B1 protein as a functional PGFS in endometrial cells.
PGFS activity of AKR1B1. A, AKR1B1 activity was measured in vitro using 10–100 μg of purified recombinant enzyme in presence of 40 μm PGH2. PGs were extracted and analyzed by TLC as described in Materials and Methods. B, Western blot analysis of AKR1B1 and AKR1C3 proteins after transfection with respective expression vectors in HIESC2 cells to overexpress these enzymes. C, Effect of increased AKR1B1 and AKR1C3 protein level on PGF2α production in the presence and the absence of exogenous AA.
Association between AKR1B1 and PGF2α production in endometrial explants
To confirm the physiological relevance of our observations in vitro, fresh endometrial biopsies were used to study the effect of IL-1β on AKR1B1 stimulation and PGF2α production. In endometrial explants IL-1β (10 ng/ml), stimulation elevated COX-2 and AKR1B1 protein levels (Fig. 5A) and increased PGF2α production (Fig. 5B) at variable levels. The high production of PGF2α in explant 2 appears to be related to a particularly high expression of COX-2.
Regulation of PGF2α production by AKR1B1 in endometrial explant cultures. Three different endometrial explants were prepared from fresh biopsies and treated with IL-1β. A, Effect of IL-1β (10 ng/ml) treatment for 6 h on COX-2 and AKR1B1 protein levels analyzed by Western blot. B, PGF2α production after stimulation with IL-1β (10 ng/ml) for 6 h for the corresponding endometrial explants.
Inhibition of PGFS activity of AKR1B1 with ponalrestat
Endometrial cells were stimulated with IL-1β in the presence of increasing concentrations of ponalrestat (Statil), a functional aldose reductase inhibitor developed to inhibit the conversion of glucose to sorbitol by AKR1B1 (21). A dose-dependent inhibition of PGF2α production was observed in the two cell types (Fig. 6A) but appeared more complete in stromal than epithelial cells. When endometrial explants were treated with IL-1β at 10 ng/ml in the presence of ponalrestat at 100 μm, PGF2α production was blocked below unstimulated levels even in explant 5, which did not respond to IL-1β (Fig. 6B).
Effect of the ARI ponalrestat (Statil) on PGF2α production in endometrial cells and endometrial explants. A, Endometrial epithelial (HIEEC-22) and stromal (HIESC-2) cells were grown to confluence and treated with increasing doses of Statil in presence of IL-1β at 10 ng/ml. After 6 h the culture medium was recovered for analysis of PGF2α by enzyme immunoassay. Results are expressed as a percent of production in the presence of IL-1β alone and represent the mean ± sem of three experiments. B, Three endometrial explants (different from those in Fig. 5) have been stimulated by IL-1β at 10 ng/ml for 6 h in the presence or absence of Statil (100 μm). Culture medium was harvested for the measurement of PGF2α expressed in terms of picograms per microgram of proteins.
Discussion
PGs are important regulators of female reproductive function and contribute to gynecological disorders. Normal menstruation depends on an equilibrium between vasoconstrictors such as PGF2α and vasodilators such as PGE2 (22). Excessive production of contracting prostaglandins create an ischemia-reperfusion response causing painful menstruation or dysmenorrhea, whereas increased vasodilatation leads to abundant menstrual bleeding (23). Nonsteroidal antiinflammatory drugs (NSAIDs), the most widely used drugs on the market, are efficient to treat menstrual disorders at different degrees. However, these drugs act nonspecifically at an early step of biosynthesis common to all PGs. Whereas the short duration of NSAIDs needed to treat menstrual pain should not lead to serious side effects, long-term treatment with NSAIDs such as naproxen leads to serious gastrointestinal side effects (24). It was also reported that the chronic use of any NSAID can impair ovulation and reduce fertility (25).
Because of its notorious role on inflammation and pain, the biosynthetic pathway leading to PGE2 has been studied extensively, but that of PGF2α is poorly documented. We studied for the first time in the human endometrium two potential PGFSs, AKR1B1 and AKR1C3, and their functional association with PGF2α production. It has been reported that production of PGF2α is higher in late secretory and menstrual periods of the menstrual cycle (26). Accordingly, both AKR1B1 and AKR1C3 enzymes are present in the endometrium throughout the menstrual cycle. By contrast with AKR1B1 present in both stromal and glandular epithelial cells and modulated in parallel with endometrial PGF2α production, AKR1C3 is absent in stromal cells, as was reported previously (9). The absence of the only currently accepted human PGFS, AKR1C3, in stromal cells (Fig. 2) was surprising because we and others have shown that human endometrial stromal cells produce high levels of PGF2α (11, 12, 27). Because of a similar finding in the bovine endometrium and our identification of AKR1B5 as a potential PGFS in that system (7), we hypothesized that the corresponding human enzyme AKR1B1 could possess PGFS activity in the human endometrium. We have shown in previous studies that indeed AKR1B1 was expressed in human endometrial cells and modulated in parallel with PGF2α production, but its PGFS activity remained to be demonstrated unequivocally.
In a global study of endometrial transcriptome, Talbi et al. (13) have documented the expression of 54,600 genes using high-throughput DNA array, including significant alterations in expression of COX and AKR genes. In the present study, we confirm the expression of AKR1B1, AKR1C3, COX-1, and COX-2 mRNAs with significant variation among samples and some regulation during the cycle. This is not surprising, given that biopsies were obtained from women with benign endometrial pathologies. Moreover, PG release is rapidly triggered by microbial infections, which may have been present in the endometrium or introduced during the sampling procedure. Finally, the potential modulation of mRNA during the cycle may represent a variation in the relative proportion of epithelial and stromal cells for genes like AKR1C3 exhibiting heterogeneous expression between the cell types. At the protein level, COX-2 expression is increased in both endometrial cell types at the end of the proliferative period together with high expression of AKR1B1 (Fig. 2). This is consistent with the proposed role of PGF2α in the initiation of menstruation and its association with menstrual pain at the beginning of the cycle.
We demonstrated the ability of the purified recombinant human AKR1B1 to release PGF2α and metabolize PGH2 in vitro in the presence of nicotinamide adenine dinucleotide phosphate oxidase (Fig. 4A). As for the bovine isoform, the human AKR1B1 can metabolize PGH2 and form PGF2α with a high efficiency. In fact, AKR1B1 uses PGH2 at concentrations within the physiological range, whereas it was shown to process glucose only at supraphysiological concentrations found primarily in diabetes (10). Recently, using different recombinant AKR proteins, Kabututu et al. (28) have shown that AKR1B1 is 20 times more potent than AKR1C3 in producing PGF2α from its precursor PGH2. We also show here that transfection of endometrial cells to increase AKR1B1 stimulates PGF2α release. By contrast, down-regulation of AKR1B1 using siRNA (Fig 3) significantly reduced the production of PGF2α in IL-1β-stimulated cells. Similarly, ponalrestat (Statil), a known ARI, dose dependently reduced PGF2α production in the same cells (Fig 6).
The physiological relevance of the results with endometrial cells was verified using endometrial explants treated with IL-1β (Fig. 5) and Statil (Fig. 6). Note that the pattern of response is similar to that of endometrial cell lines but exhibits higher variability. This may be attributed to the different pathophysiological status of the tissues. Endometrial explants have been used to study apoptosis during menstruation (29) and PGE2 release in women with abnormal uterine bleeding (30).
We have found that transfection of epithelial cells with AKR1B1 increased the production of PGF2α, whereas knocking down its expression with specific siRNA reduced it (results not shown). We have also confirmed the PGFS activity of AKR1C3 after transfection of endometrial stromal cells and found slightly increased PGF2α production (results not shown). Because AKR1C3 is expressed only in epithelial cells that represent only a small fraction of endometrial functionalis and because this enzyme is neither modulated during the cycle nor stimulated by IL-1β, its contribution to the release of endometrial PGF2α is probably marginal. This is supported by our observation that AKR1C3 is much less efficient than AKR1B1 to trigger increased PGF2α production after overexpression of the two enzymes (Fig. 4).
IL-1β is an important regulator of endometrial PG production that also induces apoptosis in epithelial cells of the endometrium (31), thus contributing to the initiation of menstruation (32). Interestingly, a cDNA microarray study of 15,164 sequence-verified clones identified AKR1B1 as an important gene up-regulated by IL-1β in human endometrial cells (14), supporting our observation that it is a key inducible endometrial protein (11). This finding was confirmed in the endometrial gene profiling experiment reported by Talbi et al. (13) and a cluster analysis of the genes considered in the present study using data available on Geo Profile, indicating that AKR1B1 and IL-1b followed a similar expression profile.
Together these results suggest that AKR1B1 is likely the functional PGFS responsible for PGF2α production in the human endometrium. AKR1B1 is a well-known and widely studied enzyme, but its contribution to PG production had never been suspected until we proposed that the corresponding bovine AKR1B5 had PGFS activity (7). AKR1B1 has been traditionally associated with reduction of glucose and diabetes-induced oxidative stress (33). Accordingly, AKR1B1 knockout mice have been used to study the pathogenesis of various diseases associated with diabetes mellitus such as cataract, retinopathy, neuropathy, and nephropathy but did not present any obvious reproductive phenotype (34). The mouse may not be the best model to study human aldose reductase because two distinct genes and proteins, AKR1B3 and AKR1B7, appear to exert redundant activity similar to AKR1B1 (28). Interestingly, transgenic mice overexpressing human AKR1B1 were more prone to myocardial ischemic injury (35), whereas knockout mice appeared protected against cerebral ischemic injury (36). These observations suggest that AKR1B1 is involved in the regulation of vascular tone by mechanisms distinct from those of glucose metabolism. Interestingly, ischemia is a documented response to PGF2α (37), and menstrual pain or uterine ischemia is induced by exogenous PGF2α (38).
The different PGs induce a wide variety of responses mediated principally through distinct G protein-coupled receptors (5). In the vascular system, thromboxane A2 and PGI2 exert opposing action on coagulation and vascular tone to regulate hemostasis (37). In the reproductive system, a similar opposing action is often observed for PGE2 and PGF2α (39). There have been reports showing that some terminal synthases are preferentially associated with a specific COX such as mPGES-1 with COX-2 or mPGES-2 with COX-1 (40), but no such association was found for PGFSs (41). In this respect, acetylsalicylic acid, the first marketed NSAID (aspirin), exhibits a slight preference for COX-1 and platelets, thus yielding preferential inhibition of thromboxane A2 over PGI2 in the vascular system (42). Similarly, the recently developed coxibs such as BEXTRA and VIOXX are COX-2 selective and have proven extremely efficient to reduce pain and inflammation induced by PGE2. Unfortunately, the chronic use of these drugs was associated with increased risk of heart failure, whereas other common NSAIDs such as ibuprofen, affecting both COX-1 and COX-2, appeared safer (43). Therefore, acting at the level of terminal synthases responsible for the release of specific PG isotypes appear to be promising for the control of the release of undesirable PGs and allowing actions of desirable PGs (44).
AKR1B1 was first identified as a key enzyme of the polyol pathway strongly associated with complications of metabolic disorders such as type 2 diabetes and more recently as a detoxification enzyme involved in the reduction of a wide range of carbonyl compound including benzaldehyde derivatives, quinones, sugars, and many lipid peroxidation end-products such as 4-hydroxy trans-2-nonenal and acrolein (10). Because of the known association of AKR1B1 with pathological conditions, ARIs such as ponalrestat were developed but most were found inefficient. The present finding that AKR1B1 is a functional PGFS liberating the bioactive PGF2α metabolite was unexpected and is highly challenging. Indeed, the newly discovered enzyme activity will have to be considered as contributing to the pathophysiological processes associated with aberrant expression of AKR1B1. This result opens entirely new avenues to study the mechanisms underlying menstrual disorders and other pathologies. Because the conversion of glucose into sorbitol by AKR1B1 may exhibit enzyme dynamics different from biosynthesis of PGF2α, there may be existing ARIs that have gone through toxicity tests and different phases for human use that could represent a new class of medications to treat menstrual disorders.
This work was supported by Canadian Institutes for Health Research and a Wyeth-Ayerst/Canadian Institutes for Health Research postdoctoral fellowship (to E.B.).
Disclosure Summary: E.B., P.C., E.M., and M.A.F. are inventors on international patent application PCT/CA2008/002012 owned by Laval University. S.B.-K., N.H., P.Y.L., and M.L. have nothing to declare.
Abbreviations:
- AA,
Arachidonic acid;
- AKR,
aldoketoreductase;
- ARI,
aldose reductase inhibitor;
- COX,
cyclooxygenase;
- HSD,
hydroxysteroid dehydrogenase;
- mPGES-1,
microsomal PGES;
- NSAID,
nonsteroidal antiinflammatory drug;
- pAB,
polyclonal antibody;
- PG,
prostaglandin;
- PGFS,
PGF synthase;
- PGHS,
PGH synthase;
- qRT-PCR,
quantitative RT-PCR;
- siRNA,
small interfering RNA;
- TLC,
thin-layer chromatography.
