The Prostaglandin Pathway is Activated in Patients who Fail Medical Therapy for Benign Prostatic Hyperplasia (BPH) with Lower Urinary Tract Symptoms (LUTS)

Abstract

Background:

Little is known about how benign prostatic hyperplasia (BPH) develops and why patients respond differently to medical therapy designed to reduce lower urinary tract symptoms (LUTS). The Medical Therapy of Prostatic Symptoms (MTOPS) trial randomized men with symptoms of BPH and followed response to medical therapy for up to 6 years. Treatment with a 5α-reductase inhibitor (5ARI) or an alpha-adrenergic receptor antagonist (α-blocker) reduced risk of clinical progression, while men treated with combination therapy showed a 66% decrease in risk of progressive disease. However, medical therapies for BPH/LUTS are not effective in many patients. The reasons for non-response or loss of therapeutic response in the remaining patients over time are unknown. A better understanding of why patients fail to respond to medical therapy may have a major impact on developing new approaches for the medical treatment of BPH/LUTS. Prostaglandins (PG) act on G-protein-coupled receptors (GPCRs), where PGE₂ and PGF₂ elicit smooth muscle contraction. Therefore, we measured PG levels in the prostate tissue of BPH/LUTS patients to assess the possibility that this signaling pathway might explain failure of medical therapy in BPH/LUTS patients.

Method:

Surgical BPH (S-BPH) was defined as benign prostatic tissue collected from the transition zone (TZ) of patients who failed medical therapy and underwent surgical intervention to relieve LUTS. Control tissue was termed Incidental BPH (I-BPH). I-BPH was TZ obtained from men undergoing radical prostatectomy for low-volume, low-grade prostatic adenocarcinoma (PCa, Gleason score ≤7) confined to the peripheral zone. All TZ tissue was confirmed to be cancer-free. S-BPH patients divided into four subgroups: patients on α-blockers alone, 5ARI alone, combination therapy (α-blockers plus 5ARI), or no medical therapy (none) prior to surgical resection. I-BPH tissue was sub grouped by prior therapy (either on α-blockers or without prior medical therapy before prostatectomy). We measured prostatic tissue levels of prostaglandins (PGF_2α, PGI₂, PGE₂, PGD₂, and TxA₂), qPCR levels of mRNAs encoding enzymes within the PG synthesis pathway, cellular distribution of COX1 (PTGS1) and COX2 (PTGS2), and tested the ability of PGs to contract bladder smooth muscle in an in vitro assay.

Results:

All PGs were significantly elevated in TZ tissues from S-BPH patients (n=36) compared to I-BPH patients (n=15), regardless of the treatment subgroups. In S-BPH versus I-BPH, mRNA for PG synthetic enzymes COX1 and COX2 were significantly elevated. In addition, mRNA for enzymes that convert the precursor PGH₂ to metabolite PGs were variable: PTGIS (which generates PGI₂) and PTGDS (PGD₂) were significantly elevated; non-significant increases were observed for PTGES (PGE₂), AKR1C3 (PGF_2α ), and TBxAS1 (TxA₂). Within the I-BPH group, men responding to α-blockers for symptoms of BPH but requiring prostatectomy for PCa did not show elevated levels of COX1, COX2, or PGs. By immunohistochemistry, COX1 was predominantly observed in the prostatic stroma while COX2 was present in scattered luminal cells of isolated prostatic glands in S-BPH. PGE₂ and PGF_2α induced contraction of bladder smooth muscle in an in vitro assay. Furthermore, using the smooth muscle assay, we demonstrated that α-blockers that inhibit alpha-adrenergic receptors do not appear to inhibit PG stimulation of GPCRs in bladder muscle. Only patients who required surgery to relieve BPH/LUTS symptoms showed significantly increased tissue levels of PGs and the PG synthetic enzymes.

Conclusions:

Treatment of BPH/LUTS by inhibition of alpha-adrenergic receptors with pharmaceutical α-blockers or inhibiting androgenesis with 5ARI may fail because of elevated paracrine signaling by prostatic PGs that can cause smooth muscle contraction. In contrast to patients who fail medical therapy for BPH/LUTS, control I-BPH patients do not show the same evidence of elevated PG pathway signaling. Elevation of the PG pathway may explain, in part, why the risk of clinical progression in the MTOPS study was only reduced by 34% with α-blocker treatment.

1. INTRODUCTION

Except for increasing age and prostate volume, no clinical variable reliably predicts whether a given patient will respond to medical therapy for benign prostatic hyperplasia (BPH) or lower urinary tract symptoms (LUTS). The prevalence of BPH in 60-year-old men is >50%; by 85 years of age it is >90%¹. BPH patients experience a spectrum of symptoms that are gauged by the International Prostate Symptom Score (IPSS). Standard medical treatments for BPH/LUTS include steroid 5α-reductase inhibitors (5ARI) that target steroid-5α-reductase type I and/or type II (SRD5A1 and SRD5A2). This blocks production of dihydrotestosterone (DHT) from testosterone (T) to shrink prostatic volume. Alpha-adrenergic receptor antagonists (α-blockers) are another therapy that increase voiding efficiency by reducing smooth muscle tone.

To study BPH response to medical therapy, the Medical Therapy of Prostatic Symptoms (MTOPS) randomized 3,047 men >63 years of age with BPH to various medical treatment regimens and followed them for up to 6 years¹. Treatment with either a α-blocker or a 5ARI only reduced risk of clinical progression by 34%, while combination of 5ARI (finasteride) plus a α-blocker (doxazosin) reduced the risk of clinical progression by 66%¹. It is not known why so many patients failed to respond or progressed after initially responding to medical therapy. Since finasteride inhibits only SRD5A2, a dual inhibitor, dutasteride, which blocks both type I and II isoforms has been developed and tested in combination with a α-blocker in the ComBAT study². In this study, patients on a combined regimen of α-blocker (tamsulosin) and 5ARI (dutasteride) had a 41% reduced risk of clinical progression compared to tamsulosin monotherapy. There was no comparison group without medication. Therefore, although dutasteride suppresses serum DHT levels more than finasteride, their clinical efficacy in reducing urinary tract symptoms is thought to be similar.

Several other medication classes are also available for male LUTS. Overactive bladder (OAB) can be caused by bladder outlet obstruction or bladder dysfunction. In patients with OAB, anti-cholinergic agents are frequently recommended, but these medications are often associated with intolerable side effects resulting in discontinuation of the medication³. As an alternative, a β3-adrenoceptor agonist has been used to treat OAB. A recent meta-analysis indicated that combining alpha-adrenergic blockade with mirabegron, a beta-adrenergic agonist that induces smooth muscle relaxation, was a safe and effective treatment regimen for OAB⁴. A novel pathway driving therapeutic failure in BPH was suggested by patients reporting relief of LUTS after taking erectile dysfunction (ED) drugs, specifically cGMP phosphodiesterase type 5 (PDE5A) inhibitors (PDE5-I)⁵. Clinical trials of patients using sildenafil, vardenafil, and tadalafil showed an improvement in LUTS due to BPH^6–8. A review of multiple studies concluded that these PDE5-Is combined with α-blockers provided a modest but significant benefit to patients, resulting in FDA approval of tadalafil for treatment of BPH⁹. Although limited data exist on the pharmacodynamics of PDE5-I, recent research shows that they do regulate smooth muscle contractility¹⁰. However, addition of dutasteride and tadalafil have only marginally improved the treatment of BPH/LUTS, suggesting that new targets must be pursued.

Since prostaglandins (PGs) were first discovered in seminal fluid, they were named after the prostate gland¹¹. Subsequently, it was recognized that the seminal vesicles are the primary source of prostaglandins found in seminal fluid¹². PGs are produced in many tissues where they function in vasodilation, inflammation, and smooth muscle contraction acting in autocrine or paracrine manners. PGs are synthesized from arachidonic acid that is converted to PGH₂ by cyclooxygenases (COX1 and COX2, also known as prostaglandin-endoperoxide synthase PTGS1 and PTGS2, respectively). PGH₂ serves as the precursor for the synthesis of all the other PGs. COX1 levels are usually static, resulting in baseline PG synthesis, whereas COX2 is an inducible enzyme that when expressed raises local PG concentrations.

PGs act on a subfamily of G protein-coupled membrane receptors (GPCRs) and a myriad of functions include mediating smooth muscle contraction. Recent work has suggested that besides inhibition of alpha-adrenergic receptors, BPH patients may benefit from inhibition of other non-adrenergic signaling pathways that also regulate smooth muscle tone¹³. Nonsteroidal anti-inflammatory drugs (NSAIDs) that inhibit COX1/COX2 have been shown to improve urinary symptoms in BPH patients^14,15. A clinical trial of patients with large prostate volumes (>40 cc) treated with a COX2 inhibitor and finasteride showed significant short-term improvement in LUTS at 4 weeks. These effects were transient and not significant by 24 weeks of therapy¹⁶. This effect was not seen in single treatment arms with either finasteride or COX2 inhibitor alone. The authors suggested that combination therapy induced a rapid, short-term improvement until the 5ARI finasteride substantially blocks DHT levels over the long term. However, this result is still remarkable since a COX2 inhibitor synergized with a 5ARI to provide an immediate response that neither drug could achieve in single agent therapy¹⁶. Urinary PG concentrations have been measured in men without prostate cancer and compared to prostate volume and LUTS severity (IPSS)¹⁷. A significant inverse correlation between PGE-M levels (the major urinary metabolite of PGE₂) and prostate volume was seen only in men undergoing treatment for BPH; no correlation with LUTS was detected¹⁷. Other than this prior study, thorough investigations of the role of PGs in BPH/LUTS have not be reported.

Since activation of GPCRs by PGs can cause contraction of smooth muscle, we investigated whether synthesis of PGs was elevated in BPH patients who failed medical treatment for LUTS. We compared transition zone (TZ) BPH tissue from patients requiring surgery for LUTS (Surgical-BPH: S-BPH) to control TZ tissue (Incidental-BPH: I-BPH) obtained from men undergoing radical prostatectomy for limited volume prostate cancer. We found that all the PGs, COX1, and COX2 levels were significantly elevated in S-BPH samples compared to I-BPH. Finally, in vitro PGE₂ and PGF_2α induced contraction of smooth muscle from urinary bladder muscularis consistent with a previous report¹⁸. However, contraction induced by PGs could not be reversed by a α-blocker. Taken together, these data indicate that PG signaling may explain why some patients fail current medical therapies for BPH. A better understanding of why patients fail medical therapy may have a major impact on developing new pharmaceutical approaches for the treatment of BPH/LUTS.

2. EXPERIMENTAL PROCEDURES

2.1. Human Prostate Tissue Collection

Two sources of biologic tissue were collected: I-BPH is from the transition zone (TZ) of men undergoing radical prostatectomy for low-volume, low-grade PCa confined to the peripheral zone. We selected samples in which the risk of malignancy is low [Gleason Score 6 or 7 (3+3 or 3+4)] and localized in the peripheral zone (PZ) of the prostate to minimize any cancer field effects on the TZ. Using radical prostatectomy patients to harvest control samples in any volume beyond a biopsy core is an accepted method to obtain TZ^19–22. S-BPH is benign prostatic tissue from patients who failed medical therapy and underwent surgical treatment for BPH, such as Holmium laser enucleation of the prostate (HoLEP) or open/simple prostatectomy. A pathologist confirms that the samples from the TZ contains only benign tissue with no cancer present. All samples are snap frozen. For analysis, approximately 1 gram of TZ is powdered under liquid nitrogen, thoroughly mixed, and used for tissue levels of PGs and qPCR. Formalin-fixed TZ tissue from the same patients were used for IHC.

2.2. Prostaglandins Measurements in Tissue

Prostaglandins were measured by the Eicosanoid Core Laboratory at Vanderbilt University Medical Center. Snap-frozen powdered tissue was used to measure prostaglandins in S-BPH (n=36) and I-BPH (n=18) patients. Each powdered tissue sample was extracted with 0.4mL of an ice-cold methanol solution containing the cyclooxygenase inhibitor indomethacin. The extracted sample was spun in a microcentrifuge at 10,000rpm for 2 min to pellet protein. 100μL of the tissue extract was diluted with 500μL deionized water containing the internal standard mix (1ng each deuterated PG [Cayman Chemical, Ann Arbor, MI USA]). The sample was vortexed and then purified on an Oasis MAX uElution plate (Waters Corp., Milford, MA) as follows: sample wells were first washed with methanol (200μL) followed by 25% methanol in water (200μ). Following washing, samples were loaded into individual wells and washed with 600μL 25% methanol. Eicosanoids were eluted from the plate with 30μL 2-propanol/acetonitrile (50/50, v/v) containing 5% formic acid into a 96-well elution plate containing 30μL water in each well.

Samples were analyzed on a Waters Xevo TQ-XS triple quadrupole mass spectrometer connected to a Waters Acquity I-Class UPLC (Waters Corp., Milford, MA USA). Separation of analytes was obtained using an Acquity PFP column (2.1 × 100mm) with mobile phase A being 0.01% formic acid in water and mobile phase B acetonitrile. Eicosanoids were separated using a gradient elution beginning with 30% B going to 95% B over 8 minutes at a flow rate of 0.25mL/min. Quantified levels of PGs are expressed as ng/mg of prostate tissue.

2.3. qPCR

If snap-frozen tissue was available for samples from which PGs were measured, they were also used in qPCR analysis [S-BPH (n=36) and I-BPH (n=15)]. Total RNAs from tissues were extracted using Trizol (Gibco-BRL), and residual genomic DNA was removed by DNase I (Invitrogen) treatment. The RNAs were reverse transcribed using random primers and Superscript II (Gibco-BRL) according to the manufacturer’s protocol. Real-time primer sets designed specifically to amplify each target genes were listed in S-Table 1. Real-time PCR reactions were carried out in a 20μl volume using a 96-well plate format and fluorescence was detected utilizing the Bio-Rad I-Cycler IQ Real-time detection system. Gene expression was normalized to β-actin by the 2^-ΔΔCt method²³. The values plotted represent the mean of at least three individual samples ± SD. Enzyme’s mRNA measured were Cyclooxygenase 1 [COX1 or prostaglandin-endoperoxide synthase (PTGS1)]; COX2 (PTGS2); Prostaglandin I2 Synthase (PTGIS); Prostaglandin D2 Synthase (PTGDS); Prostaglandin E Synthase (PTGES); Thromboxane A Synthase 1 (TBXAS1); and Aldo-Keto Reductase Family 1 Member C3 (AKR1C3).

2.4. Immunohistochemistry

Tissue was fixed in 10% buffered formalin for 24 h, processed, and embedded in paraffin. Immunohistochemistry (IHC) was performed at room temperature using manufacturer’s protocol. Primary antibodies used for IHC were COX1 (Abcam, ab109025, 1/500), and COX2 (Invitrogen, MA5–14568, 1/500 ThermoFisher Scientific).

2.5. In vitro Smooth Muscle Assay

Male C57/Bl6 mice and Sprague Dawley rats 12–18 weeks of age were euthanized and lower urinary tracts were rapidly collected. Tissue preparations were placed between platinum-coated electrodes in 15 ml tissue baths containing oxygenated modified Krebs solution (pH 7.4; NaCl, 137.7; NaH₂PO₄, 1.4; KCl 4.7; CaCl₂, 2.5; MgCl₂, 1; NaHCO₃, 16.3; and glucose, 7.8; all concentrations in mM) maintained at 37°. Tissue preparations were secured to the bath and to force displacement transducers (Grass FT-03, Grass Instruments, West Warwick, RI), and stretched gently with a tension of 0.5 to 1.5 mN, optimized for each tissue type. Tissues were allowed to equilibrate for 1 hour. PGE₂ and PGF_2α (0.01 to 100 μM) were tested in a concentration-dependent manner. Responses to electrical field stimulation (EFS) were generated using a Grass S88 stimulator (10 V, 0.5 ms duration, pulse train 5 s, at an increasing frequency of 1–50 Hz with 3 minutes interval). EFS-induced responses were nerve-mediated and may mimic some aspects of physiological conditions. At the end of experiment, tissues were exposed to 80 mM KCl Kreb’s solution.

2.6. Statistical Analysis

Where appropriate, experimental groups were compared using two-sample t-test, with significance defined as p <0.05. Analysis was performed using SigmaPlot software.

3. RESULTS

3.1. Study Population

The median age of patients with S-BPH was 67.5 years and for patients with I-BPH was 61.6 years (Table 1). At the time of surgery, S-BPH men were either on no medication (none: n=7), steroid 5α-reductase inhibitor (5ARI: n=7), alpha-adrenergic receptor antagonist (α-blockers: n=13), or combination α-blockers plus 5ARI (n=9). At the time of surgery, I-BPH men were either on no medical therapy (n=8) or on α-blockers (n=7). Some of the patients in the I-BPH group were treated for LUTS with α-blockers but treatment with 5ARI or the combination of both α-blockers plus 5ARI was not medically warranted and would be unethical to place these individuals on unnecessary treatment for the purpose of this study. Therefore, the I-BPH population is limited to patients on no medical treatment or only α-blockers. The median IPSS for S-BPH was 18, for I-BPH on α-blockers was 12.5, and I-BPH on no medical therapy (None) was 17.

Table1:

Study Population Characteristics. Surgical BPH (S-BPH) was defined as benign prostatic tissue collected from the transition zone (TZ) of patients who failed medical therapy and underwent surgical intervention to relieve LUTS. Control tissue was termed Incidental BPH (I-BPH). I-BPH was TZ obtained from men undergoing radical prostatectomy for low-volume, low-grade prostatic adenocarcinoma (PCa, Gleason score ≤7) confined to the peripheral zone. Medical treatment at the time of surgery was None; α-blocker (alpha-adrenergic receptor antagonist); 5ARI (5α-reductase inhibitor); or α-blocker + 5ARI. I-BPH population is limited to patients on no medical treatment or only α-blockers since other medical treatment was unwarranted and therefore unethical to use for the purpose of this study. N/A not applicable.

Patient type	S-BPH	I-BPH	p-value
Number of Patients	36	15	N/A
Median Age	67 (interquartile range = 10)	59 (interquartile range = 8)	0.02*
Medical treatment for BPH/LUTS
None	7 (19%)	8 (53%)	0.009*
α-blocker	13 (36%)	7 (47%)
5ARI	7 (19%)	0
α-blocker + 5ARI	9 (25%)	0
IPSS
0–7	Median = 3 (n=3)	5 (n=1)	0.5
8–19	Median = 13 (n=18)	Median 14 (n=6)
>20	Median = 28.5 (n=14)	Median 26.5 (n=2)
Number of Patients stratified by Prostate Volume
<40g	n=1 (35 g)	n=1 (36 g)	0.2
40–59g	n=6 (mean 48 g)	n=6 (mean 52 g)
>60g	n=25 (mean 134 g)	n=8 (mean 84 g)
Gleason Score	N/A	6 or 7 (3+3, n=12; 3+4, n=3)	N/A

^*p <0.05.

3.2. Prostaglandin levels are significantly higher in S-BPH compared to I-BPH tissue

Prostaglandins were measured in snap-frozen prostatic tissue from S-BPH (n=36) and I-BPH (n=15) patients. Since the precursor of all PGs (PGH₂) is very unstable, it was not measured. Statistical analysis demonstrated that tissue PGF_2α, PGI₂, PGE₂, PGD₂, and TxA₂ (metabolite TxB₂ is measured) were significantly (p=0.001) elevated in S-BPH compared to I-BPH tissue (Fig 1A). Patients were then grouped by medical treatment and PGs levels were compared within S-BPH or within I-BPH groups (Fig 1B–F). Some PG levels significantly varied over different S-BPH treatment groups. For example, PGF_2α was significantly higher in the 5ARI group (Fig 1C: p=0.033) compared to the control group (no medical therapy) and PGE₂ was significantly higher in the 5ARI group compared to no treatment (Fig 1B: p=0.026). There was no significant difference within I-BPH treatment subgroups, but the trend for some PGs (PGE₂, PGI₂, and PGF_2α) was to be higher in I-BPH patients taking α-blockers. PG levels in men with I-BPH being treated successfully for LUTS with α-blockers were still 8- to 10-fold lower than in patients in the S-BPH group who failed medical therapy.

Figure 1:

Prostaglandins levels in S-BPH and I-BPH. Prostaglandins were measured in snap-frozen prostatic tissue and values were expressed as ng/mg of tissue. A) Total levels of PG from S-BPH (n=36) and I-BPH (n=15) patients regardless of treatment; Levels of PG in S-BPH and I-BPH patients based upon treatment: B) PGE₂ C) PGI₂ D) PGF2α E) PGD2 and F) TxB2. Treatment groups for S-BPH patients were None (no medical therapy: n=7); α-B (α-blockers: n=13); 5ARI (5-α-reductase inhibitor: n=7) and α-B + 5ARI (n=9). Treatment groups for I-BPH patients were None (n=8) or α-B (n=7). Error bars (SD) and P values comparing groups are added over the bar graphs.

3.3. The prostaglandin biosynthesis pathway is elevated in S-BPH compared to I-BPH tissue

By qPCR, mRNA encoding COX1 (PTGS1) and COX2 (PTGS2), enzymes involved in PG synthesis, were significantly (p<0.001 and 0.002, respectively) elevated in S-BPH (n=36) compared to I-BPH (n=15) (Fig 2A). Within the I-BPH group, we compared COX1 and COX2 in men on no treatment for BPH/LUTS to those being treated with α-blockers. There were no significant changes in COX1 and COX2 mRNA levels with or without α-blockers (Fig 2B). Likewise, within the S-BPH group, we compared COX mRNA levels among the different treatment subgroups (Fig 2C). Only COX2 levels were significantly (p=0.003) lower in patients on α-blockers compared to no treatment (Fig 2C). Elevated levels of COX1 and/or COX2 will increase PGH₂ synthesis, a commitment step toward PG synthesis dependent upon expression of downstream enzymes. Accordingly, we measured the mRNA of enzymes that convert precursor PGH₂ to its final products. The differences between S-BPH and I-BPH varied from significant for enzymes such as PTGIS (p<0.006) for PGI₂, PTGDS (p<0.02) for PGD₂, near significant for PTGES (p<0.06) for PGE_2, and non-significant for AKR1C3 (p<0.35) and TBxAS1 (P<0.265) for TxA₂ (Fig 3). Within S-BPH or I-BPH groups, there were no significant differences among the different medical therapy subgroups (Supplementary S-Fig 1). Thus, significant increases in levels for the COX1 and COX2 enzymes (Fig 2A) result in a commitment to synthesize PGs and correspond to the increased levels of PGs seen in S-BPH versus I-BPH tissue.

Figure 2:

COX1 and COX2 mRNA levels in S-BPH and I-BPH. COX1 and COX2 mRNA levels were measured by qPCR and showed A) that they were significantly (p<0.001 and 0.002, respectively) elevated in S-BPH (n=36) compared to I-BPH (n=15); B) that neither showed a significant difference in I-BPH between None (no medical therapy) and α-B (α-blockers); C) the S-BPH COX2 mRNA levels were significantly (p=0.003) lower between patients on α-blockers compared to None. Treatment groups for S-BPH patients were None (no medical therapy: n=7); α-B (α-blockers: n=13); 5ARI (5-α-reductase inhibitor: n=7) and α-B + 5ARI (n=9). Treatment groups for I-BPH patients were None (n=8) or α-B (n=7). Error bars (SD) and P values comparing groups are added over the bar graphs.

Figure 3:

Prostaglandin biosynthesis pathway. Elevated levels of COX1 and/or COX2 will increase PGH₂ synthesis, a commitment step toward PG synthesis dependent upon the expression of the appropriate downstream enzymes. Thus, we measured the mRNA of enzymes that convert precursor PGH₂ to its final products. The downstream enzyme mRNA for PTGIS and PTGDS were significantly elevated in S-BPH compared to I-BPH but PTGES, AKR1C3, and TBxAS1 did not show significant difference between these two patient populations. S-BPH (n=36) and I-BPH (n=15) patients. Error bars (SD) and P values comparing groups are added over the bar graphs.

3.4. COX1 and COX2 cellular localization in S-BPH and I-BPH tissue

By immunohistochemistry (IHC), we determined that COX1 was expressed in stroma and occasionally with the basal myoepithelial layer of prostate glands. Very rarely, COX1 was observed in luminal cells, but when present in luminal cells it was often associated with focal acute and/or chronic inflammatory infiltrates. In contrast, COX2 was expressed in basal and luminal epithelial cells, but not stroma (Fig 4) Similar to results with COX1, COX2 staining was stronger in areas where inflammatory cells were adjacent to epithelium compared to areas without marked inflammatory infiltrates. COX2 staining of epithelium in S-BPH tissue was always higher than in I-BPH tissue. IHC comparison of S-BPH to I-BPH confirmed that COX1 and COX2 immunoreactivity appeared higher in S-BPH compared to I-BPH tissue, regardless of treatment group. This result is consistent with the elevated mRNA levels for these enzymes seen in S-BPH patients.

Figure 4:

COX1/2 expression in benign prostate hyperplasia. A-D) Hematoxylin and eosin (H & E) stains. E-H) COX1 expression. I-L) COX2 expression. C-D) arrows indicate areas of inflammation in S-BPH and I-BPH. E) COX1 is expressed predominantly in the stroma and to a lesser extent in the epithelium of S-BPH. F) COX1 expression is limited and low in the stroma of I-BPH. G, H) when inflammation occurs (arrows) adjacent epithelium expresses higher COX1 in both S-BPH and I-BPH. I) COX2 is expressed in the epithelium of S-BPH. F) COX2 expression is low in I-BPH. K-L) when inflammation occurs (arrows) adjacent epithelium expresses higher COX2 in both S-BPH and I-BPH.

3.5. In vitro smooth muscle assay for prostaglandin activity

Smooth muscle in the prostate and bladder respond similarly to known stimuli. Previous reports show that PGs cause contraction of smooth muscle in the bladder¹⁸, but possible crosstalk between the inhibition of alpha-adrenergic receptors by α-blockers and PG stimulation of G-protein-coupled receptors has not been reported. We tested the ability of PGE₂ and PGF_2α to cause contraction of smooth muscle in the bladder wall in vitro. PGE₂ (Fig 5A) and PGF_2α (Fig 5B) induced contractions of the bladder body in mice in a concentration-dependent manner. PGE₂ and PGF_2α-induced responses were unaffected by prazosin, a α-blocker (Fig 6A & 6B respectively). Prazosin may have failed to block PG-induced contraction since low levels of expression of alpha-adrenergic receptors have been reported in the murine bladder body wall²⁴. Electrical field stimulation (EFS) is a nerve-mediated reaction that mimics some aspects of the physiological responses of the bladder wall to nerve stimulation. As expected, prazosin also failed to affect EFS in the mouse bladder body (Fig 6C). Contractile responses of rat bladder to PGE₂ (S-Fig 2A) and PGF_2α (S-Fig 2B) were less than those of mouse bladder but did induce contraction in a concentration-dependent manner. Like in the mouse bladder body, prazosin did not affect EFS-induced contraction of rat bladder body (data not shown). Neither PGE₂ nor PGF_2α induced contraction of the bladder neck (trigone), urethra, or prostate in rats (data not shown). Phenylephrine (a α-adrenergic receptor agonist) induced contractions of rodent bladder neck (S-Fig 3A), urethra (S-Fig 3B), and prostate (S-Fig 3C) but not the bladder wall (data not shown). Prazosin did not block EFS-induced response in the bladder neck (S-Fig 4A) but did attenuate responses seen in the urethra and prostate (S-Fig 4B and 4C). These data indicate that PG receptors may not play a role in the rodent bladder neck, urethra, and prostate, and that alpha-adrenergic receptors do not appear to be important physiologic regulators of urinary bladder contractile function. Prazosin did not moderate PGE₂, PGF_2α, or EFS-induced contraction of bladder body, suggesting that both alpha-adrenergic receptors and GPCRs do not crosstalk or that alpha-adrenergic receptors levels are too low in the rodent bladder wall to mediate a response.

Figure 5:

Mouse smooth muscle in the bladder contracts in respond to PG stimuli. Representative tracings reveal that PGE₂ (A) and PGF_2α (B) induced contractile responses of the mouse bladder body in a concentration-dependent manner (0.01 −100 μM). Arrows indicate applications of PGE₂ (A) or PGF_2α (B).

Figure 6:

Prazosin, a α-blocker, does not inhibit PG induced contractions. Prazosin at 1 and 10 μM did not affect A) PGE₂ or B) PGF_2α induced contractile responses of the mouse bladder body. C) Prazosin (10 μM) did not affect electrical field stimulation (EFS) induced responses in the mouse bladder body. Tension was measured in grams (g) of force. Hertz (Hz).

DISCUSSION

We explored prostaglandin synthesis by the prostate as a possible explanation for the failure of α-blockers and/or 5ARIs to reduce LUTS in men with BPH. Prostaglandins induce smooth muscle contraction via GPCRs and therefore may be a potential mechanism whereby patients fail current medical therapy for BPH/LUTS. The MTOPS study demonstrated that there was only a 34% reduction in symptomatic progression of LUTS in BPH patients on α-blockade or 5ARI monotherapy, but combination therapy reduced symptomatic progression by 66%. A reduced risk of progression was also found for patients on combination therapy in the more recent ComBAT study.² There may be additional mechanisms that account for treatment failure, especially in medication non-responders. Although we cannot prove the mechanism of insensitivity to medical therapy, our data are consistent with the hypothesis that PG signaling may contribute to treatment failure.

All of the patients we tested who failed medical therapy and required surgery for LUTS showed significantly elevated levels of each PG we assayed in prostatic tissue. In addition, increased mRNA expression and immunoreactivity of COX1 (PTGS1) and COX2 (PTGS2), gatekeepers for PG synthesis, were observed in men who failed medical therapy for BPH. PGs can affect bladder tone and induce contractions of smooth muscle in vitro²⁵. By in vitro testing, we show that α-adrenergic blockade with prazosin fails to inhibit PGE₂ and PGF_2α-induced contractions of bladder wall in mice and rats. Electrical field stimulation (EFS) is a nerve-mediated reaction that may mimic some aspects of physiological responses to alpha-adrenergic receptors. However, prazosin also failed to affect EFS in the rodent bladder neck, while it did block EFS contractions in the urethra and prostate. Phenylephrine, a α-adrenergic agonist, induced contractions in bladder neck, urethra, and prostate, but not in the bladder wall. These data indicate that prazosin is unlikely to moderate PGE₂ and PGF_2α signaling effects on contractile functions of bladder wall smooth muscle. A previous study demonstrated that in human, rat, and macaque bladders, PGE₂ and PGF_2α induced contractions, but there were species differences in the degree of the response, with the rat bladder showing the lowest response¹⁸. The lack of crosstalk between α-blockers on alpha-adrenergic receptors and GPCRs may be a consequence of distinct signal transduction pathways and/or non-overlapping distribution of receptors, with low prevalence of alpha-adrenergic receptors in the bladder wall. Regardless, these results still suggest a mechanism (PG-induced signaling and smooth muscle contraction) of resistance to α-blockers in the bladder wall. Patients who fail 5ARI treatment show the same significant increase in PG pathway signaling, suggesting a common mechanism for failure of either α-blockers or 5ARI therapy. In men responding to α-blockers (but undergoing prostatectomy for PCa), levels of three prostaglandins (PGF_2α, PGI₂, and PGE₂) in the TZ tended to be higher (but not significantly so) compared to TZ tissue from patients not undergoing treatment for BPH/LUTS. These results suggest that current medical therapy induces or selects for regional PG overexpression in compensation for locally reduced smooth muscle tone. However, it should be noted that PG levels in I-BPH patients treated with α-blockers are still around 8- to 10-fold lower than levels seen in S-BPH patients who failed medical therapy. It is also possible that PG overexpression is driven in BPH through ischemia. Ischemia in animal models causes prostate hyperplasia,^26,27 and higher vascular resistive indices have been observed in humans with BPH compared to controls.²⁸ Ischemia including hypoxia-inducing factors are known to stimulate the prostaglandin biosynthesis pathway,²⁹ providing another possible link between BPH pathophysiology and PGs.

By IHC examination of prostatic tissue, COX1 and COX2 showed stronger staining in S-BPH patients compared to I-BPH patients, regardless of treatment group. Although it has been reported that COX1 expression is limited to stromal elements in prostate^30,31, we also noted its expression in the basal myoepithelial layer of a subpopulation of prostate glands. Moreover, in rare cases, COX1 immunoreactivity was observed in luminal cells in areas of acute and/or chronic prostatitis. COX2 was focally expressed in basal and luminal epithelial cells, and appeared upregulated in glands adjacent to inflammatory infiltrates. COX1 and COX2 both generate PGH₂ from arachidonic acid, the precursor substrate for all subsequent prostaglandin synthase enzymes, and are the critical regulatory enzymes for PG signaling. Although both enzymes control PG synthesis, the inducible nature of COX2 makes this enzyme a critical regulator of PG synthesis.

Three of the PG synthase enzymes we assayed (PTGIS [PGI₂]; PTGDS [PGD₂]; and PTGES [PGE₂]) were greater in S-BPH tissue compared to I-BPH tissues; two others showed no significant difference (AKR1C3 [PGF_2α] and TBxAS₁ [TxA₂]). These data are consistent with the high levels of PGs seen in all S-BPH samples. PGE₂ can cause contraction or relaxation of smooth muscle and can be pro- or anti-inflammatory depending upon which of its four receptors are present. An important physiologic role for PGF_2α is smooth muscle contraction to induce labor in gravid uteri; it appears to have a limited role, if any, in inflammatory responses. PGD₂ can relax or contract smooth muscle and function as a vasodilator, while PGI₂ is primarily a vasodilator but can induce contraction of the bladder body smooth muscle²⁵. TxA₂ can also cause contraction of smooth muscle and stimulate smooth muscle proliferation (see review³²).

Although chronic inflammation is associated with the severity and progression of BPH/LUTS³³, its pathogenetic role, if any, is unresolved. Since prostate volume is associated with chronic inflammation³³, a potential growth-promoting effect of pro-inflammatory factors such as cytokines and prostaglandins has been suggested^31,34,35. Areas of inflammation are also associated with increased COX2 within luminal epithelial cells³¹, consistent with our data. Together, results from prior studies and our current data suggest that inhibition of COX2 by nonsteroidal anti-inflammatory drugs (NSAIDs) may be of benefit for BPH/LUTS. A meta-analysis of three randomized, placebo-controlled clinical trials showed that NSAIDs improved LUTS in BPH patients^14,15. Another separate clinical trial that specifically selected patients with large prostate volumes (>40 cc) combined a COX2 inhibitor with finasteride and showed significant improvement in LUTS over short term treatment (4 weeks), but not long term (24 weeks), when compared to COX2 inhibitor or finasteride alone¹⁶. These investigators suggested that combination therapy induced a rapid, short-term improvement until the finasteride significantly block DHT over the long term. Although COX2 inhibition did not benefit patients in the long term, this result is still remarkable since a COX2 inhibitor synergized with a 5ARI to provide an immediate response that neither drug achieved in monotherapy treatment arms. In the MTOPS study, each patient had at least one baseline prostate biopsy specimen and additional tissue was biopsied for up to 4.5 years¹. In a follow-up analysis from baseline to time of progression, MTOPS samples revealed that men with acute or chronic inflammation at baseline showed a significant increase in prostate size and risk of BPH progression compared to men without inflammation³⁶. In addition, this report found that use of anti-inflammatory medications (NSAID and/or systemic steroids) were associated with an increased risk of progression³⁶. However, NSAID use was not separated from steroid treatment in this analysis, and inhibition of COX enzymes per se on progression of BPH cannot be evaluated with this data. Overall, these studies indicate that inflammatory pathways may play a critical role in prostate hypertrophy and the severity and progression of LUTS. Our data show that elevation of COX1 and COX2 and increased PG synthesis is observed in patients who fail medical therapy with α-blockers, 5ARI, or a combination of these drugs. Whether this increase in the synthesis of PGs is a cause or effect of failure of medical therapy remains to be determined. However, inhibition of COX2 shows promise in improving LUTS^15,16, suggesting that targeting this pathway may be of clinical benefit. Furthermore, blocking the PGE₂ EP4 receptor with an EP4 antagonist improved bladder over activity associated with prostatic inflammation in the rat³⁷. These studies suggest that there may be therapeutic potential in blocking PG synthesis in patients with BPH/LUTS. Delineating the interconnection between inflammatory signaling and the failure of current medical therapy for BPH/LUTS will define new pathways for drug targeting. Testing of drugs to block specific PGs maybe useful in the treatment of BPH/LUTS patients that fail medical therapy.

Supplementary Material

tS1

S-Table 1: Human qPCR Primer Sets. Cyclooxygenase 1 [COX1 or prostaglandin-endoperoxide synthase (PTGS1)]; COX2 (PTGS2); Prostaglandin I2 Synthase (PTGIS); Prostaglandin D2 Synthase (PTGDS); Prostaglandin E Synthase (PTGES); Thromboxane A Synthase 1 (TBXAS1); and Aldo-Keto Reductase Family 1 Member C3 (AKR1C3)

fS1-4

S-Figure 1: No significant differences in PG levels within the different medical therapy subgroups. Treatment groups for I-BPH patients were None (n=8) or α-B (n=7). Treatment groups for S-BPH patients were None (no medical therapy: n=7); α-B (α-blockers: n=13); 5ARI (5-α-reductase inhibitor: n=7) and α-B + 5ARI (n=9). Error bars (SD) and P values comparing groups are added over the bar graphs.

S-Figure 2: Rat smooth muscle in the bladder contracts in respond to PG stimuli. Representative tracings reveal that PGE₂ (A) and PGF_2α (B) induced contractile responses of rat bladder body in a concentration dependent manner (0.01–100 μM). Arrows indicate applications of PGE₂ (A) and PGF_2α (B). Tension was measured in grams (g) of force. Hertz (Hz).

S-Figure 3: Phenylephrine induces rat smooth muscle contractions. Representative tracings reveal that phenylephrine, an α-adrenergic receptor agonist, induced contractile responses of rat bladder neck (A), urethra (B), and ventral prostate (C) in a in a concentration dependent manner (0.01–100 μM). Arrows indicate applications of phenylephrine. Tension was measured in grams (g) of force.

S-Figure 4: Electrical field stimulation (EFS) induced response in rat are altered by prazosin. Prazosin, a α-blocker, did not affect EFS stimulation of rat bladder neck (A) but did attenuate the response of rat urethra (B) and ventral prostate (C). Tension was measured in grams (g) of force. Hertz (Hz).