Modulation of platelet cyclic nucleotide content by PGE1 and the prostaglandin endoperoxide PGG2.

The effects of prostaglandin E1 and prostaglandin G2, the prostaglandin endoperoxide, on platelet cyclic nucleotide concentrations were measured in platelet rich plasma (PRP), and in washed intact platelets. PGE1 was found to be a potent stimulator of platelet cAMP levels in both PRP and washed cells, and to inhibit aggregation in both systems. PGE1 did not change platelet cGMP levels in either PRP or washed cells. PGG2 which is a potent inducer of platelet aggregation, did not affect either the basal cAMP or the basal cGMP concentration. However, PGG2 was found to antagonize the increases in cAMP content in response to PGE1 in both PRP and washed platelets. The addition to our system of a cyclic nucleotide phosphodiesterase inhbitor, theophylline, did not change our findings. It is suggested that PGG2 may induce platelet aggregation by inhibiting PGE1-stimulated cAMP accumulation.     

Pertussis toxin reverses prostaglandin E2- and somatostatin-induced inhibition of rat parietal cell H(+)-production.

In enzymatically dispersed enriched rat parietal cells we studied the effect of pertussis toxin on prostaglandin E2 (PGE2)- or somatostatin-induced inhibition of H(+)-production. Parietal cells were incubated in parallel in the absence (control cells) and presence of pertussis toxin (250 ng/ml; 4 h). [14C]Aminopyrine accumulation by both pertussis toxin-treated and control cells was used as an indirect measure of H(+)-production after stimulation with either histamine, forskolin or dibutyryl adenosine 3',5'-cyclic monophosphate (dbcAMP) alone and in the presence of PGE2 (10(-9)-10(-7) M) or somatostatin (10(-9)-10(-6) M). PGE2 inhibited histamine- and forskolin-stimulated [14C]aminopyrine accumulation but failed to alter the response to dbcAMP. Somatostatin was less effective and less potent than PGE2 in inhibiting stimulation by histamine or forskolin and reduced the response to dbcAMP. Pertussis toxin completely reversed inhibition by both PGE2 and somatostatin on histamine- and forskolin-stimulated H(+)-production but failed to affect inhibition by somatostatin of the response to dbcAMP. After incubation of crude control cell membranes with [32P]NAD+, pertussis toxin catalysed the incorporation of [32P]adenosine diphosphate (ADP)-ribose into a membrane protein of molecular weight of 41,000, the known molecular weight of the inhibitory subunit of adenylate cyclase (Gi alpha). Pertussis toxin treatment of parietal cells prior to the preparation of crude membranes almost completely prevented subsequent pertussis toxin-catalysed [32P]ADP ribosylation of the 41,000 molecular weight protein.(ABSTRACT TRUNCATED AT 250 WORDS)     

The effect of pyridoxal-5-phosphate on the inhibition of platelet aggregation and adenosine-3'-5'-cyclic monophosphate accumulation in human platelets

Platelet adenosine-3'-5'-cyclic monophosphate (c-AMP) was determined in platelet rich plasma and washed platelets by a modification of the c-AMP protein binding assay. Studies were performed in the presence and absence of platelet aggregation inducers (adenosine diphosphate and thrombin) and inhibitors (pyridoxal-5-phosphate (PALP), prostaglandin E1 (PGE1), theophylline and papaverine). At 70% inhibitory concentration of platelet aggregation (PA) induced by papaverine or theophylline, a small but significant increase in platelet c-AMP level was found. With PALP however the inhibition of PA was not associated with a significant increase in the c-AMP level. PALP or theophylline potentiated greatly both the inhibition of PA and c-AMP accumulation induced by PGE1. Yet PALP potentiated the inhibition of PA induced by theophylline without increasing platelet c-AMP level. Our results indicate that the inhibition of PA by PALP is not mediated by c-AMP accumulation. However in the presence of PGE1, the increment in PA inhibition, is mediated by further indirect activation of the adenyl-cyclase.     

Cooperative interactions of protein kinase C and cAMP-dependent protein kinase systems in human promyelocytic leukemia HL60 cells

Interactions of protein kinase C (PKC) and cAMP-dependent protein kinase (PKA) systems were investigated in HL60 cells. It was found that the differentiating effects of 12-O-tetradecanoylphorbol-13-acetate (TPA) were potentiated by dibutyryl cAMP (dbcAMP) or prostaglandin E2 (PGE2). In addition, dbcAMP or PGE2 inhibited TPA-induced binding of PKC to plasma membrane, leading to decreased protein phosphorylation, and promoted subsequent redistribution of enzyme to the nuclear membrane region. The findings are consistent with the hypothesis that PKC and PKA systems regulate cooperatively the phenotypical differentiation of leukemic cells.     

PGE2 and dibutyryl-cAMP enhance the response of rat granulosa cells to FSH

Granulosa cells isolated from immature Sprague-Dawley rat ovaries produce progesterone (31.7 pg/micrograms cell protein) in response to an acute FSH stimulus (5 micrograms/ml NIH-FSH-S11, 2 H). After culture for 48 h in the absence of hormones (control culture), progesterone production by the granulosa cells in response to FSH is significantly reduced (2.9 pg/micrograms cell protein). Cells cultured with prostaglandin E2 (PGE2, 1 microgram/ml) or dibutyryl-cAMP (dbcAMP, 1 mM) exhibited a discernibly greater steroidogenic response to FSH (12.5 and 53.4 pg/microgram cell protein, respectively) than that of control cultures. Therefore the presence of PGE2 or dbcAMP in the culture medium helps to maintain the steroidogenic capacity of granulosa cells in culture. It is probable that this capacity is maintained at a locus distal to the production of cAMP by FSH. Paradoxically, granulosa cells cultured with PGE2 produce less cAMP in response to FSH stimulation than cells in control cultures (15.9 vs. 250.3 fm/micrograms cell protein). This may be due to a suppressive effect of prior exposure to PGE2 on the subsequent activity of adenylate cyclase when the FSH is introduced and a concomitant elevation of phosphodiesterase activity.     

Effect of gonadotropins and prostaglandin on cumulus mucification in cultures of intact follicles

Follicles of 28 day-old pregnant mare serum gonadotropin (PMSG)-primed rats, were cultured for up to 24 hours in the presence or absence of ovine gonadotropins, highly purified rat gonadotropins, dibutyryl cyclic AMP (dbcAMP), methylisobutylxanthine (MIX), choleratoxin (CT), or prostaglandin E2 (PGE2). The morphology of the cumulus-oocyte complexes isolated from these follicles was subsequently examined with the light microscope. Cumulus mucification was studied under the different culture conditions using scanning electron microscopy (SEM) and the hyaluronidase sensitivity test. The features of the cumulus-oocyte complexes in the control cultures did not change throughout the incubation period, while complexes from follicles incubated with LH, FSH, dbcAMP, MIX, CT, or PGE2 changed their appearance and accumulated extracellular mucoid material. Treatment of these cumuli with hyaluronidase resulted in lysis of the extracellular mucus and dispersal of the cumulus masses. The results of this study agree with our earlier observation that the maturation of the cumulus-oophorus, which occurs in vivo following the LH surge, can be induced in vitro by either gonadotropins or cAMP. Prostaglandin E2 did not affect cumulus cells, unless incubated enclosed by their follicles. This suggests that this hormone may influence the cumulus cells indirectly, probably via other components.     

Effect of cyclic AMP on human blood platelets: induction of pseudopod formation and protein phosphorylation

A variety of techniques, including immunofluorescence, electron microscopy and biochemical analysis, were used to examine shape changes and cytoskeletal reorganization of human blood platelets during treatment with N6,O2-dibutyryl adenosine 3',5'-cyclic monophosphoric acid (dbcAMP), and agent known to elevate the intracellular level of cyclic AMP (cAMP). Cytochemical analysis shows that the unstimulated platelets have a discoid shape with no obvious membrane projections. Platelets treated with dbcAMP produce pseudopod-like structures containing cytoskeletal proteins such as actin and microtubules. Biochemical analysis reveals that a 125,000 dalton phosphoprotein (P-125) is preferentially recruited into cytoskeletal fractions of platelets treated with dbcAMP. This protein, which is one of the substrates for cAMP-dependent kinase(s) and/or is closely associated with the cytoskeleton, may play an important role in regulating the shape changes and cytoskeletal reorganization that occur during the early stages of platelet activation.     

Interleukin-10 and cAMP-elevating agents cooperate to induce suppressor of cytokine signaling-3 via a protein kinase A-independent signal

We have recently shown that IL-10 represents an efficient stimulus for suppressor of cytokine signalling (SOCS)-3 mRNA expression in human neutrophils and PBMC. Herein, we identify cAMP-elevating agents such as prostaglandin E2 (PGE2), PGE1, forskolin, dibutyryl cAMP (dbcAMP) and cholera toxin as a novel class of agonists able to induce SOCS-3 mRNA and protein expression in human leukocytes, cooperating with interleukin-10 (IL-10) in such activities. While PGE2 or dbcAMP prolonged the stability of SOCS-3 mRNA isolated from IL-10-treated leukocytes, inhibitors of cAMP-dependent protein kinase A (H89, KT5720, and St-Ht31 peptide) did not influence the action of PGE2/dbcAMP and/or IL-10 on SOCS-3 mRNA and protein expression, implying that their effect are mediated through a PKA-independent pathway. Taken together, our data identify cAMP-elevating substances as a novel class of agonists able to trigger SOCS-3 expression, and suggest that SOCS-3 might be involved in the regulatory effects of cAMP-elevating substances.     

Suppression of LH-stimulated prostaglandin and progesterone accumulation in rat granulosa cells by isoquinolinesulfonamide protein kinase inhibitors

Rat granulosa cells were incubated with isoquinolinesulfonamide inhibitors of protein kinases A and C and/or LH, dibutyryl cAMP (dbcAMP), tetradecanoylphorbol acetate (TPA), cholera toxin, or forskolin for 5 h. H7 (25 microM) was observed to inhibit LH, cholera toxin or dbcAMP stimulation of prostaglandin (PGE), and progesterone accumulation. H7 produced inhibition when added as little as 2 min before and as long as 1 h after LH. HA1004 was ineffective against LH or cholera toxin stimulation of PGE or progesterone at up to 100 microM. H9 blocked some LH and forskolin responses at 25 microM, but required a 50 microM concentration to minimally affect TPA stimulation. Cytotoxicity was not observed at the concentrations and times of isoquinolinesulfonamides tested. H7 and H9, therefore, suppress LH stimulation of granulosa cell functions in a dose- and time-dependent manner consistent with inhibition of protein kinases A and/or C, and consonant with a requirement for such kinases in LH action.     

Differentiation of HL-60 myeloid leukaemia cells is associated with a transient block in the G2 phase of the cell cycle

The human promyelocytic leukaemia cell line HL-60 can be induced to differentiate towards mature granulocytes by treatment with dibutyryl cyclic adenosine-3',5'-monophosphate (dbcAMP). Differentiation begins within 16-24 h of treatment and is associated with a time- and dose-dependent accumulation of cells in the G0/G1 phase of the cell cycle with a concomitant decrease in the number of cells in the S and G2 + M phases. Using acridine orange staining, we found that the RNA content of the cells also decreased following differentiation. Stathmokinetic analysis of HL-60 cell populations following dbcAMP treatment showed no effect on the total number of cells in the G0/G1 or S phases, or the rate of progression of cells through these cell cycle compartments. In contrast, dbcAMP was found to induce a transient arrest of the cells in the G2 phase. We also found that differentiation induced by dbcAMP did not require progression of the cells through the cell cycle. Cells arrested in either G1/S by hydroxyurea or G2 + M by colcemid eventually expressed markers of mature granulocytes. These results demonstrate that dbcAMP modulates cell cycle progression. However, these cell cycle changes alone are insufficient to induce granulocytic differentiation of HL-60 cells.     

Obligatory action of polypeptide growth factors for the IL-1-mediated prostaglandin E2 production in fibroblasts. Potential role of growth factors in modulation of tissue response to IL-1

Our studies show that in connective tissue cells, induction of PGE2 synthesis in response to IL-1 requires costimulation with platelet-derived growth factor (PDGF) or fibroblast growth factor (FGF). In cells incubated in medium containing fresh serum, IL-1 induced a dose-dependent synthesis of PGE2. However, when the cells were incubated in medium containing low serum or platelet poor plasma (lacking PDGF), IL-1 alone failed to induce PGE2 synthesis. PGE2 synthesis was restored when platelet poor plasma was supplemented with PDGF. Addition of PDGF or FGF together with IL-1 resulted in a 14- and 66-fold stimulation of PGE2 synthesis, respectively. Stimulation was dependent on the concentration of both IL-1 and the growth factor. PGE2 synthesis was also dependent on the synthesis of new proteins. In cells simultaneously treated with IL-1 and PDGF, PGE2 synthesis was initiated after a lag of 2 to 3 h, proceeded first with a rapid rate for 6 h, and then with a slower rate through 24 h. PGE2 synthesis during the latter, slower phase was greatly enhanced by pretreatment with PDGF, but not by pretreatment with IL-1. PDGF pretreatment also resulted in maintenance of 10- to 12-fold higher cell surface IL-1-binding during this phase. These data provide evidence for potentially novel interactions between PDGF and IL-1 activities, one of which is the modulation of IL-1 receptors by PDGF. Furthermore, these studies suggest that by virtue of their effect on IL-1 activities, PDGF and FGF may play additional roles in connective tissues, including an indirect role in inflammatory processes.     

Interaction of receptors for prostaglandin E1/prostacyclin and insulin in human erythrocytes and platelets.

Prostaglandin E1/I2 and insulin receptors of human erythrocyte and platelet are capable of modulating each other's activity. This modulation of the receptor activity and number in one system by a second receptor system in human platelet and erythrocyte seems to be beneficial. Insulin increases the PGE1 binding to platelets and thereby enhances the platelet antiaggregatory action of prostaglandin by increasing cyclic AMP levels. Similarly, PGE1 increases insulin binding to human erythrocyte, and thereby reduces the optimum concentration of insulin for a maximal reduction in membrane microviscosity. During ischemia the reduced response of platelets to the inhibitory effect of PGE1 or PGI2 relates to the impaired PGE1/I2 receptor activity. Treatment of these platelets with insulin at physiological concentrations can normalise the PGE1/I2 receptor activity. This review focuses on the relationship between the two receptor systems in human blood cells.     

Evidence for a role of myosin phosphorylation in the initiation of the platelet shape change response

When platelets were stimulated with ADP to cause shape change without aggregation or secretion, myosin 20,000-Da light chain phosphorylation was rapid and appeared to precede slightly the shape change response. While the shape of the platelets remained spheroidal, myosin phosphorylation was transient and after 2-5 min returned to the same level as that of unstimulated cells. Phosphorylation of the 47,000-Da platelet protein was minimal under these conditions. The phosphorylation time course was not altered by the addition of indomethacin or allowing the cells to aggregate. The dose-response curve of myosin phosphorylation very closely paralleled that of shape change with a midpoint at 0.7 microM ADP. ATP, a competitive antagonist of ADP, inhibited both shape change and myosin phosphorylation with the same concentration of ATP causing 50% inhibition of each response. Similarly, when platelets were stimulated with either 15-hydroxy-9,11-azo-prostadienoic acid or collagen, myosin phosphorylation slightly preceded shape change. These results suggest that myosin phosphorylation is required for the initial change in platelet shape but is not necessary for maintenance of the spherical shape.     

Direct evidence for the interaction between papaverine and alpha 2-adrenergic receptors in canine platelets

The effects of papaverine on specific [3H]-yohimbine binding to canine platelet alpha 2-adrenergic receptors and on the platelet aggregation were assessed and compared with those of verapamil. Both compounds concentration-dependently inhibited [3H]-yohimbine binding with KI values for respective compounds of 0.39 +/- 0.05 microM (n = 3) and 15 +/- 0.19 microM (n = 3). In the presence of either compound KD values in Scatchard analysis of the equilibrium ligand binding increased in concentration-dependent manner, whereas Bmax did not change, indicating competitive inhibition of the ligand binding by these compounds. (-)-Epinephrine (3 microM) potentiation of adenosine diphosphate (ADP, 0.1 microM) aggregation was inhibited by papaverine with IC50 of 11 +/- 3.6 microM (n = 4). In the same experiments verapamil inhibited the platelet aggregation with lower IC50 (3.1 +/- 0.87 microM, n = 4) in comparison with that for papaverine. These results suggest that papaverine, like verapamil, inhibits physiological response of canine platelets through alpha-adrenergic receptor stimulation by direct interaction with the receptors.     

Signaling pathway involved in the inhibitory effect of FTY720P on the Na<Superscript>+</Superscript>/K<Superscript>+</Superscript> ATPase in HepG2 cells

The Na⁺/K⁺ ATPase modulates the activity of many transporters in the liver, and maintains the ionic constancy of the intracellular milieu, preserving thus normal functioning of hepatocytes. Previous work showed that FTY720P, a sphingosine one phosphate receptor agonist used in the treatment of multiple sclerosis, exerts in HepG2 cells, an inhibitory effect on the activity of the ATPase, mediated via PGE2. This study is an attempt to identify the signaling molecules involved downstream of the prostaglandin. The activity of the ATPase was assayed by measuring the amount of inorganic phosphate liberated in presence and absence of ouabain, a specific inhibitor of the enzyme. The effect of FTY720P and PGE2 disappeared completely in presence of PF-04418948, a blocker of EP2 receptors, RpcAMP, an inhibitor of PKA, PD98059, an inhibitor of ERK, as well as in presence of PTIO, a nitric oxide synthase inhibitor, but was mimicked by butaprost, an EP2 agonist, dbcAMP, a cell permeable cAMP analogue, and SNAP1,a nitric oxide generator. PGE2 and dbcAMP increased the expression of phosphorylated ERK but not total ERK. This increase did not appear however in presence of PTIO, indicating that PKA is upstream of NO. It was concluded that FTY 720P induces PGE2 release which activates NOS leading to NO production and ERK activation. ERK then inhibits directly or indirectly the Na⁺/K⁺ ATPase.     

Les injections intracaverneuses de prostaglandine E1: un reel progres dans l’efficacite et la securite des injections intracaverneuses

Intracavernosal injections of vaso-active agents were a major breakthrough in the investigation and treatment of impotence. However the first generation agents (papaverine ± phentolamine, phenoxybenzamine) result in a too high rate of complications (prolonged erection, priapism, corporeal fibrosis,), what requires strict precautions, making their use cumbersome and even stressing. A tremendous development of the pharmacological research arises in order to find agents not resulting in those risks. This article reviews the litterature concerning one of them, Prostaglandin E1 (PGE1), and compares its effects in near 4000 patients to those of papaverine ±phentolamine in over 5000. Experimental data suggests that PGE1 might act in normal erection, and its local tolerance might be better than that of Papaverine, specially as regards the risk of fibrosis. The studies without intraindividual comparison done in men tend to confirm this better tolerance: during the workup period, PGE1 induces prolonged erection in 2.7% of men (requiring treatment in only 0.3%) compared to 9.5% with papaverine and 5.3 % with papaverine + phentolamine (p<0.001); during auto-intracavernosal therapy, the rate of prolonged erections is 0.8% with PGE1 compared to 8.7% with papaverine, and the rate of fibrosis is 0.4% versus 3.5 to 5.4% with papaverine (p<0.001). PGE1 induces slightly more erections compatible with penetration than papaverine ± phentolamine (77% vs 73 %). However it induces pain in 23% of the cases, either at the time of injection, or, more specifically, during erection. This pain is intense in only 3.9% of the cases. General tolerance is excellent. The studies with intra-individual comparison confirm the statistically significant reduction of the risk of prolonged erection requiring treatment (1.7% with PGE1 versus 4.3% with papaverine and 6% with the combination, p<0.05) and the increase in efficacy (75% versus 64% with the combination), though papaverine is superior to PGE1 in a minority of the cases. By reducing the risks, PGE1 allows directly using the effective dosage as regards the diagnostic injections. Its higher effectiveness reduces the false-negative rate in the psychogenic patients, and widens the possibilities of the intracavernosal therapy.     

Revue rapide: 381 sujets impuissants,explores en 2 ans,dont 207 avec Prostaglandine E1

During the last two years, 381 impotent patients have been tested, 174 with papaverine and 207 with prostaglandin E1 (PGE1: Alprostadil). More positive responses were obtained using PGE1 (76.8%) than papaverine (44.8%), and some patients whose response to an initial PGE1 test was negative demonstrated a positive effect after a second injection. Two non-responders to PGE1 showed a positive response to papaverine+phentolamine. Tolerance of PGE1 was extremely good, provoking fewer prolonged erections than papaverine. Pain following intracavernosal injection of PGE1 was distressing in 10% of cases, all young men with psychogenic impotence. A therapeutic trial of repeated injections, carried out in 49 selected patients with mild organic dysfunction but strong psychogenic inhibition, typically showed a rapid clinical improvement. Moreover, 16 patients with cavernosal sclerosis and vascular impotence who were treated by injections two or three months apart (instead of having a prosthetic implant) showed a recovery of sexual activity. Consequently, during these two years only 35 patients (9.1%) were treated surgically, 21 by the implantation of hydraulic prostheses. The significant benefits of PGE1, especially the ease of self-administration, demonstrate the need for new therapeutic trials.     

Prostacyclin (PGI2) inhibits the development in human platelets of ADP and arachidonic acid-induced shape change and procoagulant activity

Platelets which change shape from discs to spheres concomitantly develop platelet procoagulant activity which is independent of and precedes aggregation or the release reaction. Since prostacyclin (PGI₂) is known to be potent inhibitor of platelet aggregation and releae, the effect of PGI₂ on platelet shape change and the development of platelet procoagulant activity was measured. Platelet shape change (percent discs and spheres) was assayed by a light transmission technique. Platelet procoagulant activity was assayed using recalcified clotting times measured concurrently (by aggregometry) with platelet shape assays. PGI₂ inhibited the development of platelet shape change and procoagulant activity induced by the addition of ADP (0.7 μM); the 50% inhibitory dose of PGI₂ was 2 nM. PGI₂ also inhibited arachidonic acid (0.3–1.2 mM) induced platelet shape change and procoagulant activity; the 50% inhibitory dose of PGI₂ was 2.3 nM. Thus, physiologic concentrations of PGI₂ inhibit platelet shape change and prevent the development of sphering associated procoagulant activity.     

Involvement of a calcium-independent pathway in plasmin-induced platelet shape change

Plasmin-induced platelet activation is considered to be a cause of reocclusion after thrombolytic therapy with plasminogen activators. However, little is known regarding its mechanism and regulation, particularly with respect to the initial step shape change. We here demonstrate that a Ca2+-independent pathway is involved in plasmin-induced human platelet shape change, and that Rho-kinase plays an important role in this pathway. When the increase in cytosolic Ca2+ was prevented by an intracellular Ca2+ chelator, 5,5'-dimethyl-BAPTA, plasmin-induced platelet shape change was partially inhibited but still occurred. In the presence of 5,5'-dimethyl-BAPTA, a specific Rho-kinase inhibitor, Y-27632, completely inhibited the shape change. Phosphorylation of myosin light chain, a key regulator of platelet shape change, was completely inhibited by Y-27632 in 5,5'-dimethyl-BAPTA-treated platelets. Although plasmin caused tyrosine phosphorylation of the 80 kDa protein during the shape change, it did not seem to have a critical role. cAMP-elevating agents inhibited plasmin-induced shape change in 5,5'-dimethyl-BAPTA- or Y-27632-treated platelets with similar efficiency. These results indicated that plasmin causes platelet shape change by activating Ca2+-dependent and Ca2+-independent-Rho-kinase-dependent pathways, both of which are sensitive to cAMP.     

Interet diagnostique d’une injection intracaverneuse de 20 μg de Prostaglandine E1 dans l’impuissance

Intracavernous injection of 20 μg of prostaglandin E1 (PGE1) was carried out in 130 impotent patients. The erectile response was compared to the results of arteriological investigations including nocturnal penile tumescence and rigidity monitoring (NPTR) in 59 patients. The response of 60 patients positively categorized as exclusively psychogenic or vasculogenic was also compared to the pattern of the response to 80 mg of papaverine observed in a previous study by the same authors. The PGE1 test may not discriminate psychogenic from wholly organic patients since its results are not correlated to those of NPTR. It helps for the screening of vasculogenic impotence. Lack of response or a partly rigid response is consistent with this actiology but is not specific for it. A fully response makes it unlikely. Compared to papaverine, PGE1 induces less non rigid responses in psychogenic patients (15% versus 35% with papaverine) and more fully rigid responses in vasculogenic patients (respectively 12% and 5 %). Consequently the specificity of the PGE1 test is higher but its sensitivity lower than that of papaverine so that there is no clear difference in the effectiveness of the tests. Nevertheless the PGE1 test should be preferred, because it is safer. Prolonged erections occured in only 5 patients, and all ceased spontaneously. However 4 presented severely painful erections.