BACKGROUND Impotence is the consistent inability to achieve or sustain an erection of sufficient rigidity for sexual intercourse. It has been estimated that approximately 10 million American men are impotent (R. Shabsigh et al.,"Evaluation of Erectile Impotence,"Urology, 32: 83-90, 1988 ; W. L. Furlow,"Prevalence of Impotence in the United States,"Med. Aspects Hum. Sex., 19: 13-16,1985).

A number of causes of impotence have been identified, including vasculogenic, neurogenic, endocrinologic and psychogenic. Vasculogenic impotence, which is caused by alterations in the flow of blood to and from the penis, is thought to be the most frequent organic cause of impotence. Common risk factors for vasculogenic impotence include hypertension, diabetes, cigarette smoking, pelvic trauma and the like. Neurogenic impotence is associated with spinal-cord injury, multiple sclerosis, peripheral neuropathy caused by diabetes or alcoholism and severance of the autonomic nerve supply to the penis consequent to prostate surgery.

Erectile dysfunction can also be associated with disturbances in endocrine function resulting in low circulating testosterone levels and elevated prolactin levels.

More specifically, penile erection requires (1) dilation of the arteries that regulate blood flow to the lacunae of the corpora cavernosum, (2) relaxation of trabecular smooth muscle, which facilitates engorgement of the penis with blood, and (3) compression of the venules by the expanding trabecular walls to decrease venous outflow.

Trabecular smooth muscle tone is controlled locally by adrenergic (constrictor), cholinergic (dilator) and nonadrenergic, noncholinergic (dilator) innervation, and by endothelium-derived vasoactive intestinal polypeptide (VIP), prostanoids, endothelin and nitric oxide (NO). High sympathetic tone (noradrenergic) is implicated in erectile dysfunction, and, in some patients the disorder can be successfully treated with noradrenergic receptor antagonists (Krane et al., New England Journal of Medicine, 321: 1648,1989).

Under normal physiological conditions, following sexual stimulation, release of nitric oxide from penile non-adrenergic, non-cholinergic nerves and the endothelium activates guanylyl cyclase and induces intracellular cGMP synthesis in erectile tissue trabecular smooth muscle cells. Increased cGMP levels reduce intracellular calcium concentrations, inhibiting smooth muscle contractility and thereby initiating the erectile response.

Sildenafil citrate (Viagra@), which is a phosphodiesterase type V inhibitor was introduced to the US market in 1998, for management of erectile dysfunction, by oral ingestion. It is currently the preferred treatment.

Phosphodiesterase type V (PDE type V) is the predominant enzyme responsible for cGMP hydrolysis in trabecular smooth muscle relaxation, resulting ultimately in restoration of basal smooth muscle contractility and penile flaccidity. Sildenafil citrate is a potent PDE type V reversible and selective inhibitor which blocks cGMP hydrolysis effectively at low concentrations. Under conditions of excessive adrenergic tone or impaired neurovascular status, following sexual stimulation, sildenafil acts to enhance NO-mediated (nitric oxide) smooth muscle relaxation, resulting in improved penile erection in men with erectile

dysfunction. In the absence of sexual stimulation, there is minimal NO released from the parasympathetic NANC nerves and the cGMP level is low. The penis is flaccid as the input from the sympathetic nervous system exceeds that from the parasympathetic NANC nervous system.

Accordingly, the mode of action of the oral drug is to maintain an erection by blocking cGMP hydrolysis. Thus, it has been accepted that in the absence of sexual stimulation, sildenafil citrate has no activity on corporal smooth muscle relaxation. Therefore, Viagra has a poor response rate in patients who have lost neural penile input due to, for example, neurogenic causes or radical prostatectomy.

Other methods of treating erectile dysfunction are described in: U. S. Pat. No.

4,127,118 to Latorre ; U. S. Pat. No. 5, 439,938 to Snyder et al. ; Virag et al., Angiology- Journal of Vascular Diseases, February, pp. 79-87,1984 ; Brindley, Brit. J Psychiat., 143: 332-337,1983 ; and Steif et al., Urology, XXXI: 483-485,1988 ; PCT Publication No.

WO 91/16021, U. S. Pat. No. 4,801,587 to Voss et al., and U. S. Pat. Nos. 5,242,391; 5,474,535 ; 5,686,093 and 5,773,020 to Place et al.; U. S. Pat. No. 6,037,346 to Doherty et al. ; US Pat. No. 5,236,904 to Gerstenberg et al. However, these methods have generally been considered less desirable than Viagra@.

A need exists for an alternative treatment. Prior to the introduction of the phosphodiesterase inhibitors (PDE), one general mode of administration was intracavernosal injection. However, such treatments, for male impotence have not been considered fully effective and require concurrent sexual stimulation. The currently available dopamine agonists, with few exceptions, have found limited use in the treatment of erectile dysfunction because of their peripheral side effects. These effects include nausea and vomiting, postural hypotension, arrhythmias, tachycardia, dysphoria, psychosis, hallucinations, drowsiness and dysidnesias (See, e. g., Martindale The Extra Pharmacopeia, 31 st Ed., pages 1151-1168).

In some cases, particularly where the dysfunction is psychogenic or neurogenic and severe atherosclerosis is not involved, injection of papaverine, a smooth muscle relaxant, or phenoxybenzamine, a non-specific blocker and hypotensive, into a corpus cavernosum has been found to cause an erection sufficient for vaginal penetration. However, treatment of impotence with papaverine or phenoxybenzamine often results in priapism, a locking up of an erection for a long period of time, typically a few hours and sometimes longer than 24 hours.

Priapism is a serious, deleterious side effect of treatment of erectile insufficiency with these drugs. Priapism is usually painful, irreversibly damages erectile tissue, and, to be relieved requires bleeding or pharmacologic intervention (e. g., injection of a sympathomimetic drug such as adrenaline). Even if priapism does not occur with use of papaverine, such use is associated with a painful, burning sensation in the first two or so minutes after the injection and there are indications that repeated use of papaverine causes undesirable, extensive intracavemous fibrosis.

A need exists for an intracavernosal agent, which is safe and effective in causing erection without significant pain.

Also widely used to treat impotence are penile implants, whereby an erection sufficient for vaginal penetration can be caused mechanically. In recent years, implants have been employed, especially in cases where injections of papaverine are ineffective, which are usually cases of severe atherogenic impotence. However, treatment of impotence with penile implants also entail serious disadvantages. Such treatment requires surgery and entails total destruction of the erectile tissues of the penis, forever, precluding normal erection.

The neuropeptide, human vasoactive intestinal peptide (hereinafter referred to as "VIP"), is thought to be associated with erections in normal males) i. e. males not suffering from erectile insufficiency). Injection of VIP into a corpus cavernosum of a normal male,

without subjecting the male to sexual stimulation, causes only a slight swelling (slight tumescence) of the penis but not an erection.

It would be useful to have a method of treating individuals who do not typically respond to oral Viagra treatment, for example, individuals who suffer from neurogenic impotence or other causes resulting in loss of penile neural input such as those individuals who have undergone penile prostatectomy.

SUMMARY OF THE INVENTION We have now surprisingly discovered that intracavernosal injection of a PDE type V inhibitor such as sildenafil causes penile erection, in the absence of sexual stimulation.

Thus, individuals exhibiting a poor response rate to an oral dose of a PDE type V inhibitor such as sildenafil, can be treated by local administration to the individual of a pharmaceutical formulation containing sildenafil citrate. Administration of the pharmaceutical formulation is carried out within the context of a predetermined dosing regimen such that the agent is effective in the treatment of erectile dysfunction. Drug delivery is preferably administered via intracavernosal injection. The effective dose ranges from about 5-60 mg, preferably 10-50 mg, still more preferably about 10-40 mg, even more preferably about 15-40 mg and still more preferably about 20-30 mg.

In another aspect of the invention, a method is provided for treating an individual who has lost penile neural input and who does not respond to oral doses of sildenafil, the method comprising local administration to such an individual a pharmaceutical formulation containing a PDE type V inhibitor such as sildenafil citrate without the need for prior sexual stimulation. Administration of the pharmaceutical formulation is carried out within the context of a predetermined dosing regimen such that the agent is effective in the treatment of erectile dysfunction, as described above. Drug delivery is preferably administered via intracavernosal injection.

In another aspect a kit is provided to assist an individual in drug administration to carry out the method of the invention. Generally the kit will include the following components: a pharmaceutical formulation comprising the sildenafil citrate to be administered; a device for effective delivery of the pharmaceutical composition ; a container housing the pharmaceutical formulation during storage and prior to use; and instructions for carrying out drug administration in a manner effective to treat erectile dysfunction.

BRIEF DESCRIPTION OF THE DRAWINGS Figures 1A and 1B. Effect of L-NAME and ODQon sildenafil-induced relaxation in rabbit corpus cavernosum smooth muscle. Rabbit corpus cavernosum tissue strips exposed to vehicle (saline or 50% DMSO). Tissue strips were then exposed to increasing concentrations of sildenafil citrate by cumulative addition and the relaxation response was recorded. In parallel experiments, tissue strips were first incubated for 30 minutes with 0.6 mM L-NAME (n=7) (A) or 3 um ODQ (n=4) (B), and the relaxation response to increasing doses of sildenafil was determined. Data are the mean + SEM.

Figures 2A, 2B, 2C and 2D. Effects of intracavernosal sildenafil. Varying doses of sildenafil (0.3-1.3 mg) were administered to anesthetized adult male New Zealand White rabbits by intracavernosal injection. We measured the peak ICP and normalized it to the SAP (A). A normalized ratio of 1.0 indicates full erection and 0.27 indicates a flaccid penis. We also determined the time to reach peak ICP in minutes (B), the duration of erection in hours (C), and changes in SAP (mm Hg) with sildenafil administration (D). Data represent the mean of four separate experiments (*P < 0.5, relative to saline).

Figure 3. Effects of intracavernosal L-NAME and sildenafil on ICP. Anesthetized adult male New Zealand White rabbits were subjected to PNS to induce penile erection.

Following detumescence, rabbits were administered 0.2 mL of 3 mM L-NAME by intracavernosal injection and subjected to PNS after 10 minutes. Some rabbits received

intracavernosal administratio of L-NAME followed by sildenafil (1.3 mg) in the absence of PNS. The data represent the mean of four separate experiments (*P zu 0.5 relative to PNS in the absence of L-NAME).

DETAILED DESCRIPTION OF THE INVENTION We have discovered a method for treating erectile dysfunction in a male individual for whom oral drug treatment of sildenafil citrate is ineffective. The method comprises locally administering to the individual, an effective amount of a pharmaceutical composition comprising a Type V phosphodiesterase inhibitor.

The preferred class of individuals for treating erectile dysfunction are male individuals, wherein the individual does not respond to physical sexual stimuli. For example, individuals whereby the cause of erectile dysfunction is due to loss of neural input to the penis by neurogenic impotence or other causes, such as radical prostatectomy.

The preferred phosphodiesterase inhibitor is sildenafil citrate, wherein administration is effected via intracavernosal injection without the need for prior or concurrent sexual stimulation.

The preferred dose range of sildenafil citrate is preferably less than 60 mg, more preferably less than 50 mg, even more preferably 10-40 mg, still more preferably 15-40 mg, and most preferably 20-30 mg.

The term"erectile dysfunction"as used herein, is intended to include any and all types of erectile dysfunction, including: vasculogenic, neurogenic, endocrinologic and psychogenic impotence; Peyronie's Syndrome; priapism; premature ejaculation; and any other condition, disease or disorder, regardless of cause or origin, which interferes with at least one of the three phases of human sexual response, i. e., desire, excitement, and orgasm (see Kaplan, Disorders of Sexual Desire, New York, N. Y.: Brunner Mazel Books Inc., 1979).

Most preferably it includes the class of individuals having lost penile neural input and hence

cannot respond to treatments requiring concurrent sexual stimulation for the treatment of erectile dysfunction.

The terms"active agent,""drug"and"pharmacologically active agent"are used interchangeably herein to refer to a chemical material or compound that induces a desired effect. In the preferred embodiments herein, the terms refer to a PDE type V inhibitor such as sildenafil citrate which is capable of being delivered locally, preferably by intracavernosal injection. The preferred PDE type V inhibitor is sildenafil citrate.

The term"intracavernosal"as used herein refers to a mode of drug administration and involves injection into one or both corpora of the corpora cavernosal tissue of the penis.

"Carriers"or"vehicles"as used herein refer to carrier materials suitable for local drug administration. Carriers and vehicles useful herein include any such materials known in the art which is nontoxic and does not interact with other components of the composition in a deleterious manner.

By an"effective"amount of a drug or pharmacologically active agent is meant a nontoxic but sufficient amount of the drug or agent to provide the desired effect, i. e., treatment of erectile dysfunction.

Preferably, sildenafil citrate is administered to an individual prone to erectile dysfunction by any known method of injection.

Sildenafil citrate is designated chemically as 1- [ [3- (6, 7-dihydro-1-methyl-7-oxo-3- propyl-lH-pyrazolo [4,3-d] pyrimidin-5-yl)-4-ethoxyphenyl] sulfonyl]-4-methylpiperazine citrate. A commercial form of this is Viagra@, the citrate salt of sildenafil, which is sold by Pfizer. The process for preparing Sildenafil has been described in U. S. Patent Nos. 5,955,611 and 6,066,735 to Dunn et al.

Examples of type V phosphodiesterase inhibitors include, but are not limited to, zaprinast, MY5445, dipyridamole. Other type V phosphodiesterase inhibitors are disclosed

in PCT Publication Nos. WO 94/28902 and WO 96/16644 and US Patent No.: 6,037,346 to Doherty et al.

Salts, esters, amides, prodrugs, and other derivatives of the active agents may be prepared using standard procedures known to those skilled in the art of synthetic organic chemistry and described for example, by J. March, Advanced Organic Chemistry ; Reactions, Mec1hanisms and Strgsture, 4 Ed. (New York, Wiley-Interscience, 1992). See also U. S.

Patent No. 6,037,746. For example, acid addition salts are prepared from the free base using conventional methodology, and involves reaction with a suitable acid. Generally, the base form of the drug is dissolved in a polar organic solvent such as methanol or ethanol and the acid is added thereto. The resulting salt either precipitates or may be brought out of solution by addition of a less polar solvent. Suitable acids for preparing acid addition of salts include both organic acids, e. g. acetic acid, propionic acid, glycolic acid and the like, as well as inorganic acids, e. g. hydrochloric acid, hydrobromic acid, sulfuric acid, nitric acid, phosphoric acid, and the like. An acid addition salt may be reconverted to the free base by treatment with a suitable base. Conversely, preparation of basic salts of acid moieties which may be present on a phosphodiesterase inhibitor molecule are prepared in a similar manner using a pharmaceutically acceptable base such as sodium hydroxide, potassium hydroxide, ammonium hydroxide, calcium hydroxide, trimethylamine or the like. Preparation of esters involves functionalization of hydroxyl and/or carboxyl groups which may be present within the molecular structure of the drug. The esters are typically acyl-substituted derivatives of free alcohol groups, i. e., moieties which are derived from carboxylic acids of the formula RCOOH where R is alkyl. Esters can be reconverted to the free acids, if desired, by using hydrogenolysis or hydrolysis procedures. Amides and prodrugs may also be prepared using techniques known to those skilled in the art or described in the pertinent literature. For example, amides may be prepared from esters using suitable amine reactants, or they may be

prepared from an anhydride or an acid chloride by reaction with ammonia or a lower alkyl amine. Prodrugs are typically prepared by covalent attachment of a moiety which results in a compound that is therapeutically inactive until modified by an individual'metabolic system.

The preferred method of administration is via intracavernosal injection.

Intracavernosal injection can be carried out by use of a syringe or other suitable device. An example of a hypodermic syringe useful herein, that can be used for simultaneous injection in both corpora is described in U. S. pat. No. 4,127,118 to Latorre. The injection is made on the dorsum of the penis by placement of the needle to the side of each dorsal vein and inserting deep into the corpora.

For intracavernosal injection, the active agent to be administered is incorporated into a sterile liquid preparation, typically a solution or suspension in an aqueous or oleaginous medium. The solution or suspension may be formulated according to techniques known in the art using suitable carriers, dispersants, wetting agents, diluents, suspending agents or the like. Among the acceptable vehicles and solvents that may be employed are water, isotonic saline, vegetable oil, fatty esters and polyols.

We have demonstrated that a PDE type V inhibitor, such as sildenafil citrate, relaxes human corpus cavernosum strips in organ baths in the absence and presence of nitric oxide synthase inhibitor. This indicates that sildenafil at those levels is acting on smooth muscle by a mechanism other than that mediated by NO. Animal studies using intracavernosal injection of sildenafil caused penile erection in the absence of sexual stimulation. These observations, indicate that the present method results in a new and novel mechanism of sildenafil action in corpus cavernosum.

The preferred time of injection is at least 5 minutes prior to sexual contact, more preferably at least 15 minutes.

The preferred dose range that would correct male erectile dysfunction without causing priapism is about 5-60 mg, preferably 10-50 mg, more preferably 15-40 mg, still more preferably 20-30 mg.

One preferred class of individuals to be treated by the present method are those individuals prone to erectile dysfunction or only partially respond to physical sexual stimulation and do not respond effectively to oral treatment by Viagra (g.

Another preferred class of individuals to be treated are individuals prone to erectile dysfunction who cannot respond to physical sexual stimuli.

Yet another group of individuals who can be treated by the present method are individuals who have lost, at least some, penile neural input due to neurogenic or other causes relating to loss of neural input to the penis. For example, individuals who have undergone radical prostatectomy.

The invention also encompasses a kit for patients to carry out the present method.

The kit contains the pharmaceutical formulation to be administered, a device for administering the formulation, preferably a syringe and hypodermic needle as described supra (such as Latorre), a container, preferably sealed, for housing the drug and device during storage and prior to use, and instructions for carrying out drug administration in an effective manner. The formulation may consist of the drug in unit dosage form. The kit may also contain multiple formulations of different dosages of the same agent.

It is to be understood that while the invention has been described in conjunction with the preferred specific embodiments thereof that the foregoing description as well as the examples that follow are intended to illustrate and not limit the scope of the invention. Other aspects, advantages and modifications within the scope of the invention will be apparent to those skilled in the art to which the invention pertains.

EXAMPLES Example 1 Solubilized Viagra was injected intracavernosally into the penis of several anesthetized animals which resulted in sustained penile erection.

Example 2 In vivo Studies New Zealand white male rabbits (3.5-4 kg) will be sedated and anaesthetized with subcutaneous ketamine and xylazine (35 mg/kg and 5 mg/kg respectively). Lidocaine (4 mg/kg) is used for local infiltration at the base of the penis and the midline of the neck.

Sharp and blunt dissection is used to expose the left corpus cavernosum and the left carotid artery. A 21-gauge butterfly needle is placed in the corpus cavernosum and a 20-gauge angiocath in the carotid artery. These are connected to a Grass Model 7400 physiologic recorder for the simultaneous monitoring of the intracavernosal and systemic arterial pressures. A 27-gauge butterfly needle is placed in the left corpus cavernosum approximately 7 mm proximal to the 21-gauge needle. Sildenafil is dissolved in normal saline and injected directly into the left corpus cavernosum through the smaller needle. The injection volume is standardized at 0.2 ml to ensure minimal volume effect on the intracorporal pressure. The effect of sildenafil on the intracorporal pressure, the duration of the pressure increase and the change in arterial pressure with intracavernosal administration are then evaluated. Sildenafil (0.3 mg, 0.6 mg, 1.0 mg and 1.3 mg) is administered intracavernosally and saline is used as a control.

The time interval between injection and the peak intracavernosal pressure and the duration of erection are read directly from the data charts. The normalized intracavernosal pressure rise (peak intracavernosal pressure subtract the pre-injection intracavemosal

pressure divided by the systemic arterial pressure at the time of the peak intracavernosal pressure) is calculated.

Example 3 In vitro Studies Organ chambers can be used to evaluate the effect of sildenafil on rabbit corpus cavernosal tissue. These tissues are obtained and prepared for organ bath studies as previously described. The effects of sildenafil on smooth muscle relaxation in the absence or presence of nitric oxide synthase inhibitors will be determined.

Based on our preliminary studies, the administration of sildenafil at 0.25 mg/kg is effective in causing erections. This corresponds to an intracavernosal doses of 20 mg.

Intracavernosal sildenafil at this concentration may result in reduced levels of sildenafil in the systemic circulation and minimize the noted adverse effects observed with oral Viagra such as headaches, flushing, visual effects and gastro-esophageal reflux.

All references are hereby incorporated by reference in their entireties.

Example 4 Materials and Methods Studies in Vitro Unless specified, all reagents were purchased from Sigma Chemical Co., St. Louis, Mo.

Penile corpus cavernosum tissue strips were obtained from male New Zealand White rabbits (3.5-4.0 kg ; Pine Acres Rabbitry, Builford, Vt.) and prepared for organ bath studies, as previously described (Kim et al, 1991). Briefly, rabbit cavernosal tissue strips were mounted onto force transducers (Model FT03 ; Grass Instruments, Quincy, Mass.) and immersed in 25- mL baths of physiological salt solution (118.3 mM NaCI, 4.7 mM KC1, 0.6 mM MgS04,1.2 mM KH2P04, 2.5 mM CaCl2, 25 mM NaHCO3, 0.026 mM CaNa2EDTA, and 11.1 mM glucose) maintained at 37°C and aerated with 5% C02, 19% 02, and 76% N2. Tissues were

progressively stretched until optimal isometric tension was reached. All protocols included 3 gm indomethacin to inhibit synthesis of prostaniods and 10 llm bretylium to inhibit adrenergic neurotransmission. Cavernosal tissue strips were contracted with 1 um phenylephrine and subjected to electrical field stimulation (2,5, and 10 Hz) and then exposed to 1, um acetylcholine. After extensive washing, tissue strips were contracted again with phenylephrine and treated with either vehicle (distilled, deionized H20) or 0.6 mM of the NOS inhibitor N'-nitro-L-arginine methyl ester (L-NAME) for 30 minutes. Tissue strips were then subjected to electrical field stimulation and exogenous acetylcholine. Tissues treated with L-NAME did not relax to either electrical field stimulation or acetylcholine, indicating effective blockade of NO synthesis (unpublished). Tissue strips were then exposed to increasing concentrations of sildenafil citrate (a generous gift from Dr. Farid Saad, Jenapharm GmbH, Jena, Germany) by cumulative addition. In a separate protocol, cavernosal tissue strips at optimal isometric tension were treated with vehicle (50% DMSO [dimethyl sulfoxide]) or 3 urn of the guanylyl cyclase inhibitor, lH- [l, 2,4]-oxadiazolo- [4,3- a]-quinoxalin-I-one (ODQ ; Biomol Research Laboratories, Plymouth Meeting, Pa.) for 20 minutes prior to performing sildenafil dose responses.

Cyclic GMP Measurements Strips of rabbit corpus cavernosum tissue were dissected out, rinsed in ice-cold physiological salt solution, and placed into tissue culture dishes containing Dulbecco modified Eagle medium supplemented with 0.05% bovine serum albumin, penicillin, and streptomycin (GIBCO-BRL, Rockville, Md.). Tissue strips were equilibrated at 37°C in a C02 tissue culture incubator for 1 hour with one change of media halfway through the equilibration time.

Vehicle or sildenafil (100nM final concentration) was added to the media. After 30 minutes, tissue strips were quickly removed, blotted, and frozen in liquid nitrogen. Frozen tisues were pulverized and homogenized in ice-cold 1 N percholoric acid using a glass-glass tissue

grinder. Crude homogenates were processed for CGMP assay as prviously described (Kim et al, 1993). Cyclic GMP accumulation was quantified by radioimmunoassay (Biomedical Technologies, Inc., Stoughton, Mass.). All results were normalized to total soluble tissue protein content.

In Vivo Studies New Zealand White male rabbits (3.5-4.0) were sedated with subcutaneous ketamine (35 mg/kg) and xylazine (5 mg/kg), and anesthesia was maintained with intravenous pentobarbital (20 mg, as required: James Brudnick Co., Malden, Mass.). Lidocaine (4 mg/kg) was used for local infiltration at the base of the penis and the midline of the neck.

Surgical dissection was performed to expose the left corpus cavernosum and the left carotid artery. In the experiments in which access to the pelvic nerve was required, the nerve was localized on the antero-lateral rectal wall and a bipolar electrode was placed in the corpus cavernosum and a 20-gauge angiocatheter was introduced into the carotid artery. These were connected to a model 7400 physiologic recorder (Grass Instruments) for the simultaneous monitoring of intracavernosal pressure (ICP) and systemic arterial pressure (SAP). A 27- gauge butterfly needle was placed in the left corpus cavernosum approximately 7-8 mm proximal to the 21-gauge needle. Sildenafil (0.3,0.6,1.0, and 1.3 mg) was injected directly into the left corpus cavernosum through the smaller needle. Injection volumes were standardized to 0.2 mL to ensure minimal volume-related changes on ICP. The effect of sildenafil on ICP, the duration of the pressure increase, and the change in SAP with intracavernosal administration were measured.

Some rabbits were injected intracavernosally with 0.2 mL of 3 mM L-NAME 10 minutes prior to pelvic nerve stimulation (PNS) or sildenafil injection. The pelvic nerve was stimulated for 30 seconds with a train of square waves at an amplitude of 10 volts, a pulse duration of 0.8 msec, and frequency of 32 Hz. The ICP/SAP ration for PNS alone was

compared with that of L-NAME treatment followed by PNS or 1.3 mg of sildenafil. Three data parameters were analyzed. The time interval between injection and the peak ICP (time to peak) and the duration of erection were read from the data charts. The normalized peak ICP (peak ICP divided by SAP) was calculated. A normalized ratio of 1.0 indicates that ICP equals SAP and that the penis was fully erect. In the flaccid state, the normalized ration was 0.27. These studies were approved by the Institutional Animal Care and Use Committee at the Boston University School of Medicine.

Statistics Data were analyzed by analysis of variance, followed by a two-tailed Student's t test (with appropriate Bonferroni corrections) for comparison between two groups. Data were considered statistically significant when the P value was less than or equal to 0.5.

In Vitro Studies In rabbit corpus cavernosum, tissue strips contracted with phenylephrine, and sildenafil caused dose-dependent relaxation at concentrations above 0.1 u M (Figure 1A).

Little or no effect was observed at lower concentrations. This relaxation response may in part be mediated by the basal release of NO from the NANC nerves and the endothelium lining the lacunar spaces of the corpus cavernosum ; To these this, tissue strips were treated with L- arginine derivative, L-NAME, to inhibit neuronal and endothelial NOS activity. The efficacy of 0.6mM L-NAME was verified by its ability to inhibit relaxation to electrical field stimulation and to exogenous acetylcholine, as described in"Materials and Methods."It is interesting that treatment with L-NAME did not alter the relaxation to sildenafil (Figure 1A).

In a separate series of experiments, we treated rabbit cavernosal tissue strips with the guanylyl cyclase inhibitor, ODQ, to determine if the relaxation to sildenafil was mediated by cGMP production. Addition of 50% DMSO vehicle significantly enhanced sildenafil- induced relaxation, relative to saline vehicle (Figure 1). Treatment with 3 u. M ODQ

increased smooth muscle tone (mean =41%) and attenuated, but did not prevent, sildenafil- induced relaxation (Figure 1B). In cavernosal tissue strips, basal levels of cGMP were determined to be 0.54 0.09 pmol/mg prot (n = 6). Incubation with 100 nM sildenafil resulted in a twofold increase in cGMP (1.10 i 0.11 pmol/mg prot) over a 30-minute time interval.

In Vivo Studies We further tested the effects of high concentrations of sildenafil on penile erections in an in vivo animal model ICP and SAP were recorded to objectively assess the degree of penile tumescence. In the flaccid state, ICP was 27% of SAP. Injection of saline alone did not result in an erection, whereas intracavernosal injection of sildenafil caused a dose- dependent increase in the ICP/SAP ratio. At doses of 1 mg and greater, ICP approximated SAP, indicating that the penis was fully erect (Figure 2A). The time to achieve maximal stable ICP was dose-dependent and was prolonged for higher doses of sildenafil (Figure 2B).

The duration of the erection resulting from intracavernosal injection of sildenafil also increased in a dose-dependent fashion (Figure 2C). One of 4 animals had an erection lasting greater than 5 hours when administered 1 mg of sildenafil. At the 1.3 mg dose, 3 of 4 animals had erections lasting greater than 5 hours. Intracavernosal sildenafil had no significant effect on SAP at all doses tested (Figure 2D).

To determine if intracavernosal injection of sildenafil caused erections via the NO/cGMP pathway or by an independent mechanism, we administered 0.2 mL of 3 mM L- NAME intracavernosally 10 minutes prior to either PNS or intracavemosal sildenafil (1.3 mg). As shown in Figure 3, PNS caused full erection that was significantly reduced by intracavernosal administration of L-NAME. Intracavernosal injection of sildenafil after L- NAME administration caused a full erection that was indistinguishable from that obtained by administration of sildenafil alone.

Discussion Under normal physiological conditions, following sexual stimulation, release of NO from the penile parasympathetic NANC nerves and the endothelium activates guanylyl cyclase and induces intracellular cGMP synthesis in erectile tissue smooth muscle cells.

Increased cGMP levels activate protein kinase G and reduce intracellular calcium concentrations, inhibiting smooth muscle contractility and thereby initiating the erectile response. Because the NO/cGMP pathway is a key modulator of penile erection, agents that inhibit cGMP hydrolysis or increase cGMP synthesis will be useful therapeutic agents for treatment of erectile dysfunction. PDE type 5 is the predominant enzyme responsible for cGMP hydrolysis in trabecular smooth muscle. Activation of PDE type 5 terminates cGMP- mediated smooth muscle relaxation, resulting ultimately in restoration of basal smooth muscle contractility and penile flaccidity. Sildenafil citrate, a potent, reversible, and selective PDE type 5 inhibitor effectively blocks cGMP hydrolysis and has been shown to be a safe and effective oral drug in the treatment of erectile dysfunction (Goldstein et al, 1998). Based on the basic science and clinical data published in the literature, the activation of the NO pathway be sexual stimulation is required for sildenafil to be effective in facilitating erections. This premise is based on the ability of oral sildenafil to inhibit PDE type 5.

In this study, we observed that sildenafil relaxed corpus cavernosum smooth muscle in organ baths in the absence of significant NOS activity. This relaxation response also persisted in the presence of ODQ at concentrations shown to completely inhibit cGMP production (Moro et al. 1996). The relaxation response to sildenafil was, interestingly, potentiated by the addition of 50% DMSO vehicle, relative to saline vehicle. Organ bath experiments using isolated rabbit aorta suggest that DMSO causes vascular smooth muscle relaxation by inhibiting calcium mobilization (Jackson et al, 1979). Similar mechanisms may

cause the potentiation observed in our studies. The partial attenuation of sildenafil-induced relaxation by ODQ is most likely due to the inhibition of guanylyl cyclase basal activity.

Basal production of cGMP may accumulate to exert a significant modulating effect on cavernosal smooth muscle tone in the presence of PDE 5 inhibition. This is supported by our data that demonstrate a twofold increase in CGMP levels in the presence of sildenafil.

Whether this basal cGMP production is due to stimulation by NO remains unclear because we did not treat cavernosal tissue strips with both L-NAME and ODQ. Nevertheless, sildenafil-induced cavernosal smooth muscle relaxation is not blocked by guanylyl cyclase inhibition. Thus, these data suggest that sildenafil facilitates smooth muscle relaxation by a mechanism that is different from the previously accepted dogma (ie, PDE 5 inhibition).

Further, our experiments in anesthetized animals have documented that sildenafil administered intracavernosally caused erections in the absence of PNS as well as in the presence of L-NAME. The inhibition constant of sildenafil for PDE 5 in human corpus cavernosum tissue extracts is approximately 3.5nM (Moreland et al, 1999). The plasma level of sildenafil following oral administration of a 100 mg. Viagra tablet is 560 ng/mL (0.84suM, Center for Drug Evaluation and Research, 1998). In the rabbit model, intracavernosal injection of 1 mg. Of sildenafil translates into a local plasma concentration of 2. 1 mM within the penile corpora cavernosa (assuming a rabbit penis volume of 0.7 mL). This estimate of the intracavernosal concentration of sildenafil is 6000000 times greater than the K of this drug and 2500 times greater than the serum level following oral administration of the highest available does of Viagra. This disparity, in conjunction with our data, strongly suggests that pathways other than inhibition of PDE 5 activity mediate intracavernosal sildenafil-induced erections.

Although high concentrations of sildenafil (M) have recently been shown to enhance intracellular cyclic adenosine monophosphate (cAMP) accumulation in isolated human cavernosal tissue segments, this effect was ultimately attributed to an indirect effect of PDE 5 inhibition (Stief et al, 2000). The authors reasoned that an elevation in cGMP could inhibit PDE type 3 and lead to the accumulation of cAMP, which is also known to cause vascular smooth muscle relation. This mechanism is supported by numerous studies in peripheral blood vessels (Maurice and Haslam, 1990; Pyne et al, 1996; Pelligrino and Wang, 1998).

However, unlike human penile erectile tissue, rabbit corpus cavernosum appears to lack PDE type 3 (Qiu et al, 2000). Thus, our results in the rabbit model cannot be explained by the inhibition of PDE 3 by cGMP. We suggest that sildenafil, at high doses, causes smooth muscle relaxation by a novel mechanism independent of the NO/cGMP pathway.

Further anecdotal evidence for the effects of intracavernosal sildenafil is obtained from a case report of a 47-year old man who did not respond to oral sildenafil (personal communication, I. Goldstein). Following a self-administered intracavemosal injection of a crushed Viagra tablet (100 mg) suspended in saline, he experienced immediate pain in both corpora and developed a tender erection over the next 2 hours. A rigid and painful erection ensued. He sought medical attention after 6 days of priapsim, and was managed with intracavernosal saline lavage, alpha-adrenergic agonist therapy, a Winters shunt, and an Al- Ghorab shunt. Despite these therapies, he developed complete erectile dysfunction. Tri-mix (phentolamine 1 mg/mL, papaverine 30 mg/mL, and PGE, 100 llg/mL) was ineffective, and he elected to have a penile prosthesis inserted. In addition to any confirmed or potential mechanisms of sildenafil action, it is reasonable to speculate that the insoluble talc filler contained in the tablet may have obstructed the subtunical venules, resulting in veno- occlusion and priapism. It is unlikely that trauma to a cavernosal artery causing high-flow

priapism was responsible for the erection. Thus, further mechanistic studies are needed to elucidate the induction of cavernosal smooth muscle relaxation by high doses of sildenafil.

We propose that the effects of intracavernosal sildenafil may be related to its structural similarity with cGMP. It is reasonable to suggest that at high doses, sildenafil may be interacting as a cGMP analogue with proteins (other than PDE 5) that are regulated by cGMP (eg, protein kinase G). Whereas the NO/cGMP pathway is considered to be one of the major physiological pathways modulating erectile function, the role of other pathways cannot be ruled out. Pathways involving G-protein coupled receptor agonists such as PGE (Caverject and Edex) or adenylate cyclase activators (forskolin) mediate smooth muscle relaxation and cause erection via the cAMP pathway. Furthermore, intracavernosal papaverine, a nonspecific, low-affinity PDE inhibitor, causes smooth muscle relaxation and penile erection, presumably by a mechanism independent of the cyclic nucleotide pathway (Kim et al, unpublished observations). It is possible that intracavernosal sildenafil at high concentrations is acting in a similar manner. This study demonstrates that the biological activity of sildenafil may be multifaceted and other pathways must be considered to account for its clinical effect on erectile tissue. Because intracavernosal sildenafil did not adversely affect SAP in this animal model, sildenafil is useful as a intracavernosal agent for the treatment of male erectile dysfunction.

The following references are incorporated herein by reference.

References: Ballard et al., Effects of sildenafil on the relaxation of human corpus cavernosum tissue in vito and on the activities of cyclic nucleotide phosphodiesterase isozymes,"J Urol. 1998: 159: 2164-2171.

Carter et al.,"Effect of the selective phosphodiesterase type 5 inhibitor sildenafil on erectile function in the anesthetized dog,"J. Urol. 1998: 160: 242-246.

Center for Drug Evaluation and Research. Viagra (Sildenafil)."Joint Clinical Review"for NDA-20-895, Appendix A23, page 150through Appendix A33.6; page 169, Table 123 "Pharmacokinetic parameters for Sildenafil and UK-103320 (Study 148-228)."FDA web site (http://www. fda. gov/cder/consumerinfo/viagra/default. htm): 1998.

Chuang et al.,"Sildenafil, a type-5 cGMP phosphodiesterase inhibitor, specifically amplifies endogenous cGMP-dependent relaxation in rabbit corpus cavernosum smooth muscle in vitro,"J : Urol. 1998; 160: 257-261.

De Mey,"Opportunities for the treatment of erectile dysfunction by the modulation of the NO axis-alternatives to sildenafil citrate,"Curr. Med.. Res. Opin., 14: 187-202.

Goldstein et al.,"Oral sildenafil in the treatment of erectile dysfunction. Sildenafil study group. N. Engl. J Med. 1998; 338 : 1397-1404.

Jackson et al.,"Influence of dimethyl sulfoxide (Me2SO) on vascular smooth muscle,"Arch Intern Pharmacodyn Ther. 1979; 237: 4-15.

Jeremy et al.,"Effect of sildenafil, a type-5 cGMP phosphodiesterase inhibitor, a papaverine on cyclic AMP levels in the rabbit corpus cavernosum in vitro,"Br. J from 1997; 79: 958- 963.

Kim et al.,"A nitric oxide-like factor mediates nonadrenergic-noncholinergic relaxation of penile corpus cavernosum smooth muscle,"J : Clin. Invest., 1991 ; 88 : 112-118.

Kim et al.,""Alpha-adrenergic receptor blockade by phentolamine increases the efficacy of vasodilators in penile corpus cavemosum,"Int J Impotency Res. 2000; 12: S26-S30.

Kim et al.,"Oxygen tension regulates the nitric oxide pathway. Physiological role in penile erection,"J Clin. Invest. 1993; 91: 437-442.

Lincoln et al.,"Pleiotropic regulation of vascular smooth muscle tone by cyclic GMP- dependent protein kinase,"Hypertension, 1994; 23: 1141-1147.

Maurice et al.,"Molecular basis of the synergistic inhibition of platelet function by nitrovasodilators and activators of adenylate cyclase: inhibition of cyclic AMP breakdown by cyclic GMP."Mol Pharmacol. 1990; 37: 671-681.

Moreland et al."Sildenafil, a novel inhibitor of phosphodiesterase type 5 in human corpus cavernosum smooth muscle cells,"Life Sci. 1998; 62: PL309-PL318.

Moreland et al.,"Sidenafil citrate, a selective phosphodiesterase type 5 inhibitor: research and clinical implications in erectile dysfuction,"Trends Endocrinol. Metab., 1999; 10: 97- 104.

Moro et al.,"cGMP mediates the vascular and platelet actions of nitric oxide: confirmation using an inhibitor of the soluble guanylyl,"Proc Natl Acad Sci USA, 1996; 93: 1480-1485.

Pelligrino et al.,"Cyclic nucleotide crosstalk and the regulation of cerebral vasodilation," ProgNeuorbiol, 1998; 56: 1-18.

Pyne et al.,"cGMP signal termination,"Biochem Soc Trans., 1996; 24: 1019-1022.

Stief et al.,"The effect of the specific phosphodiesterase (PDE) inhibitors on human and rabbit cavernous tissue in vitro and in vivo,"J UroL, 1998; 159: 1390-1393.

Qiu et al.,"Rabbit corpus cavernosum smooth muscle shows a different phosphodiesterase profile than human corpus cavernosum,"J : Urol., 2000; 164: 882-886.