Acetylcholine (Ach), l-NMA, l-arginine, bradykinin, arachidonic acid, indometacin, reduced nicotinamide dinucleotide (NADH), reduced nicotinamide dinucleotide phosphate (NADPH), diethyldithiocarbamate (DDTC), catalase and SOD were purchased from Sigma (Saint-Quentin-Fallavier, France). H₂O₂ was from Merck (Noisy, France). Diphenyl-iodonium (DPI) was from Calbiochem (France).

A microelectrode was built using a microcarbon fibre (8–12-μM diameter, Le Carbone Lorraine, France) inserted in a glass microelectrode and stretched in a microforge according to previous works [20]. It was commercially available (MFC1, Tacussel, Lyon). The reference and auxiliary electrodes were respectively Ag/AgCl and platinum. All measurements were performed with three electrode systems: one apparatus (PRG5, Tacussel, France), which was automated with relay cards and AD/DA cards, inserted in a microcomputer or an integrated apparatus (Autolab, PGSTAT 10, Roucaire, France). These two apparatus gave identical results. RDPV was performed with the following parameters: pulse height = 0.1 V, pulse delay = 0.1 s, potential swept = 0.04 V s⁻¹. The starting potential was −0.5V vs. Ag/AgCl. The swept were made in reduction. No Faraday cage was needed for measurements according to nA current range. Standardizations for H₂O₂ was realized in vitro in phosphate buffered saline (PBS) at 20 °C at oxygen atmospheric pressure at −1.02V vs. Ag/AgCl (range 1–200 μM). With these microelectrodes, the sensitivity limit was μM for H₂O₂. Other compounds can be detected in reduction. For NO the limit of detection was μM at −1.37V vs. Ag/AgCl. Superoxide was also detected at μM at −0.79V vs. Ag/AgCl. Oxygen was undetectable under 1 mM at −0.67V vs. Ag/AgCl. The observed signal in vivo for RDPV was the NA peak at −1.66V vs. Ag/AcCl, as previously described [1].

Male adult rats (Sprague Dawley) were 432–544 g (Janvier, Le Genest-Saint-Isle, France). They were anesthetized with urethane (1.45 g kg⁻¹, ip) and were on a homemade heating table during measurements. The penis was exposed and a small incision (1–2 mm) was realized. Then the voltammetric microelectrode and a micro glass-device for intracavernous injection of pharmacological agents were introduced with micromanipulators (Prior, Phymep, France) in the corpus cavernosum near the voltammetric microelectrode under binocular observation (Zeiss, France). Each drug was dissolved in PBS at 1 mM (37 °C); 40–50 μl were injected each time. DPI was dissolved in dimethylsulphoxide (DMSO). Intracavernosal injection of a compound was made at least in three different animals. Vasodilatation of penis was monitored using laser Doppler flowmetry. Briefly, a microprobe was put on the penis near the microelectrode (PF418 master probes and Periflux 4001 master, Perimed, France). It gave on line the blood flow in flow units. In a first series (n=10 rats) RDPV was realised every minute with simultaneous laser Doppler flowmetry after intracavernous injection of acetylcholine (50 μl, 100 μM); Then 30 min later, a mixture of acetylcholine plus l-NMA was injected ic (50 μl Ach 100 μM, l-NMA 1 mM). In a second series (n=36 rats), RDPV was realised every minutes after intracavernous injection of the same mixture of acetylcholine and l-NMA; then at 30 min a compound (n=12) was injected ic. In each series RDPV was monitored 2 h in vivo. Each rat received one compound and this compound was tested on three rats.

RDPV was realized on line in rat corpus cavernosum. The microelectrode tip was located in the extra vascular space in contact with the endothelial cells of the sinus. The microelectrode detected compounds at the contact of its active tip. When the microelectrode tip was removed of the cell surface contact the signal immediately disappeared. Measurements corresponded to punctual concentrations at the interface of the microelectrode tip and endothelial cells membranes. A substance was detected when its concentration was superior to the sensitivity limit of RDPV for this molecule. Before penis-induced vasodilatation, no signals were observed in RDPV, indicating either an absence of electroactive compounds or an insufficient concentration of electroactive products. This is highly favourable for the study of vasodilatation by voltammetry. Simultaneous injection of acetylcholine and l-NMA produced vasodilatation followed by the laser Doppler apparatus and appearance of a new peak. Its precise potential localization permitted to attribute this signal to H₂O₂. It was observed that molecules had the same peak potential either in vitro or in vivo. According to the fact that only one endogen molecule, NA, have been observed in vivo as electroactive in reduction on carbon microelectrode [18–20], the identification of H₂O₂ is quasi certain. Injection of H₂O₂ near the microelectrode tip confirmed this identity. The fact that injected H₂O₂ induced vasodilatation was another important point. The third proof was found by the injection of catalase, which was specific for H₂O₂. Catalase gave an immediate disappearance of the endogenous peak. Microelectrodes were easily standardized in vitro for H₂O₂ in PBS without oxygen interferences according to an endogen level below the sensitivity limit. The observed concentrations in rat corpus cavernosum (0.115 μM) during vasodilatation were close to the concentrations observed in rat brain for H₂O₂ using microdialysis [24,25]. The stability, the repeatability, and the high current values of these peaks with always the same potential permitted to follow and quantify H₂O₂ concentrations in the extracellular space without any problems.

The origin of H₂O₂ was studied by injection of pharmacological products. H₂O₂ production in vivo was well known and different enzymatic origins could be responsible of its secretion. Xanthine oxidase, eNOS, P450 enzymes and NAD(P)H oxidases were candidates to synthesize H₂O₂ in vivo in various cells. In endothelial cells eNOS and NAD(P)H oxidases have been described to produce H₂O₂. Precedent work on eNOS demonstrated that O2•− was not produced at a detectable level by voltammetry in pure enzymatic preparation, even in the absence of the substrate. The enzyme, eNOS, was not the source of extracellular release of H₂O₂ in this system, as observed in endothelial cells [1,7]. Intracavernous injection of DDTC, a SOD inhibitor, gave H₂O₂ peak disappearance along with the appearance of a new peak. This new peak corresponded to O2•−, which was identified by its peak potential. This was the first time that O2•− was observed in vivo by RDPV. This was due to the high concentration value of H₂O₂ in vivo according to the equimolar transformation of O2•− into H₂O₂ by SOD. SOD injection had no effect on H₂O₂ peak in vivo. These facts could be explained by the fact that O2•− was produced inside the cells and that H₂O₂ was only present in the extracellular fluid where is located the microelectrode [26]. Another point was the presence of extracellular SOD, which immediately scavenged O2•− to H₂O₂ without needing an exogenous source of SOD. DDTC experiments showed that O2•− was the source of H₂O₂. DDTC can penetrate into endothelial cells. It is not the case for injected SOD. Substrates and inhibitor of NAD(P)H oxidases were tested. DPI, a specific NAD(P)H oxidases inhibitor, inhibited immediately this production. This indicated an enzymatic synthesis for H₂O₂ by NAD(P)H oxidases. NADH and NADPH injections enhanced the H₂O₂ peak, indicating NAD(P)H oxidases system for H₂O₂ production. The vasodilatation measured by laser Doppler flowmetry supported the idea that H₂O₂ was a vasodilatator in the corpus cavernosum bed. This was observed in other rat vascular tissues. H₂O₂ was found in pulmonary arterial walls where it is secreted by endothelial cells using NAD(P)H oxidases localized in the cells membrane [27,28]. It was observed in calf or bovine pulmonary arteries a cyclic guanosine monophosphate (cGMP)-mediated relaxation occurring through an increased intracellular H₂O₂ generation. It was also observed in rat aorta [29]. NA was not the sole agent of penile vasodilatation, H₂O₂ is another mediator of vasodilatation using cGMP pathway. Two molecules of the prostanoid system were tested. Indometacin and arachidonic acid had no effects on the H₂O₂ peak, eliminating the prostanoid pathway for its production.

In normal physiological conditions the l-arginine pathway was the sole used for vasodilatation in rat corpus cavernosum. When this system was blocked, another system using endothelial NAD(P)H oxidases was used with production of H₂O₂, a potent physiological vasodilatator. Intracavernous injection of l-arginine made H₂O₂ disappear with simultaneous apparition of NA, the product of eNOS. The presence of l-arginine imposed the synthesis of NA. NA is the mediator of vasodilatation in normal conditions [1]. There was a switch mechanism between the two systems depending of the availability of eNOS to synthesize NA. An absence of l-arginine or the inhibition of eNOS switched the eNOS system to the NAD(P)H oxidases system. The ability of voltammetry to monitor NA, H₂O₂ and O2•− was very useful for continuous on line measurements in tissue. In mice, H₂O₂ was presented as EDHF and it could be the case in rat corpus cavernosum. NA was presented as EDRF in the same tissue; if true, voltammetry could monitor simultaneously EDRF and EDHF in a living animal. Rat corpus cavernosum was a very useful model for such studies.