4.1. Oxidation by Chemicals and Biomimetics
Since the NO production in the human body originates from the oxidation of
l-arginine by NOS, this oxidation route and the factors affecting the oxidation were investigated. In 1998, Koikov et al. studied the oxidation of three series of compounds including oximes and amidoximes at a concentration of 2 × 10
−4 M with K
3Fe[(CN)
6] at pH 12 [
57]. Under these conditions, oximes were not able of releasing NO and this phenomenon was explained by the acceptor or donor properties of the substituents (hydroxyphenyl and pyridine rings) and by the acidity of the oxime functions. However, many tested amidoximes showed the capacity to release NO in the presence of K
3Fe[(CN)
6] and methyl and phenyl amidoximes released 25% and 10% of NO, respectively. It was observed that replacing the phenyl ring by a pyridine ring allows to increase twice the amount of released NO and that the amidoximes exhibit quite a similar potential of NO release than methylamidoxime. The structures of these amidoximes are shown in
Figure 6. The most important release was observed for the pyridine-2,6-diamidoxime for which the amount of released NO reached up to 40%. Benzonitrile was also identified as the major product present after the oxidation of benzamidoxime.
To better understand the products formed during the oxidation of amidoximes, Vadon-Le Goff et al. reported the use of various oxidants and biomimetic systems to study the reaction products as well as the causes and factors affecting the outcome of the reaction [
4]. The authors evaluated the potential of NO release of 4-chlorobenzamidoxime in the presence of different oxidants. Amides or nitriles were identified as the main products but the presence of a dimeric product in the mixture was also detected (
Figure 7).
It was demonstrated that with oxidants like Pb(OAc)
4 and Ag
2CO
3, amidoximes are selectively oxidized into nitriles, while amides were selectively formed using oxidants capable of transferring one oxygen atom like H
2O
2,
t-BuOOH or
m-CPBA. Using
m-CPBA, 4-chlorobenzamidoxime was oxidized into a mixture of dimeric products and amide, the latter being the major component. This oxidant, as well as
t-BuOOH, associated with catalytic amounts of iron-porphyrin generates preferentially the nitrile in less than 1 h (ca. 80% of the amidoxime oxidized) (
Table 5, lines 1–3). This likely originates from the formation of highly reactive iron-oxo species which oxidize the amidoxime into the nitrile. This result is in accordance with those described in the literature related to the oxidation of amidoximes by hemoproteins. Horseradish peroxidase, able of generating iron-oxo species in the presence of H
2O
2, oxidizes amidoximes to nitriles and a dimeric product. On the other hand, the oxidation by CYP450 generates exclusively the amide product instead of the nitrile. This was explained by the presence of the superoxide O
2•
− radical anion or of a protein-metal-O
2 complex instead of the presence of the iron-oxo intermediates (see
Section 4.2 for the oxidation by CYP450).
Amidoximes and oximes can also be oxidized via the amidoximate/oximate anions in the presence of singlet oxygen [
58]. The authors demonstrate that amidoximes and oximes are inert to
1O
2 but once converted to amidoximates and under photooxygenation, amidoximes can be oxidized into the corresponding amides along with nitriles as minor products (
Table 5, entry 4). The use of
1O
2 is similar to the oxidation of
N-hydroxyguanidines by the NOS and clarifies the biological oxidation routes (
Figure 8, red part).
To better understand the oxidation process of amidoximes, oxidants like 2-iodobenzoic acid (IBX) and the IBX/tetraethylammonium bromide (TEAB) combination were also used [
12]. The authors demonstrate that IBX can selectively oxidize benzamidoxime (
Figure 8, blue part) into amide (best yield of 83% using an amidoxime:IBX ratio of 1:1 and only 10% of the nitrile). This method was extended to various amidoximes and similar results were obtained. Nevertheless, the use of an amidoxime:IBX:TEAB molar ratio of 1:2:2 gave 90% of the nitrile and only 5% of the amide. By using this ratio with a series of amidoximes, similar results were obtained by the authors (
Table 5, entries 5 and 6).
4.2. Biological Pathways of Amidoximes and Oximes Oxidation
As it is well known, NO is synthesized in vivo by the NOS enzymes. In recent years, many scientists got interested in finding new ways to oxidize amidoximes and oximes in order to increase the NO concentration in the organism. A few studies demonstrated that C=N-OH moieties present in compounds like amidoximes and oximes can be oxidized by CYP450 enzymes [
59,
60]. In 1998, a study showed the key role of CYP450 in the oxidation of amidoximes, oximes and
N-hydroxyguanidines along with their oxidation products [
5]. In the presence of rat liver microsomes, NADPH and O
2, amidoximes and oximes can be oxidized and release NO and NO-related aerobic products like NO
2– and NO
3−. The presence of both NADPH and O
2 and of the active microsomes was demonstrated to be essential for the oxidation, indicating that this reaction is enzymatic. Moreover, the authors demonstrated that the CYP450 present in the microsomes are responsible for this reaction by using either the CYP450 inhibitor Miconazole which caused the inhibition by ca. 80–90% of the reaction or Dexomethasone (DEX), a CYP450 inducer, treated microsomes. The oxidation of various amidoximes and oximes generates both NO
2− and NO
3−, especially in the presence of DEX-treated microsomes which boosts the oxidation. This release was always accompanied by the formation of amides and nitriles in the case of amidoximes and of ketones and nitroalkanes in the case of ketoximes.
In 2004, another study used CYP450 to test the oxidation of the oximes. Hence, Mäntylä et al. incubated buparvaquone–oxime and its derivatives with NADPH and either untreated rat liver microsomes or treated ones with various CYP450 inducers. These buparvaquone’s prodrugs were studied for their ability to be oxidized and releasing the corresponding buparvaquone alongside NO which was quantified via the NO
2- species produced [
61]. All of the oximes were able to be oxidized in the presence of CYP450 and release buparvaquone bearing a C=O function effective against leishmaniasis. The presence of NO was also believed to have a role against this disease. The solubility of the oximes prodrugs was further improved by the synthesis of their corresponding phosphate prodrugs. These latter were able to release the corresponding oximes-buparvaquone [
62].
Recently, we evaluated the capacity of mono- and bis-amidoximes to release NO by a CYP450-mediated oxidation (
Figure 9) [
38]. Amidoximes were also incubated with NADPH and untreated microsomes. Only aromatic mono-amidoximes showed an important NO release while the aliphatic mono-amidoxime and the bis-amidoxime bearing both an aromatic and aliphatic amidoxime released only small NO quantities. The inadequate size or shape of the latter compounds and/or the lower reactivity of the aliphatic amidoximes compared to aromatic ones hinders their metabolization by rat liver microsomes. It was of interest to determine if these amidoximes are able to release NO once incubated with human cells not originating from the liver. For that purpose, these compounds were first demonstrated to be cytocompatible with human vascular smooth muscle cells (HVSMC) and were later incubated with these cells for 1 h to evaluate their ability to release NO. This test showed that all of the amidoximes are able to release NO and to increase the NO storage through nitrosothiol (RSNO) formation inside the cells. It is worth mentioning that the bis-amidoxime exhibited a high potential for the generation of RSNO, very close to that of the aromatic mono-amidoximes, but at a twice lower concentration.
Other reports focused on finding the capacity of the NOS in oxidizing amidoximes and oximes. Acetoxime can be oxidized by CYP450 and NADPH in the presence of metal complexes into NO
2– and reactive NO species [
63]. However, this study clearly proved that acetoxime cannot be oxidized by the NOS II enzyme and cannot even inhibit its activity, which suggests a low affinity between NOS and this ketoxime. This demonstrates that the oxidation pathway of acetoxime in the cells involves only CYP450. In another study, four amidoximes were incubated with NOS I and II and NADPH and the NO formation was monitored by the Griess assay in order to study the amidoximes recognition by NOS [
64]. All the amidoximes were inactive in the presence of NOS and did not show any NO release. Moreover, they are also very bad inhibitors of NOS I and II. The corresponding
N-hydroxyguanidines analogs showed high affinity, indicating that the NOS better recognizes molecules bearing the –NH-C(R)=NH moiety than C=NOH. The latest study related to that topic was conducted in 2002 [
65], and a series of
N-hydroxyguanidines, amidoximes, ketoximes and aldoximes where tested for their NO release capacity in the presence of recombinant NOS II. The presence of an unmodified
N-hydroxyguanidine function is mandatory for the oxidation by the NOS. Thus, none of the amidoximes and oximes could be oxidized by this enzyme.
As shown by these studies, CYP450 seems to be responsible of the oxidation of both amidoximes and oximes while NOS does not appear to have a role in these oxidations. As mentioned above, the oxidations by CYP450 afforded as major products amides in the case of amidoximes and ketones in the case of oximes alongside with the release of NO, nitrites and nitrates [
63,
64,
65]. The mechanism of amidoximes and oximes oxidation by CYP450 that generates amides/ketones is explained by the formation of the superoxide radical anion (O
2•
−) which is a dissociation product of the CYP450-Fe(II)-O
2 complex, this latter was shown to be easily transformed to CYP420-Fe(II)-O
2 during the oxidation [
60,
63,
64,
65]. The formation, even in small amounts, of a nitrile and nitroalkane mixture originates from the presence of an iron-oxo complex as it is the case for the biomimetic reactions generating nitriles. Many CYP450 isoforms have been identified as responsible of the oxidations of these products like CYP4501A
1, CYP4502B
1, and CYP4502E
1 [
5,
63].
These results demonstrate that NOS are not optimal for the evaluation of the NO release from amidoximes. It is preferable to use the CYP450 pathway that shows good results with both amidoxime and oxime functions.
4.3. In Vivo and In Vitro Biological Responses to NO Release from Amidoximes/Oximes Oxidation
Along with the discovery of the capacity of amidoximes and oximes to release NO in the presence of biological extracts and cells, many scientists started to test these products on biological tissues. Indeed, the in vivo generation of NO has gained a lot of attention among others due its capacity of vasodilatation and thus reducing platelet aggregation and thrombosis formation.
The first major works that appeared on that topic were published by Rehse et al. in 1997 and 1998 [
66,
67]. Seven aryl azoamidoximes and their effects on thrombosis inhibition, blood pressure and inhibition of platelet aggregation were investigated [
66]. Five of the seven amidoximes inhibited the formation of thrombus in mesenteric rat vessels and even two of these amidoximes inhibited the thrombus formation by more than 20% (
Figure 10). The remaining two amidoximes bearing a methoxy or a fluoro group on the
para position did not show any significant effect on the thrombus formation in both arterioles and venules. The inactive 4-methoxyphenylazo-methanamidoxime had also no significant action in lowering the blood pressure during 2, 4, 6, and 24 h. The highly active 4-chlorophenylazo-methanamidoxime decreased the blood pressure for 24 h with a maximum efficiency of 16% ± 10% at 6 h. It is noteworthy that both amidoximes were able to release NO in the presence of DEX-treated rat liver microsomes and NADPH, but only the second amidoxime was used by the cells in order to lower blood pressure and reduce thrombi formation. The last test conducted with these amidoximes was the platelet aggregation. All of the studied amidoximes exhibited low or poor activity after calculating their IC
50 following the Born test. This means that these amidoximes are not able of being oxidized in these conditions.
In another series of experiments, 17 compounds containing different aliphatic, aromatic, and bis-amidoximes were tested [
67]. The results were in accordance with the previous study since the majority of the amidoximes showed a poor effect or even a lack of activity towards the inhibition of platelet aggregation which proves once more that the platelet rich plasma is not convenient for amidoximes oxidation. The most effective amidoxime was the aromatic one bearing a chlorine atom on the para position and an ethene group between the phenyl and the amidoxime function. The presence of this ethene group raised the efficiency for the platelet aggregation inhibition. Bis-amidoximes displayed a good activity for the inhibition of thrombus formation in arterioles and venules and the bis-amidoxime bearing an ethene group exhibited the highest percentage of thrombus inhibition. All other amidoximes also show some activity and it was noticed that adding a group on the para position may increase the thrombi formation inhibition, the key parameter being the lipophilicity rather than the electronic variations. The amidoximes exhibiting the highest activity in reducing the thrombus formation were also tested for decreasing the blood pressure but their activity was weak (the highest effect was observed for the bis-amidoxime with a blood pressure lowering of only 5%) (
Figure 10).
These studies were not limited to amidoximes but also to azide oximes. These compounds showed better antiplatelet effect than the azide amidoximes and the amidoximes since the majority of the studied azide oximes exhibited a capacity to inhibit platelet aggregation with the oxime bearing the nitrophenyl function having an IC
50 of 2 µmol/L in the Born test [
68]. Similarly to amidoximes, the electronic variance and the lipophilicity had no major role in changing the efficiency of the oximes, however strong electron withdrawing groups like a nitro function seem to increase the oxime activity. All of the oximes showed also 10–20% efficiency in thrombus formation inhibition and two of them exhibited values higher than 20%. Additionally, azide oximes allowed a decrease of the blood pressure that may reach 10–15%. This parameter is not connected to the thrombus formation inhibition since the compound exhibiting the highest antithrombotic activity did not have the highest capacity of blood pressure lowering. Overall, the azide oximes showed to be more effective than the azide amidoximes and other studied amidoximes (
Figure 10).
In another study and in order to evaluate the relaxation capacity of amidoximes, Jia et al. used formamidoxime in the presence of a tracheal ring previously contracted by carbachol, a cholinomimetic drug. They observed a relaxation effect on the tracheal ring after the incubation along with an accumulation of the cyclic guanosine 3′,5′-monophosphate (cGMP) level after incubation with the tracheal smooth muscle cells [
69]. The relation between the production of cGMP and the relaxation of the ring was proved by using a cGMP inhibitor on both the tracheal rings and the smooth muscle cells. This inhibitor prevented the relaxation of the rings and the formation of the cGMP in both cultures. After the detection of NO in the incubation media of the tracheal smooth muscle cells with formamidoxime, the authors proposed that after the amidoxime oxidation, the released NO activates the formation of the cGMP which induces the relaxation of the tracheal ring. As previously, a NOS inhibitor was used with the cultured cells and the tracheal rings but it did not inhibit the production of cGMP or the relaxation of the rings. On the contrary, the cGMP production and the ring relaxation were both inhibited after the use of a CYP450 inhibitor, 7-ethoxyresorufin. Similarly, the cGMP production was inhibited after the use of Miconazole on the cell cultures. These results provided evidence that the pathway of the amidoxime oxidation in the tracheal smooth muscle cells and tracheal ring is through CYP450 and not NOS, and more specifically by CYP4501A
1 being a strong substrate of the 7-ethoxyresorufin.
This study using formamidoxime got a lot of attention and other researches were conducted to see if this compound, along with other amidoximes and oximes, exhibited the same effects in the rat aorta. Vetrovsky et al. tested a series of amidoximes and one oxime in the presence of endothelium-denuded rat aorta (
Figure 11) [
70]. All of the compounds caused an endothelium-independent relaxation higher than the NOHA with 4-chlorobenzamidoxime being the most active. It should also be noted that formamidoxime (
Table 6, entry 1, in vitro studies) caused an important rat aorta relaxation close to that of the 4-chlorobenzamidoxime. The authors demonstrate that the presence of electron donating or withdrawing groups has a limited influence on the aorta relaxation since benzamidoxime, 4-methoxybenzamidoxime, and 4-
n-(hexyloxy)benzamidoxime were slightly less potent than 4-chlorobenzamidoxime and 4-nitrobenzamidoxime. Additionally, the lipophilicity increase in 4-
n-(hexyloxy)benzamidoxime compared to 4-methoxybenzamidoxime did not play a key role on the relaxation effect. Finally, the use of an oxime instead of an amidoxime function caused a decrease in the relaxation capacity.
Using the same tests but focusing on oximes like acetaldoxime, acetoxime and formaldoxime (
Table 6, entries 2–4, in vitro studies), it was shown that all of these oximes are more potent to induce relaxation in rat aorta than hydroxyguanidine, formaldoxime being the most potent [
71]. Similar to amidoximes, the lipophilicity increase from acetaldoxime to acetoxime does not significantly affect the induced relaxation. For the previously studied amidoximes and oximes, cGMP and NO were involved in the aorta relaxation since the addition of the guanylate cyclase inhibitor and NO-scavengers completely inhibited the relaxation. The use of NOS inhibitors and CYP450 inhibitors like proadifen did not alter the relaxation effect, indicating that other oxidation pathways are involved with these amidoximes and oximes. However, the use of the 7-ethoxyresorufin inhibited the relaxation of the aorta rings in the presence of 4-chlorobenzamidoxime and oximes but the authors suggested that this is not due to an inhibition of the CYP4501A
1 but to an inhibition of a different NADPH-dependent reductase pathway. In a following study [
6], the aliphatic oximes and formamidoxime (
Table 6) were tested in vivo to evaluate the blood pressure decrease in conscious chronically cannulated rats in which the endogenous NO synthesis was blocked. In the precedent in vitro studies, the three tested oximes and the formamidoxime caused an aorta relaxation, contrariwise, in vivo only formamidoxime and formaldoxime were capable of lowering the blood pressure, with the amidoxime being the most active. This result was different from the in vitro experiments where the formaldoxime was the most potent. The inhibition of the guanylate cyclase in vivo caused the inhibition of the blood pressure reduction when using formaldoxime, which indicates that NO is responsible for the whole effect while with formamidoxime NO is responsible for one third of the effect, suggesting that with the amidoxime not the whole blood pressure decrease is caused by the NO release. Finally, hydrophilic substances were found to be more active in vivo than in vitro.
More recently, other studies evaluated the potency of some oximes to induce a vasorelaxation in superior mesenteric arteries isolated from rats. It was demonstrated that
E-cinnamaldehyde oxime was the most potent and able of causing a NOS and endothelium-independent vasorelaxation. The NO release and relaxation activity of this compound were only partially inhibited by 7-ethoxyresorufin, a CYP4501A
1 specific inhibitor, but not by other CYP450 nonspecific inhibitor, which shows also that the oxime oxidation is catalysed by a NADPH-dependent reductase pathway in the superior mesenteric arteries just like the studies in rat aortic rings. The relaxation pathway is also caused by the activation of cGMP with a key role played by a type of K
+ channel and the reduction of the Ca
2+ influx [
72]. The role of the K
+ channels was demonstrated in another study published in 2014 after the incubation of an oxime with rat aortic rings in the presence of different channel blockers [
73]. The relaxation was observed when the enzyme and the aortic rings were incubated without the channel blockers. Furthermore, when administered to conscious rats, the oxime caused the reduction of blood pressure. The aorta and superior mesenteric artery rings relaxations were similar in the presence and in the absence of the endothelium after the incubation with the oxime, indicating that NOS were not responsible of the oxidation. The relaxation, as for the previously studied compounds, is achieved by the release of NO and the activation of the cGMP pathway alongside the activation of the K
+ channels.
In addition, amidoximes can also be used for their capacity of lowering the intraocular pressure (IOP) in rabbits due to the NO release followed by the formation of cGMP [
74,
75]. In 2006, Oresmaa et al. tested various imidazole amidoximes for their IOP-lowering abilities. Two of them (
Figure 12) were able to produce NO and increase the cGMP formation when incubated with porcine iris-ciliary bodies but only the methyl form was able to reduce the IOP when administered intravitreally in rabbits. The esterification of the active amidoximes caused the loss of their biological activity.
All of these studies demonstrate that when amidoximes or oximes are incubated with cultured cells such as tracheal smooth muscle cells [
69] or with biological tissues such as aortic/tracheal rings [
69,
70], the generation of NO followed by the relaxation of the tissues is observed. This relaxation was proved to be caused by an increase in the cGMP accumulation after the activation of the guanylate cyclase by the released NO. Moreover, even in biological tissues, the NOS pathway does not seem to play a role in the oxidation of these products [
69,
70,
71]. However, in some biological tissues like aortic rings and superior mesenteric arteries, CYP450 did not seem to play a role in the oxidation but it was suggested to be another NADPH-dependent reductase pathway [
70,
71,
72]. Independently from the enzyme involved in the oxidation of amidoximes and oximes, it is now well established that the related NO release causes an activation of the guanylate cyclase with an accumulation of cGMP. This is followed by an activation of the BK
Ca channels which causes an efflux of the potassium ions and thus a decrease in the calcium channels activity [
72,
73]. This pathway causes a relaxation in the targeted tissue along the detected effects such as blood pressure decrease.