Nitric Oxide Releasing-dendrimers : an Overview

Douglas Wagner Franco , Ph.D. in analyt ical chemistry in 1972 at Universidade de São Paulo (USP), Brazil. Postdoctoral research at: 1. Stanford University USA, Prof. Henry Taube (1976); 2. University of California at Santa Barbara USA, Prof. Peter C. Ford (1982); 3. Institute de Recherches sur la Catalyse – France, Prof. Igor Katchenko (1984); 4. Università degli Studi di Parma – Italia, Prof. J. Paolo Chiusoli (1985); 5. York University – Canada, Prof. Alfred B. P. Lever (1990). Full Professor at University of São Paulo since 1986. Current research interest addresses the chemistry of ruthenium nitrosyl complexes, and the chemistry of the Brazilian sugar cane spirit (cachaça),

Due to the properties outlined above, an exogenous NO deliver is a potential therapeutic agent, inasmuch deficiencies in NO biosynthesis are related to some diseases (Baylis, 2008;Luiking, Engelen, Deutz, 2010).NO can be administered in gaseous form (respiratory route) in specific cases (Serafim et al., 2012), but this compound is known to react with a wide number of molecules (proteins, thiols, heme proteins, metals and others) and also can be oxidized in solution to NO 2 by O 2 (Fukuto et al., 2012;Toledo, Augusto, 2012).In addition, the NO lifetime in biological media varies from 10 ms to 1 sec (Thomas et al., 2001;Flores-Santana et al., 2009).Furthermore local lack or excess of NO can be deleterious for biological systems.Therefore, strategies to produce stable NO carriers systems have been developed, being much of the work focused on the synthesis of nitric oxide donors (NO-donors) (Wang et al., 2002;Serafim et al., 2012) or scavengers (Fricker et al., 1997;Davies et al., 1997).
Most of synthesized NO-releasing compounds have been used to help in understanding the physiologic aspects related to NO and to describe its therapeutic properties.
In this context, next topics deal specifically with functionalized dendrimers, tailored for NO delivery and their possible applications.

Dendrimers as a platform for NO transport
Dendrimers are molecules of the dendritic materials group, which is also composed by dendronized and hyperbranched polymers, dendrons and dandrigrafts (Carlmark et al., 2009;Astruc, Boisselier, Ornelas, 2010).The main difference between dendrimers and dendrons regarding to the other dendritic polymers is related to their shape: dendrimers and dendrons are perfectly branched (Figure 2).
The concept of dendrimer-like compound was first described by Vögtle and coworkers (Buhleier, Wehner,   Vögtle, 1978), who synthesized polypropyleneimine (PPI), employing branching trough repetitive growth approach.The first dendrimer synthesis was carried out independently by two groups of researchers: Tomalia and coworkers (1985) and Newkome and coworkers (1985), whose "baptize" their molecules as "starburst polymer" and "arborol", respectively.Dendrimers are generally described as hyperbranched molecules, with a well defined structure, possessing three main architectural components (Figure 2B): core, interior layers (radially branched) which define dendrimers generations and end-groups (surface or exterior) (Tomalia et al., 1985).
Taite and West ( 2006) have synthesized a poly(ethyleneglycol)-lysine (PEG-lys) dendrimer in which multiple diazeniumdiolate NO-donors were generated by  the reaction of PEG-lys with NO gas (Table I, compound 10).This reaction converted ~83 ± 9% of amines in diazeniumdiolate NO-donors.Benini, McGarvey and Franco (2008) have functionalized PAMAM dendrimers of generation 0, 2 and 3 with the complex K[Ru(edta)Cl].The functionalization was performed through a peptide bond between one of the carboxylate arms of the edta ligand and the superficial amines of dendrimers.This anchored chloro complex in aqueous solution yields the respective aquo species [Ru(edta)H 2 O] -attached to the dendrimer surface.The reaction of this system with NO gas (6 h) yielded the NO analogous species [Ru(edta)NO] -, which remain anchored on PAMAM (Figure 5 and Table I, compounds 11-13).The percentage of superficial groups of PAMAM functionalized with ruthenium complexes were 100% for G0, 75% for G2 and 90.6% for G3 (Benini, McGarvey, Franco, 2008).According to the reported data (Benini, McGarvey, Franco, 2008), compound 13 was able to storage ~1.4 µmol NO mg -1 .
Following a similar approach, Sun et al. (2012) functionalized PPI dendrimers of generation 2 and 5 with PEG, PO or SO, and then reacted the products with NO gas (10 atm, in sodium methoxide solution in methanol) to obtain compounds of similar structure of 18 and 21 for PEG, 22 and 25 for PO and 26 and 29 for the SO PPI functionalized dendrimer (Figure 6).The NO storage capacity of these compounds was in the range of 1.8-3.8µmol NO mg -1 and the conversion efficiency of amines into diazeniumdiolates, between 19-35% (Sun et al., 2012).Antibacterial activity of these compounds was tested against Grampositive and Gram-negative pathogenic bacteria (Sun et al., 2012), and it will be discussed in next sections.
Nitric oxide release in these dendrimer-diazeniumdiolates derivatives occurs through a hydrolytic degradation mechanism, initiated by protonation at the amine nitrogen (Davies et al., 2001;Hrabie, Keefer, 2002;Keefer, 2011).Compound 1 and 7 (Figure 3) were able to maintain the sustained-release of NO for more than 16 h, followed by 6 (~14h), being these values the longest observed among analogous compounds (dendrimers 2-5) and small alkyl diazeniumdiolates (Stasko, Schoenfisch, 2006).The great increase on the NO release duration time, in comparison with small alkyl diazeniumdiolates, was attributed to the dendritic effect (Stasko, Schoenfisch, 2006), which resulted in an enhancement on the diazeniumdiolate stability.This effect was also correlated with the increase in the NO release half-life (t 1/2 ) of NO-releasing dendrimers in comparison to the structural equivalent small molecules (Stasko, Schoenfisch, 2006).
In the other series of diazeniumdiolate PPI dendrimers (14-33, Figure 6), compounds 14-17 released NO for ~25 h, 18-21 for ~21 h, 22-25 for ~22 h, 26-29 for ~17 h and 30-33 for ~22 h (Lu et al., 2011).For compounds 14-33, the NO release kinetics presented a wide range of t 1/2 values, which varied from 18 to 293 min (Table 1).This broad range of values for t 1/2 was explained based on the dendrimers exterior modifications.Faster NO release (in general) observed for compounds 18-25 in comparison to others of this class (14-17 and 26-33) was attributed to the presence of PEG (18-21) or PO (22-25) groups on the dendrimers surface (Figure 6), which are hydrophilic and consequently facilitate the diazeniumdiolate degradation and the NO release (Lu et al., 2011).This explanation was consistent with the measured t 1/2 values for NO release from compounds 26-33, which have more hydrophobic surface groups at the exterior of dendrimer (SO or ED, Figure 6), thus resulting in higher t 1/2 values than the ones observed for compounds 18-25.
It is interesting to note that despite the foregoing, the longest t 1/2 values for NO release in the series of diazeniumdiolate PPI dendrimers were observed for the compounds 14-17, which have a hydrophilic cyano group on the dendrimer surface.According to the authors (Lu et al., 2011), cationic groups on the neighboring of diazeniumdiolates are responsible for the additional stabilization of this class of NO-donors, thus increasing the t 1/2 values for NO release.
Also, Taite and West ( 2006) have synthesized another diazeniumdiolate dendrimer (10) that release NO for over 60 days period under physiological conditions, with the majority of the release occurring within the first 10 days.
According to Stasko, Fischer, Schoenfisch (2008) S-nitrosothiol (RSNO) derivative dendrimers (8, 9) are able to release nitric oxide by two main pathways (Figure 8): (A) transition metal-mediated catalytic decomposition, based on Cu + /Cu 2+ redox couple and (B) photo-initiated decomposition.In the last pathway, irradiation of compounds 8 and 9 using a broad-spectrum white light resulted in homolytic cleavage of the S-N bond, generating a thiyl radical (RS • ) and also liberating NO (Stasko, Fischer, Schoenfisch, 2008).The rate constant for NO release was dependent on the structure of the nitrosothiol (SNAP or NACysNO) and on the NO release triggers (copper or light).This resulted in t 1/2 values for NO dissociation in the range of 7-106 min for 8 and 1.5-2.3min for 9 when triggered by copper (0.2, 0.6 or 1 mM of Cu 2+ ), and 49-97 min for 8 and 34-200 min for 9 when triggered by light (60, 100 or 200 W) (Stasko, Fischer, Schoenfisch, 2008).
It is interesting to highlight that the chemical properties of ruthenium nitrosyl complexes were not altered after their attachment to the PAMAM dendrimers (Benini, McGarvey, Franco, 2008).The potential for the nitrosyl reduction (NO + /NO 0 ) in the compounds 11-13 was around -0.32 V vs SCE (saturated calomel electrode), values which are very close to the ones measured for the complex [Ru(edta)NO] -(-0.34V vs SCE).These redox potentials are accessible for the reducing agents as NADH, NADPH and cysteine present in biological media, thus enabling the in vivo delivery of NO from these metal nitrosyls.Also, the rate constants for NO release (k -NO ) in compounds 11 and 12 after one-electron reduction were in the range of 1.9-2.9x 10 -3 s -1 , which is in agreement with the previously FIGURE 9 -Nitric oxide release from ruthenium nitrosyl dendrimer (Benini, McGarvey, Franco, 2008).
All together, the general features of the nitric oxide release mechanisms for all NO-donors presented above (diazeniumdiolates, S-nitrosothiol and metal nitrosyl complexes) were maintained when these compounds were attached to dendrimers.

Applications of NO-releasing dendrimers
Most of NO-releasing dendrimers discussed above were tested in different biological assays (Taite, West, 2006;Benini, McGarvey, Franco, 2008;Stasko, Fischer, Schoenfisch, 2008;Johnson et al., 2010;Sun et al., 2012).Stasko, Fischer and Schoenfisch (2008) described the ability of S-nitrosothiol derivative dendrimer 9 in the inhibition of thrombin-mediated platelet aggregation.According to the authors, the mechanism occurs via transnitrosation (Hogg, 2000;Walsh et al., 2007;Bell, Shah, Gordge, 2007;Stasko, Fischer, Schoenfisch, 2008), a process in which a nitrosothiol transfers the nitroso group to a free thiol (present in proteins on the platelet surface).Compound 9 inhibited platelet aggregation regardless of nitrosothiol (SNAP) concentration (100, 50 or 25 mM), and at 25 mM led to a reduction of 62% in the aggregation compared to 17% for 25 mM of the analogous small molecule (similar S-nitrosothiol -SNAP -not attached on dendrimer).Best results for 9 in comparison to the "free" SNAP were attributed by the authors to the increase on the S-nitrosothiols local concentration during the transnitrosation process, due to the anchoring of the dendrimer on the cell surface (Stasko, Fischer, Schoenfisch, 2008).
Taite and West ( 2006) tested the capacity of compound 10 on the inhibition of platelet adhesion to thrombogenic surfaces.PEG-lys-NO dendrimers (10) reduced platelet adhesion in approximately 81% regarding to the analogous dendrimer that did not carry NO (PEG-Lys) (Taite, West, 2006).Another assay was performed to verify the ability of compound 10 on the regulation of vascular cell proliferation (Taite, West, 2006).First, cell viability was assessed using culture of bovine aortic endothelial cells (BAECs) and rat aortic smooth cells (SMSc), which were exposed to NO-releasing dendrimers (10, PEG-Lys-NO) and to ones that do not release NO (PEG-Lys).According to the authors, the NO-releasing dendrimer 10 is non toxic for BAECs (100 ± 1% of cell viability) and SMCs (99 ± 3% of cell viability).The same test performed with PEG-Lys, resulted in cell viability of 89 ± 3% for BAECs and 87 ± 6% for SMCs.As reported (Taite, West, 2006) this result showed that even the PEG-Lys low toxicity is reduced when NO is linked to the dendrimer (PEG-Lys-NO, 10).They also demonstrated the ability of compound 10 to selectively target an inflamed endothelium (compound 10 modified with a targeting ligand specific for inflamed endothelium -Sialyl Lewis X), as an example of the significant therapeutic potential of such NO-releasing dendrimer.
Johnson et al. (2010) reported the activity of Snitrosothiol dendrimer 9 (G4-SNAP) in reducing ischemia/ reperfusion injury.Using the same concentration of S-nitrosothiol (2 mM of SNAP) in G4-SNAP and in free SNAP (not attached on dendrimer), the infarct percentage (% infarct) was respectively of 3.48 ± 0.46% and 4.41 ± 1.45% (not statistically different).These values were lower than the exhibited by control (similar molecules that do not delivery NO: G4-NAP and NAP), in which the % infarct was around 11% (Johnson et al., 2010).
Acute toxicity of 9 was evaluated on human umbilical vein endothelial cells (HUVEC) and pulmonary artery endothelial cells (CPA-47), using two different assays: propidium iodide uptake (PI) and the release of lactate dehydrogenase (LDH).Results demonstrated the low cytotoxicity of compound 9 in two concentrations (4 and 40 mM) after two hours of exposure: the HUVEC viability was in the range of 89.1-92.7% and the CPA-47 in the range of 87.7-92.3% (in both PI and LDH tests) (Johnson et al., 2010).
The same authors (Johnson et al., 2010) also evaluated the kinetics aspects of nitric oxide release from compound 9 (300 nM) initiated by different concentrations (10, 2, 1, 0.5 mM) of glutathione (GSH).The higher rate for NO release was achieved with 0.5 mM of GSH, resulting in a maximum flux of 1429 ppb NO mg -1 s -1 after 45 min.In addition, reperfusion experiments performed with GSH (0.5 mM) and G4-SNAP together (varying the concentration between 23 pM-31 nM) resulted in an optimal therapeutic dose of 0.23 nM of G4-SNAP (i.e., 15 nM of SNAP).This result represents a 133 fold lower dose in comparison to the free SNAP optimal dose (2 mM), and led to only 2.34 ± 0.9 % infarct (Johnson et al., 2010).Benini, McGarvey and Franco (2008) evaluated the relaxation effects of ruthenium nitrosyl dendrimers (11 and 13) in denuded normotensive rat aortic rings (precontracted with noradrenaline).Slow delayed relaxation started after 15 min (14.1 ± 6.2% induced by compounds 11 and 13, at a concentration of 3 mM of the ruthenium nitrosyl) and the maximum relaxation (36.8 ± 6.5%) was achieved in the second hour.This is on agreement with the behavior observed for [Ru(edta)NO] -in solution, for which the maximal relaxation effect in the second hour was 62.3 ± 16.5%.Compound 11 was also evaluated regarding to its trypanocidal activity in vitro against drug-resistant strain (Y strain) of Trypanosoma cruzi.At the concentration of 1.0 mM (in relation to ruthenium nitrosyl), 100% of typomastigote forms of the parasite were killed, in comparison to the 89% of activity exhibited by [Ru(edta)NO] -not attached to the dendrimer (Benini, McGarvey, Franco, 2008).The trypanocidal effect was attributed to NO release, once the non NO-releasing complex [Ru III (edta) (H 2 O)] -did not exhibited any activity (Benini, McGarvey, Franco, 2008).Sun (2012) evaluated the antibacterial activity of compounds 18, 21, 22, 25, 26, 29 (see Figure 6 and Table 1) and each respective precursor (dendrimer without diazeniumdiolate moiety -non NO-releasing dendrimers, described for simplicity as 18a, 21a, 22a, 25a, 26a, 29a) against Pseudomonas aeruginosa (Gram-negative), standard Staphylococcus aureus (Gram-positive) and antibiotic-resistant (methicillin) S. aureus (MRSA).In general, the concentration of NO-releasing dendrimers needed to kill completely P. aeruginosa was lower than the one required by the respective non NO-releasing dendrimers.Similar results were observed for Gram-positive S. aureus, except for the dendrimer functionalized with PO, in which bactericidal efficacy of compound 22a (non NO-releasing dendrimer) was greater than for 22 (NOreleasing dendrimer).This behavior was attributed to the negative zeta potential (ζ) value for the NO-releasing 22 (-14.0± 5.4 mV) in comparison to the positive value for 22a (7.1 ± 0.6 mV) (Sun, et al., 2012).Once bacterial membrane is negatively charged, interactions with molecules with more positive ζ are expected to be enhanced, improving the efficiency of these compounds (Sun et al., 2012).NO-releasing dendrimers (18, 21, 22, 25, 26, 29) also presented greater bactericidal efficacy against antibiotic-resistant S. aureus (MRSA) than the respective non NO-releasing species (18a, 21a, 22a, 25a, 26a, 29a) (Sun et al., 2012).Nitric oxide bactericidal activity was attributed to NO byproducts such as peroxynitrite (ONOO -) and dinitrogen trioxide (N 2 O 3 ) which would drive the oxidative and the nitrosative stress (Sun et al., 2012).
The same authors (Sun et al., 2012) also evaluate the antibacterial activity of compounds 18, 21, 22, 25, 26, 29 on respect to the dendrimer size (G2 vs. G5) and exterior functionality (PO, SO and PEG) (Figure 6).In general, dendrimer G5 (21, 25, 29) were more effective than G2 (18, 22, 26), and this behavior was attributed to the high density of diazeniumdiolates groups, which led to NO release close to the bacteria cells.In respect to the exterior functionality, dendrimer with SO on the surface (Figure 6) were the most effective, followed by PO and PEG (bacterial activity SO > PO > PEG) (Sun et al., 2012).Authors attributed this result to the association between the dendrimer surface and the cytoplasmic membrane of the bacteria (negatively charged), which is dependent on hydrophobic and electrostatic interactions, and thus favoring SO groups (more positive zeta potential and hydrophobicity) on the surface of dendrimers relative to PO and PEG (Sun et al., 2012).
According to the data described above, in general, the NO-releasing dendrimers exhibit low cytotoxicity and the biological assays showed that the activity of the different NO-donors was similar or improved after the attachment to dendrimers.

CONCLUSION
The combination of classical and new NO-donors with different materials allows to expand the field of possible applications for the NO molecule, constituting an attractive therapeutic option.In general these strategies aim to solve the main challenge in NO-donor therapy: release a desired amount of NO at specific targets.
In this context, dendrimers are molecules which hold interesting properties to be used on the development of NO-carrier systems.The possibility of surface functionalization of dendrimers with one or more molecules of interest can improve biocompatibility, solubility and specificity of these systems, allowing the delivery of nitric oxide to targeted cells or tissues.Furthermore, according to the type of NOdonor conjugated with the dendrimer (metal nitrosyl, diazeniumdiolates, S-nitrosothiols and new ones), the mechanism of NO release can be modified to fit a desired situation, like fast and large amount of NO released, or a low and continuous NO flow.Also, the mechanism involved in the nitric oxide release can be chosen according to the purpose, from spontaneous to triggered.
The diverse NO-releasing dendrimers described in this review offers advantages and disadvantages, and exhibit inherent characteristics regarding to the NO-release mechanism.This fact allows the choice of the most appropriate NO-donor dendrimer and acts as a guide for the development of new NO-donors carriers.Also, the diverse biological applications tested so far with NO-releasing dendrimers show the potential uses of this system in different therapies and the new possibilities.
Another aspect that deserves attention is on regard to NO scavengers dendrimers.This topic has only briefly mentioned here as a consequence of a lack of literature on this regard.Certainly, synthetic aspects and biological experiments aiming septic shock control will be quite welcome.
In this context, the use of dendrimers as NO carriers is a quite promising subject of research to be explored.

FIGURE 7 -
FIGURE 7 -NO release from diazeniumdiolate under physiological conditions.