Nitric oxide release from polydimethylsiloxane-based polyurethanes

INTRODUCTION

The role of nitric oxide (NO) in physiology, neuroscience and immunology has been recognized to the extent that it was named by the journal Science as “Molecule of the year” in 1992 (1). In addition to its vasodilatory function, NO has also been shown to inhibit smooth muscle cell proliferation (2), inhibit platelet aggregation (3) and adhesion (4), and to attenuate leukocyte function (5). The molecule also plays an important role in angiogenesis, as it inhibits apoptosis and increases endothelial proliferation and migration (6). A review on the effects of NO in the cardiovascular system may be found in an article by Strijdom et al (7).

With all these potentially therapeutic attributes, it is not surprising that NO is used not only in clinical settings as inhaled gas for treatment of primary pulmonary hypertension, hypoxia and acute respiratory distress syndrome (8), but also increasingly in experimental medical device applications (9). Examples of devices include vascular grafts (10, 11), oxygen sensors (12, 13), extracorporeal circuits (14, 15), stents (16) and glucose sensors (17). A large number of NO donors are available (18), and many of them have been incorporated in prosthetic materials such as silicone rubber (14), polyurethane (PU) (19), polyvinyl chloride (10), polymethacrylate (20) and expanded polytetrafluoroethylene (ePTFE) (9).

Two classes of NO donors are commonly used in biomaterials, namely diazeniumdiolates (also called NONOates) and S-nitrosothiols. The S-nitrosothiols are generally formed by reaction of nitrous acid with the parent thiol, and are reported to require copper-mediated decomposition, reaction with ascorbate or cleavage by light to release NO. The former (NONOates) are synthesized by reaction of amines with NO gas to form relatively stable compounds that spontaneously release NO on contact with bodily fluids, and most of the work in the field of NO donors in biomaterials has concentrated on the use of the NONOates (9). Although primary and secondary amines may be used (21), secondary amines are more often used, as their diazeniumdiolates tend to be more stable and give greater NO yields (22, 23). A list of donors used in NO-eluting materials that include a range of secondary amines is given in Table I. In order to maximize the number of secondary amines, combined with the requirement to have a nontoxic, high-molecular-mass material for incorporation into the PU polymer, linear polyethyleneimine (LPEI) was chosen as donor in the current study. The diazenium-diolization of a secondary amine, as exemplified by LPEI, is shown in Figure 1.

NO donors may be non-covalently incorporated (by dispersion) into the polymer matrix or coating (28). More recently, more lipophilic donors have been developed (27), or the donors were covalently coupled to the polymer backbones (19, 27, 33). When the donors are incorporated into the polymers, there is also a choice between diazeniumdiolation of the donor before incorporation (28), or incorporating the donor and then exposing the polymer to NO gas to NONOates it in situ (14). In the latter case, the diffusion of the NO gas through the base polymer is of importance, and this method holds the advantage over pre-exposure in that the device may be processed into its final shape and then exposed to NO before use. On the other hand, pre-exposure of the donors would require careful handling of the diazeniumdiolate and doped polymer during processing and storage, as NONOates can have very short half-lives (9).

The current study thus describes the incorporation of LPEI in PU films containing different amounts of polydimethylsiloxane (PDMS) in their soft segments and subsequent NONOation by NO gas exposure to determine whether long-term NO elution can be achieved. In addition, the hypothesis that the higher diffusivity of diatomic gasses through PDMS over PU (39, 40) would result in increased NO release with increased PDMS content (e.g., N₂ has a larger diffusion constant in PDMS than in PU). Finally, the effect of LPEI doping and elution on the mechanical properties of the PU was determined, mainly to ensure that the loading with LPEI and NO did not have an unacceptable adverse effect on these properties.

MATERIALS AND METHODS

RESULTS

DISCUSSION

The current paper discusses the effect of soft-segment composition on the mechanical and NO release properties of PDMS-based PUs. Specifically: (i) the PUs (without LPEI) showed increased modulus, decreased ultimate stress and decreased ultimate strain with increased PDMS content, (ii) PU films doped with LPEI could be cast from solution and exposed to NO gas at elevated pressure to form NO donors in situ, (iii) the inclusion of LPEI and exposure to NO gas had little or no effect on the mechanical properties of the films, (iv) the total amount of NO released in vitro generally increased with increased PDMS content of the PU, (v) the release of NO occurred over an extended time period after an initial burst release within the first 48 hours and (vi) the release rate of NO from the films exceeded that reported for endothelial cells for up to 120-240 hours depending on the polymer formulation.

The decreases in ultimate tensile stress and strain with increased PDMS content of the PU soft segments were similar to those observed by Gunatillake et al (42). The stiffness (Young’s moduli) of the materials also showed similar increases with increased PHMO content, which can be ascribed to the compatibilizing effect of the PHMO in the soft segments, as there is a large difference in the solubility parameters between PDMS and the other reagents constituting the PU. The overall lack of significant differences between the mechanical properties of the control (PU-50-NO) and NO gas exposed (PU-50+NO) samples (without LPEI incorporation) shows that the gas itself did not cause damage to the polymer that was sufficient to discern at this level. The inclusion of LPEI in the polymers similarly did not dramatically influence the mechanical properties: apart from one instance of decreased modulus, 2 samples actually showed small increases in ultimate tensile strength that may be attributed to a reinforcing effect of the added polymer, as well as variation in sample properties often observed with tensile specimens. As the further exposure of the LPEI-loaded films to NO gas also had mostly a small and nonsignificant effect on all 3 of the mechanical parameters investigated, the overall effect of donor incorporation and in situ diazeniumdiolation of the donor on the mechanical properties of the films was concluded to be small and deemed acceptable. The lack of significant differences in mechanical properties of samples before and after exposure to PBS also indicates that the mechanical integrity of the materials is not compromised by exposure to the eluant in the short term, and is not expected to be compromised by in vivo exposure, as these polymers were shown to exhibit good biostability, especially the formulation with 80% PDMS in the soft segment (43).

The choice of LPEI as donor was based on a number of factors: it (i) is relatively nontoxic (especially compared with small dialkyl diamines; e.g., N,N’-dibutyl-1,6-hexanediamine [DIB] has an LC50 of 220 mg/m³), (ii) has a high secondary amine content and (iii) has a high molecular mass that should slow elution rate from the matrix. Trials with N,N’-dimethyl-1,6-hexanediamine (DIM) and DIB as secondary amines were abandoned after blooming of the former to the surface, while the latter was essentially fully eluted from the polymer films within 12 hours. In comparison, semiquantitative determination of LPEI release showed that only approximately 25% of the loaded amount eluted after 100 hours (data not shown).

The formation of the diazenium diolate was not shown directly, but their formation and the subsequent release of NO is indirectly shown by the highly elevated values measured for NO release in when LPEI was present in the PU polymer; PU samples containing no LPEI showed release of less than a tenth of those amounts. The latter can be explained by the absorption of the NO gas in the PU material itself. Although the decomposition of diazeniumdiolate NO donors have been shown to be pseudo-first order (27), the release profile of NO from the films was not consistent with first order (lnM_t α t) (44) kinetics (R² <0.95-0.86 for 00+50+NO through 100+50+NO). The release also did not follow the Higuchi release model based on Fickian diffusion (M_t α t^1/2) (45) (R² <0.89-0.59 for 00+50+NO through 100+50+NO), as NO release is dependent not only on the decomposition of the NONOate and diffusion of the released NO from the polymer, but also on the rate of ingress of the eluant. The data did, however, give very good least square fits over the entire 800-hour period using a modified Korsmeyer-Peppas power model (M_t α tⁿ) (46), with R² values ranging from 0.951 to 0.999 (Tab. II), while even better fits (0.987 to 0.999) could be achieved for the 60%-100% PDMS samples when an x-shifted power law (M_t = k’(t-a)ⁿ) was applied, where the x-shift could account for the loss of NO in these samples while being transferred into the elution medium. Despite this good fit to the power curve, the diffusional exponent (n) did not correspond to the values (0.5<n<<1) for thin films obtained by Ritger and Peppas for anomalous non-Fickian transport (46). The short half-life of NONOates in the presence of water, and the high diffusion rates of the NO in the polymer (due to its small size) are offered as a possible explanation for the low value of the exponent in this case, while the increase in NO release with increased PDMS content can be ascribed to the increased diffusivity of PDMS toward diatomic gases, although the gas transport of NO was not specifically tested.

Direct comparison with historical NO release values found in the literature for similar systems is hampered by the lack of consistency of units and often the lack of normalization of released amounts to unit mass or area. Table I has thus been compiled to, wherever possible, normalize the data presented in the literature and present the data in comparable units, as well as give an indication of the release profiles achieved. Thus, the current release of 93-225 µmol/g of film compares favorably with others in the field (see Tab. I). For example, Mowery et al achieved 11 and 125 µmol/g NO elution from PU material loaded with LPEI and DIM respectively (28), where the diazeniumdiolates were prepared in solution before incorporation, while 2.5 to 25 µmol/g could be achieved by incorporating diazeniumdiolated secondary amine-containing chain extenders covalently into the PU backbone (37). In terms of elution per unit area, the 2-6 µmol/cm² also compares well with values of 0.001 and 0.02 µmol/cm² for PUs containing S-nitrosothiol donors (35, 36) and even with the 8-20 µmol/cm² achieved from PVC materials using 4%-8% noncovalently incorporated DIB (27). NO fluxes are even harder to compare as they were determined at different times during the elution process, but the fact that relatively high fluxes in the order of 1 pmol·cm^-2·s^-1 could be achieved in the current experiment over such a long timeframe (120-800 hours), compared with the double digit time points reported in most prior cases (Tab. I: Flux at time t [flux t] and time point of flux measurement [flux time], is encouraging.

The increase in NO release capacity with increased PDMS content may be ascribed to the increased diffusivity of the PDMS soft segment, as the non-LPEI-loaded samples did not show significant differences in NO storage/release capacity.

Although the total release capacity of the device may be of importance, the rate at which the release is maintained and the period for which it is maintained are probably the more important measure. Resting endothelial cells are reported to release NO at a rate (Flux: F) of 0.83 pmol·cm^-2·s^-1, while that of a stimulated endothelium is in the 2.7-6.8 pmol·cm^-2·s^-1 range (24), and release rates of synthetic devices are often compared with these benchmarks. In the current study, the release rate was above that of the stimulated endothelium for at least the first 24 hours, while levels above base value were maintained for 120-240 hours. It is known that high release rates (10-20 pmol·cm^-2·s^-1) can decrease bacterial adhesion to surfaces (17, 32), and that even at levels well below (0.002 pmol·cm^-2·s^-1) that of the endothelium, platelet and smooth muscle cell adhesion are decreased, and endothelial cell adhesion increased (19). The current films maintained initial rates greater than that shown to be required for bacterial resistance (in the initial stages when this will be most required in an implant), and maintained values above those shown to be sufficient for the other advantageous cellular attachment effects for the full investigation period of more than 33 days.

The encouraging in vivo results by Fleser et al (10), namely that of increased thrombus free area and patency when using NO-releasing PVC coatings on Dacron vascular grafts in a sheep arteriovenous shunt model, shows the potential of this technology. It thus seems only a question of time for the fulfilment of the promise of NO-releasing materials becoming the “artificial endothelium” (24) and of being used in the vascular grafts of the future (47).

CONCLUSIONS

Increased PDMS content in the segmented PU Elast-Eon™ resulted in increased NO release when the LPEI donor was incorporated into the PU and subsequently exposed to NO gas. LPEI and NO incorporation did not result in significant losses in mechanical properties. Extended release rates above that of the cultured endothelium could be achieved for extended periods using this system, which holds much promise for use in vascular graft and other applications where thrombogenicity requires suppression.