Transdermal delivery of nitric oxide from diazeniumdiolates

Journal of Controlled Release 51 (1998) 153–159

Transdermal delivery of nitric oxide from diazeniumdiolates Daniel J. Smith b

a ,b ,

*, Maia L. Simmons a

a Department of Chemistry, The University of Akron, Akron, OH, USA Department of Biomedical Engineering, The University of Akron, Akron, OH 44325 -3601, USA

Received 10 April 1997; received in revised form 29 July 1997; accepted 30 July 1997

Abstract Adverse physiological effects can often interfere with the use of nitric oxide (NO) as a therapeutic agent, especially when it is systemically generated from prodrugs. NO which is generated and delivered site-specifically by transdermal donors may be useful in the treatment of parasitic, bacterial or viral skin infections without causing systemic side effects. Three diazeniumdiolates (formerly ‘‘NONOates’’), including two water soluble compounds, (Z)-1-[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]-diazen-1-ium-1,2-diolate (DETA-NO) and (Z)-1-[N-(3-aminopropyl)-N-(3-ammoniopropyl)amino]diazen-1-ium-1,2-diolate (DPTA-NO), and one insoluble compound, DPTA-NO grafted to dextran microspheres (DPTANO–g-dextran) were used to transdermally deliver NO to rats. Dextran microspheres were obtained by simultaneously grafting DPTA-NO to dextran and cross-linking dextran with CNBr in an oil–water emulsion. Suspended in hydrogel, DETA-NO, DPTA-NO, and DPTA-NO–g-dextran were applied three times to depilated rats at 4 day intervals. Results show that metabolic urinary nitrate levels increase with time (24–48 h), reach a maximum, and return to baseline by the fourth day. DPTA-NO applications produced an average maximum nitrate level of 94.2 mmol / day634.2 mmol S.D. compared to the average maximum nitrate level of 39.8 mmol / day68.6 mmol S.D. obtained from treatment with DETA-NO. These results suggest that DPTA-NO delivered NO more efficiently than DETA-NO. When DPTA-NO–g-dextran microspheres were used as the NO donor, results comparable to DPTA-NO were observed with an average maximum nitrate level of 14.9 mmol / day63.0 mmol S.D.. These and other conclusive data indicate that, via these diazeniumdiolates, NO can be delivered effectively through rat skin.  1998 Elsevier Science B.V. Keywords: Nitric oxide, NO; Transdermal delivery; NONOate; Diazeniumdiolate

1. Introduction Although transdermal delivery of glyceryl trinitrate has commonly been used to treat numerous cardiovascular conditions, and to arrest premature labor and prolong gestation in humans, systemic side effects including hypotension and headache have been observed [1]. Incorporating specificity in the *Corresponding author. Tel.: 11 330 9727414 (Office); fax: 11 330 9727370.

mode of nitric oxide (NO) delivery can selectively target without eliciting undesirable physiological effects. Site-specific transdermal delivery presents a unique way of transporting NO to a targeted area. Potentially, this method can prevent the occurrence of side effects; in addition to bringing about desirable therapeutics, such as treating epidermal or dermal viral infections [2]. For this transdermal application, a relatively new class of prodrugs, the diazeniumdiolates (formerly NONOates), was found to be a suitable NO donor.

0168-3659 / 98 / $19.00  1998 Elsevier Science B.V. All rights reserved. PII S0168-3659( 97 )00161-2

154

D. J. Smith, M.L. Simmons / Journal of Controlled Release 51 (1998) 153 – 159

Scheme 1.

These compounds, as shown in Scheme 1, spontaneously form two equivalents of NO in an aqueous environment [3]. NO release is first order and the rate is dependent on the temperature, pH and type of nucleophile in the diazeniumdiolate. Because NO is a small, neutral molecule, its lipophilic nature should enable it to be easily transported through the lipophilic outermost barrier of the epidermis, the stratum corneum. Recently, Subczynski et al. reported that NO permeates lipidbilayer membranes in vitro [4]. These same lipid bilayer membranes comprise blood vessel walls and their permeabilities allow NO to easily enter the bloodstream. In the bloodstream, NO binds to hemoglobin and is subsequently oxidized to nitrate, the primary oxidation product of NO in vivo [5]. To date there is no reported evidence that NO can be delivered transdermally to living organisms. A need for in vivo experimentation provided the motivation for these studies. In our studies, specific delivery of NO was achieved by topically applying a delivery device containing diazeniumdiolate and hydrogel to depilated rat skin. Upon hydro-activation of the diazeniumdiolate, NO was spontaneously generated at the application site. Subsequently, urine was analyzed by chemiluminescence, and urinary nitrate levels were calculated and compared to baseline levels. Herein, we demonstrated effective, direct transdermal delivery of NO generated from soluble or insoluble diazeniumdiolate-containing patches.

2. Materials and methods

2.1. Materials Materials used in these studies were obtained from the following sources: analytical grade potassium nitrate (99.999%), sodium nitrate and nitrite, and cyanogen bromide from Aldrich (Milwaukee, WI, USA); dextran (512 kD) and arginine (Sigma, St.

Louis, MO); vanadium (III) chloride (Johnson Matthey /Alfa, Ward Hill, MA, USA); nitric oxide (Matheson, Twinsburg, OH, USA); Water-lock A100 (Grain Processing Corporation, Muscatine, IA, USA); sterile Bioclusive  TM (Johnson and Johnson, Arlington, TX, USA); Vetrap (3M Company, St. Paul, MN, USA); AIN-76 low nitrate diet (ICN Biochemicals, Cleveland, OH, USA); and Zipwax (Lee Pharmaceuticals, S. ElMonte, CA, USA). (Z)-1[N-(2-aminoethyl)-N-(2-ammonioethyl)amino]diazen-1-ium-1,2-diolate (DETA-NO) and (Z)-1-[N(3-aminopropyl)-N-(3-ammoniopropyl)amino]diazen-1-ium-1,2-diolate (DPTA-NO) were kindly supplied by Larry Keefer from the Laboratory of Comparative Carcinogenesis, Frederick Cancer Research and Development Center (Frederick, MD, USA). Dense silicone sponge saddles with medical adhesive were generously provided by Variseal (Parkman, OH, USA). Male Sprague Dawley rats (75–99 g) were purchased from Zivic Miller (Zelienopole, PA, USA).

2.2. Experimental method Each rat was given clean bedding, distilled deionized water, and custom low nitrate chow containing 2% arginine ad lib. To facilitate urine collection, rats were placed in metabolic cages, and then randomly assigned to either control or treatment groups. Urine was collected and used to determine baseline urinary nitrate levels at 24 h intervals for 5–8 days prior to the transdermal studies. A 5 ml volume of 3 M HCl was added daily to each urine collection vial, to inhibit bacterial growth. After treatment with diazeniumdiolates, urine collection continued at 24 h intervals. Treatment compounds were applied as follows: rats were anesthetized with Nembutal (45 mg / kg i.p.), the mid-dorsal section shaved, any remaining hair removed with Zipwax, the skin cleaned with an isopropanol-soaked pad, and a silicone saddle applied to the back. Rectangular rat saddles were made from extruded closed-cell silicon (7 cm36 cm, 3.1 cm diameter, center hole 1 cm deep). The contoured saddles were placed on the shaved section of the rat and secured by medical adhesive. Approximately 3.0 g of hydrogel vehicle, prepared by swelling 300 mg of water-lock in 40 ml of water,

D. J. Smith, M.L. Simmons / Journal of Controlled Release 51 (1998) 153 – 159

was added to each saddle well. Control saddle wells were filled with hydrogel (3.0 g) only, while experimental saddle wells received diazeniumdiolates dissolved or suspended in 3.0 g of hydrogel. After topical application, the silicon wells were covered with Bioclusive film, and the midsection was wrapped with Vetrap to further secure the saddle. Saddles were removed during reapplication, and the wells were wiped with gauze pads. In separate control experiments, selected quantities of sodium nitrate or nitrite dissolved in hydrogel were applied in the same manner.

2.2.1. Nitrate analysis All urine samples were assayed in duplicate for nitrate using a Monitor Labs Model 8440 Nitrogen Oxides Analyzer (Data Systems, San Diego, CA, USA) as previously described [6]. Briefly, urine samples were injected into a solution of 0.4 M vanadium (III) chloride and 1.5 M HCl at 958C. The NO gas generated from the reduction of nitrate or nitrite was purged from the impinger reaction vessel with helium and detected by chemiluminescence. The instrument was calibrated daily with standard potassium nitrate solutions. 2.2.2. DPTA-NO–g-dextran synthesis Dextran (3.6 g) dissolved in 15.6 ml of water and 1.5 g of DPTA-NO were mixed with 10 M NaOH (5 ml). The solution was then poured into a Waring Blender, containing light mineral oil (300 ml). After mixing for 15 min, cyanogen bromide (5 g) dissolved in 35 ml water was added and blended for another 2 min. The mineral oil was removed from the microspheres with three, 300 ml washes of petroleum ether. The hydrated microspheres were then washed four times with 1 l of 35% aqueous ethanol and centrifuged each time, for 5 min, at 5000 rpm. The supernatant from the fourth wash gave no absorbance at 250 nm (UV abs. for NO bearing amines), and therefore, was considered free of residual DPTA-NO. The microspheres were finally dehydrated by washing with 1 l of EtOH, filtered and vacuum dried at room temperature, yielding 3.0 g of free flowing white powder. 2.2.3. No release from DPTA-NO–g-dextran The in vitro NO release profile was determined

155

using the nitrogen oxide analyzer. The sampling chamber consisted of a gas impinger bottle fitted with a 2-way valve. One end of the valve was connected to the analyzer, and the other end to a flow meter and helium gas tank. The flow of helium gas was 10 psi and the flow meter was set at 150 ml / min. The sampling chamber was filled with 50 ml of phosphate-buffered saline (pH 7.4) and degassed; then 10 mg of DPTA-NO–g-dextran was added. With the valve closed, NO from the beads was allowed to accumulate in the headspace. When the valve was opened, the accumulated NO was purged, as helium flushed into the detector, and analyzed by chemiluminescence. Samples were collected periodically until no more NO could be detected. The analyzer was calibrated with known samples of potassium nitrate.

3. Results Before any transdermal procedures were performed, in vitro first order NO release profiles for DPTA-NO and DETA-NO were determined in nonbuffered hydrogel at 328C. The half-lives of DPTANO and DETA-NO were 36 h and 62 h, respectively as compared to 3 h and 20 h, measured at 378C and pH 7.4, as reported by Keefer et al. [7]. These considerable differences may be reflective of a higher initial gel pH and a lower temperature. All experimental animals were placed on a nitratedeficient diet at least seven days prior to the transdermal study. Baseline urinary levels for these animals averaged 7.4 mmol / day ranging from 4 to 8 mmol / day. (Figs. 1–4). Control animals in each study received either hydrogel alone (Figs. 1–3) or hydrogel with cross-linked dextran (Fig. 4). Urinary nitrate levels for these animals varied slightly, but again averaged 7.4 mmol nitrate / day. The possibility that nitrate or nitrite (formed when NO generated in solution is oxidized) would cross the epidermis instead of or in addition to NO was eliminated by conducting the following experiment. Specific quantities of NaNO 3 and NaNO 2 , 306 or 612 mmol, were dissolved in hydrogel, and topically applied to rats (n52) with subsequent monitoring of the urinary nitrate levels for five days (Fig. 1, insert). No significant changes were observed in the urinary

156

D. J. Smith, M.L. Simmons / Journal of Controlled Release 51 (1998) 153 – 159

Fig. 1. Average urinary NO 2 3 profiles for DETA-NO (n53) and control (n53) treated rats. Topical applications of 0.3 mmol DETA-NO dissolved in hydrogel or hydrogel alone (control) were made at the beginning of days 0, 3, 6, 9, and 12 as designated by the arrows. Insert: 2 2 average urinary NO 2 3 profile for rats treated on day 0 with a topical application of 306 mmol NO 3 (n52) or 612 mmol NO 3 (n52) dissolved in hydrogel.

nitrate levels and similar results were obtained for nitrite (data not shown). In two separate preliminary studies, 50 mg (0.3 mmol) of DETA-NO or 200 mg (1.0 mmol) of DPTA-NO dissolved in gel matrix was topically applied to three rats. As a result of applying five, 50

mg doses of DETA-NO to the epidermis at 3 day intervals, five significant peaks appeared within 24 to 48 h, as shown in Fig. 1. Essentially, each peak returned to baseline levels of urinary nitrate after 3 days. The peak heights ranged from 18.2 mmol to 34.2 mmol nitrate / day with an average maximum

Fig. 2. Average urinary NO 2 3 profiles for DPTA-NO (n53) and control (n53) treated rats. Topical applications of 0.3 mmol DPTA-NO dissolved in hydrogel or in hydrogel alone (control) were made at the beginning of days 0, 3, 6, 9, and 12 as designated by the arrows.

D. J. Smith, M.L. Simmons / Journal of Controlled Release 51 (1998) 153 – 159

157

Fig. 3. Comparative average urinary NO 2 3 profiles for DETA-NO (n52), DPTA-NO and control (n52) treated rats. Topical applications of 0.6 mmol DETA-NO or DPTA-NO dissolved in hydrogel or hydrogel alone (control) were made at the beginning of days 0, 4, and 8 as designated by the arrows.

Fig. 4. Average urinary NO 2 3 profile for DPTA-NO–g-dextran microsphere (n52) or dextran microsphere (n52) treated rats. Topical applications of hydrogel containing 200 mg of DPTA-NO–g-dextran microspheres (70 mmol NO) or dextran microspheres (control) were made at the beginning of days 0, 4, 8, and 12 as designated by the arrows.

158

D. J. Smith, M.L. Simmons / Journal of Controlled Release 51 (1998) 153 – 159

peak height of 24 mmol / day. Five larger doses of DPTA-NO (200 mg) also produced five urinary nitrate peaks as shown in Fig. 2. Except for the first application, these did not return to baseline after 3 days; however, peak heights were observed 24 to 48 h after application. The levels were correspondingly higher, averaging 255.0 mmol / day, broadly ranging from 88.4 to 395.0 mmol / day. Visual observation of the animals’ skin surfaces showed no obvious inflammation or redness; therefore, no histological studies were done. A second set of experiments was designed to directly compare transdermal NO delivery profiles. DETA-NO or DPTA-NO (0.6 mmol) was applied three times at 4 day intervals to a separate group of rats (n52). Delivery profiles for this comparative study, as shown in Fig. 3, resulted in three peaks each. Maximal peak heights were again reached in 24 to 48 h post-treatment. Comparable NO release from DETA-NO and DPTA-NO was observed with the first application. However, DPTA-NO caused progressively increasing peak heights while little change was observed with DETA-NO. DPTA-NO treatment produced an average maximum peak height of 94.2 mmol nitrate / day634.2 mmol S.D. (ranging from 48.6 to 144.8 mmol / day). DETA-NO produced an average maximum peak height of 39.8 mmol nitrate / day68.6 mmol S.D. (ranging from 26.1 to 49.3 mmol / day). Because each DETA-NO and DPTA-NO peak returned to baseline, actual NO deliveries could be calculated by summing the points in each peak. When each actual delivery was divided by the theoretical yield, the average total percent NO delivered by DETA-NO and DPTA-NO was estimated to be 6.1 and 17.5%, respectively. To eliminate the possibility that DPTA-NO directly entered the skin instead of NO, DPTA-NO was chemically grafted to a solid microsphere support. Grafting of DPTA-NO to dextran was achieved through cyanogen bromide activation of dextran in an oil in water emulsion. Control microspheres (dextran only) were made similarly without incorporating DPTA-NO. The grafting procedure produced DPTA-NO–g-dextran in 75% yield after removal of all soluble DPTA-NO; the NO release profile was first order with a t 1 / 2 of 3.9 h (pH 7.4; 228C). Grafted DPTA-NO has a slightly shorter halflife than soluble DPTA-NO; 3.9 h compared to 4.7 h.

The total amount of NO released from the microspheres was 0.35 mmol NO / mg microsphere, and based on this value, the grafting efficiency was calculated as 2%. The microspheres rapidly swelled in water and were stable to hydrolysis with no observable release of soluble NONOate or dextran after one week. Based on these results, the microspheres were used in the following transdermal experiment. DPTA-NO–g-dextran (200 mg containing 70 mmol of NO) or dextran microspheres (200 mg) were applied to animals and hydrated with vehicle (hydrogel). The experiment consisted of two animals in each group, with four applications, at 4 day intervals. Results of the urinary nitrate profile from these animals are shown in Fig. 4. Peaks are observed after each application with maximum levels of nitrate reached after 48 h. Maximum average peak heights ranged from 12.2 to 20 mmol nitrate / day, averaging 14.95 mmol nitrate / day63.0 mmol S.D. Here, the average total percent NO delivered was 16.4%. Although there was considerable variability in these profiles, comparable NO delivery was achieved from either soluble or grafted DPTA-NO.

4. Discussion At this point, effective delivery of NO has been clearly demonstrated; although, the depth of NO penetration has yet to be determined. NO was spontaneously generated within a hydrogel / prodrug containing patch from the proton catalyzed decomposition of either diazeniumdiolate, DETA-NO or DPTA-NO. Following the decomposition of the diazeniumdiolate, NO was transferred through the skin by diffusion and, subsequently was cleared from the rat primarily as urinary nitrate. Most likely, with a half-life of about 4 s, NO decomposed to nitrate before it could have adverse effects elsewhere. A possible explanation for the elevation of urinary nitrate could have been that NO within the patch decomposed to nitrate or nitrite and then entered the skin. However, the control experiments showed that relatively high levels of nitrate and nitrite, incorporated into the patch, did not enter the skin. Using similar logic, the prodrug itself may have been transdermally delivered before decompos-

D. J. Smith, M.L. Simmons / Journal of Controlled Release 51 (1998) 153 – 159

ing, but ultimately formed nitrate. But, when DPTANO was chemically linked to a solid support, i.e., cross-linked dextran, the nitrate levels still increased similarly as when DPTA-NO was used alone, as reflected by the average total NO deliveries. DETA-NO and DPTA-NO were chosen for these experiments because of their stability and efficiency of NO generation (2 equivalents of NO / mole diazeniumdiolate). Interestingly, profiles for these two compounds showed very little difference; however, the average maximum peak height was higher for DPTA-NO (94.2 mmol nitrate / day compared to 39.8 mmol nitrate / day for DETA-NO). Treatment with either produced a delayed response, followed by a rapid rise to a maximum which returned to baseline 4 to 5 days post-application. The 24 h delay commonly observed may be indicative of slow NO or nitrate clearance from the skin because of cellular or extracellular retention; or alternatively, there may be a delay in the delivery of NO from the prodrug. However, the delayed delivery from the prodrug is unlikely since it was recently shown that when NO was delivered from an insoluble polymeric diazeniumdiolate to a full thickness wound on a rat, an immediate increase in urinary nitrate level was observed [8]. The urinary nitrate profiles obtained from treatment with equal doses of these diazeniumdiolates were similar, suggesting that the delivery of NO was independent of the half-life of the donor. A number of factors may be responsible for this observation. First, the assay may be too insensitive to detect the changes between these two compounds. Second, the state of gel hydration may influence the rate of NO release from either DPTA-NO or DETA-NO. Often the gel containing the diazeniumdiolate became dry after remaining on the animal 3 to 4 days. This dehydration may have slowed NO delivery. Third, since NO has limited water solubility (2 mM), it may initially saturate the hydrogel before travelling to the skin. Later, when the gel dehydrates, it may be lost primarily by diffusion out of the gel as a gas. Clearly, future studies implementing the diazeniumdiolate / hydrogel patch will require an improved delivery system. Incorporating a gas impermeable backing would potentially prevent significant NO gas escape. In addition, adjusting the pH of the donor / vehicle and reducing the hydrogel thick-

159

ness may accelerate NO generation as well as shorten the time required for NO to reach the skin. Such optimizations will be the focus of later experiments. Potentially, our delivery device may help solve one of the major problems associated with administering NO as a therapeutic agent. Because NO prodrugs are delivered systemically to treat localized areas, broad or system-wide effects normally result. Topical application of diazeniumdiolates would allow elevated levels of NO to be transported transdermally to a specific site, without adversely affecting other NO-sensitive cell types, tissues, or organs. This type of delivery could be useful in targeting various skin conditions resulting from NO production impairment or viral and bacterial pathogens.

References [1] C. Lees, S. Campbell, E. Jauniaux, R. Brown, B. Ramsay, D. Gibb, S. Moncada, J.F. Martin, Arrest of preterm labour and prolongation of gestation with gyceryl trinitrate, a nitric oxide donor, Lancet 342 (1994) 1325–1326. [2] G. Karupiah, Q. Xie, M.L. Buller, C. Nathan, C. Duarte, J.D. MacMicking, Inhibition of viral replication by interferon-linduced nitric oxide synthase, Science 261 (1993) 1445– 1448. [3] J.A. Hrabie, J.R. Klose, D.A. Wink, L.K. Keefer, New nitric oxide-releasing zwitterions derived from polyamines, J. Org. Chem. 58 (1993) 1472–1476. [4] W.K. Subczynski, M. Lomnicka, J.S. Hyde, Permeability of nitric-oxide through lipid bilayer-membranes, Free Radic. Res. 24(5) (1996) 343–349. [5] A. Wennmalm, G. Benthin, A. Edlund, L. Jungerstein, N. Kieler-Jensen, S. Lundin, U.N. Westfelt, A.S. Peterson, F. Waagstein, Metabolism and excretion of nitric oxide in humans. An experimental and clinical study, Circ. Res. 73 (1993) 1121–1127. [6] J.P. Bulgrin, M. Shabani, D. Chakravarthy, D.J. Smith, Nitric oxide synthesis is suppressed in steroid-impaired and diabetic wounds, Wounds 7 (1995) 48–57. [7] L.K. Keefer, R.W. Nims, K.M. Davies, D.A. Wink, ‘‘NONOates’’ (1-Substituted Diazen-1-ium-1,2-diolates) as NO donors: Convenient nitric oxide dosage forms, Methods Enzymol. 268 (1996) 281–293. [8] M. Shabani, S.K. Pulfer, J.P. Bulgrin, D.J. Smith, Enhancement of wound repair with a topically applied nitric oxidereleasing polymer, Wound Repair Regenerat. 4 (1996) 353– 362.