Glucocorticoid–Dextran Conjugates as Potential Prodrugs for Colon-Specific Delivery: Hydrolysis in Rat Gastrointestinal Tract Contents

Glucocorticoid-Dextran Conjugates as Potential Prodrugs for Colon-Specific Delivery: Hydrolysis in Rat Gastrointestinal Tract Contents ANDREWD. MCLEOD'S, DAVIDR. FRIENDS,AND THOMASN. TOZER*~ Received September 21, 1993, from the *Department of Pharmacy, School of Pharmacy, University of California,

San Francisco, CA 94 143-0446, and #Controlled Release and Biomedical-Polymers Department, SRI International, Accepted for publication May 2, 1994@'. §Current Address: 333 Ravenswood Avenue, Menlo Park, CA 94025. Douglas Pharmaceuticals Ltd, P.O. Box 62-523, Central Park, Auckland 6, New Zealand. Abstract 0 Chronic colitis, e.g., ulcerative colitis and Crohn's disease, is presently treated with glucocorticoids and other antiinflammatoty agents. Side effects limit chronic glucocorticoid therapy. The dose, and consequently the side effects, may be reduced by using prodrugs that selectively deliver drug to the colon. We previously synthesized glucocorticoid-dextran conjugates in which dexamethasone and methylprednisolone were attached to dextran (weight-average molecular weight = 72 600) using dicarboxylic acid linkers (succinate and glutarate). In the present study, the hydrolysis kinetics of the hemiesters (hemiester = glucocorticoid + linker) and dextran conjugates were determined after incubation at 37 "C in diluted luminal contents of the gastrointestinal (GI) trace of male Sprague-Dawley rats. The hemiesters were rapidly hydrolyzed in the proximal small intestine (e.g., dexamethasonehemiglutarate ft12 = 0.5 h). This rate decreased progressively down the GI tract (til2 = 4.8, 54, and 68 h in distal small intestine, cecum, and colon, respectively). The enzyme responsible for hemiester hydrolysis, apparently a type-A alkaline carboxylesterase, is probably of host origin because its activity is highest in the small intestine where bacterial count is low. The dextran conjugates resisted hydrolysis in upper GI tract contents but were rapidly degraded in cecal and colonic contents where the bacterial count is high. The dextran conjugate tested, methylprednisolonesuccinatedextran, was easily hydrolyzed by an endodextranase, indicating that substrate specificity is not lost upon the attachment of glucocorticoid. The results of this study indicate that dextran conjugates may be useful in selectively delivering glucocorticoids to the large intestine for the treatment of colitis.

Introduction Chronic colitis, e.g., ulcerative colitis and Crohn's disease, is currently treated with glucocorticoids and other antiinflammatory agents.' Administration of glucocorticoids, e.g., dexamethasone and methylprednisolone, by the oral and intravenous routes produces systemic side effects, including adrenosuppression, immunosuppression, Cushinoid symptoms, and bone resorption.' In theory, selective delivery of drug to the colon could lower the required dose and hence reduce systemic side effects.2 Thus, colon-specificdelivery of glucocorticoids may be useful for maintenance therapy of chronic colitis. The bacterial count in the colon is higher than that of preceding sections of the gastrointestinal (GI)tract by many orders of magnitude in humans and other animals3 Enzymes of the colonic bacteria are believed to release the pharmacologic active component from natural and synthetic prodrugs, e.g., senna glycosides4and ~ulfasalazine.~ The ideal prodrug is both stable and unabsorbed from the upper GI tract, yet completely releases drug in the colon where the drug is fully absorbed.2 Dextran-naproxen was recently reported by @

Abstract published in Advance ACS Abstracts, June 15, 1994.

1284 /Journal of Pharmaceutical Sciences Vol. 83, No. 9, September 1994

Larsen et al. t o possess these ideal characteristic^.^^^ In this prodrug, the carboxylic acid of naproxen is directly attached to dextran by an ester bond. The prodrug remains intact and is unabsorbed from the upper GI tract.7 Naproxen is completely released, presumably by a combination of dextranases and esterases, and absorbed in the lower GI tract.6 Glucocorticoids do not possess carboxylic acid groups for direct attachment to dextran. Polymeric prodrugs can be prepared using linkers, molecules with functional groups that permit attachment of drug t o the polymer.8 This approach has been previously used in the synthesis of drug-dextran conjugates, for example, benzyl alcohol-carbonate-dextran,g metronidazole-succinate-dextran,1° and metronidazole-glutarate-dextran.lO Dextran conjugates using dicarboxylicacid linkers have not, however, been tested for colon-specific drug delivery. In a previous paper we described, with structures, the attachment of dexamethasone and methylprednisolone to dextran using succinic acid." In addition, dexamethasoneglutarate conjugate was synthesized to determine whether a longer linker has an effect on drug release. Chemical stability of these conjugates was examined a s a function of pH. As expected for esters, the conjugates exhibited specific acidhasecatalyzed hydrolysis with less stability at both extremes of pH. Dextran conjugates were also incubated a t 37 "C in isotonic buffer a t pH 6.8 to measure the chemical degradation expected to occur in the intestinal tract in wiuo. The conjugates had degradation half-lives ranging from 75 to 105 h. In preliminary studies in which dexamethasone-succinatedextran was incubated with rat GI tract contents, the majority of enzyme-mediated drug release was demonstrated t o occur in contents of the large intestine.ll In another study we have shown that these conjugates are effective in treating colitis in a rat model.12 The present investigation focuses on the mechanism of drug release during incubations of the hemiesters and dextran conjugates with contents of the rat GI tract. Incubations were performed in the contents of the stomach, proximal small intestine (PSI),distal small intestine (DSI), cecum, and colon. In addition, the nature of the hydrolytic enzymes present in rat cecal contents was investigated. These experiments included determination of the effects of pH, centrifugation, homogenization, anaerobic conditions, and esterase inhibitors on hydrolysis kinetics. In a further group of experiments, dextran and methylprednisolone-succinate-dextran were incubated with purified dextranase. Degradation was monitored by size-exclusion chromatography to determine if the attachment of glucocorticoid to dextran alters substrate specificity for dextranase.

Experimental Section Materials-Methylprednisolone, methylprednisolone-hemisuccinate, and dexamethasone were generous gifts from Upjohn (Kalamazoo, MI). Calcium chloride, dextran (weight-average molecular weight

0022-3549/94/1200-1284$04.50/0

0 1994, American Chemical Society and American Pharmaceutical Association

= 72 600; number-average molecular weight = 43 400), dextranase (1,6-a-~-glucan-6-glucanohydrolase, EC 3.2.1.11, produced by Penicillium sp.), dextran sulfate (weight-average molecular weight = 5000), diethyl p-nitrophenyl phosphate (E-600), EDTA, lysozyme, ovalbumin, L-phenylalanine, (phenylmethy1ene)sulfonyl fluoride (PMSF), physostigmine sulfate, prednisone, Sephadex (G-75-120,bead diameter 40-120 pM), soybean trypsin inhibitor, Triton X-100, and trypsin were obtained from Sigma (St. Louis, MO). Blue dextran (approximate molecular weight = 2 000 000) was from Pharmacia LKB (Piscataway, NJ). Acetonitrile, methanol, methyl tert-butyl ether, pentane, and Tris were HPLC grade or equivalent (Fisher Scientific, Pittsburgh, PA). Glucocorticoid-hemiesters (dexamethasone-hemisuccinate, dexamethasone-hemiglutarate, and methylprednisolone-hemisuccinate) were reacted with dextran to form covalently-linked conjugates: dexamethasone-succinate-dextran (DS-Dextran); dexamethasone-glutarate-dextran (DG-Dextran) and methylprednisolone-succinatedextran (MPS-Dextran)." Glucocorticoid content (mg per 100 mg of dextran conjugate), measured by HPLC after alkaline hydrolysis, were as follows: DS-Dextran, 8.6; DG-Dextran, 5.7; and MPS-Dextran, 10.5. Apparatus-HPLC was performed on the following equipment: Shimadzu SCL-6A system controller, LC-6A pumps, SPD-6AV variable wavelength detector (Kyoto, Japan); Fisher Scientific 5 pM octadecylsilane Econosphere 250-nm x 4.6-mm column (Pittsburgh, PA); Hamilton PRP-1 guard column cartridge (Reno, NV); Waters WISP 710B autoinjector (Milford, MA); and a Hewlett-Packard HP3396A integrator (Avondale, PA). The mobile phase consisted of 35% acetonitrile and 65% buffer (50 mM trisodium citrate adjusted to pH 4.1 with phosphoric acid). A flow rate of 1 m u m i n and a detection wavelength of 242 nm were used. Other equipment included an Orion pH meter (model 231, Cambridge, MA) and a Brinkmann Homogenizer (Sybron Corp., Cantiague, NY)fitted with a PolytronAggregate blade assembly (Kinematica, Switzerland). Analysis-Samples (200 pL) were withdrawn at predetermined intervals using a wide-bore pipet tip and added to chilled tubes containing 200 pL of saturated aqueous sodium chloride and 60 pL of 6% phosphoric acid or 200 p L of chilled 10% trichloroacetic acid. These conditions were found to prevent further ester hydrolysis.'l After addition of internal standard (either dexamethasone, methylprednisolone, or prednisone) the samples were extracted by vortexing with either a mixture of methyl tert-butyl ether:pentane (6:4) or methyl tert-butyl ether for 20 s. After centrifugmg for 2 min at lOOOg, the organic phase was removed and evaporated at 50 "C in a stream of nitrogen. The residue was dissolved in 100 pL of methanol and 50 pL was analyzed by HPLC. Incubations with Rat GI Tract Contents-Male SpragueDawley rats (180-220 g ) were used throughout these experiments and were fed a standard diet (Purina Rodent Chow, no. 50-01, RalstonPurina, Richmond, IN). Rats were decapitated and the GI tract was removed and chilled within 10 min. The contents of each tissue (stomach, proximal small intestine (PSI), distal small intestine (DSI), cecum, and colon) were removed and diluted to 15% w/v with chilled isotonic buffer. Stomach contents were diluted with acetate buffer (prepared by mixing 0.15 M sodium acetate with 0.3 M acetic acid to achieve pH 4.4, a typical value for rat stomach content^).^ Contents further down the GI tract were diluted with phosphate buffer (prepared by mixing 0.1 M Na2HP04 and 0.15 M NaHZP04 to achieve pH 6.8, a typical value for r a t intestinal content^).^ The latter buffer is subsequently referred to as "pH 6.8 buffer". The diluted contents (2 mL) were placed in glass culture tubes and warmed to 37 "C in a shaking water bath. Stock solutions of glucocorticoid-hemiesters and dextran conjugates (all containing 1.5 mM glucocorticoid) were prepared in pH 6.8 buffer and warmed to 37 "C. The reactions were initiated by adding 1 mL of the stock solutions to the diluted contents (pH of the diluted stomach contents was increased to 4.7 by addition of the stock solution, a pH within ~ the range of values normally measured in the rat ~ t o m a c h ) .Thus, the final concentration of glucocorticoid during the incubations was 500 pM. Hydrolysis of Glucocorticoid-Hemiesters in Rat Cecal Content-Effect of pH-Cecal content, obtained from rats as previously described, was diluted to 20% w/v with 0.9% sodium chloride solution. Large particulate matter was removed from the dispersion by centrifuging at 500g for 0.5 min. The supernate (250 pL) was added to tubes containing 200 pL of 0.25 M citratelphosphate buffer at the

following pHs: 4.3,5.1,5.9,6.9, and 7.9. The tubes were warmed to 37 "C and 50 pL of 500 pM glucocorticoid-hemiester solution was added. Due to faster kinetics the substrate concentration was smaller (50 pM) than in the dextran conjugate experiments (500 pM). Centrifugation a n d Homogenization-A 15% suspension of rat content was prepared using chilled pH 6.8 isotonic phosphate buffer. The suspension was centrifuged at 15000g for 2 min and after aspirating the supernate the pellet was resuspended with buffer. The process was repeated twice. The final pellet was reconstituted with buffer to the same volume as the pooled supernates (3.4 mL). In additional experiments, the cecal content suspension was homogenized (setting 7) for 1 min and compared to a nonhomogenized control. One milliliter of DS-Dextran solution (containing 1.5 mM dexamethasone in pH 6.8 buffer) was added to 2 mL of each of the suspensions obtained in the experiments described above. AnaerobiclAerobicIncubation-In these experiments cecal contents were removed anaerobically in a nitrogen-filled glovebag. All solutions had been degassed by stirring under reduced pressure for 12 h and subsequent purging with nitrogen. The cecal contents were dispersed to 15% w/v using chilled pH 6.8 isotonic phosphate buffer. Two milliliters were placed in each of six vials. After warming to 37 "C, 1mL of DS-Dextran solution (containing 1.5 mM dexamethasone in pH 6.8 buffer) was added. During sampling, three vials were exposed to the atmosphere while those remaining were kept under anaerobic conditions. Enzyme Inhibitors-The effect of esterase inhibitors on the hydrolysis of dexamethasone-hemisuccinate in cecal contents was studied. The following ranges of carboxylesterase inhibitor concentrations were used: diethyl p-nitrophenyl phosphate (0.001-100 pM); (phenylmethy1ene)sulfonylfluoride (0.01-800 pM); physostigmine sulfate (0.01-100 pM) andp-hydroxymercuribenzoic acid sodium salt (1-4000 pM). To examine the effect of metal ions,13 EDTA (0.6-50 mM) and CaClz (2.5-250 mM) were added. To determine whether the enzyme was alkaline phosphatase,14J5incubations were performed with the addition of phosphate (5-20 mM) and L-phenylalanine (1040 mM). The alkaline phosphatase experiments were performed a t pH 7.4 in Tris buffer (20 mM). To determine whether the esterase was membrane-bound, the effect of Triton X-100 (0.8-25 g/L) was also studied.13 Cecal content was diluted to 20% w/v with isotonic pH 6.8 phosphate buffer. Next, the test compounds were dissolved in buffer, and 200 pL was added to 250 pL of the diluted cecal content. The tubes were incubated at 37 "C for 5 min before the addition of 50 pL of 500 pM dexamethasone-hemisuccinate solution. Incubation of MPS-Dextran with Dextranase-Size-exclusion chromatography was used to monitor the breakdown of dextran and MPS-Dextran during incubation with a purified endodextranase.16 Solutions of dextran or MPS-Dextran (both 5% w/v) were prepared in 0.2 M pH 6 citrate/phosphate buffer. Dextranase was added to give a final activity of 5 units/mL. The solution was incubated at 37 "C, and 1-mL samples were withdrawn at 0.5, 6, 15, 25, and 40 min. The reaction was stopped by immersing the samples in boiling water for 1 min, followed by rapid cooling on ice. One hundred microliters was injected onto the column (30-mL bed volume packed with Sephadex G75-120) and eluted with 0.2 M citratelphosphate buffer at pH 6. Two milliliter fractions were collected and diluted 2-fold with buffer. The methylprednisolone content of each fraction was measured by absorbance a t 242 nm. Polysaccharide content was measured by adding 2 mL of anthrone reagent (35 mg of anthrone in 100 mL of concentrated sulfuric acid) to 1mL of each fraction. The solution was allowed to cool for 20-30 min before measuring the absorbance at 630 nm. A standard curve of log molecular weight vs elution volume was constructed using methylprednisolone-hemisuccinate, dextran sulfate, lysozyme, soybean trypsin inhibitor, trypsin, and ovalbumin. Void volume was 7 mL, calculated using blue dextran. D a t a Fitting and StatisticalAnalysis-Initial rate kinetics were used to calculate k3, the rate constant describing the hydrolysis of glucocorticoid-hemiesters.ll This method required a high substrate concentration (500 pM) because product formation was measured in % The slope of the product the initial stages (generally ~ 5 conversion). concentration vs time curve divided by the initial substrate concentration was k3. Equations expressing concentrations of glucocorticoid and glucocorticoid-hemiester as a function of time were derived according to

Journal of Pharmaceutical Sciences / 1285 Vol. 83, No. 9, September 1994

15 7 Glucocorticoid

I

tk3 >Fl

Conjugate Dexmn

I

Hemiester

0

50

100

1 5 0 mln

Figure 1-Glucocorticoid may be released directly from the dextran conjugate by hydrolysis of the glucocorticoid-linker ester bond ( k t ) or by the sequential hydrolysis of the linker-dextran (k2) and glucocorticoid-hemiester (k)bonds.

T

--z

10-

T

MPS

10-

0

Y

10-

0

50

100

1 5 0 min

1 00

150 min

10' Stomach

PSI

DSI

Cecum

Colon

Buffer (PH 6.8)

Figure 2-Hydrolysis rate constant (k3) of glucocorticoid-hemiesters MPS = methylprednisolone-hemisuccinate, DS = dexamethasone-hemisuccinate, and DG = dexamethasone-hemiglutarate. Data are mean + SD (n = 5 animals)on a semilogarithmicscale. The asterisk (*) indicates significant (p < 0.05) difference

compared to buffer hydrolysis rate constants.

the model shown in Figure 1. The equations were fitted t o the data obtained from incubations of dextran conjugates in cecum and colon contents using Minim, a nonlinear regression program for the Apple M a c i n t 0 ~ h . lIt ~ was not possible t o accurately estimate k l and kz for incubations in stomach, PSI, and DSI contents due to the slow hydrolysis rate of dextran conjugates a t these sites. Analysis of variance and t-tests were used to determine statistical differences.

Results and Discussion Figure 2 shows that the hemiesters are rapidly hydrolyzed in the proximal small intestine (PSI) and that the rate declines progressively further down the GI tract. Taking the data from dexamethasone-hemiglutarate, the corresponding half-lives in the contents of the PSI, DSI, cecum, and colon were 0.5, 4.8,54,and 68 h, respectively. A similar enzyme distribution has been reported for another esterase.ls The enzyme distribution suggests that the esterase is either secreted in bile or released from the small intestinal mucosa. Release of glucocorticoid from the three conjugates is shown in Figure 3. Maximal glucocorticoid concentrations were only approximately 8 pM during the 160-min incubations, and no GI tract content appeared to consistently give faster hydrolysis rates. These results show minor, nonspecific release of glucocorticoid from the three conjugates during the incubations. Release of glucocorticoid-hemiester, shown in Figure 4,is dramatically different. Fastest hydrolysis consistently occurs in contents of the cecum and colon with approximate hemiester concentrations of 20-30 pM. The lack of glucocorticoid and glucocorticoid-hemiester release during incubations with PSI and DSI contents indicates that dextran protects both ester bonds from hydrolysis by esterases in the PSI and DSI. This is an important requirement in order to 1286 / Journal of Pharmaceutical Sciences Vol. 83, No. 9, September 1994

0

50

Figure 3-Release of glucocorticoid in contents of the GI tract during incubation of 500 VM (equivalentconcentration of glucocorticoid) solutions of (A) MPS-Dextran, (B) DS-Dextran, and (C) DG-Dextran. Stomach (0),PSI (A), DSI (O), cecum (O),colon (A). Data are mean k SEM (n = 5 animals).

avoid premature release of drug during passage through the upper GI tract after oral administration. Curvature of the graphs suggests that hemiester production slows slightly with time, even though less than 10% of the total glucocorticoid has been released. This may be a result of some glucocorticoid sites on the dextran being less accessible t o enzymatic attack. The rate constants, k l and kz, which describe hydrolysis of the three conjugates in contents of the cecum and colon are shown in Figure 5. Most drug is released as glucocorticoidhemiester (via k2) and virtually no drug is hydrolyzed directly from the conjugates (via kl). Indeed, there is no significant difference between k l when measured in large intestinal contents and that measured in pH 6.8 buffer. These results agree with those shown in Figure 3 and suggest that chemical and not enzymatic hydrolysis is responsible for the direct release of glucocorticoid from the dextran conjugates. Conversely glucocorticoid-hemiester is released a t a faster rate in contents of the cecum and colon by means of an enzymatic reaction. These findings were surprising as one would expect that the dextran-linker ester bond would be more sterically hindered. One explanation for these results is hydrolysis of the glucose-linker bond by a glucosidase after previous

,i 30

0

~

~

0

.

~

6x10.'

(A)

T

T

50

100

1 5 0 min

CI

c

0

1

50

100

150 min

z

Ed H

kl k2

Cecum

Colon

Buffer

Cecum

Colon

Buffer

I

T

T

6x10''

kl .k2

50 100 1 5 0 min 0 Figure 4-Release of glucocorticoid-hemiester in contents of the GI tract during incubation of 500 pM (equivalentconcentration of glucocorticoid) solutions of (A) MPS-Dextran, (B) DS-Dextran, and (C) DG-Dextran. Stomach (0), PSI (A), DSI (O), cecum (O),colon (A). Data are mean & SEM ( n = 5 animals).

Cecum Colon Butter Figure 5-Rate constants for the diluted contents obtained after fitting the model shown in Figure 1 to (A) MPS-Dextran, (B) DS-Dextran, and (C)DG-Dextran hydrolysis data. Data are mean f SD (n = 5 animals) on a semilogarithmic scale. The asterisk (*) indicates significant (p c 0.05) difference compared to

hydrolysis rate constants in buffer.

disintegration of the dextran backbone by endo- and exodextranases.16 These data were obtained during incubations in 10-fold-diluted GI tract contents; therefore, the rates are expected to be considerably faster with undiluted contents in uivo. The effect of incubation pH on the hydrolysis of glucocorticoid-hemiesters (k3, h-l, mean from three replicates) in cecal constants was as follows: 123 for DG at pH 4.3, 5.1, 5.9, 6.9, and 7.9 were 0, 0.009, 0.020, 0.052, and 0.023 h-l; k3 for DS were 0.013, 0.031, 0.072, 0.1145, and 0.088 h-l; k3 for MPS were 0,0.035,0.081,0.101, and 0.057 h-I. The data all show maximal hydrolysis at pH 6.9, indicating that the cecal enzyme was an alkaline esterase.l3 This pH profile is quite different from hydrolysis of glucocorticoid-hemiesters in hamster liver microsomes described by Hattori et ~ 1 These . ~ authors found that glucocorticoid-hemiesters were hydrolyzed fastest at pH 5.5 whereas glucocorticoid-acetates were hydrolyzed fastest a t pH 8. The pellet obtained by centrifugation of cecal contents rapidly hydrolyzed DS-Dextran, but the supernate had little activity. This observation, together with the lack of activity in PSI and DSI contents where the bacterial count is much lower,3 suggests that bacterial enzymes in the large intestine are responsible for hydrolysis of the dextran conjugates. Anaerobic conditions did not affect the hydrolysis of DS-

Dextran in cecal contents; however, homogenization increased the release rate of dexamethasone-hemisuccinate. Similar results were observed during incubations of dexamethasone,8-D-glucoside in guinea pig cecal content^.'^ Hepaticz0Yz1and intestinal m u c o ~ a carboxylesterases ~~,~~ have been described; however, there is a lack of information on esterases in the large intestine.24 In most cases the enzyme isolated has been classified as carboxylesterase-B, due to inhibition by organophosphates. In the present study, hydrolysis of dexamethasone-hemisuccinate was unaffected by the organophosphates diethyl p-nitrophenyl phosphate and (phenylmethy1ene)sulfonylfluoride or the cholinesterase inhibitor physostigmine, at all concentrations shdied. Hydroly~sis was, however, reduced to 33%of control by the addition of 4 mM p-hydroxymercuribenzoic acid sodium salt. Inhibition by p-hydroxymercuribenzoic acid sodium salt but not by organophosphates indicated that the enzyme has a cysteine rather than serine at the active site. The addition of EDTA and Ca2+had no effect at all concentrations, indicating that metal ions are not required. Glucocorticoid-hemiester hydrolysis was unaffected by phosphate and L-phenylalanine, eliminating the possibility that the enzyme is an alkaline phosphatase. Triton X-100 increased the hydrolysis rate by 31%at 0.8 g/L but then decreased the hydrolysis rate by 50% Journal of Pharmaceutical Sciences / 1207 Vol. 83, No. 9, September 1994

n

0

10

30

20

40

50mL

this high molecular weight fraction containing a higher proportion of methylprednisolone appears to be less susceptible to hydrolysis by dextranase. This hydrolysis-resistant fraction may explain why the production of glucocorticoidhemiester slows slightly during incubation of the dextran conjugates with rat cecal and colonic contents. This study shows that these glucocorticoid-dextran conjugates resist breakdown by high levels of esterase in the contents of the small intestine but release glucocorticoidhemiester in contents of the cecum and colon where the bacterial count is high.3 The faster release rate of glucocorticoid-hemiester in the large intestine may explain the efficacy of these conjugates in treating experimental colitis while causing little adrenal supression in the rat.12

References and Notes a. -

a 51

c

8 0

0 0

10

30

20

40

50mL

c E

1. Hanauer, S. B.; Kirsner, J . B. In Inflammatory Bowel Disease; 3rd ed.; Kirsner, J. B., Shorter, R. G., Eds.; Lea and Febiger: Philadelphia, 1988; pp 449-450. 2. McLeod, A. D.; Tozer, T. N. In Oral Colon-Specific Drug Delivery; Friend, D. R., Ed.; CRC Press: Boca Raton, 1992; pp 85-115. 3. Williams-Smith, H. J . Path. Bact. 1965, 89, 95-122. 4. Pepercorn, M. A.; Goldman, P. J . Pharmacol. Exp. Ther. 1972, 181, 555. 5. Hardcastle, J . D.; Wilkins, J. L. Gut 1970, 11, 1038-42. 6. Larsen, C.; Harboe, E.; Johansen, M.; Olesen, H. P. Pharm. Res. 1989, 6, 995-999. 7. Harboe, E.; Larsen, C.; Johansen, M.; Olesen, H. P. Pharm. Res. 1989. 6. 919-923. 8. McLeod, A. D.In Oral Colon-SpecificDrug Delivery; Friend, D. R., Ed.; CRC Press: Boca Raton, 1992; pp 213-233. 9. Weibel, H.: Nielsen. L. S.: Larsen,, C.:. Bundgaard. H. Acta Pharm: Nord. 1991,'3, 159-162. 10. Larsen, C.; Kurtzhals, P.; Johansen, M. Acta Pharm. Suec. 1988, 25, 1-14. 11. McLeod, A. D.; Friend, D. R.; Tozer, T. N. Int. J . Pharm. 1993, 92, 105-114. 12. McLeod, A. D.; Fedorak, R. N.; Friend, D. R.; Tozer, T. N.; Cui, N. 1994, 106, 405-413. Submitted for publication in Gasteroenterology . 13. Negre, A,; Karm, S.; Sable-Amplis, R.; Sicard, R.; Dang, Q. Q.; Rogalle, P.; Douste-Blazy, L.; Salvayre, R. Comp. Biochem. Physiol. 1988, 91B, 79-83. 14. Fleisher, D.; Johnson, K. C.; Stewart, B. H.; Amidon, G . L. J . Pharm. Sci. 1986,75,934-939. 15. Kim, S. H.; Shidoji, Y.; Hosoya, N. Jpn. J . Exp. Med. 1986,56, 251-55. 16. Wheatley, M. A.; Moo-Young,M. Biotech. Bioeng. 1977,19,219I

=

- 0

10 105

30

20 10'

10'

40 10'

50mL

MOI. wt.

Figure 6-Size-exclusion chromatography of dextran (A) and MPS-Dextran (B, C) after incubation with dextranase for 0.5 (0),6 (A), 15 (a), 25 (O),and 40 (A) min. Graphs A and B show carbohydrate content and graph C shows methylprednisolone content. The molecular weight scale is indicated under the x-axis.

-0-

zaa.

17. Purves, D. R. Minim 1.8a, Department of Pharmacology, University of Otago, P.O. Box 913, Dunedin, New Zealand. 18. Hanninen, 0.;Lindstrom-Seppa, P.; Pelkonen, K. Arch. Toxicol. 1987.60.34-36. a t 25 g/L. Hydrolysis rate enhancement and inhibition by 19. Tozer, T.'N.; Rigod, J.; McLeod, A. D.; Gungon, R.; Hoag, M. K.; Friend, D. R. Pharm. Res. 1991,8,445-454. increasing concentrations of Triton X-100 have been observed 20. Mi, B.; Kaur, S.; James, E. C.; Parmar, S. S. Biochem. Pharwith other esterases.13 macol. 1985, 34, 1881-1886. In summary, the enzyme has cysteine rather than serine 21. Hattori. K.: Kamio. M.: Nakaiima. E.: Oshima. T.: Satoh. T.: at the reactive site, is not dependent on heavy metal ions, Kitaeawa. H. Biochem.'PharkacoZ: 1981.30. 2O5lL2056. 22. Cam';bell; C. J.; Chantrell, L. J.; Eastmond, R. Biochem. and is membrane associated. The enzyme is neither a Pharmacol. 1987,36, 2317-2324. cholinesterase nor an alkaline phosphatase. These results 23. Inoue, M.: Morikawa, M.; Tsuboi, M.; S u-~ u r a .M. J m . J . indicate that the enzyme is a type-A carbo~ylesterase,~~ Pharmacol. 1979,29, 9-16. probably from the mucosa of the small intestine. 24. Larsen, C.; Jensen, B. H.; Olesen, H. P. Acta Pharm. Nord. 1991, 3, 41-44. Dextranase (5 units/mL) rapidly hydrolyzed both dextran 25. Walker, C. H.; Mackness, M. I. Biochem. Pharmacol. 1983,32, and MPS-Dextran as indicated by the change in molecular 3265-3269.

weight distribution by size-exclusion chromatography, Figure 6. Dextran was completely hydrolyzed into smaller oligosaccharides within 40 min, graph A. MPS-Dextran, however, had some high molecular weight conjugate remaining after 40 min, graphs B and C. This fraction (up to 12 mL elution volume) was 7% of the total area in the carbohydrate chromatogram (graph B) whereas it was 18% of the total area of the methylprednisolone-contentchromatogram (graph C ). Thus,

1288 /Journal of Pharmaceutical Sciences Vol. 83, No. 9, September 1994

Acknowledgments We wish to acknowledge the assistance of Leon Braswell and Patrick Culhane, Department of Anesthesia, University of California, San Francisco (UCSF). Financial assistance was received from the UCSF Department of Pharmacy Graduate Support Fund and the New Zealand Pharmacy Education and Research Foundation.