- Email: [email protected]
A morphine-triggered delivery system useful in the treatment of heroin addiction
ClinicalMaterials I3 (1993) 109-119
A Morphine-Triggered Delivery System Usefill in the Treatmint of Heroin Addiction K. V. Roskos, J. A. Tefft & J. Heller* Controlled
Release and Biomedical
Polymers
Department,
SRI International,
Menlo
Park,
California
94025,
USA
Abstract: The ultimate objective of this work is to develop a device that can be triggered by morphine to release naltrexone. Two device configurations are described. In one configuration, naltrexone is dispersed in cellulose acetate phthalate microspheres which are then spray-coated with trilaurin. In the other configuration, naltrexone is dispersed in an n-octyl half ester of methyl vinyl ether and maleic anhydride copolymer and the mixture fabricated into a disk which is then coated with trilaurin. The microspheres are designed to release naltrexone abruptly while the disks are designed to release naltrexone at a constant rate over a two week period. The microspheres, or the disk along with a reversibly inactivated lipase are plased inside a semipermeable membrane that allows free passage of morphine and naltrexone but excludes the higher molecular weight components of the device. Reversible inactivation of lipase is achieved by covalent attachment of morphine and complexing with morphine antibody. Activation of the device occurs by diffusion of morphine into the device and displacing the lipasemorphine conjugate from the antibody. The activated lipase then removes the trilaurin protective coating, thus triggering naltrexone release.
INTRODUCTION
Unfortunately, readdiction is an ever present danger, even for individuals that have been opiate-free for a long time. For this reason, the development of a device that would guard against this possibility has been under development in our laboratories for a number of years. In this device, the narcotic antagonist naltrexone is dispersed in ai rate-controlling bioerodible polymer surrounded by an enzymedegradable protective coating which prevents release of naltrexone from the device. The coated device is in turn surrounded by a reversibly inaLctivated enzyme. This enzyme in its inactivated st,ate is unable to degrade the coating, but once activated is able to rapidly remove the protective coating. The reversible inactivation of the enzyme is achieved by first preparing an enzyme-morphine conjugate and then complexing the conjugate with a morphine antibody.6 Because the antibody is a fairly large molecule, access of the substrate to the enzyme’s active site is sterically inhibited, thus effectively rendering the enzyme inactive. Triggering of naltrexone release is initiated by the appearance of morphine in the tissue surrounding the device, diffusion into the
Heroin addiction remains a very serious, worldwide problem which is now widely prevalent in all segments of our society.’ Although an effective cure remains; elusive, the use of narcotic antagonists such as naltrexone is a promising approach.2’3 Such antagonists are able to block the euphoric effect of heroin and by maintaining an effective opiate-receiptor blockade, readdiction is not possible. Currently, naltrexone is administered in oral dosage form which is effective for two to three days.4 However, because discontinuation of naltrexone therapy is without side-effects, maintenance of therapy requires strong motivation. Therefore, a therapy which removes the decisionmaking, process from the patient is highly desirable and sustained release formulations that are implanted deeply enough in the tissues so that removal by the patient is impossible are currently under development.5 *To whom
correspondence
should be addressed.
109 ClinicalMaterials O267-6605/93/$6.000
1993 Elsevier
Science Publishers
Ltd, England
oskos, J. A. Tefft, 9.
110
device, and dissociation of the enzyme-haptenantibody complex, thus rendering the enzyme active. The development of a morphine-triggered naltrexone delivery system has been under development in our laboratories for a number of years.7-“’ and after considerable experimentation we have selected a lipase-degradable trilaurin as the protective coating surrounding a rate-controlling bioerodible polymer with dispersed naltrexone. Because such a device contains a number of antigenic compounds, it is surrounded by a macroporous membrane that is freely permeable to morphine and naltrexone but which excludes the enzyme and the morphine antibody. Two device configurations are currently under investigation. In one configuration, the ratecontrolling polymer with dispersed naltrexone is fabricated into a disk, which is then coate trilaurin. l2 Following activation, this device is designed to release naltrexone over a two-week period. In the other configuration, the rate-controlling polymer with dispersed naltrexone is fabricated into microspheres which are then coated with trilaurin. ’3 Following activation, this device is designed to begin the release of naltrexone in no more than one hour. A schematic representation of both configurations is shown in Fig. 1. In this manuscript we will summarize the current status of development of these devices.
ment of the two specific
A device designed
devices,
to release naltrexone for two weeks
and maleic an previously to ~~~e~~~ su comitant release of an n
RESULTS AND DISCUSSION s&lble
Regardless of the final configuration, development of these devices requires the development of four separate components which are then assembled into a final device. These components are (1) ratecontrolling bioerodible polymer, (2) enzyme-
the
rate sf
a&y1
group
ePosPon
on t&e
ester portion of the co~o~y~e~~ so that selection of
MICROPOROUS MEMBRANE PROTECTIVE COATING NALTREXONE CORE
ANTIBODY BLOCKED ENZYME
NALTREXONE RELEASE FOR TWO WEEKS
NALTREXONE RELEASE IN LESS THAN ONE HOUR
Fig. 1. Schematic representations of triggered naltrexone delivery systems. (A) naltrexone dispersed in polymer disk; (B) naltrexone dispersed in microspheres.
1
6a
3
Fig. 2. EfTect of size of ester group in
esterified cupolymdrs
of methyl vinyl ether maleic anhydride copolymers on rate of release of hydrocortisone. isks at pH7.4 and 37°C. Drug loading 1Owt %.
Morphine-triggered delivery system
d
OY 0
*
I
*
20
TIME
I
40 -
-
I
60
forheroin addiction
111
1
80 TIME
minutes
Fig. 3. Release of naltrexone from cellulose acetate phthalate microspheres in release medium at pH 7.4 and 37 “C.
the proper alkyl group allows control over the rate of polymer erosion and concomitant rate of drug release. ‘The variation in rate of drug release with changes in size of alkyl group is shown in Fig. 2. An additional benefit of this copolymer is that erosion occurs with no polymer backbone cleavage. e eroded polymer remains as a high molecular weight material which is retained within the device thus simplifying toxicological testing. For this device we have selected the n-octyl half ester. For the rapid release system we have selected a cellulose: acetate phthalate polymer principally because of the ease with which microspheres could be prepared by a coacervation-phase separation method (described later). Also, when naltrexonecontaining microspheres were placed in a pH 7.4 buffer, rapid release as shown in Fig. 3 was observed. The structure of cellulose acetate phthalate is shown below:
_
minutes
Fig. 4. Enzymatic hydrolysis of various triglyceride emulsions by 2.5 pg C. viscosum; (H) trilaurin, (0) trimyristin, (O)tripalmitin, (0) tristearin.
system as the most promising combination. Triglycerides are excellent candidates for a hydrophobic protective coating because they are low melting, highly water-insoluble natural products that are rapidly hydrolyzed by the enzyme: hpase to yield glycerol and the corresponding fatty acid.r6 The rate of hydrolysis depends upon the particular lipase used and on the chain length of the fatty acid component as well as its degree of unsaturation. Figure 4 shows the rate of hydrolysis of a number of triglycerides using the hpase C. vi.scosum. Clearly, triglycerides based on trilaurin and trimyristin hydrolyze at considerably higher rates than tripalmitin or tristearin. Based on these data we have selected trilaurin as the most suitable triglyceride. The structure of trilaurin, is shown below: F&-0-CO-(CH&---CH3 TH -0-CO-(CH&
,--43-i,
CH2-0-CO-(CH2),,--Cl-l3
Reversibly inactivated lipase
-A
b--C-CH3 ki
$Z==O
Enzyme-degradable
protective coating
After first investigating a lysozyme-partially deacetylated chitin system7 and an amylase-starch hydrogel we have selected a tridvceride-linase svstem.15 & “d I
Commercially available hpase preparations include lipases from several sources and of varying specificity and hydrolytic activity. After screening a number of lipases, we have selected C. viscosum as our prime candidate. The amino acid composition of this enzyme, determined on a Beckman 6300 amino acid analyzer by the Protein Structure Laboratory at the University of California at Davis, is shown in Table 1. The analysis does not detect tryptophan or cysteine residues and glycoproteins cannot be detected due to the acid hydrolysis step. Further, threonine and serine residue calculations are typically 5% low due to instability during hydrolysis.
Table 1. Amino
Amino Acid Asp Thr Ser ClU
GUY Ala Val Met Ileu Leu Tyr Phe LYS
l!Iis Arg Pro
acid compositional analysis of lipase from C. viscosum’ after acid hydrolysi@ nmjlnject
Mole %
Near Integer
2.0722 2.2520 1.6810 1.5268 2.3161 2.3526 Y.7493 0.0820 0.6066 1.8812 0.5997 0.5924 0.4301 0.4872 0.6239 0.7840
10.34
29 31 23 21 32 33 24
Il.24 8.39 762
11.56 11.74 8.73 0.41 3.00 9.39 2.99 2.96 2.15 2”43 3.11 391
I a 26 8 8 6 7 9 11
a Fluka Biochemicals, Buchs, Switzerland. bPerformed on a Beckman 6300 amino acid analyzer. This analysis does not detect tryptophan or cysteine residues; glycoprotein cannot be detected by this method because it is done by acid hydrolysis; the detection of threonine and serine residue calculations are typically 5% Pow owing to the instability of these residues during hydrolysis.
Enzyme conjugation was carrie out as described previously for amylaset5 y preparing a mixed anhydride from Q3-carbo methylmorphine and then reacting the mixed anhydride with the enzyme. Reaction occurs via the e-amino Hysine groups of the enzyme. The conjugation reaction is shown below:
R-COOH
h-NH,
+
*
II Cl--C-0-iBu
Enz-NH-C--R
Fig. 5. Average number of morphine molecuies per enzyme as a function of the molar ratio of enzyme to ~~rbox~et~y~~o~~~~e as determined by 3H assay
about
6’7 morphines are attached io each in exceknt agreement with the number abk fysine residues
ion of enzymatic activity conjugation is essential for rapid device Michaelis constants were native C. ~k~~~lp2 and the ese constants are the native hpase an
during
-----+
+ iBu--OH
+
where R-COOH is 03-carboxymethylmorphine. The extent of enzyme derivatization was quantified by using 3H-carboxymetbylmorphine. Analysis of the reaction product showed that the number of morphine residues attached per molecule of enzyme was a linear function of the molar ratio of conjugating agent to enzyme, up to a molar ratio of 100 : 1. This relationship is shown in Fig. 5. The number of morphine molecules per enzyme reaches a maximum at 2 200 : 1 molar ratio of 03-carboxymethylmorphine mixed anhydride to lipase and to assure that a maximum number of morphine molecules is attached to the enzyme, a molar ratio of 200 : 1 was used. At that ratio,
Fig. 6. Effect of substrate concentration on rate of enzymatic hydrolysis of trilaurin emulsions for 15 pg native and conjugated C. viscosum; (D) native enzyme, (a) 208 : I con&ate,
Morphine-triggered delivery system for heroin addiction 2 x 105
2
2
Et
1XlQ5
0
10
SUBSTRATE
20
30
40
CONCENTRATION
50
60
- mglml
Fig. 7. Hanes-Woolf plots of substrate concentration [S] versus ratio of [S] and the reaction velocity (w) for the enzymatic hydrolysis of trilaurin emulsions for 15 pg native and conjugated C. viscosum; (0) native enzyme, (u) 200 : 1 conjugate. K,,, = 2.0
mmol/ liter). Thus, the activity of lipase is not significantly altered by the conjugation procedure. To inlhibit enzyme activity, the enzyme-morphine conjugate was complexed with a rabbit morphine antibody. The ability of lipase-morphine conjugates to interact with morphine antibodies was initially assessed with immunodiffusion disks.18 This double diffusion method is also used as a rapid confirmation of the formation of enzyme-morphine derivatives during the conjugation reaction. In this method, morphine antiisera and lipase-morphine conjugate solutions are placed in wells cut in agar and allowed to diffuse toward each other. Upon interaction bletween antibody and antigen, a precipitate forms in the region of the agar where they meet, and these opaque precipitant lines can be readily visualized in the gels. Table 2. Inhibition of 200 : 1 C. vixosum lipase-morphine conjugate
Molar ratio (ZgG : enzyme)
Inhibition (%)
113
The activity of 200 : 1 C. viscssum lipase-morphine conjugate with and without antimorphine antibodies was determined as a function of increasing antisera concentration using the titrimetric assay described previously. These experiments were performed after incubation of the lipase conjugate with rabbit morphine antibodies for 60 min at 37 “C. As seen in Table 2, when increasing molar ratios of IgG were added to the conjugate, an increase in the percent inhibition was noted. Although reproducibility was poor, it is clear that in these experiments high inhibition of enzyme activity can only be achieved at high antibody concentrations. Unfortunately, a high concentrati.on of antibodies on the enzyme is not desirable b’ecause the antibody-blocked enzyme in a successful device must not only have virtually no residual activity but can also be reactivated with very low concentrations of external morphine. Since reactivation occurs by morphine displacement of the enzyme conjugate from the antibody, a high number of antibodies on the enzyme will result in a number of non-proiductive displacements, that is, morphine molecules will be consumed by displacing antibodies that are :not close to the enzyme-active site. Thus, an optimized device should contain a miminum number of antibodies, and ideally, antibodies should only be bound close to the enzyme-active site. Reactivation studies were conducted with several of these inhibited samples by adding an appropriate volume of freshly prepared aqueous morphine sulfate for final morphine concentrations of 1 x 1W5 M, 1 x 10e6 M, or I x IV7 M in the reaction vessel. Even though the reactivation was conducted with enzymes having a large excess of antibody, for the above decreasing concentrations of morphine sulfate, 70%, 804%, ,and 52.1% reactivation was noted, respectively. Current efforts are directed towards achieving reactivation with lower morphine concentrations.
1.8 : 1
19
3.0 : 1
15 21.8
Macroporous membrane
3.7 : 1
10.6 22.3
6.2 : 1
51.0 53.5
10 : 1
48.7 56.3
16 : 1
83.9 92.4
20 : 1
97.6 98.5 99-3
The semipermeable membrane shown in Fig. 1 serves a dual function. First, it excludes the passage of high-molecular-weight compounds and thus retains in its interior the conjugated enzyme, the antibody, and the bioerodible pollymer. Second, it excludes the passage of endogenous enzymes that would act on the protective coating in the device. Requirements for the membrane include: predictable molecular-weight cutoff, biocompatibility, low-protein-binding, and ease of handling for facile
114
K. V. Roskos.
device manufacture. Candidate membranes fulfilling these requirements included: polysulfone membranes (Millpore@), polycarbonate membranes (Costar-Nucleopore@ >, and other cellulosic membranes such as cellulose acetate membranes and cellulose ester asymmetric membranes in addition to regenerated cellulose membranes (Spectrum@). These membranes were evaluated for differences in permeability and time-lag characteristics for naltrexone and morphine and for their ability to exclude the lipase-morphine conjugate and the bioerodible polymer. All of the cellulosic and the poly(sulfone) membranes fulfilled the requirements. However, the poly(carbonate) membrane did not meet the molecular-weight cutoff minimum.
J. A. Teft,
J. ~-eI~e~
Fig. 8. Cumulative release of naltrexorie from an n-o&y! half ester of methyl vinyl ether/maleik: anhydride copolymer ar pH 7.4 and 37 “C. Polymer disks 7 mm diameter containing 10 wt % naltrexone; (a) uncoated disk in presence of 150 pg lipase, (m) coated disk in presence of 950i;g lipase, ( coated disk in absence of lipase of 5 ml.
kinetics
Device development
of release of
Sustained delivery devices
Fabrication of these devices involves fabricating naltrexone-containing film, punching disks from the film and coating the disks with trilaurin. Polymer films containing 10 wt % naltrexone were prepared by dissolving naltrexone-free base in a 3 : 7 ratio of 2-ethoxyethyl acetate and 4-methyl-2-pentanone (1.62 g and 3.78 g, respectively). Following complete drug dissolution, 0.6 gn-octyl half ester of methyl vinyl ether-maleic anhydride copolymer was added to the drug solution. Films were cast in Teflonlined plates and the solvent allowed to evaporate slowly in a chamber saturate with a 3 : 7 ratio of 2-ethoxyethyl acetate and 4-methyl-2-pentanone for approximately 10 days and then vacuum dried overnight. Disks (7mm d meter) were punched using a specially designed sk puncher heated to 142 “C. The disks were then coated using the following procedure: a stainless steel mold containing eight asymmetric holes (1.1 cm top x 3 mm depth x 0.9 cm bottom) was sandwiched between two sheets of Kapton@ polyimide release film and placed between platens heated to 31 “C in a Carver Press. Powdered triglyceride (100 mg) was added to the bottom of the mold cavities, fallowed by polymer disks. The remaining portion of the mold was then filled with additional triglyceride (100 mg). Following pressing at 24 kpsi for 30min, the coated disks are easily removed from the mold. Release of naltrexone from coated and uncoated disks in the presence and absence of lipase is shown in Fig. 8. As expected, the trilaurin coating is highly effective in preventing release of naltrexone from
sks was consid
the action of
the surface of the disk, this retard erosion of
~xt~erne~y viscous.
This
Morphine-triggered delivery system for heroin addiction
115
prevent the euphoric effect of subsequent administrations. Rapid delivery devices
_l_. 0
,
.
,
6
-
-.
12 TIME
16
24
- hours
Fig. 9. Cumulative release of naltrexone from an n-octyl half ester of methyl vinyl ether/maleic anhydride copolymer at pH 7.4 and 37 “C. Coated polymer disks 7mm diameter containing 10 wt % naltrexone; 150 pg lipase added to 5 ml at time zero.
INTERNAL
Fabrication of these devices involved preparing naltrexone-containing microspheres and spray-coating the microspheres with trilaurin. Microspheres were prepared using a coacervation-plhase separation method shown in Fig. 10. This procedure produced free-flowing microspheres with size ranges between 250 and SOOpm and a naltrexone loading between 10 and 20%. An SEM of these microspheres (250-355 pm) is shown in Fig. 11. Coating gram quantities of microparticles in the 250-355pm range with a uniform protective coating was a difficult process that required con-
PHASE
‘zyzg_
add a smallamcunt of water
I?-( ) +---(Z)
stiruntilNAL is a fine dispersion
+
EXTERNAL
PHASE
stirwith an overhead stirrerat 400 rpm t
pipet into externalphase
stirwhile chbroformis
room temperature
Fig. 10.Encapsulation
protocol.
Fig. 11. Scanning electron micrograph (SEM) acetate phthalate microspheres.
of cellulose
siderable process development, and in this work we have modified a miniature air-suspension spraycoating apparatus described in t e literature.‘9’20 A schematic of the apparatus is shown in Fig. 12. The body of the miniature spray-coating apparatus is a glass column 2.5 cm i.d. and 219 cm long. A bulb of 5 1 cm internal diameter was provided in the re ’ where microparticles were exposed to the spray.
Fig. 13. SEM of a trilaurin/PEA spray-coated naitre~oncloaded cellulose acetate phthalate microsphere
top
of the colun?nnwas stainless steel s
closed by a Tefbn hd eonkam/XII mesh). The ‘base gh a standard joist,
hot air into the toll
the socket of the j
eating was varied, Polymer
Air
TMWl
Cap
Glass
,
Column Mesh Wire
Mesh Wire
01 Teflon
Fig. 12. Schematic
Cap
Patrexone release from :xilaurin/PEA cellulose acetate phthalate microspheres in release medium at p
CumuiaQive
representation of miniature apparatus.
spray-coating
spray-coated
Morphine-triggered delivery system for heroin addiction
0.5
2.5
6
117
24
Time (Hours) Fig. 15. Enzyme-mediated naltrexone release from trilaurin spray-coated cellulose acetate phthalate microspheres in release medium at pH 7.4 and 37 “C as a function of enzyme amount. Enzyme used was C. viscosumwith an average of six morphines covalently attached. Indicated amounts of enzyme added io 5 ml of buffer containing 100 mg microspheres.
incorporated into the triglyceride. Aliphatic polyesters are widely distributed in nature and are substrates for certain lipases.21122 Reproducibly stable coatings were achieved by six successive increasing sprayings, each coating containing amounts of PEA (0.5, 0.6, 0.7, 0.81, 0.9, and l*O%, respectively). A final coating of PEA alone was then applied to give added hardness to the coat. An SEM of a microsphere coated by this procedure is shown in Fig. 13. As shown in Fig. 14, when the coat.ed microspheres are immersed in a pH 7.4 buffer at 37 “C, a small amount of naltrexone is able to diffuse out, but the amount released is less than 10% throughout a total of 28 days, the desired lifetime of the delivery device. Figure 15 shows the effect of the enzyme C. viscosunv2on the protective coating. In this work, the lipase-morphine enzyme conjugate, where an average of six morphines are bound to one enzyme., was used. As shown, addition of the enzyme to a suspension of the coated CAP microspheres in pH7~4 phosphate buffer produced an almost immediate effect, and at 30min, the first time-point of the experiment, about 30% of naltrexone was released as compared to little or no release in the absence of the enzyme conjugate. The amount of naltrexone released appears to be only mloderately dependent upon enzyme concentration, at least at the two concentrations used. In previouls experiments we have found that enzyme amounts I 12.5 ,ug are ineffective, and that no further increase in the rate of drug release is noted when xtnounts are increased about 150 pug (unpublished results).
Fig. 16. Trilaurin/PEA spray-coated cellulose acetate phthalate microspheres incubated with 25 pg C. vkscosum-morphine conjugate; (a) control, (b) coating after O.Sh, (c) coating after 6 h.
The effect of C. viscosum on tbie coating can be seen in Figs 16 and 17 which shows coating degradation at 0.5 and 6h for enzyme amounts of 25 and 150 pg. Although the figures show a noticeable difference in the amount of degraded coating at the two enzyme concentrations, particularly after 6 h,
absence of external trexone is released,
1. Washton, A. w/p., Gold, use of naltrexone in
S. $r Pottash, A. C , Successfui icted physicians and business
2.
(b)
Fig. 17. Trilaurin/PEA spray-coated cellulose acetate phthalate microspheres incubated with 1.50pg C. viscosum-morphine conjugate; (a) control, (b) coating after @5h, (c) coating aFter 6h.
no significant difference in naltrexone release rate is noted, indicating that rapid release of naltrexone can take place from even a slightly degraded coating. These are encouraging results since even a slight degradation of the coating will result in rapid triggering of drug release.
3. O’Brien, C. P., Greenstein, R., Ternes, J. 6%Woody, 6;. E., Clinical ~~a~a~o~o~y of narcotic antagonists. Annalr IV. t’. Acad. Ski., 31 (19783 23240. J. 4. Chien, Y. W., Long acting parenteral drug formulations. OE.,39 41981) 106 39 5. : J,, Controlled release of nahrexone ear poly [ortho esters). J. Coonlrolled Release, 14 (1990) 29L8. M. E., Sch 6. Rubenstein, S. & Uhm, E. F’.; Homogeneous enzyme say. A new immunochemical technique. Biochepm. iophyys.Res. Commun., 47 (1972) 846-51. 7 Pangburn, S. H., Trescony, P. V. a%H&r, f., Eysozyme degradation of partially deacetylated chitin, i%s fihns and hydrogels. ~io~~te~ia~~, 3 (1982) 105--K 8. Heller, J.: Pangburn, S. H. & Penhale, D. W E-E., EJse of bioerodible polymers in self-regulateo dr systems. In CobntvolEed Release Techmlogy, ed, W. R. Good. ACS Symposium Series No. 34 Chemical Society, Washington, DC, 1981, pp~ 172~-87, 9. Heller, 4., Penhale, D. drng delivery systems. In. PFVC. 3rd International Conference on Polymers ifi Mediche, Biomedi6caE and ~~~~~~ce~ticQ1 A~~~icat~o~5~ed. L. Nidais, C. Migliaresi & E. Chielini. EYsevier Science Pubhshers, Amsterdam, The Netherlands. 1988, pp, 175--M% S., Chemically self-regulated 10. Heher, systems--A review. J. ~~~t~~~~ed~ele~se, II Heller, J., Pangburn, S. M, & Roskos, K. of a triggered naltrexone dehvery device for the treatment of narcotic addiction. In ~0ppic.sin ~~~~~~ce~t~c~~Sciences reimer, D. 1. A. Crommehn & K. K
Morphine-triggered delivery system for heroin addiction
12.
13.
14.
15.
16.
Midha. Amsterdam Medical Press, Noordwijk, The Netherlands, 1989, pp. 39949. Roskos, K. V., Tefft, J. A. & Heller, J., Development of a morphine-triggered naltrexone delivery system, J. Controlled Release, 21 (1992) 191-200. Tefft, J. A., Roskos, K. V. & Heller, J., The effect of lipase on the release of naltrexone from triglyceride-coated celluJ. Biomed. Mater. Res., 26 (1992) lose acetate microspheres, 713-2.4. Heller, J., Baker, R. W., Gale, R. M. & Rodin, J. O., Controlled drug release by polymer dissolution I. Partial esters of maleic anhydride copolymers. Properties and theory. J. Appl. Polymer Sci., 22 (1978) 1991-2009. Heller, J., Pangburn, S. H., Trescony, P. V. & Roskos, K. V., Development of enzymatically degradable protective coatings for use in triggered drug delivery systems. Biomaterials, 11 (1990) 345550. Desnuelle, P., The Lipases. In The Enzymes, Vol III, 3rd
17. 18. 19.
20.
21. 22.
119
edn., ed. P. D. Boyer. Academic Press, New York, USA, 1972, pp. 575-616. Segel, I. H., Biochemical Calculations. John Wiley and Sons, New York, USA, 1976, pp. 214-41. Ouchterlony, O., Diffusion-in-gel methods for immunological analysis. Prog. Allergy, 5 (1958) l-78. Baveja, S. K., Ranga Rao, K. V. & Singh, A., Design and evaluation of a miniature air-suspens,ion coating apparatus. J. Pharm. Pharmacol., 35 (1983) 475-6. Ranga, K. V. & Buri, P., Studies of factors affecting the Proc. Znt. Symp. bioadhesion of coated microparticles. Control. Rel. Bioactive Mater., 15 (1988) 10334. Tokiwa, Y. & Suzuki, T., Hydrolysis of polyesters by lipase. Nature, 270 (1988) 76-8. Tokiwa, Y., Suzuki, T. & Takeda, K. Hydrolysis of polyesters by Rhizopus arrhizus lipase. Agric. BioE. Chem., 50 (1986) 1323-5.
A Morphine-Triggered Delivery System Usefill in the Treatmint of Heroin Addiction K. V. Roskos, J. A. Tefft & J. Heller* Controlled
Release and Biomedical
Polymers
Department,
SRI International,
Menlo
Park,
California
94025,
USA
Abstract: The ultimate objective of this work is to develop a device that can be triggered by morphine to release naltrexone. Two device configurations are described. In one configuration, naltrexone is dispersed in cellulose acetate phthalate microspheres which are then spray-coated with trilaurin. In the other configuration, naltrexone is dispersed in an n-octyl half ester of methyl vinyl ether and maleic anhydride copolymer and the mixture fabricated into a disk which is then coated with trilaurin. The microspheres are designed to release naltrexone abruptly while the disks are designed to release naltrexone at a constant rate over a two week period. The microspheres, or the disk along with a reversibly inactivated lipase are plased inside a semipermeable membrane that allows free passage of morphine and naltrexone but excludes the higher molecular weight components of the device. Reversible inactivation of lipase is achieved by covalent attachment of morphine and complexing with morphine antibody. Activation of the device occurs by diffusion of morphine into the device and displacing the lipasemorphine conjugate from the antibody. The activated lipase then removes the trilaurin protective coating, thus triggering naltrexone release.
INTRODUCTION
Unfortunately, readdiction is an ever present danger, even for individuals that have been opiate-free for a long time. For this reason, the development of a device that would guard against this possibility has been under development in our laboratories for a number of years. In this device, the narcotic antagonist naltrexone is dispersed in ai rate-controlling bioerodible polymer surrounded by an enzymedegradable protective coating which prevents release of naltrexone from the device. The coated device is in turn surrounded by a reversibly inaLctivated enzyme. This enzyme in its inactivated st,ate is unable to degrade the coating, but once activated is able to rapidly remove the protective coating. The reversible inactivation of the enzyme is achieved by first preparing an enzyme-morphine conjugate and then complexing the conjugate with a morphine antibody.6 Because the antibody is a fairly large molecule, access of the substrate to the enzyme’s active site is sterically inhibited, thus effectively rendering the enzyme inactive. Triggering of naltrexone release is initiated by the appearance of morphine in the tissue surrounding the device, diffusion into the
Heroin addiction remains a very serious, worldwide problem which is now widely prevalent in all segments of our society.’ Although an effective cure remains; elusive, the use of narcotic antagonists such as naltrexone is a promising approach.2’3 Such antagonists are able to block the euphoric effect of heroin and by maintaining an effective opiate-receiptor blockade, readdiction is not possible. Currently, naltrexone is administered in oral dosage form which is effective for two to three days.4 However, because discontinuation of naltrexone therapy is without side-effects, maintenance of therapy requires strong motivation. Therefore, a therapy which removes the decisionmaking, process from the patient is highly desirable and sustained release formulations that are implanted deeply enough in the tissues so that removal by the patient is impossible are currently under development.5 *To whom
correspondence
should be addressed.
109 ClinicalMaterials O267-6605/93/$6.000
1993 Elsevier
Science Publishers
Ltd, England
oskos, J. A. Tefft, 9.
110
device, and dissociation of the enzyme-haptenantibody complex, thus rendering the enzyme active. The development of a morphine-triggered naltrexone delivery system has been under development in our laboratories for a number of years.7-“’ and after considerable experimentation we have selected a lipase-degradable trilaurin as the protective coating surrounding a rate-controlling bioerodible polymer with dispersed naltrexone. Because such a device contains a number of antigenic compounds, it is surrounded by a macroporous membrane that is freely permeable to morphine and naltrexone but which excludes the enzyme and the morphine antibody. Two device configurations are currently under investigation. In one configuration, the ratecontrolling polymer with dispersed naltrexone is fabricated into a disk, which is then coate trilaurin. l2 Following activation, this device is designed to release naltrexone over a two-week period. In the other configuration, the rate-controlling polymer with dispersed naltrexone is fabricated into microspheres which are then coated with trilaurin. ’3 Following activation, this device is designed to begin the release of naltrexone in no more than one hour. A schematic representation of both configurations is shown in Fig. 1. In this manuscript we will summarize the current status of development of these devices.
ment of the two specific
A device designed
devices,
to release naltrexone for two weeks
and maleic an previously to ~~~e~~~ su comitant release of an n
RESULTS AND DISCUSSION s&lble
Regardless of the final configuration, development of these devices requires the development of four separate components which are then assembled into a final device. These components are (1) ratecontrolling bioerodible polymer, (2) enzyme-
the
rate sf
a&y1
group
ePosPon
on t&e
ester portion of the co~o~y~e~~ so that selection of
MICROPOROUS MEMBRANE PROTECTIVE COATING NALTREXONE CORE
ANTIBODY BLOCKED ENZYME
NALTREXONE RELEASE FOR TWO WEEKS
NALTREXONE RELEASE IN LESS THAN ONE HOUR
Fig. 1. Schematic representations of triggered naltrexone delivery systems. (A) naltrexone dispersed in polymer disk; (B) naltrexone dispersed in microspheres.
1
6a
3
Fig. 2. EfTect of size of ester group in
esterified cupolymdrs
of methyl vinyl ether maleic anhydride copolymers on rate of release of hydrocortisone. isks at pH7.4 and 37°C. Drug loading 1Owt %.
Morphine-triggered delivery system
d
OY 0
*
I
*
20
TIME
I
40 -
-
I
60
forheroin addiction
111
1
80 TIME
minutes
Fig. 3. Release of naltrexone from cellulose acetate phthalate microspheres in release medium at pH 7.4 and 37 “C.
the proper alkyl group allows control over the rate of polymer erosion and concomitant rate of drug release. ‘The variation in rate of drug release with changes in size of alkyl group is shown in Fig. 2. An additional benefit of this copolymer is that erosion occurs with no polymer backbone cleavage. e eroded polymer remains as a high molecular weight material which is retained within the device thus simplifying toxicological testing. For this device we have selected the n-octyl half ester. For the rapid release system we have selected a cellulose: acetate phthalate polymer principally because of the ease with which microspheres could be prepared by a coacervation-phase separation method (described later). Also, when naltrexonecontaining microspheres were placed in a pH 7.4 buffer, rapid release as shown in Fig. 3 was observed. The structure of cellulose acetate phthalate is shown below:
_
minutes
Fig. 4. Enzymatic hydrolysis of various triglyceride emulsions by 2.5 pg C. viscosum; (H) trilaurin, (0) trimyristin, (O)tripalmitin, (0) tristearin.
system as the most promising combination. Triglycerides are excellent candidates for a hydrophobic protective coating because they are low melting, highly water-insoluble natural products that are rapidly hydrolyzed by the enzyme: hpase to yield glycerol and the corresponding fatty acid.r6 The rate of hydrolysis depends upon the particular lipase used and on the chain length of the fatty acid component as well as its degree of unsaturation. Figure 4 shows the rate of hydrolysis of a number of triglycerides using the hpase C. vi.scosum. Clearly, triglycerides based on trilaurin and trimyristin hydrolyze at considerably higher rates than tripalmitin or tristearin. Based on these data we have selected trilaurin as the most suitable triglyceride. The structure of trilaurin, is shown below: F&-0-CO-(CH&---CH3 TH -0-CO-(CH&
,--43-i,
CH2-0-CO-(CH2),,--Cl-l3
Reversibly inactivated lipase
-A
b--C-CH3 ki
$Z==O
Enzyme-degradable
protective coating
After first investigating a lysozyme-partially deacetylated chitin system7 and an amylase-starch hydrogel we have selected a tridvceride-linase svstem.15 & “d I
Commercially available hpase preparations include lipases from several sources and of varying specificity and hydrolytic activity. After screening a number of lipases, we have selected C. viscosum as our prime candidate. The amino acid composition of this enzyme, determined on a Beckman 6300 amino acid analyzer by the Protein Structure Laboratory at the University of California at Davis, is shown in Table 1. The analysis does not detect tryptophan or cysteine residues and glycoproteins cannot be detected due to the acid hydrolysis step. Further, threonine and serine residue calculations are typically 5% low due to instability during hydrolysis.
Table 1. Amino
Amino Acid Asp Thr Ser ClU
GUY Ala Val Met Ileu Leu Tyr Phe LYS
l!Iis Arg Pro
acid compositional analysis of lipase from C. viscosum’ after acid hydrolysi@ nmjlnject
Mole %
Near Integer
2.0722 2.2520 1.6810 1.5268 2.3161 2.3526 Y.7493 0.0820 0.6066 1.8812 0.5997 0.5924 0.4301 0.4872 0.6239 0.7840
10.34
29 31 23 21 32 33 24
Il.24 8.39 762
11.56 11.74 8.73 0.41 3.00 9.39 2.99 2.96 2.15 2”43 3.11 391
I a 26 8 8 6 7 9 11
a Fluka Biochemicals, Buchs, Switzerland. bPerformed on a Beckman 6300 amino acid analyzer. This analysis does not detect tryptophan or cysteine residues; glycoprotein cannot be detected by this method because it is done by acid hydrolysis; the detection of threonine and serine residue calculations are typically 5% Pow owing to the instability of these residues during hydrolysis.
Enzyme conjugation was carrie out as described previously for amylaset5 y preparing a mixed anhydride from Q3-carbo methylmorphine and then reacting the mixed anhydride with the enzyme. Reaction occurs via the e-amino Hysine groups of the enzyme. The conjugation reaction is shown below:
R-COOH
h-NH,
+
*
II Cl--C-0-iBu
Enz-NH-C--R
Fig. 5. Average number of morphine molecuies per enzyme as a function of the molar ratio of enzyme to ~~rbox~et~y~~o~~~~e as determined by 3H assay
about
6’7 morphines are attached io each in exceknt agreement with the number abk fysine residues
ion of enzymatic activity conjugation is essential for rapid device Michaelis constants were native C. ~k~~~lp2 and the ese constants are the native hpase an
during
-----+
+ iBu--OH
+
where R-COOH is 03-carboxymethylmorphine. The extent of enzyme derivatization was quantified by using 3H-carboxymetbylmorphine. Analysis of the reaction product showed that the number of morphine residues attached per molecule of enzyme was a linear function of the molar ratio of conjugating agent to enzyme, up to a molar ratio of 100 : 1. This relationship is shown in Fig. 5. The number of morphine molecules per enzyme reaches a maximum at 2 200 : 1 molar ratio of 03-carboxymethylmorphine mixed anhydride to lipase and to assure that a maximum number of morphine molecules is attached to the enzyme, a molar ratio of 200 : 1 was used. At that ratio,
Fig. 6. Effect of substrate concentration on rate of enzymatic hydrolysis of trilaurin emulsions for 15 pg native and conjugated C. viscosum; (D) native enzyme, (a) 208 : I con&ate,
Morphine-triggered delivery system for heroin addiction 2 x 105
2
2
Et
1XlQ5
0
10
SUBSTRATE
20
30
40
CONCENTRATION
50
60
- mglml
Fig. 7. Hanes-Woolf plots of substrate concentration [S] versus ratio of [S] and the reaction velocity (w) for the enzymatic hydrolysis of trilaurin emulsions for 15 pg native and conjugated C. viscosum; (0) native enzyme, (u) 200 : 1 conjugate. K,,, = 2.0
mmol/ liter). Thus, the activity of lipase is not significantly altered by the conjugation procedure. To inlhibit enzyme activity, the enzyme-morphine conjugate was complexed with a rabbit morphine antibody. The ability of lipase-morphine conjugates to interact with morphine antibodies was initially assessed with immunodiffusion disks.18 This double diffusion method is also used as a rapid confirmation of the formation of enzyme-morphine derivatives during the conjugation reaction. In this method, morphine antiisera and lipase-morphine conjugate solutions are placed in wells cut in agar and allowed to diffuse toward each other. Upon interaction bletween antibody and antigen, a precipitate forms in the region of the agar where they meet, and these opaque precipitant lines can be readily visualized in the gels. Table 2. Inhibition of 200 : 1 C. vixosum lipase-morphine conjugate
Molar ratio (ZgG : enzyme)
Inhibition (%)
113
The activity of 200 : 1 C. viscssum lipase-morphine conjugate with and without antimorphine antibodies was determined as a function of increasing antisera concentration using the titrimetric assay described previously. These experiments were performed after incubation of the lipase conjugate with rabbit morphine antibodies for 60 min at 37 “C. As seen in Table 2, when increasing molar ratios of IgG were added to the conjugate, an increase in the percent inhibition was noted. Although reproducibility was poor, it is clear that in these experiments high inhibition of enzyme activity can only be achieved at high antibody concentrations. Unfortunately, a high concentrati.on of antibodies on the enzyme is not desirable b’ecause the antibody-blocked enzyme in a successful device must not only have virtually no residual activity but can also be reactivated with very low concentrations of external morphine. Since reactivation occurs by morphine displacement of the enzyme conjugate from the antibody, a high number of antibodies on the enzyme will result in a number of non-proiductive displacements, that is, morphine molecules will be consumed by displacing antibodies that are :not close to the enzyme-active site. Thus, an optimized device should contain a miminum number of antibodies, and ideally, antibodies should only be bound close to the enzyme-active site. Reactivation studies were conducted with several of these inhibited samples by adding an appropriate volume of freshly prepared aqueous morphine sulfate for final morphine concentrations of 1 x 1W5 M, 1 x 10e6 M, or I x IV7 M in the reaction vessel. Even though the reactivation was conducted with enzymes having a large excess of antibody, for the above decreasing concentrations of morphine sulfate, 70%, 804%, ,and 52.1% reactivation was noted, respectively. Current efforts are directed towards achieving reactivation with lower morphine concentrations.
1.8 : 1
19
3.0 : 1
15 21.8
Macroporous membrane
3.7 : 1
10.6 22.3
6.2 : 1
51.0 53.5
10 : 1
48.7 56.3
16 : 1
83.9 92.4
20 : 1
97.6 98.5 99-3
The semipermeable membrane shown in Fig. 1 serves a dual function. First, it excludes the passage of high-molecular-weight compounds and thus retains in its interior the conjugated enzyme, the antibody, and the bioerodible pollymer. Second, it excludes the passage of endogenous enzymes that would act on the protective coating in the device. Requirements for the membrane include: predictable molecular-weight cutoff, biocompatibility, low-protein-binding, and ease of handling for facile
114
K. V. Roskos.
device manufacture. Candidate membranes fulfilling these requirements included: polysulfone membranes (Millpore@), polycarbonate membranes (Costar-Nucleopore@ >, and other cellulosic membranes such as cellulose acetate membranes and cellulose ester asymmetric membranes in addition to regenerated cellulose membranes (Spectrum@). These membranes were evaluated for differences in permeability and time-lag characteristics for naltrexone and morphine and for their ability to exclude the lipase-morphine conjugate and the bioerodible polymer. All of the cellulosic and the poly(sulfone) membranes fulfilled the requirements. However, the poly(carbonate) membrane did not meet the molecular-weight cutoff minimum.
J. A. Teft,
J. ~-eI~e~
Fig. 8. Cumulative release of naltrexorie from an n-o&y! half ester of methyl vinyl ether/maleik: anhydride copolymer ar pH 7.4 and 37 “C. Polymer disks 7 mm diameter containing 10 wt % naltrexone; (a) uncoated disk in presence of 150 pg lipase, (m) coated disk in presence of 950i;g lipase, ( coated disk in absence of lipase of 5 ml.
kinetics
Device development
of release of
Sustained delivery devices
Fabrication of these devices involves fabricating naltrexone-containing film, punching disks from the film and coating the disks with trilaurin. Polymer films containing 10 wt % naltrexone were prepared by dissolving naltrexone-free base in a 3 : 7 ratio of 2-ethoxyethyl acetate and 4-methyl-2-pentanone (1.62 g and 3.78 g, respectively). Following complete drug dissolution, 0.6 gn-octyl half ester of methyl vinyl ether-maleic anhydride copolymer was added to the drug solution. Films were cast in Teflonlined plates and the solvent allowed to evaporate slowly in a chamber saturate with a 3 : 7 ratio of 2-ethoxyethyl acetate and 4-methyl-2-pentanone for approximately 10 days and then vacuum dried overnight. Disks (7mm d meter) were punched using a specially designed sk puncher heated to 142 “C. The disks were then coated using the following procedure: a stainless steel mold containing eight asymmetric holes (1.1 cm top x 3 mm depth x 0.9 cm bottom) was sandwiched between two sheets of Kapton@ polyimide release film and placed between platens heated to 31 “C in a Carver Press. Powdered triglyceride (100 mg) was added to the bottom of the mold cavities, fallowed by polymer disks. The remaining portion of the mold was then filled with additional triglyceride (100 mg). Following pressing at 24 kpsi for 30min, the coated disks are easily removed from the mold. Release of naltrexone from coated and uncoated disks in the presence and absence of lipase is shown in Fig. 8. As expected, the trilaurin coating is highly effective in preventing release of naltrexone from
sks was consid
the action of
the surface of the disk, this retard erosion of
~xt~erne~y viscous.
This
Morphine-triggered delivery system for heroin addiction
115
prevent the euphoric effect of subsequent administrations. Rapid delivery devices
_l_. 0
,
.
,
6
-
-.
12 TIME
16
24
- hours
Fig. 9. Cumulative release of naltrexone from an n-octyl half ester of methyl vinyl ether/maleic anhydride copolymer at pH 7.4 and 37 “C. Coated polymer disks 7mm diameter containing 10 wt % naltrexone; 150 pg lipase added to 5 ml at time zero.
INTERNAL
Fabrication of these devices involved preparing naltrexone-containing microspheres and spray-coating the microspheres with trilaurin. Microspheres were prepared using a coacervation-plhase separation method shown in Fig. 10. This procedure produced free-flowing microspheres with size ranges between 250 and SOOpm and a naltrexone loading between 10 and 20%. An SEM of these microspheres (250-355 pm) is shown in Fig. 11. Coating gram quantities of microparticles in the 250-355pm range with a uniform protective coating was a difficult process that required con-
PHASE
‘zyzg_
add a smallamcunt of water
I?-( ) +---(Z)
stiruntilNAL is a fine dispersion
+
EXTERNAL
PHASE
stirwith an overhead stirrerat 400 rpm t
pipet into externalphase
stirwhile chbroformis
room temperature
Fig. 10.Encapsulation
protocol.
Fig. 11. Scanning electron micrograph (SEM) acetate phthalate microspheres.
of cellulose
siderable process development, and in this work we have modified a miniature air-suspension spraycoating apparatus described in t e literature.‘9’20 A schematic of the apparatus is shown in Fig. 12. The body of the miniature spray-coating apparatus is a glass column 2.5 cm i.d. and 219 cm long. A bulb of 5 1 cm internal diameter was provided in the re ’ where microparticles were exposed to the spray.
Fig. 13. SEM of a trilaurin/PEA spray-coated naitre~oncloaded cellulose acetate phthalate microsphere
top
of the colun?nnwas stainless steel s
closed by a Tefbn hd eonkam/XII mesh). The ‘base gh a standard joist,
hot air into the toll
the socket of the j
eating was varied, Polymer
Air
TMWl
Cap
Glass
,
Column Mesh Wire
Mesh Wire
01 Teflon
Fig. 12. Schematic
Cap
Patrexone release from :xilaurin/PEA cellulose acetate phthalate microspheres in release medium at p
CumuiaQive
representation of miniature apparatus.
spray-coating
spray-coated
Morphine-triggered delivery system for heroin addiction
0.5
2.5
6
117
24
Time (Hours) Fig. 15. Enzyme-mediated naltrexone release from trilaurin spray-coated cellulose acetate phthalate microspheres in release medium at pH 7.4 and 37 “C as a function of enzyme amount. Enzyme used was C. viscosumwith an average of six morphines covalently attached. Indicated amounts of enzyme added io 5 ml of buffer containing 100 mg microspheres.
incorporated into the triglyceride. Aliphatic polyesters are widely distributed in nature and are substrates for certain lipases.21122 Reproducibly stable coatings were achieved by six successive increasing sprayings, each coating containing amounts of PEA (0.5, 0.6, 0.7, 0.81, 0.9, and l*O%, respectively). A final coating of PEA alone was then applied to give added hardness to the coat. An SEM of a microsphere coated by this procedure is shown in Fig. 13. As shown in Fig. 14, when the coat.ed microspheres are immersed in a pH 7.4 buffer at 37 “C, a small amount of naltrexone is able to diffuse out, but the amount released is less than 10% throughout a total of 28 days, the desired lifetime of the delivery device. Figure 15 shows the effect of the enzyme C. viscosunv2on the protective coating. In this work, the lipase-morphine enzyme conjugate, where an average of six morphines are bound to one enzyme., was used. As shown, addition of the enzyme to a suspension of the coated CAP microspheres in pH7~4 phosphate buffer produced an almost immediate effect, and at 30min, the first time-point of the experiment, about 30% of naltrexone was released as compared to little or no release in the absence of the enzyme conjugate. The amount of naltrexone released appears to be only mloderately dependent upon enzyme concentration, at least at the two concentrations used. In previouls experiments we have found that enzyme amounts I 12.5 ,ug are ineffective, and that no further increase in the rate of drug release is noted when xtnounts are increased about 150 pug (unpublished results).
Fig. 16. Trilaurin/PEA spray-coated cellulose acetate phthalate microspheres incubated with 25 pg C. vkscosum-morphine conjugate; (a) control, (b) coating after O.Sh, (c) coating after 6 h.
The effect of C. viscosum on tbie coating can be seen in Figs 16 and 17 which shows coating degradation at 0.5 and 6h for enzyme amounts of 25 and 150 pg. Although the figures show a noticeable difference in the amount of degraded coating at the two enzyme concentrations, particularly after 6 h,
absence of external trexone is released,
1. Washton, A. w/p., Gold, use of naltrexone in
S. $r Pottash, A. C , Successfui icted physicians and business
2.
(b)
Fig. 17. Trilaurin/PEA spray-coated cellulose acetate phthalate microspheres incubated with 1.50pg C. viscosum-morphine conjugate; (a) control, (b) coating after @5h, (c) coating aFter 6h.
no significant difference in naltrexone release rate is noted, indicating that rapid release of naltrexone can take place from even a slightly degraded coating. These are encouraging results since even a slight degradation of the coating will result in rapid triggering of drug release.
3. O’Brien, C. P., Greenstein, R., Ternes, J. 6%Woody, 6;. E., Clinical ~~a~a~o~o~y of narcotic antagonists. Annalr IV. t’. Acad. Ski., 31 (19783 23240. J. 4. Chien, Y. W., Long acting parenteral drug formulations. OE.,39 41981) 106 39 5. : J,, Controlled release of nahrexone ear poly [ortho esters). J. Coonlrolled Release, 14 (1990) 29L8. M. E., Sch 6. Rubenstein, S. & Uhm, E. F’.; Homogeneous enzyme say. A new immunochemical technique. Biochepm. iophyys.Res. Commun., 47 (1972) 846-51. 7 Pangburn, S. H., Trescony, P. V. a%H&r, f., Eysozyme degradation of partially deacetylated chitin, i%s fihns and hydrogels. ~io~~te~ia~~, 3 (1982) 105--K 8. Heller, J.: Pangburn, S. H. & Penhale, D. W E-E., EJse of bioerodible polymers in self-regulateo dr systems. In CobntvolEed Release Techmlogy, ed, W. R. Good. ACS Symposium Series No. 34 Chemical Society, Washington, DC, 1981, pp~ 172~-87, 9. Heller, 4., Penhale, D. drng delivery systems. In. PFVC. 3rd International Conference on Polymers ifi Mediche, Biomedi6caE and ~~~~~~ce~ticQ1 A~~~icat~o~5~ed. L. Nidais, C. Migliaresi & E. Chielini. EYsevier Science Pubhshers, Amsterdam, The Netherlands. 1988, pp, 175--M% S., Chemically self-regulated 10. Heher, systems--A review. J. ~~~t~~~~ed~ele~se, II Heller, J., Pangburn, S. M, & Roskos, K. of a triggered naltrexone dehvery device for the treatment of narcotic addiction. In ~0ppic.sin ~~~~~~ce~t~c~~Sciences reimer, D. 1. A. Crommehn & K. K
Morphine-triggered delivery system for heroin addiction
12.
13.
14.
15.
16.
Midha. Amsterdam Medical Press, Noordwijk, The Netherlands, 1989, pp. 39949. Roskos, K. V., Tefft, J. A. & Heller, J., Development of a morphine-triggered naltrexone delivery system, J. Controlled Release, 21 (1992) 191-200. Tefft, J. A., Roskos, K. V. & Heller, J., The effect of lipase on the release of naltrexone from triglyceride-coated celluJ. Biomed. Mater. Res., 26 (1992) lose acetate microspheres, 713-2.4. Heller, J., Baker, R. W., Gale, R. M. & Rodin, J. O., Controlled drug release by polymer dissolution I. Partial esters of maleic anhydride copolymers. Properties and theory. J. Appl. Polymer Sci., 22 (1978) 1991-2009. Heller, J., Pangburn, S. H., Trescony, P. V. & Roskos, K. V., Development of enzymatically degradable protective coatings for use in triggered drug delivery systems. Biomaterials, 11 (1990) 345550. Desnuelle, P., The Lipases. In The Enzymes, Vol III, 3rd
17. 18. 19.
20.
21. 22.
119
edn., ed. P. D. Boyer. Academic Press, New York, USA, 1972, pp. 575-616. Segel, I. H., Biochemical Calculations. John Wiley and Sons, New York, USA, 1976, pp. 214-41. Ouchterlony, O., Diffusion-in-gel methods for immunological analysis. Prog. Allergy, 5 (1958) l-78. Baveja, S. K., Ranga Rao, K. V. & Singh, A., Design and evaluation of a miniature air-suspens,ion coating apparatus. J. Pharm. Pharmacol., 35 (1983) 475-6. Ranga, K. V. & Buri, P., Studies of factors affecting the Proc. Znt. Symp. bioadhesion of coated microparticles. Control. Rel. Bioactive Mater., 15 (1988) 10334. Tokiwa, Y. & Suzuki, T., Hydrolysis of polyesters by lipase. Nature, 270 (1988) 76-8. Tokiwa, Y., Suzuki, T. & Takeda, K. Hydrolysis of polyesters by Rhizopus arrhizus lipase. Agric. BioE. Chem., 50 (1986) 1323-5.