ELSEVIER
`
`Journal of Controlled Release 41 (1996) 237-247
`
`journal of
`controlled
`release
`
`Comparative analysis of oral delivery systems for live rotavirus
`vaccines
`
`Jacqueline D. Duncana'*, Pei Xuan Wang', Catherine M. Harrington', Dennis P. Schafer',
`Yumiko Matsuokab, Jiri F. Mesteckyc, Richard W. Compansb, Miroslav J. Novak'
`'Secretech, Inc., Birmingham, AL 35205, USA
`b Department of Microbiology and Immunology, Emory University, Atlanta, GA 30322, USA
``Department of Microbiology, The University of Alabama at Birmingham, Birmingham, AL 35294, USA
`
`Received 25 July 1995; accepted 5 January 1996
`
`Abstract
`
`In this paper a comparison of delivery systems for a live rotavirus vaccine is presented. The loss of infectivity was
`estimated during incorporation into the delivery systems, and during the subsequent processing steps in the preparation of
`poly(DL-lactide-co-glycolide) microspheres, alginate microcapsules, spray-coated non-pareil seeds, granules, and tablets.
`Incorporation of live rotavirus into DL-PLG microspheres or alginate microcapsules, as well as the application to the surface
`of non-pareil seeds resulted in a complete or significant loss of rotavirus infectivity. In contrast, stabilization of the rotavirus
`vaccine with an excipient blend of cellulose, starch, sucrose and gelatin (30:30:30:10), followed by incorporation into
`granules or tablets, produced outstanding results with only minimal losses of infectivity. Of these two delivery systems
`tablets produced better results. However, the dosage form must be modified into a formulation suitable for immunizing
`infants.
`
`Keywords: Rotavirus; Oral delivery systems; Tableting; Microspheres; Non-pareil seeds; Oral vaccination
`
`1. Introduction
`
`Rotaviruses cause severe, acute diarrhea in young
`children, and may be responsible for 1-2 million
`deaths per year worldwide [1-6]. Thus, an effective
`rotavirus vaccine would significantly reduce global
`childhood mortality [3]. Infection and reinfection
`with rotavirus are common in adults [7-9] and,
`although the symptoms are usually subclinical [10],
`
`*Corresponding author. Southern Research Institute, Pharma-
`ceutical Division, 2000 9th Avenue South, Birmingham, AL
`35205, USA. Tel.: 205 581 2000; fax: 205 581 2888.
`
`they contribute to significant loss of productivity in
`developed countries.
`The first successful attempts to immunize infants
`against rotavirus infection have employed modified
`live viruses administered by the oral route, and
`several such vaccines are now in clinical develop-
`ment. Problems remain, however, in the effective
`administration of these vaccines, particularly in the
`development of a suitable oral vaccine delivery
`system. Although the infectivity of rotavirus is
`relatively stable under various test conditions [11],
`the virus is acid-sensitive. It begins to lose infectivity
`at pH 3.5 and the outer capsid of human rotavirus
`
`0168-3659/96/$15.00 © 1996 Elsevier Science Ireland Ltd. All rights reserved
`Pll S0168-3659(96)01331-4
`
`MYLAN EXHIBIT - 1044
`Mylan Pharmaceuticals, Inc. v. Bausch Health Ireland, Ltd. - IPR2022-00722
`
`

`

`238
`
`J.D. Duncan et al. 1 Journal of Controlled Release 41 (1996) 237-247
`
`collapses below pH 3.0 [12-14]. This acid-sensitivi-
`ty affects the efficacy of orally-administered vac-
`cines, as the live viruses can be inactivated in the
`acid environment of the stomach [15]. Gastric acid
`can be neutralized by the administration of buffers
`before vaccine administration [12,16,17], but this
`procedure is time consuming, expensive [18] and
`poorly tolerated by infants. One approach to this
`problem is to encapsulate the live rotavirus and to
`coat with enteric polymers that are stable at acid pH
`and protect the virus from inactivation, but which
`dissolve at intestinal pH and release the infectious
`rotavirus in the gut. However, to our knowledge, no
`group has demonstrated incorporation of a live
`rotavirus vaccine into an enteric delivery system with
`subsequent recovery of significant levels of rotavirus
`infectivity.
`In the present study, we addressed the problem of
`oral delivery of live rotavirus vaccines by conducting
`a comparative analysis of retention of infectivity in a
`variety of solid oral dosage systems developed for
`use with drugs. Loss of infectivity was estimated
`during incorporation into the delivery systems, and
`during the subsequent processing steps in the prepa-
`ration of poly(DL-lactide-co-glycolide) (DL-PLG)
`microspheres, alginate microcapsules, granules, tab-
`lets and spray-coated non-pareil seeds. These studies
`indicated that stabilization of the virus was required
`during incorporation into the delivery system. Tab-
`leting was found to maintain the infectivity of the
`live rotavirus during the initial stages of preparation
`and during subsequent processing steps.
`
`2. Materials and methods
`
`Porcine rotavirus Gottfried strain GP46 was a gift
`from Dr. L.J. Saif (Ohio State University, Wooster,
`OH) and porcine rotavirus strain SB-lA was a gift
`from Dr. A.Z. Kapikian (NIAID, Bethesda, MD).
`Pancreatin was purchased from Gibco BRL (Grand
`Island, NY). Lactose, sucrose, starch, microcrystal-
`line cellulose (Avicel) and gelatin were purchased
`from Foremost Ingredients (Baraboo, WI); Sigma
`Chemical (St. Louis, MO); Colorcon (West Point,
`PA); FMC Corporation (Philadelphia, PA); and Geo
`A. Hormel (Austin, MN); respectively. Tableting
`excipients, acdisol, stearic acid and talc were pur-
`
`chased from FMC, Mallinckrodt Specialty Chemical
`Co. (St. Louis, MO) and Luzenac America Inc.
`(Englewood, CO). The enteric coating Eudragit
`L30D, Opadry and cellulose acetate phthalate were
`purchased from Rohm Pharma (Malden, MA);
`Colorcon; and Eastman Chemical Company (Kings-
`port, TN); respectively. Sodium alginate (Keltone
`HV) was obtained from the Kelco Division of Merck
`(San Diego, CA) and non-pareil seeds from Paulaur
`Corporation (Robbinsville, NJ). The Coomassie Blue
`Binding Assay kit was purchased from Pierce (Rock-
`ford, IL).
`
`2.1. Virus growth and purification
`
`Rotavirus was grown in roller bottles on confluent
`monolayers of MA 104 (fetal rhesus monkey kidney
`cells)
`in Dulbecco's modified essential medium
`(DMEM) supplemented with 30 µg /ml of pan-
`creatin until a cytopathic effect was evident (approxi-
`mately 2-3 days). The culture media was harvested
`and clarified by centrifugation at 4°C for 20 min at
`3700 X gav, and the virus pelleted by centrifugation
`at 4°C for 1 h at 130 000 X gay,. The pellets were
`resuspended in DMEM, overlayed on a 30% gly-
`cerol-phosphate buffered saline (PBS) cushion, and
`centrifuged in an SW41 rotor (Beckman, Palo Alto,
`CA) at 4°C for 4.5 h at 200 000 X gay. The pellets
`were then suspended in 0.5 ml PBS containing 1%
`BSA and kept at —80°C until use.
`
`2.2. Plaque assay of rotavirus infectivity
`
`Confluent monolayers of MA 104 cells in six-well
`plates were washed twice with PBS and inoculated
`with 0.2 ml of a serial 10-fold dilution of the
`samples. After absorption for 1 h at 37°C, the
`unabsorbed virus was removed and the cells over-
`layed with 4 ml per well of 0.9% agar in DMEM
`containing 30 µg/ml of pancreatin. After incubation
`for 5 days at 37°C, 3 ml of a second overlay medium
`consisting of 0.9% agar in DMEM and 0.03% neutral
`red was added, and the plaques were counted on the
`next 2 days. The results are presented as plaque
`forming units (pfu), or as log loss of pfu from the
`original dose.
`
`

`

`J.D. Duncan et al. 1 Journal of Controlled Release 41 (1996) 237-247
`
`239
`
`2.3. Scanning electron microscopy (SEM)
`
`Microspheres/granules were mounted on a speci-
`men disc and coated with a 20 angstrom layer of
`palladium gold. The coating was carried out using
`the electron microscope-500 sputter coater. After
`coating, the samples were examined and photo-
`graphed using an ISI-SX40 SEM.
`
`2.4. Microencapsulation in DL-PLG microspheres
`by solvent extraction
`
`Rotavirus-containing microspheres were prepared
`by the Southern Research Institute (Birmingham,
`Alabama). Three mg of purified rotavirus strain SB-
`1A containing approximately 1012 pfu of infectious
`particles was microencapsulated in biodegradable
`and biocompatible polymers of DL-PLG by a modi-
`fication of an emulsion-based methylene chloride
`solvent evaporation procedure described previously
`[19]. The surface morphology was evaluated by
`electron microscopy and a smooth surface of con-
`tinuous polymeric coating was confirmed. The vac-
`cine content (core loading) was estimated by dissolv-
`ing a sample of the microspheres in methylene
`chloride, extracting the rotavirus, determining the
`amount of protein and calculating the percent antigen
`by weight. The core loading of 0.55% was satisfac-
`tory. The size of the microspheres was determined
`using a Malvern light scattering device and was
`found to range from 1 to 10 µm.
`in
`A sample of microspheres was dissolved
`methylene chloride, the virus extracted twice with
`PBS, and the infectivity measured by plaque assay.
`
`2.5. Microencapsulation in cellulose acetate
`phthalate (CAP) microspheres by spray drying
`with the Buchi 190 mini-spray dryer
`
`CAP has been used widely as an enteric coating
`polymer for pharmaceutical tablets or granules. CAP
`dissolves at approximately pH 5.5, and thus it
`withstands prolonged contact with the acid contents
`of the stomach, but dissolves and releases drugs
`readily in the small intestine.
`The rotavirus strain SB-1A preparation containing
`approximately 108 pfu of infectious particles was
`resuspended in 100 µl of 1.5 M sucrose containing
`
`10 mM poly-L-lysine and mixed with 40 ml of an
`aqueous solution of CAP (25 mg /ml, pH 6.5).
`Microcapsules were produced by atomization of the
`CAP and vaccine emulsion through a Buchi 190
`mini-spray dryer (Brinkmann, Westbury, NY). About
`50% of the material was recovered as microcapsules
`of 1µm to 5 µm in diameter, as determined by
`SEM.
`The infectivity of the microencapsulated rotavirus
`was determined by in vitro analysis. Fifty mg of
`microcapsules were aliquoted into 2 tubes. One
`sample was treated with 0.1 N HCl at 37°C for 30
`min. The microcapsules did not dissolve in the acidic
`solution and appeared to be intact. The microcap-
`sules were then pelleted by low speed centrifugation
`and dissolved in 0.5 ml of simulated intestinal fluid,
`pH 7.5 (USP XXI). The other 50 mg of microcap-
`sules were dissolved in 0.5 ml of simulated intestinal
`fluid without pretreatment with acid. Both samples
`were then examined for virus infectivity by plaque
`assay.
`
`2.6. Incorporation into alginate microcapsules by
`the process of ionic gelation
`
`live
`the
`Alginate microcapsules containing
`rotavirus strain SB-1A virus were formed by chela-
`tion of the sodium alginate with calcium ions.
`Rotavirus at a concentration of 1 X 108 pfu was
`mixed with 3 ml of sodium alginate solution (1.2%
`w/v in normal saline). This suspension was then
`dripped slowly through a 19 gauge needle into a
`solution of calcium chloride (1.5% w/v in distilled
`water) while stirring at 500 revs./min. The mi-
`crocapsules containing rotavirus were collected by
`sieving, rinsed three times with a normal saline
`solution, and then dried at 4°C under vacuum.
`Half of the alginate microcapsules containing
`rotavirus were incubated at 37°C in simulated gastric
`fluid (USP XXI) for 30 min. They were pelleted and
`then suspended in phosphate buffered saline (PBS),
`pH 7.7. The other half of the alginate microcapsules
`were not exposed to gastric fluid, but were re-
`suspended in PBS, pH 7.7 until dissolved. The
`rotavirus was released from the alginate microcap-
`sules, which dissolved completely at the higher pH.
`The rotavirus infectivity of both samples was then
`determined by plaque assay. The infectivity of the
`
`

`

`240
`
`J.D. Duncan et al. 1 Journal of Controlled Release 41 (1996) 237-247
`
`live rotavirus was reduced by 2 log after exposure to
`gastric fluid, but reduced by only 0.73 log when
`exposed to PBS alone.
`
`2.7. Preparation of granules for oral delivery
`
`Granules were prepared by forming an excipient
`blend of cellulose, starch, sucrose and gelatin at a
`ratio of 30:30:30:10 in a Waring blender. One batch
`of granules was prepared by adding a solution of
`rotavirus strain SB-lA at a concentration of 4.51 X
`107 pfu per 300 mg of excipient blend, and a second
`batch prepared by adding a solution of SB-lA at a
`concentration of 3.91 X 104 pfu per 20 mg of
`excipient blend. The wet mass was dried under
`vacuum at 4°C and then ground with a small mill to
`form granules. These were passed through a series of
`sieves (300 µm-3 mm, USA standard testing
`sieves), and granules of 1-3 mm were collected and
`stored at 4°C under vacuum with desiccant.
`One batch of granules containing 4.51 X 107 pfu
`rotavirus strain SB-1A /300 mg of cellulose blend
`was enterically coated with Eudragit L30D, using a
`Wurster spray coating method in a laboratory scale
`fluid bed, STREA-1 (Aeromatic Inc., Columbia,
`MD). Eudragit is an aqueous dispersion of an anionic
`copolymer based on methacrylic acid and acrylic
`acid ethyl ester. The polymer dissolves above pH 5.5
`by forming salts with alkalis, thus affording coatings
`that are insoluble in a gastric media, but soluble in
`the small intestine. Sufficient coating was applied to
`increase the weight of the granules by 25% (w/w).
`Although a weight gain of 12-18% is considered to
`be protective by manufacturer's standards, the exten-
`sive porosity of the granules required a heavier
`coating to adequately seal the pores.
`A second batch of granules containing 3.91 X 104
`pfu of rotavirus strain SB-lA per 20 mg cellulose
`blend was enterically coated with Eudragit L30D as
`described above, except that granules received a
`protective coating of Opadry prior to application of
`the Eudragit. It has previously been determined that
`enteric polymers can inactivate some ingredients
`such as viruses, proteins and peptides. Opadry,
`which is a water-soluble hydoxypropylmethylcel-
`lulose-based polymer, was applied to the granules as
`a protective coating by the Wurster spray coating
`method in a fluid bed laboratory unit, STREA-1, to
`
`give a weight increase to the granules of approxi-
`mately 8-10%. A further weight gain of 25%
`occurred on coating with Eudragit.
`Uncoated granules, granules coated with Eudragit
`L30D alone, granules coated with Opadry alone, and
`granules coated with both Opadry and Eudragit
`L30D, were stored at 4°C under vacuum with
`desiccant. The viral infectivity was determined by
`plaque assay. Disintegration analysis was used to
`determine the effectiveness of the film coatings in
`protecting the live rotavirus vaccine from exposure
`to gastric fluid. The granules were weighed prior to
`and after exposure to simulated gastric fluid (USP;
`pH 1.2) at 37°C for 1 h, and the percent gastric
`uptake determined. The granules then were exposed
`to simulated intestinal fluid (USP; pH 6.8) at 37°C,
`and the time required for complete disintegration
`determined.
`
`2.8. Preparation of tablets
`
`Rotavirus (Gottfried strain GP46) was prepared
`for tableting using a drying procedure as described
`for the preparation of the granules except that the
`rotavirus was dried, (1) with the excipient blend of
`cellulose, starch, sucrose and gelatin at a concen-
`tration of 9.7 X 106 pfu/100 mg of excipient blend,
`or (2) the rotavirus was dried with lactose alone at a
`concentration of 6.7 X 104 pfu/20 mg lactose.
`Tablets formed from preparation 1 were composed
`of the following ingredients (% dry weight): lactose
`filler (74.70%), acdisol disintegrant (3%), stearic
`acid lubricant (1.50%), talc as a processing aid
`(1.0%), and
`the dried
`rotavirus preparation
`(19.70%). Tablets formed from preparation 2 were
`composed as follows: lactose (54.10%), acdisol
`(3%), stearic acid (1.50%), talc (1.0%), and the
`dried rotavirus preparation (40.40%). The dry in-
`gredients from each preparation were mixed well to
`form a dry blend, incorporated into a 3 mm dye and
`pressed at 550 lb pressure to produce 50 mg, 3 mm
`tablets.
`The tablets were enterically coated with Opadry
`and Eudragit L30D as described for the coating of
`the granules. The subcoating of Opadry was applied
`until the weight of the tablets increased by 6—8%.
`Subsequently, an enteric coating of Eudragit L30D
`was applied until the weight of the tablets was
`
`

`

`J.D. Duncan et al. / Journal of Controlled Release 41 (1996) 237-247
`
`241
`
`further increased by 20-25%. Tablets from the three
`processing steps including uncoated tablets, tablets
`coated with Opadry, and tablets coated with both
`Opadry and Eudragit L30D, were stored at 4°C under
`vacuum with desiccant. The tablets were evaluated
`for in vitro stability by plaque assay and disinte-
`gration testing as described for the granules.
`
`L30D, were stored at 4°C under vacuum with
`desiccant. The seeds were evaluated for in vitro
`stability by plaque assay and disintegration analysis,
`as described for the granules.
`
`3. Results
`
`2.9. Comparison of the effects of excipients on the
`drying process
`
`3.1. Comparison of delivery systems
`
`As infectivity may be lost on drying of the
`rotavirus during granulation and prior to tableting,
`the effect of the formulation of the excipient blend
`on the drying process was evaluated. The formula-
`tions included lactose alone, sucrose alone, gelatin
`alone or the excipient blend of cellulose, starch,
`sucrose and gelatin at a ratio of 30:30:30:10. The
`ingredients were suspended in distilled water and
`mixed with rotavirus strain Gottfried GP46 (1.9 X
`107 pfu) until a wet mass was formed. The wet mass
`was dried in a desiccator at 4°C under vacuum until a
`level of 5% water weight was reached, then ground
`with a small mill to form a fine, dry powder. The
`viral activity retained after processing was deter-
`mined by plaque assay.
`
`2.10. Application to the surface of non-pareil
`seeds and enteric coating
`
`As an alternative to granules and tablets, which
`involve incorporating the live rotavirus inside the
`core material of the delivery system, the rotavirus
`was applied to the surface of non-pareil sugar seeds.
`The dried, live rotavirus strain Gottfried GP46 at a
`concentration of 107-109 pfu was suspended in 100
`ml of an aqueous hydroxypropylmethylcellulose
`(HPMC) film-forming polymer that contained 1-2%
`sucrose. This was applied to the surface of 200 g
`non-pareil seeds by the Wurster spray coating meth-
`od administered in a STREA-1 laboratory unit as
`previously described. Subsequently, a sub-coating of
`Opadry was applied to an approximate weight gain
`of 6-8%, followed by an enteric coating of Eudragit
`L30D that further increased the weight of the seeds
`by 20%. Seeds from the three processing steps
`including uncoated seeds, seeds coated with Opadry,
`and seeds coated with both Opadry and Eudragit
`
`Microencapsulation of live rotavirus indicated that
`incorporation of live rotavirus into DL-PLG micro-
`spheres with solvent removal by either extraction or
`evaporation, completely destroyed all viral infectivi-
`ty (Fig. 1). Furthermore, microencapsulation of
`rotavirus in CAP polymer particles by the process of
`atomization in a Brinkmann Buchi 190 mini-spray
`dryer, completely destroyed rotavirus viral infectivi-
`ty. Incorporation of live rotavirus into alginate
`microcapsules resulted in a significant loss of infec-
`tivity of approximately 2 log after exposure of
`microcapsules to gastric fluid, but reduced infectivity
`
`0
`
`0
`
`0
`a
`0
`
`1
`
`2
`
`3
`
`4
`
`Fig. 1. Loss of rotavirus infectivity after incorporation into various
`delivery systems. (1) PLGA microspheres; (2) Buchi spray
`drying; (3) Alginate microcapsules; (4) Cellulose granules. Each
`bar represents the loss of infectivity during incorporation into
`these commonly used delivery systems. Cellulose granules show a
`comparatively lower loss of infectivity.
`
`

`

`242
`
`J.D. Duncan et al. 1 Journal of Controlled Release 41 (1996) 237-247
`
`1.
`
`3
`
`4
`
`1
`
`5
`
`0-
`
`7
`
`6
`
`5-
`
`4
`
`Recovery of rotavirus PFU/dose (log 10)
`
`Fig. 2. Effect of excipients on recovery of rotavirus infectivity
`during the drying process. (1) Control solution rotavirus; (2)
`Rotavirus dried with sucrose; (3) Rotavirus dried with gelatin; (4)
`Rotavirus dried with lactose; (5) Rotavirus dried with an excipient
`blend of cellulose, sucrose, starch, gelatin at a 30:30:30:10 ratio.
`The excipient blend shows superior retention of rotavirus infec-
`tivity.
`
`by only 0.73 log when exposed to PBS alone. The
`application of rotavirus to the surface of non-pareil
`seeds and the subsequent coating process also re-
`sulted in a 2-3 log loss (Table 1). In contrast,
`preparation of granules in an excipient blend of
`starch, sucrose, gelatin and cellulose reduced infec-
`tivity by only 0.29 log, and tableting with an
`excipient blend of starch, sucrose, gelatin and cellu-
`lose reduced infectivity by only 0.77 log (Fig. 1 and
`Table 2).
`
`3.2. Effect of the formulation of the excipient
`blend
`
`To determine the appropriate excipients to be used
`in forming granules, the live rotavirus was dried on
`several different substrates (Fig. 2). Drying of the
`live rotavirus in solution at a dose of 1.9 X 107 pfu,
`on sucrose alone, gelatin alone, or lactose alone
`resulted in a loss of infectivity of between 1.39 and
`2.45 logs. In contrast, drying on the excipient blend
`containing cellulose, sucrose, gelatin and starch
`resulted in preservation of infectivity with a log loss
`of 0.29.
`
`Table 1
`Retention of infectivity after spraying rotavirus onto the surface of sugar beads
`
`Batch no.
`
`Input (pfu/bead)
`
`Virus recovered (pfu/g beads)
`
`Infectivity lost (log 10)
`
`1
`2
`3
`
`8.5 X 104
`2.7 X 106
`4.0 X 10'
`
`3.5 X 102
`2.4 X 103
`7.5 X 103
`
`Beads were coated with HPMC, and additionally batch no. 2 was coated with Eudragit.
`
`2.4
`3.0
`2.7
`
`Table 2
`Effect of tableting and coating on rotavirus infectivity
`
`pfu/ dose
`
`pfu/tablet
`(no coating)
`
`pfu /tablet
`(HPMC coating)
`
`RV dried with lactose
`RV dried with excipient blend
`
`6.7 X 104 (2.45)4
`9.7 X 106 (0.29)4
`
`7.1 X 104 (2.43)
`3.2 X 106 (0.77)
`
`3.8 x 103 (3.70)
`4.4 X 106 (0.63)
`
`aLog loss of infectivity from original rotavirus in solution, 1.9 X 107/dose.
`
`pfu/tablet
`(HPMC and
`Eudragit coating)
`
`1.0 x 10' (4.28)
`6.4 X 105 (1.46)
`
`

`

`J.D. Duncan et al. I Journal of Controlled Release 41 (1996) 237-247
`
`243
`
`Table 3
`Recovery of rotavirus infectivity after incorporation into granules
`
`Preparation of granules
`
`Virus recovered (pfu/300 mg)
`
`Infectivity lost (log 10)
`
`Uncoated granules
`Granules coated with Eudragit:
`Treated with SIF
`Treated with SGF
`
`2.3 x 10'
`
`4.5 x 103
`
`Original rotavirus input = 4.51 X 107 /300 mg cellulose blend.
`SIF, simulated intestinal fluid.
`SGF, simulated gastric fluid.
`
`3.3. Effect of coating processes on recovery of
`rotavirus infectivity from granules
`
`Enteric coating of the granules with Eudragit
`L30D resulted in a reduction of infectivity of 1.61
`log (Table 3). There was an additional 2.39 log loss
`of infectivity after exposure of the Eudragit coated
`granules to gastric fluid, for a total cumulative loss
`of 4.00 log. Upon disintegration testing there was an
`approximate weight gain of 8-10% after exposure to
`simulated gastric fluid.
`Consequently, a second batch of granules was
`coated with HPMC, Opadry primer, prior to enteric
`coating (Table 4). In contrast to the uncoated
`granules, which exhibited a 0.30 log reduction in
`infectivity, coating with HPMC resulted in a loss of
`infectivity of 0.78 logs, and the subsequent enteric
`coating with Eudragit reduced the infectivity by 1.19
`log. After
`treatment of
`the enterically coated
`granules with USP simulated gastric fluid there was
`an additional loss of infectivity of 0.10 log. Thus, the
`total losses incurred during the coating processes and
`gastric exposure was 1.29 log. In this batch of
`granules, there was an approximate weight gain of
`4-6% on disintegration testing.
`
`0.30
`
`1.61
`4.00
`
`As losses in rotavirus infectivity were observed
`after exposure of the enterically coated granules to
`the USP simulated gastric fluid, the granules were
`examined by SEM to determine the effectiveness of
`the coatings in sealing the pores and crevices on the
`surfaces of the granules (Fig. 3). The HPMC and
`Eudragit enteric coatings smoothed the surfaces of
`the granules and rounded out the rough edges, but
`did not adequately seal the pores. Pores and crevices
`were apparent in the SEM analyses of the whole
`granules, despite the application of a 50-80% weight
`gain of Eudragit L30D.
`
`3.4. Effect of incorporation of live rotavirus into
`tablets and enteric coating
`
`Preparation of tablets after drying of the live
`rotavirus vaccine onto lactose resulted in a signifi-
`cant loss of infectivity at each step of processing
`with a total accumulative loss of infectivity of 4.28
`log (Table 2). In contrast, initial drying of the live
`rotavirus vaccine onto an excipient blend of cellu-
`lose, sucrose, starch and gelatin, followed by tablet-
`ing and coating steps, resulted in a significant
`improvement in the retention of infectivity with a
`
`Table 4
`Effect of coating processes on recovery of rotavirus infectivity from granules
`
`Preparation of granules
`
`Virus recovered (pfu/20 mg)
`
`Infectivity lost (log 10)
`
`Granules without coating
`Granules with HPMC
`Granules with HPMC and Eudragit coating
`Granules with HPMC and Eudragit coating treated with SGF before release
`
`2.0 X 104
`6.5 x 103
`3.0 X 103
`2.0 X 103
`
`All granules were treated with SIF and titer of the viruses released in the solution was measured.
`Original rotavirus input = 3.91 X 104 / 20 mg cellulose blend.
`SGF, simulated gastric fluid.
`SIF, simulated intestinal fluid.
`
`0.30
`0.78
`1.19
`1.29
`
`

`

`244
`
`J.D. Duncan et al. 1 Journal of Controlled Release 41 (1996) 237-247
`
`Fig. 3. Scanning electron microscopy of the outer surface of granules either uncoated or coated with polymer film coatings. (A) Surface of
`entire uncoated granule, (B) SEM photo showing a close-up of the surface of the previous uncoated granule, (C) Surface of an entire granule
`after applying primer coat of HPMC, (D) SEM photo showing a close-up of the surface of the previous granule coated with HPMC, (E)
`Surface of an entire granule after a primer coat with HPMC and Eudragit L30D, (F) SEM photo showing a close-up of the surface of the
`previous granule coated with HPMC and Eudragit L30D. These photos demonstrate that even though HPMC and Eudragit L3OD fill in some
`of the crevices and smooth out the rough edges of the granules, there are still pores remaining.
`
`

`

`J.D. Duncan et al. / Journal of Controlled Release 41 (1996) 237-247
`
`245
`
`ate, and ether, the process of microencapsulation in
`DL-PLG copolymer almost completely ablates the
`infectivity. This loss of infectivity could be associ-
`ated with the mechanical action, possible increases in
`temperature, prolonged exposure to organic solvent
`or the lyophilization steps involved in the microen-
`capsulation process. We, and others have demon-
`strated success using DL-PLG to encapsulate viruses,
`proteins and peptides with excellent results [20], but
`this procedure is apparently not suitable for the
`preparation of a live rotavirus vaccine (Fig. 1). Spray
`drying, commonly used to encapsulate active in-
`gredients, resulted in almost complete loss of infec-
`tivity (Fig. 1). The loading of an active ingredient
`onto the surface of non-pareil seeds also suffered
`unacceptable infectivity losses during formulation
`(Table 1). As non-pareil seeds are a sucrose formula-
`tion, this finding is consistent with our other results
`indicating that sucrose alone is not effective in
`maintaining rotavirus infectivity (Fig. 2). The algi-
`nate microcapsule delivery system produced a satis-
`factory level of rotavirus infectivity in microcapsules
`that had not been exposed to gastric acid, but after
`exposure to gastric acid, rotavirus infectivity was
`reduced by 2 log (Fig. 1). While alginate did not
`produce desirable results in our experiments, another
`scientific study has demonstrated that infectious
`rotavirus could be successfully microencapsulated in
`alginate [21]. It was determined that oral or parenter-
`al inoculation of mice with alginate encapsulated
`rotaviruses induced a virus-specific IgG and IgA
`reponse in serum and intestinal contents, which was
`of greater magnitude than that induced after inocula-
`tion with the same dose of free virus.
`Comparison of the delivery systems indicated that
`granules prepared from the excipient blend per-
`formed significantly better than the other delivery
`systems (Fig. 1 and Fig. 2). This suggested that
`stabilization of the rotavirus during the drying pro-
`cess is an important factor in maintaining rotavirus
`infectivity. Thus, it is possible that the polymers used
`for microencapsulating the rotavirus vaccine such as
`DL-PLG, CAP, and alginate do not stabilize the live
`rotavirus vaccine during the drying process. Appar-
`ently an excipient blend of cellulose, starch, sucrose
`and gelatin is capable of maintaining rotavirus
`infectivity. However, when these excipients were
`used alone, the rotavirus infectivity was decreased
`
`Fig. 4. Scanning electron microscopy of the cross-section of a
`tablet coated with both HPMC and Eudragit L30D. (A) Area of the
`tablet which displays the thickness and uniformity of the Eudragit
`L3OD enteric coat, (B) Area of the tablet which displays the
`thickness and evenness of the HPMC coating, (C) Interior of the
`cross-sectioned tablet. Tablets provide a hard, even surface which
`can be coated more uniformly than granules.
`
`cumulative log loss of 1.46. Only 0.29 log loss of
`infectivity was observed after drying of the rotavirus
`on the excipient blend, compared to 2.45 log loss
`when the rotavirus vaccine was dried on lactose (Fig.
`2 and Table 2).
`As minimal losses in infectivity were observed
`after all tableting processes of the dried rotavirus and
`excipient blend, the tablets were cross-sectioned and
`examined by SEM. The SEM microscopy revealed a
`very uniform coating of Opadry and Eudragit L30D
`around the surface of the tablet providing a more
`perfect seal and thus complete protection from
`gastric fluid (Fig. 4).
`After exposure to gastric fluid, no weight gain was
`observed in either preparation on disintegration
`testing, therefore this indicates that the tablets were
`adequately coated with Eudragit L30D.
`
`4. Discussion
`
`Using current techniques, rotavirus infectivity is
`not retained during incorporation into commonly
`used delivery systems. Although our prior work
`indicated that live rotavirus is stable in some organic
`solvents such as methylene chloride, diethylcarbon-
`
`

`

`246
`
`J.D. Duncan et al. / Journal of Controlled Release 41 (1996) 237-247
`
`significantly (Fig. 2). Consequently, the excipient
`blend was chosen to form granules for further
`analysis.
`The effect of the enteric coating processes, such as
`application of HPMC and Eudragit L30D, on the
`recovery of rotavirus infectivity after exposure to
`simulated gastric fluid was determined (Table 3 and
`Table 4). The results show an improvement in the
`retention of rotavirus infectivity after exposure to
`simulated gastric fluid when
`the granules were
`coated with both HPMC and Eudragit L30D as
`opposed to coating with Eudragit L30D alone (Table
`3 and Table 4). Apparently, a primer coat is needed
`prior to the application of the Eudragit L30D for
`protection of the live rotavirus. Some cells, proteins,
`and peptides are very sensitive to this polymer and
`precoating with HPMC affords some protection. In
`addition, the crevices and pores of the granules were
`difficult to seal, even after application of a thick coat
`of polymer. On disintegration testing the weight
`gain, after exposure to gastric fluid, indicates that a
`complete seal was not achieved even after pre-coat-
`ing with HPMC. It is apparent that after completion
`of all the coating processes, there were still some
`pores and crevices present on the surface of the
`granules (Fig. 3). Thus, it was extremely difficult to
`ensure the perfect seal required for complete protec-
`tion from the gastric fluid. This problem will be
`addressed by using more current granulation technol-
`ogy to create small, agglomerated particles that are
`hard, round and smooth, in contrast to the more
`irregular, porous granules produced by the standard
`laboratory methods used in this study. The improved
`technology will provide a surface that can be more
`uniformly coated and provide a better sea