See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/288021402

Formulation of Granules for Site-Specific Delivery of an Antimicrobial

Essential Oil to the Animal Intestinal Tract

Article  in  Journal of Pharmaceutical Sciences · February 2016

DOI: 10.1016/j.xphs.2015.10.001

CITATIONS READS

11 442

4 authors, including:

Qi Wang Joshua Gong

Agriculture and Agri-Food Canada Agriculture and Agri-Food Canada
102 PUBLICATIONS   4,538 CITATIONS    124 PUBLICATIONS   3,115 CITATIONS   

SEE PROFILE SEE PROFILE

Xiao Yu (Shirley) Wu
University of Toronto
239 PUBLICATIONS   8,434 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

compelx coacervate View project

Glucose-responsive delivery of insulin for diabetes treatment View project

All content following this page was uploaded by Xiao Yu (Shirley) Wu on 03 September 2018.

The user has requested enhancement of the downloaded file.

 Journal of Pharmaceutical Sciences 105 (2016) 1124e1133

Contents lists available at ScienceDirect

Journal of Pharmaceutical Sciences

journal homepage: www.jpharmsci.org

Pharmaceutics, Drug Delivery and Pharmaceutical Technology

Formulation of Granules for Site-Specific Delivery of an Antimicrobial

Essential Oil to the Animal Intestinal Tract
Yin-Hing Ma 1, 2, Qi Wang 1, Joshua Gong 1, Xiao Yu Wu 2, *
1
Guelph Food Research Centre, Agriculture and Agri-Food Canada, Guelph, Ontario, Canada N1G 5C9
2
Advanced Pharmaceutics and Drug Delivery Laboratory, Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, Ontario, Canada M5S 3M2

a r t i c l e i n f o a b s t r a c t

Article history: Owing to proliferation of antibiotic-resistant bacteria, the use of antibiotics for livestock growth
Received 24 April 2015 promotion is banned in many countries and alternatives to in-feed antibiotics are needed. Cinnamon
Revised 4 September 2015 essential oil exhibits strong in vitro antibacterial activity; however, direct addition of essential oils to
Accepted 9 October 2015
animal feed has limited practicality due to their high volatility, odor, fast decomposition, and poor
Available online 23 December 2015
availability in the lower intestines. To solve these problems, we formulated trans-cinnamaldehyde
(CIN) with an adsorbent powder and fatty acid via a melt-solidification technique. Core granules of an
Keywords:
optimized composition contained up to 48% wt/wt CIN. The granules were then coated with an enteric
formulation
site-specific delivery
polymer to impart site-specific release of CIN. CIN was mostly retained in simulated gastric fluid and
anti-infectives released rapidly (>80% under 2 h) in simulated intestinal fluids. Rapid CIN autoxidation into cinnamic
antioxidants acid was inhibited by adding 1% vol/vol eugenol, which maintained CIN stability for at least 1 y. The
coating granule formulation increased the antimicrobial activity of CIN against Escherichia coli K88 slightly with
oral drug delivery a minimum bactericidal concentration of 450 mg/mL for CIN in lauric acidebased granules compared
stability with 550-600 mg/mL for palmitic acidebased granules and free CIN, respectively. These results
encourage the potential use of encapsulated CIN for control of animal enteric pathogens by oral in-feed
administration.
© 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.

Introduction from inhibitory to bactericidal) against various bacterial patho-

gens at concentrations between 56 and 385 mg/L (mg/mL) in liquid
The practice of supplementing antibiotics to animal feed at media.4,6,7 In addition, CIN has shown some potential use in dairy
subtherapeutic levels to promote the growth of food-producing cows to affect rumen microbial fermentation.8 Hence, CIN is a
animals and prophylaxis has been banned across EU members1 good candidate for an alternative antimicrobial agent. Neverthe-
in an effort to curb spreading of antibiotic resistance among less, the effectiveness of direct addition of free EOs to the feed is
pathogens caused by selective pressure under intensive farming questionable due to several limiting physical-chemical factors
environments.2 Alternative compounds to in-feed antibiotics are such as the high hydrophobicity, volatility, odor, oxygen sensi-
thus pressingly needed to help maintain economical livestock tivity, and low availability of the EOs in the lower gastrointestinal
production. Many plant-derived essential oils (EOs) exhibiting tract (GIT) of animals.9,10
strong antimicrobial activity in vitro3-5 can be used as alternatives To improve the availability of EOs to exert their antimicrobial
to antimicrobial drugs, shifting the use and resistance develop- activity in vivo, it is desirable to develop a formulation that is
ment away from more medically important antibiotic drugs. Of amenable to oral delivery to the lower GIT of food-producing ani-
various EOs, trans-cinnamaldehyde (CIN), the most abundant mals. For this purpose, solid granules containing liquid EOs would
component of cinnamon oil, shows antimicrobial effects (ranging be a good choice as they can be mixed within animal feed more
readily than liquid EOs, and can be further processed to reduce
volatility and degradation of EOs, and to allow site-specific release
This article contains supplementary material available from the authors by request of EOs in the GIT by applying a suitable coating to the granules.
or via the Internet at http://dx.doi.org/10.1016/j.xphs.2015.10.001. The granulation and coating can potentially control the effects of
* Correspondence to: Xiao Y. Wu (Telephone: þ1-416-978-5272; Fax: þ1-416-978-
EOs on feed palatability for animals by reducing oil volatility and
8511).
E-mail address: [email protected] (X.Y. Wu). masking odor.11

http://dx.doi.org/10.1016/j.xphs.2015.10.001
0022-3549/© 2016 American Pharmacists Association®. Published by Elsevier Inc. All rights reserved.
 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133 1125

To obtain EO granules with high loading levels while minimizing Neusilin® US2 and UFL2 (magnesium aluminum silicate) were bulk
the evaporation and odors of the oils, we have designed a new samples obtained from Fuji Chemical Industries (Fuji Health Sci-
granule formulation in this work by soaking the oil in an adsorbent ence, Burlington, NJ), microcrystalline cellulose Avicel® grades
powder and then incorporating the powder in a melted lipid to PH102, RC-591, CL-611 were obtained from FMC Biopolymer, and
form solid granules. Various pharmaceutically acceptable powdery, white wheat bran powder (50 mesh) was provided by Hayhoe Mills,
inert adsorbents were investigated including magnesium Ltd. (Woodbridge, Ontario, Canada). Carvacrol, butylated hydroxy
aluminum silicate, microcrystalline cellulose, and wheat bran toluene (BHT) and t-butyl methyl phenol were obtained from
powder. The powder possessing the highest oil retention capability Sigma-Aldrich chemicals.
was then selected for further development of granules with a lipid
binder. Determination of Oil-Adsorbing Capacity of Powders
The success of developing solid granules containing lipid and
EO-adsorbed powder depends on many factors that influence the Two types of pharmaceutical excipient powders, that is, micro-
granule properties, for example, granule size and strength, EO crystalline cellulose powder (Avicel® PH102, Avicel® RC-591, Avicel®
loading levels, and release kinetics. These factors include the CL-611) and magnesium aluminum silicate powder (Neusilin® UFL2,
compatibility of the powder with the lipid, the melting tempera- Neusilin® US2), and a food powder (wheat bran) were tested for
ture of the lipid and its mixture with the EO-adsorbed powder, and their oil-adsorbing capacity. Approximately 0.5-1 g of a powder was
the droplet-forming properties that can affect the size and shape of accurately weighed into a 15-mL conical polypropylene screw-cap
solidified granules. Although meltable lipids have been used to tube (BD Falcon), and then, CIN was added until excess was
prepare solid lipid particles previously to encapsulate various visible. After allowing the oil to completely soak the powder for >10
drugs, such dosage forms were primarily designed for parenteral min, the excess oil was then separated by centrifugation for 10 min
drug delivery with small (nano) sizes and low drug loading levels, at 3000 g. The excess oil was poured off, and a pipette was used to
and in the absence of powdery materials.12,13-17 Therefore, we draw any remaining superficial oil after inverting the tube and
conducted in-depth studies on the molten and resolidified prop- allowing drainage. After all the excess oil was removed, the tube
erties of mixtures in relation to the composition of the 3 compo- containing powder and adsorbed oil was weighed again. The weight
nent system and rationally selected the type of lipid. Simple of adsorbed oil was obtained by weight difference. The oil loading
saturated fatty acids (FAs) with well-defined melting points from capacity was calculated by the following equation:
processed oils of palm or coconut as well as fatty alcohols were
tested, as they have a melting range from 44 C to 70 C and have weight of oil
%wt=wt oil ¼  100%
been successfully applied in melt-pelletization and melt- weight of oil þ weight of powder
granulation processes.12,18-20
Despite its good antimicrobial activity, the practical application
of CIN as an alternative antibiotic product is limited by its poor
stability. CIN is known to undergo rapid oxidation on exposure to Formulation Selection by Phase Diagram
atmosphere, converting to nonbioactive cinnamic acid (CA).
Therefore, a series of antioxidants were screened to stabilize CIN. Based on the test of oil-adsorbing capacity of powders described
Eugenol, an effective antioxidant present in natural EOs, was previously, Neusilin® US2 powder yielded the highest CIN oil
selected for further study of CIN in granules and its effect on CIN adsorbing capacity among the 6 studied powders and was selected
stability was monitored at room temperature and 4 C for 1 y. for further formulation development. Oil-loaded core granules
To protect actives from acidic pH and achieve site-specific drug were formulated with 3 components: CIN oil, adsorbent powder,
release in the lower GIT, an enteric polymer was introduced in the and FA. A phase diagram was constructed to better define compo-
core and surface of granules. The surface of core CIN granules was sition ranges that yielded: (1) at molten temperatures, a mixture in
coated using a pH-responsive polymer, Eudragit® (Evonik In- the liquid state for ease of droplet formation in suspension and (2)
dustries) L 100. Structurally, this polymer contains acidic mono- when cooled to room temperature (23 C), a solid with absence of
mers (methacrylic acid) copolymerized with nonionizable excess oil on the surface. Compositions (see Supplementary
monomers (methyl or ethyl acrylate) and is identified in the United Figure S1) were tested by making 2 g of mixtures containing
States Pharmacopeia under the general term “methacrylic acid different weight % of each component in a 10-mL glass vial. Melting
copolymers.” It dissolves in an aqueous medium at pH > 6. Here, we and solidification of the mixtures were carried out using a hot
describe a systematic investigation for the first time on the water bath and ice water bath, respectively. The CIN oil was mixed
formulation development, granule production, stability testing, and with the powder first and then blended into molten FA with a
enteric coating of CIN core granules intended for releasing CIN in spatula and subsequently solidified by cooling in a water bath.
the lower GIT regions to target intestinal pathogens. Finally, the Qualitative observations were made (see Supplementary Table S1)
antimicrobial activity of CIN granules was examined against a of the mixtures at different temperatures and classified as powder,
multidrug-resistant bacterial strain Escherichia coli K8821 as solid (homogeneous), paste, or liquid.
compared with free CIN.
Preparation of Core Granules
Materials and Methods
The CIN oil was encapsulated into spherical granules by melt-
EO compounds trans-CIN (99%) and eugenol (99%), FAsdlauric solidification in an aqueous suspension. To prepare 100 g of oil-
acid (LA; C12), myristic acid (MA; C14), palmitic acid (PA; C16), stearic containing granules, CIN oil (40-50 g) was first adsorbed with
acid (SA; C18), and the fatty alcohol palmitic alcohol (C16), were Neusilin® US2 powder (US2; 0-13 g) by slow addition and mixed
obtained from Sigma-Aldrich (Oakville, Ontario, Canada). Enteric gently in a beaker to ensure even distribution of oil. In a separate
polymers Eudragit® (Evonik Industries) L 100 and S 100 are beaker, the FA was melted in a water bath set at 45 C-67 C (i.e., up to
methacrylic acidemethyl methacrylate copolymers in the ratio 1:1 5 C higher than the melting point of the FA chosen). The oil-powder
and 1:2, respectively, obtained in powder form from Almat Phar- mixture was then added to the molten FA and mixed with a spatula
machem Inc., (Concord, Ontario, Canada). Oil adsorbent powders until it became homogeneous. To stabilize the molten droplets, an
1126 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133

individual batches were taken for determining loading and stability

of CIN contained in individual batches. Due to the oxidation of CIN
(lmax ¼ 291 nm) to CA (lmax ¼ 272 nm), a mixture of CIN and CA was
present in some cases. Thus, calibration equations of the 2 standards
were obtained with both pure compounds and mixtures and then
used in determining unknown concentrations of both compounds
simultaneously using a simple additive rule.22 Readings were
obtained on a UV-vis spectrophotometer (Lambda; Perkin-Elmer).
The EO loading efficiency in prepared granules was calculated in
the following manner:

weight of CIN added

Theoretical loading ¼
total weight of all granule materials
 100%

weight of total CIN extracted

Actual loading ¼  100%
total weight of granules

actual laoding of CIN in granules

Loading efficiency ¼
theoretical loading of CIN granules
 100 %

Figure 1. Schematic diagram for granule production process.

aqueous solution of methacrylic acidemethyl methacrylate copol- In Vitro Release of CIN From Granules
ymer (Eudragit® S 100 [Evonik Industries]) neutralized with NaOH to
pH 7.5 was used at a final concentration of 4%-5% wt/wt. The aqueous Simulated gastric fluid (SGF, adjusted with 0.1-M HCl, 0.2% wt/
dispersion medium (200-300 mL) was heated to the same temper- vol NaCl) and simulated intestinal fluid (SIF, 50-mM KH2PO4,
ature as the molten FA (51 C-52 C for LA; 66 C-67 C for PA) in a adjusted with 0.1-M NaOH) were prepared at pH of 1.2 and 6.8,
60-mL beaker with a water bath to ensure the CIN-US2-FA mixture respectively. Release of CIN was monitored by a UV detector at 2
remained molten when dispersed. Tween 80 (at 0.5% vol/vol of the wavelengths (291 and 272 nm) at 10-min intervals via flow-
aqueous phase) was found to slightly improve the round shape of through cuvettes with up to 6 replicates per run. Granules were
granules. An overhead stirrer (Caframo, Ontario, Canada) was used to tested for 2 h in SGF (~250 mL) and 6 h in SIF (~900 mL), and the CIN
disperse the oily phase into the aqueous phase at 150-375 rpm using concentrations were converted into amounts to calculate %
a stainless steel 4-bladed round-edged propeller to form a 2-phase, released. Solutions were maintained at 37 C, and stirring rate was
liquid-liquid suspension. After droplet size and shape were visually 100 rpm using a dissolution testing apparatus (ERWEKA, Germany).
acceptable (~30 s-1 min), stirring was stopped, and the beaker was
transferred to an ice-water bath to solidify the droplets into granules. Antimicrobial Activity Assay With Pure Culture in Liquid Growth
Afterward, granules were suction filtered and washed with small Media
volumes (10 mL) of distilled water, and then allowed to air dry on
aluminum pans for 45 min or until free flowing. For PA granules, Antimicrobial activity of pure compound CIN, its oxidation
higher initial temperature was used (67 C-70 C) to maintain molten product CA, and CIN-encapsulated core granules were assessed. A
state before blending with oil-soaked powder. Figure 1 summarizes study of the growth inhibition by EO compounds was performed
the steps in the granule production process. Control granules were against a pure culture of E. coli K88 strain JG280 undergoing
made with 10% US2:90% FA by physically breaking down the exponential growth at 37 C in tryptic soy broth (TSB) media as
resolidified mixture into small-sized particles. described previously.4 The 3 treatments were control (no EO
compounds), free EO, and EO granules at different EO concentra-
Enteric Coating of Core Granules tions in 5 mL of TSB inoculated with E. coli K88 at a concentration of
104 colony forming units (CFUs)/mL. The bacteria were then
Core granules were first subcoated with an ethanolic dispersion of allowed to grow with agitation at 200 rpm on an orbital shaker
Kollicoat IR (polyvinyl alcoholepolyethylene glycol graft copolymer) incubator at 37 C. The optical density (OD) of the suspensions was
to resist further evaporation of oil by acting as a sealant layer. The monitored at a wavelength of 600 nm in comparison with the
outer enteric polymer coating was applied using a fluid bed machine control OD.23 The experiment was stopped when the control OD
(Glatt laboratory fluid bed, Binzen, Germany) in 100- to 500-g batches reached 1.2-1.3 or typically between 5 and 6 h of incubation. The
of core granules. The polymer for enteric coating was Eudragit® L 100 minimum inhibitory concentration (MIC)90 of CIN against this E. coli
(Evonik Industries), which dissolves above pH 6 giving targeted K88 strain was defined as the lowest concentration of the EO
release in the jejunum-ileum regions of the intestinal tract. compound that produced a 90% inhibition of the bacteria under-
going log growth.4,24

Assay of CIN and CA Content in Granules  

ODCont  ODEO
Inhibition % ¼ 100 
ODCont
Granules were weighed (50-100 mg), and methanol (10-20 mL)
was added into glass vials. A sonication bath and shaking were used ODCont represents the OD600 of a bacterial culture grown with
to disintegrate granules to fully release CIN. Triplicate samples from the same initial 104 CFU/mL in the absence of any EO (negative
 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133 1127

Table 1
The Oil-Adsorbing Capacity of Various Powders

Powder Type Grade Oil-Adsorbing Capacity % wt/wt CIN Specific Surfacea (m2/g)

Avicel® PH102 1.532 ± 0.083 61.66 ± 0.99 1-1.35

Avicel® RC-591 0.609 ± 0.022 39.00 ± 0.84 e
Avicel® CL-611 0.676 ± 0.010 41.52 ± 0.34 e
Neusilin® UFL2 4.023 ± 0.018 80.86 ± 0.07 300-339
Neusilin® US2 5.074 ± 0.107 84.19 ± 0.28 300-339
Wheat bran e 0.970 ± 0.039 50.4 ± 1.00 e

Values represented in average ± SD (n  3).

a
Values obtained from Westermarck et al and Hentzschel et al25,26 and product bulletin.

control). The minimum bactericidal concentration (MBC) of CIN molten temperatures or at room temperature (after cooling and/or
against this E. coli K88 strain was taken as the lowest concentration resolidification). Regions of phase diagram were classified as
that produced no observable growth after transferring to fresh follows:
growth media for up to 6 h, followed by plating dilutions (20-100
mL) onto trypticase soy agar plates incubated overnight (16-24 h) to  Powder: The mixture was loose and powdery due to under-
identify the number of surviving CFUs per milliliter. These experi- saturation of the US2 and insufficient interpowder binding. At
ments were repeated at least twice. molten temperature, the mixture was not dispersible in the
aqueous continuous phase unless further liquid adsorption
Results occurred. After cooling down, there was insufficient interpowder
binding by the FA component to hold the mixture together.
Formulation and Properties of CIN Core Granules  Paste: The mixture appeared wet, weakly held together, and not
free flowing, containing oversaturation of the US2 powder by
Selection of Adsorbent Powder With Highest Capacity liquid (CIN and FA) at molten temperatures. After the mixture
Oil retention capacity of 3 types of pharmaceutical or food grade cooled, an excess of liquid CIN remained and weakened binding
powders was evaluated to find a suitable powder with the highest due to insufficient amount of FA.
oil adsorption. Table 1 compares the oil retention capacity of these  Liquid: The mixture was liquidlike, that is, free flowing under its
powders. Among the tested powders, Neusilin® US2 shows the own weight. It was easy to disperse into uniform droplets at
highest adsorbing capacity of 84%, attributable to its high specific above molten temperature.
surface area (Table 1), and was thus selected for further study.  Solid: A sufficient amount of resolidified FA to hold the mixture
Although the powders could hold a large amount of oil, once above together with little to no excess oil visible. On cooling, granules
~78% wt/wt the powder surface appeared wet with oil due to became uniformly solid; when dry, they were free flowing and
powder saturation. Furthermore, powders containing oils were round in shape due to good dispersability as droplets in the
fairly soft and weak and appeared almost like a wet paste (Fig. 2), so molten state.
a third component was added to allow formation into granules that
are a meltable FA binder. For ease of dispersibility into small droplets and forming gran-
ules, a molten mixture with liquid behavior was preferred corre-
sponding to >80% liquid and <16% powder. When the liquid
Selection of FA Binder to Obtain Sufficiently Strong Granules
saturation of US2 was approached or exceeded, the mixture
Saturated FAs were chosen as the binder for their well-defined
became more easily dispersible into droplets. But in order for
melting point (to allow quick encapsulation by rapid cooling) and
resolidification of granules to occur on cooling, sufficient FA needed
their lipophilic property. LA was initially chosen for its low melting
to be present in the mixture to bind the powder together on
point at ~44 C and for being easily resolidified at room temperature.
resolidification. The region where a paste occurred was between
When molten, LA blended easily with the oil-soaked powder and
15% and 20% US2 and 80%-84% liquid, and such mixtures were not
when cooled and dried overnight, spherical granules were obtained.
processable into droplets and resulted in large globs being formed,
However, LA-based granules were mechanically too weak to with-
even under high-speed stirring (>1000 rpm). The powder region
stand further processing in a fluid-bed coating machine due to
(blue) was where there was 80% liquid and 20% powder, and due
problems such as attrition and fusing of the granules. To address
to undersaturation of the powder, the mixture was not dispersible
these issues, 2 approaches were undertaken: blending with higher
within a liquid continuous phase, which was unfavorable for
melting temperature FAs or fully substituting with another FA type. It
forming CIN-containing droplets. The region of the phase diagram
was found that blending of different FAs yielded mixtures that were
with maximal yield of granules with high EO loading is outlined in
more difficult to solidify than pure FA. After experimenting further
red. The type of FA used to make granules did not affect the oil
with other excipients, it was concluded that the granule strength
loading properties (i.e., a similar phase diagram could be used for
was greatest when a single FA was used instead of mixtures.
the other FAs). It was observed that use of a higher melting tem-
perature FA resulted in harder granules that could be felt by
Optimizing Formulation of Core Granules Via Phase Diagram
grinding the particles between fingers because neither the oil nor
the powder possessed appreciable binding ability, and hence, the
A phase diagram was used as an aid to optimizing oil loading in
granule solidification was essentially resulting from resolidification
the final mixture and to find compositions that would allow uni-
of the FA component.
form, completely resolidified granules that were suitable for further
processing, as shown in Figure 3. The phase regions and boundaries
were drawn as a result of the processability limits of the mixtures Process Conditions and Properties of Core Granules
(see Supplementary Table and Figure S1 for actual composition
points tested). There were 4 types of phase states that were Table 2 presents the process and granule characteristics of
observed, based on the physical processability of the mixture at batches prepared under a range of conditions. From these studies,
1128 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133

Figure 4. It was found that when the US2 remained constant (10%),
increase in PA concentration from 40% to 45% produced a higher
proportion of larger granules; similarly, when the CIN concentra-
tion kept constant (45%), increasing the US2 concentration from
10% to 13% produced higher fractions of larger granules, especially
those >1400 mm. The encapsulation efficiency was similar among
the formulations tested (~93%). The formulation 48% CIN:10%
US2:42% PA yielded hard granules, with a high loading efficiency,
and had the highest weight fraction of granules in the range of 600-
1400 mm, which are suitable for processing in fluid-bed coating and
are in a product form convenient for mixing with animal feed. Thus,
this formulation was selected for further study. Formulations with
>10% US2 produced much larger fractions of granules >1400 mm
due to the apparent viscosity-enhancing effect of US2 on the
molten mixture (and other suspensions in general).

Selection of Antioxidant to Stabilize CIN in the Core Granules Against

Atmospheric Oxidation (Autoxidation)

After incorporating pure CIN in initial batches of granules, the

EO was unstable as it underwent autoxidation into CA fairly rapidly
on exposure to the atmosphere which led to loss of antimicrobial
activity as explained in the following section. As shown in Figure 5a,
the rapid conversion of CIN into CA started as soon as granules were
exposed to atmospheric oxygen overnight. Without any antioxidant
protection, >70% CIN (of the initial loading) degraded into CA
within 5 d at room temperature and within 30 d at 4 C; the rate of
decomposition was slower (Fig. 5a). To counteract this undesirable
degradation, a series of potential antioxidative compounds were
screened. Among the tested compounds (BHT, carvacrol, eugenol,
and t-butyl methyl phenol), it was found that proton-donating
compounds (phenols) were the most effective in slowing down or
inhibiting autoxidation. Eugenol was selected for further study on
account of its natural presence in cinnamon oils, primarily in the
plant leaves.27,28 It was found that a concentration of 1% vol/vol
eugenol added to pure CIN was sufficient for protection against CIN
autoxidation in granules for at least 1 y at both temperatures. After
addition of the antioxidant to CIN, granules made with PA and LA
showed the same degree of stability. Coated granules containing
CIN with 1% eugenol were also stable for at least 1 y when stored
either at room temperature or refrigerator temperatures as seen in
Figure 5b.

Properties of Coated Granules

Two types of core granules produced from LA and PA were

coated by a fluid-bed process. The first type of granules made using
LA are shown in Figure 6a. The CIN in these granules did not contain
an antioxidant, so the CIN content was low in the final product (~6%
wt/wt), and granules were solid enough to be coated due to con-
version to the solid CA. However, after inclusion of an antioxidant
(1% eugenol) in subsequent batches of LA granules, the granules
Figure 2. Pictures of Neusilin US2 at various CIN oil loading: top 0%, middle 62%, and were not strong enough to withstand the fluid-bed coating process;
bottom 78%.
so, higher melting temperature FAs were investigated and used.
The stronger granules based on PA were thus produced and coated.
The core granules contained 42% CIN and the coated granules
optimized conditions were found and then used to produce 0.8- to contained 20% CIN (with 1% eugenol) due to a weight gain of ~30%-
1-kg batch granules for subsequent coating. The shape and size of 35% from coating. Both types of granules were fairly round, and the
granules were found to be highly affected by the fluid properties of coating uniform as shown in Figure 6.
the molten mixture. Cooling of the mixture after stirring stopped
yielding of more spherical granules, whereas addition of cooling In Vitro Release of CIN From Core Granules and Coated Granules
water into the suspension disrupted droplet shape during cooling.
The sieved size distribution of finished batches prepared at fixed Figure 7a compares release profiles of CIN from uncoated core
conditions (300 rpm, 150-g batch, 300-mL continuous phase in granules prepared with different fats in SIF at pH 6.8 at 37 C. LA-
600-mL vessel), but different formulation composition is shown in based granules showed the highest rate and extent of release,
 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133 1129

Figure 3. Phase diagram for a 3-component mixture at molten (fatty acid) and cooled temperatures. Dot indicates optimized formulation. Outline indicates range of formulation
components that resulted in acceptable granules.

followed by MA, PA, SA, and cetyl alcohol (CAl). Without coating, of E. coli K88 growth (initial 104 CFU/mL) by CIN at various con-
CIN was released rapidly from granules with >80% being released centrations in a liquid medium (TSB). The low activity of CA sug-
from LA-, MA-, and PA-based granules within 1 h. SA- and CAl- gested preventative measures against oxidation of CIN to CA using
based granules showed lower extent of release probably attribut- antioxidant. From the curves in Figure 8b, the MBC of free CIN oil
able to their higher hydrophobicity. and CIN encapsulated into PA and LA granules against E. coli K88
Core granules were coated with an enteric polymer (Eudragit® were found to be 600, 550, and 450 mg/mL, respectively.
L 100 [Evonik Industries]) to suppress the release of CIN in gastric
conditions and enable the release after dissolution of the coating
polymer at pH > 6. The in vitro release profiles were determined in a Discussion
2-stage test: 2 h in SGF at pH 1.2 and then continued in SIF at pH 6.8
for 3 h. For the uncoated granules, the release test was undertaken in Importance of Formulation on the Properties of Granules
SIF. Figure 7b shows pH-dependent release of CIN from the coated
granules which is absent for the uncoated core granules. The enteric The current results indicate that the melting point of the spe-
coating prevented CIN from release in the SGF significantly with cific FA used predominately affected granule strength because
>70% of the loaded CIN being released in the SIF. Ultimately, after granules were primarily held together by intermolecular van der
dissolution of the coating material, >90% of the CIN was released. Waals forces between FA molecules. As FA chain length increases,
its melting temperature increases, and granules become harder at
Antimicrobial Activity of CIN Against E. Coli K88: MIC and MBC room temperature as compared with those made from a shorter-
Determination chain FA. As such, PA yielded harder granules than LA due to its
higher melting temperature resulting from the longer FA chain
The effectiveness of antimicrobials against bacteria growth and length (C16 vs. C12). Thus, the formulation with LA was
viability can be generally expressed in 2 forms: MIC90 and MBC. substituted with PA to obtain harder granules at room tempera-
Inhibition of growth was determined for different concentrations of ture. FA blends were not used for making stronger granules
CIN and its oxidation product CA. The MIC90 of CIN against E. coli because blends did not have a well-defined melting transition
K88 was found to be 150 mg/mL, whereas CA showed much lower temperature compared with pure FAs, possibly due to less
inhibitory activity (>1000 mg/mL). Hence, no MBC determination
for CA was carried out due to the low inhibitory activity against the
target strain as shown in Figure 8a. Figure 8a shows the inhibition

Table 2
Granule Properties and Process Characteristics

Properties and variables Value or range

Oil loading, % wt/wt (actual/theoretical) 16.9-45/25-48

Oil loading/encapsulation efficiency ~93%
RPM range 150-375
Granule batch size range 0.5-5 mm
Batch size 20-200 g
Process yield >90%
Molten:aqueous phase ratio 0.13-0.66
Formulation components ranges CIN 20%-50%
US2 0%-10%
PA 40%-60%
Optimum ratio of CIN:US2:PA 48:10:42
Figure 4. Effect of formulation composition changes on the size distribution of gran-
RPM, rotations per minute. ules while keeping batch size and stirring rate constant.
1130 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133

The behavior of the mixtures could be explained by the

saturation point of the Neusilin® powder component. As shown
in Figure 2, Neusilin® powder became saturated around 78%-80%
wt/wt liquid. Below this saturation point, the mixture still exists as
a powder irrespective of the temperature; hence, it cannot be
dispersed well in an aqueous continuous phase without taking up
water to become dispersible (not processable). Thus, when at
molten temperatures, the 3-component mixture simplifies to a 2-
phase, liquid (FA and EO) and solid (neusilin US2) mixture and the
level of powder saturation becomes dominant in affecting the
properties of the mixture. Figure 3 shows that at molten temper-
atures, the FA and CIN are both miscible, and they penetrated the
powder as 1 liquid; depending on the extent of US2 saturation, this
mixture behaved as a liquid when near or above saturation of the
powder (~84% liquid:~16% powder).
It was observed that the granule formulation affected the size
distribution primarily by the degree of liquid saturation of the
Neusilin® US2 powder (Fig. 4). Granules became larger as the
amount of US2 was increased from 10% to 13% while the degree of
powder saturation decreased. As the level of available liquid
(CIN þ molten FA) saturated the available powder (e.g., >48% CIN,
US2 10%), smaller droplets were formed which led to smaller
granules as the molten mixture behaved as a low-viscosity liquid.
In contrast, as the saturation level of US2 is decreased or if the US2
>10%, the powder exerts a viscosity-enhancing effect on the
molten liquid mixture, leading to larger droplets and granules.
This effect is one of the properties of magnesium aluminum sili-
cates, and other silicates widely used in cosmetics.30 However, if
the powder was oversaturated with oil, this effect appeared less
influential in the present method. After cooling to room temper-
ature, the FA resolidified while CIN remained liquid because no
chemical reactions among the components were expected or
observed. Due to displacement of CIN by molten FA during mixing
with US2, the maximum loading in the cooled state was about
Figure 5. (a) Stability of CIN in uncoated granules with and without antioxidant when
48%-50% or less than that without FA. Incorporating more than
stored at 23 C and 4 C. (b) Stability of coated granules with antioxidant (1% EUG) stored
at 23 C and 4 C for up to 1 y. EUG, eugenol; RT, room temperature; AX, antioxidant. 50% CIN initially led to excess superficial CIN appearing, and a
paste or liquid was obtained at room temperature instead of
granules. The loading limit was thus determined to be 50% CIN in
effective packing between the dissimilar chain length FA mole- the final formulation of granules.
cules (e.g., LA with SA). With regard to the amount of CIN that PA was chosen in formulating larger batches of granules for
could be accommodated in the granules, the independence of FA offering the best combination of higher granules strength and fast,
type could be due to similar volume being occupied by these more complete release of the encapsulated CIN, in addition to the
different FAs in both molten and solid states. Structurally, the lower temperature required for preparation vs SA. Thus, using a
different FAsdLA, MA, PA, SA (12, 14, 16, and 18 C)dand fatty higher melting FA like PA yielded harder granules than LA- and MA-
alcohol CAl (C16) differ only by a few (-CH2-) groups. When based granules yet released the encapsulated oil better than SA or
granules were made with 2 components, that is, oil and FA, the CAl. However, fatty alcoholebased granules resulted in decreased
loading maximum was about 28% wt/wt for LA, MA, PA, and SA. extent of release possibly due to the less polar nature of the fatty
The CAl, on the other hand, allowed loading up to 40%, without alcohol compared with FAs.
excess oil being visible after cooling. Including Neusilin® US2 in Coated granules exhibited only partial resistance to release in SGF
the formulation allowed increased oil loading levels up to 48% wt/ with around 27% of the CIN being released from coated granules. The
wt due to its high oil-adsorbing capacity. early release of the activity in SGF could have been caused by the
Constructing the 3-component phase diagram allowed us to migration of the volatile liquid oil from the cores to the coating layers
define optimum ratios of the mixture components for production of during the fluid-bed process. The long coating times typically used
core granules with high oil loading, good dispersibility (affected the (>6 h) as well as the complex processes occurring during film for-
size and shape of granules) at molten temperatures, and good mation31 could have contributed to this, as well as loss of the activity
mechanical strength (hardness) after solidification, which was as the constant fluidization by air could have drawn away some of
necessary to endure subsequent processing steps. This analysis led the volatile CIN from the cores. Further optimization of the fluid-bed
to an optimal weight ratio of each component of CIN:US2:PA at coating process or selecting an alternative coating process32 (e.g., dry
48:10:42, and this formulation was then used for scale-up pro- coating) could be evaluated in future experiments.
duction. Granules were coated with Eudragit® L 100 (Evonik
Industries) to render pH-dependent release behavior with fast Antibacterial Activity and Storage Stability of CIN Granules
release at pH > 6 at the jejunum, ileum, and cecum sections of the
intestines. For some applications, though, enteric coating may not E. coli K88 is an enterotoxigenic strain of E. coli that commonly
be required where a slow release into the environment is more causes piglet diarrhea during the postweaning period when ani-
desirable to reduce pathogens on the feedlot floor surface.29 mals are most susceptible to infection.33 This particular strain of
 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133 1131

Figure 6. (a) Lauric acid granules uncoated (left) and after coating (right) and (b) palmitic acid granules with 42% wt/wt CIN, coated with Eudragit® L100 and subcoated with
Kollicoat® IR.

E. coli produces K88 pili that adhere at the jejunum and ileum PA-based granules had the same antibacterial activity as the pure
epithelial regions of the pig intestine, where specific receptor sites oil and did not exhibit an enhanced effect as LA due to its higher
are expressed at the mucous surface,34 and then produce toxins melting temperature and lower solubility. The MBC of CIN ob-
that cause tissue damage and diarrhea. Furthermore, compared tained in the present study was about 2-6 times higher than
with other strains of E. coli, K88 is resistant to a spectrum of previous reports.4,7 This difference could be attributed to varia-
antibiotics but was found to be effectively inhibited by EOs like tions in some minor EO components normally found in EO prod-
CIN.4 Even though CA has little inhibitory activity against E. coli ucts from different sources, and furthermore, EO mixtures can
K88, it has good inhibitory activity against a variety of fungi35 and exhibit stronger antibacterial activity than the respective single
spoilage bacteria36 with MICs in the 1-5 mM (150-750 mg/mL) EO component.38 By using the current strategy of encapsulating
range. LA-based granules showed an overall better killing effect CIN (or other EOs) into granules and coating with an enteric
with a lower MBC occurring at 450 mg/mL compared with the MBC polymer, a sufficient concentration (MIC to MBC range) of EO can
of 600 mg/mL for free CIN oil (Fig. 8b), whereas control granules potentially be delivered to the lower intestinal tract of pigs (site of
containing only excipient (US2 and either LA or PA at the equiv- disease) while minimizing or avoiding the early absorption of EO
alent CIN concentrations ~450-1000 mg/mL) did not show signif- compounds in the upper GI tract.9 Such granule formulations
icant inhibitory activity against E. coli K88, suggesting that a could be an effective strategy, without using antibiotics, to combat
synergistic activity of the LA-CIN mixture might be happening. piglet postweaning diarrhea caused by E. coli K88.
This result could be attributed to the antibacterial effect of FAs Because the antibacterial activity of CA was determined to be
that were used in the granule formulation37 in addition to the much less than CIN against the target pathogen, it is necessary to
melting point depression effect after mixing with the CIN oil, prevent autoxidation of CIN in granules. This work has demon-
which allowed the granules to melt and disintegrate at the incu- strated that with addition of antioxidant 1% eugenol to CIN, gran-
bation temperature of 37 C (see Figure S1), allowing dispersal of ules made with PA and LA showed the same degree of stability, with
LA and CIN within the test medium. Further study into this effect or without coating, which can last for at least 1 y at room tem-
could be performed, but due to difficulty in coating LA granules, perature or refrigerator temperatures (Fig. 5b). These positive
PA-based granules were chosen for subsequent development. results prompted the exploration of in vivo performance of the
1132 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133

Figure 8. (a) Inhibitory activity of CIN vs CA toward E. coli K88 grown in TSB culture
based on change in OD600 over 6 h at 37 C with 200 rpm shaking. (b) Antibacterial
Figure 7. (a) Release of CIN in SIF (pH, 6.8) from core granules formulated with
activity of CIN oil and granules against E. coli K88 in TSB medium. Data points represent
different fatty acid types and (b) release of CIN from coated granules under 2-stage
average ± SD.
dissolution. Values represent average ± SD.

coated CIN granules in a subsequent study which will appear in Acknowledgments

another publication (Y.H. Ma et al, unpublished data, 2015).
The authors thank the Natural Science and Engineering
Conclusions Research Council (grant no. RGPIN 170460-13), Agriculture and
Agri-Food Canada (grant no. J-000235.001.04), and the Canadian
Granular formulations of CIN, a model EO compound with good Poultry Research Council for funding the research. Escherichia coli
antimicrobial activity, were successfully developed with loadings K88 strain JG280 was a gift from Dr. C. Gyles, University of
up to 50% wt/wt. The core granules were prepared by a melt dis- Guelph.
persionesolidification method using rationally selected oil-
adsorbing powder Neusilin® US2 and lipids at an optimized ratio
of CIN:US2:lipid. The enteric polymer-coated granules based on PA References
exhibited fast release of CIN at intestinal pH, suitable for site-
1. Maron DF, Smith TJ, Nachman KE. Restrictions on antimicrobial use in food
specific delivery of CIN in the lower intestines where E. coli reside animal production: an international regulatory and economic survey. Global
most. The addition of antioxidant eugenol to the granule formula- Health. 2013;9:48.
tion significantly increased CIN stability to at least 1 y. The LA- 2. Ferber D. WHO advises kicking the livestock antibiotic habit. Science.
2003;301(5636):1027.
encapsulated CIN granules were found more effective in vitro 3. Burt S. Essential oils: their antibacterial properties and potential applications in
against E. coli K88 than free CIN with about 33% lower MBC but foodsda review. Int J Food Microbiol. 2004;94(3):223-253.
were unsuccessfully coated by fluid bed, while PA-based granules 4. Si W, Gong J, Tsao R, et al. Antimicrobial activity of essential oils and struc-
turally related synthetic food additives towards selected pathogenic and
which were coated successfully exhibited an MBC similar to the free beneficial gut bacteria. J Appl Microbiol. 2006;100(2):296-305.
oil. The in vitro results obtained here indicate the potential of CIN- 5. Bakkali F, Averbeck S, Averbeck D, Idaomar M. Biological effects of essential
containing granules in combating E. coli K88 and other susceptible oilsda review. Food Chem Toxicol. 2008;46(2):446-475.
6. Michiels J, Missotten J, Fremaut D, De Smet S, Dierick N. In vitro doseeresponse
pathogens. Further in vivo studies are warranted to verify the
of carvacrol, thymol, eugenol and trans-cinnamaldehyde and interaction of
effectiveness of CIN granules. The formulations and method combinations for the antimicrobial activity against the pig gut flora. Livest Sci.
developed in this work for encapsulation of oils in solid granules 2007;109(1-3):157-160.
are relatively simple and economical and can potentially be used to 7. Michiels J, Missotten JAM, Fremaut D, De Smet S, Dierick NA. In vitro charac-
terisation of the antimicrobial activity of selected essential oil components and
encapsulate a variety of pure EOs and their mixtures or other binary combinations against the pig gut flora. Anim Feed Sci Technol.
lipophilic liquid actives. 2009;151(1-2):111-127.
 Y.-H. Ma et al. / Journal of Pharmaceutical Sciences 105 (2016) 1124e1133 1133

8. Busquet M, Calsamiglia S, Ferret A, Cardozo PW, Kamel C. Effects of cinna- 24. Hili P, Evans CS, Veness RG. Antimicrobial action of essential oils : the effect of
maldehyde and garlic oil on rumen microbial fermentation in a dual flow dimethylsulphoxide on the activity of cinnamon oil. Lett Appl Microbiol.
continuous culture. J Dairy Sci. 2005;88(7):2508-2516. 1997;24(4):269-275.
9. Michiels J, Missotten J, Dierick N, Fremaut D, Maene P, De Smet S. In vitro 25. Westermarck S, Juppo AM, Kervinen L, Yliruusi J. Microcrystalline cellulose and
degradation and in vivo passage kinetics of carvacrol, thymol, eugenol and its microstructure in pharmaceutical processing. Eur J Pharm Biopharm.
trans-cinnamaldehyde along the gastrointestinal tract of piglets. J Sci Food Agr. 1999;48(3):199-206.
2008;88(13):2371. 26. Hentzschel CM, Sakmann A, Leopold CS. Suitability of various excipients as
10. Michiels J, Missotten J, Van Hoorick A, et al. Effects of dose and formulation of carrier and coating materials for liquisolid compacts. Drug Dev Ind Pharm.
carvacrol and thymol on bacteria and some functional traits of the gut in 2011;37(10):1200-1207.
piglets after weaning. Arch Anim Nutr. 2010;64(2):136-154. 27. Friedman M, Kozukue N, Harden LA. Cinnamaldehyde content in foods deter-
11. Baser KHC, Franz C. Essential oils used in veterinary medicine. In: Baser KHC, mined by gas chromatography-mass spectrometry. J Agric Food Chem.
Buchbauer G, eds. Handbook of Essential Oils: Science, Technology, and Applica- 2000;48(11):5702-5709.
tions. Boca Raton, FL: CRC Press; 2009:881-894. 28. Cheng SS, Liu JY, Tsai KH, Chen WJ, Chang ST. Chemical composition and
12. Ukita K, Murakami T. Preparation of essential oils loaded granule by melt mosquito larvicidal activity of essential oils from leaves of different Cinna-
granulation. Drug Dev Ind Pharm. 1994;20(6):981-992. momum osmophloeum provenances. J Agric Food Chem. 2004;52(14):4395-
13. Müller RH, Ma €der K, Gohla S. Solid lipid nanoparticles (SLN) for controlled drug 4400.
deliveryda review of the state of the art. Eur J Pharm Biopharm. 2000;50(1): 29. Varel VH, Miller DN, Berry ED. Incorporation of thymol into corncob granules
161-177. for reduction of odor and pathogens in feedlot cattle waste. J Anim Sci.
14. Benita S. Microencapsulation: Methods and Industrial Applications. 2nd ed. New 2006;84(2):481-487.
York: Taylor & Francis; 2006:756. 30. Elmore A. Final report on the safety assessment of aluminum silicate, calcium
15. Wong HL, Bendayan R, Rauth AM, Li Y, Wu XY. Chemotherapy with anticancer silicate, magnesium aluminum silicate, magnesium silicate, magnesium trisi-
drugs encapsulated in solid lipid nanoparticles. Adv Drug Deliv Rev. 2007;59(6): licate, sodium magnesium silicate, zirconium silicate, attapulgite, bentonite,
491-504. Fuller's earth, hectorite, kaolin, lithium magnesium silicate, lithium magne-
16. Wong HL, Bendayan R, Rauth AM, Wu XY. Development of solid lipid nano- sium sodium silicate, montmorillonite, pyrophyllite, and zeolite. Int J Toxicol.
particles containing ionically complexed chemotherapeutic drugs and che- 2003;22:37.
mosensitizers. J Pharm Sci. 2004;93(8):1993-2008. 31. Porter SC, Felton LA. Techniques to assess film coatings and evaluate film-
17. Chattopadhyay N, Zastre J, Wong HL, Wu XY, Bendayan R. Solid lipid nano- coated products. Drug Dev Ind Pharm. 2010;36(2):128-142.
particles enhance the delivery of the HIV protease inhibitor, atazanavir, by a 32. Obara S, Maruyama N, Nishiyama Y, Kokubo H. Dry coating: an innovative
human brain endothelial cell line. Pharm Res. 2008;25(10):2262-2271. enteric coating method using a cellulose derivative. Eur J Pharm Biopharm.
18. Voinovich D, Moneghini M, Perissutti B, Franceschinis E. Melt pelletization in 1999;47(1):51-59.
high shear mixer using a hydrophobic melt binder: influence of some appa- 33. Fairbrother JM, Nadeau E, Gyles CL. Escherichia coli in postweaning diarrhea in
ratus and process variables. Eur J Pharm Biopharm. 2001;52(3):305-313. pigs: an update on bacterial types, pathogenesis, and prevention strategies.
19. Hamdani J, Moes AJ, Amighi K. Development and evaluation of prolonged Anim Health Res Rev. 2005;6(1):17-39.
release pellets obtained by the melt pelletization process. Int J Pharm. 34. Jin L, Zhao X. Intestinal receptors for adhesive fimbriae of enterotoxigenic
2002;245(1-2):167-177. Escherichia coli (ETEC) K88 in swineda review. Appl Microbiol Biotechnol.
20. Esposita E, Cortesi R, Nastruzzi C. Production of Lipospheres for Bioactive Com- 2000;54(3):311-318.
pound Delivery. Lipospheres in Drug Targets and Delivery: Approaches, Methods 35. Faria NC, Kim JH, Gonçalves LA, Martins Mde L, Chan KL, Campbell BC.
and Applications. New York, NY: CRC; 2005:23-40. Enhanced activity of antifungal drugs using natural phenolics against
21. Amezcua R, Friendship RM, Dewey CE, Gyles C, Fairbrother JM. Presentation of yeast strains of Candida and Cryptococcus. Lett Appl Microbiol. 2011;52(5):
postweaning Escherichia coli diarrhea in southern Ontario, prevalence of he- 506-513.
molytic E. coli serogroups involved, and their antimicrobial resistance patterns. 36. Roller S, Seedhar P. Carvacrol and cinnamic acid inhibit microbial growth in
Can J Vet Res. 2002;66(2):73-78. fresh-cut melon and kiwifruit at 4 and 8 C. Lett Appl Microbiol. 2002;35(5):
22. Sawyer DT, Heineman WR, Beebe JM, Reilley CN. Chemistry Experiments for 390-394.
Instrumental Methods. New York, NY: Wiley; 1984. 37. Kabara JJ, Swieczkowski DM, Conley AJ, Truant JP. Fatty acids and derivatives as
23. Si W, Gong J, Chanas C, et al. In vitro assessment of antimicrobial activity of antimicrobial agents. Antimicrob Agents Chemother. 1972;2(1):23-28.
carvacrol, thymol and cinnamaldehyde towards Salmonella serotype typhi- 38. Fratini F, Casella S, Leonardi M, et al. Antibacterial activity of essential oils, their
murium DT104: effects of pig diets and emulsification in hydrocolloids. J Appl blends and mixtures of their main constituents against some strains supporting
Microbiol. 2006;101(6):1282-1291. livestock mastitis. Fitoterapia. 2014;96:1-7.

View publication stats