Biomaterials 27 (2006) 4239–4249

Controlled release of gentamicin from calcium phosphate—

poly(lactic acid-co-glycolic acid) composite bone cement

Julia Schnieders

a

, Uwe Gbureck

b,

, Roger Thull

b

, Thomas Kissel

a

a

Department of Pharmaceutics and Biopharmacy, Philipps-University Marburg, Ketzerbach 63, 35032 Marburg, Germany

b

Department of Functional Materials in Medicine and Dentistry, University of Wu¨rzburg, Pleicherwall 2, 97070 Wu¨rzburg, Germany

Received 12 January 2006; accepted 21 March 2006

Abstract

Modification of a self setting bone cement with biodegradable microspheres to achieve controlled local release of antibiotics without

compromising mechanical properties was investigated. Different biodegradable microsphere batches were prepared from poly(lactic-co-
glycolic acid) (PLGA) using a spray-drying technique to encapsulate gentamicin crobefate varying PLGA composition and drug loading.
Microsphere properties such as surface morphology, particle size and antibiotic drug release profiles were characterized. Microspheres
were mixed with an apatitic calcium phosphate bone cement to generate an antibiotic drug delivery system for treatment of bone defects.
All batches of cement/microsphere composites showed an unchanged compressive strength of 60 MPa and no increase in setting time.
Antibiotic release increased with increasing drug loading of the microspheres up to 30% (w/w). Drug burst of gentamicin crobefate in the
microspheres was abolished in cement/microsphere composites yielding nearly zero order release profiles. Modification of calcium
phosphate cements using biodegradable microspheres proved to be an efficient drug delivery system allowing a broad range of 10–30%
drug loading with uncompromised mechanical properties.
r

2006 Elsevier Ltd. All rights reserved.

Keywords: Bone cement; Hydroxyapatite; Biodegradation; Microspheres; Antimicrobial; Drug delivery

1. Introduction

Conventional treatment of acute and chronic bacterial

osteomyelitis includes surgical removal of necrotic bone
tissue and repeated irrigations combined with high systemic
doses of antibiotic drug substances for prolonged periods
of time

[1]

. This treatment frequently causes unwanted side

effects and often fails to cure bacterial bone infections.
Therefore, drug delivery systems providing controlled
release of antibiotics were studied for local therapy as an
alternative

[2]

. Currently used local antibiotic delivery

systems are based on poly(methyl-methacrylate) (PMMA)
beads, chains or self setting PMMA bone cement

[3]

. A

disadvantage of these materials is the need for surgical
removal of the spent devices and the risk of induction of
antibiotic resistance when insufficient doses are adminis-

tered

[4]

. In addition, PMMA cements are known to cause

cytotoxic effects due to residual monomers, and tissue
necrosis due to exothermic polymerization reaction tem-
peratures exceeding 70 1C

[5]

.

More recently drug delivery systems were developed

from biodegradable and osteoconductive materials such as
degradable polymers or calcium phosphate compounds

[6]

.

Research efforts were focused on ceramic materials, e.g.
sintered hydroxyapatite or self-setting calcium phosphate
cements (CPC), as antibiotic carriers for the treatment of
bone infections due to their similarity with the mineral
phase of bone

[7,8]

. A major advantage of ceramic bone

cements is their mechanical strength after setting to
reinforce damaged bone structures and their application
as injectable cement paste

[9]

. Calcium phosphate cement

modification with various antibiotics (e.g. cephalosporin,
gentamicin, vancomycin or tetracycline) was performed in
the past by adding the water-soluble salts either to the
liquid or solid cement phase during cement mixing

[2,10–12]

. This can cause problems due to an interaction

ARTICLE IN PRESS

www.elsevier.com/locate/biomaterials

0142-9612/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:

10.1016/j.biomaterials.2006.03.032

Corresponding author. Tel.: [phone];

fax: [phone].

E-mail address:

[e-mail] (U. Gbureck).

of the water-soluble drug and the setting reaction of the
cement after adsorption of the drug molecules at the
particle surface

[10,12]

. This behavior was correlated with a

higher cement porosity and lower degree of conversion or
with the presence of chloride ions from the antibiotic salt
(e.g. tetracycline–HCl), which formed a halogenide sub-
stituted hydroxyapatite as reaction product

[13]

.

We hypothesized that interactions of antibiotics with

cement setting could be prevented by microencapsulating
drugs in biodegradable microspheres yielding cement/
polymer composites with satisfactory mechanical proper-
ties. Moreover, it was anticipated that diffusional release of
antibiotics could be controlled by the microencapsulation.
Polymers suitable for microsphere production are chitosan

[14]

, coralline hydroxyapatite-gelatin

[15]

composites, and

poly(lactic acid-co-glycolic acid) polymers (PLGA)

[16]

.

PLGA was successfully used for microencapsulation of
tetracycline

[17]

, rifampicin

[18]

or vancomycin

[19]

. PLGA

was used in this study due to its biocompatibility and
biodegradability to physiological metabolites

[20]

and the

ability to control degradation time by using different
molecular weights or compositions. Degradation times of
one month are useful to prevent antibiotic resistance due to
insufficient polymer degradation or low drug concentra-
tions, like it was described for PMMA beads

[21]

.

Spray drying was used to generate biodegradable

microspheres containing gentamicin crobefate as a poorly
soluble drug substance to prolong the release rate and to
reduce burst release

[22]

. The microspheres were then

incorporated in various ratios into an apatitic cement
matrix made from tetracalcium phosphate (TTCP) and
dicalcium phosphate anhydrate (DCPA). The setting
behavior, mechanical performance and the release kinetics
of cement/microsphere composites were investigated and
compared to the non-modified cement. Biocompatibility of
the materials was tested with L929 fibroblasts using a MTT
assay under in vitro conditions

[23]

.

2. Materials and methods

2.1. Materials

Biodegradable poly(lactic-co-glycolic acid), PLGA 50:50 with different

molecular weights and hydrophilicity were purchased from Boehringer
Ingelheim (Ingelheim, Germany) (see

Table 1

). DCPA and calcium

carbonate were acquired from Malinckrodt-Baker (Griesheim, Germany).
Gentamicin crobefate (EMD 46217) was a gift from Biomet Merck
(Darmstadt, Germany). All other chemicals used in this study were of
analytical grade. TTCP was prepared by sintering an equimolar mixture of
DCPA and calcium carbonate at 1500 1C for 18 h as described previously

[12]

followed by quenching in air and 10 min dry grinding with a planetary

ball mill (PM400, Retsch, Haan, Germany). DCPA was ground in ethanol
for 24 h followed by drying in vacuum. Both cement components were
mixed in the ball mill in an equimolar ratio with the addition of approx.
1 wt% sodium phosphate powder as setting accelerator.

2.2. Preparation of microspheres

Drug loaded microspheres were prepared using a spray drying method

[24]

. A two-fluid nozzle with a diameter of 0.5 mm was used to

manufacture placebo microspheres (MS) as well as gentamicin crobefate
loaded microspheres (GC–MS) in a Bu¨chi 190 Mini Spray Dryer (Flawil,
Switzerland). The settings were as follows: inlet temperature 48–50 1C,
outlet temperature 32–39 1C, air flow 800 Nl/h, aspirator setting 30 mbar
and feed rate 8–10 ml/min. MS batches were collected and subsequently
dried at 25 1C in vacuum to remove residual organic solvents.

Placebo MS were prepared from 5% (w/V) polymer solutions in 100 ml

dichloromethane (DCM). For the GC–MS the drug was dissolved in three
different concentrations (10, 20 and 30% (w/w) in a solution of methanol
(MeOH) and DCM (1:4) and mixed with 100 ml of a 5% (w/V) polymer
solution.

2.3. Characterization of microspheres

Particle size and size distributions of the MS were measured in

isopropanol using a laser particle sizer Horiba LA-300 (Kyoto, Japan)
based on a laser light scattering technique. Approximately 100 mg were
suspended in 200 ml isopropanol and sonicated for 15 min in an ultra-
sound bath (Sonorex RK100 H, Bandelin, Berlin, Germany). Each sample
was measured in triplicate. The weighted average of volume distribution
(D[4.3]) was used to describe the particle size.

ARTICLE IN PRESS

Table 1
Summary of characteristics from placebo and gentamicin crobefate (GC) loaded microspheres (n ¼ 3)

Microparticle batch

Polymer

M

w

(Da)

Average particles size (mm)

a

Yield (%)

b

Drug loading (%)

c

T

g

(1C)

(I) Variation of polymers
MS-1

RG 502 H

14 000

7.1

70.4

6.7

—

38.3

MS-2

RG 502

14 500

9.0

70.0

10.7

—

37.0

MS-3

RG 503 H

28 000

13.0

70.6

30.3

—

41.9

MS-4

RG 503

35 000

14.4

70.1

39.7

—

40.6

MS-5

RG 504 H

53 500

d

27.0

—

42.7

MS-6

RG 504

55 000

d

14.0

—

n.d.

(II) Variation of theoretical loading
GC-1

RG 503 H

28 000

9.2

70.3

36.0

10.0

41.4

GC-2

RG 503 H

28 000

7.7

70.5

37.4

20.0

41.6

GC-3

RG 503 H

28 000

5.4

70.2

41.4

30.0

42.8

GC-4

RG 503

35 000

10.1

71.0

34.9

10.0

42.0

a

Determined by laser diffractometry.

b

100 W

mp

/(W

drug

+W

polymer

), determined gravimetrically.

c

Theoretical loading.

d

No result due to no particle formation.

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4240

Encapsulation efficiency and drug loading were determined spectro-

photometrically after extraction from the microspheres. Briefly, 10 mg of
gentamicin crobefate loaded microspheres were dissolved in 1 ml DCM
and subsequently 5 ml of ultra pure water was added. This mixture was
placed in a rotating shaker (Rotatherm, Liebisch, Bielefeld, Germany) at
30 rpm and 37 1C for 2 h. After centrifugation the aqueous supernatant
was analyzed spectrophotometrically using calibration curves on a
Shimadzu UV 160 spectrophotometer (Shimadzu Europe Ltd/Deutsch-
land GmbH, Duisburg, Germany) at 344 nm, with a sensitivity of 2 mg/ml
and 20 mg/ml as limit of detection. Each sample was measured in triplicate.
Encapsulation efficiency was expressed as follows:

Encapsulation efficiency

¼ ð

actual drug loading=theoretical drug loadingÞ 100.

The morphology of microspheres and cement devices was analyzed by

scanning electron microscopy (SEM) using a Hitachi S 510 scanning
electron microscope (Hitachi Denshi GmbH, Rodgau, Germany). Dried
cement devices were cut using a razor-blade and then mounted on
aluminum stubs using double-sided adhesive tape. The samples were
sputter coated three times with a gold layer at 25 mA in an argon
atmosphere at 0.3 mPa for 2 min (Sputter Coater S 150, Edwards/Kiese,
BOC Edwards GmbH, Kirchheim, Germany) and analyzed with regard to
surface morphology with accelerating voltage of 25 kV. Photographs were
taken using a Pentax MX with a Pentax M 40 mm (1:2.8) objective
(Tokyo, Japan).

Glass transition temperatures (T

g

) were measured using a DSC 7

differential scanning calorimeter (Perkin-Elmer, Wiesbaden Germany).
Placebo and gentamicin crobefate loaded microsphere samples were sealed
in aluminum pans and heated twice under nitrogen atmosphere. The
resulting thermograms covering a range of 10–80 1C were recorded at
heating rates of 10 1C/min. The second run was used for T

g

calculation

referring to the midpoint temperature. Calibration of the system was
performed using an indium standard.

2.4. Preparation of cement cylinders

Cement cylinders were prepared as described previously

[12]

. Briefly

placebo or gentamicin crobefate loaded MS were mixed with the cement
for 5 min using mortar and pistil in a 1:9 ratio(10% (w/w). Cement
cylinders with 6 mm in diameter and 12 mm in height were obtained at a
powder-liquid-ratio (P/L) of 3.3 using ultra pure water. Samples were
prepared by mixing 800 mg of cement for the placebo devices and 80 mg of
GC–MS added to 720 mg of cement for loaded devices with the required
liquid volume in a nitrile rubber mixing container on a vibratory shaker
Thermolyne Maxi Type 37600 (Barnstead International, Dubuque, Iowa,
USA) for 15 s. The cement paste was transferred into stainless steel molds
(6 mm diameter), and biaxially compressed using a self-made cantilever
device at a pressure of 2.7 MPa for 5 s, followed by a load of 700 kPa for
2 h at 37 1C and 100% humidity to form cylinders with a height of 12 mm

[25]

. Specimen were removed from the moulds and stored in an incubator

(GFL Type 1008, Burgwedel, Germany) at 100% humidity and 37 1C for
additional 22 h prior to testing.

2.5. Characterization of cement cylinders

Mechanical strengths of wet samples were measured at a crosshead

speed of 1 mm/min after 24 h hardening at 37 1C, using a static mechanical
testing device Zwick 1440 (Zwick, Ulm, Germany) with a 5 kN load cell.
The initial setting time of the cements was measured in a humidity
chamber at 37 1C and 490% humidity and in normal laboratory
atmosphere (20–23 1C and 50–60% humidity), respectively using the
Gilmore needle test with a needle of 113.98 g and 2.117 mm diameter
according to ASTM standard

[26]

.

X-ray diffraction patterns of the set cements were recorded on a D 5005

diffractometer (Siemens, Karlsruhe Germany). Data were collected from
2y ¼ 202401 with a step size of 0.021 and a normalized count time of 1 s/
step using Cu K

a

radiation. The phase composition was analyzed by

means of the Joint Committee for Powder Diffraction Studies (JCPDS)
reference patterns for TTCP (PDF Ref. 25-1137), HA (PDF Ref. 09-0432)
and DCPA (PDF Ref. 09-0080). Quantitative phase compositions of the
materials were calculated by means of total Rietveld refinement analysis
with the TOPAS software (Bruker AXS, Karlsruhe, Germany). As
references database structures of TTCP, HA and DCPA were used
together with a Chebychev forth-order background model and a Cu K

a

emission profile.

2.6. Gentamicin crobefate release from microspheres and
composites under in vitro conditions

Drug release was determined by suspending 10 mg of drug loaded

microspheres or one cement device in 5 ml of PBS (pH 7.4). The particles
were wetted prior to suspension using PBS containing 0.05% Myrj 52. The
15 ml Pyrex-vials were placed in an orbital shaker maintained at 37 1C and
rotated at 30 rpm (Liebisch, Bielefeld, Germany). Samples were withdrawn
at regular time intervals after centrifugation with 5000 rpm for 10 min
using a laboratory centrifuge Sigma 203 (Sigma, Osterode am Harz,
Germany) and replaced by fresh medium. The samples were analyzed
spectrophotometrically at 344 nm with a Shimadzu UV 160 instrument
(Shimadzu Europe Ltd/Deutschland GmbH, Duisburg, Germany).

2.7. In vitro cytotoxicity testing in L929 fibroblasts

In vitro cytotoxicity of the polymers was evaluated using an MTT assay

as described previously

[23,27]

. Briefly, L 929 cells were seeded into 96-well

microtiter plates (Nunclon

TM

, Nunc, Wiesbaden, Germany) at a density

of 8000 cells/well. After 24 h the culture medium was replaced with 100 ml/
well of serial dilutions of polymer stock solutions in antibiotic-free
DMEM (n ¼ 8). Polymer extracts were prepared by using up to 10 mg MS
per milliliter medium. The extracts were incubated at 37 1C for 24 h in a
Rotatherm apparatus. If necessary the extracts were neutralized and then
sterilized by filtration. After an incubation period of 24 h MTT (3-(4,5-
dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide) was dissolved in
phosphate buffered saline at 5 mg/ml and 20 ml was added to each well
reaching a final concentration of 0.5 mg MTT/ml. After an incubation
time of 4 h un-reacted dye was removed by aspiration and the purple
formazan product was dissolved in 200 ml/well dimethylsulfoxide and
quantified by a plate reader (Titertek Plus MS 212, ICN Biomedicals,
Eschwege, Germany) at wavelengths of 570 and 690 nm. The relative cell
viability related to control wells containing cell culture medium without
polymer was calculated by [A] test/[A] control 100, with [A] as the
concentration of viable cells. Poly(ethylene imine) 25 kDa was used as a
positive control.

2.8. Calculations and analysis of data

Data were collected in a Microsoft

s

Excel 2000 database and results

were presented as means and standard deviations of at least 3 experiments
using the Origin

s

7.0 software. Significance between the mean values was

calculated using ANOVA one-way analysis (GraphPad InStat 3.06,
GraphPad Software, USA). Probability values p

p0:05 were considered

as significant.

3. Results and discussion

3.1. Microspheres preparation and characterization

The microencapsulation of gentamicin crobefate using

biodegradable PLGA was performed with a spray drying
method as described previously

[24]

. Spray drying gen-

erates spherical particles with high encapsulation efficien-
cies compared to other encapsulation methods, such as

ARTICLE IN PRESS

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4241

solvent evaporation

[28]

. High encapsulation efficiency and

drug loading are requirements for composite formation
with apatitic calcium phosphate cement as described later.
Also particle sizes

o15 mm were easily attained with

narrow size distributions. The admixed MS will reside in
the porous structure of the cement matrix with pore sizes in
the 1–20 mm range, and to avoid destruction they need to
be in the size range of 5–15 mm.

As shown in

Table 1

placebo and gentamicin crobefate

loaded microspheres with a weight average of volume
distribution (D[4.3]) in the range of 5–15 mm were obtained
by spray drying. Monomodal particle size distributions
were observed in all cases (

Fig. 1

). As expected, PLGA

molecular weight affected both yields and particle sizes
drastically. Microsphere yield increased from 6.7% up to
39.7% by increasing the M

w

and also drug loading affected

the yield in a positive manner. The yields of placebo and
gentamicin crobefate loaded MS were in the expected range
for small laboratory spray dryers

[29,30]

. With PLGA in

the M

w

range of 14–35 kDa both placebo and drug loaded

particles yielded sizes in the range of 7.1

70.4 and

14.4

70.1 mm for the unloaded and 9.270.3 and

5.4

70.2 mm for the gentamicin-loaded MS, respectively.

The particle size was affected by the viscosity of the feed
suspension and both GC–MS and placebo MS showed
comparable weighted averages of volume distribution
(D[4.3])

and

monomodal

size

distributions.

Above

M

w

435; 000 Da the feed became too viscous and hence

insufficient microsphere formation was noted.

The encapsulation efficiency of GC–MS prepared by

spray drying was found to be approximately 100% (w/w),
e.g. the effective drug loading from batch GC-4 was
98.6

72.2% and similar to results obtained for ampicillin.

By contrast, microencapsulation of gentamicin sulfate
yielded

much

lower

encapsulation

efficiencies

of

23.6

71.0% using a spray drying method at 10% loading

level

[31]

. In this study gentamicin sulfate was incorporated

into the feed not as solid but as an emulsion. Our results
demonstrate that spray drying is a suitable technique for
the preparation of GC–MS with appropriate size and high
encapsulation efficiency.

To verify the spray-drying results regarding micro-

sphere size and distribution and to characterize particle
morphology, SEM was employed and micrographs are
presented in

Figs. 2 and 3

. Placebo particles produced

with polymers with a lower M

w

showed all a smooth

surface and spherical particles morphology (

Fig. 2A–D

).

MS prepared from PLGA with M

w

435; 000 showed

strong agglomeration and discrete particle formation
could not be observed (

Fig. 2E and F

). This observa-

tion is not unexpected since homogenous spray forma-
tion is strongly influenced by the viscosity of the feed
suspension.

Based on results with different PLGA regarding MS

sizes and morphology, GC–MS were manufactured with
RG 503 H and RG 503, because this particle size range was
considered as useful for the preparation of composites

[32]

.

With both types of PLGA the preparation of drug loaded
MS by spray drying was easily achieved. The SEM
micrographs shown in

Fig. 3

allow an overview of the

morphology of GC-loaded particles as a function of
different drug loadings (10–30% (w/w)). All loaded batches
showed spherical morphology with smooth particle sur-
face, demonstrating that GC was completely entrapped
with the polymer matrix and unencapsulated gentamicin
was not attached to the surface. In comparison, SEM
images from gentamicin sulfate-loaded MS published
recently look similar, but do also show some indentations
on the surface, possibly caused by the gentamicin sulfate

[31]

. The average size of particles was comparable to

D[3,4]. Based on these results GC–MS were regarded
suitable for the preparation of CPC.

To minimize interactions between the hydrophilic

gentamicin sulfate and PLGA, we selected a more

ARTICLE IN PRESS

(A)

1

10

100

0

2

4

6

8

10

12

14

16

18

20

Volume (%)

Particle size [µm]

(B)

Volume (%)

Particle size [µm]

MS-1
MS-2
MS-3
MS-4

1

10

100

0

2

4

6

8

10

12

14

16

18

20

GC-1
GC-2
GC-3
GC-4

Fig. 1. Particle size distribution according to laser light scattering
measurements of spray dried microspheres: (A) Placebo (MS) particles
using matrix polymers with different molecular weights and (B)
gentamicin crobefate (GS) loaded microspheres with different drug
loadings.

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4242

lipophilic salt combination, namely gentamicin crobefate
(

Fig. 4

). Using micro-encapsulation it was anticipated that

GC would not interfere with the setting process of the bone
cement. On the other hand gentamicin sulfate was shown
to interact with PLGA carboxylic end groups

[31]

. The

distribution of gentamicin crobefate in PLGA micro-
spheres did not affect the glass transition temperatures
(T

g

) as shown in

Table 1

suggesting that GC–MS consist of

a solid dispersion morphology. The T

g

values for placebo

MS ranged from 41.5 to 42.0 1C and drug-loaded particles
showed similar values.

3.2. In vitro release properties of gentamicin crobefate
loaded microspheres

Spray dried MS produced from RG 503 H and loaded

with different amounts gentamicin crobefate, were used to
investigate the drug release profiles as function of drug
loading. For local treatment of bone infections, gentamicin
implant and microspheres were studied either using the
sulfate salt form

[33]

or the free base. While MS containing

gentamicin base showed good encapsulation efficiencies,
interactions between drug and PLGA caused stability

ARTICLE IN PRESS

Fig. 2. SEM micrographs of spray dried placebo microspheres (MS) batches made of six different 5% (w/v) polymer solutions: (A) MS-1: RG 502 H, (B)
MS-2: RG 502, (C) MS-3: RG 503 H, (D) MS-4 RG 503, (E) MS-5: RG 504 H, (F) MS-6: RG 504.

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4243

problems. With gentamicin sulfate-loaded particles, drug
stability was not an issue, however, initial drug release
(burst release) resulting from insufficient encapsulation and
high water solubility of gentamicin sulfate were more
problematic. GC, a poorly water-soluble salt of gentami-
cin, is commercially available as combination with
gentamicin sulfate in collagen mats for the prevention of
bone and tissue infections

[22]

. We hypothesized that this

gentamicin salt might posses more controlled release
characteristics than the sulfate, as it was shown in the
literature

[24]

.

The gentamicin crobefate release from PLGA micro-

spheres was determined before and after the addition to the

calcium phosphate bone cement. As shown in

Fig. 5

, the

drug release from PLGA microspheres under in vitro
conditions was characterized by a triphasic drug release
kinetic (

Fig. 5A

) as often described in the literature

[34]

.

The initial burst ranged from 6.5

74.0% up to 26.270.4%

in the first 24 h depending on drug loading. As expected
higher drug loading caused higher drug bursts, but the
values are much lower than those in gentamicin sulfate
microparticles where bursts up to 60% were obtained at
lower loadings

[31]

. The second phase (plateau-phase)

lasted approximately 7–12 days with only low doses
released during this period of time followed by the erosion
phase, in which polymer degradation occurred and drug
release could be observed. The plateau-phase decreased
with increasing drug loading, in accordance with the
generally postulated pore diffusion mechanism.

By contrast, GC release from MS/cement composites

showed nearly a zero-order kinetic, which was character-
ized by a slower, but linear release over 100 days without
initial drug burst (

Fig. 5B

). Burst release was strongly

reduced for all composites to values below 2% in the first
24 h as shown in

Fig. 6

. This can be explained by the lower

water solubility of GC and the slower drug diffusion out of
the cement matrix. The linear release profile can be
attributed to the embedding of the MS into the cement

ARTICLE IN PRESS

Fig. 3. SEM micrographs of spray dried microspheres (MS) batches loaded with different amounts of gentamicin crobefate (GC) and made of RG 503 H
and RG 503: (A) GC-1: 10% (w/w) GC in RG 503 H, (B) GC-2: 20% (w/w) GC in RG 503 H, (C) GC-3: 30% (w/w) GC in RG 503 H and (D) GC-4: 10%
(w/w) GC in RG 503.

H

2

O

3

PO

OMe

OMe

O

O

E

Fig. 4. Chemical structure of crobefate.

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4244

matrix. Another reason might be the adsorption of the
released drug on the cement matrix, which might also
prolong the release kinetics

[35]

. All these factors con-

tributed to the linear kinetic profile compared to the
triphasic release profile from the plain drug loaded
particles. Extremely high release rates of antibiotics were
observed in studies published in the literature, where
55–95% of drug release occurred between 3 and 7 d,
independent of the type of the cement, namely apatite or
brushite

[33]

. To our knowledge, this is the first example of

an antibiotic cement composite demonstration both low
reduced drug burst and linear release kinetics.

The minimal inhibition concentration (MIC) of genta-

micin in vitro could be reached at all time points during the
in vitro release study. All composites released much higher
concentrations of GC than the required concentration to
eradicate bacteria above the MIC of

o1–2 mg/ml (

Table 2

).

These results demonstrate the feasibility a drug delivery
system based on cement/MS composites.

3.3. Mechanical properties of hardened microsphere/cement
composites

Ceramic materials, such as hydroxyapatite have received

increasing interest recently

[8]

. The addition of antibiotics

to calcium phosphate bone cements has been reported
earlier

[13]

. The main drawback of this approach seems to

be the negative effects of antibiotics on mechanical
properties of cement composites, causing a decrease in
compressive strength. We speculated that isolation of
antibiotic by microencapsulation into a biodegradable
polymer should prevent interactions between the cement
and antibiotic during the setting phase and hence lead to
composites with sufficient mechanical strength.

Properties of the composites such as setting time and

compressive strength with and without the addition of
antibiotic-loaded MS are displayed in

Fig. 7A

. The MS free

cement showed an initial setting time of 5 min. Addition of
unencapsulated gentamicin crobefate (1% (w/w)) abolished
cement setting, since GC is a very effective inhibitor of HA
crystal growth. This is in agreement with other gentamicin

ARTICLE IN PRESS

0

1

2

3

4

5

Cumulative release (mg)

GC-1: CPC + 10% (w/w) GC
GC-2: CPC + 10% (w/w) GC
GC-3: CPC + 10% (w/w) GC

0

20

40

60

80

100

0

1

2

3

4

5

Cumulative release (mg)

Time (d)

GC-1: 10 % GC in RG 503 H
GC-2: 20 % GC in RG 503 H
GC-3: 30 % GC in RG 503 H

(A)

0

20

40

60

80

100

Time (d)

(B)

Fig. 5. In vitro release profiles from poly(lactide-co-glycolide) micro-
spheres before (A) and after (B) embedding in cement devices. Three
different drug loadings of gentamicin crobefate (GC) are shown (-’-)GC
10% (w/w), (-K-)GC 20% (w/w), (-m-)GC 30% (w/w) (n ¼ 3).

GC-1 10 %

GC-2 20 %

GC-3 30 %

0

5

10

15

20

25

30

Burst release (%)

Gentamicin release from MS
Gentamicin release from MS after
embedding into cement

Fig. 6. Gentamicin crobefate burst release from biodegradable micro-
spheres with 10%, 20% and 30% drug before and after embedding in
calcium phosphate cement. Values obtained after 24 h of incubation
(n ¼ 3).

Table 2
Gentamicin crobefate (GC) concentrations released from microspheres
after admixing to calcium phosphate bone cement (n ¼ 3)

Batch

GC released
after 8 h

x-fold
higher then
MIC
(1–2 mg/ml)

GC released
after 50 d

x-fold
higher then
MIC
(1–2 mg/ml)

GC-1 (10%)

0.4

70.2

1/5–1/2.5

79.3

710.4

80–40

GC-2 (20%)

3.0

70.2

3–1.5

128.0

73.0

130–65

GC-3 (30%)

7.3

71.2

7–3.5

124.0

711.6

120–60

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4245

salts

[12]

. By admixing gentamicin sulfate in increasing

amounts to an apatitic cement, setting of the cement was
achieved but at increased setting times up to 26 min. In
contrast, no change of setting times was noted for the
addition of placebo MS batches MS-3, MS-4 or drug-

containing batches GC-1 and GC-2. A slightly, but still
acceptable increase of the initial setting time for GC-3 to
8 min was observed. In the literature methods were
described, which use prolonged setting times as a means
to improve fitting of the cement paste to the defect site. The
setting time, however, should not exceed 12–15 min for
reasons of clinical handling

[36]

.

At the same time, the compressive strength of cements

mixed with MS of up to 30% (w/w) GC was not affected
compared to the control group with 52.08

75.1 MPa

(

Fig. 7C

). The compressive strength increased significantly

(p

p0:05 and pp0:005) up to 70 MPa by the addition of

drug-loaded MS to the cement, independent of the batch.
No significant differences between the three GC-loaded
groups could be seen. The compressive strength was
affected by the polymers of the admixed particles with
regard to the M

w

. A significant (p

p0:05) influence on

compressive strength was seen by using different polymers.
With increasing the polymer molecular weight an increase
in compressive strength of the tested cement samples was
observed (

Fig. 7B

). It is known from the literature, that

PLGA with high molecular weights are used for tissue
engineering due to their better stability, also in combina-
tion with CPC

[37]

. The limit of MS addition with respect

to mechanical properties is reached at about 20% (w/w)
drug-loaded particles to the cement which then resulted in
a decrease of compressive strength to 28.13

711.37 MPa

(data not shown) after addition of 30% (w/w) drug-loaded
particles.

After exhaustive setting, the cement/MS composites were

analyzed using an X-ray diffraction technique to study the
conversion to HA. Results of the phase analysis according
to Rietveld refinement analysis of the cement after setting
are given in

Table 3

. All materials converted to nanocrys-

talline hydroxyapatite (76–86% (w/w)) within 24 h with
crystal sizes of approximately 13–17 nm. The latter results
are based on peak broadening in X-ray diffraction analysis.
A minor phase of non-reacted TTCP (10–22%) could be
detected while the DCPA reactant was nearly completely
consumed such that the precipitated HA is thought to be of
calcium deficient composition.

Finally, the cement/MS composites were analyzed using

SEM to investigate the structure and integrity of MS
embedded in the cement composite. Additionally, the
degradation of the biodegradable MS was demonstrated
after incubation under in vitro conditions. Therefore, cross
sections of the composites before and after the in vitro
study were taken. While initially the MS are not damaged
and remain intact (

Fig. 8A

), they were completely degraded

after 100 d of incubation and are not visible at the cross
section site (

Fig. 8B

). This confirms that GC–MS released

their drug load completely.

3.4. Biocompatibility study under in vitro conditions

Different batches of placebo MS (3,4,5,6) and GC

containing MS (GC-4) were subjected to biocompatibility

ARTICLE IN PRESS

GS 3%

GS 2%

GS 1%

GC

CPC

GC-3

GC-2

GC-1

MS-4

MS-3

0

5

10

15

20

25

Setting time (min)

GC-1

GC-2

GC-3

CPC

GC

GS

0

10

20

30

40

50

60

70

80

90

**

*

**

Compressive strength (Mpa)

MS-1

MS-2

MS-3

MS-4

MS-5

0

10

20

30

40

50

60

70

80

90

**

Co

m

p

re

ssive str

eng

th

(MP

a)

(A)

(B)

(C)

Fig. 7. Basic characteristics of cement devices after mixing with placebo
and antibiotic loaded microspheres: (A) setting time, (B) compressive
strength with different polymers used for microsphere batches, (C)
compressive strength regarding different drug loadings. MS ¼ placebo
particles, GC ¼ Gentamicin crobefate-loaded particles, CPC ¼ calcium
phosphate cement, GS ¼ gentamicin sulfate, added only as solution.
(*

p0.05; **p0.005).

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4246

testing under in vitro conditions. The materials were
incubated with cell culture medium according to USP
XXV for 24 h. Serial dilutions of these extracts were
studied in a standard cell line of mouse fibroblasts (L 929)
with positive and negative controls. Cell viability was
tested using a MTT assay after 24 h of incubation as shown
in

Fig. 9

. Even at extract concentrations as high as 10 mg

microspheres/ml incubation medium both placebo and
drug-loaded MS did not show signs of impaired cell
viability. A slightly increase of cell viability was observed
for GC-4 batch due to a growth stimulatory effect of the
added gentamicin. These results are not unexpected since
both PLGA MS and calcium phosphate cement are FDA
approved and GC is used clinically

[38]

.

4. Conclusions

The feasibility of a self-setting bone cement composite

containing biodegradable gentamicin crobefate-loaded
microspheres was demonstrated in this study. Drug-loaded
microspheres were produced by spray drying with a
monomodal distribution and a high encapsulation effi-
ciency of nearly 100%, which reduced initial drug burst.
Gentamicin crobefate release was controlled by a combina-
tion of pore diffusion and polymer erosion, which resulted
in zero-order release kinetics for cement/microsphere
composites. Moreover, no change of the setting properties

ARTICLE IN PRESS

Table 3
Phase composition of cements according to Rietveld refinement analysis

Sample

Microparticle batch

Polymer

Drug loading (%)

HA % (w/w)

TTCP % (w/w)

DCPA % (w/w)

Size HA (nm)

Ref. (no GC)

—

—

—

84.8

13.9

1.3

17.2

70.3

Variation of polymers
CPC+MS-3

MS-3

RG 503 H

—

76.1

22.8

1.1

16.1

70.5

CPC+MS-4

MS-4

RG 503

—

79.0

19.5

1.5

15.3

70.3

Variation of theoretical drug loading
CPC+GC-1

GC-1

RG 503 H

10

85.0

14.7

0.3

14.8

70.3

CPC+GC-2

GC-2

RG 503 H

20

85.9

9.7

4.5

13.6

70.3

CPC+GC-3

GC-3

RG 503 H

30

86.0

9.8

4.2

14.4

70.4

CPC+GC-4

GC-4

RG 503

10

79.0

18.4

2.6

15.4

70.4

By the addition of free gentamicin crobefate to the CPC, no setting occurred, therefore no results in phase composition are shown here.

Fig. 8. SEM micrographs of calcium phosphate cement devices embedded
with spray dried microsphere batch GC-1made of RG 503 H (PLGA)
polymer with 10% (w/w) gentamicin crobefate: (A) before and (B) after
100 d of incubation in PBS, 37 1C, 30 rpm.

0

2

4

6

8

10

0

20

40

60

80

100

120

Cell Viability (%)

Concentration of Microparticles (mg/ml)

MP-3
MP-4
MP-5
MP-6
GC-4

Fig. 9. Cytotoxicity of placebo and gentamicin crobefate containing
microspheres to L 929 fibroblasts as measured by MTT assay. Cells were
incubated with increasing concentrations of particle-extracts for 24 h. No
significantly differences were detectable (n ¼ 8).

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4247

and mechanical performance of the cement was observed
due to an effective decoupling of cement setting and drug
release. Composites made from calcium phosphate cement
and drug-loaded microspheres could be of interest in
prevention and treatment of bone infections. A significant
advantage of this drug delivery system providing local
antibiotic release would be its biodegradability avoiding
secondary surgery.

Acknowledgments

The authors wish to thank Biomet Deutschland GmbH,

Berlin, for providing the gentamicin crobefate. The authors
would also like to thank the European Society of
Biomaterials for the Student poster award 2003 and the
related grant cordially.

References

[1] Lew DP, Waldvogel FA. Osteomyelitis. Lancet 2004;364:369–79.
[2] Joosten U, Joist A, Frebel T, Brandt B, Diederichs S, von Eiff C.

Evaluation of an in situ setting injectable calcium phosphate as a new
carrier material for gentamicin in the treatment of chronic
osteomyelitis:

studies

in

vitro

and

in

vivo.

Biomaterials

2004;25:4287–95.

[3] Wahlig H, Dingeldein E, Bergmann R, Reuss K. Experimental and

pharmacokinetic studies with gentamicin PMMA beads (author’s
transl). Zentralbl Chir 1979;104:923–33.

[4] Neut D, van de Belt H, van Horn JR, van der Mei HC, Busscher HJ.

Residual gentamicin-release from antibiotic-loaded polymethyl-
methacrylate beads after 5 years of implantation. Biomaterials
2003;24:1829–31.

[5] Santin M, Motta A, Borzachiello A, Nicolais L, Ambrosio L. Effect

of PMMA cement radical polymerisation on the inflammatory
response. J Mater Sci Mater Med 2004;15:1175–80.

[6] Nelson CL. The current status of material used for depot delivery of

drugs. Clin Orthop 2004;1:72–8.

[7] Blom EJ, Klein-Nulend J, Wolke JG, Waas van MAJ, Driessens

FCM, Burger EH. Transforming growth factor-b1 incorporation in a
calcium phosphate bone cement: material properties and release
characteristics. J Biomed Mater Res 2001;59:265–72.

[8] Bohner M, Gbureck U, Barralet JE. Technological issues for the

development of more efficient calcium phosphate bone cements: a
critical assessment. Biomaterials 2005;26:6423–9.

[9] Gbureck U, Barralet JE, Spatz K, Grover LM, Thull R. Ionic

modification of calcium phosphate cement viscosity. Part I:
hypodermic injection and strength improvement of apatite cement.
Biomaterials 2004;25:2187–95.

[10] Takechi M, Miyamoto Y, Ishikawa K, Nagayama M, Kon M,

Asaoka K, et al. Effects of added antibiotics on the basic properties
of

anti-washout-type

fast-setting

calcium

phosphate

cement.

J Biomed Mater Res 1998;39:308–16.

[11] Armstrong MS, Spencer RF, Cunningham JL, Gheduzzi S, Miles

AW, Learmonth ID. Mechanical characteristics of antibiotic-laden
bone cement. Acta Orthop Scand 2002;73:688–90.

[12] Gbureck U, Probst J, Thull R. Surface properties of calcium

phosphate particles for self setting bone cements. Biomol Eng
2002;19:51–5.

[13] Ratier A, Gibson IR, Best SM, Freche M, Lacout JL, Rodriguez F.

Setting characteristics and mechanical behavior of a calcium
phosphate

bone

cement

containing

tetracycline.

Biomaterials

2001;22:897–901.

[14] Lim ST, Forbes B, Berry DJ, Martin GP, Brown MB. In vivo

evaluation of novel hyaluronan/chitosan microparticulate delivery

systems for the nasal delivery of gentamicin in rabbits. Int J Pharm
2002;231:73–82.

[15] Sivakumar M, Panduranga Rao K. Preparation, characteriza-

tion and in vitro release of gentamicin from coralline hydroxy-
apatite-gelatin

composite

microspheres.

Biomaterials

2002;23:

3175–81.

[16] Prior S, Gander B, Lecaroz C, Irache JM, Gamazo C. Gentamicin-

loaded microspheres for reducing the intracellular Brucella abortus
load in infected monocytes. J Antimicrob Chemother 2004;53:
981–8.

[17] Bittner B, Maeder K, Kroll C, Borchert HH, Kissel T. Tetracycli-

ne–HCl-loaded poly(

DL

-lactide-co-glycolide) microspheres prepared

by a spray drying technique: influence of gamma-irradiation on
radical formation and polymer degradation. J Control Release
1999;59:23–32.

[18] O’Hara P, Hickey AJ. Respirable PLGA microspheres containing

rifampicin for the treatment of tuberculosis: manufacture and
characterization. Pharm Res 2000;17:955–61.

[19] Gavini E, Chetoni P, Cossu M, Alvarez MG, Saettone MF,

Giunchedi P. PLGA microspheres for the ocular delivery of a
peptide drug, vancomycin using emulsification/spray-drying as the
preparation method: in vitro/in vivo studies. Eur J Pharm Biopharm
2004;57:207–12.

[20] Vert M, Li S, Garreau H. More about the degradation of LA/GA-

derived matrices in aqueous media. J Control Release 1991;16:
15–26.

[21] Neut D, van de Belt H, Stokroos I, van Horn JR, van der Mei HC,

Busscher HJ. Biomaterial-associated infection of gentamicin-loaded
PMMA beads in orthopedic revision surgery. J Antimicrob Che-
mother 2001;47:885–91.

[22] Cornell CN, Tyndall D, Waller S, Lane JM, Brause BD. Treatment

of experimental osteomyelitis with antibiotic-impregnated bone graft
substitute. J Orthop Res 1993;11:619–26.

[23] Mosmann T. Rapid colorimetric assay for cellular growth and

survival: application to proliferation and cytotoxicity assays.
J Immunol Methods 1983;65:55–63.

[24] Weber T. Herstellung und Charakterisierung von bioabbaubaren

Mikropartikeln zur lokalen Antibiotika-Therapie. Department of
Pharmaceutics and Biopharmacy. Marburg, Philipps-University;
1991.

[25] Chow LC, Hirayama S, Takagi S, Parry E. Diametral tensile strength

and compressive strength of a calcium phosphate cement: effect of
applied pressure. J Biomed Mater Res 2000;53:511–7.

[26] C266-99 A- S. Standard test method for time of setting of hydraulic

cement paste by Gilmore needles; 2002.

[27] Fischer D, Li Y, Ahlemeyer B, Krieglstein J, Kissel T. In

vitro cytotoxicity testing of polycations: influence of polymer
structure on cell viability and hemolysis. Biomaterials 2003;24:
1121–31.

[28] Blanco-Prieto M, Lecaroz C, Renedo M, Kunkova J, Gamazo C.

In vitro evaluation of gentamicin released from microparticles. Int
J Pharm 2002;242:203–6.

[29] Maury M, Murphy K, Kumar S, Shi L, Lee G. Effects of process

variables on the powder yield of spray-dried trehalose on a laboratory
spray-dryer. Eur J Pharm Biopharm 2005;59:565–73.

[30] Masters K. Spray drying handbook, 5th ed. New York; 1994.
[31] Prior S, Gamazo C, Irache JM, Merkle HP, Gander B. Gentamicin

encapsulation in PLA/PLGA microspheres in view of treating
Brucella infections. Int J Pharm 2000;196:115–25.

[32] Giunchedi P, Conte U. Spray-drying as a preparation method of

microparticulate drug delivery systems: an overview. STP Pharma Sci
1995;5:276–90.

[33] Baro M, Sanchez E, Delgado A, Perera A, Evora C. In vitro-in vivo

characterization of gentamicin bone implants. J Control Release
2002;83:353.

[34] Siepmann J, Faisant N, Benoit JP. A new mathematical model

quantifying drug release from bioerodible microparticles using Monte
Carlo simulations. Pharm Res 2002;19:1885–93.

ARTICLE IN PRESS

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4248

[35] Queiroz AC, Santos JD, Monteiro FJ. Development of a system to

adsorb drugs onto calcium phosphate materials. J Mater Sci Mater
Med 2005;16:641–6.

[36] Barralet JE, Grover LM, Gbureck U. Ionic modification of calcium

phosphate cement viscosity. Part II: hypodermic injection and
strength improvement of brushite cement. Biomaterials 2004;25:
2197–203.

[37] Jain RA. The manufacturing techniques of various drug loaded

biodegradable poly(lactide-co-glycolide) (PLGA) devices. Biomater-
ials 2000;21:2475–90.

[38] Gruessner U, Clemens M, Pahlplatz PV, Sperling P, Witte J, Rosen

HR. Improvement of perineal wound healing by local administration
of gentamicin-impregnated collagen fleeces after abdominoperineal
excision of rectal cancer. Am J Surg 2001;182:502–9.

ARTICLE IN PRESS

J. Schnieders et al. / Biomaterials 27 (2006) 4239–4249

4249