Evaluating Release Kinetics from Alginate Beads Coated with
Polyelectrolyte Layers for Sustained Drug Delivery
Markus Witzler, Sarah Vermeeren, Roman O. Kolevatov, Razan Haddad, Martin Gericke,
Thomas Heinze, and Margit Schulze*
Cite This: https://doi.org/10.1021/acsabm.1c00417 Read Online
ACCESS Metrics & More Article Recommendations *sı Supporting Information
ABSTRACT: Current approaches in stem cell-based bone tissue
engineering require a release of bioactive compounds over up to 2
weeks. This study presents a polyelectrolyte-layered system
featuring sustained release of water-soluble drugs with decreased
burst release. The bioactive compounds adenosine 5′-triphosphate
(ATP), suramin, and A740003 (a less water-soluble purinergic
receptor ligand) were incorporated into alginate hydrogel beads
subsequently layered with different polyelectrolytes (chitosan,
poly(allyl amine), alginate, or lignosulfonate). Drug release into
aqueous medium was monitored over 14 days and evaluated using
Korsmeyer-Peppas, Peppas-Sahlin, Weibull models, and a
Langmuir-like “Two-Stage” model. Release kinetics strongly depended on both the drug and the polyelectrolyte system. For
ATP, five alternating layers of poly(allyl amine) and alginate proved to be most effective in sustaining the release. Release of suramin
could be prolonged best with lignosulfonate as polyanion. A740003 showed prolonged release even without layering. Applying
polyelectrolyte layers significantly slowed down the burst release. Release curves could be best described with the Langmuir-like
model.
KEYWORDS: drug release, biopolymers, polyelectrolytes, layer-by-layer encapsulation, release kinetics, lignosulfonate
1. INTRODUCTION
Current research in bone tissue engineering focuses not only
on new scaffold materials that help bone formation but also on
using drugs, growth factors, or receptor ligands to prevent
infections and stimulate bone cells.1−3 Purinergic P2X and P2Y
receptors can play a vital role in differentiation of human adult
mesenchymal stem cell (hMSC). For example, extracellular
ATP, one of the most prominent purinergic receptor ligands, is
known to guide stem cell differentiation including embryonic,
hematopoietic, and neuronal stem cells.4 Very recently, P2
receptors were reported to trigger the hMSC differentiation
toward endothelial and smooth muscle cell lineages.5 Addi￾tionally, osteogenic differentiation of hMSC can be triggered
and guided by specifically targeting P2X and P2Y receptors
with their respective agonists and antagonists.6−8 The P2X7
receptor is one of these crucial bone regulating receptors
mediating both osteoblast and osteoclast activity. An effective
way to up- or downregulate this receptor is the use of
A740003, a highly selective, competitive receptor antago￾nist.9,10 Sustaining and controlling the release of active
ingredients like drugs, growth factors, or receptor ligands is
one of the major issues in current drug release research.
Depending on the drug, the release system, the site of
application, and the target region, there are countless ways to
prolong a drug release.1,11,12 One example is the encapsulation
of the active compounds within hydrogel beads, ranging from
nanometer to millimeter scale.13,14 These particles are usually
made of a polymeric matrix, and the drug is released via
diffusion or matrix degradation. Both synthetic polymers (e.g.,
poly(acrylic acid), poly(ϵ-caprolactam)) and biopolymers (e.g.,
poly(lactic acid), alginate, chitosan) are frequently used in this
context.2,15−17 Parameters influencing the release rate are,
among others, hydrophobicity of drug and matrix and matrix
swellability.18,19 A widely employed biopolymer is sodium
alginate, a marine polysaccharide obtained from brown algae,
which consists of blocks of α-L-guluronic acid and β-D￾mannuronic acid.20 It is biocompatible, biodegradable, water￾soluble, and has the ability to form stable hydrogels upon
contact with divalent cations such as Ca2+ or Sr2+. Alginate is
used for many medical purposes, ranging from scaffold material
in tissue engineering to wound dressing and drug delivery
devices.16 Another marine polysaccharide is chitosan, the
deacetylated form of chitin, which is a glucosamine built up
Received: April 8, 2021
Accepted: August 18, 2021
www.acsabm.org Article
© XXXX American Chemical Society A

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
Downloaded via HELMHOLTZ-ZENTRUM HEREON on August 31, 2021 at 20:17:59 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
from β-(1 → 4)-linked D-glucosamine and N-acetyl-D-glucos￾amine units. Chitosan is insoluble in water above pH = 6,
however, in slightly acidic conditions its amino groups are
protonated, yielding a water-soluble cationic biopolymer that is
often used in drug delivery systems due to its abundance,
biocompatibility, and biodegradability.16,21 Another biopoly￾mer that started to gain attention for medical purposes or
release applications is lignin and its water-soluble derivative
lignosulfonate.22−24 Lignin is one of the most abundant
biopolymers and currently several research groups are
investigating its structure as well as its valorization beyond
energy production.25,26 Recent publications on lignosulfonate
suggest that it is noncytotoxic and generally applicable for
biomedical and drug release purposes.27−29
One major challenge in drug release research is the sustained
release of hydrophilic drugs into aqueous media over the
course of more than a few days.30,31 Layer-by-layer assembling
of polyelectrolytes is a promising approach for ionic drug
molecules.32,33 Alternating layers of anionic and cationic
polymers are deposited on a template, which can be a drug￾loaded capsule or bead, or even the solid drug itself.34−37 Due
to electrostatic barriers of the layers and other molecular
interactions, the active ingredient is then released more slowly.
This technique is already being used in various applications,
either to actively prolong the release or benefit from additional
properties of the layers such as antibacterial properties or
better cell-adhesion.38−40
The objective of this ongoing project is the development of
novel materials for a sustained 7- to 10-day release of drugs
that are suitable for guided osteogenic stem cell differentiation.
Three different drugs were selected to study their respective
release kinetics (see structural formulas in Figure 1).
Adenosine 5′-triphosphate (ATP), although being an active
nonselective P2 receptor ligand, was chosen as a model drug
only because it is widely available, easy to detect, and is a
structural key component of many other P2 receptor ligands.41
Suramin is an active nonselective P2 receptor ligand, which is
both larger than ATP and carries different chemical moieties.
A740003 is an active P2X7 antagonist that is an uncharged and
less water-soluble molecule.9 Our previous research on the
release from hydroxyapatite/agarose hybrid scaffold showed a
release of ATP and suramin of around 3−4 days.42
Here, we present the encapsulation of the three afore￾mentioned drugs in alginate beads and subsequent layering
with polyelectrolytes. This has the potential to overcome the
high diffusional release of the hydrophilic drugs, thereby
altering the kinetic release profile from a high, almost exclusive
burst release behavior, toward a Two-Stage profile with a
prolonged release after a burst phase. The kinetic evaluation of
this two-phase profile using different mathematical models is
another central part of this research.
2. EXPERIMENTAL SECTION
2.1. Chemicals. Sodium alginate (MW ≈ 170 000 g/mol) and
adenosine 5′-triphosphate disodium salt (≥98%, ATP) have been
purchased from Carl Roth, Germany. Acetic acid (p.a.), calcium
chloride dihydrate (p.a.), dimethyl sulfoxide (for synthesis), and
poly(allyl amine) hydrochloride (99%) have been obtained from
VWR Chemicals, Germany. Chitosan (85% deacetylated) came from
Heppe Medical Chitosan, Germany, magnesium lignosulfonate (MW
≈ 15 000 g/mol) from Chemische Werke Zell-Wildshausen,
Germany, suramin hexasodium salt (99.6%) from Merck, Germany,
and A740003, > 96%) was purchased from Alomone Labs, Jerusalem,
Israel. All chemicals were used without further purification. Ultrapure
water (Millipore, 18.2 MΩ) was used throughout the study.
2.2. Bead Preparation. Beads were prepared using a laboratory￾scale coaxial spraying prototype (Figure 2). Sodium alginate was
dissolved in water (1%w/w, viscosity η = 0.4 Pa s) and the solution
was loaded with the respective amount of drug. A perfusor syringe (50
mL) was mounted into a syringe pump operating at a flow rate of 8
mL min−1 transferring 10−30 mL of the solution (depending on drug
and the desired amount of beads) through a blunt 23G cannula.
Pressurized air (volumetric flow rate of 4 standard liter per minute)
was used to create fine droplets. Upon contact with a curing solution
containing calcium chloride (1 mol L−1
) and the respective drug
(same concentration as in the alginate solution), the alginate cross￾linked and formed spherical hydrogel beads. The volume of the curing
solution was five times the volume of alginate solution. The distance
between tip and curing solution was 50 mm, the curing took place at
2−8 °C for 4 h.
2.3. Polyelectrolyte Layering. Polyelectrolyte layering was
performed using an immersive approach that is illustrated in Figure
3. In order to create one set of polyelectrolyte layers, the beads were
separated from the curing solution, briefly rinsed with water and
transferred to a 5 mL vial, coated with 1 mL of polycationic solution
Figure 1. (A) Structural formula of ATP (adenosine 5′-triphosphate),
M = 507.18 g mol−1
. Used as ATP disodium salt, M = 551.14 g mol−1
(B) Structural formula of A740003 (N-[1-[[(Cyanoamino)(5-
quinolinylamino)methylene] amino]-2,2-dimethylpropyl]-3,4-dime￾thoxybenzeneacetamide), M = 474.55 g mol−1
. (C) Structural formula
of suramin (8,8′-{Carbonylbis[imino-3,1-phenylenecarbonyl-imino(4-
methyl-3,1-phenylene)carbonylimino]}di(1,3,5-naphthalenetrisul￾fonic acid)), M = 1297.29 g mol−1
. Used as suramin hexasodium salt,
M = 1429.17 g mol−1
Figure 2. Setup of bead preparation via laboratory-scale coaxial
spraying. Sodium alginate solution (loaded with or without drug) is
fed by a syringe pump through a cannula and pressurized air is used
for fine droplet creation. Parameters were chosen to obtain
homogeneous particle sizes between 600 and 1200 μm.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
containing either chitosan (2%w/w, pH 5.1) or poly(allyl amine) (1%
w/w, pH 5.1) and vortexed for 1 min. After that, the beads were
separated by filtration, briefly rinsed with water (pH 6.0), and coated
with 1 mL polyanionic solution containing lignosulfonate or alginate
(2%w/w, pH 5.2 and 5.9, respectively). After vortexing for 1 min, the
beads were separated and washed again with water. The steps were
repeated to get the desired number of coating layers.
The specimens were labeled as follows: First, the core bead alginate
(Alg), then, in case of layered systems the number of coating layers
followed by the polycation chitosan (C) or poly(allyl amine) (P) and
the polyanion lignosulfonate (L) or alginate (A). As an example,
Alg5PL describes an alginate beads layered with 5 layers of poly(allyl
amine) and lignosulfonate.
2.4. Bead Characterization. Beads were characterized with
respect to size and surface charge (ζ-potential). For size
determination, static light scattering (SLS; Bluewave, Microtrac
Europe, Germany, Microtrac FLEX v.11.0.0.4 software) was used.
For the SLS experiments, approximately 200 mg of beads were used
and sonicated for 1 min immediately before the measurement. Digital
microscope imaging for morphology evaluation was performed on a
digital microscope (VHX-7000, Keyence, Japan) without any further
sample preparation. For surface charge measurements, a Stabino
Charge Titration System (colloid-metrix, Germany, with Stabino
Control v.2.03.03 software) was used on a small part (100−200 mg)
of the samples dispersed in 5 mL of water after each addition of either
the polycation or anion. The surface charge was determined by
measuring the streaming potential in water at pH 5.5 caused by a
moving piston periodically shearing the ionic double layer.
2.5. Drug Loading. Beads were loaded with either ATP or
suramin by adding the drug to the alginate solution and curing
solution before coaxial spraying. Final drug concentrations were 100
μmol L−1 and 50 μmol L−1
, respectively. A740003 was dissolved in 3
mL of DMSO and homogeneously mixed with the alginate or curing
solution to obtain a final concentration of 50 μmol L−1
determination of the drug loading efficiency, an accurately weighed
amount (approximately 500 mg) of rinsed beads with removed excess
water was added to 10 mL of 50 mM sodium carbonate solution and
stirred for 24 h at room temperature to reverse alginate gelation. The
solution was then centrifuged for 5 min at 4000 rpm and 1 mL of the
supernatant was used for drug concentration measurements with
UV−vis absorbance (UV-1650PC, Shimadzu, Germany) at wave￾lengths λ = 255, 313, and 282 nm for ATP, suramin, and A740003,
respectively. Drug loading efficiency was calculated according to eq 1:
n drug loaded (%) 100% found in bead
theoretical (1)
2.6. Drug Release. For drug release measurements, a known
amount of beads (around 200 to 300 mg) was placed into 20 mL of
water. At given time intervals, 1 mL of medium was collected and
replaced with 1 mL of water maintaining a constant release volume.
Drug concentrations in the sample were measured via UV−vis
absorbance (UV-1650PC, Shimadzu, Duisburg, Germany) at wave￾lengths λ = 255, 313, and 282 nm for ATP, suramin, and A740003,
respectively. The release was monitored for up to 14 days, taking the
restocking of medium during sampling into account for calculations.
All release measurements were performed in triplicates. For one set of
Alg and Alg5PL, Earle’s Balanced Salt Solution (no glucose) instead
of water was used as release medium.
2.7. Kinetic Release Evaluation. In order to evaluate release
kinetics, all curves were fitted to different kinetic models using
nonlinear curve fitting in OriginPro 2016, OriginLabs, U.S.A. (built-in
Levenberg−Marquardt algorithm, least-squares, max. 500 iterations,
tolerance of 10−6
, no weighting). A common model used in drug
Therein, M is the amount of drug released at time t, M∞ is the
maximum amount of drug released, M0 is the initial amount of drug,
ideally being zero, and k is the release rate. Two other models relying
on power law were employed to fit the data; the Korsmeyer-Peppas
model (eq 3) and the Peppas-Sahlin model (eq 4).43,44
Here, K describes the release velocity constant and the exponent n
is related to the release mechanism. The Peppas-Sahlin model
expands the Korsmeyer-Peppas model with a term that accounts for
relaxational mechanisms:44 here K1 and K2 are velocity constants for
Fickian diffusional contribution and relaxational contribution,
respectively. Contrary to the Korsmeyer-Peppas model, the exponent
m is a fixed value and describes the Fickian diffusional release
exponent of the Korsmeyer-Peppas model with its optimum value
depending on the geometry. For spheres, the exponent equals m =
0.43.
The Weibull model (eq 5) is used for describing the release curve
mathematically, without correlating parameters that describe the type
of transport:45
= ×[ − ] = ∞ − − M M e qa 1 and ( )/ tT a b b (5)
Therein, M∞ again describes the maximum amount of drug
released, T is a time-scaling parameter describing the latency of the
release system The other parameters contain information about
scaling (a) and shape of the curve (b), which can be condensed to a
curve factor q, allowing better comparison of the fitting. In addition to
the above-mentioned drug release models, a new model termed the
“Two-Stage model” (eq 6) was employed to fit the data for a Two￾Stage drug release:
The parameter B corresponds to the amount of drug released
during burst release, S to the amount of drug released after burst
release. Together, B and S describe the total amount of drug released
Figure 3. Illustration of layering alginate beads with a polyelectrolytic coating layer. Step 1: Cover alginate beads with polycationic solution. Step
Vortex for 1 min, then separate liquid phase and briefly rinse with water. Step 3: Cover beads with polyanionic solution. Step 4: Vortex for 1 min,
then separate liquid phase and briefly rinse with water. For subsequent coating layers, steps 1 to 4 were repeated.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
at an infinite time point: M∞ = B + S. The release constants KB and KS
are describing how fast the individual process occurs. In detail, they
are the times until half-maximum of the respective process: the lower
the value, the faster the corresponding process. The fitting constraints
are that all parameters (B, S, KB, and KS) cannot be negative.
For goodness of fit evaluation, AICc is being used (eq 7). AICc is
the AIC corrected for small sample sizes (N/P < 40).
Therein, N is the number of data points, and P is the number of
parameters of the fitting model. The models are evaluated by their
RSS and get a penalty term depending on their complexity.
3. RESULTS AND DISCUSSION
Alginate beads were produced by gelation of a drug-containing
solution of sodium alginate with calcium ions (Section 4.1).
Subsequently, the beads formed were coated with different
polyelectrolytes using an immersion technique. The layered
beads consist of the alginate core and five coatings, each
consisting of one polycation layer and one polyanion layer
applied to the beads (Section 4.2). The naming convention of
the samples is described in the Experimental Section. Pristine
alginate beads were optically transparent and colorless. Beads
layered with poly(allyl amine) and alginate remained trans￾parent. Beads layered with lignosulfonate became slightly
opaque and light brown in color as expected for a lignin based
material (Figure 4E−H).
3.1. Bead Size. Bead size distributions were measured by
static light scattering (SLS, see Figure 4A-D). The sizes of
pristine alginate beads were in the range of around 700−1100
μm, peaking around 900 μm. Layering did not change bead
size, with all size distributions peaking between 860 and 905
μm. Digital micrographs of the beads confirmed the size
measurements by SLS.
3.2. Surface Charge. Surface charge measurements after
each layering step showed alternatingly positive and negative
values confirming that the respective polyelectrolytic layer were
applied to the bead (Figure 5). The pristine alginate beads
showed a ζ-potential around −10 mV. Subsequent modifica￾tion with chitosan layers induced a change to a positive ζ-
potential around +10 and +20 mV, while additional alginate
layers had a reverse effect (ζ-potential around −10 mV). Beads
with an outer lignosulfonate layers showed a stronger anionic
ζ-potential between −15 and −25 mV and beads with a
cationic poly(allyl amine) layer had a ζ-potential between +20
and +30 mV.
3.3. Drug Loading. Drug loading efficiency differed for the
various drugs, the results are shown in Figure 6. The
hydrophilic drugs ATP and suramin could be encapsulated
with amounts varying between 60−70%, whereas the less
hydrophilic P2X7 receptor ligand A740003 was encapsulated
up to about 90%. These values were used as a reference for the
maximum amount of drug in the beads in drug release
measurements. The difference in drug loading efficiency can be
directly related to A740003 being less water-soluble, and thus
being less affected by washing steps.
3.4. Drug Release. The raw release curves (cumulative
fraction of drug released vs time) are shown in Figure 7. Both
hydrophilic drugs ATP (A) and suramin (B) exhibited a strong
burst release within the first hours, most likely partly due to
fast diffusion through large open pores and drug molecules
being present in the curing solutions and hence being near the
Figure 4. Bead size distributions (A−D) and digital optical micrograph images (E−H) of alginate beads (A,E) and beads layered with five coating
layers of chitosan/lignosulfonate (B,F), poly(allyl amine)/ lignosulfonate (C,G), and poly(allyl amine)/ alginate (D,H). Scale bar represents 250
μm.
Figure 5. Results of the surface charge measurements (ζ-potential) of
layered alginate bead systems, tested after application of different
numbers of polyelectrolytic layers. Odd numbers indicate cationic
layer, even numbers indicate anionic layer. Tested systems: Chitosan/
lignosulfonate (CL), poly(allyl amine)/lignosulfonate (PL), and
poly(allyl amine)/alginate (PA). Mean ± SD, n = 5.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
surface.46 However, the release in a later stage could be
sustained depending on the layering material. For ATP,
pristine alginate beads (Alg) did not sustain the release,
showing high burst release of around 80% in the first day, with
the rest completely being released after 1 week. Layering the
beads decreased both burst release to around 50% and
subsequent release of another 35% in the following 13 days for
the poly(allyl amine)/alginate (Alg5PA) system. For suramin
(B), pristine alginate beads did not sustain the release
significantly, although in general they showed less burst release
than in case of ATP. Layering with chitosan/lignosulfonate
(Alg5CL) resulted in a strongly decreased burst release
(around 30%) and a slowed release after that. In case of
poly(allyl amine)/lignosulfonate (Alg5PL), after 14 days only
50% of the total amount of drug had been released from the
system. The release curves of the less hydrophilic A740003
(C) showed only a weak burst release and then a very slow
ongoing release was observed. Pristine alginate beads showed a
burst release of only around 20% in the first day, followed by a
steady release over the following 2 weeks. Layering the beads
with chitosan or poly(allyl amine) and lignosulfonate or
alginate, decreased the burst release to below 10% and slowed
down the subsequent release of A740003 even more. Again,
the burst release is most likely due to the drug being present in
the outer layer as well.
Release kinetics evaluation is crucial for understanding the
delivery processes and for tailoring the release to the desired
outcome. Over the years, several mathematical models for drug
release have been established: from power-law-based models
like Korsmeyer-Peppas43 or Peppas-Sahlin44 (eqs 3 and 4), a
first-order release model, or Higuchi’s equation47 toward more
complex mathematical fittings such as the Weibull model.45
Another approach for a Two-Stage release process, which has
not yet been described in the literature for drug release, is
variation of the Langmuir model (eq 6). Due to the model’s
inherent reflection of two stages, it can be used to
mathematically describe release curves that exhibit an initial
burst release phase followed by a slower or sustained release
after that. Each stage’s half-life time KB or KS can be used to
assess the transport mechanism by comparing it to literature
values. The model takes the complete release curve into
account, instead of being limited to only part of the overall
release or somehow averaging it. However, no model can be
applied universally; they all have their limitations, prereq￾uisites, or drawbacks.45 For example, both Peppas models are
valid for only up to 60% of the respective release, but they yield
important information on the release mechanism. For spheres,
a purely Fickian diffusion results in an exponent of n = 0.43,
while non-Fickian processes results in other values (anomalous
transport: 0.43 < n < 0.85; Case I: n = 0.85; Super Case II: n >
0.85). In the Peppas-Sahlin model, the exponent is fixed to the
value of Fickian diffusion. Other models like Weibull and Two￾Stage are more useful in describing the release curve
mathematically and providing other valuable information
such as the maximum amount of drug and time-related
coefficients, which can be used to compare data sets.45,48,49
The following mathematical models have been fitted to the
Figure 6. Drug loading values for adenosine 5′-triphosphate (ATP),
suramin (SUR), and A740003 (A74) in alginate and layered alginate
beads: Five alternating coating layers of chitosan and lignosulfonate
(CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine)
and alginate (PA). Mean ± SD, n = 3.
Figure 7. Release curves of adenosine 5′-triphosphate (ATP; A), suramin (B), and A740003 (C) from different alginate (Alg) bead systems: Five
alternating coating layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine) and alginate (PA).
Release over 14 days. (Mean ± SD, n = 3).
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
experimental data: the Korsmeyer-Peppas and the Peppas￾Sahlin models for kinetic evaluation, and the Weibull and Two￾Stage models for mathematical curve fitting. In order to
evaluate the goodness-of-fit and compare the different models,
residual analysis and Akaike’s Information Criterion Test
(AIC) have been used, since common methods like R2 or
reduced χ2 are not suitable for nonlinear curve fitting.50 In
general, a lower Residual Sum of Squares (RSS) suggests a
better fit of the model to the data, but does not include the
simplicity of a model. The AIC, or the corrected version for
small sample sizes (AICc, eq 7), may help to identify the most
parsimonious model among those tested on the experimental
data.45,48 Since a minimum value of AICc is desired, a simpler
model with fewer parameters is preferred. Models are being
evaluated by their RSS and get a penalty term for complexity.
When comparing different models, the lowest value for AICc
suggests a well-fitting yet simple enough model. In most cases,
the Two-Stage model resulted in the best fit, with lowest RSS
and AICc. Values for the amounts of burst and sustained
release could be properly determined as well. The first-order
model (eq 2) consistently underestimated the maximum
amount released and did not represent the overall shape of
Figure 8. Release kinetics fitted to experimental data of adenosine 5′-triphosphate (ATP) release from alginate (Alg) and layered alginate beads
according to different kinetic models: Korsmeyer-Peppas (A), Weibull (B), and Two-Stage (C). A is fitted to 60−70% of the respective release, due
to model constraints. Layering with five alternating coating layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL),
or poly(allyl amine) and alginate (PA). Error bars have been omitted for clarity (see Figure 7).
Table 1. Kinetic Models Fitted to Experimental Data of Adenosine 5′-Triphosphate (ATP) Release from Different Alginate
(Alg) Bead Systemsa
model
ATP material Korsmeyer-Peppas Peppas-Sahlin Weibull Two-Stage
Alg K = 0.66 ± 0.01 h−n K1 = 1.85 ± 0.06 h−0.43 M∞ = 1.00 ± 0.08 B = 0.68 ± 0.02
n = 0.09 ± 0.02 K2 = −1.31 ± 0.10 h−0.86 T = 0.00 ± 0.10 h S = 0.30 ± 0.02
q = 0.6 ± 0.4 h KB = 0.031 ± 0.008 h
KS = 18 ± 8 h
RSS = 5.02 × 10−3 RSS = 3.62 × 10−4 RSS = 1.56 × 10−2 RSS = 9.00 × 10−3
AICc = −36.68 AICc = −17.67 AICc = −94.91 AICc = −103.73
Alg5CL K = 0.75 ± 0.02 h−n K1 = 1.43 ± 0.04 h−0.43 M∞ = 0.94 ± 0.10 B = 0.78 ± 0.06
n = 0.21 ± 0.02 K2 = −0.71 ± 0.04 h−0.86 T = 0.080 ± 0.007 h S = 0.16 ± 0.05
q = 0.1 ± 0.1 h KB = 0.08 ± 0.02 h
KS = 71 ± 16 h
RSS = 9.70 × 10−4 RSS = 4.83 × 10−4 RSS = 3.20 × 10−2 RSS = 2.37 × 10−2
AICc = −12.74 AICc = −16.22 AICc = −83.46 AICc = −88.23
Alg5PL K = 0.42 ± 0.01 h−n K1 = 0.59 ± 0.03 h−0.43 M∞ = 1.0 ± 0.5 B = 0.51 ± 0.04
n = 0.18 ± 0.02 K2 = −0.15 ± 0.02 h−0.86 T = 0.1 ± 0.3 h S = 0.36 ± 0.04
q = 16 ± 7 h KB= 0.11 ± 0.05 h
KS = 104 ± 14 h
RSS = 1.12 × 10−2 RSS = 6.33 × 10−3 RSS = 5.22 × 10−2 RSS = 1.87 × 10−2
AICc = −57.93 AICc = −54.54 AICc = −75.59 AICc = −92.03
Alg5PA K = 0.27 ± 0.02 h−n K1 = 0.305 ± 0.006 h−0.43 M∞ = 1.0 ± 0.5 B = 0.50 ± 0.03
n = 0.18 ± 0.02 K2 = −0.043 ± 0.002 h−0.86 T = 0.2 ± 0.1 h S = 1.5 ± 1.3
q = 87 ± 32 h KB = 0.8 ± 0.2 h
KS = 1348 ± 286 h
RSS = 2.90 × 10−2 RSS = 1.08 × 10−3 RSS = 2.86 × 10−2 RSS = 1.49 × 10−2
AICc = −63.32 AICc = −81.33 AICc = −93.16 AICc = −95.63
Five alternating coating layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine) and alginate
(PA). Mean ± SD, n = 3.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
the experimental data (see Figure SI) or did not completely fit
to the data, which is also reflected in the rather high RSS (see
Table SI). Hence, the first-order model was not considered for
further investigation. However, it is available in the Supporting
Information (SI) for the sake of completeness.
3.5. Evaluation of ATP Release. Curve fittings for ATP
are displayed in Figure 8A−C and the respective values can be
found in Table 1. The ATP release curves demonstrated a
strong burst release for the different bead materials within the
first 6 h, especially for the pristine alginate (Alg) and the
chitosan/lignosulfonate (Alg5CL) samples. Here, more than
75% of ATP was released during the initial burst phase. The
poly(allyl amine)/lignosulfonate (Alg5PL) and poly(allyl
amine)/alginate (Alg5PA) samples still exhibited burst release,
although it was less pronounced. Applying the mathematical
models results in better fits (lower AICc) for both Weibull and
Two-Stage models. The Weibull model accurately estimated
the maximum amount of released drug (around 100%), as well
as a time latency T close to zero and an increasing time factor q
for all layered systems. The Two-Stage model accurately
estimated the maximum amount as well, which could be
derived from the sum of B (amount of burst release) and S
(amount of sustained release). One exception was Alg5PA,
where the maximum amount was too high as the experimental
data did not show a flattening of the release curve and the
model hence predicted an almost linear second stage and a too
high S value. With respect to release rates, a higher value of K
indicates a slower release. KB is several orders of magnitude
smaller than KS in all cases, indicating a fast initial burst release,
followed by a more sustained phase.
High burst release of ATP in all drug release systems limits
the applicability of both Peppas models. When applying the
Korsmeyer-Peppas model to the ATP release data, all release
exponents were far below the optimum exponent of 0.43 for
spheres. This indicates a diffusional release mechanism and
also a very strong burst release,51 which is supported by the
appearance of the graph (Figure 8). This model only applies to
the first 60% of the respective release, i.e., the burst release
region of all systems. Similar information could be obtained
from the Peppas-Sahlin model. Originally, K1 indicates
diffusional and K2 indicates relaxational contribution.44 While
K1 decreases from 1.85 h−0.43 in Alg to 0.3 h−0.43 in Alg5PA,
indicating a slowed down diffusion process, K2 also decreases
in value, but is always negative. A negative K2 might indicate
inhibited release, which can also be explained by a Two-Stage
release with strong burst and almost constant release after
that.51 AICc was rather high for both Korsmeyer-Peppas and
Peppas-Sahlin models, while the Weibull and Two-Stage
models on the other hand could be fitted with much lower
AICc, suggesting better fitting to the data. While the Weibull
model correctly yielded the total release of ATP, the Two￾Stage model provided further insight: The burst release of each
system could be calculated individually from the sustained
release stage, providing actual numeric values for each stage.
Using the parameters KB and KS, the times until half-maximum
of the respective stage, provides good information on how fast
each process was. The smaller the value, the faster the process
becomes. The values of KB and KS increased over all systems
from Alg to Alg5PA, indicating that both burst and sustained
release processes became slower. Additionally, the burst release
values B of the Two-Stage model reveal why the Peppas
models only fitted poorly: The burst release was always higher
than 50%, which is why fitting only up to 60% did not
represent the real release behavior. Li et al. correlated drug
release mechanisms and half-life t50 of the release.33 ATP
release data fit very well to this chart: diffusion-based burst
release occurs with KB below 1 day, while the sustained phase
II has KS above 4 days, suggesting electrostatic interaction or
even beginning degradation of the system, although the latter
was not observable within 14 days. The rather small molecule
ATP (Figure 1A) was released very fast and showed high burst
release regardless of the layer system (Figure 8). This is
comparable to many other small, hydrophilic drugs like
doxorubicin or paracetamol released from alginate microbead
systems.52,53 A study, where ciprofloxacin was loaded to hollow
poly(allyl amine)/poly(methacrylic acid) capsules, showed a
reduction of the released amount of drug from 70% to 40% in
the first 12 h.54 In another example, four alternating
polyelectrolyte layers of poly(allyl amine) and poly-
(styrenesulfonate) on alginate microbeads effectively decreased
the release of sodium benzoate and zosteric acid from over
95% to below 60% after 24 h and prolonging the overall release
to around 72 and 120 h, respectively.55 In a previously
published study, we investigated the release of ATP from
delivery devices made of agarose and hydroxyapatite and could
sustain the release to around 4 days.42 Compared to these data,
Figure 9. Release kinetics fitted to experimental data of suramin release from alginate (Alg) and layered alginate beads according to different kinetic
models: Korsmeyer-Peppas (A), Weibull (B), and Two-Stage (C). A is fitted to 60−70% of the respective release, due to model constraints.
Layering with five alternating coating layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine)
and alginate (PA). Error bars have been omitted for clarity (see Figure 7).
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
coating the beads with poly(allyl amine), a densely charged
polycation, proved to be very effective in both decreasing burst
release to about 50% and altering the phase II release toward a
zero-order release behavior after that.
3.6. Evaluation of Suramin Release. Models fitted to the
experimental data are displayed in Figure 9A−C and the values
are summarized in Table 2. For suramin, Korsmeyer-Peppas
exponents are higher than for ATP, especially in case of Alg
and Alg5PA, where there a less distinct burst release was
observed and the Peppas model resulted in a good fit Figure
9A).
For Alg, the release exponent of 0.38 is very close to the
“ideal” value of 0.43 for spherical systems. All release processes
followed a diffusion mechanism, which was additionally
supported by the fact that Peppas-Sahlin constants showed
higher values in K1 than in K2. Again, K2 had negative values,
indicating inhibited release in the second stage. However, the
values tended to be closer to zero when compared to ATP. As
with ATP, the Weibull and Two-Stage models yielded better
fits, the latter usually having the lowest AICc. The Weibull
model predicted M∞ accurately with small latency times and
increasing q values where the release was slowed down. This
was especially the case for Alg5PL, which showed a burst
release up to 30% and a constant release after that. However,
the Weibull model failed to accurately fit the Alg5CL sample.
This could be solved by applying the Two-Stage model.
Additionally, the Two-Stage model provided insight into the
ratio of burst and sustained release and how fast the respective
processes occurred. By adding B and S, the maximum amount
of drug released at an infinite time point can be calculated.
While all other systems add up to around 100%, Alg5PL had a
maximum of only 55%, with around 40% being released within
the observed 14 days. In general, coating with polyelectrolytes
helped to slow down suramin release and decrease burst
release. However, pristine alginate beads and Alg5PA systems
remain comparable, the alginate coating did not have as much
an effect on sustaining release than lignosulfonate coating.
The release of the structurally larger but also well-soluble
suramin (Figure 1C) was slower than that of ATP and it
somewhat resembles that of vancomycin from alginate/
chitosan bead systems.34 Vancomycin is a negatively charged
steroid of similar size to suramin, which is why the release is
prolonged most when dense positively charged bead systems
are being used. Using KB and KS, the release mechanisms can
be estimated. KB below a value of 2 days again indicates fast
diffusion during burst release phase, while the phase II half￾times of the layered systems are much higher (KS > 7 days),
suggesting a higher drug-matrix interaction. Possible effects
include electrostatic and hydrophobic interactions as well as
matrix degradation.33 As shown in Figure 9, the systems with
lignosulfonate as polyanion had the largest impact on
sustaining the release. Some studies investigating adsorption
processes for pollutant removal and chromatography have
described enhanced interaction and adsorption of aromatic
compounds on columns modified with sulfonic acid groups
due to electrostatic and π−π-interactions.56−58 This might be
comparable to the present suramin/lignosulfonate system,
where both the drug and the layers bear aromatic and sulfonate
moieties. Hence, their sulfonic acid-π interactions could sustain
the release of the drug similarly to retention in chromatog￾raphy. Compared to our previously published systems,42 the
layered bead systems, with Alg5CL in particular, perform
better in terms of sustained release than agarose/hydrox￾yapatite composite materials. Two other examples by Nie et al.,
Table 2. Kinetic Models Fitted to Experimental Data of Suramin Release from Different Alginate (Alg) Bead Systemsa
model
suramin material Korsmeyer-Peppas Peppas-Sahlin Weibull Two-Stage
Alg K = 0.15 ± 0.02 h−n K1 = 0.14 ± 0.02 h−0.43 M∞ = 1.0 ± 0.1 B = 0.16 ± 0.02
n = 0.38 ± 0.04 K2 = −0.003 ± 0.005 h−0.86 T = 0.0 ± 0.2 h S = 0.92 ± 0.03
q = 53 ± 4 h KB = 0.08 ± 0.07 h
KS = 47 ± 8 h
RSS = 2.00 × 10−2 RSS = 2.27 × 10−2 RSS = 4.79 × 10−2 RSS = 1.68 × 10−2
AICc = −60.00 AICc = −58.58 AICc = −91.74 AICc = −110.55
Alg5CL K = 0.29 ± 0.01 h−n K1 = 0.45 ± 0.04 h−0.43 M∞ = 0.6 ± 0.2 B = 0.35 ± 0.06
n = 0.08 ± 0.01 K2 = −0.13 ± 0.02 h−0.86 T = 0.0 ± 0.4 h S = 0.80 ± 0.05
q = 6 ± 1 h KB = 0.10 ± 0.06 h
KS = 386 ± 49 h
RSS = 2.11 × 10−3 RSS = 1.49 × 10−2 RSS = 1.82 × 10−1 RSS = 4.84 × 10−2
AICc = −84.70 AICc = −46.84 AICc = −67.68 AICc = −91.52
Alg5PL K = 0.22 ± 0.01 h−n K1 = 0.32 ± 0.02 h−0.43 M∞ = 1.2 ± 0.5 B = 0.29 ± 0.03
n = 0.25 ± 0.02 K2 = −0.09 ± 0.02 h−0.86 T = 0.08 ± 0.01 h S = 0.26 ± 0.08
q = 11585 ± 2284 h KB = 0.29 ± 0.03 h
KS = 174 ± 17 h
RSS = 7.50 × 10−4 RSS = 1.51 × 10−3 RSS = 7.77 × 10−3 RSS = 5.81 × 10−3
AICc = −62.20 AICc = −56.60 AICc = −115.30 AICc = −120.24
Alg5PA K = 0.29 ± 0.02 h−n K1 = 0.34 ± 0.02 h−0.43 M∞ = 1.0 ± 0.2 B = 0.60 ± 0.04
n = 0.27 ± 0.04 K2= −0.05 ± 0.01 h−0.86 T = 0.2 ± 0.1 h S = 0.5 ± 0.2
q = 33 ± 6 h KB = 1.1 ± 0.2 h
KS = 990 ± 22 h
RSS = 3.38 × 10−2 RSS = 1.49 × 10−2 RSS = 3.32 × 10−2 RSS = 1.84 × 10−2
AICc = −46.91 AICc = −55.10 AICc = −90.59 AICc = −100.59
Five alternating coating layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine) and alginate
(PA). Mean ± SD, n = 3.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
where suramin was incorporated into poly(lactic acid) (PLA)/
poly(lactic-co-glycolic acid (PLGA) core/shell microspheres,
showed comparable results.59 When loaded into PLGA shell,
70% of suramin was released within 3 days and 100% was
released after 30 days. When loaded into the PLA core, a
slower release behavior could be observed, where 70% was
released only after 15 days. While the release mechanism
(diffusion) is the same, in PLA/PLGA systems the rate￾controlling mechanism is different. PLA and PLGA are more
hydrophobic, but start to degrade via ester hydrolysis upon
contact with water. Once enough ester bonds have been
cleaved, water permeates further into the system causing
swelling, further degradation, and ultimately drug diffusion
through the water-filled pores.46,60 The Alg5 systems of the
present study on the other hand do not degrade over the
course of 14 days and the release rate is controlled through
electrostatic interactions between drug and layers.
3.7. Evaluation of A740003 Release. Models fitted to
the experimental data are displayed in Figure 10A−C and the
values are displayed in Table 3. In case of A740003, a less
water-soluble P2X7 receptor antagonist, the Peppas models
also yielded decreasing diffusion constants for the layer
systems. Relaxational contribution was close to zero and the
exponent suggests purely diffusional but inhibited release. A
Figure 10. Release kinetics fitted to experimental data of A740003 release from alginate (Alg) and layered alginate beads: Korsmeyer-Peppas (A),
Weibull (B), and Two-Stage (C). A is fitted to 60−70% of the respective release, due to model constraints. Layering with five alternating coating
layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine) and alginate (PA). Error bars have been
omitted for clarity (see Figure 7).
Table 3. Kinetic Models Fitted to Experimental Data of A740003 Release from Different Alginate (Alg) Bead Systemsa
model
A740003 material Korsmeyer-Peppas Peppas-Sahlin Weibull Two-Stage
Alg K = 0.15 ± 0.01 h−n K1 = 0.185 ± 0.008 h−0.43 M∞ = 1.0 ± 0.4 B = 0.23 ± 0.01
n = 0.27 ± 0.05 K2= −0.032 ± 0.002 h−0.86 T = 0.24 ± 0.04 h S = 0.52 ± 0.04
q = 60880 ± 2329 h KB = 0.38 ± 0.09 h
KS = 737 ± 79 h
RSS = 6.33 × 10−4 RSS = 5.65 × 10−4 RSS = 4.07 × 10−3 RSS = 1.04 × 10−3
AICc = −14.87 AICc = −37.62 AICc = −75.88 AICc = −88.64
Alg5CL K = 0.138 ± 0.006 h−n K1 = 0.179 ± 0.008 h−0.43 M∞ = 1.0 ± 0.1 B = 0.18 ± 0.02
n = 0.24 ± 0.04 K2= −0.038 ± 0.004 h−0.86 T = 0.22 ± 0.12 h S = 0.40 ± 0.13
q = 8.14 × 105 h KB = 0.23 ± 0.07 h
KS = 639 ± 81 h
RSS = 3.49 × 10−5 RSS = 1.70 × 10−4 RSS = 5.18 × 10−3 RSS = 9.99 × 10−4
AICc = −29.36 AICc = −44.82 AICc = −72.97 AICc = −92.73
Alg5PL K = 0.084 ± 0.002 h−n K1 = 0.096 ± 0.008 h−0.43 M∞ = 1.0 ± 0.4 B = 0.102 ± 0.004
n = 0.16 ± 0.02 K2= −0.018 ± 0.002 h−0.86 T = 0.23 ± 0.08 h S = 0.09 ± 0.06
q = 3.13 × 1013 h KB = 0.15 ± 0.03 h
KS = 443 ± 47 h
RSS = 2.82 × 10−5 RSS = 5.43 × 10−4 RSS = 3.30 × 10−4 RSS = 1.50 × 10−4
AICc = −30.42 AICc = −37.86 AICc = −106.03 AICc = −115.51
Alg5PA K = 0.031 ± 0.002 h−n K1 = 0.037 ± 0.002 h−0.43 M∞ = 0.058 ± 0.001 B = 0.047 ± 0.005
n = 0.29 ± 0.05 K2 = −0.006 ± 0.001 h−0.86 T = 0.17 ± 0.07 h S = 0.012 ± 0.004
q = 1.3 ± 0.3 h KB = 0.4 ± 0.1 h
KS = 18 ± 2 h
RSS = 3.03 × 10−5 RSS = 2.43 × 10−5 RSS = 3.35 × 10−5 RSS = 2.89 × 10−5
AICc = −30.07 AICc = −56.50 AICc = −133.48 AICc = −135.24
Five alternating coating layers of chitosan and lignosulfonate (CL), poly(allyl amine) and lignosulfonate (PL), or poly(allyl amine) and alginate
(PA). Mean ± SD, n = 3.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
burst release was observed for all three systems, however, it
was far less pronounced than for ATP and suramin (23% for
pristine alginate, down to 5% for Alg5PA). The Weibull model
could be fitted to an infinite release time for Alg, Alg5CL, and
Alg5PL resulting in very high time factors q, while Alg5PA
reaches a maximum of around 6% with a q of 1.3 h. The results
show that for a less soluble drug, layering significantly reduces
the release of the drug while an unmodified bead is sufficient
enough for a sustained release, where up to 75% would be
released. The data for Alg5PA suggest that for the layered
systems only the drug on the surface or in the outside layer
releases very quickly, hence the small burst release, while
diffusion from the core is almost inhibited.
The P2X7 receptor ligand A740003 (Figure 1B) has not
been the subject of release studies yet. The presented data
suggest that its lower water solubility in comparison to the
other drugs is the main factor in sustaining release. While the
permeability of unmodified alginate beads is sufficient to
achieve a prolonged release of up to 40%, layering with
polyelectrolytes eventually led to an effectively halted release in
case of Alg5PL and Alg5PA with only a small burst release. KB
of below 1 h suggest dissolution of drug on the surface or
diffusion from the outermost layer, while the very high KS
indicate high drug-matrix interaction or poor solubility.33 The
small percentage released during sustained release phase in
Alg5PL and Alg5PA might be due to the densely charged
poly(allyl amine) layer. Supporting this theory, Alg5CL still
showed a slow release, probably due to the less densely
charged layers of chitosan and lignosulfonate. According to Li
et al., a less water-soluble drug would be released mainly upon
matrix degradation, which did not occur during the 14 day
period.33 In literature, the release of poorly water-soluble
paclitaxel from hydrolytically degradable PLA/PLGA core/
shell microbeads was reported with a similar release behavior
compared to A740003 from alginate.59 Around 30−40% of
paclitaxel was released from the PLA core within 14 days
showing only very limited burst release. As discussed for
suramin, PLA/PLGA systems are more hydrophobic and their
rate-controlling step is matrix degradation, which is com￾parable to the less water-soluble A740003 in Alg5 systems.
3.8. Release into Culture Medium. In order to estimate
release behavior into culture medium instead of water,
uncoated alginate beads and Alg5PL beads were subjected to
glucose-free Earle’s Balanced Salt Solution (EBSS). This
medium was selected for its mineral composition, which is
similar to Dulbecco’s Modified Eagle Medium (DMEM), a
common culture medium for osteogenic cell lines, but without
vitamins and amino acids. The Alg5PL system was selected for
this prospective measurements as it showed sustained release
for all three tested drugs, uncoated Alg beads were used as
reference. In Figure 11, the release from Alg and Alg5PL into
EBSS is shown together with the modeled Two-Stage release
into water (dotted lines, including their 95% confidence
levels).
For ATP (A), release into EBSS is similar to that in water,
either completely within or very close to the Two-Stage
model’s confidence levels. This is due to ATP’s small size and
fast release. The presence of other ions in the medium does
not have a significant influence on the release of ATP. For
suramin (B), release from pure alginate beads into EBSS is
slightly faster, albeit comparable, to that into water. From
Alg5PL, the release into EBSS is faster than into water but also
more steady with a less pronounced burst release. The
maximum amount released into EBSS after 12 days is about
60%, compared to around 40% into water. However, Alg5PL
still significantly sustains suramin release compared to
uncoated alginate beads. For A740003 (C), the results are
similar: release from both Alg and Alg5PL into EBSS is faster
than into water, but still lower than the more water-soluble
drugs. The maximum amount released after 12 days rises from
around 35% to 55% for Alg and from around 10% to 20% for
Alg5PL. These data, especially where release after day 1 is
almost exclusively moderated by matrix degradation, suggest
that the ions contained in EBSS lead to slow matrix
degradation over a longer period of time. In all, change of
release medium from water to a culture medium does not have
a heavy influence on the release of the model drugs.
4. CONCLUSIONS
Different alginate beads layered with polyelectrolytes were
developed using a fast and facile layer-by-layer procedure for a
sustained release of drugs suitable for guided osteogenic stem
cell differentiation. The release of the three compounds
adenosine 5′-triphosphate (ATP), suramin, and A740003
strongly depended on both system and drug and demonstrates
the importance of individual drug delivery devices. These LbL￾based systems offer new possibilities to deliver active
substances in a controlled and sustained way. Since P2X7
Figure 11. Release data of ATP (A), suramin (B), and A740003 (C) from Alg and Alg5PL (alginate beads coated with 5 layers of poly(allyl
amine)/lignosulfonate) systems into EBSS (Earle’s Balanced Salt Solution). Model curves of the Two-Stage (TS) model with their 95% confidence
levels show the respective release into water (as discussed in the previous sections). Release over 12 days (Mean ± SD, n = 3).
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
receptor expression has been found to orchestrate the
osteogenic differentiation of hMSCs and also influences the
equilibrium of osteoblasts/osteoclasts at different stages,
sustained delivery of corresponding ligands such as A740003
will be a key to trigger and guide bone formation. Providing
the P2X7 antagonist and therefore enhancing bone formation
over a longer period of time could lead to faster bone
regeneration. First promising results for the application in
culture medium EBSS show that release of the model drugs is
still sustained despite the influence of the ions contained in the
culture medium. Combining tailored release systems with
scaffolds for tissue engineering and evaluating their potential in
vitro and in vivo is part of ongoing research. With respect to
kinetic evaluation, the well-established kinetic models could be
applied within their respective limits. In those cases, where the
burst release stage was followed by an almost linear sustained
release, the Weibull model failed to accurately represent the
release data. Instead, the uncommon yet easy-to-use Two￾Stage model yielded the lowest AICc, suggesting it is the most
suitable model to describe the release of ATP, suramin, and
A740003 from alginate beads with and without additional
polyelectrolyte layers. Due to its inherent coupling of two
stages, this model was able to individually fit both burst and
sustained release stages over the whole time frame and yielded
the stages half-life times, which can be used to assess the
release mechanism. Getting more accurate results makes it an
interesting candidate for future release evaluation and
prediction.
■ ASSOCIATED CONTENT
*sı Supporting Information
The Supporting Information is available free of charge at

https://pubs.acs.org/doi/10.1021/acsabm.1c00417.

First-order release models fitted to experimental data of
adenosine 5′-triphosphate (ATP), suramin, and
A740003; model parameters for first-order release
model; and Peppas-Sahlin release model fitted to the
experimental data (PDF)
■ AUTHOR INFORMATION
Corresponding Author
Margit Schulze − Department of Natural Sciences, Bonn￾Rhein-Sieg University of Applied Sciences, 53359 Rheinbach,
Germany; orcid.org/0000-0002-8975-1753;
Email: [email protected]
Authors
Markus Witzler − Department of Natural Sciences, Bonn￾Rhein-Sieg University of Applied Sciences, 53359 Rheinbach,
Germany; Institute of Organic and Macromolecular
Chemistry, Center of Excellence of Polysaccharide Research,
Friedrich-Schiller-University Jena, 07743 Jena, Germany;
orcid.org/0000-0002-0468-3051
Sarah Vermeeren − Department of Natural Sciences, Bonn￾Rhein-Sieg University of Applied Sciences, 53359 Rheinbach,
Germany
Roman O. Kolevatov − School of Chemistry, Raymond and
Beverly Sackler Faculty of Exact Sciences, Tel Aviv University,
Tel Aviv 6997801, Israel
Razan Haddad − Department of Pharmaceutical Technology,
Faculty of Pharmacy, Jordan University of Science and
Technology, Irbid 22110, Jordan
Martin Gericke − Institute of Organic and Macromolecular
Chemistry, Center of Excellence of Polysaccharide Research,
Friedrich-Schiller-University Jena, 07743 Jena, Germany
Thomas Heinze − Institute of Organic and Macromolecular
Chemistry, Center of Excellence of Polysaccharide Research,
Friedrich-Schiller-University Jena, 07743 Jena, Germany;
orcid.org/0000-0001-7726-6593
Complete contact information is available at:

https://pubs.acs.org/10.1021/acsabm.1c00417

Author Contributions
Conceptualization, M.W. and M.S.; methodology, M.W. and
S.V.; formal analysis, M.W., S.V., R.K., and R.H.; investigation,
S.V., R.K., and R.H.; resources, M.S.; writingoriginal draft
preparation, M.W.; writingreview and editing, M.W., M.G.,
T.H., and M.S.; visualization, M.W. and S.V.; project
administration, M.S., M.G., and T.H.; funding acquisition,
M.S. All authors have read and agreed to the published version
of the manuscript.
Funding
This work was funded by the German Federal Ministry of
Education and Research (BMBF) programs “Ingenieur
Nachwuchs 2015″: “PersoImplant” (grant no. 13FH019IX5)
and the Ministry of Culture and Science of the German State
of North Rhine-Westphalia (MKW NRW) program “FH Zeit
für Forschung 2016″: “TestMedgO” (grant no. 1609FHZ027).
Parts of the work were supported by “NRW-Nahost”
scholarship program for R.K. and R.H.
Notes
The authors declare no competing financial interest.
■ ABBREVIATIONS
AICc, (corrected) Akaike’s Information Criterion
Alg, alginate hydrogel beads
Alg5CL, alginate hydrogel beads layered with 5 layers of
chitosan and lignosulfonate
Alg5PA, alginate hydrogel beads layered with 5 layers of
poly(allyl amine) and alginate
Alg5PL, alginate hydrogel beads layered with 5 layers of
poly(allyl amine) and lignosulfonate
ATP, adenosine 5′-triphosphate
CL, coating layer of chitosan and lignosulfonate
DMSO, dimethyl sulfoxide
PA, coating layer of poly(allyl amine) and alginate
PL, coating layer of poly(allyl amine) and lignosulfonate
PLA, poly(lactic acid)
PLGA, poly(lactic-co-glycolic acid)
RSS, residual sum of squares
SLS, static light scattering
UV−vis, ultraviolet/visible light
■ REFERENCES
(1) Kim, Y. H.; Tabata, Y. Dual-Controlled Release System of Drugs
for Bone Regeneration. Adv. Drug Delivery Rev. 2015, 94, 28−40.
(2) Witzler, M.; Büchner, D.; Shoushrah, S.; Babczyk, P.; Baranova,
J.; Witzleben, S.; Tobiasch, E.; Schulze, M. Polysaccharide-Based
Systems for Targeted Stem Cell Differentiation and Bone
Regeneration. Biomolecules 2019, 9 (12), 840.
(3) Xu, X.-Y.; He, X.-T.; Wang, J.; Li, X.; Xia, Y.; Tan, Y.-Z.; Chen,
F.-M. Role of the P2X7 Receptor in Inflammation-Mediated Changes
in the Osteogenesis of Periodontal Ligament Stem Cells. Cell Death
Dis. 2019, 10 (1), 20.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
(4) Jacobson, K. A.; Balasubramanian, R.; Deflorian, F.; Gao, Z. G. G
Protein-Coupled Adenosine (P1) and P2Y Receptors: Ligand Design
and Receptor Interactions. Purinergic Signalling 2012, 8 (3), 419−436.
(5) Zhang, Y.; Babczyk, P.; Pansky, A.; Kassack, M. U.; Tobiasch, E.
P2 Receptors Influence HMSCs Differentiation towards Endothelial
Cell and Smooth Muscle Cell Lineages. Int. J. Mol. Sci. 2020, 21 (17),
6210.
(6) Ottensmeyer, P.; Witzler, M.; Schulze, M.; Tobiasch, E. Small
Molecules Enhance Scaffold-Based Bone Grafts via Purinergic
Receptor Signaling in Stem Cells. Int. J. Mol. Sci. 2018, 19 (11), 3601.
(7) Zippel, N.; Limbach, C. A.; Ratajski, N.; Urban, C.; Luparello,
C.; Pansky, A.; Kassack, M. U.; Tobiasch, E. Purinergic Receptors
Influence the Differentiation of Human Mesenchymal Stem Cells.
Stem Cells Dev. 2012, 21 (6), 884−900.
(8) Kaebisch, C.; Schipper, D.; Babczyk, P.; Tobiasch, E. The Role
of Purinergic Receptors in Stem Cell Differentiation. Comput. Struct.
Biotechnol. J. 2015, 13, 75−84.
(9) Honore, P.; Donnelly-Roberts, D.; Namovic, M. T.; Hsieh, G.;
Zhu, C. Z.; Mikusa, J. P.; Hernandez, G.; Zhong, C.; Gauvin, D. M.;
Chandran, P.; Harris, R.; Medrano, A. P.; Carroll, W.; Marsh, K.;
Sullivan, J. P.; Faltynek, C. R.; Jarvis, M. F. A-740003 [N-(1-
{[(Cyanoimino)(5-Quinolinylamino) Methyl]Amino}-2,2-Dimethyl￾propyl)-2-(3,4-Dimethoxyphenyl)Acetamide], a Novel and Selective
P2X 7 Receptor Antagonist, Dose-Dependently Reduces Neuropathic
Pain in the Rat. J. Pharmacol. Exp. Ther. 2006, 319 (3), 1376−1385.
(10) Wang, N.; Agrawal, A.; Jørgensen, N. R.; Gartland, A. P2X7
Receptor Regulates Osteoclast Function and Bone Loss in a Mouse
Model of Osteoporosis. Sci. Rep. 2018, 8 (1), 3507.
(11) Wang, Y.; Newman, M. R.; Benoit, D. S. W. Development of
Controlled Drug Delivery Systems for Bone Fracture-Targeted
Therapeutic Delivery: A Review. Eur. J. Pharm. Biopharm. 2018,
127, 223−236.
(12) Dorati, R.; DeTrizio, A.; Modena, T.; Conti, B.; Benazzo, F.;
Gastaldi, G.; Genta, I. Biodegradable Scaffolds for Bone Regeneration
Combined with Drug-Delivery Systems in Osteomyelitis Therapy.
Pharmaceuticals 2017, 10 (4), 96.
(13) Uyen, N. T. T.; Hamid, Z. A. A.; Tram, N. X. T.; Ahmad, N.
Fabrication of Alginate Microspheres for Drug Delivery: A Review.
Int. J. Biol. Macromol. 2020, 153, 1035−1046.
(14) Li, Y.; Zhou, M.; Pang, Y.; Qiu, X. Lignin-Based Microsphere:
Preparation and Performance on Encapsulating the Pesticide
Avermectin. ACS Sustainable Chem. Eng. 2017, 5 (4), 3321−3328.
(15) Fu, Y.; Kao, W. J. Drug Release Kinetics and Transport
Mechanisms of Non-Degradable and Degradable Polymeric Delivery
Systems. Expert Opin. Drug Delivery 2010, 7 (4), 429−444.
(16) Spicer, C. D. Hydrogel Scaffolds for Tissue Engineering: The
Importance of Polymer Choice. Polym. Chem. 2020, 11 (2), 184−219.
(17) Siepmann, J.; Elkharraz, K.; Siepmann, F.; Klose, D. How
Autocatalysis Accelerates Drug Release from PLGA-Based Micro￾particles: A Quantitative Treatment. Biomacromolecules 2005, 6 (4),
2312−2319.
(18) Castelli, F.; Pitarresi, G.; Giammona, G. Influence of Different
Parameters on Drug Release from Hydrogel Systems to a
Biomembrane Model. Evaluation by Differential Scanning Calorim￾etry Technique. Biomaterials 2000, 21 (8), 821−833.
(19) Kamaly, N.; Yameen, B.; Wu, J.; Farokhzad, O. C. Degradable
Controlled-Release Polymers and Polymeric Nanoparticles: Mecha￾nisms of Controlling Drug Release. Chem. Rev. 2016, 116 (4), 2602−
2663.
(20) Stevens, M. M. Biomaterials for Bone Tissue Engineering.
Mater. Today 2008, 11 (5), 18−25.
(21) Soares, P. I. P.; Sousa, A. I.; Silva, J. C.; Ferreira, I. M. M.;
Novo, C. M. M.; Borges, J. P. Chitosan-Based Nanoparticles as Drug
Delivery Systems for Doxorubicin: Optimization and Modelling.
Carbohydr. Polym. 2016, 147, 304−312.
(22) Witzler, M.; Alzagameem, A.; Bergs, M.; Khaldi-Hansen, B. E.;
Klein, S. E.; Hielscher, D.; Kamm, B.; Kreyenschmidt, J.; Tobiasch, E.;
Schulze, M. Lignin-Derived Biomaterials for Drug Release and Tissue
Engineering. Molecules 2018, 23 (8), 1885.
(23) Al-Rashed, M. M.; Niknezhad, S.; Jana, S. C. Mechanism and
Factors Influencing Formation and Stability of Chitosan/Lignosulfo￾nate Nanoparticles. Macromol. Chem. Phys. 2019, 220 (1), 1800338.
(24) Tang, Q.; Zhou, M.; Li, Y.; Qiu, X.; Yang, D. Formation of
Uniform Colloidal Spheres Based on Lignosulfonate, a Renewable
Biomass Resource Recovered from Pulping Spent Liquor. ACS
Sustainable Chem. Eng. 2018, 6 (1), 1379−1386.
(25) Rumpf, J.; Do, X. T.; Burger, R.; Monakhova, Y. B.; Schulze, M.
Extraction of High-Purity Lignins via Catalyst-Free Organosolv
Pulping from Low-Input Crops. Biomacromolecules 2020, 21 (5),
1929−1942.
(26) Morales, A.; Labidi, J.; Gullón, P. Assessment of Green
Approaches for the Synthesis of Physically Crosslinked Lignin
Hydrogels. J. Ind. Eng. Chem. 2020, 81, 475−487.
(27) Sipponen, M. H.; Lange, H.; Crestini, C.; Henn, A.; Ö
sterberg,
M. Lignin for Nano- and Microscaled Carrier Systems: Applications,
Trends, and Challenges. ChemSusChem 2019, 12 (10), 2039−2054.
(28) Piombino, C.; Lange, H.; Sabuzi, F.; Galloni, P.; Conte, V.;
Crestini, C. Lignosulfonate Microcapsules for Delivery and Con￾trolled Release of Thymol and Derivatives. Molecules 2020, 25 (4),
(29) Hu, J.; Zhu, Y.; Tong, H.; Shen, X.; Chen, L.; Ran, J. A Detailed
Study of Homogeneous Agarose/Hydroxyapatite Nanocomposites for
Load-Bearing Bone Tissue. Int. J. Biol. Macromol. 2016, 82, 134−143.
(30) Hoare, T. R.; Kohane, D. S. Hydrogels in Drug Delivery:
Progress and Challenges. Polymer 2008, 49 (8), 1993−2007.
(31) Vrignaud, S.; Benoit, J. P.; Saulnier, P. Strategies for the
Nanoencapsulation of Hydrophilic Molecules in Polymer-Based
Nanoparticles. Biomaterials 2011, 32 (33), 8593−8604.
(32) Wang, S.; Liu, R.; Fu, Y.; Kao, W. J. Release Mechanisms and
Applications of Drug Delivery Systems for Extended-Release. Expert
Opin. Drug Delivery 2020, 17 (9), 1289−1304.
(33) Li, J.; Mooney, D. J. Designing Hydrogels for Controlled Drug
Delivery. Nat. Rev. Mater. 2016, 1 (12), 16071.
(34) Unagolla, J. M.; Jayasuriya, A. C. Drug Transport Mechanisms
and in Vitro Release Kinetics of Vancomycin Encapsulated Chitosan￾Alginate Polyelectrolyte Microparticles as a Controlled Drug Delivery
System. Eur. J. Pharm. Sci. 2018, 114, 199−209.
(35) Tang, Q.; Zhou, M.; Qiu, X.; Yang, D. Effects of Cationic
Cetyltrimethylammonium Bromide on the Aggregation Behavior of
Sodium Lignosulfonate (NaLS) in Concentrated Solutions and
Preparation of Uniform Lignosulfonate-Based Colloidal Spheres. J.
Agric. Food Chem. 2020, 68 (35), 9451−9460.
(36) Kurapati, R.; Groth, T. W.; Raichur, A. M. Recent
Developments in Layer-by-Layer Technique for Drug Delivery
Applications. ACS Appl. Bio Mater. 2019, 2 (12), 5512−5527.
(37) Xu, L.; Chu, Z.; Wang, H.; Cai, L.; Tu, Z.; Liu, H.; Zhu, C.; Shi,
H.; Pan, D.; Pan, J.; Fei, X. Electrostatically Assembled Multilayered
Films of Biopolymer Enhanced Nanocapsules for On-Demand Drug
Release. ACS Appl. Bio Mater. 2019, 2 (8), 3429−3438.
(38) Su, L.; Zhang, S.; Jiang, G.; Pang, J.; Wang, D.; Shi, J.; Fang, G.
Layer-by-Layer Self-Assembly of a Lignin-Poly(Vinyl Alcohol) Based
Polyelectrolyte with a Conductivity Method. J. Appl. Polym. Sci. 2017,
134 (9), 1−9.
(39) Fragal, E. H.; Cellet, T. S. P.; Fragal, V. H.; Companhoni, M. V.
P.; Ueda-Nakamura, T.; Muniz, E. C.; Silva, R.; Rubira, A. F. Hybrid
Materials for Bone Tissue Engineering from Biomimetic Growth of
Hydroxiapatite on Cellulose Nanowhiskers. Carbohydr. Polym. 2016,
152, 734−746.
(40) Damanik, F. F. R.; Brunelli, M.; Pastorino, L.; Ruggiero, C.;
Van Blitterswijk, C.; Rotmans, J.; Moroni, L. Sustained Delivery of
Growth Factors with High Loading Efficiency in a Layer by Layer
Assembly. Biomater. Sci. 2020, 8 (1), 174−188.
(41) Jacobson, K. A. P2X and P2Y Receptors. Tocris Biosci. Sci. Rev.
Ser. 2010, 33, 1−15.
(42) Witzler, M.; Ottensmeyer, P. F.; Gericke, M.; Heinze, T.;
Tobiasch, E.; Schulze, M. Non-Cytotoxic Agarose/Hydroxyapatite
Composite Scaffolds for Drug Release. Int. J. Mol. Sci. 2019, 20 (14),
3565.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX
(43) Ritger, P. L.; Peppas, N. A. A Simple Equation for Description
of Solute Release I. Fickian and Non-Fickian Release from Non￾Swellable Devices in the Form of Slabs, Spheres, Cylinders or Discs. J.
Controlled Release 1987, 5 (1), 23−36.
(44) Peppas, N. A.; Sahlin, J. J. A Simple Equation for the
Description of Solute Release. III. Coupling of Diffusion and
Relaxation. Int. J. Pharm. 1989, 57 (2), 169−172.
(45) Bruschi, M. L. Strategies to Modify the Drug Release from
Pharmaceutical Systems, 1st ed.; Elsevier: Amsterdam, 2015
DOI: 10.1016/C2014-0-02342-8.
(46) Fredenberg, S.; Wahlgren, M.; Reslow, M.; Axelsson, A. The
Mechanisms of Drug Release in Poly(Lactic-Co-Glycolic Acid)-Based
Drug Delivery SystemsA Review. Int. J. Pharm. 2011, 415 (1−2),
34−52.
(47) Siepmann, J.; Siepmann, F. Modeling of Diffusion Controlled
Drug Delivery. J. Controlled Release 2012, 161 (2), 351−362.
(48) Paolino, D.; Tudose, A.; Celia, C.; Di Marzio, L.; Cilurzo, F.;
Mircioiu, C. Mathematical Models as Tools to Predict the Release
Kinetic of Fluorescein from Lyotropic Colloidal Liquid Crystals.
Materials 2019, 12 (5), 693−716.
(49) Lee, P. I. Modeling of Drug Release from Matrix Systems
Involving Moving Boundaries: Approximate Analytical Solutions. Int.
J. Pharm. 2011, 418 (1), 18−27.
(50) Spiess, A. N.; Neumeyer, N. An Evaluation of R2as an
Inadequate Measure for Nonlinear Models in Pharmacological and
Biochemical Research: A Monte Carlo Approach. BMC Pharmacol.
2010, 10, 1−11.
(51) Ford, J. L.; Mitchell, K.; Rowe, P.; Armstrong, D. J.; Elliott, P.
N. C.; Rostron, C.; Hogan, J. E. Mathematical Modelling of Drug
Release from Hydroxypropylmethylcellulose Matrices: Effect of
Temperature. Int. J. Pharm. 1991, 71 (1−2), 95−104.
(52) Boi, S.; Rouatbi, N.; Dellacasa, E.; Di Lisa, D.; Bianchini, P.;
Monticelli, O.; Pastorino, L. Alginate Microbeads with Internal
Microvoids for the Sustained Release of Drugs. Int. J. Biol. Macromol.
2020, 156, 454−461.
(53) Treenate, P.; Monvisade, P. In Vitro Drug Release Profiles of
PH-Sensitive Hydroxyethylacryl Chitosan/Sodium Alginate Hydro￾gels Using Paracetamol as a Soluble Model Drug. Int. J. Biol.
Macromol. 2017, 99, 71−78.
(54) Anandhakumar, S.; Debapriya, M.; Nagaraja, V.; Raichur, A. M.
Polyelectrolyte Microcapsules for Sustained Delivery of Water-Soluble
Drugs. Mater. Sci. Eng., C 2011, 31 (2), 342−349.
(55) Wang, Q.; Zhang Newby, B. Layer-by-Layer Polyelectrolyte
Coating of Alginate Microgels for Sustained Release of Sodium
Benzoate and Zosteric Acid. J. Drug Delivery Sci. Technol. 2018, 46
(April), 46−54.
(56) Horak, J.; Lindner, W. Contribution of Sulfonyl-Aromatic and
Sulfonic Acid-Aromatic Interactions in Novel Sulfonyl/Sulfonic Acid￾Embedded Reversed Phase Materials. J. Chromatogr. A 2008, 1191
(1−2), 141−156.
(57) Luo, X.-P.; Fu, S.-Y.; Du, Y.-M.; Guo, J.-Z.; Li, B. Adsorption of
Methylene Blue and Malachite Green from Aqueous Solution by
Sulfonic Acid Group Modified MIL-101. Microporous Mesoporous
Mater. 2017, 237, 268−274.
(58) Zhang, H.; Li, A.; Sun, J.; Li, P. Adsorption of Amphoteric
Aromatic Compounds by Hyper-Cross-Linked Resins with Amino
Groups and Sulfonic Groups. Chem. Eng. J. 2013, 217, 354−362.
(59) Nie, H.; Fu, Y.; Wang, C. Paclitaxel and Suramin-Loaded Core/
Shell Microspheres in the Treatment of Brain Tumors. Biomaterials
2010, 31 (33), 8732−8740.
(60) Bode, C.; Kranz, H.; Fivez, A.; Siepmann, F.; Siepmann, J.
Often Neglected: PLGA/PLA Swelling Orchestrates Drug Release:
HME Implants. J. Controlled Release 2019, 306, 97−107.
ACS Applied Bio Materials www.acsabm.org Article

https://doi.org/10.1021/acsabm.1c00417

ACS Appl. Bio Mater. XXXX, XXX, XXX−XXX