Content uploaded by Anne Mari Aukan Rokstad
Author content
All content in this area was uploaded by Anne Mari Aukan Rokstad on Dec 03, 2015
Content may be subject to copyright.
Alginate microbeads are complement compatible, in contrast to polycation
containing microcapsules, as revealed in a human whole blood model
Anne Mari Rokstad
a,
⇑
, Ole-Lars Brekke
b
, Bjørg Steinkjer
a
, Liv Ryan
a
, Gabriela Kolláriková
c
, Berit L. Strand
d
,
Gudmund Skjåk-Bræk
d
, Igor Lacík
c
, Terje Espevik
a
, Tom Eirik Mollnes
e,f
a
Department of Cancer Research and Molecular Medicine, Norwegian University of Science and Technology, Trondheim, Norway
b
Department of Laboratory Medicine, Nordland Hospital, Bodø and University of Tromsø, Tromsø, Norway
c
Department of Special Polymers and Biopolymers, Polymer Institute of the Slovak Academy of Sciences, Bratislava, Slovakia
d
Department of Biotechnology, Norwegian University of Science and Technology, Trondheim, Norway
e
Institute of Immunology, Oslo University Hospital Rikshospitalet and University of Oslo, Oslo, Norway
f
Research Laboratory, Nordland Hospital, Bodø and University of Tromsø, Tromsø, Norway
article info
Article history:
Received 5 November 2010
Received in revised form 17 February 2011
Accepted 9 March 2011
Available online 12 March 2011
Keywords:
Alginate
Microcapsule
Microbead
Complement
Biocompatibility
abstract
Alginate microbeads and microcapsules are presently under evaluation for future cell-based therapy.
Defining their inflammatory properties with regard to humans is therefore essential. A lepirudine-based
human whole blood model was used as an inflammation predictor by measuring complement and leuko-
cyte stimulation. Alginate microbeads were complement-compatible since they did not activate comple-
ment as measured by the soluble terminal complement complex (sTCC), Bb or the anaphylatoxins C3a
and C5a. In addition, alginate microbeads were free of surface adherent leukocytes. In contrast, microcap-
sules containing poly-L-lysine (PLL) induced elevated levels of sTCC, Bb, C3a and C5a, surface active C3
convertase and leukocyte adhesion. The soluble PLL induced elevated levels of sTCC and up-regulated leu-
kocyte CD11b expression. PMCG mic rocapsules containing poly(methylene-co-guanidine) complexed
with sodium alginate and cellulose sulfate triggered a fast sTCC response and C3 deposition. The PMCG
microcapsules were still less activating than PLL-containing microcapsules as a function of time. The
amounts of anaphylatoxins C3a and C5a were diminished by the PMCG microcapsules, whereas leukocyte
adherence demonstrated surface activating properties. We propose the whole blood model as an impor-
tant tool for measuring bioincompatibility of microcapsules and microbeads for future applications as
well as determining the mechanisms leading to inflammatory reactions.
Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
1. Introduction
Cell-based therapy using alginate containing microspheres has
been suggested for the treatment of hormone deficiencies [1] as
well as brain cancer [2]. The long-term function in experimental
animals is, however, often hampered by overgrowth reactions
leading to reduced graft performance [1]. The factors contributing
to graft failure of the encapsulated cells are only partly understood.
There is currently a need for experimental models relevant to hu-
mans reflecting the complexity of host factors that are encountered
upon transplantation. Human blood contains most of the cells and
effectors of the inflammatory machinery, thus it could be used as a
source. The critical steps required to mimic the physiological
in vivo situation using whole blood lie in the sampling, anticoagu-
lation and incubation conditions, which need to be fully controlled.
Biomaterials in direct contact with blood induce immediate
inflammatory responses through plasma cascades like the comple-
ment, coagulation and contact systems, with subsequent interplay
with inflammatory cells [3]. The level of activation is closely re-
lated to the surface properties of the materials. The activation of
complement may thus be a sensitive indicator of the ability of a
biomaterial to trigger inflammatory reactions. In the present work
we have used the novel whole blood model anti-coagulated with
the hirudin analog lepirudin [4] to study different microcapsules
containing alginate with a focus on complement and leukocyte
activation. Lepirudin specifically inhibits thrombin in the coagula-
tion cascade while not affecting the complement cascade and
inflammatory cells [4]. In this way it has been possible to study
the mutual interactions between complement and leukocytes, as
well as between other branches of the inflammatory network.
Complement is a major pro-inflammatory system that acts up-
stream of the leucocyte and cytokine responses. The complement
system consists of around 30 plasma and membrane-bound pro-
teins. The central event of complement activation is cleavage of
1742-7061/$ - see front matter Ó 2011 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved.
doi:10.1016/j.actbio.2011.03.011
⇑
Corresponding author. Tel.: +47 72825353; fax: +47 72571463.
E-mail address: anne.m.rokstad@ntnu.no (A.M. Rokstad).
Acta Biomaterialia 7 (2011) 2566–2578
Contents lists available at ScienceDirect
Acta Biomaterialia
journal homepage: www.elsevier.com/locate/actabiomat
the C3 protein into the opsonin C3b and the anaphylatoxin C3a. C3
is cleaved by the C3 convertase from the classical/lectin pathway
(C4bC2a) or from the alternative pathway (C3bBb), depending on
the activators. The activators on a biomaterial surface may be ad-
sorbed IgG inducing classical activation or various forms of C3b
(e.g. conformationally changed C3 [5] and/or spontaneously hydro-
lyzed C3b analog C3(H
2
O)) inducing alternative pathway activa-
tion. Irrespective of the initial event leading to C3b deposition,
this essential step initiates amplification by the alternative path-
way convertase, leading to escalated C3 activation. Surface bound
C3b is further assembled with the C3 convertases to form the C5
convertases (C4bC2aC3b and C3bBbC3b), which cleave C5 to C5a
and C5b. The anaphylatoxin C5a is the most potent inflammatory
mediator of complement, whereas C5b is the staging point for for-
mation of the terminal C5b-9 complement complex (TCC) on cell
membranes or in solution, sC5b-9 (sTCC) [6]. The ability of bioma-
terials to trigger complement activation seems to be directly re-
lated to whether C3b is able to form covalent links to the surface
hydroxyl or amino groups. This linkage may be formed directly
on biomaterial surfaces [7] or through adsorbed proteins exposing
hydroxyl or amino groups [8]. In addition, the amount of bound
water to the surface polymers might be important for their com-
plement activating abilities [9].
To our knowledge, complement reactions to alginate micro-
beads (Ca
2+
and Ba
2+
crosslinked alginate) and microcapsules (an
alginate core with a polycation/polyanion complexed membrane)
have not been addressed before. By using the lepirudin-based
whole blood model the ability of various types of microspheres
to activate complement could be studied under identical condi-
tions. Alginate microbeads [10,11] and PMCG microcapsules [12–
14] considered for future pancreatic islet transplantation were
evaluated. In addition, poly-L-lysine (PLL) containing microcap-
sules were included to further elucidate the mechanisms behind
their inflammatory reactivity [15–18].
Thus, the aim of the present study was to compare the inflam-
matory potential of different alginate microspheres, as their ability
to activate complement and leukocytes using the human lepirudin
anti-coagulated whole blood model.
2. Materials and methods
2.1. Study design
The study included five different types of microspheres: Ca/Ba
beads (gelled in 1 mM BaCl
2
/50 mM CaCl
2
), Ba beads (gelled in
20 mM BaCl
2
), APA microcapsules (Ca beads coated with PLL and
alginate), AP microcapsules (Ca beads coated with PLL) or PMCG
microcapsules, formed by polyelectrolyte complexation between
sodium alginate (SA)/cellulose sulfate (CS) with polycation
poly(methylene–co-guanidine) hydrochloride (PMCG) and calcium
cations. All microspheres were made with ultrapure alginate (spec-
ified in Section 2.2). In addition, dissolved UP-MVG was evaluated
as the alginate source for the microbeads.
2.2. Reagents and equipments
Ultrapure sodium alginates acquired from FMC BioPolymer AS
(NovaMatrix, Sandvika, Norway) were used: Pronova UP-MVG
(67% guluronic acid, intrinsic viscosity 1105 ml g
–1
, endotoxin
<43 EU g
–1
), Pronova UP-LVG (66% guluronic acid, intrinsic viscos-
ity 830 ml g
–1
, endotoxin <100 EU g
–1
) or Pronova UP100 M (Mac-
rocystis pyrifera, 44% guluronic acid, intrinsic viscosity 908 ml g
–1
,
endotoxin <26 EU g
–1
). The protein content was less than 0.3% for
all alginates, as specified by the manufacturer. Cellulose sulfate
(CS) sodium salt was from Acros Organics (Geel, Belgium), PMCG
hydrochloride supplied as a 35% aqueous solution was from
Scientific Polymer Products Inc. (Ontario, NY). Sodium chloride,
calcium chloride, barium chloride and sodium citrate of analytical
grade were from Merck (Darmstadt, Germany). Poly-L-lysine
hydrochloride (P2658, lot No. 96H5502), Tween 20, zymosan,
phosphate-buffered saline (PBS) with calcium and magnesium,
ethylenediaminetetraacetic acid (EDTA), paraformaldehyde and
LDS-751 were all purchased from Sigma–Aldrich (St Louis, MO).
Other reagents, were mannitol (HPLC degree, BDH Analar, VWR
International, Pool, UK), sterile saline (0.9% NaCl), non-pyrogenic
(B. Braun, Melsungen, Germany), lepirudin (Celgene Europe, Bou-
dry, Switzerland). Antibodies had the following specifications:
fluorescein isothiocyanate (FITC) control antibody (BD555057),
anti-CD14 FITC (BD347497) and anti-CD11b phycoerythrin (PE;
BD347557), all from Becton Dickinson (San Jose, CA). In addition,
FITC-conjugated rabbit anti-human C3c (F0201, Dako, Glostrup,
Denmark), FITC-conjugated poly rabbit anti-mouse (F0261, Dako,
Glostrup, Denmark), anti-human C5b-9 (clone aE11, Diatec, Oslo,
Norway) and biotinylated 9C4 [19] were used. For blood sampling
the following equipment was used: Nunc 1.8 and 4.5 ml polypro-
pylene vials (Nunc, Roskilde, Denmark) and a BD vacutainer top
(Becton Dickinson, Plymouth, UK).
2.3. Microsphere preparation
2.3.1. Alginate microbeads and microcapsules
Ca/Ba beads and Ba beads as well as APA and AP microcapsules
were made as described previously using 1.8% UP-MVG alginate
[20]. The gelling solution varied according to the type of micro-
sphere: Ca/Ba beads, 1 mM BaCl
2
/50 mM CaCl
2
in 0.15 M mannitol;
Ba beads, 20 mM BaCl
2
in 0.15 M mannitol; APA and AP, 50 mM
CaCl
2
in 0.15 M mannitol. PLL was used at a concentration of
0.05% and incubation was for 10 min. As the outer coating for the
APA microcapsules a solution of 0.1% Pronova UP100 M in 0.15 M
mannitol was used. The microspheres were made with sterile solu-
tions and under strictly sterile conditions, using autoclaved equip-
ment and a sterile hood for all steps.
2.3.2. PMCG microcapsules
Before the formation of PMCG microcapsules the polyelectro-
lytes were prepared as follows. PMCG was isolated by lyophiliza-
tion, ending with a residual water content of less than 2%. CS
was purified by treatment with activated charcoal, filtration and
precipitation in acetone as described previously [21]. The residual
water content of CS was about 10%.
PMCG microcapsules were prepared as described previously
[14], except for the following concentration changes in the poly-
mer solutions. The polyanion solution contained 0.90% UP-LVG,
0.90% CS (taking into account the residual water content) in 0.9%
NaCl. The polycation solution contained 1.2% PMCG, 1% CaCl
2
,
0.9% NaCl and 0.025% Tween 20. A multi-loop reactor [22] provided
a continuous encapsulation process with a reaction time of about
40 s for polyelectrolyte complexation. Equilibration of membrane
composition was obtained by treatment with 50 mM sodium cit-
rate in 0.9% NaCl solution, pH 7.4 for 10 min. The additional coating
layer was made with 0.1% CS in 0.9% NaCl solution for 10 min.
2.4. Whole blood model
2.4.1. Preparation of microspheres and controls for experiments
It is essential that the proportions by volume between blood
and additives are equal in the whole blood experiments. The total
volume of each additive was 200
l
l. Of this, 100
l
l consisted of sal-
ine containing either 50
l
l microspheres, 900
l
g UP-MVG alginate,
10
l
g zymosan (positive control) or saline (negative control). The
amount of soluble alginate corresponds to the alginate content
A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
2567
within 50
l
l of microbeads. To these additives was added 100
l
lof
PBS (with CaCl
2
/MgCl
2
) immediately before addition of 500
l
lof
whole blood.
2.4.2. Whole blood model performance
Single experiments for each blood donor were performed as
previous described using lepirudin (50
l
gml
–1
) as the anticoagu-
lant [4]. Blood was withdrawn into low activating polypropylene
Nunc tubes (4.5 ml). Immediately thereafter, 500
l
l of whole blood
were added to the various additives in low activating sterile Nunc
tubes (1.8 ml). Avoiding blood contamination of the screw cap is
essential to avoid biased activation. The samples were incubated
for 30, 120 and 360 min in an incubator (37
o
C) under continuous
rotation. Complement activation was stopped by adding EDTA
(10 mM final concentration) and centrifugation (3000 r.p.m.,
15 min). Aliquots of plasma were stored at –80 °C before analysis.
2.5. C3 deposition
After incubation of the microspheres in whole blood (Sec-
tion 2.4) the complement cascade was stopped by addition of EDTA
(10 mM final concentration). Microspheres were harvested and
washed (3) in a wash solution (0.1% bovine serum albumin,
2 mM CaCl
2
, 0.02% sodium azide in saline). For each type of micro-
sphere one fraction was added to 50
l
gml
–1
FITC-conjugated poly
rabbit anti-human C3c (C3 deposition) and the other FITC-conju-
gated poly rabbit anti-mouse (control) antibodies. The samples
were protected from light and continuously agitated for 30 min.
Thereafter the microspheres were washed (3). The deposition of
C3 was visualized by confocal laser scanning microscopy (CLSM)
(Zeiss LSM 510, Carl Zeiss MicroImaging GmbH, Göttingen, Ger-
many) with a 488 nm laser source (BP 505–530). Identical laser
settings were used on all microspheres using PMCG and APA
microcapsules incubated for 120 min as references.
2.6. Expression of CD11b
Expression of CD11b was measured after 15 min incubation in
whole blood as described in Section 2.4. Whole blood was fixed
with 0.5% PFA in an equal volume for 4 min at 37 °C, and then
stained with PE anti-CD11b, FITC anti-CD14 and the nuclear dye
LDS-751 and analyzed using a flow cytometer (Beckman Coulter
Epics XL-MCL, Coulter Corp, FL). To exclude red cells and debris
the threshold was set at FL-3. Granulocytes and monocytes were
gated in a SSC/FITC anti-CD14 dot plot, and CD11b expression mea-
sured as mean fluorescence intensity (MFI).
2.7. Cell adherence
Microspheres were prepared as described in Section 2.4 and
incubated for 180 min. Blood was removed and the microspheres
fixed in 0.5% PFA for 20 min. In order to keep the cells attached
the blood samples were prepared without addition of EDTA, as
would otherwise have been used to stop the complement cascade.
From each sample one fraction was stained with antibodies
(7.5
l
gml
–1
in PBS) against CD14 (FITC anti-CD14) and CD11b
(PE anti-CD11b) or with control antibodies (FITC mouse IgG2b,
j
/
PE mouse IgG2a,
j
). Both combinations were incubated for
30 min in the dark and under continuous agitation, thereafter care-
fully washed and finally 0.15% PFA was added. Evaluation of the
microspheres was performed using CLSM (Zeiss LSM 510, Carl Zeiss
MicroImaging GmbH, Göttingen, Germany), with 488 nm (BP 505–
530) and 543 nm (LP 650) excitation and emission wavelengths,
respectively.
2.8. Assay of complement activation
2.8.1. sTCC
The terminal sC5b-9 complex (sTCC) was quantified by electro-
immunoassay using mAb aEll specific for C9 incorporated in the
sC5b-9 complex and biotinylated 9C4 specific for C6 in the respec-
tive complex. The assay has been described in detail previously [6]
and was performed according to a later modification [19].
2.8.2. Bb, C3a and C5a
C3a and Bb was analyzed by ELISA using kits from Quidel (San
Diego, CA). C5a was analyzed using an ELISA kit from BD Bioscience
(San Diego, CA).
2.9. Statistical methods
One-way repeated measurements ANOVA with Dunnet’s multi-
ple comparison test were used to define statistical differences be-
tween saline and the other additives at a given time point.
Differences for the various additives over time were tested using
a two-way ANOVA. The data was not normally distributed due to
the low sample numbers (n = 5), therefore the data were log trans-
formed before analysis. Differences were considered significant at
P < 0.05.
2.10. Ethics
The use of human whole blood for basal experiments was ap-
proved by the Regional Ethic Committee at the Norwegian Univer-
sity of Science and Technology. The experiments were performed
in accordance with their guidelines.
3. Results
3.1. Activation of the complement cascade detected as sTCC formation
in human whole blood
The formation of sTCC indicates activation of the complement
cascade and is suggested to be the most sensitive and specific mar-
ker of complement activation. Fig. 1A shows the time kinetics of
sTCC formation after addition of saline, Ca/Ba beads and APA and
PMCG microcapsules as well as zymosan. The kinetics of sTCC acti-
vation were significantly different (P < 0.0001) for each additive.
Fig. 1B–D shows the data for the entire panel of additives at each
time point. After 30 min incubation the amount of sTCC was simi-
lar for the Ca/Ba beads, Ba beads and saline control (Fig. 1B). Over
time, the generation of sTCC was slower for Ca/Ba beads and Ba
beads compared with the saline control (Fig. 1C and D). This re-
sulted in significantly lower values (P < 0.05) for Ba beads after
120 and 360 min incubation and for Ca/Ba beads after 360 min
(Fig. 1D). The polycation containing microcapsules, APA, AP and
PMCG, resulted in significant increases (P < 0.05) in sTCC compared
with saline at all time points (Fig. 1B–D). The APA and AP micro-
capsules showed a time-dependent increase in sTCC (Fig. 1B–D).
The PMCG microcapsules induced a rapid initial increase in sTCC,
detected after 30 min (Fig. 1B). This was followed by a slower in-
crease compared with the APA and AP microcapsules at 120
(Fig. 1C) and 360 min (Fig. 1D). The dissolved UP-MVG alginate in-
duced a small increase in sTCC after 30 min, although non-signifi-
cant (Fig. 1B). After 120 and 360 min the sTCC amounts where
slightly lower than for the saline control (Fig. 1C and D). The
amount of UP-MVG alginate (900
l
g) corresponded to the amount
of alginate in the aliquots of microbeads. The data therefore indi-
cate minor differences between dissolved and gelled alginate.
2568 A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
3.2. Alternative pathway factor B activation
The amount of Bb, an activation product of the alternative path-
way factor B, increased with time (Fig. 2A–D) in a pattern resem-
bling the sTCC data. The increase in Bb was, however, slower
compared with sTCC. The kinetics of Bb formation were signifi-
cantly affected (P < 0.0001) by each additive (Fig. 2A). In samples
containing Ca/Ba beads, Ba beads or dissolved UP-MVG alginate
the amounts of Bb were lower than for the saline control over time
(Fig. 2C and D), with a significant (P < 0.05) difference at 360 min
(Fig. 2D). The highest concentration of Bb was found in samples
containing APA and AP microcapsules. The increase was significant
for AP at 120 and 360 min (Fig. 2C), and for APA at 360 min
(Fig. 2D). A non-significant elevation of Bb was found in the sam-
ples incubated with PMCG microcapsules for 30 min (Fig. 2B).
The early induction at 30 min was followed by an evident and sig-
nificant (P < 0.05) increase after 120 min (Fig. 2C). Interestingly,
the amount of Bb was not further increased by the PMCG micro-
capsules at 360 min (Fig. 2D). The saline control, however, showed
a steady increase, which resulted in similar Bb amounts for the
PMCG microcapsules and saline control after 360 min (Fig. 2D).
3.3. Anaphylatoxin C3a and C5a release
The anaphylatoxins C3a and C5a are potent pro-inflammatory
molecules derived from the cleavage of C3 and C5. The kinetics of
C3a production are shown in Fig. 3A after addition of Ca/Ba beads
and APA and PMCG microcapsules and for the controls. The kinetics
were significantly different (P < 0.0001) for each additive (Fig. 3A).
At the specific time points only a few of the additives resulted in sig-
nificantly different C3a amounts relative to the saline control. The
trends in stimulation were still consistent with the sTCC and Bb
findings, with the exception of the PMCG microcapsules. Briefly,
addition of Ca/Ba beads and Ba beads resulted in lower concentra-
tions of C3a after 120 and 360 min (Fig. 3C and D), while addition
of APA and AP microcapsules increased C3a (Fig. 3B–D). The PMCG
microcapsules initially gave a slight elevation of C3a (Fig. 3B). How-
ever, after 120 and 360 min the PMCG microcapsules resulted in
significantly (P < 0.05) lower amounts of C3a than saline (Fig. 3A–
D). The dissolved UP-MVG alginate induced a slight but significant
(P < 0.05) increase in C3a after 30 min (Fig. 3B), but after 120 and
360 min the amount of C3a was lower than for the saline control
(Fig. 3C and D). A particular finding for the saline solution was the
pronounced increase in C3a over time. This may be due to activation
by the plastic surface [23].
The kinetics of C5a production were also significantly different
(P < 0.0001) in response to Ca/Ba beads and APA and PMCG micro-
capsules and the controls (Fig. 4A). After 30 min Ca/Ba beads and
Ba beads resulted in non-significant, slight increases in C5a levels
relative to the saline control (
Fig. 4B). The C5a increase was slower
for Ca/Ba beads and Ba beads (Fig. 4C and D) compared with saline,
resulting in significantly (P < 0.05) lower amounts after 360 min
(Fig. 4D). A rise in C5a was detected on addition of APA and AP
microcapsules, giving significant increases (P < 0.05) for APA at
all time points and for AP at 120 and 360 min (Fig. 4B–D). PMCG
microcapsules initially gave a modest, but statistically significant
(P < 0.05), increase in C5a (Fig. 4B). Over time the PMCG microcap-
sules resulted in lower C5a amounts compared with the saline con-
trol, with a significant (P < 0.05) difference after 360 min (Fig. 4D).
3.4. C3 deposition on the microsphere surface
Deposition of C3 on the different microsphere surfaces is shown
in Fig. 5. The detected C3c fragment of the C3 molecule is present
in both the native C3 molecule, the active C3b convertase and its ana-
log C3(H
2
O). Detected C3 may, therefore, represent either adsorbed
native C3 molecule or the active C3 convertase (C3b or the analog
C3(H
2
O)) on the microsphere surface. The deposition of C3 on the
surface of Ca/Ba beads and APA, AP and PMCG microcapsules is
shown at different incubation times (Fig. 5A-L). After 30 min a slight
C3 staining was detected for Ca/Ba beads (Fig. 5A) and APA (Fig. 5B)
and AP microcapsules (Fig. 5C). In contrast, the PMCG microcapsules
showed pronounced staining at 30 min (Fig. 5D). The C3 staining on
the PMCG microcapsules revealed surface irregularities and rup-
tures (Fig. 5D). No further increase in C3 staining was observed for
the PMCG microcapsules after 120 (Fig. 5H) and 360 min (Fig. 5L).
The surfaces of the APA (Fig. 5F and J) and AP microcapsules
(Fig. 5G and K) showed increased staining with time. C3 staining also
increased with time for the Ca/Ba beads (Fig. 5E and I). However, Ca/
Ba beads with low detectable staining were estimated as approxi-
mately 70–90% of the microbead population at all time points. The
C3 distribution pattern was smooth and evenly distributed on the
Ca/Ba beads (Fig. 5E and I). For the APA (Fig. 5B, F and J) and AP
(Fig. 5C, G and K) microcapsules, C3 accumulated at certain points,
resulting in spotted patterns. Further, equatorial sections of the
microspheres at 360 min showed C3 located on the surface of the
microspheres (Fig. 5M–P). The depth of penetration of C3 was esti-
mated by LSM510 software fluorescence intensity profile analysis,
as shown in Fig. A1. This analysis indicated that the Ca/Ba beads were
most permeable, since C3 was found to penetrate 10–125
l
m into
the microbeads (S1A–B). Shorter penetration depths was found for
the microcapsules, estimated at 20–40
l
m for APA (Fig. A1C and
D), 10–20
l
m for AP (Fig. A1E and F) and 20
l
m for the PMCG micro-
capsules (Fig. A1G and H). The staining was specific for C3, as demon-
strated by the negative controls with corresponding insets showing
light transmission (Fig. 5Q–T).
3.5. Leukocyte activation as measured by CD11b expression and cell
adherence to the different microspheres
CD11b is the receptor for iC3b and an early activation marker of
leukocytes. Granulocyte and monocyte CD11b expression was ana-
lyzed by flow cytometry 15 min after addition of microspheres or
controls (Fig. 6A and B). The PMCG microcapsules showed signifi-
cantly (P < 0.05) higher granulocyte CD11b expression compared
with the saline control (Fig. 6A). CD11b expression was also higher
on monocytes, although not statistically significantly so (Fig. 6B).
The APA and AP microcapsules also resulted in a slight increase
in granulocyte CD11b, although not significantly (Fig. 6A). An over-
all, moderate increase in CD11b expression was found on addition
of the different microcapsules compared with yeast zymosan.
Leukocyte adherence on microspheres was evaluated after incu-
bationin whole blood for 3 h(Fig. 6C–J). Leukocytes did not adhere to
the Ca/Ba beads (Fig. 6C) or Ba beads (Fig. 6D). However, adherent
leukocytes were found on the surface of APA (Fig. 6E and F), AP
(Fig. 6G and H) and PMCG microcapsules (Fig. 6I and J). Cells ap-
peared as small circular dots on the surface of the microcapsules
and in the surrounding area, as seen by transmitted light (Fig. 6F,
H and J). The larger fraction of the adherent cells stained positive
for CD11b, while a smaller fraction stained positive for CD14
(Fig. 6E, G and I). CD11b is present on both granulocytes and mono-
cytes in only slightly different amounts, while CD14 is expressed in
higher amounts on monocytes. The monocytes in the present study
displayed 23–50 times higher CD14 expression compared with the
granulocytic population (data not shown). This indicates that the
adherent cells were mainly granulocytes, with fewer numbers of
monocytes.
3.6. Effect of soluble PLL, CS and PMCG
The activation potential of dissolved PLL, CS and PMCG was
evaluated by measuring sTCC production and CD11b expression.
A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
2569
AU/ml
AU/ml
120 min
0
5
10
15
20
25
80
90
100
*
*
*
*
*
Saline
APA
PMCG
Zymosan
Ca/Ba Beads
360 min
300
400
*
0
10
20
30
40
100
200
*
*
*
*
*
A
B
D
C
30 min
0
1
2
3
4
5
40
60
80
*
*
*
*
AU/ml AU/ml
sTCC
0 100 200 300 400
0
50
100
150
200
250
Time/min
Saline
Ca/Ba Beads
Ba Beads
APA
AP
PMCG
UP-MVG
Zymosan
Fig. 1. sTCC concentration in plasma after incubation with various microspheres, alginate and controls in human lepirudin anti-coagulated whole blood. (A) Time-dependent
sTCC amounts after addition of Ca/Ba Beads, APA, PMCG microcapsules, saline and zymosan. The additives significantly (P < 0.0001) affect sTCC formation over time. In the
lower three panels each time point is shown (B) 30, (C)120 and (D) 360 min with additional Ba beads, AP microcapsules and dissolved UP-MVG alginate included. (B–D)
⁄P < 0.05. The baseline sTCC value measured in the sample at the start of the experiment was 0.55 ± 0.16 AU ml
–1
. Data are expressed as means ± SEM, n =5.
2570 A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
A
B
C
D
360 min
µg/ml
120 min
Time/min
Bb
0 100 200 300 400
0
5
10
15
µg/ml
0
1
2
3
4
5
30 min
0
2
4
6
8
Saline
Ca/Ba Beads
Ba Beads
APA
AP
PMCG
UP-MVG
Zymosan
0
5
10
15
*
*
*
*
*
*
Saline
APA
PMCG
Zymosan
Ca/Ba Beads
*
*
*
*
µg/ml
µg/ml
Fig. 2. Bb concentration in plasma after incubation with various microspheres, alginate or controls. The various additives and time parameters in (A)–(D) are as described in
Fig. 1. The time-dependent Bb concentration was significantly (P < 0.0001) affected by each additive. (B–D) ⁄P < 0.05. The baseline value of Bb was 0.62 ± 0.12
l
gml
–1
. Data
are expressed as means ± SEM, n =5.
A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
2571
ng/ml
120 min
360 min
Time/min
C3a
ng/ml
0 100 200 300 400
0
2000
4000
6000
8000
10000
0
2000
4000
6000
8000
0
500
1000
1500
3000
4000
5000
Saline
Ca/Ba Beads
Ba Beads
APA
AP
PMCG
UP-MVG
Zymosan
0
5000
10000
15000
20000
30 min
*
*
*
*
*
Saline
APA
PMCG
Zymosan
Ca/Ba Beads
ng/ml ng/ml
A
B
C
D
Fig. 3. C3a concentration in plasma after incubation with various microspheres, alginate or controls. The various additives and time parameters in (A)–(D) are as described in
Fig. 1. The time-dependent C3a concentration was significantly (P < 0.0001) affected by the additives. (B–D) ⁄P < 0.05. The baseline value of C3a was 120 ± 28.8 ng ml
–1
. Data
are expressed as means ± SEM, n =5.
2572 A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
C5a
Time/min
0
100
200
300
400
500
0 100 200 300 400
0
10
20
30
100
200
300
0
10
20
30
50
100
150
200
250
300
350
Saline
Ca/Ba Beads
Ba Beads
APA
AP
PMCG
UP-MVG
Zymosan
0
50
100
150
200
500
1000
1500
*
*
*
*
*
*
*
*
*
*
*
*
*
Saline
APA
PMCG
Zymosan
Ca/Ba Beads
ng/ml
ng/ml
lm/gnlm/gn
A
B
C
D
120 min
360 min
30 min
Fig. 4. C5a concentration in plasma after incubation with various microspheres, alginate or controls. The various additives and time parameters in (A)–(D) are as described in
Fig. 1. The time-dependent C5a concentration was significantly (P < 0.0001) affected by the additives. (B–D) ⁄P < 0.05. The baseline value for C5a was 7.27 ± 0.99 ng ml
–1
. Data
are expressed as means ± SEM, n =5.
A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
2573
A clear dose-dependent induction of sTCC was found by addition of
PLL (Fig. 7A). The polycation PMCG resulted in increased sTCC for-
mation at the highest dose (1000
l
gml
–1
), whereas addition of CS
showed a weak inhibitory effect (Fig. 7A). In contrast, CS gave a
clear dose-dependent elevation of CD11b expression in
granulocytes (Fig. 7B). Increased granulocyte CD11b expression
was also found after addition of PLL (Fig. 7B), although with lower
potency than CS. In contrast, dissolved PMCG showed a slight dose-
Fig. 5. Deposition of C3 on the microsphere surface after incubation in human lepirudin anti-coagulated whole blood. (A–L) 3D projections made by sectioning entire
microspheres after incubation for 30, 120 and 360 min. (M–P) Projections through the equator overlaid with transmitted light images after 360 min. (Q–T) Controls are given
in the lower panels as projections (black pictures). The inserts show transmitted light equatorial sections for visualization. Bars are 100
l
m.
2574 A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
Fig. 6. Leukocyte CD11b expression and surface adhesion after the addition of various microspheres to lepirudin anti-coagulated whole blood. Leukocyte CD11b expression
detected by flow cytometry 15 min after addition of microspheres, UP-MVG alginate and controls. (A) Granulocytes CD11b expression and (B) monocytes CD11b expression.
Results are given as means ± SEM, n =5.⁄P < 0.05. (C–J) Leukocyte adhesion on microspheres after 3 h incubation. Images taken by CLSM are presented as optical sections at
the equator and 3D projections were produced from several sections through the microspheres. (C) Ca/Ba beads section, (D) Ba beads section, (E) APA projection, (F) APA
section, (G) AP projection, (H) AP section, (I) PMCG projection, (J) PMCG section. CD11b is shown in red and CD14 in green. Bars are 100
l
m.
A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
2575
dependent reduction in CD11b expression in both granulocytes
(Fig. 7B) and monocytes (Fig. 7C). PLL was a more potent inducer
of CD11b expression than CS in monocytes (Fig. 7C). It is important
to stress the time difference between the sTCC data from 120 min
incubation and the CD11b expression data measured after 15 min
incubation.
4. Discussion
The ability of different types of microspheres to provoke comple-
ment and leukocyte activation was evaluated using a whole blood
model. The whole blood model made it possible to unmask differ-
ences in complement and leukocyte activation between the various
types of microspheres. The three types of microcapsules containing
polycations activated complement, while the two types of alginate
microbeads did not induce activation. The activation was mainly
through the alternative pathway, as the patterns of product Bb spe-
cific for this pathway corresponded well with the end product sTCC.
Also, the level of C4d, formed by activation of the classical or lectin
pathway, remained low (data not shown), confirming that activation
occurred mainly through the alternative pathway.
The low complement activation from the alginate microbeads
suggests a high degree of complement biocompatibility. The slow
increase in complement activation in the saline control is consis-
tent with activation of complement induced by the surface of the
polypropylene vials used. The lower formation of complement acti-
vation products over time by alginate microbeads shows that they
are less activating than the polypropylene vials and clearly sup-
ports their low activation property. The lower activity of the algi-
nate microbeads compared with the control could be explained
by absorption of complement components in the open structured
gel matrix, as these alginate microbeads are estimated to be per-
meable for IgG (150 kDa) [24] and proteins of up to 350 kDa [11].
The C3 protein in its native form has a molecular weight of
185 kDa, thus it should be able to diffuse into the alginate matrix.
Penetration of C3 was found in the alginate microbeads to variable
depths of 10–125
l
m (Fig. A1). Despite this absorption, we empha-
size that accessible amounts of C3 should be available for forma-
tion of C3 convertase since C3 is abundant in plasma (0.7–1.8 g l
–1
).
It is suggested that the ability of materials to bind proteins and
to induce conformational changes in the proteins is essential for
their complement activating capability [8]. Alginate is rich in car-
boxylic groups, thus negatively charged plasma proteins such as
albumin (PI 4.7–5.1), C3 (PI 5.75) and C5 (PI 4.10) will have low
affinities for the alginate matrix. C3 will, upon binding to the sur-
face and undergoing a conformational change, expose highly reac-
tive thioester [25] groups, leading to an amplification loop of the
alternative pathway [5]. From the present study we could not ver-
ify whether the deposited C3 was in its native C3 form or in one of
its activated or inactivated forms, as the anti-C3c antibody detects
any part of C3 containing the C3c fragment. However, given the
low sTCC and Bb concentrations found after incubation with algi-
nate microbeads we suggest that this staining is mainly due to
C3 present in its native form, absorbed in the permeable alginate
matrix. This is also deduced from the smooth distribution pattern
that indicates freely diffusive molecules rather than a process
involving active convertases (discussed below). The low comple-
ment activation property of the alginate microbeads was also con-
firmed by a lack of adherent leukocytes, which is consistent with
complement being the most likely candidate for leukocyte activa-
tion in this model.
In contrast to the alginate microbeads, the APA and AP micro-
capsules induced elevated levels of sTCC, Bb and anaphylatoxins
C3a and C5a (Figs. 1–4), collectively demonstrating substantial
complement activation. The anaphylatoxins C3a and C5a are po-
tent inflammatory mediators which may play an active role in
the initiation of inflammatory reactions against the APA microcap-
sules. The present data also demonstrate deposition of C3 frag-
ments on the APA surface, starting as local spots and increasing
with time, with prominent deposition after 360 min (Fig. 5). While
the detected C3c fragment is present in both native C3 and active
C3b, the present data indicate active C3b on the APA surface since:
(1) depositions started at specific points giving a spotted pattern,
indicating activation loop formation of active convertases; (2) ele-
vated levels of C3a, which is produced in equimolar amounts to
activated C3b, were found; (3) elevated levels of C5a indicated
C5 convertase formation on the capsule membrane; (4) elevated
levels of sTCC indicated formation of active convertases. Comple-
ment activation is shown to be initiated when C3b forms amide
bonds with exposed amino groups [7]. Amino groups are abundant
within the polypeptide PLL and are likely candidates for C3b sur-
face binding. Soluble PLL also induced an increase in sTCC, demon-
strating the complement stimulatory property of PLL.
0 10 10 10 10 10 10
0
10
20
30
40
PLL
PBS
CS
PMCG
sTCC (AU/ml)
A
Granulocyte
0
100
200
300
CD11b (MFI)
B
Monocyte
0
200
400
600
Polymer dose (µg/ml)
CD11b (MFI)
C
-101234
0 10 10 10 10 10 10
-1 0 1 2 3 4
0 10 10 10 10 10 10
-1 0 1 2 3 4
Fig. 7. sTCC concentrations and leukocyte CD11b expression after addition of PLL,
CS and PMCG in solution to lepirudin anti-coagulated human whole blood. (A) sTCC
after 120 min incubation, (B) granulocyte CD11b expression 15 min after addition,
(C) monocyte CD11b expression 15 min after addition. Data are given as mean-
s ± SEM, n = 3 for sTCC and n = 2 for CD11b expression.
2576 A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
Ideally the PLL should be neutralized by alginate, as previously
demonstrated for soluble polyelectrolytes [17]. However, in APA
microcapsules PLL and alginate interact at the surface [20] and
PLL is found in relatively high amounts within the 100 Å thick out-
ermost layer of [26]. Although the complex of alginate and PLL has
been shown to be evenly distributed in the outermost surface, the
degree of interaction may vary, resulting in exposure of PLL in a
stimulatory state [26]. Insufficient neutralization of PLL could
therefore explain both the stimulatory properties of PLL present
in the APA microcapsules as well as the small difference between
the AP and APA microcapsules in the present study.
Surface bound iC3b is an important ligand for CD11b/CD18 (CR3
receptor) on leukocytes [27]. CD11b positive cells were found to at-
tach to the surface of APA microcapsules, thus iC3b may have been
involved in the observed cell attachment. However, we cannot ex-
clude involvement of other ligands, including surface bound fibrin-
ogen, since this is also a known ligand for the CD11b receptor [28].
The adherent cells on APA microcapsules were mainly granulo-
cytes with a scattered distribution of monocytes. These findings
are consistent with normal inflammatory reactions, the early
stages of which are dominated by neutrophilic granulocytes. The
anapylatoxin C5a up-regulates neutrophil CD11b expression [29]
and may thus contribute to the observed cell attachment.
The PMCG microcapsules initiated the fastest complement acti-
vation, as reflected in the early detectable levels of sTCC and Bb
and deposition of C3. In combination with the sTCC data this
may reflect active convertase formation. However, no further in-
crease in C3 staining was seen after 30 min incubation, indicating
reduced convertase formation with time. This was also reflected
in the sTCC response, which only doubled between 120 and
360 min incubation, while the APA microcapsules induced an in-
crease of 6–10-fold over the same time period. The explanation
for these differences may be found in the ability of the various
polyelectrolytes to stimulate or inhibit complement activity, affect
surface properties and their ability to form stable complexes.
When forming PMCG microcapsules poly(methylene–co-guani-
dine) diffuses into the CS/alginate microbead and complexes pre-
dominantly with CS. Leakage of polyelectrolytes may occur as a
result of loose complexation or excess polymer. In the present
study PMCG in solution was found to induce an increase in sTCC.
Moreover, the soluble CS induced rapid up-regulation of leukocyte
CD11b. Thus, rapid activation by the PMCG microcapsules might be
explained by the release of polymer, with CS a likely candidate for
early leukocyte CD11b up-regulation and PMCG for rapid comple-
ment activation. In contrast, soluble CS gave a slight concentration-
dependent decrease in sTCC response compared with the control.
This finding points towards the reported complement inhibiting
activity of cellulose sulfate [30–32] and may thus explain the lower
complement stimulatory abilities of the PMCG microcapsules with
time.
A particularly interesting finding with the PMCG microcapsules
was the low amounts of the anaphylatoxins C3a and C5a detected
after the longer incubation times. The increased amount of Bb and
particularly sTCC indicates that the anaphylatoxins had been
formed. The most plausible explanation is therefore adsorption of
the highly positively charged C3a (PI 9.7) and C5a (PI 8.6) to the
negatively charged CS on the surface of PMCG microcapsules. This
implies that the fluid phase concentrations do not necessarily re-
flect the activation potential of a surface. Anaphylatoxins present
on the surface may still be biologically active and contribute to
the leukocyte adherence on the PMCG microcapsule, in addition
to the opsonic effect of bound iC3b, as previously discussed.
The complement activation profiles of alginate microbeads and
APA microcapsules in the present study correspond well with pre-
vious biocompatibility studies showing polycation containing
microcapsules to be less biocompatible [15–18,33,34]. In such
studies bioincompatibility has been measured as overgrowth reac-
tions that might have been caused by inflammatory reactions. The
complement system is a primary inductor of inflammation were its
protein effectors reacts upstream of leukocytes and cytokines
[4,23,35–37]. Complement activity may therefore be a useful
parameter for revealing bioincompatibility. The lepirudin whole
blood model could be used as a rapid ‘‘screening’’ assay to unmask
reactive surfaces, as in the case of the APA microcapsules in the
present study. The presence of platelets and coagulation proteins
are likely to provide a tougher environment in the whole blood
model compared with, for example, the peritoneal cavity, which
is a common implantation site for microspheres. Also, platelets
may contribute to enhanced complement reactions [38]. However,
complement and leukocytes are present in body fluids [39] and
blood may come into direct contact with the capsule material dur-
ing implantation, thus the same inflammatory mechanisms must
be anticipated, but perhaps at lower intensities. A sensitive model
like the one used in the present study is advantageous for safety
evaluations, since it reveals the immune incompatibility of the im-
planted material in a human model.
5. Conclusion
The present study has demonstrated the effectiveness of the
lepirudin-based whole blood model to reveal reactive surfaces by
triggering complement and activating leukocytes. Polycation con-
taining APA and PMCG microcapsules triggered complement and
leukocyte activation, while alginate microbeads consisting of only
alginate and divalent cations did not provoke complement reac-
tions. The human whole blood model seems to be a sensitive and
efficient method of revealing bioincompatibility. The method could
therefore be used to determine the safety of different microcap-
sules for transplantation purposes.
Acknowledgements
This work has been financially supported by grants from the
Chicago Diabetes Project (http://www.chicagodiabetesproject.org/
), Norwegian Cancer Society, European Commission EP-7 Beta Cell
Therapy, the Slovak Research and Development Agency under con-
tract no. APVV-51-033205 and Helse-Midt Norge. The Chicago Dia-
betes Project group is also acknowledged for discussions during
this work. Dorte Christiansen and Grethe Bergseth at the Depart-
ment of Laboratory Medicine, Nordland Hospital, are thanked for
skillful technical support.
Appendix A. Figures with essential colour discrimination
Certain figures in this article, particularly Figs. 5 and 6, are diffi-
cult to interpret in black and white. The full colour images can be
found in the on-line version, at doi:doi:10.1016/j.actbio.2011.03.
011.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.actbio.2011.03.011.
References
[1] De Vos P, Bucko M, Gemeiner P, Navratil M, Svitel J, Faas M, et al. Multiscale
requirements for bioencapsulation in medicine and biotechnology.
Biomaterials 2009;30:2559–70.
[2] Read TA, Stensvaag V, Vindenes H, Ulvestad E, Bjerkvig R, Thorsen F. Cells
encapsulated in alginate: a potential system for delivery of recombinant
proteins to malignant brain tumours. Int J Dev Neurosci 1999;17:653–63.
A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
2577
[3] Nilsson B, Korsgren O, Lambris JD, Ekdahl KN. Can cells and biomaterials in
therapeutic medicine be shielded from innate immune recognition? Trends
Immunol 2010;31:32–8.
[4] Mollnes TE, Brekke OL, Fung M, Fure H, Christiansen D, Bergseth G, et al.
Essential role of the C5a receptor in E. Coli-induced oxidative burst and
phagocytosis revealed by a novel lepirudin-based human whole blood model
of inflammation. Blood 2002;100:1869–77.
[5] Andersson J, Ekdahl KN, Larsson R, Nilsson UR, Nilsson B. C3 adsorbed to a
polymer surface can form an initiating alternative pathway convertase. J
Immunol 2002;168:5786–91.
[6] Mollnes TE, Lea T, Harboe M. Detection and quantification of the terminal C5b–
9 complex of human complement by a sensitive enzyme-linked
immunosorbent assay. Scand J Immunol 1984;20:157–66.
[7] Toda M, Kitazawa T, Hirata I, Hirano Y, Iwata H. Complement activation on
surfaces carrying amino groups. Biomaterials 2008;29:407–17.
[8] Andersson J, Ekdahl KN, Lambris JD, Nilsson B. Binding of C3 fragments on top
of adsorbed plasma proteins during complement activation on a model
biomaterial surface. Biomaterials 2005;26:1477–85.
[9] Arima Y, Kawagoe M, Furuta M, Toda M, Iwata H. Effect of swelling of poly(vinyl
alcohol) layers on complement activation. Biomaterials 2010;31:6926–33.
[10] Qi M, Strand BL, Morch Y, Lacik I, Wang Y, Salehi P, et al. Encapsulation of
human islets in novel inhomogeneous alginate-Ca
2+
/Ba
2+
microbeads: in vitro
and in vivo function. Artif Cells Blood Substit Immobil Biotechnol
2008;36:403–20.
[11] Qi M, Morch Y, Lacik I, Formo K, Marchese E, Wang Y, Wang S, et al. Prolonged
immunosupression-free survival of human islets inhomogeneous Ca
2+
/Ba
2+
-
alginate microbeads in diabetic immunocompetent mice. Unpublished work.
[12] Wang T, Lacik I, Brissova M, Anilkumar AV, Prokop A, Hunkeler D, et al. An
encapsulation system for the immunoisolation of pancreatic islets. Nat
Biotechnol 1997;15:358–62.
[13] Wang T, Adcock J, Kuhtreiber W, Qiang D, Salleng KJ, Trenary I, et al. Successful
allotransplantation of encapsulated islets in pancreatectomized canines for
diabetic management without the use of immunosuppression. Transplantation
2008;85:331–7.
[14] Lacik I, Brissova M, Anilkumar AV, Powers AC, Wang T. New capsule with
tailored properties for the encapsulation of living cells. J Biomed Mater Res
1998;39:52–60.
[15] King A, Sandler S, Andersson A. The effect of host factors and capsule
composition on the cellular overgrowth on implanted alginate capsules. J
Biomed Mater Res 2001;57:374–83.
[16] King A, Strand B, Rokstad AM, Kulseng B, Andersson A, Skjak-Braek G, et al.
Improvement of the biocompatibility of alginate/poly-L-lysine/alginate
microcapsules by the use of epimerized alginate as a coating. J Biomed
Mater Res A 2003;64:533–9.
[17] Strand BL, Ryan TL, In’t VP, Kulseng B, Rokstad AM, Skjak-Brek G, et al. Poly-L-
lysine induces fibrosis on alginate microcapsules via the induction of
cytokines. Cell Transplant 2001;10:263–75.
[18] Safley SA, Cui H, Cauffiel S, Tucker-Burden C, Weber CJ. Biocompatibility and
immune acceptance of adult porcine islets transplanted intraperitoneally in
diabetic NOD mice in calcium alginate poly-L-lysine microcapsules versus
barium alginate microcapsules without poly-L-lysine. J Diabetes Sci Technol
2008;2:760–7.
[19] Mollnes TE, Redl H, Hogasen K, Bengtsson A, Garred P, Speilberg L, et al.
Complement activation in septic baboons detected by neoepitope-specific
assays for C3b/iC3b/C3c, C5a and the terminal C5b–9 complement complex
(TCC). Clin Exp Immunol 1993;91:295–300.
[20] Strand BL, Morch YA, Espevik T, Skjak-Braek G. Visualization of alginate–poly-
L-lysine–alginate microcapsules by confocal laser scanning microscopy.
Biotechnol Bioeng 2003;82:386–94.
[21] De Vos P, De Haan BJ, Wolters GH, Strubbe JH, van Schilfgaarde R. Improved
biocompatibility but limited graft survival after purification of alginate for
microencapsulation of pancreatic islets. Diabetologia 1997;40:262–70.
[22] Anilkumar AV, Lacik I, Wang TG. A novel reactor for making uniform capsules.
Biotechnol Bioeng 2001;75:581–9.
[23] Lappegard KT, Fung M, Bergseth G, Riesenfeld J, Lambris JD, Videm V, et al.
Effect of complement inhibition and heparin coating on artificial surface-
induced leukocyte and platelet activation. Ann Thorac Surg 2004;77:932–41.
[24] Morch YA, Donati I, Strand BL, Skjak-Braek G. Effect of Ca
2+
,Ba
2+
, and Sr
2+
on
alginate microbeads. Biomacromolecules 2006;7:1471–80.
[25] Gros P, Milder FJ, Janssen BJ. Complement driven by conformational changes.
Nat Rev Immunol 2008;8:48–58.
[26] Tam SK, Dusseault J, Polizu S, Menard M, Halle JP, Yahia L. Physicochemical
model of alginate-poly-L-lysine microcapsules defined at the micrometric/
nanometric scale using ATR-FTIR, XPS, and ToF-SIMS. Biomaterials
2005;26:6950–61.
[27] McNally AK, Anderson JM. Complement C3 participation in monocyte
adhesion to different surfaces. Proc Natl Acad Sci USA 1994;91:10119–23.
[28] Wright SD, Weitz JI, Huang AJ, Levin SM, Silverstein SC, Loike JD. Complement
receptor type three (CD11b/CD18) of human polymorphonuclear leukocytes
recognizes fibrinogen. Proc Natl Acad Sci USA 1988;85:7734–8.
[29] Rinder CS, Smith MJ, Rinder HM, Cortright DN, Brodbeck RM, Krause JE, et al.
Leukocyte effects of C5a-receptor blockade during simulated extracorporeal
circulation. Ann Thorac Surg 2007;83:146–52.
[30] Schwartz-Albiez R, Adams Y, Von der Lieth CW, Mischnick P, Andrews KT,
Kirschfink M. Regioselectively modified sulfated cellulose as prospective drug
for treatment of malaria tropica. Glycoconj J 2007;24:57–65.
[31] Eisen V, Loveday C. Effect of cellulose sulphate on serum complement. Br J
Pharmacol 1970;39:831–3.
[32] Eisen V, Loveday C. In vivo effects of cellulose sulphate on plasma kininogen,
complement and inflammation. Br J Pharmacol 1971;42:383–91.
[33] Toso C, Mathe Z, Morel P, Oberholzer J, Bosco D, Sainz-Vidal D, et al. Effect of
microcapsule composition and short-term immunosuppression on intraportal
biocompatibility. Cell Transplant 2005;14:159–67.
[34] De Vos P, De Haan B, van Schilfgaarde R. Effect of the alginate composition on
the biocompatibility of alginate–polylysine microcapsules. Biomaterials
1997;18:273–8.
[35] Lappegard KT, Fung M, Bergseth G, Riesenfeld J, Mollnes TE. Artificial surface-
induced cytokine synthesis: effect of heparin coating and complement
inhibition. Ann Thorac Surg 2004;78:38–44.
[36] Lappegard KT, Bergseth G, Riesenfeld J, Pharo A, Magotti P, Lambris JD, et al.
The artificial surface-induced whole blood inflammatory reaction revealed by
increases in a series of chemokines and growth factors is largely complement
dependent. J Biomed Mater Res A 2008;87:129–35.
[37] Lappegard KT, Christiansen D, Pharo A, Thorgersen EB, Hellerud BC, Lindstad J,
et al. Human genetic deficiencies reveal the roles of complement in the
inflammatory network: lessons from nature. Proc Natl Acad Sci USA
2009;106:15861–6.
[38] Markiewski MM, Nilsson B, Ekdahl KN, Mollnes TE. Lambris JD: Complement
and coagulation: strangers or partners in crime? Trends Immunol
2007;28:184–92.
[39] Akalin HE, Fisher KA, Laleli Y. Caglar S: Bactericidal activity of ascitic fluid in
patients with nephrotic syndrome. Eur J Clin Invest 1985;15:138–40.
2578 A.M. Rokstad et al. / Acta Biomaterialia 7 (2011) 2566–2578
