Access to this full-text is provided by Wiley.
Content available from Advanced Functional Materials
This content is subject to copyright. Terms and conditions apply.
www.afm-journal.de
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (1 of 14)
Covalent Protein Immobilization on 3D-Printed Microfiber
Meshes for Guided Cartilage Regeneration
Madison J. Ainsworth, Oliver Lotz, Aaron Gilmour, Anyu Zhang, Michael J. Chen,
David R. McKenzie, Marcela M.M. Bilek, Jos Malda, Behnam Akhavan,*
and Miguel Castilho*
Current biomaterial-based strategies explored to treat articular cartilage defects
have failed to provide adequate physico-chemical cues in order to guide func-
tional tissue regeneration. Here, it is hypothesized that atmospheric-pressure
plasma (APPJ) treatment and melt electrowriting (MEW) will produce micro-
fiber support structures with covalently-immobilized transforming growth
factor beta-1 (TGFβ1) that can stimulate the generation of functional cartilage
tissue. The eect of APPJ operational speeds to activate MEW polycaprolactone
meshes for immobilization of TGFβ1 is first investigated and chondrogenic
dierentiation and neo-cartilage production are assessed in vitro. All APPJ
speeds test enhanced hydrophilicity of the meshes, with the slow treatment
speed having significantly less CC/CH and more COOH than the untreated
meshes. APPJ treatment increases TGFβ1 loading eciency. Additionally, in
vitro experiments highlight that APPJ-based TGFβ1 attachment to the scaf-
folds is more advantageous than direct supplementation within the medium.
After 28 days of culture, the group with immobilized TGFβ1 has significantly
increased compressive modulus (more than threefold) and higher glycosami-
noglycan production (more than fivefold) than when TGFβ1 is supplied through
the medium. These results demonstrate that APPJ activation allows reagent-
free, covalent immobilization of TGFβ1 on microfiber meshes and, importantly,
that the biofunctionalized meshes can stimulate neo-cartilage matrix formation.
This opens new perspectives for guided tissue regeneration.
DOI: 10.1002/adfm.202206583
M. J. Ainsworth, J. Malda, M. Castilho
Regenerative Medicine Centre Utrecht
University Medical Center Utrecht
Utrecht, The Netherlands
E-mail: m.dias.castilho@tue.nl
M. J. Ainsworth, J. Malda, M. Castilho
Department of Orthopedics
University Medical Center Utrecht
Utrecht, The Netherlands
1. Introduction
Articular cartilage is a hyaline tissue cov-
ering the end of articulating joints per-
forming primarily a mechano-protective
function; however, this tissue has lim-
ited regenerative capacity. Being a high
and regular load bearing tissue, it suers
from age and lifestyle dependent degen-
eration.[1,2] Current clinical options for
treating articular cartilage defects include
marrow stimulation through microfrac-
ture of the subchondral bone and cell-
based strategies, in particular autologous
chondrocyte implantation (ACI)[3,4] and
matrix-assisted autologous chondrocyte
implantation (MACI).[5] Cell-based regen-
erative strategies have been shown to out-
perform the gold standard microfracture
treatment, but they often still result in
the formation of a temporary (fibro)carti-
laginous tissue that does not possess the
same characteristics as the original native
cartilage and lacks durable stability.[6] As
the new tissue is low in cartilage specific
components, like collagen type II and
ReseaRch aRticle
O. Lotz, A. Gilmour, A. Zhang, M. M.M. Bilek, B. Akhavan
School of Biomedical Engineering
University of Sydney
Sydney , Australia
E-mail: behnam.akhavan@sydney.edu.au
O. Lotz, M. M.M. Bilek
School of Aerospace, Mechanical & Mechatronic Engineering
The University of Sydney
Sydney , Australia
O. Lotz, A. Gilmour, D. R. McKenzie, M. M.M. Bilek, B. Akhavan
School of Physics
University of Sydney
Sydney , Australia
A. Gilmour, M. M.M. Bilek
Charles Perkins Centre
University of Sydney
Sydney , Australia
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/./adfm..
© The Authors. Advanced Functional Materials published by
Wiley-VCH GmbH. This is an open access article under the terms of
the Creative Commons Attribution-NonCommercial-NoDerivs License,
which permits use and distribution in any medium, provided the original
work is properly cited, the use is non-commercial and no modifications
or adaptations are made.
Adv. Funct. Mater. 2023, 33,
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (2 of 14)
proteoglycans that lack the density and zonal organization of
native cartilage, it is much weaker and only provides limited
mechanical resilience. This inevitably leads to mechanical
failure and problems for the patient in the longer term. There-
fore, load-carrying ability together with adequate recruitment
and control of dierentiation to cartilage-specific cells need to
be critically addressed to provide a long-lasting, regenerative
solution for the treatment of articular cartilage damage.
Biodegradable support structures have been explored to
enhance the stability of cartilage implants, but due to the
mechanically challenging environment of the native tissue,
relatively large amounts of biomaterial are used, leaving lim-
ited space for infiltrating cells and the induction of a regenera-
tive response. Alternative implants based on cell-laden hydro-
gels reinforced with 3D-printed fibers obtained by melt elec-
trowriting (MEW), have recently shown great potential for the
fabrication of mechanically resilient constructs using a lower
volume of support biomaterial.[7–9] Interestingly, the use of only
small percentages (7%) of highly organized, micrometric sized
fibers (made of poly-ε-caprolactone (PCL)) resulted in mechan-
ical characteristics that approximated those of native cartilage.[7]
Although these constructs can provide enhanced support, their
limited capacity to promote tissue growth and uniform matrix
deposition has limited their long-term biomechanical stability.
It is known that cytokines, like transforming growth factor
beta (TGFβs), influence chondrogenic dierentiation from the
early to the final stages of development, play a role in quies-
cent chondrocyte maintenance, and prevent hypertrophy.[10–12]
Therefore, strategies that allow the immobilization of TGFβ to
the surface of MEW PCL polymers can open new perspectives
to support and guide seeded chondrocytes’ dierentiation and
subsequent cartilage tissue formation. Over the last few dec-
ades, extensive research has been performed on wet chemistry
approaches for covalent biomolecule attachment to biomate-
rials. Despite promising results, they share a number of short-
comings that make them unsuitable for regenerative medicine,
and in particular cartilage regeneration, such as long reaction
times, variable yields, side-reactions, and reagent toxicity.[13–15]
Similarly, biomolecules that are weakly bound to polymer sur-
faces through physical adsorption can be detached too quickly,
due to erosion by fluid flow, or displacement through protein
exchange.[15–17] Previously, latent TGF complexes were attached
to electrospun fibers using low pressure ammonia plasma to
elicit a chondrogenic response.[18] The approach was limited by
the lack of control over biomolecule placement and damage to
the fibers was not thoroughly investigated. Plasma is a complex
fluid with multiple species, often described as a conductive gas
thanks to its free electrons.[19] When the ions and electrons are
out of thermal equilibrium, the plasma is described as ‘cold’,
and typically forms a ‘glow discharge’. Recently, a new strategy
for the covalent attachment of biomolecules, based on atmos-
pheric pressure plasma (APP), has been introduced.[20] APP
devices typically use non-thermal-equilibrium plasma to con-
duct a variety of surface modifications, such as etching and ster-
ilization within a matter of seconds.[21–23] This is of benefit to
softer substrates that are not resistant to thermal stress. When
gas flowing through capillaries is ionized by electric fields these
devices are referred to as APP jets (APPJs).[21–23] Results have
shown that dielectric barrier discharge APP systems[20] and
APPJs[24] can be used to activate 2D polymeric surfaces to facili-
tate on-contact covalent immobilization of extracellular matrix
proteins in a single-step, reagent-free process. Furthermore, the
potential of APPJs to functionalize 3D polymeric meshes and
subsequently covalently immobilize biomolecules without the
need for wet chemistry processes has yet to be demonstrated.
Here, we hypothesize that APPJs can be used as a biofunc-
tionalization tool to enable covalent immobilization of TGFβ1
(TGF) biomolecule on 3D microfiber meshes for improved cell
interaction and guided cartilage regeneration (Figure 1). To test
this hypothesis, 3D microfiber meshes of medical grade PCL
were generated by melt electrowriting and subsequently acti-
vated by APPJ treatment for TGF covalent immobilization. The
eects of APPJ treatment on microfiber meshes were system-
atically studied for surface activation, load-carrying ability prop-
erties, and protein immobilization. Further, to determine the
potential of immobilized TGF to guide chondrogenic dieren-
tiation, an in silico chondrogenesis model was used to comple-
ment the subsequent in vitro experiments. These biofunction-
alized MEW meshes were seeded with mesenchymal stromal
cells (MSCs) and cultured in vitro over 28 days in both basal
and chondrogenic medium conditions to assess the activity of
the immobilized TGF.
2. Results
2.1. Surface Chemistry and Morphology of APPJ-Treated MEW
Microfiber Meshes
Three APPJ treatment speeds were investigated to determine
the speed at which the optimal balance between characteristic
surface modification and structural deformation is acquired.
M. J. Chen
School of Mathematical Sciences
The University of Adelaide
Adelaide , Australia
M. M.M. Bilek, B. Akhavan
Sydney Nano Institute
University of Sydney
Sydney , Australia
J. Malda
Department of Clinical Sciences
Faculty of Veterinary Medicine
Utrecht University
Utrecht, The Netherlands
B. Akhavan
School of Engineering
University of Newcastle
Callaghan, New South Wales , Australia
B. Akhavan
Hunter Medical Research Institute (HMRI)
New Lambton Heights, New South Wales , Australia
M. Castilho
Department of Biomedical Engineering
Technical University of Eindhoven
Eindhoven, The Netherlands
M. Castilho
Institute for Complex Molecular Systems
Eindhoven University of Technology
MB Eindhoven, The Netherlands
Adv. Funct. Mater. 2023, 33,
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (3 of 14)
These speeds were labeled slow (2.50 m min−1), medium
(3.05 m min−1), and fast (3.60 m min−1). The atomic concen-
trations of carbon and oxygen on meshes did not change
with APPJ treatment speeds (i.e., slow, medium, and fast)
at the resolution of the XPS survey spectra (Figure 2a,c). For
example, the average carbon atomic concentrations for the
untreated and slow APPJ-treated samples were 75.8±0.3% and
76.6±0.5%, respectively, and those of oxygen were 24.0±0.2%
and 23.4±0.5%. Note that hydrogen is not detectable by XPS,
because it has no core electrons. No significant amount of
nitrogen was detected on either untreated or APPJ-treated
meshes. The C1s high-resolution spectra were fitted with com-
pound peaks at binding energies of 284.6 ± 0.5, 286.5 ± 0.5,
287.6± 0.5, and 289.0 ± 0.5 eV, corresponding to CC/CH,
CO, CO, and COOH groups, respectively (Figure2b). The
concentration of COOH increased with all APPJ-treatment
speeds in comparison to the untreated meshes (Figure 2d).
The slow APPJ treatment speed had significantly less CC/
CH and more COOH than the untreated sample (p< 0.05),
while the dierence was not statistically significant for the fast
or medium APPJ treatment speeds. These changes in surface
chemistry are attributed to the dierent time durations for
which the MEW microfibers were exposed to plasma. Slower
treatment corresponded to longer exposure times, and therefore
more time for interactions between the surfaces and reactive
species in the plasma. Moreover, the increase of oxygen-con-
taining groups with the increase in treatment time indicates
surface oxidation as a result of APPJ treatment. The degree of
oxidation may have been small enough as to not be detected by
the lower resolution survey scans (Figure2a), but still be detect-
able by the higher resolution C1s scans (Figure2b) as a result of
the signal to noise ratio increase.[25] Variations observed in the
COOH concentration for treated samples are within the experi-
mental error of XPS measurements, in particular for small spot
size measurements performed on thin fibers. This variability
does not aect the conclusions made from the XPS results.
To further understand the eect of the APPJ treatment
speeds on the surface oxidation and hydrophilicity of MEW
meshes, the dynamic wettability was characterized (Figure2e).
All treated conditions displayed increased hydrophilicity com-
pared to the untreated, as demonstrated by water drops being
drawn into the meshes. The drops placed on fast APPJ-treated
meshes tended to remain outside the mesh (≈15 s) before
being drawn in gradually (30—60s). Therefore, after 60s from
when the drop was originally placed there tended to be a small
amount left, that disappeared by 180s. However, those on the
slow APPJ-treated samples tended to be drawn in earlier (<5s)
and quicker (<30s), but not always drawn in completely. This
meant that despite the speed with which the drops were drawn
in, there was often a proportion of the original drop left on the
mesh’s surface after 180s. The drops placed on medium APPJ-
treated samples tended to be drawn in either quickly or gradu-
ally. However, the medium APPJ-treated samples were unlike
the slow in that the whole drop tended to be drawn into the
mesh. The images selected for the medium APPJ-treated condi-
tion display a drop drawn in quickly (<30s). These qualitative
results indicate that, independent of the APPJ treatment speed,
hydrophilicity of the MEW meshes was substantially increased.
In addition, the eect of APPJ treatment speeds on MEW
meshes’ morphology and mechanical performance was studied.
Morphological changes from the untreated condition were
observed only on the slow-APPJ-treated sample (Figure 3a;
FigureS2, Supporting Information). The surface of the fibers
had increased roughness and some fibers displayed fusing.
Moreover, these changes appeared to decrease in severity
toward the core of the meshes, moving down in the out-of-
plane printing direction (i.e., direction of laid-down fibers).
Importantly, no significant dierences were observed between
the tensile moduli of the MEW meshes APPJ-treated at dif-
ferent speeds, nor between those and the MEW meshes that
were not exposed to APPJ (Figure3b). This indicates that the
morphological changes induced by APPJ seem to not result in
a significant alteration of the mechanical performance of the
meshes under uniaxial tensile loading.
2.2. TGF Immobilization and Quantification
Confirmation and quantification of TGF immobilization was
subsequently investigated using immunofluorescent imaging
and ELISA. TGF retention on microfibers was additionally
investigated using ELISA. Immunofluorescent detection of
Adv. Funct. Mater. 2023, 33,
Figure 1. Schematic illustration showing the rationale of the study and covalent immobilization of TGF on PCL MEW microfibers. a) Depiction of PCL
MEW microfiber mesh without alteration. b) Representation of MEW mesh following activation with APPJ to add reactive sites, c) Functionalized MEW
mesh with TGF covalently immobilized to microfiber surfaces.
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (4 of 14)
Adv. Funct. Mater. 2023, 33,
Figure 2. Eect of APPJ treatment on surface chemistry of PCL MEW microfiber meshes. a) XPS survey spectra taken after treatment with various APPJ
treatment speeds. c) Curve fitted XPS Cs high resolution spectra obtained for samples prepared using various APPJ treatment speeds. The spectra are
curve fitted by four components: C, CC/CH; C, CO; C, CO; C, COOH. The dashed curve indicates fit. c) Carbon I, oxygen (O), and nitrogen
(N) atomic concentrations for samples treated with various APPJ speeds. d) Atomic concentration of peak-fitted C s components for various APPJ
treatment speeds. APPJ treatment increased the prevalence of CO, CO, and COOH groups. XPS data; n= (x technical replicates). Data (c,d)
presented as mean±SD. Paired T test for significance with p-values≤.=*. e) Selected images of the dynamic wettability of meshes after various
APPJ treatment speeds at various times after drop placement. Water drops were not drawn into the untreated meshes, but were drawn into the treated
meshes. n≥ (x technical replicates). Representative images shown. Scale bar (bottom right)=mm.
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (5 of 14)
TGF confirmed the presence of biomolecule attachment to
MEW microfiber meshes (Figure 4a). Higher signal intensi-
ties are correlated with higher protein immobilization concen-
trations. Markedly, there was a significant increase in the total
pixel value of the APPJ treated meshes when compared to the
untreated ones. This increase was ≈44-fold for the 10ngmL−1
and almost twofold for the 1000ngmL−1 group. Protein within
the MEW mesh could also be detected for the untreated group
immersed in 1000 ngmL−1. This can be potentially related to
the physical adsorption of TGF protein to the hydrophobic
MEW fibers. It is expected that this physically absorbed protein
would be removed with a more rigorous washing protocol.
An ELISA assay was used to quantitatively investigate the
presence of various TGF concentrations on PCL MEW micro-
fiber meshes (Figure 4b). The higher the TGF concentration,
the higher the signal for all APPJ-treated conditions. Further,
the ELISA signal for APPJ-treated conditions with 2000, 1000,
and 500ngmL−1 TGF were all significantly greater than that for
the untreated condition even when the untreated sample was
incubated with the highest concentration of the protein solu-
tion, 2000ngmL−1 (p≤ 0.0001). The untreated conditions did
not display a signal significantly dierent from background
(p∼1). This equivalence to background indicates that the TGF
signals observed in treated conditions corresponded to cova-
lently bound TGF. Moreover, no significant dierence was
observed in ELISA signals obtained for the meshes APPJ-
treated at various speeds and incubated with the same TGF
concentration. This lack of variation with exposure time indi-
cates that the exposure time to APPJ had no significant influ-
ence on the ecacy of covalent immobilization. Although the
slower speed resulted in higher treatment intensity, as indi-
cated by surface chemistry, morphology, and wettability results
(Figures2 and 3), the eect of APPJ speed may not have been
large enough to be detected by the ELISA experiment. Alter-
natively, the density of reactive sites imparted in the weakest
treatment may have been high enough that the amount of pro-
tein bound was limited by its diusion to the surface from the
solution rather than by availability of reactive sites. Based on
the improved hydrophilic behavior, only the slow-APPJ-treated
meshes were continued for further evaluation.
Further ELISAs were conducted to determine the immobili-
zation eciency and elution of TGF on treated and untreated
scaolds. Significantly lower amounts of TGF were found to
remain in the solutions after incubation with APPJ-treated
meshes compared to those incubated with untreated meshes
(Figure 4c). The loading eciency for untreated meshes
was calculated to be 52.4 ± 3.6%, while that of the APPJ-
treated meshes was 82.1 ± 0.1%. The average density of TGF
was calculated to be 0.75 ± 0.05 ng mm−2 on untreated and
1.182±0.001 ng mm−2 on APPJ-treated meshes. ELISA meas-
urements of TGF on scaolds underscore the observation that
APPJ-treated meshes loaded significantly more TGF after incu-
bation than the untreated ones (Figure 4d). In addition, less
TGF was released from APPJ-treated meshes than untreated
meshes over a 28 day period with multiple changes of media
(Figure4c). Approximately 55.4± 3.1% of the TGF loaded onto
untreated meshes was retained after the 28 days, whereas,
81.9 ± 1.9% was retained on the APPJ-treated meshes. The
results demonstrate that the APPJ-treated scaolds increase
TGF loading eciency and retention, and therefore, APPJ-
treated meshes provide the corresponding biochemical cues at
higher rates and for longer periods of time.
2.3. In Vitro Chondrogenic Dierentiation and Neo-Cartilage
Formation in Surface Activated MEW Meshes
Activity of immobilized TGF was investigated using in vitro
experimentation. The −APPJ +TGF condition under the fol-
lowing evaluation was considered the gold standard approach
for chondrogenic dierentiation of MSCs in vitro that entails
supply of freshly thawed TGF to the cells at each media change.
Adv. Funct. Mater. 2023, 33,
Figure 3. Eect of APPJ treatment on PCL MEW microfiber mesh morphology. a) SEM images of meshes after various APPJ treatment speeds (SE
detector). n=. Scale bars=µm. b) Uniaxial tensile test results of MEW meshes without APPJ treatment and after fast-, medium-, and slow-speed
APPJ treatment. Data displayed as mean±SD. n=. One-way ANOVA with Tukey’s multiple comparisons revealed no statistical dierence (p>.).
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (6 of 14)
Preliminary in vitro optimization of protein immobilization
and subsequent MSC dierentiation and neo-cartilage tissue
formation was performed with varying concentration of protein
immersion solutions of TGF (Figure 5). In vitro results revealed
that lower concentrations of immobilized TGF (<1000ng ml−1)
did not promote suciently observed cellular infiltration and
homogenous neo-cartilage matrix production (Figure 5c),
despite the lack of significance in normalized GAG production
between treatment groups (Figure S3, Supporting Informa-
tion). Hence, the slow APPJ treatment speed in combination
with the 1000 ng mL−1 concentration for immobilization was
used for the forthcoming in vitro analysis.
It was found that the groups with TGF (whether immobi-
lized on the surface or dissolved in the medium), had consist-
Adv. Funct. Mater. 2023, 33,
Figure 4. TGF immobilization and quantification. a) Counted total pixel values to represent fluorescence intensity with accompanying immunofluo-
rescent staining of immobilized TGF ( and ngmL−) on APPJ-functionalized MEW microfiber meshes after mild detergent washing. Scale
bar=µm. n=. Data presented as mean±SD. b) Quantification of protein attachment to MEW microfiber meshes using ELISA after rigorous
detergent washing. n=. Data presented as mean±SD. c) Concentration of TGF eluted from untreated and APPJ-treated meshes after TGF immo-
bilization by incubation in ngmL− TGF containing solution. The measurement for day zero () is the residual TGF remaining in the incubation
solution after immobilization. n=. Data presented as mean±SD. d) ELISA (absorbance at nm) of immobilized TGF retained on meshes with and
without TGF immediately after incubation () and after days. The buer solution was changed every days. n=. Data presented as mean±SD.
Bonferroni’s post hoc test for significance. (ns) p>., (*) p≤., (**) p≤., (***) p≤., (****) p≤..
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (7 of 14)
ently higher metabolic activity in comparison to the groups
without TGF supplied (Figure 6a), noting that the readings
were not normalized against DNA content in order to make
the measurements on the same samples throughout the cul-
ture period. In addition, glycosaminoglycan (GAG) production
was observed to significantly increase following 28 days in cul-
ture in the +APPJ +TGF group and was considerably elevated
at the end of the culture period compared to the other groups
(Figure 6b). Clear trends of increasing GAG production were
observed throughout the culture period in the −APPJ +TGF
group. Treatment groups without immobilization or supple-
mentation in culture medium of TGF displayed GAG produc-
tion with more variability and less consistency than the other
groups over the culture period (Figure6b).
To confirm neo-cartilage ECM deposition, mechanical prop-
erties of cultured meshes were characterized under uniaxial
compression loading (Figure 6c). Compression tests carried
out on cultured meshes on day 28 of culture revealed similar
Adv. Funct. Mater. 2023, 33,
Figure 5. Eect of TGF protein supplementation strategy and concentration on chondrogenic dierentiation of seeded MSCs and neo-cartilage forma-
tion. Schematic with representative in vitro histological analysis of cartilage-like matrix deposition. a) Treatment groups with varied concentrations of
immobilized TGF to the APPJ-functionalized surfaces. b) Group activated with APPJ treatment, but without TGF supplementation. c) Group without
functionalization, but with TGF supplied x per week in the culture medium at ngmL−. d) Group without APPJ functionalization nor TGF supple-
mentation. Blue lines indicate MEW fibers, green shapes indicate TGF, and yellow shapes indicate reactive sites. Dashed circles indicate location of
displaced pellet-like structures. Arrows indicate MEW microfibers dissolved during histological processing. Scale bars=mm. n=.
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (8 of 14)
stress–strain behavior between groups cultured in the presence
of TGF (Figure 6c), and statistically (p< 0.05) a higher com-
pressive modulus in the +APPJ +TGF group compared to the
other three groups (Figure 6d). This finding confirmed the
higher density of neo-cartilage ECM components deposited by
the seeded MSCs in this group of cultured meshes. Likewise,
brightfield imaging carried out progressively on meshes in
culture, demonstrated the higher density and more uniform
matrix production in the +APPJ +TGF group (Figure6e).
Furthermore, histological analysis of 3D-cultured meshes
was performed to detect deposition of cartilage components
(Figure 7). The treatment groups without TGF supplementa-
tion (+APPJ −TGF, −APPJ −TGF) showed minimal evidence
of neo-cartilage matrix production over the culture period
(Figure S4, Supporting Information). In contrast, 3D MEW
meshes supplemented with TGF (immobilized or in medium)
confirmed that these groups were able to support neo-cartilage
matrix production. Notably, the treatment group with immo-
bilized TGF (+APPJ +TGF) exhibited production of collagen
type II and GAGs following 1 week in culture (D7), whereas
these ECM components do not appear in the sections of the
TGF supplemented in medium group (−APPJ +TGF) until the
second week of culture (Figure7). Moreover, at D28 of in vitro
culture, the +APPJ +TGF group exhibited a higher density of
GAGs produced within the sample in comparison to the −APPJ
+TGF group. Collagen type I was also detected, but was less
prominent.
Quantitative PCR analysis revealed an increased fold change
in the +APPJ +TGF and −APPJ +TGF groups in the day 11 and
14 timepoints for COL1A1, COL2A1, SOX9, and ACAN, in com-
parison to the other treatment groups (Figure 8). This finding
can be associated with the increase in matrix production
(GAGs, collagen I and II) observed most prominently in the
+APPJ +TGF group, as well as less pronounced in the −APPJ
+TGF group (Figure 7). In addition, the +APPJ +TGF group
exhibited a trend of upregulation across the COL2A1, SOX9,
and ACAN genes from D1 until D14, and subsequently dropped
at D28. COLXA1 was initially upregulated in all groups at D1,
however did not exhibit fold change throughout the remainder
of the experimental period. The chondrogenic index (COL2A1/
COL1A1) was calculated using the treatment group means and
revealed that at day 14, the +APPJ +TGF group had a chondro-
genic index of 2.32, while the −APPJ +TGF group’s chondro-
genic index was 0.96.
3. Discussion
With cell-laden approaches for cartilage regeneration, it is
imperative that the biomaterials used are able to promote depo-
sition of tissue-specific ECM components and simultaneously
ensure sucient load-carrying ability. MSC dierentiation into
cartilage cells and subsequent ECM matrix deposition can be
facilitated by endogenous growth factors, such as TGF;[10,26]
Adv. Funct. Mater. 2023, 33,
Figure 6. MSCs seeded in biofunctionalized MEW microfiber meshes with TGF undergo chondrogenic dierentiation and support neo-cartilage
formation. a) Progressive metabolic activity of cultured meshes using resazurin assay, n=. Data presented as mean±SD. b) GAG production, nor-
malized against DNA quantity, of meshes throughout the culture period, n=. Data presented as mean±SD. Two-way ANOVA with Tukey’s multiple
comparisons for significance. c) Representative engineering stress–strain curves with d) compressive modulus determined in the physiological native
articular cartilage strain region n= (individual values plotted with mean). Two-way ANOVA with Tukey’s multiple comparisons for significance. e)
Accompanying brightfield images of cultured meshes on D of in vitro culture (scale bar = µm). (ns) p>., (*) p≤., (**) p≤ ., (***)
p≤., (****) p≤..
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (9 of 14)
while load-carrying ability of cell-laden biomaterials can be pro-
vided by well-organized fiber reinforcing meshes obtained by
MEW.[7,9] Here, we discuss our findings for covalently attaching
TGF on MEW fiber meshes using APPJ, and demonstration of
the potential for this biomaterial system to support and pro-
mote neo-cartilage ECM deposition.
For controlling TGF attachment on the surface of MEW
fiber meshes, slow APPJ-treatment speed, and correspondingly
longer treatment times, was found to be beneficial. Despite a
high intra-group variability, this treatment speed resulted in
increased dynamic wettability and hydrophilicity when com-
pared to medium and fast speed APPJ-treatments. These
findings align with previous works involving APP treatment
of substrates composed of PCL[27,28] and other polymers.[13,29]
The oxygen containing groups (CO, CO, and COOH), have
previously been observed to increase on compressed, solvent-
casted, and solution electrospun PCL substrates treated with
atmospheric pressure[27,28,30] and low pressure plasmas using
gas mixtures with[27,28,31] and without[27,28,30] oxygen. However,
the proportions of species introduced vary across studies, which
is likely due to variations in treatment parameters.
In addition, the variable behavior of drops within the
medium APPJ treatment condition may indicate a transition
state between the dynamic wettabilities of the fast and slow
APPJ treatment conditions. Typically, hydrophilic enhance-
ment is due to the introduction of polar or radical groups from
treatment.[21,30,32,33] Further variability in dynamic wettability
and contact angles could be a result of factors such as dier-
ences in exact drop placement position in relation to the MEW
mesh (i.e., pore size and fiber interconnections), from drops
being in either the Wenzel or Cassie-Baxter state[34] and from
fiber roughness.[35,36] The importance of optimizing between
treatment intensity and the lack of deformation has been dis-
cussed previously.[37,38] One of the most important factors for
the transfer of thermal energy from APPJ plumes is the dis-
tance to the substrate.[39] For the purposes of this study, maxi-
mizing treatment intensity, with slower treatment speeds, was
more important than minimizing morphological changes. In
fact, morphological changes on MEW meshes had little or no
eect on microfiber mesh load-carrying ability as demonstrated
by the uniaxial tensile test.
When the surfaces of MEW PCL fibers were treated at slower
APPJ speed, a uniform distribution of protein was observed
throughout the mesh. This may be a result of increased fiber
hydrophilicity and potential CO bond breakage[25] that allowed
for rapid protein immobilization via covalent attachment of
TGF to hydroxyl functional groups at the PCL surface after
APPJ treatment. For the medium and fast APPJ treatments,
the PCL fiber surface had fewer O-containing groups present
and thus less hydrophilic behavior, likely leading to a lower
protein immobilization density. In addition, the enhanced
hydrophilicity resulted in increased solution and cellular
Adv. Funct. Mater. 2023, 33,
Figure 7. Histological analysis of neo-cartilage components (collagen types I and II, safranin-O) produced in biofunctionalized (+APPJ +TGF) and
media-supplemented (−APPJ +TGF) MEW microfiber meshes throughout days of culture. Scale bars=µm. Representative image shown. n=.
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (10 of 14)
infiltration into the APPJ-treated meshes measured using real-
time-observation, leading to a more ecient protein loading
and cell seeding process. As such, an approximately uniform
distribution of biochemical signaling and cell response was
observed throughout treated scaolds.
Interestingly, it was observed that the meshes displayed higher
protein signals with ELISA when measured 28 days after initial
TGF incubation, including buer changes every 7 days, than
when measured on day 0 (p<0.01 for the −APPJ +TGF group).
These increased signals for both conditions may be a result of
multiple factors aecting the interactions between the TGF and
the antibodies used for the ELISA. First, there may be increased
physisorption of TGF previously released from the surfaces as
a result of PCL swelling and degradation with time in solution.
During the development of the protocol for the first ELISA, it
was observed that protein adhesion on PCL meshes after 24 h
incubation at 4 °C was significantly enhanced than that after
2h at 24°C. Typically, these incubation protocols are considered
equivalent immobilization alternatives. PCL is known to be a bio-
degradable polymer,[40,41] experiencing roughening and erosion
in water,[42] and possible swelling.[43] Therefore, it is plausible to
suggest that the PCL may be degrading with the extra time in
solution, presenting new surfaces and cavities that could adsorb
and/or trap protein molecules from solution, including the TGF
and ELISA antibodies. Second, the hydrophilicity of the meshes
would aect the rates of infiltration into the meshes for biomol-
ecules and the solutions, as well as encourage dierent presen-
tations of the TGF as a result of dierent electrostatic interac-
tions with the surface. High variability in proteoglycan and GAG
production was observed in vitro for untreated meshes, con-
sistent with high variability in their protein loading eciency.
Future experiments should investigate the consistency of protein
loading eciency and bioactivity over time in order to ensure a
reliable process for end-use applications.
While the concentration of immobilized TGF is supra-physi-
ological, the APPJ-treated scaolds with no TGF (+APPJ −TGF)
Adv. Funct. Mater. 2023, 33,
Figure 8. Relative gene fold expression of chondrogenic genes against HPRT (HKG) of MSCs seeded in MEW meshes throughout days of culture.
a) Collagen type , b) collagen type , c) SOX, d) collagen type , and e) aggrecan. Data presented as mean with individual values plotted. Error bars
represent standard deviation. n=.
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (11 of 14)
Adv. Funct. Mater. 2023, 33,
or concentrations lower than 1000ngmL−1 revealed pellet-like
formations of densely packed cells not integrated within the
mesh, which exhibited high proteoglycan deposition. This for-
mation can be explained by cellular density (and in turn, MSC
stiness that can influence chondrogenesis) in combination
with the chondrogenic factors within the growth medium.[44]
Independently of the protein attachment eciency, the
immobilized TGF on MEW meshes was revealed to be active
throughout the 28 days of in vitro cell culture, as demonstrated
by its capacity to promote MSC dierentiation and deposition of
cartilage specific ECM components, like GAGs and Type II col-
lagen. This is a notable achievement for stably coupling endog-
enous TGF, and potentially other biomolecules, to 3D-printed
polymeric structures through a simple, reagent-free process.
Previous works in which chemical functional groups, such as
NH2 or COOH, are created on other polymeric materials using
depositing processes have demonstrated increases in cell adhe-
sion[45,46] and the promotion of osteogenesis.[29,46] However, no
APPJ has previously been used in combination with 3D micro-
fiber meshes, nor to promote chondrogenesis with the immobi-
lization of biomolecules such as TGF, nor to covalently immobi-
lize biomolecules without reagents or wet chemistry processes.
Future works will require further optimization of TGF con-
centrations or even combination of covalently incorporated bio-
logical cues like TGF, or other cytokines, with physical stimula-
tion, e.g., by exposing meshes to mechanical loading during in
vitro culture as it is known to boost endogenous TGF produc-
tion by chondrocytes.[7,47] To further validate the findings of this
study, an appropriate in vivo model should be implemented to
test the hypothesis, noting the need for optimization of post-
surgical mesh integration. Importantly, our findings on the
covalent immobilization of TGF biomolecules to high resolu-
tion fiber meshes enable us to avoid the limitation of using
super physiological amounts of growth factors during in vitro
cartilage culture, that would incur high costs and adverse reac-
tions, and finally overcome the limitations of wet chemistry
surface treatment strategies and existing plasma technologies.
4. Conclusion
This study demonstrates a novel strategy for the covalent
attachment of biomolecules like TGF to 3D-printed polymeric
meshes using an APPJ surface treatment. Without dimin-
ishing the load-carrying ability of the microfiber mesh and with
an increased protein loading eciency, the APPJ treatment
allowed for subsequent, single-step, reagent-free, covalent func-
tionalization of microfiber meshes with TGF. The covalently
immobilized TGF retained its bioactivity and induced increased
rates of chondrogenic dierentiation and neo-cartilage matrix
production, compared to a standard protein-in-medium
approach. These findings not only have relevance in the field of
cartilage regeneration, but also open new perspectives for the
design of cell-free, protein-functionalized materials for guided
tissue regeneration. The results presented in this paper have
important implications, including targeted cellular dierentia-
tion and enhanced bioactivity, to the rapidly evolving fields of
biofabrication and tissue engineering where the APPJ can be
used as a powerful and versatile tool.
5. Experimental Section
Microfiber Mesh Fabrication by Melt Electrowriting: MEW was
executed using a DDiscoveryTM device (RegenHU, Switzerland). Poly-
ε-caprolactone (PCL; PURASORB PC, Corbion, the Netherlands)
was melted to °C for h prior to MEW processing. PCL microfiber
meshes were fabricated using the following MEW parameters: high
voltage .kV, nozzle size G, air pressure . bar, collector velocity
mms−, collector distance mm. Square lattice-patterned meshes were
fabricated in × mm sheets with fiber spacing of and µm
(for in vitro/characterization experiments and immunofluorescent
detection, respectively). Meshes used for in vitro experiments were
sterilized prior to protein immobilization in % ethanol for min and
then treated with UV light for min per side.
Atmospheric Pressure Plasma Jet (APPJ) Functionalization of MEW
Meshes: APPJ treatment settings were used as described previously,[]
with the following modifications: The APPJ was mounted in a D printer
(FlSun i Prusa) modified in-house (FigureS, Supporting Information).
Marlin firmware was edited and uploaded with Arduino. The printer was
operated with Repetier software, and codes were made in-house using
Matlab. The resonance frequency for the system was . kHz. APPJ
treatment was conducted at three dierent speeds, slow (.mmin−),
medium (. m min−), or fast (. m min−). APPJ treatment was
conducted in parallel lines spaced mm apart. For all experiments,
treatment was conducted on both sides of the MEW meshes, using
forceps to transfer between sides. Where samples were rectangular,
APPJ treatment lines were parallel to the longer side.
X-Ray Photoelectron Spectroscopy (XPS): A Thermo ScientificTM
K-Alpha+TM spectrometer (Thermo Fisher Scientific, UK) was used for
XPS measurements; and Thermo Avantage software (version .,
Thermo Fisher Scientific, UK) was used for collecting data and analyzing
the data. The X-ray source was a monochromatic Al K-Alpha (.eV)
with . eV nominal operating voltage. Twenty scans were collected
for survey spectra with a step size of .eV; and for carbon (Cs) high-
resolution spectra with a step size of .eV. The spot size was µm.
Mesh samples with the size of mm×mm made of layers with
µm fiber spacing were used for these measurements. For the XPS
measurements, samples were mounted on D ×mm indium pieces
for assisting in z-axis precision. Samples were squashed with a metal
spatula and secured with carbon tape at each of the four corners, and
five measurements were taken per sample. Two samples were measured
per condition. Fitting of Cs high resolution spectra was conducted
using a linear background and a combination of Gaussian (%) and
Lorentzian (%) line shapes. The presented spectra are representative
for each condition. Errors displayed are standard errors of the mean
obtained from at least three data points, each corresponding to a single
measurement.
Dynamic Wettability Analysis: The dynamic wettability of APPJ-treated
samples (× mm mesh with layers and µm fiber spacing)
was investigated using µL water droplets from an Attension Theta
tensiometer (Biolin Scientific). Dynamic wettability was considered a
more eective measure of hydrophilicity for porous D scaolds than
measuring water contact angles.[,] Samples were recorded for at least
s after the droplets touched the surfaces. At least three drops were
measured per sample, and at least two samples were measured per
condition.
MEW Microfiber Morphology Analysis: PCL meshes (× mm mesh
with layers and µm fiber spacing) before and after APPJ treatment
(n=) were first coated with a thin layer of Au/Pd using a sputter coater
(SC Quorum). Scanning electron microscopy (SEM) images of
the coated PCL meshes were then obtained using a Phenom XL SEM
with the Secondary Electron (SE) detector. The chamber pressure was
kept below Pa, while an acceleration voltage of kV was applied at a
working distance of mm.
Covalent Protein Immobilization: Following APPJ functionalization of
MEW microfiber meshes, the meshes were submerged in TGF (TGFβ,
Peprotech, USA) in PBS solutions with dierent concentrations (, ,
, or ngmL−) for h at °C. and ngmL− concentrations
were used for immunofluorescent detection of immobilized TGF. All
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (12 of 14)
Adv. Funct. Mater. 2023, 33,
concentrations were used for the first enzyme linked immunosorbent
assay (ELISA) and in vitro optimization of TGF concentration for this
study. The ngmL− condition was used for the second ELISA and
the final in vitro experiments.
Immunofluorescence Detection of Immobilized Protein: Following APPJ
functionalization (fast speed) and subsequent immobilization ( and
ng mL−), MEW meshes were washed before species-specific
blocking was carried out using % normal goat serum for min.
Anti-TGF primary antibody incubation (:; Abcam) was carried
out overnight at °C. Secondary antibody incubation (:; Abcam)
proceeded the following day for h at °C in the dark. Tween
detergent (.%) was used to remove unbound antibodies from the
samples prior to mounting using Fluoromount G (SouthernBiotech).
Samples were imaged using a Leica Sp confocal microscope (n=).
Maximum projections of constant volumes (µm with µm voxel size)
were processed for visual depiction of TGF immobilization and single
images were used for the pixel value quantification. Integrated density
of grey-scale look-up table was measured using ImageJ to quantify
fluorescence intensity by calculating the total pixel value for each image.
The total pixel value of meshes that did not contain immobilized protein
was used to subtract any potential autofluorescence detected from the
PCL. Error bars shown are standard deviation of two samples.
ELISA Analysis for Quantification of Covalent Protein Immobilization:
Following APPJ functionalization, mm diameter MEW meshes were
punched and submersed for h at °C in dierent concentrations of
TGF (, , , , , or ngmL−). Untreated meshes were
submersed in or ng mL− only. Meshes were rinsed briefly in
PBS before washing with % (w/v) sodium dodecyl sulfate (SDS) for
h at °C on an orbital shaker ( rpm) to remove non-covalently
immobilized TGF. Meshes were then rinsed six times in fresh PBS
followed by blocking with % (w/v) BSA in PBS for min at °C.
Anti-TGF primary antibody (rabbit :; Abcam) incubation was carried
out for h at °C. Meshes were washed four times in wash buer
(.% Tween , .% BSA in PBS). Secondary antibody Goat Anti-rabbit
IgG HRP (:; Abcam) incubation was carried out for h at room
temperature. Meshes were washed six times in wash buer. Meshes were
then incubated with µL -Step Ultra TMB substrate (ThermoFisher
Scientific) in a clean well plate for min at room temperature in the
dark. Meshes were removed from wells prior to addition of µL M
HSO stop solution. Absorbance was read at nm using a microplate
reader (Tecan, infinite M). This protocol was developed in response
to issues creating false signals, namely topography changes due to PCL
swelling in solution and in response to SDS that revealed new unblocked
surfaces followed by non-specific physisorption of the antibodies on the
new unblocked surfaces. A single measurement was taken per sample,
and three samples were measured per condition, corresponding to
three dierent meshes. Errors shown are standard deviation using three
samples. The extent of TGF immobilization was statistically compared
with Bonferroni’s post hoc test (concentration and speed).
ELISA Analysis for Protein Loading Eciency and Elution: APPJ-treated
and untreated mm diameter MEW meshes were punched and
submersed for h at °C in µL of ,ngmL− TGF solution with
.% BSA. Two meshes were submerged per Eppendorf tube in order to
increase the signal strength. Three sets of meshes were measured per
condition. Meshes from the ‘ day’ condition were transferred to new
Eppendorf tubes after incubation and frozen at −°C for storage until
end point ELISA. Meshes from the ‘ day’ condition were transferred to
.% BSA in PBS solution and placed in incubator at °C. The buer
solution was changed after every days and eluted protein was collected.
All collected solutions were frozen at −°C for storage until end point
ELISA. Hundred microliters of ‘ day’ immobilization solution and , ,
, and day eluted protein solutions were incubated in an ELISA plate
(Immulon HB, ThermoFisher Scientific) overnight at °C. Meshes
and ELISA plate wells were rinsed two times with fresh PBS followed
by blocking with % (w/v) BSA in PBS for h at °C. All samples were
washed three times in wash buer (.% Tween , .% BSA in
PBS). Anti-TGF primary antibody (rabbit :; Abcam) incubation was
carried out for h at °C. All samples were washed three times in wash
buer. Secondary antibody Goat Anti-rabbit IgG HRP (:; Abcam)
incubation was carried out for h at room temperature. All samples
were washed three times in wash buer. Meshes were transferred to
empty wells of the ELISA plate, then µL -Step Ultra TMB substrate
(ThermoFisher Scientific) was added to all wells and incubated for
min at °C in the dark. Meshes were removed from wells prior
to addition of µL HSO stop solution. Absorbance was read
at nm using a microplate reader (Tecan, infinite M). Errors
displayed are standard errors of the mean obtained from three data
points. The extent of TGF loading eciency and elution were statistically
compared with Bonferroni’s post hoc test (time and treatment).
Protein Immobilization Density Calculation: Average protein density
was calculated by dividing the average TGF loaded onto meshes by the
surface area of those meshes. Mesh surface area was calculated by first
modeling the mesh as a collection of cylinders in a rectangular prism
arrangement, then applying a ‘square to circle’ conversion factor to
adjust the surface area to the cylindrical shape of the real meshes. Two
mm diameter meshes with layers, µm fiber spacing, and µm
fiber thickness diameter were used per data point. The conversion factor
was calculated to be . by dividing the area for a circle by that of a
square. It was assumed that one set of fibers in x and y were included
per layer, that % of the fibers are hidden by contact with other fibers
as part of the mesh layering,[] and that the fibers were smooth. Errors
described are standard errors of the mean obtained from three data
points.
Mesenchymal Stromal Cells (MSCs) Expansion and Seeding on MEW
Microfiber Meshes: Equine MSCs were isolated from bone marrow aspirate
of a healthy horse with approval from the local animal ethics committee
as previously described.[,] MSCs were thawed and expanded in
alpha-minimum essential medium (α-MEM, Life Technologies),
supplemented with % fetal bovine serum, U mL− penicillin,
µg mL− streptomycin, .m L-ascorbic acid--phosphate (ASAP)
and ng mL− basic fibroblast growth factor (bFGF, R&D Systems).
The treatment groups consisted of APPJ-functionalized meshes with
(+APPJ +TGF) and without immobilized TGF (+APPJ −TGF), as well
as untreated meshes with TGF supplemented in the medium (−APPJ
+TGF) and without TGF supplementation (−APPJ −TGF). Mesh discs
were biopsy punched (∅=mm), sterilized, and protein immobilized
where appropriate. All samples were submerged in MSC expansion
medium for h at °C prior to seeding. There were eight samples
from each treatment group per timepoint (, , , , and days of
culture) seeded with cells and one empty sample per timepoint to act
as a blank for biochemical assays. MSCs (passage ) were resuspended
.×cellsmL− in basic chondrogenic culture medium (Dulbecco’s
modified eagle medium (DMEM, Life Technologies), supplemented with
U mL− penicillin, µg mL− streptomycin, .m ASAP, :
ITS premix (Corning), and ng mL− Dexamethasone (Sigma)). The
concentrated cell suspension was seeded into dried meshes in µL
droplets (. × cells/mesh) and left for h to allow for cellular
attachment to PCL microfibers. Following microscopic visual attachment
of cells, mL of respective culture medium (i.e., basic chondrogenic
culture medium containing or excluding TGF was added to each sample
in -well plates. +APPJ +TGF, +APPJ –TGF, and −APPJ −TGF groups
were cultured in basic chondrogenic culture medium (mentioned
above). The −APPJ +TGF group was cultured in basic chondrogenic
culture medium, supplemented with ngmL− TGF (TGFβ, Peprotech,
USA), which was freshly thawed and supplied at each media change.
Samples were cultured for a day period, with media refreshed two
times per week.
Mechanical Testing: The mechanical properties of both cell free
and cell laden fibrous meshes were assessed by uniaxial tensile and
unconfined compression testing, respectively. Both tests were performed
on a MultiTest-.-dv system (Mecmesin, UK) equipped with a N
load cell. Tests were conducted at a constant rate of mmmin− at room
temperature. For tensile tests, rectangular strips (× mm) of APPJ-
treated D microfiber meshes (µm fiber spacing) at three dierent
speeds (slow, medium, and fast) were used; while for compression tests,
cylindrical cultured meshes (mm diameter, .±. mm thickness)
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (13 of 14)
Adv. Funct. Mater. 2023, 33,
following days of culture were used. Tensile and compressive
modulus were determined from engineering stress–strain curves using
a least square fitting of the slope of the stress–strain curves between
–% strain for tensile tests and –% strain for compressive tests.
At least three samples (dierent constructs) for each type of mechanical
test and group were tested.
Histology and Immunohistochemistry: After , , , , and days
of culture, samples (n=) were fixed in formalin for min and then
transferred to .% eosin (in % formalin) for general tissue staining.
Samples were subsequently embedded in % agarose to minimize
loss of samples prior to standard tissue processing and embedding
in paran. Following paran embedding, samples were cut at
µm thickness and stained with safranin-O for glycosaminoglycan
visualization,[] fast green for cytoplasm and collagen and Weigert’s
hematoxylin for cell nuclei. Immunohistochemical staining of collagen
type II was also performed on the paran sections as previously
described[] using the primary antibody II-IIB (DSHB, USA). Histology
images were made of mounted sections in x random locations using a
bright field microscope.
Biochemical Assays: Metabolic assays were performed after , ,
, , and days of culture on the same samples (n= ) using
a resazurin assay (resazurin sodium salt, Alfa Aesar, Germany). A
working solution was prepared in chondrogenic dierentiation medium
(−TGF) containing . µ resazurin sodium salt. Briefly, samples
were incubated, protected from light, for h at °C. Fluorescence was
measured in duplo with excitation at nm and emission at nm.
To quantify glycosaminoglycan (GAG) production during the culture
period, samples were taken after , , , , and days of culture
and subsequently freeze dried (n= ). Samples were then digested
using µL of papain buer consisting of . NaHPO and .
EDTA*HO with a pH of ., mixed with .unitsmL− papain solution
and . mg mL− cysteine HCl. Samples were digested overnight at
°C and then assayed for DNA and GAG content using the Picogreen
assay kit (Thermo Fisher Scientific) and dimethyl methylene blue assay
(DMMB, Sigma), respectively. Briefly, DMMB solution was prepared
in-house with a pH of . Chondroitin sulphate C was used to prepare
a standard curve with concentrations from to µgmL−. Excitation
was measured in duplo at and nm wavelengths, dividing the
nm measurement by the nm measurement before subtracting
the blank and performing subsequent analysis. GAG production was
normalized against DNA content for each sample to normalize for cell
number variation.
Quantitative Real-Time Polymerase Chain Reaction (qRT-PCR): After ,
, , , and days, cultured meshes were collected (n=) and lysed
using . mL per sample TRIzol reagent (Thermo Fisher Scientific).
mRNA was isolated using % chloroform and extracted from the
aqueous phase. Subsequent quantification was performed using a
NanoDrop ND spectrophotometer (Thermo Fisher Scientific) at
/ nm. The iScript cDNA Synthesis Kit was used to synthesize
cDNA using the manufacturer’s instructions. PCR analysis of collagen
type I (COLA), collage type II (COLA), collagen type X (COLXA),
transcription factor primer D (SOX), and aggrecan (ACAN) were
completed with a Bio-Rad CFX Real-Time PCR Detection System
(Bio-Rad) using FastGreen SYBR Green Master mix (Sigma–Aldrich).
Hypoxanthine Phosphoribosyl transferase (HPRT) was used as the
housekeeping gene reference for the expression of the target genes.
CT-values higher than the th cycle were considered undetectable and
thus not plotted. The relative fold change was determined using the
−ΔΔCT method using the mean ΔCT value of the D, +APPJ −TGF group
as the standard. Chondrogenic index was calculated by dividing COLA
fold change mean by COLA fold change mean. The primers that were
used are listed in TableS (Supporting Information) and were designed
and validated elsewhere.[]
Statistical Analysis: Statistical analysis was performed using GraphPad
Prism ... Significance was determined using p-value < . unless
otherwise stated. The dierence in atomic concentration of peak-fitted C
s components was analyzed using a paired t test. ELISA immobilization
and loading eciency measurements were analyzed for variance using a
two-way analysis of variance (ANOVA) with Bonferroni’s post hoc test,
with concentration and speed or time and treatment, respectively. The
dierence in calculated moduli was statistically analyzed using one-way
ANOVA and subsequent post hoc Tukey multiple-comparisons analysis.
The normalized production of GAGs was statistically analyzed using
two-way ANOVA and subsequent post hoc Tukey multiple-comparisons
analysis. Error bars shown are standard deviation of six samples. The
fold change of target genes was statistically analyzed using two-way
ANOVA and subsequent post hoc Tukey multiple-comparisons analysis.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
M.J.A. and O.L. contributed equally to this work. B.A. and M.C. are shared
senior authors. The authors would like to kindly acknowledge the financial
support from the Gravitation Program “Materials Driven Regeneration”,
funded by the Netherlands Organization for Scientific Research
(..), the EU’s H Marie Skłodowska-Curie RESCUE co-fund
grant (#), the Jennifer Foong Scholarship for Biomedical Research,
and an Oce of Global Engagement Partnership Collaboration Award
between the University of Sydney, Utrecht University, and the Australian
Research Council (FL; DP; and DE). This
work was also supported by the partners of Regenerative Medicine
Crossing Borders and powered by Health ∼Holland, Top Sector Life
Sciences & Health. The authors would like to thank I. Dokter for assistance
with running qPCRs and M. van Rijen for histology expertise. The DSHB
Hybridoma Product II-IIB developed by T.F. Linsenmayer was obtained
from the Developmental Studies Hybridoma Bank, created by the NICHD
of the NIH and maintained at The University of Iowa, Department of
Biology, Iowa City, IA .
Conflict of Interest
The authors declare no conflict of interest.
Data Availability Statement
The data that support the findings of this study are available from the
corresponding author upon reasonable request.
Keywords
atmospheric-pressure plasma, cartilage, melt electrowriting, protein
immobilization, stem cell dierentiation, technology convergence,
transforming growth factor beta
Received: June ,
Revised: October ,
Published online: November ,
[] Y.Krishnan, A. J.Grodzinsky, Matrix Biol. 2018, 71, .
[] J. Becerra, J. A. Andrades, E. Guerado, P. Zamora-Navas,
J. M.López-Puertas, A. H.Reddi, Tissue Eng. – Part B Rev. 2010, 16,
.
www.afm-journal.dewww.advancedsciencenews.com
© 2022 The Authors. Advanced Functional Materials published by Wiley-VCH GmbH
2206583 (14 of 14)
[] L. Peterson, M. Brittberg, I. Kiviranta, E. L. Åkerlund, A. Lindahl,
Am. J. Sports Med. 2002, 30, .
[] S.Marlovits, P.Zeller, P.Singer, C.Resinger, V.Vécsei, Eur. J. Radiol.
2006, 57, .
[] H. B.Schuette, M. J.Kraeutler, E. C.McCarty, Orthop. J. Sport. Med.
2017, 5, .
[] S. C.Mastbergen, D. B. F.Saris, F. P. J. G.Lafeber, Nat. Rev. Rheu-
matol. 2013, 9, .
[] M. Castilho, V.Mouser, M.Chen, J.Malda, K. Ito, Acta Biomater.
2019, 95, .
[] V. H. M.Mouser, R.Levato, L. J.Bonassar, D. D.D’Lima, D. A.Grande,
T. J.Klein, D. B. F.Saris, M.Zenobi-Wong, D.Gawlitta, J. Malda, Carti-
lage 2017, 8, .
[] J.Visser, F. P. W.Melchels, J. E.Jeon, E. M.Van Bussel, L. S.Kimpton,
H. M.Byrne, W. J. A.Dhert, P. D.Dalton, D. W.Hutmacher, J.Malda,
Nat. Commun. 2015, 6, .
[] G.Zhen, X.Cao, Trends Pharmacol. Sci. 2014, 35, .
[] W.Wang, D.Rigueur, K. M.Lyons, Birth Defects Res. Part C Embryo
Today Rev. 2014, 102, .
[] L. A.Fortier, J. U.Barker, E. J.Strauss, T. M.McCarrel, B. J. Cole,
Clin. Orthop. Relat. Res. 2011, 469, .
[] L. S. Dolci, A. Liguori, A. Merlettini, L. Calzà, M. Castellucci,
M. Gherardi, V.Colombo, M. L.Focarete, J. Phys. D. Appl. Phys.
2016, 49, .
[] L. S.Wong, F.Khan, J.Micklefield, Chem. Rev. 2009, 109, .
[] C.Stewart, B.Akhavan, S. G.Wise, M. M. M.Bilek, Prog. Mater. Sci.
2019, 106, .
[] S. L. Hirsh, D. R. McKenzie, N. J. Nosworthy, J. A. Denman,
O. U.Sezerman, M. M. M.Bilek, Colloids Surf., B 2013, 103, .
[] L.Vroman, A. L.Adams, J. Biomed. Mater. Res. 1969, 3, .
[] E.-H. Lim, J. P. Sardinha, S.Myers, M.Stevens, Arch. Plast. Surg.
2013, 40, .
[] F. F.Chen, M. D. Smith, In Van Nostrand’s Scientific Encyclopedia
1984.
[] M. M. M. Bilek, M. Vandrovcová, A. Shelemin, A. Kuzminova,
O.Kylián, H.Biederman, L.Bačáková, A. S. Weiss, Appl. Surf. Sci.
2020, 518, .
[] F.Fanelli, F.Fracassi, Surf. Coat. Technol. 2017, 322, .
[] C.Tendero, C.Tixier, P.Tristant, J.Desmaison, P.Leprince, Spectro-
chim. Acta Part B At. Spectrosc. 2006, 61, .
[] X.Lu, G. V.Naidis, M.Laroussi, S.Reuter, D. B.Graves, K.Ostrikov,
Phys. Rep. 2016, 630, .
[] S. K.Alavi, O. Lotz, B.Akhavan, G.Yeo, R.Walia, D. R.McKenzie,
M. M.Bilek, ACS Appl. Mat. Interfaces 2020, 14, .
[] H. Najafi-Ashtiani, B. Akhavan, F. Jing, M. M. Bilek, ACS Appl.
Mater. Interfaces 2019, 11, .
[] L. Utomo, Y. M. Bastiaansen-Jenniskens, J. A. N. Verhaar,
G. J. V. M.vanOsch, Osteoarthr. Cartil. 2016, 24, .
[] I.Trizio, F.Intranuovo, R.Gristina, G.Dilecce, P.Favia, Plasma Pro-
cess. Polym. 2015, 12, .
[] S. Surucu, K. Masur, H. Turkoglu Sasmazel, T. Von Woedtke,
K. D.Weltmann, Appl. Surf. Sci. 2016, 385, .
[] M.Cámara-Torres, R.Sinha, P.Scopece, T.Neubert, K.Lachmann,
A.Patelli, C. Mota, L.Moroni, ACS Appl. Mater. Interfaces 2021, 13,
.
[] U.Little, F.Buchanan, E.Harkin-Jones, B.Graham, B.Fox, A.Boyd,
B.Meenan, G.Dickson, Acta Biomater. 2009, 5, .
[] Z.Ma, W.He, T.Yong, S.Ramakrishna, Tissue Eng. 2005, 11, .
[] C. T.Tran, T. J.Raeber, B. J.Murdoch, A. J.Barlow, J. G.Partridge,
D. G.McCulloch, D. R.McKenzie, J. Appl. Phys. 2019, 125, .
[] K.Fricke, K.Duske, A.Quade, B.Nebe, K.Schroder, K.Weltmann,
T.vonWoedtke, IEEE Trans. Plasma Sci. 2012, 40, .
[] K.-Y.Law, H.Zhao, in Surface Wetting, Springer International Pub-
lishing, Cham 2016.
[] M. J. Hawker, A. Pegalajar-Jurado, E. R. Fisher, Plasma Process.
Polym. 2015, 12, .
[] E. R.Fisher, ACS Appl. Mater. Interfaces 2013, 5, .
[] D. P.Dowling, M.Donegan, P. J.Cullen, V. J.Law, V.Milosavljevic,
IEEE Trans. Plasma Sci. 2014, 42, .
[] A. Maei, N. Michieli, P. Brun, A. Zamuner, A. Zaggia, M. Roso,
B. Kalinic, E. Verga Falzacappa, P. Scopece, S. Gross, M. Dettin,
A.Patelli, Appl. Surf. Sci. 2020, 507, .
[] D.Mance, R.Wiese, T.Kewitz, H.Kersten, Eur. Phys. J. D 2018, 72, .
[] T. D.Brown, P. D.Dalton, D. W.Hutmacher, Prog. Polym. Sci. 2016,
56, .
[] A.Charuchinda, R. Molloy, J.Siripitayananon, N.Molloy, M.Sriyai,
Polym. Int. 2003, 52, .
[] I. Castilla-Cortázar, J. Más-Estellés, J. M. Meseguer-Dueñas,
J. L.Escobar Ivirico, B.Marí, A.Vidaurre, Polym. Degrad. Stab. 2012,
97, .
[] X. Wei, C.Gong, M. Gou, S. Fu, Q.Guo, S.Shi, F.Luo, G.Guo,
L.Qiu, Z.Qian, Int. J. Pharm. 2009, 381, .
[] M. Sarem, O. Otto, S. Tanaka, V. P.Shastri, Stem Cell Res. Ther.
2019, 10, .
[] Z. X. Xu, T. Li, Z. M. Zhong, D. S. Zha, S. H. Wu, F. Q. Liu,
W.De Xiao, X. R.Jiang, X. X. Zhang, J. T.Chen, Biopolymers 2011,
95, .
[] R. Ghobeira, C. Philips, L. Liefooghe, M. Verdonck, M. Asadian,
P.Cools, H.Declercq, W. H.De Vos, N.De Geyter, R.Morent, Appl.
Surf. Sci. 2019, 485, .
[] P. Behrendt, Y. Ladner, M. J. Stoddart, S. Lippross, M. Alini,
D.Eglin, A. R.Armiento, Am. J. Sports Med. 2020, 48, .
[] M. Castilho, G. Hochleitner, W. Wilson, B. van Rietbergen,
P. D.Dalton, J.Groll, J.Malda, K.Ito, Sci. Rep. 2018, 8, .
[] R. Levato, W. R. Webb, I. A. Otto, A. Mensinga, Y. Zhang,
M.vanRijen, R. vanWeeren, I. M.Khan, J. Malda, Acta Biomater.
2017, 61, .
[] J.Visser, D.Gawlitta, K. E. M. Benders, S. M. H.Toma, B.Pouran,
P. R.vanWeeren, W. J. A.Dhert, J.Malda, Biomaterials 2015, 37, .
[] L.Rosenberg, J. Bone Jt. Surg. 1971, 53, .
Adv. Funct. Mater. 2023, 33,
