Poly(levodopa)-modified β-glucan as a candidate for wound dressings

Levodopa (biological precursor of dopamine) is sometimes used instead of dopamine for synthesis of highly adhesive polycatecholamine coatings on different materials. However, in comparison of polydopamine, little is known about biological safety of poly(levodopa) coatings and their efficacy in binding of therapeutically active substances.

Available online 24 July 2021

0144-8617/© 2021 The Authors Published by Elsevier Ltd This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)

Poly(levodopa)-modified β-glucan as a candidate for wound dressings

Anna Michalichaa, Agata Roguskab, Agata Przekorac, Barbara Budzy´nskad, Anna Belcarza,*

aChair and Department of Biochemistry and Biotechnology, Medical University of Lublin, Chodzki 1, 20-093 Lublin, Poland

bInstitute of Physical Chemistry, Polish Academy of Sciences, Kasprzaka 44/52, 01-224 Warsaw, Poland

cIndependent Unit of Tissue Engineering and Regenerative Medicine, Chair of Biomedical Sciences, Medical University of Lublin, Chodzki 1, 20-093 Lublin, Poland

dIndependent Laboratory of Behavioral Studies, Medical University of Lublin, Chodzki 4a, 20-093 Lublin, Poland

A R T I C L E I N F O

Keywords:

β-Glucan

Poly(levodopa)-based modification

Fibroblasts

Danio rerio

Antibacterial activity

A B S T R A C T Levodopa (biological precursor of dopamine) is sometimes used instead of dopamine for synthesis of highly adhesive polycatecholamine coatings on different materials However, in comparison of polydopamine, little is known about biological safety of poly(levodopa) coatings and their efficacy in binding of therapeutically active substances Therefore, thermally polymerized curdlan hydrogel was modified via two different modes using levodopa instead of commonly used dopamine and then coupled with gentamicin – aminoglycoside antibiotic Physicochemical properties, biological safety and antibacterial potential of the hydrogels were evaluated Poly (levodopa) deposited on curdlan exhibited high stability in wide pH range and blood or plasma, were nontoxic in zebrafish model and favored blood clot formation Simultaneously, one of hydrogel modification modes allowed

to observe high gentamicin binding capacity, antibacterial activity, relatively high nontoxicity for fibroblasts and was unfavorable for fibroblasts adhesion Such modified poly(levodopa)-modified curdlan showed therefore high potential as wound dressing biomaterial

1 Introduction

Polydopamine (PDA) is a major pigment which occurs in natural

melanin (eumelanin) (Simon & Peles, 2010) It also mimics the

specialized adhesive foot protein (Mytilus edulis foot protein-5) in

mus-sels (Lee et al., 2007) Based on this phenomenon, the method of

bio-mimetic approach for the functionalization of a wide range of materials

has been developed and proposed in 2007 (Lee et al., 2007) Since then,

formation of PDA coating became very popular as a strategy of solid

substrate functionalization for a variety of technical, environment-

protecting and medical purposes PDA coatings were used for example

for modification of graphene nanosheets (Wang et al., 2013; Xu et al.,

2010), Fe3O4 nanoparticles (for drug delivery, for catalyst support,

ad-sorbents and sensors) (Liu et al., 2013; Wang et al., 2013; Zhou et al.,

2010), silica nanoparticles (Zhu et al., 2019), polymers (Hu & Mi, 2013;

Murphy et al., 2010), titanium (Steeves et al., 2016), and many other

matrices Moreover, melanin-like coatings enable the secondary

coupling reactions with different organic molecules, due to the presence

of catechol domains in their structure which can react with thiols and

amines via Michael addition or Schiff base reactions (Burzio & Waite,

2000; LaVoie et al., 2005) Therefore, the number of scientific reports

concerning this topic grows rapidly, indicating the enormous interest in this useful technique

Most scientific reports state that melanin-like coatings are formed from dopamine Very rarely levodopa is used for this purpose instead of dopamine, forming poly(L-DOPA) Both dopamine and levodopa belong

to catecholamine family Dopamine (3-hydroxytyramine; 2-(3,4-dihy-droxyphenyl)ethylamine; 4-(2-Aminoethyl)-1,2-benzenediol) contains both catechol and amino groups in its molecule Levodopa (L-DOPA; DOPA; 3,4-Dihydroxy-L-phenylalanine; L-3-Hydroxytyrosine) is closely related to dopamine (as its precursor in catecholamine synthesis pathway) and structurally differs from this compound by the presence of carboxyl group in aminoethyl moiety Both dopamine and L-DOPA polymerize in the presence of oxidants (as O2 or Cu2+ions) and in slightly alkaline buffers (e.g in 10–50 mM Tris pH 8.5), although Bernsmann et al (Bernsmann et al., 2011) reported that O2 as oxidant may be not effective in case of L-DOPA In fact, although dopamine is used much more frequently as a monomer for polycatecholamine coat-ings formation, both these compounds are commonly called “dopamine”

in numerous scientific reports In comparison with PDA, little is known about poly(levodopa) (poly(L-DOPA)) properties, both in relation to its biological safety and potential in coupling with attractive biological

* Corresponding author

E-mail address: anna.belcarz@umlub.pl (A Belcarz)

Contents lists available at ScienceDirect Carbohydrate Polymers journal homepage: www.elsevier.com/locate/carbpol

https://doi.org/10.1016/j.carbpol.2021.118485

Received 8 June 2021; Received in revised form 12 July 2021; Accepted 22 July 2021

molecules However, carboxyl group in L-DOPA, which is absent in

dopamine molecule, may affect its polymerization to poly(L-DOPA) and

also the polymer properties For example, presence of carboxyl group (of

approx 2–3 pKa value) is responsible for negative charge of L-DOPA

which increases the poly(L-DOPA) dispersity in water (

Hashemi-Mog-haddam et al., 2018) Polyolefin membranes coated with poly(L-DOPA)

showed notable content of free –COOH groups on their surfaces and also

an increased hydrophilicity, although the latter was higher for

polydopamine-coated membranes (Xi et al., 2009) Poly(L-DOPA) films

deposited on polypropylene, nylon and poly(vinylidenefluoride)

sub-strates showed significantly higher stability in strong acidic conditions

than analogous polydopamine coatings (Wei et al., 2013) L-DOPA, due

to –COOH content, was also used for the synthesis of PDE-DOPA4

monomer blocks which were further oxidized (with assistance of NaIO4)

to form quickly setting adhesive hydrogel (Burke et al., 2007)

Importantly, free carboxyl group in poly(L-DOPA) molecule could be

an additional site for interactions between poly(L-DOPA) coatings and

different ions/particles/molecules As suggested by Bernsmann et al

(Bernsmann et al., 2011), L-DOPA during its polymerization to poly(L-

DOPA) is turned to 5,6-dihydroxyindole-2-carboxylic acid – an

inter-mediate which is unable to undergo a 2,2′-branching Thus, this

carboxyl group in 5,6-dihydroxyindole-2-carboxylic acid units is

prob-ably not engaged in the formation of covalent bonds in poly(L-DOPA)

coatings Poly(L-DOPA) was already effectively used to attach paclitaxel

to core-shell Fe3O4@poly(DOPA) nanoparticles (Hashemi-Moghaddam

et al., 2018) although the exact role of free carboxyl groups in these

nanoparticles was not explained In nitrogen-doped graphene quantum

dots obtained with assistance of L-DOPA, free surface carboxyl groups

were prone to coordinate with Fe3+ions, thus facilitating the electron

transport between ions and dots and in consequence enabling the

for-mation of Fe3O4-dots hybrids (Shi et al., 2016) Thus, poly(L-DOPA)

coatings may exert other properties and allow different applications

than PDA coatings

Recently, our group synthesized PDA-modified high-set curdlan

hydrogels using thermal polymerization method Curdlan (β-1,3-glucan)

is a polysaccharide of specific gelling properties, high water sorption

capacity, significant elasticity and relatively high mechanical resistance

(Chen & Wang, 2020) It exhibits therefore high potential for design of

wound dressing materials However, due to the presence of exclusively

hydroxyl groups in repeating glucose units of curdlan backbone, it is

biologically inert and relatively insusceptible to modifications

improving its biological properties and capability to bind

therapeuti-cally active molecules (Cai & Zhang, 2017) PDA coating could be

therefore an excellent method to introduce modifiable domains for

biological improvement of curdlan We do have demonstrated that PDA-

modified hydrogels showed the ability to bind molecules containing free

amino groups (as proved for gentamicin and peroxidase) and exhibited

unchanged mechanical stability, increased soaking capacity, prolonged

antibacterial properties and Fickian-type mechanism of drug release

(Michalicha et al., 2021) These results suggested that PDA-modified

curdlan hydrogel may serve as a carrier of free amino groups-

containing molecules and be used for different purposes, e.g

antibac-terial hydrogels for wound dressings

In view of this, we hypothesized that also levodopa may effectively

form the biologically safe deposits on polysaccharide matrix and may be

used for fabrication of wound dressings Therefore, in this paper we

report the fabrication of poly(levodopa)-modified curdlan hydrogels

The hydrogels were first characterized for stability as well as biological

safety of the deposits during contact with blood, fibroblast cell line,

fibroblast primary culture and zebrafish eggs and larvae, in relation to

their possible application as wound dressings Second, the modified

hydrogels were coupled with gentamicin and their antibacterial activity

was evaluated

2 Materials and methods

2.1 Synthesis of poly( L -DOPA)-modified curdlan hydrogels

Curdlan powder (from Alcaligenes faecalis; cat No 281–80,531; DP

6790; average Mw 1100 kDa; specific rotation [A]20/D: +30 to +35),

Cl−content < 0.5%, heavy metals content (including Pb) < 0.002%, was

provided by Wako Chemicals (Japan); Tris (2-Amino-2-(hydroxymethyl) propane-1,3-diol) and L-DOPA (3,4-Dihydroxy-L-phenylalanine) by Sigma-Aldrich (USA) Control and poly(L-DOPA)-modified hydrogels were synthesized according to procedures described elsewhere (Michalicha et al., 2021), briefly:

2.1.1 With L -DOPA monomer added to curdlan suspension Before thermal Gelling (BG)

Suspension of 0.4 g curdlan in 5 ml 10 mM Tris/HCl buffer pH 8.5 was combined with 10 mg (2-LD-BG) or 20 mg (4-LD-BG) of L-DOPA, stirred 10 min until L-DOPA was completely dissolved, transferred into glass tubes (ø 13 mm) and polymerized at 93 ◦C for 15 min After cooling, hydrogel was cut into 3 mm slices and incubated in air (air oxygen as an oxidant) for 24 h at 25 ◦C to allow L-DOPA polymerization Then slices were washed 10 times in 100 ml DI H2O, frozen and lyophilized (SRK, System Technik GMBK, Germany)

2.1.2 With L -DOPA monomer added to curdlan suspension After thermal Gelling (AG)

Suspension of 0.4 g curdlan in 5 ml 10 mM Tris/HCl buffer pH 8.5 was transferred into glass tubes (ø 13 mm) and polymerized at 93 ◦C for

15 min Cooled hydrogel was cut into 3 mm slices, immersed in 5 ml

10 mM Tris/HCl buffer pH 8.5 containing 10 mg (2-LD-AG) of L-DOPA and incubated in orbital shaker for 24 h at 25 ◦C with the access to air (air oxygen as an oxidant), to allow L-DOPA polymerization Then slices were washed 10 times in 100 ml DI H2O, frozen and lyophilized

2.1.3 Control curdlan hydrogels were prepared as in Section 2.1.2 without the immersion in L -DOPA solution and further incubation

Prior to cell cultures, zebrafish, drug release and antibacterial ac-tivity experiments, all curdlan hydrogels were sterilized by ethylene oxide method in paper/plastic peel pouch (sterilization for 1 h at 55 ◦C, aeration for 20 h)

2.2 Gentamicin immobilization and quantitative analysis

Immobilization of gentamicin into modified curdlan samples was performed by incubation of lyophilized slices in 1 mg/ml gentamicin (Sigma-Aldrich, USA) in Britton-Robinson buffer pH 8.5, in proportion 33.3 ml of antibiotic solution/1 g lyophilized curdlan hydrogel slices, using DTS-4 shaker (100 rpm), 24 h at 25 ◦C, followed by 24 h at 4 ◦C Then slices were washed twice in 50 ml DI water, frozen and lyophilized

In case of pilot experiment, EDC/NHS activation of poly(L-DOPA) carboxyl groups was used for binding with gentamicin Hydrogel sliced were first soaked in 0,1 M MES buffer pH 6,5 and then incubated in

mixture of 0.1 M EDC (N-(3-Dimethylaminopropyl)-N-ethyl-carbodiimide hydrochloride; Sigma-Aldrich, USA) and 0.2 M NHS (N-

hydroxy succinimide; Sigma-Aldrich, USA) in 0,1 M MES buffer pH 6,5 for 1 h, at 25 ◦C, on plate shaker DTS-4 (ELMI, USA), 100 rpm After-wards, the samples were washed twice (10 min.) in distilled water and immersed in 1 mg/ml gentamicin (Sigma-Aldrich, USA) in 0,05 M NaHCO3 (pH 8.5), in proportion 33.3 ml of antibiotic solution/1 g lyophilized curdlan hydrogel slices, using DTS-4 shaker (100 rpm), 24 h

at 25 ◦C

Gentamicin concentration in solutions before and after incubation was evaluated according to Ginalska et al (Ginalska et al., 2004), based

on gentamicin derivatization by phthaldialdehyde (Sigma-Aldrich, USA) Amount of immobilized gentamicin was calculated from formula (1):

C i(µg/g d.w.)=

[

C b(µg/ml) –C a(µg/ml)

]

x V(ml)

where:

Ci(μg/g d.w.) – amount of gentamicin immobilized on samples (μg/g of

dry hydrogel weight)

Cb(μg/ml) – concentration of gentamicin in solution before incubation

with samples (μg/ml)

Ca(μg/ml) – concentration of gentamicin in solution after incubation

with samples (μg/ml)

V(ml) – volume of solution incubated with samples (ml)

Mg – dry weight of samples (g)

2.3 Characterization of modified curdlan hydrogels

FTIR-ATR spectra were collected using Vertex 70 IR-

spectrophotometer (4000–400 cm− 1, 64 scans, 4 cm− 1 spectral

resolu-tion; Bruker, USA) and analysed using OPUS 7.0 software (Bruker, USA)

X-ray photoelectron spectroscopy (XPS) measurements were made

on a Microlab 350 (Thermo Electron) spectrometer with non-

monochromatic AlKα (hν =1486.6 eV, power 300 W, voltage 15 kV)

radiation as an X-ray excitation source A lateral resolution was about

0.2 cm2 The high-resolution XPS spectra were acquired using the

following parameters: pass energy 40 eV, energy step size 0.1 eV A

Smart function of background subtraction was used to obtain the XPS

signal intensity The all collected XPS peaks were fitted using an

asymmetric Gaussian/Lorentzian mixed function The measured binding

energies were calibrated with respect to the energy of C 1s at 285.0 eV

Avantage-based data system software (Version 5.9911, Thermo Fisher

Scientific) was used for data processing

For evaluation of soaking capacity, lyophilized curdlan hydrogels

were incubated in 5 ml 0.9% NaCl at 37 ◦C for up to 72 h In defined time

points, the samples were withdrawn from the liquid, drained on

What-man paper to remove the excess of the liquid and weighed (XS205,

Mettler-Toledo, Switzerland) The obtained data were normalized (to

100% of initial dry weight of samples) Experiment was performed in

triplicate Statistically significant differences between non-activated and

activated sample in each modification mode were considered at

p < 0.05, according to Student’s t-test (GraphPad Prism 6 Software, San

Diego, CA)

2.4 Stability of poly( L -DOPA) deposits

Poly(L-DOPA)-modified hydrogel slices were extensively washed in

distilled water (10 times in 0.5 l; first 5 washes for 2 h, second 5 washes

for 12 h; RM 5-30 V shaker (CAT M Zipperer Gmbh, Germany),

30 rpm.) Incubation in human serum (kindly donated by Regional

Center of Blood Donation and Blood Treatment in Lublin) and in human

blood (collected after approval of Bioethics Committee at the Medical

University of Lublin, no KE-0254/258/2020) was performed (2 ml per

30 mg of dry hydrogel weight) on plate shaker DTS-4 (ELMI, USA),

100 rpm, at 37 ◦C, for up to 96 h; then washed once in distilled water

Effect of pH was tested using 0.1 M Britton-Robinson buffers pH 2, 4, 6,

8, 10 and 12 (the same liquid-to-biomaterial proportion as for human

serum and blood), for at 37 ◦C, for 7 days Macro photography of

plasma-, blood- and buffers-incubated hydrogels and post-incubation

buffers was performed using E-520 digital camera (Olympus, Japan)

2.5 Hemolysis and blood clot formation test

Human blood was collected on citrate from healthy volunteer on

approval of Bioethics Committee at the Medical University of Lublin, no

KE-0254/258/2020 Its total hemoglobin and plasma hemoglobin

con-centration were estimated on basis of reaction with Drabkin reagent and

appropriate calibration curve (using 96-well plates and Synergy H4 hybrid microplate reader, Biotek, USA) and were 136 mg% and 0.18 mg/ml, respectively For hemolysis test, lyophilized hydrogel slices (30 mg ± 2 mg) were immersed in 2 ml of blood 100× diluted in PBS

pH 7.4 without Ca2+and Mg2+ions Positive control contained 0.1% Triton X-100 while negative one: 30 mg ± 2 mg of HDPE (high density polyethylene, Sigma-Aldrich, USA) Then samples were incubated 3 h at

37 ◦C in Innova 42 incubator shaker (New Brunswick Scientific, USA),

150 rpm Erythrocytes-released hemoglobin was estimated using reac-tion with Drabkin reagent, as above For blood clot formareac-tion test, 100 μl

of whole 10 mM CaCl2-activated blood was placed onto 30 mg ± 2 mg lyophilized hydrogel slices or pieces of HDPE (negative control) Non- activated Ca2+-free whole blood (100 μl) served as positive control Then samples were incubated 15 min., 30 min or 45 min at 37 ◦C, without shaking (controls were performed individually per each time point) Then all samples were incubated with 2.5 ml of distilled water for

5 min Finally the hemoglobin content in solution was estimated using reaction with Drabkin reagent, as described above Each experiment was performed in triplicate Statistically significant differences between

negative control and various samples were considered at p < 0.0001,

according to a One-way ANOVA with post-hoc Dunnett’s test (GraphPad Prism 8.0.0 Software, San Diego, CA)

2.6 Cell culture experiments

Cell culture tests were conducted using human skin fibroblasts: normal human skin fibroblast cell line (BJ cell line, ATCC-LGC dards) and primary human dermal fibroblasts (HDFs, ATCC-LGC Stan-dards) Fibroblasts were cultured at 37 ◦C in a humidified atmosphere of 5% CO2 and 95% air BJ fibroblasts were maintained in EMEM medium supplemented with 10% fetal bovine serum (FBS, Pan-Biotech), and

100 U/ml/100 μg/ml penicillin/streptomycin mixture (Sigma-Aldrich Chemicals) HDFs were maintained in a Fibroblast Basal Medium sup-plemented with the components of Fibroblast Growth Kit-Low Serum (both purchased from ATCC-LGC Standards)

2.6.1 Cytotoxicity tests according to ISO 10993-5 standard

Fibroblast suspension with a concentration of 2 × 105 cells/ml was seeded in 100 μl into the wells of 96-multiwell plates After 24-h incu-bation, the culture medium was discarded and the monolayer of cells was exposed to the extracts of the tested samples The extracts of the materials were prepared according to ISO 10993-12 standard by placing 0.1 g sample in 1 ml of complete culture medium followed by 24-h in-cubation at 37 ◦C Non-toxic extract serving as a negative control of cytotoxicity was prepared by the incubation of complete culture me-dium in a polystyrene vessel for 24 h at 37 ◦C without any biomaterial (extract marked as PS control) Fibroblasts were maintained in the ex-tracts for 48 h and then cytotoxicity of the samples was determined by evaluation of cell metabolism using MTT assay (Sigma-Aldrich Chem-icals) and cell number using total LDH test (Sigma-Aldrich ChemChem-icals) The MTT assay was carried out based on the procedure described earlier (Przekora et al., 2014) Total LDH test was performed after cell lysis according to the manufacturer instructions Results of MTT and total LDH tests were presented as the percentage of negative control of cytotoxicity (100% viability in terms of cell metabolism and cell num-ber) Three independent experiments were conducted for both cyto-toxicity tests Statistically significant differences between negative

control (PS control) and various samples were considered at p < 0.05,

according to a One-way ANOVA with post-hoc Dunnett’s test (GraphPad Prism 8.0.0 Software, San Diego, CA)

2.6.2 Cell proliferation

Fibroblast suspension with a concentration of 1.5 × 104 cells/ml was seeded in 100 μl into the wells of 96-multiwell plates After 24-hour

incubation, total LDH test was conducted to determine cell number at starting point (time = 0 h – before addition of the extracts) The exact

number of fibroblasts was calculated using calibration curve prepared

for known concentrations of BJ cells and HDFs Then, extracts of the

biomaterials (prepared as described in section 2.6.1) were added to the

cells which were further incubated for 3 days Cell number was

deter-mined after 24- and 72-hour exposure to the extracts Three independent

proliferation tests were performed Statistically significant diffrences

between PS control (non-toxic control extract prepared by incubation of

culture medium in a polystyrene vessel) and various samples were

considered at p < 0.05, according to a One-way ANOVA with post-hoc

Dunnett’s test (GraphPad Prism 8.0.0 Software; San Diego, CA)

2.6.3 Cell adhesion to the samples

Before the experiment, all samples in the form of discs (12 mm in

diameter and 2 mm in height, weighing 30 mg ± 2 mg) were sterilized

by ethylene oxide, placed in the wells of 24-multiwell plate and soaked

in the complete culture medium Then, fibroblasts (2 × 105 cells per

sample) were seeded directly on the top surface of the materials

Fi-broblasts seeded into a polystyrene well of the 24-multiwell plate served

as control cells (PS control), revealing good adhesion and high viability

After 72-h culture of fibroblasts on the samples, their adhesion and

viability were evaluated using Live/Dead Double Staining Kit (Sigma-

Aldrich Chemicals) and confocal laser scanning microscope (CLSM,

Olympus Fluoview equipped with FV1000) Cell staining was carried out

according to the manufacturer protocol, using two fluorescent probes:

calcein-AM (green fluorescence of viable cells) and propidium iodide

(red fluorescence of dead cells)

2.7 Danio rerio embryotoxicity tests

2.7.1 Zebrafish maintenance

Danio rerio of the AB strain (Experimental Medicine Centre, Medical

University of Lublin, Poland) were maintained at 28.5 ◦C, on a 14/10 h

light/dark cycle, under standard aquaculture conditions Fertilized eggs

were collected via natural spawning Embryos were reared in E3 embryo

medium (pH 7.1–7.3; 17.4 μM NaCl, 0.21 μM KCl, 0.12 μM MgSO4 and

0.18 μM Ca(NO3)2) in an incubator (IN 110 Memmert GmbH, Germany)

at 28.5 ◦C

For the experiments, the embryos were treated with hydrogel

ex-tracts prepared by incubation of the control and poly(L-DOPA)-modified

hydrogels in sterile E3 medium (in proportion: 1 ml E3 medium/0.1 g

dry weight of hydrogel sample) at 37 ◦C, 24 h, with agitation (50 rpm,

Innova 42, New Brunswick Scientific, USA) The extracts were therefore

prepared in analogous manner as in tests of cytotoxicity (Section 2.6.2.)

as well as drug release and antibacterial activity in semi-open-loop

system (Sections 2.8.1 and 2.9.3), to enable the appropriate

compari-son between therapeutic potential and biological safety of hydrogels

Immediately after the experiment, larvae were killed by immersion

in 15 μM tricaine solution All experiments were conducted in

accor-dance with the National Institute of Health Guidelines for the Care and

Use of Laboratory Animals and the European Community Council

Directive for the Care and Use of Laboratory Animals of 22 September

2010 (2010/63/EU)

2.7.2 Fish embryo toxicity (FET) test

The collected embryos were transferred to a Petri dish with E3

me-dium and then placed in 96-well plates, 1 embryo per well To each well,

either 200 μl extract of control or poly(L-DOPA)-modified hydrogels or

the same volume of pure E3 medium was added (each group of n = 15)

The embryos were then maintained in the incubator (as described in

Section 2.7.1) Apical observations of acute toxicity in zebrafish

em-bryos 24–96 h post fertilization (hpf) were performed according to

OECD guidelines for the testing of chemicals no 236 (Organization for

Economic Co-operation and Development 2013 Test No 236: Fish

embryo acute toxicity (FET) test Guidelines for the Testing of

Chem-icals Paris, France) Coagulation, somite formation, tail detachment and

heartbeat were observed using a stereomicroscope (Zeiss Axio Vert,

ZEISS, Germany) Incidence of morphological and physiological abnor-malities e.g a lack of somite formation, scoliosis, or the pericardial oedema were observed and compared to the control embryos Image analysis was performed to determine the percentage of malformed embryos

2.7.3 Locomotor activity assay

Test was performed on 5 day post fertilization (dpf) larvae One larva per well was placed in 96 multi-well plate To each well, either 200 μl extract of control or poly(L-DOPA)-modified hydrogels or the same

volume of pure E3 medium was added (each group of n = 20) and the

zebrafish larvae were incubated in the extracts for 30 min before the test Then, EthoVision XT 15 video tracking software (Noldus Informa-tion Technology b.v., The Netherlands) was used for evaluaInforma-tion of lo-comotor activity The distance moved in 10 min period was calculated in

cm, in a light condition The results were processed by the one-way

ANOVA analysis with Dunnett’s post hoc test A p value <0.05 was

considered statistically significant All statistical analyses were per-formed using GraphPad Prism 8 (GraphPad Software, San Diego, CA, USA)

2.8 Gentamicin release

Drug release evaluation was performed as described earlier (Michalicha et al., 2021), both in semi-open-loop and close-loop sys-tems Both procedures were described below:

2.8.1 Semi-open-loop system

Samples (in triplicate) were incubated with sterile PBS pH 7.4 (proportion: 1 ml PBS/0.1 g dry weight of hydrogel sample) at 37 ◦C, with agitation (50 rpm, Innova 42, New Brunswick Scientific, USA) The extracts were collected daily and replaced by the same volume of PBS Gentamicin concentration was measured in daily collected extracts, until detectable

2.8.2 Closed-loop system

USP 4 Flow-through cell dissolution testing was performed using CE1 Sotax units (Donau Lab, Switzerland) for gentamicin-loaded control and poly(L-DOPA)-modified hydrogels, each in amount containing 2 mg of gentamicin, with 50 ml PBS pH 7.4 and 1 ml/min laminar flow rate, at

37 ◦C 3 ml of extracts were collected at defined time intervals for the estimation of drug concentration (3 ml of fresh PBS was immediately added to the units to maintain the constant volume) Cumulative drug concentrations were calculated based on the results of 4 independent experiments (each in triplicate) Korsmeyer–Peppas and Higuchi models were used to describe the drug release kinetics Korsmeyer–Peppas model was also used to determine the drug release mechanisms, ac-cording to a general equation:

Mt/M∞=ktn

where:

Mt is the amount of drug released from the composite in time t, M∞ is

the accumulated released drug amount at time t → ∞, k and n are the

kinetic constant and the release exponent, respectively

Drug release mechanism was interpreted via nonlinear regression analysis, using the Statistica 10 software

2.9 Antibacterial activity evaluation

Antibacterial activity evaluation was performed as described earlier (Michalicha et al., 2021), briefly:

2.9.1 Bacterial strains

3 reference bacterial strains (Staphylococcus aureus ATCC 25923,

Staphylococcus epidermidis ATCC 12228 and Escherichia coli ATCC

25922) were refreshed from sterile microbanks using Mueller-Hinton

Agar medium (Oxoid, USA), at 37 ◦C for 24 h Then bacteria were

transferred into Mueller-Hinton (M-H) broth medium (Oxoid, USA) and

cultured for 24 h at 37 ◦C Directly before the experiments, the

sus-pension of propagated bacteria was diluted to appropriate density The

strains were chosen on basis of available statistics of infections

occur-rence for implantable biomaterials (Khatoon et al., 2018)

2.9.2 Antibacterial activity test (based on the standard: AATCC Test

Method 100–2004)

Briefly, gentamicin-loaded hydrogels, both control and poly(L-

DOPA)-modified ones, were placed on sterile Petri dishes (in

quadru-plicate) and treated with bacterial suspension (1.5 × 105 CFU/ml) of

each strain, prepared in M-H broth (250-fold in sterile 0.9% NaCl) The

volumes of bacterial suspension were calculated to be completely

absorbed by hydrogels without the excess leaching outside Hydrogel

samples were then incubated at 37 ◦C for 24 h (T24h), then transferred to

sterile 0.9% NaCl (in volume 100-fold larger than volume of bacterial

suspension absorbed by samples) and vigorously shaken (1 min.) to elute

the bacterial cells Samples of the collected eluate were plated (in

trip-licate) onto M-H agar using automatic plater (EasySpiral Dilute,

Inter-science, France) Another set of untreated control hydrogels was soaked

in bacterial inoculate as above and immediately (T0h) subjected to

bacterial cells elution M-H agar plates with plated bacteria eluted from

hydrogel samples were incubated at 37 ◦C for 48 h CFU were then

counted for each plate and percent reduction of bacteria was calculated

by the formula (2):

where:

B – the number of bacteria recovered from the inoculated control

specimen immediately after the inoculation (T0h);

A – the number of bacteria recovered from the inoculated poly(L-

DOPA)-modified specimen 24 h after the inoculation (T24h)

2.9.3 Bacterial growth inhibition in semi-open-loop system

Lyophilized and sterilized control and poly(L-DOPA)-modified

hydrogels were incubated in sterile M-H broth (proportion: 1 ml

me-dium/0.1 g dry weight of hydrogel sample) at 37 ◦C, with agitation

(50 rpm, Innova 42, New Brunswick Scientific, USA) The extract was

collected daily and replaced by the same volume of fresh medium The

experiment was performed in triplicate, until the loss of antibacterial

properties in collected extract was observed

200 μl of each collected extract (in quadruplicate) in 96-well plate

was inoculated by 10 μl of inoculate of three bacterial strains (density:

3 × 107 CFU (colony forming units)/1 ml) and cultured for 24 h at 37 ◦C,

on plate shaker DTS-4 (ELMI, USA), 100 rpm In parallel, bacterial

growth positive control was tested for each strain in M-H broth,

inocu-lated and treated as above After 24 h, absorbance (at 660 nm) of the

cultures was measured on Synergy H4 Hybrid microplate reader

(Bio-Tek, USA) Inhibition of bacterial growth by hydrogel extracts was

calculated as a percent of growth in positive controls Beginning of

bacterial infection was noted when the density in inoculated extracts

was higher than that of untreated blank (pure M-H broth) Experiment

was performed until bacterial growth in inoculated extracts reached the

optical density of positive growth controls (final stage of infection)

Hydrogel samples were finally subjected to the test of bacterial adhesion

(as described in Section 2.9.4.)

2.9.4 Bacterial adhesion

The samples of hydrogel after the bacterial growth inhibition test

were incubated in 1 ml of bacterial suspensions (approx 1.0 × 108 cells/

ml; in M-H broth) of each strain, for 1.5 h, at 37 ◦C, without shaking

Afterwards, non-adhered bacteria were gently washed away with 0.9%

NaCl (50 ml, 4 times) Washed samples were incubated with Viability/ Cytotoxicity Assay Kit for Bacteria Live & Dead Cells (Biotium, USA) in 0.9% NaCl (according to manufacturer instructions: at R/T, 15 min, in darkness) After staining, the samples were washed in 0.9% NaCl to remove non-absorbed dye Adhered bacteria were visualized by confocal microscopy (Olympus Fluoview FV1000; Olympus, Japan)

3 Results and discussion

L-DOPA monomer was used for modification on curdlan fibers ac-cording to two different modes, as presented in procedure description: the monomer was introduced to curdlan before (BG; Before Gelling) or after (AG; After Gelling) its thermal gelling into high-set hydrogel (Fig 1a) Hypothetically, when monomer solution is added to already polymerized curdlan hydrogel (AG method), poly(L-DOPA) can be formed on the surface of triple helix curdlan network (Zhang & Edgar,

2014) Whereas when L-DOPA monomer is added to curdlan suspension before its thermal gelling (BG method), it may enter the fibers of non- polymerized curdlan and then polymerize not onto but within the further formed triple helix structure of curdlan hydrogel This difference may affect the structure and therefore properties of poly(L-DOPA) de-posits on hydrogel helical fibers The concentration of L-DOPA monomer used in this study for poly(L-DOPA) modification was 2 or 4 mg/ml Most commonly, concentration of catecholamine monomer used for poly-merization in limited to 2 mg/ml solution, because higher monomer concentrations may result in the formation of unstable coatings (Liu

et al., 2018) However, conditions of L-DOPA polymerization according

to BG method are different than standard catecholamine polymerization (by incubation of solid samples in monomer solution) Therefore, in this particular method, two L-DOPA concentrations were tested (2 mg/ml and 4 mg/ml) Uniform brownish-black color of all synthesized poly(L- DOPA)-modified curdlan hydrogels (2-LD-BG, 4-LD-BG and 2-LD-AG) indicated their efficient covering by polycatecholamine layer (Fig 2a) One of important features of hydrogels is their ability to absorb liq-uids (soaking capacity, water absorption capacity) It was found earlier that wettability of biomaterials can be improved by coating with PDA (Guo et al., 2016; Li et al., 2020; Liu et al., 2018) Therefore, this property was evaluated for poly(L-DOPA)-modified curdlan hydrogels All versions of modified curdlan hydrogels reached the total soaking capacity after 48 h Interestingly, only the value of total soaking capacity for 2-LD-AG material was increased by 13% in comparison with other samples (811% for 2-LD-AG and 707–711% for 2-LD-BG, 4-LD-BG and control curdlan hydrogels), referred to their initial dry weight (Fig 1c) Therefore, the increased soaking capacity of poly(L-DOPA)-modified curdlan seems to be related to the specific mode (After Gelling) of L- DOPA monomer introduction into curdlan matrix

Chemical structure of the samples was tested by FTIR technique, in reference to pure L-DOPA monomer Pure poly(L-DOPA) was impossible

to be obtained in relevant polymerization conditions, neither in solution nor on glass slides This phenomenon, due to the high dispersity of poly (L-DOPA), was in agreement with previous reports (Bernsmann et al.,

2011; Hashemi-Moghaddam et al., 2018) The results of FTIR analysis of produced hydrogels were similar to those earlier obtained for PDA- modified curdlan hydrogels (Michalicha et al., 2021) First, no signs of

L-DOPA, manifested by the presence of N–H stretching (3393 cm− 1), N–H bending in primary amine (1526 cm− 1), C–N stretching (1282 cm− 1) and C–H out-of-plane stretching (816 cm− 1) bands, respectively (Mohanraj et al., 2013; Zhou et al., 2013), were detected in modified hydrogels (Fig 1b) This can be caused by low quantity of poly (L-DOPA) deposited within the curdlan hydrogel: 2 or 4 mg of L-DOPA monomer added to 400 mg of curdlan corresponds to less or equal to 1%

of total dry hydrogel weight Therefore this amount of poly(L-DOPA) distributed within the curdlan hydrogel matrix was probably too low to

be detected by FTIR technique Second, spectra of 2-LD-BG, 4-LD-BG and 2-LD-AG were similar to that of control curdlan This suggests the lack of chemical changes in basal curdlan hydrogel structure

Deeper insight into chemical structure of modified hydrogels

sam-ples was provided by XPS technique which was used to determine the

elemental composition and chemical state information of the surface

elements present in the investigated samples The control and modified

samples consisted mainly of C, O and N with smaller contribution of

elements such as Ca and Si Table 1 presents the calculated chemical

composition of the investigated samples The atomic ratio of C to O at

the surface of control sample was around 1.8 and differs significantly for

the values obtained for 2-LD-BG (C/O = 5.6) and 2-LD-AG (C/O = 2.2),

while remained similar for 4-LD-BG (C/O = 1.7) The nitrogen

contri-bution in control sample is approximately 0.3 at.% For the modified

samples the nitrogen was detected only for 2-LD-BG (0.1%) and 2-LD-AG

(1.3%)

High-resolution spectra of the C1s, O1s and N1s regions obtained for

control and modified samples are presented as Supporting Information

The binding energies and assigned functional groups for each peak

within the region envelope are shown in Fig S1 The percent

contri-bution for each functional group within the C1s, O1s and N1s regions for

control and modified samples are presented in Table 1 In general, the

peak energies of each component remained similar but the relative

in-tensities of each component varied, particularly for 4-LD-BG and 2-LD-

AG samples

The C1s region was fit with four peaks assigned to C–C/C–H, C–O/

C–N, C––O and O––C–OH species with the dominant peak being

attributed to C–O and C–C species The contribution of C–O species

varied from 59.3% to 64.4% with the highest value obtained for control

sample The C–C/C–H feature increased in relative intensity upon

modification (13.5–24.9%) comparing to control sample (13.2%) The

significant difference in contribution one can see for C––O and

O––C–OH components The C––O species were not detected for 2-LD-

AG, while the O––C–OH group were not present for 4-LD-BG (Table 1) The O1s region was fit with three peaks assigned to O––C, O–C and adsorbed water; except for 4-LD-BG for which water was not detected The contribution of each species differs depending on the modification conditions The dominant peak was attributed to O–C species or all samples under investigation It reached the highest contribution for the control sample (93.7%) and decreased for modified samples (82.7–87.1%) (Table 1)

The N1s region for control and 2-LD-BG samples presented one component which can be assigned to secondary amine (R–NH–R) The

N 1s spectrum of 2-LD-AG showed three contributions arising from primary (R-NH2), secondary (R–NH–R) and tertiary/aromatic (––N–R) amine functionalities According to the literature and probable reaction mechanism of poly(L-DOPA) formation (Azari et al., 2014; Manolakis & Azhar, 2020; Zhao et al., 2019), the primary amine is associated with L-DOPA and the secondary and tertiary amine is asso-ciated with both polymerization reaction intermediates and poly(L- DOPA)

Presence of nitrogen (0.3 at.%) and carboxyl groups (2.8% of C1s) in control hydrogel can be easily explained by 1.6% content of protein in curdlan which was confirmed by Hrom´adkov´a (Hrom´adkov´a et al.,

2003) Increase of nitrogen (1.3 at.%) and carboxyl groups (13% of C1s) content in 2-LD-AG sample results from the deposition of poly(L-DOPA)

on the surface of triple helical curdlan network However, the expla-nation of results for samples obtained according to BG method is more difficult, especially for 4-LD-BG For this particular sample, neither ni-trogen nor carboxyl groups were detected by XPS although stable black color of the sample confirmed the presence of poly(L-DOPA) in hydrogel There are two possible explanations of this phenomenon, due to the specific method of samples synthesis (in which L-DOPA monomer is

Fig 1 Description (a), FTIR spectra (b) and soaking capacity (c) of synthesized curdlan hydrogel samples LDOPA – L-DOPA monomer Soaking capacity was expressed as % of hydrogels weight increase

added to hydrated curdlan suspension before its thermal gelation at

93 ◦C; Section 2.1.1.) One of them is partial thermal decomposition of L-

DOPA resulting in a loss of carboxyl group (in particular in 4-LD-BG

sample) However, thermal studies on L-DOPA showed that this

com-pound is stable in the temperatures lower than 250 ◦C (Ledeti et al.,

2016) Second explanation is more probable and is a consequence of

hydrogel synthesis method Electron microscopy of aqueous curdlan

suspension revealed the presence of fibrillary structures, with

microfibrils of 10–20 nm in width (when heating temperature was

90 ◦C) or of 30–40 nm in width (when heating temperature was 120 ◦C) (Kasai & Harada, 1980) If L-DOPA was entrapped within these micro-fibrils and poly(L-DOPA) was formed afterwards (as in BG method), it could be undetectable for XPS technique (penetration depth of the X- rays is only few nanometers) In AG method, poly(L-DOPA) is more likely

to be deposited on the surface of curdlan microfibrils than inside them Thus, it is possible that steric location of poly(L-DOPA) within curdlan

Fig 2 Stability of control and poly(L-DOPA)-modified curdlan hydrogel samples: before (BW) and after (AW) extensive washing in water (a); after incubation (24 h and 96 h) in human serum or blood (b); after incubation (7 days) in Britton-Robinson buffers pH 2–12 (c)

Table 1

Chemical composition and XPS functional group percentages for control and modified curdlan hydrogels

Chemical composition in at %

XPS functional group percentages

hydrogels obtained by BG and AG methods is different and this may

affect their properties

Stability of poly(L-DOPA) deposits on curdlan hydrogels was studied

in different conditions First, we observed that the modification was

stable after very extensive wash in water (5 l of water in toto, 2 h or 12 h

long washes) (Fig 2a) Second, effect of human serum and fresh human

blood was verified, because wound dressings are likely to contact not

only with skin but also with serum or whole blood Both blood and

serum contain a variety of enzymes which may decompose poly(L-

DOPA) polymer However, 24 h and even 96 h incubation of poly(L-

DOPA)-modified hydrogels with these liquids did not cause any visible

change of polycatecholamine deposits, as indicated by stable black color

of the samples (Fig 2b) Serum and blood themselves did not visibly

change neither (data not shown) Thus, enzymatic pattern of these body

liquids seems not to affect the stability of poly(L-DOPA) modification

Third, impact of pH on poly(L-DOPA)-modification was tested in buffers

of pH ranging from 2 to 12 After 7 days of incubation, the modified

hydrogels seemed to be stable in each pH; however, the buffers

them-selves became slightly brownish in pH 10 and 12 (Fig 2c) This

sug-gested that applied modification was slightly unstable in pH 10–12 The

stability of poly(L-DOPA) deposits in acidic and physiological pH (2–8)

which include those typical for healthy and inflamed tissue, suggests

their possible applications in design of wound dressing biomaterials

Slightly higher lability of studied deposits in alkaline pH range, which

does not appear in physiological conditions in human body, does not

seem to disqualify poly(L-DOPA)-modified curdlan hydrogels for this

purpose

Contact between wound dressing biomaterials and human blood is

likely in wound treatment Thus, it is important to define the impact of

such biomaterials on blood compatibility PDA was earlier reported to

improve the blood compatibility of reduced graphene oxide

nano-particles in comparison with bare graphene oxide (Cheng et al., 2013) It

was therefore interesting to verify the blood compatibility of poly(L-

DOPA)-modified curdlan hydrogels First phenomenon which should be

studied is hemolysis which can be monitored by the release of free

he-moglobin (HB) from erythrocytes destroyed upon a contact with tested

materials As shown in Fig 3a, the level of HB released from the blood

incubated with poly(L-DOPA)-modified curdlan hydrogels is minimally

(but statistically significantly different) higher than negative control

test Second phenomenon concerning biomaterials-blood interaction is

clot formation which is an important factor in the process of wound

closure Clot formation may be monitored by the amount of HB released

from erythrocytes which were not entrapped within the clot network It

was observed that contact of blood with pristine (control) curdlan

caused a delay in clot formation: the amount of free HB for this sample

was 2–3 times higher than for normal clot (negative) Also, the clotting

process was not complete even after 45 min However, for all poly(L-

DOPA)-modified curdlan hydrogels the clot formation was at a higher level as for normal clot (negative) The obtained results are in agreement with reports that PDA coatings demonstrate good affinity to protein adsorption and thus may cause blood coagulation (Xie et al., 2017) In this particular case, poly(L-DOPA) deposits on curdlan hydrogels normalized the blood clot formation which was negatively affected by pristine curdlan and increased the blood compatibility of this hydrogel Wound dressings designed in this way may help to reduce bleeding and increase the rate of wound healing

The proposed application of poly(L-DOPA)-modified curdlan con-cerns the medical purposes, precisely the synthesis of wound dressings Therefore, fibroblast cell lines are the first choice cultures to verify the hydrogel cytotoxicity Reference cell line and primary culture of fibro-blasts were selected for this study Lack of cytotoxicity of tested samples was expected because previous reports concerning polydopamine coat-ings suggested that they are not only neutral for living cells and or-ganisms but may even reduce the in vivo toxicity of biomaterials (Hong

et al., 2011) However, MTT cytotoxicity test, which is an indicator of cell metabolism, showed that human skin fibroblasts (both BJ cells and HDFs) exposed to the extracts of the samples had significantly reduced metabolism compared to the control cells (PS control) (Fig 4a) BJ cells had cell metabolism reduced by 50% after exposure to 2-LD-BG, 4-LD-

BG, and 2-LD-AG and by 60% after treatment with control extract (pure biomaterial) Interestingly, sample extracts decreased metabolism of HDFs by approx 30% and in the case of 2-LD-AG extract by 20% It should be noted that according to ISO 10993-5 standard, results ob-tained with BJ cells indicated cytotoxicity of the samples, whereas re-sults obtained with HDFs their non-toxicity since cell viability was higher or equal to 70% (according to ISO, exposure to 100% extracts should not reduce cell viability by more than 30%)

Surprisingly, total LDH test, which determines cell number/biomass compared to the healthy control cells, revealed that only 4-LD-BG extract was cytotoxic to the fibroblasts as both BJ cells and HDFs had reduced cell number by approx 50% (Fig 4a) According to ISO 10993-

5 standard, 2-LD-BG extract was non-toxic because it decreased cell number to 85% (BJ cells) and to 70% (HDFs) Importantly, fibroblasts exposed to 2-LD-AG extract revealed cell biomass near 100%, proving its high safety and non-toxicity Obtained results suggested that extracts of the control, 2-LD-BG, and 2-LD-AG samples negatively affected cell metabolism, but were non-toxic since cell number was decreased by less than 30% compared to the control cells

Cell proliferation assay showed significant acceleration of fibroblasts proliferation only after their exposure to the extract of 2-LD-AG (Fig 4b) Extract of 2-LD-BG increased cell number after 24-hour exposure time, but further incubation in the presence of this extract resulted in cell number decrease Importantly, in the case of HDFs a significant decrease in cell number was observed after 72-exposure to

Fig 3 Hemolysis (a) and clot formation (b) in blood incubated with control and poly(L-DOPA)-modified curdlan hydrogel samples (a) & symbol indicate

statis-tically significant results according to One-way ANOVA with post-hoc Dunnett’s test in comparison with negative test (p < 0.005); value of positive test was reduced

10 fold; (b) *, # and ^ symbols indicate statistically significant results according to One-way ANOVA with post-hoc Dunnett’s test for 15 min., 30 min and 45 min.,

respectively, in comparison with negative tests for each time point (p < 0.0001)

the 2-LD-BG and 4-LD-BG extracts Surprisingly, although 4-LD-BG

extract showed cytotoxic effect towards BJ cells reducing their

meta-bolism and number by 50% (Fig 4a), it did not significantly affect

proliferation of these cells Observed discrepancy between cytotoxicity

and proliferation assays may be explained by growth phase of the

fi-broblasts during mentioned tests Cytotoxicity test was performed using

100% confluent culture of the cells, so fibroblasts were then in a

sta-tionary phase and thus were prone to dying and detachment from the

well surface Whereas proliferation assay was set up using low density

culture of fibroblasts that were in their Log phase characterized by

healthy and well-proliferating cells In the case of HDFs, results of

pro-liferation assay were consistent with cytotoxicity test

Fluorescent staining of the fibroblasts showed that surfaces of

con-trol, 2-LD-BG and 4-LD-BG were unfavorable to adhesion of BJ cells and

HDFs There were single, spherical cells on the surface of mentioned

samples (Fig 5a) Importantly, observed round cells were viable since

they exhibited green fluorescence Surprisingly, 2-LD-AG material did

not allow for adhesion of BJ cells, but was favorable to adhesion of

HDFs It should be noted that in the case of external wound dressing applications, biomaterial surface that is unfavorable to skin cell adhe-sion is very desired since it allows to remove or exchange the dressing without disruption of the wound bed (Vivcharenko et al., 2021; Wojcik

et al., 2021) Performed live/dead staining of fibroblasts grown around the tested samples showed their non-toxicity Nevertheless, it should be noted that ratio between culture medium and sample weight (1 ml of medium per 30 mg sample) in direct contact test was different than the one in cytotoxicity test (1 ml of medium per 100 mg sample) Materials did not support cell adhesion but cells cultured around the samples were healthy, viable, well-spread and with typical morphology (Fig 5b) Moreover no dead cells were detected around the samples However, there were noticeably fewer BJ cells around the 4-LD-BG material, which is consistent with cytotoxicity test results

Another model selected to evaluate the toxicity of synthesized

hydrogels was zebrafish The zebrafish (Danio rerio) is known to show

high molecular, physiological, genetic and immunological similarity to humans and is commonly accepted as a valid alternative to mammalian

Fig 4 Cell culture experiments performed with the use of sample extracts, normal human skin fibroblastcell line (BJ), and primary human dermal fibroblasts

(HDFs): (a) Cytotoxicity tests: MTT assay (cell metabolic activity) and total LDH assay (cell number/biomass); PS control – negative control of cytotoxicity revealing

100% viability; asterisks indicate statistically significant results according to One-way ANOVA with post-hoc Dunnett’s test (*** p < 0.001); (b) Cell proliferation

assay; PS control – cells exposed to non-toxic extract prepared by incubation of culture medium in a polystyrene vessel; asterisks indicate statistically significant

results according to One-way ANOVA with post-hoc Dunnett’s test (* p = 0.03, ** p = 0.006, *** p < 0.001)

models for evaluation of toxicity and biocompatibility of novel

bio-materials In our experiments, zebrafish embryos were maintained in

extracts collected from tested hydrogels obtained in the same medium-

to-hydrogel proportion that in the tests of fibroblasts cytotoxicity

Ob-servations made during 96 h post fertilization showed the lack of the

extracts influence on coagulation, tail detachment, development of

so-mites and heartbeat in the comparison with negative E3 control Also,

there were no incidence and description of morphological and

physio-logical abnormalities of zebrafish during 4 days of incubation in the

extracts (Fig 6a, b) Moreover, extracts of control, 2-LD-BG, 4-LD-BG

and 2-LD-AG curdlan hydrogels did not influence the locomotor activity

of 5 dpf zebrafish (one-way ANOVA: F(4,99) = 0.5283; p = 0.7152,

Fig 6c) Thus, we may exclude the neurotoxic effects of the hydrogels

Interestingly, the results of our experiments conducted on fibroblasts

(suggesting cytotoxicity of tested samples, at least in some of the per-formed tests) and on zebrafish eggs and larvae (showing the complete lack of toxicity and neurotoxicity of the samples) were not consistent However, it is known that results of biomaterials biological safety in different toxicity models may vary significantly For example, it was observed that gypsum (calcium sulfate dihydrate), a popular bone void filler, is highly cytotoxic in cell culture tests due to the quick dissolution and release of significant amounts of calcium ions (Przekora et al., 2014) but simultaneously it is claimed nontoxic in vivo and accepted for

commercial use (eg 3i® Calcium Sulfate Bone Cement, Biomet Inc.,

USA; FDA approval in 2001) Therefore, slight cytotoxicity of poly(L- DOPA)-modified curdlan (in particular, 2-LD-BG sample) is not a contraindication to use the hydrogel as biomedical material because it was simultaneously found safe in a sensitive zebrafish model

Fig 5 CLSM images after live/dead staining (green fluorescence – viable cells, red fluorescence – dead cells) of fibroblasts cultured in the presence of tested samples:

(a) Fibroblast adhesion to the surface of tested samples (PS control – healthy and well attached cells grown on the surface of polystyrene well); (b) Viability of fibroblasts cultured around the tested samples