Document [original]

RESEARCH ARTICLE

www.mrc-journal.de

Polyglycerol-Based Mucus-Inspired Hydrogels

Antara Sharma, Boonya Thongrom, Sumati Bhatia, Benjamin von Lospichl,

Annalisa Addante, Simon Y. Graeber, Daniel Lauster, Marcus A. Mall,

Michael Gradzielski,* and Rainer Haag*

The mucus layer is a hydrogel network that covers mucosal surfaces of the

human body. Mucus has important protective properties that are related to its

unique rheological properties, which are based on mucins being the main

glycoprotein constituents. Mucin macromolecules entangle with one another

and form a physical network that is instrumental for many important defense

functions. Mucus derived from various human or animal sources is poorly

deﬁned and thus not suitable for many application purposes. Herein, a

synthetic route is fabricated to aﬀord a library of compositionally deﬁned

mucus-inspired hydrogels (MIHs). MIHs are synthesized by thiol oxidation to

render disulﬁde bonds between the crosslinker ethoxylated

trimethylolpropane tri(3-mercaptopropionate) (THIOCURE ETTMP 1300) and

the linear precursors, dithiolated linear polyglycerol (LPG(SH)2)or

polyethylene glycol (PEG(SH)2) of diﬀerent molecular weights. The mixing

ratio of linear polymers versus crosslinker and the length of the linear polymer

are varied, thus delivering a library of compositionally deﬁned mucin-inspired

constructs. Their viscoelastic properties are determined by frequency sweeps

at 25 and 37 °C and compared to the corresponding behavior of native human

mucus. Here, MIHs composed of a 10:1 ratio of LPG(SH)2and ETTMP 1300

are proved to be the best comparable to human airway mucus rheology.

1. Introduction

Mucus is a complex, viscoelastic hydrogel that protects the mu-

cosal surfaces of the mammalian body, such as the respiratory,

A. Sharma, B. Thongrom, S. Bhatia, R. Haag

Institut für Chemie und Biochemie

Freie Universität Berlin, Takustraße 3, Berlin 14195, Germany

E-mail: [email protected]

B. von Lospichl, M. Gradzielski

Institut für Chemie

Technische Universität Berlin

Straße des 17. Juni 124, Berlin 10623, Germany

E-mail: [email protected]

The ORCID identiﬁcation number(s) for the author(s) of this article

can be found under https://doi.org/10.1002/marc.202100303

by Wiley-VCH GmbH. This is an open access article under the terms of

the Creative Commons Attribution-NonCommercial License, which

permits use, distribution and reproduction in any medium, provided the

original work is properly cited and is not used for commercial purposes.

DOI: 10.1002/marc.202100303

gastrointestinal, reproductive, and oculo-

rhino-otolaryngeal tract surfaces. It is in-

volved in a diverse range of functions as

a result of its unique properties. It pre-

vents dehydration of the mucosa by main-

taining a hydrated layer on the epithelial

surface,[1] and exhibits shear-thinning rhe-

ological behavior that supports its lubri-

cant role by warranting a smooth passage

for passing objects.[2] In the respiratory

tract, mucus entraps inhaled pathogens, al-

lergens, and irritants that are constantly

removed by mucociliary clearance driven

by ciliary beating on airway surfaces fa-

cilitating constant mucus clearance from

the lungs.[3] Furthermore, the mucus layer

serves as the external defense system of

the body against pathogen attack in mu-

cosal immunology.[1,4,5] The mucus coat-

ing is a complex heterogenous biopoly-

mer matrix comprising water (up to 97%

by weight), mucin (typically about 5% by

weight or less), inorganic salts (≈1% by

weight), carbohydrates, proteins, antimicro-

bial peptides, and lipids. The physical be-

havior and resultant protective functions

of mucus are attributed to its constituents, particularly mucin

glycoproteins,[6] which represent more than 80% of the overall

organic composition.[7] Mucin macromolecules contain regions

of high glycosylation. The sulfate ester and sialic acid motifs

A. Addante, S. Y. Graeber, M. A. Mall

Department of Pediatric Respiratory Medicine, Immunology and Critical

Care Medicine

corporate member of Freie Universität Berlin and Humboldt Universität

zu Berlin

Charité – Universitätsmedizin Berlin

Berlin 13353, Germany

A. Addante, S. Y. Graeber, M. A. Mall

Berlin Institute of Health at Charité – Universitätsmedizin Berlin

Berlin 10178, Germany

A. Addante, S. Y. Graeber, M. A. Mall

associated partner site

Deutsches Zentrum für Lungenforschung e. V.

Aulweg 130, Gießen 35392, Germany

D. Lauster

Institut für Chemie und Biochemie

Fachbereich Biologie, Chemie, Pharmazie

Freie Universität Berlin

Arnimallee 22, Berlin 14195, Germany

www.advancedsciencenews.com www.mrc-journal.de

decorate the glycosyl end groups, which in turn act as decoys

for various virus families. In addition, the network structure of

mucus has an equally important role in mucosal immunology.[8]

The long, ﬁbrillar structure of the mucin molecules enables

intermolecular entanglement as well as various covalent and

electrostatic interactions to form a gel matrix. The network struc-

ture provides multiple obstructions in the path of an incoming

pathogen and prevents it from reaching the epithelial cell surface

by essentially aﬀecting its motility.[8,9]

The invasion by pathogens to initiate the infection process is

typically initiated by a compromised mucus barrier and thus the

subsequent proliferation of infection. Respiratory viruses, such

as the inﬂuenza A virus and SARS-CoV2 must pass through the

mucus barrier to access the underlying epithelia and initiate the

viral life cycle.[2,4,10]

However, naturally occurring mucus is ineﬀectual for antivi-

ral purposes due to its poorly deﬁned composition, the chal-

lenges of extracting of suﬃcient amounts, the arduous puriﬁca-

tion required, and performance deviation as a result of a source-

dependent composition.[11] The use of synthetic mucus oﬀers a

controlled pathway to overcome such challenges and can, there-

fore, help to understand the role of diﬀerent mucus compositions

and functional properties in health and disease. The molecular

structure, polymer length, and viscoelastic behavior can be pre-

cisely controlled while oﬀering a reduction in heterogeneity typ-

ical to natural mucus.

Synthetic platforms for surface-tethered mucins have been

designed and studied by Kramer et al.[12] An analog of the

peptide backbone of a mucin molecule bearing the N-GalNAc

branching points native to mucus was constructed using N-

carboxyanhydride polymerization. However, this model is de-

signed speciﬁcally for transmembrane mucin modeling rather

than application purposes. Mahalingam et al. devised an elegant

platform for the fabrication of a mucus gel network, wherein

the boronate-diol condensation reaction was employed to induce

a stimuli-responsive network against HIV infection.[13] This gel

mimics cervicovaginal mucus, rather than presenting a broader

mucus-mimetic approach. Other synthetic formulations have

been explored which exploit the adhesive and lubricating prop-

erties of the gel.[14] With the multivalent presentation of hydroxyl

groups on MIHs, as well as the usage of a disulﬁde linkage,

the gels will be even more customizable and closer to natural

mucus.

The aim of the present work was to synthetically fabricate lin-

ear polyglycerol (LPG) and polyethylene glycol (PEG)-based gel

systems that mimic certain characteristics of gel-forming mucus,

with the goal of attaining rheological characteristics approaching

those of naturally occurring gel-forming mucus of healthy hu-

man sputum.

In order to synthetically combine these properties, a highly wa-

ter soluble, biocompatible, multivalent, and linear core scaﬀold

was provided by thiolated linear polyglycerol[14–16] (LPG(SH)2).

The multivalent backbone provided ample opportunity for cre-

ating a burgeoning synthetic glycosyl domain, akin to mucin.

These molecules may be functionalized as required, thereby fa-

cilitating a certain inﬂuence on the rigidity. Mucus-inspired hy-

drogels (MIHs) were then realized by reacting these molecules

with ethoxylated trimethylolpropane tri(3-mercaptopropionate)

ETTMP 1300 crosslinkers by thiol oxidation to form disul-

ﬁde bonds, whereby their rheological properties were measured

(Figure 1). Thus, by an amalgamation of facile techniques, a li-

brary of rheologically diverse, compositionally deﬁned mucin-

inspired hydrogels was developed. This model not only ap-

proaches the structural properties of natural mucus but also at-

tains some of its primary functions without the hassle of lengthy

and diﬃcult reactions.

2. Results and Discussion

2.1. Design and Synthesis of Building Blocks

The network structure in natural mucus arises due to the phys-

ical and chemical properties imparted by mucin glycoproteins.

There are cysteine end groups on every mucin glycoprotein that

allow the long, ﬁbrillar molecules to elongate further, resulting in

the formation of a mucin “chain.”[1,7] This results in the physical

entanglement of the chains and a network structure arises. The

crosslinking is further secured as the chains also interact with

one another by hydrophobic and electrostatic means.[2,15]

In order to develop components for MIHs, an ideal model

would be one which allows easy modulation along the lines of

viscoelasticity and pore size, while oﬀering multiple backbone

functionalization groups to enable multivalent sugar modiﬁca-

tions. A chemically crosslinked hydrogel requires a linear struc-

ture interspersed with crosslinking points, in ratios that enable

a target ﬂexibility. A linear precursor in the MIH network would

correspond to the mucin glycoprotein structure in naturally oc-

curring gel mucus, exhibiting structural and functional similar-

ity to it. It would incorporate reactive end groups to enable chain

elongation and reaction with crosslinking agents.

The nature of crosslinking in naturally occurring gel mucus is

relevant in determining the rheological properties of the network.

Individual mucin molecules are linked to each other by disulﬁde

bonds.[7] Thus, such a reversible, redox-responsive bond should

be present in the MIH system. Furthermore, the components

should be able to react with each other under mild conditions

at room temperature.

With these tools, a library of MIHs has been created with rheo-

logical properties, network mesh sizes, and backbone structures

that approach the corresponding traits of naturally occurring mu-

cus. Homobifunctionalized linear polyglycerol chains with thiol

groups at both ends, i.e., LPG(SH)2, were prepared as analogs

of mucin glycoproteins in MIH networks. Moreover, the LPG

backbone structures display multiple hydroxyl groups that equip

these models for functionalization with various sugars or other

functional groups found in mucin to enable future modiﬁca-

tion when necessary. The structural properties of MIHs can be

regulated by altering the length of the LPG backbone or by re-

placing it with another linear structure altogether, thus facili-

tating rheological comparisons on the basis of backbone struc-

ture and rigidity. For this purpose, an LPG(SH)2backbone of

molecular weight 5 kDa was designed (Scheme 1). For com-

parison, two more MIH families were synthesized with linear

dithiofunctionalized polyethylene glycol PEG(SH)2of molecular

weights3and6kDa(Scheme 2).

www.advancedsciencenews.com www.mrc-journal.de

Figure 1. A) MIH-1a containing iodine traces; B) after iodine removal by PBS washing; C) the decrease in elasticity of MIH-1b with increasing linear

component becomes visible by eye; D) the hard, coarse, and brittle solid formed upon oxidation of pure ETTMP; E) (right) scheme representing the

network structure of MIHs composed of ETTMP 1300 and LPG(SH)2; chemical structures of ETTMP 1300 (left) and LPG(SH)2(below).

Scheme 1. Synthesis of LPG(SH)21e.

Scheme 2. Synthesis of PEG(SH)2(2c and 3c).

www.advancedsciencenews.com www.mrc-journal.de

2.1.1. Synthesis of Dithiolated LPG(SH)21e

The bromine-capped linear polyethoxyethyl glycidyl ether

(LPEEGE-Br) was synthesized by anionic ring-opening polymer-

ization following the procedure reported in the literature.[16]

LPG(SH)2was synthesized in three steps. LPEEGE-Br with

molecular weight 10 kDa 1a was hydrolyzed under basic

conditions (pH >10), resulting in the formation of linear

polyethoxyethyl glycidyl ether (LPEEGE-(OH)21b with two

terminal hydroxy groups. This was followed by a mesylation

reaction. The mesylated product 1c was reacted with thiourea to

obtain LPEEGE dithiourea 1d as the substitution product. The

nitrogen and sulfur contents were estimated by elemental analy-

sis to corroborate the degree of conversion. The ﬁnal compound

LPG5(SH)21e was obtained after acidic and basic hydrolysis

of ethoxyethyl and thiourea groups, respectively (Scheme 1).

Tris(2-carboxyethyl)phosphine (TCEP) was added to this mixture

as a reducing agent. Subsequent dialysis led to pure products. 1H

NMR analysis at each step corroborated the formation and purity

of the compound. The presence of 2 thiol groups per chain was

conﬁrmed by elemental analysis and 1H NMR studies.

2.1.2. Synthesis of PEG(SH)22c and 3c

Dithiolated PEGs with molecular weights 3 kDa PEG3(SH)2and

6 kDa PEG6(SH)2were synthesized following an approach mod-

iﬁed from the original report from Mahadevegowda and Stu-

paru (Scheme 2).[17] First homo-bifunctional dihydroxy PEGs,

i.e., PEG3(OH)22a and PEG6(OH)23a were mesylated under an-

hydrous conditions. PEG3(OMs)22b and PEG6(OMs)23b were

obtained as a white powder upon precipitation of the reac-

tion mixture. Thiolation of the thus obtained PEG(OMs)2was

then performed with thiourea under elevated temperature, fol-

lowed by basic hydrolysis and reduction by TCEP to ﬁnally yield

PEG3(SH)22c and PEG6(SH)23c as pale yellow powder after pre-

cipitation.

2.2. Synthesis of Mucus-Inspired Hydrogels (MIHs)

In order to assess whether hydrophilic linear polyglycerols can

provide mucus-like hydrogels with matching rheological charac-

teristics, a commercial homotrifunctionalized PEG thiol, ETTMP

1300 was used as the crosslinker. Gelation was induced by simply

oxidizing the thiol groups present on both of the macromonomer

components in the presence of hydrogen peroxide and NaI solu-

tion in water.[18] A 3D matrix was consequently formed with thiol-

disulﬁde interchange reactions facilitating certain characteristic

rheological properties of such networks.

Several MIH families were synthesized and compared on the

basis of their structure as well as the molecular weight of the

linear scaﬀold. Thus, three gel families were assembled using

LPG(SH)2and PEG(SH)2of varying lengths as linear precursors:

LPG5(SH)21e, PEG3(SH)22c, and PEG6(SH)23c. Furthermore,

the ratio of the crosslinker to the mucin analog component was

varied while maintaining a constant gel volume of every gel fam-

ily, allowing the rheological properties to be ﬁne-tuned.

In Table 1, all the studied gels are listed with their correspond-

ing synthetic details. Thiol oxidation is a sensitive procedure,

Table 1. Description of various MIH gel series studied. The gels are clas-

siﬁed by series depending upon the linear precursors of which they are

composed. Diﬀerent types of MIHs within each series are fabricated by

changing the ratios of ETTMP 1300 crosslinker to the linear component,

thus rendering a total of 12 gels. A total gel volume of 200 μL was ﬁxed for

every composition.

MIH series Linear precursor Ratio of ETTMP 1300 crosslinker to linear

component

1:3 1:7 1:10 1:14

MIH-1 LPG5(SH)2MIH-1a MIH-1b MIH-1c MIH-1d

MIH-2 PEG3(SH)2MIH-2a MIH-2b MIH-2c MIH-2d

MIH-3 PEG6(SH)2MIH-3a MIH-3b MIH-3c MIH-3d

with the possibility of oxidation side products such as sulfenic,

sulﬁnic, and sulfonic acids.[19] These side products may have in-

hibitory eﬀects on gel formation. To circumvent this problem, the

amount of oxidizing agent needs to be slightly decreased with in-

creasing linear component amounts in order to account for the

disparity in the overall thiol content.

The gelation process could be visually monitored as the color of

the pre-gel mixture deepened from a cloudy white solution to an

orange-brown semi-solid upon oxidation. The gels were kept for

some time so that an equilibrium state could be attained. There-

after rheology studies were conducted. The color transition in-

dicates formation of I2(Figure 1A). In order to remove the io-

dine side product from the gels, all MIHs were incubated for 24

h in phosphate-buﬀered saline (PBS) solution. A change of color

thus followed, from the iodide-characteristic orange color to an

opaque, white gel, which otherwise retained all other properties

(Figure 1B).

2.3. Rheology

2.3.1. Oscillatory Rheology

The mechanical properties of a gel can be deﬁned by its viscoelas-

tic properties, which in turn can be determined by oscillatory

rheology measurements. A study of the viscoelastic properties of

the LPG- and PEG-based scaﬀolds was done with the aim to es-

tablish relationships between the viscoelastic properties of MIHs

and their respective constitution. The results thus obtained were

compared with corresponding values in naturally occurring mu-

cus to identify systems with similar rheological properties.

The speciﬁcations of the LPG-based MIH-1 gels, and PEG-

based MIH-2 and MIH-3 gels studied are listed in Table 1.

The gels compared were based on three scaﬀolds—LPG5(SH)2,

PEG3(SH)2, and PEG6(SH)2. The ratio of the linear precursor to

the crosslinking component, ETTMP 1300, was varied in each

case. The concentration of the crosslinker was decreased steadily

while maintaining a constant amount of polymer dithiols used

overall. This then should lead to increasingly less crosslinked hy-

drogel networks.

The rheological properties of all the hydrogels were charac-

terized in terms of the storage G’ and loss G’’ moduli obtained

from oscillatory shear experiments. First, a strain-sweep test

was conducted to determine the linear viscoelastic region (LVE).

www.advancedsciencenews.com www.mrc-journal.de

Figure 2. Storage (G’) and loss (G") modulus as function of the radial

frequency 𝜔of MIH-2 A) at 25 °C; B) at 37 °C, for samples where the

crosslinker density was varied systematically.

Then oscillatory rheology experiments, speciﬁcally oscillatory

shear and strain tests, were conducted in the LVE to determine

G’ and G" as functions of radial frequency 𝜔. The oscillatory

rheology results for hydrogels with PEG3(SH)2and PEG6(SH)2

are shown in Figures 2 and 3, respectively; data for gels with

LPG5(SH)2are shown in Figure 4. Each test was conducted at

25 °C and at the physiological temperature of 37 °C.

For the PEG-based compounds (MIH-2 and MIH-3), a rather

linear increase of G’ and G’’ can be seen in a double-logarithmic

plot, where for the longer PEG spacer (MIH-3) G’’ is typically

higher than G’, which indicates that these systems are dominated

by their viscous properties. This also indicates that these systems

are not really proper gels, but at best can be classiﬁed as soft

and gel-like, i.e., a viscoelastic ﬂuid. They are not permanently

crosslinked and are relaxing under a given stress proportional

to the time left for this relaxation. Interestingly, with increasing

temperature both moduli increase substantially (by more than a

factor 10) and the increase is larger for G’, i.e., the relative elas-

tic properties of these gels increase with rising temperature. As

Figure 3. Storage (G’) and loss (G") modulus as function of the radial

frequency 𝜔of MIH-3 A) at 25 °C; B) at 37 °C, for samples where the

crosslinker density was varied systematically.

stated before, this behavior may be attributed to the fact that PEG

becomes less well soluble in water with increasing temperature

and therefore the network structure would be contracting. This

slope of the frequency increase was higher for G’ (≈𝜔1.2,at37

°C) than for G" (≈𝜔0.85, at 37 °C) and accordingly with increas-

ing frequency the gap between G" and G’ decreases. This sug-

gests that the crossover point where G’ would overtake G" would

take place at frequencies higher than the frequency range applied

here, indicating a rather short structural relaxation time of less

than 0.02 s.

Generically similar is the behavior of the PEG-based systems

with the shorter PEG spacer (MIH-3), but here the relative in-

crease of G’ with rising temperature is even much more marked.

Upon going from 25 to 37 °C, it increases by a factor of 30–200,

while G" only increases by a factor of 3–5. This also means that

at low temperature the systems are dominated by viscous prop-

erties, but this is reversed at 37 °C, where the elastic properties

become dominant. Interestingly, the scaling of both G’ and G" at

37 °C is about ≈𝜔0.75−0.8.

www.advancedsciencenews.com www.mrc-journal.de

Figure 4. Storage modulus G’ and loss modulus G" as functions of the ra-

dial frequency 𝜔of A) MIH-1 at 25 °C; B) MIH-1 at 37 °C, for samples where

the crosslinker density was varied systematically. For comparison, data of

fresh airway mucus obtained from healthy donors were also included in

(A) (averaging over three diﬀerent samples).

In contrast, for LPG5(SH)2-based MIH-1 gels, both moduli, G’

and G’’, remained rather constant, with G’ being much larger

than G’’, i.e., they behaved like typical gels with a yield stress.

For instance, for the sample with thrice the linear component as

compared to the crosslinker (MIH-1a), G’ remained about two

orders of magnitude higher than G" throughout the frequency

range, thus showing gel-like behavior. For the other gels with

less crosslinker, G’ remained also constant, but at much lower

frequency values. Compared to the PEG-based gels the increase

of viscoelastic parameters is much less pronounced here.

Even more diﬀerent is the behavior of G" that increases

markedly with rising frequency in the higher frequency range

and there even bypassing G’, although also G’ starts to increase

somewhat for frequencies above 10 Hz. Such behavior is not very

frequently seen, but has for instance similarly been reported for

carrot puree.[20] Generally, it indicates a faster relaxation process

being present within the system.

Table 2. Shear modulus G0and estimated mesh size 𝜉of LPG5(SH)2-based

MIHs at 25 and 37 °C.

MIH [crosslinker]/[linear

component]

25 °C 37 °C

Shear

modulus

Mesh

size

Shear

modulus

Mesh

size

[G0/Pa] [𝜉/nm] [G’/Pa] [𝜉/nm]

MIH-1a 1:3 1157 15 1112 15

MIH-1b 1:7 11723352

MIH-1c 1:10 8 82 19 61

MIH-1d 1:14 0.2 286 69 40

A direct comparison of the rheological properties of healthy

human airway mucus with MIHs can be very interesting in this

context. Previous experiments have shown that the shear modu-

lus of mucus in healthy lungs is ideally around 1–2 Pa.[21] To com-

pare synthesized MIHs with human mucus, rheological proper-

ties of healthy airway mucus from three individual donors were

determined and are included in Figure 4. It can be seen that hu-

man mucus also shows such an upturn above 10 Hz in a sim-

ilar fashion for G’ and G", while at lower frequency a constant

plateau is observed, with G’ ∼2 Pa. It is interesting to note that

the rheological properties of mucus are rather insensitive to the

temperature change from 25 to 37 °C (while the synthetic gels

generally show some increase especially of the storage modulus,

which can be attributed to the eﬀect of entropy elasticity for cova-

lently bound networks). Comparing the data in Figure 4, one can

conclude that MIH-1 with a ratio of crosslinker to linear compo-

nent in the range of 1:10 to 1:14 is showing rather similar rheo-

logical behavior as healthy airway mucus. Thus, the rheological

data show that tunable hydrogels can be constructed, where stor-

age and loss moduli, and to some extent even their frequency de-

pendence, can be controlled via the ratio of the linear component

to the crosslinking agent.

The marked decrease of the elastic properties with decreasing

content of crosslinker could be attributed to a decreasing num-

ber of crosslinking points within the gel network, but given the

large decrease it would not just mean less crosslinker but also

a substantially reduced crosslinking eﬃciency. In order to quan-

tify this further, we determined the shear modulus G0as the av-

erage of G’ in the plateau regime and the corresponding values

are given in Table 2. Especially at lower crosslinker content, an

increase in temperature had a similar but smaller eﬀect on the

LPG-based MIHs as observed for the PEG-based MIHs, i.e., the

elastic properties increase at the higher temperature. An explana-

tion for that behavior would be that the linking polymer chains

are less well dissolved in water with increasing temperature and

therefore they swell less and correspondingly their crosslinking

density increases. The eﬀective network mesh sizes 𝜉were then

estimated from the shear modulus G0from the simpliﬁed rela-

tion: G0=kT/𝜉3andtheyaregiveninTable2.

[22]

The mesh size 𝜉of a network (which is the inverse of the

crosslinking density) is directly linked to its rheological prop-

erties. The rough estimate obtained from the shear modulus

G0(Table 2, Figure 5) yields 𝜉values in the range of 15–80 nm,

which compares well with the experimentally observed values of

www.advancedsciencenews.com www.mrc-journal.de

Figure 5. Shear modulus G0for the MIH-1 gels as a function of the ratio

of crosslinker to linear component, measured at 25 and 37 °C.

Table 3. Crosslinking densities obtained by Equation (1).

MIHtype 25°Cμmol cm−337 °C μmol cm−3

MIH 1a 0.44 0.45

MIH 1b 0.0099 0.0043

MIH 1c 0.0062 0.0023

MIH 1d 0.021 0.000072

20–200 nm for naturally occurring mucus.[23] While the data

diﬀer depending upon the source and method of examination, a

mesh size in the range of 100 nm up to even a few micrometers

is the norm.[24] This observation also aligns with the inferred

rheological data. As shown in Table 2, for MIH-1c the eﬀective

mesh sizes at physiological temperature were about 60 nm

and thus in the same range as the desired size range. With the

tunable design of such systems, much larger pore sizes can be

achieved by substitution of backbone structures or crosslinkers

of diﬀerent lengths.

The eﬀective crosslinking density 𝜌can be estimated from the

rheological data via Equation (1):

𝜌=

G′

plateau

R.T(1)

This equation is based on the plateau modulus G’plateau ob-

tained during frequency sweep experiments, where Ris the uni-

versal gas constant, and Tis the temperature. The eﬀective rheo-

logical crosslinking density is simply the inverse of the mesh size

given in Table 2, and listed in Table 3.

The observed eﬀective crosslinking density decreases substan-

tially with lowering the content of the crosslinker. For compari-

son one can also calculate the crosslinking density, as it would

be expected for perfect polymerization, and for this assumption

one obtains values within the limits of 2 – 20 μmol cm-3,de-

pending on the gel composition. This indicates that the chem-

ical crosslinking is only eﬀective to a rather low extent and corre-

spondingly one has a much more open structure than would be

present for complete crosslinking. In addition, one has to note

that Equation (1) is just a general relation and depending on the

precise topology of the network one would have a prefactor dif-

ferent than one, which to a lesser extent may also attribute to the

observed discrepancy of values.

3. Conclusion

In this work, the endeavor was to create such a mucus-inspired

hydrogel in a facile, low-cost manner with a nature-inspired disul-

ﬁde crosslinking chemistry. For this purpose, LPG and PEG

chains of diﬀerent molecular weights were utilized to produce

dithiol-functionalized chains, which in turn were gelled with

ETTMP 1300—a tri-PEG crosslinker, in the presence of an oxi-

dizing agent. Of the formed gels, the most appropriate for the in-

tended purpose was the MIH-1c gel based on a LPG-dithiol with

an Mn of 5 kDa. At lower frequencies, the gel showed elastic-

dominant behavior, with G’ falling in the range of 5–10 Pa at

25 °C. As for all other gels, the G’ increased to 15–20 Pa in the

same frequency range as the temperature was increased to 37 °C.

Thus, we successfully demonstrated the synthesis of mucus-like

hydrogels with a range of rheological properties and achieved

with MIH-1c a hydrogel that has rheological properties compa-

rable to native human airway mucus.

4. Experimental Section

Materials:All chemicals were obtained from Merck (Darmstadt, Ger-

many) and Acros Organics (Geel, Belgium), and used without any puriﬁ-

cation. ETTMP 1300 was received as a gift from Bruno Bock (Marschacht,

Germany) and used as such. Moisture-sensitive reactions were performed

under anhydrous conditions, in ﬂame-dried glassware. Reﬂux reactions

were performed at 115 °C in an oil bath. Dialysis was performed in Spectra

Por dialysis tubing (MWCO =2000 g mol−1) (Carl Roth GmbH, Karlsruhe,

Germany). The dialysate was changed ﬁve times within 2 days. 1HNMR

spectra were recorded in AV 500 spectrometer (Bruker, Massachusetts,

USA). NMR chemical shifts were reported as 𝛿values in ppm. LPEEGE

with average molecular weights of 10 kDa (LPEEGE10) were synthesized

by following procedures reported in literature.[15]

Methods:Synthesis and Characterization of Dithiolated Linear Polyglyc-

erol LPG(SH)2:Synthesis of LPEEGE 1a: LPEEGE-Br was synthesized as

described previously.[16] LPEEGE10Br 1a (2 g, 2.38 ×10−2mmol) was

dissolved in a mixture of MeOH (18 mL) and H2O (3 mL). Sodium

methoxide (30 wt%, 200 μL, 0.88 mmol) was added to the solution

and stirred overnight at room temperature. The crude LPEEGE product

was then puriﬁed by dialysis against acetone (2 days) to remove excess

sodium methoxide as well as to remove methanol and water to aﬀord pure

LPEEGE10(OH)21b (1.9 g, 95% yield).

Synthesis of LPEEGE10 dithiourea 1d: Thoroughly dried LPEEGE10(OH)2

1b (1.9 g, 2.4 ×10−2mmol, 1 eq.) was dissolved in anhydrous dimethylfor-

mamide (DMF, 250 mL), followed by the addition of triethylamine (TEA,

25 μL, 1.3 mmol, 7.5 equivalent) to the stirring solution. The reaction ﬂask

was cooled over an ice bath, and subsequently a solution of MsCl (5.5 μL,

7.14 ×10−2mmol, 3 eq.) in DMF (1.5 mL) was added in a dropwise man-

ner to the vigorously stirring solution. The reaction mixture was stirred

over an ice bath for 1 h, after which the solution was allowed to attain room

temperature and continued to be stirred for 48 h. After solvent removal, the

crude product was dialyzed against a 3:1 chloroform–methanol mixture

for 24 h. The resultant yellow-orange-colored LPEEGE10(OMs)2(1c) had

a honey-like consistency (yield: 50%) which was used for the dithiourea

formation without any delay.

In order to aﬀord thiol-functionalized LPEEGE10(SH)2,

LPEEGE10(OMs)21c (1.9 g, 2.4 ×10−2mmol) was dried under vac-

uum, and then dissolved in mixture of 1-propanol (200 mL) and MeOH

www.advancedsciencenews.com www.mrc-journal.de

(10 mL). The reaction was reﬂuxed for 48 h after thiourea (73 mg, 9.52

×10−1mmol, 4 eq.) was added. 1-propanol was evaporated to aﬀord

crude LPEEGE10 dithiourea 1d, which was then puriﬁed by dialysis using

methanol as a solvent. Its constitution was conﬁrmed by elemental

analysis; N =0.716; C =46.66; S =0.297; H =7.578 (≈2 thiol groups per

chain). LPEEGE10 dithiourea 1d, 95% isolated yield, 1H NMR (500 MHz,

MeOD, 𝛿(ppm)): 1.17–1.30 (m, 11H), 3.88–4.11 (m, LPG backbone),

7.91 (s, 4H).

Synthesis of LPG5(SH)21e: An acidic solution of ethanol was prepared

by mixing HCl in ethanol such that the solution had a pH of 5. LPEEGE10

dithiourea 1d was dissolved in 150 mL of the prepared acidic solution and

then stirred overnight, after which the pH of the solution was brought up

to 8 using 1 m NaOH solution. Sodium hydroxide (4 mg, 0.1 mmol, 4

eq.) was added to an aqueous solution of the crude product, which was

then reﬂuxed for 48 h. The temperature of the reaction ﬂask was brought

down to room temperature. After the reaction, crude product was dialyzed

against deionized water for 72 h to ﬁnally obtain pure LPG(SH)21e in

≈95% isolated yield which was stored in the presence of TCEP.

Synthesis and Characterization: Dithiolated PEG (PEG3(SH)2): Synthesis

of Dimesylated PEG (PEG(OMs)2): Dry PEG6(OH)23a (20 g, 3.3 mmol, 1

eq.) was dissolved in anhydrous dichloromethane (DCM, 100 mL). TEA

(2.77 mL, 20 mmol, 6 equiv.) was injected in this solution. The reaction

ﬂask was then cooled over ice bath followed by subsequent dropwise ad-

dition of methane sulfonyl chloride (1.03 mL, 13.3 mmol, 4 equiv.). The

reaction was allowed to run for 1 day. Afterward the crude product was puri-

ﬁed by ﬁrst washing the reaction mixture thrice with brine, drying the DCM

layer, and then precipitating the resultant product in cooled diethylether.

The precipitate so obtained was collected and dried in vacuo over 24 h to

obtain PEG(OMs)2as a white colored powder.

PEG3(OMs)22b: 90% isolated yield, 1H NMR (700 MHz, CDCl3,𝛿(ppm)):

3.07 (3H, s), 3.63–3.76 (m), 4.36 (2H, t, J=7Hz)

PEG6(OMs)23b: 95% isolated yield, 1H NMR (500 MHz, CDCl3,𝛿(ppm)):

3.07 (3H, s), 3.48–3.78 (m), 4.37 (2H, t, J=5Hz)

Synthesis of Dithiolated PEG (PEG(SH)2): To a solution of dimesylated

PEG (PEG6(OMs)23b (19 g, 3.2 mmol, 1 eq.) in 1-propanol (100 mL),

thiourea (1.02 g, 13.3 mmol, 4 eq.) was added and the solution was re-

ﬂuxed for 24 h to obtain diisothiouronium PEGintermediate. 1-propanol

was then removed from the crude mixture and NaOH (0.53 g, 13.3 mmol,

4 eq.) and water (100 mL) were added. This solution was then reﬂuxed for

24 h. Afterward, the crude mixture was neutralized to pH 7. TCEP (1.67 g,

6.7 mmol, 2 eq.) was consequently added, and the reaction was stirred for

a further 2 h to obtain the crude ﬁnal PEG-dithiol product. The product

was puriﬁed following a precipitation procedure. To this end, NaCl was

ﬁrst added to the reaction mixture until the saturation point was attained.

The product was then extracted with DCM. Thereafter, the DCM layer was

dried with Na2SO4. Finally, the pure product was precipitated in cooled

diethylether followed by drying in vacuo so as to obtain PEG (SH)2as a

pale yellow colored powder.

PEG3(SH)22c: 79% isolated yield, 1H NMR (600 MHz, CDCl3,𝛿(ppm)):

1.59 (1H, t, J=6 Hz, 12 Hz), 2.69 (2H, quat, J=6 Hz, 12 Hz), 3.51–3.76

(m). Elemental analysis; N =0.24; C =53.43; S =2.43; H =8.52

PEG6(SH)23c: 88% isolated yield, 1H NMR (500 MHz, CDCl3,𝛿(ppm)):

1.59 (1H, t, J=5 Hz, 10 Hz), 2.69 (2H, quat, J=5 Hz, 10 Hz), 3.48–3.78

(m). Elemental analysis; N =0.13; C =54.24; S =2.02; H =8.47

Synthesis of MIHs: LPG-based and PEG-based gels were both fabricated

in a manner that resulted in gels with the same volumes. Four families

of gels were prepared, diﬀering only by their linear components, namely,

LPG5(SH)2, PEG3(SH)2, and PEG6(SH)2, listed as gel series MIH-1,MIH-

2,andMIH-3 in Table 2, respectively. Furthermore, the crosslinker to linear

component ratios within each series was varied in order to zero in on the

gels with the most promising rheological properties. Three such ratios for

each series were inspected, and thus a total of 12 gel types were ﬁnally

examined, the speciﬁcations of which are also listed in Table 2.

First, separate 50% w/v aqueous solution of the linear component was

prepared in water. Similarly, a 50% w/v solution was also prepared for

ETTMP 1300. A 10 ×10−3m NaI solution was freshly prepared. The two

gel components were then mixed in an Eppendorf tube, in amounts as

listed in Table 2, with the ﬁnal crosslinker to backbone ratios being 1:3,

1:7, 1:10, and 1:14, respectively. NaI and H2O2were added consecutively.

This cloudy pregel solution was then vortexed until a color change from

white to yellow was observed indicating the release of iodide and conse-

quent formation of the tri-iodide ion. The color deepened within the next 2

min to ﬁnally develop into an orange-brown semi-solid. The gel was incu-

bated at room temperature overnight in a closed container, and rheology

tests were conducted thereafter.

The saturated yellow color of the gel was a leftover of the oxidation

process, and easily be removed by incubation of the formed gel in PBS

buﬀer for 24 h to obtain a white gel (Figure 1B). All gels were synthesized

following the same gelation procedure.

Human Lung Sputum Samples: Collection of airway mucus samples

from healthy donors was approved by the ethics committee of the Charité

– Universitätsmedizin Berlin and written informed consent was obtained

from all participants. All donors were healthy nonsmokers. Lung sputum

was collected after induction by inhalation of hypertonic saline (NaCl 6%)

with a PARI LC PLUS nebuliser (PARI GmbH, Starnberg, Germany) ac-

cording to a standard operating procedure. Macrorheology was measured

immediately using the Kinexus Pro +rheometer (NETZSCH GmbH, Selb,

Germany) at 25 °C.

Rheological Analysis: The hydrogels were measured using a stress-

controlled MCR 501 Anton Paar rheometer with a plate–plate stainless

steel geometrical setup. A 25 mm upper rotating plate diameter was used

for all measurements with constant gap size of 0.15 mm. The measure-

ments were conducted at room temperature (25 °C) as well as physiolog-

ical temperature (37 °C). The samples were allowed to reach temperature

equilibrium for ≈5 min prior to each measurement.

Oscillatory Rheology: First, an amplitude sweep was performed to deter-

mine the limits of the LVE for each type of hydrogel analyzed in this study.

The critical deformation strain was found to be at about 1%. The shear

strain (𝛾) is deﬁned as the ratio of the deﬂection path (s)tothesheargap

(h)

𝛾=s

h(2)

Next, oscillatory shear tests were conducted within the LVE region, ap-

plying a constant strain amplitude, 𝛾A, of 1%. Frequency sweeps of the

hydrogel samples were conducted in the range 0.1–50 Hz. From the exper-

imentally determined shear stress 𝜏A, storage modulus G’ and loss mod-

ulus G’’ were determined, which are related to each other via the phase

angle (𝛿)

tan 𝛿=G′

G′′ (3)

The loss modulus G’’ describes the viscous behavior of the sample

which in turn is the result of energy dissipation due to the internal friction

between the moving layers. Conversely, the storage modulus G’ describes

the energy stored elastically within the internal material structure as it is

sheared.

Moreover, the rheological data could also be used to draw informa-

tion on the internal hydrogel structure and to roughly estimate an eﬀec-

tive mesh size, 𝜉, from the number density, 1N, of crosslinking points, the

Boltzmann constant kB, absolute temperature T, and the plateau modulus

G0:

1N=G0

kBT=1

𝜉3(4)

Hysteresis curves were produced for all frequency sweeps in order to as-

certain the reproducibility of the data. All reported results were the average

of the up- and down-ramps at each data point. The rheological properties

of all three families of LPG- and PEG-based hydrogels were investigated

on this basis.

www.advancedsciencenews.com www.mrc-journal.de

Supporting Information

Supporting Information is available from the Wiley Online Library or from

the author.

Acknowledgements

This study was supported by the Helmholtz Graduate School for Macro-

molecular Bioscience and funded by the Deutsche Forschungsgemein-

schaft (DFG, German Research Foundation)—SFB 1449—projects A01,

B03, C04, and Z02 and by the German Federal Ministry of Education

and Research (82DZL0098B1). The rheometer employed was ﬁnanced by

the Deutsche Forschungsgemeinschaft (DFG, German Research Founda-

tion) via grant GR1030/24-1. S.Y.G. is participant in the BIH-Charité Clini-

cian Scientist Program funded by the Charité – Universitätsmedizin Berlin

and the Berlin Institute of Health.

Open access funding enabled and organized by Projekt DEAL.

Conﬂict of Interest

The authors declare no conﬂict of interest.

Data Availability Statement

Research data are not shared.

Keywords

bio-inspired hydrogels, linear polyglycerol, mucus, redox responsive hy-

drogels

Received: May 12, 2021

Revised: July 23, 2021

Published online: September 13, 2021

[1] R. Bansil, B. S. Turner, Curr. Opin. Colloid Interface Sci. 2006,11, 164.

[2] R.A.Cone,Adv. Drug Delivery Rev. 2009,61, 75.

[3] M. A. Mall, J. Aerosol Med. Pulm. Drug Delivery 2008,21, 13.

[4] D. J. Thornton, Proc. Am. Thorac. Soc. 2004,1, 54.

[5] M. R. Knowles, R. C. Boucher, J. Clin. Invest. 2002,109, 571.

[6] D. Song, D. Cahn, G. A. Duncan, Langmuir 2020,36, 12773.

[7] N. A. Peppas, Y. Huang, Adv. Drug Delivery Rev. 2004,56, 1675.

[8] M. Caldara, R. S. Friedlander, N. L. Kavanaugh, J. Aizenberg, K. R.

Foster, K. Ribbeck, Curr. Biol. 2012,22, 2325.

[9] H. C. Berg, L. Turner, Nature 1979,278, 349.

[10] K. Ribbeck, Biosci. Hypotheses 2009,2, 359.

[11] C. Nowald, A. Penk, H.-Y.i Chiu, T. Bein, D. Huster, O. Lieleg, Macro-

mol. Biosci. 2016,16, 567.

[12] J. R. Kramer, B. Onoa, C. Bustamante, C. R. Bertozzi, Proc. Natl. Acad.

Sci.U.S.A.2015,112, 12574.

[13] A. Mahalingam, J. I. Jay, K. Langheinrich, S. Shukair, M. D. Mcraven,

L. C. Rohan, B. C. Herold, T. J. Hope, P. F. Kiser, Biomaterials 2011,

32, 8343.

[14] C. Werlang, G. Cárcarmo-Oyarce, K. Ribbeck, Nat. Rev. Mater. 2019,

4, 134.

[15] X. Cao, R. Bansil, K. R. Bhaskar, B. S. Turner, J. T. Lamont, N. Niu, N.

H. Afdhal, Biophys. J. 1999,76, 1250.

[16] M. Gervais, A. -. L. Brocas, G. Cendejas, A. Deﬃeux, S. Carlotti, Macro-

molecules 2010,43, 1778.

[17] S. H. Mahadevegowda, M. C. Stuparu, Eur. J. Org. Chem. 2017,2017,

570.

[18] B. D. Fairbanks, S. P. Singh, C. N. Bowman, K. S. Anseth, Macro-

molecules 2011,44, 2444.

[19] J.-P. R. Chauvin, D. A. Pratt, Angew. Chem., Int. Ed. Engl. 2017,56,

6255.

[20] H. Moritaka, S.-I. Sawamura, M. Kobayashi, M. Kitade, K. Nagata,

Biosci., Biotechnol., Biochem. 2012,76, 429.

[21] S. K. Lai, Y.-Y. Wang, D. Wirtz, J. Hanes, Adv. Drug Delivery Rev. 2009,

61, 86.

[22] M. S. Green, A. V. Tobolsky, J. Appl. Phys. 1946,17, 407.

[23] H. M. Yildiz, C. A. Mckelvey, P. J. Marsac, R. L. Carrier, J. Drug Targeting

2015,23, 768.

[24] J. Leal, H. D. C. Smyth, D. Ghosh, Int. J. Pharm. 2017,532,555.