Document [original]

Patterning Strategies of Poly(ethylene glycol) Based

Hydroxyapatite Composites for Biomedical Applications

Vorgelegt von

Dipl.-Ing. Axel Löbus

aus Leipzig

Von der Fakultät II - Mathematik und Naturwissenschaften

der Technischen Universität Berlin

zur Erlangung des akademischen Grades

Doktor der Ingenieurwissenschaften

Dr.-Ing.

Genehmigte Dissertation

Promotionsausschuss:

Vorsitzender: Prof. Dr.-Ing. Matthias Bickermann

Berichter/Gutachter: Prof. Dr. Ir. Marga C. Lensen

Berichter/Gutachter: Prof. Dr. rer. nat. Andreas Taubert

Tag der wissenschaftlichen Aussprache: 6. September 2013

Berlin 2013

D 83

Table of Contents

List of Abbreviations 5

Chapter 1: Scope and Organization of the Thesis 7

Chapter 2: Introduction 10

2.1 Surface Patterning of Materials for Potential Biomedical Application 10

2.2 Soft Lithography 17

2.3 The Fillmolding in Capillaries (FIMIC) Method 23

2.4 Atomic Force Microscopy and Force Spectroscopy as Characterization Tool

for Hydrogel Surface Patterns 27

Chapter 3: Materials and Methods 31

3.1 Materials 31

3.2 Methods 36

Chapter 4: Results and Discussion 43

4.1 AFM Characterisation of Elastically Micropatterned Surfaces Fabricated

by Fillmolding In Capillaries (FIMIC) and Investigation of the Topographic

Influence on Cell Adhesion to the Patterns 45

4.2 Blending PEG-Based Polymers to obtain a Library of New Biomaterials

and their Use in Surface Micro-Patterning by the FIMIC Method 58

4.3 Soft Lithographic Surface Patterning of Physically and Chemically Mineralized

Poly(ethylene glycol) Hydrogels for Selective Interface Interaction 66

4.4 3D Patterned Reactive Mineralized Poly(ethylene glycol) Derived Hydrogels 90

Chapter 5: Conclusion and Outlook 114

References 117

Abstract 128

Zusammenfassung 130

Acknowledgements 133

List of Publications 134

Contribution to Scientific Conferences 136

Appendix 138

List of Abbreviations

3BC - PEG-b-PPG-b-PEG

8PEG – 8arm star Poly(ethylene glycol)

ACP – Amorphous Calcium Phosphate

AFM - Atomic Force Microscopy

ATR FtIR – Attenuated Total Reflection Fourier Transformed Infrared Spectroscopy

BSA – Bovine Serum Albumin

CL - Crosslinker

CLSM - Confocal Laser Scanning Microscopy

E – Young’s Modulus

EDX - Energy Dispersive X-Ray diffraction

F/d curve – Force-distance curve

FIMIC – Fillmolding In Capillaries

G‘’ – Storage Modulus

G’ – Loss Modulus

HAp - hydroxyapatite

Mw – Molecular weight

NPs – nanoparticles

PDMS - Polydimethylsiloxane

PEG - Poly(ethylene glycol)

PETA - Pentaerythritol triacrylate

PI - Photoinitiator

PMMA - Poly(methyl methacrylate)

PPG – Poly(propylene glycol)

PS - Polystyrene

PUA - Poly(urethane acrylate)

Ra - Arithmetic Average (common measure for surface roughness)

RM - Replica Molding

RMS - Root Mean Square (common measure for surface roughness)

SBF – Simulated Body Fluid

SEM - Scanning Electron Microscopy

TEM - Transmission Electron Microscopy

Tg – Glass Transition Temperature

TGA - Thermogravimetric Analysis

UV – ultraviolet

XRD - X-ray diffraction

ε – Elongation at Break/ Compression Strain at Break

σdb – Ultimate Compressive Strength

σdf – Compressive Yield Point

σmax – Ultimate tensile Strength

σy – Yield Strength

Chapter 1

Scope and Organization of the Thesis

The idea of substituting injured or malfunctioning parts of the human body in order to regain

lost functionality is almost as old as mankind. Already in the ancient times, efforts were made

to replace teeth or fix broken or injured bones and cartilages. Those devices were designed

empirically by experience and fabricated from materials found in nature such as ivory or animal

bone. Concerted construction and investigation of what is nowadays called biomaterials only

started in the beginning of the 20th century. It was only then, when medicine emerged to the

discipline as it is known today with the development of modern implants and diagnostics. As

natural science evolved, it was discovered that the implants’ biological performance is greatly

affected not only by the type of material employed but also by its inherent surface

characteristics. Considering the complexity of natural surfaces, which were aimed to mimic, it

was revealed that each bears a specific design, intended to meet all the requirements of the

concrete purpose. However, natural materials are hardly ever of morphologically simple or

chemically pure nature, but are composites, which are structured from the macro level down to

the nanometer scale. Most famously of those might be bone and nacre, which have been vastly

examined in order to understand their outstanding properties and to give new insight of how to

mimic biological materials1. Thus, it was revealed that an intriguing combination of different

substances along with complex patterns of all relevant length scales result in the extraordinary

performance of natural composites. That extends to surface properties, determining the

interaction with the respective surrounding such as biological interfaces or ambient

environment. The most famous example may be that of the lotus leaf, a self-cleaning super

hydrophobic surface, which derives its intriguing properties solely from the micropatterned

surface structure of its leaves. Having grasped the paramount importance of patterning in order

to mimic natural composites such as bone in a successful manner, implant design was

extensively investigated in terms of physical, mechanical and chemical structures, paying

special attention to surface properties.

Surface modification and patterning is not restricted to tissue engineering, but is of essential

interest for biomedical applications such as biosensors or diagnostic devices as well. The

avoidance of nonspecific protein adsorption and consequent cell adhesion due to surface

characteristics is of common interest, but poses several challenges. Among those is the

respective determination of the isolated single influences of physical, mechanical and chemical

features to interface interaction, which is still discussed in the scientific community. The

discrimination and understanding of these effects towards protein adsorption and cell adhesion

are centered in this thesis. The main goal was the fabrication of surface substrates, which allow

attributing cellular response to designed surface properties. Thus, it was aimed to produce

platforms, which exhibit the contrast of bioinert versus bioactive sites, displaying protein

repellent and protein attracting behavior, respectively. Pursuing this objective, substrates were

manufactured, which promote spatially controlled selective cell adhesion. Such platforms allow

the attribution of single surface properties on cellular response, essential to profound

understanding of biomaterials surface reaction and necessary for the design of novel devices,

which prevent undesired interface interaction between substrate and biological environment.

This thesis is organized in four main chapters, which include “Introduction”, “Materials and

Methods”, “Results and Discussion” and finally “Conclusion and Outlook” and are in the

following briefly described. The chapter “Results and Discussion” is organized that each

subchapter represents in large parts an own publication or manuscript in preparation.

Introduction: This chapter aims to explain the plausibility of this thesis including the major

scientific principles and investigation techniques. Therefore, the importance of patterning in

biomaterials surface engineering is outlined and soft lithographic techniques as major method

are introduced. Furthermore, Fillmolding in Capillaries (FIMIC) is described in detail, as it is the

dominant surface patterning method applied in this thesis. Atomic Force Microscopy (AFM)

involving topographic imaging and Force spectroscopy are explained in a manner that aids the

understanding of the respective figures in the “Result and Discussion” chapter.

Materials and Methods: This chapter states all employed chemical substances, preparation

methods and characterization techniques relevant to this thesis in a comprehensive and

detailed manner.

Results and Discussion: This chapter contains the major findings of this thesis organized in four

subchapters. The first three subchapters (4.1-4.3) deal with different aspects of FIMIC sample

preparation and characterization and are designed in a consecutive manner in order to build up

a final result in chapter 4.3. The chapter 4.4 constitutes a subchapter of separate content,

describing the 3D formation and characterization of hydrogel composite scaffolds.

Conclusion and Outlook: This chapter summarizes the main results of this thesis, comments on

those and states potential challenges worth addressing in future research.

Note: The research project was part of a joint effort from three PhD students of the Lensen Lab;

Christine Strehmel, Zhenfang Zhang and Axel Loebus, each possessing different scientific

background. Consequently, due to the interdisciplinary nature of the here presented research,

many of the attained results could only be achieved in close collaboration with Zhenfang Zhang

and Christine Strehmel. Whenever the main intellectual contribution or experimental

processing was conducted by (one of) those two colleagues, it will be clearly referred to their

respective PhD theses2,3. In case any other person contributed significantly to the herein

displayed results, it will be clearly indicated, e.g. in the “Materials and Methods” chapter.

Chapter 2

Introduction

First of all, three definitions are introduced, which find application throughout this thesis.

However, the author is aware that diversions of these definitions exist.

A Biomaterial is nonviable material used in a medical device, intended to interact with

biological systems4.

Biocompatibility is the ability of a material to perform with an appropriate host response in a

specific application4.

Cytocompatibility can be tested by several standardized methods e.g. according to ISO10993,

“CytoTox-OneTM Homogeneous Membrane Integrity Assay” and “Trypan Blue Life/Dead

Staining”, to name a few popular examples.

2.1 Surface Patterning of Materials for Potential Biomedical Application

Surface patterning has been of paramount importance within the biomaterial science

community over the last decades5,6. In the course of exploring biomaterials and their design,

characterizing surface properties has proven to be essential, since the initial interaction taking

place at the interface implant - host may decisively impact the host’s reaction towards the

implant. Consequently, substantial effort has been invested and many systems have been

scrutinized in order to reveal influential factors, allow classification and set standards for

evaluation methods7,8. Hence, it was elaborated that (I) topography5–7, (II) chemistry8,9 and (III)

elasticity9 serve as determining factors in order to assess function and host response of

biomaterials surface patterns. The need to ascertain these key influential factors experimentally

in order to comprehend individual and combined impact was recently reviewed in a very

concise manner by Anseth and co-workers10. Surface patterning comprises physical, mechanical

and chemical structuring from the macro- down to nano-scale. Physical patterns, which

describe topographic features determine the surface’s roughness and may be of regular or

irregular nature. Mechanical patterns describe the characteristic of different adjacent surface

elasticities on investigated substrates, while chemical patterns regard to distinct chemical

surface cues leading to e.g. altering wettability or reactivity. They can be combined in almost

any desired manner. The principles of surface patterning methods are depicted in Figure 1.

Figure 1: Schematic representation of the most common principles of soft lithographic surface

patterning techniques: (a) physical patterns (topography e.g. lines, dots); (b) mechanical patterns

(elasticity e.g. gradual change, alternating); (c) chemical patterns (biological or chemical cues e.g. Au

nanodots, proteins); (d) example of a combination of patterns.

Physical patterning (topographic features) can be elegantly introduced via e.g. soft lithographic

methods such as Molding or Imprinting, which were developed by Whitesides et al. in the

1990s11. Alves and co-workers reviewed that resulting micrometer-sized structures on polymer

surfaces may promote selective cell adhesion or prevent cellular interaction as schematically

depicted in Figure 27. Thus, arrays may be designed, which spatially control protein adsorption

and cell adhesion and may even allow discrimination between different kinds of cell types for

example for application in diagnostics.

Figure 2: Schematic presentation of surface patterns employing different polymers, which promote or

prevent cellular adhesion and spreading. Modified image reprinted from reference [7.]

The patterning techniques however are not limited to pure chemical substances, but may also

include hybrid composite materials. One among many options is represented by mineralization

of polymer hydrogels employed in soft lithography. Mineralization of patterned substrates by

simple inorganic salts is conducted in various forms by nature1 and has hence drawn great

interest to this field of research elaborating profound comprehension regarding many aspects

of this issue. Among that, the biomineralization by calcium carbonates12 and calcium

phosphates13 seems best understood. Deposition of HAp by thermal spraying14, sol-gel

techniques15, electrophoretic deposition16,17 and deposition from Simulated Body Fluid (SBF)18

on metal, ceramic and polymer substrates, respectively, have elicited great interest, due to the

ease and high control in processing and the resulting elevated degree of bioactivity. Among

those techniques, deposition from SBF may be the most promising one in consequence of the

biomimetic process and inherent mild reaction conditions, which allow serving as drug carrier19

or the incorporation of e.g. osteoinductive biomolecules20. Although HAp nucleation and

deposition may be spatially controlled by varying surfaces charges21, sharp interfaces and

precise morphology tunability in order to obtain multifunctional materials containing distinct

bioaffinities, without subjecting to lavish and costly processing, remain a challenge22–27.

Latest developments describe new ways of patterning or the combination of commonly utilized

methods. Since it is impossible to summarize this vast research field in a few pages, the author

restricts to three examples, which are either related directly to the main objectives of the thesis

(Figure 3 and Figure 4), or present an intriguing, completely new approach for potential

biomedical applications (Figure 5). Figure 3 displays an approach by Luz and coworkers22 of

targeted inorganic deposition in order to obtain surface patterns of regular inorganic-organic

nature via imprinting of bioactive glass nanoparticles. It demonstrates that by controlled

incorporation of chemical cues, cell adhesion may be guided and HAp nucleation from SBF may

be spatially tuned.

Figure 3: Example of mineralization of in soft lithographic patterning: Imprinting of bioactive glass on

chitosan substrates in order to gain sites of preferential cell adhesion. Image reprinted from reference

[22].

Another example concerns the combination of physical micro- and nano-patterns by Farshchian

et al.28 as depicted in Figure 4. The feasibility of this approach is explained by the fact that

natural materials generally possess physical patterns over various length scales, which exhibit

distinct interaction characteristics to its environment. Incorporating different length scales has

been a challenge addressed by many contributions in the research community.

Figure 4: Scanning electron microscopy (SEM) micrographs of (a) primary PDMS stamp having

microgratings; (b) pre-nanopatterned PMMA substrate imprinted using a PUA/PC composite stamp; (c)

PMMA substrate imprinted using 3D nanomolding after demolding primary PDMS stamp; (d) 3D PMMA

substrate after removing the ultra-thin PDMS stamp. Image reprinted from reference [25].

A completely new approach centers the goal of cellular patterning introduced by the research

groups of Khademhosseini and Langer29 (principle detailed in Figure 5). The general idea to this

objective is to pattern different kinds of cell types in an adjacent manner in micrometer

dimensions. Thereby, the volume change due to alteration of temperature of microwells is

exploited.

Figure 5: Schematic diagram of spatially controlled patterning of two different cell types with dynamic

microwells. (a) Seeding the first cell type (red) at 24 °C when microwell structures were at swollen state.

(b) Washing microwells to rinse off undocked cells on microwell surfaces. (c) Undocked cells were

washed off the microwell surfaces. (d) Incubation at 37 °C to allow microwell structures to shrink,

resulting in more free space for the second cell type. (e) Seeding the second cell type (green) within

microwells. (f) Subsequently washing microwells to rinse off undocked cells on the surface. (g) Two cell

types were spatially distributed within microwells and further incubated at 37 °C. (h) Side and top views

of the resulting microtissues containing two spatially organized cell types. Image reprinted from

reference [26].

2.2 Soft Lithography

Lithographic methods are often employed in order to achieve complex 2D or 3D patterns. This

extremely versatile technique involves an ever growing number of different kinds of patterning

approaches for almost any substrate imaginable covering all length scales from macro down to

nano. All lithographic methods are employed in order to design surface characteristics of the

goal-substrate in any desired way. Soft lithography is among the most successful techniques,

due to its inherent versatility, ease of fabrication, accurate processing down to nanometer-

scale and mass production availability. Invented by Whitesides and co-workers in the early

1990s, it has by now evolved into a great number of different methods among which Replica

Molding (RM) may be the most famous one, since it represents the first preparation step of

many other sophisticated techniques11. An incomplete overview of most relevant lithographic

patterning techniques focusing on soft lithography is displayed in Figure 6. The methods finding

extensive application in this thesis are highlighted.

Figure 6: Lithographic patterning techniques, methods extensively applied during the course of this

thesis are highlighted in red. Replica Molding (RM) and Fillmolding In Capillaries (FIMIC).

Replica Molding (RM) may be considered as one of the most important and widely applied soft

lithographic techniques, since it is often an auxillary processing step to other techniques. RM

application is basically restricted by a curable liquid precursor solution of sufficient viscosity

allowing accurate penetration of the Silicon wafer pattern. In account to this fact, Figure 7

states details of the single processing steps. First, a Si wafer with the desired structural

dimensions is selected (Figure 7a). Next, the desired liquid precursor solution is placed and

cured via e.g. ultraviolett (UV) radiation as displayed in Figure 7b. As last step, the cured

solution is peeled off the original Si wafer and represents a negative (replica) of the original

master shown in Figure 7c, which is subsequently utilized as secondary mold for further

processing (replicating). RM denotes a process in which the negative (e.g. from PEG) from the

first replication step is employed as secondary mold for a subsequent replication step;

replication from a “soft mold”.

Figure 7: Processing steps of RM: (a) patterned Si wafer, serving as initial stamp; (b) polymer precursor

covering Si wafer and subsequent ultraviolet radiation; (c) removing of cured patterned polymer

negative (replica), which serves as secondary mold in a subsequent replication step.

With the help of RM, various kinds of surface patterns of differing magnitude such as dots,

lines, squares or any given combination can be produced in great quantities (Figure 8) as

reviewed in a concise manner by Roach and colleques30. Several polymeric formulations, such

as PEG based hydrogels or Polydimethylsiloxane (PDMS) based organosilicon materials, have

been vastly utilized.

Figure 8: Surface chemical and topographic patterning examples. (a) isolated single molecule grafted

pattern; (b) Isolated molecules grafted pattern; (c) island molecules pattern; (d) line molecules pattern;

(e) pillar topographic pattern; (f) grooves topographic pattern; (g) and (h) mixed chemical and

topographic patterns. Modified image reprinted from reference [30].

However, RM substrates may not be fabricated at any given aspect ratio. The limits of RM

processing can be mainly summarized by three prevailing effects (Figure 9) as shown by

Whitesides and coworkers11. A common problem is represented by Pairing (Figure 9a), that is

when an aspect ratio of base and height of standout posts (lines may be just as affected) exists,

which exceeds a specific material dependent value and the posts contact another. Patterns with

relatively deep patterns and narrow distances between posts/lines are particlulary vulnerable

to such defects. The second type of defect is called Sagging (Figure 9b), this is when due to the

inherent weight of the applied polymer stamp mechanical integrity is lost and the stamps

contacts the substrate in non-desired areas. Hence, the pattern gets lost. Shrinking (Figure 9c) is

a very common phenomenon due to e.g. UV radiation. That is because secondary bondings

such as Van der Waals bondings or hydrogen bonding are converted into covalent bonds, which

exhibit smaller binding distances and hence the bulk materials shrinks. For PEG based materials,

commonly 10 vol% are stated31.

Figure 9: Schematic illustration of possible deformations and distortions of microstructures in the

surfaces of polymers patterned via Replica Molding. a) Pairing; b) Sagging; c) Shrinking. Image reprinted

from reference [11].

Micro-Contact Printing is considered as a very popular soft lithographic patterning method with

a vast range of potential applications (Figure 10). The underlying principle of this easy and

straight forward bench top method is the inking of a substrate with a stamp (e.g. PDMS stamp)

during contact, a more elaborated version represents e.g. inking of liquid polymers or beads

(Figure 10b) on the substrates, which may be already chemically functionalized. Patterns down

to sub-micrometer dimensions have been realized as displayed by Zhou and coworkers32.

Figure 10: Micro-Contact Printing: (a) Schematic representation of polymer patterning; (b) Images of 96-

nm polystyrene beads deposited on the lines of patterned polyelectrolyte complexes. Modified image

reprinted from reference [32].

Another promising approach in this field is the recently developed Fillmolding In Capillaries

(FIMIC) method9 invented by the Lensen Lab, which enables the fabrication of sub micrometer-

precise Patterns of Elasticity. Those are surface patterns, which ideally are horizontally perfectly

plane in hydrated state and exhibit an alternating elasticity. In addition, the thus fabricated

FIMIC platforms may incorporate chemical functionalities, which can be introduced in a spatially

controlled manner. A more detailed description is found in the following chapter 2.3.

2.3 The Fillmolding In Capillaries (FIMIC) Method

Fillmolding In Capillaries was invented in the Lensen Lab and first reported by Diez et al.9. The

main principle of this sophisticated, soft lithographic easy bench top method is to fill pre-

prepared micropatterned hydrogel molds with a second elastomeric phase via capillary force.

Figure 11 depicts the single steps along with the dimension-determining hydrogel mold

produced via RM or simple replication from Si wafers. Figure 11a details the decisive

characteristics of a hydrogel mold, in which d represents the pattern distance, w the pattern

width and h the pattern height. Those dimensions may be in any given relation, but need to

comply with the requirements stated in chapter 2.2 in order to refrain from defects. Figure 11b

demonstrates schematically the FIMIC fabrication in 4 consecutive steps. (I) the hydrogel mold

(e.g. fabricated via RM), (II) the hydrogel mold is turned upside down on a glass or Si wafer and

a drop of the second liquid elastomeric phase is placed right in front of it. As a result of capillary

action, the liquid precursor is driven into the channels and cured under the application of

ultraviolet radiation (III). As final step, the as prepared FIMIC substrate can be released from

the glass/silicon wafer and turned upside-down; in such a way that the topographically smooth,

patterned surface faces upwards (IV). The resulting FIMIC is now ready for subsequent

investigation/application.

Figure 11: Processing Scheme of FIMIC substrates: (a) hydrogel mold with characteristic patterning

dimensions (d- distance, w-width, h- height); (b) processing steps for FIMIC fabrication: (I) as-prepared

hydrogel mold with defined distance, width and height characteristics, (II) placing of liquid polymer

precursor in order to allow penetration of empty channels via capillary force, (III) ultraviolet radiation of

filled lines, (IV) smooth final FIMIC substrate.

As important influential factors, surface tension at the liquid-vapor interface, capillary radius

and the contact angle between substrate and liquid (measure for wetting behavior of the

second phase on the hydrogel mold) were determined (Figure 12). Since all utilized hydrogel

precursors showed sufficient wetting behavior, the capillary radius, and consequently the

capillary force as driving force proved to be the most decisive factor for successful FIMIC

substrate production. It showed that small radii yield in highly defective platforms, which result

in enormous difficulties to fabricate FIMIC platforms with sub-micrometer dimensions. Hence,

most platforms used throughout this thesis had dimensions of between 10 µm and 50 µm.

Those are also the dimensions most relevant for cellular response investigations.

Figure12: Driving Force of Capillary Filling during FIMIC processing. Laplace Pressure with defining

measures.

Figure 13 depicts optical and Scanning Electron Microscopy (SEM) images of successfully

fabricated FIMIC platforms. The optical topview of a FIMIC platform at the end of filled channels

is shown in Figure 13a, which demonstrates the ease in discriminating areas of filled grooves

from unfilled ones. The cross- section recorded via SEM (Figure 13b) states the defined adjacent

nature of the two applied phases and serves as an instrument to guarantee fabrication quality.

Figure 13: Depiction of FIMIC substrates: (a) optical microscopy image, topview; (b) Scanning electron

microscopy image, cross-section.

Still, defects may occur, especially due to delamination in hydrated state as seen in Figure 14.

The main cause for this phenomenon may be incomplete curing during UV-radiation or great

differences in differential swelling behavior between the utilized hydrogel matrices. A weakness

of the FIMIC concept is the missing control of chemical interaction between the hydrogel mold

and the filled channels, which in particular in hydrated state becomes an issue. Then, the

physical interaction (e.g. adhesive forces) between the involved phases is reduced and hydrated

polymers may release from the hydrogel mold. This is less observed for materials exhibiting

little swelling degree or in pattern combinations containing materials of similar hydration

behavior.

Figure 14: Hybrid hydrogel with pattern ‘‘soft (20mm) in stiff (10mm)’’ in hydrated state; (a) and (b)

depict SEM images; showing one line coming loose in image (b). Light and dark gray are soft and stiff

hydrogel, respectively. (c) and d) depict optical micrographs, revealing a defect area on the patterned

surface in (d). Scale bars represent 50 µm. Image reprinted from reference [9].

2.4 Atomic Force Microscopy and Force Spectroscopy as

Characterization Tool for Hydrogel Surface Patterns

Atomic Force Microscopy (AFM) may be readily applied for the physical and mechanical

investigation of as prepared FIMIC platforms in dry or hydrated state. This versatile instrument

allows for qualitative and quantitative characterization of relevant surface properties. Methods

include physical measurement of topographic features as well as Force spectroscopy, which

involves the determination of surface elasticities and the proof of present mechanical surface

patterns, e.g. so called Patterns of Elasticity (Figure 15). There exist two fundamentally different

techniques finding application in the investigation of the micropatterned hydrogel surfaces.

First, the topographic imaging (Figure 15a), which allows physical surface characterization in dry

and hydrated state. However, in particular measurements conducted in swollen state serve the

purpose of the intended applications, since the main purpose of the fabricated platforms

remains the investigation upon cellular response. The topographic imaging yields 2D topview

image and cross-section at any given position. Thus, the real topography can be determined

over an area up to 100 µm x 100 µm. Due to the size of cells of approximately 40 µm in

maximum elongation, this is sufficient to scrutinize the areas, which single cells can sense.

Second, Force spectroscopy records Force-distance-curves (F/d curve) (Figure 15b), which yield

two main characteristics that are called adhesion and slope respectively. The information

obtained includes the vertical position of the tip (thus indentation) and deflection (bending of

the tip). Those two types of information are read out and converted via vertical displacement.

In the next step, the two entities, slope and adhesion are derived. Adhesion represents the

surface tip interaction via secondary interactions such as Van der Waals interaction, capillary

action of water meniscuses, electrostatic interaction. Slope determines the resistance against

elastic deformation, which represents the Young’s Modulus at the investigated spot. Force

spectroscopic methods are mainly applied to reveal regular mechanical patterns, which can be

correlated to the respective topography (Figure 15b). Those measurements on the one hand

may result in quantitative elastic analysis (F/d curve) of any chosen single surface spot, which

represents the surface tip interaction at the very end of the AFM tip. On the other hand can

Force Mapping experiments be conducted for the areas of interest via gridwise visualization of

recorded data points resulting in so-called Force maps. Precision is adjusted by the number of

recorded single F/d curves regarding a specific investigation site.

Figure 15: Atomic Force Microscopy in FIMIC characterization: (a) topographical imaging, detailing

scheme, 2D height imaging and according cross-section; (b) Force spectroscopy, detailing quantitative

determination of elasticity and force mapping.

In order to better comprehend the potentials of AFM recording detailed interaction between

AFM cantilever and investigated surface is depicted in Figure 16, as concisely displayed by

Shahin et al.33. During the recording process of one single F/d curve, the following steps of

interaction are experienced by the cantilever and are exploited for the mechanical surface

characterization: Point A defines the tip-surface interaction during the approach modus before

any interaction takes place. The cantilever is not bent and no contact between tip and surface is

yet observed. In point B, the tip jumps to the surface, due to secondary tip surface interaction.

Now, the tip is in contact with the surface, but the cantilever is not yet bent, since no surface

deformation has yet taken place. In state C, the AFM tip tries, resulting from applied pressure,

to penetrate the sample surface. The samples shows resistance towards elastic deformation

(Young’s Modulus, defined as the slope of F/d curve during C or D), the cantilever is bent in a

concave fashion. The tip deforms the substrate according to the adjusted pressure or distance

and retracts as soon as the pre-set value is reached. During the retraction modus, the stress

imposed on the sample is relieved and at point D the tip intends to separate from the sample

surface. However, due to secondary interaction in state E, the tip is held at the surface and the

cantilever bends in a convex fashion in order to retrieve the tip (adhesion force). Finally, the

cantilever retracts further and can break away as adhesion forces are overcome.

Figure 16: Details of Force Distance Curves (a) schematic approach and retract modus of an AFM tip on

investigated surface; (b) according AFM cantilever bending due to experienced interaction with the

surface during approach and retract modus. Image reprinted from reference [33].

Topographic imaging of hydrogel surfaces in hydrated state yield the challenge to determine

the exact liquid-solid interface of sample and surrounding liquid due to the soft nature of the

matter in hydrated state. This was further complicated, since regular surface patterns of low

nanometer dimensions were to be revealed. Nevertheless, the feasibility of our approach to

physically characterize patterned hydrogel surface in hydrated state was demonstrated by the

detection of horizontal regular structures down to below 50 nm. Those measurements could be

subsequently correlated with according Force maps from force spectroscopic investigations.

Hence, topographic and mechanical AFM characterization of as prepared hydrogel surfaces

proved to be vital and the instrument served as crucial investigation tool for the herein

presented PhD project.

Chapter 3

Materials and Methods

3.1Materials:

Any polymer utilized in this thesis was synthesized and processed by Zhenfang Zhang, a fellow

PhD student of the Lensen Lab.

3.1.1 Poly(ethylene glycol) (PEG) diacrylate (DA):

PEG DA (Sigma Aldrich, Mw 575 Da) is mixed with various amounts of crosslinker

(Pentaerythritol triacrylate, tech, Sigma Aldrich, Mw 298, 3 Da; 1 %, 5 % and 10 %) and one

percent of photoinitiator (2-hydroxy-4’-(2hydroxy-ethoxy)-2 methyl-propiophenone, 98 %;

Sigma Aldrich, Mw 224.26 Da) (PI) and subsequently UV cured under the exclusion of Oxygen.

3.1.2 PEG-based triblock copolymer of poly(propylene) glycol (PPG) and PEG,PEG-b-PPG-b-

PEG (3BC):

See details in the PhD thesis of Zhenfang Zhang2.

3.1.3 Polymer blends fabrication:

The gel networks were formed via ultraviolet (UV) radiation of an acrylated polymer in the

presence of a photinitiator (PI) (Irgacure 2959) and pentaerythritol triacrylate (PETA) as a

crosslinker (CL). In order to fabricate varying stiffness, pre-curing mixtures were prepared by

mixing acrylate terminated PEG or 3BC in a small amount of acetone with varying amounts of

CL (0 - 10 wt%) and 1 % PI (with respect to the amount of polymer).

3.1.4 Hydroxyapatite nanoparticles (HAp NPs):

Hydroxyapatite nanoparticles (HAp NPs) were synthesized applying the precipitation method.

Synthesis time and temperature, the two key factors, were adjusted to assure stable synthesis

conditions and to control the properties of HAp. Calcium nitrate tetra hydrate (Ca(NO3)2·4H2O,

Carl Roth GmbH & Co. KG) and diammonium hydrogen phosphate ((NH4)HPO4, Carl Roth GmbH

& Co. KG) were used as starting materials. Ammonium hydroxide ((NH4)OH) was used to adjust

the desired pH value. 5.96 g of Ca(NO3)2·4H2O were given into 200 mL distilled water and stirred

thoroughly until all Ca(NO3)2·4H2O was dissolved. A second solution was prepared by adding

1.92 g (NH4)HPO4 to 200 mL of distilled water. Both solutions were given simultaneously into a

glass reaction vessel. Subsequent addition of 10 mL 33 % (NH4)OH solution to adjust the pH of

the resulting suspension to 10. The precipitation time was chosen to be 48 h, additionally 10 mL

of 33 % (NH4)OH solution were added after 24 h due to potential evaporation of Ammonium

hydroxide. The precipitation reaction was performed at room temperature (22 °C). A constant

temperature and magnetic stirring were maintained during the whole reaction time. During

maturation the cover of the glass vessel was lifted several times to allow atmospheric exchange.

After that, the suspension was quickly filtered and washed with sufficient distilled water until all

(NH4)OH was removed. The resulting cake was dried at 75 °C for at least 24 h and grinded to a

homogeneous powder using a ball mill grinding jar.

3.1.5 PEG nHAp hydrogel composite synthesis:

PEG nHAp hydrogel composites made of PEG575 and HAp were fabricated as shown in Figure 1 in

chapter 4.3.1. As prepared salt solutions of Calciumnitrate tetrahydrate (Ca(NO3)2·4H20

analytical grade from Sigma Aldrich) and Diammonium hydrogen phosphate ((NH4)2HPO4

analytical grade from Sigma Aldrich) were added to the precursor solution of PEG575

complemented with 1 % PI at room temperature under vigorous magnetic stirring. Maturation

time was held constant at 24 h. Compositions were set in order to obtain 10 wt%, 20 wt% and

40 wt% PEG nHAp composites by weight (PEG 10nHAp, PEG 20nHAp and PEG 40nHAp

respectively). The resulting suspensions were thoroughly washed with distilled water and the

resulting suspension was transferred to a glass flask for further processing.

The following synthesis steps were performed. Salt solutions containing Ca(NO3)2·4H20 and

(NH4)2HPO4 respectively were added to acrylated PEG575 precursor in the reaction vessel. The

resulting homogeneous suspension was kept at 22 °C under constant vigorous magnetic stirring

for 24 h in order to allow for HAp nucleation. After a maturation-time of 24 h the suspension

was filtered to remove all residual water and the viscous composite gel was transferred to a

glass flask for further processing. Due to the incorporation of photo-initiator, UV curing was

enabled and thus application in softlithography. In addition, the cured composite gels proved to

be transparent and fluidic enough to be easily processed within the soft lithographic regime.

3.1.6 PEG HAp NPs hydrogel composite synthesis:

As prepared HAp NPs (synthesis see chapter 3.1.4) were given into the PEG575 precursor solution

and homogenized to result in composite hydrogels containing 20 % NPs by weight (20 wt%). In

order to achieve best distribution and least agglomeration behavior of the resulting PEG HAp

NPs hydrogel composite, the composite precursor solutions were exposed extensively to

vortexing and subsequent ultrasonic treatment prior to application.

3.1.7 8arm PEG:

8arm PEG-OH with a molecular weight of 15 KDa was purchased from Jenkem technology USA.

Acrylation-procedure:

For more details, see PhD thesis of Zhenfang Zhang2.

The product we use is 8arm PEG-acrylate. After acrylation, the molecular weight is still around

15 KDa, because the acrylation group is quite small compared with 8arm PEG.

3.1.8 8arm PEG HAp Composite (8PEG HAp) Synthesis:

Composites made of 8arm PEG and HAp were fabricated as shown in Figure 2 in chapter 4.4.1.

As prepared salt solutions of Ca(NO3)2 and (NH4)2HPO4 were added to the precursor solution of

8arm PEG complemented with 1 % PI at room-temperature under vigorous magnetic stirring.

Maturation-time was held constant at 24 h. Compositions were varied in order to receive 10 %,

20 %, and 40 % HAp – 8arm PEG composites by weight (8PEG 10HAp, 8PEG 20HAp and 8PEG

40HAp, respectively).

3.1.9 8arm PEG Composite Scaffold Synthesis:

The freeze drying process was conducted at 0.0019 mbar and -84 °C for 48 h (Alpha 2-4 L D

plus, Christ). Samples were prepared as follows. First the salt solutions containing Ca(NO3)2 and

(NH4)2HPO4 were given simultaneously into a glass reaction vessel in the presence of the

according amount of 8arm PEG. The precipitation reaction was performed over 24 h, after

which a homogeneous gel was attained. The obtained gel was swollen different times (5, 15, 45

min) in deionized water and subsequently UV cured (wavelength of lamp 365 nm) in hydrated

state over 15 min. Afterwards, the samples were frozen at -20 °C in a freezer.

3.1.10 Preparation of Simulated Body Fluid (SBF):

SBF solution was prepared as stated in detail by Müller and co-workers34. In summary, SBF

solution with a HCO3 content of 5 mmol/l was prepared by pipetting calculated amounts of

concentrated solutions of KCl (59.64 g/l), NaCl(116.88 g/l), NaHCO3 (45.37 g/l), MgSO4·7H2O

(49.30 g/l), CaCl2, TRIS(tris-hydroxymethyl aminomethane; 121.16 g/l), NaN3(100 g/l) and

KH2PO4 (27.22 g/l) into doubly distilled water. Concentrated salt solutions were added into a

1000 ml flask and filled up with double distilled H2O. The 121.16 g of TRIS was dissolved in 650

ml double distilled H2O. During stirring, the pH was adjusted to 7.6–7.7 at 25 °C, which is equal

to a pH of 7.3–7.4 at 37 °C, by adding concentrated HCl. The clear solution of TRIS and HCl was

quantitatively transfused into a 1000 ml flask and filled up with double distilled H2O. All salt

solutions were stored in polyethylene (PE) bottles in a refrigerator at a temperature of 4 °C. The

solutions were pipetted into 700 ml double distilled water in the sequence KCl, NaCl, NaHCO3,

MgSO4·7H2O, CaCl2, (TRIS + HCl), NaN3 and KH2PO4 to prevent precipitation. The pH of human

blood plasma ranges from 7.3 to 7.4 at 37 °C. Since the pH of SBF depends on temperature the

pH of SBF prepared at room temperature (21 °C) has to be adjusted between 7.75 and 7.85 in

order to obtain 7.3 and 7.4 at 37 °C. The NaN3 addition was conducted in order to inhibit the

growth of bacteria.

3.2Methods:

3.2.1 Soft lithographic methods:

3.2.1.1 Fabrication of micropatterned PEG575 composite replicas:

Micro-patterned silicon wafers were rinsed with acetone, water and isopropanol and dried

under a mild stream of nitrogen before use. Prior to the replication, the cleaned silicon masters

were fluorinated with trichloro(1H,1H,2H,2H-perfluorooctyl)silane 97% (Sigma-Aldrich). The

selected PEG575 or PEG575 composite pre-curing mixtures were dispensed on the silicon master,

covered with a thin glass coverslip and exposed to UV light (340 nm) for 15 minutes. Following

curing, the transparent polymeric film, with an inverse relief to that on the silicon master, was

peeled off mechanically. The stand-alone film (250 – 300 µm in thickness) could be handled

with tweezers.

3.2.1.2 Fabrication of Fillmolding In Capillaries (FIMIC) substrates:

More detailed description of the process is stated elsewhere9, briefly: the PEG575 replicas were

placed upside down on a glass slide and a small amount of a second liquid prepolymer-

composite (PEG nHAp or PEG HAp NPs) were carefully dispensed at the edge of the open

channels. Si masters with e.g. 20 µm grooves, 10 µm wide ridges and 10 µm deep structures

were applied for initial replica molding of elastomers. Subsequently, the viscous mixture was

allowed to fill the capillaries for 1 minute, after which the assembly was exposed to UV

radiation for 20 minutes. After exposure time was complete, the hybrid construct was easily

mechanically detached from the glass substrate and turned upside down to proceed with

surface characterization (AFM), protein adsorption experiments and cell culture. The resultant

polymer/polymer-composite FIMIC platforms were robust, free standing and transparent.

3.2.2 Characterization

3.2.2.1 Mechanical Characterization:

Rheological Measurements:

Rheology measurements were conducted using a Gemini 200 HR (Malvern Instruments) by

determining the appropriate frequency and vertical force on the sample (strain-controlled

mode). An 8 mm plate was used and measurements were taken at room temperature. Samples

were swollen min. 12 h prior to measurement. During the recording of data of swollen samples

were kept in a solvent trap to avoid loss of water during the experimental run. First, the linear

elastic range of the samples was determined with the help of the amplitude sweep. This is

observed, when Storage Modulus (G’) (indicating the elastic property of the network) and the

Loss Modulus (G”) (indicating viscous properties of the fluid) are yielding a constant plateau)

run over a certain frequency range in a horizontal and parallel fashion to each other. The value

was transferred to the frequency sweep, where the suitable values were ascertained within a

range of 0.01 to 10 Hz. 1 Hz as applied frequency and 0.0001 – 0.01 as deformation value (γ)

were chosen as appropriate parameters for all measured samples. The value of the observed

plateau was recorded and the bulk elasticity was calculated by the following equation as

described by Flory35,

E = 3 G’ (1)

where E is the Young’s Modulus and G’ is the Storage Modulus. Each material composition was

measured at least 3 times.

Tensile Test:

Tensile testing (Universal Testing Device, Zwick) was conducted in order to ascertain Young’s

Modulus (E), Yield Strength (σy), Ultimate Tensile Strength (σmax), and Elongation at Break (ε) of

the UV-cured composite hydrogels in hydrated state. Prior to measurements, dog-bone shaped

samples were immersed in deionized water for at least 12 h at 23 °C. For each set of hydrogel

composites 5 samples were prepared. Measurements were carried out at 1 mm/min

deformation velocity with a set-off of 2 N.

Compression Test:

Uniaxial Compression testing (Universal Testing Device, Zwick) was conducted in order to

ascertain Young’s Modulus (E), Compression Yield Point (σdf), Ultimate Compressive Strength

(σdb) and Compression strain at Break (ε) of the UV-cured composite hydrogel scaffolds in

hydrated state. Prior to measurements, cylindrical samples were immersed in deionized water

for at least 12 h at 23 °C. For each set of composites 5 samples were prepared. Measurements

were carried out at 1 mm/min deformation velocity with a set-off of 2 N.

3.2.2.2 Spectroscopic Methods:

RAMAN Spectroscopy:

Surface Resonance RAMAN Spectroscopy (LABRAM, HR Horiba Scientific) was conducted with

an excitation wavelength of 514 nm. Spectra of the as prepared samples were recorded

between 500 cm-1 and 1300 cm-1.

ATR Fourier Transformed Infrared Spectroscopy (ATR FtIR):

ATR FtIR measurements were conducted (Bruker Optics GmbH Equinox 55 with ATR golden gate

unit head) in order to qualitatively examine the presence of characteristic functional groups of

HAp formation in the range of 450 cm-1 to 4000 cm-1 within the composite materials and pure

HAp. The samples were vacuum-dried for 24 h prior to investigation.

3.2.2.3 X-ray diffraction (XRD):

XRD was conducted applying a powder-diffractometer (X'Pert Pro, PANalytical) in the range of 5

to 80 ° 2 theta. Samples were vacuum-dried at room temperature and grinded into powder

prior to measurement. The X-ray generator was operated at a voltage of 40 kV and 40 mA

producing CuKα radiation with a wavelength λ of 0.154 nm. A scanning speed of 2300 s degree-1

was applied for measuring each sample.

3.2.2.4 Microscopy:

Atomic Force Microscopy (AFM):

An Atomic Force Microscope (JPK instruments, Nanowizard II) was used in order to measure the

topography and surface elasticity of samples in dry and hydrated state.

Topographical Imaging:

Imaging was done in intermittent contact (hydrated samples) and contact mode (dry samples)

using silicon nitride cantilevers (PNP TR, k ≈ 0.08 N/m, f0 ≈ 17 kHz; Nanoworld Innovative

technologies) with a chromium-gold coating. Images were edited with NanoWizard IP Version

3.3a (JPK instruments). Samples measured in hydrated state were immersed for at least 12 h in

deionized water prior to measuring.

Force-Spectroscopy:

Force Mapping was conducted via qualitative comparing of surface elasticity using a set of

Force-distance curves (F/d curves) on the same scanning probe microscope (JPK instruments,

Nanowizard II). After every set of measurements, the cantilever was newly calibrated (by

applying the thermal-noise-method) before starting with the next set of force-distance

measurements. This value was then taken as surface elasticity. Silicon nitride cantilevers (PNP-

TR) with a chromium-gold coating (k ≈ 0.08 N/m, f0 ≈ 17 kHz; Nanoworld Innovative

technologies) were used. Images were edited with NanoWizard IP Version 3.3a (JPK

instruments). PNP TR tips (NanoWorld) exhibit a pyramidal tip-shape (face angle 35°), the tip-

geometry has been taken into account by applying the Bilodeau formula36 in order to fit force

distance curves. The fitting is implemented in the Nanowizard IP software and resulting values

for the E-Modulus are accordingly obtained.

Optical microscopy:

Light microscopy images were taken with an inverted Axiovert 100A Imaging microscope (Carl

Zeiss, Goettingen, Germany) using an AxioCam MRm digital camera and analyzed using the

AxioVisionV4.8.1 software package (Carl Zeiss, Goettingen, Germany).

Electron microscopy:

Scanning electron microscopy (SEM) was performed to visualize resulting patterns as well as an

in-situ electron Energy dispersive x-ray spectroscopy (EDX) in order to identify elemental

composition (in particular Ca/P ratios) and Elemental Mapping in defined areas of the

respective each sample. Prior to image recording, samples were carbon-coated and an

acceleration voltage of 8 keV was applied yielding EDX spectra at a penetration depth of

approximately 2.0 µm.

Transmission electron microscopy (TEM)( TECNAI G²20) was used to identify and quantify the

crystalline character of the respective PEG based PEG-HAp composites. In order to analyze

elemental constitution in-situ EDX was utilized. Prior to measurements samples were frozen in

LN2 and subsequently grinded with the help of a little ethanol. Following this, a copper net was

introduced and shortly after removed. As final step, the suspension on the copper net was left

to dry under ambient conditions. The instrument was operated at 200 keV acceleration voltages

with a LaB6-cathode and 0.24 nm point-resolution. TEM measurements were conducted by

Dipl. Ing. Sören Selve of the ZELMI, the Central Institution of Electron Microscopy of the

Technical University of Berlin.

Confocal Laser Scanning Microscopy (CLSM):

Confocal Laser Scanning Microscopy was conducted with the help and under supervision of

Gonzalo de Vicente Lucas, a fellow PhD student of the Lensen Lab.

Confocal and 3D images were recorded using a Leica TCS SP5 II Confocal Microscope (Leica,

Wetzlar, Germany) with a 20x and a water immersion objective 63x. The excitation wavelength

used was 496 nm and the detector was set in the range of 501 nm - 666 nm for BSA-FITC. The

applied scan speed was 10 Hz for images and 100 Hz for 3D series. Images were analysed and

processed utilizing the program Bioimage XD (Free software).

3.2.3 Cell Culture:

Cell culture experiments were conducted by Christine Strehmel, a fellow PhD student in the

LensenLab.

Mouse connective tissue fibroblasts (L-929) were kindly provided by Dr. J. Lehmann (Fraunhofer

Institute for Cell Therapy and Immunology IZI, Leipzig). L-929 cells were cultured in RPMI 1640

containing 10 % Fetal Bovine Serum (FBS) and 1 % Penicillin/ Streptomycin (PS; all PAA

Laboratories GmbH) at 37 °C and 5 % CO2 in a humidified incubator. The cells were grown in 75

cm2 cell culture flasks (SPL Life Sciences Inc.) until confluence, washed with Dulbecco’s

Phosphate Buffered Saline solution (DPBS, PAA Laboratories GmbH) and treated with Trypsin-

EDTA (PAA Laboratories GmbH).

MC3T3-E1 osteoblast like cells (further denoted as osteoblasts), originally derived from mouse

calvaria, were kindly provided by Prof. Z. Su (Beijing University of Chemical Technology, China).

MC3T3- cells were grown in 75 cm2 cell culture flasks containing Minimum Essential Medium

(Sigma- Aldrich) supplemented with 10 % FBS and 1 % PS at 37 °C and 5 % CO2 in a humidified

incubator. The cells were grown until confluence, washed with DPBS and treated with Trypsin-

EDTA.

3.2.3.1 Cell viability:

Murine fibroblast viability on smooth PEG and PEG-HAp substrates was studied using a live-

dead assay. For this purpose, PEG and PEG-HAp substrates (1 cm x 1 cm) were washed with

ethanol (70 %), rinsed in DPBS and placed in a μ-slide (Ibidi GmbH). 300 μl of a cell suspension

containing 50.000 L-929 cells were seeded onto each substrate and incubated at 37 °C, 5 % CO2

atmosphere and 100 % humidity. The viability of cells on the distinct substrates was estimated

after a 24 h incubation period. Following incubation, cells were stained with 100 μl of a vitality

staining solution containing fluorescein diacetate (stock solution 0.5 mg/ml in acetone, Sigma-

Aldrich) and propidium iodide (stock solution 0.5 mg/ml in DPBS, Fluka). Viable and dead cells

were detected by fluorescence microscopy.

3.2.3.2 Cell Morphology:

Samples were washed with ethanol (70 %), rinsed in DPBS and were placed in 24 well-plates

(Becton Dickinson). 50.000 cells/ ml were seeded on top of the samples and incubated for 24 h

at 37 °C, 5 % CO2 and 100% humidity. After 24 h, the cells were washed with DPBS to remove

unattached cells as well as remaining medium components and fixed for 30 min with 4 %

formaldehyde, pH 7 (Carl Roth GmbH). Prior to the observation of the cells with a scanning

electron microscope (SEM, Hitachi S-520), samples were dehydrated in a graded acetone series.

Finally, the samples were dried with critical point drying (CPD 030, Baltec), sputtered with gold

using a sputter coater (SCD 030, Balzers) and observed with a SEM using an acceleration voltage

of 20 kV and a working distance of 10 mm.

Chapter 4

Results and Discussion

The main findings of the thesis are presented and discussed in this chapter. The four

subchapters each address a distinct topic, but are not to be considered separately, since the

consecutive order imparts better project-understanding. Thus, the reader may follow the main

research objectives with greater ease, since the several chapters describe various applied

patterning techniques with growing materials complexity and augmenting method

sophistication. The reader should in particular note that the chapters 4.1, 4.2 and 4.3 all

comprise the soft lithographic surface patterning technique Fillmolding In Capillaries (FIMIC).

Therefore, the consecutive order of the chapters starts with the general introduction and

characterization of the FIMIC method in chapter 4.1 and concludes with the illustrating the vast

patterning potential, namely by incorporation of novel composite materials and subsequent

cellular experiments in chapter 4.3. Additionally, this thesis does not restrict to a single

patterning method, but involves surface patterning techniques (e.g. FIMIC) in chapters 4.1 – 4.3

as well as controlled formation of 3D structures for potential tissue engineering application in

Chapter 4.4.

Each subchapter represents a different publication manuscript stated in the respective front

page. To the chapters 4.1 and 4.2, the author contributed patterning processing as well as

surface and bulk characterization of the relevant substrates. Materials formation and cell

studies were conducted by colleagues of the Lensen Lab. To chapter 4.3 and chapter 4.4, the

author contributed, as author, to all relevant data presented and discussed except cell studies,

which were conducted in collaboration with Lensen Lab group members.

Chapter 4.1 addresses the fabrication of FIMIC platform with different kinds of Poly(ethylene

glycol) (PEG) based hydrogels and scrutinizes the feasibility of physical and mechanical surface

characterization via AFM topographic imaging and Force spectroscopy, respectively. This results

in so called Patterns of Elasticity, hydrogel surfaces which exhibit alternating stiffness in the

micrometer range. In addition, FIMIC substrate investigation is not only restricted to dry state,

but extends to hydrated state revealing the necessity of considering differential swelling of

hydrogels into the FIMIC design.

Chapter 4.2 addresses the issue of differential swelling, employing several PEG based polymer

blends with different hydration properties in order to equilibrate the physical surface

characteristics of the FIMIC platforms in completely hydrated state.

Chapter 4.3 addresses the application of different kinds of PEG hydroxyapatite (HAp) hydrogel

composites, namely PEG HAp NPs and PEG nHAp into the FIMIC processing. PEG HAp NPs

consists of PEG and 20 wt% of physically incorporated HAp NPs. PEG nHAp in contrast features

chemically synthesized PEG HAp hydrogel composite by in-situ nucleation of HAp in the PEG

precursor matrix. These materials were processed in a spatially controlled manner into FIMIC

substrates and the resulting chemically and mechanically micropatterned platforms were

investigated and compared upon post fabrication functionality with HAp from Simulated Body

Fluid (SBF), protein adsorption ability of the model protein Bovine Serum Albumin (BSA) and

cellular adhesion behavior of fibroblasts (L-929) and osteoblast like cells (MC3T3-E1). This is the

concluding chapter to the FIMIC processing with varying PEG based hydrogels.

Chapter 4.4 addresses the scaffold formation of 8armPEG HAp nanocomposite (8PEG HAp) via

freeze-drying technique. The composite scaffolds are characterized regarding chemistry,

morphology, mechanical performance, protein adsorption ability of the model protein Bovine

Serum Albumin (BSA) and cellular response comparing adhesion behavior of fibroblasts (L-929)

and osteoblast like cells (MC3T3-E1) on as-prepared scaffolds. The determination of

morphological properties was conducted in cooperation with the group of Prof. Dr. Christian

Rüssel from the Otto Schott Institute of the Friedrich-Schiller University of Jena, Germany.

Additional cellular experiments were conducted in collaboration with the group Prof. Su of the

Department of Materials and Engineering of the Beijing Chemistry and Technology University,

Beijing, China.

Chapter 4.1

AFM Characterisation of Elastically Micropatterned Surfaces Fabricated

by Fillmolding In Capillaries (FIMIC) and Investigation of the

Topographic Influence on Cell Adhesion to the Patterns

Abstract

The recently developed soft lithographic method Fillmolding In Capillaries (FIMIC) was utilised

in order to fabricate micropatterned substrates of Poly(ethylene glycol) (PEG) derived hydrogels

with varying elasticities, so called Patterns of Elasticity. Surface characteristics were scrutinized

via Atomic Force Microscopy regarding surface roughness and surface elasticity. Topographic

imaging demonstrated generally smooth surfaces; however the physical profile is greatly

impacted, when applying materials exhibiting different degrees of volume change during

hydration. Thus, it is displayed how the topographic landscape of the hydrogels substrates is

defined by the swelling behaviour of the constituents.

This chapter contains results from the following journal publication:

Kelleher S, Jongerius A, Loebus A, Strehmel C, Zhang Z, Lensen MC.; AFM Characterization of

Elastically Micropatterned Surfaces Fabricated by Fill-Molding In Capillaries (FIMIC) and

Investigation of the Topographical Influence on Cell Adhesion to the Patterns.; Advanced

Engineering Materials, 14, B56–B66 (2012).

Results and Discussion

4.1.1 Materials Characterisation

As initial step, the non-cytotoxic characteristics of all polymer gels prepared from PEG and 3BC

were verified utilizing a colony forming ability assay. As expected no cytotoxic effect was

observed (for details, see thesis of Christine Strehmel3).

Subsequently, the physical and mechanical properties of the employed hydrogels were

investigated. Fabricating FIMIC samples (details in chapter 2.3) with different surface and

hydration properties requires the use of tailormade prepolymers. Physical properties of

crosslinked Poly(ethylene glycol) (PEG) hydrogels may be tuned and adjusted by varying

degrees of crosslinker (CL) and photoinitiator (PI) of the polymer precursor mixture.

Throughout the processing, the amount of PI was kept constant at 1 wt%, while the notation of

the specific hydrogel formulation states the amount of CL incorporated e.g. “PEG 5%”

represents a mixture of PEG (Figure 1a) with 1 wt% PI and 5 wt% CL. Both, the elastic and the

hydration properties of the applied hydrogels were ascertained, since these characteristics are

of crucial relevance with respect to the eventual physical nature of the FIMIC platforms. That is,

because Patterns of Elasticity (patterns with adjacent areas exhibiting different elasticities) are

manufactured, of which hydration softens the mechanical properties of employed hydrogels.

Figure 1 outlines and compares the aforementioned properties for the hydrogel networks.

These measurements have been conducted in dry and hydrated state in order to scrutinize the

hydrogels in as prepared state as well as in aqueous medium, since the latter would become

essential during cell testing.

The ability of hydrogels to take up water is dependent on the network structure, the degree of

crosslinking and the hydrophilicity. Since hydrogel compositions are prepared from neat liquid

precursors, they are tightly chemically crosslinked and do not swell to the extent as if they had

been cured in a hydrated state (not necessarily equilibrium state). Figure 1c shows the water

uptake of the hydrogels after 24 h in deionised water. For hydrogels with higher amounts of CL,

the degree of the crosslinking increases and consequently the network is more tightly linked.

Therefore, the amount of water that the system can take up into the system decreases. On the

other hand, despite any lower crosslinking density, 3BC gels have a lower swelling ratio than

PEG due to the variance in chemical structure. Hence, 3BC exhibits significantly less hydrophilic

behaviour than pure PEG.

Figure 1: Chemical structures and some physical properties of the PEG-based gels: a) structure of the

two pre-polymer diacrylates PEG and 3BC; b) bulk elasticity of the 3BC materials in dry and hydrated

states; c) the swelling degree of the crosslinked hydrogels.

The degree of crosslinking also determines the elasticity of chemically crosslinked hydrogels in

that less tightly bound hydrogels exhibit smaller Young’s Moduli than those with a higher

crosslinking density. Additionally, for linear polymer systems, the mechanical properties are

impacted by the molecular weight (Mw) in the sense that the smaller Mw the higher the

elasticity. Hence, it was expected that 3BC (higher Mw) displays lower elasticity than PEG (lower

Mw). The values for the bulk elasticity of 3BC as calculated by rheological measurements are

shown in Figure 1b. Unfortunately, the values for PEG could not be ascertained, as prepared

PEG was too stiff for the experimental setup.

4.1.2 Microscopic Investigations of the FIMIC Platforms

The surface of the FIMIC substrates can be examined using optical microscopy in order to verify

the successful processing or reveal potential defects. Sometimes, the formation of scum layers

could be observed, which occurs when the pre-polymer does not only fill the designated FIMIC

channels, but also covers residual areas of the replica, creating a layer of polymer across the

replica.

Figure 2a depicts the top-view recorded via optical microscopy of a successfully fabricated

FIMIC platform and Figure 2b shows a FIMIC substrate with a connected forward line of the

meniscus indicating the presence of a scum layer. To gauge the thickness of any scum layer, real

cross sections were recorded. Figure 2c shows a cross section of an as prepared FIMIC platform

exhibiting a scum layer with approximately 1 µm in dimension. However, optical microscopy

may not be the most suitable technique to quantify, due to limited accuracy reading the results

and sample placing problems. For that reason, and to ensure that the liquid polymer had

actually entirely filled into the corners of the channels, cross section images of FIMIC samples

were recorded via Scanning Electron Microscopy (SEM). Figure 2d depicts an SEM image, which

verifies that the channels have filled completely. FIMICs were made from a replica with [w-s-d]

[50-10-10], whereas w represents groove width, s – spacing between grooves and d-depth (see

chapter 2.3).

Figure 2: Scum layer on FIMIC samples: (a) and (b) Optical images of FIMICs made from replica where

[w-s-d] is [50-10-10] showing a scum layer in (b); (c) and (d) respective optical and electron micrographs

of real cross sections showing a scum layer in (c).

4.1.3 Atomic Force Microscopy and Force Spectroscopy Investigations

Atomic Force Microscopy (AFM) was employed in order to investigate the surface topography

and the local mechanical differences of various FIMIC compositions; for that reason

topographic imaging and force mapping were utilized respectively. Both, the microscopic and

spectroscopic investigations were carried out under ambient conditions applying a liquid cell for

measurements in aqueous conditions. Topographic imaging was conducted in intermittent

contact mode, while Force spectroscopy was performed employing the same tip and recording

a set of Force-distance (F/d) curves on exactly the same area of the sample, where the

microscopic investigation had been conducted beforehand. The set of spectroscopic

measurements was subsequently employed to create so-called Force-maps (Figure 3), which

are elaborated in the following paragraph.

The surface elasticity of materials can be calculated from the F/d-curves, which represent the

elastic response of the material to indentation. Therefore, Figure 3a details the interaction-

cycle between an AFM tip and the scrutinized substrate of a single force spectroscopic run. At

large distances, there is no interaction between the tip and the substrate, and the force

remains zero. However, upon approaching the surface, the tip starts to interact with the

surface and attractive forces (such as capillary forces, Van der Waals interactions, hydrogen

bonding and electrostatic interactions) make the tip snap into contact with the surface. This

attraction can be recognised as the minimum in the F/d-curve, while the recorded F-values

correlate with the strength of interaction. These attractive interactions are adhesive forces and

can also be plotted as part of Force-maps.

The slope of the F/d-curve is a measure for the elastic response of the material upon

indentation with the AFM tip, hence representing the elasticity (Young’s Modulus)[36]. Stiffer

materials are less easily deformed than softer ones and the slope will consequently be steeper,

indicating that a larger force is needed to indent the surface. Force-maps, as depicted in Figure

3b and 3c, refer to qualitative mapping of surface elasticities (Figure 3b) and adhesive

interactions (Figure 3c) and contain valuable information about the chemical homogeneity and

elastic properties of the sample surface. Beyond that, they can reveal mechanical or chemical

surface patterns of investigated substrates as extensively exploited in this thesis. The general

colour code in Force-maps relates to; higher elastic values appear in a lighter colour and lower

values in darker colour.

Figure 3: (a) Representative force-distance curve; (b) and (c) Force-maps created according to the slope

of the F/d-curve; (b) adhesion forces between the tip and the substrate; (c) Distinction between

different identical hydrogels materials applied.

The elasticity of the two different hydrogels, both in dry and in hydrated state, was

quantitatively determined using Force spectroscopy as depicted in Figure 4. This method

enables to determine the respective values of PEG based hydrogels, which was not possible by

bulk rheological measurements. Nevertheless, the surface elasticity, as measured by AFM,

cannot be directly compared to bulk values as measured by rheology, since the two methods

are inherently different and assess different mechanical properties; AFM ascertains surface

elasticity and rheology bulk elasticity. For cell culture studies, however, the surface elasticity in

hydrated state is arguably the more important property, since cells sense the surface elasticity

before any other mechanical property. In fact, there are different opinions existing, e.g.

Buxboim et al. and Kuo et al. state as deep as the size of the magnitude of the cell itself37,38,

others limit it to a few micrometres (1 µm – 6 µm)39–41. A concise commentary of this

controvert issue from Buxboim et al. summarizes the different standpoints42.

It is obvious from Figure 4 that as prepared hydrogels from PEG possess much stiffer

mechanical properties than those of 3BC, which can be expected on the basis of the relatively

low molecular weight of the precursor. Interestingly, the elasticity of PEG in hydrated state is

drastically decreased in comparison to that in dry state (by 2 orders of magnitude) and due to

this softening effect, the PEG 0% hydrogel becomes even softer than any of the 3BC hydrogels.

The elasticity of 3BC also decreases by water uptake, but to a much smaller extent (only by a

factor of about 2). This different behaviour can be explained by the differences in water uptake;

PEG swells significantly more than 3BC.

Figure 4: Surface elasticity of the hydrogels as measured by Force spectroscopy using the AFM.

Initially, replicas from PEG 5% and filled with 3BC 5% were created as surface pattern with

sharp, clear boundaries and a distinct contrast in elasticity. The quality of the mechanical

patterning of FIMIC platforms may be recognised from the Force maps (see Figure 3b and c),

which display that PEG 5% possesses a higher Young’s Modulus and a lower tip-surface

interaction than 3BC 5%. In hydrated state, however the differences in adhesion forces are not

as pronounced as in dry state, which can be attributed to the absence of a capillary water film

between the tip and the substrate in immersed state.

The topography difference in the dry state verses hydrated state was examined using AFM

topographic imaging combined with the analysis of according cross sections (Figure 5). In the

dry state, the samples showed a topography difference of about 120 nm, which was shown to

increase significantly to around 1.0 µm in the hydrated state. The increase in the height

difference between dry and hydrated states leads to the conclusion that the topography

change is related to the differential swelling properties of the two employed hydrogels resulting

in the PEG replica protruding from the surface. Nevertheless, in order to exploit the differential

swelling in an advantageous manner, it was attempted to level out this slight indentation by

fabricating a FIMIC platform, in which the replica swells less than the filling material (Figure 5b)

addressing to counteract the depression by allowing the filler to swell to the level of the replica.

Since 3BC swells less than PEG, consequently, FIMIC substrates from a 3BC 5% replica and a PEG

5% filler were created. In this instance, AFM showed that the topography in the swollen state

had indeed decreased (Figure 5d). Thus, it proved that this strategy seems successful to create

patterned surfaces that display the least topographic difference in hydrated state, and hence in

cell culture.

Figure 5: AFM topography images and corresponding cross sections of FIMICs consisting of (a) and (b)

PEG 5% (Replica) and 3BC 5% (Filler); (c) and (d) 3BC 5% (Replica) and PEG 5% (Filler); left: dry state;

right: swollen state.

To clarify the extent of the effect of differential swelling on the FIMIC topography, substrates

with two components possessing identical swelling properties were examined. Utilizing PEG 5%

as both the replica and as the filler material, it can be observed that there is little difference in

topography between the dry and hydrated states, with a height difference of as little as 40 nm

(see SI 1). This demonstrates that for identical hydrogel compositions with equal swelling

properties, the change in topography between the replica and filled areas of the FIMIC surface

is negligible. It should be noted that in dry state, the typical cross section profile always shows a

small depression on the filler side, regardless of which material is employed as a mold and

which as the filler. This could be attributed to shrinking during UV-radiation. Table 1

summarizes the effects of differential swelling that either amplify or diminish topography.

Table 1. Exploiting differential swelling behaviour of 3BC and PEG to increase or decrease the

topography in hydrated state.

In order to test the versatility of this approach and to arrive at very well controlled micro

Patterns of Elasticity that are topographically smooth in cell culture, the fabrication of this type

of FIMIC platform(replica that swells less, filling material that swells more) was repeated.

Otherwise, contact guidance may be observed. That is, if significant topography remains,

cellular response may be attributed to topography instead of chemistry or elasticity (for more

detailed information see PhD thesis of Christine Strehmel3). Due to the selection of materials

the chemistry of both constituents were kept the same. Replicas made of 3BC 10% and filled

with 3BC 0%, where the two components exhibited slightly different swelling properties (a

water uptake of 4 % versus 6 % respectively), were produced. Again, it was observed that there

is a decrease in topography from approximately 250 nm to 100 nm (Table 1) and that the

surface profile is levelling out due to the changes in degree of swelling. The same strategy was

employed for PEG, whereas the effect of decreased topography by virtue of differential swelling

(in this case comparing a water uptake of 18 % for PEG 10% with 28 % for PEG 0%, details see

PhD thesis of Zhenfang Zhang2) was even larger, decreasing from approximately 400 nm in dry

state to approximately 125 nm in hydrated state. It should be noted however, that the

topographic differences not only vary between different samples, but also on the very same

sample. Variations can be observed between measurements on different areas of the same

sample, which also can be attributed to the standard deviation in water-uptake, which

hydrogels generally possess (more detailed information can be found in the PhD thesis of

Zhenfang Zhang2).

4.1.4 Summary

The scope and limitation of the recently developed Fillmolding In Capillaries (FIMIC) method

was investigated in order to fabricate elastically micropatterned substrates, so-called Patterns

of Elasticity. The recent investigations employing Atomic Force Microscopy (AFM) however,

have revealed that the surfaces of such FIMIC substrates are never completely smooth. A slight

topographic landscape seems inherent to this particular fabrication method. The present

results demonstrate that the topographic differences can even be enhanced, when the FIMIC

platforms are immersed in water. This is due to differential swelling of the two PEG-based

hydrogel materials.

Nevertheless, one can apply the differential swelling in a beneficiary nature to counteract

volume shrinking of applied hydrogels observed during UV-radiation. The right choice of the

two constituent hydrogel materials in FIMIC substrate production should be employed, namely

by employing material that exhibit little tendency to water-uptake as the mold and a hydrogel

that displays a higher degree of swelling as the filler. Thus, it was possible to successfully level

out the topographic differences to some extent. Although no perfect levelling could be

achieved, the presented results still prove the feasibility of this approach.

Chapter 4.2

Blending Poly(ethylene glycol) (PEG)-based Polymers in order to obtain

a Library of New Biomaterials and their Application in Surface-Micro-

Patterning by the Fillmolding In Capillaries (FIMIC) Method.

Abstract

A library of Poly(ethylene glycol) (PEG) based polymer blends with varying hydration properties

were synthesized. They were fabricated in the absence of solvents, which makes them a

suitable candidate for Fillmolding In Capillaries (FIMIC) processing. The main aim of this chapter

is to include differential swelling properties of utilized hydrogels formulations in order to

achieve FIMIC substrates without physical topography in hydrated state. This poses the

challenge of levelling out the topography by adjusting shrinking properties during ultraviolet

radiation with swelling differences of employed polymer blends. In consequence, genuine 2D

Patterns of Elasticity of binary nature could be fabricated, which would allow studying cell

migration phenomena in systematic detail.

Details in processing of pure polymers and blends can be found in the appendix of this thesis,

since it was conducted by two colleagues of the Lensen Lab (Zhenfang Zhang and Dr. Susan

Kelleher), and since my contribution to the blend fabrication was relatively limited. For a better

understanding, here, only a short introduction to the hydrogels employed is given.

Kelleher S, Zhang Z, Loebus A, Strehmel C, Lensen MC.; Blending Poly(ethylene glycol) (PEG)-

based Polymers in order to Obtain a Library of new Biomaterials and their Application in

Surface-micro-patterning by the Fillmolding In Capillaries (FIMIC) Method. Submitted to Soft

Matter

Results and Discussion

4.2.1 Materials Characterization

First, material notations with corresponding chemical properties finding application in this

chapter are stated in Figure 1. One of the main reasons for applying Poly(ethylene glycol) (here

denoted PEG1) (PEG, Mw 575 gmol-1) and 3BC1, a tri-block copolymer (3BC) of PEG and

Poly(propylene glycol) (PPG) (Mw 4400 kDa), in the fabrication of FIMIC platforms is the liquid

nature of both hydrogels at room temperature (22 °C), which provides greater ease to FIMIC

processing (for more processing details, see chapter2.3). The general idea of this project was to

achieve successful levelling out of FIMICs. Therefore, employed hydrogels had to meet two

principle characteristics: First processibility, since they are of non-linear nature e.g. multi-arm

PEG poses the challenge of conversion from dry powder-like state into liquid state. Second cell

repellence, since blends needed to contain more PEG than PPG in 3BC1 in order to ensure the

perpetuation of cell-repellent behaviour. The polymers employed are presented in Figure 1. The

table alongside shows the aggregation state of the polymer at room temperature with the

according molecular weight and the calculated theoretical percentage of PEG in the polymer;

pure PEG contains 100 % PEG and the block co-polymer contains either 30 % PEG in the case of

3BC1 or 80 % in the case of 3BC2. One should be aware that the PEG content in these blends is

essential in order to understand the physicochemical properties of the herein presented

hydrogels.

Figure 1: The general structure of the three types of polymers used, i.e. PEG, 3BC and 8PEG and the

properties of five selected examples that were investigated.

The addition of both pre-polymers with longer linear chains PEG2 (Mw 3400 Da), 3BC2 (Mw 8400

Da) and multiple arm 8PEG (Mw 15,000 Da), allows to fabricate polymer networks with very

different characteristics than those utilized before, possessing more possible crosslinking points

in the case of the multi-arm PEG or longer distances between either crosslinking points (higher

Mw between crosslinks and corresponding larger mesh sizes). Thus, one can precisely tune the

blends’ elasticity and protein and cellular adhesiveness.

Important to remember is that one of the main requirements for the FIMIC patterning method

is the need to apply liquid precursors, both to make replicas from silica masters via Replica

Molding and subsequently to fill the grooves present on patterned hydrogel replica surfaces.

The combination of specific chemistry and low molecular weight of these three new polymers

(PEG2, 3BC2 and 8PEG) means that they are, in contrast to PEG1 and 3BC1, solid at room

temperature and need to be processed into liquid state for blending and subsequent

employment in FIMIC processing. This resulted in greater handling difficulty and poses

therefore a substantial obstacle to the processing procedures. As a consequence of lavish trails,

PEG2 was eliminated from further usage.

4.2.2 Application of Blends in the Fill Molding In Capillaries Method

In order to illustrate the challenge of addressing different hydration behaviours of employed

hydrogel blends in FIMIC substrate fabrication. Figure 3 displays the three possible alternatives.

Depicted are schematic cross sections as recorded from AFM topographic imaging. Figure 3a

depicts the option ‘Filler swells more’, which refers to the former grooves standing out after

hydration. Figure 3b refers to ‘Filler swells less’, which means that after complete hydration,

the grooves still exhibit a depression. In rare cases (rather theoretically), that may also be the

case, in which the swelling difference between the utilized hydrogel blends is lower than the

volume shrinkage of the filler material during UV radiation. Figure 3c illustrates the ultimate,

however very challenging objective of a completely levelled out FIMIC substrate in hydrated

state.

Figure 3: Schematic illustration of hydrated FIMIC platforms employing hydrogel combinations with

varying degrees of hydration. (a) Filler swells more; (b) Filler swells less; (c) perfectly levelled-out

sample.

Hydrogel blends made of PEG1/3BC1 and of PEG1/8PEG were applied in order to produce FIMIC

platforms. They were chosen, since they form the most promising combination of mechanical

and protein and cell adhesion properties as well as to provide a necessary degree of

processibility. To ensure that the polymers blends are in their liquid state for FIMIC processing,

the PEG1/3BC1 blend is utilized as prepared and the PEG1/8PEG is melted and prepared using

pre-heated glassware. Two different FIMIC-pattern-dimensions were finally selected, namely

“20-10” and “10-50”.

Blend 1: PEG1/3BC1

“20-10” FIMIC samples (filler occupies a groove of 10 µm in width) were produced, employing a

blend ratio of PEG1:3BC1 (33:66). In the sample substrates, where the filler swells less than the

mold, the filler contained 10 % CL and the mold contained 5 % CL. In an approach pursuing the

opposite hydration behaviour, the filler contains 5 % CL and the mold contains 10 % CL. The

topography of the substrate surfaces of these samples was measured by Atomic Force

Microscopy (AFM) in the dry and the hydrated state.

The AFM results display that the strategy of levelling out the FIMIC substrates utilizing fillers

with greater degree of water-uptake was generally working, but no complete levelling out of

the investigated samples could be achieved (see SI 4). The contour of the FIMIC surfaces was

still concave, with the filler appearing to be kind of retracted from the mold. Hence, it was

concluded that the hydrogels made from PEG1/3BC1 blends do not take up sufficient amounts

of water in order to level out the sample entirely and thus, the effort was focused on the set of

blends fabricated from PEG1/8PEG, which exhibit greater swelling degrees. Therefore, they

promise better chance to reduce the physical patterns of FIMIC substrates to a minimum.

Blend 2: PEG1/8PEG

The hydrogel blend PEG1/8PEG was selected, since firstly, these constituents are the easiest to

handle and to process, and secondly they had a relatively high ability to take up water and

beyond that, they consist of pure PEG-constituents. Consequently, the fabricated blends

maintain the protein repellent nature of the pure substances. On PEG-surfaces, any aided

protein adsorption and subsequent cell adhesion must be attributed to the designed elasticity

pattern or to eventually remaining, undesired topographic effects. For these FIMIC substrates,

regardless if mold or filler, blends were selected that contained the same ratio of the two

components, PEG1 and 8PEG, but different amount of CL. A series of PEG1/8PEG FIMIC

platforms were fabricated, applying a mold with “50-10” pattern dimensions. Hence, the

master resulted in FIMIC substrates, where the filler spun 50 µm and the mold constituted of 10

µm lines. Seven different FIMIC samples using this blend were fabricated; for each sample, due

the maximum to water uptake ability, 0 % CL in the filler pre-polymer solution was chosen. The

hydrogel molds contained a varied amount of CL ranging from 10 % down to 0 %. This series of

substrates was produced in order to obtain FIMIC platforms perfectly levelled out. The samples

were prepared and left to swell in deionised water for 12 h before the topography was

measured via AFM. Upon measurement in the hydrated state, it was observed that in all

samples, the filler in fact protruded from the mold. Thus, FIMIC substrates of previously

unobserved convex patterning properties were formed (Figure 3). For the samples containing a

high percentage of CL (e.g. 10 %), the filler protruded up to 1 µm from the mold. Consequently,

as the CL percentage in the mold was reduced and the mold swelled more, the resulting FIMIC

platform topography diminished.

Figure 4a plots the AFM topographic cross section of the filler protrusion of all investigated

sample compositions with varying CL, showing that the decrease the in CL percentage of the

mold leads to great reduction and eventual levelling out of topographic difference. Moreover, a

schematic illustration is given for better understanding of the recorded FIMIC profiles. Figure 4b

shows the height of the filler sticking out relative to the mold, which is plotted against the CL

percentage of the mold. This work elegantly shows that one can use PEG based hydrogel blends

to make FIMIC substrates that are either convex, concave or close to level. However, Figure 4b

shows that although the series might incline a linear line with the crossing point to a height

different of zero, a completely levelled out sample might never be achieved precisely. This is

attributed to deviations in the FIMIC processing and standard deviations of water uptake

behaviour of employed hydrogels. Hence, no FIMIC platform regardless of its composition can

be absolutely identical to another (for more details, see PhD thesis of Zhenfang Zhang2). This

applies especially, if hydrated hydrogels are investigated in a low nanometer-scale. However,

considering a height difference of 100 nm over a distance of 50 µm, one can conclude that

these samples are essentially smooth and effects on protein adsorption and cell adhesion due

to physical patterns may be considered negligible. Any research known to the author deals with

topographies far above the here present data43–46. A study conducted e.g. by Kunzler et al.

demonstrates that osteoblasts prefer rougher areas of around Ra (arithmetic average) 5 µm,

while fibroblasts tend to adhere more on smoother substrates with Ra 1 µm45. Cellular

responses on nano topographic substrates were e.g. carried out by Gelbinger et al., which

pointed out that osteoclasts tend to interact more with surface roughnesses of Ra 500 nm,

while surface roughnesses of around Ra 10 nm caused no profound interaction with

osteoclasts46. Still, those are Ra values (RMS ~ 1.1 Ra) far beyond the here presented data. Apart

from that, the author is convinced that not absolute roughness values determine the

interaction cell-substrate but rather the curvature of the scrutinized substrate area. This means

that the cell sense changes in physical substrate orientation rather than absolute topographic

differences over a certain area. This means that the change in slope (referring to a surface area

cross section) is considered to be the prevailing factor for cells in sensing topography.

Figure 4: Graphs displaying the relationship between CL and topographic profile of the PEG1/8PEG

blend: (a) representative topographic AFM profiles with varying CL in the filler with according schematic

illustration; (b) topographic profile plotted over CL with according standard deviations.

4.2.3 Summary

First of all, the initial goal of levelling out FIMIC platforms with PEG based hydrogel blends,

while maintaining protein repellent properties of PEG, could be achieved. That was attained by

employing a blend fabricated from PEG1 and 8PEG, which unite the FIMIC processibility

requirements and water uptake ability in the desired manner. Hence, a model system with

varying CL could be designed in order bring physical patterns of FIMIC substrates to a minimum,

finding the matching combination regarding CL for both constituents; mold and filler. This

demonstrated the feasibility of the FIMIC method aiming at true 2D Patterns of Elasticity, in

which solely the hydrogels elasticity would determine the resulting cellular response to the

substrate and no topographic overruling effect occurs. In addition, pursuing this method, one

may even tune the elasticity according to the applied hydrogel systems in a manner that one

could even control on the type of cells that adhere to the provided binary mechanical surface

pattern.

Chapter 4.3

Soft Lithographic Surface Patterning of Physically and Chemically

Mineralized Poly(ethylene glycol) Hydrogels for Selective Interface

Interaction

Abstract

Understanding protein adsorption and cell adhesion behavior on engineered surfaces and

interfaces is essential for the successful development of novel biomaterials. Therefore, the main

influential factors, i.e. topography, chemistry and elasticity were examined, comparing (I)

physically mixed Poly(ethylene glycol)-Hydroxyapatite (PEG HAp NPs) and (II) chemically

synthesized Poly(ethyleneglycol)-Hydroxyapatite (PEG nHAp) based nanocomposite hydrogels.

Both, the intrinsic protein repellent behavior of PEG-based hydrogels and the protein attractive

properties of HAp-based ceramics were exploited in order to produce surface patterns of

bioactive and bioinert properties. The recently developed soft lithographic method, Fillmolding

In Capillaries (FIMIC), an easy bench top method with the employment of a new bioactive

nanocomposite is paired in order to produce unique chemical and mechanical micro patterns.

Aiming to focus on the impact of HAp to study preferred protein adsorption and cell adhesion,

FIMIC patterns were tuned to exhibit virtually no topography in hydrated state (approximate

height difference of 50 nm over 10 – 20 µm lateral dimensions ). Materials characterization,

investigating the chemical and morphological structure of PEG nHAp as a homogeneous

nanocomposite hydrogel containing nanometer-sized domains of crystalline HAp were obtained

via Electron microscopy, X-ray diffraction (XRD) and spectroscopic methods (FtIR, RAMAN).

Physical surface properties of either FIMIC platforms in hydrated state were ascertained with

Atomic Force Microscopy (AFM) and Energy-Dispersive X-ray spectroscopy (EDX) mapping,

confirming the spatial control of physical, mechanical and chemical constitution, respectively.

Still, only for chemically synthesized PEG nHAp hydrogel composites a pronounced level of

differential protein adsorption and locally controlled homogeneous HAp deposition from

Simulated Body Fluid (SBF) possessing distinct interfaces could be observed. The FIMIC

platforms that were processed with PEG nHAp, demonstrated a unique HAp deposition

monolayer-pattern, which can be tuned over immersion-time in SBF and result in a micro-

meter-sized pattern made of bioinert PEG and bioactive HAp globules. Ultimately, cell adhesion

experiments were conducted, in which the cell adhesion behavior of fibroblasts and osteoblasts

were scrutinized on PEG nHAp. These investigations revealed that calcium phosphate affine

osteoblasts respond directly to the FIMIC platform and align along the incorporated HAp

pattern, while in contrast, only a few fibroblasts adhere, not recognizing the surface pattern and

distributing rather randomly. Beyond that, osteoblasts display a spread morphology indicating

strong interaction with the FIMIC surface, while fibroblast exhibit a rather round morphology

attributed to little surface-cell interaction.

Loebus A, Zhang Z, de Vicente Lucas G., Strehmel C, Arafeh M, Lensen MC, Soft lithographic

Surface Patterning of Physically and Chemically Mineralized Poly(ethylene glycol) Hydrogels for

Selective Interface Interaction; to be submitted

Incentives

Spatial control of cell-adhesion on biomaterials surface is of great interest in the understanding

of cellular response in general. Consequently substantial effort has been invested and many

systems have been investigated in order to reveal influential factors, allow classification and set

standards for evaluation methods7,8. As main sources determining the final performance of

biomaterials-surface-patterns topography9,47,48, chemistry49,50 and elasticity51,52 could be

elaborated.

Physical patterning (topographic features) can be elegantly induced via e.g. soft lithographic

methods such as Molding or Imprinting, which were introduced by Whitesides et al. in the

1990s11. Another promising approach in this field is the recently developed Fillmolding In

Capillaries (FIMIC) method 9, which enables the fabrication of sub micrometer Patterns of

Elasticity; that are in hydrated state, horizontally perfectly plane surface patterns with

alternating elasticity, which simultaneously allow chemical patterning in the lateral micrometer

scale.

In this study, poly(ethylene glycol) (PEG)-based hydrogels and PEG nanocomposite materials

containing hydroxyapatite (HAp) are utilized to fabricate patterned surfaces with defined areas

of bioinert and bioactive properties, respectively. It is well known that PEG is widely recognized

for its protein-repellent behavior, which have resulted in different applications such as contact

lenses or protective coatings or drug-delivery vessels18,53–55. Moreover, PEG based hydrogels

have been successfully applied as medium the illuminate cell-repellent surface behavior

towards cell-adhesive cites9,51. In contrast, calcium phosphate based ceramics12,13,56 e.g.

Amorphous Calcium Phosphate (ACP), Tri Calcium Phosphate (TCP) or HAp often find application

in biomaterials when increase in bioactivity or improved mechanical properties of polymer

hydrogels are desired54,57–60.

ResultsandDiscussion

4.3.1MaterialsCharacterization

Polymeric composite materials may be fabricated by simple mixing of an inorganic material

phase into a polymeric network. In this project, two fundamentally different approaches were

pursued: First, the physical mixing of HAp nanoparticles (NPs) into a PEG matrix (denoted as PEG

HAp NPs) and secondly, the chemical synthesis of a PEG HAp composite (denoted as PEG nHAp)

via in-situ nucleation, in which the HAp domains nucleate and maturate within the PEG matrix.

Details to synthesis of both formulations are given in chapters 3.1.5 and 3.1.6, respectively. The

aim of the hydrogel composite synthesis via in-situ HAp nucleation was to avoid any influence

from HAp aggregation and guarantee surface accessibility of introduced chemical functionalities

in order to retain precise local control of post-gelation surface modification. Figure 1 depicts the

materials synthesis and according materials notation as well as the lucency of the hydrogel

composites and the suitability for FIMIC processing. Figure 1a displays the according details of

the fabrication process for chemically synthesized PEG nHAp composite hydrogels. Step I depicts

the Calcium- and Phosphate-salt solution addition to the photoinitiator (PI) containing PEG (Mw

575 g/mol) precursor. Step II displays the maturation stage of the hydrogel composite involving

vigorous magnetic stirring, while filtering of the aqueous hydrogel composite was conducted in

order to receive the final product as described in Step III. After step III, the composite precursor

is processible for soft lithographic regime; subsequent ultraviolet (UV) radiation is applied when

the liquid-to-solid conversion, e.g. for substrate fabrication, is desired. Three different types of

chemically synthesized hydrogel composites (PEG nHAp) were fabricated. Hydrogel composite

materials notations with according HAp content by weight, which find application in this

chapter, are depicted in Figure 1b. One should note the absence of water in any of the stated

materials compositions. Additionally, the table in Figure 1b displays the varying degrees of

lucency of the respective hydrogels composites, an important aspect for the characterization of

cell adherence patterns with optical microscopy. The ability to incorporate either hydrogel

composite in FIMIC fabrication is also given, which relates to the viscosity of the hydrogel

composite pre-cursor. In case the viscosity is too high, capillary action cannot take place or

grooves of the hydrogel molds are filled incompletely (for more details to FIMIC fabrication, see

chapter 2.3).

Figure 1: Materials: (a) synthesis process of PEG nHAp: adding of salt solutions (I), maturation (II),

filtering (III) and UV curing (IV); (b) notation of applied hydrogel composites and according HAp

percentage by weight, degree of lucency of the hydrogels and the processibility for FIMIC fabrication.

Materials characterization proving the chemical and morphological structure as a homogeneous

nanocomposite hydrogel containing nanometer-sized domains of crystalline HAp were obtained

via Electron microscopy, Elemental Mapping via EDX, X-ray diffraction (XRD) and spectrometrical

methods (FtIR, RAMAN) (Figure 2). Since IR and RAMAN spectroscopy and X-ray diffraction are

often used to qualitatively detect and discern calcium phosphate phases56, the as prepared

samples were characterized with FtIR, RAMAN and XRD analysis in order to prove distinctive

bands (FtIR), shifts (RAMAN) and reflections (XRD) for HAp formation.

FtIR measurements (Figure 2a) revealed distinctive bands at 560 cm-1 and 598 cm-1, which can

be attributed to the triply degenerated bending mode, ν4, of the O-P-O bond of the PO42- group.

Furthermore the weak peak at 474 cm-1 and the shoulder at 462 cm-1 indicate the doubly

degenerated bending mode, ν2, which can be clearly discerned at least for PEG 40nHAp. All the

other characteristic peaks e.g. 1032 cm-1, 1046 cm-1 and 1087 cm-1 (triply degenerated

asymmetric stretching mode vibration, ν3, of the P-O bond) for HAp formation are obscured by

bands from the PEG575 polymer-matrix and therefore cannot be clearly identified. This

shortcoming can be compensated by RAMAN spectroscopy.

The recorded RAMAN spectra (Figure 2b) exhibit an increasing peak at approximately 965 cm-1

(ν1PO4) with augmenting HAp content, which derive from a totally symmetric non-degenerated

stretching mode of the free tetrahedral phosphate ion. This is in agreement with literature,

where this peak is widely recognized as identification site for HAp presence56,61,62. In addition, a

broader peak at approximately 1050 cm-1 (ν3PO4) was observed, which can be assigned to

amorphous calcium phosphate-like mineral62. The decrease of the peak intensity ratio of 1050

cm-1 to 965 cm-1 with rising HAp content implies the transition from ACP to HAp. Furthermore, it

is a widely accepted fact that ACP is a precursor for HAp formation56,62. The spectrum of HAp

NPs is additionally given for comparison.

XRD measurements can be ideally used in order to verify phase purity56. In this case, it proved

the presence of reflections typically assigned to HAp at 2theta-values of 26 ° and 32 ° (Figure

2c). The intensity of the signal increases with HAp content indicated by the black arrow. PEG

10nHAp is not depicted due to very small intensity of reflections in the according spectra.

Furthermore, the spectra demonstrate that an increased incorporation of HAp also results in

enhanced crystallinity of the HAp phase. In order to reliably quantify this aspect or even state

about crystal sizes, Rietveld Refinement would be necessary63, which was not conducted in the

present study.

The morphology of as prepared PEG nHAp composite hydrogels was investigated via

Transmission Electron Microscopy (TEM) in order to identify HAp formation and comment on

HAp distribution within the PEG nHAp nanocomposites. As representative example of the three

investigated compositions PEG 20nHAp is shown and analyzed. The TEM images (Figure 2d)

reveal crystalline areas distributed throughout the composite sample, which implies a high level

of uniformity down to the nanoscopic scale. This can be attributed to the in-situ nucleation of

HAp, since thus, agglomeration as often observed for physically incorporated HAp NPs is

prevented64. The in-situ X-Ray diffraction pattern (Figure 2e) demonstrates the crystalline

character of the nucleated HAp-phase, where the recorded reflections were characterized and

quantified starting from the most centered ring (largest crystallographic dimensions) to the

outer ones. The diffraction data allowed identification of the present crystallographic net-planes

with corresponding crystallographic net-plane distances and 2 theta angles. The obtained

results are in very good agreement with the XRD spectrum, since the most characteristic

reflections at 26 ° and 32 ° were amongst others detected and according net-plane distances

calculated (see table Figure 2e). Moreover, it was found that the measured distances are in very

good agreement with literature for crystalline HAp56. In-situ EDX analysis was applied in order to

prove the presence of the elements Calcium (Figure 2g) and Phosphorous (Figure 2h). The

traces of Silicon and Magnesium can be attributed to impurities during sample preparation,

while grinding the particles. Copper can be explained by the usage of a copper TEM grid during

sample preparation.

Figure 2: Chemical and morphological characterization of PEG nHAp. FtIR spectra (a) displaying

distinctive bands at 560 cm-1 and 598 cm-1, RAMAN absorption spectra; (b) with distinctive bands at ca.

965 cm-1 and approximately 1050 cm-1 and XRD analyses; (c) distinctive reflections of HAp of the

different composite matrices at 26 ° and 32 °; (d) TEM micrograph with visible net-planes of the

composite; (e) to the electron diffraction pattern (miller indices) with according diffraction angle net-

plane distances; (f-h) elemental mapping of plane composite films with scale bar representing 20 µm (f)

micrograph (g) Ca-map (h) P-map.

In addition, rheological measurements of the hydrogels in completely hydrated state were

conducted in order to quantify the bulk mechanical properties of the hydrogel composite as

detailed in Table 1. They revealed, that after ultraviolet (UV) radiation treatment the hydrogel

composite are considerably softer than pure PEG (decrease almost 5-fold), which is tentatively

attributed to the disturbance of chemical crosslinking in the presence of HAp nano-domains65–

68. Remarkably though, it was observed that PEG 20nHAp exhibits the highest Young’s Modulus

(3.45 MPa) in comparison with PEG 10nHAp and PEG 40nHAp (2.38 MPa and 2.41 MPa

respectively). For non-UV radiated hydrogel composites, it is presumed that the HAp domains

increase the stiffness of the polymer-network54,69–71. In contrast, those crystallites may disturb

chemical bond formation during UV-curing72,73. This effect is the more dominant the smaller the

molecular weight of the polymer-matrix, since the crosslinking density increases with

decreasing molecular weight. Thus, the introduction of HAp partially limits the hydrogel

network formation and consequently the mechanical integrity, which strongly depends on the

number of chemical crosslinks formed. Hence, in the case of PEG, one observes this effect at an

elevated level leading to the here monitored result, in which small amounts of HAp interfere

potentially very strongly in the curing process. However, results imply that HAp incorporation

contributes only to a small degree to the strengthening of the composite hydrogel via physical

interaction. The incorporation of greater amounts of HAp (as for PEG 40nHAp) nevertheless,

either do not disturb significantly more the chemical bond-formation than little amounts of HAp

during UV curing, or do not elicit stronger composite hydrogels due to limited secondary

interaction processes between the organic and inorganic phase .

Table 1: Bulk elasticity obtained via rheology of the investigated PEG nHAp composite hydrogels with

varying concentrations of HAp.

The chemically synthesized composite hydrogels PEG 10nHAp and PEG 20nHAp were fluidic and

transparent, while PEG 40nHAp had a considerably higher viscosity and appeared rather

opaque, consequently PEG 20nHAp was chosen for further soft lithographic processing due to

the highest amount of HAp incorporation and the processibility. The superb degree of

patternability down to sub-micrometer range is demonstrated in Figure 3.

Figure 3: Line patterns in the sub-micrometer scale fabricated via Replica Molding applying PEG 20nHAp.

For comparison reason, a conventional, physically mixed hydrogel composite (PEG 20 HAp NPs)

(HAp Ø 20 nm) was fabricated. Therefore, HAp NPs were successfully synthesized via the wet

chemical route, which resulted in rod-like HAp NPs with a size of approximately 20 nm. Those

were physically mixed into the PEG precursor solution with a weight percentage of 20 % with

respect to the entire composite mass. In order to achieve best homogenicity of the resulting

PEG HAp NPs hydrogel composite, extensive vortexing and subsequent ultrasonic treatment

were applied. The final suspension was rather opaque exhibiting a low viscosity, which enabled

easy application in soft lithographic regimes. Surface roughness in form of the Root-Mean-

Square (RMS) measure was determined in hydrated state in order to verify that the unpatterned

plane RMS value would be in the range of the substrate used (glass or silicon wafer), hence

guaranteeing processing quality. This is displayed in Figure 4, which demonstrates that although

PEG 20nHAp exhibits expectedly a greater RMS value (approximately 2.0 nm) than that of pure

PEG (approximately 0.6 nm), ascertained values are still in very low nanometer range and to the

best of the author’s knowledge not detectable by cells. At least any literature known to the

author centers RMS far above the here stated values43–46,74 (see chapter 4.2.2).

Figure 4: Surface Roughness in form of the Root-Mean-Square (RMS) value of (a) PEG575 and (b) PEG

20nHAp in completely hydrated state.

4.3.2SurfacePatterningandCharacterization

The general idea of the intended patterning process involving FIMIC production and selective

HAp deposition from Simulated Body Fluid (SBF) is schematically depicted in Figure 5a. As

shown two different FIMIC sample platforms are fabricated and compared. Those FIMIC

platforms contain regions of pure PEG, which are in comparison stiff and bioinert as the

adjacent surface areas contain PEG HAp hydrogel composites and are rather soft and bioactive.

As will be demonstrated, both physically mixed and chemically synthesized hydrogel composites

(PEG HAp NPs and PEG nHAp respectively) were utilized in the FIMIC method in order to gain

micropatterned surface structures, exhibiting defined areas with and without HAp presence.

Subsequently, as fabricated FIMIC platforms were immersed into SBF up to 10 days in order to

assess the locally controlled HAp deposition as final patterning step. Thus, it is intended to

ascertain post-fabrication functionalization and to construct multiple-protein-affine substrates.

Those are substrates, which exhibit different degrees of protein affinity on adjacent areas (may

be in a regular pattern as for FIMIC platforms). Figure 5b depicts a schematic cross section of

the fabricated FIMIC platforms which are chemically (introduction of HAp) and mechanically

(different elasticity, see Table 1) patterned; either with PEG nHAp or with PEG HAp NPs. The

hydrogel mold consists of pure PEG (only elements are Carbon, Hydrogen and Oxygen) and are

cell anti adhesive, while the grooves contain Calcium and Phosphorus, are bioactive and are

considerably softer than pure PEG in ultraviolet radiated state.

Figure 5: (a) Scheme of the intended patterning process involving FIMIC fabrication and HAp nucleation

from SBF including PEG nHAp and PEG HAp NPs. (b) schematic depiction of FIMIC cross section.

FIMIC production (as shown in detail in chapter 2.3) is an easy bench-top method, which allows

micrometer precise fabrication of hydrogel patterns suitable for the investigation of potential

interface interactions. The application of this process with PEG HAp NPs proved to be straight

forward; however the non-homogeneous nature of this class of physically interacting hydrogel

composites demonstrated non-uniform HAp distribution as shown in Figure 6. Figure 6a displays

two optical topviews of FIMIC substrates containing HAp NPs. The according PEG HAP NPs areas

with according HAp NPs, visible as agglomerates, are indicated by the green arrows. Figure 6b

depicts a SEM topview with indicated areas of EDX elemental analysis for Calcium (Ca) and

Phosphorus (P) (areas 1 and 2). For area 1, both elements can be clearly detected, while in area

2 the absence of both is detected. Figure 6c shows a SEM cross section of Figure 6b

investigating the volume distribution of HAp NPs within the PEG HAp NPs filled grooves via EDX

analysis. For surface-near (area 3) and surface-far (area 4) areas, moreless equal amounts of

Calcium and Phosphorous can be measured implying a rather homogeneous distribution of HAp

in the micro-scale.

Figure 6: FIMICs containg PEG HAp NPs (a) optical image of FIMICs containing PEG HAp NPs with HAp

agglomerates; (b) SEM image of FIMICs containing PEG HAp NPs with according EDX analysis from the

top; (c) SEM image of cross-section of FIMICs containing PEG HAp NPs with according EDX analysis.

Even though the processed FIMIC platforms seem intact and well defined, still no final

conclusion about successful HAp incorporation may be drawn. This is in particular the case

when dealing with homogeneous HAp NPs distribution and surface accessibility in order to

allow interface interaction. Some distributed HAp NPs agglomerates are detectable via optical

microscopy, still precise control of amount and placing of HAp NPs is missing. In order to

ascertain the difference of HAp incorporation of either hydrogel composite into the patterns,

SEM micrographs with according elemental mappings, namely of Calcium and Phosphorus were

recorded (Figure 7). Distinction can be easily addressed, since elemental proof of Ca and P for

PEG HAp NPs are restricted to the agglomerates or are of so little concentration that they fall

below the threshold of detection (Figure 7a-c). On the contrary, homogeneous distribution

covering the desired areas was achieved for PEG nHAp composites verifying the concept of

homogeneously chemically functionalized 2D Patterns of Elasticity (Figure 7d-f). Thus it was

possible to apply a novel hydrogel composite via soft lithographic processing regime, gaining

spatial control of chemical patterns with micrometer precision.

Figure 7: SEM micrographs and elemental mapping (Ca, P) of as prepared FIMIC platforms (PEG575 PEG

HAp NPs/PEG nHAp) containing physically mixed and chemically synthesized PEG-HAp composites. (a-c)

physically mixed PEG HAp NPs composites (a) micrograph of FIMIC platform with PEG HAp NPs (b) Ca-

map (c) P-map; (d-f) chemically synthesized PEG nHAp composites d) micrograph of FIMIC platform with

PEG-nHAp composite; (e) Ca-map; (f) P-map; (scale bar 20 µm).

As concluding patterning step on FIMIC platforms, both sets of FIMIC substrates were incubated

in SBF in order to achieve multiple-protein-affine surfaces and to assess the post-FIMIC

fabrication patternability. Therefore, locally controlled, time-resolved HAp crystallite monolayer

deposition with spatial micrometer precision was conducted to demonstrate the selective

interface reactivity of the investigated FIMIC samples.

The general idea was to verify surface interaction potential, since HAp deposition from SBF

requires nucleation sites (in general charged surface cues), not facilitated by the herein utilized

PEG derived hydrogels18,34,75. Hence, any HAp deposition detectable can be attributed to

chemical surface patterning with reactive ions from HAp domains. Therefore, pre-patterned

FIMIC samples containing physically mixed PEG HAp NPs and chemically synthesized PEG nHAp

hydrogel composites were immersed in SBF over a 2, 4, 6, 8 and 10 day period. Figure 8 displays

the HAp crystallite deposition on PEG nHAp and PEG HAp NPs over several time spans. Figure

8a-e demonstrate the continuous homogeneous HAp deposition over time on PEG nHAp, in

which Figure 8a depicts small isolated HAp nuclei after 4 days, Figure 8b the almost dense layer

of single nuclei after 8 days and Figure 8c an uniform monolayer after 10 days. Figure 8d and

Figure 8e confirm the density and monolayer morphology of the HAp nuclei, along with

crystallite shape of the single HAp nuclei well known from literature76–78. Figure 8f displays the

HAp deposition structure after 10 days on patterned substrates containing PEG-HAp NPs

composite.

Precisely controlled spatial deposition of HAp nuclei from SBF over time on the as prepared

FIMIC hydrogel patterns demonstrate the superior tailoring ability and versatility of the

chemically synthesized hydrogel composite PEG nHAp over its physically mixed PEG HAp NPs

counterpart. In contrast, even though HAp disposition to the designed areas could be observed,

both, the non-uniformity of nucleated HAp crystallites together with the formation of large

agglomerates verify the rather limited potential of PEG HAp NPs as platform for controlled

interface interaction.

Consequently, this simple patterning method, combined with precise control of determining

parameters yield common interest, since patterning via SBF deposition so far generally

generates perturbed interfaces24, may only proceed after lavish processing27 or result from

transferred bioactive nanoparticles agglomerates22.

Figure 8: SEM micrographs of locally controlled time resolved HAp deposition from Simulated Body Fluid

(SBF) on as prepared FIMIC platforms containing chemically synthesized hydrogel composites (PEG

nHAp) and for comparison reason physically mixed hydrogel composites (PEG HAp NPs). (a-e) FIMICs

containing PEG nHAp (a) HAp nuclei after 4 days of immersion in SBF; (b) HAp nuclei after 8 days of

immersion in SBF; (c) HAp nuclei after 10 days of immersion in SBF; (d) HAp nuclei after 10 days of

immersion in SBF - close up of interface; (e) HAp nuclei after 10 days of immersion-monolayer; (f)HAp

nuclei and agglomerates on FIMIC containing PEG HAp NPs a after 10 days of immersion in SBF; (scale

bar (a), (b), (c) and (f) 20 µm; scale bar (d) and (e) 2 µm ).

4.3.3 Cell Adhesion of Fibroblasts and Osteoblasts on Fillmolding In Capillaries

(FIMIC)PlatformsPatternedwithPEGnHAp

As preceding step to cellular response to biomaterials surfaces, proteins interact and may show

more or less affinity towards the respective surface. In order to verify the anticipated

differential protein adsorption pattern on either FIMIC platforms, the model protein Bovine

Serum Albumin (BSA) was utilized to conduct protein adsorption measurements. It is well

known from literature that BSA and calcium phosphates share electrostatic interaction between

Ca2+ (HAp surface) and COO- (protein surface), in particular when HAp exhibits amorphous

characteristics 79,80 and may even impact the morphology of deposited HAp nuclei77. Hence, the

experiment was designed to ascertain the accessibility of HAp domains to interface interaction,

indispensable for successful chemical surface patterning.

As prepared FIMIC platforms with PEG HAp NPs and PEG nHAp were investigated and compared

as shown in Figure 9. As expected, patterns exhibiting chemical composites PEG nHAp not only

impart homogeneous BSA adsorption over the entire area functionalized with HAp, but possess

in addition a precisely defined spatial interface of protein-repellent and protein-adhesive areas

(Figure 9a-c). That makes the FIMIC platform a perfect tool for the investigation of selective

protein adsorption experiments as well as for spatially controlled further functionalization with

HAp affine chemical or biological cues.

Scrutinizing the adsorption pattern of proteins on composites containing PEG HAp NPs, a

distinctively different image is gained. One can clearly see the HAp agglomerates as potential

BSA adsorption sites (Figure 9d), however, no continuous HAp NPs incorporation can be

determined as no homogeneous BSA adsorption pattern on the FIMIC substrate areas

containing PEG HAp NPs can be detected (Figure 9e). This is in agreement with the findings

depicted in Figure 7a-c, where Calcium and Phosphorus of PEG HAp NPs can only be detected

locally as agglomerates; instead of a homogeneous distribution of HAp domains as for PEG

nHAp (see Figure 7d-c). In some cases even, although the agglomerates are clearly visible in the

optical image, no BSA adsorption can be noted, which can potentially be explained by a polymer

layer (cover) around the particles covering all potential sites of interaction. This can be seen in

Figure 9f, in which the inset demonstrates the absence of any BSA adsorption in the investigated

sample (inset of Figure 9f, image enlarged in appendix SI 1).

As consequence of the different BSA adsorption patterns for PEG nHAp and PEG HAp NPs,

subsequent cell experiments with osteoblasts and fibroblasts were conducted on FIMIC

platforms containing PEG 20nHAp. That formulation was selected due to the combination of

highest degree in HAp incorporation yielding highest potential for cell adhesiveness and ease of

pattern processibility.

Figure 9: Confocal Laser Scanning Microscopy (CLSM) optical and fluorescent images of FIMIC platforms

containing PEG NPs HAp and PEG nHAp hydrogel composites depicting selective adsorption of Bovine

Serum Albumin (BSA). (a-c) PEG nHAp (a) optical image; (b) fluorescent image; (c) fluorescent image

close up; (d-f) PEG NPs HAp composites (d)optical image depicting one large HAp and several small HAp

clusters; (e) according fluorescent image depicting non homogeneous BSA adsorption; (f) HAp

agglomerates visible within PEG NPs HAp but no BSA adsorption detectable (black inset).

The physical and mechanical properties of biomaterials surfaces impose great influence on

cellular behavior44,49,81, especially since cells respond to substrates exhibiting submicrometer-

topographical features46 and micrometer-sized tuned elasticity82–84. Therefore, it is of crucial

importance to control the topographic and mechanical properties and determine these in

hydrated state. This is of particular interest if, as in our case for PEG based hydrogels, one wants

to assess the effectiveness of elastic and chemical patterning, which may be overruled by

physical patterning as demonstrated by Lensen et al.48.

In order to ascertain the physical and mechanical surface characteristics of as prepared FIMIC

platforms, AFM was employed, assessing the topographic and elastic properties in dry and

hydrated state (Figure 10). Among the great challenges in order to apply this technique in

surface pattering proved to be the equilibration of the swelling level (and hence the swelling

degree of the several hydrogels applied) of the investigated FIMIC platforms. While physical

patterning in dry state exhibited values up to 300 nm in topography (Figure 10a), this was

reduced to approximately 50 nm for samples in hydrated state (Figure 10c); an observation

which is of extraordinary interest, since cellular spreading (positioning) may be dominantly

guided by physical patterning such contact guidance. However, if topographic cues are in the

sub-micrometer range, cells react chiefly to chemical ones, as demonstrated by Britland and

colleagues85. This means that if the topographic features are in the nanometer range, the

cellular response and subsequent adhesion patterns are dominated by the surface chemistry. In

addition, incorporating two different sets of hydrogel systems (in this case PEG and PEG nHAp)

result in mechanical patterning as characterized by AFM-based surface force spectroscopic

experiments (according Force-maps denote areas of soft areas to light colour and stiff areas to

dark colour) as stated in Figure 10b/d. These thus produced 2D Patterns of Elasticity exhibit

micrometer-sized spatially controlled elastic properties and can moreover impart chemical cues

(e.g. HAp), combining mechanical and chemical patterning, while simultaneously reducing

topography to a minimum. These properties make FIMIC platforms with incorporated PEG nHAp

a superb tool for scrutinizing the effect of mechanical and chemical patterning on interface

interactions in the absence of topography.

Figure 10: Topographical imaging of FIMIC platforms with PEG nHAp in dry (a) and hydrated state (c) with

according force-maps (b/d), (if not differently stated, dimensions are in µm).

Since spatial control of cell-adhesion on biomaterials surface is essential, in this approach,

locally controlled cell adhesion of two different kinds of cell-lines (I) osteoblasts (mouse

osteoblast-like MC3T3-E1) and (II) fibroblasts (mouse fibroblasts L-929) is centered.

Furthermore, it was intended to guide the cellular-adhesion along the FIMIC-structures

containing HAp, since pure un-patterned PEG exhibits protein- and therefore cellular-repellent

behavior.

In consequence of the data recorded during selective protein adsorption measurements, it was

decided to investigate only those FIMIC platforms, which were fabricated with PEG nHAp due to

the more homogeneous nature. Moreover, it was decided to work with two different kinds of

cell lines, which exhibit different affinity towards calcium phosphates in order to elucidate the

effectiveness of chemical patterning of the FIMIC substrates.

In order to compare cellular adhesion behavior of both cell types, osteoblasts and fibroblasts

were seeded on the FIMIC platforms for 48 h and subsequently imaged with optical microscopy

(data not shown) and SEM (Figure 11). The observed adhesion behavior demonstrates the

potential to control spatially guided cell-substrate interaction elicited via chemical patterning

with PEG nHAp. The highly spread morphology of the osteoblasts on the FIMIC patterns indicate

strong cell-substrate interaction and hence strong cell adhesion (Figure 11a-c). Furthermore, it

can be noticed that the osteoblasts elongate along the trenches functionalized with HAp; this

can be explicitly well noted in Figure 11a, in which three single osteoblasts adhere in parallel

along the FIMIC pattern. This in contrary to investigations conducted with fibroblasts, which do

not restrict interaction along the patterning structure (Figure 11c) and moreover exhibit round

morphology (Figure 11d) implying only little affinity to the HAp patterned FIMIC substrate.

That is very interesting to note, since areas containing HAp are softer (see force map in Figure

10d). Under the presumption that cells preferentially adhere to stiffer areas than to softer as

shown by Discher and co-workers83,84, those findings may imply the overruling of mechanical

patterning by chemical cues. Moreover can be excluded that one may potentially explain the

different response patterns by the varying affinity towards surface roughness, since Kunzler et

al. demonstrated that fibroblasts show greater tendency of adherence on smooth substrates

than osteoblasts even though the experimental setup was in micrometer-range45.

Figure 11: Scanning Electron Microscopy (SEM) images of osteoblast (a/b) and Fibroblast (c/d) –adhesion

on FIMIC substrates containing chemically synthesized PEG-20nHAp composites. White globules are

Sodium-phosphate (Na3PO4) from the nutrition solution during cell-culturing as determined via EDX

analysis.

4.3.4Summary

A new PEG - nHAp nanocomposite was successfully synthesized, characterized and micro-

patterned with the recently developed soft lithographic Fillmolding In Capillaries (FIMIC)

patterning method. In order to compare the effectiveness of chemical patterning, PEG nHAp

was compared with a conventional, physically mixed hydrogel composite (PEG HAp NPs)

regarding availability of HAp on the FIMIC surface, necessary for differential protein adsorption

and localized control of post-fabrication functionalization with a HAp monolayer deposited from

SBF. The spatial control over the HAp crystallite monolayer deposition creates an interesting tool

for multiple-protein-affine materials enabling specific interface interaction. Furthermore, the

ability to tune the deposited HAp concentration over time contributes to the precise control of

the degree of surface functionality.

Ultimately, it could be shown, that osteoblasts and fibroblast display different adhesion

behavior on the PEG nHAp FIMIC platforms. It could be demonstrated that osteoblasts, which

display a high degree of affinity to calcium phosphates, adhere along the patterning structure of

the FIMIC substrates and spread highly, while fibroblasts do not tend to align along the FIMIC

pattern and show a rather round morphology, which can be attributed to weaker cell-substrate

interaction.

It can be concluded, that the herein presented chemical hydrogel composite PEG nHAp

possesses far superior potential to successfully introduce chemical patterning to biomaterials

surfaces than the physically mixed PEG HAp NPs. This is due to its inherent homogenicity and

processibility, which result in binary bioinert-bioactive PEG - PEG-HAp 2D patterns leading to

differential cellular response depending on the cellular affinity towards HAp.

The author is convinced that the findings concerning the comparison between physical and

chemical hydrogel composite materials may be generalized for many hybrid composite systems

in biomaterials research, in particular in those in which precise control over surface properties

and interfaces of biomaterials is a key issue.

Chapter 4.4

3D Patterned Reactive Mineralized Poly(ethylene glycol) Derived

Hydrogels

Abstract

Providing decent replacement for bone or tendon defects remain one of the most challenging

and demanding issues in the field of biomaterials research. In order to meet nowadays

requirements, soft cytocompatible Poly(ethylene glycol) (PEG) based hydrogels with bioactive

HAp were developed to fabricate porous and 3D patterned scaffolds with unique

physicochemical and mechanical properties.

Post-gelation reactive 8armPEG based calcium phosphate nanocomposite hydrogels with

homogeneously dispersed crystalline hydroxyapatite (HAp) nanodomains were fabricated via a

novel in-situ synthesis procedure. Spontaneous gelation upon mixing of the Calcium and

Phosphorous salt-solutions to the PEG-precursor solution indicate the formation of strong

secondary bonding, yielding a novel self-assembly process responsible for the perfect

homogenicity observed. In-situ Transmission Electron Microscopy (TEM) X-ray diffraction

analysis revealed crystalline regions in the generally amorphous nature of the here presented

composite system. Infrared- and RAMAN-spectroscopy as well as X-ray diffraction were

conducted in order to verify the formation of HAp domains and to distinguish from other

potential calcium phosphates. Ultraviolet (UV) radiation was applied in order to stabilize the

formed nanocomposite hydrogel network under physiological conditions.

The strategy to synthesize well-defined nanocomposite biomaterials relies on the in-situ

nucleation of varying amounts of HAp within the matrix of hydrogel precursors. The effect of

HAp incorporation on as prepared PEG HAp hydrogel composite scaffolds were examined via

compression testing and cell-testing, which were fabricated via freeze-drying and subsequent

UV-radiation. Composite scaffolds exhibited interconnected 3D pore structures with pore-sizes

up to 500 micrometer allowing for vascularization and medium transport, decisive for

successful tissue ingrowth.

Different amounts of HAp were incorporated in the PEG precursor matrix, from which hydrogel

composites containing 20 wt% HAp (8PEG 20HAp) demonstrated to be most suitable for

potential application in tissue engineering. They exhibit best structural properties, resulting in

most appropriate mechanical performance and highest level of selective cellular response.

Attained data was gained in cooperation with the research group from Prof. Dr. Christian Rüssel

of the Otto Schott Institute of the Friedrich-Schiller-University Jena, Germany (SEM images and

EDX line Scan of Scaffolds) and the group of Prof. Su of the Department of Materials and

Engineering of the Beijing Chemistry and Technology University, Beijing, China (optical

microscopy of osteoblast adhesion on flat 8PEG 20HAp samples and according live-dead assay).

Loebus A+, Zhang Z+, Li Q, Strehmel C, Wisniewski W, Arafeh M, Rüssel C, Su Z, Lensen MC. 3D

Patterned Reactive Mineralized Poly(ethylene glycol) Derived Hydrogels., manuscript in

preparation

Incentives

Hydrogels are of great interest in various applications such as biotechnology, tissue-engineering

and drug delivery due to their ability to take up water, porous structure and often found

cytocompatible nature 86–91. Among others, Poly(ethylene glycol)(PEG)-based hydrogels have

been extensively explored as promising material for according applications 92–96. However,

despite the many advantages provided by PEG-based hydrogels they display insufficient

mechanical rigidity and are therefore limited concerning an application as 3D matrix regarding

tissue repair or scaffold in order to deliver cells or bioactive molecules 97. Those drawbacks

motivated in incorporation of a variety of nanoparticles (NPs), which aimed to improve the

mechanical and biological properties such as laponite, hydroxyapatite (HAp) and silicate 98–101.

HAp NPs 53,55,102–105 and their bioactive characteristics106–109 have been vastly investigated and

are known to exert a strong impact on the final polymer-HAp composites systems, which find

especial interest in hard tissue engineering related to bone or tendons 6,54,57,58,60.

In the herein presented approach, an in-situ nucleation process of HAp in a PEG-based hydrogel

matrix exhibiting spontaneous gelation of a bioactive nanocomposite hydrogel is reported that

not only overcomes the issue of homogeneous distribution of HAp NPs, which is a generally

encountered problem64, but also enables to precisely tune the mechanics, the cell response and

the pore morphology.

Scaffolds fabrication and characterization have been in the focus of the hard-tissue engineering

community for almost three decades. Modern scaffold design involves an entire library of novel

requirements that are to be satisfied 110,111. Porous interconnected structures support the

diffusion of nutrients and waste as well as enable vascularization112,113 . Such 3D fabricated

devices offer an appropriate microenvironment for cells to adhere and proliferate or

incorporate drugs to stimulate specific cellular responses111,114–118 . However, major drawbacks

encountered are inappropriate mechanical properties, toxic degradation products, ill-designed

porous structure and insufficient biological performance. Several approaches of

interpenetrating networks and composite hydrogels have been proposed in order to overcome

some of the limitations mentioned99,115,119–124. However, some of the material systems

suggested are only poorly understood or only fulfill some but not all crucial requirements in a

satisfactory manner 69.

In this study focus is laid on the fabrication of 3D macro-porous homogeneous nanocomposite

hydrogels via an in-situ gelation process of hydroxyapatite (HAp)53,55,102,105 domains utilizing

PEG-based hydrogels matrix, which results in a materials system with precisely tunable

properties in terms of bioactivity, elasticity and pore-structure. That routine avoids the

common problem of agglomeration of HAp NPs64 and furthermore enables the manufacture of

scaffolds exhibiting large enough pore in order to allow for vascularization vital to successful 3D

craft ingrowth 125,126.

Results and Discussion

4.4.1 Gel-formation

It was intended to fabricate 3D macro-porous, functionalizable homogeneous nanocomposite

hydrogels via an in-situ gelation process of hydroxyapatite (HAp) domains utilizing

Poly(ethylene glycol) (PEG)-based, namely 8PEG, hydrogel matrices. That results in a material

system with precisely tunable properties in terms of pore-structure, elasticity, bioactivity, and

post-gelation reactivity. Details to important material’s properties are given in Figure 1.

Figure 1: Materials properties of raw substances utilized in this project. Characteristics determined by

following methods:* Colony forming assay (CFA), ** live dead assay, *** L-929 cells.

Successful 8PEG hydroxyapatite (8PEG HAp) nanocomposite gelformation was observed

following the in-situ regime demonstrated in Figure 2. The salt solutions were given into

reaction vessel and gelation was observed depending on the amount of NH3 added. The

obtained mineralized hydrogels were slightly opaque, almost transparent, easy processible and

could be readily shaped to any desired form. As dipicted in Figure 2a, the pH value plays a

critical role in the gelformation based on Michael type addition. It should be adjusted to around

9.8 in order to form stable hydrogel networks (more details of chemical principles are given in

the PhD thesis of Zhenfang Zhang2, a collaegue from the Lensen Lab). Hence, increased

incorporation of HAp should also lead to accelerated gelformation, since the amount of injected

NH3 also augments. Nevertheless, this cannot be monitored, but it is rather observed that

medium concentration of HAp and consequently NH3 lead to quickest gelformation. That might

implies the interference of HAp formation with Michael type addition chemistry (more detailed

description of Michael type chemistry and reaction interpretation are stated in the PhD thesis

of Zhenfang Zhang2). Proof for potential chealation of Ca2+ and PO42- ions is not found, since

neither the according FtIR- nor RAMAN-bands show any sign of shift. One can rather assume

that the HAp nuclei form a sterical hindrance for the crosslinking points. The benefit of this

reaction are on the one hand the reaction condition, allowing incorporation of biological cues,

and on the other hand the avoidance of the common agglomeration problem, familiar to

nanocomposites consisting of an inorganic and organic phase64. The in-situ nucleation of HAp

domains in 8PEG leads to secondary interaction between both phases instead of pure

mechanical entrapment of the inorganic filler phase in the polymer matrix.

Rheological measurements were conducted in order to access the gelation time in dependence

of the HAp content (see table 1). Figure 2b depicts a representative example of a gelation time

recording, which was determined as the crossover-point of the Storage Modulus (G’) and the

Loss Modulus (G”). It was observed, that against intuition, there is no clear trend of gelation

time with rising or falling HAp content. In contrast to expectations, 8PEG 20HAp (20 wt% HAp)

showed the shortest gelation-time, which indicates strongest interaction between the salt

solutions and the precursor 8PEG network. This is worth noticing, since the concentration of

Calcium- or Phosphate-ions is considerably lower than in hydrogel composites containing 40

wt% HAp (8PEG 40HAp). It furthermore demonstrates that a minimum concentration of salt

ions must be present in order to facilitate gelformation, since hydrogels with 10 wt% HAp (8PEG

10HAp) formed after approximately 18 h only, and were very soft in comparison to the other

hydrogels with higher HAp content. Figure 2c sketches the proposed reaction scheme based on

Michael type addition chemistry, in which HAp nucleates and matures in-situ, but does not play

a vital role in the facilitation of hydrogel network formation.

Additional covalent binding of the model molecule SH-dye indicates the ability to bind chemical

and biological cues posterior to network formation and processing via residual C=C bonds

(proposed chemical mechanism can be found in PhD thesis of Zhenfang Zhang2).

Figure 2: Gelformation process (a) gelformation during the reaction: (I) adding of Ca and P salts (II)

maturation time under constant magnetic stirring (III) gelified 8PEG HAp composite hydrogel; (b)

representative crossover recording of Storage and Loss Modulus during rheological measurement with

according gelation times for the different 8PEG HAp hydrogel composites; (c) proposed gelation reaction

scheme detailing the Michael type additon chemistry of the gelation reaction.

Chemical and morphological characterization was achieved via TEM, X-ray diffraction (XRD) and

RAMAN spectroscopy (Figure 3).

In order to analyze the morphology of the as prepared nanocomposite gels, TEM imaging and

in-situ electron diffraction was conducted (Figure 3a). One can clearly detect the

crystallographic net-planes representing crystalline domains (white circle) of the composite.

Electron diffraction helped identifying the corresponding net-planes assigned from the

according Miller index reflections (002) and (211) and determining the respective net-plane

distance to 2.84 Å and 3.41 Å 56.

XRD measurements (Figure 3b) can be ideally used in order to verify phase purity 56. Here, it

proved the presence of reflections typically assigned to HAp at 26 ° 2theta and 32 ° 2theta

already ascertained via TEM diffraction verifying HAp presence in the nanocomposites. As

shown, the intensity of the reflections increases with HAp content. As a consequence, 8PEG

10HAp is not depicted due to very small intensity of reflections in the according spectra.

Furthermore, the spectra demonstrate that an increased incorporation of HAp also results in

rising crystallinity of the HAp phase. However, for presented samples, XRD investigation imply

very low crystallinity, untypical for HAp, as already indicated from in-situ TEM electron

diffraction with only two reflections visible. Therefore, the Ca/P ration of the composite

hydrogel was determined, which is generally considered a helpful tool discerning the different

calcium phosphate phases. Hence, the observed low crystallinity of investigated composite

hydrogel samples might be attributed to additional presence of Amorphous Calcium Phosphate

(ACP). Energy Dispersive X-Ray diffraction (EDX) revealed a Ca/P ration of 1.35 ± 0.09 (n = 5)

supporting the assumption of the presence of ACP, since ACP exhibits low Ca/P ration down to

1.18, as reviewed by Nancollas et al.105 . This could be explained due to the fact that ACP is a

precursor-phase for the formation of HAp during the nucleation process; moreover one should

note that even calcium deficient HAp exhibits Ca/P values only down to 1.5.

Yet, final proof is delivered by RAMAN spectroscopy; bands existing at 965 cm-1 (attributed to

HAp) and 1046 cm-1 (attributed to ACP) confirm the formation of HAp in 8PEG56,62. RAMAN

spectroscopy enabled the identification of distinctive band peaks for HAp formation as depicted

in Figure 3c. The recorded RAMAN spectra exhibit a rising peak at ca. 965 cm-1 with augmenting

HAp content, which derives from a totally symmetric non-degenerated stretching mode of the

free tetrahedral phosphate ion. This is in agreement with literature, where this peak is widely

recognized as identification site for HAp presence 56,61,62. The spectrum of HAp NPs is

additionally given for comparison. In addition, RAMAN spectroscopy allowed quantification of

residual C=C double-bond (1636 cm-1) decisive for post gelation reactivity (Figure 3d), revealing

that for 8PEG 20HAp a 73 % of its initial amount of the C-double bonds are still accessible after

gelation of the nanocomposite hydrogel. This demonstrates the superb potential to specifically

tailor this unique nanocomposite material posterior to gelation.

Figure 3: Chemical characterization of 8PEG HAp nanocomposites with varying composition. (a) TEM

image of 8PEG 20HAp depicting crystalline areas (white circle) with corresponding-in-situ X-ray

diffraction pattern; (b) Powder X-ray diffraction patterns displaying 8PEG 20HAp and 8PEG 40HAp; (c/d)

RAMAN analysis of PEG HAp composite materials, demonstrating the existence of the HAp phase; Table

in (d) states the residual quantities of C=C double bonds available for further post-gelation

functionalization.

Thermogravimetric Analysis (TGA) (Figure 4) was primarily conducted in order to verify

theoretically intended incorporated chemically bound HAp, but in addition revealed different

amounts of physically and chemically bound water for varying HAp content (inset of Figure 3d).

Interestingly, 8PEG 20HAp exhibits the greatest amount of bound water and of residual mass

after complete combustion at approximately 600 °C. This implies that composite hydrogels

containing an optimum amount of HAp (8PEG 20HAp with 20 % HAp by weight) form the most

homogeneous network, chemically binding the greatest amount of HAp. Surprisingly, 8PEG

40HAp theoretically designed with a stoichiometric amount 40 wt% HAp, exhibits a lower mass

fraction of HAp after complete combustion than 8PEG 20HAp, which indicates an excess of

Calcium and Phosphate in the salt solution interfering with hydrogel network formation of

8PEG. Otherwise, the residual HAp after complete combustion should be higher than that of

8PEG 20HAp. Apparently, no hydrogel network actively incorporating the entire amount of the

Ca2+- and PO42--ions could be established. The observed buckle at approximately 200 °C may be

due to physical transitions such as vaporization or desorption of gaseous substances. The fact

that 8PEG 20HAp displays the highest degree of mass loss at approximately 130 °C may be

explained by the greatest degree of bound water. It is presumed that this is caused by the

presence of HAp, which due to its inherent polar character can interact with molecular water

and furthermore promote adsorption of physically bound water. This is supported by the

absence of this buckle for pure 8PEG, only exhibiting polymer degradation. This assumption is

further backed by the observation that for reduced presence of HAp (8PEG 10HAp) this buckle

is hardly detectable. The monitored decline at 130 °C and leveling out at 600 °C is in coherence

with the anticipated combustion process, since HAp presence is responsible for water binding

and adsorption in the 8PEG network. TGA analysis supports RAMAN data, stating that 8PEG

20HAp seems to be the hydrogel composite exhibiting the greatest homogenicity and the

highest degree of 8PEG - HAp interaction.

100

Figure 4: Thermogravimetric Analysis (TGA) measurements of 8PEG HAp hydrogel composites with

varying HAp content.

In order to characterize the elemental distribution of Calcium and Phosphorus, Scanning

Electron Microscopy (SEM) including EDX mapping was conducted as depicted in Figure 5. As a

result, both elements were found largely homogeneously dispersed within the samples

displaying areas of greater concentration (indicated by brighter colour, respectively) than in

other areas. This was a little surprising, since perfectly homogeneous samples were expected.

The observed result may be attributed to potential interactions between Ca2+ and PO42- with

the polymeric matrix during HAp maturation from ACP. However, no huge clusters or regions

without Calcium and Phosphorus were detected. This is a characteristic, which underlines the

novelty of the synthesis routine, preventing the clustering of the inorganic phase in hydrogel

101

composites produced, which is commonly observed in composites produced by the physical-

mixing techniques64.

Figure 5: EDX Mapping of Calcium and Phosphorus demonstrates homogeneous distribution of HAp

phase within the 8PEG matrix. (Scale bar 10 µm)

The readiness of incorporated HAp domains to chemically interact with the surface was tested

and proven via the homogeneous adsorption of the model protein Bovine Serum Albumin (BSA)

over the entire PEG 20HAp substrate surface (appendix SI 1). This supports the results from

elemental mapping, since it was demonstrated that for PEG 20HAp a homogeneous elemental

distribution of Ca and P over the entire sample could be monitored. This finding is in agreement

with BSA adsorption experiments of PEG nHAp (see chapter 4.3 Figure 9.).

Mechanical characterization via tensile testing demonstrated increasing elasticity, strength and

toughness (Table 1) with augmenting HAp incorporation. This finding may be explained by ionic

interaction between the inorganic ions and the polymer matrix as well as higher NH3 addition

accompanying an increase of calcium phosphate addition.

102

Table 1: Summary of the main results of chemical and mechanical characterization of the various 8PEG

HAp nanocomposites.

Concluding cytotoxicity tests were performed in order to demonstrate the cytocompatible

nature of the here presented 8PEG HAp nanocomposite hydrogels (data not shown here, but in

the PhD thesis of Christine Strehmel3).

4.4.2 3D Patterning of 8PEG HAp Hydrogel Composites

Porous, interconnected 3D scaffolds of reactive hydrogel composites were achieved via the

freeze-drying technique. Optical microscopy revealed that different amounts of HAp enclosure

yielded varying pore sizes and pore size structures, demonstrating precise tunability upon

previous swelling and HAp content (Figure 6). Most promising results were attained for 8PEG

20HAp swollen for 15 min. Due to the time-dependent water uptake (data shown in appendix SI

2, details to experimental setup can be found in the PhD thesis of Zhenfang Zhang2), hydration

was conducted in order to obtain homogeneous 3D interconnected structures. However, since

water uptake properties for all presented hydrogel composites are in close resemblance, the

monitored variations in the pore shapes with rising hydration time may be attributed

exclusively to HAp content. Due to the bimodal pore pattern displaying pore dimensions

103

ranging around 300 µm and approximately 20 µm, vascularization and medium transport is

enabled, necessary for tissue integration into the host, providing ideal ground for successful

implant ingrowth.

Figure 6: Optical images of a set of 3D scaffolds (fabricated via freeze-drying) with different amounts of

incorporated HAp and different swelling times. (Scale bar 100 µm)

Patterned hydrogel composites exhibiting 15 min hydration time were investigated in more

detail with SEM intending to identify the concrete pore shapes and sizes (Figure 7). Already

observed bimodal interconnected pore structure for 8PEG 20HAp is verified as well as

homogeneous round shaped bimodal pore shape structure (Figure 7c,d). The additional

104

detailed line scan EDX analysis for 8PEG 20HAp revealed varying intensities of Calcium and

Phosphorus is depicted in Figure 7g, which is in agreement with EDX mapping of pure 8PEG

samples (see Figure 5) (EDX line scan conducted by Dr. W. Wisnewski, FSU Jena). However, the

persistent presence of both ions could be confirmed. Lower degree in regular pore shape

implied by optical characterization (Figure 6) for 8PEG 10HAp, which exhibit rather oval shaped

pore structures (Figure 7a,b) and irregular pore structure for 8PEG 40HAp (Figure 7e,f),

respectively, is affirmed.

105

Figure 7: Scanning Electron Microscopy (SEM) images of 8PEG HAp composite scaffolds with varying HAp

content and elemental scanning of Calcium (Ca) and Phosphorus (P). (a,b) porous structure of 8PEG

10HAp; (c,d) porous structure of 8PEG 20HAp; (e,f) porous structure of 8PEG; (g) Elemental line scan of

Ca and P of 8PEG 20HAp. Scale bars (a),(c) and (e) 500 µm, (b),(d),(f) and (g) 50 µm, respectively.

In order to demonstrate the potential of post gelation reactivity of as fabricated composite

scaffolds, a SH-dye as model molecule was employed, readily reacting with accessible C=C

106

double bonds from the 8PEG matrix (details to the chemical principles and an according

chemical reaction scheme are given in the PhD thesis of the Lensen Lab colleague Zhenfang

Zhang2). Figure 8 displays a 3D image (Figure 8a,b) with selected centering horizontal cross-

section (Figure 8c-e), implying complete functionalization even in the interior (volume cross-

section) indicating complete consumption of all residual C=C reaction sites. This proof of

principle states that for any kind of functionality (e.g. biological or chemical cues), such 3D

patterned hydrogel composites can be tailored precisely to the specific need. Furthermore, it

can be claimed that the incorporation of inorganic ions do not interfere or impede with post-

gelation functionalization. It should be noted, that the seemingly entire functionalization of the

sample scaffold with the SH-dye demonstrates the interconnected nature of the pore structure,

since medium transport facilitates the SH-dye transport and provides homogeneous

distribution. Thus, SH-dye functionalization additionally serves as a model solution to illustrate

potential biological medium transport in biological application.

107

Figure 8: Confocal Scanning Laser Microscopy (CSLM) images (a,b 3D images, c-e volume cross-section)

of a 3D patterned PEG 20HAp hydrogel composite, post-gelation functionalized with an SH-dye: (a) 3D

optical image; (b) 3D fluorescent image; (c) cross section; (d) optical image; (e) overlay of fluorescent

image and optical image. Scale bar 100 µm

As a consecutive step, it was investigated whether the model protein Bovine Serum Albumin

(BSA) adsorbs on as prepared scaffolds (Figure 9). In consequence, it could be displayed that

the composite hydrogel facilitates BSA adsorption, which is attributed to chemical interaction

between Ca2+ (HAp surface) and COO- (protein surface). Literature states in addition that BSA

adsorption in particular is monitored when HAp exhibits amorphous characteristics 79,80 and

beyond that may even impact HAp nucleation morphology77. The general idea of this

experiment is to demonstrate the chemical accessibility of the HAp domains after scaffold

formation, typical to chemical composites, in which the inorganic compound is not simply

108

entrapped within the polymer matrix, but rather covalently or electrostatically linked to the

organic material.

Figure 9: Confocal Scanning Laser Microscopy (CSLM) images of volume cross-section of Bovine Serum

Albumin (BSA) adsorption in 8PEG 20HAp hydrogel composite scaffolds. (a) Fluorescence image; (b)

Optical image; (c) overlay of Fluorescence image and Optical image. Scale bar 100 µm.

4.4.3 Chemically Cross-Linked 3D Interconnected Scaffolds

Radiation curing (e.g. UV radiation) represents another form of post gelation functionalization.

In the course of this project, UV curing was conducted of as prepared 3D structured scaffolds. In

consequence, 8PEG HAp scaffolds yield promising mechanical properties and resistance

towards chemical degradation (e.g. hydrolysis). This leads to the formation of stable chemical

crosslinks in the composite network, which durably stabilizes the system for application under

physiological conditions (e.g. during cell experiments).

Mechanical testing in hydrated state of as prepared 8PEG HAp scaffolds was conducted via

compression testing (Table 2). This method was chosen, since during potential application in

tissue engineering, 3D structured crafts would most likely be exerted to compressive loading.

109

Although no significant distinction of the different examined 8PEG HAp composite hydrogels

can be detected, still a clear trend towards strongest mechanical performance can be observed.

That surprises, since enhanced amount of HAp incorporation was expected to lead to significant

increase in mechanical strength and elasticity. However, it is still in agreement with water-

uptake measurements (see appendix SI 3, experimental details, see PhD thesis of Zhenfang

Zhang2), which demonstrate that 8PEG 20HAp exhibits the lowest degree of hydration with a

pronounced plateau indicating an equilibrium hydration state after approximately 1.5 h. That

supports the findings of compression testing, namely that out of the investigated 3D structures,

8PEG 20HAp scaffolds form the most homogeneous and robust hydrogel composite network.

Hence, mechanical testing suggests that 8PEG 20HAp exhibits the tightest hydrogel network,

which is in coherence with quantification of residual C=C double bonds by RAMAN spectroscopy

(see Figure 3c,d).

Table 2: Compression testing of UV cured, fully hydrated 3D patterned hydrogel composite samples (n =

3).

In conclusion of scaffold morphology characterization and mechanical investigation of UV cured

3D patterned 8PEG HAp hydrogel composite scaffolds, 8PEG 20 HAp was selected to ascertain

the biological performance. Therefore, protein adsorption measurements with BSA and cellular

investigation with different kinds of cell lines, namely osteoblasts (mouse osteoblast like

MC3T3-E1) and fibroblasts (mouse fibroblasts L-929) were conducted. As expected, BSA

adsorption still takes place homogeneously throughout the investigated scaffolds; however,

some agglomerations of BSA occur and hence suggested HAp agglomerations are visible (Figure

110

10). One might assume that this observation is a result from molecular restructuring during the

conversion from secondary interaction to covalent bonds during UV radiation (see appendix SI

4).

Figure 10: Confocal Scanning Laser Microscopy (CSLM) images of volume cross-section of Bovine Serum

Albumin (BSA) adsorption in UV cured 8PEG 20HAp hydrogel composite scaffolds; (a) Fluorescence

image; (b) Optical image. Scale bar 100 µm.

Preliminary to cellular investigation on UV-cured scaffolds, cellular adhesion measurements on

flat, unpatterned 8PEG 20HAp samples were performed. The surface roughness of such

samples was characterized via AFM (RMS 3.0 ± 0.7 nm over 160 µm2 in fully hydrated state) in

order to exclude topography as dominant cause for cellular adhesion. Investigation employing

osteoblast-like cells as cell line for adhesion experiments (SI 5a) and cytotoxicity testing (SI 5b)

via optical microscopy displayed a great number of spread cells over the entire sample, while

only very few dead cells (dead cells are assigned in red in Figure SI 5b) after the live-dead assay

could be monitored (viable cells are depicted in green in Figure SI 5b). This suggests strong

adhesion of osteoblasts on 8PEG 20HAp substrates underlining the recognized bioactivity of

111

HAp as well as cytocompatibility of the here presented hydrogel composite (cell adhesion

experiment cytotoxicity test conducted by the group of Prof. Su, China).

In order to show the differences of adhesion between different kinds of cell-lines exhibiting

varying affinity towards calcium phosphates, fibroblasts and osteoblasts respective adhesion

behavior and spreading pattern was monitored via SEM (Figure 12 and Figure 13 respectively).

The general idea of this approach was to verify the feasibility of HAp incorporation into the

8PEG scaffolds. A significantly greater number of adhered osteoblasts were observed as shown

in Figure 12. Almost carpet-like covering presented by osteoblasts (Figure 12a,b) is in contrast

to distributed cluster-like appearance of fibroblasts (Figure 12c) on the 3D structured

substrates.

Figure 12: Adhesion characterization via Scanning Electron Microscopy (SEM) images of UV cured 3D

patterned PEG 20HAp composite hydrogels with different cell types: (a,b) osteoblasts (mouse osteoblast

like MC3T3-E1); (c) fibroblasts (mouse fibroblasts L-929).

It was revealed that osteoblasts exhibit a great degree of spreading characterized by the flat

elongated morphology depicted in Figure 13a-c. Moreover, hardly any cell clustering can be

observed. This relates to strong interaction between cell and substrate as well as to great

affinity of the cells towards the scrutinized platform. Fibroblasts however do not exhibit the

same degree of interaction, which is confirmed by the rather round morphology and apparent

112

tendency towards cell clustering (Figure 13 d-f). This can be explained by the widely recognized

affinity of osteoblasts towards calcium phosphates. Thus, it could be displayed that chemical

functionalization by HAp incorporation yields a direct effect on cellular reaction, demonstrating

the feasibility of this approach and the potential for specific application in bone or tendon

tissue engineering.

Figure 13: Spreading analysis of osteoblasts (mouse osteoblast like MC3T3-E1) and fibroblasts (mouse

fibroblasts L-929) via Scanning Electron Microscopy (SEM) imaging of UV cured 3D patterned PEG 20HAp

composite hydrogels: (a-c) osteoblasts; (d-f) fibroblasts.

113

4.4.4 Summary

A new PEG – HAp nanocomposite was successfully synthesized, characterized and 3D

patterned, designed for potential application in tissue engineering. The 3D structured post

fabrication functionalizable star PEG HAp hydrogel composite (8PEG HAp) scaffolds were

manufactured via freeze drying. The ability to be functionalized after gelformation yields the

promising potential for the produced interconnected 3D structures to incorporate a wide range

of biological or chemical cues for targeted applications. This was proven twofold; first,

readiness of HAp accessibility via the adsorption of the model protein BSA and second, the

cleavage of a model molecule in form of an SH-dye to residual C=C double bonds verifying the

ability of post-gelformation functionalization. That confirms that even after the formation of

the porous crafts, 8PEG HAp hydrogel composite scaffolds maintain the post functionalization

characteristic. The combination of novel chemical properties and processibility to a 3D design

exhibiting pore structures necessary for vascularization and medium transport impart the

potential of concerted application in tissue engineering. In addition can be stated that the clear

distinction between observed adhesion patterns of osteoblasts and fibroblast demonstrate the

successful chemical patterning with HAp as well as the ability of the herein presented scaffolds

to distinguish between cell types possessing different affinities to calcium phosphates.

114

Chapter 5

Conclusion and Outlook

The general objective of this thesis marks the spatial control of cell - substrate interaction via

concerted patterning strategies, and in consequence, to gain the ability of guided cell adhesion

behavior on platforms for potential biomedical applications. In order to attain this purpose, two

fundamentally different patterning approaches, namely surface patterning via a soft

lithographic regime, Fillmolding In Capillaries (FIMIC) in chapters 4.1, 4.2 and 4.3 and 3D

fabrication of scaffolds for potential tissue engineering in chapter 4.4 were employed.

Nevertheless, both strategies have the same principal objective in common; controlled cellular

interaction on specifically designed substrates, chemically patterned with calcium phosphates,

which may favor osteoplastic cell type over fibroblastic ones. Chapter 4.3 and 4.4 demonstrate

the successful achievement of these goals.

As main result of the FIMIC related research of this thesis, one should consider the controlled

osteoblast adhesion on specifically designed FIMIC platform areas, while fibroblasts do not

exhibit a notable pattern of interaction. The observed effect was solely obtained due to

successful localized introduction of hydroxyapatite (HAp) nanodomains into the micropatterned

systems. Two fundamentally different composite processing routes were applied and compared

regarding homogenicity and accessibility of HAp eliciting increased osteoblast response for

controlled cellular adhesion guidance. Those routes contained on the one hand the physical

introduction of pre-prepared HAp nanoparticles (NPs) into linear Poly(ethylene glycol) (PEG)

matrices (PEG HAp NPs) and on the other hand the production of a novel chemical PEG HAp

hydrogel composite (PEG nHAp). It could be demonstrated that if both composite materials

were incorporated in the FIMIC fabrication, PEG nHAp would, due to its inherent homogeneous

nature, possess much greater potential to exert controlled cellular reaction than the physically

mixed hydrogel composite PEG NPs HAp. Consequently, produced substrates exhibiting

localized chemical functionalization were applied in cell studies with osteoblasts and fibroblasts

demonstrating the potential of the fabricated FIMIC platforms to differentiate between distinct

115

cell types possessing varying affinities to calcium phosphates. This is underlined by the fact that

to the best of the author’s knowledge, topographic influence on cell adhesion patterns can be

ruled out due to the leveled out physical nature of the presented FIMIC platform. Moreover, it

could be shown that the post fabrication patterning ability of FIMIC samples patterned with

PEG nHAp is far superior to those of PEG NPs HAp, proved via precise spatial control of HAp

nucleation from Simulated Body Fluid resulting in sharp surface interfaces on prepared

multiple-protein-affine substrates (substrates which exhibit areas with different degrees

attractiveness adjacent to another), detailed in chapter 4.3.

The second part concentrates on the 3D fabrication of post fabrication functionalizable star PEG

HAp hydrogel composite (8PEG HAp) scaffolds designed for potential tissue engineering

application. The ability to be functionalized after gelformation yields the promising potential for

the produced interconnected 3D structures to incorporate a wide range of biological or

chemical cues for targeted applications. Even after the formation of the porous crafts via freeze

drying, 8PEG HAp hydrogel composite scaffolds maintain the post functionalization

characteristic. This conjuncture of novel chemical properties and processibility to a 3D design

exhibiting pore structures necessary for vascularization and medium transport impart the

potential of concerted application in tissue engineering. In addition can be stated that the clear

distinction between observed adhesion patterns of osteoblasts and fibroblast demonstrate the

successful chemical patterning with HAp as well as the ability of the herein presented scaffolds

to distinguish between cell types possessing different affinities to calcium phosphates.

The FIMIC platforms still yield the problem of delamination in hydrated state, an issue that

remains to be solved in order to make it a multifunctional device for any desired hydrogel

combination. An inherent problem to patterned platforms via soft lithography remains the

hydration characteristic of hydrogels, which may always differ to some extent leading e.g. for

the same type of FIMIC to slight changes of topographies in nanometer range. Thus, regarding

FIMIC substrates, careful characterization via AFM is generally recommended in order to

exclude the topographic effect as driving force for cell - substrate interaction.

116

Scaffolds fabricated via freeze drying generally yield the disadvantage of unsuitability to

minimal invasive therapies. However, the author is convinced that the system of 8PEG HAp

possesses the potential to form 3D crafts with precisely controllable physical and mechanical

properties after syringe ejection.

A concluding experiment for both presented strategies might be the joint seeding of osteoblast

and fibroblast, which may demonstrate the dominant adhesion pattern of osteoblasts, or at

best over time, display an expulsion process of fibroblast due to a lower degree of interaction

with the substrate.

117

References

1. Dunlop, J. W. C. & Fratzl, P. Biological Composites. Annual Review of Materials Research

40, 1–24 (2010).

2. Zhang, Z. Synthesis and Properties of Eight-arm Poly(ethylene glycol) Based Hydrogel for

Biomaterial Research. (2013).

3. Strehmel, C. title to be determined. (2013).

4. Williams, D. F. Definitions Biomaterials. 242, 72 (Elsevier: Amsterdam, 1987).

5. Wintermantel, E. & Ha, S.-W. Medizintechnik Life Science Engineering. Science 5. Auflage,

(Springer-Verlag: 2009).

6. Ratner, B. D., Hoffman, A. S., Schoen, F. J. & Lemons, J. E. Biomaterials Science: An

Introduction to Materials in Medicine. Journal of Clinical Engineering 22, (Elsevier

Academic Press: 2012).

7. Alves, N. M., Pashkuleva, I., Reis, R. L. & Mano, J. F. Controlling cell behavior through the

design of polymer surfaces. Small 6, 2208–20 (2010).

8. Wheeldon, I., Farhadi, A., Bick, A. G., Jabbari, E. & Khademhosseini, A. Nanoscale tissue

engineering: spatial control over cell-materials interactions. Nanotechnology 22, 212001

(2011).

9. Diez, M., Schulte, V. A., Etti, P. A., Ventre, M. & Lensen, M. C. Molding Micropatterns of

Elasticity on PEG-based Hydrogels to Control Cell Adhesion and Migration. Advanced

Engineering Materials 13, B395–B404 (2011).

10. Kloxin, A. M., Kloxin, C. J., Bowman, C. N. & Anseth, K. S. Mechanical properties of

cellularly responsive hydrogels and their experimental determination. Advanced

Materials 22, 3484–94 (2010).

11. Xia, Y. & Whitesides, G. M. Soft Lithography. Annual Review of Materials Science 28, 153–

184 (1998).

12. Meldrum, F. C. & Cölfen, H. Controlling Mineral Morphologies and Structures in Biological

and Synthetic Systems. Chemical Reviews 4332–4432 (2008).

118

13. Palmer, L. C., Newcomb, C. J., Kaltz, S. R., Spoerke, E. D. & Stupp, S. I. Biomimetic Systems

for Hydroxyapatite Mineralization Inspired By Bone and Enamel. Chemical Reviews 4754–

4783 (2008).

14. Gross, K. A. & Berndt, C. C. Thermal processing of hydroxyapatite for coating production.

Journal of Biomedical Materials Research 39, 580–7 (1998).

15. Bigi, a., Boanini, E. & Rubini, K. Hydroxyapatite gels and nanocrystals prepared through a

sol–gel process. Journal of Solid State Chemistry 177, 3092–3098 (2004).

16. Boccaccini, A. R., Minay, E. J. & Krause, D. Bioglass® Coatings on Superelastic NiTi Wires

by Electrophoretic Deposition (EPD). Key Engineering Materials 314, 219–224 (2006).

17. Kmita, G., Stoch, J., Jastrze, W., Stoch, A. & Broz, A. Electrophoretic coating of

hydroxyapatite on titanium implants. Journal of Molecular Structure 596, 191–200

(2001).

18. Abe, Y., Kokubo, T. & Yamamuro, T. Apatite coating on ceramics , metals and polymers

utilizing a biological process. Journal of Materials Science: Materials in Medicine 1, 233–

238 (1990).

19. Stigter, M., De Groot, K. & Layrolle, P. Incorporation of tobramycin into biomimetic

hydroxyapatite coating on titanium. Biomaterials 23, 4143–53 (2002).

20. Liu, Y., Hunziker, E. B., Randall, N. X., De Groot, K. & Layrolle, P. Proteins incorporated

into biomimetically prepared calcium phosphate coatings modulate their mechanical

strength and dissolution rate. Biomaterials 24, 65–70 (2003).

21. Zhu, P., Masuda, Y. & Koumoto, K. The effect of surface charge on hydroxyapatite

nucleation. Biomaterials 25, 3915–21 (2004).

22. Luz, G. M., Boesel, L., Del Campo, A. & Mano, J. F. Micropatterning of bioactive glass

nanoparticles on chitosan membranes for spatial controlled biomineralization. Langmuir

28, 6970–7 (2012).

23. Matsumoto, T., Ozawa, N. & Yao, T. Micropatterning of Apatite by Using CaO-SiO2 Based

Glass Powder Dispersed Solution. Key Engineering Materials 254-256, 911–914 (2004).

24. Okada, H., Yamaguchi, S., Hasegawa, S., Neo, M. & Nakamura, T. Development of Apatite

Micropattern Test Specimen for Cell Operation. Key Engineering Materials 309-311, 663–

666 (2006).

119

25. Ozawa, N. & Yao, T. Micropattern formation of apatite by combination of a biomimetic

process and transcription of resist pattern. Journal of Biomedical Materials Research 62,

579–86 (2002).

26. Shi, J., Alves, N. M. & Mano, J. F. Thermally Responsive Biomineralization on

Biodegradable Substrates. Advanced Functional Materials 17, 3312–3318 (2007).

27. Pittrof, A., Bauer, S. & Schmuki, P. Micropatterned TiO₂ nanotube surfaces for site-

selective nucleation of hydroxyapatite from simulated body fluid. Acta Biomaterialia 7,

424–431 (2011).

28. Farshchian, B., Choi, J., Amirsadeghi, A., Lee, J. & Park, S. 3D nanomolding for lab-on-a-

chip applications. Lab on a Chip 12, 4764–71 (2012).

29. Tekin, H., Sanchez, J., Landeros, C., Langer, R. & Khademhosseini, A. Controlling spatial

organization of multiple cell types in defined 3D geometries. Advanced Materials 24,

5543–7, 5542 (2012).

30. Roach, P., Eglin, D., Rohde, K. & Perry, C. C. Modern biomaterials: a review - bulk

properties and implications of surface modifications. Journal of Materials Science.

Materials in Medicine 18, 1263–77 (2007).

31. Pezron, E. & Magny, B. Modeling of UV oligomers properties. European Coating Journal

9, 602–606 (1996).

32. Zhou, F., Zheng, Z., Yu, B., Liu, W. & Huck, W. T. S. Multicomponent polymer brushes.

Journal of the American Chemical Society 128, 16253–8 (2006).

33. Shahin, V., Ludwig, Y., Schafer, C., Nikova, D. & Oberleithner, H. Glucocorticoids remodel

nuclear envelope structure and permeability. Journal of Cell Science 118, 2881–2889

(2005).

34. Müller, L. & Müller, F. A. Preparation of SBF with different HCO3- content and its

influence on the composition of biomimetic apatites. Acta Biomaterialia 2, 181–189

(2006).

35. Flory, P. J. Principles of Polymer Chemistry. (University Press: Cornell, 1953).

36. Bilodeau, G. Regular pyramid punch problem. Journal of Applied Mechancis 59, 519–23

(1992).

37. Buxboim, A., Rajagopal, K., Brown, A. E. X. & Dennis, E. How deeply cells feel: methods

for thin gels. J. Phys. Condens. Matter. 22, (2011).

120

38. Kuo, C.-H. R., Xian, J., Brenton, J. D., Franze, K. & Sivaniah, E. Complex stiffness gradient

substrates for studying mechanotactic cell migration. Advanced Materials 24, 6059–64

(2012).

39. Maloney, J., Walton, E., Bruce, C. & Van Vliet, K. Influence of finite thickness and stiffness

on cellular adhesion-induced deformation of compliant substrata. Physical Review E 78,

041923 (2008).

40. Cortese, B., Gigli, G. & Riehle, M. Mechanical Gradient Cues for Guided Cell Motility and

Control of Cell Behavior on Uniform Substrates. Advanced Functional Materials 19, 2961–

2968 (2009).

41. Sen, S., Engler, A. & Discher, D. Matrix strains induced by cells: Computing how far cells

can feel. Cell. Mol. Bioeng. 2, 39–48 (2010).

42. Buxboim, A., Ivanovska, I. L. & Discher, D. E. Matrix elasticity, cytoskeletal forces and

physics of the nucleus: how deeply do cells “feel” outside and in? Journal of Cell Science

123, 297–308 (2010).

43. Ball, M., Grant, D. M., Lo, W.-J. & Scotchford, C. A. The effect of different surface

morphology and roughness on osteoblast-like cells. Journal of Biomedical Materials

Research. Part A 86, 637–47 (2008).

44. Curtis, a & Wilkinson, C. Topographical control of cells. Biomaterials 18, 1573–83 (1997).

45. Kunzler, T. P., Drobek, T., Schuler, M. & Spencer, N. D. Systematic study of osteoblast and

fibroblast response to roughness by means of surface-morphology gradients.

Biomaterials 28, 2175–82 (2007).

46. Geblinger, D., Addadi, L. & Geiger, B. Nano-topography sensing by osteoclasts. Journal of

Cell Science 123, 1814–1814 (2010).

47. Bartneck, M., Schulte, V. A., Diez, M., Lensen, M. C. & Zwadlo-Klarwasser, G. Induction of

specific macrophage subtypes by defined micro-patterned structures. Acta Biomaterialia

6, 3864–3872 (2010).

48. Lensen, M. C., Schulte, V. A., Salber, J. & Möller, M. Cellular responses to novel,

micropatterned biomaterials. Pure and Applied Chemistry 80, 2479–2487 (2008).

49. Lim, J. Y. & Donahue, H. J. Cell sensing and response to micro- and nanostructured

surfaces produced by chemical and topographic patterning. Tissue Engineering 13, 1879–

91 (2007).

121

50. Williams, D. F. On the mechanisms of biocompatibility. Biomaterials 29, 2941–2953

(2008).

51. Schulte, V. A., Diez, M., Hu, Y., Möller, M. & Lensen, M. C. Combined influence of

substrate stiffness and surface topography on the antiadhesive properties of Acr-sP(EO-

stat-PO) hydrogels. Biomacromolecules 11, 3375–3383 (2010).

52. Nemir, S., Hayenga, H. N. & West, J. L. PEGDA hydrogels with patterned elasticity: Novel

tools for the study of cell response to substrate rigidity. Biotechnology and

Bioengineering 105, 636–44 (2010).

53. Cazalbou, S., Combes, C., Eichert, D. & Rey, C. Adaptative physico-chemistry of bio-

related calcium phosphates. Journal of Materials Chemistry 14, 2148 (2004).

54. Neffe, A. T., Loebus, A., Zaupa, A., Müller, F. A. & Lendlein, A. Gelatin functionalization

with tyrosine derived moieties to increase the interaction with hydroxyapatite fillers.

Acta Biomaterialia 7, 1693–701 (2011).

55. Vallet-Regi, M. Calcium phosphates as substitution of bone tissues. Progress in Solid

State Chemistry 32, 1–31 (2004).

56. Uskokovic, D. P. Review Nanosized hydroxyapatite and other calcium phosphates :

Chemistry of formation and application as drug and gene delivery agents. Tissue

Engineering (2010).

57. Paxton, J. Z., Donnelly, K., Keatch, R. P. & Baar, K. Engineering the Bone-Ligament

interface using Polyethylene Glycol Diacrylate incorporated with Hydroxyapatite. Tissue

Eng Part A. 15, 1201–1209 (2009).

58. Zhou, Z., Ren, Y., Yang, D. & Nie, J. Performance improvement of injectable poly(ethylene

glycol) dimethacrylate-based hydrogels with finely dispersed hydroxyapatite. Biomedical

materials (Bristol, England) 4, 035007 (2009).

59. Zhou, Z., Yang, D., Nie, J., Ren, Y. & Cui, F. Injectable Poly(ethylene glycol)

Dimethacrylate-based Hydrogels with Hydroxyapatite. Journal of Bioactive and

Compatible Polymers 24, 405–423 (2009).

60. Fu, S., Guo, G., Gong, C., Wei, Y. & Qian, Z. Injectable biodegradable thermosensitive

hydrogel composite for orthopedic tissue engineering. 1. Preparation and

characterization of nanohydroxyapatite/poly(ethylene glycol)-poly(epsilon-

caprolactone)-poly(ethylene glycol) hydrogel nanocomposites. The Journal of Physical

Chemistry. B 113, 16518–25 (2009).

122

61. Muller, F., Muller, L., Caillard, D. & Conforto, E. Preferred growth orientation of

biomimetic apatite crystals. Journal of Crystal Growth 304, 464–471 (2007).

62. Seah, R. K. H., Garland, M., Loo, J. S. C. & Widjaja, E. Use of Raman Microscopy and

Multivariate Data Analysis to Observe the Biomimetic Growth of Carbonated

Hydroxyapatite on Bioactive Glass. Analytical Chemistry 81, 1442–1449 (2009).

63. Rietveld, H. M. & Loopstra, B. O. The Structure of Some Alkaline-Earth Metal Uranates.

Acta Crystallographica B25, 787–91 (1969).

64. Supová, M. Problem of hydroxyapatite dispersion in polymer matrices: a review. Journal

of Materials Science. Materials in Mgedicine 20, 1201–13 (2009).

65. Halvorson, R. H., Erickson, R. L. & Davidson, C. L. The effect of filler and silane content on

conversion of resin-based composite. Dental Materials 19, 327–33 (2003).

66. Howard, B., Wilson, N., Newman, S., Pfeifer, C. & Stansbury, J. Relationships between

Conversion, Temperature and Optical Properties during Composite Photopolymerization.

Acta Biomaterialia 6, 2053–2059 (2011).

67. Yang, J., Han, C.-R., Duan, J.-F., Xu, F. & Sun, R.-C. Mechanical and viscoelastic properties

of cellulose nanocrystals reinforced poly(ethylene glycol) nanocomposite hydrogels. ACS

Applied Materials & Interfaces 5, 3199–207 (2013).

68. Moraes, R., Grarcia, J., Barros, D., Liu, J. & Stansbury, J. Control of polymerization

shrinkage and stress in nanogelmodified monomer and composite materials. Dental

Materials 27, 509–519 (2012).

69. Gaharwar, A. K., Dammu, S. a, Canter, J. M., Wu, C.-J. & Schmidt, G. Highly extensible,

tough, and elastomeric nanocomposite hydrogels from poly(ethylene glycol) and

hydroxyapatite nanoparticles. Biomacromolecules 12, 1641–50 (2011).

70. Fu, S., Wang, X., Guo, G., Wei, Y. & Qian, Z. Preparation and Characterization of Nano-

Hydroxyapatite-Poly(ε-caprolactone)-Poly(ethylene glycol)-Poly(ε-caprolactone)

Composite Fibers for Tisue Engineering. Journal of Physical Chemistry C 114, 18372–

18378 (2010).

71. Pielichowska, K. & Blazewicz, S. Bioactive Polymer / Hydroxyapatite ( Nano ) composites

for Bone Tissue Regeneration. Advances in Polymer Science 97–207 (2010).

72. Vural, S., Dikovics, K. B. & Kalyon, D. M. Cross-link density, viscoelasticity and swelling of

hydrogels as affected by dispersion of multi-walled carbon nanotubes. Soft Matter 6,

3870 (2010).

123

73. Xiao, Q. & Zhang, L. C. Effective separation and alignment of long entangled carbon

nanotubes in epoxy. Journal of Materials Science 40, 6513–6516 (2005).

74. Huang, H. H., Ho, C. T., Lee, T. L., Liao, K. . & Chen, F. L. Effect of surface roughness of

ground titanium on initial cell adhesion. Biomolecular Engineering 21, 93–97 (2004).

75. Cha, C., Kim, E.-S., Kim, I. W. & Kong, H. Integrative design of a poly(ethylene glycol)-

poly(propylene glycol)-alginate hydrogel to control three dimensional biomineralization.

Biomaterials 32, 2695–703 (2011).

76. Shkilnyy, A., Gräf, R., Neffe, A. T., Hartmann, J. & Taubert, A. Unprecedented, low

cytotoxicity of spongelike calcium phosphate/poly(ethylene imine) hydrogel composites.

Macromolecular Bioscience 9, 179–86 (2009).

77. Luong, L. N., Hong, S. I., Patel, R. J., Outslay, M. E. & Kohn, D. H. Spatial control of protein

within biomimetically nucleated mineral. Biomaterials 27, 1175–86 (2006).

78. Hofmann, I., Müller, L., Greil, P. & Müller, F. A. Precipitation of Carbonated Calcium

Phosphate Powders from a Highly Supersaturated SBF Solution. Key Engineering

Materials 330-332, 59–62 (2007).

79. Mavropoulos, E., Costa, A. M., Costa, L. T., Granjeiro, J. M. & Rossi, A. M. Adsorption and

bioactivity studies of albumin onto hydroxyapatite surface. Colloids and Surfaces. B,

Biointerfaces 83, 1–9 (2011).

80. Lee, W.-H., Zavgorodniy, A. V, Loo, C.-Y. & Rohanizadeh, R. Synthesis and characterization

of hydroxyapatite with different crystallinity: effects on protein adsorption and release.

Journal of Biomedical Materials Research. Part A 100, 1539–49 (2012).

81. Suh, K. Y., Seong, J., Khademhosseini, A., Laibinis, P. E. & Langer, R. A simple soft

lithographic route to fabrication of poly(ethylene glycol) microstructures for protein and

cell patterning. Biomaterials 25, 557–563 (2004).

82. Pelham, R. J. & Wang, Y. L. Cell locomotion and focal adhesions are regulated by the

mechanical properties of the substrate. The Biological Bulletin 194, 348–50 (1998).

83. Engler, A. J., Sen, S., Sweeney, H. L. & Discher, D. E. Matrix elasticity directs stem cell

lineage specification. Cell 126, 677–89 (2006).

84. Discher, D. E., Janmey, P. & Wang, Y. Tissue Cells Feel and Respond to the Stiffness of

Their Substrate. Science 1139, (2012).

124

85. Britland, S., Perridge, C., Denyer, M., Curtis, A. & Wilkinson, C. Morphogenetic guidance

cues can interact synergistically and hierarchically in steering nerve cell growth.

Experimental Biology Online - EBO 1, 1–15 (1996).

86. Hoffman, A S. Hydrogels for biomedical applications. Annals of the New York Academy of

Sciences 944, 62–73 (2001).

87. Peppas, N. A., Hilt, J. Z., Khademhosseini, a. & Langer, R. Hydrogels in Biology and

Medicine: From Molecular Principles to Bionanotechnology. Advanced Materials 18,

1345–1360 (2006).

88. Khademhosseini, A., Langer, R., Borenstein, J. & Vacanti, J. P. Microscale technologies for

tissue engineering and biology. Proceedings of the National Academy of Sciences of the

United States of America 103, 2480–7 (2006).

89. Lee, K. Y. & Mooney, D. J. Hydrogels for Tissue Engineering. Surgery 101, (2001).

90. Sant, S., Hancock, M. J., Donnelly, J. P., Iyer, D. & Khademhosseini, A. Biomimetic

Gradient Hydrogels for Tissue Engineering. The Canadian Journal of Chemical Engineering

88, 899–911 (2010).

91. West, J. L. & Hubbell, J. A. Reactive polymers Photopolymerized hydrogel materials for

drug delivery applications. Science 25, 139–147 (1995).

92. Iza, M., Stoianovici, G., Viora, L., Grossiord, J. L. & Couarraze, G. Hydrogels of

poly(ethylene glycol): mechanical characterization and release of a model drug. Journal

of controlled release : official journal of the Controlled Release Society 52, 41–51 (1998).

93. Harris, J. Poly(Ethylene Glycol) Chemistry: Biotechnical and Biomedical Applications.

(Plenum Press: New York and London, 1992).

94. Kim, B. & Peppas, N. A. Poly ( ethylene glycol ) -containing Hydrogels for Oral Protein

Delivery Applications. Drug Delivery 333–341 (2003).

95. Schulte, V., Diez, M., Moeller, M. & Lensen, M. Surface Topography Induces Fibroblast

Adhesion on Intrinsically Nonadhesive Poly ( ethylene glycol ) Substrates.

Biomacromolecules 2795–2801 (2009).

96. Peppas, N. a, Keys, K. B., Torres-Lugo, M. & Lowman, a M. Poly(ethylene glycol)-

containing hydrogels in drug delivery. Journal of controlled release : official journal of the

Controlled Release Society 62, 81–7 (1999).

97. Kopecek, J. Hydrogel Biomaterials: A smart Future? Biomaterials 28, 5185–5192 (2008).

125

98. Haraguchi, K. Synthesis and properties of soft nanocomposite materials with novel

organic/inorganic network structures. Polymer Journal 43, 223–241 (2011).

99. Gaharwar, A. K., Rivera, C. P., Wu, C.-J. & Schmidt, G. Transparent, elastomeric and tough

hydrogels from poly(ethylene glycol) and silicate nanoparticles. Acta Biomaterialia 7,

4139–48 (2011).

100. Schexnailder, P. & Schmidt, G. Nanocomposite polymer hydrogels. Colloid and Polymer

Science 287, 1–11 (2008).

101. Wu, C.-J., Gaharwar, A. K., Schexnailder, P. J. & Schmidt, G. Development of Biomedical

Polymer-Silicate Nanocomposites: A Materials Science Perspective. Materials 3, 2986–

3005 (2010).

102. Gupta, H. S. & Guitia, F. An Effective Morphology Control of Hydroxyapatite Crystals.

Crystal Growth & Design (2009).

103. Loo, S. C. J., Siew, Y. E., Ho, S., Boey, F. Y. C. & Ma, J. Synthesis and hydrothermal

treatment of nanostructured hydroxyapatite of controllable sizes. Journal of Materials

Science: Materials in Medicine 19, 1389–1397 (2008).

104. Padilla, S., Izquierdo-Barba, I. & Vallet-Regi, M. High Specific Surface Area in Nanometric

Carbonated Hydroxyapatite. Chem. Mater. 20, 5942–5944 (2008).

105. Wang, L. & Nancollas, G. H. Calcium orthophosphates: crystallization and dissolution.

Chemical Reviews 108, 4628–69 (2008).

106. Lu, Z., Roohani-Esfahani, S., Kwok, P. & Zreiqat, H. Osteoblasts on rod shaped

hydroxyapatite nanoparticles incorporated PCL Film provide an optimal osteogenic niche

for stem cell differentiation. Jun, 1651–61 (2011).

107. Tsurushima, H., Marushima, A., Nakamura, K., Matsumura, A. & Ito, A. Enhanced bone

formation using hydroxyapatite ceramic coated with fibroblast growth factor-2. Acta

Biomaterialia 6, 2751–9 (2010).

108. Kaito, T., Myoui, A., Takaoka, K., Tamai, N. & Yoshikawa, H. Potentiation of the activity of

bone morphogenetic protein-2 in bone regeneration by a PLA-PEG/hydroxyapatite

composite. Biomaterials 26, 73–9 (2005).

109. Cai, L., Guinn, A. S. & Wang, S. Exposed hydroxyapatite particles on the surface of photo-

crosslinked nanocomposites for promoting MC3T3 cell proliferation and differentiation.

Acta biomaterialia 7, 2185–99 (2011).

126

110. Sachlos, E. & Czernuszka, J. T. Making tissue engineering scaffolds work. Review: the

application of solid freeform fabrication technology to the production of tissue

engineering scaffolds. European cells & materials 5, 29–39; discussion 39–40 (2003).

111. Freed, L. E., Novakovic, G. V., Biron, R. J., Barlow, S. K. & Langer, R. Biodegradable

Polymer Scaffolds for Tissue Engineering. Nature 12, 689–693 (1994).

112. Drury, J. L. & Mooney, D. J. Hydrogels for tissue engineering: scaffold design variables

and applications. Biomaterials 24, 4337–4351 (2003).

113. Hollister, S. J. Porous scaffold design for tissue engineering. Nature Materials 4, 518–24

(2005).

114. Cukierman, E., Pankov, R., Stevens, D. R. & Yamada, K. M. Taking cell-matrix adhesions to

the third dimension. Science 294, 1708–12 (2001).

115. Lee, J., Cuddihy, M. J. & Kotov, N. A. Three-Dimensional Cell Culture Matrices: State of

the Art. Tissue Engineering Part B: Reviews 14, 61–86 (2008).

116. Abbott, A. Biology ’ s new dimension. Nature 424, (2003).

117. Cushing, M. C. & Anseth, K. S. Hydrogel Cell Cultures. Science 316, 1133–1134 (2007).

118. Kretlow, J. D., Klouda, L. & Mikos, A. G. Injectable matrices and scaffolds for drug delivery

in tissue engineering. Advanced Drug Delivery Reviews 59, 263–273 (2007).

119. Haraguchi, K. & Takehisa, T. Nanocomposite Hydrogels : A Unique Organic Inorganic

Network Structure with Extraordinary mechanical, optical and swelling/deswelling

properties. Advanced Materials 1120–1124 (2002).

120. Moutos, F. T., Freed, L. E. & Guilak, F. A biomimetic three-dimensional woven composite

scaffold for functional tissue engineering of cartilage. Nature Materials 6, 162–7 (2007).

121. Yu, L. & Ding, J. Injectable hydrogels as unique biomedical materials. Chemical Society

Reviews 37, 1473–81 (2008).

122. Podsiadlo, P., Kaushik, A. K., Lahann, J., Ramamoorthy, A. & Kotov, N. a Ultrastrong and

stiff layered polymer nanocomposites. Science 318, 80–3 (2007).

123. Wang, Q., Mynar, J. L., Yoshida, M., Kinbara, K. & Aida, T. High-water-content mouldable

hydrogels by mixing clay and a dendritic molecular binder. Nature 463, 339–43 (2010).

127

124. Capadona, J. R., Shanmuganathan, K., Tyler, D. J., Rowan, S. J. & Weder, C. Stimuli-

responsive polymer nanocomposites inspired by the sea cucumber dermis. Science 319,

1370–4 (2008).

125. Klenke, F. M., Liu, Y., Hunziker, E. B., Siebenrock, K. a & Hofstetter, W. Impact of pore size

on the vascularization and osseointegration of ceramic bone substitutes in vivo. Journal

of Biomedical Materials Research. Part A 85, 777–86 (2008).

126. Hing, K., Best, S. M., Tanner, K. E., Bonfield, W. & Revell, P. a Mediation of bone ingrowth

in porous hydroxyapatite bone graft substitutes. Journal of Bomedical Materials

Research. Part A 68, 187–200 (2004).

128

Abstract

This thesis comprises two fundamentally different approaches, namely the soft lithographic

surface patterning Fillmolding In Capillaries (FIMIC) technique and the fabrication of 3D crafts in

order to pursue the same goal; controlled cellular adhesion behavior on the basis of localized

chemical functionalization. It is commonly accepted that the understanding of protein

adsorption and cell adhesion behavior on engineered surfaces and interfaces is essential for the

successful development of novel biomaterials. In the course of this thesis, chemical

functionalization with calcium phosphates, in particular hydroxyapatite (HAp), were conducted.

The intention to compare cellular response of different cell types, namely osteoblasts (mouse

osteoblast like MC3T3-E1) and fibroblasts (mouse fibroblasts L-929) in order to reveal guided

osteoblast adhesion on specifically chemically patterned hydrogel substrates. Therefore,

Poly(ethylene glycol) (PEG) based hydrogels and HAp were utilized in order to form various

polymer-inorganic hybrid composites. The overall objective was to exploit the intrinsic protein

repellent behavior of PEG-based hydrogels and the protein attractive properties of HAp-based

ceramics in order to produce surface patterns of adjacent bioactive and bioinert properties.

Thus, it could be demonstrated that spatially controlled introduction of HAp on micropatterned

surfaces can determine the adhesion pattern of HAp affine osteoblasts. This stands in opposite

to fibroblast, where the aforementioned effect could not be observed. In consequence, it

displays the potential of this approach to promote specific cell-substrate interaction of desired

cell phenotypes and to distinguish between several cell lines exhibiting varying affinity to a

particular chemical pattern.

In the first part of this thesis, the soft lithographic FIMIC method was utilized in the first place to

demonstrate that targeted surface patterning with HAp may control adhesion pattern and site

of osteoblast adhesion, but in contrast does not show the same effect for fibroblast. Therefore,

physically mixed (PEG HAp NPs) and chemically synthesized (PEG nHAp) PEG HAp hydrogel

composites fabricated via a precipitation reaction were produced and compared. It was

revealed via protein adsorption experiments with the model protein Bovine Serum Albumin

129

(BSA), that PEG nHAp, in consequence of their inherent higher homogenicity, reaches a far

greater bioactivity than PEG HAp NPs. This verifies the principle, in which successful chemical

patterning for a specific cell type is desired, yielding control over cellular response due to cell

substrate interaction. In this project, Atomic Force Microscopy (AFM) was applied for the

physical and mechanical characterization of as prepared FIMIC platforms in dry and hydrated

state; a necessary requirement for in-situ investigation in biological application as well as to

scrutinize whether topography represents the dominant factor for adhesion. Post fabrication

patternability with spatial and chronological resolution of HAp nucleation from Simulated Body

Fluid (SBF) over a time-span of 10 days was carried out as concluding step in the comparison

between PEG HAp NPs and PEG nHAp. It could be revealed that PEG nHAp allows precise control

regarding site and density of HAp nuclei deposition in opposite to PEG HAp NPs. The herein

presented comparison of physical and chemical hydrogel composites may serve as a model

system to demonstrate the necessity and feasibility of chemical inorganic-organic composites

towards their physical counterparts.

In the second part, porous scaffolds made of chemically synthesized star-shaped PEG HAp

(8PEG HAp) hydrogel composites via precipitation reaction and subsequent freeze-drying were

produced, which combine readiness of functionalization of biological or chemical cues with

physical and mechanical suitability for tissue replacement. The general idea was to fabricate 3D

interconnected structures, which possess post fabrication functionalization ability via residual

carbon double bonds and chemically active HAp. This was proven via the successful cleavage of

a model molecule in form of a SH dye and the homogeneous adsorption of the HAp reactive

model protein Bovine Serum Albumin (BSA). In conclusion, fibroblast and osteoblast displayed

distinct adhesion patterns regarding number of adhered cells and cell morphology. In contrast

to fibroblasts, osteoblasts covered almost the entire craft and showed strong spreading, which

is generally related to strong cell-substrate interaction. Moreover could be monitored that

scaffolds exhibited a pore size structure suitable for vascularization and medium transport,

which is indispensable for function and the successful quick ingrowth to the host, an essential

need, considering potential replacements in tissue engineering.

130

Zusammenfassung

Die vorgelegte Arbeit beinhaltet zwei fundamental unterschiedliche Herangehensweisen, Soft

Lithographie in Form von Fillmolding In Capillaries (FIMIC) und die Fabrikation von 3D

Werkzeugen um kontrolliertes Zelladhäsionsverhalten durch gezielte chemische Strukturierung

zu erreichen. Es ist weithin akzeptiert, dass das Verständnis für Proteinadsorption und

Zelladhäsionsverhalten auf Oberflächen und Grenzflächen für die Entwicklung zeitgemäßer

Biomaterialien von großer Bedeutung ist. Deshalb wurden in dieser Thesis Kalziumphosphate,

insbesondere Hydroxylapatit (HAp) mit dem Ziel verwendet unterschiedlich HAp-responsive

Zelltypen, Osteoblasten (Maus-Osteoblasten ähnlich MC3T3-E1) und Fibroblasten (Maus-

Fibroblasten L-929), auf chemisch strukturierten Hydrogel Oberflächen anzusiedeln und

vergleichend zu untersuchen. Daher wurden Poly(ethylen glykol) basierte Hydrogele und HAp

eingesetzt um verschiedene inorganisch-organische Hydrogel Komposite zu formen. Dabei war

es Ziel, das intrinsische proteinabweisende Verhalten von PEG basierten Hydrogelen und die

proteinanziehenden Eigenschaften von HAp basierten Keramiken auszunutzen um

mikrostrukturierte Oberflächen zu designen, in denen bioinerte und bioaktive Areale direkt

einander angrenzen. Auf diese Weise konnte gezeigt werden, dass die örtlich kontrollierte

Einbringung von HAp das Adhäsionsverhalten von HAp affinen Osteoblasten bestimmt. Das

steht im Gegensatz zu Fibroblasten, wo dieser Effekt nicht beobachtet werden konnte und

demonstriert das Potential dieses Ansatzes, gewünschte spezifische Zell-Substrat

Wechselwirkung zu fördern und zwischen unterschiedliche Zelltypen mit variiernder Affinität zu

chemischen Substanzen zu selektieren.

Im ersten Teil dieser Arbeit wurde die soft lithographische FIMIC Methode in erster Linie dafür

eingesetzt um das lokal kontrollierte Adhäsionsverhalten von Osteoblasten zu zeigen, ein Effekt

der für Fibroblasten nicht nachgewiesen werden konnte. Dafür wurden aus linearem PEG und

entsprechenden Salzlösungen mittels Abscheidungsreaktion physikalisch gemischte (PEG HAp

NPs) und chemisch synthetisierte (PEG nHAp) PEG HAp Hydrogel Komposite produziert und

verglichen. Es stellte sich durch Experimente zur Proteinadsorption mit dem Modell-Protein

131

Bovine Serum Albumin (BSA) heraus, dass aufgrund der inherenten überlegenen Homogenität

von PEG nHAp eine weitaus höhere Bioaktivität erreicht wird als bei PEG HAp NPs. Dabei wird

das Prinzip bestätigt, in welchem durch erfolgreiche chemische Oberflächenstrukturierung für

gezielte Zellarten Kontrolle über die Zell-Substrat Interaktion gewonnen werden kann. In

diesem Rahmen wurde Rasterkraftmikoskopie (AFM) zur physischen und mechanischen

Untersuchung der erzeugten FIMIC Plattformen in trockenem und hydratisierten Zustand

eingesetzt, da dies eine absolut notwendige in-situ Charakterisierungmethode für biologische

Anwendung darstellt. Darüber hinaus untersucht diese Methode, dass Adhäsionsmuster der

Zellen nicht durch topographische Eigenschaften des Substrates dominiert werden. Die

weitergehende Strukturierung von FIMIC Plattformen mit örtlich und zeitlich kontrollierter HAp

Disposition von Simulated Body Fluid (SBF) über eine Zeitspanne von 10 Tagen war der

abschließende Schritt in dem Vergleich zwischen PEG nHAp und PEG HAp NPs. Es zeigte sich,

dass PEG nHAp im Gegensatz zu PEG HAp NPs sich ausgezeichnet eignet um HAp lokal gezielt

und zeitlich kontrolliert nukleieren zu lassen. Der hierin präsentierte Vergleich physikalischer

und chemischer Hydrogel-basierter Verbundwerkstoffe könnte als Modellsystem dienen um die

Notwendigkeit zu demonstrieren, chemische inorganische-organische Komposite

entsprechenden Physikalischen vorzuziehen.

Im zweiten Teil dieser Arbeit wurden poröse Gerüste (scaffolds) aus chemisch synthetisierte

PEG HAp (8PEG HAp) Hydrogel Komposite aus sternartigem PEG sowie entsprechender

Salzlösungen mittels Abscheidungsreaktion und darauffolgender Gefriertrocknung hergestellt.

Diese besitzen die Fähigkeit der Funktionalisierung mit biologischen und chemischen

Reaktanten nach der Gerüstformung sowie die nötigen physikalischen und mechanischen

Eigenschaften für den Einsatz als Gewebeersatzwerkstoff. Die generelle Idee war es 3D

Strukturen zu fertigen, die auf der einen Seite nach der Formgebung noch weiter chemisch

funktionalisierbar sind und des weiteren über aktive HAp Domänen an der Oberfläche verfügen.

Dies wurde durch die demonstrierte Anbindung eines Modellmoleküles in Form eines

Modellmoleküls eines SH dye sowie die erfolgreiche homogene Adsorption des HAp reaktiven

Modelproteins Bovine Serum Albumin (BSA) erreicht. Im Anschluss zeigten Fibroblasten und

Osteoblasten unterschiedliche Adhäsionmuster in Hinsicht auf Anzahl der adhärierten Zellen

132

und Zellmorphologie. Im Kontrast zu Fibroblasten bedeckten Osteoblasten beinahe das

gesamte Substrat bei starker Zellspreitung, was gemeinhin als Zeichen starker Zell-Oberflächen

Wechselwirkung angesehen wird. Zudem verfügen die produzierten 3D Strukturen über

Porgrößen von über 100 µm, welche für Vaskularisation und Mediumtransport notwendig sind

um somit Funktion und schnelles Einwachsen bei potentieller Anwendung als Gewebeersatz zu

garantieren.

133

Acknowledgements

I would like to thank Prof. Marga C. Lensen for giving me the opportunity to carry out my

doctoral thesis in her research group. I want express my gratitude to her guidance, supervision,

time and patience throughout the past three years.

Special gratitude goes in particular to Zhenfang Zhang, with whom I conducted basically every

single project in cooperation. This cooperation is the main source of to the herein presented

results.

Furthermore, I owe gratitude to Christine Strehmel and Gonzalo de Vicente Lucas for working

with me directly in the lab and conducting experiments, which I would have not been able to

conduct myself. Moreover do I want thank them for their advice and many inspiring

conversations.

In addition, I wish to thank the rest of the group for the nice working atmosphere throughout

the time that I spend at the TU Berlin

Beyond that, I want to express my gratitude to the persons, who helped in several instrumental

analyses conducted during the course of this thesis.

I am indebted to my family for their support, patience and empathy in the past three years.

134

List of Publications

Publications employed in this thesis

Kelleher S, Jongerius A, Loebus A, Strehmel C, Zhang Z, Lensen MC.; AFM Characterization of

Elastically Micropatterned Surfaces Fabricated by Fill-Molding In Capillaries (FIMIC) and

Investigation of the Topographical Influence on Cell Adhesion to the Patterns.; Advanced

Engineering Materials, 14, B56–B66 (2012).

Publication-Manuscripts employed in this thesis

Kelleher S, Zhang Z, Loebus A, Strehmel C, Lensen MC.; Blending Poly(ethylene glycol) (PEG)-

based Polymers in order to Obtain a Library of new Biomaterials and their Application in

Surface-micro-patterning by the Fillmolding In Capillaries (FIMIC) Method. Submitted to Soft

Matter

Loebus A+, Zhang Z+, Strehmel S, de Vicente Lucas G, Lensen M C. Soft lithographic surface

patterning of physically and chemically mineralized Poly(ethylene glycol) hydrogels for selective

interface interaction. To be submitted.

Loebus A+, Zhang Z+, Li Q, Strehmel C, Wisniewski W, Arafeh M, Rüssel C, Su Z, Lensen MC. 3D

Patterned Reactive Mineralized Poly(ethylene glycol) Derived Hydrogels., manuscript in

preparation

135

Additional publications and manuscripts

Zhang Z, Loebus A, de Vicente Lucas G, Manar Arafeh M, Lensen MC. In situ formation of novel

Poly(ethyleneglycol)-based Hydrogels via Amine-Michael Type Addition with Tunable

Mechanics and Chemical Functionality. Submitted to Angew. Chem.

Zhang Z+, Loebus A+, Li Q, Strehmel C, Wang J, Su Z, Lensen MC. Calciumphosphate

Incorporation into Chemically Crosslinked Poly(ethylene glycol) based Composite Hydrogels for

Bone Tisse Engineering. Manuscript in preparation

Zhang Z, Kelleher S, Steinhilber D, Loebus A, de Vicente G, Strehmel C, Haag R, Lensen MC. Click

Chemistry as a Crosslinking Method in the Fabrication of Hydrogels with Tuneable Degradation

Properties from star-shaped and Hyperbranched Polyether Macromonomers. Manuscript in

preparation

Li Q+, Zhang Z+, Zhang X, Loebus A, Wang J, Ouyang Z, Su Z, Lensen MC. Chemically cross-linked

PEG-based hydrogel with Crystalline Domains in long-range Ordering. Manuscript in preparation

Strehmel C, Perez-Hernandez H, Zhang Z, Loebus A, Werner C, Lasagni AF, Lensen MC.

Controlling Cell Alignment and Spreading by Poly(ethylene glycol) and Ormocomp®

Microstructures on Polymer Surfaces. Manuscript in preparation.

Li Q+, Zhang Z+, Zhang X, Loebus A, Wang J, Ouyang Z, Su Z, Lensen MC. Electrospinning Fibers

through 8armPEG Hydrogel Formation Process. Manuscript in preparation

Chen J, Arafeh M, Guillet A, Felkel D, Loebus A, Kelleher S, Fischer A, Lensen MC. Hybrid

Hierarchical Patterns of Gold Nanoparticles and Poly(Ethylene Glycol) Microstructures.

Submitted to Journal of Materials Chemistry C

136

Contribution to Scientific Conferences

Poster contributions

Micro- and Nanofabrication Methods to create patterned PEG Hydrogels for Nano-

Biotechnological Applications. Axel Loebus, Susan Kelleher, Zhenfang Zhang, Marga C. Lensen.

EuroBiomat 2011, First Posterprice

Micro- and Nano- surface patterning of PEG-Nanocomposites Hydrogels for Biomedical

Applications. Axel Loebus, Zhenfang Zhang, Christine Strehmel, Marga C. Lensen. Nanomaterials

for Biomedical Technologies 2012, Frankfurt/Main

Micro- and Nano- surface patterning of PEG-Nanocomposites Hydrogels for Biomedical

Applications. Axel Loebus, Zhenfang Zhang, Christine Strehmel, Marga C. Lensen. Jahrestagung

der Deutschen Gesellschaft für Biomaterialien 2012, Hamburg

137

Appendix

In this “Appendix” the reader will find all auxiliary information. All data shown here are referred

to in the main text of the thesis and is ordered according to the respective chapter. In general,

the reader will find results of experimental processing conducted by a colleague or where the

intellectual effort is attributed to someone else. In some cases however, figures serve to

support a statement made in the main text body, but are not of essential importance for the

thesis itself.

Chapter 4.1

Acrylation procedure

PPEG-b-PPG-b-PEG (4400 Da, Sigma Aldrich) and K2CO3 were dried in a vacuum oven at 100 °C

for 4 hr. PEG-b-PPG-b-PEG (5 g) and K2CO3 (20 g) were dissolved in dry CH2Cl2 (50 mL) under N2.

The solution was then cooled to 0 °C and acryloyl chloride (15 g) was added dropwise. The

mixture was stirred at 50 °C for 2 days. The solution was filtered and poured into petroleum

ether cooled to -196 °C. The solution was stirred for 10 min and the ether decanted, leaving

behind a crude product. This crude product was dissolved in 50 mL CH2Cl2 and then washed 3

times with saturated NaCl solution 3 times. The organic layer was collected, dried with MgSO4,

filtered and the solvent was removed under reduced pressure resulting in a colourless liquid. A

typical yield from this procedure was 60 %. 1H NMR (OCH2CHCH3O 1.12 ppm, OCH2CHCH3O 3.38

ppm, OCH2CHCH3O 3.52 ppm, OCH2CH2O 3.63 ppm, (C=O)OCH2 4.30 ppm, =C-H trans 5.83 ppm,

CH=C 6.15 ppm, =C-H cis 6.42 ppm).

138

Figure 1: Atomic Force Microscopy (AFM) topographic images of PEG575 in PEG575 containing 1 % PI and 5

% CL in the mold and filler likewise. (a) dry state; (b) hydrated state.

Chapter 4.2

Acrylation-procedure for 8arm PEG:

At first, 8arms PEG-OH (15 kDa) and K2CO3 was dried in a vacuum oven at 100 °C for 4 h. Then,

8arms PEG (5g) and K2CO3 (30 times) were dissolved in 50 mL CH2Cl2 under N2 -condition.

Acryloyl Chloride (20 times) was added dropwise in a water-ice bath. The mixture was stirred at

60 °C for 4 days. The solution was filtered, and then poured into the cold petroleum ether

(cooled by water-ice). The solution was stirred for 10 min, and then separated to get the crude

product. The crude product was dissolved in 50 mL DCM and then extracted with saturated

NaCl solution for 3 times. The organic layer was collected. The solution was dried by

magnesium-sulfate overnight, then filtered to remove MgSO4 and subsequently the solvent was

removed under reduced pressure to get the final product. (white solid) Isolated yield

(72 %)1HNMR(OCH2CH2O 3.64 ppm(1496H), (C=O)OCH2 4.31 ppm(16H), =C-H trans 5.83

ppm(8H), CH=C 6.15 ppm(8H), =C-H cis 6.42 ppm(8H)).

Processing of pure polymers:

In order to process the polymers into mouldable gels, one must ensure they are 1) in a liquid

state and 2) homogenously mixed with the photoinitiator (PI) (1% PI). In the case of the two

liquid polymers (PEG1 and 3BC1), the PI is evenly dispersed throughout the mix using a small

139

amount of acetone, which is removed after mixing. Solid polymers, on the other hand, require

additional treatments for them to become processible.

Dissolving the solid polymers (PEG2, 3BC2 and 8PEG) and the PI in water allows for the

formation of very soft but strong hydrogels, depending of the amount of water used in the

process. However, the FIMIC process is not compatible with water-based gels at present, due to

either rapid dehydration of highly hydrated molds during the filling step or indeed hydration of

non-hydrated molds being filled with water-based polymer solution, both phenomena that are

followed by curling off of the molds from surface, so one has to avoid using water in our

system. For this reason solids were heated above their Glass Transition Temperature (Tg) to

yield melted polymers and to allow to handle them in liquid form and therefore enable the

molding of these materials without any solvent.

The window of processing time for this liquid processing is less than five minutes, after which

time precursors begin to solidify. Polymers that were heated to the correct temperature (above

their Tg) and were rapidly cast into molds and resulting transparent, homogenous samples.

Analysis using optical microscopy and Atomic Force Microscopy (AFM) confirmed the

homogeneity of the samples (results not shown). Highly crystalline polymers, e.g. PEG2, were

the most difficult to make into a transparent, homogeneous gel sometimes showing the

formation of spherulites upon molding (and concurrent crystallisation). This can be avoided by

working quickly with the melt and warming the molds and glassware used in the casting.

The hydrogels are easy to handle, flexible and transparent and thus ideal for using in the

eventual cell studies. After fabricating samples from five pure polymers, the swelling degree

(SD) was calculated of each sample after 24 hours in deionised water at 37 °C (SI 1).

140

Table 1 – The characteristics of the pure polymers and how well they formed transparent, homogenous

gels (+++ = very well, + = not very well).

Of the pure PEG-based gels, PEG2 swells the most after 24 hours. This can be explained by the

longer chain length compared to the other two derivatives (PEG1 and 8PEG) and therefore the

cross-linking points that are further apart. The swelling ability of the block co-polymer 3BC1 is

significantly poorer than the pure PEG counterparts, due to the chemistry of the block

copolymer, with a large section of the polymer consisting of the hydrophobic PPG. In addition,

the relatively short chains mean that 3BC1 hardly swells at all in water. Interestingly, on the

other hand, the block co-polymer 3BC2 has the highest swelling degree of all the gels, showing

that despite the presence of the PPG moieties, the longer chain length produces, upon

crosslinking, a “loose” network that is capable of taking up more water.

The swelling data is vital for the understanding the swelling degree of the hydrogels used in

order to manipulate the topography produced in the FIMIC samples and subsequently levelling

out of respective samples. From Table 1 it becomes clear that based on their large swelling

ability, 3BC2 and 8PEG are the most promising candidates to use in the strategy of employing a

filler material that swells more than the mold. Nevertheless, these two derivatives are also the

most challenging to make processible.

141

Moreover, the chemistry of these two polymers is quite different; 3BC2 contains 20% PPG

whereas 8PEG contains only PEG. In order to rule-out any chemistry differences on FIMIC

samples, two similar polymers for blending for the use in the FIMIC process are paired. Blending

building blocks give more versatility in tuning the properties of the resulting gels; the

physicochemical properties such as swelling degree lie in between those of the gels formed

from the pure constituent. By adding different amounts of crosslinker (CL), one can keep the

chemistry of two gels the same, while adjusting the stiffness. Samples of homogenous gels with

different levels of cross-linker (0 %, 5 % and 10 %) and therefore, tuneable physicochemical

properties, were fabricated.

Blending:

Blending the two liquid pre-polymers (i.e. PEG1 and 3BC1) was thought to be the least

complicated combination to physically mix. Yet, attempting to blend the two polymers in

different ratios showed that the homogenous blending reached a limit (SI 2a). Gels with a

smaller ratio of PEG1:3BC1 were better able to form a homogenous gel, with mixes containing a

higher percentage of 3BC1 were shown to form a phase-separated material. This phase-

separated material was recognised by the naked eye by its opaqueness, and optical microscopy

confirmed the presence of micrometer-sized droplets (around 50-100 µm in size; SI 2b).

Nevertheless, when the % PEG1 was smaller than approximately 40 % (3BC1 was

approximately 60 % or more) the formation of a homogenous, transparent hydrogel could be

observed (SI 2c).

SI 1: (a) Successful formation of good hydrogels depends on the ratio of PEG1:3BC1; (b) phase

142

separation gel formed from a mixture of PEG1:3BC1 (50:50); (c) a mixing ratio of PEG1: 3BC1 (33:66)

gives a transparent gel (the limit for the transparency lies at approximately 40% PEG1).

The swelling degrees of the successful, homogenous blends (i.e. PEG1:3BC1 in mixing ratios

below 40:60) as well as the swelling degree after 24 hours of the pure polymer gels were

measured (SI 3). The more PEG1 and/or less crosslinker (CL) present in the gel, the more it

swells. This graph shows that by altering these two factors one can fabricate gels within a wide

range of swelling degrees. The liquid PEG-based hydrogels have a maximum swelling ratio of

about 30%.

SI 2: The swelling degrees of pure PEG1, 3BC1 and the homogenous blend gels (described in terms of

their PEG1 content). Higher amounts of crosslinker (CL) reduce the swelling ability of the hydrogels.

Although it is able to process the individual solid PEG polymers (3BC2, PEG2 or 8PEG) to get

transparent, homogenous gels, blending of two solid PEG polymers by heating and mixing was

unsuccessful. Therefore, to add fluidity to the mixtures of solid polymers, solid PEG-derivative

with a liquid one, namely with PEG1 or 3BC1 were mixed.

SI 4 summarizes the combinations of liquid polymer with solid polymer that resulted in the

formation of homogenous, transparent gels. The best (i.e. the easiest to form, most

143

homogenous/transparent and stable within the time frame of processing) liquid/solid blends

that were fabricated involved the PEG1 liquid polymer. Mixing melts of each 3BC2 and 8PEG

with liquid PEG1 easily gave gels that were homogenous and transparent upon curing (entries 1

and 2). On the other hand, the blending of PEG1 with the solid, longer chain PEG2 was

unsuccessful in the formation of a homogenous gel as the PEG2 recrystallizes upon mixing with

the PEG1 leading to opaque gels full of crystals (entry 3). Finally, the blending of liquid 3BC1

with the two solid, pure PEG-polymers (PEG2 and 8PEG) did not result in any homogenous gel

formation as they were immiscible (entries 4 and 5).

Thus, the successful blending of the combinations of both the liquid/liquid blend PEG1/3BC2

and the liquid/solid blends of pure PEG, i.e. PEG2/8PEG or PEG1/8PEG, yielded a range of gels

with different swelling ratios. In particular, the swelling degree and crosslinking variation of the

PEG1/8PEG gels show promise in being an excellent candidate for use in the levelling out

experiment, since 8PEG itself swells more than PEG1, and the blends therefore combine good

processibility with a larger swelling range than the pure constituents. SI 4b shows the swelling

degree (SD) of blends of three different ratios of PEG1/8PEG; obviously, the more 8PEG in the

blend, the more the gel swells.

SI 3: (a) Table - The outcome of blending liquid/solid pre-polymers aiming to fabricate a homogenous

blend; (b) Chart - the swelling degrees of the PEG1/8PEG blends as a function of their composition and

amount of crosslinker (CL).

144

SI 4: Topography of FIMICs with blends of PEG and 3BC with varying amounts of crosslinker (CL) in mold

and filler.

Chapter4.3.3

SI 1: Enlarged depiction of Figure 7f. CLSM optical and fluorescent images (a) optical image, (b)

fluorescent image.

145

Chapter4.4.1

SI 1: CLSM fluorescent images of Bovine Serum Albumin (BSA) adsorption on a representative non-UV

cured flat 8PEG 20HAp sample.(a) topview; (b) cross-section.

146

Chapter 4.4.2

SI 2: Water-uptake of 8PEG HAp hydrogel composites before UV curing.

147

SI 3: Water-uptake of 8PEG HAp hydrogel composites after UV curing.

Si 4: Scheme of ultraviolet radiation curing of 8PEG hydrogels

148

Chapter 4.4.3

SI 5: Adhesion experiments and cytotoxicity testing with osteoblasts (mouse osteoblast like MC3T3-E1)

characterized with optical microscopy. (a) adhesion of osteoblasts on flat UV cured 8PEG 20HAp, (b) live-

dead assay of osteoblasts on 8PEG 20HAp.