A Flexible Approach to the Preparation of Polymer Scaffolds for Tissue Engineering

Communication

A Flexible Approach to the Preparation ofPolymer Scaffolds for TissueEngineering

Poonam Borgaonkar, Sachin Sharma, Ming Chen, Sankha Bhowmick,Daniel F. Schmidt*

Porous polyester thermoset xerogels have been produced via sol–gel chemistry as a first step inthe development of sol–gel derived tissue engineering scaffolds templated by replica moldingand/or salt leaching. The pore structure of these untemplated thermosets is tunable and can bealtered independent of or in tandem withalterations in composition. Cytocompatibilitystudies on these xerogels imply the effects ofboth pore size and materials chemistry, withfully aliphatic polyesters with large pore struc-tures allowing the growth of mammalian cells.To the best of our knowledge, this representsthe first report examining the preparation andpotential of sol–gel derived porous polymerxerogels as tissue engineering scaffolds.

Introduction

Tissue engineering combines cell biology, materials

chemistry, and processing to recreate viable tissue that

restore and maintain the function of the body. Three

dimensional biodegradable porous bodies referred to as

‘‘scaffolds’’ are often created for this purpose. Typically, an

P. Borgaonkar, S. Sharma, D. F. SchmidtDepartment of Plastics Engineering, University of MassachusettsLowell, 1 University Avenue, Lowell, MA 01854 USAFax: þ1 (978) 458 4141;E-mail: [email protected]. Chen, S. BhowmickDepartment of Mechanical Engineering/Bioengineering andBiotechnology Program, University of Massachusetts Dartmouth,285 Old Westport Road, North Dartmouth, MA 02747 USA

Macromol. Biosci. 2007, 7, 201–207

� 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

appropriate material is chosen and attempts are made to

produce a porous, cell-friendly scaffold optimized for a

specific application. Much has been done, but the state of

the art is far from perfect, and there is much more to do in

order to realize the potential of tissue engineering.[1,2]

Specifically, there is a critical need for appropriate scaffolds

to serve as the base for engineered or regenerated

tissues.[3] At the least, such scaffolds must possess a

significant amount of porosity (>30%), a large pore size (on

the order of tens or hundreds of microns), and a continuous

pore structure in order for cell infiltration. The ideal

scaffold is also mechanically robust, biodegradable,

non-toxic, immunocompatible, and with cell-directing

surface chemistry,[4] and topology/topography/surface

mechanics.[5] It will also act as drug-delivery system for

agents whose activity must be preserved during materials

preparation.[6]

DOI: 10.1002/mabi.200600212 201

P. Borgaonkar, S. Sharma, M. Chen, S. Bhowmick, D. F. Schmidt

202

One criterion not mentioned above is the need for a

flexible preparation technique. The body consists of hund-

reds of different types of tissues; no one material will fit the

bill in every tissue engineering application. It is, therefore,

critical that we have a means of rapidly developing new

materials as needs indicate. Likewise, the ability to prepare a

wide range of porous materials whose composition and pore

structures may be independently varied would be an

invaluable tool for scientists attempting to understand

the effects of such variations. It is with these aims that we

have approached the subject, and have selected organic

sol–gel chemistry as a flexible means of scaffold synthesis.

While sol–gel chemistry already sees application in the

field of tissue engineering, most of the efforts have focused

on the development of hydrogel-based systems[7,8] or the

use of inorganic sol–gel chemistry to prepare porous

bioceramics.[9,10] We report on the development of organic

sol–gel chemistry as a means of preparing tissue

engineering scaffolds with structures and chemistries

appropriate for colonization by mammalian cells. While a

single publication has been found describing the prepara-

tion of polyurethane hydrogels via organic sol–gel chemis-

try,[11] we believe this to be the first report describing the

Figure 1. Structure of hyperbranched polyester polyol Boltorn H40. Thyperbranched/dendritic polyester polyol based on a pentaerthyritol (PTheoretically, Mn ¼ 7316 g�mol�1, with 64 primary OH groups pe2800 g�mol�1 (Mw=Mn ¼ 1.8) with �24–25 OH groups per molecule.



preparation of polymer xerogels for tissue engineering

applications.

The developed systems exhibit rapid and efficient

sol–gel transitions to form open, fully interconnected

pore structures that survive ambient drying to form robust

xerogels. Step polymerization and network formation can

be made to favor gelation over precipitation, while resi-

dual reactive groups may be used to add functionality to

the material. We address the issue of materials selection by

taking a library approach, allowing materials with arbi-

trary properties (hard vs. soft, rapidly vs. slowly degrading,

etc.) to be rapidly prepared by selecting appropriate

components and reaction systems. Variations in reaction

media and conditions allow tuning of the phase separation

conditions and pore structure and accommodation of

active agents (metal precursors, dyes, biomolecules, living

cells, etc.). Additional templating agents (colloids, surfac-

tant phases) may be introduced, while porous inorganic

materials may be formed via pyrolysis or calcination. The

underlying principles at work are simple and versatile,

giving a single synthetic methodology that can address the

needs of a broad range of applications for specific, highly

functional porous materials.

he idealized structure of Boltorn H40, a pseudo-fourth generationE) core reacted with 2,2-bis(hydroxymethyl)propanoic acid (bis-MPA).

r molecule. From supplier data, typical characteristics are Mn ¼

DOI: 10.1002/mabi.200600212

A Flexible Approach to the Preparation of . . .

Boltorn H40

OH

OH

OHOH

HO

HO

HO

HO OH

Boltorn H40

OH

OH

OHOH

HO

HO

HO

HO OH

++R

O O

Cl Cl

Boltorn H40

OH

O R

O O

O

OHOH

HO

HO

HO

HO OH

Boltorn H40

OH

OH

OHOH

HO

HO

HO OH

-HCl

-HCl

Figure 2. Sol–gel reaction scheme. Schematic of the reaction that forms the basis for the organic sol–gel process described in thiscommunication (triethylamine catalyst not shown). In the case of Samples 1 and 3, terephthaloyl chloride was used and R¼C6H4 (p-phenyl).In the case of Sample 2, adipoyl chloride was used and R¼C4H8 (n-butyl).

In addition to the well-known and readily applicable

approach of salt-leaching as a means of controlling

macroscale porosity,[12] we are developing the means to

directly replicate natural porous bodies using the sol–gel

techniques reported in this communication. Past work has

shown that natural cytocompatible porous bodies (bone,

coral, sponge, etc.) can be infiltrated with wax, then

dissolved to leave the wax ‘‘negative’’, after which a

polymerization reaction may be carried out within the

porous wax to form a porous polymer replica.[13,14] Future

reports will cover these efforts in detail.

In this communication, we focus on the development of

sol–gel systems for use in the aforementioned replica

molding scheme. A variety of systems have been studied

based primarily on the formation of polyester and

polyurethane networks in solution. Of these, three have

been selected on the basis of rapid reaction, robust

materials properties and reasonable homogeneity, and

their structure and cytocompatibility reported.

Experimental Part

Materials

Commercial hyperbranched aliphatic polyester Boltorn H40

(Perstorp Corporation; Mw ¼ 5 100 g �mol�1, Mw=Mn ¼1.8, hydro-

xyl number¼ 470–500 mg KOH � g�1; see Figure 1), terephthaloyl

chloride (99þ%, Acros Organics), adipoyl chloride (98%, Aldrich),



triethylamine (99.5%, Sigma–Aldrich), and tetrahydrofuran (ACS

reagent, �99.0%, contains 250 ppm BHT as inhibitor, Sigma–

Aldrich) were used as received.

Scaffold Preparation

Sample 1: 20% w/v (total solids weight per total solution volume,

20 g per 100 ml in this case) solutions of terephthaloyl chloride in

THF and Boltorn H40 in THF were combined in the presence of

triethylamine (TEA) in a poly(propylene) vial, which was then

sealed. The overall molar ratio in the final mixture was 1:1:0.25

OH/COCl/TEA, with the catalyst concentration chosen to induce

the rapid formation of a homogeneous, colorless, opaque gel. The

gel was allowed to stand for 24 h at room temperature before

being dried in a vacuum oven at 60 8C for 48 h.

Sample 2: identical to Sample 1, except that adipoyl chloride

was used instead of terephthaloyl chloride. The overall molar ratio

in the final mixture was 1:1:0.5.

Sample 3: identical to Sample 1, except that the total solid

content of the precursor solution was 15% w/v.

Materials analyses were performed without further treatment,

with the exception of the cell culture studies. In this case, the dried

xerogels were immersed in boiling water for 1 h to ensure the

deactivation and/or removal of any remaining monomer or catalyst

and to reduce the chances of contamination by microorganisms.

Scanning Electron Microscopy (SEM)

For the purpose of characterizing the pore structures of the various

xerogels, a JEOL JSM-7401F field emission scanning electron

www.mbs-journal.de 203

https://www.researchgate.net/publication/29861964_Salt_leaching_for_polymer_scaffolds_Laboratory-scale_manufacture_of_cell_carriers?el=1_x_8&enrichId=rgreq-679bb8c1-4571-4458-8615-799f6e7bba98&enrichSource=Y292ZXJQYWdlOzY1MTA3MTE7QVM6Mjc2ODg1Mjg5MzU3MzEyQDE0NDMwMjYwMDUxMTA=


Figure 3. SEM images of polyester xerogels. Representative SEM images of Sample 1 (a), Sample 2 (b), and Sample 3 (c), all taken at amagnification of 1,000� (size bar corresponds to 10 mm) using uncoated samples imaged at a low accelerating voltage (1.0 kV) to minimizecharging.

204

microscope (FE-SEM) was used. Dried, uncoated samples were

imaged at an accelerating voltage of 1.0 kV to minimize charging.

For the purpose of imaging fibroblasts, samples were fixed in

4% glutaraldehyde (Aldrich) for 1 h at room temperature, rinsed

with distilled water (Sigma), and dehydrated in increasing

concentrations of ethanol (Aldrich). Dehydrated samples were

dried overnight in a desiccator, sputter coated with gold, and

observed using a JEOL M-5610 SEM with an accelerating voltage of

20 kV.

Mercury Intrusion Porosimetry

Porosities and pore size were determined using a PoreMaster 33

mercury intrusion porosimeter (Quantachrome Instruments). The

maximum intrusion pressure was set at 50 psi (corresponding to a

minimum pore size of �4.26 mm), while mercury surface tension

and intrusion and extrusion contact angles were set at 480

erg � cm�2, 1408, and 1408, respectively. The stem volume used for

all the tests was 0.5 cm3. Samples with masses of �50–100 mg



were cut from the xerogels following drying, weighed using an

analytical balance, then analyzed. One analysis was performed for

each sample, except Sample 2, which was analyzed twice to

confirm the reproducibility of the measurement.

Mercury intrusion is one of the most commonly applied

porosimetric methods, thanks to the ease with which the

measurement is performed and the wide range of pore sizes

detected, and operation based on a very simple principle: the

intrusion of a non-wetting fluid into a porous body requires

greater applied pressure as the pore size decreases. This is

described mathematically by Washburn’s equation, Equation (1):

Pc ¼2g

rcos u (1)

where Pc is the capillary pressure, g the surface tension of the

liquid within the capillary, r the radius of the capillary, and u the

contact angle made by the liquid with the walls of the capillary.

Mercury is the most commonly used intrusion liquid because it

displays consistent non-wetting behavior (u�1408) on almost any

surface, the exceptions being compounds with which it may react

DOI: 10.1002/mabi.200600212


(most notably gold, zinc, and selected other metals, sulfur, and

sulfur-rich compounds). While the use of the term ‘‘capillary’’

implies regular cylindrical pores, it is understood that real pores

have more complex geometries. The underlying physics remains

the same, however, in that it is always harder to infiltrate a

smaller pore than a larger pore, allowing the consistent derivation

of pore size distributions by monitoring mercury intrusion into

the sample as a function of pressure. The volume of mercury

intruded into the sample is monitored versus pressure and

converted into a plot of cumulative pore volume versus pore size.

The differential pore volume is then determined based on the

instantaneous slope (i.e., the first derivative) of the cumula-

tive intrusion volume versus pore size curve to give the familiar

differential pore size distribution, where pore volumes given at a

particular pore size are not cumulative but correspond to the

volume of pores of that size alone.

Cell Culture Studies

NIH 3T3 mouse fibroblast cells were cultured on the xerogels in

culture medium containing Dulbecco’s Modified Eagle’s Medium

(DMEM), 10% Bovine Calf Serum (BCS) and 1% penicillin/

streptomycin. The cell number was indirectly measured using

[3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-

sulfophenyl)-2H-tetrazolium] (MTS), a compound that is biore-

duced by cells into a colored formazan product that is soluble in

the tissue culture medium. This conversion is accomplished by

0

1

2

3

4

5

6

7

8

10101

Pore Si

Dif

fere

nti

al P

ore

Vo

lum

e (-

dV

/dlo

gD

, mL

/g)

Figure 4. Pore size distributions of polyester xerogels. Pore size distintrusion porosimetry; Samples 1 and 3 are based on Boltorn H40þSample 2 is based on Boltorn H40þ adipoyl chloride (20% w/v solid



NADPH or NADH produced by dehydrogenase enzymes in

metabolically active cells, allowing for detection of living cells

by monitoring of the absorbance at 490 nm. A CellTiter 96

Aqueous One Solution Cell Proliferation Assay (Promega, Madison,

WI) was used to measure the cellular activity. On day 3, each

sample was put in a fresh well and rinsed twice with

phosphate-buffered saline (PBS) followed by incubation with

20% MTS reagent in a total of 1 ml of serum free culture medium

for 1 h at 37 8C and 10% CO2. Then, five 100 ml aliquots from each

well were pipetted into five wells of a 96 well plate. The 96 well

plates were placed into a Vmax microplate reader (Molecular

Devices, Sunnyvale, CA), with the absorbance of the contents of

each well at 490 nm measured to indicate metabolically active cell

number.

Results and Discussion

A schematic representation of the structure of the Boltorn

H40 hyperbranched polyester polyol is shown in Figure 1.

Demonstrating its biocompatibility and suitably for

medical applications, this aliphatic polyester has already

attracted interest as a precursor for antiviral drugs[15] and

as a basis for dental resins.[16] Figure 2 shows the step of

polymerization reaction by which gel formation occurs. As

with inorganic sol–gel chemistry, the reactions used are

straightforward and reaction speed is readily varied from

1000010000

ze (µm)

Sample 1

Sample 2

Sample 3

ributions of selected polyester xerogels as determined by mercuryterephthaloyl chloride (20 and 15% w/v solids, respectively), while

s). All samples were prepared with a 1:1 OH/COCl molar ratio.



Figure 5. SEM images of fibroblast cells attached to Sample 2.Representative SEM images of fibroblast cells attached toSample 2 following cell culture studies. Samples were fixed withglutaraldehyde and sputter-coated with gold, with images takenat a magnification of 1,000� using an accelerating voltage of20 kV.

Table 1. Cell culture data for polyester xerogels. Results of MTScell proliferation assays performed on selected polyester xero-gels. As before, Samples 1 and 3 are based on BoltornH40þ terephthaloyl chloride (20 and 15% w/v solids, respect-ively), while Sample 2 is based on Boltorn H40þ adipoyl chloride(20% w/v solids). All the samples were prepared with a 1:1 OH/COCl molar ratio.

Sample Mean absorbance at 490 nm

1 0.1000� 0.0036

2 0.2630� 0.0092

3 0.0906� 0.0055

Background 0.1013� 0.0005

206

seconds to days by adjusting the catalyst concentration.

Still, many organic sol–gel reactions lead either to inhomo-

geneous gelation or precipitation due to early phase

separation, or the formation of dense bodies prone to

significant drying shrinkage due to little or no phase

separation. Careful control of the solvent quality as well as

the solid content is, therefore, critical for the successful

preparation of the macroporous xerogels reported here.

Changes in composition also result in changes in the

structure developed during gel formation, further empha-

sizing the need for proper control of solvation conditions in

order to prepare comparable systems in the untemplated

case. While pure solvents are used in the examples given,

mixed solvent systems can be used to fine-tune the sol–gel

process.

Scanning electron microscopy was performed on the

xerogels with representative images shown in Figure 3.

While porosity is volumetric whereas microscopy gives

images of surfaces or thin sections, these results never-

theless offer a qualitative assessment of pore size and



morphology. Smaller pores are observed in Sample 1 as

compared to Samples 2 and 3 but while Sample 1 and

Sample 2 display a granular morphology, some elongated,

almost fibrous features are observed in Sample 3,

indicating a morphological shift, thanks to the lower solid

content versus Sample 1.

Figure 4 shows the mercury intrusion pore size distri-

butions of the polyester xerogels. Based on these results,

the porosities and volume median pore sizes for each

sample are as follows: 79 vol.-%/12 mm (Sample 1), 77

vol.-%/27 mm (Sample 2), and 89 vol.-%/41 mm (Sample 3),

assuming rsolid � 1.2 g � cm�3 for Sample 2 and rsolid � 1.4

g � cm�3 for Samples 1 and 3 based on typical densities of

aliphatic and semi-aromatic polyesters, respectively. A

repeat analysis of Sample 2 (not shown) indicates some

variation in pore size (48 mm) but an almost identical level

of porosity (74 vol.-%), indicating a reasonable level of

homogeneity. From these data, it can be seen that the

sol–gel technique allows for the rapid production of

macroporous polyester xerogels produced via ambient

drying. In addition, these results demonstrate that the pore

structure may be tuned by altering the composition and/or

concentration of the precursor solutions, allowing for the

variation of composition with minimal structural varia-

tion or vice-versa.

The results of the cell culture studies are given in Table 1.

Sample 1 and Sample 3 display background level absor-

bances, indicating no cell growth on these scaffolds.

Sample 2 displays an absorbance significantly higher than

background level, however, implying metabolic activity of

3T3 fibroblasts on this scaffold following cell culturing.

This conclusion is supported by SEM images of Sample 2

following cell culture trials (Figure 5), in which numerous

fibroblast cells are seen to have attached to the polyester

xerogel.

From this, we can draw several conclusions. First, the

positive result in Sample 2 indicates a possible combination

of appropriate pore size and materials chemistry in the

Boltorn H40/adipoyl chloride system. Second, the negative

DOI: 10.1002/mabi.200600212


result in Sample 3, despite having similar pores to those

found in Sample 2, implies that the Boltorn H40/terephtha-

loyl chloride polyester thermoset is incompatible with cell

growth. Accepting this argument, the lack of cell growth in

Sample 1 can be explained by a combination of structural

incompatibility as a result of the very small pore size versus

the other xerogels and materials incompatibility, given an

identical composition to Sample 3.

Conclusion

In this work, we show that porous polyester thermosets

may be produced via an organic sol–gel technique, and

that a range of pore structures may be realized by chang-

ing the composition, solid content, and reaction condi-

tions. The cytocompatibility studies performed on these

xerogels in turn indicate that large pore materials are

appropriate for the growth of mammalian cells, and show

that the replacement of an aliphatic diacid chloride by an

aromatic diacid chloride in otherwise similar polyester

xerogels may degrade cytocompatibility, emphasizing the

importance of materials choice as well as pore structure.

Finally, this work validates the synthesis of fully organic

sol–gel derived xerogels as a useful means to produce

porous polymers with significant cytocompatibility. In

addition to more detailed studies of these and similar

sol–gel systems, we are pursuing replica molding as a

means to ensure the desired pore structure and further

expand the range of scaffolds that may be produced via the

sol–gel technique.

Acknowledgements: We thank Dr. Earl T. Ada (Manager, UMLMaterials Characterization Laboratory) for the SEM images



presented, as well as Hugh Lippincott (graduate student,Biomedical Engineering and Biotechnology Program, U. Mass.Lowell) and Shruti Shah (graduate student, Biomedical Engineer-ing and Biotechnology Program, U. Mass. Lowell) for assistance inmaterials preparation and analysis. Finally, we thank theNanomanufacturing Centers (NSF NSEC Center for High-RateNanomanufacturing/Nanomanufacturing Center of Excellence) atthe University of Massachusetts at Lowell for general support.

Received: September 25, 2006; Revised: December 1, 2006;Accepted: December 12, 2006; DOI: 10.1002/mabi.200600212

Keywords: biocompatibility; porous polymers; sol–gel chemis-try; surfaces; tissue engineering

[1] D. O. Fauza, Curr. Opin. Pediatr. 2003, 15, 267.[2] A. Vats, N. S. Tolley, J. M. Polak, J. E. Gough, Clin. Otolaryng.

2003, 28, 165.[3] P. X. Ma, Mater. Today 2004, 7, 30.[4] K. F. Leong, C. M. Cheah, C. K. Chua, Biomater. 2003, 24, 2363.[5] A. Curtis, M. Riehle, Phys. Med. Bio. 2001, 46, R47.[6] Y. Tabata, Pharm. Sci. Tech. Today 2000, 3, 80.[7] K. Y. Lee, D. J. Mooney, Chem. Rev. 2001, 101, 1869.[8] J. L. Drury, D. J. Mooney, Biomater. 2003, 24, 4337.[9] B. Ben-Nissan, Curr. Opin. Solid State Mater. Sci. 2003, 7, 283.

[10] H.-M. Kim, Curr. Opin. Solid State Mater. Sci. 2003, 7, 289.[11] P. Petrini, S. Fare, A. Piva, M. C. Tanzi, J. Mater. Sci. Mater. Med.

2003, 14, 683.[12] J. K. V. Tessmar, T. A. Holland, A. G. Mikos, ‘‘Salt Leaching for

Polymer Scaffolds: Laboratory-Scale Manufacture of CellCarriers’’, in: Scaffolding in Tissue Engineering, 1st edition,

P. X. Ma, J. Ellisseeff, Eds., CRC Press LLC, Boca Raton 2005,p. 111 ff.

[13] J. N. Weber, E. W. White, Mineral Sci. Eng. 1973, 5, 150.[14] E. W. White, J. N. Weber, D. M. Roy, E. L. Owen, R. T. Chiroff,

R. A. White, J. Biomed. Mater. Res. 1975, 9, 23.[15] J. Rojo, R. Delgado, J. Antimicrob, Chemother. 2004, 54, 579.[16] Q. Wan, S. R. Schricker, B. M. Culbertson, J. Macromol. Sci. Pure

Appl. Chem. 2000, A37, 1317.









A Flexible Approach to the Preparation of Polymer Scaffolds for Tissue Engineering

Documents

Transcript of A Flexible Approach to the Preparation of Polymer Scaffolds for Tissue Engineering