Biomaterials 25 (2004) 2319–2329

ARTICLE IN PRESS

*Correspondin

E-mail addres

0142-9612/$ - see

doi:10.1016/j.bio

Effect of PEG–PLLA diblock copolymer on macroporous PLLAscaffolds by thermally induced phase separation

Hyun Do Kima, Eun Hee Baeb, Ick Chan Kwonc, Ravindra Ramsurat Pala,Jae Do Nama, Doo Sung Leea,*

aDepartment of Polymer Science and Engineering, Center for Advanced Functional Polymers, Sungkyunkwan University,

Suwon, Kyungki 440-746, South KoreabRegenbiotech, 945A, 39-1, Korea Institute of Science and Technology, P.O. Box 131, Cheongryang, Seoul 130-650, South Korea

cBiomedical Research Center, 39-1, Korea Institute of Science and Technology, P.O. Box 131, Cheongryang, Seoul 130-650, South Korea

Received 19 June 2003; accepted 7 September 2003

Abstract

A regular and highly interconnected macroporous poly(l-lactic acid) (PLLA) scaffold was fabricated from a PLLA–dioxane–

water ternary system with added polyethylene glycol (PEG)–PLLA diblock using thermally induced phase separation (TIPS). The

morphology of the scaffold was investigated in detail by controlling the following TIPS parameters: quenching temperature, aging

time, polymer concentration, molecular structure, and diblock concentration. The phase diagram was assessed visually on the basis

of the turbidity. The cloud-point curve shifted to higher temperatures with increasing PEG content in the additives (PEG–PLLA

diblocks), due to a stronger interaction between PEG and water in solution. The addition of diblock series (0.5wt% in solution)

stabilized interconnections of pores at a later stage without segregation or sedimentation. The pore size of the scaffold could be

easily controlled in the range 50–300mm. A macroporous PLLA scaffold was used to study an MC3T3-E1 cell (an osteoblast-like

cell) culture. The cells successfully proliferated in the PLLA scaffold in the presence of added PEG-PLLA diblock for 4 weeks.

r 2003 Elsevier Ltd. All rights reserved.

Keywords: Macroporous scaffold; Thermally induced phase separation (TIPS); PLLA; PEG-PLLA diblock copolymer; MC3T3-E1 cell

1. Introduction

The incidence of organ and tissue loss or failure isincreasing steadily, whereas the traditional surgicaltreatment of implantation of a healthy organ from adonor is limited by immune rejection and the number ofdonors. As an application of tissue engineering, the useof cell transplantation is now being investigated as analternative therapeutic strategy for tissue repair andorgan replacement [1–5]. Transplanted cells culturedfrom the healthy issues of a patient can be implantedback to him or her without requiring an immunoisola-tion system. In the system reported here, the cell cultureis related to the scaffold shape because the scaffold playsan important role in the design of the specializedtemporary substrates to be grown [6–9]. Biodegradableand biocompatible synthetic polymers, such aspoly(lactic acid), poly(glycolic acid), and poly(d,l-lactic

g author. Fax: +82-31-292-8790.

s: dslee@skku.edu (D.S. Lee).

front matter r 2003 Elsevier Ltd. All rights reserved.

materials.2003.09.011

acid-co-glycolic acid), have been widely utilized as three-dimensional (3D) scaffolds [10–12]. Polymeric scaffoldsshould be porous enough to allow a high density of cellsto be seeded, and provide sufficient mechanical stabilityand well-defined networks of interconnected pores topermit ingrowth into the implanted structure [8,13]. Theoptimum pore size of the scaffold differs with the type ofcells or tissue; for example, close to 20 mm for theingrowth of fibroblasts and hepatocytes [14], 50–150 mmfor skin regeneration [15], and 100–150 mm for boneregeneration [16,17].

Poly(l-lactic acid) (PLLA) scaffolds are currentlybeing tested for hard-tissue cell transplantation, and arefabricated by various techniques, such as porogenleaching/salt leaching, emulsion freeze-drying, expan-sion in high-pressure gas, and phase separation [18–22].Recently, thermally induced phase separation (TIPS)and freeze-drying have been used to prepare 3Dmacroporous PLLA scaffolds [23–29]. The TIPS meth-od involves two thermodynamic processes: (1) nuclea-tion and growth in the metastable region of the phase

ARTICLE IN PRESS

Table 1

Diblock copolymers used

Copolymer Mna

Diblock1 PEG114–PLLA6 5000–413

Diblock2 PEG114–PLLA40 5000–2848

Diblock3 PEG114–PLLA62 5000–4532

Diblock4 PEG45–PLLA26 2000–1830

aNumber-average molecular weight as measured by 1H-NMR.

H.D. Kim et al. / Biomaterials 25 (2004) 2319–23292320

diagram [23], and (2) spinodal decomposition process inthe unstable region [24–29]. The thermodynamic state ofthe polymeric solution at the time that the solution isquenched determines the morphology of the resultingscaffold. If the solution is quenched whilst in the formerstate, a scaffold with irregular pore size and poorlyinterconnected structure will be induced, whereas ifquenching occurs in the latter state, an open, porous,and 3D structure will be induced. The open, porousmorphology can be controlled by several experimentalparameters: the quenching temperature, quenching rate,quenching period or aging time, polymer concentration,ratio of solvent to nonsolvent, molecular structure, andthe addition of porogen [28–32]. Nam and Park reportedthat an open porous PLLA scaffold with a pore size of10–50 mm resulted from the coarsening process in thelater stages of TIPS in a PLLA–dioxane–water ternarysystem [25]. They also found that the addition of aporogen such as Pluronic surfactant (F-127) in the TIPSformulation increased the pore size and controlled themorphology.

In previous papers [28–30], we reported that thePLLA–dioxane–water ternary system (dioxane:water at87:13% w/w) could be used to fabricate a highlyinterconnected macroporous scaffold with a pore sizeof 50–150 mm through a liquid–liquid spinodal decom-position mechanism at room temperature. However, thegelation due to partial crystallization of PLLA pre-vented the development of phase separation in the latestage, restricted the interconnections, and reduced thepore size during the coarsening process. The addition ofporogen to the PLLA–dioxane–water ternary systemraised the cloud point and increased the thermodynamicdriving force, resulting finally in an increase in the poresize of the scaffold.

In the present study, we investigated the effect of theaddition of PEG–PLLA diblock and PEG on themorphology of PLLA scaffolds. The morphology ofthe scaffold was investigated whilst adjusting TIPSparameters, such as quenching temperature, aging time,and additives and their concentrations. In order todetermine the optimal pore size of the PLLA scaffold,MC3T3-E1 cells were cultured in a series of PLLAscaffolds with different pore sizes, and the number ofcell was counted in a BCA protein assay.

2. Experimental

2.1. Materials

PLLA (Lacty 5000) was purchased from Shimadzu.The number-average molecular weight was 218,000(PDI, Mw/Mn=1.55). 1,4-Dioxane and deionized waterwere a good solvent and nonsolvent for PLLA,respectively. Monomethoxy poly(ethylene oxide)

(MPEG, Mn: 2000, 5000, Mw/Mn=1.1, Aldrich) wasdissolved in dried chloroform and then precipitated inn-hexane. l-lactide (Boehringer Ingelheim) was purifiedby recrystallization from thoroughly dried ethyl acetateunder a dry nitrogen atmosphere and sublimed beforeuse.

PEG-PLLA diblocks (listed in Table 1) were synthe-sized from MPEG and l-lactide and characterized asreported in a previous paper [33]. Brief synthesis methodis as follows. MPEO was introduced into a round-bottom, three necked flask with dried toluene. Residualwater in the solution was removed by azeotropicdistillation, resulting in half of toluene solution.Appropriate amount of l-lactide and Sn(Oct)2 wereadded and refluxed under dry nitrogen atmosphere for12 h for the synthesis of PEO–PLLA diblock copolymer.After the reaction was completed, the solution wasprecipitated in diethyl ether and then dried undervacuum at ambient temperature for 24 h. Mn wasmeasured by 1H-NMR and GPC.

2.2. Phase diagram

The cloud-point curve of the PLLA ternary system inthe presence of PEG or PEG–PLLA diblocks wasdetermined by visual turbidimetry. Weighted PLLA (1,3, 4.5, 5.5, or 7wt%) and PEG or PEG–PLLA diblocks(0.2, 0.5, or 1wt% in whole solution) were added in a 4-ml-vial with a 1,4-dioxane:water mixture (87%:13%w/w) as the solvent, and then dissolved by heating at63�C for 5 h whilst stirring continuously. The homo-genous PLLA solution was reheated to approximately10�C above the expected cloud-point temperature, andthen slowly cooled in steps of 1�C, equilibrating thesystem for 10min at each new temperature. The cloudpoint was determined as the temperature at which theclear solution became visually turbid. The gelation pointwas determined by inverting the vial horizontally after ithad been maintained for 10min at a constant tempera-ture, as described in a previous paper [28].

2.3. Preparation of PLLA scaffold

PLLA (4.5 or 5.5wt%) solutions containing PEG orPEG–PLLA diblocks (0.2, 0.5, or 1wt%) with a mixtureof 1,4-dioxane and water (87:13%w/w) as the solvent

ARTICLE IN PRESS

PLLA Contentration (wt%)

0 1 2 3 4 5 6 7 8

Tem

per

atu

re (

°C)

20

30

40

50

60

70PEG 5000 1wt%PEG 5000 0.5wt%PEG 5000 0.2wt%

Diblock3 1wt% Diblock3 0.5wt% Diblock3 0.2wt%

Pure PLLA

Fig. 1. Cloud-point curves of diblock3 and PEG5000.

H.D. Kim et al. / Biomaterials 25 (2004) 2319–2329 2321

were prepared. Each sample was reheated to 15�C abovethe measured cloud-point temperature, and then wasplaced in a water bath preheated to the quenchingtemperature. It was kept for 2, 10, 30, 60, or 120min atthe quenching temperature to observe the coarseningeffect. The annealed sample was directly immersed inliquid nitrogen to be fast frozen for 1 h, and then a smallhole was cut in the vial cap to release the solvents.Freeze-drying was performed in a freeze dryer at �77�Cand 7mTorr for 3 days in order to remove solvents andobtain the macroporous scaffolds. The dry scaffoldswere cut into cubes with a surgical blade (7–8� 7–8mm,thickness 2mm; 13–15mg PLLA). Prior to cell seeding,3D scaffolds were prewetted with 70% ethanol for 3 h tosterilize them and enhance their water uptake. Theethanol was removed by soaking with agitation for 1 h insix changes of phosphate-buffered saline (PBS), andthen the scaffolds were left overnight in the media.Finally, the pure PLLA scaffolds were prepared byfollowing similar procedures.

2.4. Cell culture

MC3T3-E1 cells (osteoblast-like cells derived frommouse calvaria) were kept in a-minimum essential media(a-MEM, Gibco BRL) containing 10% fetal bovineserum (FBS, Gibco BRL). Cells were trypsinised andsuspended at the concentration of 2� 105 cells per ml infresh media. Aliquots of 3-ml cell suspensions(2� 105 cells/ml) were seeded onto the top of prewettedscaffolds placed in a 60-mm petri dish. The loadedscaffolds were placed in a humidified 5% CO2 incubatorat 37�C, and the medium was changed every 2–3 days.Cell proliferation on each specimen was determinedafter 1, 3, 7, 10, 14, 21, and 28 days after seeding.

2.5. Lysis and cell count from BCA protein assay

The number of cells in the scaffolds was determinedindirectly by using BCA protein assay for total cellularprotein on each alternate culture day. The culturedscaffolds were harvested every other day and washedtwice with PBS. The washed scaffolds were homoge-nized with 500 ml of lysis buffer (60mm Tris–Cl pH 6.8,25% glycerol, 2% SDS, 1mm PMSF, 1 mg/ml Aprotinin)for 2min on ice. An aliquot of the lysate was measuredfor total protein with a commercial BCA protein assaykit (Pierce, Rockford, IL, USA) according to themanufacturer’s instructions. After incubation for 2 h,absorbance readings were performed with an ELISAreader at a wavelength of 562 nm. The amount of totalprotein was extrapolated to cell numbers by comparisonto the total protein of known number of MC3T3-E1cells treated in exactly same conditions except beingloaded in scaffolds.

2.6. Morphology characterization

The macroporous morphology of the scaffolds wasobserved using scanning electron microscopy (SEM,Hitachi S-2400). Fracture-frozen cross-sections of thescaffold were mounted on an aluminum stub coveredwith a carbon adhesive, and then coated with goldparticles.

3. Results and discussion

3.1. Cloud-point curve

Cloud-point temperatures of the PLLA–dioxane–water ternary system with various PEG–PLLA diblocksand PEG are shown in Figs. 1 and 2. As described inprevious reports [28–30], the cloud-point curve of thepure PLLA–dioxane–water ternary system graduallyincreased with increasing polymer concentration at afixed dioxane:water ratio (87%:13% w/w). The cloud-point curves of the systems containing PEG5000 ordiblock3 shifted to higher temperatures as the amountof additive increased from 0.2 to 1wt% (Fig. 1).

In Fig. 2, the cloud-point curves of solution in thepresence of various additives (at 0.5wt%) are shifted tohigher temperatures in the following order: PEG5000>diblock1>diblock2>diblock3>diblock4. It is knownthat phase separation is assisted by addition of diblockor triblock copolymer as a surfactant. This assistance isinduced by the amphiphilic effect of the addedsurfactant serving as a nucleus for phase separation. Itis affected by the molecular properties of the additives,including molecular weight, hydrophobic/hydrophilicblock ratio, and block lengths [34,35]. Obviously, an

ARTICLE IN PRESS

PLLA Concentration (wt%)

0 1 2 3 4 5 6 7 8

Tem

per

atu

re (

°C)

20

30

40

50

60

70

Sedimentation boundary

GelationSediment

PEG 5000 0.5wt% Diblock1 0.5wt% Diblock2 0.5wt% Diblock3 0.5wt%Diblock4 0.5wt%

Pure PLLAPEG 5000 0.5wt% gelation pointDiblock2 0.5wt% gelation point

Diblock3 0.5wt% gelation point Pure PLLA

A B

Fig. 2. Cloud-point curves (closed symbols) and gelation-point curves

(open symbols) of diblocks and PEG. Sedimentation line of pure

PLLA solution (A) and with additives (B).

H.D. Kim et al. / Biomaterials 25 (2004) 2319–23292322

increase in PEG content in the diblock copolymer(PEG>diblock1>diblock2>diblock3) raises the cloudpoint to a higher temperature. Increasing the PEGcontent in a diblock copolymer can enhance theinteraction between the diblock copolymer and water,because PEG exhibits stronger solubility in water thanPLLA, so that the cloud-point temperature (liquid–liquid demixing) shifts to a higher temperature.A diblock copolymer having the same PEG andPLLA content but smaller size (diblock4) shows aslightly lower cloud-point temperature than diblock3in Fig. 2.

As illustrated above, the PLLA solution showsgelation at temperatures lower than the cloud pointtemperature. This gelation resulted from the partialcrystallization of PLLA in the polymer-rich phase afterthe liquid–liquid phase separation. The pure PLLAsolution having a higher concentration than thesedimentation boundary (o4wt%, dotted line A inFig. 2) shows gelation when it was cooled to below thegelation point [28]. The gelation point graduallyincreased with increasing polymer concentration. Belowthe sedimentation boundary, the polymer solution wasseparated into two layers: a polymer-rich phase and asolvent phase. Fig. 2 shows that the gelation-pointcurves of ternary system with PEG or diblocks shifted toa slightly higher temperature. The addition of diblocksor PEG has only a small effect on the gelation of PLLA,while the cloud-point curve was greatly shifted. Thesedimentation boundary is shifted to a higher polymerconcentration (o4.5wt%, solid line B in Fig. 2) withincreasing PEG content in solution, but this change inthe sedimentation boundary is smaller than the changewhen NaCl is used as an additive [29].

3.2. Effect of quenching temperature

As described above, the porous morphology of thescaffold is determined by the thermodynamic state ofthe solution. The porous morphology was formed by aphase-separation mechanism, such as binodal or spino-dal decomposition. The morphology depended onvarious processing parameters, including quenchingtemperature, polymer concentration, solvent composi-tion, and aging time. In a previous paper [28], anoptimum process condition was 4.5 wt% polymerconcentration at a dioxane:water ratio of 87%:13%(w/w), which has been found to result in a regular PLLAscaffold with a pore size of 50–150 mm. On the basis ofthis finding, various PLLA scaffolds were prepared from4.5wt% PLLA in dioxane/water (87:13w/w) withadditives.

Fig. 3 shows SEM micrographs of a 4.5wt% PLLAsolution with 0.5wt% diblock2 made by quenching atvarious temperatures, as a function of aging time. Thephase diagram in Fig. 2 shows that three quenchingtemperatures (25�C, 30�C, and 35�C) are located in theunstable region (probably the spinodal region). There-fore, the interconnected and open porous structure isobserved even at the starting period of phase separation(o10min). The pore size of the scaffold at a quenchingtemperature of 30�C increases continuously with agingtime; after 30min the pore size is greater than 150 mm(Fig. 3b).

When the scaffold was fabricated at a higherquenching temperature (35�C, rather than 30�C), whichhas a smaller quench depth (9�C), pores developedmuch faster. The liquid–liquid phase separation is fasterat a higher temperature due to the lower viscosity of thepolymer solution, even though it has a lower drivingforce due to the smaller quench depth. In this case thepore size grew to 200 mm at 30min. At a longer agingtime (120min), the walls between the pores weredestroyed. This seems to represent the final stage ofthe coarsening process, as shown in Fig. 3c, whichmeans that the coalescence process progresses quickly atlow viscosity. In contrast, when the solution wasquenched at 25�C, the pore size and variation weresmaller even though the driving force was larger due tothe larger quench depth. In this case quenching occursbelow the gelation-point temperature, so the partialcrystallization of PLLA reduces the rate of phaseseparation and restricts the growth of pores at a longeraging time (>30min).

3.3. Effect of additives

At the starting period of phase separation, themorphology of the scaffold is determined by the initialthermodynamic driving force which is dependent on thequench depth from the cloud-point temperature. The

ARTICLE IN PRESS

(a) (b) (c)

10min

30min

60min

120min

500 m

100 m

100µm

100µm 100µm

100µm 100µm

100µm

100µ m 100µ m

100µ m 100µ m

500µ m

100µm

Fig. 3. SEM micrographs of scaffolds prepared from a 4.5wt% PLLA dioxane:water (87%:13%w/w) solution with 0.5wt% diblock2, as a function

of aging time at quenching temperatures of 25�C (a), 30�C (b), and 35�C (c).

H.D. Kim et al. / Biomaterials 25 (2004) 2319–2329 2323

cloud-point temperature is increased by the addition ofdiblocks or PEG5000, as shown in Fig. 2, whichincreases the quench depth for the same quenchingtemperature. Fig. 4 illustrates the effect of diblock1 anddiblock3 on PLLA scaffold morphology with aging timeat a quenching temperature of 30�C. The addition ofdiblock1 or diblock3 increases the thermodynamicdriving force relative to that in the pure PLLA solution.This increases the size of interconnected pores for ashort aging time (o2min), but after 10min of aging thephase separation of the pure PLLA solution proceeds tothe later stage of phase separation. After 60min, anirregular macroporous structure was observed due to acoalescence process involving the combining of larger

structures at the expense of smaller structures [36] (asshown in Fig. 4a).

The addition of diblocks to a ternary system appearsto make the porous structure stable even after a longaging time, and prevents the coalescence of the porousstructure and the sedimentation of the polymer-richphase. This is particularly clear for the systems contain-ing diblocks in Figs. 3b, 4b and c. For these systems,regular, open, and well-interconnected macropores wereeasily fabricated in the size range 150–200 mm after60min. These amphiphilic diblocks can act like asurfactant, lowering the interfacial tension that isimportant at the later stage of phase separation. At thisstage, the phase-separation kinetics is controlled by the

ARTICLE IN PRESS

2min

10min

30min

60min

120min

100 m

100 m

100µm 100µm 100µ m

100µm 100µm 100µ m

100µ m100µm100µm

100µm 100µm 100µm

100µ m100µm

(a) (b) (c)

Fig. 4. SEM micrographs of scaffolds prepared at a quenching temperature of 30�C from a 4.5wt% PLLA solution with no additives (a), 0.5wt%

diblock1 (b), and 0.5wt% diblock3 (c).

H.D. Kim et al. / Biomaterials 25 (2004) 2319–23292324

motion of interfaces driven by interfacial tension [34].The lowering of surface tension could decelerate thephase separation and enhance the stability of the porousstructure and its interconnections. The hydrophilic/hydrophobic ratio of the diblocks can be a major factorin changing the interfacial tension. The addition ofdiblock3 (rather than diblock1) with a low PEG/PLLAratio causes low interfacial tension, making it is easier toobtain the desired pore size and scaffold interconnec-tions. A system with a pore size of 150–250 mm and ahigh degree of interconnections was obtained after120min when using diblock3 (Fig. 4).

3.4. Effect of molecular structure of additives

The morphology of scaffolds prepared from thePLLA solution with PEG and diblock4 at the samePEG/PLLA ratio but with a shorter block length areshown in Fig. 5. The scaffolds were prepared from4.5wt% PLLA solution at a quenching temperature of30�C. The system with PEG added at 0.5 wt% exhibitedthe largest quench depth (18�C) at 30�C; it is located at1�C above the gelation-point temperature. Within ashort aging time (o2min), the large thermodynamicdriving force produces a large pore size. However, after

ARTICLE IN PRESS

2min

10min

30min

60min

120min

(a) (b)

100 m

100µ m100µ m

100µ m

100µ m

100µ m

100µ m 100µ m

100µ m

100µ m

100µ m

Fig. 5. SEM micrographs of scaffolds prepared at a quenching temperature of 30�C from 4.5wt% PLLA solution with 0.5wt% PEG5000 (a) and

0.5wt% diblock4 (b).

H.D. Kim et al. / Biomaterials 25 (2004) 2319–2329 2325

10min the pore growth gradually decelerated andstopped during the coarsening process due to gelation,as shown in Fig. 5a.

In comparison with diblock3, the molecular structureof diblock4 is characterized by a similar hydrophobic/hydrophilic ratio but a shorter block length. As

mentioned above, the lower the PEG content and theshorter the block length of additives, the lower thecloud-point temperature. In this case, a smaller drivingforce is produced at the same quenching temperature.The initially small pores increase gradually over a longaging time (30–60min), developing into a regular, highly

ARTICLE IN PRESSH.D. Kim et al. / Biomaterials 25 (2004) 2319–23292326

interconnected macroporous structure due to thesurfactant effect (Fig. 5b). The pore size is 150–300 mm, greater than that of the diblock3 system andof any other system at 30�C. The cloud-point tempera-tures of the diblock3 and diblock4 systems are similar.Note that diblock4 contains twice as many molecules asdiblock3 (by comparison of the molecular weights). Itshould be noted that the scaffold pore size can be furtherincreased by additives having the same PEG/PLLAratio and smaller molecular weight. The system contain-ing diblock4 had a stable interconnected structurebetween macropores, with no tendency for segregationor sedimentation even over a long aging time.

3.5. Effect of additives and PLLA concentration

The effect of additives and PLLA concentration onthe scaffold morphology was investigated. Scaffoldswere prepared at a quenching temperature of 30�C anda 60-min aging time (Fig. 6). As shown in Fig. 6a, thescaffold with 4.5wt% PLLA and 0.2wt% diblock3exhibited an irregular morphology with some closedpores, although it was better interconnected than thepure PLLA system (Fig. 4a). In addition, the pore sizewas smaller than that of the system with 4.5wt% PLLAand 0.5wt% diblock3 (see Fig. 4c). This indicates that tofabricate scaffolds with a regular, macroporous, and

100µm

100µ m

(a) (b

(c) (

Fig. 6. SEM micrographs of scaffolds prepared at a quenching temperature

4.5wt% PLLA solution with 0.2wt% diblock3; (b–d) 5.5wt% PLLA soluti

highly interconnected structure, the amount of diblockadded should exceed a certain critical concentration.

When a 5.5wt% PLLA solution with diblock3 wasquenched at 30�C, the pore size increased only slightly(from 80 to 100 mm) with the amount of diblock3 (from0.2 to 1.0wt%; Fig. 6b–d). This pore size is smaller thanthat of the 4.5wt% PLLA system (Fig. 4c) due to thehigh viscosity of the 5.5 wt% PLLA solution induced bythe high polymer concentration and partial gelation.

3.6. Cell culture

PLLA has been used in hard-tissue regeneration dueto its slow degradation. In this study, the scaffolds madeof PLLA with various pore sizes were investigated forthe application as substrates for osteoblast cell, which isa precursor of bone cell. MC3T3-E1 cells (mouseosteoblast-like cells) are able to proliferate and start todifferentiate to bone cells upon confluency. To verify ifMC3T3-E1 cells could adhere to and proliferate on thescaffolds, we monitored the changes in cell number at 1,3, 7, 14, 21, and 28 days after loading MC3T3-E1 cells tothe scaffolds.

Fig. 7 shows the results of cell proliferation assayusing aforementioned scaffolds with pore sizes from lessthan 100 mm to more than 300 mm. We found that cellnumbers were increased in all five types of scaffold until

100µ m

100µ m

)

d)

of 30�C from a PLLA solution with diblock3 after 60min of aging: (a)

on with 0.2, 0.5 and 1.0wt% diblock3, respectively.

ARTICLE IN PRESSH.D. Kim et al. / Biomaterials 25 (2004) 2319–2329 2327

21days and most abundant on scaffolds with pore size of150–200 mm. To confirm these optical data, we deter-mined the amount of total protein, which is from cellcontent and extracellular matrix protein produced from

3days 7day

(a)

(b)

(c)

(d)

(e)

100µ m

100µ m 100µ m

100µ m

100µ m100µm

100µ m

100µ m

100µ m 100µm

Fig. 7. SEM micrographs of cell proliferation in scaffolds prepared from 4

scaffold, quenching temperature of 20�C, after 30min of aging (pore sizeB1

30min of aging (pore size 100–150mm). (c) With 0.5wt% diblock1, quenchi

(d) With 0.5wt% diblock3, quenching temperature of 30�C, after 60min of

temperature of 30�C, after 60min of aging (pore size 200–300mm).

metabolically active cells, per mg scaffold (Fig. 8). Theamounts of total protein from all five types of scaffoldwere increased until 28days. The amount of protein wasincreased mostly when cells were cultured on scaffold

s 21days

100µ m

100µ m

100µ m

100µ m

100µm

.5 wt% PLLA solution as a function of culture days: (a) Pure PLLA

00mm). (b) Pure PLLA scaffold, quenching temperature of 30�C, after

ng temperature of 30�C, after 60min of aging (pore size 150–200mm).

aging (pore size 200–300mm). (e) With 0.5wt% diblock4, quenching

ARTICLE IN PRESS

0 5 10 15 20 25 30

0

2

4

6

8

10

12

14

16

18

20Diblock1Pure 30 °CDiblock3Diblock4Pure 20 °C

Cel

l Num

ber

× 10

4 / P

LL

A s

caff

old

mg

Culture days

Fig. 8. Cell proliferation curves from a BCA protein assay.

H.D. Kim et al. / Biomaterials 25 (2004) 2319–23292328

with pore size of 150–200 mm and secondly on scaffoldwith pore size of 100–150 mm. The amount of proteinwas increased least on scaffold with pore size of 100 mm.These results suggested that the optimum pore size tosupport growth of metabolically active MC3T3-E1 cellsis around 150 mm. We could speculate that this pore sizeis the maximum to hold enough number of cellsinducing cell–cell interaction and the minimum tofacilitate material transfer to provide nutrients to cells.MC3T3-E1 cells have morphology of adherent fibro-blast-like cells when growing as monolayer. However,when reached to confluency with sufficient cell-cellinteraction, they start to differentiate to bone cells.There is high possibility that the scaffold with pore sizeof 150–200 mm could be used to induce the differentia-tion of MC3T3-E1 cells to bone. To clear this issue,culturing beyond 28days and assay for differentiationmarker will be performed.

4. Conclusions

The addition of amphiphilic diblocks to a PLLA–dioxane–water ternary system could provide a newmethod for preparing macroporous scaffolds by TIPS.Using this technique it is possible to fabricate a regularand highly interconnected and macroporous PLLAscaffold without segregation or sedimentation over along aging time (>30min). The pore size of PLLAscaffold was easily controllable from 50 to 300 mm byadjusting the quenching temperature, aging time, poly-mer concentration, and composition and block length ofthe diblocks.

Cell proliferation profile was characterized by BCAprotein assay, PLLA scaffold with size of 150–200 mm,

induced the highest proliferation providing enoughnutrient transfer and cell-cell interaction.

This study, which builds on previous work by ourgroup [28–30,37], provides a promising method forcontrolling the pore size of open, regular, and well-interconnected macroporous scaffolds for the growthand culture of cells, when the scaffolds are prepared byTIPS.

Acknowledgements

This work was supported by a grant from the KoreaResearch Foundation (no. KRF-2001-005-E00006).

References

[1] Russell PS. Selective transplantation. An emerging concept. Ann

Surg 1985;201:255–62.

[2] Vacanti JP, Morse MA, Saltzman WM, Domb AJ, Perez-Atayde

A, Langer R. Selective cell transplantation using bioabsorbable

artificial polymers as matrices. J Pediatr Surg 1988;23:3–9.

[3] Mikos AG, Sarakinos G, Lyman MD, Ingber DE, Vacanti JP,

Langer R. Prevascularization of biodegradable polymer scaffolds

for hepatocyte transplantation. Polym Mater Sci Eng Prepr

1992;66:34–5.

[4] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:

920–6.

[5] Vacanti JP. Beyond transplantation. Third annual Samuel Jason

Mixter lecture. Arch Surg 1988;123:545–9.

[6] Raper SE, Wilson JM. Cell transplantation in liver-directed

gene therapy. Cell Transplantation 1993;2:381–400 (discussion

407–410).

[7] Liu DM, Dixit V. Porous materials for tissue engineering.

Switzerland: Tech Publication Press; 1997.

[8] Mikos AG, Bao Y, Cima LG, Ingber DE, Vacanti JP, Langer R.

Preparation of poly(glycolic acid) bonded fiber structures for cell

attachment and transplantation. J Biomed Mater Res 1993;

27:183–9.

[9] Mooney DJ, Mazzoni CL, Breuer C, McNamara K, Hern D,

Vacanti JP, Langer R. Stabilized polyglycolic acid fibro-based

tubes for tissue engineering. Biomaterials 1996;17:115–24.

[10] Sutherland IW. Alginates. In: Byrom D, editor. Biomaterials:

novel materials from biological sources. New York: Stockton

Press; 1991.

[11] Kaplan DL, Wiley BJ, Mayer JM. Biosynthetic polysaccharides.

In: Biomedical polymers: designed-to-degrade systems. Munich,

Germany: Carl Hanser Verlag; 1994.

[12] Freed LE, Marquis JC, Nohria A, Emmanual J, Mikos AG,

Langer R. Neocartilage formation in vitro and in vivo using cells

cultured on synthetic biodegradable polymers. J Biomed Mater

Res 1993;27:11–23.

[13] Cima LG, Vacanti JP, Vacanti C, Ingber DE, Mooney DJ,

Langer R. Tissue engineering by cell transplantation

using degradable polymer substrates. J Biomech Eng 1991;113:

143–51.

[14] Cima LG, Ingber DE, Vacanti JP, Langer R. Hepatocyte culture

on biodegradable polymeric substrates. Biotechnol Bioeng

1991;38:145–58.

[15] Wilkins LM, Watson SR, Prosky SJ, Meunier SF, Parenteau NL.

Development of a bilayered living skin construct for clinical

applications. Biotechnol Bioeng 1994;43:747–56.

ARTICLE IN PRESSH.D. Kim et al. / Biomaterials 25 (2004) 2319–2329 2329

[16] Puleo DA, Holleran LA, Doremus RH, Bizios R. Osteoblast

responses to orthopedic implant materials in vitro. J Biomed

Mater Res 1991;25:711–23.

[17] Thomson RC, Wake MC, Yaszemski MJ, Mikos AG. Biodegrad-

able polymer scaffolds to regenerate organs. Adv Polym Sci

1995;122:245–74.

[18] McGrath KP, Fournier MJ, Mason TL, Tirrell DA. Genetically

directed syntheses of new polymeric materials. Expression of

artificial genes encoding proteins with repeating -(AlaGly)3Pro-

GluGly-elements. J Am Chem Soc 1992;114:727–33.

[19] Kung IM, Wang FF, Chang YC. Surface modifications of

alginate/poly(l-lysine) microcapsular membranes with poly(ethy-

lene glycol) and poly(vinyl alcohol). Biomaterials 1995;

16:649–55.

[20] Freed LE, Vunjak-Novakovic G, Biron RJ, Eagles DB, Lesnoy

DC, Barlow SK, Langer R. Biodegradable polymer scaffolds for

tissue engineering. Biotechnology 1994;12:689–93.

[21] Harris LD, Kim BS, Mooney DJ. Open pore biodegradable

matrices formed with gas foaming. J Biomed Mater Res

1998;42:396–402.

[22] Whang K, Thomas CH, Healy KE. A novel method to fabricate

bioabsorbable scaffolds. Polymer 1995;36:837–42.

[23] Schugens C, Maquet V, Grandfils C, Jerome R, Teyssie P.

Biodegradable and macroporous polylactide implants for cell

transplantation: 1. Preparation of macroporous polylactide

supports by solid–liquid phase separation. Polymer 1996;37:

1027–38.

[24] Nam YS, Park TG. Porous biodegradable polymeric scaffolds

prepared by thermally induced phase separation. J Biomed Mater

Res 1999;47:8–17.

[25] Nam YS, Park TG. Biodegradable polymeric microcellular foams

by modified thermally induced phase separation method. Bioma-

terials 1999;20:1783–90.

[26] Lloyd DR, Kim SS, Kinzer KE. Microporous membrane

formation via thermally induced phase separation. II. Liquid–

liquid phase separation. J Membr Sci 1991;64:1–11.

[27] Schugens C, Maquet V, Grandfils C, Jerome R, Teyssie P.

Polylactide macroporous biodegradable implants for cell trans-

plantation. II. Preparation of polylactide foams by liquid–liquid

phase separation. J Biomed Mater Res 1996;30:449–61.

[28] Hua FJ, Kim GE, Lee JD, Son YK, Lee DS. Macroporous

poly(l-lactide) scaffold 1. preparation of a macroporous scaffold

by liquid–liquid phase separation of a PLLA–dioxane–water

System. J Biomed Mater Res 2002;63:161–7.

[29] Hua FJ, Nam JD, Lee DS. Preparation of a macroporous

poly(l-lactide) scaffold by liquid-liquid phase separation of

PLLA–dioxane–water ternary system in the presence of NaCl.

Macromol Rapid Commun 2001;22:1053–7.

[30] Hua FJ, Park TG, Lee DS. A facile preparation of highly

interconnected macroporous poly(d,l-lactic acid-co-glycolic acid)

(PLGA) scaffolds by liquid–liquid phase separation of a PLGA–

dioxane–water ternary system. Polymer 2003;44:1911–20.

[31] Cha BJ, Char K, Kim JJ. The effects of diluent molecular weight

on the structure of thermally induced phase separation mem-

brane. J Membr Sci 1995;108:219–29.

[32] Witte P, Dijkstra PJ, Berg AWV, Feijen J. Phase separation

processes in polymer solutions in relation to membrane forma-

tion. J Membr Sci 1996;117:1–31.

[33] Choi SW, Choi SY, Jeong BM, Kim SW, Lee DS. Thermo-

reversible gelation of poly(ethylene oxide) biodegradable

polyester block copolymers. II. J Polym Sci A: Polym Chem 1999;

37:2207–18.

[34] Kawakatsu T, Kawasaki K, Furusaka M, Okabayashi H, Kanaya

T. Late stage dynamics of phase separation processes of binary

mixtures containing surfactants. J Chem Phys 1993;99:8200–17.

[35] Kawakatsu T, Kawasaki K, Furusaka M, Okabayashi H, Kanaya

T. Phase separation dynamics in the presence of surfactant

molecules with an asymmetric shape. J Chem Phys 1995;102:

2247–53.

[36] Song SW, Torkelson JM. Coarsening effects on the formation of

microporous membranes produced via thermally induced phase

separation of polystyrene–cyclohexanol solutions. J Membr Sci

1995;98:209–22.

[37] Kim GE, Kim HD, Lee DS. Effect of added Pluronics on

fabrication of poly(l-lactic acid) scaffold via thermally induced

phase separation. Polymer (Korea) 2002;26:821–8.