Mixing a Sol and a Precipitate of Block Copolymers withDifferent Block Ratios Leads to an Injectable Hydrogel

Lin Yu, Zheng Zhang, Huan Zhang, and Jiandong Ding*

Key Laboratory of Molecular Engineering of Polymers of Ministry of Education, Department ofMacromolecular Science, Laboratory of Advanced Materials, Fudan University, Shanghai 200433, China

Received February 3, 2009; Revised Manuscript Received March 26, 2009

A facile method to obtain a thermoreversible physical hydrogel was found by simply mixing an aqueous sol ofa block copolymer with a precipitate of a similar copolymer but with a different block ratio. Two ABA-typetriblock copolymers poly(D,L-lactic acid-co-glycolic acid)-b-poly(ethylene glycol)-b-poly(D,L-lactic acid-co-glycolicacid) (PLGA-PEG-PLGA) were synthesized. One sample in water was a sol in a broad temperature region, whilethe other in water was just a precipitate. The mixture of these two samples with a certain mix ratio underwent,however, a sol-to-gel-to-precipitate transition upon an increase of temperature. A dramatic tuning of the sol-geltransition temperature was conveniently achieved by merely varying mix ratio, even in the case of a similarmolecular weight. Our study indicates that the balance of hydrophobicity and hydrophilicity within this sort ofamphiphilic copolymers is critical to the inverse thermal gelation in water resulting from aggregation of micelles.The availability of encapsulation and sustained release of lysozyme, a model protein by the thermogelling systemswas confirmed. This “mix” method provides a very convenient approach to design injectable thermogellingbiomaterials with a broad adjustable window, and the novel copolymer mixture platform is potentially used indrug delivery and other biomedical applications.

Introduction

Over the past decade, much work has been done on stimulus-responsive polymers and their applications in biochemistry,biomedicine, and related fields.1-8 Especially, stimuli-sensitivegelling polymers9-15 or in situ forming hydrogels16-20 haveattracted wide attention as potential injectable implant systemswith minimal invasiveness. Such a system enables drugs or cellsto be simply mixed with the aqueous copolymer solution at lowtemperatures and then form a semisolid matrix after injectioninto the body at a target site. Triblock copolymers poly(ethyleneglycol-b-propylene glycol-b-ethylene glycol) (PEG-PPG-PEG),called Pluronic (BASF) or Poloxamer (ICI), have been studiedby many researchers as an inverse thermogelling polymer.21,22

However, the drawbacks of nonbiodegradability and rapiderosion of the gel after administration limit their utility inbiomedical applications to a certain extent. Considerable effortshave been devoted to incorporating biodegradable segments intopolymers aiming at designing biodegradable thermogellingcopolymers. The block copolymers composed of PEG/PLGA,23,24

PEG/poly(caprolactone),25,26 PEG/poly(propylene fumarate),27

PEG/poly(propylene glycol)/polyester,12,28 PEG/peptide,29 chi-tosan/glycerolphosphate,30 and poly(phosphazenes)31 are typicalthermosensitive biodegradable copolymers exhibiting sol-geltransitions in water with increase of temperature. Amongthermogelling polymers, PEG/polyester copolymer hydrogelshave been successfully applied in drug delivery, cell therapy,tissue regeneration, and wound healing32-34 due to theirbiocompatibility and long persistence in the gel form in vivo.

Recently, many researches focused on exploiting novelthermogelling copolymers by varying their compositions orstructures.35-40 Vermonden et al. developed thermosensitivehydrogels based on poly(N-(2-hydroxypropyl) methacrylamide

lactate)-b-poly(ethylene glycol)-b-poly(N-(2-hydroxypropyl) meth-acrylamide lactate) (PHPMAmlac-PEG-PHPMAmlac) triblockcopolymers.36 The copolymers can become soluble in physi-ological conditions following the hydrolyzation of lactate groupswhich linked with the HPMAm monomers. However, thesoluble copolymers are difficult to be further degraded andcleared up in the body due to their high molecular weights(MWs). Madsen et al. exploited thermogelling poly(2-hydroxy-propyl methacrylate)-b-poly(2-(methacryloyloxy) ethyl phos-phorylcholine)-b-poly(2- hydroxypropyl methacrylate) (PHPMA-PMPC-PHPMA) triblock copolymers.38 The degradation comesfrom the cleavage of a disulfide bond in the center of the middleblock, and the degradation products are two diblock copolymers.While the composition is rather new, the degradation rate isnot easy to be controlled in medical applications and thedegraded products are hard to be further degraded. Jeong et al.reported that thermogelling PEG-g-PLGA and PLGA-g-PEGgraft copolymers can be promising carriers for protein and celldelivery.35 While their molecular design leads to synthesiscomplication, they extended the architecture of thermogellablecopolymers and confirmed their potential medical applicationsbased on an animal model. Hao and co-workers describedthermogelling poly(ε-caprolactone-co- D,L-lactide)-b-poly(eth-ylene glycol)-b-poly(ε-caprolactone-co-D,L-lactide) (PCLA- PEG-PCLA) triblock copolymers.37 Their degradable products showedrelatively low acidity while the degradation was relatively long.Ohya and co-workers synthesized partially cholesterol-substi-tuted 8-arm PEG-b-PLLA copolymers.39 The main advantageof this injectable hydrogel is its high mechanical strength toact as a cellular scaffold. In our preceding researches, asurprising end-group effect on macroscopic physical gelationof PLGA-PEG-PLGA triblock copolymer aqueous solutions hasbeen found,40 which revealed that an end-capping approachcould be employed to obtain novel thermosensitive hydrogels.The inverse thermogelling in all of these amphiphilic block

* Corresponding author. Tel.: 086-021-65643506. Fax: 086-021-65640293. E-mail: jdding1@fudan.edu.cn.

Biomacromolecules 2009, 10, 1547–1553 1547

10.1021/bm900145g CCC: $40.75 2009 American Chemical SocietyPublished on Web 04/22/2009

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g

copolymer systems originates from a subtle balance of hydro-philicity and hydrophobicity. However, the feasible compositionis very limited, which restricts its application to a large extent.For instance, only a narrow regime of MW and composition ofPEG/PLGA block copolymers could be selected to prepare athermogelling material, because a bit different MW or blockratio might lead to just a sol or a precipitate throughout thebiologically important temperature range.41 Even though PEG/PLGA block copolymers with appropriate compositions exhibitthermoreversible transition behaviors, the adjustable compositionwindow is narrow. Based on the central PEG block with MWof 1000, a series of PLGA-PEG-PLGA triblock copolymerswere synthesized in our previous work and the sol-gel transitiontemperature increased from about 10 to 25 °C by varying thePEG/PLGA ratios from 1/2.8 to 1/2.1.42 Similar reports couldalso be seen in the literature by other groups, for instance, Singhet al.43 The gel window is very important, because a samplewith a low sol-gel transition temperature is difficult to handleand inject, whereas a sample with a sol-gel transition temper-ature higher than the body temperature cannot be gelled afterinjection. Different block ratios as synthesized in this papermight even lead to loss of thermogelling ability.

In this study, we put forward a convenient approach tobroaden the gelling window of the corresponding polymers byfinding that a simple mixing of a sol and a precipitate mightlead to a reversible sol-gel transition and the gelation temper-ature in a biologically important temperature region is easilyadjusted by mix ratio. While these sols or precipitates aloneare unsuitable for encapsulation of drugs or cells, their mixturesmight, if exhibiting sol-gel transition in water and the gellingtemperature is between the body temperature and the refrigeratortemperature, be a promising candidate for injectable deliverysystems because drugs or cells could be dispersed in sols andentrapped after injection. Meanwhile, such a chemical-reaction-free approach is very meaningful for medical applications,because any “minor” chemical modification must lead to tediousefforts in warrant of an implantable material. Moreover, thisfinding is also helpful for elucidating the underlying mechanismof the physical gelation of amphiphilic block copolymers inwater, which remains as a challenging fundamental problem inthis field. At last, lysozyme was used as a model protein toexamine whether or not the thermogelling mixture systems couldencapsulate and deliver biological substance such as proteinsin a biologically active form for a long duration.

Experimental Section

Materials. Poly(ethylene glycol) (PEG; MW 1000 and 1500) andstannous 2-ethylhexanoate (stannous octoate, 95%) were purchased fromSigma. DL-Lactide (LA) and glycolide (GA) were acquired from Puracand used as received. Lysozyme and Micrococcus lysodeikticus (M.luteus) were purchased from Sigma. All other chemicals were of reagentgrade and used as received.

Synthesis of PLGA-PEG-PLGA. Triblock copolymers PLGA-PEG-PLGA were prepared following typical ring-opening polymerizationof lactide and glycolide using PEG as initiator. The detailed procedurehas been described by our previous papers.34,42 In this study, usingtwo PEG blocks (MW 1500 and 1000) and a given LA/GA ratio (4/1),a couple of PLGA-PEG-PLGA triblock copolymers with the PEG/PLGA ratios of 1/1.7 and 1/3.1 were synthesized and studied.

NMR Measurements. 1H NMR (in CDCl3) measurements wereperformed on a 500 MHz 1H NMR spectrometer (Bruker, DMX500spectrometer) to confirm the chemical structure and composition ofthe copolymers.44 13C NMR was also performed to observe the spectralchange of copolymer (20 wt % in D2O) at varied temperatures. Thesolution temperature was equilibrated for 20 min before measurements.

Gel Permeation Chromatography Characterization. Molecularweights and their distributions of copolymers were determined by gelpermeation chromatography (GPC, Agilent1100) equipped with adifferential refractometer as detector. GPC measurements were per-formed at 35 °C in THF as eluent at a flow rate of 1.0 mL/min. TheMWs were calibrated with polystyrene standards.

Determination of Sol-Gel Transition. The test tube invertingmethod45 was used to determine sol-gel transition temperatures. Eachsample with a given concentration was prepared by dissolving thepolymer in distilled water in a 2 mL vial and stored at 4 °C. The vialswith 0.5 mL polymer solutions were immersed in a water bath andallowed to reach equilibrium. The sample was regarded as a “gel” inthe case of no visual flow within 30 s by inverting the vial with atemperature increment of 1 °C per step.

Rheological Analysis. The sol-gel transition of the copolymeraqueous solutions were also investigated on a strain-controlled rhe-ometer (ARES Rheolometric Scientific) using a Couette cell (Couettediameter, 34 mm; bob diameter, 32 mm; bob height, 33.3 mm; bobgap, 2 mm). Cold polymer solutions were transferred into the Couettecell and carefully overlaid with a thin layer of low-viscosity siliconeoil to minimize solvent evaporation. During temperature sweepexperiments, appropriate strain amplitude was set according to pre-liminary tests to get both the linearity of viscoelasticity and sufficienttorque for data collection. Temperature was played with an accuracyof (0.05 °C by an environment controller (Neslab, RTε-130). Theheating rate was set at 0.5 °C/min, while the angular frequency ω wasset at 10 rad/s.

Determination of Critical Micellization Concentration (CMC). TheCMC values were determined by the solubilization method of ahydrophobic dye, 1,6-diphenyl-1,3,5-hexatriene (DPH) at 25 °C. Thehydrophobic dye solution in methanol (50 µL at 0.4 mM) was injectedinto an aqueous polymer solution (5 mL) with concentrations rangingfrom 0.001 to 0.2 wt % and equilibrated overnight at 4 °C. A UV-visspectrophotometer was used to obtain UV-vis spectra in the range of280-420 nm. The CMC value was determined by the plot of thedifference in absorbance at 378 and 400 nm (A378-A400) versuslogarithmic concentration.

Dynamic Light Scattering. Dynamic light scattering (DLS) wasperformed with a light scattering spectrophotometer (Autosizer 4700,Malvern) using a vertically polarized incident beam at 532 nm suppliedby an argon ion laser. The measurements were made at the scatteringangle of 90°. Before each measurement, the sample was treated througha 0.45 µm filter (Millipore) to remove dust. The hydrodynamic radiusof particles was calculated following the Stokes-Einstein equation.The intensity-intensity time correction function was analyzed by theCONTIN method.

In Vivo Gel Formation. Sprague-Dawley (SD) rats were anes-thetized by inhalation of dimethyl ether. Briefly, rats were enclosed ina 1 L beaker and a tampon saturated by dimethyl ether was added.The rats continuously inhaled the gas of dimethyl ether until anesthe-tized. Then, 0.5 mL of a Mixture-1 aqueous solution (25 wt %) wassubcutaneously injected into dorsal areas of rats with a 23-gauge needle.The animal experiments adhere to the “Principles of Laboratory AnimalCare” (NIH publication #85-23, revised 1985). Photographs were takenin surgery at 48 h after the subcutaneous injection into the rats.

In Vitro Release of Lysozyme. The thermogelling mixture wasdissolved in water at room temperature to make a 25 wt % solution.Lysozyme powder was loaded to a concentration of 20 mg/mL. Then1.0 mL of formation was placed into a test tube and incubated at 37°C. After 10 min, 8 mL of phosphate buffer saline (PBS) solution (pH7.4) containing NaN3 (0.025%, w/v) was added to the test tube as releasemedium. The tube was shaken at 50 rpm for the entire period of study.At sampling times, the release medium was withdrawn and replacedimmediately with fresh buffer to keep the sink condition. The amountof lysozyme in the release medium was determined by UV-vis assayat 280 nm associated with the characteristic absorption peak of lysozymein PBS solutions. The UV absorbance of the release media of drug-

1548 Biomacromolecules, Vol. 10, No. 6, 2009 Yu et al.

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g

free hydrogels at each associated degradation time was taken as blankcontrol. The amount of lysozyme in the released sample was obtainedfrom the standard curve.

Examinations of Bioactivity of Lysozyme by Enzyme ActivityAssay. The activity of lysozyme was detected by the M. luteus assay.46

Briefly, 10 mg of M. luteus stock suspension was put in phosphatebuffer (30 mL, pH 6.24) and vigorously shaken, and then a minoraddition of the bacterium or the buffer was used to adjust A450, theoptical density at 450 nm of the suspension about 1.30. A 2.5 mL ofthis M. luteus suspension was transferred into a spectrophotometer celland a portion of a diluted lysozyme sample with a known initialconcentration of enzyme was added. The rate of decrease of A450 wasmonitored by a UV spectrophotometer during the total incubation period(4 min) at 25 °C. An appropriate dilution was chosen to achieve alinear relation of absorbance change versus time, and usually theunderlying slope S450 is between 0.02 ∼ 0.04 min-1. One unit of enzymeactivity (EU) is defined as 0.001/min of absorbance change under theconditions employed.47 Set the weight concentration of lysozyme afterdilution as Clyso and the volume as V (0.5 mL in our experiments), thespecific activity of lysozyme was calculated by

EU/mg ) S450/(0.001 × V × Clyso) (1)

Unpaired Student t-test was employed to estimate significantdifference between two groups with p < 0.05 as criterion.

Results and Discussion

Synthesis and Characterization of PLGA-PEG-PLGATriblock Copolymers. PLGA-PEG-PLGA triblock copolymerswere synthesized via the ring-opening polymerization ofDL-lactide and glycolide using PEG with the terminal hydroxylgroups as initiator in the presence of stannous octoate as catalyst.The PLGA-PEG-PLGA triblock copolymers used in thisresearch are listed in Table 1. 1H NMR spectrum of PLGA-PEG-PLGA triblock copolymers along with their chemicalstructure is shown in Figure 1. The peaks at 5.20 and 1.55 ppmwere assigned to a methine proton and methyl protons of theLA units, respectively. The peaks at 4.80 and 3.60 ppm wereattributed to methylene protons of the GA units and PEG,

respectively. The methylene proton of PEG neighboring thePLGA block appeared at 4.31 ppm. Following the previousreport,44 the peaks at 4.80, 3.60, and 1.55 ppm were used tocalculate the number average MW of the PLGA-PEG-PLGAtriblock copolymer. Both PLGA-PEG-PLGA triblock copoly-mers we synthesized showed similar total MWs. GPC wasfurther performed to determine MW and its distribution. TheGPC traces of both copolymers were unimodal with a polydis-perse index (Mw/Mn) less than 1.27 (Figure 2), indicating thatpolydispersity is sufficiently low to study their physical proper-ties. Interestingly, the GPC trace of the mixture of Copolymer-1and Copolymer-2 (the mix ratio: 1/1, Mixture-1) was stillunimodal due to very close MWs of the two polymers.

Temperature-Induced Sol-Gel Transition of the Copoly-mer Mixtures. Although Copolymer-1 and Copolymer-2 usedin this study showed similar total MWs, they exhibited extremelydifferent aqueous behaviors: Copolymer-1 was soluble in waterand no sol-gel transition was observed following the increaseof temperature, while Copolymer-2 was too hydrophobic to besoluble in water in a biologically important temperature region(from 4 to 50 °C). However, their mixtures with appropriateratios were found to be well-dissolved in water and exhibitinverse thermal gelation upon an increase of temperature. Thisinteresting finding incidates that the hydrophilic Copolymer-1can enhance the solubility of the hydrophobic Copolymer-2 inwater. The physical mixing leads to a temperature-inducedsol-gel transition due to the elaborate balance of hydrophobic-ity/hydrophilicity.

Figure 3 shows the phase diagrams of the mixture aqueoussolutions, which underwent sol-gel-sol (suspension) transitionupon heat stimulus. The state “sol (suspension)” indicates thata gel is first disrupted into a flowable suspension and eventuallypolymers are precipitated. Although the gel-sol (suspension)

Table 1. PLGA-PEG-PLGA Triblock Copolymers in this Study

sample Mnd

PEGMn

dLA/GAd

(mol/mol) Mne Mw/Mn

e

Copolymer-1a 1255-1500-1255 1500 4.0 5510 1.16Copolymer-2b 1571-1000-1571 1000 4.0 5850 1.27Mixture-1c 4.0 5630 1.21

a Dissolved in water as a sol in a broad temperature region. b Insolublein water. c A mixture of Copolymer-1 and Copolymer-2 with weight mixratio 1/1 exhibiting a thermoreversible sol-gel transition upon raisingtemperature. d The number-averaged MW, Mn of the central block PEGwas provided by Aldrich. The molar ratio of lactide/glycolide (LA/GA) andMn of each PLGA block were calculated by 1H NMR. e Measured via GPC.

Figure 1. 1H NMR spectrum of PLGA-PEG-PLGA triblock copolymersin CDCl3.

Figure 2. Gel permeation chromatograms of Copolymer-1, Copolymer-2, and Mixture-1.

Figure 3. Phase diagrams of mixture aqueous solutions with theindicated weight mix ratios of the two PLGA-PEG-PLGA blockcopolymers. The transition temperature was measured by the testtube inverting method.

Mixing a Sol with a Precipitate Leads to a Hydrogel Biomacromolecules, Vol. 10, No. 6, 2009 1549

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g

transition temperatures were rather sensitive to polymer con-centration, the sol-gel transition temperatures seemed relativelyless sensitive to concentration in the gel window, which isdifferent from PEG/PCL system.25

Sol-gel transition temperature, gel window width, and criticalgelation concentration (CGC) are the key parameters whichshould be taken into consideration in designing an in situ gellingbiomaterial, and all of the information could be known fromFigure 3. The sol-gel transition temperature increased morethan 10 °C and the CGC decreased from 14 to 7 wt % as themix ratio of Copolymer-1/Copolymer-2 decreased from 2/1 to1/2; accordingly, the height of the gel window significantlyincreased as well. Our studies indicate that the sol-gel transitiontemperature could be well-tuned in a physiologically importanttemperature range (25-45 °C) by merely changing the mixratios of the two copolymers, which were, albeit of similar MWs,originally a sol and a precipitate.

The viscosity and storage modulus of samples with theindicated mix ratios as a function of temperature are shown inFigure 4. Although the viscosity (η) of Copolymer-1 increasedupon heating, its maximum value was only a bit more than 1Pa ·S in the whole temperature range measured. So, the changein viscosity of Copolymer-1 was not significant enough to forma macroscopic hydrogel upon the increase of temperature. Incontrast, while the viscosities of the mixtures were all less than0.01 Pa ·S at low temperatures, the changes in the viscositiesof the mixtures spanned more than 3 orders of magnitude withincrease of temperature, indicating a sol-gel transition. Thesol-gel transition temperature varied depending on mix ratio.Following the decrease of the ratio of Copolymer-1/Copolymer-2, the maximum value of viscosity increased while the sol-geltransition temperature decreased. The storage modulus changedin a similar way of viscosity.

The viscosity in the sol state was sufficiently low for an easyformulation of polymer. The gel phase of Mixture-1 observedat about 37 °C further confirmed that it could serve as apromising injectable biomaterial. A subcutaneous injection ofMixture-1 into SD rats was thus performed, and we found thatinjection could be well-performed even using a very thin needleof 23-gauge. The rat experiments confirmed the availability ofin situ gel formation after injection, as shown in Figure 5.

Micellization and the Proposed Hierarchical Structureof the Gel. The sol-gel transition might be attributed toaggregation of micelles instead of a “straight-forward” gelformation from molecules. Amphiphilic block copolymers are,as usual, self-assembled into core-corona-like micelles in water.The formation of the core-corona structure in water could bedetermined by the hydrophobic dye method in the presence ofpolymer. The hydrophobic dye partitioned into a hydrophobicdomain such as the core of a micelle showed stronger absor-bance at 337, 356, and 378 nm (Figure 6a). The absorbancedifference between 378 and 400 nm as a function of concentra-tion was used to determine the CMC (Figure 6b). Obviously,hydrophilic PEG homopolymers could not form micelles inwater as the PEG concentration increased. Both Copolymer-1and Mixture-1 in water were self-assembled into micelles. Figure6b shows that the CMC of Copolymer-1 at room temperatureis 4.1 × 10-4 g/mL and that of Mixture-1 is 3.1 × 10-4 g/mL.The lowering of CMC of Mixture-1 compared with that ofCopolymer-1 might be due to addition of relatively hydrophobicCopolymer-2 into relatively hydrophilic Copolymer-1.

Figure 7 shows DLS results under different mix ratios inrelatively low-concentrated aqueous solutions (0.5 wt %) at 20°C. The micelle radius of Copolymer-1 was about 10 nm at 20°C. Upon addition of otherwise insoluble Copolymer-2 intoCopolymer-1, the micelle radius and size distribution of themixtures increased. It seems that an increase of the globalhydrophobicity of block copolymer systems could enlargemicelles in our samples.

The change of micelle radius as a function of temperature at0.5 wt % is shown in Figure 8. In a temperature range of 10 ∼25 °C, the average radius of a micelle of Copolymer-1 wasalmost constant at 10-11 nm. As the temperature increased upto 30 and 35 °C, the micelle radius and size distribution slightlyincreased. At a higher temperature, the micelle radius increasedto 25 nm. At the same concentration (0.5 wt %), the averageradius of a micelle of Mixture-1 located at 11-12 nm over 10∼ 20 °C; as the temperature increased, the significant increasesin the micelle radius and size distribution were observed above

Figure 4. Viscosity η (a) and storage modulus G′ (b) of samples withthe indicated weight mix ratios in aqueous solutions (20 wt %) as afunction of temperature. The ratio “1/0” refers actually to theCopolymer-1 aqueous sol. Heating rates: 0.5 °C/min; oscillatoryfrequency: 10 rad/s.

Figure 5. In situ gel formation of Mixture-1 aqueous solution (25 wt%, 0.5 mL) in the back of an SD rat. The photograph of in vivo gelformation was taken at 48 h after the subcutaneous injection to rats.The arrow indicates a gel.

1550 Biomacromolecules, Vol. 10, No. 6, 2009 Yu et al.

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g

25 ∼ 40 °C; at 45 °C, the most of micelles were aggregated to180 nm in radius and there was a small population of micelleswith 50 nm. These results indicate that the ability of aggregationof Mixture-1 is greatly stronger than that of Copolymer-1.

Using 13C NMR of Mixture-1 in D2O (20 wt %), moreinformation of the local environments of blocks of amphiphilicpolymers could be obtained (Figure 9). Below the sol-geltransition temperature (20 °C), the sharp PEG peak and collapsed-CH3 peak of PLGA indicate that the sample form micelles inwater with PEG as corona and PLGA as core. At highertemperatures (33, 37, and 43 °C) where the copolymer formeda hydrogel, the PEG peak was a little broadened and the -CH3

peak of PLGA was just a bit increased. Such nonabrupt changesof the peaks reflect that the micelle structure was maintainedduring the sol-gel transition. At 55 °C the PEG peak furtherbroadened and the height of -CH3 peak of PLGA was pro-nouncedly increased. The NMR measurement is consistent with

Figure 6. (a) UV-vis spectra of Mixture-1 aqueous solutions contain-ing the hydrophobic dye (DPH). DPH concentration was fixed at 4µM and polymer concentration varied. For clarity, just four weightconcentrations are shown in (a). (b) CMC determination by extrapola-tion of the difference of absorbance at 378 and 400 nm. Themeasurements were done at room temperature.

Figure 7. Hydrodynamic radius of micelles in aqueous solutions (0.5wt %) measured by DLS at 20 °C. The weight mix ratios ofCopolymer-1 and Copolymer-2 are indicated.

Figure 8. Distribution of hydrodynamic radius Rh of micelles of (a)Copolymer-1 and (b) Mixture-1 with equal weight of Copolymer-1 anda precipitate Copolymer-2 in water at indicated temperatures. Polymerconcentration was 0.5 wt % in the DLS measurements.

Figure 9. 13C NMR spectra of Mixture-1 in D2O (20 wt %) at indicatedtemperatures with respect to typical condensed states, namely, 20°C (sol state), 33 °C (near sol-gel transition), 37 °C (gel state), 43°C (gel state), and 55 °C (macro-phase-separated state). Beforemeasurements, the solution was equilibrated for 20 min at the settingtemperatures. The peaks close to 70 and 16 ppm come from theethylene group of PEG and the methyl group of PLGA, respectively.The PLGA signal is very weak if these blocks are constrained in ahydrophobic micellar core.

Mixing a Sol with a Precipitate Leads to a Hydrogel Biomacromolecules, Vol. 10, No. 6, 2009 1551

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g

the global observation of a precipitate. In this case, thecore-corona structure of the polymers was destroyed due tophase mixing between PEG and PLGA, and the mixing led toan increase of PLGA signal and a broadening of the PEG peak.

The physical gel was formed hierarchically via micellarnanoparticles. Copolymer-1 can only form micelles for theabsence of physical gelation upon heating (only sols) (Figure10a), while Copolymer-2, a globally more-hydrophobic am-phiphilic block copolymer takes on macroscopic phase separa-tion in water (Figure 10b). The mixture might be a sol-likesuspension at low temperatures and the radius of micelles mightbe increased due to the enhanced hydrophobicity (Figure 10c).13C NMR experiments (Figure 9) illustrate that the micellestructure maintained in the gel state. This evidence together withthe micellar aggregation behavior (Figure 8) implies that thesol-gel transition might be attributed to the formation of apercolated micelle network (Figure 10d), which is equivalentto an infinite micellar aggregate. The availability of micellesas building block of a thermoreversible physical gel has alsobeen supported by thermogelling of core-cross-linked micellesof another amphiphilic block copolymers in our recent work.48

The driving force of the sol-gel transition of PEG/polyesterblock copolymers in water comes from nothing but thehydrophobic interaction. The “inverse” gelling arises from themore significant effect of the hydrophobic interaction uponheating. The hydrophobic interaction originates from an entropyeffect, and the influence of any entropy effect to free energymust be enhanced with increase of temperature. While the detailof the underlying thermogelling mechanism is still open, ourstudy illustrates that the balance of hydrophobic and hydrophilicblocks might play a key role to exhibit an inverse thermalgelation in water for this kind of copolymers.

In Vitro Release of Lysozyme. The inverse thermogellingin water with a tunable gelling temperature affords a candidateof an injectable medical hydrogel. Taking lysozyme as a modernprotein, we found that the mixture of a biological substancewith our biomaterial was, at low temperatures, say 4 °C, of lowviscosity for injection and an injection at room temperature orbody temperature after taking the sample from the refrigeratorwas available.

We further examined the in vitro release behaviours of themodel protein lysozyme from the mixture hydrogels (Figure 11).Not only sustained release of lysozyme but also low initial burstrelease (about 8% at the first three days) was found with bothformulations with varied mix ratios of block copolymers. Therelease of proteins from these two hydrogels exhibited differentprofiles, demonstrating that mix ratios could also be taken asan adjustable factor of release profiles. Figure 11 also illustratesthat, under an appropriate condition, an almost zero-order releasecould be achieved up to 1.5 months. In comparison to the highburst release (16-42%) and short duration (15 days, up to70-90% lysozyme release) from single-component-polymerPLGA-PEG-PLGA hydrogels,43 the mixture hydrogel candecrease the burst release to a low level (about 8% at the firstthree days) and increase the sustained release to a long duration(55 days, up to 65-80% lysozyme release). The difference maycome from that of the compositions between these two hydrogelsand, consequently, hydrogel properties such as modulus anddegradation rate (not necessarily a result of mixture versussingle-component polymer).

Table 2 provides data of specific enzyme activities oflysozyme at the 45th day during in vitro release. Compared tofresh lysozymes, just a bit lowering of enzyme activity wasfound after encapsulated into and released out of our hydrogelsafter 1.5 months at 37 °C. In contrast, naked enzymesexperiencing the same process exhibited a significant decreaseof activity. Therefore, our hydrogels not only did not damageactive proteins, but also protected them well from denaturing.

Conclusions

It is the first report on thermogelling systems prepared bymixing synthetic polymers with nonthermogelling molecular

Figure 10. Schematic presentation of the mixing approach andunderlying self-assembly. Upper images show the sol state ofCopolymer-1 and the formation of micelle in water (a) and theprecipitate state of Copolymer-2 in water (b) in a biologically importanttemperature region (from 4 to 50 °C). Bottom images show thesol-gel transition of Mixture-1 aqueous solution with increasingtemperature (c and d). The insets are optical images of the corre-sponding polymeric aqueous systems (20 wt %).

Figure 11. Cumulative release of lysozyme from mixture hydrogelswith indicated mix ratios of Copolymer-1 and Copolymer-2. Thepolymer mixture concentration was 25 wt % and the drug loadingwas 20 mg/mL. The release was performed in phosphate buffer (pH7.4) at 37 °C. Error bars represent the standard deviation (n ) 3).

Table 2. Specific Enzyme Activity of Lysozymes at the 45th Dayin Releasea

formulations specific enzyme activity (EU/mg) × 104

fresh lysozyme 5.84 ( 0.21control 0.83 ( 0.02mix ratio: 1/1 5.28 ( 0.35mix ratio: 1/2 5.41 ( 0.12a Note: The drug loading was 20 mg/mL. The case of control refers to

lysozymes experiencing the same experimental process (20 mg in releasebuffer, at 37 °C in oscillating water for 45 days). n ) 3. Significantdifferences were found between the control group and any one of otherthree groups. No significant differences were found among other threegroups.

1552 Biomacromolecules, Vol. 10, No. 6, 2009 Yu et al.

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g

parameters. The sol-gel transition temperature could be tailoredto a large extent merely by adjusting mix ratios. Micellaraggregation was thought to be involved in the sol-gel transition.The system is injectable, and the availability of physical gellingwas confirmed both in vitro and in vivo. The in vitro testsillustrate that proteins could be well-encapsulated and protectedfrom denaturing. The release rate could also be tuned by themix ratios of copolymer mixtures, and an almost zero-ordersustained release of lysozymes was achieved up to 50 days. Asa result, our study affords a facile approach to obtain, basedupon known polymers, novel injectable biomaterials potentiallyapplicable in drug sustained release, tissue engineering, and soon. We expect that this finding might also trigger extension ofgellable systems to other self-assembly amphiphiles beyondPEG/PLGA block copolymers.

Acknowledgment. The group was supported by ChineseMinistry of Science and Technology (973 Project Nos.2009CB930000 and 2005CB522700), NSF of China (Grant Nos.50533010 and 20774020), Chinese Ministry of Education (KeyGrant No. 305004), Science and Technology DevelopingFoundation of Shanghai (Grant Nos. 07JC14005 and 074319117),and Shanghai Education Committee (Project No. B112).

References and Notes(1) Hoffman, A. S. AdV. Drug DeliVery ReV. 2002, 54, 3–12.(2) Zhang, Y.; Zhu, W.; Wang, B. B.; Ding, J. D. J. Controlled Release

2005, 105, 260–268.(3) Ye, X. D.; Lu, Y. J.; Liu, S. L.; Zhang, G. Z.; Wu, C. Langmuir 2007,

23, 10366–10371.(4) Wu, D. Q.; Sun, Y. X.; Xu, X. D.; Cheng, S. X.; Zhang, X. Z.; Zhuo,

R. X. Biomacromolecules 2008, 9, 1155–1162.(5) David, G.; Simionescu, B. C.; Albertsson, A. C. Biomacromolecules

2008, 9, 1678–1683.(6) Yu, L.; Ding, J. D. Chem. Soc. ReV. 2008, 37, 1473–1481.(7) Matsumoto, S.; Christie, R. J.; Nishiyama, N.; Miyata, K.; Ishii, A.;

Oba, M.; Koyama, H.; Yamasaki, Y.; Kataoka, K. Biomacromolecules2009, 10, 119–127.

(8) Yang, C. T.; Wang, Y. L.; Yu, S.; Chang, Y. C. I. Biomacromolecules2009, 10, 58–65.

(9) Jeong, B.; Bae, Y. H.; Lee, D. S.; Kim, S. W. Nature 1997, 388, 860–862.

(10) Bhattarai, N.; Ramay, H. R.; Gunn, J.; Matsen, F. A.; Zhang, M. Q.J. Controlled Release 2005, 103, 609–624.

(11) Shim, W. S.; Yoo, J. S.; Bae, Y. H.; Lee, D. S. Biomacromolecules2005, 6, 2930–2934.

(12) Loh, X. J.; Tan, Y. X.; Li, Z. Y.; Teo, L. S.; Goh, S. H.; Li, J.Biomaterials 2008, 29, 2164–2172.

(13) Huynh, D. P.; Nguyen, M. K.; Pi, B. S.; Kim, M. S.; Chae, S. Y.;Lee, K. C.; Kim, B. S.; Kim, S. W.; Lee, D. S. Biomaterials 2008,29, 2527–2534.

(14) Zhao, Y.; Yokoi, H.; Tanaka, M.; Kinoshita, T.; Tan, T. W.Biomacromolecules 2008, 9, 1511–1518.

(15) Kirkland, S. E.; Hensarling, R. M.; McConaughy, S. D.; Guo, Y.;Jarrett, W. L.; McCormick, C. L. Biomacromolecules 2008, 9, 481–486.

(16) Yeo, Y.; Bellas, E.; Highley, C. B.; Langer, R.; Kohane, D. S.Biomaterials 2007, 28, 3704–3713.

(17) Hiemstra, C.; Zhou, W.; Zhong, Z. Y.; Wouters, M.; Feijen, J. J. Am.Chem. Soc. 2007, 129, 9918–9926.

(18) Hiemstra, C.; Zhong, Z. Y.; van Steenbergen, M. J.; Hennink, W. E.;Jan, F. J. J. Controlled Release 2007, 122, 71–78.

(19) Rhodes, N. P. Biomaterials 2007, 28, 5131–5136.(20) Matricardi, P.; Pontoriero, M.; Coviello, T.; Casadei, M. A.; Alhaique,

F. Biomacromolecules 2008, 9, 2014–2020.(21) Johnston, T. P.; Punjabi, M. A.; Froelich, C. J. Pharm. Res. 1992, 9,

425–434.(22) Vashi, A. V.; Keramidaris, E.; Abberton, K. M.; Morrison, W. A.;

Wilson, J. L.; O’Connor, A. J.; Cooper-White, J. J.; Thompson, E. W.Biomaterials 2008, 29, 573–579.

(23) Jeong, B.; Bae, Y. H.; Kim, S. W. Macromolecules 1999, 32, 7064–7069.

(24) Lee, D. S.; Shim, M. S.; Kim, S. W.; Lee, H.; Park, I.; Chang, T. Y.Macromol. Rapid Commun. 2001, 22, 587–592.

(25) Bae, S. J.; Joo, M. K.; Jeong, Y.; Kim, S. W.; Lee, W. K.; Sohn,Y. S.; Jeong, B. Macromolecules 2006, 39, 4873–4879.

(26) Hyun, H.; Kim, Y. H.; Song, I. B.; Lee, J. W.; Kim, M. S.; Khang,G.; Park, K.; Lee, H. B. Biomacromolecules 2007, 8, 1093–1100.

(27) Behravesh, E.; Shung, A. K.; Jo, S.; Mikos, A. G. Biomacromolecules2002, 3, 153–158.

(28) Loh, X. J.; Goh, S. H.; Li, J. Biomacromolecules 2007, 8, 585–593.(29) Choi, Y. Y.; Joo, M. K.; Sohn, Y. S.; Jeong, B. Soft Matter 2008, 4,

2383–2387.(30) Chenite, A.; Chaput, C.; Wang, D.; Combes, C.; Buschmann, M. D.;

Hoemann, C. D.; Leroux, J. C.; Atkinson, B. L.; Binette, F.; Selmani,A. Biomaterials 2000, 21, 2155–2161.

(31) Lee, B. H.; Song, S. C. Macromolecules 2004, 37, 4533–4537.(32) Zentner, G. M.; Rathi, R.; Shih, C.; McRea, J. C.; Seo, M. H.; Oh,

H.; Rhee, B. G.; Mestecky, J.; Moldoveanu, Z.; Morgan, M.; Weitman,S. J. Controlled Release 2001, 72, 203–215.

(33) Kim, M. S.; Kim, S. K.; Kim, S. H.; Hyun, H.; Khang, G.; Lee, H. B.Tissue Eng. 2006, 12, 2863–2873.

(34) Yu, L.; Chang, G. T.; Zhang, H.; Ding, J. D. Int. J. Pharm. 2008,348, 95–106.

(35) Jeong, B.; Lee, K. M.; Gutowska, A.; An, Y. H. H. Biomacromolecules2002, 3, 865–868.

(36) Vermonden, T.; Besseling, N. A. M.; van Steenbergen, M. J.; Hennink,W. E. Langmuir 2006, 22, 10180–10184.

(37) Jiang, Z. Q.; Deng, X. M.; Hao, J. Y. J. Polym. Sci., Part A: Polym.Chem. 2007, 45, 4091–4099.

(38) Madsen, J.; Armes, S. P.; Bertal, K.; Lomas, H.; MacNeil, S.; Lewis,A. L. Biomacromolecules 2008, 9, 2265–2275.

(39) Nagahama, K.; Ouchi, T.; Ohya, Y. AdV. Funct. Mater. 2008, 18,1220–1231.

(40) Yu, L.; Zhang, H.; Ding, J. D. Angew. Chem., Int. Ed. 2006, 45, 2232–2235.

(41) Shim, M. S.; Lee, H. T.; Shim, W. S.; Park, I.; Lee, H.; Chang, T.;Kim, S. W.; Lee, D. S. J. Biomed. Mater. Res. 2002, 61, 188–196.

(42) Yu, L.; Chang, G. T.; Zhang, H.; Ding, J. D. J. Polym. Sci., Part A:Polym. Chem. 2007, 45, 1122–1133.

(43) Chen, S. B.; Pieper, R.; Webster, D. C.; Singh, J. Int. J. Pharm. 2005,288, 207–218.

(44) Jeong, B.; Lee, D. S.; Shon, J. I.; Bae, Y. H.; Kim, S. W. J. Polym.Sci., Part A: Polym. Chem. 1999, 37, 751–760.

(45) Tanodekaew, S.; Godward, J.; Heatley, F.; Booth, C. Macromol. Chem.Phys. 1997, 198, 3385–3395.

(46) Shugar, D. Biochim. Biophys. Acta 1952, 8, 302–309.(47) Meng, F. T.; Ma, G. H.; Qiu, W.; Su, Z. G. J. Controlled Release

2003, 91, 407–416.(48) Yang, Z. G.; Ding, J. D. Macromol. Rapid Commun. 2008, 29,

751–756.

BM900145G

Mixing a Sol with a Precipitate Leads to a Hydrogel Biomacromolecules, Vol. 10, No. 6, 2009 1553

Dow

nloa

ded

by F

UD

AN

UN

IV o

n Ju

ly 6

, 200

9Pu

blis

hed

on A

pril

22, 2

009

on h

ttp://

pubs

.acs

.org

| do

i: 10

.102

1/bm

9001

45g