Polymer particles that switch shapein response to a stimulusJin-Wook Yoo and Samir Mitragotri1

Department of Chemical Engineering, University of California, Santa Barbara, CA 93106

Edited* by David A. Tirrell, California Institute of Technology, Pasadena, CA, and approved May 14, 2010 (received for review January 10, 2010)

Particle engineering for biomedical applications has unfolded theroles of attributes such as size, surface chemistry, and shape formodulating particle interactions with cells. Recently, dynamicmanipulation of such key properties has gained attention in viewof the need to precisely control particle interaction with cells. Withincreasing recognition of the pivotal role of particle shape in deter-mining their biomedical applications, we report on polymeric par-ticles that are able to switch their shape in real time in a stimulus-responsive manner. The shape-switching behavior was driven by asubtle balance between polymer viscosity and interfacial tension.The balance between the two forces wasmodulated by applicationof an external stimulus chosen from temperature, pH, or chemicaladditives. The dynamics of shape switch was precisely controlledover minutes to days under physiological conditions. Shape-switchingparticlesexhibitedunique interactionswithcells. Ellipticaldisk-shaped particles that are not phagocytosed by macrophageswere made to internalize through shape switch, demonstratingthe ability of shape-switchable particles in modulating interactionwith cells.

drug delivery ∣ carrier ∣ geometry ∣ nanotechnology ∣ phagocytosis

Interactions of polymeric particles with various cells, includingmacrophages, in the form of endocytosis and phagocytosis de-

termine the effectiveness of carriers used for drug delivery andmedical imaging (1, 2). The outcome of these interactions relieson optimal selection of key particle properties including surfacechemistry, size, and shape (3, 4). Accordingly, numerous studieshave reported on methods to synthesize materials with preciselyengineered functional attributes such as size, surface chemistry,mechanical properties, and shape to facilitate or mitigate inter-actions with various cell types (3–5).

For a given application, particle properties are optimizedthrough extensive experimentation, and a set of fixed valuesare then chosen for further development. In reality, however,the optimal values of parameters may vary with time dependingon the application. This variation has motivated the need to gaindynamic control over key particle properties so as to achieve aninteractive interface between the particles and the complex bio-logical milieu. For example, studies have reported on stimulus-responsive type control over the size of particles, and such par-ticles have been used for the triggered release of encapsulateddrugs (6–8). In another study, the surface chemistry of polymericmicelles has been controlled by using the environmental pH ofsolid tumors so as to enhance the cellular uptake and releaseof anticancer agents (9). Studies have also reported on achievingdynamic control of surface properties through the use of electricfields (10). Relatively little attention, however, has been devotedto switching particle’s shape.

Here we report poly(lactide-co-glycolide) (PLGA) particleswhose shape can be switched in real time from an elliptical diskto a sphere in response to a stimulus selected from temperature,pH, or a chemical. The time scale of switching was controlledover a wide range, from minutes to days, by appropriate selectionof polymer molecular weight, particle size, and the strength of thestimulus. We also demonstrate that these particles can modulatetheir interactions with macrophages through shape switch.

ResultsPLGA was selected as a polymer because of its biocompatibilityand proven record in polymer-based medical products (11, 12).Elliptical disk (ED) was chosen as a model particle shape, whichwas then switched to a sphere upon application of a stimulus in-cluding temperature, pH, and chemicals (Fig. 1A). Switching ofparticle shape was driven by a subtle balance between interfacialtension and polymer viscosity. Specifically, the interfacial tension(σ) between PLGA and the surrounding aqueous liquid drives therelaxation of a disk-shaped particle to an energetically favorablespherical shape (Fig. 1B and Movie S1); however, this relaxationis resisted by the high viscosity of PLGA (μ). The dynamics ofparticle relaxation can be described by solving the equation of mo-tion assuming that the polymer can be modeled as a Newtonianfluid, a reasonable assumption at temperatures above polymer’sglass transition temperature Tg (13–15). PLGA indeed exhibitsNewtonian behavior near Tg and at low shear rates(10−4–10−1 s−1), which is the relevant range of shear rates forshape switching reported here (Fig. S1). A quasi-steady Stokesequation has been numerically solved for interfacial tension-driven relaxation of ellipsoidal droplets of Newtonian fluids,and a scaling parameter describing time constant of shape change,τ, was given by (16)

τ ¼ LμLð1þ λÞ∕σ; [1]

where L is the characteristic particle size and λ ¼ μ∕μL, where μ isPLGA viscosity and μL is the viscosity of the surrounding solution,typically an aqueous medium with viscosity comparable to water.Because PLGA is highly viscous compared to water, μ ≫ μL, thusresulting in τ ∼ Lμ∕σ. By appropriately controlling the valuesofL,μ, andσ, wewere able to control thedynamics of shape changeover a few minutes (rapid change) to weeks (nearly staticshapes) (Fig. 1C).

The first demonstration of shape switching was obtained viatemperature-induced change of particle viscosity (μ) by increas-ing temperature above Tg, which in turn, was controlled throughpolymer molecular mass. We first measured the viscosity ofPLGA (ester end, molecular mass 29.8 kDa, TgðmidÞ ¼ 40 °C)as a function of temperature. These measurements indicatedthat PLGA viscosity at Tg and at low shear rates is about1.5 × 106 Pa · s, which is generally consistent with the literaturedata for other polymers (14). The viscosity of PLGA decreaseddramatically with temperature above Tg . The dependence of visc-osity on temperature was consistent with the Vogel–Fulcher–Tammann (VFT) equation, which is routinely used to describepolymer viscosities near Tg (Fig. S1) (17). EDs with variousaspect ratios (AR; the ratio of major axis to minor axis) were

Author contributions: J.-W.Y. and S.M. designed research; J.-W.Y. performed research;J.-W.Y. contributed new reagents/analytic tools; J.-W.Y. and S.M. analyzed data; andJ.-W.Y. and S.M. wrote the paper.

The authors declare no conflict of interest.

*This Direct Submission article had a prearranged editor.1To whom correspondence should be addressed. E-mail: samir@engineering.ucsb.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1000346107/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1000346107 PNAS ∣ June 22, 2010 ∣ vol. 107 ∣ no. 25 ∣ 11205–11210

ENGINEE

RING

Dow

nloa

ded

by g

uest

on

Mar

ch 1

3, 2

021

prepared from different types of PLGAs by using a previouslyreported method (Table S1) (18). For the temperature-inducedshape switch, EDs (AR ¼ 5.5) were fabricated by using PLGA-ester (PLGA with an ester [1-dodecanol] end group) of twomolecular masses (29.8 kDa, Fig. 2A, top panels, and52.7 kDa, Fig. 2A, bottom panels). At room temperature(∼25 °C) both particles maintained their ED shape for prolongedtimes (Fig. S2), indicating that the shape does not switch in aglassy state of an amorphous polymer. The temperature was thenraised abruptly to various levels. Detectable shape switch for EDsof molecular mass 29.8 kDa PLGA occurred once the tempera-ture reached 32 °C, which is the onset Tg as determined fromdifferential scanning calorimetry (Fig. S2). At a temperature of40 °C (mid-Tg), the shape was nearly completely switched inabout 4 h (Fig. 2B, Closed Squares). For EDs of molecular mass52.7 kDa, the particles gradually changed their shape to spheresover a period of 24 h upon increasing the temperature to theirmid-Tg (43 °C, Fig. 2B, Open Squares). Although the shapechange for both particles was studied at their respectivemid-Tg , the switching time increased with increasing molecularmass, because of the higher viscosity of high molecular masspolymer. The characteristic switch time (T1∕2, defined as the timerequired for a 2-fold reduction of AR) decreased rapidly with anincrease in the operating temperature, and for a 10 °C differentialbetween the operating temperature and mid-Tg , the switch wascompleted in less than 10 min (Fig. 2C and Fig. S2). Particle sizeis another important parameter to determine the shape-switchtime. A 2-fold increase or 10-fold decrease in the diameter ofthe initial sphere, compared to the baseline case of 1.6 μm, re-spectively, led to an around 3-fold increase and around 10-folddecrease in the switch time of EDs (Fig. 2A, bottom panels,

Fig. 2D and Fig. S3). These results also demonstrated that theshape-switching phenomenon also occurs in nanosized particles.

In order to validate the proposed mechanism for shape switch-ing, the switching time for PLGA (ester end, molecular mass29.8 kDa, TgðmidÞ ¼ 40 °C) was calculated by using a viscosityof 1.5 × 106 Pa · s at Tg, an interfacial tension of 5.4 mN∕m de-rived from contact angle measurements, and a particle diameterof 1.6 μm (Fig. S1). The estimated T1∕2 value under these con-ditions was in the range of 22–75 min, which is of the same orderof magnitude as the experimentally observed value of 38 min(Fig. 2B, Closed Squares).

The dependence of T1∕2 on temperature and particle size isalso consistent with the proposed mechanism of shape switch.For example, the combination of Eq. 1 and VFT theory suggeststhat the (Log [switching time]) should scale as (1∕temperature).Indeed, the data in Fig. 2C are consistent with this trend. A plotof log (T1∕2) with 1∕temperature yields a good linear fit withr2 ¼ 0.97. Similarly, Eq. 1 suggests that the switching time shouldscale linearly with the particle size. Indeed, the data in Fig 2Dconfirm this trend (r2 ¼ 0.96).

The results shown in Fig. 2 suggest that the dynamics of shapeswitching of PLGA EDs can be controlled over 1,000-fold byappropriate combination of molecular mass, temperature, andparticle size. For biological applications, the shape change,however, must occur near physiological temperatures. For thispurpose, we studied shape switch of PLGA-ester EDs at 37 °Cwitheither blends of various molecular masses or various sizes anddemonstrated that shape switch can be obtained at physiologicaltemperatures (Fig. S4).

We next investigated whether the pH of the surroundingaqueous phase can be used as a stimulus to induce shape switchby affecting the interfacial tension of PLGA particles, σ (Eq. 1).Despite the general hydrophobic nature of PLGA, its surfaceproperties can be tuned by choosing the appropriate end group,acid, or ester and by adjusting environmental pH (19–21). Spe-cifically, the surface charge of PLGA-acid (PLGA with a car-boxylic acid end group), and hence the interfacial tension ofPLGA particles, can be controlled through ionization of itsend carboxylic acid groups. At around physiological pH (∼7.4),the carboxyl groups of PLGA (pKa ¼ 3.85) are largely charged(−COO−), leading to low hydrophobicity and thus low interfacialtension. Lowering the pH protonates the acid groups (−COOH),thus increasing the hydrophobicity and induces shape switch(Eq. 1). Measured contact angles of droplets with different pHon PLGA films demonstrated that the interfacial tension ofPLGA-acid substantially decreases with increase in pH, whereasthat of PLGA-ester was not significantly changed (Fig. S5 a andb). The zeta potential of PLGA-acid spheres [molecular mass4.1 kDa, TgðmidÞ ¼ 27 °C], which was highly negative at pH 7,increased to almost zero as pH approached to 3.0, supportingthe hypothesis that acidic pH protonates the acid end group ofPLGA (Fig. S5c).

To assess the ability of pH to induce shape switch, EDs(AR ¼ 4) were prepared by using PLGA-acid (molecular mass4.1 kDa) and incubated in a buffer solution at pH 7.4 at 37 °C.A low molecular mass PLGAwas chosen for these experiments soas to maintain a relatively low viscosity at the operating tempera-ture of 37 °C. The ability of the particle to switch shape is thenlimited by the driving force, that is, interfacial tension. At pH 7.4,the EDs did not exhibit any significant shape switch. Upon low-ering the pH, however, the shape changed to sphere in an expo-nential manner (Fig. 3A). Although the hydrophilic surface ofPLGA-acid particles at pH 7.4 tends to uptake water becauseof reduced interfacial tension (22), no morphological deforma-tion of particles such as swelling was found during pH-inducedshape switch (Fig. S6), indicating that the shape switch was in-duced predominantly by the modulation of interfacial tensionwith pH change. Whereas interfacial tension may provide the

A

L

σT, pH, C

µ

B

CGoverning quasi-steady stokes equation Characteristic shape switch time

Lpuu 0

2

Fig. 1. Design of stimulus-responsive shape-switchable PLGA particles.(A) Mechanism of shape switch. The balance between viscosity of polymer(μ) and interfacial tension (σ) between the particle and surrounding mediadetermines the extent and dynamics of shape switch. When the interfacialtension overwhelms the polymer viscosity, an ED undergoes shape switchto a sphere in response to temperature (T), environmental pH, and chemicals(C). (B) Shape switching of PLGA particles (see also Movie S1). PLGA-ester EDs[molecular mass ∼5.2 kDa, TgðmidÞ ¼ 28 °C, AR ¼ 5, switched in deionizedwater at 37 °C] (Left: 0 min; Center: 2 min; Right: 5 min). (Scale bar: 5 μm.)(C) For Newtonian fluids, shape switching can be described by the quasi-steady Stokes equation, and the scaling factor for shape switch (τ) is givenby Eq. 1.

11206 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1000346107 Yoo and Mitragotri

Dow

nloa

ded

by g

uest

on

Mar

ch 1

3, 2

021

primary driving force for shape switch in low molecular massPLGA, high molecular mass PLGA may possess potential contri-butions from strain-induced chain orientation. The characteristicswitch time (T1∕2) depended on the pH, with a T1∕2 value of

38 min at pH 5.0 (Fig. 3B). Change of ionic strength of buffersolutions did not alter the dynamics of shape switch (Fig. S7a).The pH-dependent shape switching is exhibited only byPLGA-acid particles and not PLGA-ester particles. Specifically,

Fig. 2. Shape switching via temperature-induced change of polymer viscosity. (A) SEM images of shape-switching PLGA particles with different molecularmasses. PLGA-ester EDs [Top, molecular mass 29.8 kDa, TgðmidÞ ¼ 40 °C, AR ¼ 5.5 (scale bar: 5 μm)]; [Middle, molecular mass 52.7 kDa, TgðmidÞ ¼ 43 °C, AR ¼5.5 (scale bar: 5 μm)]; and [Bottom, molecular mass 29.8 kDa, TgðmidÞ ¼ 40 °C, AR ¼ 4 (scale bar: 500 nm)] were incubated at their mid-Tgs. (Scale bar: 5 μm.)(B) AR of the EDs decreased nearly exponentially to 1 (sphere) in a molecular-mass-dependent manner at their mid-Tg (closed squares: molecular mass 29.8 kDa;open squares: molecular mass 52.7 kDa). Higher molecular mass particles went through slower shape switch (T1∕2) because of their higher viscosity. (C) Tem-perature-responsive shape switching of PLGA-ester EDs (molecular mass 29.8 kDa). Whereas shape switch slowly occurred at onset Tg (32 °C), the T1∕2 remark-ably decreased at a temperature 10 °C above mid-Tg. (D) Size-dependent shape switch of PLGA-ester EDs (molecular mass 29.8 kDa). The T1∕2 increased as theinitial sphere’s particle size increased.

Yoo and Mitragotri PNAS ∣ June 22, 2010 ∣ vol. 107 ∣ no. 25 ∣ 11207

ENGINEE

RING

Dow

nloa

ded

by g

uest

on

Mar

ch 1

3, 2

021

PLGA-acid particles exhibited over 10,000-fold faster switchingat pH 4.0 compared to that at pH 7.4 (Fig. 3C). On the otherhand, PLGA-ester particles exhibited a relatively rapid shapeswitch at both pH values because of their hydrophobic surface.

On the basis of pH-sensitive shape switch, we hypothesizedthat the shape of PLGA particles can also be switched by usingexogenous chemicals. Specifically, PLGA-acid particles possess acharged surface at around physiological pH (∼7.4). We hypothe-sized that immobilization of certain chemicals on PLGA particlescan be used to enhance their surface hydrophobicity and induceshape switch through increased interfacial tension. Azure C waschosen as a model chemical because of its cationic charge as wellas the presence of a hydrophobic group (Fig. 4A). PLGA-acidparticles (molecular mass 4.1 kDa) were used in these experi-ments so as to maintain high ionization at pH 7.4 and low viscosityat 37 °C to facilitate switching. Incubation of these particles withAzure C (2 femtomoles per particle) led to shape switch of theEDs, which otherwise retain their shape (Fig. 4A and Fig. S7b).Azure C-induced shape switch was observed at very low Azure Camounts (<1 femtomole per particle) and was concentration-dependent up to about 5 femtomoles per particle, after whichthe effect appears to have saturated (Fig. 4B). The effect of Azure

C on shape switch was, to some extent, pH-dependent (Fig. 4C).Specifically, the degree of shape switch by Azure C was much low-er at pH 5.5 compared to that at 7.4, which is consistent with thehypothesis that the particle surface is weakly ionized at lower pH,thereby limiting the binding of Azure C.

The ability of shape-switching particles to modulate their in-teractions with cells was next assessed. Macrophages were usedas a model cell for these experiments. Macrophages, an essentialcomponent of the immune system, play a major role in governingthe fate of polymeric particles administered into the body fordrug delivery or diagnostic applications (23–25). It has beenshown that particles with high AR, such as EDs, have the poten-tial to mitigate phagocytosis (26). We hypothesized that the abil-ity of particles to switch shapes may be used to achieve precisecontrol over the extent and duration of mitigation of phagocyto-sis. PLGA-acid and -ester EDs (AR ¼ 5, Table S1) were opso-nized with mouse IgG and incubated with mouse peritonealmacrophages. PLGA-ester EDs switched their shape to spheresin due time, after which they were internalized (Fig. 5A,Movie S2, and Fig. S8a). As a control, PLGA-acid EDs, whichare designed not to switch shape, were not phagocytosed bymacrophages because of large AR (Fig. 5B, Movie S3, and

5 105

s) 105

s)

3

3.5

4

4.5

5

ct R

atio

, AR

Change pH from 7.4

pH = 7.4

102

103

104

witc

h T

ime,

T1/

2 (m

inut

es

102

103

104

10pH 7.4

pH 4.0

witc

h T

ime,

T1/

2 (m

inut

es

1

1.5

2

2.5

Asp

e

Time (Hours)

pH = 3

pH=5

10-1

100

101

-0.5 0 0.5 1 1.5 2 2.5 3 2 3 4 5 6 7 8

Cha

ract

eris

tic S

w

pH

10-1

100

101

Acidend

Esterend

Cha

ract

eris

tic S

w

A B C

Fig. 3. pH-induced shape switch of PLGA particles. (A) The shape-switching study was performed in a buffer solution at pH 7.4 at 37 °C by using PLGA-acid EDs[molecular mass 4.1 kDa, TgðmidÞ ¼ 27 °C, AR ¼ 4]. There was no detectable shape switch at pH 7.4 because of low interfacial tension of hydrophilic surfaceoriginating from ionized acid end groups. Shape switch occurred upon pH reduction, which led to protonation of the acid end and exhibited a near-exponential decrease of AR. Data are shown as mean� SEM (n > 30). (B) The T1∕2 of PLGA-acid EDs (molecular mass 4.1 kDa) depended strongly on pH.The T1∕2 gradually increased as pH increased from 3.0 to 4.5 and abruptly increased above pH 5.0. (C) Whereas PLGA-acid EDs showed over 10,000-fold increaseof T1∕2 from pH 4 to 7.4, PLGA-ester ED [molecular mass 6.5 KDa, TgðmidÞ ¼ 28 °C, AR ¼ 4] switched shape quickly (T1∕2 < 1 min) at both pHs, indicating thatthere was negligible change in the interfacial tension on PLGA-ester particles by pH.

Fig. 4. Chemical-induced shape switch. PLGA-acid EDs [molecular mass 4.1 kDa, TgðmidÞ ¼ 27 °C, AR ¼ 4] were incubated at pH 7.4 and 37 °C. Cationic Azure Cpossessing a hydrophobic group was used as a trigger for shape switch. (A) PLGA-acid particles, whose shape remained the same, started switching shape uponexposure to Azure C (2 femtomoles per particle), indicating that Azure C binds to particles and reduces the interfacial tension. AR decreased nearly exponen-tially. Data are shown as mean� SEM (n > 30). (b) The Azure C-induced shape switch was concentration-dependent up to 5 femtomole per particle, abovewhich the effect seemed to have saturated. (C) Fold difference in the characteristic switch time between Azure C-treated and nontreated EDs. The degree ofshape switch by Azure C was higher at high pH than at low pH because of greater availability of Azure C binding sites (i.e., ionized end carboxylic acids) on theparticles at higher pH.

11208 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1000346107 Yoo and Mitragotri

Dow

nloa

ded

by g

uest

on

Mar

ch 1

3, 2

021

Fig. S8b). The duration for which the particles resist phagocytosisdepends directly on the dynamics of shape switch, which can becontrolled over a wide range of times, from minutes to days, bychoosing appropriate parameters.

DiscussionPLGA particles reported here exhibit stimulus-dependent shapeswitching. Yang et al. have previously reported shape switchingof nanoparticles made from liquid crystalline polymers (27); how-ever, those polymers are not physiologically compatible and theshape change occurred only under nonphysiological conditions(∼95 °C) with no temporal control. Studies have also reportedon change in shape of hydrogel microparticles through their nat-ural erosion behavior (28); however, synthesis of biocompatiblepolymeric particles with stimulus-controlled, physiologically com-patible, switchable shape and their ability to modulate cellularinteractions has not been previously demonstrated.

The results presented here demonstrate that shape of particlescan be switched in real time in response to a wide range of stimuli,and the dynamics of shape switch can be controlled over a widerange of times by modulating physico-chemical properties of par-ticles, such as molecular mass, size, and surface chemistry, as wellas the strength of the external stimulus. The stimuli used here arealso amenable to in vivo applications. For example, local eleva-tion of temperature can be achieved by using ultrasound (29) orphotothermal activation mediated by gold nanoshells (30), whichcan be incorporated in the particles. pH may also be used as astimulus, especially for targeting tumors, which possess an acidicenvironment, or late endosomes and lysosomes, which also pos-sess a distinctively acidic environment (31). Shape-switching par-ticles can also be prepared at nanoscale. We prepared elongatedparticles from 160-nm PLGA nanoparticles and found that theseparticles switched their shape in 20 min upon temperature eleva-tion from 25 to 40 °C (Fig. 2A, Bottom). Nanometer-sized parti-cles changed shape in response to temperature in the same way aslarger, micron-sized particles. Shape-switching nanoparticlescould have potential applications for tumor targeting. Specifi-cally, elongated (rods or elliptical disks) particles can be usedto target tumors because of their prolonged circulation and betteraccumulation in tumors compared to their spherical counterparts

(32–34). However, they exhibit slower internalization comparedto spheres (34–37). Shape-switching nanoparticles can addressthis conflict by allowing the rods to accumulate in tumors andthen shifting them to spheres to facilitate internalization. Thisshape shift can be induced by thermal or pH triggers as men-tioned above.

The motivation to develop particles with switchable shapescomes from the fact that shape has recently emerged as a crucialdesign parameter of micro- and nanoparticles (3, 4, 33, 38, 39).Studies have demonstrated that departure from the conventionalspherical shape of particles brings unique and improved function-alities to the particulate systems, resulting in improved biologicalresponses. For example, macrophage-mediated phagocytosis isgoverned by the local geometry of the particles (40), and elon-gated particles exhibit reduced phagocytosis, longer circulationin the blood, and higher accumulation in target tissues (26, 32,34, 41). The same shape, however, impedes particle internaliza-tion into target cells compared to spheres (34–36). Such conflictsin shape requirements can potentially be addressed through“switchable particles” whose shape can be switched in responseto a stimulus.

Shape-switching particles follow the paradigm adopted bynatural cells including erythrocytes and platelets as well as cancercells, all of which undergo shape switch in response to a physicalor chemical stimulus to adapt their properties and achieve ad-vanced functions (42–44). Particles described here switch shapein one direction, and further studies are necessary to assess thefeasibility of reversible shape switch. Upon future studies focusedon cell-specific interactions and in vivo studies, shape-switchingparticles may open additional opportunities in biomedicalapplications.

Materials and MethodsParticle Preparation. The characteristics of different types of PLGA used in thisstudy are summarized in Table S1. Relatively uniform-sized PLGA sphereswere synthesized by using a membrane emulsification technique (45). EDswere prepared by using previously published stretching method with somemodifications (40). See SI Text for details of preparation of all particles.

Fig. 5. Time-lapse video microscopyclips of shape-dependent phagocytosisby macrophage. (A) A shape-switchingPLGA-ester ED (mixture of two PLGAs,AR ¼ 5) was initially attached on amacrophage and not phagocytosed.The macrophage then quickly interna-lized the particle once shape switchedto near-sphere shape. (B) Macrophagespread on a PLGA-acid ED [molecularmass 4.1 kDa, TgðmidÞ ¼ 27 °C,AR ¼ 5], which do not switch shape atpH 7.4, but could not complete phago-cytosis. All particles were opsonizedwithmouse IgGbefore theexperiments.(Scale bar: 10 μm.)

Yoo and Mitragotri PNAS ∣ June 22, 2010 ∣ vol. 107 ∣ no. 25 ∣ 11209

ENGINEE

RING

Dow

nloa

ded

by g

uest

on

Mar

ch 1

3, 2

021

Characterization of PLGAs. Contact angle studies were performed by using agoniometer, and interfacial tension of PLGA in aqueous condition was calcu-lated from the contact angles by using Young’s equation and equation ofstate approach (46). See SI Text for details. Viscosity of PLGA and zeta poten-tial of PLGA particles was measured by using methods described in SI Text.

Shape-Switch Study. The freshly prepared EDs were dispersed in deionizedwater, phosphate/citrate buffer, or Azure C solution prepared in phos-phate/citrate buffer (107 particles in 40 μL solution) in a polypropylene tube.The buffer solutions were normalized to 0.5 M of ionic strength by usingNaCl. The particle dispersion was incubated in a temperature-controlledwater bath (Isotemp Digital Model 210; Fisher Scientific). For temperature-induced switch, the temperature was adjusted to various levels. For pH-basedswitching, the pH of phosphate/citrate buffer was lowered to a specific pH byaddition of a precalculated amount of 0.1 M citric acid. At a determined timepoint, the sample was immediately cooled down to 4 °C in the ice-cold waterto stop further shape switch, and the AR of the particles was analyzed underthe Axiovert 25 microscope at 100× or by SEM.

Scanning Electron Microscope (SEM). SEM was used for confirmation of shapechange progression of the particles at various time points. During the shape-change study, particle suspension was collected and dried at 4 °C to preventfurther shape switch. To verify particle morphology, particles were coatedwith palladium (Hummer 6.2 Sputtering System; Anatech Ltd.) and imagedwith a Sirion 400 scanning electron microscope (FEI Co.) at 3 eV.

Phagocytosis.Mouse peritoneal macrophage cells (J774, ECACC) were used asmodel macrophages. The cells were cultured in DMEM (Gibco BRL) supple-mented with 100 U∕mL penicillin, 100 U∕mL streptomycin, and 10% fetalbovine serum under standard culture conditions (37 °C, 5% CO2, humidified).Cells (2 × 105 cells∕mL) were allowed to attach in dishes lined with coverslipglass in the culture medium. The dishes were placed on the Axiovert 25 mi-croscope at 100× with phase contrast filters and equipped with BioptechsDelta T Controlled Culture Dish System® (Bioptechs Inc.) to keep the cellsat 37 °C. For opsonization, EDs were incubated with 0.25 mg∕mL mouseIgG (Sigma-Aldrich) for 60 min at 37 °C. The particles were washed twice withPBS. Opsonization was verified with fluorescently tagged IgG (MolecularProbes). EDs (one particle per cell) were added to the dishes and bright-fieldimages were collected every 30 s for 20 min by a CoolSNAPHQ CCD camera(Roper Scientific) connected to the Metamorph® software. Observed cellswere randomly chosen from the entire population, thus discounting poten-tial bias because of heterogeneity in macrophage size (radius 7.5� 2.5 μm).Images were condensed into movies and analyzed for phagocytic events. Suc-cessful phagocytosis exhibited membrane ruffling at the site of attachment,blurring the crisp boundary of the membrane, and subsequent reforming ofthe membrane boundary after internalization.

ACKNOWLEDGMENTS. The authors thank Nishit Doshi, Krystyna Brzezinska,and Siyoung Choi for many discussions. The authors acknowledge supportfrom Program of Excellence in Nanotechnology from National Heart, Lungand Blood Institute (1U01HL080718).

1. Nel AE, et al. (2009) Understanding biophysicochemical interactions at the nano-biointerface. Nat Mater 8:543–557.

2. Karp JM, Langer R (2007) Development and therapeutic applications of advancedbiomaterials. Curr Opin Biotechnol 18:454–459.

3. Champion J, Katare Y, Mitragotri S (2007) Particle shape: A new design parameter formicro-and nanoscale drug delivery carriers. J Control Release 121:3–9.

4. Mitragotri S (2009) In drug delivery, shape does matter. Pharm Res 26:232–234.5. Mitragotri S, Lahann J (2009) Physical approaches to biomaterial design. Nat Mater

8:15–23.6. Griset AP, et al. (2009) Expansile nanoparticles: Synthesis, characterization, and in vivo

efficacy of an acid-responsive polymeric drug delivery system. J Am Chem Soc131:2469–2471.

7. Owens DE, 3rd, Eby JK, Jian Y, Peppas NA (2007) Temperature-responsive polymer-goldnanocomposites as intelligent therapeutic systems. J Biomed Mater Res A 83:692–695.

8. Zhang L, Guo R, Yang M, Jiang X, Liu B (2007) Thermo and pH dual-responsive nano-particles for anti-cancer drug delivery. Adv Mater 19:2988–2988.

9. Lee ES, Na K, Bae YH (2005) Super pH-sensitive multifunctional polymeric micelle.Nano Lett 5:325–329.

10. Lahann J, et al. (2003) A reversibly switching surface. Science 299:371–374.11. Putney SD, Burke PA (1998) Improving protein therapeutics with sustained-release

formulations. Nat Biotechnol 16:153–157.12. Shive MS, Anderson JM (1997) Biodegradation and biocompatibility of PLA and PLGA

microspheres. Adv Drug Deliver Rev 28:5–24.13. Macosko CW (1994) Rheology: Principles, Measurements, and Applications (Wiley-

VCH, New York).14. Hiki Y, Kobayashi H, Takahashi H (2000) Shear viscosity of inorganic glasses and

polymers. J Alloy Compd 310:378–381.15. Takahashi H, Hiki Y, Kobayashi H (1998) Study on the shear viscosity of several

inorganic glasses. J Appl Phys 84:213–218.16. Stone HA, Leal LG (1989) Relaxation and breakup of an initially extended drop in an

otherwise quiescent fluid. J Fluid Mech 198:399–427.17. Rossler E, Hess K, Novikov V (1998) Universal representation of viscosity in glass

forming liquids. J Non-Cryst Solids 223:207–222.18. Champion JA, Katare YK, Mitragotri S (2007) Making polymeric micro- and nanopar-

ticles of complex shapes. Proc Natl Acad Sci USA 104:11901–11904.19. Butler SM, Tracy MA, Tilton RD (1999) Adsorption of serum albumin to thin films of

poly(lactide-co-glycolide). J Control Release 58:335–347.20. Lam XM, Duenas ET, Daugherty AL, Levin N, Cleland JL (2000) Sustained release of

recombinant human insulin-like growth factor-I for treatment of diabetes. J ControlRelease 67:281–292.

21. Tracy MA, et al. (1999) Factors affecting the degradation rate of poly(lactide-co-glycolide) microspheres in vivo and in vitro. Biomaterials 20:1057–1062.

22. von Burkersroda F, Schedl L, Gopferich A (2002) Why degradable polymers undergosurface erosion or bulk erosion. Biomaterials 23:4221–4231.

23. Juliano R (1988) Factors affecting the clearance kinetics and tissue distribution ofliposomes, microspheres and emulsions. Adv Drug Deliver Rev 2:31–54.

24. Koval M, Preiter K, Adles C, Stahl PD, Steinberg TH (1998) Size of IgG-opsonizedparticles determines macrophage response during internalization. Exp Cell Res242:265–273.

25. Poznansky M, Juliano R (1984) Biological approaches to the controlled delivery ofdrugs: A critical review. Pharmacol Rev 36:277–336.

26. Champion JA, Mitragotri S (2009) Shape induced inhibition of phagocytosis of polymerparticles. Pharm Res 26:244–249.

27. Yang Z, Huck WT, Clarke SM, Tajbakhsh AR, Terentjev EM (2005) Shape-memorynanoparticles from inherently non-spherical polymer colloids. Nat Mater 4:486–490.

28. Hwang DK, et al. (2009) Stop-flow lithography for the production of shape-evolvingdegradable microgel particles. J Am Chem Soc 131:4499–4504.

29. Diederich CJ, Hynynen K (1999) Ultrasound technology for hyperthermia. UltrasoundMed Biol 25:871–887.

30. Lal S, Clare SE, Halas NJ (2008) Nanoshell-enabled photothermal cancer therapy:Impending clinical impact. Acc Chem Res 41:1842–1851.

31. Mukherjee S, Ghosh RN, Maxfield FR (1997) Endocytosis. Physiol Rev 77:759–803.32. Decuzzi P, Ferrari M (2006) The adhesive strength of non-spherical particles mediated

by specific interactions. Biomaterials 27:5307–5314.33. Decuzzi P, Pasqualini R, Arap W, Ferrari M (2009) Intravascular delivery of particulate

systems: Does geometry really matter? Pharm Res 26:235–243.34. Muro S, et al. (2008) Control of endothelial targeting and intracellular delivery of

therapeutic enzymes by modulating the size and shape of ICAM-1-targeted carriers.Mol Ther 16:1450–1458.

35. Chithrani BD, Ghazani AA, Chan WC (2006) Determining the size and shape depen-dence of gold nanoparticle uptake into mammalian cells. Nano Lett 6:662–668.

36. Decuzzi P, Ferrari M (2008) The receptor-mediated endocytosis of nonsphericalparticles. Biophys J 94:3790–3797.

37. Yoo JW, Doshi N, Mitragotri S (2010) Endocytosis and intracellular distribution of PLGAparticles in endothelial cells: Effect of particle geometry. Macromol Rapid Comm31:142–148.

38. Ferrari M (2008) Nanogeometry: Beyond drug delivery. Nat Nanotechnol 3:131–132.39. Simone EA, Dziubla TD, Muzykantov VR (2008) Polymeric carriers: Role of geometry in

drug delivery. Expert Opin Drug Del 5:1283–1300.40. Champion J, Mitragotri S (2006) Role of target geometry in phagocytosis. Proc Natl

Acad Sci USA 103:4930–4934.41. Geng Y, et al. (2007) Shape effects of filaments versus spherical particles in flow and

drug delivery. Nat Nanotechnol 2:249–255.42. Chien S (1987) Red cell deformability and its relevance to blood flow.Annu Rev Physiol

49:177–192.43. Jin J, Daniel JL, Kunapuli SP (1998) Molecular basis for ADP-induced platelet activation

II. The P2Y1 receptor mediates ADP-induced intracellular calcium mobilization andshape change in platelets. J Biol Chem 273:2030–2034.

44. Sahai E (2007) Illuminating the metastatic process. Nat Rev Cancer 7:737–749.45. Ito F, Makino K (2004) Preparation and properties of monodispersed rifampicin-loaded

poly(lactide-co-glycolide) microspheres. Colloid Surface B 39:17–21.46. Kwok DY, Neumann AW (1999) Contact angle measurement and contact angle

interpretation. Adv Colloid Interfac 81:167–249.

11210 ∣ www.pnas.org/cgi/doi/10.1073/pnas.1000346107 Yoo and Mitragotri

Dow

nloa

ded

by g

uest

on

Mar

ch 1

3, 2

021