Lab on a Chip - Columbia BMEorion.bme.columbia.edu/leonglab/publications/pdf/... · Lab on a Chip CRITICAL REVIEW Cite this: Lab Chip,2016,16,4482 Received 22nd September 2016,

Lab on a Chip

CRITICAL REVIEW

Cite this: Lab Chip, 2016, 16, 4482

Received 22nd September 2016,Accepted 20th October 2016

DOI: 10.1039/c6lc01193d

www.rsc.org/loc

Cell-laden microfluidic microgels for tissueregeneration

Weiqian Jiang,† Mingqiang Li,† Zaozao Chen and Kam W. Leong*

Regeneration of diseased tissue is one of the foremost concerns for millions of patients who suffer from

tissue damage each year. Local delivery of cell-laden hydrogels offers an attractive approach for tissue re-

pair. However, due to the typical macroscopic size of these cell constructs, the encapsulated cells often

suffer from poor nutrient exchange. These issues can be mitigated by incorporating cells into microscopic

hydrogels, or microgels, whose large surface-to-volume ratio promotes efficient mass transport and en-

hanced cell–matrix interactions. Using microfluidic technology, monodisperse cell-laden microgels with

tunable sizes can be generated in a high-throughput manner, making them useful building blocks that can

be assembled into tissue constructs with spatially controlled physicochemical properties. In this review, we

examine microfluidics-generated cell-laden microgels for tissue regeneration applications. We provide a

brief overview of the common biomaterials, gelation mechanisms, and microfluidic device designs that are

used to generate these microgels, and summarize the most recent works on how they are applied to tissue

regeneration. Finally, we discuss future applications of microfluidic cell-laden microgels as well as existing

challenges that should be resolved to stimulate their clinical application.

1. Introduction

Millions of patients suffer from diseased or damaged tis-sue each year. Although tissue transplantation can be usedto treat these patients, its application is limited by a se-vere shortage of donor tissue. Tissue engineering offers asolution by engineering tissues to replace the lost func-

tions.1 The extracellular matrix, a critical component ofnatural tissue that is composed of a variety of proteins aswell as both soluble and insoluble macromolecules, regu-lates tissue dynamics by influencing cellular processessuch as proliferation, differentiation, migration, and apo-ptosis through bi-directional molecular interactions withencapsulated cells.2 Currently, a popular tissue engineeringapproach is to isolate and incorporate a patient's autolo-gous cells into three-dimensional scaffolds that mimic thefunctions of the extracellular matrix. These cell-laden scaf-folds, which provide an environment for new tissue

4482 | Lab Chip, 2016, 16, 4482–4506 This journal is © The Royal Society of Chemistry 2016

Weiqian Jiang

Weiqian Jiang received dual BSdegrees in Biomedical Engineer-ing and Electrical and ComputerEngineering from Duke Univer-sity in 2013. He is currently agraduate student majoring inBiomedical Engineering at Co-lumbia University. His currentresearch focuses on microfluidicsand tissue engineering.

Mingqiang Li

Dr. Mingqiang Li received his BSdegree from the University of Sci-ence and Technology of China in2009, followed by a PhD degreefrom the Chinese Academy ofSciences in 2015. He is workingas a postdoctoral research scien-tist with Prof. Kam W. Leong atColumbia University on micro-fluidics and biomaterials.

Department of Biomedical Engineering, Columbia University, New York, NY

10027, USA. E-mail: [email protected]

† These authors contributed equally to this paper.

Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article OnlineView Journal | View Issue

http://crossmark.crossref.org/dialog/?doi=10.1039/c6lc01193d&domain=pdf&date_stamp=2016-11-09

http://dx.doi.org/10.1039/C6LC01193D

http://pubs.rsc.org/en/journals/journal/LC

http://pubs.rsc.org/en/journals/journal/LC?issueid=LC016023

Lab Chip, 2016, 16, 4482–4506 | 4483This journal is © The Royal Society of Chemistry 2016

generation, can be inserted into the diseased area of a pa-tient's body to guide the structure and function of thenew tissue.

The material comprising the scaffold determines its physi-cal, biological, and mass transport properties, which are cru-cial design variables to consider depending on the target tis-sue. Many synthetic polymers, including polyĲglycolic acid),polyĲlactic acid), and poly(lactic-co-glycolic acid), have beenused as scaffold materials, but they require surgical incisionsfor placement into the patient's body.3 In contrast, hydrogels,which are three-dimensional polymer matrices formed bycrosslinking hydrophilic homopolymers, copolymers, ormacromers, can be delivered into the body in a minimally in-vasive manner. Moreover, hydrogels are not only biocompati-ble but also structurally and compositionally similar to theextracellular matrix.4 Despite these favorable properties, en-capsulation of cells within macroscopic hydrogels often leadsto limited cell–cell contact and communication as well aspoor nutrient exchange due to a low rate of diffusion andsuboptimal distance between extracellular molecules.5

This issue can be resolved by forming hydrogel micro-spheres, or microgels, whose large surface area-to-volume ra-tio promotes effective nutrient and water transfer as well asimproves cell–matrix interactions, thereby maintaining long-term viability of the encapsulated cells.6,7 Cell-laden micro-gels have been used in tissue engineering applications asbuilding blocks for complex tissue mimics,8 co-culture sys-tems for developing three-dimensional organ models,9 andcontrolled microenvironments for directing stem cell differ-entiation.10 In all of these applications, control over the sizeand size distribution of the microgels is important as theycan influence the phenotypes of the encapsulated cells.11,12

Although many microfabrication techniques can entrap cells,they suffer from low throughput inherent in batch process-ing. A promising alternative is microfluidic techniques,which can be used to rapidly generate monodisperse micro-gels with tunable sizes simply by manipulating and control-ling the flow of multiple immiscible liquids.

Here, we review microfluidics-generated cell-laden micro-gels for tissue regeneration applications. We first brieflycover the methods for gelation and materials used to producemicrogels. We next describe typical microfluidic device de-signs used to generate microgels and strategies for incorpo-rating cells into these microgels. This is followed byhighlighting the advantages of microfluidics-generatedmicrogels over conventional hydrogels. After summarizingthe most recent works on cell-laden microgels for tissue re-generation (Fig. 1), we speculate on the potential applicationsof these microgels in other tissue engineering applications.

2. Principle of gelation

Aqueous hydrogel precursor solutions generally undergo ei-ther physical or chemical gelation to form solidified micro-gels. The gelation method is usually chosen based on a vari-ety of factors, such as the type of polymer used forcrosslinking, strategy for cell encapsulation, and specific tis-sue regeneration application. Comprehensive overviews ofcommon gelation mechanisms can be found in recent re-views.13,14 Here, we provide a brief overview of the gelationprinciples for common physical and chemical gelation strate-gies, as well as outline their respective advantages and disad-vantages when applied to the preparation of microfluidicmicrogels (Table 1).

2.1 Physical gelation

Physical gelation can be initiated by an environmental trigger,such as changes in pH or temperature, upon which macromol-ecules crosslink via noncovalent, secondary interactions suchas hydrophobic, electrostatic, and hydrogen bonding.15,16 Al-though the gelation process can be carried out under mild con-ditions, the physical hydrogel network typically has weak me-chanical strength and shows poor stability in tissues.17 Here,we briefly overview several common physical gelation methods,including electrostatic interaction (ionic crosslinking), ther-mally mediated gelation, and hydrogen bonding.

Zaozao Chen

Dr. Zaozao Chen received his BSand MS degrees from SoutheastUniversity, followed by a PhDdegree from the University ofNorth Carolina at Chapel Hill.He is working as a postdoctoralresearch scientist with Prof. KamW. Leong at Columbia Univer-sity, focusing on development ofhuman “Organs-on-a-Chip” sys-tems as disease models and drugscreening platforms.

Kam W. Leong

Kam W. Leong is the Samuel Y.Sheng Professor of BiomedicalEngineering at Columbia Univer-sity, and holds a joint faculty po-sition in the Department of Sys-tems Biology. His lab focuses onnon-viral gene delivery, directcell reprogramming, tissue-chipsystems, and genome editing. Heis the Editor-in-Chief of Biomate-rials, and is a member of theNational Academy of Engineer-ing and National Academy ofInventors.

Lab on a Chip Critical review

Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



Being one of the most frequently used physical methods,crosslinking based on electrostatic interactions can be car-ried out between either a polymer and a small molecule or

two polymers of opposite charges. For the first case, a classicexample would be the crosslinking of alginate with Ca2+ ions,which attach to the carboxylic acid groups of alginatemolecules.18–20 An example of the latter case would be ioniccomplementary peptides with alternating positively and nega-tively charged units that can self-assemble into crosslinkednetworks.15 The crosslinking process can be triggered by pHchanges, which ionize the functional groups that induce gela-tion. For instance, acetic acid has been used to trigger the re-lease of Ca2+ from Ca–EDTA complexes mixed in alginate,leading to gelation upon contact of the ions with alginatemolecules.21 However, changes in pH should be gradual andkept within a certain limit so that cell viability is not signifi-cantly reduced. An advantage of this strategy is that ionic spe-cies in the extracellular fluid can bind competitively withcomponents in the ionically crosslinked network, therebygradually degrading the hydrogel.

Another popular physical strategy is thermally mediatedsol–gel transition.22–25 For instance, gelatin solutions areemulsified at 40 °C, and the droplets are subsequently cooleddown to 5 °C to induce gelation.26 The advantage of thismethod lies in its biocompatibility because it doesn't requireany toxic reagents. However, the precursor solution must bemaintained at a temperature that is higher than the gelationpoint when operating the emulsification step in a micro-fluidic device. Otherwise, temperatures below the gelling

Fig. 1 Schematic overview of the applications of cell-laden micro-fluidic microgels for tissue regeneration.

Table 1 The advantages and disadvantages of common physical and chemical gelation strategies for fabricating microfluidic microgels

Gelation strategy Advantage Disadvantage Ref.

Physical gelationElectrostaticinteraction

• Hydrogel can be gradually degradedby ionic species in extracellular fluid

• pH-triggered crosslinking may damage cells 15

Thermally mediatedgelation

• Biocompatible • Small variations in temperature may cause dramatic increases inviscosity, making it more difficult to produce monodisperse microgels

26•No toxic crosslinking agents

Hydrogen bonding • Biocompatible • Rapid degradation of hydrogels as influx of water will dilute thehydrogen-bonded network

15,30• No chemical crosslinking agents

• Constituent polymers found naturallyin ECM• Polymer blend well-suited forinjection

Chemical gelationPhotopolymerization • Gelation can be carried out at room

or physiological temperature• Released free radicals can damage cellular components 31,

32• Photoinitiator concentration, UV light intensity, and exposure timemust be fine-tuned to minimize cell damage• Short gelation time (less than a

second to a few minutes)• Temporal and spatial control overgelation• In situ gelation, conforms to shape ofdefect

Michael addition • Requires only a small amount ofcatalyst

• Gelation occurs under basic conditions 5, 33,34• Fast gelation kinetics requires sufficient mixing of reactants to

prevent heterogeneous gelation• Relatively mild reaction conditions(no heat or light)• No degradation products

Enzymatic reaction • Enzymes used catalyze reactions thatoccur naturally in the body

• Some reactions are difficult to control 35–39

• Gelation occurs at neutral pH andmoderate temperatures• High substrate specificity, eliminatesundesired side reactions or toxicity• Cell-responsive degradation

Lab on a ChipCritical review

Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



point could lead to fluctuations in the viscosity of the poly-mer solution, making it difficult to produce monodispersedroplets.

The last physical gelation strategy is hydrogen bond-mediated crosslinking, which requires a mixture of two ormore natural polymers with compatible geometries that con-tain chemical groups capable of forming hydrogenbonds.27–29 The viscoelastic properties of the resulting poly-mer mixture are more gel-like than those of the individualpolymers, making the mixture suitable for injection.30 Themain advantage of this method is its excellent biocompatibil-ity because no chemical crosslinking agent is required andthe constituent polymers are usually found naturally in theECM. However, hydrogels assembled using this strategy maydegrade rapidly, as an influx of water will quickly dilute thehydrogen-bonded network.15

2.2 Chemical gelation

Compared to physical hydrogels, hydrogels produced bychemical gelation via covalent crosslinking show greater sta-bility, longer durability, and better mechanical properties.17

However, the use of chemical or enzymatic crosslinkers maycause cytotoxicity, and the low specificity of some chemicalreactions may lead to undesired interactions with proteinsand cells. Several frequently implemented chemicalcrosslinking strategies are presented here, including photo-polymerization, Michael addition, and enzymatic reactions.

During photopolymerization, UV light homolytically splitsphotoinitiator molecules into free radicals, which initiate thepolymerization of a liquid monomer or macromer into a co-valently crosslinked hydrogel that is mechanically strong andstable.31 The gelation process not only can be carried out rap-idly at room or physiological temperature, usually taking lessthan a second to a few minutes, but can also be controlledtemporally and spatially.32 Due to a high degree of control,photopolymerization can effect in situ gelation after solutionsof photosensitive polymers conform to irregularly shaped tis-sue defects. The disadvantage of this crosslinking strategy isthat the released radicals can react with and damage cellularcomponents. Therefore, parameters such as the type ofphotoinitiator, photoinitiator concentration, UV light inten-sity, and total UV exposure time must be finely tuned to en-sure high cell viability.

Another chemical gelation method is the milder and mostcommon type of Michael addition, thiol-Michael addition, inwhich a nucleophilic thiolate reacts with an electrophile,such as an unsaturated ester, to form a thioester linkage.5

Thiol-Michael addition only requires a small amount of cata-lysts, occurs under mild reaction conditions without the useof heat or light, and does not produce any degradation prod-ucts.33 However, one particular issue to note is that insuffi-cient mixing of reactants will result in heterogeneous gela-tion and lead to inconsistent cellular responses to thebiochemical and mechanical cues of the hydrogel.34 There-fore, gelation kinetics must be fine-tuned to achieve a bal-

ance between the mixing and gelation rates. Since gelationusually occurs on the order of minutes, droplets encapsulat-ing the reactants can first be flowed through serpentinemicrofluidic channels, which induce passive mixing, beforeundergoing complete gelation.

Finally, enzymatic reactions provide a more physiologicallyrelevant chemical crosslinking strategy, as most enzymesused for hydrogel crosslinking are responsible for catalyzingreactions that occur naturally in the human body.35–38 Forexample, one of the most commonly employed enzymes istransglutaminase, which is responsible for forming fibrinclots during wound healing. It can catalyze the formation ofcovalent bonds between the free amine groups and theγ-carboxamide groups of protein, resulting in highly stablebonds that are resistant to proteolytic degradation.39 In addi-tion to occurring at neutral pH and moderate temperatures,enzymatic reactions also avoid any undesired side reactionsor toxicity due to the high substrate specificity of the en-zymes. Moreover, cell-responsive enzymes can be designed sothat the degradation rate of the hydrogel can be coupled withcellular signals secreted during tissue regeneration.

3. Gel materials

One of the most important factors to consider when design-ing microgels is the constituent material. Since cells aresuspended in an aqueous precursor solution prior to micro-gel formation inside the microfluidic device, the gel materialmust be biocompatible and promote cell survival. Ideally, itshould also provide biochemical and mechanical cues to en-hance tissue regeneration, and possess a degradation ratethat is closely coupled with the rate of tissue growth. Cur-rently, there is no single type of material that meets all ofthese demands, but by carefully considering their respectivemerits and drawbacks (Table 2), we may judiciously applythese materials to specific tissue regeneration applications.Here, we briefly overview the properties of several gel mate-rials, and readers may refer to these comprehensive reviewsfor their detailed chemical functionalization and principle ofgelation.40,41

3.1 Natural polysaccharides

Natural polysaccharides are frequently used as materials forcell encapsulation because they are biocompatible and capa-ble of gelling under mild physiological conditions.18 How-ever, since natural polysaccharides are isolated from livingorganisms, their physicochemical properties depend on thespecific species from which they are derived and may alsoface immunogenicity issues. Here, we introduce several com-mon polysaccharides used for gel formation, including algi-nate, hyaluronic acid, agarose, dextran, and chitosan, whosechemical structures are presented in Fig. 2A.

Alginate, an inexpensive and cytocompatible polymer, isan unbranched polysaccharide comprising varying ratios andarrangements of β-D-mannuronic acid and α-L-guluronic acidresidues.42 The interaction of multivalent ions, such as Ca2+,


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



Co2+, and Ba2+, with alginate leads to the formation of ionicbridges between the ions and the carboxylic acid groups ofα-L-guluronic acid, and subsequently results in a three-dimensional network.43 The concentration of alginate can bemodulated to produce microgels with varying stiffness. How-ever, achieving stable microfluidic emulsification at moder-ately high alginate concentrations is relatively difficult due toa significant increase in viscosity that is induced by the rapidgelation of alginate.44 To mitigate this issue, the droplet for-mation and gelation processes must be separated in time,which can be accomplished using one of three general strate-gies: internal gelation, external gelation, and coalescence-

induced gelation.45 The most representative works have beenreported as well as summarized by Kumacheva andcoworkers.26,46–49

In the internal gelation approach, a precursor polymer so-lution is mixed with an initiator or crosslinking agent, whichis subsequently activated upon contact with a compound inthe continuous phase, thereby triggering gelation.48,50 In onestudy by Tan et al., droplets were formed using a dispersedphase composed of an alginate solution containing cells andCaCO3 nanoparticles as well as a continuous phase compris-ing oil with acetic acid.44 As acetic acid diffused into thedroplet, the acidity of the alginate solution increased, thereby

Table 2 The advantages and disadvantages of typical natural polysaccharides, proteins, and synthetic polymers that are crosslinked to form microgels

Biomaterial Advantage Disadvantage Reference

Natural polysaccharidesAlginate • Inexpensive • Poor mechanical strength 42–44, 64, 71

• Cytocompatible • Impurities may affect the degradation rate• Crosslinking occurs under mild conditions • Dramatic increases in viscosity induced by rapid

gelation may increase polydispersity of microgels• Dependence on diffusion of divalent ions maylead to heterogeneous crosslinking

Hyaluronic acid • Non-immunogenic • Anionic surface does not promote cell attachment 55–57, 64, 71• Biocompatible• Enzymatically degradable• Interacts with cell-surface receptors toinfluence tissue organization

Agarose • Bioinert • No active protein adsorption or cell adhesion sites 62, 64, 71• Cytocompatible • Temperature fluctuations can lead to dramatic

increases in viscosity during emulsification• Transparent• Can be readily functionalized withcell-adhesion proteins

Dextran • Chemically similar to GAGs • Weak mechanical properties 64• Hydroxyl groups can be chemically modifiedwith cell-recognition sites

• Resistant to cell adhesion and protein adsorption

Chitosan • Biocompatible • Weak mechanical properties 55, 64–67, 69• Antibacterial activity • Unstable, cannot maintain a predefined shape• Low toxicity and immunostimulatory activities • Difficult to obtain medical-grade chitosan

• Impurities may affect the degradation rate• Hydrophilic surface promotes cell adhesion,proliferation, and differentiation• Enzymatically degradable

ProteinsCollagen • Low antigenicity • Poor mechanical strength 71, 72, 76, 78, 79

• Good cell-binding properties • Rapid degradation• Good biodegradability • Crosslinking agents may be toxic

• Slow gelation process (>30 min)Fibrin •Minimal cytotoxicity • Weak mechanical properties 55, 80–82

•Biocompatible•Promotes cell adhesion•Cell-mediated degradation kinetics

Gelatin •Bioactive and cell-adhesive motifs • Weak mechanical properties 71, 83–88•MMP-sensitive degradation sites • Crosslinking agents may be toxic•Cytocompatible • Temperature fluctuations can lead to dramatic

increases in viscosity during emulsificationSynthetic polymersPEG • Biocompatible (FDA-approved) • Weak mechanical properties 5, 64, 89, 90

• Non-immunogenic • Poor degradability• Hydroxyl end groups can be converted intoother functional groups

• Resistant to cell adhesion and protein adsorption

• Proteolytic and cell-adhesion sites can bereadily incorporated

Pluronic • Biocompatible • Rapid erosion at the injection site 92–95• Non-immunogenic • Non-biodegradable

• Poor mechanical strength


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



releasing Ca2+ ions from the nanoparticles and inducing gela-tion. However, prolonged direct contact of the cells withCaCO3 nanoparticles in the alginate solution can lead tophysical damage to the cells and decrease their viability. Inorder to reduce the amount of damage, the cells and CaCO3

nanoparticles can be flowed through the microchannels intwo different alginate solutions before emulsification.51 Also,instead of mixing acetic acid with the continuous oil phase,it can be directly added after droplet formation, thereby re-ducing the exposure time of cells under acidic conditions.

For external gelation, the crosslinking agent and the algi-nate solution are initially physically separated. After dropletsencapsulating the alginate solution are formed, they comeinto contact with another liquid that contains thecrosslinking agent, which diffuses into the droplets and initi-ates gelation. Compared to internal gelation, this strategy en-ables control over the morphology of the resulting microgelthrough modulation of the amount of Ca2+ ions that diffuseinto the droplets.26 Kim et al. co-flowed an alginate solutioncontaining cells and another aqueous alginate solutionthrough a microfluidic flow-focusing device, and generateddroplets with a continuous phase comprising an oleic acidsolution containing CaCl2.

52 The alginate solution gelled asCaCl2 diffused into the droplets.

In coalescence-induced gelation, droplets containing aprecursor polymer solution and droplets encapsulating acrosslinking agent coalesce with each other one-by-one toform crosslinked microgels.53 Therefore, the productivity ofmicrogel formation depends on the probability of collisionsbetween two separate droplets.46 Using this technique, Liuet al. generated sodium alginate droplets and CaCl2 dropletsin two separate flow-focusing channels, which subsequentlyled to two circular expansion chambers where the two typesof droplets collided and merged with one another, triggeringgelation.54

Hyaluronic acid (HA), a linear polysaccharide composed ofalternating units of the disaccharide β-1,4-D-glucuronic acid-β-1,3-N-acetyl-D-glucosamine, is a non-sulfated glycosaminogly-can that is found throughout the body in connective, epithe-lial, and neural tissues.55 It plays an active role in wound re-pair, cellular signaling, morphogenesis, and organization ofthe ECM.56 Additionally, HA interacts with cell-surfacereceptors to influence tissue organization, stimulates theproduction of collagen II and aggrecan, and promotes cellproliferation.55 With excellent biocompatibility and non-immunogenicity, the enzymatically degradable HA is a bio-logically relevant material for fabricating hydrogels.57 HA-based hydrogels can be generated using a variety of gelationstrategies, including photopolymerization, Michael addition,Schiff-base reaction, thermally induced crosslinking, and co-valent crosslinking.57,58 Out of these methods, photo-polymerization and Michael addition are the most frequentlyused strategies for producing HA hydrogels. For instance, thecarboxylic acids on the HA backbone can easily be modifiedwith methacrylic anhydride to form methacrylated HA, whosenetwork properties can be controlled by varying the length ofUV exposure, HA molecular weight, concentration, and num-ber of reactive groups.56,59 Another study by Segura's groupreported the assembly of hyaluronic acid-based buildingblocks, which were crosslinked via pseudo-Michael addition,into an injectable microporous scaffold that can be loadedwith cells during scaffold formation.60

Agarose is a neutral, bioinert polysaccharide comprising thepolymers β-1,3-linked D-galactose and α-1,4-linked 3,6-anhydro-L-galactose in an alternating order.18 The gelation of agarose isthermally reversible, as it maintains a liquid form at the physio-logical temperature of 37 °C and gels upon cooling below 20°C, which is a result of the aggregation of double helices pro-duced by the entanglement of anhydro bridges on each agarosemolecule. Therefore, emulsification of agarose solutions is usu-ally performed at temperatures above 37 °C, and microgels aresubsequently formed by cooling the droplets to below the gela-tion temperature. The mechanical properties of the resultingagarose microgels can be tuned by adjusting the agarose con-centration.61 Although agarose does not have active adsorptionsites for proteins or adhesion sites for cells, cell adhesion pro-teins or collagen can be incorporated to promote attachment ofanchorage-dependent cells.62,63

Dextrans, homopolysaccharides consisting mainly of α-1,6-linked D-glucopyranose residues, are synthesized by lactic acidbacteria.64 Although they are highly hydrophilic and chemi-cally similar to glycosaminoglycans, they are resistant to celladhesion and protein adsorption. To ameliorate this issue,the abundant hydroxyl groups of dextran can be chemicallymodified to incorporate specific cell-binding sites. Dextran-based hydrogels can be generated by either physical or chem-ical crosslinking, but the more common approach is byphotopolymerization of dextran that has been functionalizedwith polymerizable groups such as methacrylates.

Chitosan, a linear polysaccharide consisting of β-1,4-linked D-glucosamine and N-acetyl-D-glucosamine units, is

Fig. 2 Chemical structures of common gelation materials. (A)Structures of natural polysaccharides. (B) Structures of syntheticpolymers.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



derived from the naturally occurring chitin and has a similarstructure to glycosaminoglycans.55 It has excellent biocom-patibility, antibacterial properties, as well as low toxicity andimmunostimulatory activities.65 Furthermore, its hydrophilicnature supports cell adhesion, proliferation, and differentia-tion.64 Chitosan hydrogels can be formed through thermallyinduced or pH-triggered gelation and enzymatically degradedby lysozymes in vivo,66 releasing harmless natural metabolitesas degradation products.67 It is often mixed with other poly-mers to improve its mechanical properties, and its aminegroups enable the conjugation of biomolecules to control itsbiochemical properties.67 Fan et al. used metal-free clickchemistry to generate an oxanorbornadiene-functionalizedchitosan and azido-functionalized hyaluronan composite hy-drogel, whose crosslinking and resulting microstructure de-termine its characteristics such as equilibrium swelling, me-chanical properties and degradation kinetics.68 However, oneparticular disadvantage is that it is hard to obtain ultra-pure,medical-grade chitosan, so impurities such as allergens mayaffect its rate of de-polymerization and biological activity.69

3.2 Proteins

Proteins are attractive materials for generating cell-ladenmicrogels because their chemical properties and fibrousstructure closely mimic those of the extracellular matrix,thereby providing a confined microenvironment that pro-motes the adherence, growth, migration, and differentiationof the encapsulated cells.70 Here, we provide a brief overviewof several frequently used proteins, including collagen, fibrin,and gelatin.

Collagen, the most abundant protein in the human bodythat helps maintain structural integrity, consists of repeating(Gly–X–Y)n units that form left-handed helical α-chains,which pack together to form triple helices that are stabilizedby specific covalent crosslinks between neighboring collagenmolecules.64 Of the numerous forms of collagen, type I colla-gen is the most commonly used due to its abundance,good biodegradability, low antigenicity after removal oftelopeptides, and cell-binding ability.71,72 Collagen has fre-quently been used as an ECM substrate in applications suchas bioartificial liver studies.73–75 However, extended applica-tion of collagen as a robust material for tissue regenerationhas been hampered by its rapid degradation rate and henceweak mechanical strength. Increasing the collagen concentra-tion will result in a modest increase in strength, but will alsorestrict cell migration and nutrient diffusion.76 Glutaralde-hyde crosslinking has been explored to enhance the mechani-cal properties of collagen-based gels, but it can cause cytotox-icity and biocompatibility issues.64 Another strategy is togenerate collagen-containing composite hydrogels, often ac-complished by incorporating photocrosslinkable hyaluronicacid into collagen gels, to increase the elastic modulus andmechanical strength.77 Another issue with collagen is that itspolymerization mechanism is thermally driven and thus re-quires exposure to elevated temperatures for at least 30 mi-

nutes to ensure complete gelation.78 Thus, an additional in-cubation step is often required after microfluidic synthesis ofcollagen-containing emulsion droplets.79

Fibrin, an important ECM component that regulateswound healing, is polymerized from a precursor plasma pro-tein called fibrinogen. Upon vascular injury, the enzymethrombin is released and cleaves peptide fragments from sol-uble fibrinogen to generate insoluble fibrin peptides thatform fibrils, which eventually assemble into a fibrin networkthat is crosslinked upon addition of the transglutaminaseFactor XIII.80 The clotting time is the main factor in deter-mining the characteristics of the resulting microstructuresuch as fiber size and porosity.80 For instance, higher throm-bin concentration leads to more rapid gelation, resulting in atighter network structure with smaller pore sizes. Sincethrombin has a Na+ binding site that controls its activationkinetics,81 addition of sodium chloride can increase the gela-tion time to several minutes.82 In addition to fibrin's excel-lent biocompatibility, cell adhesiveness, and biodegradability,the mechanical properties of fibrin hydrogels can be tunedby varying the initial concentrations of fibrinogen and throm-bin. One of the main advantages of fibrin is that it forms aprovisional matrix, which serves as a temporary scaffold thatpromotes the invasion of leukocytes and endothelial cellsuntil the diseased tissue is completely regenerated.82 Thismeans that fibrin clots are naturally designed with appropri-ate cell-mediated degradation kinetics to balance the rate oftissue regeneration and matrix remodeling, which is hard toachieve with synthetic materials or even other more perma-nent natural polymers in the ECM.

Gelatin, the denatured form of collagen, comprises a vari-ety of amino acids such as glycine and proline.18 Bioactiveand cell-adhesive motifs in gelatin promote cell attachmentand proliferation, while matrix metalloprotease (MMP)-sensi-tive degradation sites enable gelatin-containing microgels tobe easily degraded to release the encapsulated cells.7,83 Fur-thermore, gelatin is cytocompatible and shows lower antige-nicity compared to collagen.84 At room temperature, gelatinsolidifies due to its extensive physical bonds. However, it liq-uefies upon heating to above the physiological temperatureof 37 °C, making it impractical to use gelatin directly as aninjectable in situ curable gel. Although chemical crosslinkingagents such as glutaraldehyde can be used to strengthen themechanical properties of gelatin, they can lead to cytotoxic-ity.85 As a more cell-friendly alternative, methacrylate groupscan be incorporated into the amine-containing side groups ofgelatin, rendering the gel photocrosslinkable via activation ofa photoinitiator.86 An additional advantage of this strategy isthat the mechanical and chemical properties of the gel canbe tuned by adjusting parameters in the methacrylation andphotocrosslinking processes.83,87,88

3.3 Synthetic polymers

Synthetic polymers have emerged as attractive alternativescaffold materials to naturally derived ECM proteins, which


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



are limited by their immunogenicity and poor mechanicalproperties.89 Compared to natural materials, pure syntheticpolymers have lower risks of toxicity, immunogenicity, andinfection.64 Additional advantages of synthetic polymers in-clude their ability to be photopolymerized, tunable mechani-cal properties, and facile manipulation of the scaffold struc-ture, chemical composition, and degradability.90 However,synthetic polymers lack critical biological functions that bio-active natural materials possess, including cell adhesion andbiodegradation, but this can easily be compensated for byfunctionalizing them with bioactive molecules such as cell-adhesive peptides and enzyme-sensitive peptides. Here, weintroduce two common synthetic polymers, polyĲethyleneglycol) (PEG) and Pluronic, whose chemical structures areshown in Fig. 2B.

PEG is an FDA-approved hydrophilic polymer comprisinga PEG diol with two hydroxyl end groups, which can beconverted into other functional groups to generate hydrogelswith different chemical properties.89 Although PEG is bio-compatible and non-immunogenic, it has poor degradabilityand is resistant to cell adhesion and protein adsorption.5 Inorder to improve its degradability, enzyme-sensitive peptidesequences can be incorporated into PEG hydrogels to makethe degradation process responsive to cellular signals andcell-secreted enzymes.89 Likewise, addition of adhesive pro-teins such as fibronectin introduces proteolytic and cell ad-hesion sites into the resulting gel. There are three generalcrosslinking strategies for fabricating PEG hydrogels: radia-tion of PEG polymers, free radical polymerization of PEG ac-rylates, and chemical reactions such as Michael additionand enzymatic reactions.89 Out of these strategies, the mostcommonly used is photopolymerization, which has the ad-vantage of inducing in situ gelation of a PEG macromer solu-tion at physiological temperature and pH.89 Moreover, witha careful choice of photoinitiator chemistry and concentra-tion as well as light intensity, cell exposure to radicals canbe minimized, making the gelation process more cell-friendly.

Pluronic, a commercial surfactant comprising the triblockcopolymer polyĲethylene oxide-b-propylene oxide-b-ethyleneoxide) (PEO–PPO–PEO), has been used as an injectablethermosensitive hydrogel.91 Above the critical micellizationconcentration, the amphiphilic copolymer blocks can formmicelles in water, which become entangled with one anotherupon heating, thereby leading to gelation.91,92 For instance,an aqueous solution of Pluronic F127 copolymer (>20 w/v%)can be gelled in situ when heated to body temperature. Al-though Pluronic has demonstrated excellent biocompatibilityboth in vitro and in vivo, Pluronic hydrogels suffer from rapiderosion at the injection site, non-biodegradability, and poormechanical properties.92–95 Nevertheless, its biodegradabilitycan be improved by linking ester, urea, or carbonate bondsto the copolymers.96,97 Furthermore, other crosslinking strat-egies, such as photopolymerization and Michael addition,can be used to improve the stability and mechanical proper-ties of Pluronic gels.91

4. Design of microfluidic device formicrogel generation

In order to generate microgels with the desired size, geome-try, or biochemical and mechanical properties, microfluidicdevices must be carefully designed with appropriate dimen-sions as well as features for manipulating the spatial organi-zation of the constituent cells. Here, we briefly explain howgeneral microfluidic devices can be used to fabricate micro-gels, and provide schematics of typical designs.

4.1 Microfluidic devices

4.1.1 PDMS-based microfluidic devices. Poly-Ĳdimethylsiloxane) (PDMS) is one of the most widely used ma-terials for fabricating microfluidic devices.98 Several featuresmake it an attractive material for prototyping, including itslow cost, optical transparency, and biocompatibility.99 Addi-tionally, with the development of soft lithography for PDMS,not only has the prototyping process been greatly simplifiedsuch that new concepts can easily be tested in PDMS-baseddevices, but also more complicated devices could be designedwith components such as pneumatically activated valves.These valves can restrict fluid movement through pressuriza-tion and deflection of an adjacent channel, which is onlymade possible by the elastomeric nature of PDMS and cannotbe accomplished using rigid materials such as silicon orglass.100

A common PDMS device design for microgel generation isbased on the flow-focusing geometry (Fig. 3A), in which a dis-persed liquid phase flows through the middle channel and asecond immiscible continuous liquid phase flows throughthe two outer channels.101 Typically in this design, the dis-persed phase contains a gel solution mixed with cells whilethe continuous phase comprises a stream of immiscible oilthat exerts a shear force on and focuses the dispersed phasethrough a narrow cross-junction, resulting in the periodicbreak-up of the dispersed phase and formation of gel-encapsulated emulsion droplets. The main advantage of thisdesign is that the number of encapsulated cells per dropletand the droplet diameter can be controlled simply by tuningthe gel and oil flow rates. Since a droplet production rate of 1kHz per device can easily be achieved,10 parallelization of alarge number of devices may facilitate the scale-up of micro-fluidic fabrication of microgels.

Although the flow-focusing configuration is a reliablemethod for producing monodisperse spherical microgels, itis not amenable to producing gels with other geometries,which could serve as important tissue constructs that mimicthe natural tissue architecture. For instance, in a hepatic lob-ule, which is the basic unit of liver tissue, hepatocytes are ar-ranged in a cord-like structure surrounded by endothelialcells in a sinusoidal distribution.102 The complex spatial pat-terning of this tissue architecture can be recreated usingPDMS devices with pneumatically actuated microvalves(Fig. 3B), which can be selectively manipulated to precisely


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



control fluid movement within the microchannels. Duringdevice operation, flexible membranes are deflected outwardsupon application of air pressure, and effectively serve asmicro-molding barriers that keep the precursor polymer solu-tion in a desired geometric configuration until the gelationprocess is completed. By sequentially actuating these micro-valves, additional layers with complex patterns can be addedsuccessively, forming a biomimetic microgel. Despite the ver-satility that this technology affords, its fabrication processcan be tedious as it requires multilayer soft lithography tobuild the valve system.

4.1.2 Capillary microfluidic devices. Aside from commonlyused PDMS-based devices, capillary microfluidic devices offeran alternative method of generating highly monodisperseemulsion droplets with controllable sizes. Capillary micro-fluidic devices are fabricated by coaxially aligning circularand square glass capillaries on glass slides.103 First, a squarecapillary is securely attached to a glass slide. Next, a cylindri-cal capillary, usually with an outer diameter of 1–2 mm, is ta-pered to create an orifice, simply by heating and then pullingthe capillary with a micropipette puller. The size of theresulting orifice represents the inner diameter of the circularcapillary, which can be further adjusted with a microforge toprecisely define the diameter of the injection nozzle for theinner disperse phase.7 To ensure accurate coaxial alignmentof both capillaries, the outer diameter of the circular capillarymust match the inner dimensions of the square capillary,which allows the tapered circular capillary to be slid into thesquare capillary to form the microfluidic device.

By adjusting the geometric configuration of this setup,two different methods for droplet generation can be achieved

using capillary microfluidics. The first method utilizes a co-flow geometry (Fig. 3C), in which two fluids injectedthrough the circular and square capillaries flow in thesame direction to generate coaxial flow. In contrast to theco-flow configuration, the flow-focusing geometry involvesthe injection of two fluids from opposite ends of the samesquare capillary (Fig. 3D), which allows the inner fluid tobe hydrodynamically flow-focused as it is forced throughthe capillary orifice by the outer fluid. Both methods canproduce monodisperse emulsion droplets upon extrusion ofthe inner fluid from the capillary orifice under low rates offluid flow. However, if either flow rate exceeds a certainthreshold, a jet will form in which the inner fluid travelsin a long stream before droplets are generated downstream,leading to greater variation in droplet size. Compared tothe co-flow configuration, the flow-focusing setup is bettersuited for generating small monodisperse droplets that con-tain particulate suspensions in the inner phase, whichcould otherwise accumulate and block the capillary orificein the co-flow configuration.

The adoption of glass as a constituent material, as well astheir geometric configuration, provides capillary microfluidicdevices with several advantages. First, the compatibility ofglass with common organic solvents extends the range of ap-plications in which these devices can be used, such as chemi-cal and material syntheses.98 Second, the wettability of capil-lary devices can be precisely controlled by treating the innerwall with a surface modifier. For instance, treatment withn-octadecyltrimethoxysilane makes the capillary wall hydro-phobic while treatment with 2-[methoxyĲpolyethyleneoxy)-propyl]trimethoxysilane makes the wall hydrophilic.104 Fi-nally, the coaxial assembly of glass capillaries allows thesedevices to generate truly three-dimensional fluid flow, whichrobustly confines sample flow to a streamline at the center ofthe capillaries, enabling precise control over the size and uni-formity of the emulsion droplets.

4.2 Size control

Precise regulation of the size and size distribution of micro-gels is essential to the preparation of cell-laden microgels forvarious applications in tissue regeneration. The diameter ofthe microgel must be controlled within a certain range tomaintain a microenvironment that promotes high cell viabil-ity and proliferation. For instance, if the diameter of theresulting microgel were to fall below 60 μm, there may not bea sufficient number of encapsulated cells to promote cellcontact and proliferation.103 Conversely, the microgel shouldnot exceed 200 μm in diameter so that oxygen can be readilyexchanged across the gel to maintain long-term cell sur-vival.105 Furthermore, a narrow size distribution of cell-ladenmicrogels is desired in applications such as the co-culture ofmultiple cell types and the study of the role of cell confine-ment and intercellular distance in influencing cell fate,which require accurate control over the average number ofcells in each microgel.18

Fig. 3 Typical microfluidic device designs. (A) PDMS flow-focusing. (B)PDMS-based pneumatically actuated microvalves. (C) Capillary co-flow. (D) Capillary flow-focusing.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



Bulk emulsification techniques, which often utilize someform of mechanical agitation, suffer from a lack of controlover the size distribution of the microdroplets.106,107 In con-trast, microfluidic methods offer enhanced control over thedroplet size and significantly reduce the size distribution ofthe microdroplets.108–110 As a result, microfluidics provides afacile method for modulating the gel size and controlling thesize distribution within a narrow range. A common strategyfor selectively tuning the droplet diameter is to vary the flowrate ratio between the dispersed and continuous phases. Fora specific device geometry and combination of input liquids,increasing this ratio leads to larger droplet diameters. Stableand accurate droplet formation using this method requiresthe maintenance of a constant pressure difference betweenthe aqueous liquid and oil channels throughout the entiremicrofluidic fabrication process, which can easily be achievedusing syringe pumps.111 A more robust alternative strategy isto adjust the dimensions of the input microchannels. For in-stance, increasing the width of the channels enables the for-mation of droplets with larger diameters.

5. Strategy for cell deposition

Microfluidic microgels can be classified into two general cate-gories based on the spatial deposition of the cells. Dependingon the specific application, cells can be encapsulated withinthe microgel or grown on the surface of the gel (Fig. 4). Here,we explain why one strategy may be preferred over the other.

5.1 In the microgel

True recapitulation of the native cellular microenvironmentrequires a 3D matrix with distinct physicochemical proper-ties. Although macroscale 3D culture systems have been used

as alternatives to conventional monolayer cultures, they arenot feasible for long-term cell culture due to a maximum oxy-gen diffusion limit of 200 μm.112 Therefore, cell-laden micro-gels with diameters between 100 and 400 μm have been fabri-cated to ensure that the encapsulated cells stay safely withinoxygen diffusion constraints. Additionally, important factorsthat influence the behavior of cells in the microenvironment,such as extracellular matrix components, soluble biomolecu-lar signals, and biophysical cues, can be incorporated intothe microgels to provide the cells with more physiologicallyrelevant microenvironments.

Since cells are often pre-mixed with a precursor solutionbefore emulsification, the fabrication process must becytocompatible, which limits the number of appropriate ma-terials and gelation methods that can be used. For instance,if the temperature required for gelation exceeds that of physi-ological conditions or if the crosslinking agent and droplet-forming oil contain cytotoxic components, then cell viabilitycould be compromised. Nevertheless, this strategy is particu-larly useful for in vivo applications in which cell-laden micro-gels can provide a minimally invasive means of delivery to alocalized wound area as well as protect the cells from thehost immune response and external stress during injection.

5.2 On the microgel

Culturing cells on the microgel surface provides an alternativemethod of in vitro cell culture that possesses several advan-tages over conventional monolayer culture. First, due to the sig-nificantly higher surface-to-volume ratio of microgels, a greaternumber of cells can be cultured with smaller volumes of cul-ture media.113 Second, compared to those of plastic or glasssurfaces, the mechanical and chemical properties of microgelscan be tailored to replicate physiological conditions more faith-fully. Third, cell-seeded microgels can be directly used as in-jectable tissue constructs, whereas multiple extra steps are re-quired to collect monolayer-cultured cells and attach themonto microgels for subsequent injection into the body.

As described in the previous section, cell encapsulationwithin microgels enables a more accurate 3D recapitulationof the native microenvironment compared to 2D strategies.Despite this, there are several reasons why cell culture on the2D surface of microgels would be preferred instead. First,cells grown inside microgels are restricted within the innerspace of the gel, thereby limiting the ability of the cells toproliferate.113 In contrast, cells grown on the surface are nottightly confined, enabling them to proliferate more quicklyso that a greater number of cells can be collected within ashorter amount of time. Second, detachment and collectionof cells growing on the microgel surface are much easier thanextracting cells encapsulated within the microgel, which of-ten requires enzymatic or thermally mediated degradation.Finally, compared to cell-encapsulated microgels, thesemicrogels can be generated prior to cell adherence to the sur-face, thereby offering more flexibility in the type of precursorsolution and gelation conditions.

Fig. 4 Schematic illustrating two different strategies for celldeposition: inside the microgel and on the surface of the microgel.Note: the zoomed-in area of the schematic is not drawn to scale. Thesize of a typical cell (20 μm) is approximately three orders of magni-tude greater than that of a typical polymer network mesh (1–10 nm).


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



6. The advantages of microfluidicmicrogel compared with conventionalhydrogel

Conventional hydrogels, characterized by their high watercontent and broad range of physical properties, can beengineered in shape, size, and form to mimic the extracellu-lar environment of natural tissue. Due to their mechanicaland chemical versatility, hydrogels have found applicationsin medicine as soft contact lenses,114 adhesives for woundclosure,14 biosensors,115 etc. However, applications in stemcell research, cell therapy, and tissue engineering demand ad-ditional features from hydrogels such as compatibility withcells and biomimicry of the native extracellular matrix.11 Tomaintain high cell viability, there must be effective oxygenand nutrient transfer, which is largely dictated by the size ofthe hydrogel. Furthermore, since the properties and func-tions of cells are influenced by the distance between cells aswell as their structural arrangement, the size and size distri-bution of the hydrogel must be strictly controlled.116

While conventional suspension polymerization, spraying,and precipitation techniques do not provide sufficient con-trol over the dimensions of the generated particles, micro-fluidic methods offer facile control over the size, size distri-bution, shape, and morphology of the resulting particles,making them particularly useful for the generation of micro-gels with polydispersities of under 2–3%.26,109,117,118 Since en-capsulated cells require steady diffusion of nutrients and oxy-gen into the gel and waste products out of the gel, the largesize of hydrogels, which typically ranges from 500 to 1000μm, makes it difficult to support efficient transport for cellsin the inner core, thereby leading to reduced cell viability. Incontrast, the size of microfluidic microgels can be effectivelycontrolled within the ideal range of 60 to 200 μm, whichhelps maintain high cell viability due to the microgels' signif-icantly greater surface-to-volume ratio that enhances nutrientand waste exchange.

In addition to the gel size, uniformity in size distributionis another important criterion to consider, especially for ap-plications in directed-assembly tissue engineering, wherebuilding blocks composed of cell-laden gels are arranged in adesired geometry with specific shape and size to mimic func-tional tissue units.119 Since the size of stem cell aggregatescan influence the differentiation of cells into specific line-ages,120 cell-laden gels should ideally be monodisperse toproduce uniform aggregates that differentiate into the samelineage. Otherwise, biomimetic tissue units assembled frombuilding blocks with heterogeneous cell populations could re-sult in poor functionality.

7. Cell-laden microfluidic microgelsfor tissue regeneration

Microfluidic technologies have been employed to generatecell-laden microgels with varying properties and geometries

for use in a range of tissue engineering applications. Here,we overview the latest representative works that exploit theadvantages of cell-laden microfluidic microgels for tissue re-generation, and provide a brief summary in Table 3.

7.1 Cell encapsulation for 3D cell culture

One of the most frequent applications of microgels is the en-capsulation of cells for in vitro cell culture andexpansion.104,121–131 Since a 3D microenvironment morefaithfully mimics the cell's native niche, physicochemicallywell-defined 3D matrices are preferred for in vitro cell culture.Micron-sized hydrogels are attractive candidates because theynot only promote the exchange of nutrients and metabolicwaste during cell growth and tissue development but also en-able the tuning of physical and chemical properties of thegel. Furthermore, combined with microfluidic technologies,monodisperse cell-laden microgels with tailored propertiescan be generated to serve as 3D culturing units for studyinghow various microenvironmental properties and cues affectcell fate. Additionally, parallelization and automation ofmicrofluidic systems enable high-throughput screening ofmicrogel properties, which will facilitate the discovery ofin vitro conditions that more closely mimic those of thein vivo microenvironment. Readers are referred to many ex-cellent reviews on cell encapsulation within microfluidics-generated microgels for 3D cell culture.12,18,26,45,119,132–135 Wewill instead focus on other applications of cell-laden micro-fluidic microgels in tissue regeneration.

7.2 Controlled microenvironment for stem cell differentiation

Stem cells, which can renew themselves and differentiate intospecialized cell types, play a major role in the regeneration oftissues and restoration of tissue functions.136–138 Due to theability of murine embryonic stem (ES) cells to differentiateinto any type of cell in the adult body, many differentiationprotocols have been developed for differentiating ES cellsinto adult tissues for tissue engineering.139–141 However, it isoften difficult to control the tightly regulated microenviron-mental factors that induce specific differentiation pathwayswhen large quantities of stem cell-derived tissue constructsare needed. Thus, a scalable strategy for tailoring the stemcell microenvironment with important growth factors is re-quired. Microfluidics, a large-scale bioprocessing technique,shows great promise as it not only facilitates rapid nutrientand waste exchange but also reduces hydrodynamic stressesfrom stirring during suspension culture of stem cells.142

Moreover, compared with monolayer cultures, the expansionand differentiation of encapsulated stem cells can be accom-plished efficiently in a one-step process in microfluidicmicrogels.135 Furthermore, the monodispersity of micro-fluidic droplets is particularly important in microscale tissueengineering, in which the spheroid size influences the stemcell behavior and differentiation potential.143,144

Microfluidic double emulsion (DE) droplets provide well-defined, uniform, and compartmentalized microenvironments


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



for cell culture, with the oil shell serving as a selective bar-rier that enables an influx of nutrients, oxygen, and smallmolecules from the external environment while retainingbiomacromolecules in the inner core.145 In our recent work,we used microfluidic DEs as miniature bioreactors to rapidlygenerate human mesenchymal stem cell (hMSC) spheroids,which were subsequently encapsulated in alginate-RGDmicrogels and demonstrated enhanced osteogenic differenti-ation (Fig. 5A).146 Compared to previous water-in-oil (W/O)single-emulsion cultures, this platform enhances the viabilityof encapsulated cells by introducing an additional outeraqueous phase into the external oil phase, which has beenshown to be incompatible for long-term cell culture.147 More-over, by varying the cell density during the encapsulationprocess, the size of the spheroid can be precisely controlled.This is important because small spheroids lead to morehomogeneous chondrogenic differentiation while largerspheroids form more heterogeneous tissues.148 For instance,average spheroid sizes of 36, 46, 62, and 84 μm can beobtained with densities of 2, 5, 10, and 20 million cells permL in 200 μm DE droplets (Fig. 5B). Additionally, one of thegreatest advantages of this system is that it only takes 150min for the encapsulated cells to aggregate into spheroids(Fig. 5C), whereas other technologies require approximately1 to 4 days.149–152 For microgel fabrication, hMSCssuspended in an alginate-RGD solution were used as the in-ner phase during DE formation. Upon successful cell aggre-gation, the spheroids were released into a solutioncontaining calcium ions, which induced rapid crosslinkingof alginate molecules and subsequent formation of micro-gels. Assessment of the osteogenic differentiation of hMSCspheroids encapsulated in alginate-RGD microgels showedthat they contained greater quantities of calcium depositsand higher alkaline phosphatase activity after seven days in

culture (Fig. 5D), suggesting that spheroid behavior can becontrolled by modulating its microenvironment using thissystem.

Previously, there has not been any growth factor-containing microgel used for ES cell differentiation. To ex-plore this possibility, Siltanen et al. generated bioactivemicrogels (120 μm) composed of heparin and PEG, and in-corporated the growth factors FGF-2 and Nodal to enhancethe differentiation of mouse embryonic stem cells (mESCs)toward a definitive endoderm.10 Heparin-methacrylate andPEG-diacrylate macromers were mixed with 8-arm PEG-thiolinside the microchannels and emulsion droplets were pro-duced at rates of up to ∼1 kHz, with gelation occurring viaMichael addition (Fig. 6A–C). This crosslinking strategy hasthe benefits of being specific, cytocompatible, and indepen-dent from free radicals. To test the ability of the microgels toenhance endodermal differentiation, undifferentiated mESCswere encapsulated in the microgels and supplemented withFGF-2 and Nodal/Activin by following a definitive endodermdifferentiation protocol. The mESCs showed >95% viabilitytwo hours after encapsulation (Fig. 6D-a), and formed mESCspheroids that expanded in size and broke free from themicrogels (Fig. 6D-b). After 5 days of culture, data from qRT-PCR indicated that mESCs encapsulated in the microgels andsubjected to a one-time dose of FGF-2 and Nodal expressed a33- and 65-fold increase in the endoderm markers FoxA2 andSox17, respectively, compared to undifferentiated mESCs(Fig. 6E). In contrast, cells cultured in monolayers (control)on fibronectin with continuous FGF-2/Nodal supplementa-tion, 3D spheroids cultured on a matrigel, and mESCs encap-sulated in microgels without FGF-2 or Nodal expressed 2- and5-fold, 10- and 11-fold, and 3- and 7-fold increases in FoxA2and Sox17, respectively (Fig. 6E). These results indicate that,compared to standard 2D or 3D differentiation procedures

Table 3 Microfluidic microgels for tissue regeneration applications

Microfluidic devicePolymerizationstrategy Biomaterial Tissue regeneration application Ref.

Capillary flow-focusing Photopolymerization Gel-MA Injectable BMSC-laden microgels (160 μm) with prolonged release ofBMP-2 growth factor as osteogenic tissue constructs

7

PDMS flow-focusing Electrostaticinteraction

Alginate 3D core–shell microgel scaffold (169 μm) for co-culture of hepatocytesand fibroblasts as an in vitro liver model for drug screening

21

PDMS flow-focusing Michael addition Heparin, PEG Bioactive microgels (120 μm) containing growth factors FGF-2 andNodal for directed endodermal differentiation of mESCs

10


Alginate Alginate-RGD microgels for formation of hMSC spheroids (36, 46, 62,84 μm) and induction of osteogenic differentiation

146

PDMS devices withintegrated pneumaticmicrovalves

Thermal Collagen,gelatin, agarose

Cell-laden microgels (200 μm thick) of different shapes, with singleand multiple microchannels of varying shapes, for constructingbiomimetic liver lobules comprising a co-culture of hepatocytes andendothelial cells, as a 3D tissue model for in vitro drug toxicityscreening

9

PDMS flow-focusing Photopolymerization Gel-MA, silica Cell-seeded Gel-MA microgels (100 μm) with a protective silica hydro-gel shell for cell culture and injectable tissue construct

113

PDMS flow-focusing Michael addition PEG-VS + RGD,K, and Qpeptides

Injectable scaffold assembled from annealed microgel buildingblocks (30–150 μm) for accelerated wound healing

176


Alginate,collagen

Co-culture of hepatocyte–EPC composite spheroids (∼80 μm) inalginate–collagen microgels for enhancing hepatocellular functions

162


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



with continuous growth factor supplementation, encapsula-tion of mESCs in heparin microgels provides a well-defined,definitive endoderm-inducing microenvironment where aone-time dose of growth factors was sufficient to induce sig-nificantly higher expression levels of endoderm markers.

7.3 Co-culture systems as 3D organ models

Tissues are three-dimensional structures composed of a mul-titude of cells that are surrounded and separated by extracel-lular matrices.153 The spatial organization of a cell deter-mines how it interacts with other cells, which subsequentlyeffects the transmission of molecular signals involved in cellmovement and formation of boundaries in tissues. As a re-sult, tissue functions are affected by a variety of cues such asintercellular signaling and interaction between cells and theirsurrounding extracellular matrices.154,155 Therefore, in orderto fabricate artificial tissues that more realistically mimic thein vivo microenvironment, methods that can spatially patternmultiple types of cells embedded in biocompatible extracellu-lar matrices are needed.

Chen et al. used droplet-based microfluidics tocontrollably assemble 3D core–shell microgel scaffolds com-prising a hepatocyte-filled aqueous core and a fibroblast-laden alginate shell.21 A PDMS flow-focusing device was usedto generate water–water–oil (W/W/O) DEs with four phases:an inner phase consisting of hepatocytes (HepG2 cells) in acell culture medium, a middle phase composed of fibroblasts(NIH-3T3 cells) mixed with alginate solution and Ca–EDTA,an outer phase made up of oil, and an additional oil phasecontaining 0.15% acetic acid that triggers Ca2+ release fromthe Ca–EDTA complex to form a crosslinked alginate network(Fig. 7A and B). The mechanical strength and portability ofthe microgels enable them to be stored at −80 °C, thawed at37 °C, and then cultured again without damaging the struc-ture or significantly reducing the viability of the cells(Fig. 7C-a). Moreover, the alginate microgels (169 μm in di-ameter) are well-suited for long-term culture as their highpermeability facilitates nutrient and metabolite exchange. Anadditional benefit of using unmodified alginate is that itdeters cell attachment and prevents migration of the

Fig. 5 Microfluidic DE droplets serve as miniature bioreactors to rapidly generate human mesenchymal stem cell (hMSC) spheroids, which weresubsequently encapsulated in alginate-RGD microgels and demonstrated enhanced osteogenic differentiation. (A) Schematic diagram of DE drop-let generation and subsequent spheroid formation. Two PDMS flow-focusing devices are connected serially, with the first device producing W/Oemulsions and the second device supplementing an outer aqueous phase to generate W/O/W DE emulsions. Cells encapsulated in DE droplets as-semble into a single spheroid, which can then be released with or without microgel encapsulation. (B) Diameter of spheroids measured at differentdensities of encapsulated cells (n ≥ 50) (data = mean ± SD). (C) Time course images showing the formation of spheroids in 150 min. (D) Control ofthe microenvironment for spheroid culture using DE droplets: (a and b) confocal images showing phalloidin staining and immunostaining forE-cadherin and integrin α5β1 for spheroids encapsulated in alginate or alginate-RGD gels and cultured in normal media for 3 days; (c and d) im-ages of alizarin red staining and alkaline phosphatase activity staining using BCIP/NBT as a substrate for spheroids cultured in osteogenic mediumfor 7 days. Adapted from ref. 127 with permission from Nature Publishing Group.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



Fig. 6 Fabrication of heparin microgel with a microfluidic flow-focusing device. (A and B) Heparin-methacrylate, PEG-diacrylate, and 8-arm PEG-thiol undergo Michael addition crosslinking to form heparin hydrogels. (C) Microfluidic flow-focusing device for mixing and emulsifying hydrogelprecursors. Numbers represent inlets for the (1) carrier oil, (2 and 3) hydrogel precursors, and (4) collection outlet. Scale bar = 300 μm. (D) Encap-sulated mESC viability and EB formation: (a) representative live/dead staining of mESCs 2 h after encapsulation; (b) embryoid body formation. Scalebar = 200 μm. (E) Growth factor adsorption and mESC-to-DE differentiation in microgels. Definitive endoderm marker expression after 5 days ofdifferentiation with or without Nodal and FGF-2. Error bars represent standard deviation for n = 3 samples. Significance indicated at the p < 0.05level. Adapted from ref. 10 with permission from Elsevier.

Fig. 7 (A) Microfluidic construction of the 3D scaffold consisting of an aqueous core and a hydrogel shell: (a) crosslinking of the alginate networkby triggered release of Ca2+ from the Ca–EDTA complex; (b) schematic diagram of the PDMS device; (c) fabrication of core–shell droplets using W/W/O DEs as templates; (d) monodisperse core–shell droplets generated using droplet-based microfluidics. The alginate shell is labeled with fluo-rescein, as shown in the inset. (B) Simultaneous assembly of hepatocytes in the core and fibroblasts in the shell, forming an artificial liver in a drop-let. Cell viability is characterized using the calcein AM/EthD-1 staining kit. Scale bar = 100 μm. (C) Co-culture of hepatocytes and fibroblasts in thecore–shell spheroids: (a) high viability of cells encapsulated in the spheroids after being frozen at −80 °C for two weeks and then thawed at 37 °C;(b) co-culture of cells encapsulated in the spheroids for 10 days showing high cell viability; (c) the viability of cells cultured for 14 days decreasesslightly. Cell viability is characterized using the calcein AM/EthD-1 staining kit. Scale bar = 100 μm. (D) Comprehensive assays of liver-specific func-tions of the hepatocyte/fibroblast co-culture and the hepatocyte only culture: (a) albumin secretion and (b) urea synthesis of HepG2/NIH-3T3 co-culture and HepG2 culture measured over seven days. *p < 0.01. Adapted from ref. 18 with permission from the Royal Society of Chemistry.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



fibroblasts. Maintaining spatial separation between the hepa-tocytes and fibroblasts is of particular importance as randomco-cultures of these two types of cells will result in lowerliver-specific functions due to an imbalance between homo-typic and heterotypic cell–cell interactions.156 To assess theenhancement of liver-specific functions of the core–shellspheroid, two microgel samples were cultured using the sameamount of hepatocytes in the core, one with and the otherwithout fibroblasts in the shell (Fig. 7C-b, c). Compared to

their monotypic counterpart, the co-cultured spheroids dem-onstrated increased albumin secretion and urea synthesis(Fig. 7D), which are two major biomarkers of the liver usedin drug screening.157 These results suggest that this 3D core–shell microgel scaffold is a good in vitro liver model and maybe applied to rapid drug screening.

Although shear-induced droplet formation, employedby microfluidic technologies such as flow-focusing andT-junction devices, can produce monodisperse spherical cell-

Fig. 8 Flexible fabrication of cell-laden microgels using PAM. (A) Schematic illustrating the fabrication process of PAM: (a) 3D view of the PAMprocess; (b) cross-sectional view of microvalve manipulation during PAM. (B) Generation of multi-compartmental cell-laden microgels. Diagramdepicting a two-step PAM protocol. (C) Application of PAM to liver tissue engineering: (a and b) biomimetic construction of 3D liver lobule-likemicrotissue. (a) Illustration of the liver lobule, which consists of radially arranged hepatic cords lined by hepatic sinusoids. (b) Schematic delineatingthe experimental procedure for constructing the biomimetic microtissue. HepG2 cells were molded to form a hepatic cord-like network (greenpattern), and then HUVEC-C cells were co-immobilized to generate a hepatic sinusoid-like network (red pattern), which together recreated theliver lobule-like morphology (merged pattern); (c) morphological demonstration of the 3D liver lobule-like microtissue on-chip and after harvestingand deposition into media: (iii) the fabricated liver lobule-like microtissue consisted of (i) a hepatic cord-like network and (ii) a hepatic sinusoid-likenetwork. Scale bars = 500 μm; (d) analysis of APAP-induced hepatotoxicity via the biomimetic microtissue. The data are given as means ± SD andcollected from three independent experiments. *p < 0.05; **p < 0.01. Adapted from ref. 9 with permission from the Royal Society of Chemistry.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



laden microgels, it does not support the formation of othershapes, which could provide additional insight into the re-lationship between ECMs and cell fate and behavior.158 Analternative approach called pneumatic-aided micro-molding(PAM) makes it possible to not only generate cell-ladenmicrogels with precisely controlled sizes and well-definedgeometries but also fabricate single and multiple micro-channels in the microgels using various combinations ofhydrogels and homogeneous or heterogeneous cell types.9

The central components of PAM are pre-designed pneu-matic microvalves integrated into PDMS microfluidic de-vices using multilayer soft lithography, which enables pre-cise manipulation of fluid and micron-sized objects. Thebasic operating principle is as follows (Fig. 8A): cells arefirst pre-mixed with a hydrogel precursor solution andloaded into the chamber of the device; then an array ofmicrovalves is activated by external gas pressure, duringwhich the PDMS membrane of the valves deflect outwardsto form a 3D physical barrier that serves as a micro-moldto restrict the cell–hydrogel mixture; after gelation throughpolymerization, the microgels (200 μm thick) are recoveredby deactivating the microvalves to release the barrier. Basedon this principle, a two-step PAM configuration consistingof an inner ring-shaped and an outer U-shaped microvalveenables precise manipulation of heterogeneous cell typesand ECM components in 3D (Fig. 8B), which is essentialfor fabricating functional tissue constructs that can mimicthe complex organization of in vivo tissuearchitectures.159–161 As a proof of concept, two-step PAMwas used to construct 3D liver lobule-like microtissue byaligning endothelial cells (HUVEC-C) along radially pat-terned hepatocytes (HepG2) to mimic the natural architec-ture of the liver lobule in which cords of hepatocytes radi-ate from a central vein and are separated by sinusoid-likeendothelial cells (Fig. 8C-a–c). Compared to 2D and 3Dmonolayer-cultured HepG2 cells, the liver lobule-like micro-tissues exhibited lower cell viability given the same treat-ment of acetaminophen (APAP), which causes hepatotoxicityupon overdose (Fig. 8C-d). This result indicates that theHepG2 cells cultured in the microtissues were more sensi-tive to APAP toxicity, likely due to the microtissues' abilityto better maintain heterotypic cell–cell and cell–ECM inter-actions, making the microtissues more reliable 3D tissuemodels for in vitro drug toxicity screening.

In our latest work, we developed a microfluidic DEdroplet-based micro-encapsulation technology for co-culturing hepatocytes and endothelial progenitor cells(EPCs) to form composite spheroids, which were subse-quently encapsulated into alginate–collagen microgels anddemonstrated enhanced hepatocellular functions.162 DEdroplets were generated using two PDMS flow-focusing de-vices, with the inner phase comprising a 5 : 1 mixture ofhepatocytes and EPCs (HepEPCs) suspended in an alginatesolution with an optimized 1 : 1 mixture of EPC (EGM-2)and hepatocyte culture media. Following spheroid forma-tion after only 4 h in culture, the oil shell was removed

and the inner phase was exposed to a calcium chloridesolution, which induced polymerization and led to the for-mation of alginate microgels (Fig. 9A-a). The resultingmicrogels were smaller than 200 μm and contained ∼80μm encapsulated composite spheroids (Fig. 9A-b, c), whichis vital to maintaining cell viability as sizes greater than∼150 μm will induce necrosis in the spheroid core.163 Re-sults showed that EPCs significantly improved hepatocytefunctions, such as syntheses of albumin and urea, at aHepEPC co-culture ratio of 5 : 1 (Fig. 9B). Further increasesin EPC fraction led to reduced hepatocyte performance,suggesting that this synergistic effect is only manifested atcertain co-culture ratios. To examine whether conducivematrix cues would further enhance the synergistic effect,collagen I—a matrix substrate that helps maintain hepato-cellular functions—was added to the gelation solution toform alginate–collagen (Alg-col) microgels. As a result, theamount of cumulative albumin synthesized was signifi-cantly higher in the HepEPC-in-Alg-col group than those inthe Hep-in-Alg-col and HepEPC-in-Alg groups (Fig. 9C), in-dicating that matrix cues provide additional support to en-hance the functions of hepatocytes in co-culture. Overall,this DE micro-encapsulation system offers several advan-tages over existing technologies. First, production of ∼80μm spheroids is difficult to achieve using existingmethods, which usually generate microgels with sizes ofaround 500–1000 μm,164 presenting a diffusion barrier tothe detoxification process of many bioartificial livers inclinical trials.165 Second, this method enables the produc-tion of microgels containing single spheroids and preventsthe formation and fusion of multiple spheroids within thesame gel, which could ultimately lead to spheroid sizesthat are greater than the 150 μm limit for ensuring highcell viability. Finally, while other strategies require thegeneration of spheroids before micro-encapsulation, theDE technology enables facile one-step generation ofmicrogel-encapsulated hepatocyte spheroids with tunablebiochemical and mechanical properties.

7.4 Bone/cartilage regeneration

Stem cell transplantation is a promising strategy fortreating bone and cartilage injuries. Although direct injec-tion provides a minimally invasive method for deliveringcells to the repair site, it is limited by low cell retentionand engraftment, which may be attributed to mechanicalshear forces that damage the cell membrane during the in-jection process.166 To overcome this limitation, stem cellshave been suspended in hydrogels that can be injected andsolidified in situ. However, these hydrogels suffer from in-sufficient oxygen and nutrient exchange due to the largesize of the bone defects, which impedes bone regenera-tion.167 Alternatively, microgels not only maintain cell via-bility by facilitating efficient nutrient and waste exchangebut also protect the scaffold from shear stress during injec-tion. Due to the controlled sizes, monodispersity, and high


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



encapsulation efficiency of microfluidic microgels, theyhave been used to encapsulate cells for bone/cartilage re-generation applications.

Weitz's group engineered methacrylated gelatin (Gel-MA)microgels encapsulating bone marrow-derived mesenchymalstem cells (BMSCs) and the osteogenic growth factor, bonemorphogenetic protein-2 (BMP-2), that may serve as inject-able tissue constructs for enhancing bone formation and os-sification.7 A capillary flow-focusing microfluidic device wasused to generate the microgels, with the dispersed phaseconsisting of Gel-MA supplemented with the photoinitiator2-hydroxy-4′-(2-hydroxyethoxy)-2-methylpropiophenone,BMSCs, and BMP-2, and the continuous phase comprisingthe perfluorinated oil HFE-7500. The size of the W/O emul-sions was controlled by adjusting the flow rate of each of thetwo phases. After collection, the W/O emulsions were exposedto UV light (365 nm) for 20 s and polymerized into microgelsthat were approximately 160 μm in diameter. Adherent cellsreadily spread inside the microgels, and there were a largenumber of BMSCs that migrated from the interior of themicrogel to the surface after 4 weeks of culture, indicatingthat the BMSCs were actively involved in the regeneration

process. Cells released from the microgels demonstrated sim-ilar viability and morphology to those directly grown on tis-sue culture plastic. There was an early reduction in alkalinephosphatase activity, which usually signals the differentiationof BMSCs into osteoblasts. Meanwhile, the percentage of cal-cium deposits within the microgels increased to almost 70%after 4 weeks of culture, demonstrating the in vivo osteogenicpotential of the BMSC-laden microgels. Moreover, the Gel-MAmicrogels demonstrated a prolonged release of BMP-2 thatlasted for 1 week, which is advantageous for repairing bonedefects as it promotes bone formation and ossification longerthan most hydrogels do. In a model of rabbit femoral defects,Gel-MA microgels delivering both BMSCs and BMP-2 inducedthe most extensive in vivo bone formation and hence thehighest therapeutic efficacy out of the three groups: micro-gels loaded with BMSCs, with and without BMP-2, and thecontrol group of microgels alone.

7.5 Microgel scaffold for wound healing

Injectable materials have been developed as a minimally in-vasive means of implanting stem cells into wound sites for

Fig. 9 (A) (a) Micro-encapsulated spheroid production (left and middle: bright-field images taken at time 0 and 4 h; right: microgel formation afteroil removal). Scale bar = 200 μm. (b) Size distribution of DE droplets. (c) Size distribution of spheroids and microgels. (B) (a) Tracking of cell organi-zation in the composite spheroids at different co-culture ratios. Scale bar = 50 μm. (b) Cumulative albumin release (**p < 0.01 between 1 : 5 and1 : 1/3 : 1/0 : 1). (c) Cumulative urea secretion (*p < 0.05 between 1 : 5 and 1 : 1/3 : 1/0 : 1). (d) Basal CYP3A4 activity measured by a luminogenic assay(**p < 0.01 between 1 : 5 and 0 : 1). The figure legends denote co-culture cell ratios (EPC : hepatocytes). Data represent mean ± S.E.M., n = 3. (C)(a) Characterization (immunostaining against hepatocyte (albumin) and EPC (vWF) markers, live/dead staining and staining of bile canaliculi) of sin-gle or co-culture spheroids cultured under different conditions. Scale bar = 50 μm. (b) Cumulative albumin release. (c) Cumulative urea secretion.(d) Basal CYP3A4 activity measured by a luminogenic assay (*p < 0.05 between HepEPC in Alg-col and HepEPC in Alg). (e) Induction of CYP3A4activity after treatment with 10 × 10−6 M Dex for 72 h. Data represent mean ± S.E.M., n = 3. Adapted from ref. 143 with permission from John Wileyand Sons.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



regenerating healthy tissue.168 The efficacy of these materialsin promoting tissue regeneration depends on how closelycoupled the rates of material degradation and tissue develop-ment are, so that newly forming tissues can gradually replacethe materials at a similar rate.55 For example, if the degrada-tion rate of the material is too fast, an insufficient amount ofscaffolding will hamper tissue ingrowth. In contrast, a ratetoo slow may promote fibrosis and lead to improper tissuedevelopment.169 Hydrogels, whose degradation rates can eas-ily be tuned by modulating the density of the crosslinked net-work as well as the polymeric backbone, have been widelyused as injectable scaffold materials.170–173 However, thisapproach relies on material degradation prior to tissue inte-

gration, which may not be suitably applied in different deg-radation environments. A more robust strategy—thebuilding-block approach—uses microgels to form scaffoldsby assembling them into interconnected lattices withmicron-sized pores, thereby eliminating the need to tunethe degradation rate of the material for different woundsites, which possess varying physical and chemicalcharacteristics.

Khademhosseini's group cultured cardiac side population(CSP) cells on the surface of Gel-MA microgels, and coatedthe cell-seeded microgels with an additional layer of silica hy-drogel to protect the cells against external stress during injec-tion procedures.113 The Gel-MA microgels were fabricated

Fig. 10 (A) (a) Microfluidic fabrication of Gel-MA microgels. Aqueous droplets made of Gel-MA pre-gel solution, generated from a microfluidicflow-focusing device, were photopolymerized to form Gel-MA microgels. (b, c) Microscopy images of (b) the microfluidic flow-focusing devicegenerating Gel-MA droplets and (c) Gel-MA microgels fabricated by UV-initiated photopolymerization of Gel-MA droplets. (B) (a) Gel-MA microgelswere used as an in vitro platform to culture cells on the surface. (b) Microscopy images of CSP cells cultured on Gel-MA microgels. The cells ad-hered on the surface of Gel-MA microgels and proliferated over time. Scale bar = 100 μm. (c) CSP cells on Gel-MA microgels were placed on acell-adhesive surface, and their adhesion and proliferation over time were monitored. Scale bar = 50 μm. (C) (a) Schematic of the fabrication of aprotective silica hydrogel shell on a Gel-MA microgel. (b) Optical microscopy images of Gel-MA microgels and silica-coated Gel-MA microgels. (D)(a) Fluorescence images of CSP cells on Gel-MA microgels (left) and Gel-MA microgels coated with silica hydrogel (right), subjected to oxidativestress. The cells were stained with calcein-AM (green) and ethidium homodimer-1 (red) to visualize live and dead cells, respectively. Scale bar = 50μm. (b) Viability of the CSP cells was quantified by the percentage of live cells (*p < 0.05). Adapted from ref. 94 with permission from the AmericanChemical Society, Copyright 2014.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



using a PDMS-based microfluidic flow-focusing device, withthe dispersed aqueous phase consisting of Gel-MA and aphotoinitiator, and the continuous oil phase consisting ofmineral oil supplemented with a surfactant (Fig. 10A). As thedroplets exited the outlet of the device and passed throughplastic connection tubing, they were polymerized in situ un-der 850 mW UV light for 5 min. The elastic modulus of theresulting microgels, as determined by AFM-assisted nano-in-dentation, was 1.87 kPa (±0.23), suggesting that they are ca-pable of withstanding shear stress upon injection from a sur-gical syringe. CSP cells were cultured on Gel-MA microgels

with a diameter of 100 μm, which provide sufficient surfacearea for multiple cell attachment while also being smallenough in size to be injected through syringe needles(Fig. 10B-a, b). When the cell-laden microgels were placed intissue culture plates, the cells migrated to the plastic surfaceand proliferated to cover the entire surface, demonstratingthe ability of these microgel-cultured CSP cells to spread tothe surrounding environment, which is a vital step followingcell transplantation to target tissues in a wound site(Fig. 10B-c). Lastly, the Gel-MA microgels were coated with asol–gel-generated silica hydrogel (Fig. 10C), which acts as a

Fig. 11 (A) Microfluidic generation of microgel building blocks for the creation of MAP scaffolds: (a) schematic illustrating microgel formationusing a microfluidic W/O emulsion system. Pre-gel and crosslinker solutions are segmented into monodisperse droplets followed by in-dropletmixing and crosslinking via Michael addition; (b) microgels are purified into an aqueous solution and annealed using FXIIIa to generate a micropo-rous scaffold, either in the presence of cells or as a pure scaffold; (c) fluorescence images showing purified microgel building blocks (left) and asubsequent cell-laden MAP scaffold (right); (d) MAP scaffolds are moldable to macroscale shapes, and can be injected to form complex shapes thatare maintained after annealing; (e) this process can be performed in the presence of live cells. (B) MAP scaffolds promote fast wound closure inSKH1-Hrhr and Balb/c epidermal mouse models: (a) quantification of wound closure over a five day period shows statistically significant wound-closure rates for MAP scaffolds when compared with non-porous bilateral controls (ND6). Statistical significance attained using standard two-tailed t-test (*p < 0.05); (b) representative images of wound closure during a five day in vivo wound-healing model in SKH1-Hrhr mice; (c) repre-sentative images of wound closure during seven day in vivo Balb/c experiments; (d) quantification of wound-closure data from Balb/c in vivowound healing. Adapted from ref. 157 with permission from Nature Publishing Group.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



shell that protects the cells on the surface of the microgelfrom exposure to reactive oxidative species, mechanical stressduring the injection process, and the host immune re-sponse, which are common external factors responsible forlow post-transplantation cell viability.174,175 Upon exposureto an oxidative stress environment, the CSP cells protectedby the silica hydrogel shell only showed an 8% decrease inviability, compared to a 50% decrease for CSP cells withoutthe protective shell (Fig. 10D), suggesting that the cell-seeded microgels combined with a silica hydrogel shell arepromising injectable tissue constructs for applications inwound healing.

Griffin et al. fabricated microporous annealed particle(MAP) gels by using the activated enzyme Factor XIII (FXIIIa)to anneal microfluidics-generated microgel building blocksinto a scaffold containing interconnected microporous net-works.176 The microgels were produced in a PDMS flow-focusing device, with the dispersed aqueous phase composedof a mixture of bioactive 4-arm PEG-vinyl sulphone (PEG-VS),a cell-adhesive peptide (RGD), transglutaminase peptide sub-strates (K and Q), and MMP-sensitive peptide sequences, andthe continuous oil phase consisting of mineral oilsupplemented with a surfactant. Serving as building blocks,the microgels were annealed to each other via FXIIIa-mediated non-canonical amide linkage between peptides Kand Q (Fig. 11A). Prior to scaffold annealing, the FXIIIa-supplemented microgel mixture containing live cells can beinjected into a wound site and solidify into the shape of thecavity (Fig. 11A-d). After annealing of the MAP scaffolds, thecells began to spread and form extensive networks in vitro. Amurine skin wound-healing model was used to test the abilityof the MAP scaffolds to support in vivo cell migration andbulk tissue integration. After five days post-injection, 60% ofthe wound area remained in the group with MAP scaffoldscompared to 100% for the non-porous control (Fig. 11B),demonstrating that MAP scaffolds facilitate significantlymore rapid wound closure. In a seven day in vivo wound-healing experiment on BALB/c mice, the MAP scaffolds con-tributed to 39% wound closure, compared to only 19% forthe no-treatment control, 7% for the physically matched non-porous control, and 10% for non-annealing microgels,suggesting that microgel annealing is essential to rapidwound closure. Also, unannealed microporous gels fabricatedusing a porogen-based casting method led to 27% wound clo-sure, indicating that micron-sized pores greatly facilitatewound healing in vivo. Furthermore, the interconnected net-work of micropores in the MAP scaffolds promoted bulk tis-sue integration, as evidenced by the co-localization of endo-thelial cells and pericytes in the scaffolds, which signals thedevelopment of a vascular network. In summary, MAP scaf-folds have the following advantages: 1) their injectability en-ables them to completely fill wounds of any shape and size;2) their microporosity facilitates rapid cellular invasion andnetwork formation; 3) the modular nature of their assemblyallows for the fabrication of microgel building blocks with adiverse range of chemical and physical properties that can be

used for tissue regeneration in a wide range of physiologicalniches.

8. Potential applications of cell-ladenmicrofluidic microgels in tissueregeneration

The advantages of microfluidics-generated cell-laden micro-gels, such as uniform size, could make it possible to over-come existing technical challenges in 3D bioprinting. Basedon layer-by-layer deposition, 3D bioprinting offers great con-trol over the spatial patterning of functional components.Using this technology, artificial tissues with functions thatclosely mimic those of native tissue can be generated by rep-licating the intricate interaction between cells and theirmicroenvironment on the microscale. Tissue constructs canbe assembled from building blocks composed of either a cell-laden hydrogel precursor solution or cell-encapsulatedmicrospheres.

Upon activation by thermal-, photo- or pH-mediated mech-anisms, hydrogel solutions containing cells can becrosslinked in situ as they are injected through the orifice ofthe printing nozzle.175 Although the small inner diameter ofthe nozzle allows for x–y resolutions of 50 μm or less, itcauses a pressure build-up that could reach 20 psi orgreater,177,178 which leads to substantial mechanical stressthat damages the cells and reduces their physiological func-tions. Furthermore, since cells are prepared and dispersed inan aqueous precursor solution, there is a loss of cell–cell andcell–substrate contact during and after the printing process.A long period of culture, often a week or more, is requiredfor the cells to degrade the hydrogels and gradually reformcell–cell junctions. As a result, cells that require stable cell–cell and cell–substrate signaling may lose their functions andundergo cell de-differentiation or apoptosis.

The microsphere building-block approach, in which thebasic units are composed of cell spheroids or microspheresfully coated with cells, enables the generation of micro-spheres containing tens of thousands of cells. The micro-spheres are extruded through a large nozzle with an inner di-ameter of hundreds of microns, which significantly mitigatesthe amount of shear stress and leads to increased cell viabil-ity.175 Moreover, cells on the microspheres have alreadyformed stable cell–cell junctions and cell–substrate adhesion,so ECM proteins such as fibronectin, collagen, and lamininare deposited on the microspheres,162 leading to improvedcellular activity, viability, and tissue functions.

An existing problem with using microspheres as buildingblocks for 3D bioprinting is the inability to produce homoge-neous microspheres and cell spheroids. Variation in micro-sphere size results in non-uniform cell dispersion and leadsto unstable tissue structures, as microspheres with differentsizes will form unstable stacks that are prone to movementor shift, which reduces both the mechanical strength of theprinted tissue and its spatial resolution. A promising


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



alternative is microfluidic cell-laden microgels, whose mono-dispersity allows them to be closely packed into multiplelayers, thereby forming a tissue construct that has good me-chanical stability as well as homogeneous physicochemicalproperties. The ability of microfluidics to produce microgelswith varying biochemical and mechanical properties, com-bined with 3D bioprinting's control over the spatial assemblyof functional components, may facilitate the generation ofmacroscale biomimetic tissue constructs for organ transplan-tation or for in vitro drug testing.

9. Concluding remarks and futureperspectives

Cell-laden scaffolds for tissue regeneration should possessgood biocompatibility, controlled degradability, mechanicalstability, uniform size, a conducive cellular microenviron-ment with appropriate biochemical and biophysical cues,and appropriate mass transport properties. Combined with ajudicious selection of biomaterials, microfluidic technologiescan be used to generate monodisperse cell-laden microgelswith easily tunable sizes, biophysical properties, and bio-chemical cues. These microgels may serve as building blocksthat can be assembled into mesoscale tissue structures withfunctions that closely mimic those of native tissue. However,several challenges must be overcome before clinical imple-mentation of microfluidic microgels can be realized.

One of the foremost concerns is the scale-up demand ofbio-manufacturing macroscale tissue assemblies.179 For ex-ample, the average weight of an adult human liver is approxi-mately 1.56 kg,180 with around 139 million cells per gram ofliver,181 which means that nearly 220 billion cells in totalwould be required to regenerate an entire functional liver. As-suming that cell-laden microgels can be generated at a rateof ∼1 kHz,10 and that the average encapsulation density is∼25 cells per microgel, then non-stop microgel productionfrom 100 microfluidic devices running in parallel for an en-tire day would be required to achieve such a feat. In a previ-ous study by Nisisako et al., a mass-production module com-posed of 256 O/W emulsion-forming units was used toproduce monodisperse acrylic microgels at a rate of ∼0.3 kgper hour, demonstrating scale-up potential.110 However, forcomplex droplet-forming devices that demand high-resolution spatial patterning of wettability,182,183 more robustdesigns are needed to ensure controlled droplet formationduring large-scale parallelization. A few groups have devel-oped complex devices that obviate the need for localized sur-face modification of channel hydrophilicity, but they are lim-ited either by low droplet production rates or by the types ofemulsions that can be generated.184,185 Therefore, there is aneed to develop devices that can support large-scale produc-tion of microgels with complex geometries, such as core–shell morphology.

Another issue impeding the application of cell-ladenmicrogels in tissue implantation is the lack of proper vascu-larization. Biomimetic tissue constructs should be composed

of precisely organized components in 3D so that different celltypes, ECMs, and vasculatures can properly interact with eachother to maintain normal tissue functions.186 Out of thesecomponents, a well-perfused microvasculature is perhaps themost vital to the viability of the tissue construct, as cells needto be within 200 μm of a blood vessel, which facilitates effi-cient exchange of nutrients and waste as well as transport ofgrowth factors and cellular signals.187 Therefore, organizedvascular networks must be incorporated into cell-ladenmicrogels before they are injected into a defect site so thatthe regenerated tissue can readily integrate with the host vas-culature. Proper vascularization requires precise spatial andtemporal patterning of different cell types and chemical cuesthat control the vascular organization and remodeling pro-cesses.188 Various methods have been demonstrated for fabri-cating vascular networks in hydrogels, such as using gelatinas a sacrificial material to form hollow channels189 or usingPDMS molds to form collagen-filled microchannels that areseeded with endothelial cells.190 In order to reliably producemicrogels with branched vascular networks, strategies needto be developed for adapting these techniques to a miniatur-ized microfluidic device and automating the series of stepsthat are involved in the fabrication process.

The development of whole-organ decellularization tech-niques during the past decade may mitigate the vasculariza-tion dilemma and enable the fabrication of microfluidicmicrogels that can maximally recreate the complex architec-ture and functionality of native tissue. After removal of allcellular and immunogenic species, the remaining ECM of thedecellularized organ not only retains relevant biological sig-nals but also preserves the microarchitecture of the native tis-sue, including an intact vascular system that could easilyintegrate with the patient's circulatory system.191,192 Since thespecific composition and organization of the ECM depend onthe tissue source from which the ECM is derived,193 the useof tissue-specific ECMs as scaffolds for tissue regenerationapplications may result in biomimetic tissue constructs thatbetter recapitulate the functions of the native tissue. There-fore, one can envision the incorporation of tissue-specificdecellularized ECMs into a hydrogel precursor solution toproduce microfluidic microgels that not only contain cells en-capsulated within a physiologically relevant microenviron-ment but also promote the bulk integration of the vascularsystem with surrounding tissue. In concert with advances intissue engineering and microfluidic technology, the clinicalapplication of cell-laden microfluidic microgels for tissue re-generation sees a bright future.

Acknowledgements

Funding support from NIH (HL109442, AI096305,GM110494), the Guangdong Innovative and EntrepreneurialResearch Team Program (No. 2013S086), and the Global Re-search Laboratory Program (Korean NSF GRL; 2015032163) isalso acknowledged.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



References

1 R. Langer and J. P. Vacanti, Science, 1993, 260, 920–926.2 M. P. Lutolf and J. A. Hubbell, Nat. Biotechnol., 2005, 23,

47–55.3 K. Y. Lee and D. J. Mooney, Chem. Rev., 2001, 101,

1869–1877.4 B. V. Slaughter, S. S. Khurshid, O. Z. Fisher, A.

Khademhosseini and N. A. Peppas, Adv. Mater., 2009, 21,3307–3329.

5 G. D. Nicodemus and S. J. Bryant, Tissue Eng., Part B,2008, 14, 149–165.

6 R. Chen, S. J. Curran, J. M. Curran and J. A. Hunt,Biomaterials, 2006, 27, 4453–4460.

7 X. Zhao, S. Liu, L. Yildirimer, H. Zhao, R. Ding, H. Wang,W. Cui and D. Weitz, Adv. Funct. Mater., 2016, 26,2809–2819.

8 F. Yanagawa, H. Kaji, Y. H. Jang, H. Bae, D. Yanan, J.Fukuda, H. Qi and A. Khademhosseini, J. Biomed. Mater.Res., Part A, 2011, 97, 93–102.

9 C. Ma, C. Tian, L. Zhao and J. Wang, Lab Chip, 2016, 16,2609–2617.

10 C. Siltanen, M. Yaghoobi, A. Haque, J. You, J. Lowen, M.Soleimani and A. Revzin, Acta Biomater., 2016, 34, 125–132.

11 D. Seliktar, Science, 2012, 336, 1124–1128.12 S. Allazetta and M. P. Lutolf, Curr. Opin. Biotechnol.,

2015, 35, 86–93.13 J. Malda, J. Visser, F. P. Melchels, T. Jungst, W. E. Hennink,

W. J. Dhert, J. Groll and D. W. Hutmacher, Adv. Mater.,2013, 25, 5011–5028.

14 C. Ghobril and M. W. Grinstaff, Chem. Soc. Rev., 2015, 44,1820–1835.

15 T. R. Hoare and D. S. Kohane, Polymer, 2008, 49,1993–2007.

16 A. C. H. Pape and P. Y. W. Dankers, SupramolecularPolymer Networks and Gels, ed. S. Seiffert, SpringerInternational Publishing, 2015, pp. 253–279.

17 Y. Jiang, J. Chen, C. Deng, E. J. Suuronen and Z. Zhong,Biomaterials, 2014, 35, 4969–4985.

18 D. Velasco, E. Tumarkin and E. Kumacheva, Small, 2012, 8,1633–1642.

19 S. K. Leslie, A. M. Nicolini, G. Sundaresan, J. Zweit, B. D.Boyan and Z. Schwartz, J. Mater. Chem. B, 2016, 4,3515–3525.

20 L. Mazutis, R. Vasiliauskas and D. A. Weitz, Macromol.Biosci., 2015, 15, 1641–1646.

21 Q. Chen, S. Utech, D. Chen, R. Prodanovic, J. M. Lin andD. A. Weitz, Lab Chip, 2016, 16, 1346–1349.

22 H. Lee, S. Chung, M. G. Kim, L. P. Lee and J. Y. Lee, Adv.Healthcare Mater., 2016, 5, 1638–1645.

23 Y. B. Lee, Y. M. Shin, E. M. Kim, J. Lim, J.-Y. Lee and H.Shin, Adv. Healthcare Mater., 2016, 5, 2320–2324.

24 Q. H. Xu, C. L. He, K. X. Ren, C. S. Xiao and X. S. Chen,Adv. Healthcare Mater., 2016, 5, 1979–1990.

25 Y. L. Cheng, C. L. He, J. X. Ding, C. S. Xiao, X. L. Zhuangand X. S. Chen, Biomaterials, 2013, 34, 10338–10347.

26 E. Tumarkin and E. Kumacheva, Chem. Soc. Rev., 2009, 38,2161–2168.

27 M. Anthamatten, Supramolecular Polymer Networks andGels, ed. S. Seiffert, Springer International Publishing,2015, pp. 47–99.

28 F. M. Xu, H. B. Wang, J. Zhao, X. S. Liu, D. D. Li, C. J. Chenand J. Ji, Macromolecules, 2013, 46, 4235–4246.

29 Y. Chen, N. Ballard, F. Gayet and S. A. F. Bon, Chem.Commun., 2012, 48, 1117–1119.

30 A. Sionkowska, Prog. Polym. Sci., 2011, 36, 1254–1276.31 K. T. Nguyen and J. L. West, Biomaterials, 2002, 23,

4307–4314.32 N. E. Fedorovich, M. H. Oudshoorn, D. van Geemen, W. E.

Hennink, J. Alblas and W. J. Dhert, Biomaterials, 2009, 30,344–353.

33 Y. Fu and W. J. Kao, J. Biomed. Mater. Res., Part A, 2011, 98,201–211.

34 N. J. Darling, Y. S. Hung, S. Sharma and T. Segura,Biomaterials, 2016, 101, 199–206.

35 N. E. Davis, S. Ding, R. E. Forster, D. M. Pinkas and A. E.Barron, Biomaterials, 2010, 31, 7288–7297.

36 R. Jin, L. S. Moreira Teixeira, P. J. Dijkstra, Z. Zhong, C. A.van Blitterswijk, M. Karperien and J. Feijen, Tissue Eng.,Part A, 2010, 16, 2429–2440.

37 R. Jin, L. S. Teixeira, P. J. Dijkstra, C. A. van Blitterswijk, M.Karperien and J. Feijen, Biomaterials, 2010, 31, 3103–3113.

38 C. W. Yung, W. E. Bentley and T. A. Barbari, J. Biomed.Mater. Res., Part A, 2010, 95, 25–32.

39 L. S. Teixeira, J. Feijen, C. A. van Blitterswijk, P. J. Dijkstraand M. Karperien, Biomaterials, 2012, 33, 1281–1290.

40 J. Thiele, Y. Ma, S. M. C. Bruekers, S. Ma and W. T. S.Huck, Adv. Mater., 2014, 26, 125–148.

41 C. Ghobril and M. W. Grinstaff, Chem. Soc. Rev., 2015, 44,1820–1835.

42 F. Topuz, A. Henke, W. Richtering and J. Groll, Soft Matter,2012, 8, 4877–4881.

43 Y. A. Morch, I. Donati, B. L. Strand and G. Skjak-Braek,Biomacromolecules, 2006, 7, 1471–1480.

44 W. H. Tan and S. Takeuchi, Adv. Mater., 2007, 19,2696–2701.

45 R.-C. Luo and C.-H. Chen, Soft, 2012, 01, 1–23.46 H. Zhang, E. Tumarkin, R. M. A. Sullan, G. C. Walker and E.

Kumacheva, Macromol. Rapid Commun., 2007, 28, 527–538.47 D. Velasco, E. Tumarkin and E. Kumacheva, Small, 2012, 8,

1633–1642.48 H. Zhang, E. Tumarkin, R. Peerani, Z. Nie, R. M. A. Sullan,

G. C. Walker and E. Kumacheva, J. Am. Chem. Soc.,2006, 128, 12205–12210.

49 N. Raz, J. K. Li, L. K. Fiddes, E. Tumarkin, G. C. Walker andE. Kumacheva, Macromolecules, 2010, 43, 7277–7281.

50 P. Sabhachandani, V. Motwani, N. Cohen, S. Sarkar, V.Torchilin and T. Konry, Lab Chip, 2016, 16, 497–505.

51 S. Akbari and T. Pirbodaghi, Microfluid. Nanofluid.,2013, 16, 773–777.

52 C. Kim, S. Chung, Y. E. Kim, K. S. Lee, S. H. Lee, K. W. Ohand J. Y. Kang, Lab Chip, 2011, 11, 246–252.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



53 R. K. Shah, J.-W. Kim, J. J. Agresti, D. A. Weitz and L.-Y.Chu, Soft Matter, 2008, 4, 2303–2309.

54 K. Liu, H. J. Ding, J. Liu, Y. Chen and X. Z. Zhao, Langmuir,2006, 22, 9453–9457.

55 Y. Li, J. Rodrigues and H. Tomas, Chem. Soc. Rev., 2012, 41,2193–2221.

56 J. A. Burdick and G. D. Prestwich, Adv. Mater., 2011, 23,H41–H56.

57 H. Tan, C. M. Ramirez, N. Miljkovic, H. Li, J. P. Rubin andK. G. Marra, Biomaterials, 2009, 30, 6844–6853.

58 J. A. Burdick and G. D. Prestwich, Adv. Mater., 2011, 23,H41–H56.

59 L. Bian, C. Hou, E. Tous, R. Rai, R. L. Mauck and J. A.Burdick, Biomaterials, 2013, 34, 413–421.

60 E. Sideris, D. R. Griffin, Y. Ding, S. Li, W. M. Weaver, D. DiCarlo, T. Hsiai and T. Segura, ACS Biomater. Sci. Eng.,2016, DOI: 10.1021/acsbiomaterials.6b00444.

61 A. Kumachev, J. Greener, E. Tumarkin, E. Eiser, P. W.Zandstra and E. Kumacheva, Biomaterials, 2011, 32,1477–1483.

62 Y. Ling, J. Rubin, Y. Deng, C. Huang, U. Demirci, J. M. Karpand A. Khademhosseini, Lab Chip, 2007, 7, 756–762.

63 E. Tumarkin, L. Tzadu, E. Csaszar, M. Seo, H. Zhang, A.Lee, R. Peerani, K. Purpura, P. W. Zandstra and E.Kumacheva, Integr. Biol., 2011, 3, 653–662.

64 D. Puppi, F. Chiellini, A. M. Piras and E. Chiellini, Prog.Polym. Sci., 2010, 35, 403–440.

65 R. De Souza, P. Zahedi, C. J. Allen and M. Piquette-Miller,Biomaterials, 2009, 30, 3818–3824.

66 N. Bhattarai, J. Gunn and M. Zhang, Adv. Drug DeliveryRev., 2010, 62, 83–99.

67 S. T. Koev, P. H. Dykstra, X. Luo, G. W. Rubloff, W. E.Bentley, G. F. Payne and R. Ghodssi, Lab Chip, 2010, 10,3026–3042.

68 M. Fan, Y. Ma, J. Mao, Z. Zhang and H. Tan, Acta Biomater.,2015, 20, 60–68.

69 H. K. Holme, L. Davidsen, A. Kristiansen and O. Smidsrod,Carbohydr. Polym., 2008, 73, 656–664.

70 L. G. Griffith and M. A. Swartz, Nat. Rev. Mol. Cell Biol.,2006, 7, 211–224.

71 B.-S. Kim, I.-K. Park, T. Hoshiba, H.-L. Jiang, Y.-J. Choi, T.Akaike and C.-S. Cho, Prog. Polym. Sci., 2011, 36, 238–268.

72 J. Glowacki and S. Mizuno, Biopolymers, 2008, 89, 338–344.73 K. Domansky, W. Inman, J. Serdy, A. Dash, M. H. Lim and

L. G. Griffith, Lab Chip, 2010, 10, 51–58.74 C. T. Ho, R. Z. Lin, R. J. Chen, C. K. Chin, S. E. Gong, H. Y.

Chang, H. L. Peng, L. Hsu, T. R. Yew, S. F. Chang and C. H.Liu, Lab Chip, 2013, 13, 3578–3587.

75 S. A. Lee, Y. No da, E. Kang, J. Ju, D. S. Kim and S. H. Lee,Lab Chip, 2013, 13, 3529–3537.

76 D. Wallace, Adv. Drug Delivery Rev., 2003, 55, 1631–1649.77 M. D. Brigham, A. Bick, E. Lo, A. Bendali, J. A. Burdick and

A. Khademhosseini, Tissue Eng., Part A, 2009, 15,1645–1653.

78 C. Moraes, A. B. Simon, A. J. Putnam and S. Takayama,Biomaterials, 2013, 34, 9623–9631.

79 S. Hong, H. J. Hsu, R. Kaunas and J. Kameoka, Lab Chip,2012, 12, 3277–3280.

80 S. L. Rowe, S. Lee and J. P. Stegemann, Acta Biomater.,2007, 3, 59–67.

81 B. C. Lechtenberg, D. J. Johnson, S. M. Freund and J. A.Huntington, Proc. Natl. Acad. Sci. U. S. A., 2010, 107,14087–14092.

82 H. E. Davis, S. L. Miller, E. M. Case and J. K. Leach, ActaBiomater., 2011, 7, 691–699.

83 J. W. Nichol, S. T. Koshy, H. Bae, C. M. Hwang, S. Yamanlarand A. Khademhosseini, Biomaterials, 2010, 31, 5536–5544.

84 Y. Li, H. Meng, Y. Liu, A. Narkar and B. P. Lee, ACS Appl.Mater. Interfaces, 2016, 8, 11980–11989.

85 K. Sisson, C. Zhang, M. C. Farach-Carson, D. B. Chase andJ. F. Rabolt, Biomacromolecules, 2009, 10, 1675–1680.

86 Y. C. Chen, R. Z. Lin, H. Qi, Y. Yang, H. Bae, J. M. Melero-Martin and A. Khademhosseini, Adv. Funct. Mater.,2012, 22, 2027–2039.

87 H. Aubin, J. W. Nichol, C. B. Hutson, H. Bae, A. L.Sieminski, D. M. Cropek, P. Akhyari and A.Khademhosseini, Biomaterials, 2010, 31, 6941–6951.

88 H. Qi, Y. Du, L. Wang, H. Kaji, H. Bae and A.Khademhosseini, Adv. Mater., 2010, 22, 5276–5281.

89 J. Zhu, Biomaterials, 2010, 31, 4639–4656.90 J. L. Drury and D. J. Mooney, Biomaterials, 2003, 24,

4337–4351.91 C. Lin, P. Zhao, F. Li, F. Guo, Z. Li and X. Wen, Mater. Sci.

Eng., C, 2010, 30, 1236–1244.92 L. Yu and J. Ding, Chem. Soc. Rev., 2008, 37, 1473–1481.93 S. Y. Lee and G. Tae, J. Controlled Release, 2007, 119,

313–319.94 A. Sosnik and D. Cohn, Biomaterials, 2004, 25,

2851–2858.95 E. J. Ricci, L. O. Lunardi, D. M. Nanclares and J. M.

Marchetti, Int. J. Pharm., 2005, 288, 235–244.96 D. Cohn, A. Sosnik and A. Levy, Biomaterials, 2003, 24,

3707–3714.97 D. Cohn, G. Lando, A. Sosnik, S. Garty and A. Levi,

Biomaterials, 2006, 27, 1718–1727.98 G. M. Whitesides, Nature, 2006, 442, 368–373.99 S. K. Sia and G. M. Whitesides, Electrophoresis, 2003, 24,

3563–3576.100 M. A. Unger, H. P. Chou, T. Thorsen, A. Scherer and S. R.

Quake, Science, 2000, 288, 113–116.101 S. L. Anna, N. Bontoux and H. A. Stone, Appl. Phys. Lett.,

2003, 82, 364–366.102 Y. Liu, H. Li, S. Yan, J. Wei and X. Li, Biomacromolecules,

2014, 15, 1044–1054.103 R. K. Shah, H. C. Shum, A. C. Rowat, D. Lee, J. J. Agresti,

A. S. Utada, L.-Y. Chu, J.-W. Kim, A. Fernandez-Nieves, C. J.Martinez and D. A. Weitz, Mater. Today, 2008, 11, 18–27.

104 C. H. Choi, H. Wang, H. Lee, J. H. Kim, L. Zhang, A. Mao,D. J. Mooney and D. A. Weitz, Lab Chip, 2016, 16,1549–1555.

105 E. P. Herrero, E. M. M. D. Valle and M. A. Galán,Biotechnol. Prog., 2007, 23, 940–945.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



106 D. S. Tawfik and A. D. Griffiths, Nat. Biotechnol., 1998, 16,652–656.

107 A. Huebner, S. Sharma, M. Srisa-Art, F. Hollfelder, J. B. Edeland A. J. Demello, Lab Chip, 2008, 8, 1244–1254.

108 A. S. Utada, E. Lorenceau, D. R. Link, P. D. Kaplan, H. A.Stone and D. A. Weitz, Science, 2005, 308, 537–541.

109 S. Xu, Z. Nie, M. Seo, P. Lewis, E. Kumacheva, H. A. Stone,P. Garstecki, D. B. Weibel, I. Gitlin and G. M. Whitesides,Angew. Chem., 2005, 117, 734–738.

110 T. Nisisako and T. Torii, Lab Chip, 2008, 8, 287–293.111 L. Deveza, J. Ashoken, G. Castaneda, X. Tong, M. Keeney,

L.-H. Han and F. Yang, ACS Biomater. Sci. Eng., 2015, 1,157–165.

112 C. K. Colton, Adv. Drug Delivery Rev., 2014, 67–68, 93–110.113 C. Cha, J. Oh, K. Kim, Y. Qiu, M. Joh, S. R. Shin, X. Wang,

G. Camci-Unal, K. T. Wan, R. Liao and A. Khademhosseini,Biomacromolecules, 2014, 15, 283–290.

114 P. C. Nicolson and J. Vogt, Biomaterials, 2001, 22,3273–3283.

115 N. A. Peppas, J. Z. Hilt, A. Khademhosseini and R. Langer,Adv. Mater., 2006, 18, 1345–1360.

116 S. N. Bhatia, U. J. Balis, M. L. Yarmush and M. Toner,FASEB J., 1999, 13, 1883–1900.

117 J. K. Oh, R. Drumright, D. J. Siegwart and K. Matyjaszewski,Prog. Polym. Sci., 2008, 33, 448–477.

118 Z. H. Nie, S. Q. Xu, M. Seo, P. C. Lewis and E. Kumacheva,J. Am. Chem. Soc., 2005, 127, 8058–8063.

119 N. N. Kachouie, Y. Du, H. Bae, M. Khabiry, A. F. Ahari, B.Zamanian, J. Fukuda and A. Khademhosseini,Organogenesis, 2010, 6, 234–244.

120 A. Khademhosseini and N. A. Peppas, Adv. HealthcareMater., 2013, 2, 10–12.

121 Y. Tsuda, Y. Morimoto and S. Takeuchi, Langmuir, 2010, 26,2645–2649.

122 T. Rossow, J. A. Heyman, A. J. Ehrlicher, A. Langhoff, D. A.Weitz, R. Haag and S. Seiffert, J. Am. Chem. Soc., 2012, 134,4983–4989.

123 T. Rossow, S. Bayer, R. Albrecht, C. C. Tzschucke and S.Seiffert, Macromol. Rapid Commun., 2013, 34, 1401–1407.

124 Y. Shi, X. Gao, L. Chen, M. Zhang, J. Ma, X. Zhang and J.Qin, Microfluid. Nanofluid., 2013, 15, 467–474.

125 D. Steinhilber, T. Rossow, S. Wedepohl, F. Paulus, S. Seiffertand R. Haag, Angew. Chem., Int. Ed., 2013, 52, 13538–13543.

126 D. M. Headen, G. Aubry, H. Lu and A. J. Garcia, Adv. Mater.,2014, 26, 3003–3008.

127 S. Allazetta, L. Kolb, S. Zerbib, J. Bardy and M. P. Lutolf,Small, 2015, 11, 5647–5656.

128 S. Hackelbusch, T. Rossow, D. Steinhilber, D. A. Weitz andS. Seiffert, Adv. Healthcare Mater., 2015, 4, 1841–1848.

129 S. Utech, R. Prodanovic, A. S. Mao, R. Ostafe, D. J. Mooneyand D. A. Weitz, Adv. Healthcare Mater., 2015, 4, 1628–1633.

130 Y. Wang, L. Zhao, C. Tian, C. Ma and J. Wang, Anal.Methods, 2015, 7, 10040–10051.

131 R. Wieduwild, S. Krishnan, K. Chwalek, A. Boden, M.Nowak, D. Drechsel, C. Werner and Y. Zhang, Angew.Chem., Int. Ed., 2015, 54, 3962–3966.

132 B. G. Chung, K. H. Lee, A. Khademhosseini and S. H. Lee,Lab Chip, 2012, 12, 45–59.

133 S. Seiffert, ChemPhysChem, 2013, 14, 295–304.134 S. Seiffert, Angew. Chem., Int. Ed., 2013, 52, 11462–11468.135 W. Leong and D. A. Wang, Trends Biotechnol., 2015, 33,

653–666.136 P. Bianco and P. G. Robey, Nature, 2001, 414, 118–121.137 G. Q. Daley and D. T. Scadden, Cell, 2008, 132, 544–548.138 R. Passier, L. W. van Laake and C. L. Mummery, Nature,

2008, 453, 322–329.139 J. A. Thomson, J. Itskovitz-Eldor, S. S. Shapiro, M. A.

Waknitz, J. J. Swiergiel, V. S. Marshall and J. M. Jones,Science, 1998, 282, 1145–1147.

140 S. Levenberg, J. S. Golub, M. Amit, J. Itskovitz-Eldor and R.Langer, Proc. Natl. Acad. Sci. U. S. A., 2002, 99, 4391–4396.

141 J. M. Jukes, S. K. Both, A. Leusink, L. M. Sterk, C. A. vanBlitterswijk and J. de Boer, Proc. Natl. Acad. Sci. U. S. A.,2008, 105, 6840–6845.

142 J. L. Wilson and T. C. McDevitt, Biotechnol. Bioeng.,2013, 110, 667–682.

143 P. Agarwal, S. Zhao, P. Bielecki, W. Rao, J. K. Choi, Y.Zhao, J. Yu, W. Zhang and X. He, Lab Chip, 2013, 13,4525–4533.

144 L. Schukur, P. Zorlutuna, J. M. Cha, H. Bae and A.Khademhosseini, Adv. Healthcare Mater., 2013, 2, 195–205.

145 Y. Zhang, Y. P. Ho, Y. L. Chiu, H. F. Chan, B. Chlebina, T.Schuhmann, L. You and K. W. Leong, Biomaterials,2013, 34, 4564–4572.

146 H. F. Chan, Y. Zhang, Y. P. Ho, Y. L. Chiu, Y. Jung andK. W. Leong, Sci. Rep., 2013, 3, 3462.

147 F. Chen, Y. Zhan, T. Geng, H. Lian, P. Xu and C. Lu, Anal.Chem., 2011, 83, 8816–8820.

148 F. Kitagawa, S. Takei, T. Imaizumi and Y. Tabata, TissueEng., Part C, 2013, 19, 438–448.

149 S. Sakai, S. Ito, H. Inagaki, K. Hirose, T. Matsuyama, M.Taya and K. Kawakami, Biomicrofluidics, 2011, 5, 13402.

150 Y. C. Tung, A. Y. Hsiao, S. G. Allen, Y. S. Torisawa, M. Hoand S. Takayama, Analyst, 2011, 136, 473–478.

151 G. S. Jeong, J. H. Song, A. R. Kang, Y. Jun, J. H. Kim, J. Y.Chang and S. H. Lee, Adv. Healthcare Mater., 2013, 2,119–125.

152 S. Yoon, J. A. Kim, S. H. Lee, M. Kim and T. H. Park, LabChip, 2013, 13, 1522–1528.

153 E. S. Place, N. D. Evans and M. M. Stevens, Nat. Mater.,2009, 8, 457–470.

154 V. Vogel and M. Sheetz, Nat. Rev. Mol. Cell Biol., 2006, 7,265–275.

155 T. P. Kraehenbuehl, R. Langer and L. S. Ferreira, Nat.Methods, 2011, 8, 731–736.

156 S. R. Khetani and S. N. Bhatia, Nat. Biotechnol., 2008, 26,120–126.

157 B. E. Uygun, A. Soto-Gutierrez, H. Yagi, M. L. Izamis, M. A.Guzzardi, C. Shulman, J. Milwid, N. Kobayashi, A. Tilles, F.Berthiaume, M. Hertl, Y. Nahmias, M. L. Yarmush and K.Uygun, Nat. Med., 2010, 16, 814–820.

158 S. Mitragotri and J. Lahann, Nat. Mater., 2009, 8, 15–23.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online



159 Y. K. Cheung, B. M. Gillette, M. Zhong, S. Ramcharan andS. K. Sia, Lab Chip, 2007, 7, 574–579.

160 N. C. Rivron, J. Rouwkema, R. Truckenmuller, M.Karperien, J. De Boer and C. A. Van Blitterswijk,Biomaterials, 2009, 30, 4851–4858.

161 Y. Cheng, F. Zheng, J. Lu, L. Shang, Z. Xie, Y. Zhao, Y. Chenand Z. Gu, Adv. Mater., 2014, 26, 5184–5190.

162 H. F. Chan, Y. Zhang and K. W. Leong, Small, 2016, 12,2720–2730.

163 R. Z. Lin and H. Y. Chang, Biotechnol. J., 2008, 3,1172–1184.

164 A. A. Tomei, V. Manzoli, C. A. Fraker, J. Giraldo, D. Velluto,M. Najjar, A. Pileggi, R. D. Molano, C. Ricordi, C. L. Stablerand J. A. Hubbell, Proc. Natl. Acad. Sci. U. S. A., 2014, 111,10514–10519.

165 B. Struecker, N. Raschzok and I. M. Sauer, Nat. Rev.Gastroenterol. Hepatol., 2014, 11, 166–176.

166 M. A. Laflamme, K. Y. Chen, A. V. Naumova, V. Muskheli,J. A. Fugate, S. K. Dupras, H. Reinecke, C. Xu, M.Hassanipour, S. Police, C. O'Sullivan, L. Collins, Y. Chen, E.Minami, E. A. Gill, S. Ueno, C. Yuan, J. Gold and C. E.Murry, Nat. Biotechnol., 2007, 25, 1015–1024.

167 T. Kaully, K. Kaufman-Francis, A. Lesman and S.Levenberg, Tissue Eng., Part B, 2009, 15, 159–169.

168 D. E. Discher, D. J. Mooney and P. W. Zandstra, Science,2009, 324, 1673–1677.

169 J. Alijotas-Reig, M. T. Fernandez-Figueras and L. Puig, Clin.Rev. Allergy Immunol., 2013, 45, 97–108.

170 J. Patterson, R. Siew, S. W. Herring, A. S. Lin, R.Guldberg and P. S. Stayton, Biomaterials, 2010, 31,6772–6781.

171 W. E. Hennink and C. F. van Nostrum, Adv. Drug DeliveryRev., 2012, 64, 223–236.

172 J. L. Holloway, H. Ma, R. Rai and J. A. Burdick, J. ControlledRelease, 2014, 191, 63–70.

173 B. P. Purcell, D. Lobb, M. B. Charati, S. M. Dorsey, R. J.Wade, K. N. Zellars, H. Doviak, S. Pettaway, C. B. Logdon,J. A. Shuman, P. D. Freels, J. H. Gorman, 3rd, R. C.Gorman, F. G. Spinale and J. A. Burdick, Nat. Mater.,2014, 13, 653–661.

174 M. A. Laflamme and C. E. Murry, Nat. Biotechnol., 2005, 23,845–856.

175 S. V. Murphy and A. Atala, Nat. Biotechnol., 2014, 32,773–785.

176 D. R. Griffin, W. M. Weaver, P. O. Scumpia, D. Di Carlo andT. Segura, Nat. Mater., 2015, 14, 737–744.

177 C. M. Smith, J. J. Christian, W. L. Warren and S. K.Williams, Tissue Eng., 2007, 13, 373–383.

178 C. C. Chang, E. D. Boland, S. K. Williams and J. B. Hoying,J. Biomed. Mater. Res., Part B, 2011, 98, 160–170.

179 H. F. Chan, S. Ma and K. W. Leong, Regener. Biomater.,2016, 3, 87–98.

180 D. K. Molina and V. J. DiMaio, Am. J. Forensic Med. Pathol.,2012, 33, 368–372.

181 A. K. Sohlenius-Sternbeck, Toxicol. In Vitro, 2006, 20,1582–1586.

182 A. R. Abate, A. T. Krummel, D. Lee, M. Marquez, C. Holtzeand D. A. Weitz, Lab Chip, 2008, 8, 2157–2160.

183 A. R. Abate, J. Thiele, M. Weinhart and D. A. Weitz, LabChip, 2010, 10, 1774–1776.

184 T. Nisisako, T. Ando and T. Hatsuzawa, Lab Chip, 2012, 12,3426–3435.

185 M. B. Romanowsky, A. R. Abate, A. Rotem, C. Holtze andD. A. Weitz, Lab Chip, 2012, 12, 802–807.

186 Y. Zheng and M. A. Roberts, Nat. Mater., 2016, 15, 597–599.187 R. K. Jain, P. Au, J. Tam, D. G. Duda and D. Fukumura, Nat.

Biotechnol., 2005, 23, 821–823.188 J. Rouwkema and A. Khademhosseini, Trends Biotechnol.,

2016, 34, 733–745.189 A. P. Golden and J. Tien, Lab Chip, 2007, 7, 720–725.190 Y. Zheng, J. Chen, M. Craven, N. W. Choi, S. Totorica, A.

Diaz-Santana, P. Kermani, B. Hempstead, C. Fischbach-Teschl, J. A. Lopez and A. D. Stroock, Proc. Natl. Acad. Sci.U. S. A., 2012, 109, 9342–9347.

191 J. E. Arenas-Herrera, I. K. Ko, A. Atala and J. J. Yoo, Biomed.Mater., 2013, 8, 014106.

192 A. Khademhosseini and R. Langer, Nat. Protoc., 2016, 11,1775–1781.

193 S. F. Badylak, D. Taylor and K. Uygun, Annu. Rev. Biomed.Eng., 2011, 13, 27–53.


Publ

ishe

d on

20

Oct

ober

201

6. D

ownl

oade

d by

Col

umbi

a U

nive

rsity

on

01/0

7/20

17 1

6:22

:32.

View Article Online


Documents

Lab on a Chip - Columbia BMEorion.bme.columbia.edu/leonglab/publications/pdf/... · Lab on a Chip CRITICAL REVIEW Cite this: Lab Chip,2016,16,4482 Received 22nd September 2016,