Acta Biomaterialia 10 (2014) 1965–1974

Contents lists available at ScienceDirect

Acta Biomaterialia journal homepage: www.elsevier.com/locate/actabiomat

Galactosylated reversible hydrogels as scaffold for HepG2 spheroid generation Yuhan Wu, Ziqi Zhao, Ying Guan, Yongjun Zhang ⇑ State Key Laboratory of Medicinal Chemical Biology and Key Laboratory of Functional Polymer Materials, Institute of Polymer Chemistry, College of Chemistry, Nankai University, and Collaborative Innovation Center of Chemical Science and Engineering (Tianjin), Tianjin 300071, China

a r t i c l e

i n f o

Article history: Received 26 July 2013 Received in revised form 28 November 2013 Accepted 18 December 2013 Available online 29 December 2013 Keywords: Multicellular spheroids Hepatocytes Galactose ligands Reversible hydrogels Microgels

a b s t r a c t Various galactosylated scaffolds have been developed for hepatocyte culture because galactose ligands help maintain cell viability, facilitate the formation of multicellular spheroids and help maintain a high level of liver-specific functions. However, it is difficult to harvest the cell spheroids generated inside the three-dimensional scaffolds for their further biological analysis and applications. Here we developed a new galactosylated hydrogel scaffold which solidifies in situ upon heating to physiological temperature, but liquefies again upon cooling back to room temperature. The new scaffold is composed of poly(N-isopropylacrylamide) (PNIPAM) microgel and poly(ethylene glycol) (PEG). Because of the thermosensitivity of PNIPAM microgel, the mixed dispersions gel upon heating and liquefy upon cooling. PEG was added to reduce the shrinkage of the gels. Part of the PNIPAM microgel was replaced with a galactosylated one to provide a series of blend gels with various galactose ligand contents. HepG2 cells, a human hepatocarcinoma cell line, were encapsulated in the in situ-formed gels. As expected, the cell viability increases with increasing content of galactose ligands. In addition, compact multicellular spheroids were obtained in gels containing galactose ligands, while loose spheroids formed in gel without galactose ligands. The cells cultured in galactose-containing gels also exhibit a higher level of liver-specific functions, in terms of both albumin secretion and urea synthesis, than those cultured in gel without these ligands. The new galactosylated scaffold not only promotes the formation of hepatocyte spheroids, but also allows for their harvest. By cooling back to room temperature to liquefy the gel, the hepatocyte spheroids can be facilely harvested from the scaffold. The reversible galactosylated scaffold developed here may be used for large scale fabrication of hepatocyte spheroids. Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved

1. Introduction Multicellular spheroids are three-dimensional (3-D) cell culture models with important applications in drug screening [1–3], tumor studies [4] and tissue engineering [5]. It has long been recognized that two-dimensional (2-D) cell monolayers cannot accurately model the in vivo environment, in which cell–cell and cell–extracellular matrix (ECM) interactions play a significant role in modulating the differentiation, proliferation and migration of the cells. As a result, tissue-specific properties are often lost in 2-D cell monolayers [6]. In contrast, 3-D cell spheroids, with improved cell–cell and cell–matrix interactions, can preserve complex in vivo cell phenotypes. Specifically 3-D spheroids of hepatocytes and hepatic cell lines usually exhibit a higher level of liver-specific functions [7,8]. These materials can be used in bioartificial liver devices, which can temporarily replace liver functions for patients suffering ⇑ Corresponding author. E-mail address: [email protected] (Y. Zhang). http://dx.doi.org/10.1016/j.actbio.2013.12.044 1742-7061/Ó 2013 Acta Materialia Inc. Published by Elsevier Ltd. All rights reserved

from acute liver failure [9]. They have also been developed as in vitro models for drug safety testing [10]. One effective method to generate 3-D hepatocyte spheroids is the employment of galactose-modified substrata, because the asialoglycoprotein receptors (ASGPRs) on the surface of hepatocytes selectively adhere to galactose ligands, and the interaction between ASGPRs and galactose ligands can induce the formation of hepatocyte aggregates [11]. Previously, Weigel studied the binding of rat hepatocytes to flat polyacrylamide surfaces containing galactose and found that cell binding is completely inhibited by asialo-orosomucoid but not by orosomucoid or asialoagalacto-orosomucoid, suggesting that cell binding is mediated by asialoglycoprotein receptors on the cell surface [12]. Later, Kim et al. [13] observed enhanced hepatocyte adhesion and spheroid formation on galactose-modified polystyrene surfaces. Lu et al. [14] developed a simple method to immobilize galactose ligands onto poly(vinylidene difluoride) membrane. They found that hepatocytes cultured on the membrane self-assembled into hepatocyte spheroids and showed a higher level of functional maintenance. Other

1966

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

2-D galactosylated substrata for hepatocyte culture include poly(ethylene terephthalate) films [15–17] and Si3N4 membrane [18]. 3-D galactosylated scaffolds have also been developed for hepatocyte culture. For example, Yang et al. [9] synthesized galactosylated alginate gels and found that hepatocytes encapsulated show higher viability, more spheroid formation and higher liver functions than those in ordinary alginate gels. Similar results were reported from other galactosylated 3-D scaffolds, including silk fibroin [19], hydroxypropyl cellulose [10], chitosan [20] and alginate/chitosan [21]. Compared with 2-D substrata, 3-D scaffolds provide the cells an environment closer to the normal tissue environment, therefore should be a better choice for spheroid formation. Indeed, it was reported that hepatocytes organized into spheroids when cultured in some hydrogel scaffolds, even in the absence of galactose ligands [22–24]. Although a lot of galactosylated 3-D scaffolds have been reported, it is usually difficult to harvest the hepatocyte spheroids from the 3-D matrix, which is a prerequisite for their biological analysis and also for their further applications, e.g. as building blocks in organ printing [5]. Here a reversible hydrogel was developed as a new galactosylated 3-D scaffold for the generation of hepatocyte spheroids. The scaffold is composed of thermosensitive poly(N-isopropylacrylamide) (PNIPAM) microgels and poly(ethylene glycol) (PEG), which gels in situ when heated to 37 °C but liquefies again when cooled back to room temperature (Scheme 1). We demonstrated that HepG2 cells, a human hepatocarcinoma cell line, exhibit a higher viability, a greater tendency to form spheroids and a higher level of liver-specific functions in the presence of galactose ligands. Thanks to the reversibility of the system, the in situ-formed spheroids can be facilely harvested from the scaffold, simply by cooling back to room temperature. It is noteworthy that PNIPAM-based polymers have long been developed as injectable cell scaffolds, including loosely cross-linked PNIPAM and poly(N-isopropylacrylamide-co-acrylic acid) (P(NIPAM-AA)) [25–27], PNIPAM-grafted gelatin [28,29] and linear copolymer P(NIPAM-AA) [30]. Thanks to their thermosensitivity, these materials gel in situ when heated to 37 °C and therefore can facilely encapsulate the cells inside the in situ-formed scaffolds.

sodium dodecyl sulfate (SDS) and PEG (Mw = 20,000) were purchased from Acros. Lactobionic acid, acrylic acid (AA), potassium persulfate (KPS), N-(3-dimethylaminopropyl)-N-ethylcarbodiimide hydrochloride (EDC) and ethylenediamine were purchased from local providers. AA was distilled under reduced pressure. NIPAM was purified by recrystallization from a hexane/acetone mixture and dried in a vacuum. Other reagents were used as received. 2.2. Microgel synthesis Poly(N-isopropylacrylamide-co-acrylic acid) (P(NIPAM-AA)) microgels with various AA contents were synthesized by free radical precipitation polymerization, using BIS as crosslinker [31]. The total monomer in the feed was 0.014 mol throughout, while the comonomer molar ratio ((98 – x):x:2) (NIPAM:AA:BIS) was varied according to the desired AA content. The monomers and SDS (0.058 g, 2 mmol) were first dissolved in 95 ml of water. The solution was then filtered to remove any possible precipitate and transferred to a three-necked round-bottom flask equipped with a condenser and a nitrogen inlet. After being heated at 70 °C under a gentle stream of nitrogen for 1 h, 0.081 g of KPS (dissolved in 5 ml of water) was added to initiate the reaction. The reaction was allowed to proceed at 70 °C for 5 h. The resultant microgels were purified by dialysis (cutoff 8000–15,000 Da) against water for 2 weeks. The resulting microgels with AA content of 0, 2 and 10% were denoted as PN, PNA2 and PNA10, respectively. Galactosylated microgels were synthesized using Yang et al.’s method with some modification (Scheme 2) [9]. First, lactobionic acid was coupled with ethylenediamine to obtain monoamine terminated lactobionic lactone (L-NH2). Briefly, 1.0 g of lactobionic acid and 3.7 ml of ethylenediamine were dissolved in 20 ml of dimethyl sulfoxide (DMSO), to which 5.368 g of EDC was added. The mixture was stirred at 30 °C for 24 h. The product was precipitated with chloroform. After purification by repeated dissolving in DMSO and precipitating in chloroform, the product was dried in a vacuum. The P(NIPAM-AA) microgels (PNA2 or PNA10) were then coupled with L-NH2, also using EDC as a catalyst. Briefly, 5 ml of P(NIPAM-AA) microgel dispersion was diluted to 50 ml, to which 0.5 g of L-NH2 was added. After the pH of the mixture was adjusted to 5.0, 0.240 g EDC was added. The reaction was allowed to proceed at 4 °C for 1 h and then at room temperature for an additional 3 h. The final product was purified by dialysis. The galactosylated microgels derived from PNA2 and PNA10 were denoted as PNG2 and PNG10, respectively.

2. Experimental section 2.1. Materials N-Isopropylacrylamide (NIPAM) was purchased from Tokyo Chemical Industry Co., Ltd. N,N0 -methylenebisacrylamide (BIS),

** * ** *

(3)

(1)

** * ** *

***

*** ***

***

(2)

***

*** ***

***

***

***

***

*** ***

***

PEG

** * ** *

*** ** *

***

***

***

(4)

** * ** *

** *

PNIPAM microgel ***

galactosylated PNIPAM microgel HepG2 cell

Scheme 1. Generation and harvest of HepG2 spheroids using galactosylated reversible hydrogels composed of PNIPAM microgel, galactosylated PNIPAM microgel and PEG. (1) The cells are mixed with microgel dispersion and heated to 37 °C to solidify the scaffold. (2) In situ generation of multicellular spheroids during culture. (3) The cell/ scaffold is liquefied by cooling to room temperature. (4) Harvest of HepG2 spheroids by separation from microgel particles.

1967

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

CH2OH

CH2OH OH CH2OH O OH O OH

OH COOH NH2CH2CH 2NH 2

CH2OH OH O OH

OH

OH O NH2 N H

O OH

OH EDC OH

OH L-NH 2

Lactobionic acid

CH2OH CH2OH OH O OH

L-NH 2 O C

O C

NH

O OH

EDC OH

OH

O C

OH O

O C HN

N H

HN OH

P(NIPAM-AA) microgel

Galactosylated PNIPAM microgel Scheme 2. Synthesis of galactosylated PNIPAM microgel.

2.3. Preparation of pre-gel mixtures for the reversible hydrogel scaffolds The pre-gel mixtures were prepared by dissolving the lyophilized microgels and PEG in phosphate-buffered saline (PBS; pH 7.4, I = 0.154 M). The name and the final composition of the microgel/PEG blends are shown in Table 1. These mixtures gel in situ when heated to 37 °C and liquefy again when cooled back to room temperature. For samples BG0-BG5, PEG was added to reduce the shrinkage of the in situ-formed hydrogel. 2.4. Cell culture HepG2 cell (Peking Union Medical College, Beijing, China) dispersion was mixed with microgel/PEG mixtures. The final cell concentration was 3  105 cells ml1. To each well of a 48-well culture plate, 0.5 ml of the cell/microgel mixture was added. The mixtures were brought to 37 °C and gelled immediately. The cells were cultured in a media containing 90% Dulbecco’s modified Eagle medium (DMEM, Sigma–Aldrich), 10% heat-inactivated fetal calf serum and 100 units ml1 of penicillin/streptomycin. The cultures were maintained in an incubator at 37 °C with a humidified atmosphere of 5% CO2. The medium was changed every 1–3 days. The cell proliferation in the hydrogel scaffolds was measured using MTT assay (Sigma–Aldrich). After being cultured for a predeterminated period, the cells were incubated with 5 ml of MTT solution (5 mg ml1) for 4 h at 37 °C. After washing with PBS three times to remove the remaining cultivating medium, 10 ml of DMSO was added to dissolve the resultant formazan crystals. The ultraviolet absorbance of the solubilized formazan crystals was measured spectrophotometrically at 490 nm. Table 1 Composition and shrinkage degree (h(1)/h0) of microgel/PEG blends. Sample name

PN gel PNG2 gel BG0 BG1 BG2 BG3 BG5

Concentration (wt.%)

h(1)/h0

PN

PNG2

PEG

3.0 0 3.0 2.7 2.4 2.1 1.5

0 3.0 0 0.3 0.6 0.9 1.5

0 0 0.50 0.50 0.50 0.50 0.50

The appearance of the cells cultured in the scaffold was observed in situ using an Olympus LX70-140 inverted fluorescence microscope. The samples were stained with acridine orange (AO) and ethidium bromide (EB) before imaging. The excitation wavelength used for AO was 450 ± 20 nm. 2.5. F-actin staining To stain F-actin, the spheroids were first fixed in 4% paraformaldehyde for 20 min, permeabilized in 0.1% Triton X-100 for 5 min and incubated in 5 lg ml1 fluorescein isothiocyanate (FITC)-phalloidin (Sigma–Aldrich) for 20 min. They were then imaged with a fluorescence microscope at an excitation wavelength of 450 ± 20 nm. 2.6. Albumin secretion At predetermined intervals the culture medium was replaced with fresh medium without serum. The medium was collected after 24 h culture and stored at 4 °C until analysis. The amounts of albumin in the medium were determined by a sandwich enzyme-linked immunosorbent assay (ELISA) using a human albumin ELISA kit (Aquatic Diagnostics Ltd, UK) according to the manufacturer’s protocol. The optical density (OD) values at 450 nm were measured with a microplate reader (Infinite F50, TECAN). The amount of albumin was read out from a standard curve which was plotted using a reference human albumin. 2.7. Urea synthesis At predetermined intervals the culture medium was replaced with fresh medium. The medium was collected after 24 h culture. The amounts of urea in the medium were determined using a QuantiChrom Urea Assay Kit (Bioassay systems, USA) according to the manufacturer’s protocol. OD values at 405 nm were measured with a microplate reader (Infinite F50, TECAN).

0.40 0.60 0.61 0.62 0.62 0.69

2.8. Dynamic light scattering Dynamic light scattering characterization was performed on a Brookhaven 90Plus laser particle size analyzer at a scattering angle of 90°. The sample temperature was controlled with a built-in

1968

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

Peltier temperature controller. CONTIN algorithm was used for data analysis. 2.9. Turbidity The turbidity of the diluted microgel suspensions, which is represented as the absorbance at 550 nm, was measured on a TU1810PC UV–Vis spectrophotometer (Purkinje General, China) using water as a reference. The sample temperature was controlled with a refrigerated circulator. 2.10. Rheological measurements Dynamic rheological analysis of the concentrated microgel dispersions was performed on an AR2000ex rheometer (TA Instruments). Aluminum parallel plate geometry with a diameter of 40 mm was used. The sample gap was set to be 1.0 mm. The temperature was controlled by a Peltier system in the bottom plate connected with a water bath. Temperature-dependent changes in elastic (storage) modulus, G0 , and viscous (loss) modulus, G00 , were recorded in a dynamic temperature ramp test. All the rheological experiments were performed within the linear viscoelastic region. 2.11. Statistical analysis All experiments were repeated at least three times. Comparison of means was performed by one-way analysis of variance. Differences were considered significant if p < 0.05. 3. Results and discussion 3.1. Synthesis and thermosensitive behaviors of galactosylated PNIPAM microgel The scaffold designed here is based on our previous observation that the dispersions of thermosensitive PNIPAM microgels thermally gel at 37 °C in the presence of added salt [32–34] and liquefy when cooled back to room temperature [35]. To promote the generation of hepatocyte spheroids galactose ligands were introduced. For this purpose, galactosylated PNIPAM microgels were synthesized using Yang et al.’s method with some modification (Scheme 2) [9]. First lactobionic acid was coupled with ethylenediamine using EDC as a catalyst to obtain monoamine terminated lactobionic lactone L-NH2 (Figs. S1 and S2 in Supporting information). It was then coupled to P(NIPAM-AA) microgel, the copolymer microgel of N-isopropylacrylamide and acrylic acid, again using EDC as a catalyst. Using PNA10 microgel with an AA content of 10% as an example, the successful introduction of galactose moiety

A 130

1.2

Absorbance at 550 nm

Rh (nm)

B

PNG2 o VPTT~34 C PNA2 o VPTT~39 C

105

was confirmed by nuclear magnetic resonance and Fourier transform infrared spectroscopy (Figs. S3 and S4 in Supporting information). Using the same method, we previously introduced other functional groups, such as phenylboronic acid [36] and benzo-18crown-6 [37], into PNIPAM microgels. It is well-known that PNIPAM-based microgels are thermosensitive [38–40]. They are hydrophilic and highly swollen at room temperature, but become hydrophobic and shrink sharply when heated above their volume phase transition temperature (VPTT). Fig. 1A shows the size change of the microgels, which are dispersed in 0.01 M pH 7.4 phosphate buffer, with increasing temperature. For the pure PNIPAM microgel PN, its VPTT, defined as the onset of the phase transition, was determined to be 32 °C. PNA2 microgel, a P(NIPAM-AA) microgel containing 2 mol.% AA, shows a higher VPTT of 39 °C, because the carboxylic acid groups in the microgel are deprotonated at pH7.4. PNG2 microgel is the galactosylated microgel derived from PNA2 microgel. With the amidation of the carboxylate groups, the VPTT of the microgel is lowered by 5 °C. However, the VPTT of PNG2 is still slightly higher than that of PN, suggesting that the galactose moieties in the microgel act as hydrophilic groups. The VPTTs of the microgels in media with physiological pH and ionic strength were determined by turbidity (Fig. 1B). Because of the increased ionic strength in the media, the VPTT of the microgels shifts to lower temperatures. However, the VPTT of PNG2 (32 °C) is still slightly higher than that of PN (31 °C). Also because of their thermosensitivity, concentrated PNIPAMbased microgel dispersions thermally gel under appropriate conditions [32–34]. The thermal gelling behavior of concentrated PN and PNG2 microgel dispersions was studied using the dynamic rheological method. To mimic the physiological conditions, the dispersions were prepared with 0.01 M pH 7.4 PBS. Ionic strength was adjusted to 0.154 M by adding NaCl. As shown in Fig. 2, at low temperature, storage modulus G0 is lower than loss modulus G00 , indicating that the dispersions are in a liquid state. Upon heating, both moduli first decrease with increasing temperature. The reduced moduli are attributed to the shrinking of the microgel particles, which results in a decreased effective volume fraction and, in turn, decreased viscosity and elastic properties [41]. As temperature continues increasing, both G0 and G00 increase sharply. As G0 increases faster than G00 , they cross over at certain temperature. Beyond this point, G0 is always larger than G00 , indicating the formation of a physical network. The gelation temperature, Tgel, defined as the temperature where G0 and G00 cross over, was determined to be 31 °C for PN and 31.5 °C for PNG2. The Tgels of both microgel dispersions are close to their VPTT at the same pH and ionic strength (Fig. 1B), indicating that the main driving force for the gelation of the dispersions is hydrophobic interaction among the

PN o VPTT~32 C

80

55

1.0 0.8 0.6 0.4

PN o VPTT~31 C

PNG2 o VPTT~32 C

0.2

PNA2 o VPTT~36 C

0.0

30 20

25

30

35 o

T ( C)

40

45

50

25

30

35

40

45

o

T ( C)

Fig. 1. (A) Hydrodynamic radii (Rh) of the microgels dispersed in 0.01 M pH7.4 phosphate buffer as a function of temperature. (B) Turbidity of dilute microgel dispersions as a function of temperature. The media are 0.01 M pH 7.4 phosphate buffers with an ionic strength of 0.154 M. (M) PN microgel, () PNG2 microgel, () PNA2 microgel.

1969

A 1000

B 100

Dynamic Modulus (Pa)

Dynamic Modulus (Pa)

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

100 10 1

PN microgel

0.1 G' G''

0.01

10

1 PNG2 microgel 0.1

G' G''

0.01

1E-3 15

20

25

30

35

40

45

o

T ( C)

15

20

25

30

35

40

45

o

T ( C)

Fig. 2. Evolution of dynamic modulus of 3.0 wt.% microgel dispersions in PBS (pH 7.4, I = 0.154 M) with temperature increase under 0.1 Pa and 0.1 Hz. (A) PN microgel. (B) PNG2 microgel.

microgel particles [33]. Although both dispersions gel upon heating, the strength of the resulting gel is different. At 37 °C, G0 of the PNG2 gel is only 5.4% of that of the PN gel. The reduced strength of PNG2 gel may be attributed to the large, hydrophilic galactose moieties on the surface of PNG2 microgel, which hinder the hydrophobic interaction among the particles. Another possible reason may be the relatively bigger size of the PNG2 microgel when in a fully collapsed state. As shown in Fig. 1A, when both microgels are fully collapsed, the size of PNG2 is bigger than PN. Therefore the specific surface area of the PN microgel particles is larger, as is the interaction among the particles. As a consequence, PN gel is stronger than PNG2 gel. 3.2. Adjusting shrinkage degree of hydrogel scaffold The performance of a hydrogel scaffold can be influenced by various factors, including mechanical properties, the presentation of adhesive ligand and growth factor, and transport and degradation kinetics [26,42,43]. Previous studies show that the in situ formed hydrogels from PNIPAM microgel dispersions shrink with time, therefore when used as 3-D cell scaffold, their performance is always affected by the degree of shrinkage [24,44]. To better reveal the effect of galactose ligand, we will need scaffolds with the same degree of shrinkage but various amount of galactose ligands. For this purpose, a series of mixed dispersions of PN and PNG2 microgels were prepared (samples BG0-5 in Table 1). The total

concentration of microgel remains constant (3.0 wt.%) for all the samples; however, the amount of PNG2 is different, providing different amounts of galactose ligands. PEG is added (0.50 wt.%) as we previously showed that the shrinkage of the in situ formed gel can be effectively reduced by blending with PEG [44] (Scheme 1). All these mixtures gel when heated to 37 °C and shrink gradually with time. To study the shrinkage of the gels, equal amount of the pregel solutions were added to small glass vials with a diameter of 12 mm and aged at 37 °C for 24 h (Fig. 3). The ratio of the final height of the gel to the original height, h(1)/h0, was used as a measure to reflect the degree of shrinkage. In the absence of PEG, h(1)/ h0 is 0.4 for the hydrogel gelled from 3.0 wt.% PN microgel (i.e. PN gel in Table 1). This ratio increases to 0.60 for BG0 gel, which is the same concentration of PN microgel but blended with 0.5 wt.% PEG. This result is consistent with our previous report [44]. For other gels in which PN microgel is partially replaced by PNG2 microgel, an even smaller shrinkage was observed. However, if less than 30% PN microgel is replaced, i.e., the BG1-3 gels, as shown in Fig. 3, the change in shrinkage degree is negligible. For example, h(1)/h0 was measured to be 0.62 for BG3 gel, in which 30% PN microgel is replaced. For BG5 gel containing 50% PNG2, h(1)/h0 was measured to be 0.69, showing a relatively large difference from that of BG0 (Table 1). As the shrinkage degree of BG0-3 is almost the same, they were chosen for further studies. The thermogelation of the four gels was studied by rheology. As an example, the result of BG3 was shown

Fig. 3. Photographs of the microgel/PEG hydrogels aged at 37 °C for 0 min, 10 min, 30 min, 2 h, 6 h and 24 h, respectively.

1970

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

in Fig. 4. As can be seen, the thermal behavior of the blend systems is similar to the systems containing only microgels, as shown in Fig. 2. At low temperature G00 dominates G0 , but upon heating, G0 overpasses G00 , indicating the formation of a 3-D network. As shown in Table 2, the gelation temperature, Tgel, is almost the same for the all blend gels; however, the gel strength decreases slightly with increasing PNG2 content. 3.2.1. 3-D cell culture The blend gels with almost same shrinkage degree and mechanical strength were then used to study the effect of galactose ligand on the generation of HepG2 multicellular spheroids. For this purpose, the cells were seeded by simply mixing with the microgel/ PEG mixtures at room temperature. Solid cell/matrix construct formed in situ when brought to 37 °C. The viability of the cells cultured in the 3-D scaffolds was assessed using MTT assay. As shown in Fig. 5, in all gels, the number of viable cells first increase with time and then deceases slightly after a long time culture, suggesting that all gels can support the growth and proliferation of the

Dynamic Modulus (Pa)

1000 100 10 1

BG3

0.1

G' G''

0.01 1E-3 15

20

25

30

35

40

45

o

T ( C) Fig. 4. Evolution of dynamic modulus of the microgel/PEG mixture BG3 with temperature increase under 0.1 Pa and 0.1 Hz.

Table 2 Gelation temperature (Tgel) of the microgel/PEG mixtures and G0 of the resultant hydrogel at 37 °C. Sample

PN

BG0

BG1

BG2

BG3

PNG2

G0 (Pa,37 °C) Tgel (°C)

120.7 31

56.79 31

50.24 31.5

38.68 31.5

30.02 31.5

6.517 31.5

Absorbance at 490 nm

1.2 ab

1.0

ab

ab ab

ab

0.8

ab

ab ab

0.6 ab

ab

ab ab

0.4

ab

ab

ab ab

ab

ab

ab

a

a

a

ab

ab

a

a

a

a

BG0 BG1 BG2 BG3

0.2 0

5

10

15

20

25

Time(Day) Fig. 5. Proliferation of HepG2 cells cultured in the in situ-formed blend hydrogels as assessed by MTT assays. The data shown are the mean of three independent experiments. Error bars indicate the standard deviations. The superscript letters represent significant difference between groups (p < 0.05): acompared to the beginning value; bcompared to BG0 value.

cells encapsulated inside. The result is reasonable because previous research has shown that the PNIPAM polymers are not cytotoxic and have potential to be used as cell culture substrata and cell delivery vehicles [30]. Possible reasons for the slightly decreased viability after a long-term culture include the depletion of oxygen and/or nutrient because of the high cell density at this stage, and the accumulation of metabolic wastes. It is expected that cell viability will continue to decrease if the cell are cultured longer. Fig. 5 also shows that the cell viability increases with increasing PNG2 content, or in other words, galactose content, from BG0 to BG3. For the blend gels studied here, their chemical composition is similar. In addition they all have a similar shrinkage degree and mechanical strength. Therefore, the different cell viability should be mainly attributed to the different amount of galactose ligands presented in the gels. As mentioned above, cell–matrix interactions play an important role in the regulation of cell behaviors [45–47]. For anchorage-dependent cells, such as hMSCs, they require a support matrix in order to survive. In the absence of cell–matrix interactions, these cells will undergo apoptosis [47]. For example, it was observed that mouse embryonic liver cells encapsulated in bio-inert PEG hydrogel lost their viability rapidly [48]. It was thought that binding of the cells with the extracellular matrix through integrin receptors provide signals for suppression of apoptosis [49]. Therefore enhanced cell–matrix interactions may result in improved cell viability. This effect has been widely observed for many cells, such as hMSCs [50], osteoblasts [26], etc. HepG2 cells are also anchorage-dependent. As galactose ligands can be recognized by ASGPR on the cell surfaces [12,51], introduction of the ligands can promote the adhesion of hepatocytes, and so improve their viability, as evidenced by the previously developed 2-D and 3-D galactosylated cell scaffolds [9–11,14,15,17–21,52,53]. Here the enhanced cell viability in galactose-containing gels may also be attributed to the promoted anchorage of HepG2 cells to the hydrogel wall in the presence of galactose ligands. Chemical cues provided by galactose ligand may also play an important role, as the galactose ligand only interacts weakly with ASGPR receptors in the hepatocyte cell membrane [10]. The morphology of cells cultured in the 3-D scaffolds was studied using inverted fluorescence microscopy after being stained with AO/EB. Formation of multicellular spheroids was observed in all the gels. As an example, Fig. 6 shows the evolution of multicellular spheroids in BG3 gel. At the beginning, all cells exist as single cells and dispersed evenly throughout the gel matrix. After being cultured for 7 days, however, they form multicellular spheroids with a size of 100 lm (106 ± 27 lm). The spheroids grow to a larger size with time (168 ± 30 lm on day 14 and 200 ± 16 lm on day 21). The formation of HepG2 spheroids can be attributed to the relatively weak interaction between cell and substrate. Therefore the cells tend to aggregate via cell–cell interaction, instead of attaching to the hydrogel surface [6,16,22]. It is noteworthy that during the culture almost all of the cells are alive as they have uniform bright green nuclei with organized structure (Fig. 6). Although multicellular spheroids form in all gels, their morphology is different. Fig. 7 compares the morphology of the spheroids obtained in different gels after 14 days’ culture. As can be seen, the spheroids obtained in BG0 are loosely aggregated. With increasing PNG2 content in the gel, the spheroids become more compact. The result indicates that the amount of galactose ligands influences not only the cell viability, but also the morphology of the spheroids. Enhanced spheroid formation of hepatocytes has previously been observed from both 2-D [9,14] and 3-D galactosylated scaffolds [10,19,20]. It was believed that the galactose ligands provide chemical cues to the hepatocytes to reorganize into 3-D spheroids [10], possibly through the specific interaction with ASGPR on the cell surface [9].

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

1971

Fig. 6. Fluorescence images of HepG2 cells cultured in the in situ-formed BG3 hydrogel for 0 (A), 7 (B), 14 (C) and 21 days (D). The cells were double stained with AO/EB. Scale bar: 100 lm.

Fig. 7. Fluorescence images of HepG2 cells cultured in BG0, (A) BG1, (B) BG2, (C) and BG3 (D) for 14 days. The cells were double stained with AO/EB. Scale bar: 100 lm.

NIH 3T3 cells, which do not have ASGPRs [54], were also seeded and cultured in the gels. Quite different from HepG2 cells, the viability and morphology of NIH 3T3 cells are almost the same in the four gels (Figs. S6 and S7 in Supporting information). The results indicate that the presence of galactose ligands do not influence the growth of cells without ASGPRs, confirming that the enhanced viability of HepG2 cells in the presence of galactose ligands is a result of the specific interaction between the ligands and ASGPR on the cell surface. 3.3. Liver-specific functions To study the effect of galactose ligands on the liver-specific functions of the cells, albumin secretion and urea synthesis by HepG2 cells cultured in the blend gels was quantified. Fig. 8A

shows the amount of albumin secreted into the medium per day in each well on the day of analysis. A similar trend was found for cells cultured in all four gels: the albumin secretion increases first with time, but decreases gradually with longer culture. In BG0 gel without galactose ligand, albumin secretion begins to decrease on day 7. In contrast, in other gels containing galactose ligands, albumin secretion keeps increasing until day 10. In addition, albumin secretion level at the same day increases with increasing PNG2 content in the gel. At day 10, the amount of albumin secreted by HepG2 cells in BG3 is approximately three times that in BG0. The result suggests that albumin secretion of the cells is improved in the presence of galactose ligands. Urea synthesis is another important liver-specific function of HepG2 cells. The urea synthesis profile of the cells is similar to that for albumin secretion. As shown in Fig. 8B, in the first 10 days urea

1972

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

600

BG0 BG1 BG2 BG3

ab ab ab

400

ab

ab

ab

ab

200

ab ab

ab

ab

a

a

a

ab

ab a

B Urea synthesis (mg/dL/day)

Albumin secretion ( g/well/day)

A

ab

7

BG0 BG1 BG2 BG3

ab

6

ab

ab

ab

ab

ab

ab

ab

5 ab

ab

ab

a

a

a

4

ab

a

ab

a

a

a

a

3

0 0

2

4

6

8

10

12

14

16

18

20

Culture time (day)

0

2

4

6

8

10

12

14

16

18

20

Culture time (day)

Fig. 8. (A) Albumin secretion and (B) urea synthesis by HepG2 cells in blend gels as a function of culture time. Data are shown as mean ± standard deviation from three samples. The superscript letters represent significant difference between groups (p < 0.05): acompared to the beginning value; bcompared to BG0 value.

Fig. 9. Fluorescence image of the HepG2 spheroids after F-actin staining. The spheroids were obtained by culture in BG3 for 14 days. Scale bar: 50 lm.

synthesis rate of the cells increases with time and then decreases gradually thereafter. In addition, urea synthesis rate increases with galactose content in the gels, confirming again that the introduction of galactose ligands enhances the liver functions of the encapsulated HepG2 cells. Similar enhanced liver-specific functions were observed from hepatocytes cultured in other galactosylated scaffolds [9,10,19–21]. The enhanced liver-specific functions are directly linked with the enhanced viability of the cells in the presence of galactose ligands, as revealed in Fig. 5. In addition, the more compact structure of the spheroids generated in the galactosylated gels, as revealed in Fig. 7, may also play an important role. The cell–cell interaction in the more compact spheroids should be tighter, which is believed to play a key role for the maintenance of their functions. Actin filament (F-actin) distribution was previously used to characterize the relative strength of the cell–cell and cell– substratum interactions experienced by hepatocytes [10,17]. For

Fig. 10. (A) Morphology change of a 3.0 wt.% BG3 microgel dispersion: (a) the original dispersion at 25 °C, (b, c) the in situ-formed hydrogel by curing the dispersion at 37 °C for 5 min (b) and 1 day (c); (d–f) return to 25 °C for 5 min (d), 10 min (e), and 3 h (f). (B) Evolution of dynamic moduli of BG3 microgel dispersion in a heating and cooling cycle under a stress of 0.1 Pa and a frequency of 0.1 Hz. (C) Optical microscope image of the released cell spheroids. (D) Fluorescence image of the released spheroids stained with AO/EB. Scale bar: 100 lm.

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974

hepatocytes in 2-D monolayers, as the cell–substratum interaction is strong, intense actin stress fibers were usually observed throughout the cells. In contrast, in 3-D spheroids, cortical F-actin distribution was usually observed, indicating strong cell–cell interaction characteristic of hepatocytes in vivo [10,17]. Using this method we studied the cell–cell interaction in HepG2 spheroids generated in BG3. F-actin of the spheroids was fluorescently stained with FITC-phalloidin and observed using a fluorescence microscope. As shown in Fig. 9, the actin cytoskeleton of the spheroids has a predominant cortical localization. The result indicates that the cell–cell interactions in these spheroids are strong, which is in accord with their high level of liver-specific functions as revealed in Fig. 10.

1973

Acknowledgements We acknowledge financial support for this work from the National Natural Science Foundation of China (Grants Nos. 21174070, 21274068, 21228401 and 21374048), the Program for New Century Excellent Talents in University (NCET-11-0264) and the PCSIRT program (IRT1257). Appendix A. Figures with essential colour discrimination Certain figures in this article, particularly Figs. 1–10, are difficult to interpret in black and white. The full colour images can be found in the on-line version, at http://dx.doi.org/10.1016/ j.actbio.2013.12.044.

3.4. Harvest of HepG2 Spheroids Appendix B. Supplementary data The in situ generated cell spheroids can be facilely harvested by cooling the cell/scaffold constructs to room temperature. As shown in Fig. 10A, the gel liquefies when cooled back to room temperature. As mentioned above, the microgel dispersions gel upon heating, mainly because of the hydrophobic interactions among the shrunken microgel particles. When cooled back to a temperature below its VPTT, the microgel particles become hydrophilic again. With the vanishing of hydrophobic interaction among the microgel particles, the 3-D network collapses. The liquefaction of the BG3 gel upon cooling was also studied by rheology. As shown in Fig. 10B, when cooled below 30 °C, both G0 and G00 decrease sharply. As G0 decreases faster than G00 , G0 and G00 cross over again. Beyond this point, G0 is always smaller than G00 , indicating that the hydrogel returns to a liquid state. Because of the reversibility of the hydrogel scaffold, heptocyte spheroids generated inside the scaffold can be facilely harvested. For this purpose, HepG2 cells were seeded and cultured in BG3 for a predetermined time, usually 14 days, to generate spheroids. The scaffolds were then cooled to room temperature. With the liquefaction of the scaffold, spatial constraint on the spheroids is removed. The spheroids can be easily harvested by centrifugation (Scheme 1). Fig. 10C shows a picture of the harvested spheroids. This process is mild and will not result in any cell injury or DNA damage. As shown in Fig. 10D, live/dead assay of the harvested spheroids indicates that almost all cells are alive. It is noteworthy that it is very difficult to harvest the spheroids from the previously reported galactosylated scaffolds [9,10,19–21]. The facile harvest of the spheroids from the reversible hydrogel scaffold will pave the way for their further analysis and applications.

4. Conclusions In conclusion, we developed a new galactosylated hydrogel scaffold for hepatocyte culture. The hydrogel is reversible, i.e., it solidifies in situ upon heating to physiological temperature, and liquefies again upon cooling back to room temperature. The introduction of galactose ligands increases the viability of HepG2 cells cultured in the in situ-formed gels, enhances the formation of multicellular spheroids and improves their liver-specific functions in terms of both albumin secretion and urea synthesis. More importantly, the in situ-generated spheroids can be facilely harvested by cooling back to room temperature, thus paving the way for further biological analysis and application of these spheroids.

Disclosures No competing financial interests exist.

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.actbio.2013. 12.044. References [1] Xia L, Sakban RB, Qu Y, Hong X, Zhang W, Nugraha B, et al. Tethered spheroids as an in vitro hepatocyte model for drug safety screening. Biomaterials 2012;33:2165. [2] Ho VHB, Slater NKH, Chen R. PH-responsive endosomolytic pseudo-peptides for drug delivery to multicellular spheroids tumour models. Biomaterials 2011;32:2953. [3] Friedrich J, Seidel C, Ebner R, Kunz-Schughart LA. Spheroid-based drug screen: considerations and practical approach. Nat Protoc 2009;4:309. [4] Kunz-Schughart LA, Kreutz M, Knuechel R. Multicellular spheroids: a threedimensional in vitro culture system to study tumour biology. Int J Exp Pathol 1998;79:1. [5] Mironov V, Visconti RP, Kasyanov V, Forgacs G, Drake CJ, Markwald RR. Organ printing: tissue spheroids as building blocks. Biomaterials 2009;30:2164. [6] Lin RZ, Chang HY. Recent advances in three-dimensional multicellular spheroid culture for biomedical research. Biotechnol J 2008;3:1172. [7] Glicklis R, Shapiro L, Agbaria R, Merchuk JC, Cohen S. Hepatocyte behavior within three-dimensional porous alginate scaffolds. Biotechnol Bioeng 2000;67:344. [8] Kataoka K, Nagao Y, Nukui T, Akiyama I, Tsuru K, Hayakawa S, et al. An organic–inorganic hybrid scaffold for the culture of HepG2 cells in a bioreactor. Biomaterials 2005;26:2509. [9] Yang J, Goto M, Ise H, Cho CS, Akaike T. Galactosylated alginate as a scaffold for hepatocytes entrapment. Biomaterials 2002;23:471. [10] Nugraha B, Hong X, Mo X, Tan L, Zhang W, Chan P, et al. Galactosylated cellulosic sponge for multi-well drug safety testing. Biomaterials 2011;32:6982. [11] Cho CS, Seo SJ, Park IK, Kim SH, Kim TH, Hoshiba T, et al. Galactose-carrying polymers as extracellular matrices for liver tissue engineering. Biomaterials 2006;27:576. [12] Weigel PH. Rat hepatocytes bind to synthetic galactoside surfaces via a patch of asialoglycoprotein receptors. J Cell Biol 1980;87:855. [13] Kim S, Hoshiba T, Akaike T. Effect of carbohydrates attached to polystyrene on hepatocyte morphology on sugar-derivatized polystyrene matrices. J Biomed Mater Res Part A 2003;67A:1351. [14] Lu HF, Lim WS, Wang J, Tang ZQ, Zhang PC, Leong KW, et al. Galactosylated PVDF membrane promotes hepatocyte attachment and functional maintenance. Biomaterials 2003;24:4893. [15] Ying L, Yin C, Zhuo RX, Leong KW, Mao HQ, Kang ET, et al. Immobilization of galactose ligands on acrylic acid graft-copolymerized poly(ethylene terephthalate) film and its application to hepatocyte culture. Biomacromolecules 2003;4:157. [16] Du YN, Han RB, Ng S, Ni J, Sun WX, Wohland T, et al. Identification and characterization of a novel prespheroid 3-dimensional hepatocyte monolayer on galactosylated substratum. Tissue Eng 2007;13:1455. [17] Du Y, Chia S, Han R, Chang S, Tang H, Yu H. 3D hepatocyte monolayer on hybrid RGD/galactose substratum. Biomaterials 2006;27:5669. [18] Zhang S, Xia L, Kang CH, Xiao G, Ong SM, Toh YC, et al. Microfabricated silicon nitride membranes for hepatocyte sandwich culture. Biomaterials 2008;29:3993. [19] Gotoh Y, Ishizuka Y, Matsuura T, Niimi S. Spheroid formation and expression of liver-specific functions of human hepatocellular carcinoma-derived FLC-4 cells cultured in lactosesilk fibroin conjugate sponges. Biomacromolecules 2011;12:1532. [20] Feng ZQ, Chu XH, Huang NP, Wang T, Wang YC, Shi XL, et al. The effect of nanofibrous galactosylated chitosan scaffolds on the formation of rat primary

1974

[21]

[22] [23]

[24] [25]

[26] [27]

[28]

[29]

[30]

[31] [32]

[33] [34]

[35]

Y. Wu et al. / Acta Biomaterialia 10 (2014) 1965–1974 hepatocyte aggregates and the maintenance of liver function. Biomaterials 2009;30:2753. Seo SJ, Kim IY, Choi YJ, Akaike T, Cho CS. Enhanced liver functions of hepatocytes cocultured with NIH 3T3 in the alginate/galactosylated chitosan scaffold. Biomaterials 2006;27:1487. Lee J, Cuddihy MJ, Cater GM, Kotov NA. Engineering liver tissue spheroids with inverted colloidal crystal scaffolds. Biomaterials 2009;30:4687. Sakai S, Ito S, Kawakami K. Calcium alginate microcapsules with spherical liquid cores templated by gelatin microparticles for mass production of multicellular spheroids. Acta Biomater 2010;6:3132. Gan TT, Guan Y, Zhang YJ. Thermogelable PNIPAM microgel dispersion as 3D cell scaffold: effect of syneresis. J Mater Chem 2010;20:5937. Stile RA, Burghardt WR, Healy KE. Synthesis and characterization of injectable poly(N-isopropylacrylamide)-based hydrogels that support tissue formation in vitro. Macromolecules 1999;32:7370. Stile RA, Healy KE. Thermo-responsive peptide-modified hydrogels for tissue regeneration. Biomacromolecules 2001;2:185. Kim S, Chung EH, Gilbert M, Healy KE. Synthetic MMP-13 degradable ECMs based on poly(N-isopropylacrylamide-co-acrylic acid) semi-interpenetrating polymer networks, Degradation and cell migration. I. Degradation and cell migration. J Biomed Mater Res Part A 2005;75:73. Ohya S, Nakayama Y, Matsuda T. In vivo evaluation of poly(Nisopropylacrylamide) (PNIPAM)-grafted gelatin as an in situ-formable scaffold. J Artif Organs 2004;7:181. Ibusuki S, Fujii Y, Iwamoto Y, Matsuda T. Tissue-engineered cartilage using an injectable and in situ gelable thermoresponsive gelatin: Fabrication and in vitro performance. Tissue Eng 2003;9:371. Au A, Ha J, Polotsky A, Krzyminski K, Gutowska A, Hungerford DS, et al. Thermally reversible polymer gel for chondrocyte culture. J Biomed Mater Res Part A 2003;67A:1310. Jones CD, Lyon LA. Synthesis and characterization of multiresponsive coreshell microgels. Macromolecules 2000;33:8301. Gan T, Zhang Y, Guan Y. In situ gelation of P(NIPAM-HEMA) microgel dispersion and its applications as injectable 3D cell scaffold. Biomacromolecules 2009;10:1410. Liao W, Zhang Y, Guan Y, Zhu XX. Gelation kinetics of thermosensitive PNIPAM microgel dispersions. Macromol Chem Phys 2011;212:2052. Liao W, Zhang Y, Guan Y, Zhu XX. Fractal structures of the hydrogels formed in situ from poly(N-isopropylacrylamide) microgel dispersions. Langmuir 2012;28:10873. Wang D, Cheng D, Guan Y, Zhang Y. Thermoreversible hydrogel for in situ generation and release of HepG2 spheroids. Biomacromolecules 2011;12:578.

[36] Zhang YJ, Guan Y, Zhou SQ. Synthesis and volume phase transitions of glucosesensitive microgels. Biomacromolecules 2006;7:3196. [37] Luo Q, Guan Y, Zhang Y, Siddiq M. Lead-sensitive PNIPAM microgels modified with crown ether groups. J Polym Sci Part A: Polym Chem 2010;48:4120. [38] Saunders BR, Vincent B. Microgel particles as model colloids: theory, properties and applications. Adv Colloid Interface 1999;80:1. [39] Pelton R. Temperature-sensitive aqueous microgels. Adv Colloid Interface 2000;85:1. [40] Guan Y, Zhang Y. PNIPAM microgels for biomedical applications: from dispersed particles to 3D assemblies. Soft Matter 2011;7:6375. [41] Senff H, Richtering W. Temperature sensitive microgel suspensions: colloidal phase behavior and rheology of soft spheres. J Chem Phys 1999;111:1705. [42] Pek YS, Wan ACA, Ying JY. The effect of matrix stiffness on mesenchymal stem cell differentiation in a 3D thixotropic gel. Biomaterials 2010;31:385. [43] Tibbitt MW, Anseth KS. Hydrogels as extracellular matrix mimics for 3D cell culture. Biotechnol Bioeng 2009;103:655. [44] Cheng D, Wu Y, Guan Y, Zhang Y. Tuning properties of injectable hydrogel scaffold by PEG blending. Polymer 2012;53:5124. [45] Jabbari E. Bioconjugation of hydrogels for tissue engineering. Curr Opin Biotechnol 2011;22:655. [46] Streuli CH, Gilmore AP. Adhesion-mediated signaling in the regulation of mammary epithelial cell survival. J Mammary Gland Biol 1999;4:183. [47] Meredith Jr JE, Fazeli B, Schwartz mA. The extracellular matrix as a cell survival factor. Mol Biol Cell 1993;4:953. [48] Underhill GH, Chen AA, Albrecht DR, Bhatia SN. Assessment of hepatocellular function within PEG hydrogels. Biomaterials 2007;28:256. [49] Ruoslahti E, Reed J. Anchorage dependence, integrins, and apoptosis. Cell 1994;77:477. [50] Nuttelman CR, Tripodi MC, Anseth KS. Synthetic hydrogel niches that promote hMSC viability. Matrix Biol 2005;24:208. [51] Oka JA, Weigel PH. Binding and spreading of hepatocytes on synthetic galatose culture surfaces occur as distinct and separable threshold responses. J Cell Biol 1986;103:1055. [52] Lu HF, Lim WS, Zhang PC, Chia SM, Yu H, Mao HQ, et al. Galactosylated poly(vinylidene difluoride) hollow fiber bioreactor for hepatocyte culture. Tissue Eng 2005;11:1667. [53] Chua KN, Lim WS, Zhang PC, Lu HF, Wen J, Ramakrishna S, et al. Stable immobilization of rat hepatocyte spheroids on galactosylated nanofiber scaffold. Biomaterials 2005;26:2537. [54] Iwasaki Y, Takami U, Shinohara Y, Kurita K, Akiyoshi K. Selective biorecognition and preservation of cell function on carbohydrateimmobilized phosphorylcholine polymers. Biomacromolecules 2007;8:2788.