Biocompatible Silica Nanoparticles-Insulin Conjugates for MesenchymalStem Cell Adipogenic Differentiation

Dan Liu,†,§ Xiaoxiao He,*,†,§ Kemin Wang,*,‡,§ Chunmei He,†,§ Hui Shi,‡,§ and Lixin Jian‡,§

Institute of Biology, State Key Laboratory of Chemo/Biosensing and Chemometrics, College of Chemistry and ChemicalEngineering, Hunan University, Changsha 410082, China, and Key Laboratory for Bio-Nanotechnology and MoleculeEngineering, Hunan Province, Changsha 410082, China. Received April 24, 2010; Revised Manuscript Received July 5, 2010

There is increasing interest in developing bioconjugated carriers for the cellular delivery of bioactive moleculesto stem cells, since they can allow modulation of stem cell differentiation. The present study reported biocompatiblesilica nanoparticle-insulin conjugates for rat mesenchymal stem cell (RMSC) adipogenic differentiation in vitro.A systematic study was first carried out on the biocompatibility of the SiNPs with RMSCs. The cell viabilityassay was performed to screen the SiNP concentration for creating little cytotoxicity on RMSCs. Furthermore,transmission electron microscopy (TEM) and adipogenesis and osteogenesis assays revealed that the pure SiNPshad no effect on cellular ultrastructures, adipogenic differentiation, and osteogenic differentiation. Under theoptimized SiNP concentration with little cytotoxicity on RMSC and no effects on the RMSC phenotype,SiNP-insulin conjugates were prepared and used for RMSC adipogenic differentiation. Results showed that RMSCshad the ability to differentiate into adipocytes when cultured in the presence of insulin-conjugated SiNPs. Thiswork demonstrated that the biological activity of insulin conjugated to the SiNPs was not affected and the SiNPscould be used as biocompatibile carriers of insulin for RMSC adipogenic differentiation, which would help toexpand the new potential application of SiNPs in stem cell research.

INTRODUCTION

Ongoing advances in the directed differentiation in vitro ofstem cells have offered great prospects for medical research andopened new avenues for therapy in many human disorders (1-3).Until now, various strategies have been developed and employedto direct and control stem cell differentiation in vitro, includingdevelopment of defined culture milieu (4), use of extracellularmatrix (5, 6), co-culturing with fully differentiated somatic celltypes (7, 8), up-regulating the expression of connexin proteinsto enhance the gap junction formation and intercellular coupling(9), and directing stem cell differentiation through geneticmodulation (10). Among the strategies to achieve controlleddifferentiation of stem cells into well-defined lineages in vitro,bioactive molecules such as cytokines, growth factors, and smallchemicals as defined culture conditions have been widely usedto promote stem cell differentiation in vitro (11). Especially forthe adult stem cells, differentiation cannot appear spontaneouslyin the absence of cytokines and growth factors.

For stem cell differentiation in vitro using bioactive moleculesas defined culture conditions, the differentiation efficiency wouldusually be affected not only by the internal factors such asgenetic, developmental potential of stem cells, but also by thecharacteristics of exogenous bioactive molecules, such as theirbiological activity (12), solubility (13), and cell toxicity (14).

Although endogenous bioactive molecules could be producedwith recombinant DNA constructs through genetic manipulationand have successfully induced the transfected stem cell dif-ferentiation in vitro, the risk of mutagenesis and/or oncogenesisof recombinant DNA and the viral-based vectors is stillunpredictable (15). Therefore, the development of novel strate-gies to deliver exogenous bioactive molecules to the stem cellsis still regarded as a hot issue for stem cell differentiation invitro.

Recently, the wide spectrum of nanotechnologies holds greatpromise for the study of stem cell biology and the developmentof new approaches to stem cell research (16). The potentialapplications of nanotechnologies in stem cell research includemagnetic nanoparticles for stem cell isolation and sorting (17),the fluorescent nanoparticle labeling method for stem celllabeling and stem cell in vivo tracking (18), engineeringnanometer-scale materials for proliferation and differentiationof stem cells (19, 20), and nanomaterial-based carriers for theintra/extracellular delivery of genes, drugs, and bioactivemolecules for stem cell differentiation (21-23). For intra/extracellular delivery of bioactive molecules to stem cells, theuse of nanomaterial-based carriers is expected to maintain thebiological activity and stability of bioactive molecules and likelyextend the duration of exposure of cells to the bioactive mol-ecules (24). In the case of nanomaterial-based carriers forbioactive molecules to stem cells, most efforts are focused onthe development of polymeric nanoparticles. For example, anovel dexamethasone-loaded carboxymethylchitosan/poly(ami-doamine) dendrimer (Dex-loaded CMCht/PAMAM) nanopar-ticle has been designed and combined with either ceramic orpolymeric scaffolds to induce the differentiation of rat bonemarrow stromal cells in vitro, and it has demonstrated that Dex-loaded CMCht/PAMAM dendrimer nanoparticles could be usedas intracellular nanocarriers of biological agents to modulatethe early osteogenic differentiation of rat bone marrow stromalcells in vitro (22). The biocompatible and biodegradable

* Corresponding authors. Kemin Wang, State Key Laboratory ofChemo/Biosensing and Chemometrics, College of Chemistry & Chemi-cal Engineering, Hunan University, No. 1 Lushan South Road,Changsha, Hunan, 410082, PR China. Tel.: +86-731-88821566; Fax:+86-731-88821566, E-mail: kmwang@hnu.cn. Xiaoxiao He, Instituteof Biology, Hunan University, No. 1 Lushan South Road, Changsha,Hunan, 410082, PR China. Tel.: +86-731-88823073; Fax: +86-731-88823780; E-mail: xiaoxiaohe@hnu.cn.

† Institute of Biology, Hunan University.‡ State Key Laboratory of Chemo/Biosensing and Chemometrics,

College of Chemistry and Chemical Engineering, Hunan University.§ Key Laboratory for Bio-Nanotechnology and Molecule Engineering.

Bioconjugate Chem. 2010, 21, 1673–1684 1673

10.1021/bc100177v 2010 American Chemical SocietyPublished on Web 08/24/2010

poly(lactide-co-glycolide) (PLGA) particles with different sizeshave also been used as a model system to design growth-factor-releasing systems for controlling the differentiation of humanembryonic stem cells (hESCs) into the vascular lineage (23).

By comparing with polymer nanoparticles, silica nanoparticles(SiNPs) are of interest for their biocompatibility and theirmechanical properties. Because of their unique characteristics,SiNPs have been widely studied in a range of areas includingchemistry, engineering, and biomedicine, especially the ap-plication of functionalized SiNPs for cancer diagnostics andtherapy (25, 26). Recently, they are also emerging as an ideaagent for efficient stem cell labeling and tracking (27, 28). Toour knowledge, SiNPs have not previously been used as carriersof bioactive molecules for stem cell differentiation in vitro. Inthe present study, the SiNP-based carriers of insulin forSprague-Dawley (SD) rat mesenchymal stem cell (RMSC)adipogenic differentiation have been reported. The cell viabilityassay of RMSCs in the presence of pure SiNPs was firstperformed using the MTT assay and morphology test to screenthe SiNP concentration needed to create little cytotoxicity inthe RMSCs. After further confirmation that the pure SiNPsin the concentration needed to create little cytotoxicity had noeffect on cellular ultrastructures, adipogenic differentiation, andosteogenic differentiation, the SiNPs were then selected ascarriers to be conjugated with insulin, one of the importantgrowth factor supplements for RMSC adipogenic differentiation.In vitro cell studies were carried out in order to evaluate theadipogenic efficacy of adipogenic induction media containingSiNP-insulin conjugates in the control experiments. The resultsshowed that the biological activity of insulin conjugated to theSiNPs was not affected and the SiNP-insulin conjugates couldbe used for RMSC adipogenic differentiation, which would helpto expand the new potential application of SiNPs in stem cellresearch.

EXPERIMENTAL PROCEDURES

Materials. Tris(2,2-bipyridyl)dichlororuthenium(II) hexahy-drate (Rubpy), Triton X-100, and aspartic acid were purchasedfrom Sigma-Aldrich. Insulin was supplied by Invitrogen. N-(p-Maleimidophenyl)isocyanate (PMPI) was obtained from Ther-mo. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bro-mide (MTT) was supplied by Amersco. Oil red-O and Alizarinred were purchased from Dingguo Biotech. Sprague-Dawleyrat mesenchymal stem cells (SD RMSCs), trypsin-EDTA, andphosphate-buffered saline (PBS, pH 7.2-7.4) were purchasedfrom Cyagen biosciences. LG-DMEM growth medium consist-ing of low-glucose Dulbecco’s modified Eagle’s medium (LG-DMEM), 10% fetal bovine serum (FBS), 100 mg/mL strepto-mycin, 100 U/mL penicillin, and 2 mM glutamine; adipogenicinduction media containing LG-DMEM growth medium andother supplements (5 × 10-4 M 3-isobutyl-1-methyl-xanthine,1 × 10-4 M indomethacin, 1 × 10-6 M dexamethasone, 0.01mg/mL insulin); adipogenic maintenance media consisting ofLG-DMEM growth medium and 0.01 mg/mL insulin; osteogenicmedia containing LG-DMEM growth medium and other supple-ments (0.8 mM ascorbate, 0.01 M -glycerophosphate, 1 × 10-7

M dexamethasone) were also supplied by Cyagen Biosciences.Other chemicals if not specified were all commercially availableand used as received. All aqueous solutions were preparedexclusively with 18 MΩ deionized water (Barnstead ThermolyneNanopure, Garner, NC).

Preparation of SiNPs and FSiNPs. Silica nanoparticles(SiNPs) and Rubpy dye-doped silica nanoparticles (FSiNPs)were synthesized using the water-in-oil (W/O) microemulsionmethod as reported before (29). Briefly, a solution containingcyclohexane, Triton X-100, and n-hexanol (with volume ratio4:1:1) was mixed with adequate water and stirred for 5 min

(for FSiNPs, followed by the addition of Rubpy dye solutionsas the core material). In the presence of tetraethylorthosilicate(TEOS), polymerization reaction was initiated by addingconcentrated NH4OH. The reaction was allowed to continue for24 h. Then, the nanoparticles were isolated by acetone andwashed with ethanol and water several times to remove sur-factant molecules.

Culture of RMSCs in Vitro. RMSCs were seeded at adensity of 4 × 103 cells/cm2 in the T25 or T75 flasks andcultured in the LG-DMEM growth medium, containing 10%FBS, 100 mg/mL streptomycin, 100 U/mL penicillin, and 2 mMglutamine. Cultures were maintained at 37 °C in a humidifiedincubator with 5% CO2. The cells were then dissociated withtrypsin-EDTA solution and subcultured at a 1:3 ratio whenthey were grown to 80-90% confluence.

Cell Viability and Proliferation Assay. The effect of pureSiNP concentration on RMSC viability and proliferation wasfirst examined using MTT reduction and morphology testing.For cell viability and proliferation experiment using MTT assay,the RMSCs were grown in 96-well plates at an initial densityof 4 × 103 cells/well and cultured in 200 µL LG-DMEM growthmedium for 24 h. Then, the cells were exposed to the pure SiNPswith final concentrations of 0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 1, and2 mg/mL, respectively. After the cells were cultured in the mediacontaining pure SiNPs with different concentrations for a further24, 48, and 72 h, 20 µL of MTT (5 mg/mL in PBS) was addedto each well and incubated with the cells at 37 °C for 4 h.Thereafter, the medium was removed and 150 µL of dimethylsulfoxide (DMSO) was added and incubated for 10 min. Afterall crystals were dissolved, the optical density of each well wasmeasured at 570 nm using a microplate reader (Bio-Rad ELISAreader, Hercules, CA). The RMSCs untreated with SiNPs wereprovided as the normal control group, and the LG-DMEMgrowth media in wells were prepared as the blank control group.The cell viability was calculated according to the followingFormula 1.

In addition, the effects of pure SiNP concentration on themorphology of RMSCs were also investigated. Briefly, the cellswere cultured in a monolayer for 24 h and fed with LG-DMEMgrowth medium containing pure SiNPs with different finalconcentrations as mentioned above for a further 72 h. Then,the media was removed, and the cells were washed twice withPBS and observed under the microscope.

Determination of Cellular Uptake of Pure SiNPs by RM-SCs and the Effect of SiNPs on Cellular Ultrastructure. Todetermine the uptake of SiNPs by RMSCs, the well-grown cellswere incubated with 0.1 mg/mL of Rubpy-doped SiNPs(FSiNPs) suspension in LG-DMEM growth medium at 37 °Cfor 24, 48, and 72 h. Treated cells were then washed three timeswith cold PBS to remove excess FSiNPs. Then, the cells wereresuspended with PBS and further stained using DAPI as anuclear counter-stain. The fluorescent dye incorporated inFSiNPs served as a marker to determine their cellular uptakeand analyzed by laser scanning confocal microscopy.

Transmission electron microscopy (TEM) was subsequentlyperformed to investigate if the uptaken SiNPs had an effect oncellular ultrastructures. The well-grown cells were incubatedwith 0.1 mg/mL SiNPs suspension in LG-DMEM growthmedium at 37 °C for 72 h. Treated cells were then washed withPBS to get rid of excess unbound nanoparticles. Afterward, thetrypsinized cells were fixed in 2.5% glutaraldehyde for 2 h,washed with PBS, and postfixed with 2% osmium tetroxide for1 h. Then, the cells were dehydrated through a graded series ofacetone, embedded in Araldite resin, and sectioned with an

Formula 1: Cell viability % )(ASiNPs - Ablank control)/(Anormal control - Ablank control) × 100%

1674 Bioconjugate Chem., Vol. 21, No. 9, 2010 Liu et al.

ultramicrotome. The prepared ultrathin sections (70 nm) wereimaged with transmission electron microscopy (JEM-1230) atan accelerating voltage of 80 kV. The RMSCs untreated withSiNPs were also prepared for the control experiments.

Determination of the Effect of Pure SiNPs on RMSCPhenotype. Mesenchymal stem cells are multipotent stem cellswhich can differentiate into a variety of cell types in vitro or invivo, including osteoblasts, adipocytes, chondrocytes, myocytes,and endotheliums (30). In order to demonstrate the effect ofpure SiNPs on RMSC phenotype, adipogenesis and osteogenesisassays were conducted in the presence of pure SiNPs. The effectsof pure SiNPs on adipogenic differentiation were performed asfollows. RMSCs were seeded onto six-well tissue culture platesin LG-DMEM growth medium at 2 × 104 cells/cm2. The cellsgrown in culture plates were then incubated with 0.1 mg/mL pureSiNPs for 24 h and maintained in LG-DMEM growth mediumuntil post-confluence. Then, the LG-DMEM growth medium withSiNPs was replaced with LG-DMEM growth medium. The pureSiNP-incubated RMSCs were subjected to four cycles ofadipogenic induction/maintenance media. Each cycle consistedof incubation with the induction media for 3 days, followed by1 day incubation with maintenance media. After induction, oilred-O staining was performed to investigate the presence oflipid-containing vacuoles. Briefly, the induced RMSCs werewashed with PBS and fixed in a PBS solution containing 4%paraformaldehyde for 30 min, then washed with PBS twice andstained with oil red-O for 20 min. To further quantify the lipidvacuole content, the oil red O was extracted using 100%isopropanol, and absorbance was measured at a wavelength of490 nm.

To evaluate the effect of pure SiNPs on osteogenesisdifferentiation, RMSCs were plated in six-well tissue cultureplates at 3 × 103 cells/cm2 and exposed to 0.1 mg/mL pureSiNPs for 24 h. After the treatment, the pure SiNP incubatedRMSCs were subjected to osteogenic differentiation. Theosteogenic differentiation medium was replaced every three

days, and induction was continued for a four-week period. Afterfour weeks of osteogenic differentiation, RMSCs were rinsedwith PBS, fixed for 30 min in 4% paraformaldehyde, and stainedin Alizarin red solution for 7 min. For quantization by calciumassay, cells stained with Alizarin red were destained with 10%cetylpyridinium chloride (CPC10%) and the absorbance ofdestaining solution was measured at a wavelength of 562 nm.The amount of calcium was calculated according to a standardcurve of Alizarin red.

Scheme 1. Schematic Drawings of Insulin Conjugation onto theSiNPs

Table 1. Adipogenic Differentiation Treatment of RMSCs withDifferent Induction Media and Different Maintenance Media

no. 1 2 3 4

Step 1 Induction Media

complete adipogenic differentiation medium + - - -adipogenic differentiation medium without of insulin - + + +pure SiNPs - - - +SiNP-insulin conjugates - + - -

Step 2 Maintenance Media

insulin/LG-DMEM + - - -LG-DMEM - + + +Pure SiNPs - - - +SiNP-insulin conjugates - + - -

Figure 1. Transmission electron microscopy images of SiNPs (a),FSiNPs (b), and SiNP-insulin conjugates (c).

SiNP-Insulin Conjugates, Stem Cells Differentiation Bioconjugate Chem., Vol. 21, No. 9, 2010 1675

Preparation and Characterization of SiNP-Insulin Con-jugates. Under the optimized SiNP concentration with littlecytotoxicity on the RMSC and no effect on the RMSCphenotype, SiNPs as biocompatible carriers of insulin for RMSCadipogenic differentiation were investigated. Insulin was con-jugated to the surface of SiNPs via hydrogen bonds with asparticacid-modified SiNPs, as shown in Scheme 1. First, the SiNPswere modified with aspartic acid through conjugation of thehydroxyl group of SiNPs to the amino group of aspartic acidby using PMPI heterobifunctional cross-linking. The SiNPs (20mg) washed with ethanol were suspended in 2 mL of dimethylsulfoxide (DMSO). A solution containing 2.1846 mg of PMPIin 150 µL of DMSO was then added to the SiNP suspensionand incubated with gentle shaking for 1 h at 37 °C. The PMPIactivated SiNPs were washed and resuspended in 2 mL NaHCO3

solution (pH 8.3). 2 mL of 5 mg/mL aspartic acid was added

to 1 mL of 10 mg/mL PMPI activated SiNPs and reacted for12 h at 4 °C. Then, the mixture of aspartic acid modified SiNPsand aspartic acid was then thoroughly washed to removeuncoordinated aspartic acid. After the SiNPs were modified withaspartic acid, 875 µL of 4 mg/mL insulin was added to thesolutions of aspartic acid-capped SiNPs and incubated for 16 hat 4 °C. Following repetitive washing of SiNP-insulin conju-gates and removal of unconjugated insulin by centrifugation,the SiNP-insulin conjugates were redispersed in PBS for futureusage. The unconjugated insulin presented in the supernatantwas determined using the Bradford method. The amount ofconjugated insulin was calculated as Formula 2 (31), where qwas the amount of insulin conjugated onto a unit mass of SiNPs(mg/mg). Wti, Wsi, and WSiNPs represented the total amount ofinsulin added, the amount of insulin in supernatant afterconjugation, and the mass of SiNPs, respectively. To determine

Figure 2. (A) RMSC viability in the presence of pure SiNPs of different concentrations from 0 to 2 mg/mL after the treatment at 24, 48, and 72 h,respectively. (B) Cell morphology of RMSCs treated with pure SiNPs at last concentration of (a) 0, (b) 0.05, (c) 0.1, (d) 0.2, (e) 0.3, (f) 0.4, (g)0.5, (h) 1, and (i) 2 mg/mL, respectively, for 72 h.

1676 Bioconjugate Chem., Vol. 21, No. 9, 2010 Liu et al.

if the insulin was conjugated to the SiNPs, the insulin reactedSiNPs were characterized by Fourier transform infrared (FTIR)spectroscopy analysis.

Assessment of RMSC Adipogenic Differentiation withSiNP-Insulin Conjugates. RMSCs were seeded onto six-welltissue culture plates in LG-DMEM growth medium at thedensity of 2 × 104 cells/cm2 and cultured to 100% confluence.To assess the adipogenic differentiation of RMSCs treated withSiNP-insulin conjugates, the cells were exposed to four roundsof adipogenic induction with different induction media for threedays, with a resting period of one day in different maintenancemedia, respectively, as shown in Table 1. After treatment byadipogenic induction, the qualitative and quantitative analysisfor RMSC adipogenesis was studied as mentioned above. Allexperiments were carried out more than three times.

RESULTS AND DISCUSSION

Characterization of SiNPs. The SiNPs, FSiNPs and insulin-conjugated SiNPs were first characterized by transmissionelectron microscopy (JEM-1230). The morphological analysisshowed that all of them were well-dispersed spherical nano-particles and uniform in size (Figure 1a,b,c). The diameters ofSiNPs, FSiNPs, and insulin-conjugated SiNPs were 58.7 ( 1.1nm, 52.6 ( 1.2 nm, and 60.4 ( 1.4 nm, respectively. Theinsulin-conjugated SiNPs were obtained through the covalentmodification of insulin on the surface of SiNPs. The TEMimaging results indicated that the sizes of insulin-conjugatedSiNPs were a little larger than that before modification, possiblydue to the conjugation of insulin. Additionally, the sizedistribution of SiNPs before and after bioconjugation has alsobeen determined by dynamic light scattering (DLS) assay using

a Malvern Zetasizer (Zetasizer NanoZS). The results showedthat the hydrodynamic diameter of SiNPs before bioconjugationwas about 141 nm. When the SiNPs were conjugated withinsulin, the size of the insulin-conjugated SiNPs was littlechanged, with diameter of about 153 nm (Supporting Informa-tion Figure S1), which was consistent with the size change trendscharacterized by TEM imaging. While the mean diameters ofSiNPs before and after bioconjugation determined by DLS wereboth larger than those determined by TEM, the reason for thisis that the DLS values of nanoparticles in aqueous solution arealways larger than solid-state diameters in a monolayer by TEM,as reported in the literature (32).

Effect of Pure SiNPs with Different Concentration onCell Viability and Proliferation. Our previous biocompatibilitystudies of SiNPs with cancer cells and normal cells have shownthat the bioeffects of SiNPs on cancer cells and normal cellswere both concentration-dependent. The cytotoxicity increasedwith increasing SiNP concentration (33). Therefore, the effectof SiNP concentration on the metabolic activity of RMSCs, aparameter commonly used to reflect cytotoxicity, was firstassessed before using SiNPs as carriers of insulin for RMSCadipogenic differentiation. As shown in Figure 2A, addition ofSiNPs to the cell culture medium with last concentration of 0.05mg/mL did not affect the metabolic activity of RMSC cells afterexposure for 24, 48, or 72 h. Although the cell survival ratedecreased as the concentration of SiNPs increased to 0.1 mg/mL, the survival rate of the cells after exposure for SiNPs of24, 48, and 72 h remained 90 ( 6%, 84 ( 9%, and 82 ( 5%,respectively. Moreover, the difference in effects observed forthe cells treated with either 0.05 mg/mL or 0.1 mg/mL SiNPswas not more than 10% in the three time points. The resultsindicated that addition of SiNPs to the cell culture medium withlast concentrations of 0.05 mg/mL and 0.1 mg/mL did not affectthe metabolic activity and proliferation of RMSCs. As theconcentration of SiNPs increased, the cell survival rate de-

Figure 3. Confocal fluorescence images of RMSCs after treatment with 0.1 mg/mL FSiNPs for (a) 24 h, (b) 48 h, and (c) 72 h, respectively. (a1),(b1), and (c1) represented the blue fluorescence imaging of RMSCs from DAPI stained nuclear imaging; (a2), (b2), and (c2) represented the redfluorescence imaging of RMSCs from FSiNPs; (a3), (b3), and (c3) represented the merger of blue fluorescence imaging and red fluorescence imagingof RMSCs.

Formula 2: q )Wti(mg) - Wsi(mg)

WSiNP(mg)

SiNP-Insulin Conjugates, Stem Cells Differentiation Bioconjugate Chem., Vol. 21, No. 9, 2010 1677

creased. If the last concentration of SiNPs was not higher than0.1 mg/mL, the survival rate of the cells after exposure for 24 hremained around 80%. However, the treatment with SiNPs foreither 48 or 72 h significantly decreased the metabolic activityof RMSCs at concentrations starting from 0.1 mg/mL. WhenSiNP concentration was increased to 2 mg/mL, the treatmentof RMSCs at 24, 48, or 72 h caused a reduction in metabolicactivity to 52 ( 9%, 24 ( 3%, and 23 ( 7%, respectively.These data indicated that the effect of SiNPs on RMSC viabilityand proliferation was concentration-dependent. With the increaseof SiNP concentration, cytotoxicity increased gradually. TheSiNPs were biocompatible with RMSCs if the treatmentconcentration of SiNPs was not more than 0.1 mg/mL.

The morphological changes and cell confluence of treatedRMSCs with different concentrations of SiNPs for 72 h were

also observed under an inverted phase-contrast microscope.After 72 h exposure to SiNPs with last concentration of either0.05 or 0.1 mg/mL, the microscopic evaluation of RMSCsshowed no major changes in cell morphology and cell conflu-ence as compared to unexposed RMSCs (Figure 2B, a-c). Mostcells showed a fibroblast-like morphology, growing in a spindleshape and polygon shape. They displayed eddy current arrayafter 90-100% confluence. However, proliferation of RMSCshas been shown to be inhibited when the RMSCs were treatedwith high concentrations of nanoparticles starting from 0.1 mg/mL. When SiNP concentration was increased to 2 mg/mL, therewas only about 70% confluence (Figure 2B, d). With furtherincrease of SiNPs concentration, the confluence of cells thatwere involved in fusion was less than 50%, and the cellsappeared to retract and showed a greater tendency to die. The

Figure 4. Transmission electron microscopy images of (a) RMSCs without treatment of SiNPs, (b) RMSCs after treatment with 0.1 mg/mL SiNPsfor 72 h, (c) SiNPs, (d) SiNPs and mitochondria cytoplasm, (e) connotations similar to glycogen and cell organs, and (f) cell nucleolus.

1678 Bioconjugate Chem., Vol. 21, No. 9, 2010 Liu et al.

cell toxicity increased obviously. Cell morphology had a clearretraction and the trace of the dead cells separated from thedishes could be easily observed, as shown by the arrow marksin Figure 2B, e-i. The morphological analysis results alsoconfirmed that the effect of SiNPs on RMSC viability andproliferation was concentration-dependent.

After screening the SiNP concentration creating little cyto-toxicity on the RMSCs, further experiments were confirmed toinvestigate if the pure SiNPs with concentration of 0.1 mg/mLhad any effect on the cellular ultrastructure, adipogenic dif-ferentiation, and osteogenic differentiation.

Cellular Uptake of SiNPs and Effect of Uptaken SiNPson Cellular Ultrastructures. Using confocal laser scanningmicroscopy, the cellular uptake of SiNPs by RMSCs wasdetermined by the incorporated Rubpy fluorescence dye. Mergedimages of the red fluorescence from FSiNPs and the bluefluorescence from DAPI stained nuclei indicated that FSiNPscould be effectively taken up by the RMSCs and mainlyoccupied the cytoplasm after the RMSCs were treated with 0.1mg/mL FSiNPs for 24, 48, and 72 h, respectively (Figure 3).The RMSCs still maintained good shape and growth conditionsafter uptake of a large number of particles.

Figure 5. Effect of 0.1 mg/mL SiNPs on adipogenic differentiation of RMSCs. (A) Imaging of oil red-O stained lipid vacuoles of RMSCs. (B)Absorbance intensity of oil red-O stained RMSCs. (a), (b), (c), and (d) represented adipogenic differentiation medium induced SiNP pretreatedRMSCs, adipogenic differentiation medium induced RMSCs, RMSCs without adipogenic differentiation medium, and SiNP pretreated RMSCswithout adipogenic differentiation medium, respectively.

SiNP-Insulin Conjugates, Stem Cells Differentiation Bioconjugate Chem., Vol. 21, No. 9, 2010 1679

In order to further confirm the uptake of SiNPs by RMSCsand investigate the effect of uptaken SiNPs on cellular ultra-structures, TEM analyses of RMSCs treated with 0.1 mg/mLof SiNPs for 72 h were performed (Figure 4a,b). From thewhole-cell images taken, the structures of RMSC cells treatedwith SiNPs were preserved by comparison with the controlgroup without treatment with SiNPs. The uptaken SiNPs wereobviously distributed in the cells and still retained the well-dispersed spherical and uniformity in size after their internaliza-tion into cells (Figure 4c,d). The cristae and matrix of themitochondria were clearly seen (Figure 4d), and the connotationssimilar to glycogen and cell organs were distributed richly inthe SiNP-treated RMSCs (Figure 4e). The nucleolus remainedintact with nuclear bags, excrescences, and nucleolus (Figure

4f). As we know, the RMSCs are multipotent stem cells thatcan differentiate into a variety of cell types in vivo and in vitro.The rich cell organs are the necessary structural and functionalunits of RMSCs, and the connotations similar to glycogen canprovide the energy for differentiation. These cell ultrastructureimage results indicated that the treated RMSCs stayed in arelatively active period with normal cell function.

Effect of Pure SiNPs on the Adipogenesis and Osteogenesisof RMSCs. The effect of SiNPs on the ability of RMSCs todifferentiate into adipocytes and osteocytes was subsequentlyexamined. For differentiation into adipocytes, RMSCs wereseeded onto six-well tissue culture plates in growth medium at2 × 104 cells/cm2 and exposed to 0.1 mg/mL SiNPs for 24 h.Then, RMSCs were subjected to adipogenic differentiation

Figure 6. Effect of 0.1 mg/mL SiNPs on osteogenic differentiation of RMSCs. (A) Imaging of Alizarin red stained RMSCs. (B) Amount of calciummineralization. (a), (b), (c), and (d) represented osteogenic differentiation medium induced SiNP pretreated RMSCs, osteogenic differentiationmedium induced RMSCs, RMSCs without osteogenic differentiation medium, and SiNP pretreated RMSCs without osteogenic differentiation medium,respectively.

1680 Bioconjugate Chem., Vol. 21, No. 9, 2010 Liu et al.

medium for 18 days. Meanwhile, three control groups weredesigned, including the RMSC adipogenic induction withoutpretreatment of SiNPs, the RMSCs cultured in the absence ofadipogenic stimulants, and the SiNP pretreated RMSCs culturedin the absence of adipogenic stimulants. The intracellular lipidaccumulation was determined by examination with light mi-croscopy and absorbance measurement at 490 nm after Oil red-Ostaining. As shown in Figure 5A, the total amount of ac-cumulated lipids in the SiNP pretreated RMSCs was close tothe untreated RMSC group following the 18 days of adipogenicstimulation in the presence of adipogenic stimulants. Theabsorbance values lacked any significant difference betweenSiNP-treated RMSCs and untreated RMSCs during adipogenicdifferentiation (Figure 5B). However, both the RMSCs and SiNPpretreated RMSCs had no lipid vacuoles if they were notinduced by adipogenic stimulants. These results revealed thatthe presence of pure SiNPs did not interfere with adipogenicdifferentiation of RMSCs.

For differentiation into osteocytes, RMSCs were seeded ontosix-well tissue culture plates in growth medium at 3 × 103 cells/cm2 and exposed to 0.1 mg/mL SiNPs for 24 h. Then, RMSCswere subjected to osteogenic differentiation medium for 30 days.Three control groups were also designed, which included theRMSC osteogenic induction without pretreatment with SiNPs,the RMSCs cultured in the absence of osteogenic differentiationmedium, and the SiNP pretreated RMSCs cultured in theabsence of osteogenic differentiation medium. Alizarin redstaining showed that the RMSCs had good biological charac-teristics of osteocytes in the presence of SiNP treated RMSCssimilar to untreated cells following the 30 days of osteogenicstimulation (Figure 6), and both the RMSCs and SiNP pretreatedRMSCs were not stained because they were not subjected toosteogenic differentiation medium. The amount of calciummineralization by the calcium assay results also lacked signifi-cant difference between SiNP treated RMSCs and untreatedRMSCs following 30 days of osteogenic stimulation, suggestingthat the presence of SiNPs had no effect on osteogenic dif-ferentiation of RMSCs.

Characterization of SiNP-Insulin Conjugates UsingFourier Transform Infrared (FTIR) Spectroscopy. FTIRspectroscopy was used to determine if the insulin was conjugatedto the SiNPs. FTIR spectra of pure SiNPs and pure insulin as

control group were also recorded. As shown in Figure 7, acomparison of spectra with that recorded from the pure SiNPsamples (curve a) revealed the presence of prominent resonanceat 1630 cm-1, which was ascribed to the physically adsorbedwater, probably in the micropores (34). The determination ofspectra with that recorded from insulin-conjugated SiNPssamples (curve b) showed the presence of prominent resonanceat 1540 cm-1. This absorption band absent in FTIR spectrarecorded from the SiNPs samples was identified as the amideII vibrational mode, suggesting that insulin was indeed conju-gated to the SiNPs. Insulin is a very important biologically activeprotein composed of 51 amino acids. Each insulin protein hadstrong absorption bands at 1540 cm-1 and 1640 cm-1 (curvec), which were identified as the amide I and II vibrational modes,respectively. The position of these two bands was important tothe biological activity of insulin protein. The absorption bandat 1640 cm-1 (amide I) recorded from the insulin-conjugatedSiNP samples should be overlap the spectra of SiNPs andinsulin. The amount of conjugated insulin on the SiNPs was0.3437 mg/mg calculated from Formula 2.

Effect of Insulin-Conjugated SiNPs on the AdipogenicDifferentiation of RMSCs. To induce the RMSC adipogenicdifferentiation in vitro using defined culture milieu, the adipo-genic differentiation medium should be complete. Insulin is oneof the important supplements of adipogenic induction andadipogenic maintenance media. In the present study, the insulin-conjugated SiNPs for RMSC adipogenic differentiation havebeen investigated. After culturing in a different culture mediumfor a period of 18 days, oil red-O staining of the lipid vacuolesof RMSCs was carried out. As shown in Figure 8A, the RMSCshad plenty of lipid vacuoles after induction with completeadipogenic differentiation medium, which indicated the adipo-genic differentiation of RMSCs in vitro. When the insulin inthe complete adipogenic differentiation medium and mainte-nance media were both replaced with equal-quantity insulinconjugated SiNPs, the RMSCs also had plenty of lipid vacuoles.However, very few lipid vacuoles were observed for the RMSCsexposed to the adipogenic medium absent of insulin or theadipogenic medium without insulin but with 0.1 mg/mL pureSiNPs. Figure 8B showed the quantitative investigation of oilred-O staining for lipid vacuoles of RMSCs cultured in adifferent culture medium for the period of 18 days. There wereno significant differences in the absorbance values for theRMSCs exposed to the complete adipogenic differentiationmedium and the insulin conjugated SiNP adipogenic differentia-tion medium after 18 d of culturing. Low absorbance valueswere observed for the RMSCs exposed to the adipogenicmedium without insulin but with 0.1 mg/mL pure SiNPs. Theseresults obviously demonstrated that insulin is necessary for theadipogenic differentiation of RMSCs in vitro. The adipogenicdifferentiation medium without insulin could not induce theadipogenic differentiation of RMSCs even though pure SiNPshad been added. When the insulin was conjugated to SiNPs,the biological activity of insulin as a supplement of the adip-ogenic differentiation medium was still retained.

As we all know, the successful release of biomolecules fromthe nanoparticles is critical for nanoparticle-based carriers tomeet specific needs. The release of the conjugated biomoleculesfrom the nanoparticles is related to many factors, including thematerials of the nanoparticles, the interaction of nanoparticleswith biomolecules, and so on. In this work, the nature ofinteraction between insulin and SiNPs carriers would beexpected to play a major role in the release of insulin. Asdescribed above, insulin was linked to the aspartic acid-SiNPsurface coating via much weaker hydrogen bonding. In this case,the release of insulin is more facile and faster in the biologicalsystem as reported previously (35). Therefore, the release of

Figure 7. Fourier transform infrared (FTIR) spectroscopy measurementsof (a) pure SiNPs, (b)SiNP-insulin conjugates, and (c) insulin.

SiNP-Insulin Conjugates, Stem Cells Differentiation Bioconjugate Chem., Vol. 21, No. 9, 2010 1681

insulin from SiNP conjugates would be possible after uptakeby the RMSCs. Then, the potential release characteristic madethe SiNPs able to be used successfully as biocompatibile carriersof insulin for RMSC adipogenic differentiation.

The aforementioned cell viability assay of SiNPs with RMSCsand the in vitro cell studies of insulin-conjugated SiNPs on theadipogenic differentiation of RMSCs have clearly demonstratedthat SiNPs could be used as biocompatible carriers of insulinfor RMSCs adipogenic differentiation. The SiNP-based carriersare expected to keep the biomolecules’ activity and stability. Itis conceivable that this novel nanocarrier for the intracellularand controlled delivery of bioactive molecules can likely be anexcellent candidate for modulation of the cellular functions ina more effective manner ex vivo and maintenance of the cellularphenotype in vivo upon reimplantation. If substituting the insulin

to suit other biomolecules, it is believed that the SiNP basedcarrier holds great potential application in other stem celldirected differentiation in vitro and in vivo.

CONCLUSION

In this work, we demonstrated the biocompatible silicananoparticle-insulin conjugates for RMSC adipogenic dif-ferentiation. The biocompatibility of the SiNPs with RMSCswas first investigated, namely, for cell viability, proliferation,cellular uptake, cellular ultrastructures, adipogenic differentia-tion, and osteogenic differentiation. Results showed that theSiNPs with concentration creating little cytotoxicity on theRMSCs could be well uptaken by RMSCs and would not affectthe cellular ultrastructures, RMSC adipogenic differentiation,and osteogenic differentiation. Further experiments demonstrated

Figure 8. Imaging (A) and absorbance intensity (B) of oil red-O stained RMSCs after induction with different culture medium (a) complete adipogenicdifferentiation medium and insulin/LG-DMEM maintenance media; (b) adipogenic differentiation medium and maintenance media containingSiNP-insulin conjugates; (c) adipogenic differentiation medium and maintenance media without insulin; and (d) adipogenic differentiation mediumand maintenance media without insulin but with 0.1 mg/mL SiNPs.

1682 Bioconjugate Chem., Vol. 21, No. 9, 2010 Liu et al.

that insulin could be conjugated to the SiNPs and retain itsbiological activity for induction of the adipogenic differentiationof RMSCs in combination with other supplements. Therefore,SiNPs could be used as biocompatible intracellular nanocarriersof biological agents able to modulate the behavior of stem cells.

ACKNOWLEDGMENT

This work was partially supported by Program for InnovativeResearch Team of Hunan National Science Foundation(10JJ7002), Program for New Century Excellent Talents inUniversity (NCET-06-0697) and National Science Foundationof P.R.China (90606003, 20775021).

Supporting Information Available: Characterization ofSiNPs and SiNP-insulin conjugates by dynamic light scattering(DLS). This material is available free of charge via the Internetat http://pubs.acs.org.

LITERATURE CITED

(1) Gardner, R. L. (2007) Stem cells and regenerative medicine:principles, prospects and problems. C. R. Biol. 330, 465–473.

(2) Zur Nieden, N. I., Kempka, G., and Ahr, H. J. (2007) In vitrodifferentiation of embryonic stem cells into mineralized osteo-blasts. Differentiation 71, 18–27.

(3) Wernig, M., Meissner, A., Foreman, R., Brambrink, T., Ku,M., Hochedlinger, K., Bernstein, B. E., and Jaenisch, R. (2007)In vitro reprogramming of fibroblasts into a pluripotent ES-cell-like state. Nature 448, 318–324.

(4) Chevallier, N., Anagnostou, F., Zilber, S., Bodivit, G., Maurin,S., Barrault, A., Bierling, P., Hernigou, P., Layrolle, P., andRouard, H. (2010) Osteoblastic differentiation of human mes-enchymal stem cells with platelet lysate. Biomaterials 31, 270–278.

(5) Chaubey, A., and Burg, K. J. L. (2008) Extracellular matrixcomponents as modulators of adult stem cell differentiation inan adipose system. J. Bioact. Compatible Polym. 23, 20–37.

(6) Evans, N. D., Gentleman, E., Chen, X., Roberts, C. J., Polak,J. M., and Stevens, M. M. (2010) Extracellular matrix-mediatedosteogenic differentiation of murine embryonic stem cells.Biomaterials 31, 3244–3252.

(7) Qiu, C., Hanson, E., Olivier, E., Inada, M., Kaufman, D. S.,Gupta, S., and Bouhassira, E. E. (2005) Differentiation of humanembryonic stem cells into hematopoietic cells by coculture withhuman fetal liver cells recapitulates the globin switch that occursearly in development. Exp. Hematol. 33, 1450–1458.

(8) Gai, H., Nguyen, D. M., Moon, Y. J., Aguila, J. R., Fink, L. M.,Ward, D. C., and Ma, Y. (2010) Generation of murine hepaticlineage cells from induced pluripotent stem cells. Differentiation79, 171–181.

(9) Chang, J.-C., Hsu, S.-H., and Su, H.-L. (2009) The regulationof the gap junction of human mesenchymal stem cells throughthe internalization of quantum dots. Biomaterials 30, 1937–1946.

(10) Klopper, J., Lindenmaier, W., Fiedler, U., Mehlhorn, A., Stark,G. B., and Finkenzeller, G. (2008) High efficient adenoviral-mediated VEGF and Ang-1 gene delivery into osteogenicallydifferentiated human mesenchymal stem cells. MicroVasc. Res.75, 83–90.

(11) Heng, B. C., Haider, H. K., Sim, E. K.-W., Cao, T., and Ng,S. C. (2004) Strategies for directing the differentiation of stemcells into the cardiomyogenic lineage in vitro. CardioVasc. Res.62, 34–42.

(12) Ashman, L. K. (1999) The biology of stem cell factor and itsreceptor C-kit. Int. J. Biochem. Cell Biol. 31, 1037–1051.

(13) Kawakami, K., Oda, N., Miyoshi, K., Funaki, T., and Ida, Y.(2006) Solubilization behavior of a poorly soluble drug undercombined use of surfactants and cosolvents. Eur. J. Pharm. Sci.28, 7–14.

(14) Stahn, C., Lowenberg, M., Hommes, D. W., and Buttgereit,F. (2007) Molecular mechanisms of glucocorticoid action and

selective glucocorticoid receptor agonists. Mol. Cell. Endocrinol.275, 71–78.

(15) Rovira, M., Jane-valbuena, J., Marchand, M., Savatier, P.,Real, F. X., and Skoudy, A. (2007) Viral-mediated coexpressionof Pdx1 and p48 regulates exocrine pancreatic differentiation inmouse ES cells. Cloning Stem Cells 9, 327–338.

(16) Ferreira, L., Karp, J. M., Nobre, L., and Langer, R. (2008)New opportunities: The use of nanotechnologies to manipulateand track stem cells. Cell Stem Cell 3, 136–146.

(17) Jing, Y., Moore, L. R., Williams, P. S., Chalmers, J. J., Farag,S. S., Bolwell, B., and Zborowski, M. (2007) Blood progenitorcell separation from clinical leukapheresis product by magneticnanoparticle binding and magnetophoresis. Biotechnol. Bioeng.96, 1139–54.

(18) Lei, Y., Tang, H. Y., Yao, L. D., Yu, R. C., Feng, M. F., andZou, B. S. (2008) Applications of mesenchymal stem cellslabeled with Tat peptide conjugated quantum dots to cell trackingin mouse body. Bioconjugate Chem. 19, 421–427.

(19) Lee, J. B., Jeong, S. M., Kim, K. J., Cho, D. H., Kwon, I. K.,Yoon, I. C., Choi, K., Suh, J.-K. F., Park, J. H., Park, Y. D.,Chung, J. H., Choi, K. K., Kim, G. T., Choi, G. W., and Choi,Y. S. (2009) Osteogenic differentiation of human adipose-derivedstem cells (hADSCs) on a dexamethasone eluting nanofiberscaffolds. J. Tissue Eng. Regen. Med. 6, 371–379.

(20) Curran, J. M., Chen, R., Stokes, R., Irvine, E., Graham, D.,Gubbins, E., Delaney, D., Amro, N., Sanedrin, R., Jamil, H.,and Hunt, J. A. (2010) Nanoscale definition of substrate materialsto direct human adult stem cells towards tissue specific popula-tions. J. Mater. Sci. Mater. Med. 21, 1021–1029.

(21) Santos, J. L., Oramas, E., Pego, A. P., Granja, P. L., andTomas, H. (2009) Osteogenic differentiation of mesenchymalstem cells using PAMAM dendrimers as gene delivery vectors.J. Controlled Release 134, 141–148.

(22) Oliveira, J. M., Sousa, R. A., Kotobuki, N., Tadokoro, M.,Hirose, M., Mano, J. F., Reis, R. L., and Ohgushi, H. (2009)The osteogenic differentiation of rat bone marrow stromal cellscultured with dexamethasone-loaded carboxymethylchitosan/poly(amidoamine) dendrimer nanoparticles. Biomaterials 30,804–813.

(23) Ferreira, L., Squier, T., Park, H., Choe, H., Kohane, D. S.,and Langer, R. (2008) Human embryoid bodies containing nano-and microparticulate delivery vehicles. AdV. Mater. 20, 2285–2291.

(24) Edelman, E. R., Nugent, M. A., and Karnovsky, M. J. (1993)Perivascular and intravenous administration of basic fibroblastgrowth factor: Vascular and solid organ deposition. Proc. Natl.Acad. Sci. U.S.A. 90, 1513–1517.

(25) Choi, J., Burns, A. A., Williams, R. M., Zhou, Z., Flesken-Nikitin, A., Zipfel, W. R., Wiesner, U., and Nikitin, A. Y. (2007)Core-shell silica nanoparticles as fluorescent labels for nano-medicine. J. Biomed. Opt. 12, 064007.

(26) Lee, J. E., Lee, N., Kim, H., Kim, J., Choi, S. H., Kim, J. H.,Kim, T., Song, I. C., Park, S. P., Moon, W. K., and Hyeon, T.(2010) Uniform mesoporous dye-doped Silica nanoparticlesdecorated with multiple magnetite nanocrystals for simultaneousenhanced magnetic resonance imaging, fluorescence imaging, anddrug delivery. J. Am. Chem. Soc. 132, 552–557.

(27) Huang, D.-M., Hung, Y., Ko, B.-S., Hsu, S.-C., Chen, W.-H., Chien, C.-L., Tsai, C.-P., Kuo, C.-T., Kang, J.-C., Yang, C.-S., Mou, C.-Y., and Chen, Y.-C. (2005) Highly efficient cellularlabeling of mesoporous nanoparticles in human mesenchymalstem cells: implication for stem cell tracking. FASEB J. 19, 2014–2016.

(28) Lu, C.-W., Hung, Y., Hsiao, J.-K., Yao, M., Chung, T.-H.,Lin, Y.-S., Wu, S.-H., Hsu, S.-C., Liu, H.-M., Mou, C.-Y., Yang,C.-S., Huang, D.-M., and Chen, Y.-C. (2007) Bifunctionalmagnetic silica nanoparticles for highly efficient human stemcell labeling. Nano Lett. 7, 149–154.

(29) He, X. X., Wang, K. M., Tan, W. H., Li, J., Yang, X. H.,Huang, S. S., Li, D., and Xiao, D. (2002) Photostable luminescentnanoparticles as biological label for cell recognition of system

SiNP-Insulin Conjugates, Stem Cells Differentiation Bioconjugate Chem., Vol. 21, No. 9, 2010 1683

lupus erythematosus patients. J. Nanosci. Nanotechnol. 2, 317–320.

(30) Barry, F. P., and Murphy, J. M. (2004) Mesenchymal stemcells: clinical applications and biological characterization. Int.J. Biochem. Cell Biol. 36, 568–584.

(31) Yavuz, H., and Denizli, A. (2005) Immunoadsorption ofcholesterol on Protein A oriented beads. Macromol. Biosci. 5,39–48.

(32) Buscher, K., Helm, C. A., Gross, C., Glockl, G., Romanus,E., and Weitschies, W. (2004) Nanoparticle composition of aferrofluid and its effects on the magnetic properties. Langmuir20, 2435–2444.

(33) He, X. X., Wu, X., Wang, K. M., Shi, B. H., and Hai, L. (2009)Methylene blue-encapsulated phosphonate-terminated silica nano-particles: A bifunctional carrier for simultaneous in vivo imagingand photodynamic therapy. Biomaterials 30, 5601–5609.

(34) Li, L. Y., Wang, J. G., Sun, P. C., Liu, X. H., Ding, D. T., andChen, T. H. (2008) Microporous silica hollow nanospheres tem-plated by anionic polypeptide. Acta Phys. Chim. Sin. 24, 359–63.

(35) Joshi, H. M., Bhumkar, D. R., Joshi, K., Pokharkar, V., andSastry, M. (2006) Gold nanoparticles as carriers for efficienttransmucosal insulin delivery. Langmuir 22, 300–305.

BC100177V

1684 Bioconjugate Chem., Vol. 21, No. 9, 2010 Liu et al.