Journal of Controlled Release 163 (2012) 46–54

Contents lists available at SciVerse ScienceDirect

Journal of Controlled Release

j ourna l homepage: www.e lsev ie r .com/ locate / jconre l

In vivo biodistribution and pharmacokinetics of silica nanoparticles as a function ofgeometry, porosity and surface characteristics

Tian Yu a,c, Dallin Hubbard b,c, Abhijit Ray a,c, Hamidreza Ghandehari a,b,c,⁎a Department of Pharmaceutics and Pharmaceutical Chemistry, University of Utah, Salt Lake City, UT 84112, USAb Department of Biomedical Engineering, University of Utah, Salt Lake City, UT 84112, USAc Utah Center for Nanomedicine, Nano Institute of Utah, University of Utah, Salt Lake City, UT 84108, USA

⁎ Corresponding author at: Utah Center for NanomeUniversity of Utah, 5215 SMBB, 36 S. Wasatch Dr. SaUSA. Tel.: +1 801 587 1566; fax: +1 801 581 6321.

E-mail address: hamid.ghandehari@pharm.utah.edu

0168-3659/$ – see front matter © 2012 Elsevier B.V. Alldoi:10.1016/j.jconrel.2012.05.046

a b s t r a c t

a r t i c l e i n f o

Article history:Received 29 March 2012Accepted 28 May 2012Available online 6 June 2012

Keywords:Silica nanoparticlesNanotoxicityBiodistributionGeometryPorosity

The in vivo biodistribution and pharmacokinetics of silica nanoparticles (SiO2) with systematically varied ge-ometries, porosities, and surface characteristics were investigated in immune-competent CD-1 mice via theintravenous injection. The nanoparticles were taken up extensively by the liver and spleen. MesoporousSiO2 exhibited higher accumulation in the lung than nonporous SiO2 of similar size. This accumulation wasreduced by primary amine modification of the nanoparticles. High aspect ratio, amine-modified mesoporousnanorods showed enhanced lung accumulation compared to amine-modified mesoporous nanospheres. Ac-cumulation of the nanoparticles was mainly caused by passive entrapment in the discontinuous openings inthe endothelium of the liver and spleen or in the pulmonary capillaries, and was highly dependent on nano-particle hydrodynamic size in circulation. The SiO2 were likely internalized by the reticulo-endothelial sys-tem (RES) following physical sequestration in the liver and spleen. The nanoparticles that were transientlyassociated with the lung were re-distributed out of this organ without significant internalization. Pharmaco-kinetic analysis showed that all SiO2 were rapidly cleared from systemic circulation. Amine-modified ornonporous nanoparticles possessed a higher volume of distribution at steady state than their pristine coun-terparts or mesoporous SiO2. In all, surface characteristics and porosity played important roles in influencingSiO2 biodistribution and pharmacokinetics. Increasing the aspect ratio of amine-modified mesoporous SiO2

from 1 to 8 resulted in increased accumulation in the lung.© 2012 Elsevier B.V. All rights reserved.

1. Introduction

Silica nanoparticles (SiO2) have utility in a wide range of applica-tions such as biologic delivery platforms [1–3], imaging and diagnos-tic agents [4–7], and targeted therapeutic carriers [8–10]. Recentimprovements in regulating the geometry, porosity, and surface char-acteristics of SiO2 have further facilitated their biomedical applica-tions [9,11–15]. Previous studies have been conducted to show theeffect of size, pegylation, and surface charge on biodistribution andin vivo toxicity of SiO2 [16–18]. It has been reported that mesoporousSiO2 of smaller size with surface pegylation had lower capture by theRES and were more slowly degraded [16]. Organically modified silicananoparticles (ORMOSIL) with diameters of 20–25 nm exhibitedeffective clearance via the hepatobiliary route without any sign oforgan toxicity [17]. Cationic mesoporous SiO2 were excreted rapidlyby the hepatobiliary route probably due to charge-dependent serum

dicine, Nano Institute of Utah,lt Lake City, UT 84112‐5001,

(H. Ghandehari).

rights reserved.

protein adsorption [18]. Limited information however is availableabout the impact of geometry of SiO2 on biodistribution and toxicity.

Recent studies have demonstrated that geometry of nanocarriers caninfluence their circulation half-life and other pharmacokinetic parame-ters [19–22]. For example, pegylated polymericmicelles of long,filamen-tous shape persisted in the circulation up to one week after intravenousinjection, approximately 10 times longer than their spherical counter-parts [19]. It was suggested that the spherical micelles were taken upby the cells more readily than the long filaments under fluid flow condi-tions since the cellular entry of the latter was opposed by flow [19]. Cy-clic polymers composed of α-cholo-ε-caprolactone and ε-caprolactone,which had molecular weights greater than the renal threshold, showedlonger blood circulation time in mice than linear polymers of similarcomposition and comparable molecular weight [20]. This effect was at-tributed to the fact that linear polymers traverse nanopores in glomeruliby end-on motion of one chain end, while cyclic polymers transitthrough by entering the pores with two chain segments since theylack chain ends [20]. While the studies above relate to more flexiblepolymeric systems, the influence of geometry on biological system hasalso been studied for more rigid nanoparticles. For example, it wasshown that the anti-intercellular adhesionmolecule 1 elliptical polysty-rene disks (0.1×1×3 μm) had higher endothelial targeting specificity

47T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

in the lung than spheres of different sizes (0.1, 1, 5 μm) [21]. We pre-viously demonstrated that pegylated gold nanorods (10×45 nm,1.13 mV) exhibited longer blood circulation half-life and highertumor accumulation than pegylated gold nanospheres (50 nm,−27.1 mV) in ovarian tumor bearing mice [22]. These studies sug-gest that geometry and carrier architecture can influence the invivo behavior of nanoscale platforms. However much needs to be ex-amined in this area since factors such as porosity and surface charac-teristics can further influence biodistribution and pharmacokinetics.

Our previous studies on lung cancerous epithelial cells and macro-phages showed that in vitro toxicity of spherical or rod-shaped SiO2 wasmainly determined by porosity and surface characteristics irrespectiveof geometric features [23]. Further in vivo studies demonstrated that thesystemic toxicity of these nanoparticles was also mainly influencedby nanoparticle porosity and surface characteristics. MesoporousSiO2 tended to be less tolerated than nonporous SiO2 while aminemodification on mesoporous nanoparticles improved the tolerateddose threshold [24]. Geometry did not make a significant differencein the mechanism or extent of the systemic toxicity. The next logicalstep is to investigate the distribution of these SiO2 in animals to shedlight on the causes for in vivo toxicity observed in mice beyond max-imum tolerated doses (MTDs) and to relate the in vitro toxicity andcellular uptake with in vivo toxicity and biodistribution. Herein, wereport the biodistribution and pharmacokinetics of SiO2 in mice asa function of geometry, porosity, and surface characteristics.

2. Materials and methods

2.1. Materials

Nonporous silica nanospheres (Stöber) or mesoporous SiO2 with dis-tinct geometrical features (nanospheres,Meso S; aspect ratio 8 nanorods,AR8), and their amine-modified counterparts (SA, MA, 8A) were pre-pared as reported previously [23]. Briefly, nonporous SiO2 were synthe-sized by Stöber method and mesoporous SiO2 were synthesizedthrough a one-step condensation under dilute silica source and low sur-factant concentration conditions with ammonium hydroxide as the basecatalyst [23]. Monoiodinated Bolton–Hunter Reagent, 1 mCi/37 MBq inbenzene, was purchased from American Radiolabeled Chemicals (St.Louis, MO). CD-1 mouse serum was a customized bio-specimen orderfromCharles River Laboratories, fromwhich the CD-1micewere orderedfor this study. All other chemicals were of reagent grade from Sigma-Aldrich.

2.2. Pre-modification of SiO2 for radiolabeling experiments

Cationic, amine-modified SiO2 were produced by reacting thenanoparticles with 3-(aminopropyl)triethyloxysilane (APTES) at aweight ratio of 1:1 in anhydrous ethanol for 20 h at room tempera-ture as described previously [23]. To obtain anionic, slightly amine-modified SiO2, the same procedure was used except that APTESreacted with SiO2 at the weight ratio of 1:50 to make sure therewere available primary amine groups on the surface to conjugatewith monoiodinated Bolton–Hunter Reagent while the surface chargeof SiO2 remained negative. The nanoparticles were stored in ethanoland thoroughly washed in water and borate buffer immediately be-fore radiolabeling experiments.

2.3. SiO2 radiolabeling experiments

The radiolabeling protocol was adapted from an established Bolton–Hunter method, whereby the primary amine groups available on nano-particle surface formed an amide bond with N-hydroxysuccinimidegroup from monoiodinated Bolton–Hunter Reagent [25]. To react withnanoparticles, 20 μL monoiodinated Bolton–Hunter Reagent was trans-ferred to a glass vial and the solvent was allowed to dry in the air for

extended time (1 h). 10 mg of SiO2 prepared in the above section(10 mg/mL) in 0.05 M borate buffer (pH 8.5) was quickly added tothe glass vial and stirred on ice for 45 min. Then the mixture wastransferred to a dialysis cellulose ester membrane with a cutoff sizeof 3.5–5 kD (Float-A-Lyzer G2, Spectra/Por, Spectrum Laboratories,Inc.) and dialyzed against 4 L water at room temperature for 20 dayswith water changing on a daily basis. The unreacted Bolton–Hunter Re-agent was readily hydrolyzed in the aqueous medium and was removedby dialysis. The hydrolyzed product is referred to as 125I-BHR in this arti-cle. Thin layer chromatography (TLC) silica gel was used to check thepresence of unbound 125I-BHR in radiolabeled SiO2 (125I-SiO2) usingmethanol water (4:1 volume ratio) solvent as the mobile phase. The ra-dioactivity on silica gel was measured by the Packard Cobra auto-gamma counter (GMI, Ramsey, Minnesota). For SA and slightly amine-modified Stöber, dialysis did not completely remove 125I-BHR fromnanoparticles and thus an alternative centrifugation method describedbelow was used. The mixture from radiolabeling reaction was collectedinto a 2.0 mLmicrotube and spunat 15,000×g for 30 min in anEppendorfcentrifuge 5415D (Eppendorf, Hamburg, Germany). Nanoparticles wereextensively washed in water and methanol and finally in water.Then TLC method was applied to check the presence of 125I-BHR innanoparticles from each washing cycle until unbound 125I-BHR wasconfirmed to be absent in 125I-SiO2.

2.4. Serum stability of 125I-SiO2

The stability of radiolabeling on SiO2 was tested in mouse serumbefore the biodistribution study. 0.5 mg 125I-SiO2 was added to 1 mL50% CD-1 mouse serum in saline and incubated at 37 °C for 72 h.The experiment was done in triplicate with 125I-BHR in 50% mouseserum as the positive control. At the end of 72 h, an aliquot of mixturewas withdrawn by a glass capillary and analyzed by TLC. The stabilityof 125I-SiO2 was expressed as percentage of radioactivity in the origi-nal spotting site out of the total radioactivity on the plate.

2.5. Biodistribution and pharmacokinetic analysis

Animal studies were conducted under an approved protocol of theUniversity of Utah Institutional Animal Care and Use Committee(IACUC). Female CD-1 mice, 6–8 weeks old, were purchased fromCharles River Laboratories and housed in standard cages with five ani-mals per cage. All animals were acclimated to the animal facility for atleast one week prior to experimental procedures. CD-1 mice wereinjected via the lateral tail vein with 20 mg/kg SiO2 suspension in200 μL sterile saline. The SiO2 suspension was a mixture of 125I-SiO2

and SiO2 of the same type to make a radioactivity dose of 60,000 cpmper animal for pristine SiO2 (Stöber, Meso S, AR8) or a radioactivitydose of 120,000 cpm per animal for amine-modified SiO2 (SA, MA,8A). The weight content of 125I-SiO2 which contributed to the dose ofSiO2 in injection formulation was considered to be negligible. Themice were sacrificed at 5 min, 30 min, 2 h, 24 h and 72 h post intrave-nous injection. At each time point, animals were terminated by CO2 as-phyxiation and blood sampleswere collected via inferior vena cava by aheparin coated syringe immediately post euthanasia. The animals wereflushed with 20 mL sterile saline to remove blood that remained in theorgans in order to obtain accurate tissue accumulation counts based onnanoparticle tissue association or uptake rather than blood content.During necropsy, organs of interest (heart, liver, spleen, lung, kidneys,brain, stomach, small intestine, large intestine, tail, thyroid) and therest of carcass (bones, muscle and skin) were dissected and weighedfollowed by tissue radioactivity measurement by a gamma counter. Ra-dioactivity obtained from different organs was calculated as the per-centage of the injected dose per gram tissue. Compartmental analysisof the pharmacokinetic data was performed using WinNonlin Profes-sional, version 5.3 (Pharsight Corporation, CA). A two-compartmentalpharmacokinetic model was utilized with first order elimination.

48 T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

2.6. Urinary and hepatobiliary excretion studies

Tomeasure the excretion of SiO2 into urine and feces,five animals re-ceived the intravenous injection of SiO2 of each type at 20 mg/kg andwere individually housed in special single-mouse metabolic cages.Urine and feces were collected into separate tubes at 2 h, 24 h, 48 hand 72 h. The samples were immediately weighed and their radioactiv-ity was measured by a gamma counter. In order to identify the radioac-tive species in urine, the urine samples, positive controls (950 μL normalurine+50 μL 2.5 mg/mL SiO2+5 μL 15,000 cpm 125I-BHR), andnegativecontrols (950 μL normal urine+50 μL 15,000 cpm2.5 mg/mL SiO2 injec-tion formulation)were centrifuged at 15,000×g for 20 min to obtain thesupernatant. The supernatant was removed and measured by a gammacounter. The percentage of radioactivity in the supernatant out of theoverall urine sample is indicative of percentage of unbound radioactivityor degraded product in the urine.

2.7. Statistical analysis

Experiments were performed in triplicate with results present as av-erage value ormean±standarddeviation. For in vivo studies,five animals

Size by transmission electron microscopy*

(nm)

Aspect ratio*

Pore size*

(nm)

ζ-p(m

Stober 115 ± 13 1.0 N/AMeso S 120 ± 25 1.0 2.7AR8 136 ± 26 × 1028 ± 139 7.6 2.7aAmine-modified counterparts of Sotber, Meso S, and AR8 are namreported in reference [23].

Biodistribution

On ice, 45 minutes

0.05 M borate buffer, pH 8.5

MonoiodiAmine-modified SiO2

Pharmacokinetics

Fig. 1. Schematic illustration of nanoparticle selection, radiolabeling, and animal administroverall physicochemical characterization of SiO2.

were used per group and differences in in vivo data were analyzed usingone-way ANOVA by GraphPad Prism (GraphPad Software, CA). Wheredetected, Tukey's test was used to evaluate pairwise differences betweenthe groups.

3. Results

3.1. 125I-SiO2 characterization and serum stability

Nonporous nanospheres (Stöber), mesoporous nanospheres (MesoS), or mesoporous nanorods (high aspect ratio 8, AR8) were previouslysynthesized and stored in ethanol [23]. The pristine SiO2 were furthermodified with APTES to obtain their highly cationic counterparts (SA,MA, 8A). These amine-modified SiO2were directly used in radiolabelingexperiments for SA, MA, and 8A. To track the distribution of pristineSiO2 in vivo, the SiO2 were slightly modified with APTES to generateavailable primary amine groups for radioisotope conjugation whilethe anionic surface charge was maintained for comparison with highlycationic, amine-modified SiO2 (Fig. 1). The content of unbound 125I-BHR in the 125I-SiO2 product post purification was analyzed by TLC.There was minimum presence of unbound radioisotope molecules

otential*

V)ζ-potential (mV) APTES:silica 1:50

ζ-potential (mV) APTES:silica 1:1a*

-50.4 -41.0 17.0-39.4 -18.0 32.4-36.6 -21.9 36.7

ed as SA, MA, and 8A, respectively. *Values were previously

125I-SiO2

nated Bolton-Hunter Reaget

Primary amine modification

Mesoporous SiO2(Aspect ratio may be 1 or 8)

Nonporous SiO2

ation for biodistribution and pharmacokinetic studies. The insert table [23] shows the

49T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

associated with 125I-SiO2 product (Supplemental Fig. 1, SupplementalTable 1). Serum stability study on a typical 125I-SiO2, 125I-MA, demon-strated that the amide bond formed during radiolabeling reaction wasstable in 50% mouse serum at 37 °C for 72 h (Supplemental Fig. 2).

3.2. In vivo biodistribution of SiO2 in healthy mice

Biodistribution of a series of SiO2 with varied shapes, porosities, andsurface characteristics was evaluated in immune-competent CD-1 micevia bolus tail vein injection by tracing the radioactivity distribution asdepicted in Fig. 1. Results show that SiO2 of various physiochemicalproperties mainly accumulated in the liver and spleen with differentialdistribution into the lung (Fig. 2, Supplemental Table 2). To evaluate theeffect of geometry on biodistribution, spherical mesoporous SiO2 (Meso

Fig. 2. Biodistribution of SiO2 with varied geometry, porosity, and surface characteristics: A)injection at a dose of 20 mg/kg. Organ accumulation is expressed as percent of injected dpresented as mean±standard deviation (n=5).

S, MA) were compared with rod-shaped mesoporous SiO2 with aspectratio of 8 (AR8, 8A). Both Meso S and AR8 exhibited extensive lungaccumulation. This accumulation was almost eliminated with amine-modified nanospheres MA but not with amine-modified nanorods 8A.All the accumulation in the lung showed a rapid elimination from thisorgan within 24 h post injection. To examine the influence of porosityon nanoparticle biodistribution, mesoporous nanospheres (Meso S)were compared with nonporous nanospheres (Stöber). Results showthat Meso S was primarily accumulated in the lung while Stöber hadnegligible accumulation in this organ. Stöber also exhibited increasedpercentage of liver accumulation out of total recovered dose comparedwith Meso S. To analyze the surface modification effect, amine-modifiedSiO2 were compared with their pristine counterparts. It was revealedthat amine modification could efficiently reduce nanoparticle lung accu-mulation (Fig. 2B, D, F).

Meso S, B) MA, C) AR8, D) 8A, E) Stöber, and F) SA in healthy mice post bolus tail veinose per gram of tissue post euthanasia at 5 min, 30 min, 2 h, 24 h, and 72 h. Data are

50 T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

3.3. Pharmacokinetic analysis

The blood profiles of SiO2 were fitted to a two-compartmentalpharmacokinetic model. All nanoparticles studied were rapidlycleared from blood circulation within 2 h of injection followed by aslow elimination phase which indicated the slow re-distribution be-tween the blood and organs/tissues (Fig. 3). There was no significantdifference in the terminal clearance rates of various types of SiO2

(p>0.05) (Fig. 4A). Stöber showed a significantly higher volume of distri-bution at steady state (Vss) thanMeso S (pb0.05), while amine-modified,spherical SiO2 (MA, SA) exhibited a significant increase inVss compared totheir pristine counterparts (Meso S, pb0.01; Stöber, pb0.001) (Fig. 4B).

3.4. Tissue and blood partitioning of SiO2

The tissue affinity indices, calculated as the ratio of area under thecurve from a specific organ over area under the curve of blood, reflectthe affinity and capacity of nanoparticle association with the specificorgan of interest [26]. Various SiO2 showed high affinity for the spleenand low affinity for the kidneys across the board (Table 1). High aspectratio 8A showed an average higher lung affinity thanMA. Stöber and SA

Fig. 3. Two-compartmental pharmacokinetic analysis of SiO2 biodistribution: A) Meso S, B) Mnanoparticle concentration using percent injected dose per gram blood and assumes a bloo

had on average higher liver uptake than mesoporous nanoparticleswith or without amine modification. Mesoporous SiO2 showed anaverage higher affinity to the lung than Stöber. The lung exposure wasdrastically reduced by amine modification as indicated by decreasedtissue affinity indices from amine-modified nanoparticles (SA, MA,8A) compared with their pristine counterparts (Stöber, Meso S, AR8).The same trend was also observed for kidney exposure of various SiO2.

The tissue/blood concentration ratio of various nanoparticles inmajor organs, liver, spleen, lung, and kidneys, was used as an indica-tor of changes in organ penetrability and retention over time [27]. TheSiO2 across the board generally showed an increase in partitioning inthe liver and spleen over time, but the partitioning remained constantin the lung and kidneys over 72 h (Fig. 5). In the lung, mesoporousnanoparticles (Meso S) demonstrated a significantly higher lung/blood concentration ratio than Stöber (pb0.001) or MA (pb0.001)over 3 days.

3.5. Urinary and hepatobiliary excretion

Excretion of the radioactivematerial throughurinary or hepatobiliaryroutes followed a similar pattern where radioactivity was excreted

A, C) AR8, D) 8A, E) Stöber, F) SA in healthy mice. Activity in the blood is converted tod density of 1.0 g/mL.

Fig. 4. Pharmacokinetic parameters (A) clearance and B) Vss based on the two-compartmental analysis for the nanoparticles. There was no significant difference inclearance among all nanoparticles in blood (p>0.05). Amine-modified SiO2, MA orSA, exhibited significantly higher Vss than their pristine counterparts, Meso S(⁎⁎pb0.01) or Stöber (⁎⁎⁎pb0.001). St ber showed a significantly higher Vss thanMeso S (⁎pb0.05).

51T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

through urine more than feces at all time-points (2 h, 24 h, 48 h, and72 h) and the excretion peaked at 24 h post injection for both routes(Supplemental Table 3). The overall excreted radioactivity reached15%–38% of injected dose by end of study. To investigate the radioactivespecies in the excrement,we centrifugeddown the urine samples aswellas various controls mentioned in the Materials and methods section toidentify radioactive material in the supernatant (Supplemental Fig. 3).Results from control groups show that 125I-BHR did not have physicaladsorption with SiO2 and 100% was recovered in the supernatant while4% of 125I-AR8, a typical 125I-SiO2, was recovered in the supernatant.This indicates that the 40% recovered radioactivity in supernatantof urine samples from AR8 treatment in mice was most likely tobe unbound radioisotopes or small 125I-SiO2 degraded product.Similar results were found for urine samples from other nanoparti-cle treatment groups. We also collected the feces from nanoparticletreated animals, suspended in saline followed by centrifugation.Results showed that at least 36% radioactivity in feces was fromunbound radioisotopes or small degraded product. The rest ofradioactivity recovered in the pellet of urine or feces samples post

Table 1Tissue affinity indices of SiO2 of various geometries, porosities, and surface characteris-tics in major organs of CD-1 mice.

Treatment Tissue affinity index

Liver Spleen Lung Kidneys

Meso S 46.9 93.5 138.9 3.2MA 82.9 172.3 1.4 0.8AR8 80.7 193.9 41.8 3.48A 49.4 180.5 17.2 2.0Stöber 186.0 148.1 6.0 3.6SA 249.5 113.9 4.5 1.3

centrifugation could be 125I-SiO2 or their relatively large degradedproduct.

4. Discussion

In this study the effect of geometry, porosity and surface charac-teristics of SiO2 on biodistribution and pharmacokinetics upon intra-venous injection into healthy mice was evaluated. The overall effectof physicochemical parameters of nanoparticles on the studied

Fig. 5. Tissue/blood concentration ratio of various SiO2 in A) liver, B) spleen, C) lung,and D) kidneys. Meso S showed a significantly higher lung/blood concentration ratiothan St ber (⁎⁎⁎pb0.001) or MA (⁎⁎⁎pb0.001) within 3 days. Data are presented asmean±standard deviation (n=5).

Table2

Summaryof

engine

ered

SiO2withva

riou

sph

ysicoc

hemical

prop

erties

andtheirin

vitroan

din

vivo

evalua

tion

resu

lts.

Nan

oparticle

type

Physicoc

hemical

prop

erties

aIn

vitroa

Invivo

Geo

metry

byTE

M(nm)

Porosity

Surfacech

arge

inwater

DLS

size

inserum

(nm)

IC50

(μg/mL)

Cellu

lar

association

(μgSi/100

μgprotein)

Safety

bBiod

istribution

Pharmaco-kine

tics

ME

ME

MTD

Impa

ired

orga

n(s)

abov

eMTD

Geo

metry

Porosity

Surface

characteristics

Vss(L)

Stöb

er11

5Non

porous

——

121.6

73/

21.2

1.5

450

Hea

rt,lun

g,sp

leen

/Live

rLu

ng,k

idne

ys0.20

MesoS

120

Mesop

orou

s—

268.9

89/

0.7

030

Kidne

ys/

Lung

Lung

,kidne

ys0.09

AR8

136×10

28Mesop

orou

s—

N/A

74/

0.4

065

Kidne

ys/

/Lu

ng,k

idne

ys0.07

SA11

5Non

porous

++

N/A

254

/14

.81.0

450

Lung

,kidne

ys/

Live

r/

0.37

MA

120

Mesop

orou

s+++

150.3

182

/3.3

0.4

150

Lung

,kidne

ys/

//

0.23

8A13

6×10

28Mesop

orou

s+++

N/A

225

/4.1

0.8

100

Lung

,kidne

ysLu

ng/

/0.09

aCo

nten

tada

pted

from

referenc

e[23],surface

charge

isrank

edas

——

high

lyne

gative

−60

to−

40mV,—

high

lyne

gative

−40

to−

30mV,+

+mod

eratelypo

sitive

10–20

mV,+

++

high

lypo

sitive

20–40

mV.N

/Amea

nsno

tava

ilable.

Mor

Erefers

tomacroph

ages

RAW

264.7or

canc

erou

sep

ithe

lialc

ells

A54

9,/mea

nsno

tob

served

inthestud

yde

sign

.b

Conten

tada

pted

from

referenc

e[24],for

columnun

derbiod

istribution,

itmea

nsthat

theph

ysicoc

hemical

prop

erty

(ind

icated

intheco

rrespo

ndingco

lumn)

ofna

nopa

rticlesof

asp

ecifictype

(ind

icated

intheco

rrespo

ndingrow)ledto

high

eraffinity

insp

ecificorga

ns(ind

icated

inthecrossedcell)

than

nano

particlesof

correspo

ndinglydifferen

tprop

erty

(non

porous

versus

mesop

orou

s,pristine

versus

amine-mod

ified

,orna

nosp

heresversus

nano

rods

)./mea

nsthat

this

effect

isno

tob

served

inthestud

yde

sign

.

52 T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

biological systems is summarized in Table 2 to enable the relation ofbiodistribution patterns observed here with cellular uptake and toxic-ity profiles of similar nanoparticles observed previously [23,24].

The biodistribution results show that the majority of SiO2 accumu-lated in the liver and spleen post injection (Fig. 2). This is due to thefact that the discontinuous gaps in the endothelium which lines thesinusoidal walls of liver and spleen allow the passive entrapment offoreign particulates [28,29]. The continuously increased organ/bloodconcentration ratios for liver and spleen indicate that nanoparticleswere internalized post physical sequestration due to the prevalentpresence of macrophages in these organs (Fig. 5A–B). The differencein accumulation of SiO2 in various organs corroborates with our pre-vious in vitro studies which showed that the cellular response tonanoparticle exposure was cell type dependent; macrophages hadextensively higher association with the nanoparticles than epithelialcells [23]. Thus, the liver and spleen, where most macrophages reside,showed the most extensive SiO2 accumulation in the biodistributionstudy. Nonporous SiO2 (Stöber) and their amine-modified counterparts(SA) exhibited highest liver affinity among all types of nanoparticles(Table 1). These results agree with previous in vitro studies whichshowed that porosity played a predominant role in determining nano-particle cellular association; nonporous nanoparticles with or withoutamine modification had the highest cellular association among alltypes of nanoparticles [23]. The high affinity of Stöber for liver couldbe responsible for the significant increase in liver enzyme levels inplasma beyond MTDs compared with controls as shown in previous invivo toxicity studies [24]. It suggests that porosity plays an importantrole in influencing nanoparticle biodistribution pattern.

The SiO2 of various types exhibited differential accumulation inthe lung post injection (Fig. 2). The constant lung/blood concentra-tion ratio over time indicated that the accumulation of SiO2 in thelung was because of transient association with capillary rather thaninternalization (Fig. 5C). The association was balanced between SiO2

organ concentration and SiO2 blood concentration by SiO2 translocationand redistribution into other organs. Thus, the accumulation in the lungwas mostly in capillaries rather than in pulmonary cells. MesoporousSiO2 exhibited a higher lung affinity than nonporous SiO2, and aminemodification reduced lung affinity compared with the pristine SiO2

(Fig. 5C). This pattern could be related to the changes in nanoparticlehydrodynamic size in the presence of serum (Table 2). Though pro-duced with similar size as confirmed by transmission electron micros-copy (TEM), mesoporous SiO2 had significantly higher hydrodynamicsize than nonporous SiO2 in serum [24]. This is probably because indi-vidual nonporous nanoparticle was stabilized by protein adsorption asthe hydrodynamic size decreased in serum compared to that in saline,whereas the presence of protein molecules could not dissociate theslightly aggregated mesoporous SiO2 in aqueous suspension possiblydue to the enhanced inter-nanoparticle interaction of these highsurface area mesoporous SiO2. Thus, the addition of protein layer on thesurface may have contributed to the increase in hydrodynamic size ofmesoporous SiO2 in serum compared to that in saline. Mesoporousnanoparticles with relatively large hydrodynamic size in serum aremore likely to cause obstruction in vessels and canpartially explain the in-creased lung accumulation compared with nonporous SiO2 (Fig. 2A, E).The amine-modified SiO2 showed smaller hydrodynamic sizes in serumprobably due to steric stabilization from adsorbed protein moleculesthan their pristine counterparts [24], which causes lower pulmonary ac-cumulation (Fig. 2B, D, F) and decreased tissue affinity indices (Table 1).However, amine-modified mesoporous nanorods (8A) showed higherlung affinity than amine-modified mesoporous nanospheres (MA)(Table 1), demonstrating that geometry of these nanoparticles influencesbiodistribution to a certain extent. In all, lung accumulation ofnanoparticles was mostly influenced by porosity and surface charac-teristics, however elongated geometrical shape (rods versus spheres)increased accumulation in this organ for amine-modified SiO2. De-tailed investigation should be made to confirm the mechanism of

53T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

nanoparticle-induced obstruction in the environment mimicking thein vivo circulation systemwith whole blood as the medium for futurestudies.

Our previous in vivo toxicity studies showed that the onset of ad-verse reactions was mainly due to the mechanical obstruction ofnanoparticles in the vasculature that led to congestion in organsand subsequent functional failure [24]. It appears that it is the “vascu-lature impact” rather than cellular toxicity that limits silica nanopar-ticle safety in vivo. In in vitro studies, nanoparticle toxicity was mainlyinfluenced by surface characteristics; primary amine modificationsignificantly reduced cellular toxicity as shown by the increased 50%cell inhibitory concentration (IC50) values compared with pristinenanoparticles probably due to the differential subcellular localization,whereas porosity and geometry did not seem to affect the IC50 [23]. Inin vivo studies, porosity and surface characteristics influenced hydro-dynamic sizes of SiO2 in circulation, which had an important implica-tion in their vasculature impact and resultant tolerance threshold[24]. Lung and kidneys were shown to be most susceptible to nanoparti-cle obstruction in vasculature aboveMTDsprobably due to their abundantblood supply and special anatomic structures [24]. Mesoporous SiO2,which potentially had the largest hydrodynamic size in circulation asevidenced by hydrodynamic size analysis in serum, were most prone tocause vasculature obstruction and subsequent renal failure, resulting inthe lowest MTDs at 30–45 mg/kg irrespective of geometrical features[24]. Aminemodification onmesoporous SiO2 reduced the hydrodynamicsize in serum and raised the MTDs to 100–150 mg/kg [24]. NonporousSiO2 had the smallest hydrodynamic size in serum and thus reached thehighest MTDs at 450 mg/kg as observed previously [24]. These previousobservations show that porosity and surface characteristics are majorfactors that influence in vitro or in vivo toxicity of SiO2. Our current studiesevaluating the biodistribution of these nanoparticles also ascertain thepredominant effects of porosity and surface characteristics on organaccumulation.

Pharmacokinetic analysis demonstrates that majority of SiO2 of alltypes were rapidly cleared from circulation at a similar rate (Fig. 4A).Pristine SiO2 had lower Vss than the amine-modified counterpartswhile nonporous SiO2 showed higher Vss than mesoporous SiO2

(Fig. 4B). This agrees with previous in vitro studies on both macro-phages and epithelial cells that mesoporous SiO2 had lower cellularassociation than their amine-modified counterparts while nonporousSiO2 showed higher association than mesoporous SiO2 [23]. These invivo observations suggest that amine-modified SiO2 or nonporousSiO2 tended to associate and be taken up by RES in liver or spleen,leading into increased Vss.

Further excretion experiments showed that radioactivity original-ly from 125I-SiO2 dosed intravenously was found in urine and feces(Supplemental Table 3), indicating possible excretion of SiO2 ortheir degraded product. Nanoparticle accumulation in the kidneyswas low (Fig. 2) and there was limited affinity of SiO2 to this organ(Fig. 5D). Due to the very dilute radioactivity concentration in urineand low loading capacity of TLC assays, we could not quantitativelyidentify each radioactive species in the urine by TLC. Based onthe centrifugation method, it was shown that there was possiblepresence of 125I-SiO2 or their degraded product in urine even thougha certain extent of possible unbound radioisotopes from bond breakagefrom 125I-SiO2 in vivowas detected in supernatant of urine (SupplementalFig. 3). It is possible that nanoparticles were degraded into orthosilicicacid species smaller than the reported renal excretion threshold of7 nm, and were cleared through the renal route [8]. Previous studies byTang's group have suggested that intact SiO2 larger than 100 nm in sizecan be excreted in urine post intravenous injection as evidenced byTEM imaging [30]. The mechanism of large nanoparticle excretion intourine is not fully understood and warrants further studies. The examina-tion into radioactivity in feces indicated a similar fact that SiO2 was likelyexcreted through hepatobiliary route into feces as the dense silicate formpresent in the pellet of feces from centrifugation method, which agrees

with previous studies by Lo's group that reported the hepatobiliaryexcretion of SiO2 by fluorescence imaging [18]. Our results from excretionexperiments suggest that SiO2 could be biodegraded and excreted out ofbody.

Previous studies have reported the biodistribution of nanoparticulatesystems of similar size rangewith the SiO2 system evaluated in our study(100 nm and up) [31–35]. Results suggest that the biodistributionpattern varies distinctively between inorganic nanoparticles andorganic nanoparticles of similar dimensions. For example, colloidalgold nanoparticles (100 nm, 200 nm) showed high degree of uptakeby the liver, spleen and lung with limited presence in the blood [31],which is similar to the biodistribution of SiO2 used in our study. Inthe meanwhile, polymeric systems, such as chitosan nanoparticles[32], poly(lactic-co-glycolic acid) (PLGA) nanoparticles [33], or lipo-somes [34], exhibited sustained presence in blood circulation in additionto RES accumulation. It has been demonstrated that nanoparticles ofultra-low size (b100 nm in diameter) or with surface hydrophilicityevade the RES and have long circulation [35]. Hence in addition to thecore composition, size, surface functionalization and other physicochem-ical properties play crucial roles.

5. Conclusions

Of the materials tested in our study, it was demonstrated thatSiO2 biodistribution was influenced more by nanoparticle porosity,surface characteristics, and less by geometry. The nanoparticlesacross the board showed extensive distribution into liver andspleen with different concentrations in the lung. Mesoporous SiO2

accumulated in the lung to a higher extent than nonporous SiO2 ofsimilar size. Such accumulation was reduced by primary aminemodification. However, high aspect ratio amine-modified nanorodsshowed higher lung accumulation than the amine-modifiednanospheres. Results from tissue affinity indices and tissue/bloodconcentration ratio kinetic analyses suggest that tissue affinitywas mainly porosity and surface characteristics dependent.Nonporous SiO2 exhibited high affinity to the liver, and mesoporousSiO2 had higher affinity to the lung. Amine modification reduced theaffinity of SiO2 to the lung and kidneys. Two-compartmental pharmaco-kinetic analysis showed that amine-modified SiO2 tended to have higherVss than the pristine counterparts and that nonporous SiO2 exhibited ahigher Vss than mesoporous SiO2. SiO2 could be degraded and excretedout of the body by both urinary and hepatobiliary routes. This studyenables the systematic understanding of how physicochemical factorsaffect the living system and facilitates the rational design of SiO2 fortheir intended applications in the future.

Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jconrel.2012.05.046.

Acknowledgments

This research was supported by the National Institutes of Health(R01-DE19050) and the Utah Science Technology and Research(USTAR) Initiative. The authors would like to thank Dr. YongjianWang, Giridhar Thiagarajan and Shraddha Sadekar for technical adviceand their help with animal necropsy, Dr. Olinto Linares and Nate Larsonfor suggestions on pharmacokinetic analysis, and Dr. Khaled Greish forreview and critique of this manuscript.

References

[1] I.I. Slowing, J.L. Vivero-Escoto, C.W. Wu, V.S. Lin, Mesoporous silica nanoparticlesas controlled release drug delivery and gene transfection carriers, Adv. DrugDeliv. Rev. 60 (2008) 1278–1288.

[2] F. Torney, B.G. Trewyn, V.S. Lin, K. Wang, Mesoporous silica nanoparticles deliverDNA and chemicals into plants, Nat. Nanotechnol. 2 (2007) 295–300.

[3] S.H. Wu, Y. Hung, C.Y. Mou, Mesoporous silica nanoparticles as nanocarriers,Chem. Commun. (Camb) 47 (2011) 9972–9985.

54 T. Yu et al. / Journal of Controlled Release 163 (2012) 46–54

[4] C.P. Tsai, Y. Hung, Y.H. Chou, D.M. Huang, J.K. Hsiao, C. Chang, Y.C. Chen, C.Y. Mou,High-contrast paramagnetic fluorescent mesoporous silica nanorods as amultifunctional cell-imaging probe, Small 4 (2008) 186–191.

[5] I. Sokolov, S. Naik, Novel fluorescent silica nanoparticles: towards ultrabright sil-ica nanoparticles, Small 4 (2008) 934–939.

[6] S. Giri, B.G. Trewyn, M.P. Stellmaker, V.S. Lin, Stimuli-responsive controlled-releasedelivery system based on mesoporous silica nanorods capped with magneticnanoparticles, Angew. Chem. Int. Ed. Engl. 44 (2005) 5038–5044.

[7] J.K. Hsiao, C.P. Tsai, T.H. Chung, Y. Hung, M. Yao, H.M. Liu, C.Y. Mou, C.S. Yang, Y.C.Chen, D.M. Huang, Mesoporous silica nanoparticles as a delivery system of gado-linium for effective human stem cell tracking, Small 4 (2008) 1445–1452.

[8] J. Lu, M. Liong, Z. Li, J.I. Zink, F. Tamanoi, Biocompatibility, biodistribution, anddrug-delivery efficiency of mesoporous silica nanoparticles for cancer therapyin animals, Small 6 (2010) 1794–1805.

[9] C.-P. Tsai, C.-Y. Chen, Y. Hung, F.-H. Chang, C.-Y. Mou, Monoclonal antibody-functionalized mesoporous silica nanoparticles (MSN) for selective targetingbreast cancer cells, J. Mater. Chem. 19 (2009) 5737–5743.

[10] J.M. Rosenholm, V. Mamaeva, C. Sahlgren, M. Linden, Nanoparticles in targetedcancer therapy: mesoporous silica nanoparticles entering preclinical develop-ment stage, Nanomedicine (Lond) 7 (2012) 111–120.

[11] Y. Zhao, X. Sun, G. Zhang, B.G. Trewyn, I.I. Slowing, V.S. Lin, Interaction of meso-porous silica nanoparticles with human red blood cell membranes: size and sur-face effects, ACS Nano 5 (2011) 1366–1375.

[12] A. Nan, X. Bai, S.J. Son, S.B. Lee, H. Ghandehari, Cellular uptake and cytotoxicity ofsilica nanotubes, Nano Lett. 8 (2008) 2150–2154.

[13] F. Tang, L. Li, D. Chen, Mesoporous silica nanoparticles: synthesis, biocompatibilityand drug delivery, Adv. Mater. 24 (2012) 1504–1534.

[14] S.P. Hudson, R.F. Padera, R. Langer, D.S. Kohane, The biocompatibility of meso-porous silicates, Biomaterials 29 (2008) 4045–4055.

[15] K. Yamashita, Y. Yoshioka, K. Higashisaka, K. Mimura, Y. Morishita, M. Nozaki, T.Yoshida, T. Ogura, H. Nabeshi, K. Nagano, Y. Abe, H. Kamada, Y. Monobe, T. Imazawa,H. Aoshima, K. Shishido, Y. Kawai, T. Mayumi, S. Tsunoda, N. Itoh, T. Yoshikawa, I.Yanagihara, S. Saito, Y. Tsutsumi, Silica and titanium dioxide nanoparticles cause preg-nancy complications in mice, Nat. Nanotechnol. 6 (2011) 321–328.

[16] Q. He, Z. Zhang, F. Gao, Y. Li, J. Shi, In vivo biodistribution and urinary excretion ofmesoporous silica nanoparticles: effects of particle size and PEGylation, Small 7(2011) 271–280.

[17] R. Kumar, I. Roy, T.Y. Ohulchanskky, L.A. Vathy, E.J. Bergey, M. Sajjad, P.N. Prasad,In vivo biodistribution and clearance studies using multimodal organically modi-fied silica nanoparticles, ACS Nano 4 (2010) 699–708.

[18] J.S. Souris, C.H. Lee, S.H. Cheng, C.T. Chen, C.S. Yang, J.A. Ho, C.Y. Mou, L.W. Lo, Sur-face charge-mediated rapid hepatobiliary excretion ofmesoporous silica nanoparticles,Biomaterials 31 (2010) 5564–5574.

[19] Y. Geng, P. Dalhaimer, S. Cai, R. Tsai, M. Tewari, T. Minko, D.E. Discher, Shapeeffects of filaments versus spherical particles in flow and drug delivery, Nat.Nanotechnol. 2 (2007) 249–255.

[20] N. Nasongkla, B. Chen, N. Macaraeg, M.E. Fox, J.M. Frechet, F.C. Szoka, Dependenceof pharmacokinetics and biodistribution on polymer architecture: effect of cyclicversus linear polymers, J. Am. Chem. Soc. 131 (2009) 3842–3843.

[21] S. Muro, C. Garnacho, J.A. Champion, J. Leferovich, C. Gajewski, E.H. Schuchman, S.Mitragotri, V.R. Muzykantov, Control of endothelial targeting and intracellulardelivery of therapeutic enzymes by modulating the size and shape ofICAM-1-targeted carriers, Mol. Ther. 16 (2008) 1450–1458.

[22] Arnida, M.M. Janat-Amsbury, A. Ray, C.M. Peterson, H. Ghandehari, Geometry andsurface characteristics of gold nanoparticles influence their biodistribution anduptake by macrophages, Eur. J. Pharm. Biopharm. 77 (2011) 417–423.

[23] T. Yu, A. Malugin, H. Ghandehari, Impact of silica nanoparticle design on cellulartoxicity and hemolytic activity, ACS Nano 5 (2011) 5717–5728.

[24] T. Yu, K. Greish, L.D. McGill, A. Ray, H. Ghandehari, Influence of geometry, porosity,and surface characteristics of silica nanoparticles on acute toxicity: their vasculatureeffect and tolerance threshold, ACS Nano 6 (2012) 2289–2301.

[25] N. Malik, R. Wiwattanapatapee, R. Klopsch, K. Lorenz, H. Frey, J.W. Weener, E.W.Meijer, W. Paulus, R. Duncan, Dendrimers: relationship between structure and bio-compatibility in vitro, and preliminary studies on the biodistribution of 125I-labelledpolyamidoamine dendrimers in vivo, J. Control. Release 65 (2000) 133–148.

[26] S.R. Urva, B.S. Shin, V.C. Yang, J.P. Balthasar, Sensitive high performance liquidchromatographic assay for assessment of doxorubicin pharmacokinetics inmouse plasma and tissues, J. Chromatogr. B Analyt. Technol. Biomed. Life Sci.877 (2009) 837–841.

[27] D.L. Thai, L.N. Yurasits, G.R. Rudolph, J.M. Perel, Comparative pharmacokineticsand tissue distribution of the d-enantiomers of para-substituted methylpheni-date analogs, Drug Metab. Dispos. 27 (1999) 645–650.

[28] S.E. Gratton, P.D. Pohlhaus, J. Lee, J. Guo, M.J. Cho, J.M. Desimone, Nanofabricatedparticles for engineered drug therapies: a preliminary biodistribution study ofPRINT nanoparticles, J. Control. Release 121 (2007) 10–18.

[29] G. Xie, J. Sun, G. Zhong, L. Shi, D. Zhang, Biodistribution and toxicity of intravenouslyadministered silica nanoparticles in mice, Arch. Toxicol. 84 (2010) 183–190.

[30] X. Huang, L. Li, T. Liu, N. Hao, H. Liu, D. Chen, F. Tang, The shape effect of meso-porous silica nanoparticles on biodistribution, clearance, and biocompatibility invivo, ACS Nano 5 (2011) 5390–5399.

[31] G. Sonavane, K. Tomoda, K. Makino, Biodistribution of colloidal gold nanoparticlesafter intravenous administration: effect of particle size, Colloids Surf. B Bio-interfaces 66 (2008) 274–280.

[32] T. Banerjee, S. Mitra, A. Kumar Singh, R. Kumar Sharma, A. Maitra, Preparation,characterization and biodistribution of ultrafine chitosan nanoparticles, Int. J.Pharm. 243 (2002) 93–105.

[33] K.S. Yadav, K. Chuttani, A.K. Mishra, K.K. Sawant, Effect of size on the bio-distribution and blood clearance of etoposide-loaded PLGA nanoparticles, PDA J.Pharm. Sci. Technol. 65 (2011) 131–139.

[34] D. Liu, A. Mori, L. Huang, Role of liposome size and RES blockade in controlling bio-distribution and tumor uptake of GM1-containing liposomes, Biochim. Biophys. Acta1104 (1992) 95–101.

[35] S.D. Li, L. Huang, Pharmacokinetics and biodistribution of nanoparticles, Mol.Pharm. 5 (2008) 496–504.