Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
ACS Nano. Author manuscript; available in PMC 2014 January 23.
Published in final edited form as:
PMCID: PMC3899722

Engineered Design of Mesoporous Silica Nanoparticles to Deliver Doxorubicin and Pgp siRNA to Overcome Drug Resistance in a Cancer Cell Line


Overexpression of drug efflux transporters such as P-glycoprotein (P-gp) protein is one of the major mechanisms for multiple drug resistance (MDR) in cancer cells. A new approach to overcome MDR is to use a co-delivery strategy that utilizes a siRNA to silence the expression of efflux transporter together with an appropriate anti-cancer drug for drug resistant cells. In this paper, we report that mesoporous silica nanoparticles (MSNP) can be functionalized to effectively deliver a chemotherapeutic agent doxorubicin (Dox) as well as Pgp siRNA to a drug-resistant cancer cell line (KB-V1 cells) to accomplish cell killing in an additive or synergistic fashion. The functionalization of the particle surface with a phosphonate group allows electrostatic binding of Dox to the porous interior, from where the drug could be released by acidification of the medium under abiotic and biotic conditions. In addition, phosphonate modification also allows exterior coating with the cationic polymer, polyethylenimine (PEI), which endows the MSNP contemporaneously deliver Pgp siRNA. The dual delivery of Dox and siRNA in KB-V1 cells was capable of increasing the intracellular as well as intranuclear drug concentration to levels exceeding that of free Dox or the drug being delivered by MSNP in the absence of siRNA co-delivery. These results demonstrate that it is possible to use the MSNP platform to effectively deliver a siRNA that knocks down gene expression of a drug exporter that can be used to improve drug sensitivity to a chemotherapeutic agent.

Keywords: Mesoporous silica nanoparticles, Drug resistance, Doxorubicin, P-glycoprotein (Pgp), Synergistic effects, siRNA delivery


Due to its unique structure and ease with which the surface can be functionalized, mesoporous silica nanoparticles (MSNP) constitute a multi-functional platform that can be used for drug14 and nucleic acid delivery.5 For therapeutic purposes, we prefer nanoparticles that contain a phosphonate surface coating because of the ease of particle dispersal, good bio-safety index as well as the ability to adsorb cationic polyethylenimine (PEI) polymers for complexing and delivery of DNA and siRNA.5 Since the polymer attachment leaves the porous interior free for drug binding and delivery, it establishes the potential to achieve simultaneous drug and nucleic acid delivery.6 There are a number of circumstances where dual drug and siRNA delivery could achieve a synergistic therapeutic outcome. One example is the restoration of drug sensitivity in cancer cells by knockdown of genes that are involved in the resistance to one or more chemotherapeutic agents. An example is the inducible P-glycoprotein (Pgp) gene that encodes for a gene product known as the multiple drug resistance protein 1 (MDR-1).7 Pgp is constitutively expressed in normal cells such as capillary endothelial cells in the blood brain barrier but also is selectively overexpressed in carcinomas of the stomach, breast, pancreas and cervix in response to a number of chemotherapeutics agents.8 If overexpressed, Pgp could lead to drug resistance because MDR-1 contributes to the formation of a drug efflux pump that prevents the intracellular buildup of chemotherapeutic agents.9 Because Pgp overexpression is one of the major mechanisms of multiple drug resistance (MDR) in cancer cells, knockdown of Pgp gene expression by nanoparticle siRNA delivery could help to restore the intracellular drug levels to the concentrations required for induction of apoptosis and cytotoxicity. Thus, dual drug and siRNA delivery by nanoparticles can be used to overcome drug resistance in MDR cancer cells.

In order to test the utility of MSNP as a dual delivery platform, we used the drug-resistant squamous carcinoma cell line, KB-V1, to see if Pgp knockdown restores doxorubicin (Dox) sensitivity. KB-V1 cells exhibit MDR as a result of Pgp overexpression.10 In order to effectively engineer particles to deliver Dox as well as Pgp siRNA, it was necessary to demonstrate that particle functionalization by the attachment of negative (phosphonate) as well as positive (PEI) surface groups is functionally effective. Moreover, since PEI delivery of siRNA constructs to the cytosol requires intermediary lysosomal processing, we were also interested to determine whether this endocytic route is appropriate for Dox delivery. We demonstrate that Dox can be stably attached to the porous interior by a proton-sensitive electrostatic binding interaction that allows effective drug release from the acidifying LAMP-1-positive compartment. Pgp siRNA co-delivery increases intracellular Dox concentrations with improved cytotoxic killing. The improvement of Dox resistance provides proof-of-principal testing that MSNP can be engineered to provide contemporaneous drug and siRNA delivery by effective use of charge and the state of protonation or deprotonation at the particle surface.


PEI-coated MSNP effectively deliver Pgp siRNA and knock down Pgp expression in KB-V1 cells

We confirmed the status of KB-31 and KB-V1 cells lines as Dox-sensitive and Dox-resistant cell lines, with IC50 values of 0.21 and 53.0 µg/ml, respectively (Supporting information Fig. S1A). We also confirmed through immunoblotting analysis that the Pgp expression in KB-V1 was > 1,000 times that of KB-31 cells (Supporting information Fig. S1B). In order to determine if Pgp overexpression in KB-V1 cells is impacted by Pgp siRNA delivery, PEI polymers in the size range 1.8–25 kD were electrostatically bound to the phosphonate-MSNP surface. Polymer binding to our 100–120 nm size MSNP, exhibiting uniform pore sizes of 2–2.5 nm was confirmed by TEM (Fig. 1A, left). The TEM image of the particles decorated with the 10 kD polymer shows that the surface coating (arrows) did not occupy the porous interior (Fig. 1A, right). To optimize the particle dispersal for biological experimentation, the PEI-coated particles were further treated with 1 mg/ml BSA before transfer to the complete cell culture medium. BSA coating significantly improves particle dispersal in water as well as salt-containing complete DMEM (Fig. 1B). While the PEI-coated particles exhibit a high positive zeta potential value in water, this value changes to slightly negative upon dispersal in complete DMEM containing FCS (Fig. 1B). Note that similar analysis was also performed on the Dox loaded PEI-MSNP samples with or without siRNA binding, and that no significant changes were found with or without cargo loading (Supporting information S2).

Figure 1
Physiocochemical characterization of PEI coated MSNP. (A) TEM image of phosphonate-MSNP before and after coating with the 10 kD PEI polymer. The arrows indicate that the polymer decorates the MSNP surface but leaves the porous interior accessible to drug ...

In order to determine the optimal N/P ratios for Pgp siRNA delivery, we used agarose gel electrophoresis to evaluate binding of various amounts of siRNA to MSNP coated with the different polymer lengths as shown in Fig. 2A. This demonstrated that binding capacity increases with increasing size of PEI polymers, indicating that all siRNA was bound at a N/P ratio >16 (PEI 1.8kD), >16 (PEI 10 kD) and >8 (PEI 25 kD). Based on this result, we used N/P ratios of 80, 80 and 10 for polymer lengths of 1.8, 10 or 25 kD, respectively for siRNA knockdown experiments. It is worth mentioning that Dox trapped in the pores of MSNP does not influence siRNA binding to the particle surface (Supporting information Fig. S3) and also that siRNA binding does not significant change the zeta potential of MSNP (Supporting information Fig. S2). Subsequent performance of confocal microscopy utilizing dual-labeled particles (Texas red-labeled-siRNA adsorbed to FITC-labeled MSNP) demonstrated a high rate of cellular uptake in KB-V1 cells in accordance with polymer length (Fig. 2B). Image J analysis confirmed a significant increase in siRNA uptake for particles coated with the 10 and 25 compared to the 1.8 kD polymer (Fig. 2C). The reason for this high uptake is due to the positively charged amines, which facilitates strong PEI binding to and wrapping by the surface membrane.5 Please notice that the Pgp siRNA (red) and FITC-MSNP (green) co-localize (yellow, merged) in the cell, demonstrating that the nucleic acid is stably attached to the PEI-coated particle surface (Fig. 2B). We have previously demonstrated that nucleic acids bound to the PEI-MSNP surface are resistant to enzymatic cleavage.5

Figure 2Figure 2
Effective Pgp siRNA delivery and gene knockdown in KB-V1 cells. (A) Agarose gel electrophoresis of PEI-coated MSNP to which Pgp siRNA was complexed at various nanoparticle to nucleic acid (N/P) ratios. M is molecular weight marker. The Φ lane ...

To assess the efficacy of Pgp siRNA delivered by PEI-coated MSNP, Pgp expression was followed by western blotting (Fig. 2D). This demonstrated 80 or 90% reduction, in MDR-1 expression in cells treated with MSNP coated particles that contain the 10 and 25 kD polymers, respectively. This efficacy was maintained in Dox-loaded MSNP (Supporting information Fig. S4). Scrambled siRNA-PEI-MSNP (marked as “X” in Fig. 2D) was used as a negative control to rule out any impact by the siRNA delivery method. No Pgp knockdown was seen in the scrambled siRNA-MSNP group. The PEI-coated particles were also more effective than the commercially available transfection agent, Lipofectamine 2000, which is widely used in molecular biology.

An important consideration in the use of PEI is its potential cytotoxicity.5 In this regard, we have recently shown that the polymer size plays an important role in the cytotoxicity that results from proton sequestration by unsaturated PEI amines in the lysosomal compartment.5 Thus, before the performance of Dox-induced cytotoxicity, it was necessary to compare the cytotoxic potential of MSNP coated with different PEI polymer lengths as shown in the Supporting information Fig. S5. In brief, the findings were that MSNP coated with the 25 kD polymer showed considerable toxicity in the MTS assay when the particle concentration is > 25 µg/ml. By contrast, particles coated with 10 kD PEI have no toxic effects if the concentration is kept <100 µg/ml and the exposure time limited to ≤ 24hr (Supporting information Fig. S5). Consequently, we used MSNP at doses < 100 µg/ml and for exposure time periods < 24 hrs in the drug delivery studies to avoid effects on cell viability. Although no cytotoxicity was seen with the 1.8 kD PEI polymer (Fig. S5A), this particle is relatively ineffective for siRNA delivery and Pgp knockdown (Fig. 2B, C, and D).

PEI-coated MSNP efficiently bind and deliver Dox via a proton-sensitive mechanism

We have recently demonstrated that MSNP are capable of loading and releasing water-insoluble drugs (paclitaxel and camptothecin) by a phase transfer mechanism that can be reversed by ethanol washing of the particles to demonstrate the role of hydrophobicity in MSNP drug entrapment.5, 11 A key question was whether similar effective packaging of water-soluble Dox, is possible at physiological pH, where the drug (pKa = 8.2) carries a positive charge. One conceivable approach is electrostatic attachment to the negatively charged MSNP surface. To test this possibility, Dox was loaded in MSNP decorated with OH, COOH or phosphonate groups. These particles were also compared to positively charged particles decorated with amine groups. After washing of the drug-bound particles and quantification of Dox release by HCl, the loading capacities of OH-, COOH- and phosphonate-MSNP were 1.2%, 4.2% and 8.4% (w/w), respectively (Fig. 3A). By contrast, amine-decorated particles showed low (<0.1%, w/w) Dox binding capacity (Fig. 3A). The principle of electrostatic binding to a negatively charged particle surface was also demonstrated through the use of a cationic dye, Hoechst 33342, which bound to negative but not a positive MSNP surface (Supporting information Fig. S6). Importantly, electrostatic binding of Dox to the phosphonate surface was not negated by when the particles were also coated with 10 or 25 kD PEI polymers; these particles demonstrated equivalent loading capacity to the uncoated phosphonate particles (Fig. 3B).

Figure 3Figure 3Figure 3
Phosphonate-MSNP effectively binds Dox via a proton-sensitive mechanism. (A) Modeling studies using positively and negatively charged MSNP under abiotic conditions. Loading yield of Dox in MSNP with various surface modifications. A photograph of the Dox-loaded ...

Importantly, Dox could be released in a time-dependent manner from the phosphonate-MSNP or PEI-phosphonate-MSNP surface by lowering of the solution pH (Fig. 3C and D). The Dox release profile was not be affected by siRNA binding to PEI on exterior surface of the particle (Supporting information Fig. S7). This establishes the possibility that intracellular drug release should be possible if the particles are capable of gaining entrance to acidifying cellular compartments. In order to determine whether this is possible in KB-V1 cells, we studied the intracellular localization of FITC-labeled MSNP in relation to lysosomal co-staining by fluorescent labeled anti-LAMP-1 antibody. Indeed, confocal microscopy confirmed >55% co-localization of the green-labeled particles with the red-labeled lysosomes (Fig. 3E). Moreover, in a subsequent confocal study we demonstrated that released Dox from PEI-coated particles after entrance into the lysosome could reach the nucleus, which was brightly stained by the fluorescent drug (Fig. 3F, upper panel). Please notice that some of the drug was retained in particles localized in the peri-nuclear region. This visual image changed completely when cells were treated with NH4Cl; most Dox remained confined to the lysosomal compartment with little or no nuclear staining (Fig. 3F, lower panel). This suggests that the interference in lysosomal acidification by NH4Cl prevents effective Dox release to KB-V1 nuclei.

Dual delivery of Dox and Pgp siRNA partially restores drug sensitivity in KB-V1 cells

We have already discussed the relative inefficiency of free Dox to induce KB-V1 cytotoxicity as a result of the rapid rate by which the drug was being exported by the overexpressed Pgp. Intracellular Dox concentration can be determined by measuring cellular Dox fluorescence intensity in a microplate reader (Fig. 4A). The comparatively low drug uptake after treatment with free Dox was slightly improved by delivering the drug via the phosphonate-MSNP. While the total amount of intracellular drug increased when being delivered by particles coated with the 10 kD PEI polymer (Fig. 4A), little of the drug reached the nucleus as determined by confocal imaging (Fig. 4B). Interestingly, the intracellular Dox concentration increased significantly in the presence of siRNA (Fig. 4A) so that there was also a significant increase in nuclear Dox staining by 72 hr (Fig. 4B). This contrasts with most of the Dox being confined to the endosomal compartment in cells not receiving siRNA, suggesting that although PEI-MSNP may take more Dox into to cells, any released drug is rapidly extruded from the cell before reaching the nucleus (Fig. 4B). This suggests that knocking down Pgp expression (Supporting information Fig. S4) allows a sufficient quantity of the drug being slowly released from the particle to enter the nucleus where it induces cytotoxicity. A quantitative measurement of Dox release to the nucleus through the use of Image J software confirmed a statistically significant increase of the drug in KB-V1 nuclei after delivery by siRNA-PEI-Dox-MSNP as compared to free Dox or Dox loaded into the particles without siRNA (Fig. 4C). Dox delivered by PEI-MSNP in the presence of siRNA significantly enhance intranuclear Dox concentration when compared to free Dox or Dox delivered by MSNP or PEI-MSNP without siRNA.

Figure 4
Simultaneous delivery of Dox and Pgp siRNA to the nucleus leads to a synergistic increase in cellular and nuclear Dox levels in KB-V1 cells. (A) Quantitative comparison of Dox levels using a fluorescent readout of cellular drug levels 72 hrs after introduction ...

In order to reconcile these findings with improved cell killing, we used the MTS assay to compare KB-V1 cytotoxicity under incremental Dox concentrations (Fig. 5A). Based on the calculated IC50 values of the various formulations, it was possible to rank the killing efficiency as follows: siRNA-PEI-Dox-MSNP > PEI-Dox-MSNP ≈ Dox-MSNP > free Dox. Moreover, the IC50 value of the siRNA-delivering MSNP was approximately 2.5 times lower than the IC50 of free Dox or other Dox-loaded particles. This suggests an additive effect between the drug and the siRNA that are being delivered by MSNP. Although it was not possible to restore drug sensitivity to the level seen in KB-31 cells, we could clearly observe a much higher percentage of apoptotic KB-V1 cells after treatment with Dox-loaded MSNP that co-delivers siRNA as compared to free Dox or Dox delivered by non-siRNA delivering particles (Fig. 5B). We confirmed the flow cytometry data through the use of another apoptosis assay using TUNEL staining (Supporting information Fig. S8), which showed an identical trend. Failure to fully restore drug sensitivity could be due to extremely high levels of Pgp expression and/or additional drug resistant pathways in KB-V1 that were not targeted in the current research. Because of the duration of time it takes for the siRNA to exert its effect, we also analyzed cell viability after longer treatment periods (e.g. 96 hrs) but did not observe any improvement in cytotoxicity (Supporting information Fig. S9). In addition, we also treated the KB-V1 cells by co-administration of Dox and Tariquidar, which is a potent and effective Pgp inhibitor in human clinical trial, to see if we can restore KB-V1 sensitivity to that seen in KB-31 cells. While this combination significantly improves cell killing capability in KB-V1 cells, it is incapable of completely restoring Dox sensitivity to the level seen in KB-31 cells (Supporting information Fig. S10).

Figure 5Figure 5
Comparison of the cytotoxic effects of different delivery modalities of Dox in KB-V1 cells. (A) MTS cell viability assay showing that MSNP delivery of Dox concomitant with Pgp siRNA is capable of improving the induction of cytotoxicity by free Dox or ...


Here we show that MSNP can be functionalized to deliver a chemotherapeutic agent as well as Pgp siRNA to a drug-resistant cancer cell line. The functionalization of the particle surface with a phosphonate group allows electrostatic binding of Dox to the porous interior, from where the drug could be released by acidification of the medium under abiotic and biotic conditions. In addition, phosphonate modification also allows exterior coating with the cationic polymer, PEI, which endows the MSNP with the ability to contemporaneously bind and deliver Pgp siRNA. The dual delivery of Dox and siRNA in KB-V1 cells was capable of increasing the intracellular as well as intranuclear drug concentration to levels exceeding that of free Dox or the drug being delivered by MSNP without siRNA co-delivery. This reflects the ability of the siRNA to effectively knock down Pgp expression and therefore interfering in drug efflux as one of the resistance mechanisms in KB-V1 cells. Unlike hydrophobic cargo (Supporting information Fig. S11), Dox can be released from the lysosome by a proton-sensitive mechanism. While clearly effective at improving cytotoxic killing through its dual delivery capabilities, siRNA-PEI-MSNP could not restore Dox sensitivity in KB-V1 to the level seen in drug-sensitive KB-31 cells. This is likely due to the extremely high levels of Pgp expression, as confirmed in Fig. S1.

Two of the major problems in cancer chemotherapy are toxic side effects as well as development of MDR in cancer cells. Nanoparticle drug delivery is capable of overcoming both problems through tumor cell targeting as well as the capability to overcome drug resistance.9, 12 MDR can basically be divided into two distinct categories, namely pump and non-pump resistance.9, 13 Pump resistance refers to the inducible formation of membrane-bound channels or pores that actively expel a series of structural and functional distinct chemotherapeutic agents from the cell. Drug efflux significantly decreases the intracellular concentration that limits their cytotoxic potential. The key proteins involved in pump resistance are Pgp and MRP-1, while the major mechanism in non-pump resistance is activation of cellular anti-apoptotic defense pathways, including drug-induced expression of Bcl-2 protein.13 Moreover, the pump and non-pump resistance mechanisms could be mutually interactive.9 Given this background, a number of nanomaterial design strategies can be used to overcome drug resistance:

  1. The first is co-delivery of the chemotherapeutic drug with a pharmaceutical agent that interferes in pump activity or in non-pump pathways. One example is the use of verapamil as a Pgp inhibitor that have been combined with Dox, aimed at reducing cardiotoxicity of Dox as well as overcoming Pgp-mediated MDR.14 In Dox-resistant K562 leukemia cells, a liposomal Dox-verapamil complex doubled the cytotoxicity index (IC50=11.4 µM) of free Dox (IC50=23.4 µM).14 Another example is using a polymeric nanoparticle formulation of poly(ethyleneoxide)-poly(epsilson-caprolactone) to co-administer ceramide that is capable of lowering the apoptotic threshold to paclitaxel, thereby enhancing paclitaxel-induced cytotoxicity in a MDR human ovarian cancer cell line.15
  2. A second strategy is to use nanoparticles to deliver the chemotherapeutic together with siRNA that interfere in key protein expression in pump-dependent or independent drug resistance pathways. There are reports demonstrating that compared to the free drug, dual delivery with a nucleic acid is capable of enhancing the efficacy in a synergistic fashion.16, 17 Chen et al showed that MSNP served as a drug delivery vehicle to deliver Dox and Bcl-2 siRNA that can effectively silence the Bcl-2 mRNA and significantly suppress the non-pump resistance, and therefore result in enhanced cell killing capability of Dox in multidrug resistant A2780/AD human ovarian cancer cells.16 Another example is paclitaxel that was delivered by poly(d,l-lactide-co-glycolide) nanoparticles along with Pgp siRNA to the MDR murine mammary cancer cell line, JC. The dual delivery system showed significantly higher cytotoxicity in vitro and significantly greater inhibition of tumor growth in vivo than nanoparticles loaded with paclitaxel alone.17
  3. The third strategy is to deliver the drug together with a combination of siRNA's that interfere in both pump (e.g. Pgp) and non-pump (e.g. Bcl-2) mechanisms. Such an approach may be necessitated by the co-existence of pump and non-pump resistance mechanisms.9, 13 In this regard, it has been demonstrated that the increase in intracellular drug concentration as a result of the suppression of drug efflux pumps leads to almost proportional activation of anti-apoptotic cellular defense.7, 9 As one example, a cationic liposome carrier system was developed to deliver Dox contemporaneously with two species of siRNA targeting of MRP-1 as well as Bcl-2.13 This triple component (Dox, MRP-1 siRNA, and Bcl-2 siRNA) delivery system demonstrated enhanced Dox cytotoxicity in human MDR H69AR lung cancer cells which showed >100 fold enhancement in cytotoxicity.13 Similar attempts in our own experiments in KB-V1 cells where we combined Bcl-2 with Pgp siRNA and Dox did not improve the cytotoxicity or IC50 significantly (not shown).

This paper demonstrates how MSNP can be engineered to effectively deliver a drug together with siRNA. Before we discuss what we learned from KB-V1 cells, it is worthwhile summarizing the key features of our interior and exterior design modifications. First, because Dox (pKa = 8.2) is positively charged at physiological pH, we used a number of surface charge modifications to demonstrate the utility of using electrostatic charge to bind and deliver Dox from the MSNP surface (Fig. 3A, ,4B).4B). These studies demonstrated that Dox as well as the cationic dye, Hoechst 33342 (pKa = 11.9), are capable of binding to negatively charged surfaces from where both agents could be released by dropping the environmental pH (Fig. 3C–F). Thus, when the MSNP are suspended in buffered cell culture medium (pH = 7.4), the attached phosphonate (pKa = 2) and carboxylate (pKa = 5) groups are de-protonated and assume a negative charge, whereas unmodified OH-MSNP (pKa = 7) is near its isoelectric point and therefore not quite as effective for electrostatic binding (Fig. 3A). By contrast, the amine groups (pKa ~ 9) are protonated at physiological pH and exhibit a positive charge that prevents electrostatic binding of the same agents.18, 19 From a therapeutic perspective, we focused on phosphonate-MSNP due to good particle dispersibility and biocompatibility. The ability to release Dox from the interior surface by proton interference allowed us to achieve Dox delivery inside the cell in an acidifying compartment (Fig. 3F, upper panel). The role of the lysosomal proton pump is supported by the finding that NH4Cl interferes in Dox release to the nucleus (Fig. 3F, lower panel). Phosphonate attachment also facilitates the binding of cationic PEI to the particle exterior (Fig. 1A, right panel). This binding interaction is sufficiently strong to allow the polymer and attached siRNA to stay on the particle surface until entry into the lysosomal compartment. PEI may play a role in firm cellular attachment and selection of the initial endosomal compartment. Importantly, the polymer is attached to the particle surface to leave the pores accessible to Dox binding (Fig. 3B).

It is important to briefly mention the importance of selecting the correct PEI polymer size to prevent particle toxicity.5, 20, 21 It is well known that PEI exerts cytotoxic effects when used by itself or attached to the nanoparticle surface as a polyplexing agent.5 In this regard, we have recently demonstrated that the 25 kD PEI polymer as well as high doses of the 10 kD polymer can render the MSNP toxic as a result of the proton sponge effect in the lysosome.5 For these reasons, it was necessary to limit the particle dose and exposure time to within safe limits to conduct Dox and siRNA delivery with PEI 10 kD polymer (Supporting information Fig. S5). Curiously and somewhat paradoxically, siRNA delivery to the cytosol is also dependent on the proton sponge effect of the PEI-coated particle and in this case, the lysosome appears to be a key organelle in the dual drug delivery paradigm.

Above design features extend the utility of MSNP as a drug delivery platform for chemotherapeutic agents. In addition to being able to deliver hydrophobic drugs such as campthothecin and paclitaxel, we show that it is also possible to deliver water soluble drugs by a packaging and release mechanism that is quite different from the phase transition principle that is involved in hydrophobic drug delivery (see Supporting information Fig. S11). While it is not possible to release hydrophobic drugs through an acidification mechanism, camptothecin can be extracted from phosphonate-MSNP pores by ethanol treatment (Fig. S11). The release characteristics of Dox are exactly opposite, namely a release response to protons but not ethanol (Fig. 3C, and D). Taken together, these results demonstrate the dynamic feathers of MSNP for obtaining optimal drug packaging and delivery.

Our study demonstrates the feasibility of the MSNP platform to improve the cytotoxicity of Dox by co-delivery of Pgp siRNA as proof-of-principle. Downregulation of Pgp expression allowed the intranuclear Dox levels to increase above the threshold required for inducing apoptosis and cell death. However, since the KB-V1 cell line expresses very high Pgp level that may go beyond the clinically relevant expression level, it may not be possible to restore drug sensitivity to the level seen in KB-31. It is reasonable to expect that the MSNP delivery platform will be more effective in a lesser Pgp expressing cell type. In addition, we also learned from the study that the MDR phenotype is quite complex and often involve a combination of drug resistance mechanisms such as increased efflux, blocked apoptosis, decreased drug influx, increased drug metabolism, and increased DNA repair.7, 9. Preliminary research has revealed that 36 genes were either up- or down-regulated in KB-V1 cells compared to KB-31 cells. These genes can be categorized into several groups, including oxidative stress regulation (HBB, IDH), drug metabolism (HMGCS1, ACAT1), signal transduction (CGA, SGK), tumor suppression (PTEN), etc.22 It is possible that these gene products could contribute to MDR besides Pgp overexperssion. What those additional mechanisms is still uncertain but apparently does not involve Bcl-2 as inclusion of Bcl-2 with Pgp siRNA did not significantly improved cell killing. Understanding of abovementioned MDR mechanisms may provide us new opportunity to develop more therapeutic components or combinations to achieve better therapeutic effects.


MSNP can be functionalized to act as a dual delivery vehicle for Dox as well as Pgp siRNA in a drug-resistant cancer cell line. To improve the drug sensitivity of KB-V1 cells, we used phosphonate attachment to deliver the drug as well as the siRNA via a lysosomal processing pathway. This dual delivery system increased the intracellular Dox levels to the extent that it improves cytotoxic killing in this KB-V1 MDR cell line. This strategy could be an effective new approach for the treatment of cancers that develop multiple drug resistance.



Tetraethylorthosilicate (TOES, 98%), cetyltrimethylammonium bromide (CTAB, 95%), fluorescein isothiocyanate (FITC, 90%), doxorubicin hydrochloride (Dox, >98%), camptothecin (CPT, 95%), vinblastine, polyethylenimine (PEI, branched, MW 25 kD), bafilomycin A (≥ 95%), ammonium chloride, β-actin antibody, and bovine serum albumin (BSA) were from Sigma (St. Louis, MO). Polyethylenimine (branched, MW 1.8 and 10 kD) was purchased from Alfa Aesar. 3-trihydroxysilylpropyl methylphosphonate, cyanoethyltriethoxysilane, and aminopropyltriethoxysilane were purchased from Gelest. Dulbecco’s Modified Eagle’s medium (DMEM), penicillin/streptomycin, and L-glutamine were purchased from Invitrogen (Carlsbad, CA). Fetal calf serum (FCS) was from Atlanta Biologicals, Inc (Lawrenceville, GA). LAMP-1 antibody was obtained from Abcam (Cambridge, MA). siRNA for Pgp knockdown was purchased from IDT Technologies (Coralville, IA). Tariquidar (>97%) was purchased from MedKoo Biosciences, Inc. For all experiments and analyses, water was de-ionized and filtered with a 0.45 µm pore size polycarbonate syringe filter (Millipore, Billerica, MA). All chemicals were reagent grade and used without further purification or modification.

Particle Synthesis, PEI coating and interior surface functionalization

The MSNPs were synthesized according to our previously published sol-gel procedure.5, 11 Briefly, for the synthesis of unmodified MSNP (OH-MSNP), 100 mg of CTAB was dissolved in a solution of 48 mL water and 0.35 mL sodium hydroxide (2 M) and heated to 80 °C. One half mL of TEOS was added into the aqueous solution containing CTAB surfactants. For the phosphonate modification, 3-trihydroxysilylpropyl methylphosphonate was added to the mixture 15 minutes after the addition of TEOS. The CTAB surfactants were then removed from the pores by heating the particles in acidic ethanol. To perform PEI coating, 5 mg of phosphonate-modified MSNP were dispersed in a solution containing 2.5 mg PEI (1.8 kD, 10 kD, 25 kD) in 1 ml absolute ethanol. After sonication and stirring for 30 min the PEI coated particles were washed with PBS. The amount of polymer coated onto the particle surface was approximately 5 weight percentage.

Because Dox is positively charged under physiological pH, it was necessary to demonstrate that phosphonate or other anionic surface groups are effective for drug binding, including in particles that have been coated with PEI. To demonstrate this principle, we decorated particle surfaces with carboxylate (COOH) and amine groups in addition to phosphonate and silanol (OH groups). The surface functionalization was achieved by mixing organoalkoxysilanes (made up in ethanol) with TEOS before adding the mixture into the CTAB solution.23 For carboxylate modification, 50 µL cyanoethyltriethoxysilane was mixed with 500 µL ethanol and 500 µL TEOS, then added into the surfactant solution. After the surfactant removal process, the particles were further heated in a solution of 50% sulfuric acid to hydrolyze the cyanide groups into carboxylic groups. For amine modification, 50 µL of aminopropyltriethoxysilane was first mixed with 500 µL ethanol and 500 µL TEOS before adding to the surfactant solution. After 2 hrs, the solution was cooled to room temperature and the materials were washed with methanol before the surfactant removal process.

Physicochemical characterization of MSNP

PEI-coated phosphonate MSNP were characterized for size, zeta potential, and shape, respectively. The shape and porous structure were characterized using transmission electron microscopy (JEOL JEM 2010, JEOL USA, Inc., Peabody, MA). Microfilms for TEM imaging were made by placing a drop of the respective MSNP suspensions onto a 200-mesh copper TEM grid (Electron Microscopy Sciences, Washington, PA) and then drying at room-temperature overnight. A minimum of 5 images for each sample was captured and representative images included in Fig. 1A. Particle size and zeta potential in pure water, after stabilization with 1 mg/mL BSA in water, or in cell culture medium were measured by ZetaSizer Nano (Malvern Instruments Ltd., Worcestershire, UK). All the measurements were performed in 40 µg/ml MSNP suspensions in filtered water or filtered complete cell culture media at pH 7.4. The analysis was also studied on Dox loaded particles. Similar analysis was also performed on cargo (Dox and siRNA) loaded PEI-MSNP samples.

PEI-MSNP coating with Pgp siRNA and use of agarose gel electrophoresis to determine N/P ratios

An agarose gel retardation assay was used to determine the siRNA binding to PEI-MSNP. 0.1 µg siRNA was mixed with 0.4–6.4 µg amount PEI-MSNP in aqueous solution to obtain particle/nuclei acid (N/P) ratios of 4–64. 10 µL of the polyplex solution was mixed with 2 µL of 6× loading buffer and electrophoresed in a 1.5 % agarose gel containing 0.5 µg/ml ethidium bromide (EB) at 100 V for 30 minutes in Tris-boric acid (TBE) running buffer (pH = 8). Nucleic acid bands were detected by UV light (254 nm) and the photos were captured in a Bio-Rad imaging system (Hercules, CA). The results were used to calculate the threshold N/P ratios for subsequent experiments. The threshold is defined as the lowest N/P ratio value that prevents free siRNA from entering the gel. To determine the influence of Dox loading, the gel electrophoresis assay was also performed.

Cell culture

All cell cultures were maintained in 25 cm2 or 75 cm2 cell culture flasks in which the cells were passaged at 70–80% confluency every 2–4 days. The drug-sensitive KB-31 line was cultured in Dulbecco’s Modified Eagle Medium (DMEM) (Carlsbad, CA) containing 10% fetal calf serum (FCS), 100 U/ml penicillin, 100 µg/ml streptomycin, and 2 mM L-glutamine (complete DMEM medium). The MDR cell line, KB-V1 (kindly provided by Dr. Michael M. Gottesman from National Cancer Institute, NIH, Bethesda, Maryland), was maintained in 1 µg/ml of vinblastine (made up from 10 mg/ml stock in DMSO) in complete DMEM.10 While the doubling time of the parental line was approximately 14–18 hrs, the resistant cell line doubled every 25–30 hrs.

Assessment of cellular uptake of dual fluorescent labeled siRNA-PEI-MSNP by confocal microscopy

200 µg FITC-labeled PEI coated MSNP were incubated with 2.5 µg Texas red-labeled Pgp siRNA for 30 minutes to yield a N/P ratio of 80. Cellular uptake of MSNP was performed by adding 20 µg/ml of the dual-labeled particles to 8-well chamber slides. Each well contained 5 × 104 cells in 0.4 ml culture medium. Cell membranes were co-stained with 5 µg/ml Alexa Fluor 633-conjugated wheat germ agglutinin (WGA) in PBS for 30 min. Slides were mounted with Hoechst 33342 and visualized under a confocal microscope (Leica Confocal 1P/FCS) in the UCLA/CNSI Advanced Light Microscopy/Spectroscopy Shared Facility. High magnification images were obtained with the 100× objective. The signal intensity of the red channel, reflecting siRNA abundance, was calculated by Image J software (version 1.37c, NIH).

siRNA-PEI-MSNP exposure and assessment of PGP expression

Pgp siRNA-PEI-MSNP complexes were freshly prepared as described above. The siRNA duplex consists of 5’-r(CGGAAGGCCUAAUGCCGAA) dTdT (sense) and 5’-r(UUCGGCAUUAGGCCUUCCG) dTdG (antisense) strands. First, 5 µl of 250 ng/µl Pgp siRNA or scrambled siRNA was added to 100 µg PEI 1.8 kD-MSNP (N/P ratio=80), 100 µg PEI 10 kD-MSNP (N/P ratio=80), or 12.5 µg PEI 25 kD-MSNP (N/P ratio=10), in 5 µl aqueous solution. After incubation at room temperature for 30 min, the complexes were stabilized by 1 mg/mL BSA and then transferred into 1 ml complete DMEM. KB-V1 cells were plated at 2 × 105 cells per well, then exposed to the complexes for 16 hrs. Sixteen hours later, the medium was replaced with fresh DMEM containing 10% FCS and cultured for a further 56 hrs. Cells then were harvested for immunoblotting as outlined in the Supporting information S1. A commercially available cationic liposomal transfection agent (Lipofectamine 2000) was used as a positive control. Protein abundance was quantified by densitometric scanning using a laser Personal Densitometer SI and Image Quant software (Amersham Biosciences).

Optimizing Dox loading and release via a proton-sensitive mechanism in vitro

Based on its cationic charge at pH 7.4, we wanted to determine whether Dox (pKa=8.2) would bind to the negatively charged porous interior, including under circumstances where the exterior surface was occupied by PEI. 10 mg of MSNP functionalized by OH, COOH or phosphonate attachments were mixed with 1 mg of Dox in 0.25 mL water for 12 hrs. As control, we used positively charged amine-MSNP. Subsequently, the particles were collected by centrifugation and washed with water. To corroborate the electrostatic binding hypothesis, we also used the cationic fluorescent dye, Hoechst 33342 (pKa=11.9) as model cargo in the same particles. This procedure was repeated in phosphonate MSNP that were coated with 10 and 25 kD PEI polymers. In order to compare the loading yields of above particles, 1 mg Dox-loaded MSNP pellet was resuspended and sonicated in 1 mL of a heated HCl solution (pH = 5.0) for 15 minutes. After centrifugation, another 1 mL fresh HCl aqueous solution was added. All the supernatants were combined until the MSNP became colorless. At this point the pH was readjusted to 7.0 by 1 M NaOH and the fluorescence spectrum of Dox measured at excitation and emission wavelength of 485/550 nm in a microplate reader (SpectraMax M5 Microplate Reader, Molecular Device, USA). The procedure was also repeated to determine Dox loading in MSNP coated with 10 kD and 25 kD PEI polymers.

We also assessed the effects of acidification and ethanol extraction of Dox-loaded phosphonate MSNP. 1 mg Dox-loaded MSNP was suspended into 3 ml phenol red-free DMEM medium acidified to pH 5.0 or replenished with 10% (v/v) ethanol at 37 °C. The supernatants were collected at various time points and cleared by centrifugation for measurement of Dox fluorescence. Similar experiments were carried out in siRNA bound phosphonate-MSNP coated with PEI polymers. To see if the release profile will be influenced by siRNA binding, we also studied release after 12 hrs using siRNA-PEI-Dox-MSNP.

Confocal microscopy to determine MSNP lodging in an acidifying cellular compartment, including the effects of inhibiting lysosomal acidification on Dox release

KB-V1 cells grown on chamber slides were fixed, permeabilized, and labeled with our standard immunocytochemistry protocol.5 LAMP-1 staining was performed by using a 1:500 dilution of mouse-anti-human mAb (H4A3, Abcam, USA) for 16 hrs at 4 °C. This was followed by a 1:500 diluted TRITC-conjugated goat-anti-mouse secondary antibody (Santa Cruz, USA) for 1 hr at room temperature. Cell membranes and nuclei were stained with WGA 633 and Hoechst 33342, respectively. Slides were visualized under a confocal microscope (Leica Confocal 1P/FCS). Since the Dox release is a proton-sensitive process, the effect of neutralizing the lysosomal pH with NH4Cl was also investigated in KB-31 cells. These cells were initially treated with 40 µg/ml FITC-labeled MSNP that were simultaneously coated with the 10 kD PEI polymer with or without the addition of 20 mM NH4Cl. All the images of the particles and fluorescent Dox release were captured by the same confocal microscopy applying same parameter setting over a 72 hour time period.

Determination of intracellular Dox concentration

Dox uptake in KB-V1 cells was quantitatively evaluated in a microplate reader at 72 hrs. 5 × 104 cells were placed into a 96 wells plate and treated with 2 µg/ml free Dox or the equivalent amount of drug loaded into MSNP before or after PEI coating, with or without attachment of Pgp siRNA. Following the washing of the cells in cold PBS, the intracellular Dox fluorescence was detected at excitation and emission wavelength of 485/550 nm in a microplate reader (SpectraMax M5 Microplate Reader, Molecular Device, USA). Moreover, we also captured confocal images at the end of experiment. Image J software (version 1.37c, NIH) was used to analyze the nuclear fluorescence.

Assessment of cytotoxicity and apoptosis in KB-V1 cells treated with Dox-loaded particles in the absence and presence of Pgp siRNA

To measure cytotoxicity of the different Dox formulations, KB-V1 cells were treated with free Dox, Dox-MSNP, PEI-Dox-MSNP and siRNA-PEI-Dox-MSNP, respectively. For the latter two particle types, incubation time was for 16 hrs before replenishment of the old medium with fresh complete DMEM and performance of a MTS assay at 72 hrs. Based on the absorption readout at 490 nm, the IC50 of free and Dox-loaded MSNP were calculated. We also assessed the induction of apoptosis at 72 hours through the use of Annexin V-SYTOX Blue. Briefly, 5×105 cells were harvested and stained by FITC-Annexin V- SYTOX Blue working solution (Annexin V, Trevigen; SYTOX Blue, Invitrogen) at room temperature for 15 min. The cells were washed in binding buffer before performance of flow cytometry (Becton Dickinson, Mountain View, CA). Date analysis was performed by BD CellQuest. To confirm the flow data in which there may be a minor overlap of Dox with FITC-Annexin V, a TUNEL detection kit was used according to the manufacturer’s instructions to confirm the induction of apoptosis. Briefly, 72 hrs following treatment with free Dox or Dox loaded particles, cells were washed, fixed, and permeabilized before TUNEL staining. The number of TUNEL-positive cells was assessed under a fluorescent microscope (200×). At least 3 fields were counted by the same investigator to calculate the percentage of TUNEL positive cells.

Statistical analysis

Data represent the mean ± SD for duplicate or triplicate measurements in each experiment, which was repeated at least 3 times. Differences between the mean values were analyzed by two-sided Student’s t test or one way ANVOA and results were considered statistically significant at p < 0.05.

Supplementary Material


This study was funded by the US Public Health Service Grants, RO1 CA133697 and RO1 ES016746. Support for the safety assessment of MSNP was also provided the National Science Foundation and the Environmental Protection Agency under Cooperative Agreement Number EF 0830117. Any opinions, findings, conclusions or recommendations expressed herein are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the Environmental Protection Agency.


1. Lu J, Choi E, Tamanoi F, Zink JI. Light-Activated Nanoimpeller-Controlled Drug Release in Cancer Cells. Small. 2008;4:421–426. [PMC free article] [PubMed]
2. Coti KK, Belowich ME, Liong M, Ambrogio MW, Lau YA, Khatib HA, Zink JI, Khashab NM, Stoddart JF. Mechanised Nanoparticles For Drug Delivery. Nanosacle. 2009;1:16–39. [PubMed]
3. Slowing II, Vivero-Escoto JL, Wu C-W, Lin VSY. Mesoporous Silica Nanoparticles as Controlled Release Drug Delivery and Gene Transfection Carriers. Adv. Drug Deliv. Rev. 2008;60:1278–1288. [PubMed]
4. Liong M, Lu J, Kovochich M, Xia T, Ruehm SG, Nel AE, Tamanoi F, Zink JI. Multifunctional Inorganic Nanoparticles for Imaging, Targeting, and Drug Delivery. ACS Nano. 2008;2:889–896. [PMC free article] [PubMed]
5. Xia T, Kovochich M, Liong M, Meng H, Kabehie S, George S, Zink JI, Nel AE. Polyethyleneimine Coating Enhances the Cellular Uptake of Mesoporous Silica Nanoparticles and Allows Safe Delivery of siRNA and DNA Constructs. ACS Nano. 2009;3:3273–3286. [PMC free article] [PubMed]
6. Astrid D, Achim N. Effects of Particle Size and Molecular Weight of Polyethylenimine on Properties of Nanoparticulate Silicon Dispersions. J. Am. Ceram. Soc. 2001;84:806–812.
7. Gottesman MM. Mechanisms of Cancer Drug Resistance. Annu. Rev. Med. 2002;53:615–627. [PubMed]
8. Szakacs G, Paterson JK, Ludwig JA, Booth-Genthe C, Gottesman MM. Targeting Multidrug Resistance in Cancer. Nat. Rev. Drug Discov. 2006;5:219–234. [PubMed]
9. Jabr-Milane LS, van Vlerken LE, Yadav S, Amiji MM. Multi-Functional Nanocarriers to Overcome Tumor Drug Resistance. Cancer Treat. Rev. 2008;34:592–602. [PMC free article] [PubMed]
10. Ludwig JA, Szakacs G, Martin SE, Chu BF, Cardarelli C, Sauna ZE, Caplen NJ, Fales HM, Ambudkar SV, Weinstein JN, Gottesman MM. Selective Toxicity of NSC73306 in MDR1-Positive Cells as a New Strategy to Circumvent Multidrug Resistance in Cancer. Cancer Res. 2006;66:4808–4815. [PMC free article] [PubMed]
11. Jie L, Monty L, Jeffrey IZ, Fuyuhiko T. Mesoporous Silica Nanoparticles as a Delivery System for Hydrophobic Anticancer Drugs. Small. 2007;3:1341–1346. [PubMed]
12. Ferrari M. Cancer Nanotechnology: Opportunities and Challenges. Nat. Rev. Cancer. 2005;5:161–171. [PubMed]
13. Saad M, Garbuzenko OB, Minko T. Co-Delivery of siRNA and an Anticancer Drug for Treatment of Multidrug-Resistant Cancer. Nanomedicine. 2008;3:761–776. [PMC free article] [PubMed]
14. Wu J, Lu Y, Lee A, Pan XP, Yang X, Zhao X, Lee RJ. Reversal of Multidrug Resistance by Transferrin-Conjugated Liposomes Co-encapsulating Doxorubicin and Verapamil. J Pharm. Pharmaceut. Sci. 2007;10:350–357. [PubMed]
15. van Vlerken LE, Duan Z, Seiden MV, Amiji MM. Modulation of Intracellular Ceramide Using Polymeric Nanoparticles to Overcome Multidrug Resistance in Cancer. Cancer Res. 2007;67:4843–4850. [PubMed]
16. Chen AM, Zhang M, Wei D, Stueber D, Taratula O, Minko T, He H. Co-delivery of Doxorubicin and Bcl-2 siRNA by Mesoporous Silica Nanoparticles Enhances the Efficacy of Chemotherapy in Multidrug-Resistant Cancer Cells. Small. 2009;5:2673–2677. [PMC free article] [PubMed]
17. Patil YB, Swaminathan SK, Sadhukha T, Ma L, Panyam J. The Use of Nanoparticle-Mediated Targeted Gene Silencing and Drug Delivery to Overcome Tumor Drug Resistance. Biomaterials. 2009;31:358–365. [PMC free article] [PubMed]
18. Bagwe RP, Hilliard LR, Tan W. Surface Modification of Silica Nanoparticles to Reduce Aggregation and Nonspecific Binding. Langmuir. 2006;22:4357–4362. [PMC free article] [PubMed]
19. Santra S, Yang H, Dutta D, Stanley JT, Holloway PH, Tan W, Moudgil BM, Mericle RA. TAT Conjugated, FITC Doped Silica Nanoparticles for Bioimaging Applications. Chem. Commun. 2004:2810–2811. [PubMed]
20. Meng H, Xia T, George S, Nel AE. A Predictive Toxicological Paradigm for the Safety Assessment of Nanomaterials. ACS Nano. 2009;3:1620–1627. [PubMed]
21. Lin Y-S, Haynes CL. Synthesis and Characterization of Biocompatible and Size-Tunable Multifunctional Porous Silica Nanoparticles. Chem. Mater. 2009;21:3979–3986.
22. Wang J, Huang M-h, Zeng Z-x, Fang H-x, Yang M-s. Gene expression profiling in multidrug resistant KB cells using cDNA microarrays. Chin. J. Cancer Res. 2002;14:5–10.
23. Lim MH, Blanford CF, Stein A. Synthesis and Characterization of a Reactive Vinyl-Functionalized MCM-41: Probing the Internal Pore Structure by a Bromination Reaction. J. Am. Chem. Soc. 1997;119:4090–4091.