Photocontrolled miR-148b nanoparticles cause apoptosis, inflammation and regression of Ras induced epidermal squamous cell carcinomas in mice

Abstract

Despite evidence that microRNAs (miRNAs) are essential in modulating tumorigenesis, a major challenge in cancer treatment is to achieve tumor-specific selectivity and efficient yet safe delivery of miRNAs in vivo. In this study, we have developed a light-inducible silver nanoparticle nucleic acid delivery system that demonstrates precise spatiotemporal control, high cellular uptake, low cytotoxicity, escape from endosomes and release of functional miRNA into the cytosol. Using this approach, we delivered exogenous miR-148b to induce apoptosis in Ras-expressing keratinocytes and murine squamous cell carcinoma cells while avoiding cytotoxicity in untransformed keratinocytes. When administered to transgenic mice with HRas^G12V-driven skin tumors, a single dose of silver nanoparticle conjugates followed by 415 nm LED irradiation at the tumor site caused a rapid and sustained reduction in tumor volume by 92.8%, recruited T cells to the tumor site, and acted as a potent immunomodulator by polarizing the cytokine balance toward Th1 both locally and systemically. In summary, our results demonstrate that spatiotemporal controlled miR-148b mimic delivery can promote tumor regression efficiently and safely.

Introduction

Common treatments for cancer, such as chemotherapy and radiation, cause severe side effects,[1, 2] and thus the development of novel methods to selectively kill cancer cells while reducing side effects is a critical need. MicroRNAs are single-stranded short 20 to 23 nucleotide RNAs that negatively regulate the translation and stability of mRNAs.[3-5] Recent advances in cancer biology have shown the importance of miRNAs in tumor growth, invasion, angiogenesis, and immune invasion.[6-9] For example, dysregulation of let-7,[10, 11] miR-34,[12, 13] miR-148b,[14, 15] and miR-200c[11, 16, 17] promotes oncogenesis and these miRNAs are frequently downregulated in many cancers. In particular, downregulation of miR-148b is observed in non-small cell lung cancer,[18, 19] hepatic cancer,[20] gastric cancer,[21] colorectal cancer,[22, 23] pancreatic cancer,[24, 25] and skin cancer[14, 15]. Therefore, restoring these tumor-suppressing miRNAs has the potential to improve current cancer treatments. [26-29]

Multiple strategies have been developed to modulate miRNA function by delivering either exogenous miRNA mimics or miRNA inhibitors via viruses and liposomes.[26, 27, 30-32] While cationic or neutral lipid/polymer-based nanoparticles can suppress tumor growth in a variety of xenograft cancer models, none of them achieve localized miRNA delivery with precise spatiotemporal control.[28, 33] This is a significant concern as off-target gene silencing can result in severe side effects.[34] Additionally, most preclinical cancer studies of nanoparticles for miRNA mimic delivery in vivo have been conducted using xenotransplantation models in immunodeficient mice, despite the growing recognition that these models may not adequately mimic the complexity of human tumors, leading in part to the failure of many new drugs and nanoparticle based therapeutics in clinical trials.[35-38]

To overcome these issues, we have developed a light-inducible nanoparticle system for delivering functional miRNA mimics to kill cancer cells selectively. We have tested this system in vitro and in an autochthonous tumor model in which the human mutant H-Ras^G12V oncogene is expressed in the mouse epidermis from a doxycycline-inducible cassette.[39] Silver nanoparticles (SNPs) modified with furan-maleimide Diels-Alder (DA) linkages to miRNA mimics were used as miR-148b mimic delivery vehicles (SNP-DA-miR148b). In this system mimic release was triggered by DA cleavage via the retro-Diels-Alder reaction catalyzed by excitation at the silver nanoparticle localized surface plasmon resonance (LSPR) with light at a wavelength of 415 nm. We demonstrate that in vitro our silver nanoparticles successfully deliver miR-148b mimics, alter target gene expression and selectively induce apoptosis in Ras-expressing keratinocytes and murine cutaneous squamous cell carcinoma (SCC) cells. Further, we demonstrate that the delivery of exogenous miR-148b mimic can be spatiotemporally controlled in the tumor microenvironment through exposure to 415 nm LED light and cause apoptosis and tumor regression. In addition, the treatment provokes an inflammatory response at the tumor site, recruits T cells to the tumor site, and shifts the Th1/Th2 balance toward Th1-dominant cellular immunity in both tumors and spleen.

Materials and methods

Materials and Reagents.

Custom-modified fluorescein amidites (FAM)-tagged oligonucleotides (miR148b: 5’ 6-FAM 2’OMe UCA GUG CAU CAC AGA ACU UUG U C6-NH2 3’) were acquired through Trilink Biotechnologies, LLC. (San Diego, CA). 6-Maleimidohexanoic acid (90%), 2-furanmethanethiol (98%), methanol (>99.9%), isopropanol (>99.9%), dichloromethane (99.8%), N-hydroxysuccinimide (NHS) (98%), silver nitrate (99%), formaldehyde (36.5%-38%), sodium hydroxide (98%, pellets), hydroxypropyl cellulose (HPC, Mn = 80,000, 99%), antifoam A (100%), tris(2-carboxyethyl)phosphine hydrochloride solution (TCEP) (0.5M), were all purchased and used without alteration from Sigma Aldrich (St. Louis, MO). EDC (1-ethyl-3-(3-dimethylaminopropyl) carbodiimide hydrochloride), was purchased from ThermoFisher Scientific (Waltham, MA). Mounted 415nm LED lights from ThorLabs, Inc. (Newton, NJ) were used for photo-release experiments.

SNP synthesis and modification.

HPC-SNPs were synthesized as described by Qureshi et al.[40] Briefly, silver was reduced from 125mM AgNO₃ by adding 61.5 mM formaldehyde (HCOOH) to a pre-dissolved solution containing 0.5 g NaOH, 0.31 g HPC, 330 mL deionized water, and 5 μL antifoaming agent A at room temperature. Both AgNO₃ and HCOOH were added simultaneously at 0.5mL/min. The SNPs were purified by overnight dialysis and lyophilization. To conjugate miR148b to the surfaces of SNPs, the particles were first treated with the thermally-labile Diels-Alder linker.[41, 42] Briefly, the linker was prepared at room temperature over seven days, with mixing of 6-maleimidohexanoic acid (2.11g) & 2-furanmethanethiol (0.5g) in 20mL methanol and DCM (1:1). The linkers were then added to 1mL aliquots of SNPs (210ppm) for 24h at room temperature to allow for surface attachment through thiol linkage. Subsequently, the SNPs were centrifuged (10,000 rpm, 10min) and resuspended in isopropanol thrice, followed by mixing with 100mM (100μL) EDC/NHS and the respective 5’amine miRNA mimics (4μM, 30μL) to allow for covalent coupling. The amount of miRNA conjugated was determined by centrifugation and resuspension in DEPC water and displacing the ligands with TCEP treatment (10μL). Measurement of fluorescence at 525nm for FAM in the supernatant sample indicated attachment (Spectramax M5 Microplate/Cuvette Reader, Molecular Devices, PA, USA). The nanoparticle concentrations were determined by Inductively Coupled Plasma Atomic Emission Spectroscopy (ThermoFisher ICAP 7400 ICP-AES, Waltham, MA, USA), and at each surface conjugation step, the nanoparticles were measured for hydrodynamic size and zeta potential with a Malvern Zetasizer Nano ZS instrument (United Kingdom). UV-vis absorbance spectra were obtained using NanoDrop One (ThermoFisher, Waltham, MA, USA). TEM images were visualized using FEI Tecnai G2 Spirit BioTwin (FEI, Hillsboro, OR, USA).

Photothermal Release of 5’ FAM-miRNA 148b.

To measure the release rate of the miRNA molecules from SNP surfaces upon irradiation with 415nm wavelength light, the nanoparticles were set up in solution as described previously.[41] The SNPs-miR148b suspended in DI water, were illuminated at 415nm at increasing incident light output energies and collected for fluorescence measurements. The supernatants of the irradiated samples were separated by centrifugation and measured at 525nm. All numbers were normalized to the TCEP-treated sample, representative of the total miRNA loading on the SNPs.

Confocal microscopy for the cellular uptake and release.

Zeiss LSM 880 confocal microscope with FLIM (Zeiss, Oberkochen, Germany) was used to take images of the intracellular delivery of silver nanoparticles. PAM212 cells were seeded using Opti-MEM medium in 35 mm glass-bottom microwell dishes at a density of 500,000 cells per dish. After 16 hours, about 5 ppm of FAM-tagged SNP-DA-miR148b was added to the dish. The cells were allowed to incubate with SNP overnight before the light activation. 3 hours after the photoactivation, the samples were imaged at FAM (488 nm/518) and Alexa Fluor 405 (405 nm/420-480 nm) channels.

Cell Culture.

Pam212 cells (gift of Stuart Yuspa NCI) were cultured in calcium-free EMEM media (Lonza) with 10% Chelex treated FBS and 1% penicillin-streptomycin. Primary mouse keratinocytes were isolated from 1 to 3-day old newborn mice via flotation 0.25% trypsin at 4 °C overnight and maintained in EMEM (Lonza) with 0.05 mM Ca2+, 1% antibiotics and 8% Chelex-treated FBS, as described.[43] Primary keratinocytes were transduced with the v-Ha-Ras retrovirus on day 3 of the culture.

Cell transfection.

Cells were transfected with miR-148b mimics (50 nM) using lipofectamine RNAiMAx (Invitrogen), according to the manufacturer’s protocol.

Cell Viability.

Cell viability was tested using both Quant-iT PicoGreen dsDNA assay kit (Invitrogen), and Live/Dead viability kit (Invitrogen), according to the manufacturer’s protocol. Briefly, cells were transfected with SNP-DA-miR148b/NC for 24 hours, followed by 415 nm LED light for 15 mins. Live/Dead images were taken 3 days after the light activation. Cell numbers were measured using the PicoGreen kit on day 3 after the treatment.

Gene expression analysis.

Total RNA was isolated from cells using Trizol (Invitrogen) and the expression levels were measured by RT-qPCR (Thermo Fisher Scientific) using Verso cDNA Synthesis kit (Thermo scientific) and PowerUp SYBR Green Master Mix (Thermo scientific). The sequences of primers were selected from PrimerBank and analyzed using Primer-BLAST to make sure the specificity. The gene expression was normalized with 18S using the −ΔΔCt method, and the primers used in this study were listed in the supplementary data (Table S3).

Western Blot Analysis.

Cells were washed in ice-cold PBS and lysed in RIPA buffer with protease/phosphatase inhibitors. A total of 30 μg protein was separated using 12% SDS/PAGE gels and transferred to nitrocellulose using the Trans-Blot TURBO transfer system (Bio-Rad), according to the manufacturer’s instruction. The primary antibodies were purchased from Abcam (Pro-caspase-3) and Cell Signaling Technology (Cleaved-caspase-3). Actin (Millipore) was used as loading control.

Animal Studies.

Mice on a C57/BL6 background expressing either the tetracycline-regulated transactivator (tTA) driven by an Involucrin promotor or reverse tetracycline-regulated transactivator (rTA) driven by a Keratin-14 promotor were crossed with tetO-HRas^G12V mice on an FVB/n background to produce double-transgenic InvtTA/tetO-HRas^G12V (InvRas) and K14rTA/tetO-HRas^G12V (K14Ras) mice.[39] Tumors were induced in 7-week-old InvtTA/tetO-HRas^G12V mice by reducing the suppressive does of 10 μg/mL doxycycline hyclate (Sigma-Aldrich) in drinking water to 250ng/mL. Tumors were induced in K14rTA/tetO-HRas^G12V mice by dosing drinking water with 5μg/mL doxycycline. Tumor-bearing mice were anesthetized under 5% isoflurane and administered PBS, silver-nanoparticle control SNP-DA-NC (3.2 mg/Kg), or SNP-DA-miR148b (3.2 mg/Kg) intravenously via retro-orbital injection. Tumor irradiation occurred 6-hours post-injection for a duration of 23 minutes using a mounted 415nm LED (THORLABS) and controller. Mice remained anesthetized under 3% isoflurane throughout the procedure on a heating pad set to 37° C to maintain body temperature. Mice were euthanized at study end using CO₂ narcosis. All studies were performed in compliance with U.S. Department of Health and Human Services Guide for the Care and Use of Laboratory Animals and after approval by The Pennsylvania State University Institutional Animal Care and Use Committee.

Biodistribution analysis.

Tissue samples were dissolved in concentrated nitric acid overnight and evaporated using nitrogen gas. The residues were resuspended in 5 mL of 2-5% nitric acid and filtered for ICP-AES analysis.

Histological Analysis.

Tissue samples were fixed in 10% neutral buffered formalin overnight followed by 70% ethanol. Samples were then embedded in paraffin blocks and cut into 8-μm sections. Tissue sections were stained with hematoxylin and eosin (H&E) using an ST5010 Autostainer XL (Lecia Biosystems). For immunohistochemistry, tissue samples were processed as previously described.[39] Tissue sections were incubated with anti-CD45 (1:200; BD Biosciences), anti-CD4 (1:100, Invitrogen), or anti-CD8 (1:100, Invitrogen) overnight. The following day, sections were incubated with biotinylated-IgG (1:500, Vector Labs) and streptavidin-HRP. Anti-CD3 (1:100; Santa Cruz Biotechnology) detection occurred using ImmPRESS HRP Reagent Kit (Vector Labs). All antibody stains were developed with ImmPACT DAB (Vector Labs). Samples were counterstained with hematoxylin.

Flow Cytometry Analysis of Immune Response.

Tumor and spleen samples were harvested for immunological analysis. Tumor tissue digestion was performed using complete RPMI 1640 medium containing collagenase type I (0.825 mg/mL) and II (2.5 mg/mL) (Worthington-Biochem) and hyaluronidase (0.25 mg/mL) (Sigma Aldrich). Leukocytes were isolated from digested tumor and splenic tissue using a 40:80% Percoll (GE Healthcare) gradient. Leukocytes were stained with LIVE/DEAD Fixable Yellow Dead Cell Stain Kit (Life Technologies), and then incubated with antibodies against immune cell markers CD45 (APC eFluor 780, Biolegend), CD3 (APC, Biolegend), and CD4 (PECy5, Biolegend) in 1% BSA/PBS followed by overnight fixation in 4% paraformaldehyde. For intracellular cytokine staining, cells were first incubated in complete RPMI 1640 containing Brefeldin A (5 μg/mL) (Biolegend), ionomycin (1 μg/mL), and PMA (20 ng/mL) for 4 hours at 37° C. Intracellular cytokine staining was done following membrane permeabilization with incubation with 0.2% Saponin/1% BSA/PBS using antibodies against interferon-γ (IFN-γ) (PECy7, Biolegend) and IL-17 (FITC, Biolegend). Cells were analyzed using an LSRFortessa Cytometer (BD Biosciences) and FlowJo version 7. CD4 T cells were gated on Live CD45⁺, CD3⁺, CD4⁺, followed by IFN-γ (Th1) or IL-17 (Th17).

Statistics.

Data are presented as mean ± SD of at least three independent experiments or a single experiment with triplicate. Data were analyzed by either one-way or two-way ANOVA tests using GraphPad Prism 6 with Bonferroni correction. Values were considered significantly different at P < 0.05.

Results and discussion

Synthesis and characterization of SNP-DA-miR148b.

To achieve efficient spatiotemporal control, silver nanoparticles modified with furan-maleimide DA covalent linkers were selected as the vehicle to deliver miR-148b mimics. In our previous study, it was demonstrated that multiple DA compounds composed of alternate diene chemistries could produce a range of linkers with varying and specific cleavage energies.[41] Linkers composed of pyrrole-based dienes provided much less stable bonding than the furan-containing dienes, resulting in rapid release at low irradiation energies. Meanwhile, thiophene-based DA chemistries resulted in much more stable linkage, with only partial release achieved at high energies. Based on the published results,[41] the furan-maleimide linker was selected as a covalent DA linker chemistry as it has cleavage energetics that result in a bond that is stable under physiological conditions caging the miRNA mimic to the nanoparticle in a sterically inactive form, with minimal non-specific release, while also providing efficient and predictable de-caging upon irradiation. The linker was conjugated onto silver nanoparticles using well characterized thiol linkage,[41] and amine-terminated miR-148b mimic was linked to the carboxy-terminal DA molecule using EDC coupling chemistry. Those reactions, including the synthesis of the linker, were performed at room temperature resulting in high yields and low toxicity by-products.[44] When irradiated with a mounted 415 nm LED, the miR148b mimic is released via a retro-Diels-Alder reaction as a result of localized plasmon generation produced by the decay of the excited LSP, as shown in Figure 2A.

Figure 2. Characterization of SNP-DA-miR148b and miRNA release.

(A) Schematic of the release process of caged miR-148b mimics. (B) Hydrodynamic diameters of SNP, SNP-DA, and SNP-DA-miR148b. (C) UV-vis absorbance spectra of SNP and SNP-DA-miR148b. (D) TEM images of SNP-DA-miR148b. Scale bar, 0.1 μm (E) Zeta potential of SNP, SNP-DA, SNP-DA-miR148b. (F) miR-148b release profile in the absence or presence of 415 nm LED irradiation.

TEM analysis showed that SNP-DA-miR148b was spherical with an average diameter size of 58.03±15.55 nm, which is optimal for cellular uptake,[45] with narrow dispersity as determined by dynamic light scattering (DLS). Both zeta potential and DLS have demonstrated the conjugation of the linkers and miR-148b mimics with predictable increases in the hydrodynamic radius as shown in Figure 2B and an increasingly negative zeta potential, shown in Figure 2E, as a result of sequential DA and miRNA mimic addition to the silver nanoparticle. The zeta potential of SNP-DA-miR148b was −29.77±0.15 mV and demonstrated colloidal stability. The UV-Vis spectra also confirmed that the plasmon absorption maxima is at ~415 nm. The light activation was demonstrated using FAM-tagged miR-148b mimic and a step-like release function was observed at 15 min after the exposure of 415 nm LED at 0.2 W/cm² for a total of 187.2 J as shown in Figure 2F. Minimal nucleic acid release was detected without light activation, indicating little leakage and the stability of our light-responsive nanoparticle miRNA mimic delivery system.

Analysis of SNP-DA-miR148b cellular uptake and miRNA release by confocal microscopy.

To analyze the intracellular delivery of SNP-DA-miR148b, fluorescent confocal images of the murine squamous cell carcinoma line PAM212[46] were used to validate the cellular uptake and miRNA release using FAM-tagged miR-148b mimic. The blue channel was used to observe the back scattered light from the silver nanoparticle, and the green channel was used to observe the FAM-tagged exogenous miR-148b mimic release. When the FAM fluorophore on the miRNA mimic is statically conjugated on the silver nanoparticle surface before the light activation, a significant fluorescence quenching is expected based on Förster resonance energy transfer (FRET) as the emission spectrum of FAM overlaps with the plasmon absorption of silver nanoparticles resulting in coupling.[47-49] FRET is a distance-dependent energy transfer process from an excited donor, in this case the FAM fluorophore to an acceptor, in our system the silver nanoparticle, leading to fluorescence quenching for the donor. Post-photoactivation, we expect to have a weakening of the quenching effect resulting in increased FAM fluorescence due to the increased distance between FAM fluorophore and silver nanoparticles after the miRNA mimic release. This is confirmed in our experiments, where prior to photoactivation little FAM fluorescence is observed in Figure 3A and a significant increase in FAM fluorescence, particularly in the perinuclear region is seen after photoactivation in Figure 3B, which indicates cleavage DA bond, release of the miR-148b mimic and endosomal escape. Very limited FAM fluorescence was observed for both negative control and lipofectamine transfection, as shown in Figure S4. In addition, the majority of the fluorescence signal was identified in perinuclear regions of the cytoplasm for SNP-DA-miR148b treated cells, without using transfection reagents or electroporation. This result is in accordance with previous photoactivated nanoparticle antisense delivery experiments which demonstrated strong perinuclear localization of oligonucleotides after intracellular release.[50]

Figure 3. Analysis of intracellular uptake and miRNA release of FAM-tagged SNP-DA-miR148b conjugates.

(A) Fluorescence microscopic images of PAM212 cells after incubation with FAM-tagged SNP-DA-miR148b conjugates. (B) The release of miR148b after using 415 nm LED irradiation. (C) Regulation of target gene expression for PAM212 cells with lipofectamine (L), SNP-DA-Negative control (SNP-NC) and SNP-DA-miR148b (SNP-miR148b) using qRT-PCR assays on direct targets of miRNA-148b mimics. Scale bar, 50 μm. Data are shown as mean ± SD (n=3); **p < 0.01, ***p < 0.001.

Target gene regulation by SNP-DA-miR148b.

We have shown cellular uptake of silver nanoparticles by cells via endocytosis. However, successful delivery of functional miRNA mimics can be obstructed by multiple factors. For example, after light activation miRNA mimics must escape from endosomes in order to interact with RNA-induced silencing complex (RISC) and mRNA in the cytosol. Additionally, the decrease of intravehicular pH during endosome maturation and potential lysosome fusion may degrade miRNA mimics. Thus, we validated the release of miRNA mimics from endosomes and functionality of the RNA-induced silencing complex by analyzing target gene expression. ALCAM,[14, 15] ITGA5,[14, 51] ROCK1,[51, 52] NRP1,[53] and WNT1[20] were identified as direct target of miR148b based on previous literature and TargetScan[54]. As shown in Figure 3C, we confirmed that all the target genes were downregulated in the response of both lipofectamine and SNP-DA-miR148b treatment. It is widely accepted that epithelial cells are difficult to transfect,[55] but our silver nanoparticles have enhanced cellular uptake compared with lipofectamine (Figure S4D). We believe this increased transfection efficiency accounts for the greater downregulation of target gene expression by the nanoparticle coupled miR-148b mimic compared with lipofectamine. Collectively, these results show that our delivery system can efficiently deliver functional miRNA mimics to cells and modulate target gene expressions.

Cytotoxicity of SNP-DA-miR148b in vitro.

To evaluate the ability of miR-148b mimic to kill oncogenic cells, we compared its effects on untransformed and v-Ha-Ras (HRas) retrovirus[56] transduced primary mouse epidermal keratinocytes as well as the PAM212 cell line. Cell death was measured using Live/Dead cell staining and the PicoGreen assay three days after the light activation of SNP-DA-miR148b. To rule out effects due to silver toxicity and visible light-induced damage, a scrambled RNA sequence was conjugated to silver nanoparticles using the same procedure and added as a negative control group, termed as SNP-DA-NC. For comparison, SNP-DA-miR148b without light activation and the effects of 415 nm LED exposure on cells were also evaluated, as shown in Figure S5 and Figure S7. Lipofectamine RNAiMAX was used with 50 nM of miR-148b mimic as a chemical transfection control in vitro. As shown in Figure 4A and quantitated in Figure 4B, in primary keratinocytes, regardless of treatments, there was limited reduction in live cells (green calcein-AM fluorescence) caused by elevated levels of miR-148b. In contrast, significant cell death as measured by loss of calcein-AM staining and positive staining for ethidium homodimer-1 red fluorescence was observed in both lipofectamine and SNP-DA-miR148b treated Ras-expressing keratinocytes and PAM212 cells. The viability of both PAM212 cells and Ras-expressing keratinocytes was not affected by light activated SNP-DA-NC, demonstrating the excellent cytocompatibility of our delivery system. Consistent with these results, there was an increase in cleaved-caspase-3 in protein extracts isolated from SNP-DA-miR148b treated PAM212 cells (Figure 4C) compared to controls. Together, these results show that delivery of miR-148b mimics via light-activated silver nanoparticles specifically caused cell death in a squamous cell cancer line and primary keratinocytes whose sole genetic change was the expression of a v-HRas oncogene while having little impact on the viability of non-cancerous primary derived keratinocytes. Further, these results show that cancer cell death was specific for miR-148b mimic and not due to nonspecific microRNA effects or exposure to 415 nm light. For the molecular mechanism of miR148b induced apoptosis, all the five genes we selected in PCR have been demonstrated as tumor suppressor genes and either a single gene or a combination could be responsible for the specific cancer cell death.[15, 51, 53] It is also possible that miR-148b directly activate caspase-3 and function as a tumor suppressor, as others have shown in melanoma and hepatocellular carcinoma.[14, 20]

Figure 4. In vitro cell viability of primary keratinocytes, Ras-expressing (v-HRas) primary keratinocytes and PAM212 receiving no treatment (C), lipofectamine (L), SNP-DA-Negative control (SNP-NC) and SNP-DA-miR148b (SNP-miR148b).

(A) Fluorescence images of calcium AM (green, live cells) and ethidium homodimer (red, dead cells). Scale bar, 200 μm (B) Total number of cells as measured using PicoGreen assay. Data are shown as mean ±SD (n=3); (C) Western blotting of Pro-caspase-3 and cleaved caspase-3. (D) Semi-quantitative western blot analysis for Pro-caspase-3 and cleaved caspase-3. n.s., no significant difference; *p < 0.05, ****p < 0.0001.

Rapid tumor regression with SNP-DA-miR148b delivery in vivo.

The selectivity of SNP-DA-miR148b induced tumor cell killing was evaluated in vivo using a bitransgenic system in which doxycycline induction of skin targeted HRas^G12V causes cutaneous tumors to form within two to three weeks.[39] The biodistribution and pharmacokinetics of the silver nanoparticles were determined in tumor-bearing mice using measurements of silver content in target tissue by ICP-AES at 6 hours, 12 hours, and 24 hours after retro-orbital injection. As shown in Figure 5A, the silver nanoparticles were primarily distributed in the spleen, liver and lung, in agreement with previous studies.[57, 58] Most of the nanoparticles were localized within organs containing large numbers of tissue-resident phagocytes, such as dendritic cells and macrophages,[59-62] although we have not demonstrated nanoparticle uptake by these cells in this study. Additionally, enhanced splenic uptake in our animal model could be due to splenic enlargement that accompanies tumor development. Unlike xenografted tumors,[63-65] our autochthonous mouse model demonstrated much lower nanoparticle accumulation relative to normal tissues, with a peak at six hours post-injection, suggesting a reduced enhanced permeability and retention effect for SNP-DA-miR148b. Tumor vasculature, high interstitial pressure, solid stress from tumor growth, and abnormal stromal matrix may all contribute to the low accumulation in tumors.[66, 67]

Figure 5. Rapid and sustained tumor regression in vivo with intravenous delivery of SNP-DA-miR148b.

Therapeutic effects were evaluated after three different treatments followed by 415 nm LED irradiation six hours post-injection, including PBS control (PBS), SNP-DA-Negative control (SNP-NC) and SNP-DA-miR148b (SNP-miR148b). (A) The biodistribution of Ag in main tissues and tumor at 6, 12 and 24 h after intravenous administration of SNP-DA-miR148b. (B) Relative tumor volume post-treatment in different groups. (C) Changes in body weight of mice after treatments. (D) Representative images of the same mouse before and (E) after the treatment. (F) H&E Stained images for tumors (Scale bar, 100 μm) at (G) increased magnification (Scale bar, 50 μm) with inflammatory infiltration (arrows). Data are shown as mean ±SD (n=5).

SNP-DA-miR148b and control nanoparticles were administered intravenously via retro-orbital injection to tumor bearing mice. The mice were randomly separated into three groups: PBS control, SNP-DA-NC, and SNP-DA-miR148b. Based on pharmacokinetic studies, a time point of six hours post-injection was chosen for 415 nm light exposure. Using previously described optical properties of murine skin,[68] the exposure time duration of 415 nm LED was calculated according to in vitro release analysis. Therefore, tumors were irradiated at 6-hours post-injection for 23 minutes using a mounted 415nm LED for all treatments. As shown in Figures 5B, D, E, SNP-DA-miR148b and light treatment caused significant tumor reduction after two days, and no tumor regrowth was observed over seven days. In contrast, the tumor volume increased more than 5-fold for PBS and SNP-DA-NC treated mice. Importantly, seven days after treatment, there was no significant change in mean body weights for all treatment groups, and H&E stained sections of major organs (Figure S1) demonstrated no evident histological changes in SNP-DA-NC and SNP-DA-miR148b treated groups. In addition, normal kidney and renal function tests as indicated by normal BUN, ALT and AST levels for all three treatment groups (Table S2) and the absence of animal mortality (Figure S6) indicated there is little acute toxicity from the administration of SNP-DA-miR148b or control silver nanoparticles. However, while the tumor tissue of PBS and SNP-DA-NC treated mice retained a normal squamous tumor organization and histology with a clearly defined basal layer (Figure 5F), and no effect was observed with SNP-DA-miR148b without light activation (Figure S5 B, C and D), tumors from light activated SNP-DA-miR148b treated mice were severely damaged, with disruption of the normal tumor architecture and areas of apparent cell death (Figure 5G). In addition, these tumors, but not controls had major inflammatory infiltration (Figure 5G).

SNP-DA-miR148b stimulates tumor infiltration by T cells.

To further investigate the inflammatory response of SNP-DA-miR148b, we used immunohistochemistry for all leukocytes (CD45+), T cells (CD3+), helper (CD4+) and cytotoxic (CD8+) T lymphocytes. Figure 6A and Figure S8 (increased magnification) show that in tumors from PBS and SNP-DA-NC treated mice, very few CD45 or T cells were found. After light-activation of SNP-DA-miR148b, a large increase in tumor-infiltrating leukocytes was observed, which indicates a powerful yet localized inflammatory response accompanying tumor tissue destruction. Figure 6B shows that the CD45 infiltration was localized within the tumor tissue and did not occur immediately in adjacent normal skin indicating the specificity of the inflammatory and cell death response to the tumor.

Figure 6. Modulation of immune response both locally and systemically.

(A) Representative images of immunohistochemistry (IHC) within the tumor tissue. Leukocytes are identified by CD45 (arrows). T cells are identified by CD3, CD4 and CD8(arrows). (B) H&E and IHC stained (CD45) images for the tumor tissue and adjacent normal skin after light-activation of SNP-DA-miR148b. (C) Percentage of T cell subpopulations in the tumor and spleen. Scale bar, 50 μm. Data are shown as mean ±SD (n=3); n.s., no significant difference; *p < 0.05; **p < 0.0001.

To further understand how lymphocytes respond to nanoparticle treatment, we characterized the T cell-mediated inflammatory response both locally and systemically using flow cytometry. While most of the T cell subpopulations did not fluctuate significantly one day after treatment, the percentage of T helper 1 cells (Th1) dramatically increased in both tumor and spleen, as shown in Figure 6C. For Th1 cells in the tumor, the percentage of Th1 increased only in SNP-DA-miR148b light exposed tumors, but not in either the PBS or SNP-DA-NC groups, suggesting that the destruction of tumor tissues is linked to a T cell-mediated response. In comparison, the Th1 percentages in the spleen were increased in both SNP-DA-NC and SNP-DA-miR148b, which suggests that silver nanoparticles can indirectly polarize Th1 differentiation of naïve splenic T cells possibly through scavenging by either tissue-resident splenic macrophages or dendritic cells.[69-71] However, in the spleen of SNP-DA-miR148b treated mice, there was a further significant increase in the percentage of Th1 cells suggesting that delivery of miR-148b mimics to the tumor microenvironment further enhanced systemic Th1 polarization possibly due to the apoptotic tumor cells. In addition, based on a complete blood count (CBC) analysis there was no obvious systemic inflammatory response in SNP-DA-miR148b treated mice, as shown in Table S1. For all three treatments, including PBS, SNP-DA-NC and SNP-DA-miR148b, the major types of blood cells were within the normal range.[72] Based on these normal CBC values, it is very unlikely that many of the major pro-inflammatory cytokines are elevated and when combined with normal histology in non-light exposed tissues these data indicate that significant systemic inflammation or autoimmune responses are likely not occurring.

Conclusion

In brief, we developed a light-inducible gene regulation delivery system with precise spatiotemporal control in vivo with translational potential for cancer therapy. Our findings demonstrated that the delivery of exogenous miR-148b mimics via light-inducible silver nanoparticles selectively induces apoptosis in Ras-expressing keratinocytes in vitro and inhibits tumor growth in vivo. In addition, our delivery system has a specific immunomodulatory effect on T helper cells leading to an increase in Th1 cells in both tumor and spleen of SNP-DA-miR148b treated mice that this could represent an additive immune effect of both tumor tissue destruction and the silver nanoparticle itself. This strategy provides an effective approach for tumor ablation with specificity, minimized side effects to healthy tissue, and Th1-dominant cellular immunity.

Figure 1. Schematic illustration of systemic delivery of photocontrolled silver nanoparticles for miRNA cancer therapeutics.

The animated structure of SNP-DA-miR148b with miR-148b mimic conjugates on the surface of silver nanoparticles. Once SNP-DA-miR148b internalizes within the tumor cells, furan-maleimide linkers are cleaved by 415 nm LED, and cargo miRNA mimics are released, driving cancer cells apoptosis.