A combinatorial microRNA therapeutics approach to suppressing non-small cell lung cancer

Abstract

Targeted cancer therapies, while often effective, have limited utility due to preexisting primary or acquired secondary resistance. Consequently, agents are sometimes used in combination to simultaneously affect multiple targets. MicroRNA mimics are excellent therapeutic candidates because of their ability to repress multiple oncogenic pathways at once. Here we treated the aggressive Kras;p53 non-small cell lung cancer mouse model and demonstrated efficacy with a combination of two tumor-suppressive microRNAs. Systemic nano-delivery of miR-34 and let-7 suppressed tumor growth leading to a survival advantage. This combinatorial microRNA therapeutic approach engages numerous components of tumor-cell addictive pathways and highlights the ability to deliver multiple microRNAs in a safe and effective manner to target lung tissue.

Introduction

Non-small cell lung cancer (NSCLC) is essentially untreatable, likely due to accumulation of mutations that affect the RAS and p53 pathways, and is the leading cause of cancer-related deaths worldwide. Various strategies have been proposed to increase the survival of NSCLC patients including prevention, early diagnosis, and more aggressive interventions for late-stage cancers. Indeed, five-year survival rates can approach 70% with surgical resection of stage IA disease (1). However, most individuals are diagnosed at late-stage (at initial diagnosis, 55% of patients have distant metastatic disease), when their tumors and metastatic lesions have become refractory to current treatment regiments. As such, better treatments for late-stage, aggressive disease that do not solely depend on inhibiting a single biological target are needed.

Superior, targeted treatment options depend on the individual cancer genotype and the availability of tests to identify mutations susceptible to these therapies. NSCLC gene signatures have uncovered typical protein-coding gene mutations as well as subsets of microRNAs (miRNAs) that are aberrantly expressed (2,3). Remarkably, the levels of two tumor-suppressive miRNAs, let-7 and miR-34, are often the most statistically altered miRNAs in NSCLC tumor tissue (2,4–8). The reduction in let-7 and miR-34 expression is particularly relevant to the NSCLC oncogenic phenotype, as these miRNAs target key oncogenes involved in multiple stages of the tumorigenic process and in the maintenance of oncogene addiction, such as RAS, BCL2, MET, and MYC (4,6,9–12). In addition, miR-34 is a direct transcriptional target of and produces phenotypes akin to p53 (12–16).

The recent discovery that miRNAs are modulators of key signaling pathways often disregulated in disease has resulted in their emergence as powerful therapeutic agents, actively being evaluated for the treatment of multiple diseases (see Kasinski and Slack for a review (17)). These small non-coding RNAs efficiently modulate the expression of protein coding genes either through translational repression or target mRNA destabilization (18,19). Because miRNAs bind their targets with imperfect sequence complementarity, an individual miRNA is capable of affecting the expression of multiple genes. As such, the delivery of a single miRNA is analogous to a multi-drug cocktail. Likewise, multiple miRNA binding sites are regularly found in an individual target gene, decreasing the likelihood of acquired resistance due to somatic mutations. Although the effect of an individual miRNA acting on a single target may be subtle, the collective repression of tens to hundreds of genes can have a significant impact on cells and produce strong phenotypic outcomes. This has been confirmed for tumor-suppressive miR-34 and its respective target genes BCL2, MET, CDK4, and c-MYC, as well as for let-7 and its targets KRAS, c-MYC, HMGA2, and the LIN28 isoforms (4,6,9,11,12,20). While the expression of miRNA target genes can vary in different tissues and cells, the ability of a miRNA to target multiple key oncogenes makes miRNAs an attractive therapeutic tool that is potentially more powerful than agents that target a single gene.

Both let-7 and miR-34 function as tumor suppressors in NSCLC and can inhibit tumor growth in a variety of model systems when used therapeutically as single agents. Specifically, our groups and others have shown that exogenous let-7 can both prevent and treat Kras-driven lung tumors (21–23), and that miR-34 can prevent initiation and progression of Kras^G12D+/;p53^R172H/+ lung tumors and human NSCLC tumor xenografts (24–26). Additional studies showed that miRNAs are effective therapeutically even if they do not directly repress the mutant driver gene responsible for oncogenesis. Evidence comes from Kras-activated tumors that are exposed to exogenous miR-34 (25,26); although miR-34 is not expected to directly regulate Kras, there is a robust effect on Kras-activated tumors treated with miR-34.

Kras^LSL-G12D/+;p53^flx/flx genetically engineered mice accurately model NSCLC, both in disease progression, and response and resistance to conventional therapies (27–29). Since tumor formation in this model depends on two or more signaling pathways that are associated with let-7 and miR-34, we explored whether combining miR-34 and let-7 into a single therapeutic could interfere with constitutively active processes in heterogeneous cancer cells to induce greater treatment efficacy. We show that simultaneous supplementation of these two tumor suppressor miRNAs results in an even broader repression of key oncogenes and enhanced efficacy in aggressive NSCLC compared to treatment with the individual miRNAs.

Results

miR-34 and let-7 synergize in NSCLC cells in culture

To evaluate the combined efficacy of these two master regulators, seven different lung cancer cell lines were transfected with low nanomolar concentrations of let-7 or miR-34 individually, or half of the dose of each in combination. When transfected with let-7b or miR-34a alone, proliferation of cells harboring both KRAS and TP53 open reading frame (ORF) mutations (in cell lines: H358, H23, and H441) was decreased. Similarly, the combination of half doses of let-7b and miR-34a was equally or, in some cases (H441), more effective (Fig. 1A and Supplementary Fig. 1). Cell lines with only a KRAS ORF mutation (H460 and A549) or a TP53 mutation (EKVX) were less affected by the combination. These data suggest that that the use of either miRNA alone or in combination is effective in a KRAS;TP53 mutated background.

Figure 1.

miR-34a and let-7b reduce tumor cell proliferation and invasiveness in a synergistic manner. (A–C) Cells were transfected with miRNA mimics: control, let-7b, miR-34a, or half the dose of each. (A) Sulphorhodamine B (SRB) assays were performed 7 days post transfection. Data represent 5 transfections for each treatment. Error bars depict standard deviation. (B) Change in SRB stained cells from day four to five post transfection is shown as a proxy for change in proliferation, n=3 for each treatment. Error = standard deviation (C) Dose-response curves were created over a wide range of miRNA concentrations (3 – 100nM). Effect represents decreased proliferation. ▲ let-7b and miR-34a, ■ miR-34a,● let-7b (D) Invading H23 cells were stained 24 hours after seeding. Three fields of view were photographed and the average number of invading cells per field is graphically represented. (E) Human lung cancer cell lines, A549, H441, and H23 were transfected with 25 nM of miRNA mimics: control (c), let-7b (l7), miR-34a (34), or half the dose of both let-7b and miR-34a (d) or left untransfected (u). Forty-eight hours following transfection, protein lysates were prepared, resolved, transferred to PVDF, and evaluated for relevant let-7b and mir-34a target genes. Actin serves as a loading control. ND – not detected. Densitometry measurements relative to untransfected cells are shown following normalization to actin. * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001, **** p-value < 0.0001, unless otherwise indicated, based on one-way analysis of variance (ANOVA) with Tukey’s post hoc correction.

Additional analysis in H441 cells determined that the overall rate of proliferation was decreased more strongly in cells transfected with half the dose of both miRNAs (Fig. 1B), relative to cells transfected with let-7b or miR-34a individually. To determine whether co-treatment with let-7b and miR-34a results in a synergistic effect on cell proliferation, dose-response curves and combination indices were calculated from proliferation data generated from H441 and H23 cells. This analysis confirmed that let-7b and miR-34a synergize in H441 cells (Fig. 1C) but act additively in H23 cells (data not shown). In H441 cells, combination index (CI) scores were well below 1.0 when the effect size was greater than 15%, indicating a synergistic response (CI H441: at 25% growth inhibition = 0.65, at 20% growth inhibition = 0.74, at 15% growth inhibition = 0.86). To better characterize the cellular effects of treatment with the miRNA combination in H23 cells, we evaluated cell migration through a matrigel-coated membrane following transfection. H23 cells transfected with either let-7b or miR-34a individually were capable of invading equally well or better than cells transfected with a control, scrambled miRNA (Fig. 1D). In contrast, H23 cells exposed to the let-7b/miR-34a combination showed a reduction in their ability to invade (p-values of let-7b and miR-34a transfected cells relative to: control = 0.06, let-7b = 0.07, miR-34a = 0.004).

Tumor suppressor miRNAs often mediate their effects by targeting potent genes in key oncogenic signaling pathways. Therefore, we evaluated the expression of LIN28B, MET and MYC following transfection. MET is a bona fide miR-34 target (12) while MYC is reported to be repressed by both miR-34 and let-7 (Wei:2008iz 10). let-7 also represses LIN28 (20,30,31). Specifically in the cells transfected with miR-34a, MET levels were reduced, and MYC and LIN28B were downregulated in cells transfected with let-7b. These effects were also observed in cells transfected with half the dose of let-7b and miR-34a (Fig. 1E). Together, the results indicate that combinatorial treatment with let-7b and miR-34a results in decreased levels of more key oncogenes than transfecting with either individual miRNA, providing mechanistic evidence consistent with the tumor suppressive effect of these miRNAs seen above.

Superior effects of the miR-34 and let-7 combination in NSCLC mouse models in vivo

Next we determined the anti-tumor activity of the let-7b/miR-34a combination in vivo, using the Kras^LSL-G12D/+;p53^flx/flx NSCLC model. This model is the most accurate representation of human NSCLC in disease progression, and response and resistance to conventional therapies, making it an excellent model to use for our studies (27,28). To induce tumor formation, transgenes in Kras^LSL-G12D/;p53^flx/flx mice were induced specifically in the lung through intratracheal administration of 10⁶ PFU of adenoviral particles expressing Cre recombinase (Ad-Cre). Following transgene recombination, tumors developed at an accelerated rate with defined nodules present in a subset of animals five weeks following Ad-Cre delivery. Both miR-34 and let-7 levels were confirmed to be low in tumors relative to the expression in normal lung tissue. MiR-34b and miR-34c levels were reduced to less than 50% of the level found in normal lung tissue (Supplementary Fig. 2). let-7a, let-7b and miR-34a levels were unchanged at ten weeks post transgene recombination; however, we did observe a decrease in expression four weeks later for let-7b in both the lung tissue and in the whole blood (Fig. 3, F and G; compare baseline to miR-NC). We likely did not observe a decrease in miR-34a due to the overall low baseline levels of miR-34a in the whole lung (Figure 3D). Interestingly, in the tumor tissue we also observed increased expression of the oncomiRs miR-19, miR-155 and miR-21 (Supplementary Fig. 2). On average, miR-19 was increased 5-fold, miR-155 approximately 4-fold, and miR-21 over 2-fold relative to normal lung tissue.

Figure 3.

Systemically delivered miRNAs accumulate in whole blood and lung tissue, and do not elevate serum cytokines. Serum, whole blood, and lung tissue was harvested 24 or 48 hours following the final injection. (A–C) Serum was evaluated for relevant cytokines: TNF-α (A), IL-6 (B) and IL-1β (C). Serum from animals treated with LPS was included as a positive control (n=3). (D–G) RNA was extracted from lung tissue (D and F) or whole blood (E and G) and evaluated for accumulation of miR-34a (D and E) or and let-7b (F and G) by RT-qPCR. Data are shown as copies per 1 ng of total RNA. (baseline: n=7, control: n=9, let-7b: n=9, miR-34a: n=8, let-7b and miR-34a: n=8) † Statistically significant from miR-NC treated based on one-way ANOVA of log-transformed data followed by a Dunnett’s post hoc test (multiple comparison test).

We had previously identified a neutral lipid emulsion (NLE) delivery agent suitable for systemic in vivo targeting to the lung (Max Suppressor, Bioo Scientific and Mirna Therapeutics) that showed no obvious signs of toxicity (25,26). Therefore, we tested the efficacy of NLE-encapsulated let-7b and miR-34a in the therapeutically resistant Kras^LSL-G12D/+;p53^LSL-flx/flx mouse model. We treated animals ten weeks following transgene recombination, systemically administering 1 mg/kg of NLE-encapsulated let-7b, miR-34a, or half the dose of each into the tail vein. As a control, the NLE delivery agent alone was also evaluated. Animals were dosed every other day (QOD) for a total of 11 doses, a time when the control animals and some of the miR-34a or let-7b individually dosed animals became sick and died. The remaining animals were sacrificed either 24 or 48 hours following the final injection and evaluated for lung tumor burden and tumor number (Fig. 2). A reduction in tumor nodules was only apparent when animals were administered the combination of 0.5 mg/kg of let-7b and 0.5 mg/kg of miR-34a (Fig. 2C). In animals treated with the combination, tumor burden decreased from 49% to 31% (Fig. 2D) and tumor nodules decreased from an average of 17.6 to 8.25 nodules/left lobe (Fig 2C). Because some animals were too sick to withstand the stress from the injections and died, they were not included in the final histological data set and statistical power was reduced. As such we conducted a larger and more powerful study to support use of the combinatorial therapy.

Figure 2.

Systemic delivery of miRNAs encapsulated in neutral lipid emulsion (NLE) slows tumor growth. (A) Schematic of dosing schedule. (B) Animals were sacrificed 24 or 48 hours after the final injection; lungs were perfused, harvested, and processed for hematoxylin and eosin staining (H & E). Bars = 1mm (C) Tumor nodules were counted from three sections obtained from the left lobe of each animal and are shown as average nodules/section/left lobe. (D) Overall tumor burden represents total tumor area averaged from three sections obtained from each treated animal relative to the total area of the lung. (baseline: n=3, NLE: n=3, let-7b: n=5*, miR-34a: n=4, let-7b and miR-34a: n=5) (* Note: two of the let-7b sections were not included in the nodule count since the sections obtained were predominantly one large nodule.) * p-value < 0.05, unless otherwise indicated, based on one-way analysis of variance (ANOVA) with Tukey’s post hoc correction. ns = not significant.

The second in vivo study used a promising new lipid-based delivery agent, NOV340, that accumulates in the lung, liver, and spleen of systemically treated mice (AGB, unpublished data and Fig. 3, D and F). Equally important, the path to the clinic could be quickened as this delivery agent showed limited toxicity in preclinical animal studies (32) and is currently the subject of a Phase I clinical trial for MRX34, a NOV340 nanoparticle loaded with miR-34a mimics to treat patients with unresectable primary hepatocellular carcinoma or individuals with unresectable liver metastasis (NCT01829971) (33). Due to the early demise of some animals in the previous trial we started treatment of animals at seven weeks post transgene activation instead of at the previous ten weeks. This allowed control animals to survive through the completion of the study over the course of 14 total injections QOD (Fig. 4A). Similar to the previous studies, animals were administered let-7b, miR-34a or half the dose of both miRNA mimics. Formulations were administered intravenously at 1 mg/kg per injection, a dose well below levels that induce toxicities in mice (AGB, unpublished data). Either 24 or 48 hours after the final injections, animals were sacrificed and evaluated for tumor burden, signs of toxicity, accumulation of miRNAs in whole blood and lungs, and relevant markers. As expected and in agreement with previous studies, there was no evidence of whole organ toxicity or elevation in any of the serum cytokines TNFα, IL-6, and IL-1b (Fig. 3A). Systemic delivery of the individual miRNAs led to marked accumulation of miR-34a and let-7b in the lung (Fig. 3, D and F, p-value < 0.0001 for miR-34a and < 0.0001 for let-7b) and miR-34a in the whole blood (Fig. 3E, p-value < 0.0001) compared to miR-NC treated mice. It is of interest to note that let-7b levels dropped in both the lung and whole blood as the tumors progressed (Fig. 3, F and G: baseline animals relative to miR-NC treated animals in lung, p-value = 0.001, and in whole blood, p-value = 0.034). This suggests that let-7b expression decreases as the tumors progress and, furthermore, that our systemic administration can restore the levels to at least baseline in the lung. The data for miR-34a also indicate that miRNA levels correlated with miRNA dosing: animals that received a full dose (1 mg/kg) had roughly twice as much miR-34a as animals that were injected with the combination containing half the dose (0.5 mg/kg) of each miRNA. These levels were sustained at least 48 hours after the last dose. We compared levels in animals sacrificed 24 or 48 hours after the final injection (not shown), which revealed no significant changes between the two time points.

Figure 4.

Reduction in tumor size following systemic delivery of combinatorial miRNAs encapsulated in a clinically relevant NOV340 delivery agent. (A) Schematic of dosing schedule. (B) Representative H & E stain of the left lobe of animals from each treatment group (histology from every animal is shown in Supplementary Fig. 3). Bars = 1 mm (C) Quantification of tumor area relative to total lung area. (D) The volumes of individual tumors were recorded and averaged for each treatment group. Each bar represents over 280 tumors from each treatment group. Standard error of the mean is depicted. (E) Images of KI-67 staining from representative sections for each treatment group. (F) Proliferation index calculated as KI-67 positive cells/mm². (C, D, F) * p-value < 0.05, ** p-value < 0.01, *** p-value < 0.001 based on one-way analysis of variance (ANOVA) with Tukey’s post hoc correction. ns = not significant. (G) Survival data of Kras^G12D/+; p53^flx/flx treated mice (arrows indicate injection days). Survival p-values = 0.07 for control vs. let-7b and miR-34 combination, and 0.05 for control vs. miR-34a following log-rank (Mantel-Cox) test.

Histological evaluation determined that overall tumor burden was reduced in animals that were administered the combinatorial therapy (Fig. 4, B and C, and Supplementary Fig. 3). The reduced tumor burden was attributed to a decrease in individual tumor size (Fig. 4D). The combinatorial administration reduced tumor volume to less than 43% of the control tumors (p-value = < 0.0001). Notably, average lung tumor volumes from the dual treated animals were also significantly smaller than average volumes from either of the single miRNA-treated groups (relative to miR-34a: p-value < 0.05, relative to let-7b: p-value < 0.01), suggesting that combinatorial treatment with let-7b and miR-34a has a significant effect on tumor growth. In addition, we observed a decrease of Ki-67 staining in combinatorial treated tumors, indicating that the tumors exposed to miR-34a and let-7b were proliferating at a slower rate (Fig. 4, E and F).

Individual lung tumors that remained in each of the treatment arms were harvested and used to evaluate the effects of miR-34a and let-7b on respective target genes (Supplementary Fig. 4). Messenger RNA (mRNA) levels of Myc, Lin28A and Lin28B, three of the genes that were expressed at higher levels in the tumor relative to negative control treated tissue (miR-NC), were found to be downregulated by exogenous miR-34a and/or let-7b. Bcl2 and Met, gene products that did not change during lung tumorigenesis, were also reduced by the combinatorial treatment and miR-34a and/or let-7b. The results suggest that the delivery of miRNA mimics was sufficient to repress their downstream targets. Importantly, the collective response on target genes was sustained in the group treated with the combination, showing that through combinatorial miRNA administration we were able to downregulate more target genes than through the use of each miRNA alone. The relative decreases in the transcript levels in vivo were similar to the reduced protein levels from the human cell culture experiments (Fig. 1E).

miR-34 and let-7 improve survival in Kras; p53 mutant mice

Although a few lung agents with FDA approval lead to modestly increased survival in this mouse model (27), no new clinically relevant agents have shown a similar benefit (29). Therefore, because this mouse model is the most accurate representation of human response to FDA-approved therapeutics, we sought to evaluate if the new combinatorial miRNA therapy could increase the average survival of Kras;p53 mice. Transgenes were activated in the lungs of the mice as before and tumors were allowed to develop for 6 weeks before animals were separated into 4 treatment arms. Each animal received 10 mg/kg NOV340 per intravenous miRNA injection, such that miRNAs administered alone were given at 10 mg/kg, and each miRNA in the combination was used at 5 mg/kg. Animals were dosed three times per week (TIW) for a total of 8 weeks. At this time point, ~50% of the control animals had died. Kaplan-Meier survival curves, shown in Figure 4G, demonstrate that there was a clear survival advantage (~40% increase) for the animals administered the combination of half miR-34 and half let-7b (p-value = 0.07), and for the miR-34a alone group (p-value = 0.05). Animals administered the control miRNA had a median survival of 57 days following the first injection, and miR-34a treatment alone increased the median survival to 81 days. While let-7b treatment alone had no effect on increasing survival (median of 55 days), animals administered the half-dose combinatorial therapy had a median survival of 78 days. Together, these results indicate that treatment with miR-34a or the half dose combination of let-7b and miR-34a can increase average survival in this model of aggressive NSCLC.

Discussion

The study we present here is one of the first to show that a combination of two biologically relevant, tumor-suppressive miRNAs is superior in its ability to repress oncogene expression, preventing proliferation and invasion of cancer cells in culture, to inhibit tumor proliferation in vivo, and to confer a survival advantage. Half the dose of each of these miRNAs was capable of repressing relevant biological targets in cells and in vivo, yet exhibited undetectable toxicity. We have performed these timely in vivo studies using a miRNA-delivery agent already in clinical trials, which should accelerate the translation of this combinatorial miRNA therapeutic approach into the clinic. While it still needs to be explored, it is expected that the tumors exposed to the combinatorial therapy could now be more sensitive to conventional therapies based on the coordinated downregulation of both let-7 and miR-34 target genes involved in chemoresistance, relative to the individually dosed animals. Furthermore, it is expected that this miRNA combination, which hits multiple relevant biological targets, would prevent the onset of acquired resistance that has been observed with other therapies tested in this model (29).

While the study presented here shows a similar effect on survival between animals exposed to the full dose of miR-34a or those treated with half the dose of miR-34a and let-7b, this effect may be attributed to the low levels of miR-34a in the baseline treated animals. When the miRNAs are supplemented, the change in miR-34a levels is much larger than that of let-7b, which does not change strongly (Figure 3, D–G). It is possible that in this experimental system the net effect of the large change in miR-34a levels is driving the survival effect. It is however likely that the combinatorial treatment would display increased efficacy in humans by impairing the development of secondary resistance. The accumulation of mutations in human tumors may render the tumor resistant to a single agent. By simultaneously targeting multiple key factors in the oncogenic response, the combinatorial approach presented here will likely diminish the occurrence of acquired resistance. Collectively, this is a first step in the progression towards the adoption of a combinatorial miRNA approach as a less toxic and more direct method to target multiple biologically relevant pathways that tumor cells have become addicted to.

Methods

Cells Culture and Transfection

All cell lines were obtained from ATCC and were cultured in RPMI media (Invitrogen, Carlsbad, CA, USA) supplemented with fetal bovine serum and penicillin/streptomycin, following standard tissue culture procedures. Cells were maintained at 37°C in a humidified 5% CO₂ environment and passaged every 3–4 days. Cells were transfected with miRNA mimics (Ambion, Life Technologies, Grand Island, NY) using DharmaFECT1 (Thermo Scientific, Waltham, MA, USA). For SRB experiments, 1,000 cells were seeded per well in 96-well cell culture dishes. The following day either 25 or 50 nM final concentration of miRNA mimics in various combinations was complexed with 0.5 μL of DharmaFECT 1 and added to the cells in a final volume of 100 μL of serum- and antibiotic-free media. Transfection media was replaced with complete growth media four hours later. In all cases control miRNA was supplemented in to keep the entire miRNA level consistent between transfection conditions. For protein lysate retrieval, cells were seeded in 24-well plates and transfected as above except that the DharmaFECT 1 concentration was increased to 2 μL/well.

Sulforhodamine B assays

Cells cultured in 96-well plates were fixed with 10% tricholoroacetic acid in complete media and stained for 30 minutes with 0.4% (wt/vol) SRB dissolved in 1% acetic acid. Unbound dye was removed by four washes with 1% acetic acid, and protein-bound dye was extracted with 10 mM unbuffered Tris base for determination of absorbance at 490 nm using a spectrophotometer.

Trans-well migration assays

Transfected H23 cells plated at 1 × 10⁵ in 100 μL serum-free medium were assayed for migration using BD BioCoat Invasion Chambers (BD Biosciences, San Jose, CA, USA) with 8 micron pores, according to the manufacturer’s instructions. Serum-containing media was added to the outside of the chamber to act as a chemoattractant. Briefly, cells were incubated for 24 hours to allow for migration across the pores, washed with PBS, treated with 3.7% formaldehyde in PBS, washed again, and then treated with methanol. After another wash, a 1:20 dilution of Giemsa stain was used to stain the cells, before the cells were washed a final time. Migrating cells were imaged at 10X magnification. For quantification, three fields of view for each transfection were counted and averaged. Three biological replicates were performed.

Protein extraction and immunoblotting

Forty-eight hours following transfection, cells were lysed with RIPA buffer (1 x PBS, 1% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate), supernatant was collected, 50 μg of protein was resolved on 12% TGX gels (Bio Rad, Hercules, CA) and transferred to PVDF membranes. Primary antibodies for detection included Lin28B (1:500, Cell Signaling 4196), Met (1:500, Cell Signaling 3127), and Myc-Xp (1:500, Cell Signaling, 5605). Following incubation with secondary antibodies, membranes were washed and signal was acquired using West Dura (Promega) and exposure of membranes to film. Intensities were quantified using ImageJ Software (NIH) (34).

Kras;p53 animals, induction of tumor formation, and treatment

Both Kras^LSL-G12D/+ (strain 01XJ6) and Trp53^flx/flx (strain 01XC2) mice were obtained from the NCI Frederick Mouse Repository. The Kras allele was crossed into the Trp53 background, FVB.129. Newly acquired double heterozygotes were backcrossed to the Trp53 strain for four generations before homozygosing the Trp53^flx allele. Transgenes in double mutant animals were induced based on the method of DuPage et al (28). Briefly, animals were anesthetized with a mix of ketamine and xylazine and intratracheally intubated. Adenoviral particles (10⁶ PFU) encoding for Cre-recombinase were precipitated with CaCl₂ and 65 μL of precipitated virus was delivered directly to the lungs through the catheter.

In all cases miRNA mimics were delivered systemically through the tail-vein. The neutral lipid emulsion (NLE) was prepared as previously described (26). Mimics encapsulated in NOV340 were supplied by Mirna Therapeutics. For tumor burden studies, following transgene recombination, animals were administered either 1mg/kg of NLE-encapsulated miRNA mimics every other day for a total of 11 injections or 1mg/kg of mimics encapsulated in NOV340 every other day for a total of 14 injections. Twenty-four or forty-eight hours after the final injection animals were sacrificed. Whole blood and serum were obtained and lungs were perfused with saline, harvested, fixed in 10% buffered formalin and paraffin embedded according to standard procedures. Sections were stained by hematoxylin and eosin (H&E) and evaluated using a Zeiss dissection microscope, AxioCam MRc 5 camera, and AxioVision 4.7.1 imaging software. Tumor burden represents the tumor area relative to the total lung area obtained from three independent sections for each animal. For survival studies 10 mg/kg of miRNA mimics (5mg/kg of let-7b and miR-34a for the combination) were delivered systemically through the tail vein TIW for 8 weeks. Animal survival was monitored throughout the remainder of the study. Log-rank tests were performed to determine significance. All animals were housed in the Yale University Animal Facility under the guidelines held by the Institutional Animal Care and Use Committee.

Ki67 IHC

Three micron tissue sections were adhered to PLUS slides and deparaffinized at 70°C for 30 minutes followed by standard xylene and alcohol clearing steps. Slides were immersed in 90°C Target Retrieval Solution (DAKO, Denmark) for 20 minutes, followed by a gradual cooling step by placing the hot solution at room temperature for 10 minutes. Slides were then rinsed with Wash Buffer (DAKO). The next steps were followed sequentially by a rinse with wash buffer: block slides for endogenous peroxidase activity using dual enzyme block (DAKO) for 5 minutes, Fc Receptor Block (Innovex, CA) for 1 hour, Background Buster (Innovex) for 30 minutes, Ki67 primary antibody (Abcam, Cambridge, MA, cat# ab8191) at 1:500 dilution for one hour, secondary reagent (Rabbit Polymer, DAKO) for one hour, Envision+HRP (DAKO) for 30 minutes, DAB substrate for 10 minutes, and Automated Hematoxylin (DAKO) counter stain for 6 minutes. Slides were dehydrated by rinsing slides once in 70% alcohol, twice in 95% alcohol, and three times in 100% alcohol. Slides were dipped three times in xylene and coverslipped using Richard-Allen Scientific Mounting Medium (Thermo Scientific, MA). KI-67 positive cells, in 2mm sections, were counted and averaged across all animals in the treatment arm.

RNA isolation and qRT-PCR

Total RNA from whole blood and tissues was isolated using the mirVana PARIS RNA isolation kit (Ambion, TX) following the manufacturer’s instructions. Briefly, 250 μL of blood lysate (250 μL whole blood + 250 μL RNA denaturing solution) or 600 μL of tissue homogenate (1 mL RNA denaturing solution added to slightly thawed tissue sample and homogenized) was combined with an equal volume of acid-phenol, vortexed, and centrifuged at 13.2k × g for 15 minutes. One half of the resultant aqueous phase was transferred to a fresh tube and the remaining protocol was followed according to the manufacturer’s instructions. A volume of 50 μL of heated (95°C), nuclease-free water was used to elute the RNA from the column.

For the detection and quantification of miRNA from whole lungs and blood by qRT-PCR, isolated RNA was converted to cDNA using MMLV-RT (Invitrogen, CA) under the following conditions: 4°C for 15 min; 16°C for 30 min; 42°C for 30 min; 85°C for 5 min. TaqMan microRNA assays were used for expression analysis of hsa-miR-34a and hsa-let-7b (Life Technologies, CA). Following cDNA synthesis, qPCR was performed on 2 μL of cDNA on the ABI Prism 7900HT SDS (Life Technologies) using Platinum Taq Polymerase (Invitrogen) under the following cycling conditions: 95°C for 1 min (initial denature); then 50 cycles of 95°C for 5 sec, 60°C for 30 sec. Additions to the manufacturers’ reagents include DMSO (final concentration of 6%) and tetramethylammoniumchloride (TMAC; final concentration of 50 mM in both RT and PCR) to improve the slope, linearity and sensitivity of the assays.

Absolute quantitation was determined as copy number per nanogram of RNA input by extrapolation to the miR-34a and let-7b standard curves. Final values were calculated by factoring in losses during the acid-phenol extraction (carry-over of only half of the aqueous phase), elution volume, volume of RNA eluent into the reverse transcription reaction and starting mass of RNA.

For the detection and quantification of miRNAs and mRNA targets by qRT-PCR from individual lung tumors, isolated RNA was converted to cDNA using miScript II RT Kit (Qiagen). Following cDNA synthesis, qPCR was performed on 2 μL of cDNA on the Roche 480 using Qiagen’s miScript SYBR Green Assay Kit. Qiagen probes were used for expression analysis of all genes of interest. Relative quantification of miR-34a and let-7b and their targets was determined as the concentration (fold-change) of the genes of interest compared to the normal lung tissue. Rnu6B (miRNA), and Gapdh and Actin (mRNA) were used as housekeeping genes to normalize for RNA input variances into the reverse transcription reaction.

Serum cytokines

Mice serum samples were used to test for IL-6, IL-1b and TNF-α concentrations using the cytokine assay kit from R&D systems according to manufacturer’s instructions (Fluorokine MAP Mouse Base Kit Cat# LUM000, mouse IL-1b Cat# LUM401, mouse IL-6 Cat# LUM406, mouse TNF-α Cat# LUM410). Briefly, serum samples were thawed and cleared from debris by centrifugation at 1000 × g for 10 minutes prior to analysis. All samples or standards were added to a 96-well plate together with assay buffer, and cytokine beads. Plates were gently shaked for 3 hours at room temperature and spun as above. After removal of the supernatant, wells were washed. A detection antibody was added and the plates were incubated 1 hour at RT with shaking. Plates were rinsed and incubated with streptavidin-phycoerythrin for 30 minutes at RT. The wells were washed and the capture beads were re-suspended to acquire data using the Luminex 100 IS2.2 detection system. Cytokine standard curves were used to calculate the cytokine concentrations in serum samples [pg/mL].

Significance.

Lung cancer is the leading cause of cancerous deaths worldwide, with non-small cell lung cancer making up the bulk of newly identified cases. Unfortunately, current therapeutics fail to treat the majority of individuals with this disease, even when using recently developed targeted therapeutics. Primary or secondary resistance from targeted agents can be attributed to acquired mutations in the target that render the initial therapeutic less effective or ineffective and/or compensatory mutations in other pathways that allow the cancer cells to overcome the effects of the targeted therapy. Combinatorial microRNA therapeutics is a new frontier in gene therapy that has the ability to overcome these challenges by targeting multiple components of key oncogenic pathways.