|Home | About | Journals | Submit | Contact Us | Français|
A significant barrier to the successful general development of small-interfering RNA (siRNA) therapeutics is the ability to deliver them systemically to target organs and cell types. In this study, we have developed a mouse strain that will facilitate the evaluation of the efficacy of siRNA delivery strategies. This strain contains robust ubiquitous expression of firefly luciferase from germ line Cre-mediated recombination of the ROSA26-LSL-Luc allele. We show that luciferase is highly and uniformly expressed in all tissues examined. Using this mouse model, we describe a facile assay that enables the assessment of the pharmacodynamics of a systemically delivered siRNA formulation. These mice can also be used as universal donors, enabling the efficient and sensitive monitoring of cell trafficking or tissue transplantation. The primary advantage of this approach is that siRNA efficacy against a nonessential target can be easily evaluated in any tissue. This strain should generally enhance the ability to rapidly screen, compare and optimize various siRNA formulations for tissue-targeted or -enhanced systemic delivery in a preclinical development setting.
Systemic delivery of small-interfering RNA (siRNA) to target cells and organs remains a significant barrier for the successful development of this new class of therapeutics for many possible indications. A variety of approaches to this problem is currently under intensive investigation including means to augment the pharmacokinetic and distribution properties of siRNAs as well as strategies to effect cell type–specific uptake of siRNAs in vivo (for review see ref. 1). The ability to evaluate these strategies rapidly in preclinical models should facilitate the selection and optimization of the best approaches. New developments in small animal molecular imaging have significantly impacted modern therapeutic development.2 Among these is bioluminescence imaging (BLI),3 which involves the sensitive detection of photons from cells expressing a luciferase enzyme. Luciferase-mediated oxidation of the luciferin substrate in the presence of adenosine triphosphate, oxygen, and Mg2+ generates bioluminescence which is detected using a sensitive charge-coupled device camera. BLI offers several advantages over alternative molecular imaging techniques in that it is relatively simple to execute, is cost effective, and enables noninvasive serial imaging over time without harm to the experimental animal.
Several transgenic mouse models have been engineered to express a luciferase enzyme. Expression of luciferase can be restricted to certain tissues or cell types using tissue-specific promoters (e.g., refs. 4–6), and engineered luciferase enzymes can be activated in response to certain molecular processes (e.g., refs. 7–9). Luciferase can also be ubiquitously expressed in an animal using a constitutive promoter.10,11 Constitutive promoters, such as the β-actin promoter, have been used to drive luciferase expression in these transgenic mice. Higher luciferase expression may generate more signal; allowing for more sensitive detection in vivo. In each case, BLI has provided novel insights into a number of biological processes such as gene regulation, apoptosis, transplantation, cell trafficking, and spontaneous tumorigenesis.5,12–14 BLI is also routinely used to assess tumor response to therapy in both xenograft and syngeneic transplant models.15–17 Such models provide an attractive means to screen the effects of novel antineoplastic agents. They also offer the opportunity to examine site-specific effects of compounds and have been particularly useful while studying the effects of local and systemic siRNA delivery.18
In order to develop a facile screen to study the pharmacodynamics of systemically delivered siRNAs, we generated a mouse strain that has robust ubiquitous expression of luciferase from an endogenous promoter. We utilized the conditional activation of pGL3 luciferase from ROSA26-LSL-Luc mice developed by Safran et al.19 through ectopic expression of Cre-recombinase in oocytes under the control of the probasin promoter.20 We demonstrate that luciferase expression is detected in all tissues examined including lymphocytes which can be serially imaged over time when adoptively transferred into naive mice. Using our mouse model we demonstrate that we can efficiently assess pharmacodynamic properties of a novel liposomal siRNA formulation targeting luciferase. We show that a single bolus dose of this siRNA targets luciferase knockdown to the liver, largely avoiding other tissues for up to 16 days, demonstrating the potential therapeutic application of liposomal siRNA therapy.
In order to generate a mouse that ubiquitously expresses luciferase, we crossed male ROSA26-LSL-Luc and female Pb-Cre4+ mice.19,20 Pb-Cre4+ mice contain a Cre-recombinase transgene under control of the rat probasin promoter where expression is restricted to prostatic epithelium in male mice. However, it has been documented that female mice which harbor the Pb-Cre4 transgene have ectopic expression of Cre-recombinase in oocytes.20 ROSA26-LSL-Luc mice are a strain in which a pGL3 luciferase complementary DNA preceded by a LoxP-Stop-LoxP (LSL) is inserted into the ubiquitously expressed ROSA26 locus.19 Using these two strains of mice we took advantage of the ectopic oocyte expression of Cre and crossed female mice containing the Pb-Cre4 transgene that were also homozygous for the ROSA26-LSL-Luc allele (R26-LSL-Lucfl/fl) with male mice that were similarly positive for Pb-Cre4+ and R26-LSL-Lucfl/fl (Figure 1a). Cre expression in the oocyte (see inset in Figure 1a) should lead to recombination of the LSL-Luc allele such that the stop codon upstream of luciferase is removed and luciferase becomes expressed from the endogenous ROSA26 promoter. If this recombination occurs in the zygote, each daughter cell will contain two active copies of the luciferase allele and the result will be a mouse that has ubiquitous expression of the luciferase enzyme. To confirm this, we performed BLI on the offspring from this cross (Figure 1b). BLI of 4-week-old animals revealed that most mice were positive for luciferase expression. However, we also produced some Pb-Cre4+ mice that did not have ubiquitous luciferase expression indicating that not all of the oocytes had ectopic expression of Cre-recombinase. We found that Cre-recombinase was ectopically active in ~93% of the offspring (N = 43 mice). For convenience, we designated these mice as FLASH (firefly luciferase activated systemically in homozygotes). Ex vivo imaging confirmed that luciferase was expressed in all tissues examined (Figure 1b). We found that intravenous injections of D-luciferin (150 mg/kg) increased maximal signal output by approximately fivefold compared to intraperitoneal injections (Figure 1c). Maximal signal output using intravenous injection of D-luciferin occurred immediately after administration, while optimal imaging for intraperitoneal injections occurred ~15–20 minutes after D-luciferin administration as has been described previously in other models. 21 In order for us to examine whether luciferase was expressed homogeneously, we prepared lysates from a panel of tissues from both male and female FLASH mice and performed an in vitro luciferase activity assay. BLI of protein lysates from FLASH mice revealed that luciferase is expressed uniformly in all tissues examined except for the testes which have a higher expression level (Figure 1d). The reasons for the latter are unclear, however original characterization of the ROSAβgeo26 mouse documented that ROSA26 was highly active in the testes.22
Conventional BLI strategies to image immune cell trafficking or tissue transplantation require purification of the desired cell types and transduction with a luciferase reporter using a variety of gene transfer techniques.23–25 Gene transfer efficiency remains an obstacle and stable reporter expression can be variable over time. The use of a transgenic mouse that has active ubiquitous luciferase expression may bypass some of these limitations. In order to test whether we could use FLASH mice as a universal donor, we purified CD4+, CD8+, B Cells, and NK cells from FLASH mice and performed ex vivo imaging (Figure 2a). All four cell types produced detectable luciferase expression higher than that observed in previous studies.10 Signal output from these cells was relatively uniform ranging from 1.1 to 2.7 photons/s. To determine whether we could image these cells in vivo, we adoptively transferred 1 × 107 CD4+ T cells into albino C57BL6 mice (C57BL6TYRC2J mice). BLI of mice 24 hours postadoptive transfer revealed robust bioluminescence signal in all mice. Focal signals were detected in vicinity of the spleen, cervical lymph nodes, mesenteric lymph nodes, and inguinal lymph nodes (Figure 2b). Ex vivo imaging confirmed the location of CD4+ T-cell engraftment (Figure 2c). Longitudinal analyses of CD4+, CD8+, and NK cell adoptive transfer into severe combined immunodeficiency mice revealed that bioluminescence signal was increased 60 days postadoptive transfer (Supplementary Figure S1). These data demonstrate that cells derived from FLASH mice can be transplanted and sensitively imaged in vivo.
In order to determine whether FLASH mice could be used to assess the pharmacodynamics of systemically delivered siRNAs, we first examined whether a liposomal formulation of luciferase-targeted siRNA (si-Luc)26 could inhibit the expression of luciferase in mouse embryonic fibroblasts (MEFSs) isolated from FLASH mice (Figure 3a). This experiment demonstrated that we are able to reduce luciferase protein levels by ~90% using a 90 nmol/l dose of si-Luc in vitro. Importantly, a 270 nmol/l dose of control siRNA–targeting factor VII (si-FVII) did not reduce luciferase expression levels. To determine the efficacy of si-Luc in vivo, we treated FLASH mice with either si-Luc or si-FVII via single bolus intravenous injection (5 mg/kg). Ex vivo BLI of liver and lung 3 days after injection revealed that the si-Luc was able to reduce the expression of luciferase in the liver compared to si-FVII-treated mice (Figure 3b). We also detected a slight decrease in ex vivo bioluminescence from the lungs of si-Luc-treated mice. We prepared RNA and protein from these tissues and tested whether the reduced levels of luciferase expression seen ex vivo could be confirmed using quantitative reverse transcriptase-PCR and an in vitro luciferase activity assay. Luciferase RNA and protein are reduced by ~90% in the livers of mice treated with si-Luc compared to si-FVII controltreated mice (Figure 3c). Although there may be a trend toward reduced luciferase RNA and protein in the lungs of si-Luc-treated mice, this did not reach statistical significance. However it is still possible that liposomal si-Luc exerts minor effects in the lung. In order to establish that we could sensitively measure differences in luciferase activity in the linear range in these various tissues, we prepared protein lysates from a panel of tissues and performed an in vitro luciferase activity assay on a wide range of protein concentrations (Figure 3d and data not shown). In all tissues examined there was a high correlation (R2 = 0.99) between light output and protein concentration. These data demonstrate the utility of FLASH mice for analyzing siRNA efficacy in vivo.
In order to characterize the pharmacodynamic properties of liposomal si-Luc, we injected mice once either by intravenous tail vein injection or intraperitoneal injection with 5 mg/kg of si-Luc or si-FVII. We then harvested the tissue 3 days after injection and performed a luciferase activity assay. Intravenous administration of liposomal-formulated siRNAs is far superior in reducing luciferase expression in the liver (Figure 4a). We next sought to test whether we could achieve greater inhibition of luciferase expression in heterozygous mice that only contain one copy of the luciferase reporter allele (fl/wt mice). We treated both homozygous (fl/fl) and heterozygous (fl/wt) FLASH mice with si-Luc or si-FVII and prepared protein lysates 4 days after injection (Figure 4b). Inhibition of luciferase expression was not significantly different between fl/wt and fl/fl FLASH mice indicating that changes in the levels of luciferase expression within a twofold range do not affect RNA interference efficacy in this model. To determine whether there are any gender-specific differences in RNA interference between male FLASH mice and female FLASH mice, we treated both males and females with either si-Luc or si-FVII (Figure 4c). Levels of luciferase inhibition in the livers of treated male and female mice were not statistically different and we were unable to inhibit luciferase in any other tissue in female FLASH mice. To test the pharmacodynamic properties of liposomal si-Luc further, we performed a time course of luciferase inhibition in FLASH mice. We treated mice with liposomal-formulated siRNAs and prepared protein lysates from a panel of tissues at days 4, 8, and 16 after injection. Luciferase expression in the liver slowly increases from ~20% at day 4 to ~50% at day 16 after injection (Figure 4d). These data show that a single injection of siRNA is able to reduce luciferase expression in the liver significantly for 16 days, but we did not observe any significant decreases in any other tissues examined. Hence, we have demonstrated the utility of FLASH mice in assessing pharmacodynamic properties of in vivo–administered siRNAs.
Here, we describe the development of a mouse strain that contains robust ubiquitous expression of a firefly luciferase reporter allele from the endogenous ROSA26 promoter. By taking advantage of ectopic germ line expression of Cre in the Pb-Cre4 transgenic mouse, we activated luciferase expression in all tissues of ROSA26-LSL-Luc mice. While we have not definitively demonstrated that luciferase is expressed in every cell in these mice, the amount of luciferase activity in a variety of organ homogenates and the similar levels of luciferase activity on a per cell basis in several lymphocyte populations suggests that expression is ubiquitous and uniform in all cells except the testis where higher expression levels were observed. Previously developed mouse models aimed at generating luciferase-expressing mice involved transgenic approaches. Expression of luciferase in these models may vary due to position effects, transgene silencing, or promoter expression in murine tissues. For example, although it is difficult to compare directly, the LucRep mouse developed by Lyons et al.11 apparently demonstrates less signal than FLASH mice when subjected to similar imaging conditions. Moreover, in FLASH mice, luciferase is expressed from an endogenous promoter (ROSA26) in its normal genomic context, which is probably a more stable configuration for expressing consistent levels of luciferase through many generations of mice. The ROSA26 promoter was originally discovered by random retroviral gene trapping in embryonic stem cells.27 Since then it has been targeted for genetic manipulation offering researchers novel methods to study temporal and tissue-specific gene expression.19,28,29
Luciferase expression was detected in all tissues examined and at higher levels in lymphocytes than previously described mouse models.30 The level of expression was sufficiently robust to measure the trafficking of 107 CD4+ cells to various lymphoid organs in living animals. Our mouse model has several advantages as universal donors for analyzing cell trafficking or tissue transplantation. Previous studies aimed at imaging trafficking of immune cells via BLI were performed using purified subsets of cells that had been engineered to express luciferase in vitro. Inherent problems with these studies are transfection efficiencies, promoter expression in cells, and stable reporter expression over time.30,31 Cells derived from FLASH mice already contain endogenous expression of luciferase from a constitutively active promoter and mitigate the need for any transduction of transfection of reporter constructs. Nuclear imaging strategies (single photon emission computed tomography and positron emission tomography) have also been used to image immune cell migration noninvasively.25,32 Although these techniques are quantitative, they also demand sophisticated instrumentation and are relatively expensive and less sensitive. BLI of a luciferase reporter provides robust, cost–effective, and sensitive means to image these processes in vivo.
We also describe the use of FLASH mice to analyze the pharmacodynamics of systemically delivered siRNAs. siRNA therapy is rapidly emerging as an approach for the treatment of a variety of human diseases.1 The availability of animal models that enable high-throughput screening of siRNA delivery, efficacy, and safety would greatly facilitate the progression of siRNA therapy into clinical trials. Traditional assays to determine siRNA efficacy in vivo involve preparation of RNA and protein for quantitative reverse transcriptase PCR and western blot analysis. Here, we describe an alternative method using an in vitro luciferase activity assay. We demonstrate that a single intravenous injection of a liposomalformulated siRNA-targeting luciferase was able to reduce the levels of luciferase mRNA and protein by ~90% in the liver without significantly affecting expression in other tissues. Previous studies using the same liposomal siRNA formulation targeting apolipoprotein B and FVII documented similar levels of target inhibition in the liver.26,33 We found that intravenous injections were far superior in inhibiting luciferase expression in the liver relative to intraperitoneal injections. This suggests that siRNA delivery efficacy is driven primarily by Cmax (peak plasma concentration) in the liver. Reduction of luciferase expression was relatively durable, evident even 16 days after injection. Ongoing efforts are aimed at harnessing this powerful method of gene inhibition for certain liver-associated diseases.34 We were not able to inhibit luciferase significantly in any other tissues analyzed. Liposomal-formulated siRNAs used in this study are not targeted by ligand attachment, but this formulation has been optimized to target delivery to the liver.26 Efforts continue to develop ligand-based tissue- and tumor-specific delivery.35,36 The primary advantages of this model are that siRNA efficacy can be determined against a uniformly expressed and nonessential target in all tissues. One drawback of the approach described here is that the ubiquitous expression of luciferase precludes evaluation of tissue-specific effects in living animals. However, it may be possible to transplant target cells of interest, as we have shown here for lymphoid cells, into nonbioluminescent hosts so that effects of siRNAs specifically on those cells can be evaluated in vivo. We believe that FLASH mice offer the opportunity to screen such formulations rapidly for targeting efficacy. The ability to determine the pharmacodynamics of novel siRNA formulations rapidly in FLASH mice will aid in their progression as potential therapeutics.
All animal procedures were performed with approval from the University of Iowa Animal Care and Use Committee and by the authors’ Institutional Review Board. Pb-Cre4+ mice (obtained from the National Institutes of Health Mouse Models of Human Cancer Consortium) were mated with ROSA26-LSL-Luc mice (gift from William Kaelin) and backcrossed onto the C57BL6TYRC2J (albino C57BL6; Jackson labs, Bar Harbor, ME) background for six generations. N6 Pb-Cre4+, ROSA26-LSL-Lucfl/fl male mice were then mated to N6 PbCre4+, ROSA26-LSL-Lucfl/fl female mice. Offspring were genotyped by PCR for the presence of the Pb-Cre4+ and ROSA26-LSL-Luc alleles using gene-specific primers and imaged in an IVIS100 imaging system to assess bioluminescence.
All BLI was performed in an IVIS100 imaging system (Caliper Life Sciences, Alameda, CA). For BLI of mice, D-luciferin (Gold Biotechnology, St Louis, MO) was administered to each mouse via intraperitoneal or intravenous injection at a dose of 150 mg/kg. Animals were then anesthetized in a chamber with 3% isoflurane and immediately placed onto the imaging platform while being maintained on 3% isoflurane. Mice were then imaged after 2 minutes of substrate injection using a 20 cm field of view and an exposure time of 1 second and serially imaged every 4 minute for a total period of 50 minutes. Bioluminescence values were calculated by measuring photon flux (photons/s) in the region of interest surrounding the bioluminescence signal emanating from the mice. For BLI of protein lysates, samples were imaged using a 10-cm field of view and an exposure time of 10 seconds. For BLI of lymphocytes, cells were imaged using a 10-cm field of view and an exposure time of 5 minutes. Bioluminescence values were calculated by measuring photons/s for each sample.
The liposomal formulation comprising the novel lipidoid 98N12-5, a polyethylene glycol lipid (PEG-DMG), and cholesterol has been described in ref. 26). The sequence of si-FVII has been described in ref. 26 and the sequence of si-Luc is as follows: sense strand, cuuAcGcuGAGuAcuucGAT*T; antisense strand, UCGAAGUACUCAGCGUAAGT*T. Lower case letters represent 2′-O-methyl-nucleotides. The asterisks denote phosphorothioate linkages. The siRNAs were synthesized by Alnylam Pharmaceuticals (Cambridge, MA) as described in ref. 26. si-Luc was used to test luciferase inhibition in FLASH MEFS. MEFS were isolated from E13 embryos using the following protocol. Isolated embryos were eviscerated, washed in phosphate-buffered saline and minced using scissors and scalpels. After enzymatic dissociation (Trypsin/EDTA for 30 minutes at 37 °C), embryonic tissue was mixed with 18 ml of Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA) supplemented with 10% fetal bovine serum (Hyclone, South Logan, UT), 1% nonessential amino acids (Invitrogen), 1% penicillin/streptomycin and left for 10 minutes to let clumps fall. Supernatant was then plated into 10-cm dishes and incubated at 37 °C in an atmosphere containing 5% CO2. MEFS were then grown to confluence and frozen stocks (8% dimethyl sulfoxide) were prepared. For in vitro siRNA treatment, MEFS were seeded into 24-well plates at 1 × 105 cells/well. Following overnight culture, a dose–response of liposomal siRNA (10, 30, 90, and 270 nmol/l si-Luc and 270 nmol/l si-FVII) was added in triplicate to the MEFS along with a nontreated control. Fresh media was added 24 hours later. Seventy-two hours after siRNA treatment, D-luciferin (150 µg/ml) was added to each well and BLI was performed. Cell viability for each treatment was determined using a cell proliferation reagent WST-1 (Roche, Nutley, NJ) according to the manufacturer’s instructions. For in vivo administration of siRNAs, 5 mg/kg siRNA were injected via lateral tail vein injection into FLASH mice. Following euthanasia, organs were removed and analyzed using an in vitro luciferase assay. Photon flux for si-Luc-treated mice was compared to si-FVII-treated mice and data were represented as % FVII control (set at 100%).
Tissues from FLASH mice were removed and either flash frozen in liquid nitrogen for subsequent protein preparation or immediately placed (100 mg) into 1 ml of TRIZOL (Invitrogen) for RNA preparation. For RNA extraction, samples in TRIZOL were homogenized using a Powergen125 Teflon homogenizer (Fisher Scientific, Pittsburgh, PA). Samples were then subjected to RNA extraction under the manufacturer’s conditions. Purified RNA preparations were subjected to DNaseI treatment to remove any contaminating genomic DNA. For protein preparation, frozen tissues were homogenized into a powder using a mortar and pestle. Samples were then suspended in 1× reporter lysis buffer (Buffer RLB; Promega, Madison, WI) and frozen overnight at −80 °C. Samples were then subjected to three freeze–thaw cycles (37 °C—LN2) and centrifuged at 13,000 rpm for 7.5 minutes. Supernatants were analyzed for protein quantity using a Bio-Rad (Hercules, CA) protein assay using manufacturer’s conditions.
RNA samples were subjected to quantitative reverse transcriptase PCR as previously described in ref. 17. Protein samples (150 µg) were subjected to an in vitro luciferase activity assay as previously described in ref. 17.
Single cell suspensions from spleen and lymph node were resuspended in 0.83% NH4Cl (in 1 mmol/l Trizma Base) for erythrocyte lysis. Cells were washed, counted, and resuspended at 2 × 108/ml in bead buffer (phosphate-buffered saline with 0.1% bovine serum albumin and 0.02% NaN3). B cells were depleted by incubation with BioMag goat anti-mouse immunoglobulin G and goat anti-mouse immunoglobulin M paramagnetic beads (Polysciences, Warrington, PA). After B-cell depletion, the remaining cells were again resuspended at 2 × 108/ml in bead buffer and incubated with either BioMag SelectaPur CD8a paramagnetic beads for generation of enriched CD4+ T cells, BioMag SelectaPur CD4 paramagnetic beads for generation of enriched CD8+ T cells or a combination of the two for NK cell enrichment. B-cell enrichment was performed by incubation with a combination of BioMag SelectaPur CD4 and CD8a paramagnetic beads. Viability of enriched cell fractions exceeded 95% and purity was >80%. Cells were washed repeatedly and resuspended in phosphate-buffered saline for BLI and adoptive transfer. For adoptive transfer, 1 × 107 cells/mouse (C57BL6TYRC2J) were injected via the lateral tail vein.
All statistical analyses utilized analysis of variance with Bonferroni post-test.
Longitudinal BLI of adoptively transferred lymphocytes.
We thank members of the Henry laboratory for comments on the manuscript. This work was supported by a grant from the Prostate Cancer Foundation (to M.D.H.), a pilot and feasibility grant (to M.D.H.) from the University of Iowa Center for Gene Therapy (P30 DK54759), a VA Merit Award (to Z.K.B.), and by the National Institutes of Health grants RO1AA014418 (to Z.K.B.), RO1EB00244 (to D.G.A.), and a subcontract from R21CA127189 (to M.D.H.). J.R.D, A.A., and D.B. are employees of Alnylam Pharmaceuticals which is developing therapeutics based on RNA interference. This work was performed in Iowa City, Iowa, USA.