Safety evaluation of 2'-deoxy-2'-fluoro nucleotides in GalNAc-siRNA conjugates.

For oligonucleotide therapeutics, chemical modifications of the sugar-phosphate backbone are frequently used to confer drug-like properties. Because 2'-deoxy-2'-fluoro (2'-F) nucleotides are not known to occur naturally, their safety profile was assessed when used in revusiran and ALN-TTRSC02, two short interfering RNAs (siRNAs), of the same sequence but different chemical modification pattern and metabolic stability, conjugated to an N-acetylgalactosamine (GalNAc) ligand for targeted delivery to hepatocytes. Exposure to 2'-F-monomer metabolites was low and transient in rats and humans. In vitro, 2'-F-nucleoside 5'-triphosphates were neither inhibitors nor preferred substrates for human polymerases, and no obligate or non-obligate chain termination was observed. Modest effects on cell viability and mitochondrial DNA were observed in vitro in a subset of cell types at high concentrations of 2'-F-nucleosides, typically not attained in vivo. No apparent functional impact on mitochondria and no significant accumulation of 2'-F-monomers were observed after weekly administration of two GalNAc-siRNA conjugates in rats for ∼2 years. Taken together, the results support the conclusion that 2'-F nucleotides can be safely applied for the design of metabolically stabilized therapeutic GalNAc-siRNAs with favorable potency and prolonged duration of activity allowing for low dose and infrequent dosing.

Impact of Oligonucleotide Structure, Chemistry, and Delivery Method on In Vitro Cytotoxicity.

Single-stranded (ss) 2'-fluoro (2'-F)-modified oligonucleotides (ONs) with a full phosphorothioate (PS) backbone have been reported to be cytotoxic and cause DNA double-strand breaks (DSBs) when transfected into HeLa cells. However, the molecular determinants of these effects have not been fully explored. In this study, we investigated the impact of ON structure, chemistry, delivery method, and cell type on in vitro cytotoxicity and DSBs. We found that ss PS-ONs were more cytotoxic than double-stranded (ds) PS-ONs, irrespective of the 2'-ribose chemistry, inclusive of the 2'-F modification. Cytotoxicity of ss ONs was most affected by the total PS content, with an additional contribution of 2'-F substitutions in HeLa, but not HepG2, cells. The relatively mild cytotoxicity of ds ONs was most impacted by long contiguous PS stretches combined with 2'-F substitutions. None of the tested ds 2'-F-modified PS-ONs caused DSBs, while the previously reported DSBs caused by ss 2'-F-modified PS-ONs were PS dependent. HeLa cells were more sensitive to ON-mediated toxicity when transfected with Lipofectamine 2000 versus Lipofectamine RNAiMax. Importantly, asialoglycoprotein receptor-mediated uptake of N-acetylgalactosamine-conjugated ss or ds PS-ONs, even those with long PS stretches and high 2'-F content, was neither cytotoxic nor caused DSBs at transfection-equivalent exposures. These results suggest that in vitro cytotoxicity and DSBs associated with ONs are delivery method dependent and primarily determined by single-stranded nature and PS content of ONs.

Exposure to siRNA-GalNAc Conjugates in Systems of the Standard Test Battery for Genotoxicity.

Registration of pharmaceuticals requires an assessment of their genotoxic potential using in vitro and in vivo tests outlined in the International Conference on Harmonisation (ICH) guidance S2(R1). We have evaluated numerous siRNA-N-acetylgalactosamine (GalNAc) conjugates containing phosphorothioate linkages and various combinations of 2'-fluoro and 2'-O-methyl ribose modifications of multiple nucleotides in the ICH battery of assays, all of which have uniformly yielded negative results. To verify these negative genotoxicity results, in this study we confirm test article exposure using toolkit small interfering RNAs (siRNAs) representative of those in the clinic. In the Ames test, the highest uptake of the siRNA-GalNAc conjugates occurred at 1 h postdose in all bacterial strains independent of siRNA sequence or chemistry (up to ∼14,000 siRNA molecules per cell), followed by metabolic degradation of the parent siRNA at 6, 24, and 48 h postdose. siRNA-GalNAc conjugates were internalized by bacteria as assessed by protection from the addition of nucleases to the culture media following uptake and by the requirement of cell lysis for detection of the siRNA. In the in vitro chromosome aberration assay, uptake was observed in Chinese hamster ovary cells (up to ∼5,500 siRNA molecules per cell at 21 h postdose) and in CD3+ human peripheral blood lymphocytes (up to ∼500 siRNA molecules per cell at 21 h postdose). In the in vivo micronucleus assay in rat bone marrow, exposure to parent siRNA was 100-350 µg of antisense strand per gram of protein at 24 and 48 h postlimit dose of 2 g/kg. Loss of terminal nucleotides was detected in bone marrow by mass spectrometry, indicating exposure to monomer metabolites as well. Negative genotoxicity results were also confirmed in an in vitro double-strand DNA break assay in HeLa and HepG2 cells where exposure was maximized using transfection reagents. Thus negative genotoxicity assay results for siRNA-GalNAc conjugates were valid and not the result of poor or no intracellular exposure.

OSWG Recommendations for Genotoxicity Testing of Novel Oligonucleotide-Based Therapeutics.

The Oligonucleotide Safety Working Group subcommittee on genotoxicity testing considers therapeutic oligonucleotides (ONs) unlikely to be genotoxic based on their properties and on the negative results for ONs tested to date. Nonetheless, the subcommittee believes that genotoxicity testing of new ONs is warranted because modified monomers could be liberated from a metabolized ON and incorporated into DNA and could hypothetically cause chain termination, miscoding, and/or faulty replication or repair. The standard test battery as described in Option 1 of International Conference on Harmonisation S2(R1) is generally adequate to assess such potential. However, for the in vitro assay for gene mutations, mammalian cells are considered more relevant than bacteria for most ONs due to their known responsiveness to nucleosides and their greater potential for ON uptake; on the other hand, bacterial assays may be more appropriate for ONs containing non-ON components. Testing is not recommended for ONs with only naturally occurring chemistries or for ONs with chemistries for which there is documented lack of genotoxicity in systems with demonstrated cellular uptake. Testing is recommended for ONs that contain non-natural chemical modifications and use of the complete drug product (including linkers, conjugates, and liposomes) is suggested to provide the most clinically relevant assessment. Documentation of uptake into cells comparable to those used for genotoxicity testing is proposed because intracellular exposure cannot be assumed for these large molecules. ONs could also hypothetically cause mutations through triple helix formation with genomic DNA and no tests are available for detection of such sequence-specific mutations across the entire genome. However, because the potential for triplex formation by therapeutic ONs is extremely low, this potential can be assessed adequately by sequence analysis.

An RNAi therapeutic targeting antithrombin to rebalance the coagulation system and promote hemostasis in hemophilia.

Hemophilia A and B are inherited bleeding disorders characterized by deficiencies in procoagulant factor VIII (FVIII) or factor IX (FIX), respectively. There remains a substantial unmet medical need in hemophilia, especially in patients with inhibitory antibodies against replacement factor therapy, for novel and improved therapeutic agents that can be used prophylactically to provide effective hemostasis. Guided by reports suggesting that co-inheritance of prothrombotic mutations may ameliorate the clinical phenotype in hemophilia, we developed an RNA interference (RNAi) therapeutic (ALN-AT3) targeting antithrombin (AT) as a means to promote hemostasis in hemophilia. When administered subcutaneously, ALN-AT3 showed potent, dose-dependent, and durable reduction of AT levels in wild-type mice, mice with hemophilia A, and nonhuman primates (NHPs). In NHPs, a 50% reduction in AT levels was achieved with weekly dosing at approximately 0.125 mg/kg, and a near-complete reduction in AT levels was achieved with weekly dosing at 1.5 mg/kg. Treatment with ALN-AT3 promoted hemostasis in mouse models of hemophilia and led to improved thrombin generation in an NHP model of hemophilia A with anti-factor VIII inhibitors. This investigational compound is currently in phase 1 clinical testing in subjects with hemophilia A or B.

Shielding of Lipid Nanoparticles for siRNA Delivery: Impact on Physicochemical Properties, Cytokine Induction, and Efficacy.

Formulation of short interfering RNA (siRNA) into multicomponent lipid nanoparticles (LNP) is an effective strategy for hepatic delivery and therapeutic gene silencing. This study systematically evaluated the effect of polyethylene glycol (PEG) density on LNP physicochemical properties, innate immune response stimulation, and in vivo efficacy. Increased PEG density not only shielded LNP surface charge but also reduced hemolytic activity, suggesting the formation of a steric barrier. In addition, increasing the PEG density reduced LNP immunostimulatory potential as reflected in cytokine induction both in vivo and in vitro. Higher PEG density also hindered in vivo efficacy, presumably due to reduced association with apolipoprotein E (ApoE), a protein which serves as an endogenous targeting ligand to hepatocytes. This effect could be overcome by incorporating an exogenous targeting ligand into the highly shielded LNPs, thereby circumventing the requirement for ApoE association. Therefore, these studies provide useful information for the rational design of LNP-based siRNA delivery systems with an optimal safety and efficacy profile.

Degradable lipid nanoparticles with predictable in vivo siRNA delivery activity.

One of the most significant challenges in the development of clinically viable delivery systems for RNA interference therapeutics is to understand how molecular structures influence delivery efficacy. Here, we have synthesized 1,400 degradable lipidoids and evaluate their transfection ability and structure-function activity. We show that lipidoid nanoparticles mediate potent gene knockdown in hepatocytes and immune cell populations on IV administration to mice (siRNA EC50 values as low as 0.01 mg kg(-1)). We identify four necessary and sufficient structural and pKa criteria that robustly predict the ability of nanoparticles to mediate greater than 95% protein silencing in vivo. Because these efficacy criteria can be dictated through chemical design, this discovery could eliminate our dependence on time-consuming and expensive cell culture assays and animal testing. Herein, we identify promising degradable lipidoids and describe new design criteria that reliably predict in vivo siRNA delivery efficacy without any prior biological testing.

Lipopeptide nanoparticles for potent and selective siRNA delivery in rodents and nonhuman primates.

siRNA therapeutics have promise for the treatment of a wide range of genetic disorders. Motivated by lipoproteins, we report lipopeptide nanoparticles as potent and selective siRNA carriers with a wide therapeutic index. Lead material cKK-E12 showed potent silencing effects in mice (ED50 ∼ 0.002 mg/kg), rats (ED50 < 0.01 mg/kg), and nonhuman primates (over 95% silencing at 0.3 mg/kg). Apolipoprotein E plays a significant role in the potency of cKK-E12 both in vitro and in vivo. cKK-E12 was highly selective toward liver parenchymal cell in vivo, with orders of magnitude lower doses needed to silence in hepatocytes compared with endothelial cells and immune cells in different organs. Toxicity studies showed that cKK-E12 was well tolerated in rats at a dose of 1 mg/kg (over 100-fold higher than the ED50). To our knowledge, this is the most efficacious and selective nonviral siRNA delivery system for gene silencing in hepatocytes reported to date.

Biodegradable lipids enabling rapidly eliminated lipid nanoparticles for systemic delivery of RNAi therapeutics.

In recent years, RNA interference (RNAi) therapeutics, most notably with lipid nanoparticle-based delivery systems, have advanced into human clinical trials. The results from these early clinical trials suggest that lipid nanoparticles (LNPs), and the novel ionizable lipids that comprise them, will be important materials in this emerging field of medicine. A persistent theme in the use of materials for biomedical applications has been the incorporation of biodegradability as a means to improve biocompatibility and/or to facilitate elimination. Therefore, the aim of this work was to further advance the LNP platform through the development of novel, next-generation lipids that combine the excellent potency of the most advanced lipids currently available with biodegradable functionality. As a representative example of this novel class of biodegradable lipids, the lipid evaluated in this work displays rapid elimination from plasma and tissues, substantially improved tolerability in preclinical studies, while maintaining in vivo potency on par with that of the most advanced lipids currently available.

Harnessing a physiologic mechanism for siRNA delivery with mimetic lipoprotein particles.

Therapeutics based on RNA interference (RNAi) have emerged as a potential new class of drugs for treating human disease by silencing the target messenger RNA (mRNA), thereby reducing levels of the corresponding pathogenic protein. The major challenge for RNAi therapeutics is the development of safe delivery vehicles for small interfering RNAs (siRNAs). We previously showed that cholesterol-conjugated siRNAs (chol-siRNA) associate with plasma lipoprotein particles and distribute primarily to the liver after systemic administration to mice. We further demonstrated enhancement of silencing by administration of chol-siRNA pre-associated with isolated high-density lipoprotein (HDL) or low-density lipoprotein (LDL). In this study, we investigated mimetic lipoprotein particle prepared from recombinant apolipoprotein A1 (apoA) and apolipoprotein E3 (apoE) as a delivery vehicle for chol-siRNAs. We show that apoE-containing particle (E-lip) is highly effective in functional delivery of chol-siRNA to mouse liver. E-lip delivery was found to be considerably more potent than apoA-containing particle (A-lip). Furthermore, E-lip-mediated delivery was not significantly affected by high endogenous levels of plasma LDL. These results demonstrate that E-lip has substantial potential as delivery vehicles for lipophilic conjugates of siRNAs.

Safety profile of RNAi nanomedicines.

The emerging class of RNA interference (RNAi) therapeutics is a fundamentally novel approach to treating human disease by enabling the pursuit of molecular targets considered "undruggable" by small molecules and traditional protein therapeutics. A key challenge toward realizing the full potential of this technology is the safe and efficient delivery of siRNA to target tissues. The physical chemical properties of siRNAs preclude passive diffusion across most cell membranes. For systemic administration, novel delivery systems are required to confer "drug-like" pharmacokinetic and pharmacodynamic properties. Engineered nanomaterials and the emerging field of nanomedicine are important drivers of turning the promise of RNAi therapeutics into reality. The current clinical progress of systemically administered siRNA therapeutics is reviewed, with special attention to the toxicity profiles associated with RNAi nanomedicines. As a case study, the preclinical development of ALN-VSP, the first lipid nanoparticle (LNP)-formulated siRNA therapeutic to be tested in cancer patients, is reviewed to broadly highlight some of the preclinical safety challenges and areas of investigation for "next generation" LNP systems.

Comprehensive evaluation of canonical versus Dicer-substrate siRNA in vitro and in vivo.

Since the discovery of RNA interference (RNAi), researchers have identified a variety of small interfering RNA (siRNA) structures that demonstrate the ability to silence gene expression through the classical RISC-mediated mechanism. One such structure, termed "Dicer-substrate siRNA" (dsiRNA), was proposed to have enhanced potency via RISC-mediated gene silencing, although a comprehensive comparison of canonical siRNAs and dsiRNAs remains to be described. The present study evaluates the in vitro and in vivo activities of siRNAs and dsiRNAs targeting Phosphatase and Tensin Homolog (PTEN) and Factor VII (FVII). More than 250 compounds representing both siRNA and dsiRNA structures were evaluated for silencing efficacy. Lead compounds were assessed for duration of silencing and other key parameters such as cytokine induction. We identified highly active compounds from both canonical siRNAs and 25/27 dsiRNAs. Lead compounds were comparable in potency both in vitro and in vivo as well as duration of silencing in vivo. Duplexes from both structural classes tolerated 2'-OMe chemical modifications well with respect to target silencing, although some modified dsiRNAs demonstrated reduced activity. On the other hand, dsiRNAs were more immunostimulatory as compared with the shorter siRNAs, both in vitro and in vivo. Because the dsiRNA structure does not confer any appreciable benefits in vitro or in vivo while demonstrating specific liabilities, further studies are required to support their applications in RNAi therapeutics.

Systemic RNAi-mediated Gene Silencing in Nonhuman Primate and Rodent Myeloid Cells.

Leukocytes are central regulators of inflammation and the target cells of therapies for key diseases, including autoimmune, cardiovascular, and malignant disorders. Efficient in vivo delivery of small interfering RNA (siRNA) to immune cells could thus enable novel treatment strategies with broad applicability. In this report, we develop systemic delivery methods of siRNA encapsulated in lipid nanoparticles (LNP) for durable and potent in vivo RNA interference (RNAi)-mediated silencing in myeloid cells. This work provides the first demonstration of siRNA-mediated silencing in myeloid cell types of nonhuman primates (NHPs) and establishes the feasibility of targeting multiple gene targets in rodent myeloid cells. The therapeutic potential of these formulations was demonstrated using siRNA targeting tumor necrosis factor-α (TNFα) which induced substantial attenuation of disease progression comparable to a potent antibody treatment in a mouse model of rheumatoid arthritis (RA). In summary, we demonstrate a broadly applicable and therapeutically relevant platform for silencing disease genes in immune cells.

Rational design of cationic lipids for siRNA delivery.

We adopted a rational approach to design cationic lipids for use in formulations to deliver small interfering RNA (siRNA). Starting with the ionizable cationic lipid 1,2-dilinoleyloxy-3-dimethylaminopropane (DLinDMA), a key lipid component of stable nucleic acid lipid particles (SNALP) as a benchmark, we used the proposed in vivo mechanism of action of ionizable cationic lipids to guide the design of DLinDMA-based lipids with superior delivery capacity. The best-performing lipid recovered after screening (DLin-KC2-DMA) was formulated and characterized in SNALP and demonstrated to have in vivo activity at siRNA doses as low as 0.01 mg/kg in rodents and 0.1 mg/kg in nonhuman primates. To our knowledge, this represents a substantial improvement over previous reports of in vivo endogenous hepatic gene silencing.

Activation of protein kinase Czeta increases OAT1 (SLC22A6)- and OAT3 (SLC22A8)-mediated transport.

Organic anion transporters (OATs) play a pivotal role in the clearance of small organic anions by the kidney, yet little is known about how their activity is regulated. A yeast two-hybrid assay was used to identify putative OAT3-associated proteins in the kidney. Atypical protein kinase Czeta (PKCzeta) was shown to bind to OAT3. Binding was confirmed in immunoprecipitation assays. The OAT3/PKCzeta interaction was investigated in rodent renal cortical slices from fasted animals. Insulin, an upstream activator of PKCzeta, increased both OAT3-mediated uptake of estrone sulfate (ES) and PKCzeta activity. Both effects were abolished by a PKCzeta-specific pseudosubstrate inhibitor. Increased ES transport was not observed in renal slices from OAT3-null mice. Transport of the shared OAT1/OAT3 substrate, rho-aminohippurate, behaved similarly, except that stimulation was reduced, not abolished, in the OAT3-null mice. This suggested that OAT1 activity was also modified by PKCzeta, subsequently confirmed using an OAT1-specific substrate, adefovir. Inhibition of PKCzeta also blocked the increase in ES uptake seen in response to epidermal growth factor and to activation of protein kinase A. Thus, PKCzeta acted downstream of the epidermal growth factor to protein kinase A signaling pathway. Activation of transport was accompanied by an increase in V(max) and was blocked by microtubule disruption, indicating that activation may result from trafficking of OAT3 into the plasma membrane. These data demonstrate that PKCzeta activation up-regulates OAT1 and OAT3 function, and that protein-protein interactions play a central role controlling these two important renal drug transporters.

Predictive toxicogenomics in preclinical discovery.

The failure of drug candidates during clinical trials due to toxicity, especially hepatotoxicity, is an important and continuing problem in the pharmaceutical industry. This chapter explores new predictive toxicogenomics approaches to better understand the hepatotoxic potential of human drug candidates and to assess their toxicity earlier in the drug development process. The underlying data consisted of two commercial knowledgebases that employed a hybrid experimental design in which human drug-toxicity information was extracted from the literature, dichotomized, and merged with rat-based gene expression measures (primary isolated hepatocytes and whole liver). Toxicity classification rules were built using a stochastic gradient boosting machine learner, with classification error estimated using a modified bootstrap estimate of true error. Several types of clustering methods were also applied, based on sets of compounds and genes. Robust classification rules were constructed for both in vitro (hepatocytes) and in vivo (liver) data, based on a high-dose, 24-h design. There appeared to be little overlap between the two classifiers, at least in terms of their gene lists. Robust classifiers could not be fitted when earlier time points and/or low-dose data were included, indicating that experimental design is important for these systems. Our results suggest development of a compound screening assay based on these toxicity classifiers appears feasible, with classifier operating characteristics used to tune a screen for a specific implementation within preclinical testing paradigms.

Toxicogenomics strategies for predicting drug toxicity.

INTRODUCTION: The failure of pharmaceutical drug candidates due to toxicity, especially hepatotoxicity, is an important and continuing problem for drug development. The current manuscript explores new toxicogenomics approaches to better understand the hepatotoxic potential of human pharmaceutical compounds and to assess their toxicity earlier in the drug development process by means of a toxicity screen. RESOURCES: Data consisted of two commercial knowledgebases that employed a hybrid experimental design in which human drug toxicity information was extracted from the literature, dichotomized and merged with rat-based gene expression measures. One knowledgebase used gene expression from rat primary hepatocytes while the other employed whole rats. Approximately 100 compounds were used in each. METHODS: Toxicity classification rules were built using a stochastic gradient boosting machine learner, with classification error estimated using a modified bootstrap estimate of true error. Several types of clustering methods were also applied, some based on sets of compounds and others based on sets of genes. RESULTS: Robust classification rules were constructed for both in vitro (hepatocytes) and in vivo (liver) data, based on a high dose, 24-hour design. There appeared to be little overlap between the two classifiers, at least in terms of their gene lists. Robust classifiers could not be fitted when earlier timepoints and/or low dose data were included, indicating that experimental design is important for these systems. CONCLUSIONS: In light of these findings, a working compound screen based on these toxicity classifiers appears feasible, with classifier operating characteristics used to tune a screen for a specific implementation. To ensure robust and optimal performance, issues such as site variability of microarrays and generalizability of findings should be addressed as indicated.

The flounder organic anion transporter fOat has sequence, function, and substrate specificity similarity to both mammalian Oat1 and Oat3.

The flounder renal organic anion transporter (fOat) has substantial sequence homology to mammalian basolateral organic anion transporter orthologs (OAT1/Oat1 and OAT3/Oat3), suggesting that fOat may have functional properties of both mammalian forms. We therefore compared uptake of various substrates by rat Oat1 and Oat3 and human OAT1 and OAT3 with the fOat clone expressed in Xenopus oocytes. These data confirm that estrone sulfate is an excellent substrate for mammalian OAT3/Oat3 transporters but not for OAT1/Oat1 transporters. In contrast, 2,4-dichlorophenoxyacetic acid and adefovir are better transported by mammalian OAT1/Oat1 than by the OAT3/Oat3 clones. All three substrates were well transported by fOat-expressing Xenopus oocytes. fOat K(m) values were comparable to those obtained for mammalian OAT/Oat1/3 clones. We also characterized the ability of these substrates to inhibit uptake of the fluorescent substrate fluorescein in intact teleost proximal tubules isolated from the winter flounder (Pseudopleuronectes americanus) and killifish (Fundulus heteroclitus). The rank order of the IC(50) values for inhibition of cellular fluorescein accumulation was similar to that for the K(m) values obtained in fOat-expressing oocytes, suggesting that fOat may be the primary teleost renal basolateral Oat. Assessment of the zebrafish (Danio rerio) genome indicated the presence of a single Oat (zfOat) with similarity to both mammalian OAT1/Oat1 and OAT3/Oat3. The puffer fish (Takifugu rubripes) also has an Oat (pfOat) similar to mammalian OAT1/Oat1 and OAT3/Oat3 members. Furthermore, phylogenetic analyses argue that the teleost Oat1/3-like genes diverged from a common ancestral gene in advance of the divergence of the mammalian OAT1/Oat1, OAT3/Oat3, and, possibly, Oat6 genes.