An integrative genomics approach to infer causal associations between gene expression and disease.

A key goal of biomedical research is to elucidate the complex network of gene interactions underlying complex traits such as common human diseases. Here we detail a multistep procedure for identifying potential key drivers of complex traits that integrates DNA-variation and gene-expression data with other complex trait data in segregating mouse populations. Ordering gene expression traits relative to one another and relative to other complex traits is achieved by systematically testing whether variations in DNA that lead to variations in relative transcript abundances statistically support an independent, causative or reactive function relative to the complex traits under consideration. We show that this approach can predict transcriptional responses to single gene-perturbation experiments using gene-expression data in the context of a segregating mouse population. We also demonstrate the utility of this approach by identifying and experimentally validating the involvement of three new genes in susceptibility to obesity.

Persistence of seed-based activity following segmentation of a microRNA guide strand.

microRNAs are ∼ 22 nucleotide regulatory RNAs that are processed into duplexes from hairpin structures and incorporated into Argonaute proteins. Here, we show that a nick in the middle of the guide strand of an miRNA sequence allows for seed-based targeting characteristic of miRNA activity. Insertion of an inverted abasic, a dye, or a small gap between the two segments still permits target knockdown. While activity from the seed region of the segmented miRNA is apparent, activity from the 3' half of the guide strand is impaired, suggesting that an intact guide backbone is required for contribution from the 3' half. miRNA activity was also observed following nicking of a miRNA precursor. These results illustrate a structural flexibility in miRNA duplexes and may have applications in the design of miRNA mimetics.

Quantitative evaluation of siRNA delivery in vivo.

Effective small interfering RNA (siRNA)-mediated therapeutics require the siRNA to be delivered into the cellular RNA-induced silencing complex (RISC). Quantitative information of this essential delivery step is currently inferred from the efficacy of gene silencing and siRNA uptake in the tissue. Here we report an approach to directly quantify siRNA in the RISC in rodents and monkey. This is achieved by specific immunoprecipitation of the RISC from tissue lysates and quantification of small RNAs in the immunoprecipitates by stem-loop PCR. The method, expected to be independent of delivery vehicle and target, is label-free, and the throughput is acceptable for preclinical animal studies. We characterized a lipid-formulated siRNA by integrating these approaches and obtained a quantitative perspective on siRNA tissue accumulation, RISC loading, and gene silencing. The described methodologies have utility for the study of silencing mechanism, the development of siRNA therapeutics, and clinical trial design.

Mechanistically probing lipid-siRNA nanoparticle-associated toxicities identifies Jak inhibitors effective in mitigating multifaceted toxic responses.

A major hurdle for harnessing small interfering RNA (siRNA) for therapeutic application is an effective and safe delivery of siRNA to target tissues and cells via systemic administration. While lipid nanoparticles (LNPs) composed of a cationic lipid, poly-(ethylene glycol) lipid and cholesterol, are effective in delivering siRNA to hepatocytes via systemic administration, they may induce multi-faceted toxicities in a dose-dependent manner, independently of target silencing. To understand the underlying mechanism of toxicities, pharmacological probes including anti-inflammation drugs and specific inhibitors blocking different pathways of innate immunity were evaluated for their abilities to mitigate LNP-siRNA-induced toxicities in rodents. Three categories of rescue effects were observed: (i) pretreatment with a Janus kinase (Jak) inhibitor or dexamethasone abrogated LNP-siRNA-mediated lethality and toxicities including cytokine induction, organ impairments, thrombocytopenia and coagulopathy without affecting siRNA-mediated gene silencing; (ii) inhibitors of PI3K, mammalian target of rapamycin (mTOR), p38 and IκB kinase (IKK)1/2 exhibited a partial alleviative effect; (iii) FK506 and etoricoxib displayed no protection. Furthermore, knockout of Jak3, tumor necrosis factor receptors (Tnfr)p55/p75, interleukin 6 (IL-6) or interferon (IFN)-γ alone was insufficient to alleviate LNP-siRNA-associated toxicities in mice. These indicate that activation of innate immune response is a primary trigger of systemic toxicities and that multiple innate immune pathways and cytokines can mediate toxic responses. Jak inhibitors are effective in mitigating LNP-siRNA-induced toxicities.

The siRNA sequence and guide strand overhangs are determinants of in vivo duration of silencing.

The use of short interfering RNAs (siRNA) in animals for target validation or as potential therapeutics is hindered by the short physical half-life when delivered as unencapsulated material and in turn the short active half-life of siRNAs in vivo. Here we demonstrate that the character of the two 3'-overhang nucleotides of the guide strand of siRNAs is a determinant of the duration of silencing by siRNAs both in vivo and in tissue culture cells. We demonstrate that deoxyribonucleotides in the guide strand overhang of siRNAs have a negative impact on maintenance of both the in vitro and in vivo activity of siRNAs over time. Overhangs that contain ribonucleotides or 2'-O-methyl modified nucleotides do not demonstrate this same impairment. We also demonstrate that the sequence of an siRNA is a determinant of the duration of silencing of siRNAs directed against the same target even when those siRNAs have equivalent activities in vitro. Our experiments have determined that a measurable duration parameter exists, distinct from both maximum silencing ability and the potency of siRNAs. Our findings provide information on incorporating chemically modified nucleotides into siRNAs for potent, durable therapeutics and also inform on methods used to select siRNAs for therapeutic and research purposes.

Analysis of acyclic nucleoside modifications in siRNAs finds sensitivity at position 1 that is restored by 5'-terminal phosphorylation both in vitro and in vivo.

Small interfering RNAs (siRNAs) are short, double-stranded RNAs that use the endogenous RNAi pathway to mediate gene silencing. Phosphorylation facilitates loading of a siRNA into the Ago2 complex and subsequent cleavage of the target mRNA. In this study, 2', 3' seco nucleoside modifications, which contain an acylic ribose ring and are commonly called unlocked nucleic acids (UNAs), were evaluated at all positions along the guide strand of a siRNA targeting apolipoprotein B (ApoB). UNA modifications at positions 1, 2 and 3 were detrimental to siRNA activity. UNAs at positions 1 and 2 prevented phosphorylation by Clp1 kinase, abrogated binding to Ago2, and impaired Ago2-mediated cleavage of the mRNA target. The addition of a 5'-terminal phosphate to siRNA containing a position 1 UNA restored ApoB mRNA silencing, Ago2 binding, and Ago2 mediated cleavage activity. Position 1 UNA modified siRNA containing a 5'-terminal phosphate exhibited a partial restoration of siRNA silencing activity in vivo. These data reveal the complexity of interpreting the effects of chemical modification on siRNA activity, and exemplify the importance of using multiple biochemical, cell-based and in vivo assays to rationally design chemically modified siRNA destined for therapeutic use.

Improved efficacy for ezetimibe and rosuvastatin by attenuating the induction of PCSK9.

Reducing circulating LDL-cholesterol (LDL-c) reduces the risk of cardiovascular disease in people with hypercholesterolemia. Current approaches to reduce circulating LDL-c include statins, which inhibit cholesterol synthesis, and ezetimibe, which blocks cholesterol absorption. Both elevate serum PCSK9 protein levels in patients, which could attenuate their efficacy by reducing the amount of cholesterol cleared from circulation. To determine whether PCSK9 inhibition could enhance LDL-c lowering of both statins and ezetimibe, we utilized small interfering RNAs (siRNAs) to knock down Pcsk9, together with ezetimibe, rosuvastatin, and an ezetimibe/rosuvastatin combination in a mouse model with a human-like lipid profile. We found that ezetimibe, rosuvastatin, and ezetimibe/rosuvastatin combined lower serum cholesterol but induce the expression of Pcsk9 as well as the Srebp-2 hepatic cholesterol biosynthesis pathway. Pcsk9 knockdown in combination with either treatment led to greater reductions in serum non-HDL with a near-uniform reduction of all LDL-c subfractions. In addition to reducing serum cholesterol, the combined rosuvastatin/ezetimibe/Pcsk9 siRNA treatment exhibited a significant reduction in serum APOB protein and triglyceride levels. Taken together, these data provide evidence that PCSK9 inhibitors, in combination with current therapies, have the potential to achieve greater reductions in both serum cholesterol and triglycerides.

siRNA-induced liver ApoB knockdown lowers serum LDL-cholesterol in a mouse model with human-like serum lipids.

Increased serum apolipoprotein (apo)B and associated LDL levels are well-correlated with an increased risk of coronary disease. ApoE⁻/⁻ and low density lipoprotein receptor (LDLr)⁻/⁻ mice have been extensively used for studies of coronary atherosclerosis. These animals show atherosclerotic lesions similar to those in humans, but their serum lipids are low in apoB-containing LDL particles. We describe the development of a new mouse model with a human-like lipid profile. Ldlr CETP⁺/⁻ hemizygous mice carry a single copy of the human CETP transgene and a single copy of a LDL receptor mutation. To evaluate the apoB pathways in this mouse model, we used novel short-interfering RNAs (siRNA) formulated in lipid nanoparticles (LNP). ApoB siRNAs induced up to 95% reduction of liver ApoB mRNA and serum apoB protein, and a significant lowering of serum LDL in Ldlr CETP⁺/⁻ mice. ApoB targeting is specific and dose-dependent, and it shows lipid-lowering effects for over three weeks. Although specific triglycerides (TG) were affected by ApoB mRNA knockdown (KD) and the total plasma lipid levels were decreased by 70%, the overall lipid distribution did not change. Results presented here demonstrate a new mouse model for investigating additional targets within the ApoB pathways using the siRNA modality.

RNA-induced silencing complex-bound small interfering RNA is a determinant of RNA interference-mediated gene silencing in mice.

Deeper knowledge of pharmacokinetic and pharmacodynamic (PK/PD) concepts for RNA therapeutics is important to streamline the drug development process and for rigorous selection of best performing drug candidates. Here we characterized the PK/PD relationship for small interfering RNAs (siRNAs) targeting luciferase by examining siRNA concentration in plasma and liver, the temporal RNA-induced silencing complex binding profiles, mRNA reduction, and protein inhibition measured by noninvasive bioluminescent imaging. A dose-dependent and time-related decrease in bioluminescence was detected over 25 days after a single treatment of a lipid nanoparticle-formulated siRNA targeting luciferase messenger RNA. A direct relationship was observed between the degree of in vivo mRNA and protein reduction and the Argonaute2 (Ago2)-bound siRNA fraction but not with the total amount of siRNA found in the liver, suggesting that the Ago2-siRNA complex is the key determinant of target inhibition. These observations were confirmed for an additional siRNA that targets endogenously expressed Sjögren syndrome antigen B (Ssb) mRNA, indicating that our observations are not limited to a transgenic mouse system. Our data provide detailed information of the temporal regulation of siRNA liver delivery, Ago2 loading, mRNA reduction, and protein inhibition that are essential for the rapid and cost-effective clinical development of siRNAs therapeutics.

MicroRNA-like off-target transcript regulation by siRNAs is species specific.

siRNAs mediate sequence-specific gene silencing in cultured mammalian cells but also silence unintended transcripts. Many siRNA off-target transcripts match the guide-strand "seed region," similar to the way microRNAs match their target sites. The extent to which this seed-matched, microRNA-like, off-target silencing affects the specificity of therapeutic siRNAs in vivo is currently unknown. Here, we compare microRNA-like off-target regulations in mouse liver in vivo with those seen in cell culture for a series of therapeutic candidate siRNAs targeting Apolipoprotein B (APOB). Each siRNA triggered regulation of consistent microRNA-like off-target transcripts in mouse livers and in cultured mouse liver tumor cells. In contrast, there was only random overlap between microRNA-like off-target transcripts from cultured human and mouse liver tumor cells. Therefore, siRNA therapeutics may trigger microRNA-like silencing of many unintended targets in vivo, and the potential toxicities caused by these off-target gene regulations cannot be accurately assessed in rodent models.

Evaluation of efficacy, biodistribution, and inflammation for a potent siRNA nanoparticle: effect of dexamethasone co-treatment.

Despite recent progress, systemic delivery remains the major hurdle for development of safe and effective small inhibitory RNA (siRNA)-based therapeutics. Encapsulation of siRNA into liposomes is a promising option to overcome obstacles such as low stability in serum and inefficient internalization by target cells. However, a major liability of liposomes is the potential to induce an acute inflammatory response, thereby increasing the risk of numerous adverse effects. In this study, we characterized a liposomal siRNA delivery vehicle, LNP201, which is capable of silencing an mRNA target in mouse liver by over 80%. The biodistribution profile, efficacy after single and multiple doses, mechanism of action, and inflammatory toxicity are characterized for LNP201. Furthermore, we demonstrate that the glucocorticoid receptor (GR) agonist dexamethasone (Dex) inhibits LNP201-induced cytokine release, inflammatory gene induction, and mitogen-activated protein kinase (MAPK) phosphorylation in multiple tissues. These data present a possible clinical strategy for increasing the safety profile of siRNA-based drugs while maintaining the potency of gene silencing.

Application of an end-to-end biomarker discovery platform to identify target engagement markers in cerebrospinal fluid by high resolution differential mass spectrometry.

The rapid identification of protein biomarkers in biofluids is important to drug discovery and development. Here, we describe a general proteomic approach for the discovery and identification of proteins that exhibit a statistically significant difference in abundance in cerebrospinal fluid (CSF) before and after pharmacological intervention. This approach, differential mass spectrometry (dMS), is based on the analysis of full scan mass spectrometry data. The dMS workflow does not require complex mixing and pooling strategies, or isotope labeling techniques. Accordingly, clinical samples can be analyzed individually, allowing the use of longitudinal designs and within-subject data analysis in which each subject acts as its own control. As a proof of concept, we performed multifactorial dMS analyses on CSF samples drawn at 6 time points from n = 6 cisterna magna ported (CMP) rhesus monkeys treated with 2 potent gamma secretase inhibitors (GSI) or comparable vehicle in a 3-way crossover study that included a total of 108 individual CSF samples. Using analysis of variance and statistical filtering on the aligned and normalized LC-MS data sets, we detected 26 features that were significantly altered in CSF by drug treatment. Of those 26 features, which belong to 10 distinct isotopic distributions, 20 were identified by MS/MS as 7 peptides from CD99, a cell surface protein. Six features from the remaining 3 isotopic distributions were not identified. A subsequent analysis showed that the relative abundance of these 26 features showed the same temporal profile as the ELISA measured levels of CSF A beta 42 peptide, a known pharmacodynamic marker for gamma-secretase inhibition. These data demonstrate that dMS is a promising approach for the discovery, quantification, and identification of candidate target engagement biomarkers in CSF.

A cell proliferation signature is a marker of extremely poor outcome in a subpopulation of breast cancer patients.

Breast cancer comprises a group of distinct subtypes that despite having similar histologic appearances, have very different metastatic potentials. Being able to identify the biological driving force, even for a subset of patients, is crucially important given the large population of women diagnosed with breast cancer. Here, we show that within a subset of patients characterized by relatively high estrogen receptor expression for their age, the occurrence of metastases is strongly predicted by a homogeneous gene expression pattern almost entirely consisting of cell cycle genes (5-year odds ratio of metastasis, 24.0; 95% confidence interval, 6.0-95.5). Overexpression of this set of genes is clearly associated with an extremely poor outcome, with the 10-year metastasis-free probability being only 24% for the poor group, compared with 85% for the good group. In contrast, this gene expression pattern is much less correlated with the outcome in other patient subpopulations. The methods described here also illustrate the value of combining clinical variables, biological insight, and machine-learning to dissect biological complexity. Our work presented here may contribute a crucial step towards rational design of personalized treatment.

Embracing complexity, inching closer to reality.

Drugs designed against targets in presumably simple linear signaling pathways found to be associated with disease are often less effective than predicted. One reason for this is the overly simplistic view of the molecular mechanisms underlying common human diseases. This viewpoint is a consequence of biological reductionism, brought about by the need to form a basic understanding of the fundamental attributes of biological systems and by limitations in the set of tools available for analysis of biological systems. However, complex biological systems are best modeled as highly modular, fluid systems exhibiting a plasticity that allows them to adapt to a vast array of conditions. Historically, this viewpoint has long represented the ideal, but the tools needed to examine and describe this complexity were often lacking. Here we argue that the tools of biological science now allow for a more network-oriented view of biological systems and for explaining the underlying causes of disease, as well as the best ways to target disease. Ultimately, this will help to ensure that the right drug is administered to the right patient at the right time. Focusing on well-studied signaling pathways, refining the definition of disease, and identifying disease subtypes, we demonstrate a more holistic approach to elucidating common human diseases, with the potential to revolutionize treatment of these diseases.

Skin biopsies for the measurement of clinical pharmacodynamic biomarkers.

The integration of biomarkers reporting on drug pharmacodynamics or on a disease state is becoming necessary in the cost-effective design of clinical trials. Transcriptional changes in skin measured using whole-genome arrays have been useful in assessing the disease state in dermatology. Not only are skin biopsies well-tolerated and easy to obtain, but they sample a system in which many complex signaling and developmental networks are active and in which a wide variety of drug targets may be expected to play a role, either directly or indirectly. Recent advances have led to the use of transcriptional profiling of skin for making pharmacodynamic measurements in clinical trials, even for non-dermatological conditions.

High Resolution Discovery Proteomics Reveals Candidate Disease Progression Markers of Alzheimer's Disease in Human Cerebrospinal Fluid.

Disease modifying treatments for Alzheimer's disease (AD) constitute a major goal in medicine. Current trends suggest that biomarkers reflective of AD neuropathology and modifiable by treatment would provide supportive evidence for disease modification. Nevertheless, a lack of quantitative tools to assess disease modifying treatment effects remains a major hurdle. Cerebrospinal fluid (CSF) biochemical markers such as total tau, p-tau and Ab42 are well established markers of AD; however, global quantitative biochemical changes in CSF in AD disease progression remain largely uncharacterized. Here we applied a high resolution open discovery platform, dMS, to profile a cross-sectional cohort of lumbar CSF from post-mortem diagnosed AD patients versus those from non-AD/non-demented (control) patients. Multiple markers were identified to be statistically significant in the cohort tested. We selected two markers SME-1 (p<0.0001) and SME-2 (p = 0.0004) for evaluation in a second independent longitudinal cohort of human CSF from post-mortem diagnosed AD patients and age-matched and case-matched control patients. In cohort-2, SME-1, identified as neuronal secretory protein VGF, and SME-2, identified as neuronal pentraxin receptor-1 (NPTXR), in AD were 21% (p = 0.039) and 17% (p = 0.026) lower, at baseline, respectively, than in controls. Linear mixed model analysis in the longitudinal cohort estimate a decrease in the levels of VGF and NPTXR at the rate of 10.9% and 6.9% per year in the AD patients, whereas both markers increased in controls. Because these markers are detected by mass spectrometry without the need for antibody reagents, targeted MS based assays provide a clear translation path for evaluating selected AD disease-progression markers with high analytical precision in the clinic.

Biodistribution of small interfering RNA at the organ and cellular levels after lipid nanoparticle-mediated delivery.

Chemically stabilized small interfering RNA (siRNA) can be delivered systemically by intravenous injection of lipid nanoparticles (LNPs) in rodents and primates. The biodistribution and kinetics of LNP-siRNA delivery in mice at organ and cellular resolution have been studied using immunofluorescence (IF) staining and quantitative polymerase chain reaction (qPCR). At 0.5 and 2 hr post tail vein injection of Cy5-labeled siRNA encapsulated in LNP, the organ rank-order of siRNA levels is liver > spleen > kidney, with only negligible accumulation in duodenum, lung, heart, and brain. Similar conclusions were drawn by using qPCR to measure tissue siRNA levels as a secondary end point. siRNA levels in these tissues decreased by more than 10-fold after 24 hr. Within the liver, LNPs delivered siRNA to hepatocytes, Kupffer cells, and sinusoids in a time-dependent manner, as revealed by IF staining and signal quantitation methods established using OPERA/Columbus software. siRNA first accumulated in liver sinusoids and trafficked to hepatocytes by 2 hr post dose, corresponding to the onset of target mRNA silencing. Fluorescence in situ hybridization methods were used to detect both strands of siRNA in fixed tissues. Collectively, the authors have implemented a platform to evaluate biodistribution of siRNA across cell types and across tissues in vivo, with the objective of elucidating the pharmacokinetic and pharmacodynamic relationship to guide optimization of delivery vehicles.

ApoB siRNA-induced liver steatosis is resistant to clearance by the loss of fatty acid transport protein 5 (Fatp5).

The association between hypercholesterolemia and elevated serum apolipoprotein B (APOB) has generated interest in APOB as a therapeutic target for patients at risk of developing cardiovascular disease. In the clinic, mipomersen, an antisense oligonucleotide (ASO) APOB inhibitor, was associated with a trend toward increased hepatic triglycerides, and liver steatosis remains a concern. We found that siRNA-mediated knockdown of ApoB led to elevated hepatic triglycerides and liver steatosis in mice engineered to exhibit a human-like lipid profile. Many genes required for fatty acid synthesis were reduced, suggesting that the observed elevation in hepatic triglycerides is maintained by the cell through fatty acid uptake as opposed to fatty acid synthesis. Fatty acid transport protein 5 (Fatp5/Slc27a5) is required for long chain fatty acid (LCFA) uptake and bile acid reconjugation by the liver. Fatp5 knockout mice exhibited lower levels of hepatic triglycerides due to decreased fatty acid uptake, and shRNA-mediated knockdown of Fatp5 protected mice from diet-induced liver steatosis. Here, we evaluated if siRNA-mediated knockdown of Fatp5 was sufficient to alleviate ApoB knockdown-induced steatosis. We determined that, although Fatp5 siRNA treatment was sufficient to increase the proportion of unconjugated bile acids 100-fold, consistent with FATP5's role in bile acid reconjugation, Fatp5 knockdown failed to influence the degree, zonal distribution, or composition of the hepatic triglycerides that accumulated following ApoB siRNA treatment.

LRRTM3 promotes processing of amyloid-precursor protein by BACE1 and is a positional candidate gene for late-onset Alzheimer's disease.

Rare familial forms of Alzheimer's disease (AD) are thought to be caused by elevated proteolytic production of the Abeta42 peptide from the beta-amyloid-precursor protein (APP). Although the pathogenesis of the more common late-onset AD (LOAD) is not understood, BACE1, the protease that cleaves APP to generate the N terminus of Abeta42, is more active in patients with LOAD, suggesting that increased amyloid production processing might also contribute to the sporadic disease. Using high-throughput siRNA screening technology, we assessed 15,200 genes for their role in Abeta42 secretion and identified leucine-rich repeat transmembrane 3 (LRRTM3) as a neuronal gene that promotes APP processing by BACE1. siRNAs targeting LRRTM3 inhibit the secretion of Abeta40, Abeta42, and sAPPbeta, the N-terminal APP fragment produced by BACE1 cleavage, from cultured cells and primary neurons by up to 60%, whereas overexpression increases Abeta secretion. LRRTM3 is expressed nearly exclusively in the nervous system, including regions affected during AD, such as the dentate gyrus. Furthermore, LRRTM3 maps to a region of chromosome 10 linked to both LOAD and elevated plasma Abeta42, and is structurally similar to a family of neuronal receptors that includes the NOGO receptor, an inhibitor of neuronal regeneration and APP processing. Thus, LRRTM3 is a functional and positional candidate gene for AD, and, given its receptor-like structure and restricted expression, a potential therapeutic target.

Sequence variation in NPC1L1 and association with improved LDL-cholesterol lowering in response to ezetimibe treatment.

Niemann-Pick C1-like 1 (NPC1L1) is an intestinal cholesterol transporter and the molecular target of ezetimibe, a cholesterol absorption inhibitor demonstrated to reduce LDL-cholesterol (LDL-C) both as monotherapy and when co-administered with 3-hydroxy-3-methylglutaryl coenzyme A reductase inhibitors (statins). Interestingly, significant interindividual variability has been observed for rates of intestinal cholesterol absorption and LDL-C reductions at both baseline and post ezetimibe treatment. To test the hypothesis that genetic variation in NPC1L1 could influence the LDL-C response to ezetimibe, we performed extensive resequencing of the gene in 375 apparently healthy individuals and genotyped hypercholesterolemic patients from clinical trial cohorts. No association was observed between NPC1L1 single-nucleotide polymorphism and baseline cholesterol. However, significant associations to LDL-C response to treatment with ezetimibe were observed in patients treated with ezetimibe in two large clinical trials. Our data demonstrate that DNA sequence variants in NPC1L1 are associated with an improvement in response to ezetimibe pharmacotherapy and suggest that detailed analysis of genetic variability in clinical trial cohorts can lead to improved understanding of factors contributing to variable drug response.