MEDICALSCIENCES
`
`Therapeutic RNAi targeting PCSK9 acutely lowers
`plasma cholesterol in rodents and LDL cholesterol
`in nonhuman primates
`
`Maria Frank-Kamenetsky*, Aldo Grefhorst†, Norma N. Anderson†, Timothy S. Racie*, Birgit Bramlage‡, Akin Akinc*,
`David Butler*, Klaus Charisse*, Robert Dorkin*, Yupeng Fan*, Christina Gamba-Vitalo*, Philipp Hadwiger‡,
`Muthusamy Jayaraman*, Matthias John‡, K. Narayanannair Jayaprakash*, Martin Maier*, Lubomir Nechev*,
`Kallanthottathil G. Rajeev*, Timothy Read*, Ingo Ro¨ hl‡, Ju¨ rgen Soutschek*, Pamela Tan‡, Jamie Wong*, Gang Wang*,
`Tracy Zimmermann*, Antonin de Fougerolles*, Hans-Peter Vornlocher‡, Robert Langer§¶, Daniel G. Anderson¶,
`Muthiah Manoharan*, Victor Koteliansky*, Jay D. Horton†储, and Kevin Fitzgerald*§
`*Alnylam Pharmaceuticals, 300 Third Street, Cambridge, MA 02142 ; Departments of †Molecular Genetics and 储Internal Medicine, University of Texas
`Southwestern Medical Center, 5323 Harry Hines Boulevard, Dallas, TX 75390; ‡Roche Kulmbach GmbH, Fritz-Hornschuch-Strasse 9, 95326 Kulmbach,
`Germany; and ¶David H. Koch Institute for Integrative Cancer Research, Massachusetts Institute of Technology, Cambridge, MA 02139
`
`Contributed by Robert Langer, June 6, 2008 (sent for review May 7, 2008)
`
`Proprotein convertase subtilisin/kexin type 9 (PCSK9) regulates
`low density lipoprotein receptor (LDLR) protein levels and function.
`Loss of PCSK9 increases LDLR levels in liver and reduces plasma LDL
`cholesterol (LDLc), whereas excess PCSK9 activity decreases liver
`LDLR levels and increases plasma LDLc. Here, we have developed
`active, cross-species, small interfering RNAs (siRNAs) capable of
`targeting murine, rat, nonhuman primate (NHP), and human
`PCSK9. For in vivo studies, PCSK9 and control siRNAs were formu-
`lated in a lipidoid nanoparticle (LNP). Liver-specific siRNA silencing
`of PCSK9 in mice and rats reduced PCSK9 mRNA levels by 50 –70%.
`The reduction in PCSK9 transcript was associated with up to a 60%
`reduction in plasma cholesterol concentrations. These effects were
`shown to be mediated by an RNAi mechanism, using 5ⴕ-RACE. In
`transgenic mice expressing human PCSK9, siRNAs silenced the
`human PCSK9 transcript by >70% and significantly reduced PCSK9
`plasma protein levels. In NHP, a single dose of siRNA targeting
`PCSK9 resulted in a rapid, durable, and reversible lowering of
`plasma PCSK9, apolipoprotein B, and LDLc, without measurable
`effects on either HDL cholesterol (HDLc) or triglycerides (TGs). The
`effects of PCSK9 silencing lasted for 3 weeks after a single bolus i.v.
`administration. These results validate PCSK9 targeting with RNAi
`therapeutics as an approach to specifically lower LDLc, paving the
`way for the development of PCSK9-lowering agents as a future
`strategy for treatment of hypercholesterolemia.
`
`plasma PCSK9 兩 tissue LDLR levels
`
`Proprotein convertase subtilisin/kexin type 9 (PCSK9) is a
`
`member of the mammalian serine proprotein convertase
`family that typically functions in the proteolytic processing and
`maturation of secretory proteins (1, 2). PCSK9 was the first
`family member to be implicated in a dominantly inherited form
`of hypercholesterolemia (3). Mechanistic studies addressing the
`function of PCSK9 in mice and humans have demonstrated that
`overexpression or gain-of-function mutations in PCSK9 reduced
`low density lipoprotein receptor (LDLR) protein levels in liver,
`which significantly increased circulating plasma cholesterol both
`in mice and humans (4). Additional studies showed that the
`deletion of Pcsk9 in mice resulted in increased LDLR levels,
`accelerated the clearance of low density lipoprotein cholesterol
`(LDLc), and reduced circulating cholesterol levels (5). Recently,
`studies in mice have also shown that lowering PCSK9 transcript
`levels by antisense oligonucleotides resulted in reduced total
`cholesterol, LDLc, and HDL cholesterol (HDLc) in blood and
`increased LDLR levels in liver after 6 weeks of treatment (6).
`This effect was very similar to that observed in the Pcsk9⫺/⫺ mice
`
`(5). Collectively, these studies have clearly established a role for
`PCSK9 in cholesterol homeostasis.
`Validation of PCSK9 as an attractive therapeutic target for the
`treatment of hypercholesterolemia has come from genetic stud-
`ies in humans. Cohen et al. (7) first identified loss-of-function
`mutations in PCSK9 that lowered plasma LDLc in the Dallas
`Heart Study. In a larger 15-year prospective study, they dem-
`onstrated that nonsense mutations in PCSK9 reduced LDLc
`levels by 28% and decreased the frequency of CHD by 88% in
`African Americans (8). Despite this genetic validation, several
`physiological aspects of potential PCSK9-modifying agents must
`be further defined to assess therapeutic potential and benefit.
`For instance, will the acute lowering of PCSK9 (e.g., over 48–72
`h) result in LDLc lowering, and if so, will this reduction be
`associated with other potentially adverse consequences, such as
`increased liver lipids? Rodents lack cholesterol ester transferase
`protein (CETP) and carry the majority of their plasma choles-
`terol in HDL. Thus, they are not ideal models in which to
`determine whether PCSK9 silencing will only decrease LDLc
`and not HDLc. Studies in a more relevant model, such as
`nonhuman primates (NHPs), are required.
`Currently, a number of individuals with hypercholesterolemia are
`unable to reach target LDLc levels with available therapies. To
`address the efficacy of inhibiting PCSK9 via an siRNA mechanism,
`we designed and synthesized several siRNA therapeutic molecules
`to silence PCSK9 mRNA in mice, rats, NHPs, and humans. These
`siRNAs were administered by using a lipidoid nanoparticle (LNP)
`to achieve efficient hepatocyte delivery in vivo. This approach
`enabled us to study the effect of PCSK9 silencing on the levels of
`PCSK9 mRNA, plasma PCSK9 protein, hepatic LDLR protein,
`
`Author contributions: M.F.-K., H.-P.V., M. Manoharan, V.K., J.D.H., and K.F. designed the
`research; M.F.-K., A.G., N.N.A., T.S.R., B.B., M. John, J.W., and K.F. performed the research;
`A.A., D.B., K.C., R.D., Y.F., P.H., M. Jayaraman, K.N.J., M. Maier, L.N., K.G.R., T.R., I.R., J.S.,
`P.T., G.W., T.Z., A.d.F., R.L., and D.G.A. contributed new reagents/analytical tools; M.F.-K.,
`C.G.-V., J.D.H., and K.F. analyzed the data; and M.F.-K., J.D.H., and K.F. wrote the paper.
`
`Conflict of interest statement: R.L. is a shareholder and member of the Scientific Advisory
`Board of Alnylam. D.G.A. and J.D.H. are consultants of Alynylam Pharmaceuticals. Alnylam
`also has a license to certain intellectual property invented at Massachusetts Institute of
`Technology by Drs. Anderson, Langer, and colleagues. M.F.-K., T.S.R., A.A., D.B., K.C., R.D.,
`Y.F., C.G.-V., M. Jayaraman, K.N.J., M. Maier, L.N., K.G.R., T.R., J.S., J.W., G.W., T.Z., A.d.F.,
`M. Manoharan, V.K., and K.F. are employees of Alnylam Pharmaceuticals. B.B., P.H., M.
`John, I.R., P.T., and H.-P.V. are employees of Roche Kulmbach.
`
`Freely available online through the PNAS open access option.
`§To whom correspondence may be addressed. E-mail: kfitzgerald@alnylam.com or
`rlanger@mit.edu.
`
`This article contains supporting information online at www.pnas.org/cgi/content/full/
`0805434105/DCSupplemental.
`
`© 2008 by The National Academy of Sciences of the USA
`
`www.pnas.org兾cgi兾doi兾10.1073兾pnas.0805434105
`
`PNAS 兩 August 19, 2008 兩 vol. 105 兩 no. 33 兩 11915–11920
`
`Alnylam Exh. 1073
`
`

`

`1.2
`1.0
`0.8
`0.6
`0.4
`0.2
`0
`
`relative to PBS
`
`Liver PCSK9 mRNA
`
`B
`
`H
`N
`
`O
`
`NH
`
`O
`
`N
`
`NH
`
`N
`
`O
`
`NH
`
`N
`
`O
`
`NH
`
`O
`
`NH
`
`Lipidoid 98N12-5
`
`A
`
`1.2
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0
`
`D
`
`relative to PBS
`
`Liver PCSK9 mRNA
`
`PBS
`LNP-PCS-A2
`
`***
`
`***
`
`PCSK9
`
`Tc
`
`1.4
`
`1.2
`1.0
`0.8
`
`0.6
`0.4
`0.2
`0
`
`relative to PBS
`
`mRNA and serum Tc levels,
`
`C
`
` Liver PCSK9
`
`**
`
`***
`
`PBS
`
`5
`LNP-Ctrl
`(mg/kg)
`
`5
`3
`1.5
`LNP-PCS-A2 (mg/kg)
`
`LNP-Ctrl
`
`5 mg/kg
`LNP-PCS-A2
`
`7.5 mg/kg
`LNP-PCS-A2
`
`***
`
`***
`
`***
`
`***
`
`**
`
`**
`
`Day 6
`
`Day 12 Day 16 Day 20 Day 26
`Day Post-injection
`
`Lipidoid formulation and effects of PCSK9 silencing in wild-type mice. (A) Cationic lipidoid component structure of the formulation. (B) Dose-dependent
`Fig. 1.
`decrease in hepatic PCSK9 mRNA (relative to controls) 2 days after dose (n ⫽ 6 per group). (C) Liver PCSK9 mRNA and total serum cholesterol levels in mice (n ⫽
`5 per group) 3 days after a dose of 5 mg/kg LNP-PCS-A2 or PBS. (D) Duration of hepatic PCSK9 transcript silencing in mice (n ⫽ 5 per group) after a single injection
`of 5 mg/kg or 7.5 mg/kg LNP-PCS-A2. (B) one-way ANOVA with Student’s t test; (C and D) two-way ANOVA with Bonferroni test. (B–D) Each value is the group
`mean ⫾ STDEV. Asterisks represent statistical difference between PBS and PCSK9 siRNA treated groups. **, P ⱕ 0.01; ***, P ⱕ 0.001
`
`total serum cholesterol, LDLc, and HDLc concentrations in mul-
`tiple species. These in vivo studies demonstrate that PCSK9 low-
`ering by siRNA has an acute effect on plasma LDLc, but not HDLc,
`in NHPs. Our data validate PCSK9 as a target for therapeutic
`intervention by siRNA and provide a strategy for treatment of
`hypercholesterolemia.
`
`Results
`Selection and Formulation of Active siRNA Molecules Targeting PCSK9.
`A series of approximately 150 siRNAs were designed to be
`cross-species reactive through an initial bioinformatics analysis
`and screened in vitro for activity in cultured HepG2 cells. Active
`molecules PCS-A1, PCS-A2, PCS-B2, and PCS-C2 were chosen
`for further studies based on their pM IC50 values as measured in
`primary cynomolgus monkey hepatocytes [supporting informa-
`tion (SI) Table S1].
`Certain siRNAs can induce immune responses via interferons
`and proinflammatory cytokines (9, 10). The siRNAs studied
`here were designed to avoid immune stimulatory sequence
`motifs. The siRNAs selected for further study contained two
`nucleotide 3⬘ overhangs to prevent activation of the RIG-1
`pathway (11, 12). Nevertheless, the selected siRNAs were also
`tested for activation of the immune system in primary human
`blood monocytes (hPBMCs). Specifically, IFN-␣ and TNF-␣
`were measured in hPBMCs transfected with each molecule listed
`in Table S1. The parental compound PCS-A1 was found to
`induce both IFN-␣and TNF-␣. However, its chemically modified
`version, PCS-A2, and chemically modified duplexes PCS-B2 and
`PCS-C2, were negative for both IFN-␣ and TNF-␣ induction in
`these assays (Table S1 and Fig. S1 for a PCS-A1/PCS-A2 paired
`example). These results demonstrate that chemical modifica-
`tions are capable of attenuating both IFN-␣ and TNF-␣ re-
`sponses to siRNA molecules.
`LNP is a lipidoid formulation comprised of a novel cationic
`component 98N12-5 (1)䡠4HCl (Fig. 1A), cholesterol, and a poly-
`(ethylene glycol)-lipid (13). We have shown that LNP-
`formulated siRNAs that target apoB or Factor VII mediated in
`vivo silencing in liver at doses of 5 mg/kg with minimal toxicity
`and without perturbation of the endogenous miRNA biogenesis
`
`pathway (12). Here, we use LNP to formulate and test our RNAi
`therapeutics against PCSK9 in mice, rats, and NHPs.
`
`Silencing of Hepatic PCSK9 mRNA in Rodents Results in Rapid and
`Reversible Lowering of Serum Cholesterol. Pcsk9⫺/⫺ mice have
`⬇50% reduction in total serum cholesterol concentrations (5).
`To test whether acute silencing of the PCSK9 transcript by a
`PCSK9-specific siRNA would result in acutely lower serum
`cholesterol levels, we formulated a cross-species siRNA mole-
`cule PCS-A2 in LNP for study in vivo. Various doses of the
`lipidoid-formulated PCS-A2 (LNP-PCS-A2) were injected via
`the tail vein into mice and rats.
`In mice, livers were harvested to measure PCSK9 mRNA
`levels and blood was collected for total cholesterol analysis. As
`shown in Fig. 1B, LNP-PCS-A2 displayed a dose response with
`maximal PCSK9 mRNA silencing of ⬇60–70% at a dose of 5
`mg/kg. The decrease of mRNA transcript levels (at the highest
`dose) translated into ⬇30% lowering of total plasma cholesterol
`(Fig. 1C). The reduction in serum cholesterol was similar to that
`measured in mice heterozygous for a disrupted Pcsk9 allele (5
`and J.D.H., unpublished observations). Moreover, the effect on
`the PCSK9 transcript persisted for ⬇20 days, with higher doses
`displaying greater initial transcript level reduction and subse-
`quently more prolonged effects (Fig. 1D).
`Next, we studied rats, which are resistant to cholesterol
`lowering by high doses of HMG-CoA reductase inhibitors (st-
`atins) (14, 15). In rats, LNP-PCS-A2 was dosed 1–5 mg/kg, which
`resulted in a dose-dependent reduction in the PCSK9 transcript
`with 50–60% silencing at the highest dose (Fig. 2A). The mRNA
`silencing was associated with an acute 50–60% decrease of
`serum total cholesterol (Fig. 2 A and B) lasting 10 days, with a
`gradual return to predose levels by ⬇3 weeks (Fig. 2B)
`Lowering of proteins involved in very-low-density lipoprotein
`assembly and secretion (microsomal triglyceride transfer pro-
`tein; MTP or apoB) by genetic deletion, small molecule inhib-
`itors, or siRNA, results in increased liver TGs (16, 17) (T.Z.,
`unpublished data). To determine whether cholesterol lowering
`via PCSK9 inhibition alters liver lipid content, hepatic choles-
`terol and TG concentrations in livers of treated and control
`animals were quantified. As shown in Fig. 2C, there was no
`
`11916 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0805434105
`
`Frank-Kamenetsky et al.
`
`

`

`LNP-PCS-A2
`LNP-Ctrl
`
`1.2
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
` Serum Tc level †
`
`B
`
`PCSK9 mRNA
`Tc
`
`*
`
`* *
`
`** **
`
`PBS
`
`5
`LNP-Ctrl
`(mg/kg)
`
`1
`
`2.5
`LNP-PCS-A2
`(mg/kg)
`
`5
`
`0
`
`0
`
`2
`
`4
`
`6
`
`8 10 12 14 16 18 20 22 24 26 28
`Days post-injection
`
`Hepatic Cholesterol
`Hepatic TG
`
`PBS
`
`5
`LNP-Ctrl
`(mg/kg)
`
`1
`
`2.5
`LNP-PCS-A2
`(mg/kg)
`
`5
`
`PBS
`
`LNP-Ctrl
`
`LNP-PCS-A2
`
`PBS
`
`D
`
`LDLR
`
`TFR
`
`Lane 1 2 3
`
`4 5 6 7 8 9 10 11 12 13 14 15
`
`16 17 18
`
`1.6
`
`1.2
`
`0.8
`
`0.4
`
`0
`
`and serum Tc levels †
`Liver PCSK9 mRNA
`
`A
`
`2.0
`
`1.6
`
`1.2
`
`0.8
`
`0.4
`
`0
`
`C
`
`and TG levels †
`
`Liver cholesterol
`
`Fig. 2. Hepatic PCSK9 silencing, hepatic TGs, and LDLR levels in rats. (A) LNP-PCS-A2 mediated dose-dependent lowering of hepatic PCKS9 mRNA and total serum
`cholesterol 3 days after dose (n ⫽ 6 per group). Each value is the group mean ⫾ STDEV. One-way ANOVA with Student’s t test. Asterisks represent statistical
`difference between PBS- and PCSK9 siRNA-treated groups *, P ⱕ 0.05; **, P ⱕ 0.01. (B) Total serum cholesterol lowering of LNP-PCS-A2 treated rats (n ⫽ 6 per
`group) is maximal (⬇60%) by 2 days after dose and returns to baseline over ⬇21 days. PCSK9-treated groups are statistically significant (until approximately day
`16) compared with PBS and LNP-Crtl groups (one-way ANOVA, Student’s t test, with P values of ⱕ 0.05). (C) Liver TGs and cholesterol contents from treated and
`control animals (same as in A). There were no significant differences in the liver TGs (ANOVA: P ⫽ 0.824 for cholesterol content ANOVA on ranks; P ⫽ 0.935 for
`LNP-PCS-A2-treated animals vs. the LNP-Crtl- or PBS-treated control groups for TG levels). (D) Immunoblot of liver extracts from LNP-PCS-A2-, LNP-Crtl-, and
`PBS-treated rats (same as in A). Transferrin receptor (TFR) levels were used to normalize for protein loading. (Note that treated animal lane 15 was a noticeable
`outlier that did not up-regulate LDLR and on close examination also did not lower PCSK9 levels, possibly because of a misinjection). †, relative to PBS.
`
`MEDICALSCIENCES
`
`cleavage of the PCSK9 transcript via a targeted RNAi-specific
`mechanism.
`
`Efficacy of siRNA-Mediated Inhibition of Human PCSK9 in Transgenic
`Mice. Next, we tested the ability of LNP-PCS-A2 and LNP-
`PCS-C2 (PCS-C2 targets only human and NHP PCSK9 mRNA)
`(see Table S1) to silence human PCSK9 in vivo. To this end, we
`used a line of transgenic mice that express human PCSK9 cDNA
`under the apoE promoter (23). Specific PCR reagents and
`antibodies were designed that detect human, but not mouse,
`transcripts and protein, respectively. Cohorts of the humanized
`transgenic mice were injected with a single 5 mg/kg dose of
`LNP-PCS-A2 or LNP-PCS-C2, and both livers and blood were
`collected 72 h later. As shown in Fig. 4A, a single dose of
`LNP-PCS-A2 or LNP-PCS-C2 was able to decrease the human
`PCSK9 transcript levels by ⬎70%, and this resulted in ⬎500-fold
`reduction in the levels of circulating human PCSK9 protein as
`measured by ELISA (Fig. 4B). These results demonstrated that
`both siRNAs were capable of silencing the human transcript and,
`subsequently, reducing the amount of circulating plasma human
`
`LNP-PCS-A2 LNP-Ctrl
`
`PBS
`
`Predicted
`PCR band
`
`Lane
`
`1
`
`2 3 4 5 6 7 8 9 10 11 12
`
`M
`
`siRNA-mediated cleavage of PCSK9 mRNA in rats (5⬘-RACE). Rats (n ⫽
`Fig. 3.
`4 per group) were administered with 4 mg/kg LNP-PCS-A2, LNP-Crtl, or PBS and
`killed 4 days later. 5⬘-RACE detects the predicted mRNA cleavage product in
`LNP-PCS-A2- and not in LNP-Crtl- or PBS-treated animals. Eighty-seven percent
`of clones from the circled bands mapped to the predicted siRNA specific
`cleavage site.
`
`statistically significant difference in liver TG or cholesterol
`concentrations among animals administered PCSK9 siRNAs
`compared with control rats.
`The mechanism by which PCSK9 impacts plasma cholesterol
`levels has been linked to the density of LDLRs on the hepatocyte
`cell surface (5, 18–20). Pcsk9⫺/⫺ mice have 2- to 3-fold higher
`levels of liver LDLR protein compared with wild-type mice, and
`this effect is magnified by statin treatment (5). Similarly, reduc-
`tion of PCSK9 (using antisense oligos) over a 6-week period in
`high-fat-fed mice resulted in an up-regulation of LDLR levels
`(6). To investigate whether regulation of hepatic LDLRs oc-
`curred upon siRNA silencing of PCSK9 in rats, liver LDLR levels
`were quantified by immunoblot analysis after their treatment
`with 5 mg/kg LNP-PCS-A2. As shown in Fig. 2D, LNP-PCS-
`A2-treated animals had a significant 3- to 5-fold induction of
`LDLR levels compared with PBS- or LNP-Crtl-treated animals.
`Together, the rodent studies demonstrate that lowering of
`PCSK9 mRNA levels with siRNAs targeting PCSK9 in the liver
`results in an acute and durable decrease of serum cholesterol as
`a result of increased hepatic LDLR expression, and the acute
`change in LDLR expression is not associated with excess lipid
`accumulation in the liver.
`
`InVivoMechanism of PCSK9 Silencing Is siRNA Mediated. To confirm
`that the reduction in PCSK9 transcript observed in rodents was
`because of a siRNA mechanism,
`liver extracts from either
`treated or control rats were subjected to rapid amplification of
`cDNA ends (5⬘-RACE), a method previously used to demon-
`strate that siRNA-mediated cleavage occurs (21, 22). 5⬘-RACE
`analysis of liver mRNA from animals treated with LNP-PCS-A2
`revealed a product of the expected size (Fig. 3). Sequence
`analysis of cloned PCR products demonstrated that 73 of 84 of
`these products were derived from the predicted cleavage event
`at position (GAGT/TTAT). No specific bands were amplified in
`the 5⬘-RACE experiments from PBS- or LNP-Crtl-treated an-
`imals. These results demonstrate that the effect of LNP-PCS-A2
`observed on hepatic PCSK9 mRNA levels is consistent with
`
`Frank-Kamenetsky et al.
`
`PNAS 兩 August 19, 2008 兩 vol. 105 兩 no. 33 兩 11917
`
`

`

`PBS
`
`LNP-PCS-A2
`
`LNP-PCS-B2
`
`*** ***
`
`***
`
`******
`
`**
`
`***
`
`***
`
`***
`
`***
`
`***
`
`3 4 5 7 14 21
`3 4 5 7 14 21
`Day Post-injection
`
`PBS
`
`LNP-PCS-A2
`*** *** *** *** *** **
`
`LNP-PCS-B2
`
`***
`
`***
`
`***
`
`***
`
`***
`
`***
`
`3 4 5 7 14 21
`3 4 5 7 14 21
`Day Post-injection
`
`PBS
`
`LNP-PCS-A2
`
`LNP-PCS-B2
`
`***
`
`** ***
`
`***
`
`***
`
`***
`
`***
`
`***
`
`*** ***
`
`3 4 5 7 14 21
`3 4 5 7 14 21
`Day Post-injection
`
`1.4
`
`1.2
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0
`
`1.2
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0
`
`relative to pre-dose
`
`Serum LDLc
`
`relative to pre-dose
`
`Serum Tc/HDLc
`
`2.5
`
`2.0
`
`1.5
`
`1.0
`
`0.5
`
`0
`
`relative to pre-dose
`Plasma PCSK9 levels
`
`A
`
`B
`
`C
`
`Pharmacology of PCSK9 silencing in NHPs. (A) Direct LDLc measure-
`Fig. 5.
`ments of serum from cynomolgus monkeys treated with 5 mg/kg LNP-PCS-A2
`(n ⫽ 5 per group), LNP-PCS-B2 (n ⫽ 4 per group), or PBS (n ⫽ 3 per group). (B)
`Total cholesterol/HDL ratios as measured in the samples described in A. (C)
`Plasma samples from the treatment groups in A were analyzed for their levels
`of PCSK9 protein by ELISA. Values for LDLc, Tc/HDLc, and PCSK9 concentrations
`were graphed as a ratio of the average values of after dose compared with
`predose values within an animal. Those values were then combined into group
`averages. (A–C) LNP-Crtl behaved similarly on d4 and d7 as PBS (data not
`shown). PBS values are the mean of the groups ⫾ STDEV averaged over days
`3–14. Each value represents the mean of the group ⫾ STDEV. Two-way ANOVA
`with Bonferroni test was used. Asterisks represent statistical difference be-
`tween PBS- and PCSK9 siRNA-treated groups. **, P ⱕ 0.01; ***, P ⱕ 0.001.
`
`altered (data not shown). ALT and AST levels were not signif-
`icantly impacted (⬍3-fold induction over baseline in the same
`animal) with the exception of animal 6002, which had a 4- to
`5-fold increased ALT and 4- to 9-fold increased AST (Table S2).
`We note that animals in this study were a mixture of naı¨ve and
`nonnaı¨ve animals (one naı¨ve per group), with several animals
`having somewhat elevated ALT or AST at baseline (animal 6002
`was nonnaı¨ve). However, the cholesterol effects observed were
`independent of these measures as animals with no apparent
`increases in liver enzymes had similar LDLc reductions as animal
`6002 (Table S2 and Fig. S3A).
`Inasmuch as these experiments were nonterminal, we were
`unable to measure the reduction of transcript levels in the liver.
`As a surrogate for effects of siRNA treatment on PCSK9
`transcript silencing, we measured plasma PCSK9 protein levels
`in both treated and control NHPs by ELISA. As shown in Fig.
`5C, compared with predose concentrations, both LNP-PCS-A2
`and LNP-PCS-B2 treatments significantly reduced plasma
`PCSK9 concentrations. One caveat to measuring circulating
`PCSK9 levels in the context of LNP01 formulated siRNAs is that
`the delivery and silencing of PCSK9 is greatest in the liver as
`opposed to other tissues, such as the intestine, where PCSK9 is
`also expressed. It therefore remains possible that our measure-
`ments are an underestimate of liver PCSK9 protein lowering.
`
`PBS
`LNP-Ctrl
`LNP-PCS-C2
`LNP-PCS-A2
`
`2
`1
`Days after injection
`
`3
`
`2.5
`
`2.0
`
`1.5
`
`1.0
`
`0.5
`
`0
`
`B
`
`Plasma PCSK9 (mg/ml)
`
`***
`
`***
`
`PBS LNP-
`Ctrl
`
`LNP-
`PCS-C2
`
`LNP-
`PCS-A2
`
`1.4
`
`1.2
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0
`
`A
`
`relative to PBS
`
`Liver PCSK9 mRNA
`
`Silencing of human PCSK9 mRNA and protein reduction in PCSK9
`Fig. 4.
`humanized mice. (A) Transgenic mice (n ⫽ 4 per group) expressing the human
`PCSK9 full-length cDNA under the apoE promoter were dosed with LNP-
`PCS-A2 and LNP-PCS-C2, and LNP-Crtl or PBS. Both LNP-PCS-C2 and LNP-PCS-A2
`significantly lowered the human transcript as measured by quantitative PCR 3
`days after dose. Each value is the group mean ⫾ STDEV. (One-way ANOVA,
`Student’s t test, P ⱕ 0.001 between PBS- and PCSK9 siRNA-treated groups). (B)
`Circulating human PCSK9 protein levels were reduced in treated vs. control
`transgenic mice (n ⫽ 4 per group) as measured by ELISA. Each value represents
`group mean ⫾ STDEV. All time points for the PCSK9 siRNA treated groups are
`statistically significant compared with PBS (Two-way ANOVA, Bonferroni test,
`P ⱕ 0.01).
`
`PCSK9 protein. Interestingly, the human/NHP selective PCS-C2
`showed efficacy comparable to that of the cross-species PCS-A2.
`This result validates an approach where cross-species active
`siRNAs can be used in the development of RNAi therapeutic
`drugs.
`
`RNAi Silencing of PCSK9 Acutely Reduces PCSK9 Protein and Plasma
`LDLc, but Not Plasma HDLc, in NHPs. The results above demon-
`strated that siRNAs targeting PCSK9 acutely lower both plasma
`PCSK9 protein and total cholesterol levels with an effect that
`lasts ⬇3 weeks after a single dose in mice and rats. To extend
`these findings to a more relevant species with lipoprotein profiles
`that more closely resemble humans, we carried out similar
`experiments in cynomolgus monkeys.
`Animals were randomized based on day 3 LDLc levels. On the
`day of dosing (designated day 1), PBS and 1 mg/kg or 5 mg/kg
`of LNP-PCS-A2, -B2, and -Crtl were administered as a single
`infusion over 30 min. As the experiment progressed, it became
`apparent that the 1 mg/kg dose was not efficacious in reducing
`plasma LDLc (Fig. S2). We therefore dosed the PBS group
`animals on day 14 with 5 mg/kg LNP-Crtl so that they could serve
`as controls for animals that received the 5 mg/kg LNP-PCS-A2
`and LNP-PCS-B2. The PBS- and LNP-Crtl-treated animals
`behaved similarly for all measured endpoints (data not shown
`and Figs. S2 and S3A).
`As shown in Fig. 5A, a single dose of 5 mg/kg LNP-PCS-A2 or
`LNP-PCS-B2 resulted in a statistically significant reduction of
`LDLc beginning at day 3 after the dose that returned to baseline
`over ⬇14 days (for LNP-PCS-A2) and ⬇21 days (LNP-PCS-B2).
`This effect was not observed in the PBS group, the LNP-Crtl
`group, or the 1 mg/kg treatment groups (Fig. 5A and Figs. S2 and
`S3A). LNP-PCS-B2 resulted in an average LDLc lowering of
`56%, 72 h after the dose, with one of the four animals achieving
`nearly a 70% reduction in plasma LDLc compared with predose
`levels (see Fig. 5A and Fig. S3A). As expected, the lowering of
`LDLc in the treated animals correlated with a trend toward
`lower circulating apoB levels as measured by serum ELISA (Fig.
`S3B).
`Neither LNP-PCS-A2 nor LNP-PCS-B2 treatments resulted
`in a lowering of HDLc in this study. In fact, both siRNAs resulted
`(on average) in a trend toward a decreased total cholesterol/
`HDLc ratio (Fig. 5B). In addition, plasma TG levels were not
`
`11918 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0805434105
`
`Frank-Kamenetsky et al.
`
`

`

`MEDICALSCIENCES
`
`Overall, however, the reduction in blood PCSK9 protein levels
`detected here is consistent with the extent and duration of LDLc
`and circulating apoB reduction observed.
`
`Discussion
`The current standard of care for hypercholesterolemia is ade-
`quate for many patients, yet falls far short in others who are
`unable to reach target LDLc levels with currently available
`therapies. Although only recently described, PCSK9 represents
`one of the best validated targets for the reduction of LDLc.
`Humans carrying PCSK9 loss-of-function mutations have signif-
`icantly lower plasma LDLc and are remarkably protected from
`cardiovascular disease (8). Furthermore, human compound het-
`erozygous PCSK9-null individuals have been identified with very
`low LDLc (⬍20 mg/dl), but with otherwise normal health (24).
`Finally, the effects of PCSK9 have been shown to be highly
`conserved in animal models including PCSK9 transgenic and
`knockout mice.
`Here, we have shown that pharmacologic and RNAi-mediated
`reduction of PCSK9 transcript levels in liver achieves acute
`lowering of total plasma cholesterol levels in mice, rats, and
`cynomolgus monkeys. Silencing of PCSK9 mRNA in mice and
`rats was specific, and the silencing effects were proven to be
`mediated by an RNAi mechanism in rats using 5⬘-RACE.
`Delivery of the PCSK9 siRNA to the liver was facilitated by a
`lipidoid nanoparticle formulation as described in ref. 13. The
`formulated siRNAs silenced hepatic PCSK9 mRNA, resulting in
`a marked increase in liver LDLR protein levels. The same siRNA
`demonstrated silencing in mouse, rat, transgenic mouse, and
`NHP models systems.
`An unexpected finding in the rat studies was the robust total
`cholesterol reduction found with PCSK9 silencing compared
`with the well described lack of a statin effect (14). As opposed
`to other proposed LDLc-lowering targets such as MTP or apoB,
`this effect was achieved with no evidence of increased liver
`triglycerides or other untoward effects, thereby further validat-
`ing PCSK9 as the target of choice for therapeutic intervention.
`The effects of statins on plasma cholesterol are secondary to
`their effects on LDLR levels in NHPs and humans. Our results
`indicate that at least with regard to the level of PCSK9 and its
`control of the LDLR, the pathways in mice, rats, NHPs, and
`humans are conserved.
`Rodents lack CETP and carry most of their plasma cholesterol
`in HDL particles. In addition, HDL in rodents also has a
`significant amount of apoE, which facilitates clearance by LD-
`LRs. Genetic data in humans also confirms that loss of PCSK9
`reduces LDLc but has no effect on HDLc (7). We therefore
`sought to show an acute and specific effect on LDLc lowering in
`a model closer to humans, the cynomolgus monkey. In studies
`with cynomolgus monkeys, we demonstrated that a single 30-min
`infusion of two different formulated PCSK9 siRNAs resulted in
`a highly significant, acute, specific, and durable reduction of
`plasma LDLc, apoB, and PCSK9 protein levels, but not HDLc
`or TGs. Specifically, siRNA-mediated reduction of PCSK9
`mRNA and protein resulted in lowering of LDLc by ⬇50–60%
`within 48 h after administration; this reduction lasted for nearly
`3 weeks. If the ⬇50–60% LDLc reduction observed in NHP were
`translated into humans, this amount of lowering would compare
`favorably with the LDLc reductions observed with current
`cholesterol absorption inhibitors (⬇20% LDLc lowering in
`humans with ezetimibe) or HMG-CoA reductase inhibitors
`(⬇21% LDLc lowering with pravastatin and ⬇51% LDLc de-
`creases at highest 80 mg/day atorvastatin over 30 days in the
`PROVE-IT-TIMI trial) (25, 26). In addition, observations in
`mice suggest that PCSK9 lowering and statin activity may act
`synergistically (5). It will be interesting to test whether this
`observation holds true in other animal models, such as NHPs.
`Finally, the acute onset of LDLc lowering obtained with PCSK9
`
`siRNAs could prove beneficial in a subset of patients who suffer
`an acute myocardial infarction where the rapid onset lowering of
`LDLc may be desirable (27).
`Together, the results presented here validate PCSK9 lowering
`by RNAi as a therapeutic approach with high potential for
`acutely reducing LDLc and pave the way for the development of
`novel PCSK9 lowering agents for use in the treatment of
`hypercholesterolemia.
`
`Materials and Methods
`Synthesis of siRNAs Targeting PCSK9. Single-stranded RNAs were produced at
`Alnylam Pharmaceuticals. Deprotection and purification of the crude oligo-
`ribonucleotides by anion exchange HPLC were carried out according to es-
`tablished procedures. siRNAs were generated by annealing equimolar
`amounts of complementary sense and antisense strands.
`
`siRNA Formulation into LNP Lipidoid Nanoparticles. Stock solutions of lipidoid
`98N12-5 (1)䡠4HCl, cholesterol, and mPEG2000-DMG MW 2660 (synthesized by
`Alnylam) were prepared in ethanol and mixed to yield a molar ratio of
`42:48:10 (13). siRNA was incorporated in the nanoparticles at 1:7.5 (wt:wt)
`siRNA:total lipids. Resulting particles had a mean particle diameter of ⬇50 nm
`and siRNA entrapment efficiency of ⬎95%.
`
`PCSK9siRNAinVitroScreening in HepG2 Cells and Primary Cynomolgus Monkey
`Hepatocytes. For siRNA transfection experiments, HepG2 or primary hepato-
`cyte cells were seeded at 2.5 ⫻ 104 cells per well in 96-well plates. siRNA were
`transfected by using Lipofectamine 2000 according to the manufacturer’s
`protocols. Cells were lysed 24 h after transfection, and PCSK9 mRNA levels
`were quantified by using the branched-DNA-technology-based QuantiGene
`Reagent System (Panomics), according to the manufacturer’s protocols. PCSK9
`mRNA levels were normalized to GAPDH mRNA.
`5⬘-RACE was conducted as described in ref. 20 (see SI Text).
`
`InVivoRodent Experiments. All procedures used in animal studies conducted
`at Alnylam were approved by the Institutional Animal Care and Use Commit-
`tee and were consistent with local, state, and federal regulations as applica-
`ble. Mice and rats were maintained on a 12-h light/12-h dark cycle and killed
`at the end of the dark cycle. C57BL/6 mice and Sprague–Dawley rats received
`either PBS or siRNA in lipidoid formulations via tail vein injection at a volume
`of 0.01 ml/g. After dosing, animals were anesthetized by isofluorane inhala-
`tion and blood was collected into serum separator tubes by retro-orbital
`bleed. Total cholesterol in mouse serum was measured by using the Wako
`Cholesterol E enzymatic colorimetric method (Wako Chemicals). In experi-
`ments where liver mRNA levels were assessed, livers were harvested and snap
`frozen in liquid nitrogen. Frozen liver tissue was ground and tissue lysates
`were prepared. PCSK9 mRNA levels relative to those of GAPDH mRNA were
`determined in the lysates by using a branched DNA assay (QuantiGene Re-
`agent System, Panomics). LDLR protein was quantified by using 20 ␮g of liver
`membrane protein that was subjected to SDS/PAGE and transferred to nitro-
`cellulose membranes as described in ref. 20, followed by immunoblotting and
`imaging by using LI-COR Odyssey infrared imaging system (28).
`
`Studies in Transgenic Mice Expressing Human PCSK9. Transgenic mice that
`express human PCSK9 have been described in ref. 23. Human PCSK9 mRNA
`transcript was measured